Underground Distribution System Design Guide

november 2008

project 03-08

Underground Distribution System Design Guide

this page intentionally left blank

PROJECT 03-08

Underground Distribution System Design Guide Prepared by Edward S. Thomas, PE Utility Electrical Consultants, PC 620 N.West St., Suite 103 Raleigh, NC 27603-5938 and Bill Dorsett Booth & Associates, Inc. 1011 Schaub Drive Raleigh, NC 27606 for Cooperative Research Network National Rural Electric Cooperative Association 4301 Wilson Boulevard Arlington, Virginia 22203-1860

The National Rural Electric Cooperative Association The National Rural Electric Cooperative Association (NRECA), founded in 1942, is the national service organization supporting more than 900 electric cooperatives and public power districts in 47 states. Electric cooperatives own and operate more than 42 percent of the distribution lines in the nation and provide power to 40 million people (12 percent of the population).

© Underground Distribution System Design Guide Copyright © 2008, by the National Rural Electric Cooperative Association. Reproduction in whole or in part is strictly prohibited without prior written approval of the National Rural Electric Cooperative Association, except that reasonable portions may be reproduced or quoted as part of a review or other story about this publication.

Legal Notice This work contains findings that are general in nature. Readers are reminded to perform due diligence in applying these findings to their specific needs, as it is not possible for NRECA to have sufficient understanding of any specific situation to ensure applicability of the findings in all cases. Neither the authors nor NRECA assume liability for how readers may use, interpret, or apply the information, analysis, templates, and guidance herein or with respect to the use of, or damages resulting from the use of, any information, apparatus, method, or process contained herein. In addition, the authors and NRECA make no warranty or representation that the use of these contents does not infringe on privately held rights. This work product constitutes the intellectual property of NRECA and its suppliers, as the case may be, and contains Confidential Information. As such, this work product must be handled in accordance with the CRN Policy Statement on Confidential Information.

Contact:

Edward S. Thomas, PE Utility Electrical Consultants, PC 620 N.West St., Suite 103 Raleigh, NC 27603-5938 Phone: 919.821.1410 Fax: 919.821.2417

Bill Dorsett Booth & Associates, Inc. 1011 Schaub Drive Raleigh, NC 27606 Phone: 919.851.8770 Fax: 919.859.5918

Contents – iii

con t e n t s Section 1

Design of an Underground Distribution System System Components Types of UD Systems Reliability of UD Systems Design Considerations for System Operation and Maintenance Future Upgrades and Replacements Economic Comparison of System Configurations UD Loss Economics Steps for Layout of a UD System Summary and Recommendations

1 2 6 14 17 19 20 32 38 50

Section 2

Cable Selection Typical Cable Configuration Conductor Size Designations Conductor Materials and Configuration Cable Insulation Materials Insulation Fabrication Conductor Shields and Insulation Shields Cable Specification and Purchasing Cable Acceptance Summary and Recommendations

51 51 53 53 57 60 64 74 77 77

Section 3

Underground System Sectionalizing General Sectionalizing Philosophy Overcurrent Protection of Cable System Effect of Inrush Current on Sectionalizing Devices Selection of Underground Sectionalizing Equipment Faulted-Circuit Indicators Summary and Recommendations

79 79 88 96 100 105 118

Section 4

Equipment Loading Primary Cable Ampacity Pad-Mounted Transformer Sizing Summary and Recommendations

121 121 144 163

Section 5

Grounding and Surge Protection Cable Grounding System Function Factors Affecting Cable Grounding System Performance Counterpoise Application for Insulated Jacketed Cable System Ground Resistance Measurement and Calculation Underground System Surge Protection Summary and Recommendations

165 166 177 188 192 207 236

i v – C o n t en t s

c o n te n t s Section 6

Ferroresonance Allowable Overvoltages During Ferroresonance Distribution Transformer Connections Qualitative Description of Ferroresonance Ferroresonance When Switching at the Primary Terminals of Overhead and Underground Transformer Banks Ferroresonance with Cable-Fed Three-Phase Transformers with Delta or Ungrounded-Wye Connected Primary Windings Ferroresonance with Cable-Fed Three-Phase Transformers with Grounded-Wye Primary Winding and Five-Legged Core Ferroresonance with Cable-Fed Three-Phase Transformers with Grounded-Wye Primary Windings and Triplex Construction Ferroresonance in Underground Feeders Having More Than One Transformer Summary of Techniques for Preventing Ferroresonance in Underground Systems Summary and Recommendations References

239 240 241 242 252 254 260 266 270 273 276 279

Section 7

Cathodic Protection Requirements Special Note Introduction What to Protect Where to Protect Types of Cathodic Protection Systems Amount of Cathodic Protection Cathodic Protection Design with Galvanic Anodes Cathodic Protection Installation and Follow-Up Calculation of Resistence to Ground Summary and Recommendations

281 281 281 282 282 285 286 287 294 296 297

Section 8

Direct-Buried System Design Trench Construction Considerations Trench Design Components Trench Layout/Routing Considerations Depth of Burial Joint-Occupancy Trenches Summary and Recommendations

299 299 300 303 304 307 309

Section 9

Conduit System Design Conduit System Design Cable Pulling Summary and Recommendations

311 311 332 341

Contents – v

con t e n t s Section 10

Joints, Elbows, and Terminations Joints, Elbows, and Terminations for 200-Ampere Primary Circuits Joints, Elbows, and Terminations for 600-Ampere Primary Circuits Joints, and Terminations for Secondary Circuits Summary and Recommendations

343 344 353 355 357

Section 11

Cable Testing Reasons for and Benefits of Cable Testing by the User Primary Cable Tests by the User Secondary Cable Tests by the User Tests by the Cable Manufacturer Summary and Recommendations

359 359 359 369 370 372

Appendix A

Calculations for Reliability Studies Reliability Index Acceptability Criteria Calculation of Reliability Importance of Sectionalizing

373 373 374 374 375

Appendix B

Transformer and Secondary Voltage Drop Voltage Flicker

377 385

Appendix C

Sample Specification UGC2 for 600-Volt Secondary Underground Power Cable Scope General Specifications Referenced Specifications Conductor Insulation Tests Miscellaneous Markings Multiconductor Cable Assemblies

389 389 390 390 391 391 392 393 393 393

Checklist for Information Requirements Project Information Checklist

395 395

Appendix D

v i – C o n t en t s

c o n t e nt s Appendix E

Sample Specification for 15-, 25-, and 35-kV Primary Underground Medium Voltage Concentric Neutral Cable (Specification UGC1) Purpose General Specifications Referenced Specifications Conductor Conductor Shield (Stress Control Layer) Insulation Insulation Shielding Concentric Neutral Conductor Overall Outer Jacket Dimensional Tolerances Tests Miscellaneous

397 397 397 398 399 400 400 400 401 401 402 402 403

Appendix F

Allowable Short Circuit Currents for Solid Dielectric Insulated Cables

405

Appendix G

Ampacity Tables

415

Appendix H

Industry Specifications

425

Appendix I

Component Manufacturers

427

Appendix J

Cable-Pulling Examples

431

Abbreviations

435

Illust r a ti o n s – v i i

illustra t i o ns FIGURE

PAGE

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11 1.12 1.13 1.14 1.15 1.16 1.17 1.18 1.19 1.20 1.21 1.22 1.23 1.24 1.25 1.26 1.27 1.28 1.29

UD System Components Schematics for Different Types of Switchgear Flat Pad for Equipment Mounting Ground Sleeve Box Pad for Equipment Mounting Underground Substation Circuit Exit Radial Main Feeder Radial Main Feeder with Faulted Cable Section Open-Loop Feeder Open-Loop Feeder with Faulted Cable Section Radial Feeder Open-Loop Feeder in Shopping Center Multiple-Loop System Area Lighting System Loop-Feed Design of UD System Under Normal Conditions Loop-Feed Design of UD System with Damaged Cable Section Open-Loop System, 37-Lot Subdivision Open-Loop System, Single Residential Consumer Single-Phase Sub-Feeder Three-Phase Sub-Feeder Front Property Placement Back Property Placement Methods for Providing Secondary Service Road Crossing to Feed Secondary Pedestal Service and Transformer Layout for 75-Lot Subdivision Primary Cable Layout for 75-Lot Subdivision Minimum Required Working Space Sample Easement Staking Sheet for Service to a Commercial Consumer

2 3 5 5 5 6 7 8 9 9 10 11 11 12 16 16 21 22 24 25 28 28 31 40 40 42 43 47 49

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10

Jacketed Concentric Neutral Cable Bare Concentric Neutral Cable Medium-Voltage Power Cable with Tape Shield and L.C. Shield Concentric Lay Strand Options Standard Strand Arrangements for Multilayer Conductors Comparative Hot Creep vs. Temperatures for Cable Insulation Materials General Layout of a Cable Extrusion Line Typical Extrusion Methods Capacitive and Dielectric Loss Current Flow in Insulation Shield Cable Identification Markings

52 52 52 56 56 60 62 63 66 73

v i i i – Il l u st r a t i o n s

i l l us t r a tions FIGURE 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 3.20 3.21 3.22 4.1

4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 4.15 4.16

PAGE Symmetrical Current Asymmetrical Short-Circuit Current Sample Distribution Circuit with Typical Locations of Sectionalizing Devices Show Cross Section of Cable Showing Components Subject to Through-Fault Damage Example of 70-Ampere, Type “L” Recloser Curves for Cable Protection Current Limiting Fuses for Padmounted Switching Cabinets Inrush Current Resulting from Operation of Three-Phase Recloser Inrush Current Resulting from Operation of Single-Phase Recloser Trip Response for Peak-Current-Sensitive Units Trip Response for 450A and 800A FCIs Trip-Set Characteristics for Adaptive-Trip FCI FCI Placement on Overhead Feeder with Underground Segment FCI Placement on Three-Phase Underground Feeder FCI Placement for Single-Phase Open Loop FCI Placement for Underground Subdivision with Three-Phase Source Current-Reset FCI Low-Voltage-Reset FCI High-Voltage-Reset FCI Time-Reset FCI Correct Placement of FCI Sensor Incorrect Placement of FCI Sensor Reset FCI Ratio of Shield Loss to Conductor DC Loss at 90°C as a Function of Shield Resistance, 1/C 35-kV Aluminum Power Cables in Triplexed Formation Relationship Between Load Factor and Loss Factor Per Unit Thermal Resistivity vs. Moisture Content for Various Soil Types Thermal Resistivity of Soil at Various Locations Effect of Depth on Soil Temperatures as Influenced by Seasonal Temperature Variations Trefoil or Triangular Cable Configuration Flat Conductor Configuration, Maintained Spacing Direct-Buried Duct Bank Installation Using Rigid Nonmetallic Conduit Single-Phase U-Guard Installation with Vented Base Three-Phase Cable Installation Configurations 138, Typical Dead-Front, Single-Phase, Pad-Mounted Transformer Actual Load Cycle and Equivalent Load Cycle Thermal Equivalent Load Cycle Case Temperature Measurement Location—Pad-Mounted Distribution Transformer Relationship Among NEMA Starting Code Letters, Starts per Hour, and Transformer kVA per Motor HP for Transformer Thermal Considerations Maximum Motor Starts per Hour for Transformer Mechanical Considerations

82 82 86 88 90 104 107 107 108 109 110 111 111 112 112 113 114 114 115 116 116 117

124 125 127 127 128 130 130 132 136 423 145 147 147 159 160 162

Illus t r at i o n s – i x

illustra t i o ns FIGURE 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11 5.12 5.13 5.14 5.15 5.16 5.17 5.18 5.19 5.20 5.21. 5.22 5.23 5.24 5.25 5.26 5.27 5.28 5.29 5.30 5.31 5.32 5.33 5.34 5.35 5.36 5.37 5.38

PAGE Typical Distribution Transformer Core Form Design and Neutral Grounding Circuit Variation of Surge Impedance with Surge Current for Various Values of 60-Cycle Resistance Surge Characteristics of Various Ground Rods Arrester Lead Length for Two Riser Pole Installations Three-Phase Installation Showing Optimum Riser Pole Arrester Lead Connections Typical Primary and Secondary Underground Installation Schematic Diagram Showing Surge Current Paths After Lightning Arrester Discharge Maximum Jacket Voltage (Neutral to Ground) Produced by Lightning Current Surge in Ground Rod BCN Cable Riser Pole Installation Surge Arrester Discharge Paths Ground Rod Being Driven by Hydraulic Tool Resistance of Vertical Ground Rods as a Function of Length and Diameter Resistance of Multiple Ground Rods Installation of Three Rods for a Riser Pole Ground Installation of Four Rods for a Riser Pole Ground Grounding Assembly for Pad-Mounted Single-Phase Transformers Grounding Grid for Pad-Mounted Equipment Installation Installation of JCN Connection in Above-Grade Pedestal Grounding Assembly for JCN Underground Primary Cable Intermediate Grounding Assembly, Underground Primary Cable Counterpoise 60-Hz Resistance Variation with Length and Different Soil Resistivities Effect of Length on Transient Surge Impedance of Counterpoise Counterpoise Application to Reduce Jacket Voltage Earth Resistance Correct Ground Resistance Test Setup Incorrect Ground Resistance Test Setup Clamp-On Ground Resistance Tester Circuit Diagram for Multigrounded System Ground Resistance Test Setup for Clamp-On Tester Setup for Soil Resistivity Test Effects of Moisture on Soil Resistivity Effects of Salt Content on Resistivity in Soil Containing 30 Percent Moisture Coefficient K1 for Ground Resistance Calculations Grouping of Four Ground Rods with 16-Foot Spacing Grouping of Four Ground Rods with 5-Foot Spacing Types of Arresters and Their Construction Comparison of Nonlinear Characteristics of SiC and MOV Valve Elements Effect of Fast Rise Times on IR Discharge Series- and Shunt-Gapped MOV Distribution Arresters

169 171 171 173 173 174 175 175 178 180 181 182 183 183 185 185 186 187 187 188 189 190 193 193 193 195 195 195 196 198 198 201 203 203 208 209 210 210

x – Il lu s t ra t io n s

i l l us t r a tions FIGURE 5.39 5.40 5.41 5.42 5.43 5.44 5.45 5.46 5.47 5.48 5.49 5.50 5.51 5.52 5.53 5.54 5.55 5.56 5.57 5.58 5.59 5.60 5.61 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11

PAGE Dead-Front Arrester Elbow Configuration Dead-Front Surge Arresters Temporary 60-Hz Overvoltage Capability Curves—Typical MOV Distribution Arrester Typical Test Current Waveshape—Sinusoidal Wavefront Lightning Rise Time to Peak Arrester Lead Length Equal to Three Feet Arrester Lead Length Equal to 1.5 Feet Zero Arrester Lead Length Representation of Distributed Parameter Distribution Line Change in Surge Impedance at a Junction Point—Effect on Traveling Voltage Wave Traveling Wave Behavior at Junction Points Terminated with Various Surge Impedances Traveling Waves at a Cable Open-End Point Terminated by an MOV Arrester Arrester Locations Cable-End Arresters at Open Point Arrester Upstream from Open Point (Third Arrester) Two Elbow Arresters and a Feed-Through Elbow Arrester and Parking Stand Arrester Bushing Arrester and Parking Stand Arrester Elbow Arrester on Feed-Through Insert on Transformer Upstream from Open Point Bushing Arrester on Transformer Upstream from Open Point Lateral Tap Cable-End Arrester (Radial Feed Circuit) Tap-Point Arrester Typical Underground Subdivision Loop Feed with Open Point Transformer Connections for Four-Wire Wye and Four-Wire Delta Services Series RLC Circuit with Sinusoidal Excitation Cable-Fed Three-Phase Transformer Susceptible to Ferroresonance Conductor Spacings for an Overhead Line on an Eight-Foot Crossarm Equivalent Capacitance Network for an Overhead Multigrounded Neutral Line Cross Section of a Multiwire Concentric Neutral Cable Floating-Wye/Delta Transformer Bank with Fused Cutouts at Primary Terminals Three-Phase Cable-Fed Transformer with a Delta-Connected Primary Winding Voltage and Current Waveforms During Ferroresonance with a 150-kVA Delta Grounded-Wye Bank Five-Legged Wound-Type Core with Grounded-Wye Primary Windings Three-Phase Cable-Fed Transformer with a Grounded-Wye Primary Winding on a Five-Legged Core

211 212 215 217 218 219 220 221 222 223 224 225 227 230 231 231 232 232 232 232 232 232 232

242 243 245 247 247 248 253 255 255 260 262

Illus t r at i o n s – x i


6.12 6.13 6.14 6.15 6.16 6.17 6.18 7.1

PAGE

Open-Phase Voltage Waveforms with Five-Legged Core Grounded-Wye Transformers Overhead System Supplying a Cable-Fed Grounded-Wye Transformer on a Five-Legged Core Triplex-Type Wound Core with Grounded-Wye Primary Windings Cable-Fed Triplex-Core Transformer with Grounded-Wye Primary Windings Circuit with “S” Cable Sections and “N” Five-Legged Core Grounded-Wye Primary Transformers Circuit Configuration for Switching Example 6.2 Single-Line Diagram of a Portion of a UD System

262 267 269 269 270 271 274

7.8 7.9 7.10 7.11

Dissimilar Metal Effects Between Buried Metals Connected to the Neutral of an Electric Distribution Line Electric System Map Shaded to Show Corrosive Soil Locations Measurement of Potential to a Copper-Copper Sulfate Half Cell Dissimilar Metal Effects Between Copper and Steel Dissimilar Soil Effects on Buried Copper Wires Measurement of Earth Resistivity with a Four-Terminal Ground Tester Potentials of a Copper-Steel Couple Before and After Connecting a Zinc Anode Equivalent Circuit for a Galvanic Anode Connected to the Electric Neutral Anode Positioning Anode Connector Test Station Connector

8.1 8.2 8.3 8.4

Typical Trench Warning Tape Cable Route Marker Burial Depth Requirements Joint Trench Use

301 302 305 308

9.1 9.2 9.3 9.4 9.5

Typical Duct Configurations Typical Duct Line and Manhole Arrangement Typical Arrangements for System in Figure 9.2 Preferred Location of Duct Lines in Roadways Typical Manhole Configurations

316 319 319 326 326

9.6 9.7 9.8 9.9 9.10 9.11 9.12

Rectangular Manhole Construction Details Rectangular Manhole Installation Details Octagonal Manhole Construction Details Octagonal Manhole Installation Details Cable/Conduit Friction and Pulling Tension Cable Configurations in Conduit Sidewall Bearing Pressure

327 328 329 330 333 334 336

7.2 7.3 7.4 7.5 7.6 7.7

282 283 283 284 284 284 285 287 295 295 295

x i i – Il l u s t r at i o n s

i l l us t r a tions FIGURE

10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9 10.10 10.11 10.12

PAGE

Voltage Stress Concentration Voltage Stress Distribution in a Typical Premolded Joint Housing Premolded Permanent Straight Joint for Primary Cables Jacket Replacement Assembly (Method C) Premolded Permanent Wye Joint for Primary Cables Dead-Break Elbow for Primary Cables Load-Break Elbow for Primary Cables Typical 200-Ampere Elbow Accessories Heat-Shrink Jacket Seal at Elbow Premolded Indoor Termination (Slip-On Stress Cone) for Primary Cables Premolded Integral Indoor/Outdoor Termination for Primary Cables Premolded Modular Indoor/Outdoor Termination with Separate Skirts for Primary Cables Porcelain Indoor/Outdoor Terminal for Primary Cables Cold-Shrink Indoor/Outdoor Termination for Primary Cables Stick-Operable, Dead-Break Elbows Dead-Break 600-Ampere Elbow Connector and Accessories for Primary Cables Housing Assembly Joint for Secondary Cables Cold-Shrink Joint for Secondary Cables Heat-Shrink Joint for Secondary Cables Sealed Stud Termination for Secondary Cables Bus and Rubber Cover Termination for Secondary Cables Housing and Sleeve Assembly Termination for Secondary Cables

344 344 345 346 347 348 348 349 349 351 351

Test Setup for the Hot Silicone Oil Test Typical Test Setup for the Stripping Test of the Insulation Shield Typical High-Voltage Proof Tester Showing a Sectionalized Discharge Stick for Grounding the Cable

364 365

A.1 A.2

Components Affecting Outage Rate to the Consumer Sectionalized UD Area

374 376

B.1 B.2

Distance for Various Conductor Arrangements Permissible Voltage Flicker Limits

381 386

10.13 10.14 10.15 10.16 10.17 10.18 10.19 10.20 10.21 10.22 11.1 11.2 11.3

351 352 352 353 354 355 355 355 356 356 356

368

Illustr a ti o n s – x i ii


F.1

F.2

F.3

F.4

F.5

F.6

F.7

F.8

PAGE

Aluminum Conductor/Thermoplastic Insulation (PE/HMWPE)— Allowable Short Circuit Currents Based on 75°C Initial Conductor Temperature and 150°C Final Temperature Copper Conductor/Thermoplastic Insulation (PE/HMWPE)— Allowable Short Circuit Currents Based on 75°C Initial Conductor Temperature and 150°C Final Temperature Aluminum Conductor/Thermoset Insulation (TR-XLPE/EPR)— Allowable Short Circuit Currents Based on 90°C Initial Conductor Temperature and 250°C Final Conductor Temperature Copper Conductor/Thermoset Insulation (TR-XLPE/EPR)— Allowable Short Circuit Currents for 90°C Rated Insulation Based on 90°C Initial Conductor Temperature and 250°C Final Conductor Temperature Aluminum Conductor/Thermoplastic Insulation (PE/HMWPE)— Allowable Short Circuit Currents for Conductor to Not Exceed Insulation Emergency Operating Temperature Rating Based on 75°C Initial Conductor Temperature and 90°C Final Conductor Temperature Copper Conductor/Thermoplastic Insulation (PE/HMWPE)— Allowable Short Circuit Currents for Conductor to Not Exceed Insulation Emergency Operating Temperature Rating Based on 75°C Initial Conductor Temperature and 90°C Final Conductor Temperature Aluminum Conductor/Thermoset Insulation (TR-XLPE/EPR)— Allowable Short Circuit Currents for Conductor to Not Exceed Insulation Emergency Operating Temperature Rating Based on 90°C Initial Conductor Temperature and 130°C Final Conductor Temperature Copper Conductor/Thermoset Insulation (TR-XLPE/EPR)— Allowable Short Circuit Currents for Conductor to Not Exceed Insulation Emergency Operating Temperature Rating Based on 90°C Initial Conductor Temperature and 130°C Final Conductor Temperature

406

407

408

409

410

411

412

413

x i v – Ta b l e s

tables TABLE

PAGE

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11 1.12 1.13 1.14 1.15 1.16 1.17 1.18 1.19 1.20 1.21 1.22

Lamp and Ballast Characteristics—240 Volts Front Versus Rear Property Line Placement Additional Materials for an Open-Loop System Sample Spare Cable Cost, Single Residential Consumer Sample Radial System Cost, Commercial Consumer Additional Cost per Kilowatt, Open-Loop and Spare Cable Systems Single-Phase Sub-Feeder Cost Three-Phase Sub-Feeder Cost 25-kV Versus 15-kV Cable and Components Added Cost of Dual-Voltage Transformers Voltage Conversion Cost at Year 10 Voltage Conversion Cost at Year 20 Option 1—Direct-Buried Cable Option 2—PVC Rigid Conduit Option 3—Cable in HDPE Flexible Conduit Present Worth of Cable Installation Options Separate Service Cables Secondary Pedestal Sample Cable Loss Analysis Sample Secondary Cable Data Savings from Deferred Transformer Energization Savings from Deferred Transformer Installation

14 17 20 22 23 23 24 25 26 26 26 27 30 30 31 31 32 32 35 36 37 38

2.1

Dimensional Characteristics of Common Conductors (Standard Concentric-Lay) Conductor Physical and Electrical Characteristics Configurations of 4/0 AWG Aluminum Conductor RUS Insulation Thickness Insulation Shield Strippability Ratings Concentric Neutral Configurations for Common Aluminum Cables Comparison of Jacketing Material Test Data Static Coefficient of Friction for Jacketing Materials in PVC Conduit

53 54 57 59 66 67 71 72

2.2 2.3 2.4 2.5 2.6 2.7 2.8 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8

Multiplying Factors to Determine Asymmetrical Fault Currents Where Symmetrical Fault Currents Are Known Effective Cross-Sectional Area of Shield Values of T1, Approximate Shield Operating Temperature, °C, at Various Conductor Temperatures Values of T2, Maximum Allowable Shield Transient Temperature, °C Values of M for the Limiting Condition Where T2 = 200°C Values of M for the Limiting Condition Where T2 = 350°C Approximate Levels of I2t (Amperes2 x Seconds) That May Result in Destructive Transformer Failure for Internal Faults Approximate Levels of Fault Current Symmetrical (Amperes) That May Result in Destructive Transformer Failure for Internal Faults

83 91 92 92 92 92 95 95

Ta b l e s – x v

tables TABLE 4.1 4.2 4.3 4.4 4.5 4.6. 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 4.15 4.16 4.17 4.18 4.19 4.20 4.21 4.22 4.23 4.24 5.1 5.2 5.3 5.4 5.5 5.6

PAGE Ampacities for Single-Phase Primary Underground Distribution Cable— XLPE, TR-XLPE, and EPR Insulated Typical Ambient Soil Temperatures at a Depth of 3.5 Feet Ampacity for 15-kV Copper Conductor, Direct Buried, Single Circuit, 75% and 100% Load Factor Ampacity Table for 15-kV Aluminum Conductor, Direct Buried, Single Circuit, 75% and 100% Load Factor Pros and Cons of Installing Cable Circuits in Conduit Ampacity Values—15-kV Cable, Trefoil Configuration, Copper Conductor Ampacity Values—15-kV Cable, Trefoil Configuration, Aluminum Conductor Abstract of ICEA Standards for Maximum Emergency-Load and Short-Circuit-Load Temperatures for Various Insulations Correction Factors to Convert from 25°C Ambient Soil Temperature to 20°C and 30°C Correction Factors for Various Ambient Air Temperatures Typical Ampacities for Various Sizes and Types of 600-Volt Secondary UD Cable—Stranded Aluminum Conductors Average Temperatures for July and August Averaged for the Previous 10 Years Daily Peak Loads Per Unit of Nameplate Rating for Self-Cooled Oil-Immersed Transformers to Give Minimum 20-Year Life Expectancy Application of Single-Phase Distribution Transformers to Serve Residential Consumers—Sample Loading Guide Typical Watts-Per-Square-Foot Factors for Commercial Buildings Typical Electrical Load Power Factor Values Typical Electrical Load Demand Diversity Factor Values Estimated Electrical Demand (Summer) and Energy Consumption (Sample Family Restaurant) Estimated Peak Duration Transformer Loading Capability Table Typical Three-Phase Pad-Mounted Transformer Capacities— Short-Term Overload Capabilities (in kVA) Surface Temperatures Measured at Various Locations on the Cases of Pad-Mounted Transformers. Surface Contact Time to Produce Burning NEMA Starting Code Letters Surge Withstand Strengths of Polyethylene Insulating Jackets for 15-kV, 25-kV, and 35-kV Class JCN Cable 2007 NESC Ground Rod Requirements for JCN Cable Installations Spacing of Test Probes for Testing Resistance of a Single Ground Rod Spacing of Test Probes for Testing Resistance of an Electrode System Soil Resistivities for Different Soil Types and Geological Formations Effect of Temperature on Soil Resistivity

123 128 130 131 133 135 135 137 139 139 143 146 148 150 153 153 154 155 156 156 156 159 160 161

176 184 194 194 197 198

x v i – Ta b l e s

tables TABLE 5.7 5.8 5.9 5.10 5.11 5.12

5.13

5.14 5.15 5.16 5.17 5.18 5.19

5.20 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10

PAGE Ground Resistance in Varying Soil Resistivities Comparison of Protective Characteristics of Heavy-Duty Distribution Class Silicon Carbide, MOV, and Riser Pole MOV Arresters Typical Electrical Ratings and Characteristics of Dead-Front Surge Arresters Comparison of Standard Requirements for Surge Arrester Classifications Metal Oxide Surge Arrester Ratings in (kV) rms Protective Margin, 24.9-kV Underground Distribution System: 125-kV BIL Insulation, 18-kV Arresters at Riser Pole Only, 10-kA Lightning Discharge, Surge Voltage Doubled by Reflection Protective Margin, 12.47-kV Underground Distribution System: 95-kV BIL Insulation, 9-kV Arresters at Riser Pole Only, 10-kA Lightning Discharge, Surge Voltage Doubled by Reflection Recommended Arrester Locations MOV Riser Pole Arrester: Arrester Rating, 10 kV; Equipment BIL, 95 kV; Aged BIL, 76 kV MOV Riser Pole Arrester and Dead-Front Cable-End Arrester (No. 4): Arrester Rating, 10 kV; Equipment BIL, 95 kV; Aged BIL, 76 kV MOV Riser Pole Arrester: Arrester Rating, 21 kV; Equipment BIL, 125 kV; Aged BIL, 100 kV MOV Riser Pole Arrester and Dead-Front Cable-End Arrester (No. 4): Arrester Rating, 21 kV; Equipment BIL, 125 kV; Aged BIL, 100 kV MOV Riser Pole Arrester Plus Dead-Front Cable-End Arrester (No. 4) and Dead-Front Third Arrester (No. 3): Arrester Rating, 21 kV; Equipment BIL, 125 kV; Aged BIL, 100 kV Ground Resistance Testers Values for Equivalent Capacitances of an Overhead Line with 4/0 ACSR Phase Conductors and a 1/0 ACSR Neutral Conductor Representative Capacitance and Three-Phase Charging for XLPE Insulated Cables with 175 Mils Insulation Representative Capacitance and Three-Phase Charging or XLPE Insulated Cables with 220 Mils Insulation Representative Capacitance and Three-Phase Charging for XLPE Insulated Cables with 260 Mils Insulation Representative Capacitance and Three-Phase Charging for XLPE Insulated Cables with 345 Mils Insulation Phase-to-Ground Capacitance of Three-Phase Grounded-Wye Capacitor Banks Maximum Allowed Cable Lengths in 12.47-kV Systems to Limit Open-Phase Voltages to 1.25 PU Maximum Allowed Cable Lengths in 24.9-kV Systems to Limit Open-Phase Voltages to 1.25 PU Maximum Allowed Cable Lengths in 34.5-kV Systems to Limit Open-Phase Voltages to 1.25 PU Transformer and Cable Data for the System of Figure 6.17

204 209 213 214 215

219

220 229 234 234 235 235

236 237

248 249 249 250 250 251 265 265 266 272

Ta b l e s – x v i i

tables TABLE 7.1 7.2 7.3

PAGE 283 286

7.4 7.5 7.6

Typical DC Potentials in Soil Suggested DC Potentials for Cathodic Protection Calculated Resistance and Conductance to Ground of Individual Ground Rods as Related to Soil Resistivity Potentials to a Copper-Copper Sulfate Half Cell Sacrificial Anode Resistance, Output Current, and Estimated Life Conductance to Ground of BCNs with Effective Diameters as Indicated

8.1 8.2

Minimum Cover Requirements Requirements for Random-Lay Joint Trench

304 309

9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10 9.11 9.12

Classifications of Plastic Conduit PVC Duct Dimensions—Minimum Wall Thickness Comparison of Characteristics for Four-Inch Size PVC Duct PVC Duct—Impact Strength (Foot-Pounds) PVC Duct Collapse Pressure (PSI) Conduit Fill Conductor Shield Thickness Insulation Shield Thickness Concentric Neutral Thickness—Aluminum Cables Concentric Neutral Thickness—Copper Cables Secondary Cable Insulation Thickness 220-Mil Primary Cable: Minimum Size of Conduit Necessary to Accommodate Primary Underground Power Cable: 15-kV Cable— 220-Mil Insulation Wall, Concentric Neutral Construction 260-Mil Primary Cable: Minimum Size of Conduit Necessary to Accommodate Primary Underground Power Cable: 25-kV Cable— 260-Mil Insulation Wall, Concentric Neutral Construction 345-Mil Primary Cable: Minimum Size of Conduit Necessary to Accommodate Primary Underground Power Cable: 34.5-kV Cable— 345-Mil Insulation Wall Conduit Fill—Secondary Cable: Minimum Size of Conduit Necessary to Accommodate 600-Volt Secondary Underground Power Cable Recommended Dynamic Friction Coefficients for Straight Pulls and Bends Using Soap/Water or Polymer Lubricants Inside Bend Radius for 90° Schedule 40 Conduits Recommended Maximum Sidewall Bearing Pressures Cable Configuration for Various Jam Ratios Recommended Maximum Pulling Tension Stress for Pulling Eyes on Copper and Aluminum Conductors Recommended Maximum Pulling Tension Limits for Basket-Type Pulling Grips

314 314 314 315 318 320 320 320 320 321 321

9.13

9.14

9.15 9.16 9.17 9.18 9.19 9.20 9.21

288 289 290 291

322

323

324 325 333 335 337 338 339 339

x v i i i – Ta b l e s

tables TABLE 10.1 10.2

11.1

PAGE Electrical Rating of Elbows Relative Corrosion Resistance of Metal Combinations for Outdoor Terminations

350 353

11.5 11.6 11.7 11.8

Dimensions for Primary Cables to ICEA Specification S-94-649-2000 with Concentric Neutral (Concentric Stranding) Dimensions for Primary Cables to ICEA Specification S-94-649-2000 with Concentric Neutral (Compressed Stranding) Cable Diameter Tolerances Adders for Extruded Insulation Shield (Mils) to Obtain Nominal Diameter Over Insulation Shield of Cable DC Proof-Test Voltages (Conductor to Ground) for Primary Cables Insulation Thickness of Secondary Cables Manufacturers’ Voltage Withstand Tests on Completed Cable Manufacturers’ Voltage Tests on Cables Rated 0 to 600 Volts

A.1

Acceptable Outage Hours Per Year Per Consumer

374

B.1 B.2

377

B.5

Allowable Voltage Drop on a 120-Volt Base Resistance of Class B Concentric-Strand Aluminum Cable with Thermosetting and Thermoplastic Insulation for Secondary Distribution Voltages (to 1 kV) at Various Temperatures and Typical Conditions of Installation Corrections for Multiconductor Cables Comparison of Conductor Diameter and Approximate Cable Outside Diameter of Typical Single, Class B Concentric-Strand Aluminum Cables 60 Hz Reactance of Conductors in the Same Conduit

C.1 C.2

Nominal Composite Insulation Layer Thickness (Ruggedized) Nominal Insulation Thickness (Non-Ruggedized)

392 392

E.1 E.2 E.3 E.4

Extruded Conductor Shield Thickness Nominal, Minimum, and Maximum Insulation Thickness Insulation Shield Thickness for Cables with Wire Neutral Extruded-to-Fill Jacket Thickness

400 400 401 402

11.2 11.3 11.4

B.3 B.4

361 362 363 363 367 369 371 371

380 382

382 384

Ta b l e s – x i x

tables TABLE

PAGE

G.1 G.2 G.3 G.4 G.5 G.6 G.7 G.8 G.9 G.10 G.11 G.12 G.13 G.14 G.15 G.16 G.17 G.18 G.19 G.20 G.21 G.22 G.23 G.24 G.25 G.26 G.27 G.28 G.29 G.30 G.31 G.32

Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration Configuration

I.1

Cable Installation Equipment Manufacturers (Trenchers, Backhoes, Cable Plow, Guided Boring Tools, Piercing Tools, Hydraulic Pipe Pusher, Track-Mounted Cable Plows, Trench Compactors, Auger-Type Boring Tools) Cable Installation Equipment Manufacturers (Primary Circuit Joints, Elbows, and Terminations; Secondary Circuit Joints and Terminations) Manufacturers of Joint, Elbow, and Termination Accessories and Kits Partial Listing of Cable Testing Equipment Suppliers

I.2 I.3 I.4

No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No.

1—15-kV Copper 1—15-kV Aluminum 1—25-kV Copper 1—25-kV Aluminum 2—15-kV Copper 2—15-kV Aluminum 2—25-kV Copper 2—25-kV Aluminum 2, 3-Inch Type DB Conduit—15-kV Aluminum 2, 3.5-Inch Type DB Conduit—25-kV Aluminum 3—15-kV Copper 3—15-kV Aluminum 3—25-kV Copper 3—25-kV Aluminum 4—15-kV Copper 4—15-kV Aluminum 4—25-kV Copper 4—25-kV Aluminum 5—15-kV Copper 5—15-kV Aluminum 5—25-kV Copper 5—25-kV Aluminum 6—15-kV Copper 6—15-kV Aluminum 6—25-kV Copper 6—25-kV Aluminum 6, 6-Inch Type EB Conduit—15-kV Aluminum 6, 6-Inch Type EB Conduit—25-kV Aluminum 7—15-kV Copper 7—15-kV Aluminum 7—25-kV Copper 7—25-kV Aluminum

415 415 416 416 416 416 417 417 417 417 418 418 418 418 419 419 419 419 420 420 420 420 421 421 421 421 422 422 422 422 423 423

427 428 429 429

x x – Ex am p l e s

e x a m p l es EXAMPLE

PAGE

1.1 1.2 1.3

Cable Loss Calculations Calculating Losses on Secondary Cables Typical Costs Associated with Transformer Losses

35 36 37

3.1 3.2 3.3

Device Rated in Maximum Asymmetrical Current Capacity Device Rated for Maximum Circuit X/R Ratio Determine Minimum Shield Size for Known Through-Fault Current

83 84 93

4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9

Comparing the Ampacity of Trefoil and Flat-Spaced Configurations Single-Phase UD Cable Ampacities Emergency Overload Rating Cable in Protective Riser Three-Phase Substation Exit Ampacity Average Daily Temperature Selection for a Summer-Peaking Utility Selection of Maximum Permissible Transformer Per-Unit Loading Pad-Mounted Transformer Sizing for New UD Residential Consumers Sizing Commercial Transformers Dedicated Transformer Load

131 140 141 141 146 149 151 157 160

5.1 5.2

No Counterpoise Added (Switches S1, S2, and S3 Open) Attaching a 100-Foot Counterpoise to the Riser Pole Ground Rod and the Other End to a Remote, Smaller Resistance (Switch S2 Closed; S1 and S3 Open) Continuous or Full-Length Counterpoise (Switches S1 and S3 Closed; S2 Open) A Single 8-Foot × 3/4-Inch Ground Rod Driven in Soil with a Resistivity of 250 Ohm-M Two 8-Foot × 3/4-Inch Ground Rods Placed 5 Feet Apart Two Rods Spaced 16 Feet Apart Group of Four Rods Increase in Rod Length Change in Soil Resistivity The Effect of a Two-Layer Soil with a Top-Layer Resistivity of 250 Ohm-M and a Bottom-Layer Soil Resistivity of 50 Ohm-M Counterpoise of #2 AWG Conductor Buried 30 Inches Deep for a Distance of 100 Feet More Conductive Soil Counterpoise Burial Depth Protective Margin Calculation for Riser Pole Application— Industry Standard 4 kA/µs Average Rise Time for Lightning Strokes Assumed MOV Riser Pole Arrester: Arrester Rating, 10 kV MOV Riser Pole Arrester and Dead-Front Cable-End Arrester (No. 4): Arrester Rating, 10 kV MOV Riser Pole Arrester: Arrester Rating, 21 kV

191

5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11 5.12 5.13 5.14

5.15 5.16 5.17

191 191 201 202 202 203 204 204 205 206 206 206

217 234 234 235

Exa m p l e s – x x i

exa m p l e s EXAMPLE 5.18 5.19

PAGE MOV Riser Pole Arrester and Dead-Front Cable-End Arrester (No. 4): Arrester Rating, 21 kV MOV Riser Pole Arrester Plus Dead-Front Cable-End Arrester (No. 4) and Dead-Front Third Arrester (No. 3): Arrester Rating, 21 kV

235 236

6.1 6.2

Maximum Lengths of Cable Circuit Possible Energizing Multiple-Transformer System with Single-Pole

264 272

7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8

Measuring Earth Resistivity Calculating the Neutral Conductance to Ground Per 1,000 Feet of Cable Determining Required Shift in Potential Calculating Required Anode Output Current Selecting Anode Types, Sizes, and Numbers Estimating Neutral Conductance to Ground of BCN Cable Determining Required Shift in Neutral Potential Determining Output Current and Anodes Required

284 288 289 289 291 292 292 293

11.1

Diameter Calculation

363

B.1 B.2 B.3 B.4

Transformer Voltage Drop Calculation Secondary Cable Resistance and Reactance Complete Secondary Voltage Drop Calculation Voltage Flicker Calculation

379 383 385 387

G.1

Ampacity Reduction for Direct-Buried Versus Conduit Encasement for Flat-Spaced Installation Increase in Ampacity for Duct Bank Installation When Type EB Conduit is Used Versus Schedule 40

G.2

J.1 J.2

Cable Pulling Example 1: Maximum Straight-Pull Distance for Three 25-kV Cables Installed in Five-Inch PVC Conduit Cable Pulling Example 2: Feasibility of Pulling Three 25-kV Cables into a Six-Inch PVC Conduit

417 422

431 432


Design of an Underground Distribution Sy s t e m – 1

1 In This Section:

Design of an Underground Distribution System

System Components Types of UD Systems

Economic Comparison of System Configurations

Reliability of UD Systems

UD Loss Economics

Design Considerations

Steps for Layout of a UD System

Future Upgrades and Replacements

Summary and Recommendations

Since their introduction, underground distribution (UD) systems have proved generally popular with electric consumers. Although some of this popularity is due to aesthetics—eliminating pole lines and overhead conductors and “ugly” tree trimming—greater reliability is the greater attraction. Consumers facing outages due to wildlife, falling tree limbs, and ice storms think underground systems more desirable. Unfortunately, many of the present UD systems are less reliable and have more operational problems than do comparable overhead distribution systems. To reverse this trend, cooperatives must undertake several comprehensive steps: 1. Specify high-quality materials and components, 2. Stipulate every safety provision to ensure reliability of the system, 3. Design efficient systems that will have the lowest reasonable cost for both installation and operation, and 4. Plan carefully to minimize problems during construction and provide for future operation and replacement of these systems.

This section gives the engineer guidelines for designing a high-quality UD system. Before starting a design, the engineer must have comprehensive knowledge of the components of a UD system. Next, the engineer must understand how these components can be configured to form different types of UD systems and the special design concerns of each. During the design process, the engineer must consider the following: • • • • •

UD system safety, UD system reliability, UD system operation and maintenance, Future upgrades or replacement, The economics of different system configurations, and • The economics of UD losses. The final design task is layout of the UD system. On completing this task, the engineer will have a final plan and staking sheets to give to construction crews.

2 – Se c t io n 1

1 anyone enters. This requireIn the past, some UD systems ment increases the time were total underground systems A typical UD system needed to access the equipwith all components located consists of buried ment and, thus, also increases below ground. Placing transthe duration of any outage. formers, sectionalizing devices, cables and Because of these problems, and switches below ground repad-mounted a total underground system is quires buried vaults. Because equipment. impractical and unreliable. A water often accumulates in more reliable system consists these vaults, the equipment of underground cables and has to be suitable for operation pad-mounted equipment under water. Moisture also ac(transformers, sectionalizing devices, and celerates the corrosion of this equipment and switches). The pad-mounted equipment is leads to premature equipment failure. placed on the surface instead of below ground. This type of system is very difficult to operate As a result, the equipment is easier to operate and maintain. Maintenance and operation of the and subject to fewer corrosion problems. This equipment usually require a person to enter the type of UD system, with its major system comunderground enclosure. If the enclosure is full ponents, is shown in Figure 1.1. of water, the water must be pumped out before

System Components

Cable Termination Surge Arrester

Underground Cable Riser

Cable Terminations

Pad-Mounted Switchgear/ Junction Cabinet

Pad-Mounted Transformer Flat Pad

Dead-Front Surge Arrester

Box Pad Ground Line

Underground Cable, Secondary Voltage

Cable Splice Ground Electrode

Ground Electrode Underground Cable, Primary Voltage

FIGURE 1.1: UD System Components.

Service Ground


1 configurations possible, this component provides the engineer with many options in the design of a UD system.

UNDERGROUND CABLE The most extensive component of a UD system is the underground cable. The primary-voltage (15-, 25-, or 35-kV class) cable carries power from a source to the primary bushing of a transformer. The secondary-voltage (600-Volt class) cable carries power from the secondary bushings of the transformer to the consumer. Section 2, Cable Selection, describes cable construction and gives guidelines for specifying high-quality cable.

CABLE TERMINATIONS AND JOINTS Cable terminations and joints are other important components of a UD system. The joints provide a way to connect two underground cables. The terminations provide a way to connect underground cables to transformer bushings, switches, fuses, and other devices. Section 10 describes the different types of terminations and how to use them on a UD system.

PAD-MOUNTED EQUIPMENT The main types of pad-mounted equipment are transformers, protective devices, and switching devices. Pad-mounted transformers function the same as those overhead. Pad-mounted switchgear usually functions as a combination of switches and sectionalizing devices. For example, a single enclosure can provide switching on the main feed and fusing on two taps off the main feed. Figure 1.2 shows the schematics for several types of switchgear. Section 3, Underground System Sectionalizing, reviews the different types of pad-mounted switchgear. Because of the many

COMPARTMENT

-2

COMPARTMENT

-2

COMPARTMENT

-3

COMPARTMENT

SURGE ARRESTERS AND GROUNDING ELECTRODES Surge arresters are used to protect underground systems from overvoltages induced by lightning and other transients. To operate effectively, arresters must be properly connected to the cable grounding system. The grounding system must have ground electrodes that are in optimum contact with the soil. Examples of ground electrodes are:

-4

kV

Ampere, RMS Fuse

Nom.

COMPARTMENT

-1

COMPARTMENT

-1

COMPARTMENT

PME-5

PME-4

COMPARTMENT

-1

14.4

17.0

25

27

BIL

MVA 3-Phase Sym. at Load Dropping Rated Voltage

Max

Cont.

95

200

600

600

350

125

200

600

400

540

PMH-6

COMPARTMENT

-3

COMPARTMENT

-4

COMPARTMENT

COMPARTMENT

-2

COMPARTMENT

-1

COMPARTMENT

PME-9

-2

Max

Short-Circuit

Mini-Rupter

-3

COMPARTMENT

-4

COMPARTMENT

-2

COMPARTMENT

-1

COMPARTMENT

PME-10

-3

COMPARTMENT

-4

COMPARTMENT

-2

COMPARTMENT

-1

COMPARTMENT

PME-11

FIGURE 1.2: Schematics for Different Types of Switchgear. Adapted from S&C Electric Company, 2005.

-3

COMPARTMENT

-4

-2

COMPARTMENT

-1

PME-12

4 – Se c t io n 1

1 • • • •

Driven ground rods, Buried counterpoise wires, Semiconducting jacketed cables, and Metallic water or sewer systems.

Figure 1.1 shows driven ground rods as the ground electrodes. Detailed information on cable grounding systems and surge protection is contained in Section 5.

compartments of transformers, fuse cabinets, or switchgear. If the settling is severe, the pad may not support all the equipment weight. If some of the equipment weight is transferred to the attached cables, this settling can damage transformer bushings, connectors, and switch terminals.

Types of Equipment Mountings The most basic type of equipment mounting is a flat, or monolithic, pad. The flat pad provides a EQUIPMENT MOUNTINGS uniform surface for mounting equipment and has Equipment mountings provide a flat, rigid suropenings for cable access into the equipment enface for supporting pad-mounted equipment. It closure as shown in Figure is very important to mount the 1.3. Because this pad is placed bottom edge of pad-mounted directly on the ground, there equipment flush to the flat surThe soil beneath is limited space for cable trainface of the supporting pad. ing and cable terminations. Doing so prevents persons the pad must be However, this type of pad is from poking a wire or other well compacted. usually adequate for singleobject into the interior comphase pad-mounted transformpartment of pad-mounted ers and small single-phase equipment and meets the sectionalizing devices. requirements of American National Standards Some types of cable installations require more Institute/Institute of Electrical and Electronics space than is available with a flat pad. For exEngineers (ANSI/IEEE) C57.12.28 (Standard for ample, large-diameter cables are stiffer and have Pad-Mounted Equipment-Enclosure Integrity) a larger minimum bending radius than do smalland ANSI/IEEE37.74 (Standard Requirements diameter cables. Thus, the large-diameter cables for Subsurface, Vault and Pad-Mounted Load require more space for cable training. Another Interrupter Switchgear and Fused Load-Interconsideration is cold weather. Low temperatures rupter Switchgear for Alternating Current Sysmake cables stiffer and more difficult to install tems Up to 38 kV). The former code has or operate. Providing additional cable space become a standard for specifying tamper-resishelps minimize these problems. Therefore, cotant pad-mounted equipment enclosures. This operatives in areas with extended periods of tamper-resistant design helps prevent vandalism cold weather may prefer using a ground sleeve to utility equipment and protect the public from (“basement”) or a box pad instead of a flat pad. contact with energized parts. A ground sleeve or box pad also provides the The equipment must also attach securely to extra space needed for large-diameter cables. the mounting surface to prevent it from being Typical installation of a ground sleeve is moved or tipped over by people, animals, lawn shown in Figure 1.4. The ground sleeve is inmowers, or vehicles. Secure attachment is particstalled below the ground surface, with the ularly important when polyethylene pads are equipment mounting surface elevated two to used. The pad’s slick surface makes it easy for three inches above final grade. This type of an unsecured piece of equipment to slide. mounting provides additional space for cables Another important factor in a stable installabelow grade, but is suitable for equipment with tion is proper soil compaction beneath the pad. only one entry compartment such as three-phase Without proper compaction, the soil will settle pad-mounted transformers and junction cabinets. and erode, leaving the pad with little support. Ground sleeves are generally limited in their When this happens, pads can tilt or warp (if ability to support heavier pieces of equipment. made of polyethylene) and expose the interior


1

FIGURE 1.3: Flat Pad for Equipment Mounting.

FIGURE 1.5: Box Pad for Equipment Mounting.

provides plenty of room to work with the cables. This type of pad is ideal for supporting pad-mounted switchgear that has multiple cable entry compartments. Pad Materials Manufacturers offer a varied selection of pad materials, including the following: • • • •

Steel-reinforced concrete, Fiberglass-reinforced concrete, Fiberglass, and Polyethylene.

Because these materials have very different properties, the engineer must carefully select the material type suitable for the intended application. The material and pad design must have the strength required to support the FIGURE 1.4: Ground Sleeve. Source: Nordic equipment weight. This is of particular conFiberglass Inc., Warren, Minn., 2002. cern with box pads, because all the equipment weight is supported by the outside pad walls, The third type of mounting and is especially important, is a box pad (see Figure 1.5). for example, when box pads The box pad is placed in the are used for transformers Pad material must ground rather than on the sur500-kilovolt amperes (kVA) face, with typically three to six and larger. Care must be exbe suitable for the inches exposed above grade. ercised in making sure the intended application. A perimeter lip supports the box pad manufacturer clearly pad-mounted equipment. The states the strength rating of remaining space is open and the box pad walls.

6 – Se c t io n 1

1 Also of concern are polyethylene pads with wooden braces. A puncture through the polyethylene allows water to enter the pad and rot the wooden braces. When the wooden braces rot, part of the pad strength is lost, and warpage results. A second property to review is the performance of the material outdoors where it is exposed to frost and ultraviolet radiation. The pad materials must not break down or crack from ultraviolet exposure or frigid conditions. Cracks or material breakdown lead to a loss of mechanical strength.

Types of UD Systems

A final property to review is pad buoyancy. Some of the polyethylene pads tend to float and can overturn pad-mounted equipment. Therefore, these pads would not be suitable for use in areas that are subject to flooding. In summary, pads must be of a design that will have long-term durability under adverse conditions, meet system operating needs, and maintain equipment security. All these factors must be balanced when selecting a pad design for a particular UD system.

SUBSTATION CIRCUIT EXITS Disconnect Switches Underground cable is often used for substation circuit exits from distribution substations. Underground circuit exits help reduce congestion on poles just outside a substation, making the area around a substation more attractive and workCable able. As an added benefit, underground substaSurge Termination Arrester tion circuit exits are protected from ice loading, wildlife contacts, and vehicle damage, and, thus, may be more reliable than overhead exits. Neutral In most cases, each underground substation circuit exit will terminate on a riser pole and feed overhead circuit conductors. Therefore, this type of UD system consists of underground primary-voltage cable, cable terminations, surge arresters, and grounding electrodes. The conduit, cable terminations, surge arresters, grounding electrodes, and disconnect switches are commonly Riser Vent referred to as a riser assembly. See Figure 1.6. When designing underground substation circuit exits, the engineer must be particularly concerned Undergroung Circuit with reliability. If the underground cable fails, the Exit Cable circuit outage interrupts power to many consumers. Ground Placing the cable in a conduit system or concreteElectrode encased duct bank helps protect it from mechanical damage. Section 9 contains information on FIGURE 1.6: Underground Substation duct bank installations. Another way to improve Circuit Exit. reliability is to install a spare cable or provide backup capability from another source. AlDesign concerns for A special concern for unthough spare cables or backup derground circuit exits is cable substation circuit exits options do not change the risk ampacity. These cables carry of cable failure, they do reduce are reliability, system large loads and may operate the power restoration time if close to their ampacity rating. growth, and ampacity. only one cable is damaged. Therefore, the engineer must


1 carefully determine the cable operating conditions, system growth, and the resulting ampacity. MAIN FEEDERS Underground cable can serve as a distribution main feeder. A main feeder is that portion of a distribution circuit between the substation and the first in-line overcurrent protective device. The protective device in the substation clears a fault on a main feeder. Therefore, a main feeder fault causes an outage to the entire circuit. Because most faults on an underground main feeder are cable failures and are permanent, power to the circuit may remain off until the cable is repaired. The utility engineer must consider this characteristic when designing a main feeder, particularly when deciding between a radial or open-loop feeder. The engineer must also determine the maximum load to be carried by the main feeder in order to select a cable with adequate ampacity

and choose the 200-ampere or 600-ampere class of cable terminations. Section 4 provides detailed information on cable ampacity, and Section 10 provides information on the types of cable terminations. Radial Main Feeder The radial main feeder has one source and delivers power to a load area along a single path. This feeder can also serve several load areas by using a junction box or sectionalizing switch with fused taps. This type of arrangement is shown in Figure 1.7 and may have the following components: • Underground primary-voltage cable, • Cable terminations, • Pad-mounted junction box or sectionalizing switch, • Surge arresters, and • Grounding electrodes.

To Load Area Junction Box or Switching Cabinet Junction Box or Sectionalizing Switch Primary Voltage Cable

To Load Area Substation

Junction Box or Sectionalizing Switch To Load Area

FIGURE 1.7: Radial Main Feeder.

8 – Se c t io n 1

1 Power On

Open Load-Side Switch

Open Power On

Junction Box or Sectionalizing Switch Fault

Power Off

Substation

Power Off

FIGURE 1.8: Radial Main Feeder with Faulted Cable Section.

The junction box or sectionalizing switch provides sectionalizing of the load areas and limited sectionalizing of the main feeder. For example, consider a fault in the second line section as shown in Figure 1.8. This fault trips the protective device at the substation and interrupts power to all consumers on the faulted circuit. The cooperative can restore power to the first load area by placing the faulted cable(s) in a parking stand, or by opening the load-side switch on the first sectionalizing switch to isolate the faulted cable. Figure 1.8 shows this option. Because the radial feeder has no alternative source or path, the cooperative cannot restore power to the other consumers until crews repair the cable fault. It is possible to improve the reliability of a radial system by installing the cable in a concrete-encased duct bank or in a conduit system, or by installing a spare cable or conduit in the trench. A concrete-encased duct bank provides

substantial mechanical protection from dig-ins and should be considered in areas congested with other underground utilities. A conduit system provides limited mechanical protection. However, it does decrease outage time by allowing the cooperative to replace a section of faulted cable without disturbing the earth surface. This saves substantial time, particularly when the main feeder is located beneath a roadway. The spare cable or conduit provides no mechanical protection but does decrease restoration time if only one cable is faulted. Because the costs of these installation methods vary significantly, each cooperative must weigh the advantages of these more expensive installations against their costs. Under any circumstance, the simple radial does have limited operational flexibility and should not be used to serve a large number of consumers. Information on comparative system reliability may be found in Appendix A.


1 stand. In a sectionalizing switch, leaving one of the switches open creates an open point. The open-loop feeder (see Figure 1.9) provides much higher system availability than does the radial system. With an open-loop system, utility crews can isolate a faulted cable section and restore power to all consumers. A cable fault in the second line section interrupts power to all consumers on that circuit. After isolating the faulted cable section, as shown in Figure 1.10, crews can feed the first section from Substation No. 1 and remaining line sections from Substation No. 2. Because crews can restore power to all

Open-Loop Feeder In dense load areas, an underground main feeder may tie together two substations. A main feeder may also tie two circuits from the same substation. This type of arrangement would operate as an open-loop system. The components of this system are the same as those of a radial system. However, the open-loop feeder has two sources, unlike the radial feeder that has only one source. Each source provides power along a single path to the designated open point in a junction box or a sectionalizing switch. In a junction box, the open point results from placing one set of cables in a parking

Substation No. 2

Substation No. 1 Looped-Primary Circuit

N.O. N.O. = Normally Open Point Three-Phase, Pad-Mounted Transformer Sectionalizing Switch

FIGURE 1.9: Open-Loop Feeder.

Substation No. 1

Substation No. 2 Looped-Primary Circuit

Fault N.O. N.O. Three-Phase, Pad-Mounted Transformer

N.O. = Normally Open Point

Sectionalizing Switch

FIGURE 1.10: Open-Loop Feeder with Faulted Cable Section.

1 0 – Se c t io n 1

1 load areas before repairing the cable fault, the outage time is much shorter than with a radial feeder. As a result, it is not critical to install the cable in a concrete-encased duct bank or conduit. However, as already noted, in areas congested with underground utilities, the concrete-encased duct bank will help protect cables from dig-ins. Again, it is important to judge the benefits of installing duct bank or conduit against the additional cost. An open-loop feeder also requires that the designer consider the ampacity of the feeder cables while serving all possible loop segments, which may dictate the use of a larger cable size than otherwise needed. Open-loop feeders provide much more operating flexibility than do simple radial feeders. System reliability considerations generally dictate open-loop feeders as the preferred design.

deliver power to consumers. Therefore, sections of cable on a sub-feeder often terminate in padmounted transformers. The sub-feeder can have several configurations ranging from a simple radial feeder to a complex multiloop feeder. Radial Feeder The simplest type of load area feeder is a radial feeder. The radial feeder is usually the most practical way to serve a single consumer. However, a single consumer with critical loads, such as a hospital or police station, often requires a more reliable system. Methods for improving reliability include the following: • Changing to an open-loop configuration, • Adding a spare cable or conduit to the trench, and • Placing the cable in a conduit or duct bank.

SUB-FEEDERS The radial feeder can be extended to serve The more common underground feeder is the submultiple consumers as shown in Figure 1.11. feeder, also called a load area feeder. This type However, a cable fault interrupts power to all of feeder has at least one stage of sectionalizing consumers beyond the fault location. For exambetween it and the protective device at the subple, a fault between transformers T1 and T2 restation. As a result, a fault on a sub-feeder does sults in a power outage to not interrupt power to the entransformers T2 through T5. tire circuit and, thus, affects The power remains off until fewer consumers than does a A cable fault on a the cable is repaired. As the similar fault on a main feeder. sub-feeder affects number of consumers increases, The two types of feeders it becomes more practical to also have different functions. fewer consumers than consider an open-loop system. The basic function of a main does a similar fault The subsection Economic feeder is to deliver power to on a main feeder. load area feeders. The main Comparison of System Confunction of a sub-feeder is to figurations, which comes later

Single-Phase, Pad-Mounted Transformers

Fault Riser Pole

T1 Power On

FIGURE 1.11: Radial Feeder.

T2 Power Off

T3

T4

T5

Design of an Underground Distribution Sy s t e m – 1 1

1 Three-Phase Feeder

Riser Pole

Riser Pole

Three-Phase, Pad-Mounted Transformers

Normally Open Point

FIGURE 1.12: Open-Loop Feeder in Shopping Center.

Riser Pole

Riser Pole

Sectionalizing Switch

Sectionalizing Switch N.O.

N.O.

N.O. Legend N.O.

Three-Phase, Pad-Mounted Transformer Single-Phase, Pad-Mounted Transformer

N.O.

Normally Open Point

N.O.

FIGURE 1.13: Multiple-Loop System. in this section, provides information on the economics of radial versus open-loop systems. Open-Loop Feeder As mentioned earlier, the open-loop feeder has two sources and, therefore, provides better system availability. Large subdivisions or commercial shopping areas are ideal applications of open-loop systems. Figure 1.12 shows an openloop feeder in a shopping center. Utility crews can isolate any section of faulted cable and restore power to all transformers. This feature makes the open-loop feeder a preferred design

for UD systems serving multiple or critical consumers. An open-loop feeder also requires that the designer consider the ampacity of the primary cables and devices while serving all possible loop segments, which may dictate the use of a larger cable size than otherwise needed. Multiple-Loop Feeder In heavy load areas, multiple-loop feeders are necessary to improve sectionalizing and to allow the coordination of overcurrent protective devices. A typical multiple-loop system is shown in Figure 1.13. This type of system usually has a sub-

1 2 – Se c t io n 1

1 feeder that serves as an open-loop system between two sources. The sectionalizing switches on the sub-feeder have fused taps that serve other open-loop feeders. This arrangement provides excellent system availability. It also speeds up fault location because the large load area has been sectionalized into small load groups. A multiple-loop feeder also requires that the designer consider the ampacity of the feeder cables and devices while serving all possible loop segments, which may dictate the use of a larger cable size than otherwise needed. TRANSFORMER AND SECONDARY SYSTEMS Pad-mounted transformers and underground secondary-voltage cable constitute the final segment of a UD system. To properly design this part of the system, the engineer must first select the appropriate equipment rating and cable ampacity. Section 4 provides information for making these selections. Second, the engineer must consider reliability. Most secondary cable faults are the result of mechanical damage to the cable. Utilities can minimize mechanical damage by following the proper installation techniques described in Section 9 and by specifying cable with an abrasion-resistant

Lighting Package

Pole Cable Riser

Ground Electrode

Underground Secondary-Voltage Cable

FIGURE 1.14: Area Lighting System.

or self-healing insulating jacket (see Section 2). Cable dig-ins by other utilities or consumers also damage cable. To minimize dig-ins by consumers, cable should be installed two to three feet off the property line. Doing so helps prevent cable damage if the consumer installs a fence on the property line. Another method for minimizing dig-in damage is to use conduit. The conduit offers some mechanical protection, particularly from hand digging. As noted, the cooperative may particularly want to use conduit in areas congested with other utilities. A third design concern with secondary systems is voltage drop and voltage flicker. The engineer must design a system that provides the consumer with acceptable voltage levels throughout the day and during motor starting. Appendix B lists the acceptable voltage levels and gives methods for calculating voltage drop and flicker. STREET AND AREA LIGHTING Public safety and consumer convenience require street and area lighting in the area served by a large percentage of underground projects. Most cooperatives furnish this service, so the engineer must make accommodations in underground systems to include it. The engineer needs to develop a plan at the start of the project for eventual (if not actual) street and area lighting. Conduits and pedestals can then be installed at strategic locations that will minimize future trenching in lawns or around consumer facilities. This type of UD system is shown in Figure 1.14. It uses a combination of overhead components (poles and a lighting package) and underground components (underground secondary-voltage cable, surge arresters, and grounding electrodes). Street and area lights are generally self-contained units with an integral photoelectric cell for control. These standard light packages usually operate from 120-Volts single phase or 120/240Volts single phase. The cooperative may want to consider using the same lighting package that it uses in overhead areas. Doing so will avoid unnecessary duplication of stock and minimize confusion during installation and maintenance. If the lighting package requires a 120-Volt, two-wire power supply, service may be provided through a two-wire duplex underground


1 current will travel along the cable. If the cooperative has a lighting conductors and be large amount of underground Metallic lighting poles propagated into the secondary street lighting, purchasing a must be grounded of the transformer and into all twisted duplex cable with a connected services. In areas ruggedized insulation system and bonded to the with intense lightning activity, will be most economical. This system neutral for the cooperative should consider cable will essentially comply installing secondary lightning with the secondary cable speclightning protection arresters on each transformer ification presented in Appenand for public safety. that serves a lighting installation. dix C. When this duplex is Where aesthetics are of prime used, the conductor may be importance, cooperatives may either copper or aluminum. choose to install metal lighting poles. In such When aluminum is used, the size should not be cases, the height of the fixture mounting should smaller than No. 6 American Wire Gauge (AWG). not be compromised; it should be installed in Satisfactory performance may be achieved with accordance with standard practices for the particcopper conductors as small as No. 10 AWG. In ular type of light and the size of the area to be areas where deep frost lines are routine, larger lighted. With metal poles, the pole interior may aluminum conductors, possibly No. 2 AWG, generally be used as a raceway to conceal the might be considered as a minimum gauge. conductor along its entire length. In these cases, In cases of infrequent use or where ruggedisunlight resistance will not be required on Type zed duplex cable is not readily available, UF cables if the cables are shielded from sunType UF (underground feeder) commercial light along their entire length. Metal poles will cable may be substituted. This cable should be still require adequate grounding to avoid probpurchased only with copper conductors No. 10 lems with lightning surges. Metallic poles should AWG or larger. The Type UF cable must be also be directly connected to this same groundrated as sunlight-resistant. Otherwise, the cable ing system, which is also positively connected to may deteriorate where it is exposed to sunlight the neutral of the secondary supply conductors. between the pole riser conduit and the bottom If the poles are direct buried, they generally of the lighting support bracket. have an insulating coating for corrosion protecLighting packages may be installed on wood tion. If direct-buried poles are installed or if the poles at a height appropriate for the size of the poles are installed on poured concrete foundalamp and the area to be lighted. On wood tions, a ground rod is also recommended. If poles, polyvinyl chloride (PVC) conduit should poles are installed on a metal screw anchor be used to protect the cable riser. Schedule 40 base, the ground rod may be eliminated. PVC is recommended as a minimum. U-guards The main limitation on the layout of street are not recommended because irregularities in lighting conductors is voltage drop. As most wood poles may allow the smaller cable used contemporary lighting systems are either merfor lighting service to protrude or be pinched cury vapor, metal halide, or high-pressure between the U-guard and the pole surface. Each sodium systems, the most critical case is during wood pole installation must be equipped with a starting of the most distant light. This is the time pole-grounding conductor (No. 6 AWG copper) of highest current draw and lowest power factor. that is attached to a driven ground rod. This is The magnitude and power factor of the starting particularly important because street and area current depend on the type of ballast, as does lights are often among the highest objects in a the acceptable voltage range for satisfactory opsubdivision served by an underground system. eration. Table 1.1 gives examples of typical light In cases of lightning strikes, the lightning must characteristics. It is obvious that the regulator have a relatively low impedance path into the ballasts offer a substantial advantage in allowing earth. If pole grounding conductors are not inlong runs of small secondary voltage conductors stalled, a much larger portion of the lightning

1 4 – Se c t io n 1

1 TABLE 1.1: Lamp and Ballast Characteristics—240 Volts. Source: General Electric Lighting Systems Product Catalog 1985.

Lumens

Allowable Voltage Fluctuation

Operating Current (amperes)

Starting Current (amperes)

Power Factor

Allowable Voltage Dip

175-watt mercury vapor, normal power factor reactor ballast

7,950

240V±5%

1.6

2.6

55%

20%

400-watt mercury vapor, normal power factor reactor ballast

21,000

240V±5%

3.4

5.1

54%

20%

400-watt mercury vapor, regulator ballast

21,000

240V±13%

2.1

0.9

90%

50%

100-watt high-pressure sodium, normal power factor reactor

9,500

240V±5%

1.6

1.9

34%

10%

250-watt high-pressure sodium, normal power factor reactor

27,500

240V±5%

2.8

3.6

42%

10%

100-watt high-pressure sodium, high power factor reactor

9,500

240V±5%

0.6

0.9

90%

10%


27,500

240V±5%

1.4

2.4

90%

10%


50,000

240V±5%

1.9

3.8

90%

10%

250-watt metal halide floodlight, normal power factor reactor ballast

20,500

240V±10%

1.3

1.0

90%

10%

400-watt metal halide floodlight, normal power factor reactor ballast

36,000

240V±10%

2.0

1.7

90%

10%

Size and Type

Reliability of UD Systems

without unstable lamp operation. Moreover, all types of high-pressure sodium and metal halide lamps are more sensitive than are mercury vapor lamps to voltage dips. Therefore, all lighting circuits should be designed for a voltage drop of

no more than 10 percent when the largest probable lamp is started. Consideration should also be given to selecting 240-volt ballasts as opposed to 120-volt units, because they draw less current and generally create decreased operating losses.

One of the most important advantages of a welldesigned UD system is greater reliability for consumers compared to an overhead system. UD lines and equipment are located where they are not vulnerable to most of the common hazards that cause outages on overhead facilities, such as trees, weather, some animals, and vehicles. However, material or design defects in a UD system may reverse the reliability advantage of underground distribution. In fact, many early UD systems installed by cooperatives and other utilities turned out to be less reliable than comparable

overhead systems. These experiences have made it clear that reliability engineering is a necessary part of UD system design. MEASUREMENT OF RELIABILITY Reliability is usually measured in two ways. The first is the frequency of interruptions occurring at a particular point on a system, referred to as the interruption rate or outage rate. Outage rates are measured in outages per year. The second measure is the average duration of an interruption, also referred to as the restoration time.


1 Outage duration is usually measured in hours. A combination of these two measurements yields the percentage of availability for a particular location on a distribution system. A simple index of reliability used by many utilities is hours of outage per year, per consumer. For this discussion, outages are considered to be sustained interruptions. Reliability calculations of this type usually do not consider momentary interruptions that are successfully cleared by automatic circuit reclosing operations. This analysis considers only those outages that require manual intervention to restore service. Furthermore, almost all faults attributable to underground system components are permanent. System reliability undeniably affects many aspects of a cooperative’s service. Although the order of importance may vary with individual situations, the results of distribution system outages include the following: • Consumer dissatisfaction; • Consumer financial losses resulting from interrupted production, equipment damage, or other causes; • Impairment of other cooperative facilities; • Costs to the cooperative of service restoration; and • Lost cooperative revenue. All these factors have a serious impact on satisfactory cooperative system operation. Engineers, therefore, must be aware of the basic principles of reliability assessment so they can achieve satisfactory but economical UD system designs. Appendix A provides a method for calculating UD system reliability. Comprehensive reliability analysis also considers the number of consumers or kVA of load each outage affects. Thus, facilities serving many consumers (or kVA) may need to be designed for higher reliability than should facilities serving few consumers (or kVA). The analysis presented in this manual, however, does not consider this parameter because most cooperative UD systems are fairly uniform in design and consumer concentration. There is generally no need to discriminate in design quality between some parts of the system and others.

CABLE FAILURE RATES In the mid-1980s, the failure rates for commonly used UD primary cables were unacceptable. The failure rates for cross-linked polyethylene (XLPE) and high-molecular-weight polyethylene (HMWPE) cables were approaching 0.02 and 0.08 per mile per year, respectively. Furthermore, studies revealed that these failure rates were continuing to worsen as the cables aged. The most common causes of failure were electrochemical treeing of the insulation layer and corrosion of the exposed neutral conductors. In December 1987, the Rural Electrification Administration (REA), currently called Rural Utility Services (RUS), responded to the cable failure problem by issuing a revision of Bulletin 50-70 (U-1), REA Specification for 15-kV and 25-kV Primary Underground Power Cable. The main specification changes were the following: • Removing all HMWPE cable from approval, • Increasing minimum insulation thickness to 220 mils for 15-kV cable and to 345 mils for 25-kV cable, and • Requiring cable to be jacketed. At that time, RUS did not disapprove the use of XLPE cable. Nevertheless, concerns about XLPE were raised in studies, leading to the bulletin’s revision. As a consequence of these experiences in the 1980s, cooperatives should procure new cable with the requirement that the revised RUS specifications be met. Any XLPE cable acquired should also be tree retardant (TR-XLPE). As a result of recent vastly improved quality control in cable manufacturing processes, both TR-XLPE- and ethylene propylene rubber- (EPR) insulated cables provide improved reliability. Industry tests are continuing to develop information on the expected failure rates for different insulation systems. RUS is currently preparing an even further refined U-1 specification to reflect these continuing cable insulation improvements. Section 2 discusses cable selection in detail. LOOP-FEED DESIGN The time spent to locate an underground cable fault, excavate to the point of its failure, and

1 6 – Se c t io n 1

1 Riser Pole

Riser Pole

T1

T6

T2

T5

Legend Single-Phase, Pad-Mounted Transformer N.O.

Normally Open Point T4

T3 N.O. Transformer T4 Parking Stand Surge Arresters

X3 X1 X2

Copper Ground Conductor To T3

To T5

To Ground Rod

FIGURE 1.15: Loop-Feed Design of UD System Under Normal Conditions.

Riser Pole

Riser Pole

T1

T6 Damaged Cable Section

Legend

T2

T5

T3

T4

Single-Phase, Pad-Mounted Transformer

Transformer T5

Transformer T6

Parking Stand

Parking Stand

Surge Arresters

Surge Arresters

X3

X3

X1

X1 X2

Copper Ground Conductor

To Ground Rod

To T4

install a UD cable repair joint is typically much longer than that required to perform a comparable repair on an overhead line. Therefore, if the overhead type of radial distribution system configuration were used for UD, the restoration time for most UD outages would be much longer than is typical on overhead systems. This difficulty is overcome by using loop-feed design for UD systems. Under loop-feed design, each cable run serving several pad-mounted transformers is connected with a power supply point on both ends (see Figure 1.15). This formed loop is opened at some point to allow use of radial overcurrent protection methods and to prevent unwanted power transfers through the cable. If the cable fails, a repair crew can disconnect both ends of the failed cable section and close the circuit at the normal open point (see Figure 1.16). These actions promptly restore service to all consumers on the cable run. The damaged cable can then be repaired or replaced later without causing additional outage time. It must be noted that it is vitally important for loop-feed UD systems to be fed from two sources of the same feeder circuit out of a substation, with no switching or sectionalizing devices in between. Having the two sources fed from different feeder circuits could cause unexpected high-power flow through the UD system if the sources were tied together during switching operations on the UD loop. These high current levels could result in exceeding cable and/or termination current-carrying ratings, or could create outages on source fusing devices. Furthermore, on single-phase UD looped systems, it is vitally important that both sources be connected to the same phase for safe operation.

X2 Cable Fault

To Ground Rod

To Riser Plate

Front View Showing Isolated, Damaged Cable Section

FIGURE 1.16: Loop-Feed Design of UD System with Damaged Cable Section.

UD SYSTEM RELIABILITY STUDY Well-designed UD systems can provide improved reliability relative to overhead systems. However, to achieve high reliability, the cooperative needs to apply the specialized engineering knowledge gained from many years of experience with underground power distribution. This knowledge covers the field performance records of different types of cables, the proper application of surge arresters, appropriate sectionalizing, and loop-feed designs, all of which are treated by this manual.


1 Design Considerations for System Operation and Maintenance

The cooperative’s involvement with a UD system does not end after installation; the cooperative must operate and maintain the system throughout its life. Because many components of a UD system are difficult to access, operation and maintenance of the system can also be difficult. For example, it is difficult to access a pad-mounted transformer that is surrounded by shrubbery or located too close to fences or buildings. Likewise, it is difficult to repair a faulted cable that is buried beneath landscaped areas or utility buildings. The engineer needs to be aware of these problems when considering whether to place facilities along the front or rear property line and also must consider the effect of joint-use trench on operation and maintenance activities. FRONT VERSUS REAR PROPERTY LINE PLACEMENT One of the fundamental choices in UD system design is whether to locate facilities along the front property line or along the rear property line. Usually, this is a joint decision between the utility and the consumer or developer. Consumers or developers will have some authority because they must normally give the utility an easement that allows the installation of underground facilities. Often the consumers or developers believe that pad-mounted equipment detracts from the

TABLE 1.2: Front Versus Rear Property Line Placement. Location

Advantages

Disadvantages

Placement along front property line

1. More accessible for operation and maintenance

1. More unsightly to consumer

2. Usually more accessible for installation

2. Greater potential for dig-ins 3. Potential for damage from vehicles

3. Often reduces outage time 4. Reduces cable replacement costs Placement along rear property line

1. Consumers preference for equipment in backyard

1. Often requires more tree/ brush clearing

2. Possible more economical installation if lots share rear property lines

2. Difficult to access for operation and maintenance 3. Usually higher cable replacement costs

appearance of the property so they prefer the utility to locate facilities along the rear property line instead of in front of their houses. However, equipment along the rear property line is usually difficult to access and thus difficult to operate and maintain, particularly when there is no service alley or backyards are fenced and have no access gate large enough to accommodate a trencher or backhoe. In addition, the rear property line is not usually cleared of trees and may not be to final grade when cable is installed. As a result, preparing the rear of the lot for cable installation can be more costly and time-consuming than preparing the front property line. The installation cost also depends on the subdivision layout and the location of other underground utilities. An economic comparison of front versus rear property line installation is covered later in this section under Economic Comparison of System Configurations.

A final consideration is the power restoration time following an outage. When facilities are located on the front property line, it is much faster for utility crews to check for tripped fault indicators and to perform cable switching to isolate the faulted cable section. It is important that the utility engineer inform the consumer of this advantage of front-line placement. Table 1.2 summarizes the advantages and disadvantages of front and rear property line placement. The engineer can be guided by this table in selecting the cable route. In most cases, placement along the front property line is more advantageous. However, subdivision layout, the location of other utilities, or consumer relations may require placement along the rear property line. In these cases, installing the cable in conduit or installing a spare conduit allows the utility better access when cables have to be repaired or replaced. JOINT-USE TRENCH In some areas, the space allocated for underground utilities is very limited. In these areas, the utilities may agree to place facilities in a common trench. Within this common (joint-use) trench, the different utilities usually maintain a minimum separation of 12 inches. The 2007 National Electrical Safety Code (NESC), Section 354,

1 8 – Se c t io n 1

1 does allow the random separation (less than 12 inches) of some utilities. Section 8 of this guide, Direct-Buried System Design, contains information on the NESC requirements and installation guidelines for joint-use trench.

dimensions and arrangement of all utility lines. The utility that opens the trench must abide by these dimensions. Second, the contract should define who is responsible for installing the facilities. If each utility installs its own facilities, then the contract needs to state the required notification period before opening and backfilling a trench. If the other utilities receive proper notification but fail to send crews, the contract should stipulate any consequences, such as those below:

Operational Precautions Before agreeing to share a common trench, the cooperative should consider the potential for operational problems. Each utility will have to maintain its own facilities, which may require • Will the trench be closed or covered crossing other utilities to reach its facilities. To temporarily? minimize the risk of damaging other facilities • Will the delinquent utility be charged? during excavation, operation crews need a • Will a closed trench be reopened? drawing that shows a trench cross section and the location of all facilities within the trench. It Third, the contract should state who is reis also helpful to show the presence of joint-use sponsible for acquiring easements and any pertrench on the operating map for the area. mits. The utility that opens the A joint-use trench with rantrench should require copies dom lay of electric, telephone, of the easements and permits and cable television (CATV) Joint-use trench with before starting construction. cables creates additional operrandom separation Also before construction, any ating problems. This type of existing underground facilities trench requires telephone and often creates must be located. The contract CATV personnel to work next operating problems. must identify who is responsito power cables. Jacketed ble for requesting the location power cables resemble teleof these utilities. Special backphone cables. The cables must fill and compaction needs be well marked to prevent must be addressed. If select backfill is required, mistaken identity. The NESC requires all directthe contract should identify the party responsiburied, jacketed, primary-voltage cable to have a ble for acquiring the backfill material and decide specific marking on its jacket. This marking is how the additional cost will be shared among shown as Figure 350-1 in the 2007 NESC. The the utilities. NESC also has special requirements for bonding Fourth, the contract should address shared and grounding of electric, telephone, and CATV costs. These costs include the following: systems using random separation in Section 354D. Grounding and bonding are discussed • Cost to open and close the trench, further in Section 8. • Cost of the service if one utility installs all facilities, Typical Contractual Arrangements • Penalties for reopening a trench, Joint-occupancy trenches require tremendous • Penalties for temporarily covering or coordination and cooperation from each utility barricading an open trench, and involved. To help structure these efforts and • Cost adders for select backfill. provide proper agreements on liability, the cooperative must prepare a contract for joint trench Fifth, the contract should state that it is transuse. This contract would be similar to the conferable to a new owner. This transferability is tract for joint pole use. particularly important for joint-use contracts with First, the contract should address construcCATV utilities. tion concerns. It must state the required trench


1 Future Upgrades and Replacements

install dual-voltage transformers and sectionalizThe cooperative engineer can improve the deing devices rated for the higher voltage level. sign of a UD system by anticipating and providThe economics of these changes depend on the ing for future system upgrades. Changes to a subdivision layout and the number of years besystem in established yards, parking lots, or fore the voltage conversion. Before making roadways are very expensive. If trenching meththese design changes, the enods are used, the utility must gineer needs to do an ecoalso restore the soil surface. nomic study similar to the one Such restoration could include A UD system design described in Future Voltage reseeding grass, repaving, or should provide for Conversions under the Ecopouring new concrete sidenomic Comparison of System walks or driveways. Trenching future upgrades. Configurations subsection bein established yards also tends ginning on the next page. to create conflicts with property owners. The engineer can help avoid these problems THREE-PHASE VERSUS by planning for future conversions to three-phase SINGLE-PHASE INSTALLATION circuits and higher voltage levels. The engineer Most large subdivisions are developed in stages can also plan for future cable replacements by over time. For these types of subdivisions, the considering the use of conduit systems. engineer should determine if a three-phase feeder is required. A three-phase feeder is often helpful for balancing a large amount of singleFUTURE VOLTAGE CONVERSIONS phase load and for providing better sectionalizMany utilities are converting to higher distribuing. The future subdivision plans may show a tion system voltages to decrease line losses, imclubhouse or sewer lift station. These types of prove circuit voltage profile, and increase loads are often three-phase and, thus, require a system capacity. These conversions are typically three-phase primary circuit. scheduled to occur over an extended time. The If the engineer thinks the subdivision will engineer will, therefore, need to refer to the eventually require three-phase power, he should long-range work plan to locate those areas desconsider installing a three-phase feeder instead ignated for future voltage increases. For UD sysof a single-phase feeder. It is much easier to intems in these areas, the engineer needs to adapt stall three cables initially than to install one inithe design to minimize material and equipment tially and two later. The subsection immediately changeout at the time of voltage conversion. following, titled Economic Comparison of SysA simple design change involves installing tem Configurations, presents an economic comcable and cable terminations that are rated for parison of an initial versus delayed installation the higher voltage level. These two components of a three-phase feeder. The engineer can perwill operate properly at the lower voltage and form a similar economic comparison for the UD will not have to be changed when the voltage system he or she is designing. level is increased. This simple design change eliminates the need to replace all the underground primary voltage cable—a very expensive DIRECT-BURIED VERSUS and time-consuming task. An economic evaluaPLACEMENT IN CONDUIT tion under the subsection Economic Comparison At some point, most cables need to be replaced of System Configurations (next page) shows that because of a cable failure or external damage. the cooperative will save money by initially inReplacing cable in a conduit system is less exstalling the higher voltage cable. pensive than replacing direct-buried cable and Voltage conversion also requires an increase does not disturb the ground surface. However, in the insulation level of pad-mounted transthe initial installation costs are higher than those formers and sectionalizing devices. To avoid fufor direct-buried cable. To determine which systure changeouts, the cooperative can initially tem is more economical, the engineer needs to

2 0 – Se c t i on 1

1

Economic Comparison of System Configurations

perform an analysis similar to the one described later in the Direct-Buried Versus Cable in Conduit subsection. This evaluation is difficult because it must quantify the expected life of the cable. A conduit system can provide some benefits that are difficult to assign a value to. A conduit system does provide some mechanical protection to the cable and, therefore, could help prolong cable life in areas with rocky soils or areas

congested with other utilities. If a dig-in should occur, however, the conduit system will be more difficult to repair. Conduit systems may also require larger cable sizes to offset de-rating factors as a result of cable heating. Conduit can also protect cable from gophers and prairie dogs; therefore, conduit use in rodent-infested areas will likely prolong cable life.

To design an underground distribution system, the cooperative engineer needs to compare various system configurations. Points of comparison include the following:

costs require use of a carrying charge. The carrying charges are annual payments needed to support construction funds, including loan interest, taxes, and insurance. The examples in this section use a carrying charge of 12 percent. However, when doing comparisons, a carrying charge should be selected appropriate to current economics. Only a few examples consider an inflation rate. The inflation rate used is three percent per year and is not included in the carrying charges. Again, an appropriate value needs to be selected. The installed-material costs used in these examples can vary significantly from region to region. Therefore, the examples should be used as guidelines only. Economic decisions should be based on the cooperative’s own cost data and not on the costs shown in this subsection.

• • • •

Service reliability, Present and future load requirements, System maintenance and operation, and Economics.

Economics is not usually the deciding aspect when comparing different configurations. However, being aware of the different system costs can help the cooperative engineer make economically sound design decisions. The following examples compare several system configurations and show suitable methods for calculating the relative economics of each. Some of these economic evaluations compare only initial costs—the purchase cost of the materials and the installation cost for placing these materials into service. Other evaluations consider initial and future costs—operating, maintenance, and replacement costs. Evaluations that consider future

TABLE 1.3: Additional Materials for an Open-Loop System.

Item

Additional Quantity

Installed Unit Cost

Installed Total Cost

Single-Phase Riser Assembly, 25 kV

1

$ 460.00

$ 460.00

Trench and Backfill

500 ft

3.00

1,500.00

1/0 AWG A1, 25-kV Underground Cable

500 ft

2.50

1,250.00

Elbow Terminator

1

63.00

63.00

Elbow Arrester

1

237.00

237.00

Feed-Through Standoff

1

175.00

175.00

TOTAL

$ 3,685.00

LOOP VERSUS RADIAL As noted earlier in this section, an open-loop system provides better system availability than a comparable radial system does. However, an open-loop system requires additional underground facilities—at a minimum, those listed in Table 1.3. This table also shows the additional costs of these materials. The single-phase riser assembly listed in Table 1.3 includes all materials (conduit, cable terminations, surge arresters, and fused disconnect switches) for terminating underground cable on a riser pole. This assembly does not include the pole. The riser assembly in subsequent tables is defined in the same way. Because an open-loop system always requires more materials than a similar radial system, the initial cost is greater than that of a radial system. In the following examples, this cost difference is calculated for several types of underground systems.


1 Subdivisions Subdivisions usually have a high consumer density. A cable failure here can interrupt power to many consumers. As noted, power can be restored to these consumers much faster on an open-loop system than on a radial system. Fortunately, most subdivision layouts can be easily adapted to the installation of an open-loop system by extending the underground cable from the last transformer to a second riser pole or underground feeder source. To illustrate this, a 37-lot subdivision is shown in Figure 1.17. The approximate cost for a radial system is $37,145. The additional materials for an open-loop system are highlighted and are consistent with those listed in Table 1.3. This increases the project cost by $3,685, an additional cost of approximately $100 per lot. Assuming a carrying charge of 12 percent and an amortization period of 20 years, this $100 investment has a levelized annual cost of $13.40 per lot. To provide a more reliable electric system through a loop design, the cooperative will spend $13.40 a year for each consumer in the subdivision.

This incremental cost for an open-loop system could decrease considerably for a large subdivision. For example, consider a 100-lot subdivision with lot sizes similar to those in Figure 1.17. The cost for installing underground facilities will also be similar, about $1,000 per lot, so the project cost would be $100,000 for a radial system. If an open-loop system can be established with 500 or fewer feet of cable, then the additional cost remains $3,685. However, instead of $100 per lot, the cost is $36.85 per lot, with a levelized annual cost of only $5.20 per consumer. In both of these examples, the cooperative can provide a more reliable system with an additional investment of 10 percent or less. This improvement will increase consumer satisfaction and promotes a better relationship between the cooperative and the consumer. Single Residential Consumer It is usually practical to install an open-loop system for a subdivision. In contrast, an open-loop system to serve a single residential consumer may not be practical. For example, consider a

560'

N.O. 50

560'

AY ' KW 520

AS

DC

OL

50

50

50

460'

460'

50 kVA 50

NEW DOVER ROAD 50

400'

50

400'

50

0'

400'

37.5

Legend Single-Phase, Pad-Mounted Transformer Single-Phase, Primary Voltage, UD Cable Three-Phase Overhead Line N.O.

Normally Open Point

FIGURE 1.17: Open-Loop System, 37-Lot Subdivision.

28

5'

ROW

ROW

OW)

00' R

35 (1

SR 14

ROW

ROW

2 2 – Se c t i on 1

1 single residential consumer served by 500 feet of 1/0 AWG Al primary underground cable. Figure 1.18 shows this radial system and also highlights the materials required for an open-loop system. The radial system costs $4,792. Conversion to an open-loop system requires the same materials listed in Table 1.3, at a cost of $3,685. Here, an open-loop system costs an additional 77 percent

500’

500’

N.O.

Riser Pole

Riser Pole

Legend Single-Phase, Pad-Mounted Transformer Single-Phase, Primary Voltage, UD Cable N.O.

Normally Open Point

FIGURE 1.18: Open-Loop System, Single Residential Consumer.

TABLE 1.4: Sample Spare Cable Cost, Single Residential Consumer.

Item

Additional Quantity

Installed Unit Cost



500 ft

$ 2.50

$ 1,250.00

Elbow Terminator

1

63.00

63.00

Feed-Through Standoff

1

175.00

175.00

Elbow Arrester

1

237.00

237.00

25-kV O.D. Termination

1

66.00

66.00

Cutout

1

73.00

73.00

Riser Pole Arrester

1

89.00

89.00

TOTAL Note. O.D. = outside diameter

$ 1,953.00

of the cost of the radial system, a substantial increase over the 10 percent additional cost for the subdivision. A more economical system for a single customer would be a spare cable placed in the same trench and on the same riser pole. The cost of a spare cable with terminations and arresters is shown in Table 1.4. This reduces the additional cost to $1,953, which is 41 percent of the total project cost. However, this system is less reliable than is an open-loop system with separate trenches. Because the spare cable is in the same ditch as the normal feed cable, both cables are exposed to simultaneous damage during a dig-in. The openloop system in Figure 1.18 has two separate trenches; therefore, a dig-in will usually damage only one cable. Likewise, placing both cables in the same riser exposes both cables to damage whenever the pole is damaged. Instead of serving only one consumer, a single transformer may serve several consumers. Although the cost for an open-loop or spare-cable system will be the same, the cost is divided among more consumers. If the transformer is serving six consumers, then the cost drops to $614 per consumer for an open-loop system, and $326 per consumer for a spare-cable system. For these situations, the cooperative must decide if the benefits of improved reliability make the open-loop or spare-cable system a practical choice. Factors entering into this decision should include the type of customer and the difficulty of effecting repairs in a timely manner. Commercial Consumers Commercial consumers are a very diverse group, ranging from small single-phase consumers to large three-phase consumers. For this reason, there is not a single simple example to show an economic comparison of a loop versus radial system. Instead, the cooperative engineer needs to examine each case to determine the cost of the desired level of reliability. As a guideline for this evaluation, the following example will compare the costs of a radial system, an open-loop system, and a spare-cable system The example in Table 1.5 considers a 500-foot radial feed to a 300-kVA pad-mounted transformer. This system provides a 277/480-Volt four-wire


1 TABLE 1.5: Sample Radial System Cost, Commercial Consumer.

Item

Quantity

Installed Unit Cost


1

$ 1,332.00

$ 1,332.00

500 ft

3.00

1,500.00

1,500 ft

2.50

3,750.00

Three-Phase Riser Assembly, 25 kV Trench and Backfill 1/0 AWG A1, 25-kV Underground Cable 24.9/14.4-kV – 480/277-V, 300-kVA Transformer

1

6,505.00

6,505.00

Elbow Terminator

3

63.00

189.00

Elbow Arrester

3

237.00

711.00

Bushing Inserts

3

57.00

171.00

Transformer Pad

1

364.00

364.00

TOTAL

$ 14,522.00

TABLE 1.6: Additional Cost Per Kilowatt, Open-Loop and Spare Cable Systems. Consumer Type

Open-Loop System

Spare Cable System*

Residential

$3,685/259 kW = $14.22/kW

$1,953/15 kW = $130.02/kW

Commercial

$8,007/225 kW = $35.59/kW

$1,953/225 kW = $8.68/kW

*Spare cable system usually practical only for single transformer installations.

service to a three-phase consumer. Table 1.5 shows the cost of a radial system to be $14,522. An open-loop system requires an additional riser assembly, a second trench, and a separate three-phase run of underground primary cable. These additional materials will cost $8,007, thus increasing the radial system cost by 55 percent. A second option places one spare cable in the same trench as the radial feed. This single spare cable does not provide total redundancy for the three-phase cable, but would be useful if one phase of the circuit faulted. The spare cable has terminations and arresters at each end. The cost of this option is $1,953, as shown in Table 1.4. Instead of 55 percent, this option is only a 13 percent increase over the radial cost. As noted before, this system is less reliable than an open-loop system because the spare cable could be damaged by a fault in an adjacent cable, a dig-in, or damage to the riser pole.

It is often helpful to consider the cost per kilowatt (kW). Doing so also provides a way to compare residential and commercial costs. For example, assume this three-phase installation has a load of 225 kW and the 37-lot subdivision has a diversified load of 7 kW per lot for a total of 259 kW. For purposes of comparison, each single residential consumer is assumed to have a peak (non-diversified) load of 15 kW. Table 1.6 compares the added cost per kilowatt for installing an open-loop system and a spare cable system. Providing an open-loop system for the single commercial consumer costs 2.5 times that for the residential consumer in a small subdivision. However, for the installation of a spare cable, a single residential installation costs nearly 15 times as much per kilowatt. If the open-loop three-phase system serves several commercial consumers, then the additional cost per kilowatt would decrease. An open-loop system that serves three 225-kW deliveries has an additional cost of $11.86/kW instead of $35.59/kW. Therefore, an open-loop system for several three-phase consumers is more practical than is an open-loop system for a single three-phase consumer. Other Options to Consider In addition to an open-loop design and spare cable design, other options may be considered for service to particular consumers or for some cooperatives whose underground installation environment requires other strategies. In some cases, it may be economically prudent to install the primary cable in duct to a single commercial or residential customer to simplify cable replacement in case of failure. This would, of course, depend on the length of the primary cable lateral, the likelihood of future paving over the cable route, and the rock or debris content of the primary cable route excavation. A similar strategy would be to place an empty capped duct alongside the primary cable in the trench, or to install a cable-in-conduit system for selected installations. These options, in addition to the spare cable installation, may be the most economical in the long-term because retrenching a cable route after the site has developed is many times more expensive that the original trenching.

2 4 – Se c t i on 1

1 BRIDG WAY EHAM

Y WA TH

OU

W

NE

RM YA

OL DC

T.

AY KW AS

DC

EA ST

CHARINGTON CT.

ELM

NEW DOVER ROAD

IP

IP

se Pha

ROW

o Tw

Riser Pole

Riser Pole Legend

ROW

Single-Phase Sectionalizing Cabinet

IP

.

OW)

00' R

435 (1

SR 1

ROW

ROW

1/0 AWG, 25-kV, UD Cable

FIGURE 1.19: Single-Phase Sub-Feeder.

TABLE 1.7: Single-Phase Sub-Feeder Cost.

Item

Quantity

Single-Phase Riser Assembly, 25 kV

2

Installed Unit Cost $

460.00

Installed Total Cost $

920.00

Trench and Backfill

1,600 ft

3.00

4,800.00


1,600 ft

2.50

4,000.00


2

2,608.00

5,216.00

Cabinet Pad

2

217.00

434.00

1/0 AWG Terminations

8

66.00

528.00

TOTAL

$ 15,898.00

THREE-PHASE VERSUS SINGLE-PHASE The decision to install three-phase facilities instead of single-phase is usually based on the following: • Three-phase load requirements, • Load balancing, and • Sectionalizing and protection requirements.

Although economics is not the only deciding factor, it is useful to know the cost difference between installing single- and three-phase systems. As an example, consider a 1,600-foot underground sub-feeder of 1/0 AWG A1 25-kV cable. Figure 1.19 shows a single-phase sub-feeder with two single-phase sectionalizing cabinets. The sectionalizing cabinet allows the sub-feeder to feed through and also provides two fused taps. This cabinet costs about $2,600. The total project cost is $15,898, or $9.94 per foot, as shown in Table 1.7. The cost for a similar three-phase sub-feeder increases considerably. Figure 1.20 shows a three-phase sub-feeder with two three-phase sectionalizing cabinets. The three-phase sectionalizing cabinet has three-phase group-operated switches on the incoming and outgoing subfeeder cables. It also has two sets of three-phase fused taps. This cabinet costs $10,000, and the total project cost is $41,752 or $26.10 per foot. Table 1.8 shows these costs. For this example, a three-phase sub-feeder is 2.6 times the cost of a single-phase sub-feeder.


1 BRIDG WAY EHAM

Y WA TH

OU

W

NE

RM YA

OL DC

T.

AY KW AS

DC

EA ST

CHARINGTON CT.

ELM

NEW DOVER ROAD

3P

3P

se Pha

ROW

o Tw

Riser Pole

Riser Pole Legend 3P

.

ROW


OW)

00' R

35 (1

SR 14

ROW

ROW


FIGURE 1.20: Three-Phase Sub-Feeder.

TABLE 1.8: Three-Phase Sub-Feeder Cost.

Item

Additional Quantity

Installed Unit Cost


Three-Phase Riser Assembly, 25 kV

2

$ 1,332.00

$ 2,664.00

Trench and Backfill

1,600 ft

3.00

4,800.00


4,800 ft

2.50

12,000.00

Three-Phase Sectionalizing Cabinet

2

10,216.00

20,432.00

Cabinet Pad

2

400.00

800.00

1/0 AWG Terminations

16

66.00

1,056.00

TOTAL

$ 41,752.00

This comparison rather conclusively demonstrates that the decision to install a three-phase sub-feeder should not be made lightly. However, if future development plans may require the addition of a three-phase feeder along the same route within a few years, the comparison changes dramatically. The delayed installation of the underground three-phase line will require

trenching in established (landscaped) yards. Trenching in an established yard is very costly. The cooperative must remove sod and obstructions (fences, shrubbery, and utility buildings) before trenching. After trenching, the cooperative will need to replace sod or reseed. All this work increases the trenching cost from $3 per foot to $8 per foot. Assume a three-phase feeder is installed five years after the single-phase feeder is installed. The trenching cost is $8 per foot, an increase of $5 per foot over the cost shown in Table 1.8. Therefore, the trench and backfill cost increases by $8,000, and the total future cost to install this three-phase feeder is $49,752. With a carrying charge of 12 percent, this cost has a present worth of $28,229. This cost added to the cost of a single-phase sub-feeder adds up to a present worth of $44,127. These results show that initially installing a three-phase sub-feeder costs less than does delaying the three-phase installation. In addition, the conversion of the feeder will require consumer outages that would have been avoided if the three-phase installation had

2 6 – Se c t i on 1

1 been made initially. Therefore, if future loads may require a three-phase sub-feeder, the cooperative should strongly consider installing it as part of the initial installation.

Where a future voltage conversion is possible, it is wise to install 25-kV instead of 15-kV cable.

FUTURE VOLTAGE CONVERSION Conversion to a higher distribution system voltage requires an increase in the insulation

TABLE 1.9: 25-kV Versus 15-kV Cable and Components. 25-kV Unit Cost

15-kV Unit Cost

Unit Cost Increase

Quantity

Total Cost Increase

1/0 AWG A1 Underground Cable

$ 2.28/ft

$ 1.69/ft

$ 0.59/ft

4,045 ft

$ 2,387.00

Elbow Terminator

$ 36.00

$ 24.00

$ 12.00

18

216.00

Bushing Insert

48.00

25.00

23.00

18

414.00

Riser Terminator

30.00

30.00

0.00

2

0.00

Item

TOTAL

$ 3,017.00

TABLE 1.10: Added Cost of Dual-Voltage Transformers.

Item 50-kVA Transformer 37.5-kVA Transformer

25-kV Unit Cost

15-kV Unit Cost

Unit Cost Increase

Quantity

Total Cost Increase

$ 1,393.00

$ 1,160.00

$ 233.00

8

$ 1,864.00

1,349.00

1,066.00

$ 282.00

1

283.00

TOTAL

$ 2,147.00

TABLE 1.11: Voltage Conversion Cost at Year 10.

Item

Labor Cost

50-kVA, 7.2-kV Transformer

$ 94.00

Unit Cost

Unit Salvage

Quantity Installed Removed

Total Cost Increase

$

0.00

$ (580.00)

—

8

$ (3,888.00)

37.5-kVA, 7.2-kV Transformer

94.00

0.00

(533.00)

—

1

(439.00)

50-kVA, 14.4-kV Transformer

94.00

1,574.00

—

8

—

13,344.00

37.5-kVA, 14.4-kV Transformer

94.00

1,525.00

—

1

—

1,619.00

TOTAL

$ 10,636.00

level of system components. For an underground distribution system, these components are the following: • • • • •

Underground primary cable, Cable terminators, Pad-mounted transformers, Sectionalizing equipment, Transformer bushing well inserts, and • Surge arresters. The changeout of these components at the time of voltage conversion is very expensive and requires either a long outage or a series of shorter outages. This is particularly true of cable replacement in established subdivisions. Recent surveys show that the labor for cable replacement often costs $8 per foot or more. In an attempt to avoid the excessive cost of cable replacement, 25-kV cable and terminations could be installed initially. Doing so does increase the initial material cost over that for 15-kV cable and components. However, the labor cost remains the same. In light of the relatively low incremental cost for higher voltage cables and accessories, it is generally advisable to install a cable suitable for any distribution voltage expected for the area. For an economic analysis, consider the 37-lot subdivision of Figure 1.17. Table 1.9 shows the increase in material cost to be $3,017, or $0.75 per foot. Determining the future value of this additional investment requires use of a compound amount factor.


1 former changeout cost of $10,636 has a present worth of $3,425. The present worth Quantity Unit Total Cost of installing dual-voltage Item Labor Cost Unit Cost Salvage Installed Removed Increase transformers is $2,147, which 50-kVA, 7.2-kV $ 115.00 $ 0.00 $ (348.00) — 8 $ (1,864.00) makes it the economical Transformer choice. 37.5-kVA, 7.2-kV 115.00 0.00 (320.00) — 1 (205.00) Table 1.12 shows a similar Transformer analysis for a voltage conversion at 20 years instead of 50-kVA, 14.4-kV 115.00 1,938.00 — 8 — 16,424.00 10 years. The assumed inflaTransformer tion rate is three percent per 37.5-kVA, 14.4-kV 115.00 1,877.00 — 1 — 1,992.00 year and the salvage value on Transformer the removed transformers is TOTAL $ 16,347.00 30 percent. For a carrying charge of 12 percent, the present worth facIf one assumes a voltage conversion in 10 tor at 20 years is 0.1037. The transformer changeyears and a carrying charge of 12 percent, the out cost of $16,347 has a present worth of $1,695. compound amount factor is 3.1058. The initial On the basis of this analysis, it is more economiinvestment of $3,017 has a future value of cal to change out the transformers in the future $3,017 × 3.1058 = $9,370, or $2.32 per foot. This rather than install dual-voltage transformers. amount is approximately equal to the present These economic analyses show that it is imcost of 1/0 AWG Al, 25-kV underground cable. It portant to plan for future voltage conversions. If is very unlikely that this amount will cover even a voltage conversion is planned within 10 years the purchase cost of the cable in 10 years. For a of the initial installation, then the cooperative voltage conversion in 20 years, the initial investshould install 25-kV cable, 25-kV components, and ment of $3,017 has a future value of $3,017 × dual-voltage transformers. For conversions occur9.6463 = $29,103, or $7.19 per foot. This amount ring after 10 years, the cooperative should install is less than the present labor cost ($8) for cable 25-kV cable and components. To see if dual-voltreplacement. Therefore, in areas where a future age transformers are economically feasible, the voltage conversion is possible, installing 25-kV cooperative engineer will need to do an analysis cable instead of 15-kV cable is a wise investment. similar to that shown in Tables 1.10 and 1.12. Another option to consider is installing dualvoltage transformers along with the 25-kV cable FRONT VERSUS REAR PROPERTY and components. The dual-voltage transformers are As covered earlier in this section, consumers and more costly than the 7.2-kV transformers. Howthe utility often disagree about placement of elecever, the labor cost to install either transformer tric facilities. The utility often prefers to place fais the same. For the 37-lot subdivision, the total cilities along the front property line where they material cost increase for installing dual-voltage are easier to maintain and operate, thus providtransformers is $2,147, as shown in Table 1.10. ing better reliability. In contrast, consumers often Again, consider a voltage conversion 10 years prefer placing facilities along the rear property after the initial installation. Table 1.11 shows the line. This conflict will rarely, if ever, be solved cost at the time of conversion assuming an inflaby an economic analysis. However, cost is altion rate of three percent per year and a 50 perways an aspect to consider. The following examcent salvage on the removed transformers. ples show a method to compare the cost of Determining the present worth of this total refront-lot versus rear-lot placement of facilities. quires a present worth factor. For a carrying The economics of front versus rear placecharge of 12 percent and a conversion at 10 ment will vary significantly depending on the years, this factor is 0.322. Therefore, the transsubdivision lot layout. The 37-lot subdivision of TABLE 1.12: Voltage Conversion Cost at Year 20.

2 8 – Se c t i on 1

1 EHAM

605

BRIDG

’

AY HW UT

MO

WAY

W

NE

R YA

50

50

OL DC

'

400

T.

DC

EA ST

640

'

ELM

400'

605

50

'

5 37. CHARINGTON CT.

AY KW AS

5 37.

420' 50 NEW DOVER ROAD

ROW

Legend Single-Phase, Pad-Mounted Transformer

.

OW)

00' R

35 (1

SR 14

ROW


ROW

ROW

600-V Service Cable

FIGURE 1.21: Front Property Placement.

BRIDG EHAM

AY HW UT

MO

WAY

W

NE

R YA

OL AY KW

AS

DC

' 300 75 75 240 37.5

CHARINGTON CT.

'

160

180'

400'

'

50

T.

DC

A TE

S ELM

50 NEW DOVER ROAD

ROW


.

1/0 AWG, 25-kV, UD Cable 600-V Service Cable

FIGURE 1.22: Back Property Placement.

ROW

OW)

00' R

35 (1

SR 14

ROW

ROW


1 Figure 1.17 shows front-lot placement. For this particular subdivision, placement along the rear lot lines will actually require more cable and increase the total project cost. For this type of subdivision, front-line placement is a practical choice. In contrast, Figures 1.21 and 1.22 show a subdivision where lots share back property lines. With this type of lot arrangement, placement along the rear property lines requires less cable and fewer transformers. In this particular example, placement along the front property line requires 1,886 additional feet of cable and one additional padmounted transformer. These extra materials increase the project cost by $11,496, or 75 percent. Most subdivisions will be a combination of the two extremes shown in Figures 1.17 and 1.22. Because subdivision layouts differ, an accurate comparison of costs requires a case-by-case study. Calculating the installed project cost is straightforward; however, it is difficult to calculate the cost advantage of operating and maintaining facilities along the front-lot lines. As a result, it is impossible to set a dollar amount on the reliability and operational convenience gained by placing facilities along the front-lot lines. DIRECT-BURIED VERSUS CABLE IN CONDUIT Many utilities are now replacing underground cable that was installed only 15 to 20 years ago. Much of this cable is direct buried. To replace it will require opening a new trench or tunneling with long-distance boring equipment. Both of these methods are expensive: • Trench and backfill labor costs are about $8 per foot. • Long-distance boring costs are about $9 to $10 per foot. These costs will vary significantly depending on soil conditions, other utility congestion, landscape, and homeowner obstacles. One way to reduce these cable replacement costs is to install cable in a conduit system. When the cable is in a conduit system, the replacement cost is the cost of pulling out the failed cable and pulling in the new cable plus the cost of the new cable. The soil does not have to be disturbed and other utilities do not have to be located and avoided. Cost savings are tremendous. The following example compares the cost of direct-buried, PVC rigid conduit with high-density

polyethylene (HDPE) flexible cable in conduit. Because the price of conduit and cable fluctuates, it is important that the cooperative engineer perform an economic analysis based on this example but using current costs. In addition, the cost for replacement of direct-buried cable will vary greatly. If $8 per foot is not reasonable, the engineer needs to insert an appropriate cost. This example uses the 37-lot subdivision of Figure 1.17. Tables 1.13, 1.14, and 1.15 show the present, 25-year replacement, and 30-year replacement costs for the three options. This long-term analysis includes an inflation rate of three percent per year. Therefore, the cost to replace direct-buried cable will be as follows:

$14.00 per foot ($8.00 per foot × 1.75) at 25 years $15.20 per foot ($8.00 per foot × 1.90) at 30 years

These costs are shown in Table 1.13 as Trench, Backfill, Restore Surface. A present worth factor needs to be used to compare these three options. For a carrying charge of 12 percent, the single payment present worth factor from standard tables for 25 years is 0.0588 and for 30 years is 0.0334. For example, for HDPE flexible conduit, the 25- and 30-year replacement costs are the following:

$24,675 + .0588($20,185) = $25,862 $24,675 + .0334($21,924) = $25,407

Table 1.16 summarizes the present worth for each option. For cable replacement at 25 years, a flexible conduit system has the lowest present worth and is the most economical choice. For cable replacement at 30 years, a direct-buried system is the most economical. However, a small change in the cost for cable replacement can affect the economic choice. For example, if the cost to replace direct-buried cable is $10 per foot instead of $8 per foot, then the 25-year cost is $10(1.75) = $17.50 per foot, and the 30-year cost is $10(1.90) = $19.00 per foot. Thus, the total values in Table 1.13 change to $88,505 at 25 years, and $96,069 at 30 years.

3 0 – Se c t i on 1

1 TABLE 1.13: Option 1—Direct-Buried Cable.

Present Cost

Item

Quantity

Unit Installed Cost

Total Installed Cost

Trench and Backfill

4,045 ft

$ 3.00/ft

$ 12,135.00


4,045 ft

2.50/ft

10,113.00

TOTAL 25-Year Replacement

$ 22,248.00

Trench, Backfill, Restore Surface

4,045 ft

$ 14.00/ft

$ 56,630.00


4,045 ft

4.38/ft

17,717.00


$ 74,347.00

Trench, Backfill, Restore Surface

4,045 ft

$ 15.20/ft

$ 61,484.00


4,045 ft

4.75/ft

19,214.00

TOTAL

$ 80,698.00

TABLE 1.14: Option 2—PVC Rigid Conduit.

Present Cost

Item

Quantity

Unit Installed Cost


Trench and Backfill

4,045 ft

$ 3.00/ft

$ 12,135.00

2-Inch Conduit

4,045 ft

1.55/ft

6,270.00


4,045 ft

2.60/ft

$ 10,517.00


$ 28,922.00

Remove Cable From Duct

4,045 ft

$ 0.44/ft

$ 1,780.00


4,045 ft

4.55/ft

18,405.00


$ 20,185.00


4,045 ft

$ 0.48/ft

$ 1,942.00


4,045 ft

4.94/ft

19,982.00

TOTAL

For the 30-year replacement, these values result in a present worth of $22,248 + .0334($96,069) = $25,457. Therefore, the flexible conduit is the economical choice for replacement at 30 years. For this reason, it is important for the engineer to select an appropriate cable replacement cost for the economic analysis. Of course, this economic analysis could not assign a monetary value to the following: • Consumer inconvenience and irritation that results from trenching across established lawns, and

$ 21,924.00

• Added cable protection provided by a conduit system. Another consideration for this analysis is the type of native soil. If the soil is rocky, it is not suitable for backfill of a direct-buried cable. In this case, select fill material must be used for a two-inch minimum of cable bedding and a fourinch cable cover. The cost of this select fill material can substantially increase the initial project cost for a direct-buried system. In contrast, the use of a conduit system, flexible or rigid, protects the cable from rocky soils; in most cases, select


1 TABLE 1.15: Option 3—Cable in HDPE Flexible Conduit.

Present Cost

Item

Quantity

Unit Installed Cost


Trench and Backfill

4,045 ft

$ 3.00/ft

$ 12,135.00

Cable in Conduit

4,045 ft

3.10/ft

12,540.00


$ 24,675.00


4,045 ft

$ 0.44/ft

$ 1,780.00


4,045 ft

4.55/ft

18,405.00


$ 20,185.00


4,045 ft

$ 0.48/ft

$ 1,942.00


4,045 ft

4.94/ft

19,982.00

TOTAL

$ 21,924.00

get accurate results, each cooperative will need to conduct a similar analysis using its cost data.

TABLE 1.16: Present Worth of Cable Installation Options. Present Worth Installation Method

25-Year Replacement

30-Year Replacement

$ 26,620.00

$ 24,943.00

PVC Rigid Conduit

30,109.00

29,654.00

HDPE Flexible Conduit

25,862.00

25,407.00

Direct Buried

backfill is not required. Therefore, the initial project cost for the two conduit systems will not increase. Adverse soil conditions can quickly shift system economics to favor conduit installations. Although this analysis is based on a small 37-lot subdivision, the results show that a conduit system can be an economical choice. The prices for conduit, trenching, surface restoration, and longdistance boring vary from region to region. To

SEPARATE SERVICES VERSUS SECONDARY PEDESTALS In a residential subdivision, a single pad-mounted transformer often provides electrical service to several consumers. Service may be provided directly from the transformer or from a secondary pedestal. Figure 1.23 shows both methods. The arrangement that uses a secondary pedestal is less reliable than direct service from the transformer. A cable fault on the secondary cable will interrupt power to multiple consumers. In contrast, a cable fault on an individual service will interrupt power to that consumer only. This analysis compares the initial installation cost only. Table 1.17 lists the cost of providing separate services as shown in method A of

Secondary Pedestal #6

200’

10’

#6

4/0

Method A—Seperate Services

FIGURE 1.23: Methods for Providing Secondary Service.

4/0

’

4/0

50

4/0 Transformer

50

250’

150’ Transformer

’

150’

4/0

Method B—Secondary Pedestal

3 2 – Se c t i on 1

1 TABLE 1.17: Separate Service Cables. Quantity (ft)

Installed Unit Cost


Trench and Backfill

300

$ 3.00/ft

$ 900.00

4/0, 600-V Triplexed Cable

400

1.25/ft

500.00

No. 6, 600-V Triplexed Cable

200

0.25/ft

50.00

Item

TOTAL

$ 1,450.00

TABLE 1.18: Secondary Pedestal. Quantity (ft)

Installed Unit Cost


Trench and Backfill

300

$ 3.00/ft

$ 900.00

4/0, 600-V Triplexed Cable

300

1.25/ft

375.00

10

0.25/ft

2.50

Secondary Pedestal

1

172.00

172.00

Insulated Connectors

3

14.00

42.00

Item

No. 6, 600-V Triplexed Cable

TOTAL

UD Loss Economics

Figure 1.23. The separate services do share a common trench along the front property line. Method B of Figure 1.23 shows the use of a secondary pedestal. This cost is shown in Table 1.18. As this example shows, the use of separate service cables is often the economical choice for lots located on the same side of the road as the transformer if the lots are developed at the same time. However, the use of a secondary pedestal across the road from the transformer may be the economical choice since it requires trenching or tunneling across the road in only one location.

$ 1,491.50

The inevitable loss of some of the power delivCOST OF LOSSES ered through underground cables is an expense In a sample analysis of the cost of losses on for the cooperative. Optimal economic design of distribution primary lines, the Distribution the system requires that this expense be known System Loss Management Manual provides cost and evaluated. figures for a typical cooperative. The sample Cable losses are classified as either load-decooperative purchases wholesale power at $10 pendent or non-load-dependent. For UD cables, per kW per month at a 100 percent ratchet, and most of the loss is load-dependent; it is only in the wholesale energy rate is $0.03 per kilowattunusual circumstances that non-load-dependent hour (kWh). loss becomes significant. The cost of losses is Non-load-dependent losses are constant as derived from a combination of peak-load long as the cable is energized. Load-dependent demand costs and accumulated losses change with the square annual energy costs. A thorof the loading level, which ough coverage of the types of makes it difficult to determine For UD cables, losses and their costs to cooptheir average level. A quantity eratives is contained in the referred to as a loss factor is most power loss is National Rural Electric used to estimate the average load dependent. Cooperative Association’s of load-dependent losses when Distribution System Loss their peak value is known. Management Manual A value of 0.3 (30 percent) (NRECA Research Project 90-7). is suggested as typical for


1 primary distribution lines when calculating loss factors. The cost per peak kilowatt for line losses for the sample cooperative is then determined as follows:

Annual Demand Cost per kW of Peak Losses $10/kW/month × 12 months = $120/kW Annual Energy Cost per kW of Non-Load-Dependent Peak Losses 8,760 hours × $0.03/kWh = $263/kW

losses are calculated by the formula shown in Equation 1.1. Primary Cable Sheath Losses The normal UD practice is to ground cable sheaths at both ends. When this is done on three-phase cable runs, a small amount of circulating current will be induced in the cable sheaths. The flow of this current produces a small loss in the sheaths, calculated as shown in Equation 1.2. XM is determined using Equation 1.3.

Equation 1.1 Annual Energy Cost per kW of Load-Dependent Peak Losses 0.3 × 860 hours × $0.03/kWh = $79/kW Total Annual Cost per kW of Non-Load-Dependent Peak Losses $120/kW + $263/kW = $383/kW Total Annual Cost per kW of Load-Dependent Peak Losses $120/kW + $79/kW = $199/kW

WR=3 I2 R L where: WR = Total loss, in watts I = Load current, in amperes R = Phase conductor resistance, in ohms per kilofoot (kft) L = Circuit length, in kft

Equation 1.2 The resulting expense per kilowatt of loss can be used to quickly estimate the savings that will result from using UD designs that operate at lower losses. The loss savings can be compared with the annual carrying charges on the extra investment costs required to achieve lower losses. This type of economic comparison is discussed in detail in the Distribution System Loss Management Manual. CABLE SYSTEM LOSSES An essential step in the economic evaluation of losses is calculating the expected electrical losses for alternative designs. For a primary UD cable, losses occur in the conductor, sheath, and dielectric, and as a result of cable charging current. Primary Cable Conductor Losses The losses resulting from load current interacting with the conductor resistance (I2R losses) are by far the most significant losses in primary UD cables. For a run of three-phase cable, these

WS = where: WS = I = RS = L = XM =

3 I2 RS L X2M R2S + X2M

Total sheath loss, in watts Load current, in amperes Sheath resistance, in ohms per kft Circuit length, in kft Sheath reactance, in ohms per kft

Equation 1.3 XM = 0.05292 log10

S rM

where: XM = Sheath reactance, in ohms per kft S = Center-to-center spacing, in mils, for equilaterally spaced cables rM = Mean radius, in mils, to the sheath for each cable

3 4 – Se c t i on 1

1 Equation 1.4

Equation 1.5 WD =

8.28 E2 L εt cosφ log10 2T+D D

= Total three-phase dielectric loss, in watts E = Line-to-ground operating voltage, in kV L = Circuit length, in kft = Dielectric constant of the insulation εt cosφ = Insulation power factor, per unit T = Insulation thickness, in mils D = Conductor diameter, in mils

where: WD

Primary Cable Dielectric Losses Voltage stress on cable insulation produces a slight heating effect that leads to power losses. These dielectric losses can be calculated using Equation 1.4. The formula in Equation 1.4 shows that dielectric losses are directly proportional to the product of εt and cosφ. Cable engineers refer to the product εt cosφ as the cable loss factor. This use of the term loss factor is completely different from the use of loss factor earlier in this section. Dielectric losses are a consequence of the cable being energized and are, therefore, continuous; whereas the more common use of the term loss factor deals with losses due to the resistance of the conductor and, therefore, vary with the magnitude of the load being carried by the cable. Primary Cable Charging-Current Losses The capacitance of an underground cable draws charging current that interacts with the conductor resistance to produce a small loss. If the cable is delivering current to low power factor load, the charging current will be beneficial because its leading nature will cancel out some of the lagging load current. Therefore, chargingcurrent losses are of concern for only unloaded cables or those carrying unity power factor loads. The procedure for calculating charging-current losses begins with determining the cable capacitance per phase with Equation 1.5.

C=

7.354 εt log10 2T+D D

where: C = Cable capacitance, in nanoFarads (nF) per kft εt = Dielectric constant of the insulation T = Insulation thickness, in mils D = Conductor diameter, in mils Next, charging current per kilofoot of cable length is calculated with Equation 1.6. Equation 1.6 IC = 0.000377 C E where: IC = Charging current, in amperes per kft C = Cable capacitance, in nF per kft E = Line-to-ground operating voltage, in kV Finally, the charging-current loss is calculated as shown in Equation 1.7. Equation 1.7 WC = R I2C L3 where: WC = Total three-phase charging current loss, in watts R = Phase conductor resistance, in ohms per kft IC = Charging current, in amperes per kft L = Circuit length, in kft Data for Cable Loss Calculations Many items of technical data are needed on the cables involved to calculate losses from the above formulas. Physical measurements such as diameter and insulation thickness are usually shown on manufacturers’ catalog sheets. Basic electrical data such as voltage, amperes, and resistance are known from the system or can


1 be sure that the correct values are known, it is usually necessary to contact engineering specialists on the staff of the manufacturer of each specific cable type. There are often large differences in values for dielectric constant and power factor among various cable types. The spread in values is especially pronounced for the power factor. In addition, the cable power factor often varies substantially with cable temperature. It is recommended that, if comparisons are being

easily be found from catalog sheets or standard references. The insulation dielectric constant, εt, and power factor, cosφ, are sometimes difficult to determine. Manufacturers’ data sheets often do not give these parameters. For pure materials such as TR-XLPE, the information may be obtained from standard references. However, most modern insulation types contain additives that affect dielectric constant and power factor. To

EXAMPLE 1.1: Cable Loss Calculations. This example contains typical data; however, don’t use the sample data in actual-case calculations. For actual situations, consult the cable manufacturer to get accurate data on the cable being used. Table 1.19 shows data and loss calculation results for a typical three-phase cable run. Three insulation types are represented at two different temperatures. TABLE 1.19. Sample Cable Loss Analysis. @ 25° C Insulation Type Sample Data

TR-XLPE

Low-Loss EPR

@ 50° C High-Loss EPR

TR-XLPE

Low-Loss EPR

High-Loss EPR

(E) Line-to-Ground Operating Voltage in kV

7.2

7.2

7.2

7.2

7.2

7.2

Conductor Size

1/0 A1

1/0 A1

1/0 A1

1/0 A1

1/0 A1

1/0 A1

(D) Diameter in mils

373

373

373

373

373

373

(T) Thickness in mils

220

220

220

220

220

220

(rM) Mean Radius in mils

430

430

430

430

430

430

(S) Center-to-Center Spacing in mils

1,180

1,180

1,180

1,180

1,180

1,180

(R) Resistance in Ω/kft

0.20

0.20

0.20

0.20

0.20

0.20

(RS) Sheath Resistance in Ω/kft

0.60

0.60

0.60

0.60

0.60

0.60

(εt) Dielectric Constant of the Insulation

2.35

2.9

3.27

2.35

2.9

3.27

(cosφ pu) Insulation Power Factor per Unit

0.06

0.25

2.0

0.06

0.30

3.25

(L) Circuit Length in kft

4.0

4.0

4.0

4.0

4.0

4.0

(I) Load Current in Amperes

60

60

60

60

60

60

Conductor Loss, Watts

8,640

8,640

8,640

8,640

8,640

8,640

Concentric Neutral Loss, Watts

38.7

38.7

38.7

38.7

38.7

38.7

Dielectric Loss, Watts

715

3,679

33,184

715

4,414

53,924

Charging Loss, Watts

0.3

0.4

0.5

0.3

0.4

0.5

9,394

12,358

41,863

9,394

13,093

62,603

TOTAL LOSS, Watts *Insulation Data Courtesy of the Okonite Company

3 6 – Se c t i on 1

1 EXAMPLE 1.2: Calculating Losses on Secondary Cables. This example illustrates how the losses on secondary cables are calculated. Sample data are shown in Table 1.20. TABLE 1.20: Sample Secondary Cable Data. Voltage of Circuit Circuit Length

120/240-V, single-phase

loaded. In this case, a resistance of 0.167 ohms per kft is given by reference tables. Load on the neutral is assumed to be negligible. Therefore, the conductor distance is 300 feet, and the total resistance is 0.05 ohms. Losses at peak load are calculated as follows:

150 feet

Conductor

No. 1/0 AWG, aluminum

Peak Load

85 amperes

Loss Factor

20%

The conductor resistance is obtained from standard references. A conductor temperature of 25°C is assumed for underground secondary cables that are not heavily

made among cable types, the engineer should use only written data obtained from the manufacturer of that cable type. An excellent source of this data is the cable manufacturer’s Insulated Cable Engineers Association (ICEA) Qualification Report for the particular cable construction. Once the figures are obtained, compare the data from different sources to confirm the reasonableness of the information for a particular cable type. When requesting data from cable manufacturers, be as specific as possible about the data being requested. Ask the manufacturer for the data from ICEA qualification tests. Losses should be quoted for a specific temperature, such as 40°C. The loss figures in Table 1.19 show that sheath, dielectric, and charging-current losses are negligible compared with conductor load-current losses, except in the case of high-loss EPR. However, under light-load or other unusual conditions, the relative values of the three types of losses may become more significant. Charging-current losses, for example, may become significant for extremely long cable runs because these losses increase with the cube of the circuit length. Another important consideration is that small loss differences among alternative cable types can accumulate to a significant expense if an extremely large amount of cable is placed in ser-

WR = I2 R = 852 × 0.05 = 361 watts Annual energy losses are determined by using the loss factor: Energy Losses = 0.2 × 8,760 hours × 361 watts = 632,472 watt-hours = 632 kWh

vice. The dielectric loss differential between normal EPR cable and TR-XLPE cable is approximately 0.22 kW per circuit mile from the results shown on the table. Because this loss is nonload-dependent, the annual loss expense per mile as calculated above is typically $84 per mile (0.22 kW/mile × $383/kW). For 100 circuit miles of installed cable, this expense comes to $8,400 per year, which no longer seems insignificant. However, in a total economic evaluation, the cost of additional dielectric losses ($84 per mile) must be compared with any additional life expectancy that might be available from the higher loss insulation system. Appendix D of NRECA CRN Project 90-8 provides a method for evaluating cable losses and life expectancy in the purchasing process. Secondary Cable Losses For secondary UD cables, losses other than load-current-related conductor I2R losses are truly insignificant. Loss control methods for application to secondary designs are the same as described in the NRECA Distribution System Loss Management Manual for either overhead or underground situations. Appendix B to that manual gives annual kilowatt-hour losses for a selection of conductor sizes and loading levels.


1 approximately 20 percent less than this example PAD-MOUNTED TRANSFORMER LOSSES are available from manufacturers. Use of the The losses on pad-mounted transformers used higher efficiency transformer will save about $40 on UD systems are a significant expense. Close annually, which is enough to attention to the management amortize about $300 in initial of losses on any type of transinvestment cost at a 12 percent former is essential to a loss Pad-mounted carrying charge rate over a 20control program. transformer losses are year period. Thus, if the higher As with all types of transefficiency transformer can be formers, losses on pada significant expense. purchased for less than a $300 mounted transformers are of price premium over the samtwo distinct types. The first ple transformer, then it is a category, core losses, is not better economic choice in the long run. load dependent and represents a continuous exThe Distribution System Loss Management pense whenever the transformer is energized. Manual provides thorough coverage of the issue The second category, winding losses, comprises of transformer losses and the means to control load-dependent losses that become especially the associated expenses to the extent feasible. expensive during peak loads. Higher efficiency transformers with losses DEFERMENT OF TRANSFORMER ENERGIZATION New housing developments often require the construction of the electric UD system well beEXAMPLE 1.3: Typical Costs Associated with Transformer Losses. fore most living units are built and occupied. When energized transformers are installed beConsider a 50-kVA pad-mounted transformer having 140 watts of core losses and fore there are consumers to serve, the non-load490 watts of winding losses at nameplate load. If this unit is loaded to 60 kVA at peak dependent or no-load losses on the transformers load, the winding losses will be as follows: represent an expense that is uncompensated by revenue. This expense can be avoided by keepWinding Losses = (60 ÷ 50)2 × 490 watts = 706 watts ing the transformers de-energized until they are needed. Service to street lights can be concenWith the annual cost figures given for losses at the beginning of this subsection, the trated in a small number of transformers to allow annual costs associated with each type of loss can be calculated as follows: the de-energization of most of the units in areas not yet occupied. Core Loss Cost = $383/kW × 0.140 kW = $54 Installing a de-energized transformer requires Winding Loss Cost = $199/kW × 0.706 kW = $140 the use of a feed-through stand-off bushing which, in most cases, costs about $150. Because The total annual cost of the losses associated with operating this transformer this bushing can be reused elsewhere after the is $194. transformer is placed in service, the special bushing cost is equivalent to $20 annually at a 12 percent carrying charge rate over a 20-year period. Despite this expense, the avoidance of TABLE 1.21: Savings from Deferred Transformer Energization. core losses represents a net savings, as shown by Table 1.21. Annual Loss Feed-Through Annual If hundreds of units are involved, the savings Size Core Losses Cost at Device Net associated with deferred energization could ex(kVA) (watts) $383/kW Annual Cost Savings ceed $8,000 annually. 25 82 $ 31.00 $ 20.00 $ 11.00 For 50- and 100-kVA installations, larger savings can be achieved by deferring the installation 50 140 54.00 20.00 34.00 of each transformer not needed for immediate 100 260 100.00 20.00 80.00 service by placing a pedestal containing a feed-

3 8 – Se c t i on 1

1 TABLE 1.22: Savings From Deferred Transformer Installation. 25 kVA

50 kVA

100 kVA

Transformer Price

$ 750.00

$ 1,000.00

$ 1,750.00

Deferred Transformer Carrying Charges at 12% (Transformer Price x 0.12)

$ 90.00

$ 120.00

$ 210.00

31.00

54.00

100.00

Total Deferred Cost

121.00

174.00

310.00

Temporary Equipment Annual Cost

120.00

120.00

120.00

1.00

54.00

190.00

Deferred Annual Core Loss Cost (from Table 1.21)

Net Annual Savings

through device at the future transformer location. The cost of the pedestal and device is about $330, which represents a $40 annual cost at a 12 percent carrying charge rate (0.12 × 330 = $40). A nonrecoverable labor cost of about $160 is incurred for installing the temporary feed-through pedestal and removing it later. If the average deferment time is two years, this cost is $80 annually. Therefore, the cost for exercising this deferment option is $120 annually ($40 + $80). On the plus side, the annual carrying charges on a transformer are avoided along with the cost of core losses. The overall results are summarized in Table 1.22 for three common transformer sizes.

Steps for Layout of a UD System

To help the engineer with layout of a UD system, this subsection describes eight design steps: STEP 1: Get the required information. STEP 2: Arrange the service and transformer

These results show that deferred installation of transformers is not significantly beneficial for 25-kVA units. However, the net savings can be substantial in the case of larger units. If hundreds of units are involved, the savings may exceed $25,000 annually. Simply routing the cable aboveground at future transformer locations and looping it back into the trench without cutting it can achieve still larger savings. An enclosure is then installed to protect the above-ground loop. When the time comes for a transformer to be installed, the cable is de-energized and cut to prepare for the installation of the elbows and transformer. However, special care must be used to avoid excessive cable bending with this type of installation, and the extra switching that may be required during the final transformer installation does represent an additional expense. CONCLUSION Electrical losses on UD systems represent an expense that should be managed to reduce costs. When alternative UD system designs are considered, it is necessary to estimate the amount of these losses and their costs. The techniques given here and in the NRECA Distribution System Loss Management Manual provide the necessary calculation methods.

STEP 1. GET THE REQUIRED INFORMATION Before any design work can be started, the engineer must get certain information from the consumer or developer, including the following:

layout. STEP 3: Calculate the consumer load and select

proper equipment ratings. STEP 4: Select the primary cable route. STEP 5: Locate sectionalizing equipment. STEP 6: Visit the project site. STEP 7: Obtain all easements. STEP 8: Prepare staking sheets.

• Site plan with defined lots and utility easements, • Load and voltage requirements, • Project schedules, • Location of other underground utilities, • Reliability needs, and • Final grading plans.


1 barrier wall or an oil absorption For subdivisions, it is very bed around the transformer. important to get a copy of the For subdivisions, In contrast, the typical resisubdivision plat. This map get a copy of the dential load does not require shows the lot arrangements the transformer to be next to and is necessary for designing site plan and the house. Rather, the transthe layout of underground farecorded plat. former can be in a central locilities. Appendix D contains a cation and provide service to form to use when collecting several consumers. this information. The engineer can begin to arrange this service This information is rarely gathered in one and transformer layout after receiving the subdibrief conversation. Rather, it is usually compiled vision plat. Several studies have shown that the through several conversations and meetings with most economical arrangement uses the least consumers, the developer, contractors, and other number of transformers. This design, in turn, utility representatives. It is the engineer’s duty to means longer service conductor lengths and persevere until all required data are collected in more consumers per transformer. However, final form. Although the engineer can plan many voltage flicker at the consumer’s delivery point aspects of the project on the basis of preliminary often limits the service conductor length. Deinformation, a final design should not be released pending on lot size, limiting the service conducuntil all information is collected and verified. tor length may reduce the number of consumers Otherwise, the project may encounter unnecesper transformer. Another limiting factor is the sary construction difficulties, fail to meet conspace in the secondary compartment of the sumer expectations, or use materials inefficiently. transformer. Most single-phase pad-mounted In any of these cases, the cost to the cooperative transformers have space for connecting a maxiand its consumers will be greater than for a mum of eight secondary/service conductors. well-designed system. This includes secondary conductors used to feed street and area lights. STEP 2. ARRANGE THE SERVICE In some layouts, a transAND TRANSFORMER LAYOUT former may serve some lots loCommercial and industrial cated across the street. A consumers usually have heavy Service conductor convenient way to serve sevloads that can include largelength is often eral lots with only one road horsepower motors. To limit crossing is from a secondary the voltage drop and flicker limited by pedestal. The secondary associated with these loads, voltage flicker. pedestal is supplied by a sinthe transformer should be near gle secondary cable from the those consumers’ delivery transformer (see Figure 1.24). points. Often the transformer Unfortunately, this decreases service reliability. is placed near the building. The engineer should A cable fault on the one secondary cable interuse good judgment and experience in determinrupts power to all the attached consumers. ing the minimum allowable distance between However, the time required to replace the failed the transformer and the building. Factors affectcable will be shorter if the cable is in a conduit. ing the distance will include building use, fire It is also advisable to have cable in conduit for rating of the exposed wall, presence of wall any roadway crossing to eliminate future street openings, vehicle traffic, and other public safety cutting and provide additional protection against considerations. If the utility or the building dig-ins. owner concludes that additional protection is Secondary pedestals are not the ideal method warranted, such enhancements might be achieved for serving consumers on the same side of the by increased separation, use of “less flammable” road as the transformer. Each of these consumers fluid in the transformer, or installation of a

4 0 – Se c t i on 1

1 should have a separate service cable from the transformer. This improves reliability, is often more economical than installing a secondary pedestal (see Economic Comparison of System Configurations earlier in this section) and also eliminates maintenance of the secondary pedestal. Figure 1.25 shows a service and transformer layout for a 75-lot subdivision. This layout features 13 transformers that serve an average of six consumers each. Transformers located along the front property lines serve the perimeter lots. The interior lots share back property lines; therefore, it is more economical to serve these lots from transformers located along the rear property lines. This combination of front and rear property line placement is often the most economical layout. Because of criteria other than economics, the cooperative may allow transformer placement along the front property line only or the rear property line only. Table 1.2 lists these other criteria and compares the advantages and disadvantages of front versus rear line placement.

Secondary Pedestal

4/0

AY MW GEHA BRID

Transformer

T.

DC

A TE MS

EL

FIGURE 1.24: Road Crossing to Feed Secondary Pedestal.

AY MW

GEHA

W

NE

BRID

AY HW UT

MO

R YA

OL

Note: The three shaded lots indicate the worst locations for voltage drop and flicker.

T.

DC

C AD

AS

E ST

CHARINGTON CT.

AY KW

ELM

NEW DOVER ROAD

ROW

Legend Single-Phase, Pad-Mounted Transformer Secondary-Voltage Cable Streetlight

FIGURE 1.25: Service and Transformer Layout for 75-Lot Subdivision.

ROW

OW)

00' R

35 (1

SR 14

ROW

ROW


1 STEP 3. CALCULATE THE CONSUMER LOAD AND SELECT PROPER EQUIPMENT RATINGS From the information collected in Step 1 and the service and transformer layout of Step 2, the engineer can calculate the expected consumer loads. On the basis of the calculated load, the engineer will select the following:

service cable impedance. Reducing the load current or the circuit impedance reduces the voltage drop. As load current is usually a fixed value, the engineer must find ways to reduce the circuit impedance. The engineer can reduce the transformer impedance by selecting the following:

• A secondary cable with adequate capacity, • A transformer with sufficient kVA for the diversified consumer load, and • A primary cable with ampacity based on the expected operating conditions.

• A unit with a lower impedance, or • A unit with a greater kVA rating.

Information for making these selections is contained in Section 4, Equipment Loading. In reviewing the total primary current for the load to be served, the engineer must select primary cables with the proper ampacity ratings. However, when decisions are made concerning these total primary load currents, care must be taken to also maintain load balance among phases on the feeders serving these loads. After making these selections, the engineer must check for voltage drop and voltage flicker at the consumer’s delivery point. Appendix B contains equations for calculating voltage drop and flicker. The voltage drop must not exceed the maximum values in Table B.1. Likewise, the magnitude and frequency of the voltage flicker must be within the permissible levels shown in Figure B.2. For a subdivision layout, it is not necessary to calculate these values for each consumer. Instead, the engineer should determine a few worst cases and perform the calculations for these only. For voltage drop, the worst cases are the longer secondary/service cables served from transformers having a greater number of connected consumers. For voltage flicker, the worst cases are a combination of longer secondary/ service lengths, larger motors, and smaller transformer sizes. Figure 1.25 highlights the worst cases for voltage drop and voltage flicker. If the calculated voltage drop exceeds the limits in Table B.1, the engineer must modify the design. Voltage drop is a product of load current and circuit impedance. For voltage drop at the consumer’s delivery point, the circuit impedance consists of the transformer and the secondary/

However, these methods are usually not cost-effective. A low-impedance transformer typically costs more than a standard unit and requires the utility to stock standard and nonstandard transformers. A transformer with a greater kVA rating costs more and also has higher core (no-load) losses. For residential services, it is more practical to lower the secondary/service cable impedance rather than the transformer impedance by doing the following: • Shortening the cable length, • Increasing the cable size, or • Paralleling cables. By placing the transformer closer to the consumer’s delivery point, the engineer can shorten the secondary/service cable length. Although the primary cable length is increased, this approach is often economical for single deliveries, particularly those with large secondary/service cables. The larger secondary/service cables can cost more than primary cable. If it is not practical to place the transformer closer, the engineer can increase the secondary/service cable size or can parallel two smaller cables. But instead of serving a single delivery, a transformer in a subdivision will serve multiple deliveries. Therefore, shortening the secondary/service cable lengths in a subdivision requires installing additional transformers. The cost of installing and operating these additional transformers may be greater than the cost of increasing the secondary/service cable size. In subdivisions, therefore, it may be more economical to increase the cable size rather than shorten the cable length.

4 2 – Se c t i on 1

1 cable route should be the Reducing the cable impedmost efficient way to serve ance also reduces the voltage Offset the primary all the transformers. For proflicker during motor starting. cable route at least jects with multiple transformFor large motors, this method ers, an open-loop feeder is may not limit the voltage 1 to 2 feet from any preferred. The primary cable flicker to the permissible levels property line. route should be offset at least shown in Figure B.2. For situaone to two feet from any tions involving polyphase moproperty line. Property owners tors, a consumer may use a often place fences along their property lines and starting method that reduces the motor inrush could damage buried cable placed on the propcurrent. The engineer needs to review large moerty line. tors and the proposed starting methods to see if The route should also minimize conflict with the arrangements will cause problems on the other buried utilities. One way to accomplish electric system or for other consumers. One this is to establish a utility corridor. Within the method of particular concern is the use of an corridor, each utility occupies its allocated electronic “soft” starter. Unlike conventional space, which allows each utility to know the apmethods, this type of reduced voltage starting proximate location of other utilities. A utility corproduces harmonics on the electric system. The ridor requires a wider easement than the usual harmonics result from chopping the voltage sine 10-foot easement for electric facilities only. Utiliwave to reduce the voltage at the terminals of ties may find this concept works well in subdivithe motor. sions where the developer has defined a wide utility easement on the subdivision plat. STEP 4. SELECT THE PRIMARY CABLE ROUTE Some developers may ask the cooperative to After locating the transformers and services, the locate its facilities within the street right-of-way. engineer must select a primary cable route. The

AY MW

GEHA

W

NE

BRID

Y WA TH

OU

RM YA

DC

OL

T.

AS

DC

AY KW

EA ST

CHARINGTON CT.

ELM

NEW DOVER ROAD

ROW


.

Single-Phase, Primary Voltage, UD Cable

FIGURE 1.26: Primary Cable Layout for 75-Lot Subdivision.

ROW

OW)

00' R

35 (1

SR 14

ROW

ROW


1 Equipment Pad

Clear Working Space

10' 0"

Although this is convenient for the developer, it usually creates future problems for the cooperative. These private roads are often released to the local city or state road system. These governments have rules about utilities located within the road right-of-way. Most require the utility to file a right-of-way encroachment, to bury cables at a specified depth, and to meet very high compaction levels during trench backfill. When the governing body decides to widen the road, the cooperative may have to relocate its facilities at its expense. The cooperative can avoid these conflicts by locating its facilities on a private right-of-way off the edge of the city, county, or state right-of-way. Finally, the selected cable route should minimize the number of road crossings. A faulted cable section under a road is difficult to repair or replace unless the cable is in a conduit. A conduit with cable or a spare conduit placed beneath the road allows the cooperative to replace the cable without disturbing the road surface. This method is acceptable for use with directburied, jacketed primary cable. It should be noted that placing the cable in the conduit will reduce the cable ampacity. Figure 1.26 shows a primary cable route for the 75-lot subdivision. This particular cable route has two road crossings.

FIGURE 1.27: Minimum Required Working Space.

of the equipment and 10 feet out from the equipment door as shown in Figure 1.27. Padmounted switchgear often has two sets of doors and, therefore, requires working space on both sides of the equipment. Another concern is damage from vehicles. Cars are likely to bump and damage equipment located in STEP 5. LOCATE parking lots. If the equipment SECTIONALIZING EQUIPMENT The minimum working cannot be relocated, the coopAfter selecting the primary erative may have to install cable route, the engineer can space extends out some type of barricade around locate the sectionalizing 10 feet from the the equipment. However, this equipment, which includes barricade must not block the riser poles, junction cabinets, equipment door. equipment doors or obstruct fuse cabinets, and switchgear. the required working space. These devices are used to proEquipment located along vide sectionalizing at desired streets and at intersections can be damaged by points within the UD system. Section 3, Undersnow removal equipment, particularly if the ground System Sectionalizing, describes the equipment is covered by snow. Another highdesirable locations for sectionalizing devices. risk area is a crop field. Tall crops can obscure Utility personnel have to operate and mainthe equipment, making it invisible to someone tain these devices; therefore, the equipment operating farm equipment. These high-risk areas needs to be in accessible locations. Operating must be avoided or adequate protective methods pad-mounted equipment requires enough workmust be used to minimize the chance for equiping space to move elbow terminators with hot ment damage. sticks. The minimum working space is the width

4 4 – Se c t i on 1

1 problems on moderate slopes. On more severe slopes, different trenching equipment and techniques will need to be used, along with an anchored or encased conduit and more aggressive erosion control techniques. Installing pad-mounted equipment on sloped terrain requires careful excavation to provide a level terraced surface for a monolithic pad • Sloped terrain, Visit the project or the use of a compartmental • Corrosive soils, style pad, even if the slope is site to identify • Rocky soils, moderate. For more severe • Sandy soils, problem terrain. slopes, the use of retaining • Unstable soils, or walls of stone or timber will • Flood plains. be needed along with molded or pre-cast ground sleeves of During the site visit, the engineer should look sufficient height to span the difference in elevafor these and other adverse terrain types along tion from the high side to the low side. Rememthe proposed cable route and at proposed ber to establish grades in such a manner that equipment locations. Ideally, the engineer will erosion of the soil down to the transformer is relocate the cable or equipment to avoid the minimized. Also provide for adequate level opproblem terrain. Unfortunately, relocation is not erating area in front of the equipment. always practical, and the engineer must adapt Although it is not always possible, the underthe design to reduce installation and mainteground designer should try to avoid sloped nance problems. Methods for adapting a design areas for the installation of conductors and deare described under the subheadings below. vices, or at least use the more moderate slopes This step is very important because it identiwhenever practical. fies problems before construction. If these problems require relocating cable or equipment, the Corrosive Soils engineer can easily modify the preliminary layTerrain features that indicate potentially severely out. Changing the location of equipment and corrosive soils are the following: cable during construction is very time-consuming and, therefore, more expensive. • Swamps, • Streams, Sloped Terrain • Poorly drained areas, or Installation of cable and equipment on sloped • Visible alkali (mineral salts). terrain presents a number of problems whose severity usually increases with the degree of These soils can corrode unprotected, buried slope. Trenching across sloped terrain is difficult neutrals and ground conductors. One way to because of problems controlling the mechanized protect neutral conductors is to prevent them trenching equipment safely while achieving a from contacting the soil by using jacketed cable. stable excavation whose sides are vertical. However, the counterpoise and/or ground elecTrenching up or down sloped terrain also has trodes must remain in contact with the soil and control and safety issues with the trenching be protected by another means. For information equipment and, additionally, introduces probon corrosion protection, refer to Section 7, lems with both surface erosion of the backfill Cathodic Protection Requirements. That secand tunneling erosion around the cable or contion explains how to determine if soils are corduit. Careful attention to tamping and comrosive and what types of cathodic protection paction, along with installing a stable ground are needed. cover, such as sod, will generally address these STEP 6. VISIT THE PROJECT SITE After completing the preliminary layout, the engineer must visit the project site to view the terrain. Certain types of terrain can make cable installation and equipment placement difficult or impractical. Examples of problems with terrain are the following:


1 Rocky Soils Rocky soils are often characterized by protruding boulders or rocks lying on the surface. Visible rock usually indicates underlying rock. To confirm the presence of underlying rock, the cooperative can make test borings with an anchor auger. Grading by the developer can also show signs of underlying rock. Rock along the cable route slows installation and increases project cost; therefore, the cooperative should reroute to avoid rocky areas. If rerouting is not practical, the cooperative will have to use special equipment that can penetrate rock. Because the rock is difficult to penetrate, it may be hard to maintain the required burial depth. If cable cannot be placed at the minimum depth, the cooperative must provide supplemental protection such as cable placement in Schedule 40 PVC conduit, rigid steel conduit, or conduit encased in concrete. The supplemental protection must meet the requirements of the 2007 NESC, Section 352 D.2.b. A final consideration in rocky soils is damage to the underground cable, particularly the cable jacket. Rocks directly contacting the cable can damage the jacket. One way to protect the cable is to use conduit or a cable-in-conduit assembly. Either of these can be installed by trenching, plowing, or tunneling. Another option for protection in a trench is to use a select backfill for a cable bed and covering. Sandy Soils Sandy soils can cause problems in at least three different ways: • Difficulty opening a trench, • Wind erosion of sand from under equipment, and • Sandblasting of painted metal surfaces. Sandy soils shift easily from the wind and can undermine the support of pad-mounted equipment. In these areas, pads with ground sleeves or basements provide better support and more security than a flat pad does. They also help prevent exposure of cables that enter the equipment. Alternatively, the wind can blow sand and cover pad-mounted equipment, making it difficult to

access and operate. This condition is improved by using silt fencing or shrubbery as a wind block. However, installing a wind block does increase the initial project cost and future maintenance expenses. Windy conditions in a sandy environment provide nature’s own sandblasting machine, making it difficult to keep paint on padmounted equipment. After the wind-blown sand removes the paint, the exposed metal quickly corrodes, especially in coastal environments. One solution to this damage is to use stainless steel (or other noncorrosive) equipment cabinets. This adds substantially to initial cost, but maintenance will be much more practical and economical. Another option is to use overhead primary with underground services as the only underground facilities. Placing transformers on poles provides extra distance from the ground and may eliminate the problems caused by blowing sand. Sandy soils have little cohesion and usually will not hold a trench open for cable placement. In addition, these soils are often in areas with a high water table. As a result, trenches fill with water and are difficult to excavate. When trenchers are used in these conditions, they are often equipped with a cable chute. Another acceptable installation method is to use a cable plow. Increased burial depths (an additional six to 12 inches) should be considered because covering can be blown away. Unstable Soils Some examples of unstable soils are the following: • • • •

River banks, Natural springs, Unsecured embankments, and Steep grades.

These soils shift easily and are also prone to washing. Washing can erode trenches and undermine the support of pad-mounted equipment. Trench erosion can reduce the soil cover and possibly expose a buried cable. Cables may also be exposed where soil has washed away from an equipment pad. If the washing is severe, the

4 6 – Se c t i on 1

1 building codes forbid the placement of strucequipment could shift enough to damage transtures in flood plains, these areas can usually be former bushings or cable terminations. traversed with cable. Flooding has little effect on Washing can also have the opposite effect. Inburied cable and should present problems only stead of undermining pads, it can deposit large if the cable fails while the cable route is flooded. amounts of soil around a piece of equipment. If the cable section is part of an open-loop sysProlonged contact with soil deposits causes the tem, the flooded section can be isolated. Howmetal housing of the equipment to corrode. ever, if the cable section is part of a radial Such corrosion can lead to premature equipsystem, the engineer should consider providing ment failure and possible access to the interior an alternative feed. compartments. The soil deposits can also block The cooperative may have to place equipthe equipment doors, making it difficult to mainment in areas subject to floodtain the equipment. ing. Dead-front, pad-mounted Unstable soils can also transformers and dead-front, make installation difficult. If Unstable soils oil-insulated switching cabigrades or embankments are can make nets can operate during occatoo steep or if soils are too sional immersion. However, wet, construction personnel installation difficult. these devices must be supwill have problems maneuverported by pads that will not ing a trencher or a cable plow. float. Otherwise, the device Wet soils also tend to collapse may be displaced, possibly causing a system back into the open trench, making it difficult to outage or exposing the interior compartments. maintain proper depth for cable burial. Air-insulated switching cabinets will fail if subTo eliminate these types of problems, the enmerged in water and, therefore, must not be gineer should avoid routing cable or placing used in areas subject to flooding. equipment on steep slopes. If this is not practical, doing the following minimizes erosion: STEP 7. OBTAIN ALL EASEMENTS • Proper compaction and crowning of the The cooperative must get an easement from trench, all affected property owners before installing • Replanting of the slope, or any underground facilities. By definition, an • Use of equipment pads with ground sleeves easement is a right afforded a person to make or basements. limited use of another’s real property. This easement gives the utility the legal right to enter the If the potential for trench erosion is severe, the property and access a right-of-way strip. For engineer should consider placing the cable in underground facilities, this right-of-way must conduit or installing a cable-in-conduit assembly be a minimum of 10 feet wide—five feet on and possibly encasing the conduit with concrete. each side of the centerline of the electrical faciliThe engineer should also avoid placing ties. The 10-foot width provides enough space equipment at the bottom of steep slopes. If to operate a trencher or other piece of installaequipment must be placed in these areas, the tion equipment. The easement must define the cooperative will need to construct a water and width and boundaries of this right-of-way strip. soil block to prevent soil accumulation around These rights-of-way should also be shown and the equipment. recorded on the plat. To reduce misunderstandings between the cooperative Flood Plains Get an easement and its property owner memThe best way to evaluate for bers, the easement must be possible flooding is to check before installing any specific in defining the cooptopographical maps that locate underground facilities. erative’s rights. As a minimum, flood plains. Though most


1 Sample Easement

Project No. ________________________ Drawn by ___________________________

STATE OF ______________________________ COUNTY OF ____________________________ KNOW ALL MEN BY THESE PRESENTS, that __________________________________________________, _________________________________________________________________________________________ hereinafter called “Grantor” (whether one or more), in consideration of the sum of One Dollar ($1.00) and other good and valuable considerations, does hereby grant unto__________________________________, its successors, and assigns, hereinafter called “Grantee,” the right, privilege, and easement to go in and upon that certain land of Grantor (hereinafter “premises”) situated in said County and State, bounded by lands of: _________________________________________________________________________________________ _________________________________________________________________________________________ _________________________________________________________________________________________ _________________________________________________________________________________________ _________________________________________________________________________________________ and over and across said premises within a right-of-way strip having a width of _____ feet on each side of a centerline determined by the centerline of the electrical facilities as installed, to construct, maintain, and operate underground lines and conduits with other necessary apparatus and appliances, either above ground or below ground, to include transformers and service connections, for the purpose of transporting electricity and for the communications purposes of Grantee and its licensees. The following rights are also granted to Grantee: to enter said premises to inspect said lines, to perform necessary maintenance and repairs, and to make alterations and additions thereto; and to clear the land within the right-of-way strip and to keep it clear of trees, structures, or other obstructions; and to clear that land outside the right-ofway strip and to keep the area within 10 feet of said door clear of trees, structures, or other obstructions. All underground facilities are to be installed in accordance with the provisions of Grantee’s Underground Distribution Installment Plan, __________________________________, receipt of a copy of which is acknowledged by Grantor. This right-of-way is given to permit the construction of electrical facilities presently proposed. Facilities at other locations and future extensions of presently constructed facilities are not permitted by this agreement. The foregoing notwithstanding, Grantee may relocate its electrical facilities and right-of-way strip over the premises to conform to any future highway or street relocation, widening, or improvement. IN WITNESS WHEREOF, the Grantor has hereunto set his hand and seal, or, if corporate, has caused this instrument to be signed in its corporate name by its fully authorized officers and its seal to be hereunto affixed by authority of its Board of Directors, this _______ day of __________________, 20___. Witness________________________________________________________________________(SEAL) ________________________________________(SEAL) ________________________________________(SEAL) ________________________________________(SEAL) ______________________________________________ (Corporate Name) ATTEST:____________________________________ By_________________________________________ _________ Secretary

FIGURE 1.28: Sample Easement.

_________ (President)

4 8 – Se c t i on 1

1 the cooperative must have the right to install, operate, maintain, and replace the electrical facilities located within the right-of-way strip. These activities require the right-of-way to be clear of trees, structures, and other obstructions. Rights-of-way that were clear during the installation of underground facilities will likely become obstructed as property owners erect fences, storage buildings, and landscaping. Because these obstructions must be cleared to repair or replace the underground facilities, the easement must specifically define the cooperative’s right to clear the right-of-way. Another area of conflict is clear space in front of the doors of transformers and sectionalizing cabinets. As noted, maintenance of these devices requires a clear working space 10 feet out from the door (see Figure 1.27). The consumer may consider these devices unattractive and try to hide them with landscaping or a surrounding structure. As a result, the cooperative cannot maintain the device. These conflicts are more easily resolved if the easement states that the area within 10 feet of the door of any transformer or cabinet will be kept clear of any obstructions. Because the easement is a legal document, it must be filled out completely and correctly, including getting the signatures of all property owner members of a particular tract or the signatures of appropriate corporate officers, if owned by a corporation. The easement must be notarized and filed with the appropriate municipal, parish, or county authority in which the property lies. Figure 1.28 shows a sample easement. Obtaining and recording an easement can be time-consuming, particularly if one underground project involves multiple property owners, thus requiring multiple easements. To avoid this problem in a subdivision, the cooperative is wise to get one easement from the developer before any lots are sold. This way, the cooperative needs only one easement for all the planned underground facilities in the subdivision. The cooperative will also benefit if the following occur: • The right-of-way strip is shown and recorded on the plat.

• The subdivision restrictions define the cooperative easement as transferable to new owners. STEP 8. Prepare Staking Sheets The final step is preparing a staking sheet. For smaller projects, the staking sheet provides enough space for a sketch of the required work. Figure 1.29 shows a staking sheet for underground service to a commercial consumer. For larger projects, the engineer will have to attach a separate construction drawing. For subdivision installations, the engineer can show the required work on a subdivision plat. This construction drawing should show the trench, equipment, and street lighting locations and note any conduit or temporary pedestal installations. The construction drawing could also have details showing how far to offset equipment from the property line and the location of other underground utilities. Underground staking sheets provide important project information to several departments within the cooperative. These departments must be able to easily interpret the staking sheet. The staking sheet is used to generate a materials list. Purchasing and materials personnel use this list to order and stock the necessary materials. Scheduling personnel will use the staking sheet to estimate the manpower and equipment needed to construct the project. Staking personnel use the sheets to physically mark the field locations of equipment and trenches. While in the field, staking personnel may have to adjust the layout for conflicts with other utilities or for terrain problems. Any changes made in the field must be shown on the staking sheet. After personnel have staked the project, construction crews will use the staking sheets for information on installing the underground facilities. If the construction crews modify the layout, they must also modify the staking sheet. The staking sheet must agree with the as-built project because these sheets are the basis for the cooperative’s mapping system. Accurate staking sheets produce accurate system operating maps and accurate permanent plant and accounting records.

FIGURE 1.29: Staking Sheet for Service to a Commercial Consumer. Source: Piedmont Electric Membership Corporation, Hillsborough, N.C.

H

H UT SO WY 86

Desi g n o f a n Un d e rg ro u n d Di s t r i b u t i o n Sy s t e m – 4 9

1

5 0 – Se c t i on 1

1 Summary and Recommendations

1. Equipment mountings provide support for pad-mounted equipment. Flat pads are sometimes suitable for single-phase padmounted transformers and small singlephase fuse cabinets. 2. Cable wells used with a flat pad provide more space for cable training and are suitable for three-phase pad-mounted transformers and junction cabinets. 3. A box pad is useful to support pad-mounted switchgear and for installations on slopes or hillsides. 4. The main types of underground systems are the following: • • • • •

circuit exits, main feeders, sub feeders, transformer and secondary systems, and street and area lighting.

5. In designing a UD system, safety, reliability, cable ampacity, equipment ratings, voltage drop, and voltage flicker must be considered. 6. Placing facilities along the front property line makes them more accessible for operation and maintenance. 7. A joint-use trench often creates operating problems. To minimize these problems, the

8.

9.

10.

11.

12.

location of a joint-use trench must be shown on all operating maps. Joint-use trench with other utilities requires a contractual arrangement among involved parties. System upgrades should be planned by considering future voltage conversions, three-phase cable installation, and conduit installations. The UD design can be improved by comparing the economics of different system configurations. The UD system should be designed to minimize cable and pad-mounted transformer losses. The steps for layout of a UD system are as follows: STEP 1: Get the required information. STEP 2: Arrange the service and trans-

former layout. STEP 3: Calculate the consumer load and STEP 4: STEP 5: STEP 6: STEP 7: STEP 8:

select proper equipment ratings. Select the primary cable route. Locate sectionalizing equipment. Visit the project site. Obtain all easements. Prepare staking sheets.

Cable Se l e c t i o n – 5 1

2 In This Section:

Typical Cable Configuration

Cable Selection

Typical Cable Configuration Conductor Size Designations

Conductor Shields and Insulation Shields

Conductor Materials and Configuration

Cable Specification and Purchasing

Cable Insulation Materials

Cable Acceptance

Insulation Fabrication


The heart of any underground system is the cable that carries power from the source to the load. The cable must incorporate the most important characteristics of the ideal utility system: low initial cost and high reliability. Experience with early UD cables has forcefully shown that the lowest-cost cable that can be successfully placed into operation is not necessarily the best choice. It is necessary to pay close attention to the design and manufacture of all cables. This section provides an introduction to the technical aspects of electric distribution cables.

It addresses the designs and materials most effective in delivering reliable and economical service. The variety of cable features available for the various applications is also addressed. Recommendations are included for conditions generally encountered on rural electric systems. The main components of cables reviewed include the conductor and the insulation system (including shielding). The concentric neutral and jacket options for primary voltage cables are also covered.

The main types of cables used on rural electric systems are primary voltage (15- to 35-kV class) shielded cables and secondary voltage (600-volt class) unshielded cables. The higher voltage cables are used on systems rated 7.2/12.5 kV, 14.4/ 24.9 kV, and 19.9/34.5 kV. Such cables are classed by the phase-to-phase voltage of the system on which they are intended to operate. For instance, cable designed for application on a 7.2/12.5-kV system will be rated 15 kV, regardless of whether it is in a single-phase or a three-phase circuit.

To gain an overview of cable design, the engineer should consider the components of the system. The focus should be on single-conductor cable because it is the predominant type of cable used in rural and suburban distribution systems in North America. Typical system voltages are 7.2/12.5-kV, 14.4/24.9-kV, and 19.9/34.5-kV grounded wye. Most of the cables on these systems are of concentric neutral design. Generally, the major cable components are the following:

5 2 – Se c t i on 2

2 • • • • • •

Conductor, Conductor shield, Insulation, Insulation shield, Concentric neutral, and Jacket.

These are illustrated in Figure 2.1. Figure 2.1 represents a typical primary cable used in underground distribution and is the configuration currently recommended. Variations of this design may be better suited to particular types of installations. Figure 2.2 shows the arrangement of an underground cable design widely used from the mid1960s to the late 1980s. It is identical in most respects to the cable in Figure 2.1, except that it does not have a jacket over the concentric neutral. It was most often installed as a direct-buried cable, and exposure of the concentric neutral to the surrounding earth provided an excellent system ground. However, this cable design fell into disfavor because of substantial corrosion problems affecting the concentric neutral. Loss of the neutral wires led to an open neutral circuit, posing serious operational, reliability, and public safety

Conductor Extruded Conductor Shield

Insulation

Extruded Insulation Shield Encapsulated Neutral Conductors

problems. In addition, loss of neutral conductors caused deterioration of the semiconducting insulation shield and consequent cable failure. Use of bare concentric neutral cable is not approved by RUS for use on the systems of its borrowers and has essentially been discontinued except in cases where there are no corrosive conditions and special permission has been obtained. Another special case of the medium-voltage, single-conductor cable is illustrated in Figure 2.3. The main difference from the cable in Figure 2.1 is that the concentric neutral is replaced by a longitudinal corrugated (L.C.) shield or a copper tape shield. A separate neutral conductor thus must be installed with a circuit to handle return currents. The purpose of the L.C. shield or tape shield is to provide a path for capacitive currents and, thus, ensure an even voltage gradient within the cable. The major advantage of this configuration is in circuits where loads are relatively high (≥ 300 amperes). See Section 4 for more information on sheath currents and cable ampacity. The following subsections, which describe individual components of underground cables in more detail, provide an understanding of desirable features for various applications.

Conductor Extruded Conductor Shield

Extruded Conductor Shield

Insulation

Insulation

Extruded Insulation Shield Bare Neutral Conductors

Extruded Insulation Shield

Metallic Tape Shield

Jacket

Jacket

FIGURE 2.1: Jacketed Concentric Neutral Cable. Source: Okonite Company, 2006.

Conductor

FIGURE 2.2: Bare Concentric Neutral Cable. (Not RUS accepted.) Source: Okonite Company, 2006.

FIGURE 2.3: Medium-Voltage Power Cable with Tape Shield and L.C. Shield. Source: Okonite Company.


2 U.S. standards use two systems for designating conductor size. The oldest of these is the AWG, which was formerly known as the Brown and Sharpe wire gauge. This system is typically used on conductors up through those with a diameter

Conductor Size Designations

Equation 2.1 A = πr2 where: A = Area in square inches π = 3.1416 r = Radius (in inches) For Area in kcmil, use Radius in 1/1,000 inch.

TABLE 2.1: Dimensional Characteristics of Common Conductors (Standard Concentric-Lay). AWG

kcmil

6*

Area

Diameter (in.)

mm2

in.2

26.24

13.3

0.0206

0.162

2*

66.36

33.6

0.0521

0.258

1/0*

105.60

53.5

0.0829

0.325

2/0*

133.10

67.4

0.1045

0.365

4/0*

211.60

107.0

0.1662

0.460

—

250.00**

127.0

0.1967

0.575

—

350.00**

177.0

0.2749

0.681

—

500.00**

253.0

0.3927

0.813

—

750.00**

380.0

0.5891

0.998

—

1,000.00**

507.0

0.7854

1.152

* Solid ** Stranded

Conductor Materials and Configuration

MATERIALS Since the first cable system, only two conductor materials have played a significant role: copper and aluminum. These materials have appeared in a variety of alloys, tempers, and configurations. In the late 1960s, some utilities briefly experimented with sodium as a conductor

of 0.460 inch (4/0 AWG). The second system is the circular mil designation, which is always used on conductors larger than 4/0 AWG. However, circular mil designations may also be applied to conductors of 4/0 AWG and smaller. The AWG originated in the mid-19th century. Each step in this gauge approximates the successive steps in the wire drawing process. Empirical history sets the two endpoints: 4/0 AWG, with a diameter of 0.460 inch, and No. 36 AWG, with a diameter of 0.0050 inch. There are 39 equally divided steps between these two sizes. A few approximate relationships may be useful: • Each increase of three-gauge numbers doubles the area and the unit weight, and also halves the dc resistance. • Each increase of six gauge numbers doubles the diameter. • Each increase of 10 gauge numbers multiplies the area and unit weight by 10, and also divides the dc resistance by 10. The circular mil system is based on the definition of a circular mil (cmil) as being the area of a wire with a diameter of one mil (0.001 inch). Area is calculated as shown in Equation 2.1. One cmil is (0.0005)2 π or 7.854 × 10-7 inch2. It follows that 1,000 circular mils or 1.0 kcmil (formerly MCM) is equal to 7.854 × 10-4 inch2 in solid wire. Therefore, a 4/0 AWG wire, which has a diameter of 0.460 inch, has a circular mil equivalency of 211,600 cm and an area of 0.1662 inch2. The AWG and circular mil systems are now limited to U.S. and Canadian use. European designations are based on metric units of square millimeters (mm2). Table 2.1 shows AWG, circular mil, and metric designations for common conductor sizes used in North American distribution cables.

material; however, it was not cost-effective because of the special precautions required during installation and maintenance. Copper was the first material to play a major role in cable construction. With a volume resistivity of 1.673 × 10-7 ohm-meters (ohm-m) in its pure (99.999 percent) state, it compared favorably with

5 4 – Se c t i on 2

2 TABLE 2.2: Conductor Physical and Electrical Characteristics. Copper Soft Drawn

Medium Drawn

Rated Tensile Strength

—

Conductivity (% IACS)

100

Aluminum Hard Drawn

1/2 Hard (H14/H24)

3/4 Hard (H16/H26)

Hard Drawn (H19)

42–60 ksi

49–67 ksi

15.0–20.0 ksi

17.0–22.0 ksi

24.5–29.0 ksi

96.7–97.7

97.2

61.0

61.0

61.0

Note. ksi = thousands of pounds per square inch

other metals. Supplies were abundant and it could be economically fabricated. Connections were simple to make and corrosion resistance was good. However, with the rapid development of the aluminum industry in the first half of the 20th century, aluminum became cost-effective for applications in which physical size was not critical. To take advantage of this economic benefit, the electric industry developed methods to overcome some of the other physical disadvantages of aluminum. These disadvantages included higher susceptibility to flexural fatigue, the high resistivity of natural surface oxides, and cold flow (creep). For economic reasons, cables now used on underground systems are predominantly aluminum. The use of this metal leads to a larger cross-sectional area and, consequently, greater overall cable dimensions, but, in most cases, the additional cost of other project components— such as larger size conduit—does not outweigh the present economic advantage of aluminum conductors. Aluminum conductors have a volume resistivity of 2.655 × 10-7 ohm-m. Comparing this resistivity with the previously mentioned copper volume resistivity shows that, for equal crosssectional areas, aluminum will have 1.59 times the resistance of the same-size copper conductor. To simplify the comparison of various conductors, the industry uses a measure of relative conductivity that compares a particular metal to annealed electrolytic copper. This measure is referred to as the International Annealed Copper Standard (IACS). The volume resistivity of annealed copper is defined as 1.724 × 10-7 ohm-m at a temperature of 20°C (68°F). As the tensile strength of materials increases, the conductivity decreases. As an example, harddrawn copper has experienced an increase in

tensile strength because of work hardening during the drawing process and its conductivity has fallen to 97.2 percent IACS. By comparison, 1350H19 aluminum has a conductivity of about 61 percent IACS. The lower conductivity is mainly caused by the inherently higher volume resistivity of pure annealed aluminum. See Table 2.2 for a comparison of common conductor materials. Because thermal capacity of conductors and cables is a function of the heat generated by internal conductor losses, the ampacity of the higher conductivity copper conductors of equal size is approximately 1.6 times that of matching aluminum conductors. Of course, other significant elements determine the exact cable ampacity. These are discussed more extensively in Section 4 of this manual. CONDUCTOR TEMPER Both copper and aluminum conductors are available in various tempers that designate the relative hardness of the metal. Whereas overhead conductors have generally used harder metal to increase tensile strength and reduce sags, underground conductors have tended to use the lower tempers, because high tensile strength was not usually required. Most copper power cables have used soft-drawn copper for its greater flexibility. This flexibility not only makes fabrication easier but also improves installation handling, especially for larger cables. Where high tensile strength is needed for cable pulling, special installations might use harder tempers. However, this would only be where high unit stresses would be imposed on the cable conductor during installation or perhaps during cable life. Examples include mineshaft riser cables or cables for extremely long pulling


2 distances in duct. Such cables would require customized design for their particular circumstances and are beyond the scope of this manual. Aluminum conductors in power cables are generally furnished in the 3/4-hard temper. This provides a reasonable level of tensile strength, while not introducing excessive ductility that would lead to creep problems in making durable connections. As the conductor cross section increases to 750 kcmil or greater, there is some acceptance of aluminum conductors in the 1/2-hard temper. This gives adequate tensile strength while maintaining a higher degree of flexibility. All characteristics of aluminum conductors, especially tensile strength, must be considered when specifying a cable. The specifying engineer must consider the mechanical stresses on the cable during installation and service. More information on conductor characteristics can be found in reference books. Nationally accepted specifications for electrical conductors are found in American Society for Testing and Materials (ASTM) standards. Copper wire is covered by ASTM Specifications B-1, B-2, and B-3. Aluminum wire is covered by ASTM Specification B-230. Methods for measuring the most important characteristics of these and other materials can be found in other related ASTM standards. Aluminum conductors used in underground cable are addressed in other ASTM standards, including B-231 (concentric lay conductors) and B-400 (compact round conductors). CONDUCTOR ALLOY Aluminum conductor material also is designated by an alloy number. The alloy designation derives from the description of aluminum alloys in other applications in which such characteristics as high tensile strength are required. However, because high electrical conductivity (low resistivity) is the single most important aspect of underground cable conductors, pure aluminum is generally used. The alloy designation for electrical aluminum is EC. It was formerly designated as Alloy 1350. The same aluminum nomenclature system includes designations for temper. These are also shown in Table 2.2. For example, 3/4-hard temper has a classification of H16 or H26. The difference between H16 and H26 tempers is that

the H16 alloy is only strain-hardened. The H26 alloy has the same general characteristics, but it has been partially annealed after strain hardening. Copper conductors are almost universally supplied as pure copper. Pure copper provides the highest conductivity and, therefore, the highest efficiency. Because pure copper in its various tempers provides adequate mechanical strength for cable applications, there is generally no need for alloyed copper conductors. CONDUCTOR CONFIGURATION The wire and cable industry offers the electric utility industry a wide variety of standard conductor configurations, including solid conductor, various stranding arrangements, and filled-strand conductors. Each configuration has its own advantages. The engineer selecting a cable design must consider these alternatives and select the option that produces the best cable for the particular application. Elements significantly affected by the conductor configuration include the following: • Flexibility during installation (cable bending and racking), • Flexibility during operations (elbow switching), and • Longitudinal water migration. Though the decision on conductor configuration alone will not provide the solution to any of these problem areas, it is a vital part of the larger process of selecting a cable that will provide high reliability and economy. The simplest configuration is the solid, singlestrand conductor. Solid conductor is preferred in smaller cable sizes because of its absolute waterblocking capability. Because there are no voids to fill, there will be no continuing migration of water through the insulation system. Perhaps more important, if moisture does penetrate the insulation, it cannot migrate along the cable conductor to other areas of the cable. The inhibition of moisture migration is extremely important in reducing insulation deterioration problems so prevalent in underground cables. As is well known, the stiffness of cable increases as conductor diameter increases. Cable

5 6 – Se c t i on 2

2 stiffness will increase in proportion to the square of the diameter of the solid conductor. Therefore, a point will be reached at which the cable will become unmanageable, especially where bending in confined spaces is required to

operate load-break connectors. The solution is the use of Use solid or strandstranded conductors. The filled conductors smaller diameter of the individual strands lowers the total for reliability. force required to achieve the necessary bending. The reasonable upper limit for solid conductors with 3/4-hard aluminum conductors has generally been found to be 2/0 AWG. Above that size, stranded conductors are advised. Several options in stranded conductors are available, including conventional concentric lay, compressed strand, and compact configurations. Some of these are illustrated in Figure 2.4. The simplest stranded configuration is the Concentric Stranded Conductor, 37-Wire conventional concentric round stranding that uses multiple layers of circular wires. Each layer of wires is laid in the opposite direction. The predominant combinations for conventional stranded cable are 1 + 6 = 7, 1 + 6 + 12 = 19, and 1 + 6 + 12 + 18 = 37. These are illustrated in Figure 2.5. The first option, concentric round stranding, Compressed-Strand Concentric Conductor, 37-Wire obviously produces interstices (voids) between the surfaces of the individual wires. These interstices have two important effects. First, for a given equivalent metallic cross section of conductor, the outside diameter of a stranded cable will be greater than for an equivalent solid conductor. Second, the voids are continuous along Compact Concentric-Strand Conductor, 37-Wire the cable and provide an excellent path for moisture migration. In conventional stranding, FIGURE 2.4: Concentric Lay Strand Options. the conductor metal will occupy only 76 to 78 percent of the area enclosed by a circle drawn around the outside of the conductor. The number of wires in a concentric stranded conductor is defined in ASTM standards as the class of the conductor. Details are contained in ASTM Standards B8 (copper) and B231 (aluminum). An examination shows that the im1 6 12 18 24 30 36 42 proved flexibility of higher stranding comes at Number of Wires Per Layer the expense of larger diameter. In addition, the stranded conductors weigh more because the outer layers must be longer than the conductor axis. Table 2.3 compares the various stranding characteristics of a common single size of aluminum conductor. FIGURE 2.5: Standard Strand Arrangements for Multilayer Conductors.


2 TABLE 2.3: Configurations of 4/0 AWG Aluminum Conductor. Stranding Class Number of Wires

Cable Insulation Materials

Individual Wire Diameter (in.)

Overall Diameter (in.)

Weight lb./1,000 ft

DC Resistance Ω/mile @ 20°C

Solid

1

0.4600

0.460

194.7

0.4228

A, AA

7

0.1739

0.522

198.7

0.4311

B

19

0.1055

0.528

198.5

0.4311

C

37

0.0756

0.529

198.5

0.4311

D

61

0.0589

0.530

198.7

0.4311

The second stranding option, compressed strand, is an improvement on the conventional strand arrangement. This configuration is accomplished by drawing the completed conventional concentric round strand to compress the outer layer of strands after fabrication (see Figure 2.4). The result is some reduction in diameter and some reduction in the interstices of the outer strand layer. This configuration also gives a smoother, more nearly cylindrical surface. In compressed strand, the conductor metal will occupy 81 percent to 83 percent of the area of a circle that encompasses the overall diameter. Compressed strand reduces the diameter between one-half and three percent. One disadvantage is some loss of flexibility because the outside layer is slightly more rigid. The third conductor type (see Figure 2.4) is the compact round design. With this design, the conductor is drawn after each layer is applied, which greatly reduces the interstices on each layer and brings the metallic cross section up to 92 to 94 percent. The cable diameter is reduced by about nine percent when compared with the same crosssectional area in a concentric round configuration.

FILLED-STRAND (SEALANT) CONDUCTOR As noted in the previous subsections on conductor configurations, the useful service life of underground cables has been reduced by moisture in the insulation system. This has been particularly true where moisture has been present in the conductor interstices and, thus, had access to the conductor/conductor shield interface. Therefore, it is important to stop the migration of any moisture that may find its way into the conductor. Impeding moisture migration is most economically accomplished by the addition of a strandfilling material during manufacture to fill all voids within the conductor. The material must be compatible with the conductor and the semiconducting strand (conductor) shield. Often, this requirement means the strand filler will be essentially the same as the strand shield except for plasticizers added to improve viscosity. The strand filler is often applied to each of the inner layers during the stranding process. If this approach is used with proper controls, the interstices should be filled while the outside of the conductor is left clean.

OVERVIEW OF CABLE INSULATION MATERIALS Early cable insulation materials were mainly natural rubber compounds. Paper insulation was introduced for power cables about 1890. Butyl rubber was introduced in 1944 for distribution cable systems. The trend toward placing electric distribution lines underground was significantly aided in the 1960s by the wide acceptance in the United States

of high-molecular-weight polyethylene-insulated cables. Cables constructed of HMWPE were introduced in 1948. These had several advantages over the butyl rubber primary voltage cables predominant in industrial applications. In the early 1960s, EPR (ethylene propylene rubber) insulated cables became available for distribution systems. However, the industry considered EPR cables to be premium-priced cables and they

5 8 – Se c t i on 2

2 improvement in cable life expectancy as predid not gain wide acceptance, especially in the dicted by accelerated testing methods. TR-XLPE underground distribution market where initial proved to generally be the superior compound cost was the governing factor before the imporand gained much wider acceptance than did tance of long cable life was recognized. TR-HMWPE. In fact, TR-HMWPE is no longer About 1963, cross-linked polyethylene (XLPE) manufactured. The tree-retardant characteristic cables became available for distribution installaof the initial TR-XLPE compound was acquired tions in both concentric neutral and convenby adding organic compounds to the basic polytional power configurations. The initial ethylene material. advantage of XLPE cable was that, like EPR, it is As a result of escalating polyethylene cable a thermosetting material with a higher allowable failure rates, EPR cables have seen wider accepoperating temperature. It was 1975 before XLPE tance in UD installations. Since the 1960s, these cable equaled HMWPE cable in market share for cables have also enjoyed technical improvedomestic utilities. HMWPE cable continued to be ments in insulation compounds and fabrication popular because of its better technical charactertechniques. History and accelerated life tests istics and lower cost. HMWPE possessed a low have shown EPR to be equal or superior to condielectric constant, along with high dielectric temporary TR-XLPE compounds. Without quesstrength and very good insulation resistance. It tion, insulating compounds will continue to also cost less than XLPE. In 1966, some utilities improve. Continuing tests will evaluate the reported failures of HMWPE-insulated cables. longevity of different cable compounds and The failure rate was about one per 1,000 milecable fabrication methods. Cooperative engiyears. By 1970, the reported HMWPE failure rate neers must use all available information when had reached about two per 1,000 mile-years and selecting a cable for purchase. was considered to be a significant problem. One hundred percent insulation wall thickSoon thereafter, the failure rate for HMWPE canesses are 175 mils (4.4 mm) for 15 kV, 260 mils bles rapidly escalated and reached almost eight (6.6 mm) for 25 kV, and 345 mils (8.8 mm) for per 1,000 mile-years by 1982. Recognition of 35 kV. These insulation wall thicknesses are premature insulation breakdown in HMWPE caspecified by the ANSI/ICEA and are referred to bles contributed to the rapidly increasing accepas the 100 percent level. Many cable users spectance of XLPE as an insulating material. About ify an increased wall thickness, as discussed 1975, the reported failure rate of XLPE cables below, and use this minimum 100 percent insureached one per 1,000 mile-years. In about 1980, lation wall thickness only for upgrading or retrothe failure rate of XLPE cables rapidly escalated, fitting projects in which duct sizes are restricted just as HMWPE insulation did earlier. and conduit fill may be exceeded. Because of concerns with the failure of Polymer insulation thicknesses are often inHMWPE and XLPE insulations to deliver the excreased to 133 percent or 173 percent of the valpected design life, cable insulation manufacturues listed above. The choice of insulation ers began searching for methods to improve the thickness depends on the system connection (eilife of the product. The initial major developther delta or wye connected), the system protecment was tree-retardant polyethylene (TR-PE) tion available, and the desire for longer cable compounds, so named because it resisted the life. Standards state that the growth of electrochemical 100 percent insulation level is “trees” which led to insulation satisfactory for any system failure. These have been introReview the results of where faults can be cleared duced in both high-molecularaccelerated cable life within one minute, which apweight polyethylene plies to most installations on (TR-HMWPE or TR-PE) and tests when selecting grounded systems. For deltacross-linked polyethylene cable insulation. connected or ungrounded sys(TR-XLPE). These compounds tems, 133 percent insulation have exhibited a substantial


2 thickness is commonly chosen. In addition, the 133 percent insulation level is recommended by standards where fault-clearing times on wyeconnected systems are in excess of one minute but less than one hour. The additional insulation thickness will also reduce the electrical stress within the insulation and, hence, prolong cable life, which many utilities find advantageous. One disadvantage of an increase in insulation thickness is that the additional insulation volume increases the opportunity for contamination. However, this is not a realistic concern for modern cable manufacturing facilities. Also, the additional insulation, shield, and jacket materials needed because of the increased diameter will increase the final installed cable cost. This is due to the increased cost of the cable, the increased pulling and training effort, and the increase in duct size required. Finally, 173 percent insulation is used for cables on a system, usually delta or resistance-grounded, which may have a clearing time of more than one hour. It should be noted that the performance of 175-mil direct buried distribution cables on 12.5/ 7.2 kV systems proved unsatisfactory in early underground systems. This was due to treeing of the insulation which could in part, be attributed to the higher voltage stresses present in the 175-mil insulation. This was particularly true in smaller (e.g., #2 AWG) conductor sizes. For this reason, RUS mandates the use of 133 percent insulation thickness (220 mil) for 15 kV class cables. RUS is currently refining its Specifications for Underground Primary Cables in Bulletin No. 1728F-U1, which updates and supersedes former Bulletin 50-70 (U1), dated December 22, 1987. In this new bulletin, RUS adopted the insulation thickness shown in Table 2.4 and these will be specified in the pending bulletin.

TABLE 2.4: RUS Insulation Thickness. Voltage Class (kV)

Insulation Thickness (mils)

Thickness Level (%)

15

220

133

25

260

100

35

345

100

INSULATION MATERIAL CHARACTERISTICS An individual selecting a particular cable insulation should be familiar with the basic physical and electrical characteristics of various materials. Each of these characteristics affects the suitability of an insulation material for a particular application. Selecting a cable construction involves compromise as most materials have different strong and weak points. Physical characteristics of the insulating layer affect the resistance of a cable to mechanical damage under normal operating conditions. Situations imposing mechanical stresses on cable include the following: • Soil pressure in direct-burial installations, • Sidewall pressure on cables pulled into conduit, • Flexure during switching operations for elbow-connected apparatus, • Expansion/contraction in ducts, and • External clamping action on risers. Some of the pertinent physical properties are listed below. Hot Creep This is a measure of the plasticity of a material at elevated temperatures. It shows the ability of an insulating material to resist deformation at elevated operating temperatures. For thermosetting insulations, the hot creep is generally measured at 130°C (266°F), which is the maximum emergency operating temperature. The hot creep is determined by measuring the tensile stress (pounds per square inch, or psi) needed to stretch the insulation sample to 200 percent of its original length. See Figure 2.6 for a relative comparison of the hot creep of HMWPE (thermoplastic), XLPE (thermosetting), and EPR (thermosetting). High-Temperature Aging Characteristics Electrical insulation in power cables must retain good physical properties after being subjected to high temperatures. High-temperature aging evaluations usually compare tensile strength and elongation remaining after seven days (168 hours) of exposure to temperatures ranging from 120°C to 180°C (248°F to 356°F).

6 0 – Se c t i on 2

2 ELECTRICAL CHARACTERISTICS OF INSULATION MATERIALS The electrical characteristics of cable insulation are just as important as the physical characteristics. After all, if a cable is mechanically durable but will not withstand the applied voltage, the cable no longer serves its intended purpose. Electrical characteristics include insulation resistance, insulation power factor, and dielectric constant. Basic electrical characteristics of cable insulation are discussed extensively in Section 4, Equipment Loading.

100% Hot Modulus

EP

XLPE HMWPE

20

75 90 130 Temperature (°C)

250

FIGURE 2.6: Comparative Hot Creep vs. Temperatures for Cable Insulation Materials. Adapted from ANSI/ICEA T-28-562.

Insulation Fabrication

All contemporary cables use extruded dielectric insulation. The manufacturing processes generally are similar for different insulation materials and different voltage classes. The most complex manufacturer’s process involves primary voltage cables that have not only extruded insulation but also extruded conductor shields and extruded insulation shields. Secondary cables have similar construction methods, but employ only an insulating layer or, in the case of “ruggedized” styles, possibly two layers. Many aspects of the manufacturing process are very important. Some of these are the following: • Purity of the insulation material, • Lack of voids in the insulation and shields, • Smoothness of the conductor shield and conductor, • Adhesion between the conductor shield and the insulation, • Cleanliness of the conductor shield-insulation interface, • Smoothness of the insulation outer surface, • Adhesion between the insulation and the insulation shield, • Cleanliness of the insulation–insulation shield interface, • Maintenance of uniform dimensions and concentricity along the cable, and

• Inclusion of agglomerates, gels, and ambers. Failure to adhere to any of these requirements at any point in the manufacturing process will lead to defective cable that is unsuitable for utility applications. MATERIAL HANDLING One of the most important requirements of cable manufacturing is cleanliness of the raw materials. The cable manufacturer receives insulating and shielding materials, particularly polyethylene compounds, as pellets. These pellets must be handled very carefully at both the cable plant and at the insulation manufacturing plant to ensure there is no contamination. Quality control tests that meet, or exceed, industry standards must be made on each batch of pellets to ensure cleanliness. In addition to normal quality control sampling, some plants use optical scanning to continuously sample pellets before they enter extruding equipment. This sampling is beneficial because contaminated pellets are rejected before being extruded into the cable. Resin suppliers now employ online pellet inspection devices. Some manufacturers inspect 100 percent of their product. From this, a new generation of XLPE and TR-XLPE materials that bear designations of extra clean, ultra clean, or


2 super clean has emerged. However, a precise definition of each designation based on per-unit volume contamination is not available, nor is a comparison between compound manufacturers. The opaque nature of EPR does not permit a similar determination of cleanliness. Cable manufacturers, in turn, have implemented materials-handling systems to prevent contamination during the course of manufacture. For example, Class 1000 clean rooms have been installed in most cable manufacturing plants and separate handling facilities for insulation and semiconductor materials have been implemented. Supersmooth semiconducting shields were first introduced in 1988, resulting from better dispersion of the acetylene carbon black in the polymer base. Better dispersed semiconducting shields provide for a much smoother interface between the insulation and the shields, leading to much longer service life. Utility acceptance of the cleaner and smoother compounds has been rapid, as most utilities specified these materials in 2004. Polyethylene manufacturers have focused on material purity, improvement in the compounding and process design, and quality assurance and quality control improvements. In addition, delivery systems have dramatically improved over the past 15 years. Using dedicated reactors, upgrading reactor clean out and defouling procedures, and monitoring each run for ambers and gels have improved manufacturing technology. Increasing the raw material cleanliness, filtrating all process air and water, and operating under a sealed loop strategy have helped to ensure a better product. In addition, handling systems now use gravity feed and dense phase, as well as dedicated, sealed rail cars in good condition. Polyethylene is manufactured by compound suppliers and shipped in pellet form to the cable manufacturers for extrusion onto the full-sized cable. Contamination is possible at any step along the way. Most manufacturers carry out optical pellet inspection, but usually only about two percent of the total amount of compound is inspected. Needless to say, many contaminants are missed, as recent statistics suggest that even the cleanest compound can contain contaminants above 12 mils, and these may be removed with

100 percent pellet inspection. Ideally, the pellet inspection should take place as close to the manufacturer’s extruder head as possible and not contribute to further contamination. Currently, pellet inspection devices are available for use at the cable manufacturer’s plant. The inspection devices remove loose contaminants and surface contaminants as well as pellets containing embedded contaminants. All models come with a self-enclosed air filtration system that provides a Class 1000 environment under a plastic curtain surrounding the unit. Removal of contaminants starts at three mils and optimizes at 12 mils. Inspection of EPR is more difficult, as the material is opaque. Tape inspection devices can also be used for surface inspection of extruded EPR sample tapes. Also available are inspection devices for gels in polymers and for small defects in interfaces. Although interfacial inspection does not occur until after the cable is manufactured, this latter device does provide an opportunity to identify, locate, and remove interfacial problems before shipment. Although inspection for contaminants is important, it is also important to eliminate all possible sources of contamination during the manufacturing of not only the insulation system but also the conductor and insulation shields. This means controlling the contact of possible contaminants, especially airborne dust particles, to raw insulation materials or to the cable during extrusion. Materials should be exposed as little as possible to the ambient air in the plant. In addition, cable interface surfaces should, similarly, have minimum possible exposure to an uncontrolled environment during the extrusion process. EXTRUSION AND CURING PROCESSES During cable manufacture, the various shields and insulating layers are extruded over the conductor. The raw material is melted and the liquid polymer is pumped into a die that applies a continuous and uniform layer around the conductor. The material is then cured at the proper temperature for the proper time. This process is repeated for various layers until the desired cable configuration is achieved.

6 2 – Se c t i on 2

2 Expediency and quality in cable manufacture can be achieved if the extrusion of different layers is performed simultaneously. The industry uses multiple simultaneous extrusion processes. Figure 2.7 shows the general layout of a cable extrusion line. The conductor enters the process from the pay-off reel. The conductor first passes through the extrusion heads, where the shields and insulation are applied. The cable then enters the curing tube, where the extruded polymers are cured at a temperature between 218°C (425°F) and 293°C (560°F). Pressure in the curing tubes is also maintained between 150 and 300 psi. This temperature and pressure is maintained long enough for cross-linking to take place in the insulation and/or shields. After curing, the cable enters a cooling zone, commonly referred to as a water bath or quenching. However, some new production lines use dry gas cooling. The methods used to cure and cool the cable during manufacture are the subject of much research. Older systems used high-pressure steam for curing, which led to higher water content (5,000 parts per million) within the insulation. It is suspected that this insulation water content may contribute to the development of water

trees within polyethylene. Some newer equipment uses dry nitrogen as a heat transfer agent in the curing tube, which eliminates insulation contact with water until it has solidified. The result is lower water content (200 ppm) in the insulation. The few cable production lines that use dry gas for both curing and cooling achieve even lower water content (50 ppm). The significance of the lower water content is still the subject of continuing investigation. It is believed that the very lowest water contents are maintained in service only if the cables are completely sealed from moisture. However, the industry has widely accepted the desirability of dry nitrogen gas curing, especially for polyethylene-based cables. Steam curing is the oldest cross-linking or vulcanizing method employed in any continuous vulcanizing (CV) plant. In steam curing, the freshly extruded cable passes down the center of a long vulcanizing tube filled with saturated steam at about 20 atmospheres (300 pounds per square inch gauge (psig)) pressure and temperature of about 215°C (419°F). The cured insulation is then cooled under pressure by cold water. Most EPRs are still made with steam curing in a CV catenary process.

Extrusion Area – Conductor Shield – Insulation – Insulation Shield

Curing Tube

Bare

Water Cooling

o duct Con

Insulated Cable

r Take-Up Reel

Pay-Off Reel

FIGURE 2.7: General Layout of a Cable Extrusion Line.


2 Conductor Shield

Insulation

Insulation Shield Insulated Conductor with

Bare Conductor

Conductor Shield Added

Insulated Conductor

Insulated Conductor

Insulation Shield

Second Pass

First Pass (a) 2 Pass or Dual-Tandem Method

Insulation Conductor Shield

Insulation Shield

Bare Conductor

Conductor Shield Added

Insulated Conductor with Insulation Shield

(b) 1 + 2 Triple-Tandem Method

Conductor Shield

Bare Conductor

Insulation Shield

Insulated Conductor with Insulation Shield

Insulation (c) 3-in-1 Triple Method

FIGURE 2.8: Typical Extrusion Methods.

are EPR users who gain little Dry curing, on the other advantage in the dry cure techhand, consists of an electriTrue triple-tandem nology. Most utilities that speccally heated tube filled with extrusion ify EPR insulation request high-purity nitrogen gas at steam curing or do not specify about 10 atmospheres (150 is preferred. a curing method at all. psig) pressure. The infrared For UD cable production, energy emitted by the hot the triple extrusion and the dry tubes is transferred to the cure technology with the catenary arrangement cable components. The cable surface temperais most common. ture can be as high as 300°C (572°F). The cured Extrusion heads are continuing to evolve. The cable is cooled by passing through a cooling simplest head configuration is the two-pass (or section containing water under the same presdual-tandem) process shown in Figure 2.8(a). A sure as the curing section to prevent void formadisadvantage of this arrangement is the open tion in the insulation. A dry cured insulation space between the application point for the concontains voids in the order of 100/mm3, 1 to 10 ductor shield and that for the insulation. The µm (micrometers) in size, whereas steam curing conductor shield surface can be contaminated generates voids of 105/mm3, 1 to 50 µm in size. by airborne particles. In addition, the cable Sixty percent of the investor-owned utilities now must be returned to a separate extrusion line specify dry curing. Of the remainder, 33 percent

6 4 – Se c t i on 2

2

Conductor Shields and Insulation Shields

for installation of the insulation shield. This is also an opportunity for contamination of a critical interface surface. A major improvement in cable extrusion is the development of the 1 + 2 triple-tandem arrangement. Here, the insulation and the insulation shield are extruded simultaneously as shown in Figure 2.8(b). Though there is still a chance for airborne contamination between the conductor shield head and the insulation/insulation shield head, there is no chance of contamination on the insulation surface. The latest extrusion configuration is the true triple-head unit. All three compounds are extruded simultaneously in one location in a

completely enclosed head [see Figure 2.8(c)]. Simultaneous extrusion eliminates the opportunity for contamination of any interface surface. Today, the preferred extrusion method is the triple crosshead line or the true triple-head extruder. This line features one common crosshead connecting three extruders, so that the insulation and the semiconductive shields are extruded simultaneously over the conductor. With its successful development and commercialization, the triple crosshead is now generally accepted in the industry because it minimizes the chance of damage and contamination at the shields and insulation interfaces. Most utilities now specify this extrusion method.

Conductor shields and insulation shields share the function of providing a uniform cylindrical surface on either side of the cable insulation, which allows the most uniform possible distribution of electrical stress. The conductor shield is particularly important in reducing stress concentrations caused by stranded conductors or imperfections on the conductor surface. The insulation shield eliminates nonuniform voltage gradients in the insulation caused by irregular contacts with grounded objects. By producing a more uniform electrical stress distribution, shields allow thinner insulation sections to be used with more predictable results. Before the general acceptance of extruded dielectric cables, the conductor shield and the insulation shield both usually consisted of carbonloaded cotton tape. These tapes improved the surface contour of conventional stranded conductors and were generally suitable for use with paper and rubber insulation compounds. With the advent of extruded polyethylene dielectrics, extruded shields gained favor. These could be applied at a lower cost and produce a more uniform surface than could semiconducting cloth tapes. This more uniform surface was particularly important for gaining cable reliability with polyethylene cable insulation. Present practice in extruded insulation cables uses extruded conductor and insulation shields almost exclusively. The preferred material is a semiconducting version of the material used for

the cable insulation. For instance, if the cable is insulated with cross-linked polyethylene, a semiconducting XLPE would be applied for both the conductor shield and the insulation shield. Similarly, cables insulated with ethylene propylene rubber could have a semiconducting EPR compound or a similar compound, such as ethyl vinyl acetate (EVA), as shielding material. This combination produces the greatest insulation system component compatibility. It is particularly important to have very similar coefficients of thermal expansion to minimize the generation of thermal stresses within the cable at extreme operating temperatures. Other combinations may be used if elasticity and tensile strength characteristics are compatible. Most manufacturers use EVA for these shields. CONDUCTOR SHIELD For maximum effectiveness, the conductor shield should be firmly bonded to the cable insulation to minimize voids at the interface between these two components. Because this zone has the highest electrical stresses in the cable and voids will produce insulation deterioration under electrical stress, it is particularly important to have the minimum number of possible voids in this location. The extruded conductor shield material should strip freely from the conductor without leaving residue to facilitate cable splicing. Otherwise, particles of semiconducting polymer might be left inside electrical connections that would


2 unacceptably impair conducavoid stress concentrations or tivity within the connections. corona-producing voids. Most electrochemical ANSI/ICEA Specification Therefore, an extruded semitrees begin at voids S-94-649-2000 allows the conconducting insulation shield is ductor shield/insulation interinstalled to evenly distribute or protrusions near face to have protrusions of electrical stresses. An extruded the conductor shield/ seven mils (0.18 mm) into the shield of a compatible material insulation interface. conductor shield and five mils will tightly adhere to the insu(0.127 mm) into the insulation, lation, even when the cable is if standard conductor shield bent or compressed. The thicknesses are used. Voids of shield will also remain in close up to three mils (0.076 mm) are allowed at this contact when the cable is operated at extremely interface. Research on cable failures has shown design temperatures. that most electrochemical trees begin at voids or The cable insulation shield must have unvaryprotrusions near the conductor shield/insulation ing conductivity characteristics to serve as an efinterface. Tree inception at these points is befective shield and produce a uniform, equipotencause of the extremely high electrical stresses in tial surface. In addition, the insulation shield these regions and because these irregularities must carry the cable capacitive currents between serve as stress amplifiers when they produce a the insulation shield interface and the grounded nonuniform electrical field. The cable industry metallic shield tape or conductors. This capabilihas, therefore, developed the concept of a suty is particularly important where a concentric persmooth conductor shield that produces an neutral configuration is used with conductors extruded conductor shield with a much more spaced around the cable circumference. Under uniform cylindrical surface. Protrusions into the these conditions, the insulation shield must carry cable insulation are reduced in size and quanticurrents transversely as well as radially. The conty. Typical interface irregularities for these imcentric neutral configuration makes the distance proved conductor shield materials are approxitraveled by the capacitive currents greater and mately one percent of the size found in convenmakes shield uniformity even more important tional shields. This significantly reduces the (see Figure 2.9). Current concentrations under number of tree initiation sites in the section of the concentric neutral strand also make it imporinsulation with the highest electrical stresses. Betant to keep shield resistivity low. cause reducing irregularities and voids in this The cable insulation shield must maintain area will yield longer cable life, the cable purgood contact with the insulation, yet be easy to chaser should strongly consider using the adremove during splicing. If the insulation shield is vanced conductor shield systems with improved firmly bonded to the insulation, this bond will smoothness. Such materials may be slightly produce ideal electrical properties, but it will more expensive, but the total life-cycle cost of make splicing much more difficult. Firm bonding the cable may be lower because the cable failwill require cutting the shield from the insulaure rate may be reduced. tion, which must be done very carefully to keep a uniform cylindrical outer surface on the insulation. Therefore, where splicing or terminations INSULATION SHIELD are required frequently, the insulation shield The cable insulation shield forms a cylindrical should be free-stripping. Removal should leave semiconducting surface on the outside of the inno residue on the insulation surface. The cable sulation, which is essential to avoid nonuniform specifier should note any special conditions of electrical stresses in the insulation. Although it is cable use, such as low splicing temperatures, theoretically possible to place a uniform conthat may require special stripping characteristics. ducting metallic shield directly outside the cable However, to maintain acceptable electrical perinsulation, it is impractical to achieve and mainformance, certain minimum stripping force will tain the continuous intimate contact required to

6 6 – Se c t i on 2

2 Concentric Neutral Strand

Semiconducting Insulation Shield

Capacitive Current Flow

Concentric Neutral Strand

Insulation

Strand Shield

Conductor

Semiconducting Insulation Shield

FIGURE 2.9: Capacitive and Dielectric Loss Current Flow in Insulation Shield.

TABLE 2.5: Insulation Shield Strippability Ratings. Minimum Removal Tension (lb.)

Maximum Removal Tension (lb.)

EPR

3

18

TR-XLPE

6

16

Discharge Resistant

0

16

Cable Insulation Type

be required. If the minimum bonding is not maintained, insulation-damaging corona might be produced at the insulation interface, especially in cable bends. Because good adherence is necessary for satisfactory electrical performance, the installation crews may have to warm the insulation shield to an acceptable temperature for splicing and termination. If such conditions are frequently encountered, the cable specifier may wish to cite special conditions in the cable specification and call for special low-temperature stripping tests. However, the specifier should always remember that long-term performance of the cable is the most important criterion and special installation techniques may be needed under low-temperature conditions. Pending RUS 1728F-U1 specifications for primary cables call for minimum and maximum tension ratings for “strippability” of insulation shields, as shown in Table 2.5. Slightly different limits for stripping tension are used in the sample cable specification contained in Appendix E. If a cable system is going to be used in an installation requiring especially high reliability and few splices or terminations, the specifier may use a firmly bonded extruded insulation shield. Doing so will produce optimum electrical performance. If long cable pulls are used, less extra labor will be needed to make splices. However, before starting installations of this type, crews must be specially trained and proper tools must be obtained to make satisfactory splices. Firmly bonded insulation shields should never be used on underground residential systems where cables are frequently terminated. CONCENTRIC NEUTRALS AND CONDUCTIVE METALLIC SHIELDS Shielded cable systems require not only a semiconducting insulation shield but also a conductive metal shield to function properly. The metal shield is in intimate contact with the semiconducting insulation shield. The major functions of the conductive metal shield are the following: • To serve as a grounding means for the semiconducting insulation shield to keep all sections at constant potential,


2 • To serve as a path for to one-half (reduced neutral) currents generated by of the phase conductivity. This Defects in the shield capacitive coupling enables the cable to function system will cause between the central without a separate neutral conconductor and the ductor. Mechanical as well as cable failures. system neutral or the electrical considerations genersurrounding earth, ally mandate that concentric • To serve as an interceptor neutral conductors be copper, of system fault currents in case of a even if the central cable conductor is aluminum. dielectric failure, Table 2.6 shows concentric neutral sizes often • To provide a grounded metallic object used on distribution cables. between the energized conductor and the Full-capacity concentric neutrals are most cable exterior, and often used on smaller cables that are applied in • To serve as a system neutral (in some cases). single-phase circuits. Having full conductivity in the neutral reduces circuit voltage drop. The sysAll these functions are extremely important. tem neutral-to-earth voltage under both normal loads and fault conditions is reduced as well. The conductive shield’s failure to properly perReduced neutral capacities are most often form any of these functions will lead to either a used on three-phase circuits, particularly in the cable failure or a malfunction of the electric dislarger conductor sizes. Doing so is feasible and tribution system in which the cable is installed. desirable because of the following: A wide variety of conductive shield configurations have been developed for use on cable systems. Examples of typical shield configurations • In a three-phase circuit, three neutrals are are given below. connected in parallel, which reduces the cross section required to produce a full-capacity system neutral to one-third on each cable. CONCENTRIC NEUTRALS • In a three-phase system, system neutral return Concentric neutral conductors serve a dual role current should be near zero, thereby reducing as a conductive cable shield and a circuit neuthe cross-sectional area required to maintain tral. To fulfill this second function, the shield low system losses and neutral-to-earth volt(neutral) has a much larger cross section than is ages under reasonably balanced conditions. typical with flat tape, drain wire, or L.C. shield • In a three-phase cable system with interconconfigurations. Typical concentric neutral cables nected neutrals, losses in the cable neutral are will have a neutral conductivity equal to that of caused by circulating currents. All other factors the central conductor (full neutral), or one-third

TABLE 2.6: Concentric Neutral Configurations for Common Aluminum Cables. Typical Neutral Configuration Conductor Size

Full Capacity

One-Third Capacity

One-Sixth Capacity One-Twelfth Capacity

#2 AWG Aluminum

10 × 14 AWG

6 × 14 AWG

N/A

N/A

1/0 AWG Aluminum

16 × 14 AWG

6 × 14 AWG

N/A

N/A

4/0 AWG Aluminum

13 × 10 AWG

11 × 14 AWG

N/A

N/A

350 kcmil Aluminum

20 × 10 AWG

18 × 14 AWG

14 × 16 AWG

N/A

500 kcmil Aluminum

N/A

16 × 12 AWG

20 × 16 AWG

10 × 16 AWG

750 kcmil Aluminum

N/A

20 × 9 AWG

30 × 16 AWG

10 × 14 AWG

6 8 – Se c t i on 2

2 being equal, these losses are lower where there is less neutral conductivity. A cable with a one-third neutral has 53 percent of the losses of a cable with a full-capacity neutral if the cables are spaced 7.5 inches center to center. The circuit ampacity of full-neutral cables in three-phase circuits is also reduced because of these shield losses. This problem is significant in larger cable sizes, particularly if the cables are not closely grouped. For instance, on a 350-kcmil circuit carrying 390 amperes, the losses on a circuit with 7.5-inch cable spacing will drop from 12 kW/1,000 feet to 6.4 kW/1,000 feet if a one-third neutral is used instead of a full-capacity neutral. Elevated losses and reduced ampacities are not generally a problem on three-phase circuits of 1/0 AWG aluminum or smaller if the cables are grouped in a single trench. Additional information on circuit ampacity rating for various neutral configurations is given in Section 4. Longitudinally Corrugated Shield The L.C. shield has been developed as a way to provide greater conductivity in larger cables. The shield generally consists of a copper sheet that is installed with its major axis parallel to that of the cable. The sheet is then folded around the cable and sealed to itself on the opposite side. Circumferential corrugations are fabricated in the resulting tube to add flexibility and ensure that the shield will uniformly bend with the cable. The seal applied between the two sides of the copper is usually an adhesive elastomer. The tube generally does not have a metal-to-metal connection with the cable insulation shield at this point because allowance must be made for the cable insulation to thermally expand during operation at elevated temperatures. Not only is the temperature change higher in the insulation than it is in the cable shield, but all dielectrics have a substantially higher coefficient of thermal expansion than that of copper. Because the metallic shield must have good contact with the semiconducting insulation shield to function effectively, a tight fit must be maintained at all times. Therefore, the insulation expansion is accommodated by flexibility in the elastomeric seal. The return of the shield to intimate contact as the cable cools is assisted by the external insulating jacket.

L.C. shields are commonly available in five-mil thickness but, for applications requiring additional fault current capability, shield thicknesses of eight or 10 mils can be furnished. However, L.C. shields should be sized to carry expected system neutral currents. Use of L.C. shields as the system neutral will require evaluation of available system fault currents and protective system clearing times. Accessories such as shield (neutral) bonding clamps must also be carefully evaluated for long-term continuous current and fault current capacity. The L.C. shield does provide a limited degree of resistance to water vapor transmission. It is clearly superior to concentric neutral configurations for water vapor transmission. It is somewhat better than helically applied copper tape shields because the length of the straight joint is less than the helical joint. Moreover, the elastomer at the lap point does provide a better seal, although, under static pressure, the elastomeric seal cannot be depended on to prevent moisture from migrating into the cable insulation. Flat Copper Tape This is perhaps the oldest conductive shield configuration. The tape generally consists of a five-mil (0.005-inch) thick copper tape helically applied over the semiconducting insulation shield. The tape is usually installed with a 12.5 percent overlap. Tape shields may be fabricated from bare copper or may be tinned copper. Because of the small cross section, the conductivity of flat copper tape shields is relatively low compared with the central cable conductor. Equation 2.2 gives the effective cross-sectional area of an overlapped tape.

Equation 2.2 A = 4bdm ×

W 2(W – L)

where: A = Cross-sectional area, in cmils b = Tape thickness, in mils dm = Mean diameter, in mils W = Width of tape, in mils L = Overlap of tape, in mils


2 CONCENTRIC NEUTRAL CONFIGURATIONS As experience has been gained with underground installations under a variety of conditions, the utility industry has developed several specialized variations of the basic concentric neutral configurations. Each of these arrangements has an advantage for a particular set of installation conditions. Bare Concentric Neutral The first widely accepted concentric neutral cables were of a bare concentric neutral (BCN) configuration. In this design, the concentric neutral strands were laid over the semiconducting insulation shield and no jacket was applied. When the cable was directly buried, this arrangement had the advantage of exposing the concentric neutral conductors to the surrounding soil. The result was a very effective ground, especially where soil resistivity was low. The low resistance between neutral and earth meant more of the system neutral current could return to the source by way of the earth, thereby reducing current in the concentric neutral and circuit voltage drop. Furthermore, the low resistance between the neutral and earth reduced neutral-to-earth voltages during both normal operations and fault conditions. The bare concentric neutral is also considered the best possible arrangement for personnel safety in case of a dig-in. The neutral size ensures the ability to adequately conduct fault currents until protective devices operate. The higher conductivity of the concentric neutral will produce lower voltages on the neutral at the fault location. The low resistance between the neutral and earth will significantly reduce the touch potential at the dig-in site. Most important, the concentric neutral physical arrangement ensures the object penetrating the cable will have established a good neutral connection before contacting the energized center conductor. In light of all the advantages of BCN cables, it is unfortunate that there are major durability problems with this design under many installation conditions. These problems are all related to corrosion of the exposed cable neutral. In many cases, the neutral had a significantly reduced cross section after only a few years of service. In other cases, the neutral was completely corroded and the only neutral current path was through

ground rods. This condition was totally unsatisfactory from the standpoints of system safety and reliability. Therefore, the use of BCN cable has been discontinued except in very special conditions. Jacketed Concentric Neutral Because of the very serious problems experienced with BCN cables, the electric utility industry began using the jacketed concentric neutral (JCN) configuration. This configuration has most of the major advantages of the BCN design except for continuous contact of the neutral with earth. The jacketed configuration reduces access of moisture and corrosive agents to the neutral. Insulating jackets also interrupt the flow of galvanic corrosion currents between the neutral and other metallic objects. JCN design has achieved wide acceptance as a solution to the concentric neutral corrosion problem. However, the cooperative engineer must give special attention to system grounding if jacketed cables are used. Cable identification also acquires additional importance, as jacketed cables are approximately the same dimension and general appearance as many communication cables and water lines. See Section 5 in the Design Manual for detailed information on system grounding. Flat-Strap Concentric Neutrals Flat-strap concentric neutrals, not to be confused with flat-tape metallic shields, consist of helically applied flat copper straps. These straps are about 0.020 to 0.025 inches (20 to 25 mils) thick and about 0.150 to 0.175 inches wide. The straps are applied so they abut each other and provide 90 percent metallic coverage over the outside of the cable. Conductivity of flat-strap neutrals is generally equal to that of the energized conductor. Flat-strap concentric neutrals have found greatest acceptance in areas where rodents damage direct-buried cables. The complete metallic coverage on a cable was originally believed to lessen damage from gophers. However, using this type of cable to lessen rodent damage has had mixed results. Recent research shows that rodent damage is more effectively limited by increasing the diameter of the object. Therefore, flat-strap neutrals should not be depended on to prevent rodent damage.

7 0 – Se c t i on 2

2 Experience with low-voltage insulated cables has shown Always specify that aluminum conductors can copper for concentric be extremely susceptible to corrosion, even if they are inneutral conductors. sulated from the surrounding environment. Because cable jackets are not absolutely moisture proof, even an encapsulated aluminum neutral conductor may be subject to long-term Concentric Neutral Materials deterioration from moisture migration. It is unOther Than Copper wise to consider aluminum neutral conductors The predominant material in concentric neutrals for primary cables, even in a jacketed configurahas always been copper. For many years, the tion, when the only advantage to be gained is generally accepted wire for bare concentric neuslight savings in initial material cost. trals was copper with a tin or tin-lead alloy coatAnother approach that was used for a limited ing. As experience has been gained with a wide time to try to solve the bare concentric neutral variety of materials, engineers have determined corrosion problem was the use of a composite that the coating of the copper concentric neutral copper/steel conductor. The particular configuraconductors was not necessary and, in some castion used a copper center core for conductivity, es, actually led to higher corrosion rates. It is with a heavy steel coating completely surroundgenerally believed that, in the early days of coning the copper. For durability during periods of centric neutral cable manufacture, tinned copper atmospheric exposure, the steel was galvanized. concentric neutrals gained wide acceptance beThis cross-sectional arrangement offered the defcause most flat-tape metallic shields were tinned inite advantage of having steel exposed to the on jacketed cables. In some cases, that was a earth in the direct-buried cables instead of copholdover from cables on which butyl rubber inper. The exposed steel greatly simplified the apsulation was used and tinning was needed to plication of cathodic protection systems to the avoid corrosion. Also, tinned copper was used neutral. However, the conductor used in this on earlier cables because of the prevalence of neutral construction did carry a premium price. soldered connections, and the coated copper faUtilities also experienced difficulty in applying cilitated soldering of these thin shields. Because this cable to existing systems that already had concentric neutral cables never employ soldered extensive exposure of bare copper concentric connections and butyl rubber is no longer used neutrals. Systems containing this cable configufor insulation, the need for coating neutral wires ration required sacrificial anodes or impressed has disappeared. Bare copper wires are now voltage rectifiers applied to provide protection uniformly accepted as the preferred material for to the neutral. For additional information on the concentric neutrals, whether bare or jacketed. principles of cathodic protection, see Section 7. During the mid-1970s, a few utilities briefly experimented with aluminum concentric neutral cables. These were applied in a bare configuraCABLE JACKET tion. Although some laboratory studies showed In most cables, the cable jacket is the outermost that the aluminum neutrals would resist many layer of material that serves as a barrier to moistypes of soil-induced corrosion, field experience ture and mechanical damage. Therefore, it is improved quite the opposite. The very complex inportant to optimize the design and materials of teractions present on an interconnected neutral the jacket to obtain maximum performance in passing through a variety of soils led to early these important areas. failure of these cables. It became obvious that For many years, all power cable designs includaluminum should never be used as an exposed ed a jacket. However, with the advent of the exconcentric neutral in direct-buried or conduit tensive underground residential programs, electric cable installations. utilities began installing bare concentric neutral Flat-strap neutral cables should be jacketed. The thickness of the flat strap is less than the diameter of the neutral wires. Therefore, the complete cable diameter will be less. This is an advantage where space is limited.


2 cables. This design eliminated the cable jacket so that the BCN could establish conductive contact with the earth in a direct-buried installation. Engineers eventually learned that the accelerated failure rate of UD cable was largely caused by cable moisture and/or concentric neutral corrosion. Both of these factors were able to strongly influence UD cable life because of the lack of a high-quality cable jacket. It is worth noting that, although U.S. utilities installed BCN UD cables, European and Japanese utilities continued to install only jacketed cables. These utilities have experienced much higher distribution system cable reliability than has been typical in the United States. Recognizing this, the U.S. electric utility industry now mainly uses jacketed cables. These may be conventional power cables with flat-tape or drainwire shields, or they may be JCN cables. Jackets can be either insulating or semiconducting. Under any circumstances, the jacket material is very important. Desirable characteristics include abrasion resistance, flexibility, and low moisture permeability. If cable is being pulled

into a conduit system, a low coefficient of friction with the conduit material is desirable. Today, most utilities specify an outer jacket. A wide variety of chemical components have been used successfully for cable jacketing. The material most desirable for jacketing is linear low-density polyethylene (LLDPE). This material has the best balance of properties for use on underground utility cables. Table 2.7 shows a comparison of important properties of various compounds. The table shows that polyethylene is preferable in almost all categories except fire resistance. In directburied applications and outdoor conduit installations, this compromise is acceptable. Low chlorine content is an advantage because hydrogen chloride may result from these compounds at the emergency operating temperature of 130°C (266°F). This gas, particularly in conjunction with surrounding moisture, will be detrimental to XLPE and EPR insulating compounds as well as copper neutrals or other metallic shield materials.

TABLE 2.7: Comparison of Jacketing Material Test Data.

Polyethylene (PE)

Semiconducting Polyethylene*

Polyvinyl Chloride (PVC)

2,730 620

1,700 450

1,920 350

Moisture Transmission 7 Days in 70°C (158°F) Water • Grams/m2/24 hours

0.8

1.5

>10

Flame Resistance 20 Min. at 70,000 Btu/Hr • Cable Tray Fire Test

Fail

Fail

Fail**

Low Temperature Properties Cold Bend Test • Temperature Passed (°C)

-40

-50

-10

0

0

22.0

350

350

160

Physical Properties • Tensile Strength (psi) • Elongation

Chlorine Content (%) Thermal Stability • Initial Temperature of Decomposition (°C)

* Based on Union Carbide 7708. ** PVC can be specially compounded to pass the Cable Tray Fire Test.

7 2 – Se c t i on 2

2 TABLE 2.8: Static Coefficient of Friction for Jacketing Materials in PVC Conduit. Polyvinyl Chloride 0.69

Cross-Linked High-Molecular-Weight Polyethylene Polyethylene (XLPE) (HMWPE) 0.75

0.42

Linear Low-Density Polyethylene (LLDPE) 0.42

Another important characteristic of jacketing materials is the coefficient of friction in common pulling situations. Table 2.8 shows the static coefficient of friction of various jacket materials in PVC conduit. Jacket materials used on utility systems should always be sunlight-resistant. Very few installed utility cables have no part of the cable ever exposed to sunlight. Therefore, most cable jacketing compounds will be colored black to eliminate sunlight penetration and, thereby, enhance the natural durability of the basic jacket compound. Jacket Configurations There are two main physical arrangements for cable jackets. The first significant jacket configuration is the encapsulating jacket. This arrangement surrounds the concentric neutral conductors with the jacketing compound. The jacket is extruded directly over the concentric neutral strands. The jacket material fills all areas between concentric neutral strands and establishes close contact with the semiconducting insulation shield. Adequate jacket thickness is placed over the outside of the strands to minimize the chance of strand exposure during installation. The advantage of this encapsulated neutral design is that no spaces exist between neutral strands to allow movement of moisture along the cable. Therefore, any penetration will allow moisture in only one small spot, and probably will expose only one neutral strand at this location. Limiting moisture exposure to only one strand of the concentric neutral will reduce the potential for loss of neutral continuity. The second jacket configuration is an extruded jacket that overlays the metallic shield or concentric neutral. In this arrangement, the jacket is often separated from the tape shield,

drain wires, or concentric neutral by a nonadhering tape. This tape keeps the two layers entirely separate. Where drain wires or concentric neutrals are used under the jacket, this method leaves an annular (ring-shaped) space between the semiconducting insulation shield and the outside jacket. Although this space does contain the metallic wire shield, the spaces between strands become a reservoir for moisture that may enter the jacket through gradual absorption, manufacturing defects, or installation-induced damage. This space also provides an excellent path for migration of moisture along the length of the cable. This moisture is extremely detrimental to the cable by its promotion of electrochemical treeing in the insulation. This moisture also facilitates corrosion attacks on metallic shield strands. Although this jacket configuration is satisfactory for use with metal tape shields, it should not be used with concentric neutral cables that will frequently be exposed to moisture. Semiconducting Jackets The use of insulating jackets on direct-buried cable improves most performance characteristics, with one major exception. Use of an insulating jacket deprives the concentric neutral of its conductive contact with the surrounding earth, thereby relegating all system neutral grounding to driven rods or other electrodes installed along the circuit route. To improve cable grounding with its attendant benefits, a semiconducting cable jacket was introduced. The jacket consists of a semiconducting compound that is extruded in an encapsulating jacket (embedded neutral) configuration. The constructed cable has a radial resistivity of less than 100 meter-ohms and is, therefore, comparable to the conductivity of most soils. This ensures neutral-to-earth current transfer comparable to that of a BCN design. The improvement of conductivity provided by semiconducting jackets between the concentric neutral and the surrounding earth is a significant improvement in overall UD system design. However, there are some disadvantages to the semiconducting jackets. These disadvantages are principally associated with the greater moisture transmission rate of the semiconducting polyethylene compound. The first semiconducting jackets


2 had moisture transmission rates approximately 12 times that of LLDPE. At that level, moisture could penetrate the jacket and collect adjacent to the concentric neutral strands. There the moisture had the potential to serve as an electrolyte, forming a galvanic cell between the copper neutral and the carbon in the semiconducting jacket. This could result in deterioration of the neutral. Another aspect of the semiconducting jacketed cable design concerns the possibility of mechanical damage to the jacket during installation, exposing the neutral conductors directly to the soil. In this case, there is the potential for the galvanic attack to be more severe because the ratio of exposed surface areas of the carbon to copper is much greater. There also previously existed concern that the galvanic cell existing between the semiconducting jacket and interconnected subterranean steel objects might be detrimental to the steel. Examples of such objects are anchors, telephone pedestals, and water piping. Tests have been conducted by NEETRAC to demonstrate that accelerated

Printed Data Clear Space

3H

3H

3H

H

Symbol for Communication Cable Printed Data Clear Space

3H

3H

3H

H

Symbol for Supply Cable H = Height of printed characters; determined by cable manufacturer

FIGURE 2.10: Cable Identification Markings. Source: ANSI/IEEE C2 (NESC).

deterioration of interconnected steel is not a significant problem. In summary, utilities should carefully consider all aspects of the system performance before installing semiconducting jackets on direct-buried cable. Though the advantage of lower system resistance to remote earth is desirable and immediate, the potential subtle negative effects are longterm and may have an effect on the useful life of the cable. The utility should consider the particular circumstances of the proposed installation conditions and weigh the merits of each cable jacket option. Cable Jacket Marking External marking of jacketed cable is necessary and serves three major purposes. The first is to provide information on the cable’s characteristics. The conductor size, type and thickness of insulation, and voltage rating must be included. The manufacturer’s name and the year of manufacture must also be included. All these markings must be durable and indented into (or embossed onto) the jacket. The second purpose is to make individual cable identification and accounting easier by applying sequential footage markers to the outside of the jacket. These markings should be applied with the general cable information listed above. These markings, along with reel label data, tell the installer how much cable remains on a reel. The sequential footage markings also help identify a particular cable that may be exposed in the midpoint of a multiconductor run. The third important purpose of external markings is to identify JCN cables as high-voltage cables. If unmarked, JCN cables are indistinguishable from jacketed communications cables. This difference must be made clear to personnel of all utilities. Previous efforts have involved the application of three red stripes in the cable surface. Other schemes have used various patterns of raised ribs on the cable surface. To assist in solving this problem, the NESC (ANSI Standard C2) requires that all electric supply cables have a standard lightning bolt symbol included in the external marking. This symbol is illustrated in Figure 2.10. As with all other exterior markings, it must be durable and indented into (or embossed onto) the cable surface.

7 4 – Se c t i on 2

2 Cable Specification and Purchasing

Acquisition of satisfactory cable starts with preparing an adequate specification document that fully describes the cable needed. As the preceding topics in this section have shown, there are many options from which to choose. The specification must describe the following: • The cable that will best fulfill system requirements, • The quality control tests that are expected during and after manufacture, and • The packaging and shipping methods to be used. In short, all items of importance to the purchaser must be described either directly or through reference to other industry-standard specifications. Reference to industry-standard specifications can greatly simplify the specification-writing process for both the purchaser and the supplier. Perhaps the most notable examples of widely accepted U.S. cable specifications are those prepared under the auspices of the ANSI/ICEA. ANSI/ICEA Specification S-94-649 covers cables insulated with thermoplastic, cross-linked, and ethylene propylene rubber. This specification is for shielded cables rated five through 46 kV. Within these specifications, there are references to various detailed specifications, such as National Equipment Manufacturers Association and ASTM specifications. Another major specification that affects rural electric cooperatives is RUS Bulletin 1728F-U1. The RUS U1 specification makes extensive reference to ANSI/ICEA Specification S-94-649-2000. U1 is oriented specifically to UD cables up to 35 kV and optional semiconducting outer jackets. As of the writing of this manual, this RUS Bulletin 1728F-U1 is still pending final approval. Compliance with these commonly accepted electric industry specifications assures the purchaser that the manufacturers will be familiar with the general requirements and should have designs and quality control procedures in place to meet the purchaser’s needs. SAMPLE CABLE SPECIFICATIONS The first step when buying any cable is to determine the specific requirements of the project

being considered. These requirements can range from routine cable purchases for use in small-capacity, single-phase extensions to specialized cables for substation feeder exits, underwater installations, or other unusual applications. Appendix E contains sample specifications for primary cable. Appendix E addresses cables with both EPR and TR-XLPE insulation. These specifications incorporate many of the features that have been discussed and recommended in this manual. Appendix E shows features to include in specifications for the purchase of single-conductor, medium-voltage cable suitable for rural systems. These specifications are compatible with, and in some cases exceed, the requirements of pending RUS Bulletin 1728F-U1. Because these are general specifications, they are particularly oriented toward the routine cable purchase. These specifications may not include special features needed in a particular project. Therefore, the engineer must closely review these specifications and change them as needed to meet any unusual requirements of a particular project. Appendix C is a sample specification for secondary single-conductor and triplex cables. Three types of insulation are included: standard crosslinked polyethylene, ruggedized cross-linked polyethylene, and self-sealing insulated cables. Because many secondary cable failures are caused by insulation cuts during installation, these tougher insulations are required for reliability. The use of ruggedized secondary cable is recommended. Self-sealing secondary cables contain a viscous material between the outer layer of conductor strands and the inner surface of the insulation. When the insulation is disrupted, the viscous insulating material flows into the cut and restores the integrity of the insulation. This stops the entrance of moisture into the cable and arrests the progress of the typical secondary cable failure. TYPICAL SPECIAL REQUIREMENTS There are certain areas in which purchasers commonly change the specifications to meet their particular needs. Neutral Size One item that affects both the initial and the operating costs of an underground cable is the concentric neutral conductivity. If the neutral selected


2 for three-phase installations is too large, both the initial cost and the circulating current losses will be higher. However, on single-phase installations, a larger concentric neutral is needed to carry the neutral return current that may be near the magnitudes of the current in the energized conductor. On single-phase installations, a reduced neutral capacity could produce higher neutral-to-earth voltages and higher losses because of the lower conductivity of the neutral conductor. Conceivably, the reduced-capacity neutral could even be thermally overloaded as the cable approaches normal rated capacity. For these reasons, RUS requires a full-capacity neutral in single-phase installations and allows a one-third (or greater) capacity concentric neutral on three-phase cable installations. This approach ensures that there will be concentric neutral conductivity at least equal to the phase conductor conductivity in both single-phase and threephase installations. The cooperative engineer should consider the typical use of the cable that is being bought when deciding whether to use full-capacity or reduced-capacity neutrals. Length Each purchaser will have different requirements for the length of cable on reels to use on routine installations. Requirements will vary with terrain, the type of equipment used to install cables, and the typical distance between termination points. The cables should be bought in the longest lengths practical for the field crews to use so as to leave less scrap at the reel ends. Constraining factors will be the width and diameter of reels that the cable transport and installation equipment can accommodate. The cooperative engineer must also consider the weight of the full reel when deciding on the standard reel size. As with all other aspects, it is helpful to select the same maximum reel sizes that other cooperatives choose, especially if there is a group purchase arrangement. Doing so makes stocking easier for manufacturers and distributors and consequently reduces the cost for the cooperative. Cold Weather Bending Utilities operating underground systems in cold climates have experienced a variety of flexibility

problems with cables caused by the low temperatures. To lessen these problems, the specifier can insert a section requiring a cold bend qualification test. This test will indicate the probability the cable will fail during bending or movement at low temperatures. It is not a measure of cable flexibility. In most cases where the cable operating temperature is always above -17°C (0°F), cable bending problems are not significant. Feeder Cable Shielding Section 4 of this manual shows that high-capacity three-phase cable installations incur much higher losses when high-conductivity concentric neutrals are used. Induced currents that circulate between the neutrals of the three phases cause these losses. Lower conductivity neutral/shield arrangements reduce these losses. Such arrangements not only can reduce the economic loss associated with circulating currents, but also can increase cable ampacity by cutting the amount of heat generated in the neutral/shield. Substation exits or other large feeders generally have better load balance with lower neutral currents. Therefore, reduced concentric neutrals will have adequate thermal capacity, especially if they are supplemented by a separate neutral conductor. Where a high-capacity feeder is being installed, the engineer should give particular attention to the size of the neutral and/or shield specified on the cable. The engineer must also check the magnitude and duration of fault currents on the system when selecting a particular neutral/shield arrangement. Fault current duration is usually not a problem on 200-amp-class single-phase circuits because full-capacity neutrals are used and circuit reclosing is not a factor. However, the other extreme is substation feeder exit cables where there is a desire to reduce neutral capacity to minimize circulating current losses and increase ampacity. In these locations, the fault currents are higher, overcurrent protective devices operate more slowly, and reclosing is often used. All these elements contribute to higher neutral/shield temperatures under cable fault conditions. The neutral/shield component of underground substation feeder exit cables and express feeders must also carry fault currents for all down-line faults. An additional neutral conductor located in the same

7 6 – Se c t i on 2

2 trench or conduit with the insulated cables can supplement this capability. The engineer should pay particular attention to this set of conditions when selecting a reduced neutral size. CABLE PURCHASING PRACTICES Vendor Prequalification Because cable is one of the keys to a reliable and cost-effective underground distribution system and some types of cable defects are not obvious at the time of manufacture and will be recognized only years later, all cable needs to be manufactured by reliable producers. It is in the cooperative’s best interest to review the qualifications of vendors and select those that have a proven capability to produce a high-quality insulated conductor. Prequalification of vendors ensures that all parties quoting on a cable order have a proven ability to produce a high-quality cable meeting a particular specification. Prequalification avoids situations in which a vendor with questionable qualifications submits an unrealistically low price. Under these circumstances, the utility is typically required to honor the bid, which may lead to additional long-term cost through premature cable failure. It is only logical that if most of the utility industry is carefully prequalifying vendors, those found unqualified by others will have lower prices and better lead times because of lower demand for their products. This possibility makes it even more important to participate in an effective vendor prequalification program. Group Purchase One way to simultaneously improve cable prices and quality is to engage in group purchasing of cable. This practice has several advantages to both the vendor and the cooperative. Larger quantities (more than 50,000 feet) often lead to better overall quality control. During the initial part of a cable manufacturing run, larger orders mean that the front and tail ends of a particular run can be scrapped. This additional cost for nonqualifying material is then spread over a larger order, thereby reducing the unit price. Active quality control is an important part of any utility purchasing program. This quality control should include factory visits during major

cable purchases to review factory production and testing procedures. To be effective, an individual familiar with cable production and testing methods must be present. Because the expense of this observation is essentially the same for large or small orders, large orders greatly reduce the incremental unit cost for observation. Moreover, with group purchasing, there is a greater chance that a staff engineer from one member of the group will have (or be able to develop) the expertise necessary to effectively perform this function. Group purchasing and larger orders will always lead to a lower unit price. Because all the cable bought under a group plan will be according to a single specification and of the same construction, the manufacturer can achieve economies through the following: • Volume purchases of required material; • Longer, more efficient runs in wire drawing operation; • Longer, more efficient runs in cable extrusion operation; and • Wider distribution of fixed costs associated with a single order. Group purchasing of large cable quantities has a minimum effect on delivery practices. Manufacturers will usually ship parts of the larger order to destinations specified by group members at no extra cost. In some cases, groups have negotiated warehousing arrangements with manufacturers for release of cable on a designated schedule throughout a year. This arrangement reduces the cash flow burden on the cooperative. It also gives the manufacturer additional flexibility by allowing the major production runs to be scheduled at more convenient times. Another advantage to group purchasing on a standardized specification is the feasibility of having a single distribution point where the group maintains a cable stock. The ability to receive large orders coupled with reduced warehouse space requirements at the individual group members’ sites may make this approach reasonable in some cases. This option is particularly attractive when group purchase and stocking of other utility materials is also practiced.


2 Cable Acceptance

After a cooperative has analyzed its cable needs, written a comprehensive specification, and followed good purchasing procedures, one critical step remains before installation can begin. This step is the acceptance and inspection of the cable delivered by the manufacturer. Cable acceptance involves several simple and inexpensive steps that can yield big dividends. The cooperative engineer must follow these steps to make sure that a quality product is delivered to installation crews.

STEP 3. CHECK DIMENSIONAL TOLERANCE Make a simple measurement of basic cable dimensions on one reel of each cable size in a shipment to confirm that labeling is correct. Measure these dimensions:

STEP 1. VISUALLY INSPECT FOR SHIPMENT DAMAGE Visually inspect cable reels for any damage that may have occurred in transit. Signs of possible damage include impressions or nicks on the outside layer of cable or the reel lagging. If possible, this inspection should take place while reels are still on the delivery vehicle.

Section 11, Cable Testing, gives further information on allowable dimensional tolerances.

STEP 2. CHECK TAGS Visually check each reel to determine that it has proper tags and labels as described in the specifications. Make sure that information on the reel tags agrees with purchase-order information. For example, be sure that wire size, insulation thickness, neutral configuration, and jacket description all conform to the specifications and purchase order. Cable length should fall within the bounds described by the purchase order. If cable was ordered cut to specific lengths, the engineer should check the tag and sequential jacket markings (if available) to be sure that enough length is available for the required run.


Cable systems are one of the most important parts of any underground system. Special care must be used in selecting both primary and secondary cables. Some important points follow: 1. JCN cable must be used for most underground installations. Insulating jackets are preferred. 2. Aluminum central conductors are the economical choice for most underground situations. 3. Solid conductors up to No. 2/0 AWG may be used to eliminate longitudinal moisture migration.

• Conductor size and stranding, • Insulation thickness, • Concentric neutral wire size and number of strands, and • Jacket thickness.

STEP 4. CONDUCT CABLE ACCEPTANCE TESTING Once on each order or once for each 50,000 feet of cable, the cooperative should conduct a complete set of dimensional and electrical performance tests on the cable to make sure it complies with the purchase specifications and referenced industry standards. These tests include the following: • • • •

Conductor shield resistivity test; Insulation shield resistivity test; Dimensional analysis of all components; Microscopic examination for voids, contaminants, and shield interface protrusions; and • Insulation shield stripping test. An outside laboratory will need to help with these tests. Section 11 gives additional information on these tests.

4. All stranded conductors should have strand filling in interstices to eliminate longitudinal moisture migration. 5. Modern TR-XLPE or EPR cables offer reliability superior to that of earlier cables of HMWPE or XLPE. 6. Vendor quality control and manufacturing cleanliness are essential to the production of reliable cable. 7. In heavily loaded three-phase circuits, reduced neutrals will cut losses caused by circulating neutral currents. Reduced neutrals

7 8 – Se c t i on 2

2 will also increase circuit ampacity, particularly where phases are separated. 8. A comprehensive cable specification must be used and received materials inspected for compliance.

9. Initial cost, cost of dielectric losses, and cable life expectancy must be evaluated when making purchasing decisions.

Underground System Section a l iz i n g – 7 9

3 In This Section:

Underground System Sectionalizing

General Sectionalizing Philosophy Overcurrent Protection of Cable System Effect of Inrush Current on Sectionalizing Devices

General Sectionalizing Philosophy

The final design and continuous reliable performance of an electrical distribution system depend on many engineering elements. Protective device coordination, overcurrent protection, overvoltage protection, voltage regulation, and service continuity are just a few of the elements that are incorporated. This section addresses the coordination of overcurrent protective devices in underground distribution systems and the coordination of these protective devices with protective devices on interconnected overhead portions of the system. This section is not intended to provide a comprehensive procedure for planning and operating a protection program. Furthermore, the procedure for calculating system fault current is beyond the scope of this section. An excellent reference for designing protection systems and calculating faults is Electrical Distribution System Protection by Cooper Power Systems (1990). Many excellent computer programs are also available for fault current calculation. PURPOSE OF SECTIONALIZING Limit Magnitude of Damage and Injury Short-circuit currents subject a system to both mechanical and thermal stress. Mechanical stress begins at the same time as the initiation of the

Selection of Underground Sectionalizing Equipment Faulted Circuit Indicators Summary and Recommendations

fault current and is at its maximum level during the first few cycles when the asymmetrical fault is at a maximum. The ability of system components to withstand mechanical stress is mainly a function of design. Where the maximum available fault exceeds the withstand capability of the system component, the only solutions are the following: • Replace the component with a heavier duty unit, • Modify the circuit configuration to reduce the maximum available fault, or • Use current-limiting protective devices to reduce the let-through current. Thermal stress is a function of the energy released in a system component during a fault that results in rapid heat buildup. The magnitude of energy involved is proportional to current squared multiplied by time (I2t). The traditional approach to reducing thermal damage is to reduce the amount of time a fault is allowed to exist through the careful selection of protective devices and device settings. Where maximum fault levels are so high that the operating time of the protective device must be reduced to an

8 0 – Se c t i on 3

3 coordinated properly, the fault location should be between Optimize reliability by the device that has operated sizing equipment for and the next load-side device. If the maximum number of maximum faults and protective devices that can Contain Fault Damage using enough feasibly be installed are used, One objective of protective protective devices. the length of line between deequipment is to limit damage vices will be relatively short. at the actual fault site. It is This design approach will reoften impossible or impractical strict the amount of line that to completely eliminate its ocmust be searched for a fault. Thoughtful placecurrence. Through the use of protective devices, ment of devices will also help locate faults. For fault current magnitude and fault duration are example, consider a point at which three taps reduced. This reduces, but may not eliminate, branch off a circuit. If a fuse were placed in the damage to the rest of the system from throughmain circuit just before the taps branch off, opfault currents. Thus, most damage is contained eration of the fuse would show that a fault had within the actual location of the fault. occurred in one of the three taps but it would not show which specific tap. However, if a fuse Maximize System Reliability were placed at the beginning of each of the and Power Quality three branches, operation of one of the fuses Adherence to the following guidelines will maxiwould show which of the three taps contained mize system reliability. the fault. Installing the additional fuses in this situation would also improve consumer reliabil• Purchase system components that will withity by reducing the number of consumers interstand maximum calculated through-fault rupted by a fault. currents. Of course, there are practical limitations on • Locate and size protective devices so the the number and location of devices that can be smallest possible portion of the system is deplaced on a circuit. The judicious use of fault inenergized for a permanent fault. dicators between protective devices will help • Size protective devices so they do not permapinpoint a fault location. The application of fault nently open for temporary faults. This indicators is presented later in this section. Fault guideline applies mainly to overhead portions indicators are especially useful where a circuit of a system, as faults on underground systems may sometimes be backfed. In this situation, are usually permanent. protective devices may not coordinate properly and more Additional reliability may be than one device may operate achieved for critical loads by Wise placement of during a fault. Wisely placed use of an automatic transfer fault indicators would be espeswitching arrangement. These protective devices cially useful to narrow down arrangements are expensive and indicators will aid the fault location. and require two or more independent sources of power. in locating faults and OVERVIEW OF FAULTS minimizing outage Aid in Determining The IEEE Standard Dictionary size and duration. Fault Location of Electrical and Electronics Terms (2000) lists several Proper coordination and placedifferent definitions of the ment of protective devices will word fault. The first two help system operators deterdefinitions listed are relevant here: mine a fault location. If protective devices are impracticably short interval, then current-limiting devices can be used to reduce the fault current and the duration.

Underground System Section al iz i n g – 8 1

3 • “A wire or cable fault is a partial or total local failure in the insulation or continuity of a conductor.” • “A component fault is the physical condition that causes a device, a component, or an element to fail to perform in a required manner; for example, a short circuit, a broken wire, or an intermittent connection.” All faults within these two definitions fall within one of two major categories: an open circuit or a short circuit. An open circuit is any circuit in which the normal continuity of the circuit is interrupted. The IEEE dictionary defines a short circuit as “an abnormal connection (including an arc) of relatively low impedance, whether made accidentally or intentionally, between two points of a different potential.” Within the same definition, there is a note that the term fault or short-circuit fault is used to describe a short circuit. Open circuits typically do not lead to damage to the electrical system. In addition, normally available protective sectionalizing devices used on electrical distribution systems do not typically detect open circuits. Frequently, the word fault is associated with its short-circuit definition only, and is used interchangeably for short circuit. Throughout the rest of this section, the word fault will be used to mean short circuit. Although protective relays that detect open circuits to some degree are available (and others are currently being developed), they are outside the scope of this section. Description of Faults Some of the phenomena associated with a fault are listed below. • Very little current flows past a fault point, leading to loss of service to loads beyond the fault. • Voltage at the fault and beyond decreases significantly. The voltage between the generation source and the fault decreases proportionally to the inverse of the line impedance. • Faults typically lead to current levels that exceed the thermal rating of conductor and

other system components, causing damage within a fraction of a second. • The abnormal low-impedance path can include nonutility property or human beings, causing damage, injury, and even fatalities. Causes of Faults Causes of common mechanical failures of underground cables are dig-ins, rodent damage, and improper handling and installation. This last cause includes sharp bending of cable, excessive pulling force during installation, driving vehicles over laid cable, walking on cable in a trench, placing or leaving rocks in a position to cause future cable damage, and allowing nails in reels to damage cable. Principal causes of electrical faults to underground systems include lightning, insulation treeing, and thermal insulation failure caused by overloading. In addition, during single-phase faults on three-phase circuits, the phase-to-neutral voltage on the two unfaulted phases can sometimes increase to a level that can approach the normal phase-to-phase voltage. This increased voltage on the unfaulted phases stresses the insulation and can lead to failure. Failure of splices and elbows is also either electrical or mechanical failure, depending on the cause. For a comparison of the sectionalizing of overhead and underground systems, it is useful to examine the many causes of faults on overhead distribution lines. Some of the more common causes are: • • • • •

Lightning, Squirrels or large birds, Extreme weather conditions, Tree limbs or trees falling on the lines, and Vehicular damage.

Although the intent of this section is to focus on the protection of underground systems, overhead lines in many instances are connected either on the source side or, less frequently, on the load side of underground lines. In these cases, the protective devices often protect mixed line sections. Also, underground devices on systems served by overhead feeders must coordinate with those devices protecting the overhead portions of the system.

8 2 – Se c t i on 3

3 Symmetrical Versus Asymmetrical Faults The terms symmetrical currents and asymmetrical currents refer to the symmetry of the peaks of the current waves about the zero current line. A symmetrical current is symmetrical about the zero current line, as shown in Figure 3.1. Such current symmetry would typically be found in a system under normal operating conditions. During an asymmetrical current, the current wave is not symmetrical about the zero current line and can be completely above or below the zero line. Figure 3.2 shows a typical current curve immediately before and after a fault initiation. As the curve shows, the current is symmetrical before the fault initiation. Immediately after the fault initiation, the current is asymmetrical for approximately the first three cycles before returning to a symmetrical waveform. The degree of asymmetry in the current curve immediately after the initiation of a fault depends on two considerations. The first is the

time within a cycle that the short circuit occurs. If the fault is initiated during a voltage peak, then the resulting fault current will be totally symmetrical. If the fault is initiated near a voltage zero, then the initial fault current will be highly asymmetrical. As the point on the voltage curve moves from the voltage zero point to the maximum voltage point, the degree of current asymmetry decreases accordingly. The other consideration that affects the degree of asymmetry of a fault current is the reactance/resistance (X/R) ratio of the equivalent impedance circuit at the fault location. A high X/R ratio means the inductance of the circuit is greater than the resistance. The higher the X/R ratio is, the greater the asymmetry of the initial fault current is, all other conditions being constant. Using a standard symmetrical component notation, Equation 3.1 shows the X/R ratio for a three-phase fault. Equation 3.2 shows the X/R ratio for a single-phase fault. The positive sequence impedance data (X1 and R1) and zero sequence impedance data (X0 and R0) should be available from a system fault study.

Equation 3.1 Three-Phase Fault X Ratio = X1 ÷ R1 R FIGURE 3.1: Symmetrical Current.

Total Asymmetrical Current

where: X1 = Positive sequence reactance R1 = Positive sequence resistance

DC Component AD Component

Equation 3.2 Single-Phase Fault X Ratio = [(2 × X1) + X0] ÷ [(2 × R1) + R0] R

FIGURE 3.2: Asymmetrical Short-Circuit Current.

where: X1 R1 X0 R0

= = = =

Positive sequence reactance Positive sequence resistance Zero sequence reactance Zero sequence resistance


3 TABLE 3.1: Multiplying Factors to Determine Asymmetrical Fault Currents Where Symmetrical Fault Currents Are Known. X/R Ratio

“Maximum RMS” Factor for 1/2 Cycle, Mrms*

1.0

1.002

1.5

1.015

2.0

1.042

2.5

1.078

3.0

1.116

4.0

1.189

5.0

1.253

6.0

1.305

8.0

1.383

10.0

1.438

15.0

1.522

20.0

1.569

40.0

1.646

100.0

1.697

* Multiply per-phase symmetrical rms short-circuit current by Mrms to obtain momentary per-phase asymmetrical rms fault current.

The rate at which a fault current decays from its asymmetrical waveform to an essentially symmetrical waveform also depends on the X/R ratio. A circuit that has a low X/R ratio (one that is mostly resistive) will decay very quickly. A circuit with a high X/R ratio (one that is highly inductive) will take much longer to decay. Typical protective devices such as fuses, breakers, and reclosers are rated in maximum symmetrical fault-interrupting capability, although some fuses may be rated for maximum asymmetrical fault-interrupting capability. In addition, they will have either a maximum asymmetrical current interrupting capability or a maximum

symmetrical current interrupting rating and a corresponding maximum X/R ratio for the circuit in question. Likewise, switches and sectionalizers will have a close-and-latch rating expressed as amperes symmetrical with a maximum X/R ratio. The asymmetrical rating is based on the rms (root mean square) value of the maximum asymmetrical fault during the first half cycle of fault current. The X/R rating shows that the device is able to successfully interrupt or close into the maximum asymmetrical fault current expected for a system with the following: • A maximum available fault current less than or equal to the symmetrical current rating of the device, and • An X/R ratio less than or equal to the rating of the device. Where an X/R ratio is used to show the maximum asymmetrical interrupting rating of a device, this value is usually fairly conservative. In other words, most distribution system X/R ratios would be expected to be less than the rating of this device and fall within its capabilities. Table 3.1 should be useful where devices are rated in asymmetrical currents or where devices are rated in maximum X/R ratios and the actual X/R ratio exceeds the rated value. EXAMPLE 3.1: Device Rated in Maximum Asymmetrical Current Capacity. The calculated maximum symmetrical fault on a system is 8,000 amperes. The X/R ratio at this location is 10 and the fuse being considered for this location has a symmetrical interrupting rating of 8,600 amperes and an asymmetrical interrupting rating of 12,000 amperes. The multiplying factor Mrms is 1.438 for an X/R ratio of 10.0. The maximum asymmetrical fault for this location is 1.438 × 8,000 amperes, or 11,504 amperes. The maximum symmetrical fault of 8,000 in this location is less than the interrupting rating of 8,600 amperes, and the maximum asymmetrical fault of 11,504 amperes is less than the asymmetrical interrupting rating of 12,000 amperes; therefore, the device is acceptable.

8 4 – Se c t i on 3

3 EXAMPLE 3.2. Device Rated for Maximum Circuit X/R Ratio. In this application, the location being considered has a maximum available symmetrical fault current of 2,500 amperes with an X/R ratio of 20. The device being considered is a recloser with a maximum interrupting rating of 3,000 amperes symmetrical and a maximum circuit X/R ratio of 12. The Mrms factor for the circuit X/R ratio of 20 is 1.569. The Mrms factor of 1.569 times the maximum symmetrical fault current of 2,500 amperes yields a maximum asymmetrical fault current for the circuit of 3,922. Although Table 3.1 does not list an X/R ratio of 12, interpolation can be used to calculate an Mrms factor, which, although not exact, will be within acceptable limits.

transformers operating in parallel if such an arrangement is possible and usual. • A bolted fault (both three-phase and phase-toground) is applied at each location to be evaluated. A bolted fault has zero fault resistance (or reactance). The system engineer should take some precautions when calculating maximum faults:

• Do not calculate maximum faults for system configurations that cannot actually exist Equation 3.3 because of operating restrictions. • When determining the interrupting capability Mrms for X/R of 12 = of devices, use the maximum expected fault, (12 – 10) even if it would occur only under unusual or × (1.522 – 1.438) + 1.438 = 1.4716 (15 – 10) emergency conditions. • When considering the coordination of devices, calculate the maximum fault under The Mrms value of 1.4716 × 3,000 amperes symmetrical equals an asymmetrinormal conditions. In other words, devices cal interrupting rating of 4,415 amperes. The maximum fault conditions of 2,500 should be coordinated under normal system amperes symmetrical and 3,922 amperes asymmetrical are less than the device configuration. It may not be possible to coorratings. Therefore, the recloser is acceptable. If the circuit’s X/R ratio had been dinate devices under emergency conditions 12 or less, there would have been no need to calculate the respective asymmetrical fault current. (such as when a circuit is backfed from a nearby substation). • Calculate both maximum three-phase and Maximum Available Fault phase-to-ground faults. This must be done beThe maximum available fault current is used to cause phase-to-ground faults typically exceed determine if the interrupting capacity of a device three-phase faults in and near delta-to-wyeis adequate. The maximum connected substation transfault current is also the current formers, whereas three-phase magnitude at which the coorfaults typically exceed phaseMaximum available dination of devices is checked to-ground faults further out on fault current should for adequate time clearance. the circuit. Furthermore, some be used to check Maximum faults should be caldevices have different operatculated for both three-phase ing characteristics for phaseinterrupting ratings. faults and single-phase-toto-ground faults than for threeground faults. Maximum faults phase faults. Another reason are calculated using those confor calculating both types of ditions that will lead to the maximum available faults is that most systems have single-phase faults. Typical conditions are as follows: taps for which only phase-to-ground faults should be used when devices are coordinated. • The maximum fault is available from the power When coordinating devices on vee-phase lines, supplier. In this case, the power supplier is calculate phase-to-phase-to-ground faults. operating its system with maximum generation and with its transmission system interconMinimum Available Fault nected to result in a maximum available fault. The term minimum available fault current does • Substation transformers and buses are internot accurately describe the desired value. The connected to produce the maximum available actual minimum fault current on any circuit apfault. A common example is two substation proaches zero. For example, if a broken conductor


3 falls on dry sand or a dead, bone-dry tree, the effective fault resistance approaches infinity, causing a fault that approaches zero amperes. However, the concept of a minimum fault current actually involves calculating the minimum fault current that can be expected during most of the faults on a system. The variables that typically affect the calculated minimum fault are the following: • Available fault current from the source utility or transmission system, which is mainly controlled by the amount of generation online and the transmission system and bus configuration; • The configuration of the distribution system and substation buses; and • The fault resistance, which is the resistance between the faulted conductor and the return path that must be added to the known impedances of the source, transformers, circuit, and other system components.

imum faults, with 10 ohms giving more conservative results. Where circuits are composed of interconnected sections of underground and overhead, it may be necessary to make two sets of fault calculations using the underground fault resistance in one run and the overhead fault resistance in the other run. It is also important to note that site conditions vary widely between utilities and within each distribution system. This variability should always be considered when determining the system standard protection parameters.

DESIRABLE LOCATIONS FOR SECTIONALIZING DEVICES Beginning of UD Cable It is normally desirable to place sectionalizing devices at the beginning of underground cables, that is, any location where a transition from overhead to underground cable takes place or in a substation or step-down transformer where the underground circuit originates (see Figure 3.3). Doing so will minimize restoration time and help distinguish between overhead and underAlthough the effects of the first two variables ground faults. should not be discounted, they frequently either Faults on overhead lines are usually temporary do not vary significantly from the maximum and are best protected by reclosing devices such fault configuration or are not available in the as breakers or reclosers. Since faults on underminimum fault configuration. The third variable ground lines are usually permanent; they are best (fault resistance) usually has the greatest influprotected by nonreclosing devices such as fuses. ence on the difference between the maximum Of course, there are exceptions to this recomand minimum faults. mendation, such as where a circuit is mostly Many field measurements made on utility sysoverhead with a short section of underground tems in the 1930s were used to develop a plot of (for instance, under a river, highway, transmission apparent fault resistances versus a percentage of line, or airport glide path). Coordinating a fuse faults at that resistance level. The results showed with in-line reclosers on the source side and the that the median level of fault load side of the fuse might be resistance was 25 ohms and impossible. In this case, rethe average level was 35 duced protection of the underReclosing is not an ohms. A commonly used value ground line section is more advantage on a totally of fault resistance for overhead desirable than frequent operacircuits is 40 ohms. For substation of the fuse caused by temunderground system, tions of greater than 5,000-kVA porary faults on the load-side as most faults are base capacity operated in the overhead line. 15-kV distribution class, a valTo compensate for the repermanent. ue of 30 ohms is often used. duced protection of the underThese values are for faults that ground line section, the engioccur on the overhead portion of the system. For neer could design the system with a spare cable faults on underground systems with concentric (or cables), install the primary cable in conduit, neutrals or metallic shields, some parties recomor both. This reduces the time needed to restore mend a value of zero to 10 ohms to calculate minservice in case of a failed cable.

8 6 – Se c t i on 3

3 N.O.

Main Substation

115 kV–12.5/7.2 kV GRD WYE

10 miles from Main Substation

5 miles from Main Substation N.O.

N.C.

To Next Substation

N.C.

N.O.

N.C.

N.O.

Legend Overhead Line N.O.

Underground Line Breaker or Recloser Fuse Distribution Transformer Switch N.C. – Normally Closed N.O. – Normally Open

FIGURE 3.3: Sample Distribution Circuit with Typical Locations of Sectionalizing Devices Shown.


3 Another solution is to establish an alternate circuit route to the area that would allow the underground section to be de-energized for repair or maintenance without extended loss of service. Using properly installed fault indicators along with solid blade disconnects at each end of the cable will help operating personnel differentiate a cable fault from an overhead fault. End of Underground Cable Where Continued as Overhead The general use of underground cable followed by a load-side overhead line, other than short underground feeder exits at substations, opens up additional sectionalizing difficulties. Optimum fault protection of such an arrangement is almost impossible to achieve, mainly because underground faults are usually permanent and can cause widespread damage to cable insulation if not quickly and permanently interrupted. On the other hand, overhead faults are usually temporary. Overhead lines can also be subjected to faults for longer periods without extensive damage. A summary of the problems associated with this type of arrangement follows. 1. Underground lines are protected by fuses, single-shot sectionalizers, and other singleoperation devices. 2. Overhead lines are protected by reclosers or breakers that reclose two or three times. The purpose of reclosing is to test for the clearing of temporary faults. Reclosing is often successful in avoiding a sustained outage. When there is a permanent fault, the recloser or breaker will lock out after the third or fourth interruption, or a downstream fuse (or sectionalizer) will operate to isolate the permanent fault. 3. If a recloser or breaker is installed at the beginning of an overhead line that is fed by an underground line, the underground line will be subjected to multiple through-faults because of the reclosing action of the recloser or breaker. The cumulative fault duration could lead to thermal damage of the cable and any fuse protecting the cable. Alternatively, if the underground line is protected by a fuse, then any temporary faults would

cause a blown fuse and an unnecessary outage and service call. 4. A recloser or breaker installed at the beginning of the underground line to coordinate with the load-side recloser or breaker could lead to extensive cable damage during faults internal to the cable system. There have also been occurrences of self-clearing cable faults that have allowed reclosing devices to reset between arcing events, thereby substantially prolonging the duration of faults on the cable system and making cable damage much more extensive. This type of fault is typically caused by a concentric neutral that is badly corroded or fault damaged. The fault impedance would be quite high and may require a significant time interval to establish an arc after being extinguished. Taps Off Main Feeders and Sub-Feeders Typically, it is desirable to install sectionalizing devices at the beginning of taps off a main feeder or sub-feeder. Such devices will prevent service on the main feeder or sub-feeder from being interrupted if there is a fault on the tap. This is also a good location because devices can be readily installed in the switching cabinet. Transformers Pad-mounted transformers must be fused to protect the system from transformer failures and secondary faults. It is necessary to keep fuse sizes small enough to limit the energy and duration of any transformer fault that does occur. Proper transformer fusing reduces the chance of a transformer catastrophically failing. Other Locations Where long underground feeders exist, it may be necessary to install in-line sectionalizing devices at one or more locations between the beginning and end of the feeder. This is particularly the case where several heavily loaded taps are located along the length of the feeder. A feeder cable fault near the end of the feeder would interrupt service to only some, rather than all, of the taps. In-line sectionalizing is also recommended where the feeder is so long that the maximum fault currents at the beginning and

8 8 – Se c t i on 3

3 Overcurrent Protection of Cable System

end of the cable differ appreciably. In this instance, the optimum device at the beginning of the cable might not operate for a fault at the end

of the cable. An in-line device should be sized to operate for a lower fault than for the device at the beginning of the cable.

PHASE CONDUCTOR AND NEUTRAL PROTECTION General Effects of Faults on Cable Damage of underground cable because of fault currents falls into two general categories. The first involves burning at the fault location. The heat produced by the arc between the phase conductor and neutral, cable shield, or other return path can damage all cables and components near the fault. The second category of damage is that caused by a through-fault—that is, the fault current flowing through the cable between the source and the fault location. This through-fault current increases the temperature of the phase conductor and concentric neutral or metallic shield. Although it may not damage the conductors, the elevated temperatures generated by the higher I2R losses can damage those cable materials that contact the metallic conductors. Those materials include conductor and insulation shields, the primary insulation,

and the cable jacket (see Figure 3.4). If maximum through-faults fall below the levels shown on the emergency operating temperature rating graphs in Appendix F, then insulation damage should not occur.

Phase Conductor Jacket

Concentric Neutral/ Metallic Shield

Insulation Shield Insulation

Strand Shield

Strand Fill

Locations Susceptible to Overheating Damage from Fault Currents

FIGURE 3.4: Cross Section of Cable Showing Components Subject to Through-Fault Damage.

Current Paths During a fault, current will always flow from the source through the phase conductor to the fault location. The current can then return through several paths with varying percentages of the current flowing in each path. These paths can include the metallic shield, the concentric neutral, a separate ground wire, a metallic duct system, and earth. Where jacketed cable is involved, the fault current in the concentric neutral or metallic shield may split and flow both toward the source and in the opposite direction from the source until it reaches external grounding connections. Short-Term Effects Short-term effects of faults on cable typically involve obvious burn damage around the fault. In severe faults, there may be enough thermal damage from through-fault current to cause failure of splices, elbows, transformer internal buses, and cable. Poorly made splices and other connections are especially susceptible to thermal damage. Long-Term Effects Long-term effects of faults on cable include deterioration of insulation, conductor and insulation shields, splices, and fittings because of overheating or mechanical forces from large through-faults. The exact effects on the various components vary; however, the potential results are the same. At some point, one of these components may break down as a result of normal voltage stress or normal load current, causing a fault. Another possibility is that, during a later fault, a component that was weakened during previous faults will fail because of through-fault currents, leading to failure at another location. If cables that have been subjected to severe through-faults repeatedly


3 fail, all the cable may need to be replaced.

for commonly used sizes of TR-XLPE or EPR aluminum caUse thermal damage bles. For the sample 3,000-amcurves when sizing pere short-circuit condition, Application of Thermal the total clearing time of the Damage Curve protective devices. recloser falls well below the for Insulation System damage time of all the conducThe main effect on cable tor sizes shown. caused by a through-fault is If a more conservative approach is desired, damage to the conductor shield and main insuthe cables can be sized to protect against exlation from the heating of the outer surface of ceeding their emergency operating temperatures the conductor. In the process of sizing sectionalinstead of the higher short-circuit temperature izing devices to protect cable, thermal damage ratings. There are several reasons for considering curves must be developed for the cables in use this more conservative approach. First, the cable on a system. Figures F.1, F.2, F.3, and F.4 of Apmay have been installed in a manner that rependix F show maximum short-circuit currents sulted in outside mechanical forces continuously for insulated aluminum and copper conductor acting on the cable. Examples of this would incables. The horizontal axis represents short-circlude rock backfill in the trench and residual cuit current and the vertical axis represents time sidewall pressure in conduit sweeps. limitations. There are separate curves for differAlso, the temperature rise calculations used as ent conductor sizes. Figures F.3 and F.4 are the basis for the Appendix F curves consider only based on TR-XLPE or EPR insulation, each of current in the central conductor. Single-phase which has a maximum short-circuit temperature faults through concentric neutral cable will have of 250°C. The appropriate graph should be used heat generated by both the inner central conducto develop applicable thermal damage curves tor and the outer concentric neutral. This will refor the size cables being used. These curves are sult in an insulation temperature higher than calvery conservative; they make no allowance for culated by the standard equations. The emergency heat transfer through the conductor shield and operating (or overload) temperature for XLPE, insulation. When cable is protected with a fuse TR-XLPE, and EPR Classes I, II, and IV insulaor other nonreclosing device, the fuse total clear tions rated for 90°C normal operation is 130°C curve should fall to the left and below the ther(266°F). The emergency overload temperature mal damage curve. When a multiple-operation for Class III XLPE, TR-XLPE, and EPR insulations device—such as a recloser—is used, the total rated for 105°C is 140°C. Figures F.5 through F.8 time to which a cable is subjected to a fault show allowable fault current durations for the should fall below the thermal damage curve. For conductor to reach the 130°C limit. example, if a 70-ampere Type “L” four-shot Figures F.1 through F.4 contain cable damage (2A2C) recloser is used at a maximum fault curtime-current curves on the basis of the cable rent level of 3,000 amperes, the recloser will opshort-circuit temperature rating. This is a less erate twice with a clearing time of 0.03 seconds conservative approach which fully stresses the for each operation and then twice again, with a cable insulation under ideal installation conditions. clearing time of 0.07 seconds each. When using an allowable short-circuit rating, the The total time to which the cable will be suballowable temperature for thermoplastic (HMWjected to the maximum fault is as follows: PE, etc.) cables is 150°C. Thermoset (TR-XLPE, EPR, etc.) cables with a nominal operating limit (2 × 0.03 seconds) + (2 × 0.07 seconds) of 90°C have a maximum short circuit tempera= 0.20 seconds ture of 250°C. The more conservative approach of limiting fault durations such that conductor temperatures only reach the emergency operating Figure 3.5 shows the recloser time-current temperature rating is recommended. curves plotted along with cable-damage curves

9 0 – Se c t i on 3

3 Aluminum/XLPE/EPR Short-Circuit Temperature Rating 60

3,600

50

3,000

40

2,400

30

1,800

1,200

10 9 8 7 6

600 540 480 420 360

5

300

4

240

3

180

2

120

1 .9 .8 .7 .6

60 54 48 42 36

.5

30

.4

24

.3

18

2A & 2B

.2

12

.1 .09 .08 .07 .06

6.0 5.4 4.8 4.2 3.6

B

.05

3.0

.04

2.4

.03

.02

1.8

A

Type L Recloser

1.2

.01 50,000

40,000

30,000

20,000

6,000 7,000 8,000 9,000 10,000

5,000

4,000

3,000

2,000

600 700 800 900 1,000

500

400

300

.6 200

100

Time (Cycles, 60-Hertz Basis)

Time (Seconds)

750

500

350

250 4/0 3/0

2/0

1/0

#1 #2

20

Current (Amperes)

FIGURE 3.5: Example of 70-Ampere, Type “L” Recloser Curves for Cable Protection. Neutral Protection When a concentric neutral is full size or equivalent to the phase conductor in ampacity or when the concentric neutral is a reduced-size neutral

but multiple phases have neutrals operating in parallel, it is usually not necessary to review the protection of the neutral. Where a jacketed reduced concentric neutral, tape shield, or longitudinally

Underground System Section al iz i n g – 9 1

3 corrugated shield is used, the system engineer should further Heating of the review the effects of a throughI t neutral may be a A= fault on the neutral and the M materials in contact with the limiting factor where concentric neutral or shield. where: A = Metallic shield cross-sectional the neutral is less The through-fault capability area, in circular mils than full size or the of connections in the neutral I = Short-circuit current in shield, path should also be examined. in amperes cable is jacketed. In those instances in which a t = Time of short circuit, in seconds separate ground wire is run M = Constant; see Tables 3.5 and 3.6 parallel to the insulated cables, the current in the concentric neutrals or shields is typically negligible. The only portion of the TABLE 3.2: Effective Cross-Sectional Area of Shield. Adapted from concentric neutral or shield that is subject to Okonite Company, Engineering Data for Copper and Aluminum thermal damage is that portion between a fault Conductor Electrical Cables, 1998. and the nearest ground point in a jacketed sysFormula for Calculating A tem. Where the reduced concentric neutral or Type of Shield (See Notes 1 and 2) shield is jacketed and carries the majority of the return fault current for a phase-to-ground fault, 2 1. Wires applied either helically, as a braid or nds the formulas and procedures in the following taserving, or longitudinally with corrugations bles and equations should be applied. 2. Helically applied tape, not overlapped 1.27 nwb Although several other metals are sometimes employed as sheath/shield material (see Tables 100 3. Helically applied flat tape, overlapped (See Note 3) 3.5 and 3.6), copper is by far the most com4bdm 2(100 – L) monly used. Equation 3.3 gives the minimum effective cross-sectional area of metallic shield 4. Corrugated tape, longitudinally applied 1.27 [π (dis + 50) + B] b required for a given fault time. Table 3.2 shows Note 1. Meaning of Symbols the corresponding formulas for calculating the A = Effective cross-sectional area of shield effective cross-sectional area of various types of B = L.C. tape overlap, in mils (usually 375) sheaths/shields. Table 3.3 shows the approxib = Thickness of tape, in mils mate normal operating temperature of the shield dis = Diameter over semiconducting insulation shield, in mils for various steady-state conductor operating temdm = Mean diameter of shield, in mils peratures for cables rated five through 69 kV. ds = Diameter of wires, in mils Table 3.4 shows the maximum allowable trann = Number of serving or braid wires or tapes sient temperatures for shields in contact with L = Overlap of tape, percentage various materials. w = Width of tape, in mils Tables 3.5 and 3.6 give the M values for use Note 2. The effective area of composite shields is the sum of the effective areas of the in Equation 3.3. As shown by the tables, the M components. For example: The effective area of a composite shield consisting of a helically applied tape and a wire serving is the sum of the areas calculated values are constants and depend on the shield from formula 2 (or 3) and formula 1. material, the normal operating temperature of Note 3. The effective area of thin, helically applied overlapped tapes depends also on the shield, and the maximum allowable transient the degree of electrical contact resistance of the overlaps. Formula 3 may be temperature of the shield. These tables are very used to calculate the effective cross-sectional area of the shield for new cable. conservative; no allowance is made for heat transAn increase in contact resistance may occur after cable installation during service fer through the jacket or through the insulation exposed to moisture and heat. Under these conditions, the contact resistance semiconducting shield and the main insulation. Equation 3.3

may approach infinity where formula 2 would apply.

9 2 – Se c t i on 3

3 TABLE 3.3: Values of T1, Approximate Shield Operating Temperature, °C, at Various Conductor Temperatures. Source: Aluminum Electrical Conductor Handbook, 1989. Shield or Sheath Temperature °C at Conductor Temperature Rated Voltage (kV)

105

100

95

90

85

80

75

70

65

5

100

95

90

85

80

75

70

65

60

15

100

95

90

85

80

75

70

65

60

25

100

95

90

85

80

75

70

65

60

35

95

90

85

80

75

70

65

60

55

46

95

90

85

80

75

70

65

60

55

69

90

85

80

75

70

65

60

55

50

Note. The maximum conductor temperature should not exceed the normal temperature rating of the insulation used.

TABLE 3.4: Values of T2, Maximum Allowable Shield Transient Temperature, °C. Source: Aluminum Electrical Conductor Handbook, 1989. Cable Material in Contact With Shield

T2, °C/°F

Cross-linked (thermoset)

350

Thermoplastic

200

Deformation-Resistant Thermoplastic

250

Note. The temperature of the shield is limited by the material in contact with it. For example, a cable having a cross-linked semiconducting shield under the metallic shield and a cross-linked jacket over the metallic shield will have a maximum allowable shield temperature of 350°C. With a deformation-resistant thermoplastic jacket, it will be 250°C.

TABLE 3.5: Values of M for the Limiting Condition Where T2 = 200°C. (Thermoplastic Materials = HMWPE, LLDPE, PVC.) Source: Aluminum Electrical Conductor Handbook, 1989. Shield Operating Temperature (T1), °C Shield Material

100

95

90

85

80

75

70

68

60

55

50

Aluminum

0.039

0.040

0.041

0.042

0.043

0.044

0.045

0.046

0.047

0.048

0.049

Copper

0.059

0.061

0.062

0.063

0.065

0.066

0.068

0.070

0.071

0.073

0.074

TABLE 3.6: Values of M for the Limiting Condition Where T2 = 350°C. (Thermosetting Materials = XLPE, EPR.) Source: Aluminum Electrical Conductor Handbook, 1989. Shield Operating Temperature (T1), °C Shield Material

100

95

90

85

80

75

70

68

60

55

50

Aluminum

0.057

0.057

0.058

0.059

0.060

0.060

0.061

0.062

0.063

0.063

0.064

Copper

0.087

0.087

0.088

0.089

0.091

0.091

0.092

0.093

0.094

0.096

0.097


3 EXAMPLE 3.3: Determine Minimum Shield Size for Known Through-Fault Current. Determine the size copper wire shield required to carry a fault current of 10,000 amperes for 10 cycles for a 15-kV XLPE cable having an XLPE insulation shield and a deformation-resistant thermoplastic overall jacket.

STEP 1. Determine the approximate shield operating temperature for 90°C conductor temperature (which is the maximum temperature for normal operation of XLPE-insulated cables). From Table 3.3,

T1 = 85°C

STEP 2. Determine the maximum allowable shield transient temperature for the cable materials in contact with the shield, which in this case is deformation-resistant thermoplastic. From Table 3.4,

T2 = 250°C

STEP 3. Determine the M value for a copper shield with T1 equal to 85°C and T2 equal to 200°C. From Table 3.5,

M = 0.063 where T2 = 200°C

From Table 3.6,

M = 0.089 where T2 = 350°C

Interpolation of these values for M yields M where T2 = 250°C: M=

250 – 200 × (0.089 – 0.063) + 0.063 350 – 200 M = (0.3333) × (0.026) + 0.063 M = 0.072

STEP 4. Calculate the required shield cross section for a fault duration of 10 cycles (0.167 seconds). Applying Equation 3.3,

STEP 5. Determine the number and size of the wires necessary to equal or exceed 56,758 circular mils. Table 3.2 shows that the effective crosssectional area of a wire shield is equal to nds2, or the number of wires multiplied by the circular mil area of each wire. The number required for any specific wire size is simply the total cross section calculated in Step 4 divided by the individual wire circular mil area and rounded up to the nearest whole number:

A=

10,000 0.167 = 56,758 circular mils 0.072

Number of 14 AWG wires = 56,758 ÷ 4,110 = 13.8 (Use 14)

Similarly, Equation 3.3 may determine the number of any other wire size.

9 4 – Se c t i on 3

3 inside a three-phase transformer or between the primary phase lead and ground inside a singlephase transformer. The next highest fault is when the primary windings short; the magnitude of this fault depends on the impedance of the windings between the fault location and the primary leads. The lowest magnitude of fault occurs because of a short in the secondary windings. The more windings between the fault location and the primary side of the transformer, the lower the fault current. The rupture can result from the energy released within the tank and the resulting pressure. The energy, which is typically measured in joules, is proportional to the magnitude of the fault cur• Where the maximum load expected on the rent squared multiplied by the time duration of cable is much less than the capacity of the the fault in seconds (I2t). Because tank rupture cable, the protecting device can be reduced in is usually caused by failure of the transformer size, improving protection of the cable as long winding, the transformer will need to be disas other coordination criteria can still be met. carded or opened for repairs. Therefore, a com• Where emergency overloads of the cable can mon solution to preventing tank rupture is to be routinely expected, the fuse characteristics place a partial-range, current-limiting fuse under should be reviewed to make sure the overthe oil. Although operating such a current-limitload capability of the fuse is in line with the ing fuse will require opening up the transformer expected overload on the cable. tank to replace the fuse, this is not a problem • In the areas where the cold-load pickup is because the tank will have to be opened anyway. substantially more than the maximum load In addition, a dry-well canister or clip-mounted, current or where the duration of the cold-load partial-range, current-limiting fuse will provide pickup is long, it may be necessary to increase the same result. Either of these can also be full fuse sizes on the basis of operating experience. range. The disadvantage of using a full-range, current-limiting fuse is that it will operate for Whatever the situation, the fuse or device curve all levels of fault current and is much more exshould be kept below the thermal damage curve pensive to replace than an expulsion fuse. The of the cable in question. This is rarely a problem use of a bayonet fuse in seexcept where a fuse might be ries with an under-oil current protecting several cables or limiting fuse can overcome several sections of decreasingTransformers can many of these disadvantages, size cable. If the system engirupture as a result of since the replaceable element neer encounters such a large internal faults. opens for low-level faults or problem, the obvious solution overloads, and the currentis to insert additional fuses limiting element opens for wherever a conductor size high-level faults. change occurs. Standard Practices Most fuses begin to melt at approximately twice their continuous rating and series coil-operated oil circuit reclosers also tend to trip at approximately twice their continuous rating. For these types of devices, it is typical to match the continuous rating of the recloser or fuse to the continuous rating of the cable. For electronically controlled reclosers or relayed circuit breakers, the equivalent continuous rating would be about one-half the trip rating. This general rule would not be used in the following situations:

PROTECTION AGAINST PAD-MOUNTED TRANSFORMER TANK RUPTURE Internal Faults as Cause of Rupture Of the very small percentage of transformer tanks that fail by rupture, most rupture because of internal faults. The magnitude of fault current is highest for a fault between the primary leads

Philosophy and Theory of Rupture Prevention The basic philosophy of rupture prevention is to prevent ruptures for any and all fault conditions. The consequences of a rupture are as follows: • Release of oil and the consequent environmental damage,


3 • Ejection of flaming oil and metal parts into the air surrounding the transformer with possible damage to equipment and surroundings, and • The possibility of transferring the fault onto the incoming primary lines. There are no standards for the ability of padmounted transformers to withstand internal pressure from a particular level of fault current. Generally, pad-mounted transformers have a higher withstand value than overhead transformers because of the superior energy absorption capability of a rectangular tank compared with a cylindrical tank. Tables 3.7 and 3.8 show some possible fault levels that can be used as general guidelines for the fault level at which an overhead or

Equation 3.4 IS = (IA2t) × (18.75 + 105 cos θ) where: IS = Symmetrical fault current that will result in known I2t level IA = Known I2t level that may result in destructive transformer damage θ = Arctan (X/R) TABLE 3.7: Approximate Levels of I2t (Amperes2 x Seconds) That May Result in Destructive Transformer Failure for Internal Faults. System Voltage

Overhead Transformers

Pad-Mounted Transformers

15 kV

1.2 × 105

5.0 × 105

25 kV

6.6 × 104

3.0 × 105

35 kV

5.0 × 104

1.0 × 105

pad-mounted transformer will rupture. Equation 3.4 represents an approximate formula for calculating the symmetrical fault current that will result in a known I2t level. This formula was solved for selected X/R ratios at the transformer rupture levels shown in Table 3.7. With the results presented in Table 3.7, the maximum current that overhead and pad-mounted transformers can withstand at typical distribution voltage levels and selected X/R ratios was derived and is shown in Table 3.8. However, these levels are by no means an authoritative guide. Consult the manufacturer of the particular brands of transformers in use on a cooperative’s system for their withstand capability. Practical Prevention/Reduction of Ruptures Pressure-Relief Valves The pressure inside a transformer tank will increase because of extended periods of overload or low-level faults that are not cleared by the protecting fuse. If unchecked, these pressures can increase to levels high enough to severely deform the tank and damage bushing seals. A pressure-relief valve will release these slow buildups of pressure, thus avoiding the development of high internal pressures and tank damage. However, a high-level or internal fault builds the pressure too fast for the pressure-relief valve to be effective. In these cases, the pressure-relief valve cannot protect the tank from damage caused by excessive pressure. Secondary Breakers Secondary breakers act no faster than do expulsion fuses. In particular, the minimum clearing time for a secondary breaker is approximately 0.8 cycles, or the same as a fuse. More important,

TABLE 3.8: Approximate Levels of Fault Current Symmetrical (Amperes) That May Result in Destructive Transformer Failure for Internal Faults. Overhead Transformers (X/R Ratio)

Pad-Mounted Transformers (X/R Ratio)

System Voltage

2.5

5

10

20

2.5

5

10

20

15 kV

2,600

2,200

1,900

1,700

5,400

4,400

3,800

3,500

25 kV

2,000

1,600

1,400

1,300

4,200

3,400

3,000

2,700

35 kV

1,700

1,400

1,200

1,100

2,400

2,000

1,700

1,500

9 6 – Se c t i on 3

3 most ruptures are caused by internal faults that would not be cleared by secondary breakers.

rupting the maximum available fault current. A common cause of tank rupture is degeneration of oil into combustible gases as the result of a sustained secondary fault that eventually causes an internal expulsion fuse to operate. The fuse ignites the combustible mixture and a violent tank rupture can result. Because this type of failure occurs when an expulsion fuse ignites the gas mixture, the use of current-limiting fuses and pressure-relief valves (to vent gas as it is generated) will help reduce this type of violent failure.

Expulsion Fuses Internal fuses typically have a maximum interrupting rating of 3,500 amperes asymmetrical for the weak-link type of fuse rated 7.2-kV phaseto-ground. Internal fuses rated 14.4-kV phase-toground typically have a maximum interrupting rating of 2000 amperes asymmetrical for the weak-link type of fuse. These interrupting ratings vary from manufacturer to manufacturer and should be checked for the particular fuse. Current-Limiting Fuses Even lower interrupting ratings are typical of Current-limiting fuses are nonexpulsion fuses three-phase transformers, where phase-to-phase and generally have a maximum interrupting ratfaults may occur. Three-phase 25-kV transforming of about 10,000 to 50,000 amperes symmetriers with internal weak-link expulsion fuses may cal current. The maximum interrupting rating have an asymmetrical interrupting rating as low varies depending on the manufacturer, model, as 600 amperes. Where the maximum available and size of the fuse. On the majority of underfault level exceeds the rating of the fuse, an exground systems, a current-limiting fuse capable ternal expulsion fuse with of interrupting maximum fault greater interrupting rating or a currents at all or almost all full-range current-limiting fuse locations should be available. Current-limiting should be installed in series Be sure that the maximum with the internal weak link. load current is less than the fuses can protect As with all fuses, the maxicontinuous current rating of against tank rupture. mum clearing time for faults the largest current-limiting fuse within the interrupting rating available. of the fuse is 0.8 of a cycle. If Manufacturers of currentthe maximum I2t let-through limiting fuses have available graphs or tables indicating the maximum I2t current as read from Table 3.8 or calculated 2 from Equation 3.4 is less than the I t required let-through. To protect against tank rupture, the maximum total clearing I2t of the fuse must to rupture the transformer, then an expulsion fuse could prevent tank rupture. It is understood be less than the I2t withstand capability of the that the external fuse must be capable of interprotected transformer.

Effect of Inrush Current on Sectionalizing Devices

TRANSFORMER MAGNETIZING INRUSH CURRENTS When a transformer is first energized, the only magnetic field in the transformer is that caused by any residual flux. For a very short time after the transformer is first energized, the current flow will be relatively large until the steady-state flux level is reached. The size of this magnetizing inrush current depends partially on the residual flux in the core and the impedance of the source. Also controlling the size of the magnetizing inrush current is the point on the volt-

age curve of the source at the time the transformer is energized. If the transformer is energized when the supply voltage is zero, the inrush current will be at a maximum value if there is no residual flux within the core. If the transformer is energized when the supply voltage is at a maximum level, the inrush current will be zero. Estimating Magnetizing Inrush Current Level Calculating the maximum available inrush current for a particular transformer is not feasible.


3 These calculations require detailed design data Fuses that are not usually available for the transformer The main problem associated with fuses comes in question. There are many rules of thumb for from using an undersized fuse. Using an undertransformer inrush current levels. Most of these sized fuse on a large pad-mounted transformer use 0.1 second as the maximum duration for is a fairly common practice, particularly where which the inrush current will flow before dying the present load is much less than the capacity out. One rule of thumb uses of the pad-mounted trans12 times the transformer baseformer and where coordinarated full-load current for tion with the protective device Undersizing protective transformers greater than 3 at the source prevents use of MVA in size. For transformers the size fuse that is normally devices can lead to less than or equal to 3 MVA in used for full capacity. If the tripping because size, the maximum magnetizfuse falls below the magnetizing inrush current is generally ing inrush current point, the of magnetizing considered to be eight times fuse may have to be either ininrush current. the base-rated full-load curcreased in size or replaced rent. The 3-MVA level used in with another fuse of the same this rule of thumb is the threesize but a slower speed. phase MVA capacity of the transformer or of It is critical at this point to recheck coordinathe transformer bank if three single-phase tion of the new fuse with the source-side detransformers are used. vices. If this fuse is on a large transformer bank This magnetizing inrush current is shown on on a rural system, this coordination is difficult. a time-current coordination curve as a single Fuses are particularly troublesome when underpoint on the 0.1-second axis at the appropriate sized, as the magnetizing inrush current may inrush current level. All protective devices not cause the fuse to operate the first few times located on the source side of this transformer the transformer is energized. However, over a should have curves with all points on the period of time the fuse is gradually damaged, curve located either above or to the right reducing the effective size of the fuse. This of the magnetizing inrush point. damage can lead to the eventual failure of the fuse for no apparent reason. Current-limiting Effects on Devices fuses are generally not affected if they are parThe main problem associated with magnetizing tial-range fuses. Full-range current-limiting fuses inrush current is the unnecessary operation of would be affected if undersized to the point protective devices. This problem typically results that the magnetizing inrush current falls above from choosing devices with operation curves that the operation curve. fall below and to the left of the magnetizing inrush point. When a coordination protection scheme is Breakers established, not only should devices protecting The area of concern for breakers is the instantasingle transformers be reviewed for their approneous setting. This setting must be above the priate size and relationship to magnetizing current level of the magnetizing inrush current inrush currents, but tap fuses or feeder protecbecause the operation time of an instantaneous tive devices also should be investigated. This is unit is less than 0.1 second. A clear indication of particularly true where these devices protect an improperly set instantaneous level is a breaker loads—such as industrial parks—that may have with a reclosing relay operating once instantaseveral large transformers. These transformers neously when a transformer is energized and appear to be one large transformer from the then closing on the second operation, which is a perspective of the protective device when a time delay curve. If the breaker relay settings are dead feeder is energized. Below are some of the sized so the operation curve falls above the magproblems associated with particular protective netizing inrush point, breakers typically are not devices. affected by the magnetizing inrush current.

9 8 – Se c t i on 3

3 Reclosers The main problem with reclosers results from the initial fast curves being set below the magnetizing inrush point. This problem is similar to the one with a breaker in that a recloser will operate on the fast curves where a large transformer is located on the circuit and then close in and stay closed when operating on the time-delay curves. Again, the solution is to simply set the fast curves above the magnetizing inrush point. In addition, reclosers with electronic controls that have instantaneous trip or lockout accessories must have the instantaneous current setting above the magnetizing inrush current level. Sectionalizers Sectionalizers can be armed by magnetizing inrush current; that is, the sectionalizer sees the high current level as a load-side fault, which is then interrupted by a source-side recloser. In other words, the normal attenuation of the magnetizing current appears to be a recloser operation to the sectionalizer. Some of the new sectionalizers are able to differentiate between a magnetizing inrush current and a true fault current.

COLD-LOAD PICKUP CURRENTS On a typical distribution system that has been energized long enough that the system has reached a steady-state condition, not all the load-producing devices will be on at any one time. Appliances such as air conditioners, heating systems, refrigerators, and water heaters normally cycle on and off. Therefore, at any instant, a percentage of these devices will be in their off cycles. For example, a circuit that has a 2,000-kW load on it may have 500 kW in continuous load such as lights and 3,000 kW in cyclical devices, of which only half are energized at any one time. (Note that these values are used as an example and not intended to show normal values on a system.) If this circuit is de-energized for an extended period (e.g., 30 minutes) and the system is then energized, all the cyclical loads will be in an energized state or will go to an energized state upon resumption of the source voltage. This energized state occurs because the parameters that are used to operate these devices— such as air temperature or water temperature (in the case of a water heater)—exit the acceptable range and, therefore, initiate operation of the applicable device. In this example, the loads seen upon re-energizing the circuit are 3,500 kW. The load experienced by a system after the resumption of service following an extended outage period is the cold-load pickup. Caution should be taken when re-energizing a feeder after an extended outage because it may be difficult to distinguish between cold-load pickup and an uncorrected fault.

Application of Sectionalizing Devices Sizing protective devices or their curves to avoid their operation as the result of magnetizing inrush currents is usually simple. The curves should be chosen so they are located either above or to the right of the magnetizing inrush point. For example, a magnetizing inrush point of 0.1 seconds and 5,000 amperes simply shows that any Cold-load pickup point on the curve for which can cause protective the operation level is less than Estimating Cold-Load 5,000 amperes should be Pickup Currents devices to trip. greater than 0.1 second. Any The magnitude of cold-load point on the protective device pickup varies depending on curve that is less than 0.1 secthe type of load served and ond in operating time must have a current level of the time of year. In most areas, the cold-load greater than 5,000 amperes. As cautioned earlier, pickup during the spring and fall is less than several large pad-mounted transformers should during the summer or winter because many of be treated as one transformer for those instances the cyclical devices such as heaters or air condiin which circuit protective devices or station tioners do not operate during these periods. feeder breakers may be used to energize the Cold-load pickup also clearly depends on the group of transformers. geographical location of the utility in question.


3 In the Southeast and Southwest, the cold-load pickup during the summer is quite significant because of the air-conditioning load. In northern states, the cold-load pickup during the winter is probably the most significant. However, the cold-load pickup during the winter also depends on the percentage of electric as compared with nonelectric heating systems. Rule of Thumb 3.1 Where large amounts of resistive heating or air conditioning are in use, the cold-load pickup may be estimated as the following: • Two times full load current for 30 minutes, and • Three times full load current for 30 seconds. These are rules of thumb and may vary. The three most important variables the operator can expect concerning the amount of cold load to be picked up upon service restoration are: • length of outage, • type of load, and • weather conditions. Unless the outage is at a time of extreme temperature, an outage of less than 15 minutes will not allow enough time for most of the thermostats to call for heating or cooling. The practice of putting a time-delay relay on compressor start after an outage is becoming fairly common. This design approach reduces the initial inrush upon line energization but does not reduce the 30-minute load requirement of Rule of Thumb 3.1. Most cooperatives should have an idea of the cold-load pickup on their systems based on experience. Furthermore, the cold-load pickup in a system will, of course, vary from one circuit to another depending on the type of load on that circuit. For example, a circuit feeding an all-electric housing development will have a higher cold-load pickup than will a feeder into a residential neighborhood where the main heating methods are oil, propane, or natural gas. Also, some feeders with large loads using large motors, such as irrigation systems or crop-drying systems, may have lesser values of cold-load pickup because these systems may have to be manually restarted.

Effects on Devices In general, where the time-current curves for a device fall below the cold-load inrush points, the protective devices will operate for cold-load pickup. In general, it is desirable to choose devices or particular curves for those devices so the curves fall above or to the right of the coldload inrush points. In those instances where other restraints prevent this choice, it may be necessary to segment the system to pick up load after an extended outage. This segmentation is done by opening the feeder that suffered the outage at different points, picking up a section at a time starting at the end of the feeder nearest the source, and allowing each section to remain energized for long enough for the load to return to its steady-state level before energizing the next section. The effects of cold-load inrush on different types of devices are addressed below. Breakers The breaker may operate if the cold-load pickup is large enough. Where an instantaneous relay is associated with the breaker, it may be that a cold-load pickup will trip the breaker once on instantaneous trip with the breaker then reclose and provide service from that point on. The solution here is to simply increase the pickup level of the time-delay curve on the breaker or, in the case of an instantaneous pickup, to increase the pickup level. In some instances, it may be acceptable to have an instantaneous pickup that trips once on cold-load pickup. Reclosers Reclosers are similar to breakers in that they will trip if the cold-load pickup points on the timecurrent curves are above the recloser curves. This is particularly true for the fast curves on the recloser. In those instances in which the fast curves fall below the cold-load pickup points but the time-delay curves do not, the recloser may trip once or twice on the fast curves and then lock in. Those reclosers with electronic controls may have instantaneous trip devices that should be set above the cold-load pickup current level. Older, electronically controlled reclosers have an accessory that temporarily doubles the amount of current required to trip the recloser.

1 0 0 – Se c t io n 3

3 Newer electronic controls have a variety of cold-load pickup adjustments. Any standard curve can be used for the cold-load pickup curve, along with any trip level. Other features that may be available are a time delay after which the curve returns to the normal curve, additional time and current adjustments to the curve, and cold-load pickup curves for phase, ground, negative sequence, and other types of system conditions. Sectionalizers The cold-load pickup current may be sufficient to trigger the sectionalizer. In other words, the cold-load pickup current will appear as a fault to the sectionalizer. However, sectionalizers also require a sharp reduction in current following the actuating current to register as an operation of the source-side protective device. Sectionalizers also have a reset time. In most instances, cold-load pickup current will, at best, cause one count on the sectionalizer; in those instances in which the current decreases slowly, the sectionalizer may not even note any counts. For those sectionalizers that are set for two or more operations before tripping, cold-load pickup typically will not be a problem. Fuses If the cold-load pickup is sufficiently large, it will blow the fuse, interrupting service to all consumers beyond the fuse. In many instances,

Selection of Underground Sectionalizing Equipment

REVIEW OF OVERCURRENT PROTECTION METHODS Fuses The main advantages of fuses are that they are: • • • •

Inexpensive, Compact, Require little maintenance, and Are easy to replace.

Moreover, a current-limiting fuse is the only readily available device that effectively limits fault current and, thus, reduces the destructive failure of transformers and capacitors. The disadvantages of fuses are as follows.

the cold-load pickup current will be insufficient to cause immediate operation of the fuse, but will damage the fuse. Subsequent cold-load pickups will further damage the fuse until it eventually blows either during a future cold-load pickup or sometimes simply during times of high load level. The solution is to increase the size of the fuse or to replace the fuse with a fuse of the same size but with a slower operating curve. However, because of the long duration of cold-load pickup currents, the slower speed fuse will generally not work. When larger fuses do not coordinate with source-side devices and cold-load pickup is not expected to occur very frequently, the time-current curve of the fuse can slightly overlap the cold-load pickup points. Application of Sectionalizing Devices Where possible, the device curves should be set above or to the right of the cold-load pickup points on the time-current curves. In addition, the pickup level for instantaneous relays or accessories should be set above the highest coldload pickup current level. In some instances in which other criteria prevent increasing the pickup level or curves, it may be acceptable for reclosers and breakers to trip on their instantaneous or fast curves before locking in permanently. In those instances, it is very important that all cooperative personnel are aware of that possibility.

• The number of sizes and types is limited. • The total clear curves and minimum melt curves overlap at high fault current levels for fuses with current ratings that are close to each other. • The maximum current-interrupting rating is limited, especially with expulsion fuses. • Expulsion fuses produce hot gases and by-products. • Fuses do not have any reclosing capability. • Fuses have no ability to sense low-level ground faults. • Fuses cause “single-phasing” on three-phase circuits.

Underground System Section a l iz i n g – 1 0 1

3 • Fuses cannot be controlled or monitored by Supervisory Control and Data Acquisition (SCADA) systems.

• The types of relays that may be used to control the breakers are available in a wide variety of characteristics. • The relays (typically inverse overcurrents on In general, the main applia distribution circuit) may be cation for fuses is on radial varied over a wide range of taps that do not require simultime dial settings and pickup Fuses are the most taneous three-phase protection levels to accommodate most frequently used and that are not subject to fresystem conditions and to allow quent temporary faults. Fuses changes as the load increases protective device on particularly lend themselves to over time. In addition to overan underground protecting underground circurrent functions, many of the cuits. The inability of fuses to electronic relays provide alsystem. reclose is not a limitation on most any known relay function underground circuits and within one relay. Some of transformers, because faults on these functions include over/ this type of system tend to be permanent. Reunder voltage, over/under frequency, sensiclosing on this type of system simply increases tive earth, directional power and/or current, the amount of fault damage. Using current-limitimpedance, negative sequence, reclosing, and ing fuses on pad-mounted transformers is very sync check. Other features may include fault beneficial when the maximum fault level is location, a wide variety of metering functions, enough to cause destructive failure of the transevent recording, and communications. Proformer for internal faults. The primary condigrammable logic functions can be used to detions that limit the use of expulsion fuses at fine the sequence of responses to almost any certain locations are the following: type of event. • Reclosing relays are available where breakers protect portions of overhead line; • Where the maximum fault current exceeds instantaneous relays are available to provide the fault-interrupting capability of commonly high-speed operation during high fault levels. available expulsion fuses, and • Breakers can be purchased with maximum • Where the maximum load current exceeds interrupting capability that exceeds that availcurrent ratings of expulsion fuses (typically able in most reclosers. 200 amperes). • Breakers are rated for more operations Another shortcoming of fuses is that their between maintenance than are reclosers. • Breakers interrupt all three phases curves do not always coordinate well with upsimultaneously. stream breakers or reclosers. For this reason, at • Breakers are available with ground certain locations, particularly on heavily loaded feeders, a breaker or recloser rather than fuses trip protection. may be needed to coordinate with substation The disadvantages of breakers are: breakers or reclosers. Circuit Breakers Most circuit breakers on underground distribution systems are found in substations, although it is possible to install breakers on platforms on overhead portions of the system or in metal or fiberglass enclosures. Some of the advantages of breakers follow.

• They require separate relays that add to the total expense. • They require much more space than fuses do. • They require an outside power source (typically a battery). • Their relays must be calibrated initially and periodically.

1 0 2 – Se c t io n 3

3 • They are harder to operate and maintain than reclosers and, particularly, fuses. • They are significantly more expensive than other available devices.

Three-phase reclosers and breakers are used for the following: Three-phase protection,

• Three-phase protection is desired. • Ground fault protection is desired. • It is advantageous to use SCADA for both control and status reporting of reclosers.

Reclosers Where three-phase or sinReclosers are available in both High load current, gle-phase reclosers are used single-phase and three-phase on underground circuits, it is versions. The single-phase seHigh fault current, simple to disable the reclosing ries-trip versions do not reand feature and have one-shot opquire an outside power source eration of the recloser. and are frequently used on Ground fault distribution lines, although they are seen more frequently Sectionalizers protection. on overhead than on underSeveral types of sectionalizers ground systems. The interruptare currently available in both ing rating of most single-phase reclosers is overhead versions and those that can be intypically less than or equal to that of most distristalled in pad-mounted enclosures. For a secbution fuses. The main advantage of a recloser tionalizer to work properly, it must be set for is that it does reclose; however, as indicated earone less operation than its companion recloser lier, this is not considered an advantage on an or breaker. In other words, for a permanent underground system. fault, a recloser located beThree-phase reclosers can tween the sectionalizer and be supplied with a groundthe source senses a fault, Sectionalizers are fault-sensing unit, which is an opens, recloses, and continues not subject to faultadvantage. This is a particular to open and reclose until the interrupting limitations. advantage on circuits with fault is cleared or it trips for a large load where the minimum maximum number of times phase-to-ground fault may be (usually four) and locks out. on the same order of magniThe properly coordinated sectude as the maximum load current. Three-phase tionalizer senses a fault condition, counts each electronically controlled reclosers are also easily recloser operation, and locks out just before the changeable in pickup level and operating recloser goes through its final close operation. curves. The sectionalizer has, thus, isolated the fault beThree-phase reclosers with electronic control yond it, allowing the recloser to successfully reare available with a wide range of SCADA accesclose and continue service to the rest of the sories. Reclosers are usually less expensive than system. On an underground system, it is desirbreakers and come with all controls included. able to have the sectionalizer set for only one The electronic reclosers do require an outside operation to limit the exposure of the underpower source, typically 120 volts ac, although dc ground system to through-fault damage and posversions are available. Reclosers are typically sible safety problems. The source-side recloser is used as a circuit protective device inside a subset for two or more operations to lock out. The station and on main three-phase lines where the total number of operations for the recloser defollowing apply: pends on whether the majority of the system is overhead or underground. • The load current exceeds the rating of A problem inherent in many of the older sectypical fuses. tionalizers is that they tend to count magnetizing

Underground System Section a l i z i n g – 1 0 3

3 inrush or cold-load pickup currents as faults and, therefore, lock out unnecessarily. Sectionalizers are available that are capable of distinguishing between faults and inrush currents such as magnetizing or cold load. These sectionalizers use different methods for doing so. One criterion is to check for loss of voltage on the line. A true recloser operation de-energizes the line, allowing the voltage to fall to zero. Another method is to require that the load current drop to essentially zero after the high inrush current. Again, the operation of a recloser results in zero current while the recloser is open as opposed to inrush or cold-load pickup for which current drops to a normal level. Other features are also available in existing and new sectionalizers that reduce nuisance tripping. Some of the advantages of sectionalizers are as follows:

circuits are readily available. Each incoming/outgoing circuit will pass through a solid bus, a fuse, a switch, or a combination fuse/switch. Almost any kind of circuit arrangement can be accommodated by a switching enclosure or enclosures. See Figure 3.6. Switches and fuse/switch combinations may be designed for de-energized switching duty only or they may be equipped with an arc suppression device that allows opening and closing the switches under load up to a maximum rated current level. This interrupting rating may be equal to or less than the maximum continuous current rating of the switch or combination fuse/switch. Extreme care should be taken to avoid opening or closing a switch that is carrying current in excess of the interrupting rating. Therefore, a design engineer should never apply an interrupting device in a location where load will exceed its rating. Different types of fuses are available. An • They are less expensive than reclosers expulsion fuse is the most commonly available or breakers. type. This fuse is frequently supplied with a si• They do not interrupt faults and, therefore, lencer that eliminates or reduces venting when can be used in areas with higher available the fuse operates and also muffles any sounds. fault currents than can fuses. (The shortA silencer is a necessity where fault currents are time-current withstand capability of the relatively high in magnitude and the resulting sectionalizer, however, must be greater exhaust gases, if released within the enclosed than the available fault current.) space of a pad-mounted compartment, can be disastrous. Sectionalizers are also useful where coordinaAt some locations, available faults exceed tion between devices is tight, as they have no the maximum interrupting capacity of expulsion time-current curve. Heavily loaded taps often fuses. In addition, high-level faults can lead to cannot allow another level of coordination. the disruptive failure of load-side devices such as transformers. Full- or partial-range currentPROTECTIVE EQUIPMENT IN limiting fuses are available for PAD-MOUNTED ENCLOSURES use in these locations. Figure Fuses and Switches 3.6 shows an assortment of Fuses and switches are comMost protective current-limiting fuses that are bined here because both are used in pad-mounted often found in the same enclodevices, with the switchgear. sure, although enclosures can exception of breakers, Another solution for high be purchased with switches fault current level areas is a only or fuses only. In addition, are available in partial-range current-limiting fuses combined with integral pad-mounted form. fuse in conjunction with an load-interrupting mechanisms expulsion fuse. The expulsion that provide the dual function fuse will operate for low- to of a fuse and switch in one moderate-level faults without damaging the device can be purchased. Pad-mounted enclomore expensive current-limiting fuse. Both fuses sures with up to four or more incoming/outgoing

1 0 4 – Se c t io n 3

3 Vacuum switches are also available from some manufacturers. These switches have contacts in a vacuum bottle which increases the interrupting capacity of the switch to handle higher ranges of fault current. In addition, the duty cycle of the contacts is greatly increased by the vacuum. In the past, oil switches were available, but these have essentially been replaced by vacuum switches. The operation of all types of switches can be controlled by several different means:

FIGURE 3.6: Current Limiting Fuses for Pad-Mounted Switching Cabinets. Courtesy of Hi-Tech Electric (T&B), 2007.

will operate for high-level faults, with the current-limiting fuse limiting the length and magnitude of the fault and consequently limiting the total magnitude of energy expended at the fault. Sometimes the voltage withstand characteristics of blown partial-range current limiting fuses mandate the simultaneous operation of both fuses so that the open circuit created by the expulsion fuse removes voltage from the partialrange current limiting fuse. Another type of protective device is the electronic fuse, which is actually a hybrid device. A control module uses electronic circuitry to sense a fault, initiate tripping, and control the timecurrent characteristics of the device. An interrupting module interrupts the fault under the control of the control module. The interrupting module also has current-limiting capabilities. This device is available in a range of pickup levels and time-current curves. Continuous current ratings up to 600 amperes and maximum symmetrical current interrupting capability up to 40,000 amperes are available. The various timecurrent curves available with this device can often provide better coordination with adjacent devices than can traditional thermal fuses. Another advantage is that the continuous rating of the larger modules exceeds that available in current-limiting fuses.

• These switches can be simply opened or closed manually at the switch location by using hot sticks in energized switches. • Switches can be equipped with stored-energy operators for local operation. These can be spring-operated or battery-operated. Storedenergy operators generally have better switching ratings. • Automatic switch operators are also available. These may have current-sensing controls with or without inverse time-current curves. When equipped with inverse time-current curves, these vacuum switches can then be coordinated with source-side and load-side devices such as fuses, reclosers, and other vacuum switches. Reclosers Single-phase and three-phase hydraulic and three-phase electronically controlled reclosers are available for pad-mounted enclosures. Vacuum interrupters are typically used for increased fault-interrupting capability and increased service life. Hydraulic reclosers with a limited number of curves and current trip levels are available, as are electronically controlled units with an extensive number of curves and current levels. Faultinterrupting capability varies with the current interrupting level of the hydraulic units and is typically 12,000 amperes or higher for the electronically controlled units. Some manufacturers have overhead SF6 gas-insulated reclosers available that, if not yet available for pad-mounted enclosures, may be available in the future. Sectionalizers At least one manufacturer makes a single-phase sectionalizer that is designed for installation in a pad-mounted enclosure. This sectionalizer is


3 intended to work in conjunction with an upstream recloser or breaker and is available in one, two, or three counts before operating configuration. A sectionalizer that is designed to differentiate between a true fault current and a current spike caused by magnetizing inrush or cold-load pickup should be chosen. Additional information on the application of sectionalizers may be found in Electrical Distribution System Protection by Cooper Power Systems (1990). Live-Front Vs. Dead-Front Typical air-insulated, air-break pad-mounted switchgear is available in either live-front or dead-front styles. The older, traditional live-front style uses standard outdoor porcelain or polymer terminations, or stress cones for terminating cable. A removable barrier just inside the doors provides some level of protection of personnel. Once removed, the lineman is easily exposed to the energized parts. Dead-front gear generally limits access to energized parts by the use of modular elbow-type terminations. Both types of switch are generally operated by external handles on source positions, but often must be operated with insulated sticks on fused positions. Only dead-front style switchgear is currently approved for new construction by RUS. REMOTE OPERATION OF SECTIONALIZING EQUIPMENT Reason for Remote Operation There are several reasons to remotely operate a recloser or switch. One reason is to redistribute load. Doing so might be a response to load conditions on the distribution system or to remove load from a transformer or other piece of equipment that is scheduled for maintenance or replacement. Another reason is to isolate a faulted portion of the system. Switches can be opened to isolate

Faulted-Circuit Indicators

Faulted-circuit indicators (FCIs) can be used to locate a faulted section of underground primary cable. FCIs sense the passage of a fault current and display a fault condition. The

the system section that is suspected of containing the fault. The recloser or vacuum switch that opened to isolate the fault is then closed to reestablish service to the remainder of the system. Yet another reason is to retry a recloser or vacuum switch after a lockout caused by an overcurrent condition. On an underground system, this practice is not typically routine; however, there are circumstances in which this would be applicable. One instance is where the underground circuit feeds overhead taps that are unfused. Another instance is where the suspected faulted section has been removed manually and the locked-out device is remote from the fault location. Another instance is when cold-load pickup current or a switching surge is the suspected cause of the overcurrent condition. Devices That Can Be Remotely Operated Devices that can be remotely operated are electronically controlled reclosers, vacuum or oil switches, circuit breakers, and load-break-type switches with motor operators or other types of power operators. These devices must typically be ordered with a remote open-and-close accessory, although such an accessory may be field-installed. Precautions in Remote Operation The most serious danger in remotely closing a device is the possibility of energizing a line or equipment that is in contact with human beings. These could be cooperative personnel working on the line or members of the general public who are in contact accidentally, such as through an automobile that has damaged a pad-mounted transformer. They could also be individuals who have tampered with an enclosure. Another danger is re-energizing a faulted line or transformer that will lead to increased equipment damage and possible human injury.

FCIs sense fault current and display fault conditions

faulted line section will be located between the last indicator showing a fault condition and the first indicator showing a normal condition. Field personnel responding to a power

1 0 6 – Se c t io n 3

3 application problems can be outage can trace the status of corrected through a better the FCIs and quickly identify The FCI can be understanding of how an the faulted line section. They a valuable FCI works and its limitations. can then isolate this line secIn addition, some manufacturtion and promptly restore fault-locating tool. ers now supply FCIs with an power. array of automatic timed reset Without FCIs, field personoptions, which can greatly renel must search for the fault by duce or eliminate problems associated with false sectionalizing and reclosing on the fault until the tripping. faulted line section is located. This latter method The following information gives guidelines for of fault locating is time-consuming and can proper selection and application of FCIs. When cause cable insulation deterioration. properly specified and applied, the FCI is quite When properly specified and applied, FCIs reliable and can be a valuable fault-locating tool. provide the following advantages: • • • • • •

FALSE TRIPPING An FCI has a sensor to detect the current magnitude present in a cable. A current that exceeds the trip rating of an FCI causes the display to show a faulted condition. Unfortunately, the sensor cannot distinguish between fault current, inrush current, and backfeed current. The indicator simply responds to any current that Inrush and backfeed exceeds its trip rating. As a result, inrush and backfeed curcurrents that exceed rents that exceed the trip the trip rating cause rating cause false tripping.

Reduced outage time, Reduced crew and equipment cost, Reduced stress on system components, Reduced blowing of expensive fuses, Improved system reliability, and Improved consumer relations.

RELIABILITY OF FAULTEDCIRCUIT INDICATORS Older designs of FCIs have been plagued with operational and application problems. As a result, they have acquired a false reputation with some utilities as being unreliable. In response, manufacturers have improved the design of FCIs, and IEEE has approved a guide for testing FCIs (Standard 495). These efforts have helped to eliminate some of the operational problems. For example, FCIs are available with the following:

• Rugged current sensors that operate in accordance with IEEE Standard 495, • An inrush restraint feature to minimize false trips caused by inrush currents, • Sensitive current resets and low-voltage resets for use on lightly loaded circuits, and • Sensors suitable for three-phase use where cables are close together. An operational problem that persists is false tripping caused by backfeed currents. This condition is reviewed in the next subsection. Many

tripping. Inrush Currents Inrush current is a higher than normal current that occurs when a distribution circuit is energized. The inrush current decays to the normal current value after some time. The types of inrush currents and their decay times are explained above in the subsection Effect of Inrush Current on Sectionalizing Devices.

When power is restored to a de-energized line, an inrush current will flow through the cable. If this inrush current exceeds the trip rating of an FCI, the FCI will show a fault condition. Manual reset units will continue to show a fault condition until they are reset by hand. However, automatic resetting units will change back to a “NORMAL” indication when the inrush current decays to the normal load current level. In this situation, only the manual reset units continue to show a false trip condition.


3 Three–Phase Recloser

Fault

A-Phase B-Phase C-Phase

FCI 1 Inrush Current

FCI 2 FCI 3 FCI 4

LEGEND

Load 1

FCI, normal indication FCI, fault indication

FIGURE 3.7: Inrush Current Resulting from Operation of Three-Phase Recloser.

FCI 3 Load 1 FCI 2 Single-Phase Recloser

FCI 1

Inrush Current

Fault FCI 4

Inrush Current

FCI 5 LEGEND

FCI, normal indication FCI, fault indication

Load 2

FIGURE 3.8: Inrush Current Resulting from Operation of Single-Phase Recloser.

Two other situations produce false tripping and obscure a fault location. The first is when a three-phase recloser or breaker protects the underground cable. For example, a fault on phase A trips the FCIs on phase A. The recloser or breaker opens and interrupts power to all three phases. When the recloser recloses, phases B and C experience inrush current. If this current exceeds the FCI trip ratings, then those FCIs will show a “FAULT” condition. Usually the recloser locks open before the FCIs can reset. The outage crew now finds FCIs tripped on all three phases. Figure 3.7 illustrates this phenomenon. The second situation is when a single-phase recloser protects a main line with one or more laterals. A fault on the main line trips the FCIs along the main line. During reclosing, some of the laterals may experience inrush that exceeds

the FCI trip rating. Again, the falsely tripped FCIs remain in “FAULT” indication following recloser lockout. Figure 3.8 illustrates this situation. It is difficult to predict the magnitude of inrush current. Therefore, it is difficult to choose an FCI trip rating that is greater than the unknown inrush value. For this reason, most manufacturers offer an inrush restraint feature on their FCIs. Typically, this feature disables the trip response for 15 to 60 cycles following the energization of cable. The 15- to 60-cycle delay allows the inrush current to decay to its normal load value. The inrush restraint feature increases the cost of the FCI by about 35 to 40 percent. This additional cost is easily justified on underground systems that “see” the cycling action of a source-side recloser. Backfeed Currents Backfeed currents continue to produce false trips and resets of FCIs. However, unlike inrush currents, backfeed currents can remain on the system for long durations. Therefore, a timedelay feature will not alleviate the problem. To address this situation, the cooperative engineer needs to be aware of situations that likely produce backfeed currents. Backfeed currents can occur on three-phase circuits when a single-phase fault is cleared by a single-phase protective device. For example, a fuse will clear a cable fault on one phase while the other two phases remain energized. Any load-side capacitors connected to the faulted phase may discharge into the fault. If the circuit impedance is low enough, this discharge current could be large enough to trip FCIs located between the fault and the capacitor bank. More common backfeed currents result from a delta-connected motor load on a grounded-wye, grounded-wye transformer. For example, consider an underground system that serves several three-phase transformers. A cable fault in the first cable section is cleared by a fuse. The other two phases remain energized and continue to supply partial power to any delta-connected motor loads. The motors produce backfeed currents along the underground cable to the fault location. If the current level is high enough, it will falsely trip the FCIs between the cable fault and

1 0 8 – Se c t io n 3

3 the delta-connected motor load. 800 amperes could trip for any All FCIs on the faulted phase current in the range of 720 to The FCI trip rating may show a “FAULT” indication. 880 amperes. Therefore, it is should be close to These same backfeed curimportant to select an FCI that rents and voltages can also remains sensitive to the minithe available minimum produce false resets. Because mum fault current throughout fault current level. the FCI trip level is usually its range of trip ratings. hundreds of amperes and reset Conductor size also affects current level is usually less trip ratings. The FCI sensor than three amperes, false reset is a more likely mounts around an underground cable and problem than is false tripping. A feedback voltsenses the magnetic field produced by the flow age can also exist on the faulted phase. These of current. This magnetic field is a function of voltage levels can reach 50 percent of the northe radial distance from the conductor. The mal line-to-ground voltage for a grounded-wye, larger the radial distance, the weaker the maggrounded-wye transformer. For grounded-wye netic field. FCIs are typically calibrated at a spedelta transformers, this voltage can reach 86 percific cable diameter. If the actual cable diameter cent of the normal line-to-ground voltage. Most is less, then the trip rating is reduced. Likewise, low-voltage reset units have a minimum reset a large cable diameter increases the trip rating. voltage that is lower than 86 percent of the The manufacturer should be asked to supply the nominal voltage. Therefore, these units would not be suitable for grounded-wye, delta trans1 .9 .8 formers with delta-connected loads. Because .7 .6 grounded-wye, delta-connected transformers .5 should not be installed on a distribution system, .4 this situation should not occur frequently. .3 .2

.1 .09 .08 .07 .06 .05 1,000 1,200 1,400

600 800

.03

400

.04 200

Time (Seconds)

.02

.01 .009 .008 .007 .006 .005 .004 .003 .002

4,000

3,000

2,000

500 600 700 800 900 1,000

400

300

200

.001

100

SELECTING A TRIP RATING Load and Fault Current Magnitudes The trip rating of an FCI is the current magnitude that causes the FCI to display a fault condition. An ideal trip rating is low enough to sense the minimum available fault current and high enough to ignore load, inrush, and backfeed currents. To meet this criteria, the FCI trip rating should be close to the available minimum fault current level. If the available fault current level is unknown, manufacturers suggest a trip rating of two-andone-half to three times the expected load current. At long distances from the substation, the available fault current drops substantially. As a result, the available fault current may get close to the magnitude of the load current. Again, the trip rating should be close to the fault current magnitude. However, the margin between the trip rating and the inrush and backfeed currents is decreased. Thus, the FCI is more susceptible to false tripping. The accuracy of the trip rating also affects selection. Most FCIs have an accuracy of ±10 percent. For example, an FCI with a trip rating of

Current (Amperes, RMS)

FIGURE 3.9: Trip Response for Peak-CurrentSensitive Units.


3 cable diameter at which the FCI is calibrated and a correction curve for other cable diameters.

30E

100E

800A FCI

15E 10

450A FCI

Coordination with Current-Limiting Fuses Some FCIs are peak-current sensitive and will operate within two milliseconds for any current that exceeds the trip rating. Figure 3.9 shows the response time of peak-sensitive units. The peakcurrent devices will coordinate with all types of fuses, including current-limiting fuses. Proper coordination means that the FCI will trip before the fuse clears the fault. If the total clear time of the fuse is faster than the FCI response time, the FCI will not show a fault condition.

Time (Seconds)

1

0.1

0.01

0.001 10

100

10,000

1,000 Current (Amperes)

LEGEND

Fuse Minimum Melt Curve

Fuse Total Clear Curve

FCI Trip Response Curve

FIGURE 3.10: Trip Response for 450A and 800A FCIs.

If the FCI is not the peak-current type, its trip response time is a function of the current magnitude. Figure 3.10 shows the time-current characteristics for this type of FCI. Note the difference in the trip response time for the two types. For example, look at the 800-ampere curve of Figures 3.9 and 3.10. The peak-current-sensitive FCI has a response time of two milliseconds. The other FCI has a response time of 0.3 seconds (300 milliseconds). These slower devices should be compared with the time-current curves for the source-side protective device. For proper coordination with link-type fuses, the FCI curve must be to the left of the total clear curve of the fuse at the minimum fault current value. For example, refer to Figure 3.10. For a minimum fault current of 1,000 amperes, a 450-ampere FCI coordinates with a 30E and a 100E fuse. The FCI should also coordinate with a source-side current-limiting fuse. To coordinate, the FCI must trip at the letthrough peak-current level before the fuse clears the fault. For most current-limiting fuses, the clear time is approximately three milliseconds. As shown in Figure 3.10, a 450-ampere FCI will coordinate with a current-limiting fuse that has a let-through current of 1,100 amperes or greater. Adaptive-Trip FCI The adaptive-trip FCI does not have a specified trip rating. Instead of tripping at a predetermined current magnitude, this device responds to a sudden increase in current followed by a loss of current. Figure 3.11 shows the increase in current magnitude required to set the trip mechanism. For example, consider a sensor type B shown in Figure 3.11. To set the trip mechanism, the FCI must see an increase of 130 amperes within a 50-millisecond time or 100 amperes within an 80-millisecond or greater time. The trip mechanism will release and show a fault indication only if the line current drops to zero. If the line current does not drop to zero within 60 seconds, the trip-set condition will reset to normal. This trip-set and trip-release sequence prevents the FCI from showing a false trip as a result of motor starting load or cold-load pickup. Like the other types of FCIs, the adaptive-trip FCI must be checked for coordination with upstream protective devices.

1 1 0 – Se c t io n 3

3 Sensor Type 100 80 60 40

20

10 8 6 4

2

Time (Seconds)

1 .8 .6 .4

M L BD

WHERE TO LOCATE FCIS For an exact section of faulted cable in an underground system to be located, an FCI must be placed at the source end of each cable section. Most cable sections terminate in some type of pad-mounted equipment. Because this equipment also provides easy access to the cable, the location is ideal for FCIs. The following subsections show several types of underground systems and the placement of FCIs. Underground Segments of Overhead Feeders Overhead feeders may occasionally have segments of underground cable. These underground segments are often installed to avoid overhead line clearance problems. Some applications of underground segments are the following:

.2

0.1 0.08 0.06 0.04

• • • •

Lake or river crossings, Highway crossings, Transmission line crossings, and Airport glide path crossings.

Because these underground segments are part of a main feeder, they are usually not fused. 0.01 Rather, a set of solid-blade disconnects is placed 0.008 at each end of an underground cable section. 0.006 A set of FCIs at each cable end will enable 0.004 workers to determine if a fault has occurred on 0.002 the underground segment. The set of FCIs on the source side will show a “FAULT” indication for a 0.001 fault on the underground cable or on the outgo10 100 1,000 10,000 Current (Amperes) ing overhead feeder. The second set of FCIs on Fisher Pierce Fault Indicator the load side will show a “normal” indication for Model 1547 Adaptive Trip Time Current Curves a fault on the underground cable and a “FAULT” (5A Base Current) indication for a fault on the overhead feeder. This arrangement is shown in Figure 3.12. FIGURE 3.11: Trip-Set Characteristics for Adaptive-Trip FCI. Another consideration for this application is Courtesy of Fisher Pierce Division of Thomas & Betts. whether to use a three-phase FCI or three singlephase FCIs. The three-phase FCI After the circuit is re-enerhas three current sensors and gized, this FCI will adjust to one display. The display shows Locate FCIs at the the line current within 60 seca “FAULT” indication for a fault source end of each onds. During this 60-second on any of the three phases. This period, the FCI is in trip recable section. indicator is suitable when the straint. This feature helps preunderground cable is sectionalvent false trips caused by ized with single-phase devices. upstream reclosers. In addiThe single-phase sectionalizing tion, the FCI continuously readjusts itself for device will be open on the faulted phase, thus changes in the nominal line current. showing which underground cable is faulted. 0.02


3 Recloser

FCIs

Underground Line Segment

FCIs

FIGURE 3.12: FCI Placement on Overhead Feeder with Underground Segment. In contrast, a three-phase sectionalizing device will open on all phases, regardless of which phase is faulted. A three-phase FCI will show a “FAULT” indication; however, it does not indicate which phase. For this type of application, it is better to use three single-phase FCIs. Here, only the FCI on the faulted cable will show a “FAULT” indication. The use of three single-phase FCIs also works well on underground circuit exits from a distribution substation. In many cases, these circuit exits are protected by a three-phase sectionalizing device. If the sectionalizing device has indicators to show the faulted phase, a set of FCIs is needed on the load end of the underground segment only. However, if the protective device does not have phase indicators, a set of FCIs must be placed at each end of the underground segment. Some areas may have very long segments of underground cable. These segments may contain above-ground sectionalizing points or grounding points. Placing an FCI at these locations will locate the exact faulted cable section.

Three-Phase Underground Feeders The most extensive type of underground feeder connects two substations. During normal operation, this feeder has an open point with each side being fed by a different substation. In this application, the FCIs are placed on the circuit exits and on either the incoming or outgoing cables in each sectionalizing cabinet. Figure 3.13 shows this arrangement. Another consideration for this type of system is the choice of a trip rating. To select a proper trip rating, the cooperative engineer must consider the load and fault currents during normal and alternate feeds. If possible, a trip rating should be selected that will respond to the fault current available during normal and alternate feeds. Another option is to use an adaptive-trip FCI. As this FCI adapts to different line current levels, it responds properly during normal and alternate feeds. A third consideration is the use of a threephase FCI or three single-phase FCIs. As covered in the preceding subsection, a three-phase FCI is suitable only when the feeder is protected by single-phase sectionalizing devices. If the devices are three-phase, the only way to identify the faulted phase is to use a single-phase FCI on each cable, unless the three-phase protective device has an individual target for each phase.

Switchgear 1

Switchgear 3

Substation A

Substation B FCIs

FCIs

Switchgear 2

FIGURE 3.13: FCI Placement on Three-Phase Underground Feeder.

1 1 2 – Se c t io n 3

3 Riser Pole

Riser Pole

N.O.

LEGEND

Single-Phase, Pad-Mounted Transformer FCI N.O. Normally Open Point

FIGURE 3.14: FCI Placement for Single-Phase Open Loop.

Underground Residential Subdivisions An underground residential subdivision usually consists of single-phase transformers and cable operated as an open-loop system. Figure 3.14 shows this system with one FCI for each transformer. This arrangement should work properly regardless of the location of the loop open point. Large subdivisions can be more complicated. These subdivisions often contain multiple singlephase loops and may contain a three-phase underground sub-feeder. In addition to being placed at each transformer, FCIs must also be placed in each switching, sectionalizing, or junction cabinet. Figure 3.15 shows FCI placement in a large subdivision. If SW1 and SW2 were three-phase junction cabinets without fused taps, then FCIs must also be placed on each load-side cable. This arrangement lets field personnel open the cabinet and determine which phase has the faulted cable.

Riser Pole

Riser Pole

Switching Cabinet

Switching Cabinet

N.O.

N.O.

N.O.

LEGEND

Three-Phase, Pad-Mounted Transformer Single-Phase, Pad-Mounted Transformer FCI N.O. Normally Open Point N.O.

FIGURE 3.15: FCI Placement for Underground Subdivision with Three-Phase Source.


3 SELECTING A RESET METHOD Manual Reset The manual-reset type is the simplest and least expensive FCI. It typically costs half that of the automatic-resetting types. As expected, there are trade-offs for this reduction in cost. First, service personnel must reset this FCI in the field. Any tripped indicators that service personnel miss will continue to show a “fault” indication. During a future outage, these indicators will confuse crews and probably increase the time required to locate the faulted cable section. If this becomes a common occurrence, crews will soon ignore the fault indicators. Failure to reset an FCI is more likely on an underground than on an overhead system. On an underground system, the FCIs are usually located inside pad-mounted enclosures. After a crew locates the faulted line section, they must open all enclosures located before the faulted cable section and reset each FCI. During afterhours power restoration or during inclement weather, this step may be neglected. This device has two other limitations: • No coordination with current-limiting fuses, and • No remote indicator.

FIGURE 3.16: Current-Reset FCI. The unit has a flag display housed inside a clear viewing window. Courtesy of Fisher Pierce Division of Thomas & Betts.

Because it operates more slowly, this FCI cannot be used on underground systems protected by current-limiting fuses. Without remote indication, crews cannot determine the indicator status without opening each enclosure. FCIs are, thus, less desirable when used on an underground system placed along the front property lines. For these reasons, the use of manual-reset FCIs is not recommended. Automatic Reset FCIs are also available with automatic reset. After tripping, these devices can sense when the cable is re-energized and will then reset to a “NORMAL” indication. Because the reset is automatic, these devices are more likely to show correct indication than is the manual-reset type. As a result, the automatic-reset FCIs can be a more reliable fault-locating tool. Manufacturers offer many types of automatic reset. The costs of these different types are very similar. These types have different applications based on their limitations. Each type of automatic reset and how it is best used is described below. Current Reset Current reset is the most common type of automatic reset. The device uses the same sensor to detect fault and load current (see Figure 3.16). After tripping, this device resets to “NORMAL” when it detects the return of load current in the cable. The load current must be higher than the reset current level. The standard reset current levels are three amperes, 1.5 amperes, and 0.1 ampere. Before selecting a current-reset FCI, determine the normal load current. On 35- and 25-kV systems, the normal load current in a single-phase residential subdivision may be less than three amperes. For example, a load of 30 kW on a 24.9/14.4-kV system has a current of about two amperes. An FCI with a three-ampere reset level would never reset. The lower reset levels, 1.5 amperes and less, are very sensitive and can be susceptible to the magnetic fields of nearby cables. These stray fields can lead to false tripping and resetting in the following applications:

1 1 4 – Se c t io n 3

3 • Single-phase junction cabinets, • Single-phase fuse cabinets, • Three-phase junction cabinets, and • Three-phase switchgear.

primary to the secondary side of the transformer. As a safety feature, this sensor has a lumped resistance probe and 30-kV insulated cable. The resistance probe will limit the fault Current-reset FCIs current if there is a primarycan be placed in all Some of these FCIs can be to-secondary insulation system equipped with magnetic shieldfailure. types of pad-mounted ing to prevent this problem. The low-voltage-reset FCI is equipment. The current-reset FCIs reideal for lightly loaded circuits quire only a current source to where the load current is not reset. Therefore, these devices high enough to reset a currentcan be placed in all types of pad-mounted reset FCI. This FCI is not affected by the magnetic equipment and enclosures. fields of nearby cables during reset; therefore, this device would be suitable for a lightly loaded three-phase circuit. The current sensor to detect Low-Voltage Reset fault current would not have to be as sensitive as The low-voltage-reset FCI is equipped with a a sensor that must also detect load currents of less probe that connects to the secondary voltage than three amperes to reset. terminal of a transformer (see The more sensitive sensors reFigure 3.17). The current senquire magnetic shielding to sor has contact with the priThe low-voltage-reset minimize the effect of nearby mary circuit neutral. When the cables. This is described in FCI senses the proper amount FCI is ideal for lightly more detail in the Current of voltage between the secloaded circuits. ondary terminal and the circuit Reset subsection on page 113. neutral, it will reset. Most units For three-phase use, it is imhave reset voltages of 120 volts portant to know the minimum or 277 volts nominal and can be used in singlereset voltage. This value should be high enough phase or grounded-wye, grounded-wye threeto prevent a false reset caused by a feedback phase transformers. voltage. This effect is described in the Backfeed The voltage sensor will likely cross from the Currents subsection earlier in this section.

Figure 3.17: Low-Voltage-Reset FCI. Courtesy of E.O. Schweitzer Manufacturing Division of SEL.

FIGURE 3.18: High-Voltage-Reset FCI. Courtesy of Fisher Pierce Division of Thomas and Betts.


3 voltage exceeds five kilovolts, High-Voltage Reset an FCI will falsely reset. The high-voltage-reset FCI Time-reset devices mounts on the capacitive test do not respond Time Reset point of an elbow terminator The time-reset FCI resets to (see Figure 3.18). A primary to feedback voltage “NORMAL” after a specified voltage level of five kilovolts or current. time, regardless of the circuit or greater for a period of conditions (see Figure 3.19). about three minutes will reset Therefore, it is very important the FCI. These devices can be to select a time period that is used only on elbow terminaCorrect FCI sensor long enough for crews to retors with capacitive test points. spond and check the status of Care must be used on these placement is the FCIs. If the time period is devices to ensure moisture necessary for too short, the FCI can reset beprotection. fore the faulted cable section For use with three-phase proper operation. is located. These units use a systems, these devices must be lithium battery to keep the specified with magnetic shieldreset time during the power ing. Without this shielding, an outage and to power a flashing LED or beeping FCI can show a false trip or reset caused by curtype of fault indicator. Most batteries have a carents in nearby cables. Another concern on pacity of 800 flashing or beeping hours during a three-phase systems is the chance of feedback 10-year operating life. At the end of 10 years, voltage on the faulted phase. If this feedback most manufacturers recommend replacing the battery. If the unit does not have a replaceable battery, it must be replaced with a new unit. Because these devices will not reset because of feedback voltage or currents, they can be very helpful in some three-phase applications.

FIGURE 3.19: Time-Reset FCI. This unit is battery powered and has an LED flashing light display. Courtesy of Fisher Pierce Division of Thomas & Betts.

SENSOR INSTALLATION Proper Placement on Cable During a phase-to-ground fault, fault current flows through the conductor and a portion returns along the neutral. In a concentric neutral cable, the resulting magnetic field of the neutral tends to cancel the magnetic field of the conductor. If an FCI is installed directly over the concentric neutral, it may not detect the fault current because the magnetic field is canceled or reduced. A second problem occurs on a three-phase system. During a phase-to-ground fault, current can flow in the concentric neutral of the unfaulted phases. An FCI mounted directly over the concentric neutral can sense this current. If the current is large enough, it will falsely trip the fault indicator. Correct placement of the FCI minimizes these problems. Correct placement can be done in one of two ways. The first method is to train the concentric neutral conductors back over themselves on the

1 1 6 – Se c t io n 3

3 current. Therefore, an FCI can be placed directly over a shielded cable without adversely affecting the operation of the FCI.

Concentric Neutral Must Be Looped Back Through Sensor Core to Cancel Effect of Current in Neutral

FIGURE 3.20: Correct Placement of FCI Sensor. Adapted from Yeh, 1990.

FIGURE 3.21: Incorrect Placement of FCI Sensor. Adapted from Yeh, 1990.

cable. The FCI is then installed over the portion of the cable where the neutral conductors are overlaid. The second method is to train the neutral conductors to the outside of the FCI. The FCI is placed on the cable above the concentric neutral conductors. Figures 3.20 and 3.21 illustrate correct and incorrect FCI placement. For a shielded cable with 5- or 10-mil tape, the impedance of the tape shield is large enough that it carries very little fault current. Instead, the neutral will carry most of the fault

Effect of Adjacent Conductor Current The FCI current sensor responds to the magnetic field that results from a fault current flowing through the underground cable. When underground cables are close together, these magnetic fields can overlap. These conditions exist in threephase pad-mounted transformers, sectionalizing cabinets, and junction cabinets. A sensor that is not magnetically shielded can sense the magnetic field of adjacent conductors. A fault current on one conductor can produce a magnetic field strong enough to trip the FCIs on the other two conductors. This false indication can be avoided by not using unshielded current sensors in threephase, pad-mounted equipment. Three-phase applications require the use of shielded sensors. A shielded sensor forms a complete magnetic circuit around the conductor to which it attaches, effectively shielding the sensor from nearby magnetic fields. However, some closed-core sensors are designed to detect very low current flow, as low as 0.1 ampere. These sensors are extremely sensitive to low magnetic fields and, thus, susceptible to false trips and resets. These sensors cannot be used in three-phase equipment. IEEE Standard 495 requires a test for the effect of adjacent current-carrying conductors. The test must verify that the indicator will continue to show “NORMAL” when the sensor is at the manufacturer’s specified distance from an unshielded cable carrying a fault current. The sensor must not be affected by orientation. FAULT INDICATION To be of any use, an FCI must show—by a visual display, a radio frequency (RF) output, or other means—that a fault condition occurred. Figure 3.22 shows an FCI with an RF signal output. Figures 3.19 and 3.17, respectively, show FCIs with remote LED and visual flag displays. An RF FCI eliminates the need to look for the unit. This is a definite advantage in areas where snow or vegetation may obscure a visual display. RF FCIs are also significantly more effective


3 display that was previously used had an indicator arrow that pointed toward the fault. This type of FCI provides some advantage in large subdivisions because crews can first check an FCI in the middle of a cable run and trace the fault from there instead of from the dip pole. The directional feature is also useful if cable circuits are operating in parallel. Some models of the directional FCI must be connected to a secondaryvoltage bushing or an elbow test point in order to establish the direction of fault current flow. When this type of FCI requires secondary voltage, it is suitable for use in pad-mounted transformers only; extreme care must be used in correctly connecting the secondary leads to establish the proper polarity. There are other models of directional FCIs that do not require a voltage connection. Visual displays can be mounted on the sensor FIGURE 3.22: Typical Radio Transmitter Unit or can be supplied with a lead to allow remote to Accommodate Up to 12 FCIs. Courtesy of mounting. To view a display that is mounted on E.O. Schweitzer Manufacturing Division of SEL. the sensor, outage crews must open the transformer or switchgear, which requires unlocking when the cable is relatively inaccessible, such as a padlock and releasing the captive bolt. Then under bridges, in subterranean vaults, or in diffithe cabinet must be restored to a secure condicult terrain. Another advantage tion. This process can be timeis that the fault-locating consuming, especially when process is faster because the crew is under pressure to An FCI indicates a crews do not have to open locate the fault and restore fault condition by pad-mounted equipment. service. The more usual kind of inMounting the display rea visual display or dication is the visual display. motely on the enclosure wall other signal. Common types include the reduces the time spent identiflag display, the LCD readout, fying the faulted section of and the LED flashing light. A cable. The display can thus be flag display and LCD readout are typically viewed without opening the piece of equipment. housed behind a clear viewing window that This mounting method does require installing a ranges from one to three inches in diameter (see viewing window on the enclosure. Most padFigure 3.16). In contrast, the size of the flashing mounted equipment can now be ordered with light is only ¼-inch in diameter (see Figure provisions for mounting FCI remote indicators. 3.19). This size of opening is definitely easier to For existing equipment, remote mounting kits install in a manner that maintains the integrity of are available. the equipment enclosure. The flashing light is A viewing window for a flag display must easily seen at night but can sometimes be diffibe large enough to expose its face, usually a cult to see in bright sunlight. An internal battery one- to three-inch diameter circle. A circle is cut powers the flashing light display. through the enclosure wall. This opening is then Some FCIs have directional capability. These covered with a piece of Plexiglas®. The Plexiglas are useful in locations where fault current might provides some protection from impact and entry flow in either direction. One type of directional into the enclosure. Remote mounting of the

1 1 8 – Se c t io n 3

3 flashing LED is possible with a fiber optic cable and requires only a ¼-inch hole. The display mounts directly through the hole; there is no Plexiglas cover. Remote displays allow restoration crews to trace fault indicators faster. This reduces outage time and improves system reliability. However, a determined vandal could break through the Plexiglas and gain entry into pad-mounted equipment. The flashing light indicator presents less risk of forced equipment entry. However, the cooperative engineer should investigate the durability of this device to be sure that it is very difficult to damage or remove. A ¼-inch hole is large enough to probe an object into the padmounted enclosure. In areas subject to vandalism, a display mounted on the sensor or a remote flashing light display should be considered. In other areas, remote displays of either type are beneficial. Acoustic annunciation is another specialized type of FCI output. This type of FCI has a battery-powered speaker that emits a distinctive tone after the passage of a fault. Application of acoustic FCIs is generally limited to locations where the equipment could be obstructed by snow or vegetation, thus limiting the effectiveness of visual indicators. Acoustic indicators are usually time-reset with provisions for manual reset during circuit restoration. Another type of FCI output is a contact suitable for input to a distribution SCADA system. This approach might be useful in congested areas, such as shopping centers, where there are many fault indicators and an opportunity for communication circuits to connect several FCIs to a common SCADA remote terminal unit. A final concern is that the display maintains its state during normal handling in the field. IEEE Standard 495 requires an impact resistance test. This test requires the display to maintain its

indication state when the transformer lid is slammed open or shut. This is particularly important for indicators with mechanical flags. OTHER CONSIDERATIONS Fault Current Withstand FCIs are exposed to high fault currents. To be reliable, an FCI must continue to operate properly after being exposed to these high current levels. The cooperative engineer should specify that all FCIs meet the Short-Time Current Test of IEEE Standard 495. Maximum Continuous Current An FCI must be capable of operation when exposed to the maximum continuous load current. Indicators with fixed pickup settings will give false indications if the load current exceeds their rating. Adaptive FCIs have the ability to accommodate increasing load currents, but, in some cases, these changes in trip characteristics may impair coordination with system overcurrent protection. Environmental Requirements An FCI must operate in harsh environments including direct sunlight, earth burial, and intermittent or continuous water submersion. An FCI must also operate under a varying range of temperatures. IEEE Standard 495 requires that FCIs operate properly in an ambient temperature range of -40 to 85°C. In addition, this standard requires the following design tests to ensure that FCIs will function in their harsh environments: • • • • •

Temperature cycling test, Water submersion test, Outdoor weathering of plastics test, Salt spray test, and Immersion corrosion test.



1. Fault current values should be available from system fault current study. 2. Sometimes the maximum interrupting rating of a protective device is rated in asymmetrical amperes but only a symmetrical fault current rating is available. Use Equations 3.1 and 3.2 and Table 3.1 to convert from symmetrical to asymmetrical. 3. When minimum fault is calculated, a fault resistance of zero to 10 ohms for underground cable and 30 to 40 ohms for overhead line is recommended. Zero ohms for underground and 30 ohms for overhead are less conservative and should be used only within the restrictions noted in the Minimum Available Fault subsection and subject to good engineering judgment and knowledge of the system. 4. All load-carrying components should be rated to withstand maximum through-fault currents on the system. If this is not possible, current-limiting fuses or circuit reconfiguration should be used to limit the fault. 5. Proper location of protective devices will limit fault damage and the number of consumers affected by the fault and also help locate the fault. Recommended locations are the following: (a) In substations, (b) At the beginning of underground cable, (c) At transitions from underground to overhead, (d) On taps off main feeders and sub-feeders, (e) On transformers, and (f) Within long feeders. 6. Use the cable damage curves in Appendix F to determine if a protective device protects a cable against through-fault damage. The short-circuit curves are normally used; however, the emergency overload curves can be used for a more conservative approach or where the cable is normally operated near its continuous ampacity limit. 7. Where the neutral/shield is reduced in size or is jacketed, the temperature increase in the shield during faults may be more critical than the temperature increase in the phase

conductor. Equation 3.3 and Tables 3.2 through 3.6 can be used to evaluate the temperature increase in the concentric neutral or shield during faults. 8. Table 3.8 shows fault levels that may lead to destructive transformer failure for internal faults. If actual withstand levels of I2t values are known for a particular transformer, Equation 3.4 should be used to calculate a corresponding maximum symmetrical fault level. Current-limiting fuses should be used to protect against destructive transformer failure in high-fault areas. 9. The magnetizing inrush current point for a transformer is estimated as follows:

Transformer Size Three-Phase Single-Phase >3 MVA

>1 MVA

≤ 3 MVA

≤ 1 MVA

Magnetizing Inrush Current 12 × base-rated full-load current for 0.1 seconds 8 × base-rated full-load current for 0.1 seconds

Protective device curves should fall to the right of and above this point to prevent unnecessary tripping. 10. A good rule of thumb for cold-load pickup current is the following: (a) Six times full-load current for one second, (b) Three times full-load current for up to 10 seconds, and (c) Two times full-load current for 100 seconds up to 15 minutes. Frequently, these points may be modified on the basis of the type of load and local climate. Protective device curves should fall to the right of and above these points to prevent unnecessary tripping. This coordination may not always be possible. 11. Several types of protective devices are available for use on an underground system. Most of these are available in a pad-mounted type enclosure. Several of these devices can be operated remotely.


Equipment L o a d i n g – 1 2 1

4 In This Section:

Equipment Loading

Primary Cable Ampacity


Pad-Mounted Transformer Sizing

For an underground distribution system to be operated reliably and efficiently, the two major system components—cables and transformers— must be sized properly. The current rating or ampacity of primary and secondary cables must be selected to economically serve the load over the lifetime of the installation. To meet this requirement, cables must supply the load during peak periods without overheating and within acceptable voltage limits. Pad-mounted transformer

Primary Cable Ampacity

kVA ratings must be selected to carry highly diverse loads with peaks that may exceed the transformer rating. Transformers must be designed to carry these temporary overloads while lasting 20 years or more. By reviewing the conditions that affect primary and secondary cable ampacity and the ability of transformers to carry overloads for short periods, the engineer will have the tools to design the best UD system to meet various system requirements.

thermal operating limit of the cable. Voltage A simple definition of ampacity is the amount of drop is often the deciding element in very long current that a cable can carry under a specific cable runs. For short runs and large currents, set of circumstances. When current flows ampacity is usually the limiting element. through a cable, losses in the form of heat are Maximum insulation tempergenerated in the conductor ature is not the only consideraand insulation. The ability of tion for an underground the cable to transfer this heat Ampacity = Current circuit. Soil temperature to the surrounding environRating of Cable around direct-buried cable or ment sets the actual ampacity conduit should also be considof the cable. ered. If cable temperature rises The maximum conductor to an excessive level, the suroperating temperature limits rounding soil may dry out, causing a large inthe allowable loading of UD cable, although crease in soil thermal resistivity. If the condition loading the cable to the maximum operating persists for an extended period, it can lead to temperature of the insulation will not shorten thermal instability of the soil, which will cause its life. However, voltage regulation and flicker higher cable temperatures and shorter cable life. can limit circuit loading to a value less than the

1 2 2 – Se c t io n 4

4

Equation 4.1

where: TC = RT = RC = I

=

In light of these aspects that affect the rating The actual computations are quite involved, of a cable, a more exact definition of ampacity but engineers will rarely find it necessary to calcanbe formulated. The ampacity rating of a culate ampacity ratings for the cables in their incable is the amount of current (in amperes) that ventories because ampacities for a large range of will cause the temperature of the conductor to cable sizes and installation conditions have alrise from the stated ambient temperature to, but ready been calculated. The ICEA created Publinot above, the rated operating temperature of cation No. P-46-426, Power Cable Ampacities, the insulation under specific conditions that afVolumes I and II, dated 1962. These tables are fect the rate at which heat is removed from the now quite dated and are valuable only for the surface of the cable. On the basis of this definiinstallation conditions and parameters they tion, the basic procedure for calculating cable describe. A newer publication, ICEA P-53-426, ampacity will be explained. which was issued in 1976, addressed UD-style A method to accurately compute ampacity cables and, in particular, the effect of shield under various installation and operating condilosses on ampacities in single-conductor cables tions was first published in 1957 in a technical and temperatures in the earth surrounding paper by Neher and McGrath titled “The Calcuburied cables and ducts. lation of the Temperature Rise and Load CapaAlthough these publications have served the bility of Cable Systems.” This basic procedure is industry well over the years, new insulation still used today to calculate cable ampacity. It is compounds and manufacturing processes have used to calculate the maximum conductor temmade the older tables of limited use. The Insuperature as limited by the lated Conductor Committee of rated operating temperature of the IEEE compiled more upthe insulation. The conductor dated cable ampacity tables Use ampacity tables current required to produce and published IEEE Standard to pick cable ratings. the temperature change can be 835-1994, which lists cables calculated with Equation 4.1. from 600 volts to 500 kV, in ducts, in air, and in directburied situations, with virtually all combinations of single-phase, vee-phase, three-phase, and multiple circuits. An abstract of TC = I2 RC RT these tables is reproduced as Table 4.1. Two-conductor, concentric neutral power cables consist of one insulated central conductor Change in conductor temperature in degrees Celsius caused and one copper concentric neutral conductor by current-produced losses (T conductor/T ambient) applied helically over the insulation. They are Effective thermal resistance between the conductor and used on single-phase or three-phase primary ambient soil, in °C-cm/Watt underground distribution systems with operating Effective electrical resistance of the conductor, in voltage up to 35 kV. micro-ohms per ft If an application arises that is not covered by Conductor current, in kiloamperes these ampacity tables, IEEE 835-1994 or Appendix G should be consulted. Cable vendors can The change in conductor temperature, TC, is also supply cable ampacity ratings for the special given for a particular installation being considinstallations. Also, PC-based ampacity programs ered. Once RC and RT are calculated, Equation calculate ampacities for most cable installation 4.1 can be solved for cable ampacity: arrangements and types of cable. Such programs also help to perform sensitivity analyses in which different parameters can be varied to determine –T T their effect on the ampacity of the cable. UnfortuI = conductor ambient RC RT nately, these programs are sometimes expensive


4 TABLE 4.1: Ampacities for Single-Phase Primary Underground Distribution Cable—XLPE, TR-XLPE, and EPR Insulated. Conductors Rated 15 kV, 90°C, 100% LF Copper Aluminum

Conductor Size AWG or kcmil

Buried*

In Duct*

Duct in Air**

Buried*

In Duct*

Duct in Air**

4

200

121

91

156

94

71

2

260

155

118

203

121

92

1

297

176

135

232

137

105

1/0

339

200

154

264

156

120

2/0

387

227

176

302

177

137

3/0

442

258

201

344

201

156

4/0

504

293

230

393

228

179

250

—

—

—

437

255

200

300

—

—

—

488

288

226

* Two-conductor full-concentric-neutral cable in direct burial at an ambient temperature of 25°C, 100% load factor, and soil thermal resistivity rho-90. ** Two-conductor full-concentric-neutral cable in conduit in air at an ambient temperature of 40°C, 100% load factor, full sun, no wind.

The multiplying correction factors for load factors of 50% and 75% are as follows: Correction Factors 75% Load Factor

50% Load Factor

Cable Rating kV

Buried

In Duct

Buried

In Duct

15

1.08

1.04

1.16

1.07

Continuous loading at maximum rating may lead to moisture migration away from the cables and increased soil thermal resistivity, and a condition of thermal runaway may occur. See “Power Cable Ampacities,” ICEA Publication No. P-46-426; IEEE Publication No. S-135, Section 5, page XIII; or ICEA Publication No. P-53-426 (Reaffirmed 1982), NEMA Publication No. WC 50, page VI, Section E.3, and IEEE Standard 835-1994. Adapted from ICEA S-66-524, NEMA WC 7 (12/84), page 83, and ICEA S-68-516, NEMA WC 8 (reaffirmed 1982), Part 8, page 7, and modified to 25°C ambient earth temperature by multiplying by 0.9636.

and their purchase cannot be justified by most cooperatives for occasional use. CONDITIONS AFFECTING CABLE AMPACITY The maximum ampacity of a concentric neutral UD cable depends on the ability of its surrounding environment to dissipate the heat generated by internal losses. Losses occur physically within the cable in its conductor, insulation, and neutral. Losses in the insulation and neutral may or may not be negligible, depending on the type of

insulation and elements associated with the installation of the cable. Heat flows outward from where the losses are generated toward the jacket. When heat flows through the thermal resistance of the various elements between the conductor and the surrounding soil, it causes a thermal gradient. The temperature gradient, when added to the ambient temperature of the soil (or air), equals the final conductor temperature. This conductor temperature must not exceed the operating temperature of the cable insulation system.

1 2 4 – Se c t io n 4

4 them and the central conductor. Because safety practices require the neutral to be grounded at multiple points along its length, the induced voltage will cause current to flow in a threephase application, adding to the total system loss. Generally, the greater the neutral resistance for cable sizes below 1,000 kcmil, the less the losses will be because of the proportional decrease in current magnitude. This effect is graphically shown in Figure 4.1. It is not usually necessary to calculate the resistance of the concentric neutral because it is expressed as a fraction of the known conductor ac resistance. For example, full and 1/3 are the two concentric neutral resistance values specified in RUS Bulletin 1728F-U1 for primary cable. A full neutral means the neutral and phase conductors have the same resistance, whereas 1/3 means the concentric neutral resistance is three times the resistance of the central conductor. Non-current-dependent losses are caused by losses in the dielectric and charging current loss. The dielectric loss is present any time the cable is energized; the value of the loss is proportional to the square of the voltage. These losses are

Electrical Losses One condition that affects cable ampacity is the magnitude of electrical losses. When a cable is energized and current flows, losses in the form of heat are produced in the conductor and its surrounding insulation and coverings. The rate at which the heat is removed from the cable determines the temperature rise within the dielectric and, thus, the ampacity of the cable. Electrical losses can be divided into two types: current dependent and non-current dependent. Currentdependent losses are caused by current flow in the central and concentric neutral conductors. Non-current-dependent losses are due to the presence of the electrical field within the cable dielectric. They are a function of voltage and are present any time the cable is energized. Current-dependent losses are ohmic losses in the conductor and concentric neutral and vary as the square of the current. Losses in the central conductor represent the main heat-generating component and are directly related to its ac resistance. Losses in the cable concentric neutral occur when voltage is induced on the neutral wires because of the mutual reactance between

0.4 FULL 0.35 1,000 kcmil 750 kcmil 500 kcmil 350 kcmil 4/0 AWG

0.3 FULL

Ysc

0.25 0.2

1/3

0.15

FULL 1/3

1/6

0.1

1/3

FULL 1/3

.05

1/6

1/3

1/6

FULL

0 0

25

50

75

100

125

150

175

200

225

250

275

300

325

350

Rs (Microohms per Foot)

FIGURE 4.1: Ratio of Shield Loss to Conductor DC Loss (Ysc ) at 90°C as a Function of Shield Resistance (Rs), 1/C 35-kV Aluminum Power Cables in Triplexed Formation. Source: ICEA Publication No. P-53-426.

375


4

Per Unit Loss Factor

caused by the in-phase components of voltage Load Factor/Loss Factor and current induced in the dielectric. A second element that affects cable ampacity is Charging current losses are caused by the flow the load factor/loss factor of the load. The maxiof charging current and are separate from the real mum temperature rise of a cable depends on the power flow through the cable. Charging current shape of the load curve and the thermal resisis a function of cable capacitance and is present tance of the heat transfer path. A cable will have any time the cable is energized. Loss calculations a smaller temperature rise if the load varies over involving charging current are, therefore, done a 24-hour period than if the peak load was apat 100 percent loss factor. Losses are equal to plied for the whole day. The effect of a load facthe charging current squared times the ac resistor less than unity is recognized in ampacity and tance of the cable. Because charging current is temperature rise calculations by using loss factor. proportional to voltage, losses caused by it are The loss factor is the ratio of the average losses proportional to the square of the voltage. to the peak losses over a specified period. The Although dielectric losses must be considered IEEE ampacity tables are based on loss factors when setting ampacity ratings for UD cables and determined on the basis of losses for the average are factored into the ampacity tables, their effects maximum load over a one-hour period. Ampacare more pronounced at transmission voltages. ity tables are based on the projected load factor Equations to calculate conof the circuit. Load factor is ductor losses, dielectric loss, defined as the ratio of the avcable capacitance, charging erage load to the peak load. Loss factor compares current, and charging current The relationship between average losses with loss for underground cables load factor and loss factor peak losses. are found in Section 1, Equadepends on the shape of the tions 1.1 through 1.7. load duration curve. Because losses vary as the square of the current, the value of the loss factor can vary between the extreme limits 1.0 of load factor and load factor squared. Figure 4.2 shows this relationship, with curves A and B representing the theoretical limits between 0.8 which the relationship can vary. Typical load A curves for any distribution feeder will fall be0.6 tween the two curves. The loss factor is always less than the load factor except where they are C both unity. This condition occurs when there is 0.4 a constant load on the cable. The loss factor cannot be calculated directly from load factor because losses are proportional 0.2 to the square of the current and the resistance, B whereas the load factor depends only on the current (assuming constant voltage). Note that both factors are related to time. Observations by 0 0.8 1.0 0.2 0.4 0.6 many utility engineers over the years have rePer Unit Load Factor sulted in a relationship between the two values Curve A: Loss Factor = Load Factor that gives a reasonable value of loss factor in Curve B: Loss Factor = (Load Factor)2 terms of its corresponding load factor. The relaCurve C: Loss Factor = 0.2 Load Factor + 0.8 (Load Factor)2 tionship can be expressed by the empirical formula shown in Equation 4.2, which is normally FIGURE 4.2: Relationship Between Load Factor and Loss Factor used for rural feeders. Per Unit.

1 2 6 – Se c t io n 4

4 type of soil (its texture and mineral content), the moisture content, and the structural arrangement of the soil particles. Generally, the higher the Loss Factor=0.2 Load Factor+0.8(Load Factor)2 moisture content, the lower its thermal resistance and the better its heat-dissipating ability. Certain clay soils tend to dry out and become baked when heated beyond a certain temperature; this drives away the moisture and may permanently This equation is shown as curve C in Figure 4.2. increase their thermal resistivity. Clay is also an A more thorough discussion of load factor example of a soil that shrinks when dry, thereby and loss factor may be found in the NRECA losing contact with the cable, which creates an Distribution System Loss Management Manual, air layer between the soil and the cable surface pages 17–20. and adds an extreme thermal resistance to the Because the load factor of a cable determines heat dissipation path. As the thermal resistance its ampacity value, consideration must be given around the cable increases, the cable temperature to future load factors during the expected life of rises. If the cable temperature stabilizes at a safe the cable. Choosing a load factor of 100 percent level, the soil is considered stable. If the temperagives the minimum ampacity value, with all other ture continues to rise above an acceptable level, conditions being equal. An improperly high load the soil is considered thermally unstable. factor could lead to the choice of cable one or Figure 4.3 shows the variation of thermal retwo sizes larger than necessary. Knowing the efsistivity with moisture content for various types fect of the other conditions on cable ampacity will of soils. As the moisture content is reduced, reallow the engineer to make a more informed desistivity rises slowly until a critical moisture level cision about the value of load factor chosen. is reached; the thermal resistivity then starts to increase at a much higher rate. Electric Power Soil Thermal Resistivity Research Institute-sponsored research has shown Soil thermal resistivity (rho) is an important elethat, at high moisture levels, water fills the gaps ment that affects cable ampacity. The tendency between soil particles, which increases the effecof soil surrounding a buried cable to hinder the tive cross-sectional area available for heat transflow of heat from the cable or conduit surface is fer, thus reducing the thermal resistivity of the a fundamental property called soil thermal resissoil (Boggs, Chu, and Rhadhakushna, 1980). tivity, expressed in degrees Celsius-centimeter As the moisture migrates away from the cable per watt (°C-cm/watt). Rho is important in the surface for any reason, heat conduction takes selection of load capabilities of UD cables. In place through a solid soil matrix. Within the masome instances, more than one-half the total trix, the particles have only point-to-point conconductor temperature rise is caused by imtact with each other. The paired heat flow through the ability of different soils to earth. Rho can be measured dissipate heat under these along a specific route to help High soil thermal conditions is determined by in selecting the proper cable resistivity reduces the particle size distribution size. However, measurements (packing efficiency) of the soil are usually difficult and timecable ampacity. and, to a lesser extent, by the consuming to perform. Most shape of the particles. Figure utilities assume soil properties 4.3 shows that, for well-graded soils such as that have led to reliable performance in the past. limestone screenings (a quarry waste by-prodSelecting an ampacity value is complicated furuct), thermal resistivity is basically constant ther because rho depends on many conditions down to low moisture contents of approximately that are not constant through the soil profile and two percent. Below this moisture level, the thercan change with the seasons of the year. mal resistivity is shown to quickly increase. The thermal resistivity of a soil depends on the Equation 4.2


4 210 Crushed Shale Silty Sand

180

Thermal Resistivity (°C-cm/Watt)

Ottawa Sand 150 Critical Moisture Content =

120

90

60 Fire Valley Sand Stone Screenings 30

0

1

2

3

4

5

6

7

8

9

10

Percentage Moisture Content

FIGURE 4.3: Thermal Resistivity Versus Moisture Content for Various Soil Types. Source: Boggs, Chu, and Rhadhakushna, 1980.

160 150 140

The ampacity tables in Appendix G list cablesoil interface temperatures alongside the current values. These interface temperatures show that soil drying around a hot cable can lead to an increase in soil thermal resistivity and increased soil and conductor temperatures. Interface temperature is the temperature attained by the outside surface of directly buried cable, directly buried duct, or concrete encasement. Utility engineers commonly rate cables on the basis of this method. Field studies suggest an interface temperature of 50°C to 60°C be used for clay and loam soils and 35°C for sandy soils (Arman et al., 1964). Interface temperatures have been used in the past to rate cables because no other simple, dependable method exists. Thermal efficiency of the soil depends mainly on its moisture content. In most areas of the United States, soil moisture varies with the seasons. Usually during the cooler months, January through May, rain keeps the soil well saturated. The warmer months have less rainfall and the soil dries out. Because thermal resistivity and water content of the soil are interrelated, it is reasonable to assume that these two properties will vary with seasonal and climate factors as well. Figure 4.4 shows measured variation of soil thermal resistivity at four locations on a monthly basis. The resistivity is shown to generally increase during the hot/ dry months of August, September, and October.

Thermal Resistivity (°C-cm/Watt)

130 120

Soil thermal resistivity changes with moisture content.

110 100 90 80 70 60 50 40

Jan

Feb

Mar

Apr

May

Jun Jul 1952

Aug

Sep

Oct

FIGURE 4.4: Thermal Resistivity of Soil at Various Locations. Source: EEI Underground Systems Reference Book, 1957.

Nov

Dec

In most cases, for a soil of a particular type and a fixed water table level, the moisture content increases with depth. The greater the depth, the less the change in moisture during the year. A higher water content generally leads to a lower thermal resistivity. Years ago, experiments were made that investigated the differences in temperature rise for equally loaded cable buried at intermediate depths from three to 20 feet. As expected, the results showed a lower rho and less temperature

1 2 8 – Se c t io n 4

4 20 Depth Below Surface (Feet) 15 Air

1.5

Temperature (°C)

3 6.5

10

16 5

0

–5

Dec

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Months

FIGURE 4.5: Effect of Depth on Soil Temperatures as Influenced by Seasonal Temperature Variations. Source: EEI Underground Systems Reference Book, 1957.

rise at 20 feet compared with three feet. However, the increase in cable ampacity could never offset the extra cost of deeper burial. Standard industry practice is approximately three feet as an acceptable minimum depth for almost all installations outside urban areas. Unfavorable native soil conditions near the surface can be overcome for short runs by using a good thermal backfill in the vicinity of the cable. Ambient Soil Temperature Ambient soil temperature affects UD cable ampacity and must be considered when using ampacity tables. Every ampacity table has been computed for a specific ambient temperature. The temperature rise of the cable is added directly to the

TABLE 4.2: Typical Ambient Soil Temperatures at a Depth of 3.5 Feet. Source: ICEA Publication No. P-46-426. Temperature, °C Location

Summer

Winter

Northern United States

20 to 25

2 to 15

Southern United States

30 to 35

10 to 20

ambient temperature. The ambient temperature is the normal soil temperature at the burial depth of the cable that would exist if the cable were not there. The change of ambient temperature below the earth’s surface is caused by seasonal exchange of solar energy between air and earth. The earth acts like a heat sink in the summer and returns heat to the air in the winter. Measurements show that soil temperature decreases with depth in summer and increases with depth in winter. Figure 4.5 shows that the temperature change follows essentially a sinusoidal curve that changes with the seasons. The cycle does not vary much from year to year. Cyclical temperature changes below ground vary considerably from place to place and must be known for the specific location being considered. If it is not feasible to make temperature measurements at the site, usable temperature ranges may be obtained from the state Department of Agriculture or the agricultural school of a state university. Table 4.2 gives typical temperature ranges that may be used when site-specific data are not available.


4 Daily variations in air temperature produce negligible changes in ambient earth temperatures below one foot. Investigations have shown that, at depths below 36 inches, ambient soil temperatures lag the air temperature by about two weeks to one month because of the high specific heat of the soil. Cable Configuration and Circulating Current Various aspects of installation can affect the amount of current a cable can carry. For singlephase primary UD cable, dielectric loss is usually considered negligible when ampacity is calculated. Therefore, the current rating of most singlephase UD applications is limited by current-related losses in the conductor and neutral, plus the heat-sink quality of the surrounding soil. In a balanced three-phase application using concentric neutral cable, there is no return current because the phase currents vectorially add to zero at the load. No return current means the magnetic field outside the concentric neutral of each phase is not totally canceled out. Load current flowing in the other two phases will cancel some of the magnetic field produced by current in one phase. Because the net magnetic field around the phase is not completely canceled, it produces a voltage difference along the length of the concentric neutral. In the same way, voltage differences are produced along the concentric neutrals of the other two phases. Safety standards require that the concentric neutrals of all jacketed UD cables be grounded and connected together at both ends of the cable run, and at as many intermediate points as required by the NESC. This necessary grounding of the neutrals at more than one point creates a cross connection which short-circuits the voltage between them and allows circulating currents to flow. The circulating currents produce heat. This heat, when added to the mutual heating effect of the other conductors in a trench, decreases the ampacity of the cable circuit. Voltage differences—and, thus, neutral losses—are proportional to the mutual reactance of the cable system. The most common way to reduce mutual reactance is to place the cables closer together. However, the axial spacing cannot be reduced below one cable diameter, so

mutual reactance will always exist. Another point that must be considered when spacing cables close together is the mutual heating effect caused by the three conductors. Mutual heating will decrease the load-carrying ability of the system. Another way to reduce circulating current losses is to increase the resistance of the concentric neutral. This may be done by reducing the number or size of the wires. Industry practice is to list concentric neutral sizes in relation to the resistance of the central conductor. For example, a cable with 1/3-neutral would have a concentric neutral resistance three times the phase conductor. Engineers recognize that the concentric neutral physically protects the cable. For this reason, cable is usually purchased for standard applications with a concentric neutral made up of at least six No. 14 AWG copper wires. The preceding discussion shows that a threephase installation is more involved in terms of ampacity and that more factors limit its ampacity than for a single-phase circuit. In addition to the conductor losses and the thermal quality of the soil, the arrangement of the phases in relation to each other affects the total system losses and, thus, the circuit ampacity. INSTALLATION CONFIGURATIONS Physical Arrangement of Phases Example 4.1 shows how the physical arrangement of the phase conductors can affect ampacity. Observations from Ampacity Tables The following general observations can be made from reviewing the 1962 ICEA ampacity tables and IEEE Standard 835-1994 for different installation configurations: • Circulating current losses decrease and ampacities increase with increasing concentric neutral resistance. • The smaller the phase conductor, the smaller the variation of the circuit ampacity with neutral resistance. • For large conductors, there is a large variation of ampacity with neutral resistance. • The variation of ampacity with concentric neutral resistance is generally greater for spaced than for trefoil configurations.

1 3 0 – Se c t io n 4

4 EXAMPLE 4.1: Comparing the Ampacity of Trefoil and Flat-Spaced Configurations. Consider two direct-buried, three-phase primary circuits using concentric neutral jacketed cable. Circuit 1 shown in Figure 4.6 is in a closely spaced trefoil or cloverleaf configuration. Circuit 2 is a flat configuration with “maintained spacing” of approximately 7.5 inches between phases, as seen in Figure 4.7. For easier comparison of the two installations, excerpts from their ampacity tables are listed in Table 4.3.

36”

FIGURE 4.6: Trefoil or Triangular Cable Configuration.

Examination of the two configurations of Table 4.3 shows that, for conductor sizes of 350 kcmil and larger, the trefoil arrangement produces fewer losses and greater ampacity as the load and load factor of the circuit increase. For 4/0 AWG and smaller conductors, the spaced configuration gives greater current-carrying capability.

FIGURE 4.7: Flat Conductor Configuration, Maintained Spacing.

36”

Use flat spacing for small conductor installations.

A table similar to 4.3 can be made for aluminum conductors. Table 4.4 shows that, for aluminum conductor sizes up to 500 kcmil, the flat-spaced configuration gives greater ampacity values than does the trefoil. For the larger conductor sizes, the trefoil configuration gives higher ampacity ratings because circulating current losses are greater when flat spacing is used.

A

B

7.5”

C

7.5”

TABLE 4.3: Ampacity Table for 15-kV Copper Conductor, Direct Buried, Single Circuit, 75% and 100% Load Factor.* Trefoil Configuration (Amperes)

Flat-Spaced Configuration (Amperes)

Conductor Size

75% Load Factor

100% Load Factor

75% Load Factor

100% Load Factor

4/0 (1/3 neut)

404

360

432

377

350 (1/3 neut)

519

460

516

448

500 (1/3 neut)

609

535

572

496

750 (1/3 neut)

696

608

635

548

1,000 (1/6 neut)

814

706

705

605

* IEEE Standard 835-1994 Note. Soil rho = 90, Conductor Temperature = 90°C, Ambient Soil Temperature = 25°C

Continued


4 EXAMPLE 4.1: Comparing the Ampacity of Trefoil and Flat-Spaced Configurations. (cont.) TABLE 4.4: Ampacity Table for 15-kV Aluminum Conductor, Direct Buried, Single Circuit, 75% and 100% Load Factor.* Trefoil Configuration (Amperes)

Flat-Spaced Configuration (Amperes)

Conductor Size

75% Load Factor

100% Load Factor

75% Load Factor

100% Load Factor

1/0 (full neut)

216

194

241

214

1/0 (1/3 neut)

216

195

247

218

4/0 (1/3 neut)

318

284

361

311

350 (1/3 neut)

417

370

446

387

500 (1/3 neut)

502

442

513

443

750 (1/3 neut)

604

529

575

498

1,000 (1/6 neut)

716

623

675

582

* IEEE Standard 835-1994 Note. Soil rho = 90, Conductor Temperature = 90°C, Ambient Soil Temperature = 25°C

Conclusions from Ampacity Tables After a comparison of the IEEE Standard 8351994 for trefoil against spaced arrangements with short-circuited and multigrounded concentric neutrals, the following conclusions can be drawn: • When neutral losses are low in both cases, the ampacity of the spaced configuration will be more than the trefoil arrangement because of the effect of lower mutual heating. • When the circulating current losses of the trefoil are measurably greater than the spaced configurations, their ampacities will be essentially the same. • For larger size cables, it is generally better to keep them as close together as possible because the higher circulating currents of the spaced cables provide greater losses and lower ampacities than does the mutual heating effect of the trefoil configuration. Note that if single cables are installed in a spaced configuration in individual steel conduit, the same fields that produce losses in the concentric neutrals will also cause eddy currents and unacceptable heating of the steel. The

losses of the conduit added to the other losses of the circuit will reduce the ampacity even more. In some cases, steel conduit may reach temperatures adequate to cause cable failure by melting. For this reason, nonmagnetic conduit must be used for high-ampacity circuits where phases are enclosed in individual conduits. The preceding discussion will prove useful in comparing closely spaced with spaced threephase circuits. When an installation specification calls for either a trefoil or maintained spacing (flat horizontal configuration), close attention should be paid to the spacing when the cable is laid. Otherwise, inattention to detail could lead to a marginal installation after much effort has gone into selecting the right cable and configuration for the project. CONDUIT APPLICATIONS Installation in Conduit or Duct In this manual, the terms conduit and duct are used interchangeably to mean nonmetallic, nonmagnetic tubes made primarily of polyvinyl chloride or polyethylene. A duct bank means one or more runs of conduit which are usually encased in concrete that extends the full length of the run.

1 3 2 – Se c t io n 4

4 36”

Paved Driveway

36”

Paved Driveway

18”

FIGURE 4.8: Direct-Buried Duct Bank Installation Using Rigid Nonmetallic Conduit.

Heat flow through these thermal resistances causes the temperature of the insulation to rise above ambient temperature. The air between the cable and the inner conduit surface is the main reason why heat is not absorbed by the soil as efficiently as with direct burial, and why a cable in conduit has less ampacity. The concept can be more easily understood by comparing typical thermal resistivity of the various materials. For example, the thermal resistivity of air is 4,000°C-cm/watt, PVC conduit is approximately 480, and soil is approximately 90. It should be remembered this same principle might apply to cables installed with a vibratory plow. In stiff soils the earth may not heal itself tightly back against the cables, leaving air pockets. Consideration might be given to de-rating certain plowed-in cables to cable-in-conduit ratings. The air space acts essentially like an insulating blanket to impede the flow of heat to the surrounding soil. Once an air interface exists, heat transfer Pros and Cons of Cable is not solely by conduction Conduit Installations in Conduit directly from the cable surThe total thermal circuit of a = face to soil; rather, it is cable in conduit can be visualLower Cable Ampacity mostly by radiation and conized as four thermal resisvection into the air space. tances in series: The inside diameter of the conduit should be as small as possible for bet1. Thermal resistance from the conductor ter heat flow. However, the inside diameter of surface to the outer jacket surface, a conduit has little effect on the final tempera2. Thermal resistance from the jacket surface ture reached by the insulation for the typical to the inner surface of the conduit wall, conduit sizes used by utilities. For this reason, 3. Thermal resistance of the conduit material, ampacity tables do not list different ampacities and for different conduit sizes. 4. Thermal resistance of the soil. Laying cable in conduit is being done by many electric utilities. Articles in the technical press have shown that a few systems are justifying using conduit for all UD installations because cable replacement is much easier. In northern climates, it is strongly recommended that conduit be used because digging trenches in frozen ground can be costly and very time-consuming. Conduit can also provide some protection against dig-ins. A sealed conduit system is also useful to keep water away from cables to reduce insulation treeing. However, effective sealing is very difficult to achieve in practice. Underground duct is also used to protect the cable from rodent damage. Cables are usually installed in ducts where they pass under roadways, sidewalks, or other paved areas (see Figure 4.8). Because cable in conduit has less load-carrying ability than direct-buried cable does, conduit applications will be reviewed in more detail.


4 TABLE 4.5: Pros and Cons of Installing Cable Circuits in Conduit. Pros

Cable conductor temperatures in a riser application depend on the following four factors:

Cons

Cable easily replaced (if not fused or frozen)

Higher installed cost

Greater physical protection (for identical cables)

Lower ampacity

Longer life (for identical cables) Provision for load growth (replace with larger cable)

The minimum conduit size required to hold one or more primary cables depends on several elements dictated by the installation. The inside diameter should be large enough to accommodate any movement by the cable(s) caused by thermal expansion. Certain installations may dictate choosing a large conduit diameter to allow a higher ampacity cable to be installed later. When a single cable or a bundle of three JCN cables is being pulled, the conduit must be large enough to allow unimpeded passage. Table 4.5 summarizes the pros and cons of cable circuits in conduit. If more circuits are added to an existing duct bank or trench, the ampacity of all circuits must be reduced.

1. 2. 3. 4.

Number of cables in the vertical conduit, Venting method, Solar radiation, and Riser inside diameter.

When three cables are placed in a single riser, mutual heating will affect cable conductor temperature. At a conductor temperature of 90°C, three cables can have 30 percent less ampacity than can a single cable in the same riser. Additional tests have shown that the heat generated by three cables in a riser will always run hotter than the direct-buried portion of the same circuit. The higher temperature in the riser means the rating of the composite circuit is limited by the riser segment in three-phase, direct-buried applications. A vertical riser can be installed in one of three ways that will affect circuit ampacity. The installation configurations are listed from higher to lower ampacity values:

1. Open at the top and vented at the bottom, 2. Open at the top and closed at the bottom, or Conduit in Air for Riser Pole Applications 3. Closed at both top and bottom. Another cable installation element that needs to be considered in underground applications is Proper venting will greatly increase ampacity the transition from underground to overhead at when compared with a completely closed riser a riser pole (Hartlein and Black, 1983). Utilities that prevents natural air circulation around the usually place cable in vertical conduit for procable. Closing the top reduces the convective tection. It must be determined if this short secheat transfer capability from tion in a protective riser will the cable surface and inside be the limiting factor in a surface of the riser. The precable installation. Venting risers at ferred installation configuration There are no simple estabis to allow the free flow of air lished methods to rate the top and bottom through the riser, which is obriser portion of a cable circuit. increases ampacity. tained with an open top and a Usually, engineers assume that vent at the base. Laboratory underground direct-buried tests have shown that at a load cable runs cooler than does factor of 100 percent, a properly vented riser can the cable section in the riser. Appropriate de-ratincrease ampacity between 10 and 25 percent ing factors based on field and laboratory experiwhen compared with a completely closed riser. ence are then applied to reduce the circuit The vent configuration needs to comply with the ampacity when a riser is present. This method is NESC and good engineering judgment. based on the principle that the current rating of Direct exposure of the riser to the sun will dea total cable system is limited by the cable segcrease the ampacity of a cable in a vertical riser. ment that operates at maximum temperature.

1 3 4 – Se c t io n 4

4 that contain up to three curIncident solar radiation per rent-carrying conductors (with unit area is equal to 900 Sun loading will neutral) are kept more than watts/m2. This is a typical decrease riser one conduit diameter apart value for a sunny, midsummer (from surface to surface), no day throughout the United ampacity. correction of ampacity ratings States. The influence of solar need be made because of muheating in a riser application tual heating. This is because operating at a load factor of air circulation, even at an elevated ambient tem100 percent can reduce the current rating of perature of 40°C, will prevent convection heating cable by 15 percent for a completely closed riser from reducing the cable ampacity, so that it’s comand five percent for properly vented risers beparable to that of an isolated conduit in free air. tween day and night. Because ampacity tables The effect of the vertical conduit run on unlist conduit-in-air ampacity values at an ambient derground cable ampacity ratings is best shown air temperature of 40°C (104°F), solar de-rating by comparing its ampacity with four types of factors need be applied only during the hottest underground configurations. The comparison is days of the summer. Note that Table 4.1 gives shown in Table 4.6 for copper conductors and ampacity correction factors for ambient air temTable 4.7 for aluminum conductors. peratures other than 40°C and different conducTables 4.6 and 4.7 list the ampacity of single tor temperatures. This table can be used for riser and double three-phase circuits in trefoil pole applications. arrangement made up of two-conductor, singleRadiation heat transfer plays a large role in phase UD cables. Conductor sizes range from total heat dissipation from cable and riser sur4/0 AWG to 1,000 kcmil. The cables are direct faces. For maximum efficiency, riser material buried, buried in conduit, and vertical conduit in should be a light color (gray) to reflect some of air for the riser pole application. Conditions that the sun’s rays and to allow heat to be given off relate to the underground and riser pole installaby the riser surface at a higher rate. tions are shown below the ampacity values. The Cable in a large-diameter, vented riser will four circuit configurations are shown in Figure consistently operate at a lower temperature than 4.10 as configurations 1 and 4, and 3 and 5. will the same cable carrying the same current in The two tables show that, for the buried cona smaller conduit because the larger opening alduit installation configurations, the riser does lows more airflow through the riser. In addition, not limit the ampacity of the double circuit, but the larger surface area increases heat dissipation will limit single-circuit applications (shaded by convection and radiation. Conductor tempercells). For the direct-buried conditions, the vertiature differences between large- and small-diamcal run limits the installation for the cable sizes eter risers can range from 2°C to 15°C. and configurations shown in the shaded portion of the tables. Note that for the single-circuit, diSubstation Exits rect-buried case, the riser limits ampacity for all Substation exits are generally the highest ampaccable sizes and for both types of conductor. ity application of underground primary cable on ICEA and IEEE ampacity tables for conduit-inutility systems. Because cable in conduit has less air applications are different from their underload-carrying ability than does direct-buried ground counterparts. For air installations, the cable, both configurations applied in vertical ristables assume there is no wind and no sun loaders will be reviewed in more detail. ing and that the conduit is not ventilated. As Consider the condition in which two cable cirnoted in a previous subsection, Conduit in Air cuits in trefoil arrangement end on a double-circuit riser pole. When referring to IEEE 835-1994 for Riser Pole Applications, solar radiation was ampacity tables to select cable current ratings for found to decrease riser ampacity, whereas ventriser pole applications, use the tables for trefoil ing increases riser ampacity over the rating of a cable in isolated conduit in air. When conduits riser closed at both ends. Therefore, to prevent


4 TABLE 4.6: Ampacity Values—15-kV Cable, Trefoil Configuration, Copper Conductor.* Direct Buried** Cable Size

1 Circuit

2 Circuits

Riser Pole Vertical Conduit in Air

4/0

360

301

267

289

256

350

460

381

338

370

326

500

535

440

409

439

383

750

608

496

455

502

434

1,000

706

573

554

599

512

Underground

Buried in Conduit** 1 Circuit

2 Circuits

Riser Pole

90°C conductor temperature


100% load factor

40°C average air temperature

25°C ambient earth temperature

No solar radiation, venting, or wind

* Ampacity values are from IEEE Standard 835-1994. ** Circuit configurations are shown in Figure 4.10 as configurations 1 and 4, and 3 and 5, respectively.

TABLE 4.7: Ampacity Values—15-kV Cable, Trefoil Configuration, Aluminum Conductor.* Direct Buried** Cable Size

1 Circuit

2 Circuits

Riser Pole Vertical Conduit in Air

1/0

195

165

145

156

140

4/0

284

238

212

228

202

350

370

307

276

298

262

500

442

365

347

364

318

750

529

432

411

437

379

1000

623

506

508

528

452

Underground

Buried in Conduit** 1 Circuit

2 Circuits

Riser Pole



100% load factor

40°C average air temperature

25°C ambient earth temperature

No solar radiation, venting, or wind

* Ampacity values are from IEEE Standard 835-1994. ** The above circuit configurations are shown in Figure 4.10 as configurations 1 and 4, and 3 and 5, respectively.

the de-rating of riser pole (conduit-in-air) ampacity values resulting from solar effects, the riser must be open at the top and vented at the bottom. Conduit-in-air ampacity tables do not list dif-

ferent current values for different average air temperatures. Instead, ampacities are listed for 40°C, which is considered conservative for most installations. The tables also do not show am-

1 3 6 – Se c t io n 4

4 pacity variations caused by different load factors This added feature is particularly important if a for cables suspended in conduit exposed to air. U-guard backing plate is used, because this There is no load factor variation because there is arrangement has few, if any, significant air gaps no heat-sink cooling effect for the conductor/air (see Figure 4.9). For a double-circuit riser pole, system as exists for buried cable. In a riser, as U-guard should be placed on opposite sides of the load increases to a peak, the conductor temthe pole to prevent mutual heating and minimize perature increases much more quickly than if it the chance of simultaneous damage from vehicwere buried in soil. Thus, for riser applications, ular impact. the load factor is considered to be 100 percent, Regardless of whether the riser is conduit or because the thermal time constant for the U-guard, venting should only be installed where cable/air system is very short. it is required to obtain sufficient cable circuit amInstead of circular conduit, some utilities use pacity. This is because the venting fixtures are more U-guard for riser pole applications. U-guard, as expensive than normal conduit and require addithe name implies, is a U-shaped section with tional installation effort. Venting fixtures also slightflanged edges that is attached to a pole with lag ly reduce the security of the riser installation and bolts. It is used to cover and are more subject to damage by provide suitable protection for outside impact. the cable. U-guard usually Whether conduit or URiser vents should does not need to be vented at guard is used as a riser, it be installed only the base because an air space is recommended that a 90° is assumed to exist between elbow with a separate end where necessary for the pole and flanges to allow bell be installed three feet increased ampacity. enough air entry to produce a below grade level. This instalchimney-cooling effect. Howlation will ensure that the ever, many cooperatives install cable is at the proper depth a vent at the base of U-guard to ensure optimum near the riser pole. The elbow/bell end combiairflow and increased cable ampacity for all nation helps prevent cable damage during three-phase and most single-phase installations. pulling. After the cable is installed, it also helps protect against dig-ins around the base of the pole and minimizes conduit pressures on cable if soil shifts.

FIGURE 4.9: Single-Phase U-Guard Installation with Vented Base.

Emergency Overload Ratings For years, cables have been rated for operation at a maximum temperature of 90°C. When these type cables are loaded above a conductor operating temperature of 90°C for XLPE, TR-XLPE, and EPR insulations, the cable is considered overloaded. Many utilities and cable manufacturers are now specifying and rating cables for 105°C as a standard overload temperature. Cooperatives should weigh carefully the use of this rating, and consider as well the maximum operating temperature rating of cable terminations and joints. Overloading the cable will heat its insulation above its maximum operating temperature limits. The insulation temperature limits have been set by standards to maintain the integrity of the insulation for an increased life expectancy. Emergency


4 TABLE 4.8: Abstract of ICEA Standards for Maximum Emergency-Load and Short-Circuit-Load Temperatures for Various Insulations. Normal Operating Temperature (°C)

Emergency-Load Temperature (°C)

Short-Circuit Temperature of Cable Conductor (less than 30 seconds) (°C)

Thermosetting TR-XLPE and EPR

105

140*

250

Thermosetting TR-XLPE and EPR

90

130*

250

HMWPE

75

95

150

Insulation

* Operation at the emergency overload temperature of 130°C (266°F) and 140°C (284°F) shall not exceed 100 hours in any 12 consecutive months nor more than 500 hours during the lifetime of the cable. Note. Lower temperatures for emergency overload conditions may be required because of other types of material used in the cable and in the joints and terminations or because of cable environmental conditions.

operating temperature limits apply only to the infrequent higher loading of a line caused by an unplanned outage of a nearby cable or load sharing for a nearby substation. Standards state that the emergency overload conductor temperature of 130°C (or 140°C for the 105°C rating) should not be exceeded for more than 100 hours in any 12 consecutive months nor for more than 500 hours during the lifetime of the cable. Cable aging accelerates with high temperatures and accumulates over time in a way similar to aging in transformers. For these reasons, emergency overload ratings always specify both a temperature and a time limit for events over the lifetime of the cable (Aluminum Association Inc., 1989). The ratings have been derived from industry operating experience and could change as newer and better insulation materials become available. Emergency overload ratings are set by ICEA, NEMA, and ANSI/IEEE standards. Table 4.8 lists the emergency overload temperatures for the two types of insulation specified by the RUS, plus temperatures for outdated HMWPE when used as an insulation material. AMPACITY TABLES Table 4.1 lists the ampacities for single-phase UD cable direct buried and in conduit for copper and aluminum conductors. The three-phase ampacity tables and associated ampacity ratings for underground distribution cables provided in Appendix G are based on the following conditions:

• • • •

60-Hz. 15 kV, 25 kV, and 35 kV. Load factors of 75 percent and 100 percent. Three, two-conductor, concentric neutral, single-phase, primary UD cables. Installation configurations are shown in Figure 4.10.

• Conductors I Class B Stranding I Copper and Aluminum I 1/3 Concentric Neutral (1/6 for 1,000 kcmil) • Conductor Sizes I 1/0 AWG Solid Conductor—Aluminum Only—Full Neutral I 4/0 AWG Class B Stranding—Copper and Aluminum I 350 kcmil Class B Stranding—Copper and Aluminum I 500 kcmil Class B Stranding—Copper and Aluminum I 750 kcmil Class B Stranding—Copper and Aluminum I 1,000 kcmil Class B Stranding—Copper and Aluminum • Cable Specification I To Meet RUS Cable Specification 1728F-U1: N Insulation: EPR or TR-XLPE N Insulating Outer Jacket N Insulation Thickness: 220 mil @ 15 kV 260 mil @ 25 kV 345 mil @ 35 kV • All concentric neutrals are shorted and grounded at several points in the circuit, as per the National Electrical Safety Code.

1 3 8 – Se c t io n 4

4 36”

B

7.5”

7.5”

18”

Configuration 7

30”

30”

Configuration 6

36”

5”

5”

19”

18”

7.5”

Configuration 5

C

7.5”

7.5”

19”

A

Configuration 4

36”

Configuration 3

36”

Configuration 2

36”

Configuration 1

7.5”

19”

26.5”

19” 19” Duct Bank

19” 26.5” Duct Bank

FIGURE 4.10: Three-Phase Cable Installation Configurations.

• Ratings include dielectric loss and induced ac losses. • Conduit I Conduit used in Configurations No. 3, 5, 6, and 7 is Schedule 40, PVC conduit. Maximum fill requirements are 40 percent for three cables in a conduit per pending RUS Specification 1728F-U1. • Temperature Limitations I Ambient Soil = 25°C I Conductor = 90°C I Neutral (assumed) = 80°C I Conduit (assumed) = 70°C • Thermal resistivity (rho) of various materials was assumed as follows: I Soil = 90°C-cm/Watt I Insulation and Extruded Shields = 400°C-cm/Watt I Conduit and Duct = 480°C-cm/Watt I Concrete = 85°C-cm/Watt

• The ampacities for 15-kV class cable were calculated with 15 kV as the operating voltage. If 12.47 kV is used, the ampacities will be marginally higher (<1%). Adjustments for Changes in Parameters If the engineer needs to make certain changes in parameters to match them with actual site conditions or to do a sensitivity analysis on various parameters, the following formulas may be used.

Adjustment for Changes in Ambient Soil Temperature The ampacities in Appendix G are based on an ambient temperature of 25°C. To correct ampacities based on maximum conductor temperatures for different ambient temperatures, use the formula shown in Equation 4.3. The factors shown in Table 4.9 may be used


4 to correct ampacities based on maximum conductor temperatures for earth ambient temperatures of 20°C or 30°C.

Adjustment for Changes in Ambient Air Temperature To find ampacities for ambient air temperatures other than those found in the individual tables, multiply table values by the correction factors shown in Table 4.10. Equation 4.3 Tc – Ta' ×I Tc – 25

I' = where: Tc I Ta' I'

= = = =

Maximum conductor temperature from ampacity table Ampacity shown for Tc at ambient earth temperature of 25°C New ambient earth temperature Adjusted ampacity for ambient earth temperature Ta'

TABLE 4.9: Correction Factors to Convert from 25°C Ambient Soil Temperature to 20°C and 30°C. Ambient Temperature (°C)

Correction Factor (Maximum Conductor Temperature) 75°C

90°C

20

1.049

1.037

30

0.949

0.960

TABLE 4.10: Correction Factors for Various Ambient Air Temperatures. Source: Okonite Company, Engineering Data for Copper and Aluminum Conductor Electrical Cables, Bulletin EHB-90, 1990.

Adjustment for Change in Conductor Temperature The ampacities (I') for conductor temperatures other than those included in the tables (e.g., emergency conductor temperatures) can be approximated using the formula in Equation 4.4. When Tc' is greater than Tc, Equation 4.4 will give conservative values because it is based on the ratio of direct-current losses at the two temperatures, whereas the ratio of the ac conductor and concentric neutral losses to dc conductor losses decreases with increasing conductor temperature. For example, the ampacity at 110°C conductor temperature may be as much as five percent higher and at 130°C as much as 10 percent higher than values calculated from Equation 4.4. Deviations from true ampacities will depend on the conductor size, concentric neutral size, and installation configuration. Equation 4.4 is more precise for smaller conductors and higher resistance concentric neutrals (ICEA P-53-426, p. VII, May 1976). Figure 4.10 shows the seven cable installation configurations whose ampacities have been listed in the ampacity tables. Note: The ampacities listed in Appendix G are based on a conductor temperature of 90°C and an ambient soil temperature of 25°C. On the basis of these assumptions, many of the calculated current values may exceed the maximum permissible earth interface temperatures for various types of soils. Experience has shown that interface temperatures of 50°C and 60°C should be

Equation 4.4

Conductor Temperature (°C)

Ambient Air Temperature

I' =

30°C

35°C

40°C

45°C

50°C

75

0.97

0.92

0.86

0.79

0.72

85

1.06

1.01

0.96

0.90

0.84

90

1.10

1.05

1.00

0.95

0.89

100

1.17

1.12

1.08

1.03

0.98

110

1.23

1.19

1.15

1.11

1.06

125

1.31

1.27

1.24

1.20

1.16

130

1.33

1.30

1.27

1.23

1.19

150

1.42

1.39

1.36

1.33

1.30

Tc' – Ta τc + Tc ×I × Tc – Ta τc + Tc'

where: I' = Ampacity for conductor temperature Tc, in amperes Tc' = New or emergency conductor temperature, in °C Ta = Ambient earth temperature, in °C τc = Magnitude of the difference between 0°C and the zero electrical resistance of copper (234.5°C) or aluminum (228.1°C) Tc = Maximum conductor temperature from ampacity table, in °C

1 4 0 – Se c t io n 4

4 satisfactory for many types of soils. Unless the soil properties and moisture content of a particular installation are known, ampacity values should be chosen from the “Amperes at 60°C”

columns, rather than those from the “Amperes” columns. The three following examples illustrate the concepts covered in this section.

EXAMPLE 4.2: Single-Phase UD Cable Ampacities. A cooperative is planning to stock UD cable to meet the growing demand for new 12.47-kV underground installations. This cable will be used with 200-ampere class accessories. The cable will also be used to replace any faulted feeder sections on an as-needed basis. The conductor cable with the most installed circuit miles on the system is 1/0 aluminum. With this in mind, the engineer decided to check the ampacity of 1/0 cable for typical installations that exist on the system to find which cable sections could limit the current rating of an entire cable run. The cooperative direct buries its single-phase cable in all instances except for road crossings and riser pole installations. Go to the beginning of this section to find the ampacity rating for underground installations. Assume an operating conductor temperature of 90°C, soil rho = 90°C-cm/watt, and ambient soil temperature of 20°C. Using Table 4.1, find the ampacity of direct-buried TR-XLPE 1/0 aluminum cable: Ampacity = 264 amperes at 100% load factor As most single-phase circuits do not operate at 100 percent load factor, determine the cable rating at 75 percent and 50 percent load factor using the correction factors contained in Table 4.1: 75% LF = 1.08 × 264 = 285 amperes 50% LF = 1.16 × 264 = 306 amperes The cooperative’s standard installation practice for road crossings is to pull cable through conduit to speed cable change out if it fails. From Table 4.1, the cable rating in direct-buried conduit is as follows: 156 amperes at 100% LF 162 amperes at 75% LF (1.04 × 156) 167 amperes at 50% LF (1.07 × 156) It is assumed that the under-road section is long enough so there is no additional cooling effect from the direct-buried cable on either side of the road.

Because the ampacity ratings given in Table 4.1 are for an ambient soil temperature of 25°C, higher values can be expected if the soil temperature is actually 20°C. As Equation 4.3 indicates, the cable ampacity at 20°C can be found by multiplying the existing values by the correction factor: 90 – 20 = 1.0377 90 – 25 Cable ampacity for a soil temperature of 20°C is as follows: Load Factor

Direct Buried

In Duct

100%

274

162

75%

298

168

50%

318

173

For the riser pole cable section, the ampacity value is found in Table 4.1 under the “Duct in Air” column: Ampacity = 120 amperes at 40°C ambient and 100% LF This ampacity value is based on a riser that is closed at the top and bottom with no sun loading and no wind. Previous discussions have shown that venting conduit at top and bottom and leaving the top of U-guard open can increase riser ampacity, whereas solar radiation can reduce its rating. So if sun loading is considered, the riser must be properly vented or a de-rating factor should be applied to the 120-ampere riser rating. Note that solar de-rating will be a factor only for summer loading and when the temperature exceeds 100°F. This analysis shows that the riser pole limits the rating of the total underground circuit. At 100 percent load factor and 20°C ambient soil, direct-buried 1/0 cable ampacity of 274 amperes would be reduced by 57 percent and the 162-ampere rating of cable in duct would be reduced by 26 percent.


4 EXAMPLE 4.3: Emergency Overload Rating Cable in Protective Riser. From previous load studies and demand measurements, the engineer knows that the load factor on his heavily loaded loop-feed circuits has never exceeded 75 percent. Given this fact, determine the emergency overload rating of the cable in the protective riser. The conditions necessary to produce maximum current at a riser include a loop-feed installation with the open point near the center of the loop and a cable failure near the opposite riser pole. These conditions are relatively rare and represent an emergency situation that should last for only a short time. From Table 4.10, the conduit-in-air correction factor for an emergency overload conductor temperature of 130°C is 1.27 for an ambient

air temperature of 40°C. Therefore, Emergency Riser Rating = 1.27 × 120 = 152 amperes Because the cable is in a riser, no ampacity increase is allowed for 75 percent load factor. Comparing this value with the 75 percent load factor ampacity of the direct-buried and buried duct sections of the cable run shows the duct portion is overloaded by 10 percent and the directburied sections are well within their ratings. Note that, for simple radial feed circuits, the 90°C conductor ampacity rating of a riser should never be exceeded.

EXAMPLE 4.4: Three-Phase Substation Exit Ampacity. The same cooperative from Example 4.2 is planning to install a new 12/16/20-MVA transformer in an existing substation. The addition is needed to support expected load growth in the area and will replace an existing overloaded transformer. Four 12.47-kV feeders will be needed. A 600-ampere recloser will protect each feeder. Because of congestion around the substation, four underground circuit exits that will terminate on two double-circuit riser poles are planned. The cable for each underground exit must be sized to carry, under emergency ratings, the full load of one other circuit in case of a cable failure. Find the appropriate size cable for the application. Assume two three-phase circuits to a riser pole will be installed in two separate trenches. Each of the circuits to a given riser pole will provide emergency backup for the other. Each circuit will be in a single conduit in trefoil configuration, similar to configuration 3 of Figure 4.10. Maximum conductor temperature will be limited to 90°C, soil thermal resistivity (rho) will be 90°C-cm/watt, and load factor will be 75 percent. Maximum feeder loading assuming balanced feeders is approximately 260 amperes. For the contingency condition of a failed cable, the maximum short-time loading would be as follows: 2 × 260 = 520 amperes The smallest cable size to meet the emergency overload current can be found by first calculating the emergency correction factor for a conductor temperature of 130°C from Equation 4.4,

For aluminum conductor, I'130 =

130 – 25 228.1 + 90 × I90 = 1.198 × I90 × 90 – 25 228.1 + 130

For copper conductor, I'130 =

130 – 25 234.5 + 90 × I90 = 1.199 × I90 × 90 – 25 234.5 + 130

Find a copper or aluminum cable from the ampacity tables in Appendix G for configuration 3 (single direct-buried conduit with three conductors) whose 90°C, 75 percent load factor rating when multiplied by 1.2 gives a value approximating 520 amperes (520 ÷ 1.2 = 433 amperes). Cable ampacity ratings at 130°C conductor temperature are as follows: For Copper

For Aluminum

500 kcmil

750 kcmil

439 × 1.2 = 527 amperes

437 × 1.2 = 524 amperes

Continued

1 4 2 – Se c t io n 4

4 EXAMPLE 4.4: Three-Phase Substation Exit Ampacity. (cont.) The emergency rating of both cables is greater than the 520-ampere emergency requirement. Even if the two conduits had been installed within 18” of each other (Configuration 5), the single circuit in a trench ampacity table is the correct configuration to use in this instance because only one circuit will be energized during the emergency condition. Next, the riser pole current rating should be checked to see if it will limit the cable application. From Tables 4.6 and 4.7, the corresponding riser pole ratings are 409 amperes for copper and 411 amperes for aluminum. Both values are less than their respective buried conduit ratings (439 copper and 437 aluminum). The riser cable 130°C emergency overload rating at 40°C ambient air temperature can be found from Table 4.10.

Because the riser emergency rating is less than the buried conduit emergency rating, the riser cable section is the limiting element in the application. The application is a valid emergency situation if it is understood the overload condition will not exceed 100 hours in any 12-month period or 500 hours over the planned life of the installation. This substation exit application is covered by standards because it is an unplanned outage of a nearby cable. Note that the riser must be open at the top and vented at the bottom to provide additional ampacity above the values given in the tables and to compensate for any de-rating caused by solar heating.

500 kcmil copper = 1.27 × 411 = 522 amperes 750 kcmil aluminum = 1.27 × 409 = 520 amperes

• Ambient soil temperature, SECONDARY CABLE AMPACITY and Secondary cables carry power Primary and secondary • Installation configuration: at utilization voltage level from cable ampacities are I Direct buried the pad-mounted distribution I In duct. transformer low-voltage termiaffected by the nals to the service entrance The appropriate secondary same conditions. point for each consumer. The cable size is selected based on many conditions that affect the the amount of load the cable ampacity of primary cable also will serve. In a later subsection apply to secondary cable installations. Among titled Transformer Sizing for Single-Phase Transthese conditions are the following: formers for New Residential Loads, a procedure is • • • •

Maximum insulation operating temperature, Conductor resistance, Load factor, Soil thermal resistivity,

Equation 4.5

where: kVA1φ kVA3φ kVL-L I1φ I3φ

= = = = =

Single-Phase: I1φ =

kVA1φ kVL-L

Three-Phase: I3φ =

kVA3φ 3 kVL-L

Total load for single-phase applications Total load for three-phase applications Voltage line-to-line, in thousands of volts Single-phase current, in amperes Three-phase current, in amperes

outlined to determine the appropriate transformer size on the basis of the number of connected loads and the diversified demands of each load. Once the expected load and secondary operating voltage are known, the required ampacity for balanced loads can be determined from Equation 4.5. Once the secondary current load is calculated from Equation 4.5, ampacity tables can be consulted to select the proper cable size. Table 4.11 gives the allowable thermal loading for the most common secondary cable sizes in a buried environment for 100 percent load factor, 90°C maximum conductor temperature, 20°C ambient soil temperature, and 90°C-cm/watt soil thermal resistivity. After a secondary cable size is selected from ampacity tables, the application must be checked to ensure that voltage drop and flicker are within acceptable limits. These calculations are covered in detail in Appendix B.


4 TABLE 4.11: Typical Ampacities for Various Sizes and Types of 600-Volt Secondary UD Cable—Stranded Aluminum Conductors. Phase Conductors

Code Word

Size (AWG or kcmil)

Strand

Neutral

Insul. Thick. (mils)

Size (AWG or kcmil)

Strand.

Dimensions (mils) Insul. Thick. (mils)

SinglePhase Cond.

Complete Cable

Ampacity (amps)* Wt. per 1,000 ft (lb.)

Direct Burial

In Buried Conduit

DUPLEX Bard

8

7

60

8

7

60

262

524

76

70

55

Claflin

6

7

60

6

7

60

299

598

104

95

70

Delgado

4

7

60

4

7

60

345

690

138

125

90

TRIPLEX Vassar

4

7

60

4

7

60

345

745

191

125

90

Stephens

2

7

60

4

7

60

403

870

249

165

120

Ramapo

2

7

60

2

7

60

403

870

278

165

120

Brenau

1/0

19

80

2

7

60

512

1,106

387

215

160

Bergen

1/0

19

80

1/0

19

80

512

1,106

441

215

160

Converse

2/0

19

80

1

19

80

555

1,199

478

245

180

Hunter

2/0

19

80

2/0

19

80

555

1,199

535

245

180

Hollins

3/0

19

80

1/0

19

80

603

1,302

581

280

205

Rockland

3/0

19

80

3/0

19

80

603

1,302

651

280

205

Sweetbriar

4/0

19

80

2/0

19

80

658

1,421

709

315

240

Monmouth

4/0

19

80

4/0

19

80

658

1,421

796

315

240

Pratt

250

37

95

3/0

19

80

732

1,581

853

345

265

Wesleyan

350

37

95

4/0

19

80

831

1,795

1,118

415

320

Holyoke

500

37

95

300

37

95

956

2,065

1,544

495

395

Rider

500

37

95

350

37

95

956

2,069

1,597

495

395

QUADRAPLEX Tulsa

4

7

60

4

7

60

345

833

255

120

85

Dyke

2

7

60

4

7

60

403

973

342

155

115

Wittenberg

2

7

60

2

7

60

403

973

371

155

115

Notre Dame

1/0

19

80

2

7

80

512

1,236

534

200

150

Purdue

1/0

19

80

1/0

19

80

512

1,236

589

200

150

Syracuse

2/0

19

80

1

19

80

555

1,340

657

225

170

Lafayette

2/0

19

80

2/0

19

80

555

1,340

713

225

170

Continued

1 4 4 – Se c t io n 4

4 TABLE 4.11: Typical Ampacities for Various Sizes and Types of 600-Volt Secondary UD Cable—Stranded Aluminum Conductors. (cont.) Phase Conductors

Code Word

Size (AWG or kcmil)

Strand

Neutral

Insul. Thick. (mils)

Size (AWG or kcmil)

Strand.

Dimensions (mils) Insul. Thick. (mils)

SinglePhase Cond.

Ampacity (amps)*

Complete Cable

Wt. per 1,000 ft (lb.)

Direct Burial

In Buried Conduit

QUADRAPLEX (cont.) Swarthmore

3/0

19

80

1/0

19

80

603

1,456

798

250

195

Davidson

3/0

19

80

3/0

19

80

603

1,456

868

250

195

Wake Forest

4/0

19

80

2/0

19

80

658

1,588

974

290

225

Earlham

4/0

19

80

4/0

19

80

658

1,588

1,061

290

225

Slippery Rock

350

37

95

4/0

19

80

831

2,006

1,544

385

305

*Ampacity: 90°C conductor temperature, 20°C ambient temperature, rho 90, 100% load factor Note. Excerpted from Southwire Product Catalog, Section 16, pages 2, 3, and 4, 2003.

Pad-Mounted Transformer Sizing

PAD-MOUNTED TRANSFORMERS LOADING FOR NORMAL LIFE EXPECTANCY The distribution transformer is an essential comIn service, a transformer is not loaded continuponent of the underground distribution system. ously at rated kVA and at a constant temperature. Besides providing transformation from primary Instead, it goes through a daily load cycle with a to secondary voltages, it provides an area for short-time peak load occurring usually during the primary and secondary cable terminations, hottest part of the day. A varying load poses switching and surge protection equipment, and challenges in optimizing a transformer’s full-load overcurrent protective devices, all housed within capability without shortening its useful life. The the transformer enclosure. Because of increased capability of pad-mounted distribution transformUD usage, special pad-mounted distribution ers to carry loads under conditions other than transformers were developed with the above those used to establish nameplate ratings will be features. The term pad comes from the transreviewed later in this section. Because loading formers usually being located guides are based on the averon concrete slabs or pads age winding temperature rise (Fink and Beaty, 1987). above ambient, the load-carryLoading considerations Figure 4.11 shows a typical ing ability of a pole-type transfor pole- and padsingle-phase, pad-mounted former is basically the same as transformer with its cover that of a pad-mounted transmounted transformers open. Two primary bushing former. Standards make no are the same. wells are shown at the upper distinction between the two. left for use with load-break elAdditional information on bows. This dead-front configuloading distribution transformration allows a low-height design to be used in ers can be found in ANSI/IEEE Standard C57.91. residential areas and provides greater safety for Standards assign a distribution transformer a operating personnel. The secondary 240/120rated output that is expressed in kVA. The transvolt bushings with copper studs are shown on former is designed to carry this rated load continthe right. uously over its expected lifetime at an ambient


4 on the ability to transfer internal heat to the atmosphere. The capability of the cooling system to rid the transformer of heat depends mainly on the temperature differential between the tank and the ambient air because most pad-mounted transformers do not have forced-air cooling. The ambient air temperature is the most important element in determining how much load a transformer can carry because the temperature rise of the insulation for any load is added to the ambient temperature to determine the actual operating temperature of the transformer. To select daily peak loads from published loading guides, the engineer must predict what the temperature will be during the peaks. The probable ambient temperature for any future month can be estimated from historical weather data from the cooperative’s service area. ANSI/IEEE C57.91 gives two methods to predict temperature for the month involved. One is used to select loads for normal life expectancy and uses average daily temperature (defined as the average of all daily highs and all daily lows) for the month, averaged for several years. The other uses the maximum daily temperature (defined by the standard as the average of the high and low of the hottest day) for the chosen month, averaged over many years. Whenever these two methods are used, it is understood that, during any one day, the maximum temperature may exceed the average values found above. To be conservative, these FIGURE 4.11: Typical Dead-Front, Single-Phase, Pad-Mounted temperatures should be increased by 5°C beTransformer cause insulation does not recover fully after it is overheated. The maximum temperature over a 24-hour period should not exceed the average temperature of 30°C (86°F), without exceeding temperature by more than an average winding tempera10°C, which provides an adeture rise of 65°C. Under these quate safety margin. According conditions, insulation deterioAmbient temperature to the above factors, the estiration and transformer loss of mated temperature should not life are considered normal. Inand load factor set be exceeded for more than a dustry experience has shown transformer loading. few days per month; however, that normal life expectancy if it is, the transformer will not under these conditions should be adversely affected by the be at least 20 years. small incremental loss of life. Heat gain within a transformer is caused by Standard loading tables give ambient temperano-load and winding losses. Keeping the tempertures in 10°C intervals. Estimating peak loads ature rise of the windings below 65°C depends

1 4 6 – Se c t io n 4

4 EXAMPLE 4.5: Average Daily Temperature Selection for a Summer-Peaking Utility. The procedure to select the average daily temperature for loading distribution transformers is shown in this example. ABC Cooperative is located in the Southeast. As part of an operations review process, the manager and engineer decided to establish a formal procedure to select the proper size pad-mounted distribution transformers for an expected surge in underground installations in its service area.

TABLE 4.12: Average Temperatures for July and August Averaged for the Previous 10 Years. Month

Average (°F)

Average (°C)

July

80.6

27.0

August

81.0

27.2

Average of Temperatures

80.8

27.1

The first step in determining the maximum load each transformer can carry is to select an approximate ambient temperature that would be expected on the peak day. This summer peaking cooperative obtained the average July and August temperatures for the previous 10 years from the Weather Bureau of the U.S. Department of Commerce. Table 4.12 averages the temperatures found. Adding 5°C to the average, as recommended by ANSI/ IEEE C57.91, gives 27.1°C + 5°C = 32.1°C, which normally should be used for any transformer loading studies. The standard also specifies that the maximum temperature over a 24-hour period should not exceed the

average temperature selected by more than 10°C. Thus, the maximum temperature should not exceed 27.1°C + 10°C = 37.1°C (98.8°F). However, the engineer found that, in the hot summer of 1987, on many days the temperature reached 100°F (37.8°C) or more. If an actual maximum daily temperature in recent memory has been greater than 10°C above the maximum temperature averaged for the previous 10 years, it is suggested that that higher temperature be used in your calculations. To allow for the probability of 100°F days occurring, the engineer increased the 32.1°C average temperature selected previously by 0.7°C—the difference between the calculated plus-10°C maximum (37.1°C) and the actual high temperature (37.8°C)—thus using 32.8°C as the temperature to be used in the study.

Other items that can affect between the given temperapad-mounted transformer tures in a table is allowed. Altitude, tank finish, cooling are altitude and tank Peak loads obtained in this and ventilation affect finish. At higher altitudes, the way are accurate enough for air is not as dense; this dethe ambient temperatures depad-mounted creases cooling efficiency. rived from the above example. transformer cooling. Above 3,300 feet, a transHowever, extrapolation beyond former kVA rating should be the range of values shown in reduced approximately 0.4 the tables is not recommended. percent for each 330 feet of additional altitude. The engineer can perform the same type of The ability of a transformer to radiate heat is ambient temperature analysis for winter months affected by its paint finish. Some metal flake if the transformers are experiencing winter paints, like aluminum, reflect heat from direct peaks. The standard does not deal with the elecsunlight quite effectively; however, they do not tric heating load that will be greatest during the allow heat to escape as efficiently. Because most coldest days of the month, so results will be transformer heat is produced internally, metalconservative. When the ambient temperature based paints actually increase the temperature study produces a result below 0°C, the loading rise in most instances (Lee 1973). The subject of limits from the 0°C columns should be used inpaint finish is mentioned only in connection stead of extrapolation.


4 lasting from a few minutes to a few hours. A similar cycle is repeated every 24 hours. This characteristic allows the transformer to be operated at loads exceeding its continuous kVA rating during short peaks. Two main characteristics of the transformer permit the overload to be carried without decreasing normal life expectancy. The first characteristic is the thermal time constant, which ensures that the internal oil temperature increases slowly after a rapid change in load. This fact is important because of the limitation that the winding hot-spot temperaLOAD CHARACTERISTICS ture places on the ability of the The normal load duration transformer to carry an overload curve of a typical padShort peak overloads without insulation damage. For mounted transformer with a step change in load, the conmore than one service concan be carried ductor temperature at the hottest nected to it consists of a relawithout loss of spot in the winding increases to tively low load during most of transformer life. its maximum value very quickthe day, with one peak load ly. However, hot-spot and total conductor temperature are held down until the thermal time constant is exceed150 ed, which could be three to 10 hours, depending 140% Peak Load on preload conditions. Pad-mounted transformers are now designed to operate continuously at 70% Initial Load 100 a winding hot-spot temperature of 110°C. The second characteristic is the thermal aging of transformer insulation. Hot-spot temperatures 1 Hour 50 above 110°C can be carried for short times withActual Load out shortening the expected life of the transformer, as long as they are followed by longer 0 12 6 12 6 12 periods of operation below 110°C. Elevated temAM PM Time (Hours) peratures do not cause insulation failure, but only increase the rate of its deterioration when FIGURE 4.12: Actual Load Cycle and Equivalent Load Cycle. they are prolonged. It follows that a pad-mounted transformer lightly loaded before a peak will have a lower hot-spot 150 temperature than one carrying full load before 137% Peak Load the same peak. Therefore, the shape of the load curve over a 24-hour period can greatly affect Transformer Rating 100 what peak load may be carried by a transformer. If a daily load duration curve for a single 50% Initial Load transformer was plotted from data collected by 50 an interval demand recorder, it would be similar 2 Hours to the curve in Figure 4.12. To use loading guides provided in the stan0 12 6 12 6 12 dard, change the actual load duration curve into AM PM Time (Hours) a thermally equivalent, simple rectangular load cycle as shown in Figure 4.13. FIGURE 4.13: Thermal Equivalent Load Cycle. Load as Percentage of Transformer Rating

Load as Percentage of Transformer Rating

with refinishing transformers in the field. The engineer should be careful that the paint selected is a standard pad-mounted transformer finish with good radiating properties. Proper ventilation should always be considered when siting a pad-mounted transformer and after installation to allow the cooling system to function at peak efficiency. Care should be taken to allow for air to circulate freely around the unit at all times.

1 4 8 – Se c t io n 4

4 This conversion is done by deriving the values for the initial load and the peak load. These values may be approximated by the formulas shown in Equations 4.6 and 4.7 (ANSI/IEEE C57.91-1981). Equation 4.7 shows the formula for the equivalent peak load.

Estimating the duration of the peak has considerable inConsider preload fluence over the rms magnitude of the peak load. Caution conditions when should be used to not overesloading transformers. timate on-peak time. If the duration is overestimated, the rms peak load may be far below the maximum peak demand. After the equivalent peak load has been determined, a loading guide—such as the one in Equation 4.6 Table 4.13—may be used to pick a transformer size to supply the expected daily loading. It can also be used to determine whether or not an exEquivalent Initial Load = 0.29 L12 + L22 + L32 + ... L122 isting transformer will supply the listed daily peak loads and a 20-year minimum life exwhere: L1, L2, etc. = Average load by inspection for each pectancy. The ambient temperature to use in the 1-hour interval of the 12-hour period loading guide is the average daily temperature preceding the peak transformer load determined using the procedure outlined in a previous subsection, Loading for Normal Life Expectancy. Equation 4.7 The preload level given in the tables is based on the transformer nameplate rating and is not a 2 2 2 2 L1 t1 + L2 t2 + L3 t3 + ...Ln tn Equivalent Peak Load = percentage of peak load. Example 4.6 illustrates t1 + t2 + t3 + ...tn this principle. Note that even under 0°C ambient conditions where: L1, L2, . . . = The various load steps as a percentage, that might apply for winter-peaking studies in per unit, actual kVA, or current cold-climate areas, a maximum loading above t1, t2, . . . = Respective durations of these loads

TABLE 4.13: Daily Peak Loads Per Unit of Nameplate Rating for Self-Cooled Oil-Immersed Transformers to Give Minimum 20-Year Life Expectancy. Continuous Equivalent Load as Percentage of Rated kVA Preceding Peak Load 50% 75% 90% Ambient (°C) Ambient (°C) Ambient (°C)

Peak Load Time in Hours

0

10

20

30

40

50

0

10

20

30

40

50

0

10

20

30

40

1

2.52

2.39

2.26

2.12

1.96

1.79

2.40

2.26

2.12

1.96

1.77

1.49

2.31

2.16

2.02

1.82

1.43

2

2.15

2.03

1.91

1.79

1.65

1.50

2.06

1.94

1.82

1.68

1.52

1.26

2.00

1.87

1.74

1.57

1.26

4

1.82

1.72

1.61

1.50

1.38

1.25

1.77

1.66

1.56

1.44

1.30

1.09

1.73

1.62

1.50

1.36

1.13

8

1.57

1.48

1.39

1.28

1.18

1.05

1.55

1.46

1.36

1.25

1.13

0.96

1.53

1.44

1.33

1.21

1.02

24

1.36

1.27

1.18

1.08

0.97

0.86

1.36

1.27

1.17

1.07

0.97

0.84

1.35

1.26

1.16

1.07

0.95

Note. For transformer operation above 50°C or below 0°C, contact manufacturer. Peak loads shown assume 0.0137% per day loss of life for normal life expectancy. The ambient temperature to use in the table is the average temperature over a 24-hour period, with the maximum temperature not exceeding the average temperature by more than 10°C. Excerpted from Table 5, page 20, ANSI/IEEE C57.91-1981.


4 EXAMPLE 4.6: Selection of Maximun Permissible Transformer Per-Unit Loading. Load-pattern studies of pad-mounted transformers in a certain area revealed that typical 12-hour preload levels were 50 percent of peak load levels. The peak loads had a duration of two hours and the ambient temperature for the area was calculated at 30°C. The engineer needs to estimate the maximum permissible per-unit (pu) loading for the transformers to maintain normal life expectancy. The per-unit loading shown in Table 4.13 under 50 percent preload, 30°C ambient, and two-hour peak duration is 1.79 pu. However, if this loading level is permitted, the preload level will become 0.5 × 1.79, or 0.9 of the transformer nameplate rating. Therefore, conservatism requires that the engineer take the per-unit loading from the tabulated figures under 90 percent preload conditions, which will lead to a maximum loading of 1.57 pu. This change shifts the preload level to about 79 percent. A load somewhat higher than 1.57 pu is permissible. However, it should not be higher than the 1.68 pu figure shown under the 75 percent preload conditions. By interpolation, the engineer can estimate a final result of 1.63 pu loading.

If a transformer is being severely overloaded for extended periods, its life expectancy is being shortened and excessive conductor losses will be increasing operating costs. These conditions can be allowed during emergencies, but, under most conditions, they should not be continued for a sustained period of time. The NRECA Loss Management Manual thoroughly covers the issue of loss-optimal loading of distribution transformers. Engineers should definitely consult this manual before establishing final policies on loading pad-mounted transformers.

TRANSFORMER SIZING FOR SINGLE-PHASE TRANSFORMERS FOR NEW RESIDENTIAL LOADS Transformer loading is further complicated because loading levels are difficult to estimate for transformers serving residential consumers. Engineers have tried different methods to estimate the 1.8 pu cannot be justified from the tables if prepeak kVA load of a group of single-family living load conditions are 50 percent or more of peak units. However, many varying circumstances, load. This maximum exists because the 90 persuch as the sizes and types of electrical applicent preload level is the largest tabulated. This ances used, cause the load-estimating procedure analysis shows that some of the very large perto become somewhat complicated. Table 4.14 unit values shown in ANSI/IEEE Standard C57.91 shows a sample load-estimating guide for a tables are not particularly practical. southeastern utility. Cooperative engineers should It is not practical or economical to conduct an not use it to estimate transformer loading on in-depth study on every transformer suspected their own systems because diversity factors, of being overloaded. In fact, for small transformloads, and demands are different for every utilers, the cost of an individual detailed analysis ity’s service area. For example, the resistance could exceed the price of the transformer. Howheating diversity factors in this method apply to ever, if an overload is expected on a large threea semicoastal southern climate and may not acphase, pad-mounted transformer, investigation curately reflect conditions in other climates. would obviously be warranted. Also, the air conditioning efficiencies in your Loading levels applied to transformers should area may differ from those used in the developbe kept within those of Standard C57.91-1981. ment of this chart. The table is included as an Doing so protects not only the transformer example to demonstrate that similar tables windings but also ancillary components on the would be useful or can be developed from martransformer. Manufacturers design items such as keting and load research data. bushings, internal connections, Table 4.14 can be used to and fuse protection assuming estimate the diversified dethat the transformer loading mand for a group of totally Use a loading will not exceed Standard electric homes. The first step is guide developed C57.91-1981 levels. This coorto determine the number of dinated design is noted in consumers connected to the for your particular ANSI/IEEE C57.12.00-2000, transformer and select the corservice area. Section 4.2, and C57.91-1995, responding diversity factors Section 8.2.1. from Chart 1. The second step

1 5 0 – Se c t io n 4

4 TABLE 4.14: Application of Single-Phase Distribution Transformers to Serve Residential Consumers—Sample Loading Guide. Diversity, Load, and Demand Charts Chart 1

Chart 2

Diversity Factor D

Standard House Loads (kVA) Typical Residence Size (Square Feet)

Number of Consumers in Group (X)

Total Electric (TE)

Air Conditioning (A/C)

1

1.00

1.00

2

0.85

0.85

3

0.74

0.83

4

0.66

0.80

5

0.61

0.77

6

0.57

0.75

Chart 3

7

0.54

0.73

8

0.52

0.72

Equivalent kVA Demand for Houses With Resistance Heat

9

0.50

0.71

10

0.49

0.70

11

0.47

0.70

12

0.46

0.69

13

0.45

0.69

14

0.43

0.68

15

0.42

0.68

16

0.41

0.67

17

0.39

0.67

18

0.38

0.66

19

0.38

0.66

20

0.37

0.65

is to find the base kVA load for one consumer using Chart 2. The chart row labeled “TE” gives the base total electric load related to house size. The “A/C” row gives the air-conditioning load for the air-conditioner sizes shown. The equivalent kVA demands for various resistance strip heaters are listed in Chart 3. To find the load for a group of consumers, multiply the kVA values from Charts 2 and 3 by the appropriate diversity

Type of Load

1,500

2–3,000

5,000+

TE

4.3 kVA

5.7 kVA

7.5 kVA

Typical Air Conditioner Size (Tons) Type of Load

3

4

5

A/C

3.8 kVA

5.1 kVA

6.3 kVA

kW Rating

kVA Demand

5.0

5.0

7.5

6.5

10.0

8.0

15.0

10.5

20.0

14.0

Note. Values in the charts were excerpted from the South Carolina Public Service Authority (Santee Cooper) Distribution Engineering Reference Manual dated February 2, 1987.

factors from Chart 1. Diversity factors depend on the number of consumers in the group. To determine whether transformer size is set by the summer or winter load, do the calculation with air-conditioner load and then with resistance heat load. Equation 4.8 gives the total load (LX) for X identical consumers. Example 4.7 clarifies the procedure.


4 Equation 4.8 LXSummer = X[(TE Load)(DX(TE)) + (A/C Load)(DX(A/C))]kVA LXWinter = X[(TE Load)(DX(TE)) + (Heat Load)(DX(A/C))]kVA where: LX X TE Load DX(TE)

= = = =

Total load for X identical consumers, in kVA Total consumers in group Base total electric house load from Chart 2, in kVA Diversity factor D for X consumers from Chart 1, TE column A/C Load = Base air-conditioner load from Chart 2, in kVA Heat Load = Base resistance heat load from Chart 3, in kVA = Diversity factor D for X consumers from Chart 1, DX(A/C) A/C column

Example 4.7 assumes the transformer full-load rating, corrected for ambient temperature, can be up to 140 percent of its summer loading. The ability to carry more load in the winter is justified because the heating load factor is much lower than the cooling load factor for the assumed transformer service area. Cooler ambient temperature in winter also increases transformer loading capabilities. Each cooperative must set its own percentage loading limit based on experience. Before the transformer is installed, its size should be checked to see if it meets cooperative voltage drop and flicker criteria. These calculations are covered in Appendix B, “Transformer and Secondary Voltage Drop.”

EXAMPLE 4.7: Pad-Mounted Transformer Sizing for New UD Residential Consumers. Assume four totally electric, 1,500-sq.-ft. homes are to be fed from the secondary of a pad-mounted transformer in a new subdivision. All homes have identical electrical appliances, three-ton (36,000-Btu) air conditioners, and 7.5-kW resistance heaters. Select the transformer size that will serve the summer and winter loads and has a 20-year life expectancy. Pad-mounted transformers to choose from are rated 25, 37.5, and 50 kVA. First, select the diversity factors from Chart 1: X = 4 consumers in groups D4(TE) = 0.66 D4(A/C) = 0.80 Second, choose the base TE load and A/C load for a single house from Chart 2: TE Load = 4.3 kVA A/C Load = 3.8 kVA From Equation 4.8, the total summer load is 23.52 kVA, as calculated: Summer L4 = 4 [(4.3)(0.66) + (3.8)(0.80)] = 4[2.84 + 3.04] = 23.52 kVA A 25-kVA transformer is the proper size to choose, as no new houses will be added to the transformer. The total winter load is calculated the same way by replacing the airconditioning load with the strip heater load from Chart 3. The A/C

diversity factor is applied to the heating load in this instance. Because the ambient temperature will be lower in the winter, it is assumed the transformer will carry up to 140 percent of its summer peak load for short periods without undue loss of life. For the winter peak, the TE load component of the total load is the same as before: TE Load = (4)(4.3)(0.66) = 11.35 kVA The 7.5-kW strip heating component of total demand is then 4(6.5)(0.8) = 20.8 kVA Total winter diversified demand is equal to Winter L4 = 11.35 + 20.8 = 32.15 kVA The ratio of winter to summer load is then Ratio =

32.15 = 137% 23.52

Because the ratio is below 140 percent, the transformer size will be set by the 23.52-kVA summer load. The 25-kVA unit is still the proper transformer to install. (Note: Keep in mind this example is based on a methodology used by a southeastern U.S. utility and should be modified for use in other climates.)

1 5 2 – Se c t io n 4

4 Another important concern is initial loading versus future loading when load growth is expected. For many UD areas, significant load growth is not expected for individual transformers because the number of living units per transformer is set in the development plans. The modern trend in housing construction is to install all heavy appliances and heating, ventilating, and air-conditioning (HVAC) equipment in a dwelling before initial occupancy, so any growth beyond the initial level is expected to be insignificant. Even when engineers expect load growth, they seldom accurately know the rate of growth. Although complicated formulas exist for economic sizing of transformers based on load growth, use of these formulas is meaningless if the growth rate is not accurately known. A simple procedure is recommended, such as sizing the transformer for the load that is estimated to be present 10 years in the future. TRANSFORMER SIZING FOR THREE-PHASE TRANSFORMERS FOR NEW COMMERCIAL AND INDUSTRIAL LOADS Three-phase transformers—required to render service to commercial and industrial consumers—represent a significant investment for the average cooperative. As such, care should be taken in selecting transformers sized to minimize cost and losses, while providing reliable service. Sizing transformers for these type installations is not an exact science and requires sound judgment and previous experience, similar to the philosophy involved in sizing single-phase transformers. Local geographical and climatological conditions must be considered, as they play a significant role in sizing equipment. This subsection presents three generally accepted methods of sizing transformers that most cooperatives and utilities have used over the years: 1. Previous demands on similar loads, 2. Watts-per-square foot demand factors, and 3. Diversified connected load analysis. It is suggested that an analysis be made using all three methods, if possible, as a crosscheck to validate the final selection of a properly sized

unit. Further analysis using both knowledge of specific types of loads and experience anticipating the likelihood of growth in consumer demand is recommended. Method I: Previous Demands on Similar Consumers Many commercial establishments are part of large company chains that establish new facilities (or franchises) based on similar building footprints, using the same makeup of electrical devices. Convenience stores, supermarkets, drug stores, fast-food restaurants, and discount department stores, for example, have branch stores that result in very similar demands, provided geographical influences are similar. The only differences in some of these installations are whether or not natural gas, propane, or other heat source is either available or economically feasible. A starting point (or a double check) in sizing transformers for these type loads is to contact other cooperatives (or IOUs) to obtain historical demand data (both summer and winter peaks) for similar stores of the same relative size. Care should be taken for loads greater than 300 to 400 kW, as even similar stores can operate differently because of local usage patterns. Care should also be taken to evaluate power factors of loads for larger units, if such information is available from meter readings. If power factor readings (or both kW and kVAr readings) are available, then the transformer size can be selected to account for power factor by using either of the following formulas: 1. kVA2 = kW2 + kVAr2 or kVA = kW2 + kVAr2 2. Power Factor = kW kVA

or kVA =

kW PowerFactor

Method II: Watts-Per-Square-Foot Method Electrical demands for commercial and industrial buildings can be analyzed by evaluating typical watts-per-square-foot factors that have been established by utilities and design professionals over the years. Although these factors can vary over different geographical areas of the country as a result of climate factors and building practices, the basic values listed in Table 4.15 are typical of most areas of the continental United States.


4 TABLE 4.15: Typical Watts-Per-Square-Foot Factors for Commercial Buildings. Watts per Square Foot* Type Facility

Winter

Summer

9.2

6.3

Offices (less than 100,000 square feet)

10.0

8.3

Offices (more than 100,000 square feet)

7.7

6.4

Churches

9.7

6.2

Convenience Stores

13.0

12.7

Department Stores

6.9

5.6

Medical Clinics

11.3

8.6

Grocery Stores

10.1

10.4

Restaurants (fast-food)

45.8

41.5

Restaurants (fast-food/gas)

28.0

25.4

Restaurants (family)

27.3

21.9

Variety Stores

10.2

7.1

Schools

10.2

5.6

Motels

7.6

4.6

Banks

*All-electric, unless otherwise noted

TABLE 4.16: Typical Electrical Load Power Factor Values. Type Facility

Approximate Power Factor*

Restaurants

85%

Grocery Stores

85–90%

Office Buildings

90%

Retail Department Stores

90%

Residential Loads

95%

Lighting (HID)

95%

Motors That Operate at Full Load

80–85%

Motors That Operate at Less Than Full Load

50%

Sawmills

65%

Industrial Plants With Heavy Motor Load *If consumers have their own capacitors, higher values will result.

65–70%

Keep in mind that these factors are typical of loads analyzed in many areas of the country and can vary somewhat. They are a good approximation to be used as a double check of other analytical methods. Also remember to convert the calculated kW to kVA using power factor information. If the developer of the new facility cannot provide valid power factor information, Table 4.16 will assist in this effort. Again, these power factor values are typical of a number of cases sampled. Method III: Summation of Diversified Connected Loads The most analytical method available to predict the actual demand of a new consumer’s installation is to total the individual connected loads and apply diversity factors to multiple quantities of similar loads, and to the different types of load, to predict the effective actual demand. The philosophy here is that not all connected loads will operate simultaneously, as a result of cycling off and on by some automatic system (a thermostat, for example), or as a result of the operation inherent with the facility. It is important to gather all the connected load information from a consumer, both the types of loads and the quantities of similar electrical devices. For example, a restaurant may have five roof-top air conditioner units, one for each of five zones of the interior space. Further discussion with the consumer may indicate use of a demand-side monitoring system that cycles the HVAC units, so no more than three units can run at any one time. While this type of system will reduce the demand at any given time to the load imposed by three units, the result may be an increase in the customer’s load factor, which will tend to increase the required size of the transformer. Multiple kitchen appliances may, as well, be used in shift operations, with only portions of the devices operating together. The more information that may be gathered about how electrical devices will be operated, the more accurately an anticipated demand can be calculated. Following are other items to be taken into account while accumulating electrical load data for diversification:

1 5 4 – Se c t io n 4

4 • The larger of heating or air conditioning should be used, but not both. • Do load controllers limit the quantity of any devices running simultaneously? • Do certain devices, such as dishwashers, operate only on off-peak times, such as at the end of a shift? • How many portable appliances are planned to be connected to convenience outlets? • At what temperatures are refrigeration units to be operated (e.g., coolers versus refrigeration units versus deep freezers)? • Are all exterior lights to come on through photosensitive control? • Are water heaters multiple-element or load-controlled? • Is any capacity currently listed on electrical drawings as “spare” to be actually used in the near future, or not at all? Table 4.17 is a listing of typical types of loads for commercial/industrial applications, and the TABLE 4.17: Typical Electrical Load Demand Diversity Factor Values. Type of Equipment

Demand Diversity Factor

Air Conditioning (less than 100 tons) Note: 1 ton = 1.5 kW

75%*

Air Conditioning (more than 100 tons) Note: 1 ton = 1.0 kW

75%*

Electric Heating

75%*

Computers

75%

Electric Cooking Appliances

35–40%

Lighting

70–80%

Miscellaneous

35%

Motors (less than 40 Hp)**

40%

Motors (more than 40 Hp)**

25%

Receptacle Load Refrigeration Water Heating “Spare”*** * Use the larger of heating or cooling, but not both. ** Does not necessarily apply to industrial applications *** Consider “spare” only for specific needs.

10–15% 60% 40–50% 0%

typical diversity that is generally taken with respect to actual peak demands. Table 4.18 is a typical listing of the electrical connected loads associated with a new restaurant and how the loads can be tabulated to apply diversity factors so that an anticipated peak demand can be computed. Note that this demand should include the larger of heating or cooling loads, or, if necessary, a separate winter peak demand (with heating loads) and summer peak demand (with cooling loads) can be computed. Air handling units should be included in both listings. Once the kVA demand is determined, decide how large the transformer should be based on the sizes available. Some decisions will be fairly easy, whereas others fall into a gray area when demand could fit the top range of one size or the bottom range of another. The proper transformer size to be used for a calculated demand should be selected on the basis of the transformer’s ability to withstand short-term overload conditions, just as was discussed with single-phase units on residential applications. Consistent with the per-unit loading guide discussed in this section (Table 4.13), three-phase transformers are capable of similar short-term overloads (again, depending on the duration of the short-term peak and the relative loading level of the transformer for the period of time before the overloading condition). Table 4.19 lists typical commercial/industrial consumers and the duration typically found for short-term overloads. It is essential that information be obtained from the consumer to substantiate these peak durations or to determine that shorter or longer overload periods should be used. Once this information has been determined, the overload capacity of standard transformer sizes should be reviewed, based on local ambient temperature ranges. Listed in Table 4.20 is a typical cooperative’s overload factors, both summer and winter, based on ANSI/IEEE C57.911981 Table 5 (Table 4.13 in this manual). Note that the table lists both summer and winter overload factors, based on the typical ambient temperatures of the winter and summer months. In Table 4.20, 10°C has been chosen for the winter ambient, and 40°C has been chosen for the summer ambient. As a method of practical conservative


4 TABLE 4.18: Estimated Electrical Demand (Summer) and Energy Consumption (Sample Family Restaurant). Item No. 1A 1B 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

Load Description

Quantity

Unit Load (kW)

Total Load (kW)

Demand Diversity

Total Demand (kW)

Roof-Top Air Conditioning Units* Heat Pump Strip Heaters ** Baked Potato Oven Potato Warmer Heat Lamps Warming Tray Pie Safe Coffee Maker Soda Fountain Ice Machine (Continuous Use) Ice Machine (Infrequent Use) Cooler Range Hood Freezer Cold Table Iced Tea Maker Microwave Warming Tray Toaster Refrigerator Cooler Table Steamer Dishwasher Prep Cooler Beverage Cooler Vegetable Cooler Outside Freezer Outside Lighting Outside Lighting Kitchen Lighting Dining Area Lighting Dining Area Lighting Heat Lamps/Warming Tray Coolers Under Bar Coolers Under Bar Television Neon Signs Video Game Roadway Sign Total

5 2 1 1 2 1 1 2 2 1 1 1 1 1 1 2 1 1 1 1 1 2 1 1 1 1 1 2 2 20 14 6 1 2 1 1 4 1 1

8.12 15.34 11.00 0.80 0.25 1.50 0.73 1.80 0.25 1.70 7.50 1.25 0.90 1.09 1.73 1.52 1.50 1.59 2.83 1.08 0.76 0.75 8.61 2.89 3.76 3.76 2.38 0.40 1.00 0.16 0.10 0.13 1.56 0.84 0.42 0.45 0.05 0.16 0.80

40.60 30.68 11.00 0.80 0.50 1.50 0.73 3.60 0.50 1.70 7.50 1.25 0.90 1.09 1.73 3.03 1.50 1.59 2.83 1.08 0.76 1.50 8.61 2.89 3.76 3.76 2.38 0.80 2.00 3.20 1.40 0.75 1.56 1.68 0.42 0.45 0.20 0.16 0.80 151.17

0.60 0.00 1.00 1.00 1.00 1.00 1.00 0.50 0.50 1.00 0.00 0.60 1.00 0.60 1.00 0.50 0.00 1.00 1.00 0.60 0.60 0.50 1.00 0.60 0.60 0.60 0.60 1.00 1.00 1.00 1.00 1.00 1.00 0.50 0.50 1.00 1.00 1.00 1.00

24.36 0.00 11.00 0.80 0.50 1.50 0.73 1.80 0.25 1.70 0.00 0.75 0.90 0.65 1.73 1.52 0.00 1.59 2.83 0.65 0.45 0.75 8.61 1.73 2.25 2.26 1.43 0.80 2.00 3.20 1.40 0.75 1.56 0.84 0.21 0.45 0.20 0.16 0.80 83.10

* Load controlled ** Winter use only

1 5 6 – Se c t io n 4

4 TABLE 4.19: Estimated Peak Duration. Type of Business

Time (Hours)

Type of Business

Time (Hours)

Fast Food

8

Restaurants

4

Grocery Stores

8

Hotels

4

Large Office Buildings

8

Small Office Buildings

4

Large Retail Stores

8

Small Retail Stores

4

8

Schools

4

Other Commercial

4

Convenience Stores Industrial Plants

24*

* The peak durations may be less, but use this number with the loading table unless the customer can provide information that is different.

TABLE 4.20: Transformer Loading Capability Table.*

Peak Duration (Hours)

Summer Loading Capability** (% of kVA Rating)

Winter Loading Capability*** (% of kVA Rating)

4

130%

166%

8

113%

146%

24

97%

127%

* From ANSI/IEEE C57.91-1981 Table 5, based on 90% prior loading ** Based on 40°C ambient *** Based on 10°C ambient

engineering practice, 90 percent prior loading has been chosen for a safety factor. If the actual prior loading can be substantiated, then 75 percent or 50 percent prior loading per-unit values could be used. On the basis of these per-unit overload factors, the standard sizes of pad-mounted transformers in Table 4.21 can carry short-term overloads as listed for respective winter and summer ambient conditions. Care should be taken to not underestimate duration peaks and to apply proper ambient temperatures. Sizing transformers is not an exact science. However, by using the guidelines in this section, along with gaining experience from the local ambient conditions, one can become more effective in sizing transformers and the process will become less confusing. Some of the main keys to sizing a transformer are the following: • Understanding what affects a transformer’s loading capability (ambient temperature, load cycles, etc.). • Properly estimating the load (similar accounts, diversity, watts/square foot, etc.). Estimating the load is the largest factor in sizing a transformer correctly. If this part is completed correctly, most of the work is done.

TABLE 4.21: Typical Three-Phase Pad-Mounted Transformer Capacities—Short-Term Overload Capabilities (in kVA).*

Transformer Nameplate

4-Hour Peak Overload Summer Winter Capacity** Capacity***



75

97.5

124.5

84.8

109.5

72.8

95.3

112

145.6

185.9

126.6

163.5

108.6

142.2

150

195.0

249.0

169.5

219.0

145.5

190.5

225

292.5

373.5

254.3

328.5

218.3

285.8

300

390.0

498.0

339.0

438.0

291.0

381.0

500

650.0

830.0

565.0

730.0

485.0

635.0

750****

975.0

1,245.0

847.5

1,095.0

727.5

952.5

1,000****

1,300.0

1,660.0

1,130.0

1,460.0

970.0

1,270.0

1,500****

1,950.0

2,490.0

1,695.0

2,190.0

1,455.0

1,905.0

2,000****

2,600.0

3,320.0

2,260.0

2,920.0

1,940.0

2,540.0

2,500****

3,250.0

4,150.0

2,825.0

3,650.0

2,425.0

3,175.0

* Based on ANSI/IEEE C57.91-1981 Table 5, with 90% prior loading ** Based on 40°C ambient *** Based on 10°C ambient **** Overload factors for some of these units may be limited as a result of fusing limitations at primary voltages of 12.5/7.2 kV (or less).


4 • Estimating peak demand duration (unless it can be obtained from the consumer) and determining the loading capability of a transformer using the loading tables. This step becomes very important when the estimated load falls between two transformer sizes. • Using an appropriate power factor to correlate kW load to kVA load in calculations on consumer’s load profiles. Every time a transformer is sized correctly, the cooperative’s capital investment has been minimized.

MAXIMUM TRANSFORMER CASE TEMPERATURES Effect on Public Safety In today’s litigious society, some cooperatives may be less concerned with their pad-mounted transformers burning up than they are with having someone burned by touching the case of an overloaded unit. This concern is legitimate and must be addressed. However, the possible problem can be a manageable risk once it is put into proper perspective. Most people know not to touch the hood of a car that has been sitting in the sun on a hot summer day. A person touching

EXAMPLE 4.8: Sizing Commercial Transformers.

Example A A customer has requested service for a convenience store that is 24,000 square feet. The customer has provided the following load information: • • • • • • •

Neglect the electric heat load because the summer load is the dominant load. The total diversified load is 247.5 kW. Assuming a power factor of 0.9, the kVA demand of this store is 247.5/0.9 = 275.0 kVA

Lighting: 80 kW Electric Heat: 60 kW Air Conditioning: 60 tons Water Heater: 18 kW Refrigeration: 160 kW Fans: 10 kW Miscellaneous: 20 kW

From the chart, a convenience store has a peak duration of eight hours. From the transformer loading table, a transformer with a peak of eight hours can be loaded to 113 percent of its nameplate rating in the summer months. Now, 150 kVA × 1.13 = 169.5 kVA. The calculated load, 275 kVA, exceeds the loading capability of a 150-kVA transformer, so a 300-kVA transformer for this customer should be installed, which agrees with the similar account recommendation.

In reviewing other similar accounts, it has been determined that a 300 kVA or 500 kVA transformer may be needed. Therefore, other methods must be used to help make this choice.

As another check, the watts/square foot method suggests the following load:

The following load has been determined by means of diversity factors:

24,000 sq ft × 12.7 watts/sq ft = 304.8 kW

Load Description

Load (kW)

Diversity Factor

Actual Demand (kW)

Lighting

80.0

0.80

64.0

Electric Heat

60.0

0.75

45.0

1.5

0.60

67.5

Water Heater

18.0

0.50

9.0

Refrigeration

160.0

0.60

96.0

Fans

10.0

0.40

4.0

Miscellaneous

20.0

0.35

7.0

378.0

—

If a power factor of 0.9 is assumed, the watts/square foot method gives 304.8 kW/0.9 = 338.7 kVA

Air Conditioning

TOTAL

This approximation may seem a little high for this store compared with the other methods, but this load would still not exceed the loading capability of a 300-kVA transformer (300 kVA × 1.13 = 339 kVA). Therefore, after three different methods are considered, the conclusion is to install a 300-kVA transformer to serve this customer.

247.5

Continued

1 5 8 – Se c t io n 4

4 EXAMPLE 4.8: Sizing Commercial Transformers. (cont.)

Example B A customer has requested service for an office building that has 355,750 square feet. The customer has provided the following load information: • • • • • • • •

Neglect the electric heat load, because the summer load is the dominant load. The total load is 2,311.7 kW. Assuming a power factor of 0.9, the kVA demand of this office building is 2,311.7 kW/0.9 = 2,568.6 kVA

Lighting: 500 kW Electric Heat: 100 kW Air Conditioning: 1,250 tons Cooking: 288 kW Receptacles: 1,480 kW Computer Equipment: 600 kW Motor Load (larger than 40 Hp): 840 Hp total Motor Load (smaller than 40 Hp): 99 Hp total

Let’s look at another method before making a final decision. The watts/square foot method suggests the following load: 355,750 sq ft × 7 watts/sq ft = 2,490 kW If a power factor of 0.9 is assumed, the watts/square foot method gives

A similar account could not be found for this office building. Therefore, other methods must be used to help size the transformer.

2,490 kW/0.9 = 2,767 kVA

The following load has been determined by means of diversity factors:

This approximation is very close to the diversity approximation. Now, the decision must be made regarding what size transformer is to be installed.

Load Description

Load (kW)

Diversity Factor

Actual Demand (kW)

Lighting

500.0

0.80

400.0

Electric Heat

100.0

0.75

75.0

1,250.0

0.75

937.5

288.0

0.40

115.2

1,480.0

0.15

222.0

Computer Equipment

600.0

0.75

450.0

Motor Load (larger than 40 Hp)

0.746

840.00

157.0

Motor Load (smaller than 40 Hp)

0.746

99.00

30.0

Air Conditioning Cooking Receptacles

TOTAL

4,818.5

—

The peak duration for this office building can be estimated to be eight hours, unless otherwise stated by the customer. On the basis of an eight-hour peak duration, the transformer can be loaded to 113 percent of its kVA rating. For a 2,500-kVA transformer, it can be loaded to 2,825 kVA during peak times. Therefore, a 2,500-kVA transformer should be installed to serve this customer.

2,311.7

a hot transformer case will likewise naturally jerk away from it on contact. Estimating Case Temperature It is almost impossible to predict the case temperature of a pad-mounted transformer under load because many factors contribute directly and indirectly to the surface temperature: • Preloading, • Present load,

• Solar effect, • Wind direction and velocity, • Location of the unit near structures or shrubbery, • Ambient temperature variation, and • Part of case involved. To better understand the problem, a major manufacturer took three of its single-phase, lowprofile units to the test floor and measured case temperatures at full load and at a sustained


4 overload. As expected, the temperature varied widely from one part of the case to another. Figure 4.14 shows the top and front views of a pad-mounted transformer. The circled numbers one through eight denote the locations of various temperature measurements. Table 4.22 lists the

3

4

2 1 5 Inside Cabinet 6 Top of Cabinet Top View 1

6

7 Top Oil 2

4

8 Inside Oil

Front View

FIGURE 4.14: Case Temperature Measurement Location—Pad-Mounted Distribution Transformer.

temperatures and their locations for the three pad-mounted transformers of different sizes. As a point of reference when viewing the table, consider that the manufacturer designs single-phase units of 100 kVA and less to carry approximately 180 percent load for six hours with normal life expectancy. The three designs listed are based on the industry standard of a 65°C rise or less for 100 percent load. An additional calibration point is that the units are designed not to exceed 125°C top oil temperature with a 25°C ambient temperature at the higher continuous loads. Tank Temperature Burn Probability The specter of possible harm from hot pad-mounted transformer surfaces was first raised in the technical press in 1972 (Tarplay). There was a flurry of activity focused on the problem. However, the best qualitative thermal data on the subject were developed by E. I. duPont de Nemours and Company, Inc., over many years through research associated with its protective clothing activities. From curves developed by duPont and the NASA space program, Table 4.23 was developed (Lee, 1973; NASA, 1964). Table 4.23 shows that the probability of a person’s receiving a first- or second-degree burn under normal loading conditions is very small. Another point to consider is that the contact time to produce a second-degree burn is about 2.5

TABLE 4.22: Surface Temperatures Measured at Various Locations on the Cases of Pad-Mounted Transformers. Rise Above Ambient Temperature, °C (36°C Ambient) 25 kVA

37.5 kVA

50 kVA

Measurement Locations

100% Load

180% Load

100% Load

180% Load

100% Load

150% Load

1

9.5

27.0

12.5

34.0

16.5

34.0

2

13.0

44.5

24.0

60.5

31.0

61.0

3

16.5

44.0

24.5

60.5

33.5

55.5

4

16.0

43.0

21.5

50.5

28.0

54.0

5

17.5

—

13.0

37.5

33.5

38.5

6

5.0

14.0

5.5

16.5

8.5

16.0

7

25.5

67.5

35.5

87.0

43.5

84.0

8

26.5

68.0

35.5

87.0

46.5

84.0

Source: ABB Power T&D Company, Inc., Underground Distribution Transformer Division.

1 6 0 – Se c t io n 4

4 people’s ideas about burns have been formed by their personal experience with boiling water, which maintains skin contact and, thereby, causes more severe burns.

TABLE 4.23: Surface Contact Time to Produce Burning. Time in Seconds

Ambient Temperature (°C)

Case Temperature (°C)

Pain

Blister

36

69

33.0

70.0

88

7.5

19.0

36 36 36 36

DEDICATED TRANSFORMER LOADS Many cooperatives that serve farming communities with large irrigation loads and oil fields with 95 6.0 13.0 producing wells have applications in which a 110 3.0 8.0 transformer is dedicated to supply power to one 115 2.5 6.5 motor that is the total load on the transformer. Selecting the proper size three-phase transformer for times that of the pain level. this type of application is a Under these conditions, a perTransformer cases straightforward process deson’s normal reflex should opget hot but don’t scribed by Example 4.9, exerate in plenty of time to pull cerpted from the ABB Distribaway from the hot surface because burns. ution Transformer Guide. fore being burned. Although In Example 4.9, the motor the skin may become redselected had a starting current dened, it generally will not within the range of typical NEMA Code G moblister. The body’s natural protection system will tors. If voltage drop is a problem, 100-Hp monormally protect against burns up to about tors with smaller starting currents can be pur149°C (300°F). A temperature of 149°C seems chased, provided that the starting torque characquite high not to produce a burn in all instances. teristics are satisfactory for the load being driven. The problem is perception. Unfortunately, most

EXAMPLE 4.9: Dedicated Transformer Load.

STEP 1: Determine the locked-rotor kVA of the motor. NEMA standards specify starting code letters for squirrel cage induction motors that correspond to the kVA per horsepower required to start the motor. The series of curves of Figure 4.15 graphically show the relationship between the motor size, the locked-rotor requirements of the motor, and the transformer thermal capability. Table 4.24 is based on the locked-rotor code letters, but it can be used for any motor by selecting the curve that corresponds to the locked-rotor kVA/Hp of the motor for which the transformer is being sized.

100 70 50 40

Transformer kVA per Motor Hp

Determine the minimum kVA size three-phase transformer to power a 100-Hp, three-phase, 124-ampere full-load current, 460-volt squirrel cage induction motor with a locked-rotor current of 725 amperes. The motor will be driving a center pivot irrigation system. Service to the site will be through an underground three-phase cable at 12.47 kV with a minimum length of 1,600 feet.

30 20

V S P M K H GF E D C B A

10

T R N L J

7 5 4 3 2

1

1

2

3

4 5

7

10

20

30 40 50

70

100

200 300 400 500 700 1,000

Starts per Hour

FIGURE 4.15: Relationship Among NEMA Starting Code Letters, Starts per Hour, and Transformer kVA per Motor Hp for Transformer Thermal Considerations Continued


4 EXAMPLE 4.9: Dedicated Transformer Load. (cont.)

TABLE 4.24: NEMA Starting Code Letters. Code Letter

Locked-Rotor kVA per Hp

A

0.00–3.15

B

3.15–3.55

C

3.55–4.00

STEP 2: Determine the number of starts per hour planned for the motor under normal operating conditions. The load is a water pump driving a center pivot irrigation system. These systems are usually run for weeks at a time after they are started. Assume one motor start per hour.

D

4.00–4.50

STEP 3: From Table 4.24, select the curve letter that corresponds to the locked-rotor kVA/Hp of the motor.

E

4.50–5.00

Therefore,

F

5.00–5.60

G

5.60–6.30

H

6.30–7.10

J

7.10–8.00

K

8.00–9.00

L

9.00–10.00

M

10.00–11.20

N

11.20–12.50

P

12.50–14.00

R

14.00–16.00

S

16.00–18.00

T

18.00–20.00

U

20.00–22.40

V

22.40 and up

If the starting kVA or starting code letter is unknown, the locked-rotor kVA of the motor may be calculated with Equation 4.9.

STEP 4: Enter Figure 4.15 on the X-axis at the correct starts per hour for the motor being applied. Move up to the intersection of the starts per hour and the correct locked-rotor code letter curve and read the kVA of the transformer required per horsepower of motor from the Y axis. Because motor starts per hour equals one, intersection of curve G with the Y axis equals 1.5 kVA/Hp. STEP 5: Multiply the kVA/Hp by the rated horsepower of the motor to find the smallest transformer to be used in the application. Sizing the transformer with this procedure is conservative because it assumes that the voltage maintained at the motor terminals during starting is the rated voltage of the motor. Therefore, 1.5 kVA/Hp × 100 Hp = 150 kVA

Equation 4.9 Locked-Rotor kVA =

3 × VR × IS 1,000

where: VR = Rated phase-to-phase voltage of motor IS = Motor starting current at rated voltage Therefore,

Locked-Rotor kVA =

578 kVA/100 Hp = 5.78 kVA/Hp = Letter G

(1.732)(460)(725) = 578 kVA 1,000

STEP 6: Most motors started across the line require approximately 80 percent of rated voltage at their terminals under locked-rotor conditions to successfully start. After the transformer has been sized so it can withstand the starting pulse caused by the motor, check the voltage regulation of the system from the substation transformer through the secondary terminals of the distribution transformer to see if there will be enough voltage to start the motor. RUS Bulletin 160-3 describes the procedure to make the voltage drop calculation, plus other useful information.

1 6 2 – Se c t io n 4

4

Maximum Allowable per Unit Pulse

Also consider the transformer impedance: the lower the absolute impedance (ohms), the less regulation across the transformer, particularly during the motor starting sequence when reactive current predominates. Lower absolute impedance of the transformer can be accomplished in two ways: (1) a transformer of the selected capacity and the lowest available percentage impedance (%Z) can be installed, or (2) the transformer capacity (kVA) can be increased. While the latter choice may be the more expensive of these two options, it will always be less expensive than lowering impedance of the primary system. In this example, the primary objective was to ensure that the transformer kVA size was adequate to start the motor. Mention was made of the number of times per hour the motor would be started, but this was not really considered because it was assumed the motor would be started only infrequently. The “Starts per Hour” axis in Figure 4.15 is concerned mainly with limiting the thermal stress imposed on the transformer by the motor during frequent starts. Another important consideration in multistart applications is the effect of the magnitude and duration of the starting current pulse on the transformer. Each time a motor starts, it essentially puts a controlled secondary fault on the transformer. The transformer must be sized to

Equation 4.10 n=

4

4.25 Ip

where: n = Number of starts per hour IP = Pulse current per unit of transformer rated current

withstand the mechanical and thermal stresses imposed by this duty. Extensive data have been gathered by manufacturers and utilities about pulse duty on transformers. The conclusion is that, if the current pulses per hour exceed those shown in Equation 4.10, the transformer will fail prematurely because of the repeated mechanical stresses placed on the core and coils. Figure 4.16 shows the curve for Equation 4.10. Look back at the previous pad-mounted transformer sizing example (Example 4.9) to determine the number of starts per hour limitation to ensure normal life expectancy of the 150-kVA transformer selected:

Transformer rated current = 180 amperes at 480 volts Motor starting current = 725 amperes IP = 725/180 = 4.0 pu of transformer rated current

10 9 8 7 6 5

0.1 to 4

10 Sta rts/Ho

ur

3

2

10 to

1,000

1 0.1

0.2

0.3

0.4

0.5

0.6 0.7 0.8 0.9 1

2

3

4

5

6

7

8 9 10

Number of Current Pulses per Hour

FIGURE 4.16: Maximum Motor Starts per Hour for Transformer Mechanical Considerations.


4


By entering the curve on the X-axis at 4 pu, one can see that the allowable number of starts should be limited to less than 1.25 per hour. It should also be noted that there are some motor applications that impose significant thermal and mechanical stresses on transformers without multiple starts per hour. This is particularly true for motors serving loads that may

cause the motor to approach stall conditions. Examples include rock crushers and feed mills. In these cases, the same basic calculations should be run for the particular motor using the current drawn by the motor near the torquebreakdown curve. The results should then be evaluated considering the frequency of the expected stall conditions.

1. Ampacity is defined as the ability of a cable to carry maximum current under a specific set of conditions. 2. Cable ampacity can be calculated, but, in most instances, single-phase and three-phase cable ampacities are selected from ampacity tables. 3. The maximum ampacity of UD cable is set by the operating temperature of its insulation and depends on the ability of its surrounding environment to dissipate the heat generated in the conductor, concentric neutral, and insulation. 4. The maximum temperature rise of a cable depends on the shape of the load duration curve, which depends on the relationship between the loss factor and load factor of the circuit. Current values listed in ampacity tables are always calculated using a corresponding load factor. 5. Cable ampacity is affected by the ability of surrounding soil to dissipate heat generated within the cable. This fundamental property is called soil thermal resistivity. 6. Soil thermal resistivity depends on the type of soil, its moisture content, and the structural arrangement of the soil particles. 7. Soil thermal resistivity depends mainly on moisture content that is seasonally variable. 8. Ambient soil temperature affects ampacity because the insulation temperature rise is added directly to it to determine the maximum cable conductor temperature. 9. The ampacity of three-phase installations is reduced as a result of mutual heating between the phases and losses in grounded concentric neutrals resulting from circulating currents. 10. Losses in grounded concentric neutrals of three-phase applications are affected by the physical arrangement of the individual phases.

11. Cables placed in conduit have less ampacity than do direct-buried installations. 12. Direct-buried cables should be de-rated when they are installed in vertical riser pole applications. 13. Risers should be open at the top and vented at the base to maximize ampacity and to counteract solar heating effects. 14. For three-phase circuits buried in conduit, the riser usually is not the element that limits load. 15. Ambient air temperature is the most important element in determining how much load a pad-mounted transformer can carry over its expected lifetime (30 years minimum). 16. Transformer daily peak loads should be selected from loading guides after predicting what the temperature will be during the peaks. 17. Two methods should be used together to predict temperature for the month involved: (a) Average of all daily highs and all daily lows for several years, and (b) Average of the high and low of the hottest day over many years. 18. Transformer thermal time constant and thermal aging characteristics of its insulation allow short-time peak overloads to be carried without decreasing normal life expectancy. 19. Equivalent initial load and equivalent peak load must be calculated to perform loading studies. 20. Preload conditions should be considered when loading transformers. Preload levels given in loading guides are based on transformer nameplate rating and are not a percentage of peak load.

1 6 4 – Se c t io n 4

4 21. For cold weather conditions, a maximum loading above 1.8 per unit cannot be justified for preload conditions above 50 percent of peak load. This means many per-unit figures above 2.0 per unit given in ANSI/IEEE C57.91-1981 tables do not apply in practical situations. 22. Load-estimating guides based on load diversity and demand should be used to estimate peak kVA transformer load for groups of residential consumers. A loading guide developed specifically for the geographical region surrounding a cooperative’s service area should be used.

23. The surface temperatures of pad-mounted transformer cases can exceed 60°C during peak loading on sunny days. However, tests have shown that a person’s normal reflex action in response to touching a hot surface should prevent burning under normal conditions. 24. Pad-mounted transformers for dedicated motor loads should be properly sized based on motor locked-rotor kVA and the number of starts per hour.

Grounding and Surge Prot e c t i o n – 1 6 5

5 In This Section:

Grounding and Surge Protection

Cable Grounding System Function Factors Affecting Cable Grounding System Performance Counterpoise Application for Insulated Jacketed Cable

When cooperatives first started installing primary underground distribution systems, they used BCN cable, which, at that time, was an industry standard and a very effective way to provide good system grounding. Unfortunately, these cables failed long before the end of their expected life because of electrochemical treeing in the insulation layer that was accelerated by moisture and high-voltage stress. In addition, because of similar electrochemical action, the corrosion and disappearance of the bare concentric neutrals was also a major problem. A solution to these problems was the addition of an outer jacket over the concentric neutral of the cable. This jacket can take the form of an insulating jacket or a semiconducting jacket. RUS cable specifications were changed in 1987 to require an electrically insulating jacket to be applied over the cable. The jacket provides physical protection for the cable and helps prevent moisture contact with the insulation layer. The jacket also insulates the concentric neutral from direct contact with soil. Unfortunately, this feature reduces the performance of the cable grounding system.

System Ground Resistance Measurement and Calculation Underground System Surge Protection Summary and Recommendations

The function of the cable grounding system is to keep the cable as close to earth potential (“grounded”) as practicable at all times—during both normal and abnormal operating and under fault conditions. Proper grounding minimizes the effects of lightning surges on underground components after the surges are discharged by lightning arresters. Several factors affect the performance of the grounding system. Low riser pole ground resistance and the application of counterpoise wires reduce jacket voltages. There are also various methods to measure and calculate system ground resistance. Protection of the underground distribution system from lightning surges that originate on overhead lines is crucial. The application of riser pole arresters and lead length must be considered. Traveling waves on underground systems affect protection methods and dead-front arrester locations of different cable configurations. Through careful arrester location, higher protective margins than suggested by standards can be achieved. Refer to IEEE for assistance in applying distribution arresters.

1 6 6 – Se c t io n 5

5 Cable Grounding System Function

Before the function of the cable grounding system can be explained in detail, the term ground needs to be defined as used in this section. A ground is a current-carrying connection that connects a piece of equipment or a circuit to earth. The purpose of the connection is to maintain a point in the circuit or on the equipment as close to earth potential as possible. A ground is made up of a ground conductor, a bonding connector, its ground electrode(s), and the soil surrounding the electrode. The most common types of ground electrodes are: • • • •

Driven ground rods, Buried counterpoise wires, Cables with bare concentric neutrals, Concentric neutral cables with semiconducting jackets, • Metallic water or sewer systems, and • Rebar in reinforced concrete in manholes and vaults. Note that a pole butt ground applied to protect a distribution pole from lightning damage is not considered an effective ground electrode. For discussion purposes, the cable grounding system consists of the grounding circuit and the neutral circuit. The difference between the two circuits is that the neutral circuit is expected to carry current under normal operation, and the grounding circuit isn’t. The grounding circuit is made up of ground electrodes, ground conductors, and all connections. The neutral circuit includes the cable concentric neutral and any connections to it, and may include a separate neutral conductor. Under ideal circumstances, the grounding system maintains all points connected to it at earth potential during all normal and abnormal operating and fault conditions. For this ideal goal to be met, all connections between it and the earth must have a resistance of zero ohms; in reality, a zero resistance ground cannot be obtained. By using low-resistance conductors and electrodes, the design engineer can minimize the resistance of the metal circuit up to and in the earth. However, the engineer has no control over the resistivity of the soil in direct contact with the electrode, which is usually the most significant aspect in

determining the actual ground electrode resistance. According to Ohm’s Law (V = I×R, voltage equals current times resistance), if the ground resistance is relatively high at the point of a lightning current surge or a system fault, extremely high voltages can result. A low ground resistance will discharge lightning strokes with a lower probability of system disturbance. A good ground will improve the chances for rapid operation of protective relays and fuses to clear faults and limit personal injury and equipment damage. A good ground will also lower the voltage existing between grounded objects, such as transformer cases, and the nearby earth surface during fault conditions. The magnitude of the ground resistance can be found by measuring the resistance of the surrounding soil to the flow of current. This resistance is usually associated with driven ground rods and, in theory, can be calculated with Equation 5.1.

Equation 5.1 R=ρ where: R ρ L A

= = = =

L A

Ground resistance, in ohms Soil resistivity, in ohm-m Length of the current path, in meters Area of current path, in square meters

The easiest and best method to find the value of ground resistance is to measure it with a ground resistance tester. The reading is obtained directly in ohms. Soil resistivity is most accurately measured with a four-point earth resistance tester. Soil resistivity can vary widely over a small geographical area and is affected by the type of soil, moisture content of the soil, and soil ambient temperature. More information on field measurement of ground resistance, soil resistivity measurements, and the various elements that affect soil resistivity may be found in a later subsection, System Ground Resistance Measurement and Calculation.


5 This low-impedance path PUBLIC SAFETY shunts most of the fault curA well-designed, -constructed, Proper grounding rent through the grounded and -maintained grounding increases system neutral. system is essential to the operOver the years that UD sysation of any electrical distribupersonal safety. tems have been in place, they tion system to maintain all have established an excellent common points connected to safety record. One reason is it as close to ground potential that a good grounding system exists, resulting, in as practicable. Proper grounding of a four-wire, part, from the use of bare concentric neutral wye-connected, effectively grounded system cable that provides a large neutral surface in diprovides the following functions: rect contact with the soil. However, because of corro• Limits voltage across line-tosion, changes in the water ground insulation, Pay attention to table, changes in facilities, and • Provides a path to shunt how JCN installations the increasing use of JCN surge currents from the are grounded. cable, more careful attention system, should be paid to the installa• Allows ground faults to be tion of the grounding system. isolated quickly, • Reduces the shock hazard RETURN CURRENT PATH to people by reducing touch voltages during The typical underground distribution system is a faults on electrical equipment cases and three-phase, four-wire wye with multigrounded frames to safe levels, and neutral, which satisfies the definition of an effec• Improves the likelihood that ground faults tively grounded system. The neutral circuit must will be isolated quickly. be a continuous metallic path along the route of the primary feeder and must extend to every Unlike an overhead system in which equipconsumer’s location. For this requirement to be ment is physically raised above public areas, met, the concentric neutral of jacketed cable most UD systems have equipment enclosures must be grounded at each distribution transmounted on the ground within easy public acformer, at frequent intervals (specified below) cess. If a phase conductor contacts an enclosure, where no transformers are located, and at driven no dangerous voltages should exist because the ground rods at each user’s service entrance. Beenclosure could be touched by a member of the cause the concentric neutral is multigrounded, it general public or the cooperative’s crews. To is connected in parallel with the earth, which decrease the chances of a shock, ensure that the forms a relatively low resistance path to the flow enclosure is connected to the lowest possible of current. Under normal operating conditions, ground resistance. Another way to reduce touch residual current caused by unbalanced phase-tovoltage on pad-mounted equipment is to install neutral loads on primary circuits returns to the a buried counterpoise system around the system. neutral of the substation transformer along this One way someone could accidentally come parallel path. In no instance, even under emerinto contact with an energized conductor is by gency conditions, should the earth ever be used digging into a cable. All RUS-accepted UD prias the only path for the return of normal load mary cable is manufactured with concentric neucurrent on a distribution system. tral wires that provide some electrical protection For typical overhead rural distribution lines, it for someone digging into it. The theory is that has often been assumed that 40 percent of the the metal digging tool would first contact the return current is carried by the neutral with 60 grounded neutral wires and then the conductor, percent returning through the earth. However, thereby creating a low-impedance path between the current division will vary depending on earth the conductor and the concentric neutral wires.

1 6 8 – Se c t io n 5

5 For secondary single-phase, three-wire, resistivity and the size of the neutral, especially 120/240-volt systems, the two energized conin the case of JCN underground systems where ductors plus the grounded neutral from the the neutral is grounded only by ground rods or transformer are run to the user’s service enby counterpoise wires. If the neutral is the same trance where the neutral is again connected to a size as the phase conductor, which is usually the driven ground rod. The user’s ground circuit is case for single-phase underground circuits, the directly connected to the grounded neutral of current in it will be almost as large as the phase the transformer to ensure that no potential difcurrent. As the size of the concentric neutral is ferences can exist between the two systems. Efreduced, the greater the current flow in the fective grounding is especially important to earth. However, this change in current distribuprotect 120-volt equipment connected across tion does not have a linear relationship to the two halves of the 240-volt transformer secratio change in the neutral size. On single-phase ondary. The solid neutral connection holds the primary circuits, RUS specifies that the concenneutral at a point halfway between the 240-volt tric neutral and phase conductor must have the conductors. If the user’s neutral becomes isosame conductivity. lated from the transformer neutral point, unbalIn a perfectly balanced three-phase system, anced voltages across the equipment will result. no neutral or ground currents flow. However, as The voltages across the two 120-volt legs will stated previously, unequal phase-to-neutral split in proportion to the impedance of the load loads will cause an unbalanced current to flow on each side of the circuit, possibly causing in the return path. Normal practice is to try to burned-out light bulbs or damaged appliances. keep loads balanced for the system to operate efficiently. For this reason, the concentric neutral size in a three-phase circuit can be much smaller NEUTRAL CIRCUIT FUNCTION than the phase conductor. Cooperatives may opUNDER FAULT CONDITIONS erate three-phase systems with three cables On distribution circuits, the principal means of specified at 1/3 neutral each, or 100 percent of fault protection are the overcurrent relay and the conductivity of a single-phase conductor. fuse. For these types of devices to sense a shortMost engineers recognize that a 1/6 neutral, circuit condition and act quickly to interrupt the with a combined three-phase fault, the fault current magniconductivity of 50 percent of tude must be considerably the conductivity of one phase higher than the maximum load Reducing neutral conductor, is enough for most current. The most probable losses increases operating systems. Reducing type of fault on an underthree-phase circuit the size of the neutral has the ground circuit is the singleadditional benefit of reducing line-to-ground (SLG) fault. ampacity. the circulating currents inSimply stated, the amount of duced in the concentric neufault current depends on the trals when they are grounded following: and connected to each other, which increases cable ampacity and reduces losses. • The impedance of the source, The grounding and neutral circuits also pro• The voltage at the source, vide a way to ground the neutral of both three• The line impedance from the source to the phase and single-phase pad-mounted distribupoint of fault, tion transformers. The transformer neutral is • The impedance to ground at the point of connected to the cable concentric neutral and fault, and both are tied to at least one ground rod. The • The impedance of the fault. tank should be grounded at two points by separate connections to ensure that it cannot become Fortunately, in UD systems, unlike overhead, ungrounded through accident or corrosion. the cable concentric neutral is usually involved


5

LV HV LV

Core

Wingdings LV = Low Voltage HV = High Voltage

with driven rod(s). It is recomin an SLG fault, which allows mended that No. 4 AWG copthe maximum available fault A low-impedance per ground wire be used to current to flow. A large fault neutral path allows bond no larger than 400-kcmil current ensures that protective cables with 1/3 neutral. No. 2 devices act quickly and posifast protective AWG will be sufficient to bond tively to protect equipment device operation. to 1/3 neutral 500- to 1,000from excessive damage and kcmil cable. reduce the possibility that anyThe neutrals of three-phase one will be harmed. circuits should be connected together and groundAnother function of the neutral circuit is to ed to keep them at or near ground potential. Unprovide a low-resistance ground at a padder fault conditions, interconnected neutrals and mounted transformer or other equipment locagrounding will reduce the probability of arcing tion. A low resistance is needed to reduce the between the concentric neutral of a faulted cable chance of a dangerous touch potential for an and other nearby neutrals, or other grounded SLG fault in the transformer. The multigrounded metallic paths. This procedure also reduces the neutral in parallel with ground rod(s) at the lodanger to personnel who may be working in a cation will provide the necessary protection manhole or enclosure when a cable fault occurs under all except the most unusual conditions. by keeping metallic objects at the same potential. JCN cable, with an insulated jacket, must be The secondary low-voltage neutral circuit is grounded at least four times per mile for delibgrounded at the pad-mounted transformer seconderate-separation areas, and at least eight times ary and at the service entrance of a consumer. At per mile for random-separation areas. (See the the point of delivery (the meter), another metallic 2007 NESC, Rules 96C and 354D3c.) Cables with ground is required from the breaker panel to a bare concentric neutrals or with a semiconductmetallic water pipe or a suitable made electrode. ing jacket (meeting NESC Rule 94B5) may emThe grounds are necessary to prevent excessive ploy the concentric neutral as a made electrode voltages from developing between plumbing fixand the grounding requirements for the cable tures and appliances connected to the household are met, provided the installation complies with wiring system. Another contingency corrected in 2007 NESC Rule 354D2. If the required number part by the neutral grounding scheme is the posof grounds to the JCN (insulated jacket) is not sibility of a fault between the high- and low-voltobtained at sufficient transformer locations, the age windings of the transformer. In this scenario, cable neutral must be connected to ground primary voltage could be impressed on the fitrod(s) at intermediate points. In three-phase tings of 120/240-volt appliances, causing a fire. If runs, the neutrals of all three cables must be the secondary winding is grounded at the transconnected together with No. 4 or No. 2 AWG former, a high-voltage insulation failure involvcopper grounding conductor and tied to earth ing the secondary winding will immediately be shorted to ground by the center tap of the winding or by the core, blowLV ing the primary fuse and isolating the cirFault cuit from the source. The transformer ground thus prevents dangerous primary HV voltage from existing on the secondary conductors. See Figure 5.1. Secondary Core Transformer Ground

Neutral Service Ground

House Ground

FIGURE 5.1: Typical Distribution Transformer Core Form Design and Neutral Grounding Circuit.

SURGE PROTECTION GROUNDING Interest in the transient response or surge impedance of tower footings and driven ground rods began in the early 1930s

1 7 0 – Se c t io n 5

5 when engineers were trying to improve the outage rates of transmission lines. The main cause of outages was found to be direct lightning strokes to phase conductors. The protection method devised at the time required new line designs based on shielding the conductors from direct strokes through a combination of shield wires connected to ground conductors plus adequate phase-toground insulation. When lightning strikes the shield wire, the surge current is diverted to ground. It was found that a low surge impedance at the base of the structure is required to make the scheme work. Otherwise, a large surge current will produce a voltage at the top of the tower greater than the basic impulse insulation level (BIL) of the insulator string, causing a backflash to the conductor (Westinghouse T&D Reference Book, 1964). This same principle applies to the dissipation of surge currents in underground systems. Therefore, it is important to know the value of protection obtained from grounds when they are required to carry lightning discharge currents. It is, thus, necessary to establish the relationship between what will be called the surge impedance (ZSURGE) of a ground rod and its measured 60-Hz resistance (R60-Hz ) and determine how this difference does or does not affect lightning arrester protection. It is also necessary to understand the effect of lightning discharge path surge impedance on the protection and operation of underground systems using JCN cable.

Previous field and laboratory tests have shown that the surge impedance of a ground rod or a group of driven rods is defined as the ratio of peak voltage to peak current, and that ZSURGE, in ohms, is less than the 60-Hz measured values. Results also show that the surge impedance decreases considerably with increasing current. The actual magnitude of ZSURGE depends on many different elements (Bellaschi, Armington, and Snowden, 1942): • • • • •

Soil resistivity, Soil critical breakdown gradient, Surge current magnitude, Surge current waveshape (rate of rise), and Ground rod length, number, and configuration.

Ground rod resistance is usually expressed as the measured 60-Hz value; however, transmission and distribution line lightning performance depends on the impulse or surge value of the ground rod impedance. In jacketed cable installations, the cable jacket “sees” a voltage which is the sum of the IZSURGE (current × surge impedance) of the ground electrode plus the downlead component that is due to the surge current flowing into ground at the riser pole. The magnitude of the surge impedance at the base of the pole also determines how much surge current is diverted to the JCN and flows to remote connected grounds. In soils of low or medium resistivity, driven ground rods can usually obtain adequate grounding. For these grounds, Difference Between 60-Hz the surge impedance is less Grounding and Surge than the 60-Hz (R60-Hz) resisGrounding ZSURGE decreases tance value. The decrease can Ground rods are the most with increasing be shown by plotting ZSURGE common type of electrode lightning current used on utility distribution sysagainst the peak current as tems. The magnitude of their shown in Figure 5.2, which magnitude. surge impedance (ZSURGE) and depicts the ZSURGE of various the elements that affect it are grounds for peak surge curof major concern. Counterrents ranging up to 12 kA. Inpoise wires are also used to lower ground resisspecting the curves shows that, for clay soils tance. Because their initial effect on grounding with relatively low resistivity, ZSURGE will be less depends on the surge impedance of a buried than R60-Hz, but not to the extent exhibited by wire, they are covered in the subsection, sandy soils with much higher resistivity. For instance, the top curve represents an eight-foot Counterpoise Application for Insulated rod driven into ordinary sand with a measured Jacketed Cable, later in this section.


5 120 60-Cycle Resistance

100

ZSURGE (Ohms)

80

60 Rods In Sand

40

20

Rods In Clay

0 2,000

4,000

6,000

8,000

10,000

12,000

Peak Surge Current (Amperes)

FIGURE 5.2: Variation of Surge Impedance with Surge Current for Various Values of 60-Cycle Resistance. Source: Westinghouse T&D Reference Book, 1964, page 593.

60-Hz resistance of 120 ohms. At peak surge currents above 6 kA, it can be seen that ZSURGE is less than 40 ohms, a 67 percent decrease. For grounding resistances of 10 ohms or less, the surge impedance is not appreciably smaller than the 60-Hz resistance value. Different kinds of soil and types of ground can also be compared by looking at the surge characteristic of grounds shown in Figure 5.3. Here, the ratio of surge impedance to 60-Hz resistance (ZSURGE/R60-Hz) is plotted against peak surge current. In this figure, curve 2 represents a 10-foot galvanized steel rod one inch in diameter driven into moist clay with a 60-Hz resistance measured at 27.5 ohms. Curve 1 shows four of the same rods as shown in curve 2, spaced in a square 10 feet apart with a measured R60-Hz of 9.7 ohms. As the surge current increases above 12 kA, the ZSURGE/R60-Hz ratio of the single rod is less than 0.4, while the four rods in parallel will not have a ratio substantially below 0.7 at higher currents. To summarize, • The surge impedance (ZSURGE) of a ground rod or ground rod group is defined as the ratio of peak voltage to peak current.

1.0

0.8 Ratio of ZSURGE to R60-Hz

1. Four 10-ft Rods in Parallel, in Clay 2.

0.6

10-ft Rod in Clay 0.4 8-ft Rod in Sand 0.2

8-ft Rod in Gravel & Stones with Clay Mixture 8-ft Rod in Stones with Clay

0 0

2

4

6

8 10 12 Peak Surge Current (Kiloamperes)

14

16

18

FIGURE 5.3: Surge Characteristics of Various Ground Rods. Source: Bellaschi, Armington, and Snowden, 1942, page 353.

1 7 2 – Se c t io n 5

5 • ZSURGE is always less than or equal to the measured 60-Hz resistance of the ground rod(s). • ZSURGE decreases with increasing surge current magnitude. • The proportional reduction of ZSURGE is less for grounds of low resistance than it is for grounds of high resistance.

There are also various ways to reduce the magnitude of discharge currents on the neutral circuit.

Arrester Leads Lightning is a current generator. Surge arresters are applied at riser poles to protect cables from lightning-induced overvoltages by shunting the surge current to ground. Surge voltages produced by a lightning flash are a function of the Arrester Discharge Paths current magnitude, its rate of rise, and the disSurge arresters are applied on distribution lines charge path impedance. The arrester is confor two main reasons: nected to the overhead conductor and the pole ground conductor. The dis1. To shunt lightning current charge path that determines surges to ground, which the voltage impressed across reduces the magnitude of Keep arrester leads cable insulation is the arrester surge voltages propagating short to maximize and its connecting leads that on overhead and underprotection. carry lightning current in paralground systems, and lel with the cable termination. 2. To limit overvoltages on This concept is illustrated in protected equipment. Figure 5.4. Two riser pole installations are shown; the lightning discharge paths are highlighted. For the first application to be effective, there Pole 1 represents the desirable connection must be a low surge impedance to ground. In where no current flows through leads L1 and L2. the second application, ground resistance is not a consideration because the voltage across Cable phase insulation will “see” only the arequipment is limited to the arrester discharge rester discharge voltage. Pole 2 is not desirable voltage plus the voltage drop produced by the because the level of protection provided by the arrester lead(s). However, other elements must arrester is reduced when lead voltages L1 and L2 be considered when arresters are applied to proare added to the arrester discharge voltage. tect JCN cable. Arrester lead length must be considered in At the riser pole on wye-connected distribucalculating protective margin when evaluating tion systems, the arrester down lead is concurrent rate of rise. The protective margin is the nected to the pole ground conductor, the difference between the arrester discharge voltages multigrounded system neutral, and the concenplus the lead L di/dt drop and cable withstand tric neutral of the jacketed cable. Because prilevel, where di/dt is the change in current with mary and secondary neutrals are tied together at time expressed as kA/µs (kiloamperes per microthe pad-mounted transformer, the JCN provides second). Protection standards suggest using an a direct path for discharge currents to flow to average rate of rise of 4 kA/µs. Tests have shown the neutrals of premises that the transformer that the conductor normally used for leads has serves. The amount of surge current that flows an inductance, L, of about 0.4 µH/ft. The lead on the various neutrals is determined mainly by lengths connecting the arrester to the terminathe surge resistance of the pole ground. Surge tion will contribute approximately 1.6 kV/ft to voltages induced by discharge currents can damthe total voltage across the insulation if they carage the cable jacket and consumer appliances. ry lightning surge current. The 1.6 kV/ft figure is Various arrester discharge paths that occur at a based on an average probable rise time. Field riser pole have an effect on cable insulation proinvestigations have shown that this figure will be tective margin, cable jacket neutral-to-ground exceeded 30 percent of the time. Some applicavoltage rise, and how current surges on the section engineers believe 6 kV/ft or higher should ondary neutral can damage consumer equipment. be used. To minimize the effect of current rate


5 Lead L1

Lead L1

Cable Termination Lead L2

Lead L2

JCN Cable

L1 + L2 = 0 (Desired)

L1 + L2 = Lead Length (Should Not Be Used)

• Objective is to make certain no lightning current flows in the leads connected to the cable termination. Pole 1

Pole 2

FIGURE 5.4: Arrester Lead Length for Two Riser Pole Installations.

of rise, the leads should be kept as short as possible and arresters with low discharge voltages should be used. See Figure 5.5. The effect of lead length on protective margins will be covered in more detail in the Surge Arrester Application Factors subsection later in this section.

FIGURE 5.5: Three-Phase Installation Showing Optimum Riser Pole Arrester Lead Connections.

Pole Ground Conductor After a surge arrester operates to protect cable insulation, some engineers assume no additional damage will happen to other system components. This assumption is not always true. Once lightning current goes through an arrester, it flows into the neutral and ground circuits, causing overvoltages on neutral-to-ground insulation. This is especially a problem with electronic equipment (controllers, RTUs, etc.) that might be on the pole. Special methods should be considered to limit or eliminate problems this condition can and will cause. Figure 5.6 shows a typical underground primary installation fed from a riser pole and padmounted transformer. The direct-buried jacketed

1 7 4 – Se c t io n 5

5 cable and below-grade connections are also shown. Figure 5.7 shows the same installation except drawn in a way to highlight the various arrester discharge paths: • • • •

Pole ground conductor, Cable jacketed JCN, Counterpoise, and Overhead multigrounded system neutral.

After the lightning current passes through the arrester, it splits among the various paths. The respective surge impedances of the conductors and the surge impedance of the pole ground determine how the current initially divides. Resulting currents flow to both the local ground and remote grounds.

Jacket Voltages Local ground in this instance is the riser pole ground rod. When the pole ground conducts surge current, it produces a ground potential rise when measured relative to a remote refer-

ence point. The condition could be compared to ground potential rise in a substation during a ground fault. Because the cable concentric neutral is tied to the ground rod, any transient voltage produced by the surge event is transferred directly to it. The cable jacket, applied to protect the concentric neutral from environmental damage, also insulates it from ground, which means the total ground potential rise is disseminated across the jacket. The magnitude of the peak ground potential rise can be estimated as the peak current times the surge impedance of the riser pole ground rod(s). Laboratory tests have shown that peak jacket voltage occurs at a distance where the electric field strength around the ground rod and the ground potential rise approach zero. The concept can be better understood by referring to Figure 5.8. The area outside the circle represents where ground potential rise is zero. The ground rise is maximum at the center of the circle where the ground rod is located. A jacketed cable starts with its concentric neutral attached to the rod

Phase Conductor

Multigrounded System Neutral Counterpoise

Loop Feed Pad-Mounted Transformer

Continuous Counterpoise To Next Transformer JCN Cable

Connections Shown Below Grade For Clarity

Pole Ground Transformer Ground

FIGURE 5.6: Typical Primary and Secondary Underground Installation.

Triplex Secondary Cable Service Ground


5 ½

½

Phase Conductor

Cable Pothead

MOV

Multigrounded System Neutral Jacketed Concentric Neutral Pole Ground Conductor

Pad-Mounted Transformer

ZSURGE

Groundline

Insulating Jacket

LL

RN

Loads R1

RS

R2

RS

RL

Consumer’s Breaker Panel

R3

RPole ZSURGE

RTX

RService

Continuous Counterpoise Wire to 1st Transformer

FIGURE 5.7: Schematic Diagram Showing Surge Current Paths After Lightning Arrester Discharge.

and extends radially from the center. It ends at a point that is not affected by surge current flowing in the center ground rod. Measuring the voltage rise at points A and B from a remote reference gives maximum voltage at A and zero volts at B. (The ground rise is measured by dri-

ving a two-foot spike in the ground at each point.) Because the concentric neutral of the JCN cable is tied to the ground rod, the peak ground potential rise is transferred on the neutral to point B, where maximum voltage-toground exists across the jacket. Laboratory tests

At Point B: Ground Potential Rise V=0 V = Max Jacket Voltage Maximum Ground Potential Rise at Point A V = Max Jacket Voltage at Point A V=0

A

B Cable Start

Cable End

V

Ground Rod Outside the Circle Represents the Area of Maximum Jacket Voltage

FIGURE 5.8: Maximum Jacket Voltage (Neutral to Ground) Produced by Lightning Current Surge in Ground Rod.

1 7 6 – Se c t io n 5

5 show that maximum jacket voltage occurs within for the most commonly used jacket thicknesses. 50 feet of the riser pole. Laboratory tests have This analysis shows that the neutral on the also shown that lower jacket voltages will be JCN cable will not be at ground potential when measured at the end of the cable. Cable start and a surge occurs. As with an overhead system, the cable end voltages should not be the same, beneutral-to-ground voltage can reach dangerous cause the cable neutral potential is produced by levels during surges. the current in the two grounds and their respective surge resistances (GE Research Project, 1990). Jacketed Concentric Neutral The ground potential rise and the maximum Any lightning current that does not propagate jacket voltage are a function of the down-lead along the other paths attached to the arrester current and the surge impedance of the riser pole down lead will flow on the concentric neutral. ground rod. Increasing either of these quantities The JCN current magnitude depends on the will lead to higher jacket voltages. If the ground surge impedances of all connected paths. Slowrise exceeds jacket withstand strength, a jacket front waves and 60-Hz currents do not “see” the puncture will occur, allowing moisture to enter surge impedances of the JCN and the other the cable. Over time, this condition could lead paths. The 60-Hz measured resistances and imto loss of one or more of the neutral conductors pedances will be seen instead. The 60-Hz imto corrosion. pedances of each path are lower than their surge Unfortunately, no standards exist that define impedance values. If the paths are connected to the withstand strength of 50- and 80-mil jackets ground resistances lower than or equal to the most commonly used on underground cables. pole ground, a small change in the pole ground The only voltage test required by standards is resistance can mean a large current increase on the AC Spark Test that is used mainly as a qualthe concentric neutral and other paths. The path ity control check during the jacket extrusion with the lowest ground resistance will receive process. An 80-mil polyethylene jacket must most of the current. withstand 7.0 kV applied beAnother look at Figure 5.7 tween an electrode on the outshows that any increase in caside surface of the jacket and ble neutral current is transMinimize jacket the concentric neutral for not ferred directly to the neutral of less than 0.15 seconds. Laborathe pad-mounted transformer voltage with low tory tests have shown that new because of the cable insulating riser pole ground polyethylene insulating jackets jacket. Any current discharged have a surge (1.5 × 40 µs by a dead-front surge arrester rod resistance. waveform) withstand strength applied on the primary termiof about 2,500 volts/mil at nals of the transformer will also 20°C. After being in service, add to the contribution from this value drops to about 1,200 volts/mil after the JCN. If the transformer ground is much lowmoisture permeates the jacket. On the basis of er than the service ground, most of the lightning these figures, Table 5.1 lists withstand strengths current on the neutral will flow to earth at the transformer ground rod. If the reverse is true, most of the current will flow on the service neuTABLE 5.1: Surge Withstand Strengths of Polyethylene Insulating tral and to the ground at the service entrance. Jackets for 15-kV, 25-kV, and 35-kV Class JCN Cable. Damaging overvoltages can be induced on loads Jacket Thickness* New Jacket Insulation Aged Jacket Insulation R1, R2, and R3 connected inside the residence under this condition as a result of surge current 50 mil 125 kV 60 kV components flowing in the service neutral. 80 mil 200 kV 96 kV The surge impedance that has the greatest effect on current division between discharge paths 95 mil 240 kV 114 kV and surge voltages on the secondary is the pole * Jacket thickness over neutral wires ground. Keeping this resistance as low as practi-


5 cable means minimum lightning energy on the underground system neutral. The transformer ground must be a minimum resistance because some service grounds are tied to underground metallic water systems. The most economical way to obtain good grounds in the above two instances is by multiple ground rods, deep-driven rod(s), or the addition of counterpoise.

Counterpoise A continuous counterpoise conductor is shown connected to both ends of the jacketed cable in Figures 5.6 and 5.7. It is buried with the cable and represents another arrester discharge path at the riser pole. Laboratory tests have confirmed that, applied as shown, counterpoise will reduce the jacket voltage up to 50 percent under surge conditions. Adding counterpoise also improves the 60-Hz grounding of the riser pole arrester and cable neutral. Direct connection to the JCN decreases surge current transfer to the transformer neutral. Note that counterpoise is used only for JCN applications and is not required when BCN or semiconducting jacketed cable is installed. How counterpoise reduces jacket voltage and improves 60-Hz grounding is explained in more detail in the subsection, Counterpoise Application for Insulated Jacketed Cable, later in this section.

Factors Affecting Cable Grounding System Performance

Equation 5.2 Z= where: Z L C h

L 2h = 138 log ohms C r

= = = =

Surge impedance of conductor Inductance of conductor (Henries) Capacitance of conductor (Farads) Height of conductor above ground, in feet r = Radius of conductor in feet 138 = Constant from L and C values in Henries and Farads per mile

Overhead Multigrounded System Neutral The overhead system neutral presents two discharge paths for lightning current once it passes through the arrester. Surge current will flow in both directions away from the riser pole. The surge impedance of the two paths is approximately 500 ohms each, calculated from Equation 5.2 for a single aerial conductor with ground return. As can be seen, the surge impedance is determined only by the height of the conductor

above ground and its size (Westinghouse T&D Reference Book, 1964). Reducing the surge impedance of the neutral would be desirable as an additional way to reduce the amount of surge current diverted to the underground neutral/ground system. Unfortunately, its wire size is set by system requirements and reducing the height above ground is not an option. For these reasons, the overhead neutral is not a major factor in mitigating the effects of surges on the underground system. However, it is a vital part of the overhead neutral/ground system that acts with arresters to prevent lightning surges from propagating long distances from the strike point. It should be noted here that some lightning strikes are of such a magnitude that distribution voltage systems cannot be effectively protected from them. However, the majority of lightning outages and damage are caused by induced lightning strokes (approximately 95 percent), which can almost always be eliminated by effective lightning protection (including arrester protection, line configuration, and system BIL).

UNDERGROUND CABLE SYSTEM CONFIGURATION The function of the cable grounding system is to keep its entire length at ground potential at all times. Its ability to perform this function under fault and surge conditions is determined by the resistance of its electrical connections to ground. Ground resistance can be approximated

by calculation. The resistance of an actual installation can be found only by measurement. The type of cable used—BCN, jacketed, or semiconducting jacketed—will determine the effectiveness of the grounding system in performing its intended function. Getting a low ground resistance can be difficult and is highly site-specific. A question often

1 7 8 – Se c t io n 5

5 asked about system grounding is, “How low does the ground resistance have to be before it is considered a good ground?” Answering the question with a specific ohmic value is difficult because many variables are involved in an application. A low riser pole ground reduces the jacket voltage on jacketed cable. A low padmounted transformer ground—compared with the service ground—reduces surge voltage on consumer appliances. For JCN applications, the riser pole ground rod resistance should approach 10 ohms, if practical, whereas the transformer ground can have a higher value. The system configurations of bare concentric neutral, semiconducting jacketed, and jacketed concentric neutral cables affect grounding system performance. Because ground rods are the

Lightning Current

Overhead Phase Conductor

MOV Arrester Multigrounded Neutral

Bare Concentric Neutral UD Cable

Surge Current on BCN Dissipated in Earth

Ground Rod

FIGURE 5.9: BCN Cable Riser Pole Installation Surge Arrester Discharge Paths.

predominant way to obtain grounds at riser poles, intermediate points, and transformers, this subsection reviews elements affecting their resistance and required quantities. Soil resistivity also directly affects the resistance of a ground electrode. Bare Concentric Neutral Cable Direct-buried, BCN cable is considered the ideal configuration for a multigrounded neutral on a four-wire grounded-wye distribution system. Maximum continuous contact area between the system neutral and soil ensures an effectively grounded system. Correct operation of surge arresters is ensured under all conditions. Effective grounding limits neutral-to-ground voltages during faults and surge events, which reduces stress on cable insulation. The highest degree of public safety is also obtained. Unfortunately, corrosion problems associated with the BCN cable configuration preclude its continued use in new installations. Solid grounding by the BCN means the riser pole ground rod resistance has little effect on cable system surge protective levels. BCNs on direct-buried cable provide an effective path to ground under most conditions. The concept is illustrated in Figure 5.9. The overall ground resistance measured along the cable is significantly lower than the driven ground. With two arrester discharge paths available, a poor riser pole ground merely means more surge current flows on the BCN, where it quickly goes to ground. Although no longer in use by cooperatives, BCN cable relieved most but not all grounding concerns for direct-buried systems. Putting the cable in nonmetallic conduit led to a lack of continuous grounding and problems associated with poor grounding. Burying the exposed neutral in soil with different resistivities caused the neutral to corrode to the point where it was lost completely. Besides the reduction in grounding efficiency, open neutral wires caused localized electric field stresses. Over time, the insulation shield deteriorated, causing primary cable faults. The neutral wires of BCN cables were also more susceptible to damage during cable pulling and installation. In recent years, all utilities have experienced premature failures with direct-buried BCN cables.


5 grounded only at both ends of Resulting investigations found the cable. This type of system the primary causes to be elecInsulated jacket installation will decrease trochemical treeing in cable reduces grounding grounding quality when cominsulation and BCN corrosion. pared with a bare neutral conAccelerated tree growth was system performance. figuration. For example, conpinned to moisture in the insider two 1/0 AWG, singlesulation layer and high-voltage phase, direct-buried cable runs stress. As noted, these findings of jacketed and BCN cables, 1,000 feet long, in led the RUS to change Bulletin 50-70 (U-1) to resoil of 100 ohm-m resistivity. The resistance-toquire insulating jackets and thicker phase insulation ground of the bare neutral cable, assuming a caon all underground cables. Addition of the jacket ble effective diameter of 1 inch, is as follows: is a change from the BCN system configuration. Semiconducting Jacketed Cable According to tests conducted by General Electric Company for NRECA and various utilities, the concentric neutral-to-ground voltage of semiconducting jacketed cable is essentially independent of riser pole ground rod resistance and arrester discharge current. The semiconducting jacket acts like a BCN to provide good system grounding characteristics for underground installations. To provide good grounding, the semiconducting jacket must have a radial resistivity of less than 100 ohm-m (see 2007 NESC Rule 354D2c). If this jacket resistivity requirement is met, intermediate grounding for the cable run is not required. Unfortunately, this is not true for an insulating jacket; additional effort must be made to approach the same grounding system performance level achievable with semiconducting and BCN cable. Insulated Jacketed Cable An insulating, protective jacket provides many benefits. An exterior jacket provides mechanical protection for the neutral during pulling and installation. The jacket isolates the copper neutral from contact with corrosive soils. This isolation prevents galvanic cell formation and inevitable neutral corrosion. A protective jacket offers significant mechanical protection to the insulation shield and primary insulation. It also delays moisture from reaching and damaging the insulation layer, increasing cable life. However, insulating the neutral from ground has some drawbacks. The most important is that the performance of the grounding system is reduced. Jacketed cable installations less than 1,000 feet long would normally have their neutrals

1.15 siemens per 1,000 ft = 0.87 ohms for a 1,000-foot cable (from Table 7.6)

If the jacketed neutral is grounded by single 10-foot ground rods at each end with diameters of 3/4 inch, each rod would have a resistance of the following: 32.14 ohms or 0.0311 siemens (from Equation 5.9) To meet safety codes, the BCN cable must be connected to ground rods at each end as well. Adding the two ground rods to the BCN cable gives a total ground resistance for the installation. Note: Conductance (siemens), which is the reciprocal of resistance (ohms), will be used in the calculation to avoid the cumbersome formula for three resistances in parallel. Conductances of individual grounds in parallel can be combined by simple addition: 0.0311 + 0.0311 + 1.15 = 1.2122 siemens =

1 ohms = 0.8249 1.2122

For this particular example, the JCN cable installation has resistance equal to the two ground rods in parallel or 16.07 ohms; therefore, the JCN cable has the following: 16.07 ohms ÷ 0.8249 ohms = 19.48 times the ground resistance of a BCN cable installation

1 8 0 – Se c t io n 5

5 In the preceding case, both systems would be • Pad-mounted transformer locations, and considered adequately grounded according to • Service entrances. code. However, for longer runs or in higher resistivity soil, the jacketed cable would not be adGround rods normally carry high current only equately grounded. Additional driven grounds after faults or lightning arrester operation. could be added at both ends. For longer runs, Ground rods must be driven into undisturbed the NESC requires at least four grounds in each soil. They should not be placed in the hole with mile, not counting rods at indithe riser pole or driven into vidual services. To meet this backfill around an installation requirement, the jacketed neusite. Loose soil will not proDrive ground rods tral must be attached to vide the necessary rod interinto undisturbed soil. grounds at intermediate points face contact required for good along the route. grounding. Rods should be driven at least 2 feet from structures, concrete foundations, and poles with DRIVEN GROUND RODS ON THE UD SYSTEM steel reinforcing to prevent the possibility of arcGround rods are the predominant type of made ing from the rod. See Figure 5.10. electrode on underground distribution systems. Almost any metallic material may be used to They are mainly used at the following: manufacture ground rods. Copper-clad and galvanized steel are most common. The measured • Riser poles, resistance of the rod in the ground is the most • Jacketed cable intermediate grounding points, important feature to consider. Rod material has • Cable joints, little effect. Economics and corrosion considerations normally determine which rod material is selected. The ground resistance of driven rod(s) is affected by various elements. There are several ways to improve existing ground resistance. Only the measured 60-Hz resistance will be considered here because surge impedance has already been reviewed. The number of rods necessary for good grounding practice and required by the NESC is discussed here. Specific equations for calculating rod ground resistance for various configurations and examples are given later in the System Ground Resistance Measurement and Calculation subsection. When multiple ground rod sections are stacked on top of each other, a problem that can affect the ground rod resistance generally occurs. This problem is the lack of good soil contact. Because of the larger diameter of the coupling, the bottom ground rod is often the only rod making full contact with soil. The first coupling opens up a hole larger than the ground rod body and subsequent ground rod bodies make very little contact with the soil, and are definitely not in contact with undisturbed soil. FIGURE 5.10: Ground Rod Being Driven by This lack of contact with the soil (disturbed or Hydraulic Tool.


5 undisturbed) can make a big difference in the resistance reading observed. As time passes and the soil fills in around the ground rod body, the resistance values will change and most likely improve. The time required for this improvement is dependent on soil porosity, soil plasticity, and the amount of moisture in the soil. Ground Resistance of Driven Rods The ground resistance of a rod (or group of rods) is found by measuring it with a ground resistance tester. Resistance calculations can be made for specific installations and ground rod configurations to estimate what the resistance will be. Any theoretical calculations must start with the basic equation in Equation 5.1 or its equivalent. Equation 5.1 shows that the ratio between the length and area of the current path must be multiplied by the soil resistivity, ρ. The resistivity then affects the ground resistance of any electrode system, such as a single ground rod, BCN, or

1000

5/8"

3/4" 1-1/4"

Resistance (Ohms)

100

10

substation ground mat in the same way. Soil resistivity depends on soil composition. Experience has shown that resistivity can vary widely over a relatively small area. This variation throughout the soil volume cannot be modeled easily in ground resistance calculations. All formulas developed in this section for ground electrode resistances assume soil resistivity is constant throughout its volume. This restriction must be considered when the results from formulas are interpreted. Elements that affect soil resistivity are given later in this section. The three primary factors that affect the ground resistance of ground rods that the engineer can influence are the following: 1. Length, 2. Rod number, and 3. Spacing.

Resistance Variation with Depth How the resistance of a single ground rod varies with length can best be demonstrated by considering its resistance formula expressed by Equation 5.3. (Note that this formula assumes full contact of all rod sections to the soil.) Resistance does not decrease directly with length. The actual variation can be seen in Figure 5.11, which plots resistance against rod length. The curves are drawn for an earth resistivity of 250 ohm-m. A handy approximation that generally can be used is that doubling the rod length lowers the resistance by only 40 percent. For example, assume an eight-foot rod with a diameter of 5/8 inch has a measured resistance of 90 ohms. Doubling the length to 16 feet will reduce the resistance to about 54 ohms, 90 – (0.4 × 90),

Equation 5.3

1

R= 1

10

100

Length of Ground Rod (Feet)

FIGURE 5.11: Resistance of Vertical Ground Rods as a Function of Length and Diameter (Soil Resistivity = 250 Ω-m).

1,000

4L ρ In –1 (ohms); where L » a a 2πL

where: ρ = Soil resistivity, in ohm-m L = Rod length, in meters a = Rod radius, in meters

1 8 2 – Se c t io n 5

5 which agrees closely with the 5/8-inch curve of Figure 5.11. Doubling the length again to 32 feet would give a resistance of about 32 ohms, or 54 – (0.4 × 54). As the rod length keeps increasing, the law of diminishing returns applies. Additional length produces a very small reduction in ground resistance. For the 5/8-inch rod of the above example, this point of diminishing returns occurs at about 40 or 50 feet.

Multiple Rods in Parallel Reduced ground resistance can be obtained by paralleling rods to increase the cross-sectional area of the current path. Two identical rods driven into soil some distance apart will not have one-half the resistance of a single rod. The actual ground resistance will be about 60 percent. The reduction is about 40 percent for three rods in parallel and 33 percent when four rods are used. These relationships hold true Resistance Variation with for rods spaced about the same Diameter distance apart as their length. Use a longer rod, Another way to lower ground When multiple rods are acnot multiple rods, resistance is to use a larger ditually applied, the separation to lower ground ameter rod. Doubling a rod’s distance should be at least diameter reduces its resistance twice the length of one rod. resistance. by less than 10 percent. The The increased separation is multiple rod diameter curves needed to get the most useful in Figure 5.11 show the effect, effect of rod spacing. Figure which is minimal. Normal rod diameters used 5.12 shows that, for rods spaced greater than on distribution systems are 5/8 inch and 3/4 20 feet apart, the reduction of the ground resisinch. In most instances, increased rod diameter tance falls off rapidly. For example, assume a is considered only when encountering hard soil single rod 10 feet long with a measured ground or for driving deep rods connected to substation resistance of 60 ohms. Four rods spaced 20 feet ground mats. apart would have an equivalent resistance of

100-Ft Spacing

20-Ft Spacing

40-Ft Spacing

Resistance of Multiple Grounds

20%

25% 10-Ft Spacing 30% 5-Ft Spacing 40% 50% 60% 70% 100%

0

1

2

3

4

5

6

7

8

9

10

Number of Ground Rods

FIGURE 5.12: Resistance of Multiple Ground Rods (Single Rod Equals 100 Percent).


5 0.3 × 60 = 18 ohms. As the multiple rods or a deepseparation distance approaches driven rod should be used. The NESC governs infinity, the resistance of the Long rods can be hard to ground rods for JCN four rods will equal 15 ohms. drive in soil with a high rock If geological conditions content. Multiple rods can cable installations, permit, a single, deep-driven take up a lot of area. If the where practical. rod should be used instead of decision is made to install a multiple rods to lower ground rod grid, the rod arrangement resistance. For example, if four is less important than the 10-foot ground rods are placed 2 L apart (where separation distance. The conductor length and L is the length of the rod), they will have about number of connections should be kept to a the same resistance as a 40-foot rod. If the sepaminimum to tie the rods to the pole ground ration distance is less than 2 L, the deep rod will conductor. Two types of multiple rod grounding provide a much lower resistance than 4 × 10 layouts are shown in Figures 5.13 and 5.14 for a feet of rod placed close together. riser pole application. The above statements are based on a homogeneous soil profile. A deep rod will be expected Number of Driven Rods to reach the permanent water table beneath the The NESC (ANSI C2) does not specify the earth. The resistivity at this level will be considnumber of ground rods at specific locations on erably lower than near the surface. The advanunderground cable systems. It also does not rectage of the deep rod will be more pronounced in ommend what the ground resistance should be this case. Another benefit is that soil resistivity at at any specific location. But the NESC does have greater depths will not vary as much because of certain requirements for JCN installation groundchanges in temperature and moisture content as ing methods that apply to BCN installations as will resistivity near the surface. well. See the summary in Table 5.2. Specific Site conditions will normally dictate whether locations for driven rods are the following:

Riser Conduit Pole Ground Conductor

Riser Conduit Pole Ground Conductor

Vent

Vent

2L . min

2' min.

2L min.

Cable L

Cable

2' min.

L

2L min.

FIGURE 5.13: Installation of Three Rods for a Riser Pole Ground. Source: Parrish, 1982.

FIGURE 5.14: Installation of Four Rods for a Riser Pole Ground.

1 8 4 – Se c t io n 5

5 TABLE 5.2: 2007 NESC Ground Rod Requirements for JCN Cable Installations. Location Riser Poles

Pad-Mounted Transformers

Joints/Intermediate Grounding Points

Rule

Comment

92B2b(1)

Concentric neutral must be connected to surge arrester grounds where cables are connected to overhead lines

94B2a

If a driven rod is used, minimum length is eight feet and minimum diameter is 5/8 inch for steel and 1/2 inch for copper-clad. Longer rods or multiple rods may be used to reduce ground resistance.

94B2b

Minimum spacing between multiple rods is six feet.

94B2c

Driven depth not less than eight feet, with exceptions.

93C7 and 314

Concentric neutral and pad-mounted transformer and other equipment cases must be connected to a ground rod.

94B2c (exception)

If the rod is placed within the pad-mounted enclosure or pedestal, driven depth can be 7-1/2 feet.

96C

Concentric neutral must be connected to ground rods at least four times per mile (service grounds not included).

354D3c

For random separation with communications cables, grounding interval is eight times per mile (service grounds not included).

Note. Consult the specific NESC rules cited in the text to avoid any misunderstandings caused by condensing the rules in this table.

• • • •

Riser poles, Pad-mounted transformers, Joints/intermediate grounding points, and Service entrances. Pertinent NESC sections are the following:

• Section 9: Grounding Methods for Electric Supply and Communication Facilities, and • Part 3: Safety Rules for the Installation and Maintenance of Underground Electric-Supply and Communication Lines.

Riser Poles Rule 92B2b(2) says that a grounding conductor must be connected at the termination points of a nonjacketed cable. Additional grounding points for jacketed cable are recommended in 92B2b(3) since the neutral is not exposed and is not providing a ground connection. For a typical UD installation, the first termination point is the riser pole. If a driven rod is used, Rule 94B2a says that the total length may not be less than eight feet. Minimum rod cross-sectional areas are also given. If longer or multiple rods are needed, a minimum six-foot spacing is required.

A counterpoise is also considered a made electrode if the following conditions are met: • The bare wire is No. 6 AWG or larger, • The length is greater than 100 feet, and • The counterpoise is laid in the same trench as the buried cable (Rule 92B3). There is no suggested value for ground resistance at the riser pole. It is noted in Rule 96C that multigrounded systems extend over a large area and depend on a number of electrodes for grounding purposes; therefore, no specific values are imposed for the resistance of individual electrodes. As already mentioned, the lowest practical ground resistance should be obtained at the riser pole. As explained previously, low resistance ensures a low jacket voltage and prevents excessive surge currents flowing to remote transformer and service grounds. For minimal effect on the system, a good goal is to have the lowest ground at the riser pole. The next highest ground should be at the first pad-mounted transformer. The highest ground resistance compared with the previous two should be at the service entrance.


5 Pad-Mounted Transformers This subsection covers only padNote 1 mounted transformers. However, sugH1B gestions or recommendations are valid for any aboveground enclosure. X3 H1A X1 Rule 314 says that conductive parts Ground Strap must be grounded, including cases of X2 pad-mounted devices. Because the neuJumper #4 Copper tral is brought to the transformer, it must Tank Grounds be connected to a ground electrode according to Rule 96C. If a rod is used #4 Copper Ground Rod Clamps Ground Wire within the footprint of a pad-mounted Tamp Well Under Pad 7'6" min. enclosure, an exception to Rule 94B2c Note: states that its driven depth may be re1. Tie concentric neutrals together before tap to ground loop to ensure same conductivity as cable neutral. duced to not less than 7-1/2 feet. Pad-mounted transformers are norGUIDELINE ONLY mally grounded with one driven rod, except in areas of high soil resistivity FIGURE 5.15: Grounding Assembly for Pad-Mounted Single-Phase Transformers. where up to four rods might be needed. The resistance of the transformer ground should be less than the ground at the consumer’s service entrance to ensure neuTop View tral surges are not transferred to wiring inside the residence. Most cooperatives do not have control over the value of the service ground. The engineer should make a survey of existing grounds in the area. After a representative value is found, a Note 3 target for transformer and riser pole grounds can Opening be determined. Figure 5.15 shows a typical grounding assembly for a single-phase, pad-mounted transformer. A continuous ground conductor loop is shown that ensures solid grounding if one connection fails. Two clamps are shown for the ground rod. Front View Pad These are recommended to prevent a high-resistance contact when two wires are connected with 18" min. one clamp and to maintain ground electrode effectiveness if one connection is defective. 8'0" Two, three, or four rods are sometimes used min. to obtain the proper ground resistance at a transformer. The separation distance between rods Notes: should be kept to at least twice the burial depth 1. Place minimum of one ground rod at each corner to obtain low ground resistance of grounding grid. Minimum distance between ground unit assemblies = 6'0". when possible. In some instances, installing a 2. Grounding grid 1/0 AWG bare copper buried 18” minimum below ground. Run wire four-point grounding grid will obtain a low under pad to opening and allow 5'0" for grounding live front switch/fuse enclosures. ground resistance and minimize the touch po3. Place ground wire a minimum of 24" away from the side or sides of pad that a person tential between case and ground. Figure 5.16 would stand on to operate the equipment. The ground wire may be placed within 12" of the other sides. shows a typical layout. The ground conductor GUIDELINE ONLY should be a continuous wire connected to two points on the transformer. FIGURE 5.16: Grounding Grid for Pad-Mounted Equipment Installation. Transformer Installation (Front View)

1 8 6 – Se c t io n 5

5 Figure 5.17 shows an instalJoints/Intermediate lation that could be used at a Grounding Points BCN and jacketed cable joint or interThe 2007 NESC does not call semiconducting mediate neutral connection to for ground rods to be installed jacketed cables ground. These cable connecat direct-buried joints if the don’t need tions are aboveground to preconcentric neutral is effectively vent water from entering the intermediate grounds. grounded. However, jacket where the neutral is Rule 92B2b(3) recommends opened and sealed. An ideal additional connections besituation is shown in which a tween the concentric neutral continuous ground conductor is used to bond and ground for JCN systems. It also requires that the neutrals together and to make up the ground the neutral be grounded at each cable joint that is loop to and from the ground rod. not otherwise insulated to the voltage expected Figure 5.18 shows a direct-buried installation under normal conditions. Because jacketed cable that could also be used at a jacketed cable joint systems are not as well grounded as BCN sysor intermediate neutral connection to ground. All tems, any joint or splice should be used as a three neutrals are tied to ground by separate means for connecting the proper number of driconductors attached to ground rods. Two ven ground rods to improve system grounding. jumpers are added between the cable phases to Rule 96C says that JCN must be grounded at provide a continuous grounding loop, so one least four times per mile, not including grounds failed connection will not affect grounding. The at individual services. Rule 354D says that, for connection to the concentric neutrals is made random-lay installations with communication casimilar to the installation shown in drawing bles in the same trench, there shall not be less UM48-3 of RUS Bulletin 1728F-806 (D-806) than eight grounding installations in each mile, dated June 2, 2000. This connection should be not including the service grounds. Intermediate properly sealed around the concentric neutral to grounding is not required for BCN cables or prevent moisture entrance. semiconducting jacketed cables with jacket raFigure 5.19 shows a direct-buried intermediate dial resistivity less than 100 ohm-m. grounding assembly using in-line ground connectors. The principle is to strip the jacket from JCN Cable Joint a short piece of cable, wrap a braid brazed to a connecting rod around the concentric neutral, and seal the connection against moisture. The device holds promise as a quick and simple way to make an intermediate grounding point in a cable run. Extreme care should be used with this type of connection below ground so the jacket is adequately resealed to prevent moisture ingress. The installations shown in Figures 5.17 and 5.19 could be connected to three adequately spaced ground rods, if required for system Ground Rod grounding. Note that rods should be installed with an inter-rod distance equal to two rod lengths for a reasonable degree of effectiveness.

GUIDELINE ONLY

FIGURE 5.17: Installation of JCN Connection in Above-Grade Pedestal.

Service Entrance NESC Rule 250-84 requires one driven ground rod at the service entrance to a residence. Ground resistance is to be 25 ohms or less. If the desired resistance is not obtained, one rod must be added.


5 See Note 1

See Note 3

See Note 2

0” 10’ um im Min

0” 10’ um im Min

Notes: 1. #2 Thru 4/0 conductor—use #4 stranded copper ground wire, 500 kcmil conductor—use #2 stranded copper ground wire. 2. Engineer to specify number and length of ground rods. 3. Moisture seal around connections to the jacketed cable neutral. Use solid copper inside and extended through moisture seal. 4. Four grounds per mile minimum. More required with high ground resistance.

GUIDELINE ONLY

5. Use this grounding assembly only with proper sealing on concentric neutrals that prevent moisture permeating the insulation.

FIGURE 5.18: Grounding Assembly for JCN Underground Primary Cable.

In-Line Connecting Rod

Grounding Conductor Solid Copper (Continuous) #2–#4 as Required

Compression Connector Moisture Seal

Ground Rod(s)

Notes: 1. #2 AWG to 400 kcmil conductor— use #4 AWG solid copper ground wire. 500 kcmil to 1,000 kcmil conductor— use #2 AWG solid copper ground wire. 2. Engineer to specify number and length of ground rod(s). 3. Adequate moisture seal must be provided around connections to jacketed cable neutral. 4. It is recommended that connections to JCN be made above ground in an enclosure when feasible to preserve moisture integrity of jacket.

FIGURE 5.19: Intermediate Grounding Assembly, Underground Primary Cable.

As noted, the cooperative usually has no control over the ground resistance at the meter base. To lessen the probability that incoming surges on a JCN cable will cause damage to voltagesensitive consumer equipment, ensure that the service ground should have a value larger than the transformer ground. It is not practical for a cooperative to check every service ground in its territory to determine its relative resistance value compared with other system grounds. However, if problems arise because of failed equipment in the residence, the service ground would be a logical component to investigate. One installation that will provide a resistance to ground lower than the distribution transformer ground is a service neutral tied to the metal casing of a domestic water well. In this instance, trying to reduce the system ground would not be practical. Secondary metal oxide varistor (MOV) arresters with low discharge voltages are a possible solution. The arresters should be installed as close to the protected equipment as possible, preferably in the meter base rather than at the transformer. Also, the consumer should provide sensitive electronic equipment, such as personal computers, with individual protection.

1 8 8 – Se c t io n 5

5 Counterpoise Application for Insulated Jacketed Cable

Counterpoise is not frequently There is a method to estidiscussed in connection with mate the ground resistance of Use counterpoise BCN underground cable sysa counterpoise installation. only for insulated JCN tems. It is more often associVarious aspects affect the ated with transmission line ground resistance of the concable installations. tower-footing surge resistances ductor. A counterpoise preand line outage rates caused sents a surge impedance to by lightning. RUS requires cothe flow of lightning current. operatives to install cable with an insulating The impedance is different from the steady-state jacket. With increasing use of this cable, system ground resistance. Surge impedance affects riser ground quality is reduced in comparison with pole grounding and jacket overvoltage protecthe quality that could be had with BCN and tion. It is recommended that a continuous counsemiconducting jacketed cable. A counterpoise terpoise be installed from the riser pole to the is one method that will improve ground quality first transformer in the system. when insulated JCN cable is used. It is a conductor buried in the ground as a practical means COUNTERPOISE GROUND RESISTANCE to reduce ground resistance at a desired locaThe steady-state, or R60-Hz, resistance to ground tion. Lower ground resistance results from inof a counterpoise electrode can be calculated creasing the earth area in contact with the using Equation 5.4. grounding system. Installation of a counterpoise is particularly simple on underground systems Equation 5.4 because a trench is usually being opened. R=

50

40

Ω) 60-Hz Resistance (Ω

Counterpoise Wire 5/16” Diameter, 3-Strand, Galvanized, Annealed Iron Wire. Burial Depth = 30” 30

500 Ω-M

20 250 Ω-M 100 Ω-M

10

0 0

20

40

60

80

100 120 140 160 180 200 220 240 260 280 300 Length (Feet)

FIGURE 5.20: Counterpoise 60-Hz Resistance Variation with Length and Different Soil Resistivities.

where: ρ L a d

ρ 2L In –1 for d < L πL ad

= Soil resistivity, in ohm-m (Ω-m) = Conductor length, in meters (m) = Conductor radius, in meters (m) = Burial depth, in meters (m)

Figure 5.20 shows how the resistance of a #4 AWG copper wire varies with length in soils of different resistivities. Results are shown for burial depths of 30 and 42 inches. When a counterpoise is used only to improve surge arrester grounding, counterpoise lengths greater than 300 feet are not generally considered to be cost-effective. Counterpoise can be extremely helpful where upper layer soil resistivity is less than that of the soil below. When rock layers prevent driving rods of a suitable length to the proper depth, counterpoise may provide a workable alternative. So that the ground resistance does not vary widely during the year, special care should be taken to bury counterpoise below a stable moisture level. Burying below the frost line must also be considered. An analysis of Equation 5.4 shows that depth


5 does not dramatically affect counterpoise resistance. However, any increase in soil resistivity will increase the ground resistance proportionally. COUNTERPOISE SURGE IMPEDANCE When lightning current travels along a conductor, the resistance it encounters is the surge impedance, not the steady-state resistance. Surge

Equation 5.5 L ohms C

ZSURGE =

where: ZSURGE = Counterpoise surge impedance, in ohms L = Conductor inductance, in Henries/unit length C = Conductor capacitance, in Farads/unit length

150 140

Z = 150-Ω Initial Surge Impedance R = 10-Ω 60 Hz Resistance

130

Curves: Counterpoise Length 1. 1,000 ft 2. 750 ft 3. 500 ft 4. 250 ft

120 110 100

Surge Impedance (Ω)

90 80 70 60 50

1

40

2 3

30 4 20 10 0

60 Hz Resistance 1

2 3 Microseconds (µs)

4

5

FIGURE 5.21: Effect of Length on Transient Surge Impedance of Counterpoise.

impedance is usually designated by the symbol ZSURGE and is expressed by Equation 5.5. Transient or surge current initially “sees” the surge impedance of the conductor, whether it is hung in the air, buried in the ground, or run vertically on a riser pole. A horizontal buried counterpoise has an initial surge impedance that depends slightly on soil conditions and is assumed to be about 150 ohms. As the wavefront of a current surge travels along the conductor, more and more of its length helps to shunt the current to ground. The final result is that, after a series of reflections, the surge impedance reduces to the steady-state resistance, R60-Hz. The decay time depends on the length of the counterpoise and the propagation speed of the surge. Depending on the dielectric constant of the soil, a surge travels at less than one-half the speed of light (the speed of light is assumed to be 1,000 feet per microsecond). Tests have shown that a 1,000-foot counterpoise with an initial 150-ohm surge impedance will reach a resistance equal to its steady-state value in about six microseconds (6 µs). A shorter counterpoise of 250 feet will have the same 150-ohm initial surge impedance, but its steady-state resistance will occur in onefourth the time (1.5 µs). Curves one to four of Figure 5.21 show the relationship (Westinghouse T&D Reference Book, 1964). REASONS FOR COUNTERPOISE USE Counterpoise is buried with jacketed cable to reduce ground resistance at the point of application. Connection to the insulated cable neutral improves grounding of the neutral and reduces overall system ground resistance. If the counterpoise wire is run from the riser pole to the first transformer, it provides a parallel path for lightning currents to flow to ground. The additional path diverts surge current from the pole ground and JCN. Less surge current in the ground rod decreases neutral-to-ground voltage and, thus, the jacket voltage. Lower jacket surge voltages will reduce the probability of jacket puncture over time. Less current on the JCN means less current flowing to the transformer neutral. How the surge impedance of counterpoise affects the pole ground and the jacket voltage is shown by Figure 5.22. Two possible counterpoise configurations are shown. One is a continuous

1 9 0 – Se c t io n 5

5 MOV Arrester Riser Pole

Cable Termination S1 Cable Jacket R Dead-Front MOV Arrester

L


Service Loads

150' 100'

ZSURGE Counterpoise RG 25

S2

S3 RG

ZSURGE /R60–Hz Counterpoise

RT

RS

Legend: R L I RG

= = = =

Riser pole ground conductor resistance Riser pole ground conductor inductance Surge current in riser pole ground conductor Riser pole ground rod resistance

RG/25

= Remote ground rod resistance

ZSURGE

= Counterpoise surge impedance

RT

= Transformer ground rod resistance

RS

= Service entrance ground rod resitance

FIGURE 5.22: Counterpoise Application to Reduce Jacket Voltage.

by the addition of counterpoise will be less than counterpoise connected directly to the JCN at the no-counterpoise case. the top of the pole and extending to the first A full-length counterpoise connected to the transformer. The other is connected to the riser pad-mounted transformer pole ground rod and extends neutral puts its surge imto a remote ground (RG/25) that pedance in parallel with the measures at least 25 times less Always connect transformer ground rod, cable than the pole ground (RG). counterpoise at the concentric neutral, and the In the examples in this secservice neutral. The connection, it is assumed an incoming top of the riser pole. tion ensures the jacket voltage lightning surge with a set magwill be less at the transformer nitude and rate of rise will than at the riser pole. The produce a certain voltage parallel impedance reduces the surge current across the cable jacket, with no counterpoise apflowing on the JCN from the riser pole and the plied. Any changes in the jacket voltage caused


5 EXAMPLE 5.1: No Counterpoise Added (Switches S1, S2, and S3 Open). When a surge arrester conducts, lightning current will split between the pole ground conductor and the JCN in proportion to their respective surge impedances. The surge impedance of the concentric neutral depends on the geometry of the cable and the dielectric constant of the jacket material. It will fall somewhere between the 35-ohm cable surge impedance and the 150-ohm counterpoise surge impedance. If the pole ground has a surge impedance of less than 15 ohms, most of the current will be diverted to the ground rod. One component of the peak jacket voltage at the sending end of the cable is then equal to the ground potential rise caused by surge current flow through the ground rod. The jacket voltage at the transformer or receiving end of the cable will not be the same as the sending end because the voltage on the cable neutral is determined by the respective currents flowing in each ground and the resistance of each ground. The receiving end voltage will always be the smaller of the two. Another component of the neutral-to-ground voltage at the riser pole is the L di/dt voltage of the pole ground conductor. For the configuration depicted in Figure 5.22, the total neutral-to-ground voltage can be represented by Equation 5.6.

EXAMPLE 5.2: Attaching a 100-Foot Counterpoise to the Riser Pole Ground Rod and the Other End to a Remote, Smaller Resistance (Switch S2 Closed, S1 and S3 Open). This case represents laying counterpoise terminated in a ground rod or running a connection to an existing electrode to decrease 60-Hz grounding. Initially, the counterpoise will present a 150-ohm impedance in parallel with the riser pole ground rod. For surge currents peaking in 0.5 to eight microseconds, the slight decrease in the ground resistance will reduce ground potential rise by a factor depending on the difference between the magnitude of the riser pole ground (RG) and the 150-ohm surge impedance of the counterpoise. This counterpoise installation will not reduce jacket voltages very much, but still should be considered to improve system grounding.

Equation 5.6 Vng = I(R + RG) + L di/dt where: Vng I R RG L di/dt

= = = = = =

Riser pole neutral-to-ground voltage, in volts Current in riser pole ground conductor, in amperes Riser pole ground conductor resistance, in ohms Riser pole ground rod 60-Hz resistance, in ohms Pole ground conductor inductance, in Henries Surge current rate of rise, in amperes per second

For a standard 8 × 20 µs current waveform, maximum di/dt occurs during the initial part of the wavefront. Laboratory tests have shown that the L di/dt component is usually less than the IR component and will peak before the surge current waveform peaks. Therefore, the peak neutral-to-ground voltage and, thus, the peak jacket voltage is caused mainly by the surge current magnitude: I (R + RG). Because R is less than RG, the jacket peak voltage can be accurately represented by the product of the pole ground conductor current and the surge impedance of the ground rod. However, for steep-front currents peaking in less than two microseconds, the L di/dt voltage could exceed the IR component in the case of a low down-lead current. The same could happen for a high di/dt and low ground surge impedance. In both cases, the L di/dt component would predominate and produce peak jacket voltage.

EXAMPLE 5.3. Continuous or Full-Length Counterpoise (Switches S1 and S3 Closed, S2 Open). In Figure 5.22, the counterpoise is shown connected in parallel with the jacketed concentric neutral at both ends of the cable. With the counterpoise run to the top of the riser pole, its surge impedance is connected directly in parallel with the surge impedance of the concentric neutral and the down-lead conductor. This connection will reduce the current to the pole ground. As a consequence, it also lowers both the ground potential rise and the L di/dt component of the down-lead voltage. Connecting a continuous counterpoise at the riser pole ground rod, as explained in Example 5.2, will not give the same effect. Test data have shown that connection of continuous counterpoise to JCN cable near the riser pole arrester will reduce jacket voltages by up to 50 percent for fast-front waveforms and 35 percent for slow-front waveforms (General Electric, July 1990).

1 9 2 – Se c t io n 5

5 amount of current on the service neutral. The counterpoise will also divert transformer MOV arrester current from the service neutral in case of a high transformer ground resistance (RT). It also improves the 60-Hz ground resistance at the pad-mounted transformer. Although the previous discussion mentioned only direct-buried JCN cable, very similar results will be obtained for JCN cable installed in conduit. RECOMMENDATIONS FOR JACKETED CABLE 1. Obtain a low ground resistance (10 ohms or less is desired) at the riser pole for any jacketed cable installation. This strategy is best to reduce jacket voltages for all types of surges, fast-front or slow-front. The inductance of the pole ground conductor cannot be reduced. A measured ground resistance of 10 ohms or less is desired at riser poles. 2. Counterpoise will reduce jacket voltages to some extent, regardless of the riser pole ground resistance.

System Ground Resistance Measurement and Calculation

3. The counterpoise must be attached to the insulated JCN at the top of the riser pole to obtain optimum jacket voltage reduction. 4. A continuous counterpoise should be installed to the first transformer for every underground installation, if practicable. 5. If a full-length counterpoise cannot be justified economically, counterpoise of 100 to 300 feet should be installed at the riser pole, depending on soil resistivity and condition. 6. The conductor is to be random-lay in the same trench as the cable. The counterpoise must be surrounded by soil. A driven ground rod is used to terminate the counterpoise conductor. The rod is counted toward the four- or eight-grounds-per-mile NESC requirement. 7. Counterpoise will not significantly reduce touch potentials on jacketed cable installations. Therefore, proper safety procedures must be followed.

equal thickness; therefore, the shell nearest the rod has the smallest surface area and consequently the greatest resistance. The farther the shell is from the rod, the greater the surface area, which results in a lower resistance in the shell. At some remote point, an additional shell does not significantly add to the earth The main component resistance surrounding the of ground resistance rod. The final shell is considered the effective resistance is resistance of the area and depends on the driearth surrounding ven depth and the diameter of the ground rod. the ground rod.

FIELD MEASUREMENT OF SYSTEM GROUNDS To correctly measure the resistance of a system ground, the engineer needs to understand ground resistance. Ground resistance consists of the following: • Resistance of the ground rod, • Resistance of the contact between the ground rod and the soil directly in contact with the rod, and • Resistance of the body of earth surrounding the ground rod. The resistance of the ground rod and the contact resistance are usually extremely small compared with the earth resistance. To understand earth resistance, think of the ground rod as being surrounded by concentric shells of earth (see Figure 5.23). These shells have

To measure ground resistance, use a three-point or clamp-on ground resistance tester.

Three-Point Meter A three-point ground resistance tester can measure the ground resistance of the following: • A single ground rod, • Multiple ground rods, and • Small grids of ground conductor.


5

Current

Current

FIGURE 5.23: Earth Resistance.

X

P

C

Grounding Electrode Under Test

Test Probe

X

Grounding Electrode Under Test

Test Probe

P

These measurements must be made before the ground rod or grid is connected to the system ground. Unfortunately, the tester cannot practically measure the ground resistance of a long counterpoise. This ground resistance tester has three terminals as shown in Figure 5.24. The current (C) terminal and the potential (P) terminal are each connected to separate test probes. The third terminal (X) is attached to the grounding electrode that is being tested. Figure 5.24 shows the correct test setup. The tester injects a current through test probe C and grounding electrode X. The resulting potential drop is measured between test probe P and grounding electrode X. The resistance reading shown on the test is the ground resistance of the electrode. This test procedure is known as the Fall-of-Potential Method. For additional information on this test method, see IEEE Standard 81. During this test, it is important to space the test probes and electrodes correctly. If the three electrodes are too close together, then the effective resistance areas of probes C and X will overlap (see Figure 5.25). This overlapping produces inaccurate resistance readings. For readings to

Test Probe

P

C

X

C

Effective Resistance Areas Overlap

Effective Resistance Areas Do Not Overlap

62% of D

Test Probe

D

Resistance of Test Probe C

Resistance of Grounding Electrode

FIGURE 5.24: Correct Ground Resistance Test Setup.

Resistance

Resistance

D

Distance

FIGURE 5.25: Incorrect Ground Resistance Test Setup.

1 9 4 – Se c t io n 5

5 TABLE 5.3: Spacing of Test Probes for Testing Resistance of a Single Ground Rod. Source: AEMC Corp., 1990. Depth of Driven Rod (ft)

Distance to P (ft)

Distance to C (ft)

6

45

72

8

50

80

10

55

88

12

60

96

18

71

115

20

74

120

30

86

140

TABLE 5.4: Spacing of Test Probes for Testing Resistance of an Electrode System. Source: Biddle Instruments, 1981. Maximum Dimension (ft)

Distance to P (ft)

Distance to C (ft)

2

40

80

4

60

100

6

80

125

8

90

140

10

100

160

12

105

170

14

120

190

16

125

200

18

130

210

20

140

220

40

200

320

60

240

390

80

280

450

100

310

500

120

340

550

140

365

590

160

400

640

180

420

680

200

440

710

be correct, the spacing must be increased so the effective resistance areas do not overlap. Table 5.3 lists the recommended distances for probe C when testing a single ground rod. Also listed are the distances to the P probe. Probe P is placed at 62 percent of the distance from the ground rod to the C test probe. As shown in Figure 5.24, the 62 percent method should place the potential probe outside the effective resistance area of the other two electrodes. The preferred placement for P is in a straight line between C and X. To test multiple ground rods or small grids, increase the distance to the C probe. Table 5.4 provides a list of recommended spacing for the C and P test probes. The maximum dimension is the diagonal distance across the electrode system area. For example, if four rods form a square with 20-foot sides, the electrode system area is 20 feet × 20 feet. This area has a diagonal of approximately 28 feet. Using Table 5.4, choose the next highest maximum dimension, which is 40 feet. The table shows that P should be at 200 feet and C at 320 feet. The preferred placement for P is still at 62 percent of the total distance and is in a straight line between C and the electrical center of the electrode system. Most three-point meters have a resistance range of 0 to 500 ohms and are accurate for test probe resistance values of up to 5,000 ohms. Most newer models have an indicator to signal the operator if the test probe resistance values are excessive or if there is a lack of continuity between the leads and the test electrode. Clamp-On Meter Another type of meter used to make ground resistance measurements is the clamp-on ground resistance tester shown in Figure 5.26. This instrument clamps around a ground rod or ground conductor and displays a resistance reading. Unlike the three-point test, this measurement is made with the ground rod or conductor still connected to a multigrounded system. The tester contains a constant voltage source that induces a current into the test ground. This current is detected and used to determine the resistance.


5 Figure 5.27 shows a circuit diagram for a multigrounded system with the clamp-on tester in place. Rx represents the ground being measured. R1 through Rn represent the remaining grounds in a multigrounded system. Because the parallel combination of R1 through Rn is much smaller than Rx, most of the voltage drop is across Rx. Therefore, the resistance reading on the meter is basically the value of Rx. To work properly, the meter must be clamped on a ground rod or conductor that has only one return path to the neutral. If the meter is clamped onto a ground loop, the induced current will circulate around the loop and the meter will show a very low resistance reading. Placing the clamp below the loop or disconnecting one side of the loop forces the induced current to flow through the test ground (Rx). Ground loops are often inside pad-mounted transformers. Here, several ground conductors and one or more ground rods are bonded together. Clamping the meter around the ground rod and below the common attachment point should allow an accurate ground resistance reading of the rod. This test setup is shown in Figure 5.28. This meter has a resistance range of two to 1,990 ohms and a ground current range of zero to 1.99 amperes. If the ground current exceeds 1.99 amperes during the test, ground resistance measurements are not possible.

FIGURE 5.26: Clamp-On Ground Resistance Tester. Source: AEMC Corporation, 1992.

Ground Resistance of Multigrounded System I E

RX

R1

R2

Rn–1

Rn

Ground Resistance of Ground Rod that is being Tested

FIGURE 5.27: Circuit Diagram for Multigrounded System.

Front View of Transformer H1B X3

H1A

Ground Strap

X1

X2 Copper Ground Wire

Tank Grounds Copper Ground Wire Clamp-On Ground Resistance Meter (See Note 1)

Ground Rod Clamps Note: 1. For best reading, clamp meter onto the ground rod itself, below the point where ground conductors are attached.

FIGURE 5.28: Ground Resistance Test Setup for Clamp-On Tester.

1 9 6 – Se c t io n 5

5 conducted at substation sites or along transmission lines. If Knowing the soil the cooperative engineer does resistivity helps in not have soil data for the area of underground cable installathe design of an tion, then soil resistivity meaadequate grounding surements may be necessary. system. After the engineer collects soil data from different areas, he may be able to assign approximate resistivity values throughout the service territory. It will probably become apparent that each different soil type present in the service area has a relatively narrow range of resistivity. These approximate values can be used instead of measuring the soil resistivity for every underground system that is to be installed.

SOIL RESISTIVITY MEASUREMENTS In addition to thermal resistivity (discussed in Section 4 of this manual), soil has an electrical resistivity. The electrical resistivity is the resistance of a unit cross-sectional area of soil per unit length and is expressed by Equation 5.7.

Equation 5.7 ρ=R× where: ρ R A L

A L

= Soil resistivity, in ohm-m = Resistance, in ohms = Cross-sectional area, in meters2 = Length, in meters

Soil resistivity directly affects the resistance-to-ground of a grounding electrode. Knowing the soil resistivity for a particular site allows the engineer to design adequate grounding for the underground cable system. The cooperative may have soil resistivity data from tests

C1 P1 P2 C2

C1

P1

P2

b

Small-Sized Electrodes a

a

FIGURE 5.29: Setup for Soil Resistivity Test.

a

Four-Point Meter Measuring soil resistivity requires use of a fourpoint earth resistance tester. This tester is similar to the three-point tester and can be used to measure the resistance-to-ground of a ground electrode. However, the threepoint tester will not measure soil resistivity. The four-point Use a four-point tester is more sensitive than the earth resistance three-point tester, measuring values as low as 0.01 ohms. As tester to measure evident from the name, the soil resistivity. four-point tester has four terminals instead of three (see Figure 5.29). Measuring soil resistivity requires that four test probes be driven in the ground. The test probes must be equally spaced and in a straight line as shown in Figure 5.29. It is important that all test probes are driven to the same depth. A depth of six to 18 inches is acceptable. Equally important, the test probes must have good soil contact. Loose test probes can lead to erroneous readings C2 because of high probe resistance. The tester continues the test setup by placing test leads from the four terminals to the four test probes. The two current terminals (C1 and C2) connect to the two outer test probes. The two potential terminals (P1 and P2) connect to the two inner test probes. Figure 5.29 also illustrates these connections. The tester injects a current into the two outer probes and measures the


5 corresponding potential drop between test probes P1 and P2. Using these two values, the tester determines the resistance. This resistance value is what the operator reads when making soil resistivity measurements. (For more information on this test method, refer to IEEE Standard 81-1983.) Most four-point testers give indication of high probe resistance. If an operator gets this indication, he should first check to see if test probes are secure in the ground. If test probes are loose, the operator needs to drive rods deeper or relocate one or more rods. If the tester still shows high probe resistance, then the operator needs to pour water around each test probe to help reduce the test probe resistance so measurements can be made. Test accuracy will not be affected if the probe spacing significantly exceeds the diameter of the wetted area. The resistance value shown on the four-point tester is a function of the apparent soil resistivity.

This apparent resistivity is the average resistivity for a block of soil with a depth equal to the spacing between the test probes. For example, if the test probe spacing is five feet, then the resistance reading is the average resistivity to a depth of five feet. To get a complete soil profile, the operator needs to take measurements at various probe spacings. Elements Affecting Soil Resistivity Several elements affect soil electrical resistivity, including the following: • • • •

Soil type, Moisture and chemical content, Temperature, and Seasonal variations.

Different soil types have different resistivity values, as shown in Table 5.5.

TABLE 5.5: Soil Resistivities for Different Soil Types and Geological Formations. Adapted from IEEE Standard 81-1983. Earth Resistivity (ohm-m)

Quaternary

Cretaceous Tertiary Quaternary

Carboniferous Triassic

Cambrian Ordovician Devonian

Precambrian and Combination w/Cambrian

1 Sea Water Loam 10 Unusually Low

Clay Chalk

30 Very Low

Chalk Trap Diabase

100 Low

Shale

300 Medium

1,000 High

3,000 Very High

10,000 Unusually High

Limestone

Shale

Sandstone

Limestone Sandstone

Sandstone

Dolomite

Quartzite

Coarse

Slate

Sand and

Granite

Gravel in

Gneiss

Surface Layers

1 9 8 – Se c t io n 5

5 Typically, the earth surface is composed of layers of different soil types. These soil types have varying resistivities; therefore, soil resistivity measurements often show different values at different depths.

Moisture and chemical content dramatically affect soil An increase in resistivity. The moisture dismoisture and salt solves the naturally occurring salts in the soil. The resulting content decreases electrolyte improves the consoil resistivity. duction of current through the soil and, thus, reduces the soil resistivity. As the moisture content increases, the soil resistivity decreases. This decrease is rapid until the moisture content reaches 20 percent to 30 percent (see Figure 5.30). The amount of dissolved salt that is present in the soil also affects the resistivity. As the salt content increases, the resistivity decreases. However, the decrease in resistivity is minimal after the salt content reaches five percent. The graph of Figure 5.31 shows the effect of salt in soil that contains 30 percent moisture. A third element affecting soil resistivity is temperature. Temperatures above freezing have little effect on resistivity. However, as the temperature drops below freezing, the soil resistivity increases rapidly. To illustrate this effect, Table 5.6 shows

10,000 5,000

Soil Resistivity (Ω Ω-m)

1,000 500

100 50

0

5

10

15

20

25

30

35

40

45

Percentage Moisture

FIGURE 5.30: Effects of Moisture on Soil Resistivity. Adapted from IEEE Standard 80-1986.

Soil resistivity varies as a result of seasonal changes.

10,000 5,000

TABLE 5.6: Effect of Temperature on Soil Resistivity. Adapted from Biddle Instruments, 1981.

Soil Resistivity (Ω Ω-m)

1,000 500

Temperature

Resistivity

100

°C

50

20

68

72

10

50

99

1

2

3

4

5

6

7

8

9

10

Percentage Salt

FIGURE 5.31: Effects of Salt Content on Resistivity in Soil Containing 30 Percent Moisture. Adapted from IEEE Standard 80-1986.

°F

(ohm-m)

0

32 (water)

138

0

32 (ice)

300

-5

23

790

-15

14

3,300


5 how temperature affects the resistivity of sandy loam that contains 15.2 percent moisture. At the freezing point, the resistivity more than doubles. Soil temperature and moisture content usually vary throughout the year. As a result, soil resistivity also varies throughout the year. These changes must be considered when the grounding for an underground system is designed. The ground resistance of a counterpoise increases if the soil around it freezes during the winter months. Table 5.6 shows that the soil resistivity increases from 138 to 300 ohm-m at the freezing point. This change produces a proportional change in the ground resistance of the counterpoise. A counterpoise with a ground resistance of 38 ohms in the summer could increase to 83 ohms when the ground freezes. The counterpoise should thus be buried below the frost line. Likewise, ground rods should be driven to a depth that is below the frost line. In some areas, the summer months are often dry. As the soil around the grounding electrode dries out, its ground resistance increases. If the loss of moisture increases the soil resistivity by 50 percent, then the ground resistance of the electrode will also increase by 50 percent. Resistance increase caused by loss of soil moisture is a major concern. A reduction of moisture content from 25 percent to 15 percent can cause electrode resistance to triple. Extending ground rods into an area with permanent moisture content can minimize this problem. A rod that extends into the water table has a more stable ground resistance. Because seasonal changes can affect soil resistivity, it is important to note the temperature and the soil moisture content of the soil at the time of a four-point soil resistivity test. This information will help the engineer design a grounding system that performs effectively throughout the year. Equation for Deriving Resistivity from Resistance Reading The measurements made with a four-point tester are resistance values. These resistance values must be converted to soil resistivity measurements using Equation 5.8.

Equation 5.8 ρ=

4πaR 2a a 1+ – 2 2 2 a + 4b a + b2

where: ρ = Soil resistivity, in ohm-m R = Resistance readings, in ohms a = Spacing between test probes, in meters b = Depth of test probe, in meters

If the depth (b) of the probes is small (five percent of the probe spacing), Equation 5.8 reduces to ρ = 2πaR The resistivity values can be plotted against the test probe spacings. This information is needed to determine an appropriate soil model. SIMPLIFIED DESIGN OF GROUNDING SYSTEM USING RESISTIVITY DATA Resistance-to-ground (ground resistance) calculations can be used when a grounding system is designed. Using these calculations, the engineer can compare the ground resistance of several ground electrode configurations: • A single ground rod, • Groups of ground rods, and • Counterpoise.

The ground resistance of an electrode is a function of the soil resistivity and the electrode geometry. Therefore, the calculations require Use ground resistance knowledge of the soil recalculations to sistivity. If the soil resistivcompare grounding ity is unknown, the engineer can get this insystems. formation from a soil resistivity test as described earlier in this section.

2 0 0 – Se c t io n 5

5 Equation 5.9

Equation 5.12 R=

where: R ρ L a

= = = =

ρ 4L In –1 2πL a

Ground resistance, in ohms Soil resistivity, in ohm-m Ground rod length, in meters Ground rod radius, in meters

R= where: R ρ L a A

ρ 4L ρ L2 2 L4 1– 2 + In –1 + 3s 4πL a 4πs 5 s4

where: R ρ L a s

= = = = =

Ground resistance, in ohms Soil resistivity, in ohm-m Ground rod length, in meters Ground rod radius, in meters Distance between ground rods, in meters

Equation 5.11 R=

ρ 4L 4L s s2 s4 + In + In –2 + – 2 4πL a s 2L 16L 512L4

where: R ρ L a s

= = = = =

Ground resistance, in ohms Soil resistivity, in ohm-m Ground rod length, in meters Ground rod radius, in meters Distance between ground rods, in meters

The most simple grounding electrode is a single ground rod. Equation 5.9 (Dwight, 1936) provides the ground resistance of a single rod. The ground resistance of two ground rods in parallel separated by a distance, s, is given in Equations 5.10 and 5.11 (Dwight, 1936). If the spacing between the rods is greater than the rod length (s > L), use Equation 5.10. If the spacing is less than the length (s < L), use Equation 5.11.

2

Ground resistance, in ohms Soil resistivity, in ohm-m Ground rod length, in meters Ground rod radius, in meters Area occupied by ground rods, in meters2 K1 = Constant obtained from Figure 5.32 n = Number of rods in the group

Equation 5.10 R=

ρ 4L 2K L In –1 + 1 n – 1 2πnL a A = = = = =

Equation 5.13 R=

ρ 2L L 2d d2 d4 + In In –2 + – 2 + 4 L 2L 2πL a d L

where: R ρ L a d

= = = = =

Ground resistance, in ohms Soil resistivity, in ohm-m Length of counterpoise, in meters Radius of counterpoise, in meters Depth of counterpoise burial, in meters

Equation 5.14 R= where: R ρ L a d

= = = = =

ρ 2L In –1 πL ad

Ground resistance, in ohms Soil resistivity, in ohm-m Length of counterpoise, in meters Radius of counterpoise, in meters Depth of counterpoise burial, in meters

The equation becomes more complicated for groups of ground rods that are connected. In addition to individual rod geometry, the area (A) occupied by the group of rods and a coefficient (K1) affect the equation. The coefficient K1 is related to the geometry of the rod group and can


5 1.40 1.35 1.30 1.25 A Curve A: For Depth h = 0

Coefficient K1

1.20

K1 = –0.04x + 1.41 1.15

Curve B: For Depth h =

Area 10

K1 = –0.05x + 1.20

1.10

B

Curve C: For Depth h =

1.05

Area 6

K1 = –0.05x + 1.13 C

1.00 0.95 0.90 0.85

1

2

3

4

5

6

7

8

Length-to-Width Ratio, X

FIGURE 5.32: Coefficient K1 for Ground Resistance Calculations. Adapted from IEEE Standard 80-1986.

EXAMPLE 5.4: A Single 8-Foot × 3/4-Inch Ground Rod Driven in Soil with a Resistivity of 250 Ohm-M. To calculate the ground resistance, use Equation 5.9:

R=


where: ρ = 250 Ω-m L = 8 ft (0.3048 m/ft) = 2.44 m a = .5 (.75 in.)(0.0254 m/in.) = 0.0095 m

By substituting the values,

R=

250 Ω-m 4 × 2.44m In –1 = 96.8Ω 2π(2.44m) 0.0095m

be obtained from the graph of Figure 5.32 or from the associated equation. Equation 5.12 provides the ground resistance of a group of ground rods (Schwarz, 1954). Equation 5.13 provides the ground resistance of a counterpoise (Dwight, 1936). In underground system applications, the length of the counterpoise is often much greater than the depth of burial. For these cases, a simpler equation, Equation 5.14, provides a suitable approximation of the ground resistance value. Examples 5.4, 5.5, and 5.6 illustrate how the ground resistance of a grounding system changes for different configurations. Increasing the number of ground rods is one way to decrease the ground resistance. However, the spacing between the ground rods affects how much the ground resistance decreases. As the separation increases, the ground resistance decreases.

2 0 2 – Se c t io n 5

5 EXAMPLE 5.5: Two 8-Foot × 3/4-Inch Ground Rods Placed 5 Feet Apart. Because the spacing is less than the ground rod length (s < L), use Equation 5.11: R=

ρ 4L 4L s s2 s4 In + In –2 + – + 4πL a s 2L 16L2 512L4

where: ρ = 250 Ω-m L = 8 ft = 2.44 m

a = 0.375 in.(0.0254 m/in.) = 0.0095 m s = 5 ft (0.3048 m/ft) = 1.52 m

By substituting the values, R=

250 Ω-m 4 × 2.44m 4 × 2.44m 1.52m (1.52m)2 (1.52m)4 In + In –2 + – + = 57.7Ω 4π(2.44m) 0.0095m 1.52m 2 × 2.44m 16(2.44m)2 512(2.44m)4

The addition of a second rod reduced the ground resistance from 96.8 to 57.7 ohms, approximately 40 percent. If the two rods are spaced further apart, the ground resistance becomes even lower.

EXAMPLE 5.6: Two Rods Spaced 16 Feet Apart. For a spacing of 16 feet, use Equation 5.10: R=

ρ 4L P L2 2 L4 In –1 + 1– 2 + 3s 4πL a 4πs 5 s4

where: ρ = 250 Ω-m L = 8 ft = 2.44 m

a = 0.375 in.(0.0254 m/in.) = 0.0095 m s = 16 ft (0.3048 m/ft) = 4.88 m


2 (2.44m)4 250 Ω-m 4 × 2.44m 250 Ω-m (2.44m)2 In –1+ 1– + = 52.2Ω 2 3(4.88m) 5 (4.88m)4 4π(2.44m) 0.0095m 4π(4.88m)

The increased spacings reduced the ground resistance from 57.7 to 52.2 ohms. The distance between the rods can be increased until there is no mutual resistance effect. The ground resistance of the two rods is equal to the parallel combination of two individual rods. Because two identical rods have the same ground resistance, the parallel resistance is one-half the single rod ground resistance. For the two eight-foot by 3/4-inch ground rods, the limiting ground resistance value is

1 R = R1 2 where: R1 = 96.8 Ω By substituting the values, 1 R = (96.8 Ω) = 48.4 Ω 2 This is not a significant improvement from the 52.2 ohms at a 16-foot spacing.


5 EXAMPLE 5.7: Group of Four Rods. Using a grouping of four ground rods gives a more dramatic improvement. For this example, use the layout of Figure 5.33.

16 Feet (4.9 Meters)

Using Equation 5.12,

R=

ρ 4L 2K L In –1 + 1 n – 1 2πnL a A

where: ρ = 250 Ω-m L = 2.44 m a = 0.0095 m

16 Feet (4.9 Meters)

FIGURE 5.33: Grouping of Four Ground Rods with 16-Foot Spacing.

2

n = 4 K1 = 1.375 (obtained from Figure 5.32) A = (4.88 m)2


250 Ω-m 4 × 2.44m 2(1.375)(2.44m) 2 In –1 + 4 – 1 = 29.8Ω 2π(4)(2.44m) 0.0095m (4.88m)2

5 Feet (1.5 Meters)

Increasing Increasing the the number of ground number of ground rods decreases decreases the the rods ground ground resistance. resistance. The area occupied by the rods also affects the ground resistance. A smaller area results in a higher ground resistance. For example, consider the arrangement of Figure 5.34. Here,

5 Feet (1.5 Meters)

FIGURE 5.34: Grouping of Four Ground Rods with 5-Foot Spacing.

R=

250 Ω-m 4 × 2.44m 2(1.375)(2.44m) 2 In –1 + 4 – 1 = 42.2Ω 2π(4)(2.44m) 0.0095m (1.52m)2

The soil resistivity test may Another way to reduce show that the soil has two layground resistance is to Increasing the rod ers with different resistivity increase the rod length. length decreases the values. If a driven rod is in (See Example 5.8.) contact with the two layers, then Soil resistivity influences ground resistance. its ground resistance will differ ground resistance. Ground from the ground resistance in resistance is directly proporhomogeneous soil. If the lower tional to soil resistivity; for exlayer has a lower resistivity than the upper layer ample, a 20 percent decrease in soil resistivity does, then driving a rod into the lower layer redecreases the ground resistance by 20 percent. duces the ground resistance of the rod. (See Example 5.9.)

2 0 4 – Se c t io n 5

5 EXAMPLE 5.8: Increase in Rod Length.

EXAMPLE 5.9: Change in Soil Resistivity.

Use Equation 5.9 to calculate the ground resistance of a 16-foot ground rod:

If the soil resistivity is 100 instead of 250 ohm-m, the resistance of a single eight-foot rod changes from 96.8 (Example 5.4) to 38.7 ohms. This resistance can be calculated in two ways. The first uses Equation 5.9 with ρ = 100 ohm-m:

R=


where: ρ = 250 Ω-m L = 16 ft (0.3048 m/ft) = 4.88 m a = 0.0095 m

R=

The second method calculates R based on its direct proportionality to ρ:

By substituting the values,

R=

R100 100 Ω-m = R250 250 Ω-m

250Ω-m 4 × 4.88m In –1 = 54.0Ω 2π(4.88m) 0.0095m

Doubling the rod length decreased the ground resistance from 96.8 (Example 5.4) to 54.0 ohms, a 44 percent reduction. If the ground rod length is 24 feet (7.32 m), then the following results:

R=

100Ω-m 4 × 2.44m In –1 = 38.7Ω 2π(2.44m) 0.0095m

250Ω-m 4 × 7.32m In –1 = 38.2Ω 2π(7.32m) 0.0095m

This is a reduction of 61 percent. However, this calculation has used the assumption that the average soil resistivity is constant for eight-foot, 16-foot, and 24-foot rods. This is generally not the case. Soil resistivity often decreases substantially between the surface and a depth of 24 feet.

Equation 5.15 ρa =

L(ρ1ρ2)

ρ2H + ρ1(L – H)

where: ρa = Apparent resistivity, in ohm-m ρ1 = Soil resistivity of top layer, in ohm-m ρ2 = Soil resistivity of bottom layer, in ohm-m L = Ground rod length, in meters H = Thickness of top soil layer, in meters Calculating the effect of a two-layer soil requires the use of an apparent soil resistivity, ρa, as defined in Equation 5.15. Equation 5.15 is valid only when the ground rod is in contact with both soil layers (IEEE Standard 80-1986). The apparent resistivity, ρa,

where: R250 = 96.8 Ω R100 = 96.8 Ω

100 = 38.7 Ω 250

Table 5.7 shows the ground resistance of a single eight-foot by 3/4-inch ground rod driven in varying soil resistivities. TABLE 5.7: Ground Resistance in Varying Soil Resistivities. Ground Resistance R (ohm)

Soil Resistivity (ohm-m)

10

26

15

39

25

65

50

130

75

195

100

260

500

1,300

1,000

2,600

This table shows how difficult it is to achieve a low ground resistance with a single eight-foot ground rod. replaces the soil resistivity, ρ, in Equations 5.9 through 5.12.


5 EXAMPLE 5.10: The Effect of a Tw0-Layer Soil with a Top-Layer Resistivity of 250 Ohm-M and a Bottom-Layer Soil Resistivity of 50 Ohm-M. The top-layer thickness is 5 feet. Using Equation 5.15, ρa = where: ρ1 ρ2 L H

= = = =

L(ρ1ρ2)

ρ2H + ρ1(L – H)

250 Ω-m 50 Ω-m 2.44 m 5 ft (0.3048 m/ft) = 1.52 m

Substituting the values yields the following: ρa=

2.44m(250 Ω-m)(50 Ω-m) =99.7 Ω-m (50 Ω-m)(1.52m) + 250 Ω-m (2.44m – 1.52m)

The value ρa replaces ρ in Equation 5.9:

R=

ρa 2πL

In

4L –1 a

where: ρa = 99.7 Ω-m L = 2.44 m a = 0.0095 m Substituting the values yields the following: R=

99.7Ω-m 4 × 2.44m In –1 = 38.6Ω 2π(2.44m) 0.0095m

Rod contact with the more conductive lower layer reduced the ground resistance of a single eight-foot rod from 96.8 (Example 5.4) to 38.6 ohms. The lower layer is even more effective if a longer ground rod is driven. For example, a 16-foot (4.88 m) rod changes ρa to ρa =

4.88m(250 Ω-m)(50 Ω-m) =66.6 Ω-m (50 Ω-m)(1.52m) + 250 Ω-m (4.88m – 1.52m)

Equation 5.9 yields the following:

R=

66.6Ω-m 4 × 4.88m In –1 = 14.4Ω 2π(4.88m) 0.0095m

The presence of a more conductive lower layer reduced the ground resistance of a 16-foot rod to 14.4 from 54.0 ohms, as calculated in Example 5.8.

2 0 6 – Se c t io n 5

5 EXAMPLE 5.11: Counterpoise of #2 AWG Conductor Buried 30 Inches Deep for a Distance of 100 Feet. The burial depth is much smaller than the counterpoise length (d < L). Therefore, use Equation 5.14: R= where: ρ L d a

= = = =

ρ 2L In –1 πL ad

Examples 5.11, 5.12, and 5.13 calculate the ground resistance of different counterpoise configurations. A more conductive soil (a lower ρ) also reduces the ground resistance, as seen in Example 5.12. The depth of burial (see Example 5.13) is another element that affects ground resistance.

250 Ω-m 100 ft (0.3048 m/ft) = 30.48 m 30 in. (0.0254 m/in.) = 0.76 m 1/2 (0.292 in.)(0.0254 m/in.) = 0.0037 m

A more conductive soil reduces the ground resistance.

Substituting the values yields the following R=

EXAMPLE 5.12: More Conductive Soil.

250Ω-m 2 × 30.48m In –1 = 15.8Ω π(30.48m) (0.0037m)(0.76m)

As the length of the counterpoise increases, the ground resistance decreases. For example, if the counterpoise length is increased to 200 feet (60.96 m), then

R=

250Ω-m 2 × 60.96m In –1 = 8.8Ω π(60.96m) (0.0037m)(0.76m)

If the soil resistivity is 100 ohm-m instead of 250 ohm-m, a 100-foot counterpoise will have a ground resistance of: R=

Doubling the counterpoise length reduced the ground resistance by 44 percent.

100Ω-m 2 × 30.48m In –1 = 6.3Ω π(30.48m) (0.0037m)(0.76m)

This is a 60 percent reduction from 15.8 ohms (Example 5.11). Because soil resistivity and ground resistance are directly proportional, a 60 percent reduction in ρ produces a 60 percent reduction in R.

EXAMPLE 5.13: Counterpoise Burial Depth. If the burial depth is increased to 60 inches (1.52 m) and the counterpoise length remains at 100 feet (30.48 m), then, using Equation 5.13, R=

ρ 2L L 2d d2 d4 In + In –2 + – 2 + 4 L 2L 2πL a d L

where: ρ = 250 Ω-m L = 30.48 m

a = 0.0037 m d = 1.52 m

Substituting the values yields the following: R=

250Ω-m 2 × 30.48m 30.48m 2 × 1.52m (1.52m)2 (1.52m)4 In + In –2 + – + = 14.1Ω (30.48m)2 2 × (30.48m)4 2π(30.48m) 0.0037m 1.52 30.48m

Doubling the burial depth decreases the ground resistance by only 1.7 ohms, or about 11 percent.


5 Underground System Surge Protection

Protecting underground distribution from lightning surges originating on overhead lines is crucial. Analyzing the effect of transient voltages and currents is one of the most complex subjects in distribution engineering. Accurate solutions to overvoltage problems often require higher mathematics and sophisticated computer simulations. The methods presented in this manual are approximations and should be viewed as such. Most of the recommendations and protective measures reviewed are based on complicated analyses. In most instances, only the results are given. For more information on why certain recommendations were made, consult the references listed at the end of this manual. OVERVIEW In general, the equipment that must be protected from lightning surges on an underground system is the same as on an overhead system: • Transformers, • Switchgear, and • Cable insulation. Installed cable is the most significant cost of an underground system. Unfortunately, underground cable insulation is not self-restoring like overhead insulation because it is not surrounded by air. An underground cable fault is much more expensive to locate and repair than a fault on an overhead line, which, most of the time, will be cleared by a recloser operation. Limiting the number and severity of surge voltages will prolong cable life. Therefore, it is important to provide the proper surge protection. Underground feeders consist of radial taps from overhead distribution circuits. Normal practice is for a cable run to have pad-mounted transformers connected along its route. A typical installation could be a short run of a few hundred feet or a long rural feeder terminated in an open point, usually at a transformer. A lightning surge traveling down the cable from the overhead line will double in magnitude at the open point as it reflects back on itself. As the reflected surge propagates back toward the riser pole, the doubling effect is transferred throughout the length of the cable. Because some underground

installations are protected only at the riser pole, the traveling wave phenomenon points out the problem inherent in protecting underground equipment by locating arresters as close as possible to the protected equipment. Protection of underground systems served from overhead lines is complex. The critical case generally involves lightning striking the line within one span of the riser pole. For high-current-magnitude, fast-front surges, the pole top will flash over, diverting the surge current to the pole ground conductor, bypassing the arrester entirely. In theory, a riser pole arrester need operate only for low-magnitude, slow-wavefront lightning currents left on the overhead line after an insulator flashover (Parrish, 1982). But voltage doubling plus reflections on the cable requires a riser pole arrester with the best available characteristics. Normally, a single arrester at the riser can keep voltages below the BIL withstand of 12.47-kV cable and equipment. Dead-front arresters now available will reduce reflected surge voltages by up to 50 percent. They should be used at 12.47 kV to reduce insulation voltage stress. However, 25-kV systems require openpoint arresters, because equipment BIL does not double as the system voltage doubles, resulting in reduced protective margins. Historical outage data at 25 kV has shown additional dead-front arresters are justified at cable taps and additional transformer locations. (See Table 5.14 for recommended arrester ratings and locations.) Most cooperatives install open-point arresters on both 12.47/7.2 kV and 24.94/14.4 kV. Several sources of transient overvoltages of system origin must be considered when surge arresters are applied. Overvoltages caused by neutral displacement during line-to-ground faults and voltage regulation are addressed later in this section. Overvoltages caused by ferroresonance and possible solutions are presented in Section 6 of this manual. Capacitor switching and currentlimiting fuse operation are two other possible surge sources. These sources do not cause severe extra duty for arresters applied for underground lightning protection because of their infrequent occurrence. Shunt capacitors are not needed on most cooperatives’ underground feeders. Current-limiting fuses usually do not

2 0 8 – Se c t io n 5

5 cause harmful voltage levels because of circuit impedances and the nonlinear magnetizing impedance of pad-mounted transformers. In addition, both of these sources can be minimized by proper equipment selection. Any discussion of lightning protection requires some knowledge of lightning phenomena and its electrical characteristics. Volumes of literature have been written on the subject since electricity was harnessed for domestic use. A thorough review of the subject is beyond the scope of this manual. However, an NRECA publication entitled Lightning Protection Manual for Rural Electric Systems, NRECA Research Project 92-12, by D.E. Parrish, offers an excellent starting point and an extensive bibliography for further reading.

Line Connector Porcelain Insulator

External Gap

Gasket seal Compression Spring

Porcelain Housing

Valve Element Gap Assembly Gasket & Seal Ground Connector Externally Gapped Silicon Carbide Valve Arrester

Line Connector Gasket Seal Gap Assembly Valve Elements Porcelain Housing Compression Spring Ground Lead Disconnector Ground Connector Internally Gapped Silicon Carbide Valve Arrester

Gapless MOV Arrester

FIGURE 5.35: Types of Arresters and Their Construction.

SURGE ARRESTER SELECTION Surge Arrester Types The first device connected between line and ground to protect power circuits from lightninginduced overvoltages was the simple air gap. To prevent breaker operation every time the gap flashed over, the device had to be able to interrupt an arc at a current zero. This need led to the development of the first expulsion arrester. Continued research to develop a better arrester that would protect large power transformers led to the gapped silicon carbide (SiC) valve arrester. Silicon carbide provided so many advantages over previous designs that some pundits said no further improvements in the device were needed. Researchers who continued testing SiC arresters found that, when hit with steep-front waves, gapped designs exhibited an undesirable characteristic. For waves with very fast rise times, a gap requires a considerably higher voltage to break down, which adds a sharp spike to the protective characteristic. Solving this problem required elimination of the gap, which was possible with the discovery of metal oxide as a valve material. The result was the introduction of the gapless metal oxide varistor (MOV) surge arrester in the late 1970s. It was one of the most significant advances in the history of overvoltage protection, as proven by the wide acceptance of metal oxide technology by electric utilities since the mid-1980s. Manufacturers generally no longer make SiC arresters. Metal oxide and SiC distribution arresters have some similarities in construction (see Figure 5.35). The only real difference besides valve element composition is the gap assembly in the SiC unit. Figure 5.35 also shows an externally gapped SiC arrester. The external gap will increase the rated sparkover voltage of an arrester, but helps to reduce outages caused by arrester failures. While this was an important factor with silicon carbide arresters, the overvoltage protection benefits of not using external gaps for riser pole applications far outweighs any perceived outage rate reduction. MOV designs are more efficient and offer better protective margins at the same voltage ratings. For these reasons, only MOV arresters will be considered in later discussions of underground system surge protection.


5 TABLE 5.8: Comparison of Protective Characteristics of Heavy-Duty Distribution Class Silicon Carbide, MOV, and Riser Pole MOV Arresters.* Arrester Rating

Maximum Discharge Voltage for 8 x 20 µs Discharge Current Wave (kV Peak)

kV rms

HD SiC (20 kA)

HD MOV (20 kA)

RP MOV (20 kA)

9

40

34

27

10

45

37

29

18

81

68

53

21

94

79

62

*Characteristics shown are for Cooper Power Systems arrester types EL, AZL, and AZR400. Note. HD = Heavy Duty RP = Riser Pole

Riser pole arresters represent a small percentage of the total arrester population on a system and should not contribute significantly to extended outage times. Because most underground risers are fused ahead of the lightning arrester, failure of the ground lead disconnector to operate properly will affect only the underground feeder. The overhead circuit will not suffer an outage in a properly coordinated system. The benefits of not using an external gap far outweigh any perceived reduction in quality of service that might occur.

Voltage

Gapless MOV Arrester In an MOV arrester, the valve elements are made Table 5.8 lists the maximum discharge voltage from zinc oxide (ZnO). The valve elements or for three MOV heavy-duty distribution class disks are about 90 percent ZnO and are comarresters for a 20-kA, 8 × 20 µs bined with a variety of other discharge current wave. The materials to determine the first two arresters are standard electrical characteristics of the MOV arresters have heavy-duty distribution class varistor. The ingredients are better protective SiC and MOV arresters. The first mixed and then pressed third is a special MOV arrester into disks at extremely high characteristics than with better characteristics depressure. They are then fired do SiC arresters. veloped especially for riser in a kiln into a ceramic resistor pole applications. with a very nonlinear volt-amBecause increased protective pere characteristic. margins can extend cable life and reduce equipThe MOV valve elements are very nonlinear. ment failures, external gaps should not be used The 60-Hz leakage current is in the low milli-amon MOV arresters for cable circuit protection. pere range at normal line-to-ground voltage, which eliminates the need for a series gap to insulate the arrester from ground. The sharp knee of the voltDischarge Voltage ampere curve means that the disks go into conduction at a precise voltage level and stop conducting when the voltage drops below that level. 25ºC MOV A series gap is, therefore, not needed to interSilicon rupt power follow current after a current surge Carbide Normal Line-topasses through the arrester. The MOV arrester Neutral Voltage 75ºC eases out of conduction after the surge voltage Power 125ºC Surge passes, without allowing hundreds of amperes Follow Current Current of power follow current to flow, as does an SiC arrester. The MOV also eases into conduction without producing a sharp voltage spike at the start of a lightning surge. This valve element property represents a significant advantage over Leakage Current SiC technology in equipment overvoltage protec< 0.1 Amp 100–500 Amps 1–100 kA tion. The nonlinear characteristics of SiC and Current MOV valve elements are compared in Figure FIGURE 5.36: Comparison of Nonlinear Characteristics of SiC and MOV 5.36, which shows the extreme nonlinearity of Valve Elements. Source: Kershaw, Gaibrois, and Stump, 1989. the MOV (Kershaw, Gaibrois, and Stump, 1989).

2 1 0 – Se c t io n 5

5 1.8 1.6 Silicon Carbide Arrester

1.4

Voltage (Per Unit)

1.2 1.0

Metal Oxide Varistor Arrester

0.8 0.6 0.4 0.2 0

0.1

1.0

10

100

Time (Microseconds)

FIGURE 5.37: Effect of Fast Rise Times on IR Discharge.

Shunt-Gap Module Insulated Terminal Cap

MOV Discs Gap(s) Series-Gap

Spacer

Steel Coil Spring

Coil Spring Desiccant

Isolator Series-Gap Design

Isolator Shunt-Gap Design

FIGURE 5.38: Series- and Shunt-Gapped MOV Distribution Arresters.

A second MOV arrester quality that makes it better suited for cable protection is its protective characteristic when subjected to current surges with fast rise times. A standard 8 × 20 µs current wave is used to represent lightning in arrester testing. It is considered by many experts much too slow to accurately model a lightning surge. Fast wavefronts of one to three microseconds are not uncommon for lightning strokes. The response of an arrester to steep-front waves should be considered in the arrester selection process for cable protection. Figure 5.37 illustrates the effects of fast rise time surges for both MOV and SiC arresters as a multiplier of the arrester discharge voltage (Niebuhr, 1982). Inspection of the curves shows that the increase in arrester discharge voltage under steep-front waves is more severe for SiC than for MOV arresters. Moreover, the sparkover characteristics of a gapped silicon carbide arrester will increase for steep-front surges. Insulation protection will be reduced accordingly. A gapless MOV arrester will not exhibit this behavior.

Internally Gapped MOV Arresters As discussed in the previous subsection, temporary overvoltages (TOV) are a primary concern when gapless MOV arresters are applied because they cannot tolerate voltages above the MOV valve-on voltage for long periods. Two manufacturers have taken different approaches to solve this problem. One uses a series resistance-graded gap structure and a reduced stack of MOV valve elements. The design provides increased TOV capability and improved protective characteristics over gapless designs. The other uses an increased number of disks for more overvoltage capability. More valve elements reduce the leakage current during expected temporary overvoltage conditions, thus preventing thermal runaway. Spark gaps are then used to short the extra disks during a surge event and give increased protective levels. Comparing the two gapped MOV arrester designs with a heavyduty gapless design shows that a 20 percent reduction in discharge voltage can be obtained. Figure 5.38 shows a cutaway view of the two gapped arresters for riser pole applications and the location of the gaps and disks.


5 According to the manufacturers, the two different gapped arresters provide the following improvements over a gapless design:

Arrester engineers have known about these problem Consider gapped areas for years from experiMOV arresters ence with SiC arresters and early gapped MOV station where temporary class arresters. The presence overvoltages occur. • Lower discharge voltage of gaps should not prevent cocharacteristics, operatives from considering • Higher temporary overvolttheir use to take advantage of age capability, and increased protective margins and better tempo• Increased thermal capacity. rary overvoltage capability offered by these designs for specific applications. Although these designs have been in the field for some time, application engineers still have a Dead-Front Lightning Arresters few concerns about adding gaps to MOV arresters: Dead-front arresters were developed to solve the major problem inherent in the protection of UD • Gap sparkover variability resulting circuits: locating arresters as closely as possible from erosion, to the protected equipment. Various arrester de• Contamination affecting sparkover level, signs and accessories provide convenient, eco• Changes in arrester characteristics with nomical, and reliable means to connect them to time, and UD systems. Previously, gapped SiC arresters • Stability of internal seals and gaskets in the were used in live-front applications in padpresence of ozone caused by gap operation mounted transformer enclosures with limited inside the housing. success. In addition to safety considerations required for live-front operation and the added expense of larger cabinets, the gapped arrester relies on a spark-gap operation to protect the 6 5 equipment. Gap operation is sensitive to ground 7 4 planes in enclosures; unless this effect is addressed in the overall design of the protection scheme, the arrester sparkover level will be affected. The advent of gapless MOV arresters 3 eliminated the concern with gap operation and 2 made the development of dead-front arresters feasible for underground equipment protection. 1 MOV arresters for underground use are called dead-front arresters because a semi-conducting grounded shield is molded around the insulation and valve elements. Dead-front arresters have 8 been in service for a number of years and their 9 10 configurations have been standardized for interchangeability. Typical installations are in padFeatures: mounted transformer enclosures, entry cabinets, 1. Metal oxide valve elements 6. Locking ring 2. Semiconducting moulded shield 7. Operating eye vaults, and switching enclosures. 3. Rubber insulation 8. Grounding eye A cutaway view of a dead-front design mated 4. Probe 9. Stainless steel end cap to a load-break elbow connector is shown in 5. Insert interface 10. Ground lead Figure 5.39. The electric utility industry uses three basic FIGURE 5.39: Dead-Front Arrester Elbow Configuration. dead-front arrester configurations:

2 1 2 – Se c t io n 5

5 1. Elbow Arrester. This is an and reducing clutter within the MOV arrester used with a transformer. Dead-front arresters load-break elbow connecare applied close to tor. It is used mainly to Dead-front MOV arresters connect directly to a transwith elbow connectors, bushing protected equipment. former bushing. In other arresters, and parking stand applications, it is used with arresters are available from a feed-through device or a manufacturers in all three voltfeed-through insert. age classes: 15, 25, and 35 kV. Figure 5.40 shows 2. Parking Stand Arrester. This is an arrester the three types of dead-front arrester designs. IEEE/ANSI C62.11, the IEEE Standard for combined with an insulated parking bushing Metal Oxide Surge Arresters for AC Power Cirfor mounting on a transformer or switching cuits, now covers the operating characteristics of enclosure parking stand. Typical use is at dead-front arresters. Dead-front arresters are the open-point transformer of a loop-feed classified as light-duty arresters capable of passdistribution circuit to park the disconnected ing the following tests: elbow connector. This arrangement helps reduce transformer faceplate overcrowding. • 40-kA high-current withstand test; 3. Bushing Arrester. The bushing arrester con• 75-ampere, 2,000-µs low-current, figuration combines an MOV arrester with a long-duration test; and load-break bushing insert for mounting direct• 5-kA duty cycle test. ly to the pad-mounted transformer bushing well. It is used at the end of a radial circuit The continuous ambient temperature requireor at the open point of a loop. Surge protecments are -40°C to 65°C, whereas the temporary tion is provided while increasing operability maximum arrester temperature is 85°C. For comparison, ambient temperatures set by the standard for overhead arresters are -40°C to 40°C continuous and 65°C maximum. Thus, under normal service conditions, underground arresters may operate at temperatures 25°C higher than overhead arresters. The higher operating temperature requirement is intended to address MOV arrester stability at high application temperatures inside pad-mounted transformer enclosures. Because metal oxide dead-front arresters are considered light-duty devices, they do not have Dead-Front Elbow Arrester the same protective characteristics as heavy-duty and specially designed arresters for riser pole protection. Expected current magnitudes on underground circuits are not as severe as those on overhead circuits because the riser pole arrester operation reduces the surge magnitude on the protected underground cables. Therefore, matching characteristics are not required, and underground arresters are designed with discharge voltages approximately 20 to 40 percent higher than arresters used for riser pole applications at Bushing Arrester Parking Stand Arrester the same surge current magnitudes. Their discharge voltages are listed only for current surges FIGURE 5.40: Dead-Front Surge Arresters.


5 TABLE 5.9: Typical Electrical Ratings and Characteristics of Dead-Front Surge Arresters. Duty Cycle Voltage Rating (kV)

MCOV (kV)

Equivalent Front-of-Wave (kV)*

1.5 kA

3

2.55

13.4

10.6

11.5

12.1

13.2

15.1

6

5.10

26.5

21.2

23.0

24.6

27.0

32.5

10

8.40

39.8

31.7

34.5

36.3

41.0

47.3

12

10.20

46.1

36.5

39.7

42.3

46.9

55.3

15

12.70

58.0

46.2

50.3

53.5

59.5

70.6

18

15.30

69.8

56.0

61.0

65.2

72.6

86.5

21

17.00

83.0

67.0

74.0

77.0

88.5

105.1

24

19.50

96.5

78.0

85.0

90.0

103.4

123.3

27

22.00

99.4

78.5

87.0

93.5

104.5

130.3

Discharge Voltage (kV) ** 3 kA 5 kA 10 kA

20 kA

* Equivalent front-of-wave voltage is the expected discharge voltage of the arrester when tested with a 5-kA current surge peaking in 0.5 µs. ** Maximum discharge voltage for an 8/20 µs surge current.

surge current carried by the up to 20 kA versus 100 kA for riser pole arrester (Osterhout, some heavy-duty designs. Coordinate riser pole 1989). This condition can norTable 5.9 lists typical ratings and dead-front mally be achieved by keeping and characteristics for deadthe riser pole ground leads as front arresters. The voltage ratarresters for proper short as possible and using arings and maximum continuous current sharing. resters with the same voltage operating voltage (MCOV) are rating. If the riser pole leads standard values, whereas the are substantially longer than protective characteristics repthe normal three-foot elbow arrester lead, the resent industry values compiled from various dead-front arrester can be overloaded, especially manufacturers. for fast-front incoming surges. When dead-front arresters and riser pole arresters are applied on UD feeders, they are considered to be in parallel when subjected to surge Arrester Performance Classes currents. Proper coordination is required to enThere are four basic classifications of lightning sure the larger riser pole arrester takes the bulk arresters: of the surge current so the discharge capacity of the lighter duty underground arrester is not ex1. Station, ceeded. Current sharing between the two de2. Intermediate, vices depends on the discharge voltage of each 3. Distribution, and arrester when subjected to the same surge cur4. Secondary. rent and the total impedance between them. The total impedance is the cable surge impedThese classifications differ in voltage rating, ance plus the ground leads of both arresters. For thermal capacity, protective characteristics, stanproper current sharing, the dead-front arrester dard tests, and whether pressure relief is required. should see between 10 and 20 percent of the For the most part, the major difference between

2 1 4 – Se c t io n 5

5 TABLE 5.10: Comparison of Standard Requirements for Surge Arrester Classifications. Arrester Class Characteristic or Feature Approximate Protective Characteristics (at 10 kA)*

Distribution (1-30 kV) 3.5 pu

Intermediate (3-120 kV)

Station (3-684 kV)

3.0 pu

2.7 pu

Current Discharge Requirements High Current, Short Duration

65 kA (ND) 100 kA (HD)

65 kA

65 kA

Duty Cycle

22-5 kA (ND) 20-10 kA & 2-40 kA (HD)

5 kA

10 kA (>550 kV) 50 kA (550 kV) 20 kA (800 kV)

Low Current, Long Duration

20-75 A (ND) 20-250 A (HD)

Transmission Line Discharge Test Required

High Current

Not Required

16.1 kA

40–65 kA

Low Current

Not Required

400–600 A

400–600 A

Pressure Relief

* In pu of arrester rating Note. ND = Normal duty; HD = Heavy duty

each arrester class is the physical size of the disk or block. A larger diameter block reduces IR discharge voltage and greatly increases energy absorption capability and reliability. The proper choice of arrester class depends on the system voltage, protected equipment insulation level, and the size and cost of the equipment. The first three arrester classes can be used on a distribution system because their voltage ratings overlap. Distribution class arresters are usually used on feeders, whereas intermediate and station class arresters are used in substations. Because the various classes tend to overlap, the easiest way to distinguish among them is to know the different standard tests performed for each class. Table 5.10 lists the tests required by ANSI/IEEE Standards C62.1 and C62.11. For the protection of underground circuits at riser poles, MOV heavy-duty distribution class arresters are normally used. The greater current discharge requirements of the HD designation inherently mean these arresters will have a lower

discharge characteristic than a normal-duty distribution class arrester. Various manufacturers have developed what is called within the industry a “riser pole” class of arrester. This class is not recognized by standards. To obtain better protective characteristics, the manufacturers have essentially taken intermediate class blocks and packaged them in different housings. In short, taking blocks with better characteristics and placing them in distribution class housings of porcelain or polymer construction results in better protective characteristics at a reduced cost. The distribution class housing surrenders the pressure-relief capability of intermediate class units, but their electrical characteristics are kept. The units are also much easier to mount on distribution crossarm structures. These arresters should be strongly considered for any underground application because they provide better protective margins. SURGE ARRESTER APPLICATION FACTORS Voltage Rating The voltage rating of an MOV arrester is based on its operating duty-cycle test. The duty-cycle test defines the maximum permissible voltage that can be applied to the arrester and still have it discharge its rated current. For an MOV arrester to be applied, the voltage rating must be above the maximum expected line-to-ground voltage at which the arrester will have to discharge a current surge. Under most conditions, the maximum voltage will occur on the unfaulted phases of a three-phase circuit during a single-line-to-ground fault. The explanation above alludes to an important quality of metal oxide arresters. Because of the thermal properties of metal oxide, they can dissipate current surges at higher system voltages than would be seen under normal operating conditions. Figure 5.36 shows that the ZnO current characteristic will shift to the right with increasing temperature. The shift also produces increased leakage current. The material will remain stable as long as a surge event does not increase the temperature to a point where increasing leakage current causes thermal runaway (Kershaw, Gaibrois, and Stump, 1989). Increasing normal line-to-neutral voltage above the knee of the leakage-current characteristic


5 curve for a sustained period will also cause thermal runaway. For this reason, the selection of an MOV arrester is based on the MCOV applied

across the arrester during normal service. For arresters on An MOV arrester effectively grounded systems, rating is set by its the MCOV is based on the nominal system line-to-neutral MCOV. voltage, plus a continuous overvoltage operating factor. For well-regulated systems, this TABLE 5.11: Metal Oxide Surge Arrester factor is often considered to be five percent. HowRatings in (kV) Root Mean Square. Source: ever, the engineer selecting MOV arresters should ANSI/IEEE C62.11-1987. review the system operating characteristics to deDuty-Cycle Voltage MCOV termine the applicable factor. The MCOV requirement would be 7.56 kV for a 12.47-kV nominal 3 2.55 system voltage (1.05 × 12.47 kV ÷ 3) under 6 5.10 most conditions. A more thorough discussion of MOV arrester voltage rating selection is given in 9 7.65 the next subsection, Line-to-Ground Faults. 10 8.40 The standard voltage rating and MCOV for all distribution class arresters are shown in Table 5.11. 12 10.20 Another consideration in the application of 15 12.70 MOV arresters is temporary overvoltages, such 18 15.30 as the single-line-to-ground fault mentioned previously. A metal oxide arrester will operate suc21 17.00 cessfully as long as it is not required to dissipate 24 19.50 more than its design energy level. Operating time above arrester voltage rating is set by the 27 22.00 amount of energy the arrester must dissipate 30 24.40 during the event. If the overvoltage on the arrester is reduced to its MCOV before it gets too hot, thermal runaway will not occur and the arrester will not fail. Temporary overvoltage capability curves are published by each manufacturer for its products. Typical overvoltage curves are shown in Figure 5.41. Similar curves should be considered when No Pr an arrester is subjected to system overvoltage ior Du ty conditions. Prior D Two system conditions that can affect the uty voltage rating selected for a riser pole MOV arrester application are overvoltages caused by:

1.50

60-Hz Voltage—Per Unit Arrester Rating

1.40 1.30 1.20 1.10 1.00

60ºC Ambient

0.90 0.80 0.1

• Line-to-ground faults, and • Voltage regulation.

1.0

10 Permissible Duration (Seconds)

100

FIGURE 5.41: Temporary 60-Hz Overvoltage Capability Curves— Typical MOV Distribution Arrester.

1,000

Line-to-Ground Faults Selecting the proper MOV arrester voltage rating is based on experience and on calculated overvoltage values for the unfaulted phases of threephase circuits during a single-line-to-ground

2 1 6 – Se c t io n 5

5 Equation 5.16 Arrester Rating ≥ (Line-to-neutral voltage) × (Range A factor 1.05 × 1.2)

doubt, some cooperatives are now using a factor of 1.35 instead of 1.25, which would lead to the selection of the next higher arrester duty-cycle voltage rating at 12.47 and 25 kV (10 kV and 21 kV rather than 9 kV and 18 kV).

Voltage Regulation RUS Bulletin 1724D-112, The Application of Capacitors on Rural Electric Systems, calls for maximum service voltages to be no greater than five percent above nominal for Range A voltage limits. The Pick the next higher bulletin does not limit voltage • Minimum service voltage = fluctuations on feeders. Capac114 Volts, and arrester rating itor operation at light loads, • Maximum service voltage = when unsure about lightly loaded underground 126 Volts. primary cables, or voltage regovervoltage duration. ulator malfunction can lead to The equivalent primary voltsystem voltages up to 10 perage values are as follows: cent above nominal without the cooperative’s immediate knowledge. A 10 • Minimum service voltage = 6,840 Volts, and percent voltage increase on a 12.47-kV feeder • Maximum service voltage = 7,560 Volts. with a 9-kV riser pole arrester would lead to about a four percent MCOV overvoltage on the The 1.2 in Equation 5.16 represents a safety arrester (120 × 1.1 × 60 ÷ 7.65 kV = 1.035 pu of factor of 20 percent. Equation 5.16 represents MCOV rating). The problem is that these feeder the maximum voltage rise on the unfaulted overvoltages may be present for hours, not secphases of a loaded circuit. The voltage rating is onds. The long-term stability of metal oxide then equivalent to 1.25 times the nominal linevalve elements has been proven at the MCOV, to-neutral system voltage. The calculation for a not at these possible higher voltages. Prior duty 12.47-kV system is 7,200 × 1.25 = 9 kV. on the arrester will also increase thermal aging. This application rule is very conservative for Range A voltage levels must not be exceeded SiC arresters. The maximum calculated voltage is on feeders if 9- and 18-kV MOV arresters are inequal to the arrester rating, and arrester sparkstalled on the system. If higher voltages are known over far exceeds the rating. For an SiC arrester to occur for sustained periods, the next higher to operate, a transient voltage surge will have to arrester duty-cycle voltage rating should be used occur while the 60 Hz overvoltage due to neu(10 kV and 21 kV). tral shift exists. The same rule is not as conservative for metal Protective Margin oxide arresters unless it is known that the sysThe level of protection provided by an arrester is tem is truly effectively grounded. For riser pole called the protective margin. It can be defined as applications with BCN cable, effective groundthe percentage that insulation strength exceeds ing can be assumed with a reasonable degree of the maximum surge voltage allowed by the accuracy. With the installation of jacketed cable, surge arrester and its leads. Insulation strength is effective grounding might not be achieved. An commonly referred to as the BIL, which is based accurate estimate of the voltage on the unfaulted on the industry standard lightning voltage wavephases can be calculated by factoring in ground shape of 1.2 × 50 µs. The standard lightning curwire size, ground rod spacing, fault resistance, rent waveshape is an “8 × 20” wave, meaning earth resistivity, and system impedance. The calcuthat its rise time to peak is 8 µs with a time-tolated overvoltage is then compared with the MOV half value of 20 µs. Figure 5.42 shows the current arrester temporary overvoltage curves. When in fault. The most common application rule for open-wire, four-wire wye, effectively grounded systems is shown in Equation 5.16. The Range A factor is 1.05 when voltage limits are as follows:


Current (kA)

5 I

0µs

I/2

10µs

20µs

100µs

Time

FIGURE 5.42: Typical Test Current Waveshape: Sinusoidal Wavefront.

Equation 5.17 PM(%) =

BIL –1 × 100 IR + LV

where: BIL = Equipment BIL, in kV IR = Arrester discharge voltage, in kV LV = Lead voltage, in kV

EXAMPLE 5.14: Protective Margin Calculation for Riser Pole Application: Industry Standard 4 kA/µs Average Rise Time for Lightning Strokes Assumed. Assume a 12.47-kV riser pole installation is protected by a 10-kV MOV arrester connected with a two-foot lead. Arrester IR discharge voltage for a 10-kA, 8 × 20 current surge is 26.5 kV. A lead voltage of 1.6 kV/ft will be assumed. Equipment BIL is 95 kV.

PM(%) =

95 –1 × 100 26.5 + 2(16)

PM(%) = [3.20 –1] × 100 PM = 220%

waveform. This test wave is used to establish comparative IR discharge voltage data shown in catalogs of arrester characteristics. For example, a heavy-duty MOV riser pole arrester with a 9-kV rating might have a maximum discharge voltage of 24.5 kV when impulsed with a 10-kA, 8 × 20 µs current wave. This current wave produces approximately a 10-kA/8 µs = 1.25 kA/µs average

rise time, which is much slower than a typical stroke discharged by an arrester. As noted earlier in this section, the change in current magnitude with time is sometimes expressed as di/dt. To calculate protective margin, add the L di/dt inductive voltage drop in the arrester leads carrying surge current to the arrester discharge voltage. The inductance of solid wire used for connections is a constant and is typically assumed to be 0.4 microhenries per foot (µH/ft). Research data have shown that the average rise time for a typical lightning stroke, which is defined as the rate of current increase per microsecond, is closer to 4 kA/µs than the 1.25 kA/µs value mentioned above. Using a di/dt of 4.0 kA/µs gives a voltage of 4.0 × (0.4) = 1.6 kV/ft of lead, which must be added to the discharge voltage to calculate the total protective margin. For more information on this subject, see the IEEE guide on arrester lead length calculations. The protective margin for a cable installation may be calculated using the following basic formula in Equation 5.17. Comparing the above margin with the 20 percent recommended industry standard shows the installation is more than adequately protected. After other elements are considered, it will be shown that the above level of protection is optimistic in most cases.

Rate of Rise The standard 8 × 20 µs current waveshape used for testing was intended to represent a typical lightning stroke. Recent field tests and recorded data show much greater variation than that represented by the standard. Surge current rise times vary with each lightning discharge. According to recent data compiled on electrical parameters of lightning return strokes, typical rise times can vary anywhere from 0.1 to 30 µs with current magnitudes greater than 110 kA. Probability data from recorded lightning strokes are shown in Figure 5.43 and are more representative than the 8 × 20 µs wave. Researchers have reported rise times higher than 10 kA/µs for more than 50 percent of stroke currents. Maximum rise rates greater than 75 kA/µs have been recorded. Standards recognize that some fast-front surge conditions produce current waves peaking in less than eight microseconds. For surges that

2 1 8 – Se c t io n 5

5 0

Percentage Probability

Percentage Probability That Time to Peak Will Equal or Be Less Than the Time Shown on the X Axis

17%

20

40 60

57%

80

100

1

2

3

4

5

6

Time to Peak (Microseconds)

FIGURE 5.43: Lightning Rise Time to Peak.

7

a voltage wave across the arrester peaking in 0.5 µs. The resulting peak voltage is the value listed in tables of arrester characteristics. The question of what rate of rise to use with the equivalent FOW characteristic is still an open debate among protection engineers. The consensus has narrowed the rate down to between 10 and 20 kA/µs, depending on experience and the frequency of lightning in the area. Multiplying by 0.4 µH/ft gives a lead wire voltage drop of between four and eight kilovolts per foot. Because FOW characteristics for MOV arresters are higher than standard discharge 8 9 voltages, and high rate-of-rise surges produce greater voltages per foot of lead, fast-front surge conditions put maximum stress on cable insulation. When protective margins are calculated, some application engineers recommend that FOW protective levels for MOV For fast-front waves, arresters be used along use the FOW with lead length voltage of six kilovolts per foot. characteristic to

peak in two microseconds or less, the insulation strength is the Chopped Wave Withstand (CWW). For oil-filled equipment, including pad-mounted transformers, the CWW is 15 calculate protective Lead Length percent higher than the BIL. margin. Discharge voltages from surge The insulation strength of arresters travel through a cable cable, unlike oil-filled equipat half the speed of light. ment, does not increase as the When a traveling wave rise time of the applied surge reaches a point of high impedance such as an voltage decreases. For the purpose of insulation open point on a loop, it reflects on itself. This coordination, the CWW of a cable is considered reflection doubles the voltage at the open point equal to its BIL for all surge voltage waveshapes. and along the cable as the incoming and reFor the comparison of various arrester characflected waves overlap. Voltage doubling on the teristics under these conditions, an Equivalent cable system must be considered when protecFront-of-Wave (FOW) protective level was derived tive margins of arrester installations are calcuspecifically for gapless MOV arresters. It denotes lated. To minimize the surge voltage entering a the fact that a surge current with a faster rise time cable, install an arrester with low discharge charto peak than the standard 8 × 20 µs test wave will acteristics at the riser pole, connected with the produce a higher discharge voltage in an MOV shortest leads possible. arrester. This increased voltage is more severe for Tables 5.12 and 5.13 compare protective marSiC arresters. For example, referring to Figure gins on 24.9- and 12.47-kV systems using three 5.37, assume a time-to-peak of one microsecond. different arrester types. The BIL margin percentThe curves show that the IR characteristic of an ages are calculated using the industry standard SiC arrester increases approximately 30 percent, 10-kA, 8 × 20 µs waveshape, which is assumed while the MOV increase is about 10 percent. to produce 1.6-kV/ft inductive voltage drop in The equivalent FOW characteristic for an the series arrester leads. The CWW insulation MOV arrester is the arrester discharge voltage withstand is based on a 10-kA surge current for current pulses having a time-to-peak of that produces a discharge voltage that peaks in about 0.6 µs. This current waveshape produces


5 0.5 µs. This type of wave produces the discharge voltages with kV peaks shown in the Arrester Data FOW columns. A rise time of 15 kA/µs will be assumed to produce a six-kilovolt per foot voltage drop in the leads to represent severe fast-front lightning strokes. The IR discharge and lead length voltages are added and then multiplied by two to represent the voltage doubling effect caused by reflections. Inspecting the tables shows that, for the standard 8 × 20 µs waveshape and three-foot leads, the 12.47-kV system has 20 percent or better margins. Only the special riser pole MOV can provide this level of protection on the 24.9-kV system. For fast-front surges, the protective margins drop drastically when lead length effects are included. In the 24.9-kV system, eliminating the arrester lead length entirely results in a 20 percent margin for the riser pole MOV. The other arresters provide no protective margin when lead effects are considered. To provide the greatest protective margins for underground cables, keep the arrester discharge path (lead length) as short as possible in all installations. Arrester lead length is the combined line and ground lead length in series with the arrester and in parallel with the cable’s termination. Figure 5.44 shows an installation that corresponds to the three-foot-lead examples in Tables 5.12 and 5.13.

Lead = 18” Lightning Arrester

Lead = 18” JCN Cable

FIGURE 5.44: Arrester Lead Length Equal to Three Feet.

TABLE 5.12: Protective Margin, 24.9-kV Underground Distribution System: 125-kV BIL Insulation, 18-kV Arresters at Riser Pole Only, 10-kA Lightning Discharge, Surge Voltage Doubled by Reflection. Arrester Data

Protective Margin (%)*

10 kA IR (kV Peak)

Zero Lead Length

Arrester Type

8 × 20

FOW**

8 × 20

FOW**

8 × 20

FOW**

8 × 20

FOW**

Heavy-Duty SiC

69

80

-9

-22

-12

-30

-15

-36

Heavy-Duty MOV

60

66

4

5

0

-17

-4

-26

Riser Pole MOV

48

52

30

20

24

2

18

-11

*Protective Margin (%) =

1.5-Foot Lead

BIL 2 × (LPL + LV)

LV = lead voltage = feet × 6 kV/ft for FOW LV = lead voltage = feet × 1.6 kV/ft for 8 × 20

3-Foot Lead

–1 × 100

LPL = Lightning Protective Level LPL = FOW or 8 × 20 for 10-kA IR (kV Peak)

**Based on 10-kA current impulse that results in a discharge voltage peaking in 0.5 µs

2 2 0 – Se c t io n 5

5 TABLE 5.13: Protective Margin, 12.47-kV Underground Distribution System: 95-kV BIL Insulation, 9-kV Arresters at Riser Pole Only, 10-kA Lightning Discharge, Surge Voltage Doubled by Reflection. Arrester Data

Protective Margin (%)*

10 kA IR (kV Peak)

Zero Lead Length

Arrester Type

8 × 20

FOW**

8 × 20

FOW**

8 × 20

FOW**

8 × 20

FOW**

Heavy-Duty SiC

35.0

42

36

10

27

-9

20

-22

Heavy-Duty MOV

30.0

33

58

44

47

13

36

-7

Riser Pole MOV

24.5

27

94

76

77

32

62

6

*Protective Margin (%) =

1.5-Foot Lead

BIL 2 × (LPL + LV)

LV = lead voltage = feet × 6 kV/ft for FOW LV = lead voltage = feet × 1.6 kV/ft for 8 × 20

3-Foot Lead

–1 × 100

LPL = Lightning Protective Level LPL = FOW or 8 × 20 for 10-kA IR (kV Peak)

**Based on 10-kA current impulse that results in a discharge voltage peaking in 0.5 µs

Lead = 0”

Cable Termination

Lead = 18”

FIGURE 5.45: Arrester Lead Length Equal to 1.5 Feet.

Remember, keep lead length short to increase protective margin.

Figure 5.45 shows a similar installation except the line connection is taken to the arrester and then to the termination. All arrester line lead is eliminated because the wire carrying surge current through the arrester is not in parallel with the termination. Figure 5.46 shows how arrester lead length can be virtually eliminated by modifying the installation of Figure 5.45. The arrester is mounted between the termination and the pole ground conductor, so the pole ground conductor can be carried directly to the base of the arrester. The connection makes the arrester ground lead length zero with respect to the concentric neutral of the jacketed cable. Because no surge current flows in either line or ground leads, the surge voltage across the termination is limited to the discharge voltage of the arrester and represents the “zero lead length” examples in the tables. The easiest way to remember how to make the best connections can be summarized as fol-


5 Connect to the arrester first, then to the cable, to minimize lead length. Lead = 0”

Cable Termination

Lead = 0”

FIGURE 5.46: Zero Arrester Lead Length.

Equation 5.18 V=

1 ≈ 3 × 1010 cm/second = 984 ft/microsecond LC0

where: L = Inductance, in Henries per unit length C0 = Capacitance, in Farads per unit length in free space

Equation 5.19 V=

1 3 × 1010 cm/second 984 ≈ = ft/microsecond LC k k

where: k = Insulation dielectric constant (typical values from 2 to 4)

lows. Carry the line and ground connections to the arrester terminals first, and then to the conductor and ground terminals of the cable termination. This procedure will ensure the leads are kept as short as physically possible to make full use of the protective margin provided by riser pole arresters (Hubbell/Ohio Brass Co., Hi-Tension News, 1989). The engineer must recognize the extreme importance of arrester lead arrangement on effective overvoltage protection for underground systems. Furthermore, the engineer must communicate this importance to installation crews, who have final control over this item. Improper arrester lead arrangement can cancel the advantages of even the most advanced arresters and lead to premature failures on underground systems. TRAVELING WAVES ON UNDERGROUND DISTRIBUTION SYSTEMS A lightning stroke to an overhead line will cause a transient condition to occur. This rapid voltage buildup caused by the discharge of energy from a charged cloud is not transferred instantaneously to all points on the overhead line or connected cable. In fact, the surge requires a finite time to propagate down the line. The surge movement is in the form of a traveling wave. The traveling wave characteristics are determined by the distributed nature of the capacitance and inductance of the line. Its propagation speed is also set by line characteristics. For overhead lines, the velocity of wave propagation, V, is calculated as shown in Equation 5.18. The calculated value of 984 feet/µs is approximately the speed of light. For overhead lines in open air, the line conductor acts only as a guide for the electromagnetic disturbance and the velocity of propagation is near the speed of light.

2 2 2 – Se c t io n 5

5 For underground cables, the Figure 5.47 is the classic electromagnetic wave does not representation of a transmisSurges travel at travel through air, but through sion or distribution line with approximately half the cable insulation. The vedistributed L and C paramelocity in this instance depends ters. Also depicted in the figthe speed of light on the L and C of the cable, ure is the current I, which in cables. which are determined by its represents the charging curinsulation material and physirent produced by the voltage cal dimensions. Equation 5.19 surge as it travels along the shows the calculation for the velocity of the line. The current waveshape is the same as the wave in a cable with a dielectric constant of “k.” voltage, and these parameters are related by the The capacitance of the cable is increased in prosurge impedance of the line: portion to the dielectric constant of the insulation. Therefore, the velocity in cable is where k V L is the dielectric constant of the cable insulation. I= ZSURGE = ZSURGE C Therefore, the velocity in cable is 1 V= LkC0 where k is the dielectric constant of the cable insulation.

Two types of cable insulation—TR-XLPE and EPR—are normally used by cooperatives for underground applications. The approximate dielectric constants for these two insulation materials are typically 2.3 and 3.0, respectively. If 984 ft/µs is assumed to be the speed of light, surge propagation speeds within the two cables are as follows:

TR-XLPE: V = 649 ft/µs EPR: V = 568 ft/µs

Voltage Surge V L C

L C

I

C I

V=

L

1 LC

L C

C

I

ZSURGE =

L C

FIGURE 5.47: Representation of Distributed Parameter Distribution Line.

For overhead lines: ZSURGE = Surge impedance = 500 ohms (400- to 600-ohm range) For underground cables: ZSURGE = 35 ohms (20- to 60-ohm range) The line charging current should not be confused with lightning surge current, which will not flow until a discharge path to ground is formed. Once a traveling wave is initiated, it will continue along a line until its energy is dissipated or until a change in surge impedance occurs. Changes in surge impedance that are important for cable protection occur at overhead/underground connections (riser poles), open cable end points, and midpoint cable taps. The magnitude of the traveling voltage and current waves is changed at the junction points. Waves divide in proportion to the equivalent surge impedance at the junction according to Kirchhoff’s laws. This division gives rise to reflected and refracted (continuing) portions of the incident wave. Equations 5.20 and 5.21 may be used to calculate the traveling wave voltages and currents where a line terminates on an equivalent surge impedance. Figure 5.48 shows the effect of a traveling voltage wave meeting a change in surge impedance at a junction. After the incident wave encounters the discontinuity, three components of the wave exist:


5 Junction Point (JP)

Incident V1

Refracted V2

Reflected V3 Z1

I1 = V1 /Z1 I2 = V2 /Z2 I3 = –V3 / Z1

Z2

V2 = V1 + V3 I2 =I1 + I3

V1 /Z1 – V3 /Z1 = V2 /Z2 V1 /Z1 – V3 /Z1 = (V1 + V3) /Z2 Equation 5.20 The Reflection Coefficient (K) V3 /V1 = (Z2 – Z1)/(Z2 + Z1) = K

Equation 5.21 The Refraction Coefficient is Then: V2 /V1 = 2Z2 /(Z1 + Z2) = 1 + K

Where: V1, I1 V2, I2 V3, I3 Z2

Z1

= = = =

Incident Voltage and Current Approaching Junction Point Refracted Voltage and Current Continuing Beyond the Junction Point Reflected Voltage and Current from the Junction Point Equivalent Surge Impedance Beyond the Junction Point (Z2 = Parallel Impedance at all Lines Connected to the Right of the Junction Point) = Equivalent Surge Impedance to Incident and Reflected Waves

FIGURE 5.48: Change in Surge Impedance at a Junction Point—Effect on Traveling Voltage Wave. 1. Incident wave (V1), 2. Refracted wave (V2 ), and 3. Reflected wave (V3). Traveling voltage waves are illustrated in Figure 5.49. The (a) view of Figure 5.49 shows a

traveling wave (in rectangular form for simplicity) approaching a junction point (JP) on impedance path Z1. Various surge impedance conditions beyond point JP are shown in the other four views (Kershaw, 1970).

2 2 4 – Se c t io n 5

5 (a)Traveling Wave V1 Incident on Junction Point (JP) on Impedance Path Z1

(e) Z2 = 0.1Z1 Represents an Overhead Line Dead-Ending at a Riser Pole

V1

JP V1

JP

Z1

Z2 Represents equivalent surge impedance beyond junction point

Z2

V1 V2 V3

(b)

Z1 = Z2, All of the Voltage Is Refracted

V3 Z1

V1 = V2

V1

V2 Z1

V2 Z2

Shows Progression of Waveforms: Incident (V1), Refracted (V2), Reflected (V3)

Z2

Equation 5.22 (c)

Z2 = 0 Represents a Ground or Short Circuit

V2 = V1

Voltages Cancel at a Short Circuit

V1

2Z2 2(0.1) = V1 Z1 + Z2 1.1

V2 = 0.182 V1 V3 Z1

Z2

Equation 5.23 Z2 = ∞ Represents an Open Circuit

(d)

V3 = V1

I = 0 at All Times at an Open point

Z2 – Z1 0.1Z1 – Z1 = V1 Z2 + Z1 0.1Z1 + Z1

V3 = –0.818 V1

V3 V1

Z1

Z2

In this case a voltage wave of 18% of the incident value continues on the cable, while 82% of the wave is reflected back toward the source, cancelling a like portion of the incident wave.

May also be assumed to model the response at an end-of-line transformer. Transformer HV windings represent a small capacitance at transient frequencies. Voltage doubling still occurs; however, the reflected wave front would have a different shape.

FIGURE 5.49: Traveling Wave Behavior at Junction Points Terminated with Various Surge Impedances.

Cable Open-End Point Terminated by Nonlinear Resistance (Gapless MOV Arrester) Until now in the discussion of wave behavior at a junction point, only differences in surge impedance magnitude have been considered. An MOV

arrester response at an open point presents an interesting case because it is a nonlinear resistance. The incident voltage wave starts to double as previously described for an open point until the MOV valve elements start to conduct


5 V1

Incoming Wave Before Valve-On

V1 At Valve-On

VT Valve-On Voltage

V2

VT After Valve-On

V1 Valve-On Voltage

V2 VT V1 Peak Voltage = VT = V2 + V1 Peak V2

Valve-On Voltage

FIGURE 5.50: Traveling Waves at a Cable Open-End Point Terminated by an MOV Arrester.

(Figure 5.50). At this point, the excess voltage is short-circuited to ground through the arrester. It is also assumed the IR discharge voltage equals the valve-on voltage and remains constant throughout the surge event. The reflected wave, V3, is positive and adds to the incoming wave, V, until the arrester starts to conduct. Voltage at the junction point is then canceled as the negative portion of the wave is reflected upon the incoming wave. However, before wave cancellation starts, it is preceded by the positive reflected portion of the wave, which adds to the incoming wave. The positive reflection adds about one-half the valve-on voltage to the incoming wave. The peak voltage, VT, then travels back

toward the source superimposed on the incoming wave, V1 (Cooper Power Systems, 1990). The preceding analysis shows that an MOV arrester at a cable open-end point will prevent voltage doubling and transfer of the overvoltage to the sending end of the cable by reflections. Voltage doubling does not occur, but the reflected voltage is increased by one-half the valve-on voltage of the arrester. The percentage of voltage increase over the limited riser pole let-through voltage depends on the IR characteristics of the two arresters. PROTECTION METHODS ARRESTER LOCATIONS The decision of where to place arresters on underground systems for equipment protection is based on how the cable is configured and how its conductor is terminated. The previous subsection on traveling waves showed that a change in surge impedance, whether caused by tapping a cable or an open point, would cause reflections. The equivalent surge impedance at the discontinuity sets the magnitude of the reflected wave. At an open point, the voltage doubles back toward the source, subjecting the entire cable length to the overvoltage. If the cable has one or more open-ended lateral taps, the reflected waves added together can produce more than twice the riser pole let-through voltage. Proper location of dead-front or elbow surge arresters will offer increased protective margins at all points within the underground system. Engineers should consider using them even on 15-kV systems to reduce overvoltage magnitudes and prolong cable life. The engineer cannot calculate protective levels at each piece of underground equipment without transient analysis software, but general rules will produce adequate protection for most commonly encountered situations. To provide an idea of the effectiveness of various protection schemes without doing sophisticated traveling wave analyses, this subsection will evaluate several schemes that utilities use. The effectiveness of the schemes will be determined by comparing their protective levels at various locations to the level provided by a single riser pole arrester. The following seven overvoltage protection schemes will be considered:

2 2 6 – Se c t io n 5

5 1. Riser pole arrester; 2. Riser pole arrester and cable-end arrester; 3. Riser pole, cable-end, and an arrester applied at the first transformer on the source side of the open point (third arrester); 4. Riser pole arrester protecting a cable with a lateral tap; 5. Lateral tapped cable with riser pole and open-end arresters; 6. Lateral tapped cable with riser pole, openend arresters, and an arrester at the tap point; and 7. Riser pole arrester and under-oil arresters at every transformer. The seven protection schemes are shown in Figure 5.51. Riser Pole Arrester Only (Figure 5.51, No. 1) For 15-kV class systems and below, a single arrester at the riser pole will generally provide adequate protective margins for cable-connected equipment. As system voltages increase to 25 kV, equipment insulation levels, unfortunately, do not double as well. For 25-kV systems with 125-kV BIL using 18- or 21-kV arresters, the protective margins for a riser-pole-only arrangement can be nonexistent. In this case, arresters must be added to the open-end points. The maximum voltage stress that the entire cable and connected equipment can be roughly calculated using Equation 5.24, which shows the doubling effect. Equation 5.24 shows that, if care is taken to reduce arrester lead length, the major contributing factor to cable voltage stress is the doubling of the riser pole discharge voltage. Using a SiC

heavy-duty arrester on a 7.2/12.5-kV system could lead to cable voltages extremely close to new 15-kV equipment strength (95-kV BIL). Using a specially designed riser pole MOV arrester instead of an SiC design should reduce maximum cable surge voltage (VC ) by 40 to 60 percent. A reduction of this amount is important when aged insulation is considered or when fast-front surge currents enter the system. Riser Pole and Cable-End Arrester (Figure 5.51, No. 2) Placing an arrester at the open point terminates the cable with a low impedance when the arrester conducts. The low arrester impedance generates a negative reflected wave that works to reduce the voltage at the open point and along the entire cable length. The maximum system surge voltage is given by Equation 5.25. Figure 5.50 shows that maximum voltages will always occur away from the cable-end arrester. The dead-front arrester eliminates cable-end voltage doubling and limits the open-point voltage to its protective level. The reflected voltage (VT) appears as a triangular spike that is superimposed on the incident voltage wave and travels back toward the riser pole. The peak of the spike is approximately equal to 50 percent of the dead-front arrester discharge voltage at a current level of 1.5 kA. As the spike returns to the riser pole, it subjects most of the cable run to surge voltages that exceed the protective levels of the arresters at either end of the cable. However, the overvoltages are less than the doubling of the riser pole arrester

Equation 5.25 Equation 5.24

VC = VRP + VL + VC = 2(VRP + VL)

where: VC = Maximum cable and equipment surge voltage, in kV VRP = Riser pole arrester discharge voltage, in kV VL = Lead voltage drop, in kV

1 VOP 2

where: VC = Maximum cable and equipment surge voltage, in kV VRP = Riser pole arrester discharge voltage, in kV VL = Lead voltage drop, in kV VOP = Open point arrester discharge voltage, in kV


5 1. Single-Phase UD Feeder

4. Single-Phase Feeder with Lateral Tap

Open Point Jacketed Neutral

Tap Point

Cable Neutral

Riser Pole Arrester Only

Open Point

Lateral Tapped Cable with Riser Pole Arrester 5. Single-Phase Feeder with Lateral Tap

2. Single-Phase UD Feeder

Open Point

Open Point

Lateral Tapped Cable, Riser Pole, and Open-End Arresters

Riser Pole Plus Cable-End Arrester

6. Single-Phase Feeder with Lateral Tap

3. Single-Phase UD Feeder

Open Point

Riser Pole, Cable-End, and Third Arrester

Open Point

Lateral Tapped Cable, Riser Pole, Open-End Arresters, and Tap-Point Arrester

7. Single-Phase UD Feeder Under-Oil Arrester

Open Point

Riser Pole Arrester and Under-Oil Arresters at Every Pad-Mounted Transformer

FIGURE 5.51: Arrester Locations.

2 2 8 – Se c t io n 5

5 let-through voltage. Simulations and laboratory tests have shown that maximum cable surge voltage will be reduced by 25 percent with the addition of cable-end arresters. In most instances, 25-kV system protective margins obtained with this arrester configuration are less than the recommended 20 percent level. If protective margins of 20 percent or more are desired, additional arresters must be added to protect cable and equipment remote from the two end points. Riser Pole, Cable-End, and Third Arrester (Figure 5.51, No. 3) Additional surge protection can be provided by adding a third arrester between the two ends of the cable. The function of the third arrester is to suppress the voltage spike reflected from the cable-end arrester. For limiting surge voltage exposure to a minimum, the most effective location for the arrester is the first transformer on the source side of the cable-end arrester. The separation distance between the two protective devices must be at least 200 to 300 feet for the third arrester to effectively suppress the reflected wave. If the first transformer upstream from the open point is fewer than 200 feet away, the third arrester can be applied at the next upstream transformer, leaving the first unit with reduced protection. The maximum system surge voltage with third arrester protection can be calculated using Equation 5.26. Equation 5.26 shows that, except for the cable section between the third and cable-end arresters, the maximum system surge voltage is limited to the protective level provided by the riser pole arrester (Lat, 1987).

Equation 5.26

Lateral Tapped Cable with Riser Pole Arrester (Figure 5.51, No. 4) Cooperatives sometimes tap a radial cable system to provide service to nearby loads. Tapping the cable produces parallel cable runs where surge voltages can propagate independently. A tapped configuration will produce higher cable voltages than will a simple radial system because multiple traveling waves can add and subtract in complex ways. Assume a voltage surge enters the tapped system at the riser pole with no other arresters applied. When the surge reaches the tap point, it sees an equivalent surge impedance of approximately 15 to 20 ohms. The impedance is the parallel combination of the surge impedance of each cable leg. Because of the discontinuity, portions of the incident wave will be reflected back toward the riser pole and simultaneously refracted onto the two cable legs. When the two refracted voltage waves reach the respective cable ends, they will double and travel back to the tap point, where they will again be reflected and refracted. Because of the unequal travel times on the cable sections, the multiple reflections and refractions will ultimately lead to an increase in cable-end voltage. The voltage increase can be up to 30 percent more than the voltage doubling normally experienced on a radial cable run protected only by a riser pole arrester. Lateral Tapped Cable, Riser Pole, and Open-End Arresters (Figure 5.51, No. 5) Installing MOV arresters at both open points will reduce their voltages to the protective level of the arresters. Arresters at the end points will also keep the tap-point voltage within reasonable magnitudes. In one laboratory test, the tap-point voltage was 13 percent higher than the maximum midspan voltage on a radial system under the same conditions (riser pole and cable-end arrester).

VC = VRP + VL where: VC = Maximum cable and equipment surge voltage, in kV VRP = Riser pole arrester discharge voltage, in kV VL = Lead voltage drop, in kV

Lateral Tapped Cable, Riser Pole, Open-End Arresters, and Tap-Point Arrester (Figure 5.51, No. 6) Placing an arrester at the tap point will tend to further reduce voltages along the tapped feeder. This is due to the shunting effect of the arrester to limit the initial surge as well as the reflected


5 waves. A more definite statement cannot be made unless a specific example is analyzed in detail. An arrester at the tap point should not be considered as adequate protection for the two open cable ends. Tests have shown that positive wave reflections can act to more than double the cable-end voltage, when compared with a tapped system protected by a riser pole arrester only. The following general conclusions can be drawn from investigations into the effects of cable taps on surge voltage magnitudes:

or under-oil designs, which are very expensive to add to existing installations. However, some utilities are considering under-oil arresters for every new or replacement transformer installation as a way to prolong cable life. Under-oil arresters are good; however, one must consider the cost to replace an under-oil arrester when it fails, and some will fail. Replacing under-oil arresters can be very expensive. It does, however, remain to be seen whether this overall scheme will prove to be a cost-effective approach.

• A primary tap will increase surge voltage magnitudes above levels that will exist without a tap. • The surge voltage magnitude increase will typically be 10 to 30 percent. • Taps located close to the riser pole tend to produce greater surge voltage magnitudes. • Multiple taps do not appear to produce surge magnitudes significantly greater than a single tap (Ros, 1988).

RECOMMENDED ARRESTER LOCATIONS AND RATINGS The information presented above has shown that many factors affect arrester protective margins. It is not possible to consider all factors in an application because they can change for many different reasons. Experience has shown that the recommendations in Table 5.14 should be used for radial feeders and tapped laterals for conservative underground protection. The recommended arrester locations given in Table 5.14 are based on the application of riser pole MOV arresters with 10-kV and 21-kV dutycycle voltage ratings. The MCOV for these arresters is 1.17 pu and 1.18 pu of nominal line-toneutral system voltage. The recommended voltage ratings are one step above the 9-kV and 18-kV ratings that can be used on effectively grounded neutral circuits that have close voltage regulation (Range A voltage levels). Distribution systems are susceptible to longterm overvoltages caused by the following:

Riser Pole Arrester and Under-Oil Arresters at Every Pad-Mounted Transformer (Figure 5.51, No. 7) The ultimate surge protection scheme is to provide arresters at every convenient and accessible point on the underground system. Besides the riser pole, possible locations could be tap points, sectionalizing points, and pad-mounted transformers. The arresters have to be dead-front

TABLE 5.14: Recommended Arrester Locations. Voltage

Feeder Configuration

Arrester Locations

12.47 kV

Radial

Riser Pole Open Point

25 kV

Radial

Riser Pole Open Point Third Arrester Near Open Point*

12.47 kV

Tapped Lateral

Riser Pole Open Points

25 kV

Tapped Lateral

Riser Pole Open Points Tap Point*

*Optional application

• • • • • • • •

Line-to-ground faults, Poor voltage regulation, Line voltage regulator malfunctions, Ferroresonance, Fixed shunt capacitors, including long cables, Circuit backfeed, Load rejection, and Other system contingencies.

The lower-rated arresters provide additional protective margin, especially at 25 kV. However, the higher-rated arresters are recommended to prevent premature arrester failures as the installation ages. The following examples will show

2 3 0 – Se c t io n 5

5 H1A

H1B

Elbow Arrester

PRACTICAL DEAD-FRONT ARRESTER INSTALLATIONS The introduction of dead-front MOV arrester designs and their accessories for use inside UD enclosures has introduced flexibility into underground system protection. Product evolution now allows choices in the selection of optimum protection schemes based on cost and protection levels. In the previous discussion of protection methods and arrester locations, four locations were suggested for the installation of dead-front arresters:

Feed-Through

Elbow Arrester

To Riser Pole

To Riser Pole

1. On the cable end at the open-point transformer between two sections of a loop-feed circuit, 2. At the first upstream transformer from the open point, 3. At a tap point, and 4. At the cable end of a lateral tap (radial-feed circuit).

(a) Two Elbow Arrester and a Feed-Through

H1A

H1B

Elbow Connector Parking Stand Arrester Elbow Arrester

To Riser Pole

To Riser Pole (b) Elbow Arrester and Parking Stand Arrester

H1A

H1B

Bushing Arrester Insulating Cap Parking Stand Arrester To Riser Pole

To Riser Pole (c) Bushing Arrester and Parking Stand Arrester

FIGURE 5.52: Cable-End Arresters at Open Point.

that using the recommended arrester locations and voltage ratings will result in higher protective margins than those suggested by standards.

The following will describe practical ways to physically connect arresters at all four locations using load-break elbow-type connectors and elbow, bushing, and parking stand arresters described previously in the subsection, Dead-Front Lightning Arresters. To get flexibility with different arresters and components, purchase only transformers with bushing wells. Cable-End Arrester at Open Point Three configurations can be used for this installation. The configuration chosen will depend on operating practices and available space inside the pad-mounted transformer cabinet. Figure 5.52(a) shows an open-point transformer with an arrester attached to each cable end. The arrangement uses two elbow-type arresters and a feed-through mounted on the parking stand. This installation takes up the most room on the transformer faceplate. Figure 5.52(b) shows a different approach that can be taken at the open-point transformer. It uses an elbow arrester and a parking stand arrester to reduce overcrowding by eliminating the feed-through device. The third configuration, Figure 5.52(c), allows increased operational flexibility and reduces overcrowding by using bushing and parking


5 H1A

stand arresters. Operational flexibility is obtained because the open point can be closed by moving the parked cable to H1B without removing the parking stand arrester. Once a cable fault is repaired, the elbow connector is easily placed back on the parking stand arrester to reestablish the open point. The bushing arrester on H1A requires less space than an elbow arrester/feedthrough bushing insert combination mounted in the same location.

H1B

Feed-Through Bushing Insert

Elbow Arrester To Riser Pole

To Open Point (a) Elbow Arrester on Feed-Through Insert

H1A

H1B

Bushing Arrester

Elbow Connector

To Riser Pole

To Open Point (b) Bushing Arrester Only

FIGURE 5.53: Arrester Upstream from Open Point (Third Arrester).

H1B

X3 X1

H1A

X2

Primary Source

FIGURE 5.54: Two Elbow Arresters and a Feed-Through.

Alternate Source

Arrester Upstream from Open Point Two arrester configurations may be used to provide additional protective margins at 25-kV and above by clipping the voltage spike generated by operation of the open-point arrester. An elbow arrester or a bushing arrester may be applied. Figure 5.53(a) is a schematic of how an elbow arrester combined with a feed-through bushing insert can be mounted on the transformer faceplate. To reduce clutter inside the enclosure, mate a bushing arrester to the source-side cable as shown in Figure 5.53(b). Figures 5.54 through 5.58 show the five installation configurations discussed above. Lateral Tap Cable-End Arrester For lateral taps off all underground feeders, arresters should be placed at open points to prevent reflections from increasing surge voltages above levels that would exist without the tap(s). Figure 5.59 shows the desired ways arresters can be applied to two- and single-bushing transformers at the end of radial-feed circuits. To add surge protection to a two-bushing loop-feed unit, an elbow arrester or a bushing arrester must be connected to the unoccupied terminal. For the radial-feed transformer, the least-cost application is to add a bushing arrester. Tap Point Arrester For tapped lateral feeder configurations 25 kV and above, an arrester should be added at the tap point as well as on the open points. Many connection methods can be used, but the one shown in Figure 5.60 offers a simple, low-cost approach to establish a tap point, have loadbreak switching capability, plus add an arrester. The cable-to-cable connections can be made by

2 3 2 – Se c t io n 5

5 H1B

using a four-point load-break junction bolted to the inside surface of a suitable pad-mounted enclosure. The three cables are connected together using load-break elbow connectors attached to three terminals of the junction. An elbow surge arrester is installed on the fourth terminal to complete the installation.

X3 X1

H1A

X2

Primary Source

Alternate Source

FIGURE 5.55: Elbow Arrester and Parking Stand Arrester.

H1B

X3 X1

H1A

X2

Primary Source

Alternate Source

UNDERGROUND SURGE PROTECTION EXAMPLES The five examples in this subsection are based on a typical underground loop feed to a subdivision with an open point between the two laterals. Only one underground radial with four pad-mounted transformers will be investigated. The system is protected by a riser pole MOV arrester with discharge characteristics that are readily available within the industry. Dead-front arresters are used at strategic locations for increased protective margins. Arrester lead lengths are assumed to be one foot at the riser and three feet at the pad-mounted transformer locations. Various surge current waveshapes and magnitudes are used to evaluate the effectiveness of the different recommended protective schemes. The many calculations were made by a traveling-wave computer program (Cooper Power Systems’ UDSURGE™). A simplified schematic of the system is shown in Figure 5.61.

FIGURE 5.56: Bushing Arrester and Parking Stand Arrester.

H1B

H1B

X3 X1

H1A

Primary Source

X3 X1

H1A

X2

To Open Point

FIGURE 5.57: Elbow Arrester on Feed-Through Insert on Transformer Upstream from Open Point.

Primary Source

X2

To Open Point

FIGURE 5.58: Bushing Arrester on Transformer Upstream from Open Point.


5 Two-Bushing Pad-Mounted Transformer

H1A

Single-Bushing Pad-Mounted Transformer

H1B

H1A

Bushing Arrester

Elbow Arrester

To Tap Point

To Tap Point

FIGURE 5.59: Lateral Tap Cable-End Arrester (Radial Feed Circuit).

Pad-Mounted Enclosure Surge Voltage Magnitudes Calculated at Riser Pole and the 4 Pad-Mounted Transformer Locations Conduit

Four-Point Load Break Junction

Pad #1 1,000'

Pad #2 400'

Pad #3 400'

Pad #4 400'

Open Point

Elbow Arrester To Riser Pole

To Open Point

Tap Line

FIGURE 5.60: Tap-Point Arrester.

Riser Pole Arrester Protecting Jacketed Cable Underground Lateral: 12.47 kV and 25 kV Arrester Ratings: 10 kV and 21 kV Total Arrester Lead Length = 1 Foot

Dead-Front Arrester Locations Ratings: 10- and 21-kV 3-Foot Leads

FIGURE 5.61: Typical Underground Subdivision Loop Feed with Open Point.

Tables 5.15 through 5.19 summarize the surge voltages calculated by the computer program at the riser pole and the four transformer locations. Different surge current characteristics are used to illustrate how variable lightning characteristics can affect equipment protective margins. Varying the current rate-of-rise and magnitude

and calculating the voltage at all nodes enables most of the variables that go into the protective margin (lightning variability, arrester characteristics, lead length, BIL deterioration, reflections, and so forth) to be considered. In this way, for this particular system, recommended arrester locations can be evaluated on their merits.

2 3 4 – Se c t io n 5

5 EXAMPLE 5.15: MOV Riser Pole Arrester: Arrester Rating, 10 kV. Table 5.15 considers a 12.47-kV system protected by a 10-kV MOV riser pole arrester. ANSI Standards suggest a 20 percent margin for an 8 × 20 µs surge at 10 kA. Most protection engineers realize this suggestion does not consider many of the variables mentioned in the previous paragraph and is used mostly on overhead systems. To provide added security in underground applications, many engineers double the current magnitude to 20 kA. If a 20 percent or greater margin is obtained, the system is considered adequately protected. Under these conditions, Table 5.15 shows protective margins of 46 and 31 percent, respectively, for aged insulation. Further examination reveals that, for the two fast-front, high-magnitude current surges depicted in the last two rows of the table, the margin is reduced below 20 percent and actually becomes negative for the worst case. Protective margin calculations at any of the other locations on the radial feeder are simple to make. The listed surge voltage magnitudes include lead voltage drop. Protective margin is then simply as shown in Equation 5.27.

TABLE 5.15: MOV Riser Pole Arrester: Arrester Rating, 10 kV; Equipment BIL, 95 kV; Aged BIL, 76 kV. Surge Current Characteristics Riser Pole

Surge Voltage Magnitudes (kV) Padmount Padmount Padmount No. 1 No. 2 No. 3

Padmount No. 4

8 × 20 µs 10 kA

26.9

50.1

51.4

51.9

52.1 (46% margin)

8 × 20 µs 20 kA

29.7

56.0

57.4

57.9

58.2 (31% margin)

1 × 50 µs 20 kA

36.4

64.1

64.2

64.4

71.9 (6% margin)

1 × 50 µs 50 kA

54.8

87.8

88.0

88.4

108.1 (-30% margin)

Note. Percent margins in parentheses are for aged insulation BIL.

Equation 5.27 PM(%) =

BIL –1 × 100 Surge Magnitude

EXAMPLE 5.16: MOV Riser Pole Arrester and Dead-Front Cable-End Arrester (No. 4): Arrester Rating, 10 kV. The system in Table 5.16 is protected with a riser pole arrester and a cable-end arrester. Placing the arrester at the end of the cable prevents voltage doubling and keeps a minimum 39 percent margin for aged insulation throughout the entire length of the cable in the worst case. Cableend arresters are strongly recommended at 12.47 kV to protect equipment insulation from fastfront, high-magnitude lightning surges.

TABLE 5.16: MOV Riser Pole Arrester and Dead-Front Cable-End Arrester (No. 4): Arrester Rating, 10 kV; Equipment BIL, 95 kV; Aged BIL, 76 kV. Surge Current Characteristics

Riser Pole


Padmount No. 4

8 × 20 µs 10 kA

26.7

38.2 (99% margin)

37.1

34.9

28.3

8 × 20 µs 20 kA

29.6

41.6 (83% margin)

40.2

36.8

29.4

1 × 50 µs 20 kA

36.4

43.5

43.5

43.7 (74% margin)

36.9

1 × 50 µs 50 kA

54.8 (39% margin)

54.6

54.5

54.3

37.2 (104% margin)



5 EXAMPLE 5.17. MOV Riser Pole Arrester: Arrester Rating, 21 kV. The summary of surge current magnitudes for a 25-kV lateral protected by an MOV riser pole arrester rated 21 kV is shown in Table 5.17. The standard 8 × 20 µs waveform with 10- and 20kA magnitudes produces negative margins for aged insulation and less than 20 percent margin for new insulation. This example reiterates that a riser pole arrester cannot protect a 25-kV radial cable with an open-point termination.

TABLE 5.17: MOV Riser Pole Arrester: Arrester Rating, 21 kV; Equipment BIL, 125 kV; Aged BIL, 100 kV. Surge Current Characteristics

Riser Pole


Padmount No. 4

8 × 20 µs 10 kA

56.6

103.5

106.5

107.9

108.6 (-7% margin aged) (16% margin new)

8 × 20 µs 20 kA

62.3

115.6

119.0

120.2

120.5 (-17% margin aged) (4% margin new)


EXAMPLE 5.18: MOV Riser Pole Arrester and Dead-Front Cable-End Arrester (No. 4): Arrester Rating, 21 kV. The example in Table 5.18 represents a 25-kV system with arresters located at the riser pole and open point. The arresters limit the voltage to acceptable levels at both cable ends. Voltage magnitudes on the interior cable section cause inadequate margins for the 8 × 20 µs, 20-kA case and both fastfront, high-current cases. The higher voltages in the middle of the cable are caused by the addition of the dead-front arrester valve-on voltage to the reflected voltage traveling back toward the sending end of the cable. For further information, refer to the earlier explanation of traveling waves.

TABLE 5.18: MOV Riser Pole Arrester and Dead-Front Cable-End Arrester (No. 4): Arrester Rating, 21 kV; Equipment BIL, 125 kV; Aged BIL, 100 kV. Surge Current Characteristics

Riser Pole


Padmount No. 4

8 × 20 µs 10 kA

55.9

79.8 (25% margin)

77.5

73.0

57.0

8 × 20 µs 20 kA

61.8

85.1 (18% margin)

82.4

75.8

58.1

1 × 50 µs 20 kA

66.4

90.8 (10% margin)

90.9

91.2

72.0

1 × 50 µs 50 kA

89.5

98.6

98.8

99.2 (1% margin aged) (26% margin new)

70.4 (42% margin)


2 3 6 – Se c t io n 5

5 EXAMPLE 5.19: MOV Riser Pole Arrester Plus Dead-Front Cable-End Arrester (No. 4) and Dead-Front Third Arrester (No. 3): Arrester Rating, 21 kV. The example in Table 5.19 shows the addition of a dead-front arrester to the next transformer upstream from the open point. This arrester at transformer No. 3 cancels the valve-on voltage spike from the open-end arrester. The three arresters working together provide acceptable protective margins along the entire cable length.

TABLE 5.19: MOV Riser Pole Arrester Plus Dead-Front Cable-End Arrester (No. 4) and Dead-Front Third Arrester (No. 3): Arrester Rating, 21 kV; Equipment BIL, 125 kV; Aged BIL, 100 kV. Surge Current Characteristics

Surge Voltage Magnitudes (kV) Riser Pole

Padmount No. 1

Padmount No. 2

Padmount No. 3

Padmount No. 4

8 × 20 µs 10 kA

55.9

64.2 (55% margin)

61.5

57.0

57.0

8 × 20 µs 20 kA

61.2

68.0 (47% margin)

64.7

57.9

58.0

1 × 50 µs 20 kA

66.4

69.7

67.3

66.6

71.9 (39% margin)

1 × 50 µs 50 kA

89.5 (12% margin aged) (40% margin new)

89.3

89.0

68.7

70.3



Some protection engineers recommend riser pole and open-end arresters at 25 kV. A two-arrester protection scheme is adequate for most lightning conditions. However, field tests have shown fast-front high-magnitude surges can occur 20 percent of the time (see Figure 5.43.)

To protect against these lower probability events, the three-arrester scheme is recommended for underground installations. It is a conservative approach that balances increased arrester costs against increased MOV arrester and cable life.

1. The purpose of the grounding system is to maintain all points connected to it at earth potential under various conditions. 2. The grounding system consists of the grounding and neutral circuits. The grounding circuit is made up of ground electrodes, ground conductors, and all connections. The neutral circuit includes the JCN and all connections to it. 3. The return current path must be a continuous metallic circuit along the entire route of energized conductor(s). The earth should never be used as the only path for the return of normal load current. 4. Under fault conditions, the neutral circuit provides a low resistance path to ensure

fast operation of protective devices. It also prevents dangerous touch potentials on equipment cases and frames. 5. A ground rod has a 60-Hz measured resistance and a surge impedance (ZSURGE). ZSURGE is defined as the ratio of the peak voltage to current on the rod caused by a lightning discharge. 6. ZSURGE is always less than, or essentially equal to, the rod’s 60-Hz resistance value. 7. When a riser pole arrester conducts, lightning surge current flows on the following components: a. Arrester leads, b. Pole ground conductor,


5 c. Jacketed cable neutral, d. Counterpoise, and e, Overhead multigrounded system neutral.

8.

9.

10.

11.

The surge currents produce undesirable effects that, except for arrester lead length, are reduced by a low ground rod resistance when compared with the surge impedances of the various paths. Low ground rod resistance will reduce jacket voltage and the amount of surge current flowing on the JCN to the transformer and service neutrals. Continuous counterpoise connected to the JCN at the pole top and extended to the transformer ground rod will reduce jacket voltage up to 50 percent. Take special measures to adequately ground JCN cable installations when compared with semiconducting jacketed and BCN cable systems. Ground rods are the primary means to reduce ground resistance on JCN cable installations. The three factors that affect ground resistance are the following: a. Length, b. Number of rods, and c. Spacing.

12. Where possible, use longer rods, not multiple rods, to lower ground resistance. 13. When multiple rods are used because of rocky soil, space them at least two rod lengths apart. 14. The required number of driven rods for a JCN cable installation is set by the NESC. Table 5.2 summarizes ground rod rules and requirements. a. Power cable only: four rods per mile. b. Random lay: eight rods per mile. c. Counterpoise is considered a made electrode. 15. Ideally, the riser pole ground resistance should have the lowest value, followed by the transformer ground rod, and then the service ground.

16. Counterpoise will reduce jacket voltages. The counterpoise should be attached at the cable termination for best results. 17. Continuous counterpoise should be installed to the first transformer, if practical. 18. If full-length counterpoise is not justified, 100- to 300-foot lengths should be used. 19. An ideal ground has a low ground resistance value. To measure the ground resistance, one of the testers listed in Table 5.20 should be used.

TABLE 5.20: Ground Resistance Testers. Type of Grounding System

Clamp-On* 3-Point** Meter Meter

4-Point** Meter

Single Ground Rod

X

X

X

Multiple Ground Rod

X

X

X

Counterpoise

X

* Measurement must be made with the ground under test connected to a multigrounded system. ** Measurement must be made before connecting the ground under test to the system ground.

20. Soil resistivity directly affects ground resistance. Therefore, an engineer will need the soil resistivity value before designing a grounding system. If this information is not available, the soil resistivity should be measured using a four-point earth resistance tester. 21. Counterpoise and ground rods should be installed below the frost line. Doing so helps prevent an increase in ground resistance caused by frozen soil. 22. If possible, counterpoise and ground rods should be placed in an area with permanent moisture content. When the surrounding soil dries out, the ground resistance of the counterpoise or ground rod increases. 23. Ground resistance calculations should be used to compare different ground system configurations. If the ground resistance value is high, it should be decreased by:

2 3 8 – Se c t io n 5

5 a. Increasing the length of the ground rod or counterpoise, or b. Increasing the number of ground rods. 24. MOV arresters should be used for riser pole applications because they provide better protective margins than do similar SiC designs. 25. Series and shunt-gapped MOV riser pole arresters have better temporary overvoltage capability and slightly better protective characteristics than gapless models. They should be considered in areas where inadequate voltage regulation occurs. 26. Dead-front lightning arresters should be applied close to protected equipment on underground systems to increase protective margins. 27. Light-duty dead-front arresters should be coordinated with riser pole arresters so their discharge capability is not exceeded because of current sharing. Short riser pole leads and duplicate voltage ratings help ensure proper current sharing. 28. Selection of MOV arrester voltage rating is based on the MCOV the arrester sees during normal service. 29. Maximum voltage rise on the unfaulted phases of a loaded three-phase circuit and voltage regulation on distribution feeders above five percent can cause long-term overvoltages on MOV arresters. When unsure about overvoltage duration, choose the next higher MOV arrester rating (10 kV and 21 kV rather than the usual 9 kV and 18 kV). 30. Protective margin depends on protective characteristics of the arrester, lightning surge current magnitude, and equipment BIL. 31. Protective margin is calculated using Equation 5.17:

PM(%) =

BIL –1 × 100 IR + LV

32. Standards recommend using an average rise time (di/dt) of 4 kA/µs when calculating lead

33.

34.

35.

36.

37.

voltage. When this value is multiplied by 0.4 µH/ft, it gives 1.6 kV/ft lead voltage, which is the value to use with 8 × 20 µs arrester data. Recent studies have shown this value should be somewhere between four to eight kilovolts per foot when using arrester FOW characteristics to calculate protective margin. Arrester lead lengths must be kept as short as physically possible to obtain the maximum protective margin. For riser pole installations, this is accomplished by making connections to the arrester terminals first, and then to the conductor and ground terminals of the cable termination. It is important to remember how an incident traveling voltage wave reacts when it meets a change in surge impedance at a junction point such as an open point, midpoint cable tap, or MOV arrester. If the junction point is an open circuit (infinite surge impedance), the reflected voltage is positive and produces a voltage doubling effect. If this line is terminated in a short circuit, the reflected voltage is negative, which cancels the incoming wave. For a line terminated by an MOV arrester, voltage doubling does not occur, but the reflected voltage is increased by one-half the arrester valve-on voltage. Necessary and optional arrester locations that will minimize cable and transformer overvoltages should be used. After the decision is made where to place the arresters, elbow, bushing, and parking stand dead-front arresters should be physically connected at the following: a. Open-point transformer between two sections of a loop-feed circuit, b. First upstream transformer from the open point, c. Tap point, and d. Open end of a lateral tap.

38. MOV arrester voltage ratings of 10 kV and 21 kV may be used instead of 9-kV and 18-kV units, if problems are encountered with overvoltages.

Ferroreso n a n c e – 2 3 9

6 In This Section:

Ferroresonance

Allowable Overvoltages During Ferroresonance Distribution Transformer Connections Qualitative Description of Ferroresonance Ferroresonance When Switching at the Primary Terminals of Overhead and Underground Transformer Banks Ferroresonance with Cable-Fed Three-Phase Transformers with Delta or Ungrounded-Wye Connected Primary Windings Ferroresonance with Cable-Fed Three-Phase Transformers with Grounded-Wye Primary Winding and Five-Legged Core Ferroresonance with Cable-Fed Three-Phase Transformers with Grounded-Wye Primary Windings and Triplex Construction Ferroresonance in Underground Feeders Having More Than One Transformer Summary of Techniques for Preventing Ferroresonance in Underground Systems Summary and Recommendations References

were used instead of bare overhead conductors Before the use of primary voltages above 15 kV for primary circuits—operating at any voltage in overhead systems, and before the use of level—ferroresonance occurred during the medium-voltage power cables for primary distriswitching of the cable circuit and the distribubution circuits, engineers designing and operattion transformers connected to them. ing distribution systems were not concerned Ferroresonance in underground systems rewith ferroresonance. However, when 24.9-kV sults from single phasing in and 34.5-kV phase voltage levthree-phase primary circuits els were introduced for overwith distribution transformers, head distribution systems, Single phasing in which establishes configuraferroresonance occurred durthree-phase primary tions where the capacitances ing the switching of small circuits can cause of the primary circuit and the transformer banks at their prinonlinear inductances of the mary terminals. Correspondferroresonance. transformers are arranged so ingly, when shielded cables

2 4 0 – Se c t io n 6

6 that nonlinear resonance can newer low-loss transformers, occur. Single-phase conditions and has generated updated Modern low-loss occur during the normal operferroresonance avoidance transformers are much ation of fused or nonfused disguidelines. The findings of this connects, elbow connectors, research, including the new more susceptible to single-pole reclosers, and singuidelines, have been incorpoferroresonance. gle-pole sectionalizers when rated into this section. circuits and their connected Field experience has shown transformers are energized and that overvoltages occurring de-energized. Single phasing also occurs if a sysduring ferroresonance can cause failure of both tem component fails in a way that produces an metal oxide and gapped silicon carbide surge aropen conductor condition. Ferroresonance may resters, distribution transformers, cables, elbow cause very high overvoltages that damage connectors, splices, and equipment connected equipment and cause failures. to the secondary side of the distribution transThose responsible for the design and operaformer, including consumer appliances, computtion of rural distribution systems need to be faers, and electronic home entertainment equipmiliar with ferroresonance to prevent extremely ment. System designs and transformer connechigh ferroresonant overvoltages from occurring tions that are prone to ferroresonance should be during single-phase conditions in the primary avoided wherever possible. If the system design three-phase systems. or topology does not eliminate the chance of Experience and previous guidelines for avoidferroresonance under all possible switching coning ferroresonance are not always good indicaditions, operating personnel must be able to rectors of conditions in which ferroresonance may ognize when ferroresonance may occur during occur. Most guidelines predate the present widesingle-pole switching of cable circuits with conspread evaluation of losses by utilities in the nected transformer(s) and know how to sectiontransformer procurement process. No-load losses alize and switch the system so that ferroresohave a direct effect on the ferroresonance susnance will not occur. ceptibility of a transformer, and the substantial This section provides the system designer decrease in transformer losses in recent years with information needed to design a system makes the transformers of today much more in which ferroresonance is less likely. It identisusceptible to ferroresonance than those in use fies the distribution transformer connections that when previous ferroresonance guidelines were are highly susceptible to ferroresonance during developed. A major investigation of ferroresosingle-phase switching. Under some circumnance in modern grounded-wye pad-mounted stances, it may not be possible to design a systransformers was completed in 1992. This investem in which ferroresonance is prevented for tigation, sponsored by the Distribution Systems any switching procedure or sequence selected Testing, Application, and Research (DSTAR) by operating personnel. However, certain switchconsortium, of which NRECA is a member, has ing procedures and sequences will minimize the obtained results showing that some previous chance of ferroresonance during normal switchguidelines about ferroresonance are not valid for ing operations.

Allowable Overvoltages During Ferroresonance

Most rural primary distribution systems operating at nominal phase voltages up to and including 35 kV (line-to-ground voltages up to 20 kV) are multigrounded neutral systems. Each primary feeder in these systems, whether overhead or underground, and whether single-phase, veephase, or three-phase, has a neutral conductor

that is grounded at least four times per mile. As noted in Section 5, codes in some states require that the neutral conductor be grounded more frequently than four times per mile. When ground faults occur on the primary feeder of these systems, the voltage between any unfaulted phase and the neutral conductor


6 will rise above the nominal line-to-neutral voltage for the system. For overhead construction with conductors on an eight-foot or longer crossarm, or primary feeders with concentric neutral cables, the voltages from the unfaulted phases to the neutral conductor in a typical rural system will not exceed 1.25 times the nominal line-to-neutral voltage (1.25 per unit or pu). If the primary feeders employ spacer cable or armless construction, which is not common on most rural systems, the voltage from an unfaulted phase to the neutral conductor can exceed 1.25 pu, rising as high as 1.46 times nominal value (1.46 pu). The 1.25-pu voltage present during ground faults is the basis for selecting the duty cycle voltage rating of the surge arresters applied on most rural distribution systems. Arrester duty cycle ratings are at least 1.25 times the system

Distribution Transformer Connections

nominal line-to-neutral voltage. For example, on a 12.47/7.2-kV rural system, the duty cycle voltage rating of the surge arrester is either nine or 10 kV. The wide acceptance of this application guide for surge arrester voltage rating acknowledges that the equipment connected from phase to neutral on rural distribution systems is subjected to and can tolerate temporary line-toneutral overvoltages of 1.25 times nominal. This 1.25-per-unit overvoltage is also the upper limit on the temporary overvoltages that can be permitted during single-phase switching in rural distribution systems. The application tables and equations in this section for determining the allowable cable lengths during the switching of cable circuits and connected transformers are based on limiting temporary overvoltages to 1.25 pu.

The transformer connections shown in FigFerroresonance in distribution systems occurs ures 6.1(c), (d), (e), and (f) are used to supply during the single phasing of circuits, usually four-wire delta secondary systems, operating at underground cable circuits, and their connected a nominal voltage of 240/120 volts. The opendistribution transformers. Whether ferroresodelta/open-delta connections and the open-wye/ nance is possible—as well as the maximum open-delta connections usually are made from allowed length of a circuit with a connected two single-phase distribution transformers, altransformer that can be switched with singlethough some “three-phase” transformers have pole switches without exceeding 1.25-pu voltage been made with the open-wye/open-delta con—depend on the connections of the distribution nections. The delta/delta and the ungroundedtransformer primary windings. Certain winding wye/delta connections are found in some connections are highly susceptible to ferroresothree-phase transformers, and are used in connance, whereas other winding connections necting single-phase transformers into threeprevent ferroresonance under all practical phase banks. In the four-wire delta secondary conditions. systems, the three-phase, three-wire load is supFigure 6.1 shows the more common transplied phase-to-phase at 240 volts, and the sinformer connections found in rural distribution gle-phase, three-wire 120/240-volt lighting load systems. is connected across the secondary winding with The delta/grounded-wye connections and the the center tap and secondary grounded-wye/grounded-wye neutral conductor. connections in Figures 6.1(a) As discussed in detail later and (b) are used to supply With certain distribuin this section, transformers four-wire wye secondary systion transformer with ungrounded primary tems, operating at nominal winding connections, windings (delta, open-delta, voltages of either 208Y/120 and ungrounded-wye) are volts or 480Y/277 volts. These ferroresonance is highly susceptible to ferrowinding connections are used very likely during resonance during single-phase in three-phase transformers switching in underground sysand in banking three singlesingle phasing. tems. In contrast, three-phase phase transformers.

2 4 2 – Se c t io n 6

6 P

P

S

S

Neutral

P

S

Neutral Neutral

(a) Delta/Grounded-Wye

P

(b) Grounded-Wye/Grounded-Wye

S

P

S

Neutral

(d) Ungrounded-Wye/Delta

(c) Delta/Delta

P

S

Neutral

(e) Open-Delta/Open-Delta

Neutral

(f) Open-Wye/Open-Delta

P = Primary S = Secondary

FIGURE 6.1: Transformer Connections for Four-Wire Wye and Four-Wire Delta Services.

Qualitative Description of Ferroresonance

transformers or transformer banks with the grounded primary windings (grounded-wye or open-wye) are less susceptible to ferroresonance

and may even prevent ferroresonance from occurring, depending on construction of the three-phase transformer or transformer bank.

DEFINITION Ferroresonance is a complex electrical phenomenon in electrical circuits having at least one nonlinear inductor and at least one linear capacitor that is fed by one or more voltage sources having a sinusoidal waveshape. The nonlinear inductor is a saturable circuit element such as an iron core transformer. When ferroresonance occurs from a switching operation to energize or de-energize a circuit, an initial transient response may eventually settle into a sustained steady-state response. In general, the steady-state voltage and current waveforms are not sinusoidal like those of the source voltage. There can be more than one steady-state response mode in a specific circuit. The steady-state mode may depend on the initial or transient conditions in the circuit. Ferroresonance can be a chaotic phenomenon, meaning that a switching event repeated identically on the same circuit yields results that are substantially different. The circuit may never settle into a steady-state condition and may erratically jump from one mode to another indefinitely.

First, review the response of the series resistive-inductive-capacitive (RLC) circuit with linear parameters, and second, look at the effect of a nonlinear inductor in the circuit. With this background, it is then possible to consider the effects of this phenomenon on the distribution system. RESONANCE IN THE LINEAR INDUCTIVE-CAPACITIVE CIRCUIT Figure 6.2 shows a series RLC circuit, in which the resistor, inductor, and capacitor are linear. Linear means that the resistance, inductance, and capacitance of the elements do not change with time, current, or any other parameter. The source is a sine wave voltage with a peak magnitude of VM, having a frequency of ω radians per second. In the 60-Hz system, the radian frequency is 377 radians per second. The circuit is energized by closing switch S1 at time zero. Following switch closure, the current in the circuit and the voltage across each element consist of a steady-state response and almost always a transient response. The transient response decays with time to zero, leaving just the


6 S1

XL = ωL

R

VMsin(wt+θ)

XC = 1/ωC

XL = Inductive reactance

ω = Frequency of the system, in radians per second XC = Capacitive reactance

FIGURE 6.2: Series RLC Circuit with Sinusoidal Excitation.

steady-state response. The steady-state response continues as long as the circuit is connected to the source. When there is no trapped voltage on the capacitor and no current in the inductor, the point on the source voltage wave at which switch S1 is closed (closing angle θ) determines if there is a transient response and the initial magnitude of that response. Just two closing angles do not produce a transient response (occurring, of course, at the zero crossings). With linear parameters in the circuit, only one steady-state response is possible, and it is independent of the closing angle and initial conditions at switch closure, such as capacitor voltage and inductor current. In steady-state conditions, after the transient response subsides, the current and the voltages vary sinusoidally with time at the same frequency as the source voltage.

Equation 6.1: RLC Current Response. Irms = where: Irms = Vrms = R = L =

Vrms R2 + (ωL – 1/ωC)2

rms value of the current rms value of the source voltage Resistance of the resistor, in ohms Inductance of the inductor, in Henries C = Capacitance of the capacitor, in Farads ω = Frequency of the system, in radians per second

Equation 6.1 gives the rms value of the current in the circuit. In Equation 6.1, ωL is the inductive reactance, XL, and 1/ωC is the capacitive reactance, XC, with both reactances having units of ohms. When the inductive reactance ωL is equal to the capacitive reactance, 1/ωC, the denominator of the equation has a minimum value equal to R, the circuit resistance, and the current in the circuit has a maximum value equal to Vrms/R amperes. Also, at this point, the input impedance to the circuit of Figure 6.2 is purely resistive and the circuit is in resonance. If the inductance L and capacitance C are constant, the frequency at which resonance occurs is called the resonant frequency. The resonant frequency, designated as ω0 in radians per second, is given by Equation 6.2. During resonance, the voltage across the resistor is at its maximum possible value and equal to the source voltage, Vrms. The magnitudes of the voltage across the inductor and capacitor at

Equation 6.2: Resonant Frequency. ω0 =

1 radians/second LC

where: ω0 = Resonant frequency, in radians per second L = Inductance of the inductor, in Henries C = Capacitance of the capacitor, in Farads

Equation 6.3: Resonant Voltage. L C VL = Vrms R where: VL = Voltage across the inductor and capacitor at resonance L = Inductance of the inductor, in Henries C = Capacitance of the capacitor, in Farads R = Resistance of the resistor, in ohms Vrms = rms value of the source voltage

2 4 4 – Se c t io n 6

6 resonance are equal to each other, the value given by Equation 6.3. From Equation 6.3 it can be seen that the voltage across the inductor and capacitor at resonance in a series RLC circuit can be greater than the source voltage. If there were no resistance in the circuit, the capacitor and inductor voltages would be infinite at the resonant frequency, but the resistance prevents this. In addition, when there is resistance in the circuit, the voltage across the inductor has its maximum at a frequency that is somewhat above the resonant frequency, and the capacitor voltage has its maximum at a frequency that is slightly below the resonant frequency. In the linear circuit of Figure 6.2, the initial conditions and the closing angle have no effect on the steady-state response, including whether resonance does or does not occur. But if the inductor is nonlinear, because of the presence of an iron core, the initial conditions and closing angle do affect the probability of resonance occurring. The resulting response in circuits with iron core inductors is called ferroresonance. FERRORESONANCE IN THE NONLINEAR INDUCTIVE CAPACITIVE CIRCUIT When the inductor in the series RLC circuit of Figure 6.2 is nonlinear and the circuit is energized by closing switch S1, there is both a transient response and a steady-state response. The transient response decays to zero, leaving just the steady-state response, but the time for this to happen with the nonlinear circuit frequently is much greater than in a linear circuit. Generally, two steady-state responses are possible in a simple single-phase circuit with one nonlinearity, with the response determined by the closing angle and the initial conditions. However, the probability of the two possible steady-state responses is not the same or easily definable. Simple equations can be written and solved for both the transient and steady-state responses of the linear RLC circuit when energized from a sinusoidal voltage source. For example, see Equation 6.1 for the steady-state current. However, when the inductance is nonlinear, the equations describing the circuit do not have a simple solution. Most studies of ferroresonance in power systems have been performed with ei-

ther full-scale testing, transient network analyzer (TNA) studies, or digital transient programs. Graphical techniques give an approximate solution for the fundamental frequency component of the response of the ferroresonant circuit, giving some insight into the phenomenon (Rudenberg, 1970, Chapter 48). Graphical techniques also show that two steady-state solutions are possible for many ferroresonant circuits. The two steady-state responses in the singlephase ferroresonant circuit are called the normal mode and the ferroresonant mode (Feldman and Hopkin, 1974). The ferroresonant mode is characterized by substantial saturation of the nonlinear inductor, high capacitor voltages, and relatively high currents. When the steady-state ferroresonant mode occurs, the current and voltage waveforms in the circuit are periodic but not sinusoidal like the source voltage. Also, the peak values of the inductor and capacitor voltages can be higher than the peak value of the source voltage, just as in a linear circuit that is in resonance. Ferroresonant voltage waveshapes can be classified into three types of repetitive patterns or modes. Three steady-state ferroresonant modes are, thus, possible (Germany, Mastero, and Vroman, 1974): 1. Fundamental, 2. Subharmonic, and 3. Higher harmonic. With fundamental ferroresonance, the currents and voltages are badly distorted, but the component at the system frequency is the greatest. In subharmonic ferroresonance, the current and voltage waveforms repeat at intervals of two or more fundamental-frequency cycles. In subharmonic ferroresonance, the currents and voltages contain a large component whose frequency is less than the frequency of the supply system. And with higher harmonic ferroresonance, the response quantities include a large component whose frequency is higher than that of the supply voltage. All three of these responses have been observed during ferroresonance in cable-fed five-legged core, groundedwye/grounded-wye transformers used on the systems of RUS borrowers (Smith, Swanson, and Borst, 1975).


6 Relatively, the normal mode is characterized by lower values of flux, current, and voltage than occur in the ferroresonant mode. In addition, with sinusoidal voltages applied, the responses in the normal mode are more or less sinusoidal. The final ferroresonant mode that has been observed in some nonlinear circuits is one in which the responses are nonperiodic. That is, a steady-state response never develops. This mode can occur in distribution systems during ferroresonance. When both the ferroresonant mode and normal mode responses are possible in circuits as in Figure 6.2 and there is no trapped charge on the capacitor or flux in the nonlinear inductor, the point on the voltage wave at switch closing determines the response mode. A range of closing angles give the ferroresonant mode response and a range of closing angles give the normal mode response.

Surge Arresters Fused Switches Shielded Cable Circuit

φA

Pad-Mounted Transformer H1

X1 No Load X2

φC

H3

X3

φB

Cable Capacitance Neutral Conductor

MGN Feeder (Overhead or Underground)

H2

Cable Shield and Concentric Neutral Riser Pole or Switching Enclosure

FIGURE 6.3: Cable-Fed Three-Phase Transformer Susceptible to Ferroresonance.

Ferroresonance in distribution circuits can occur if a capacitor is placed in series with a nonlinear inductor. This condition is present when one or two phases of the primary line are open and there are unloaded transformers downstream from the open conductor. The capacitance can be either upstream or downstream of the transformer as long as both are downstream from the open-phase point. SWITCHING OPERATIONS PRODUCING FERRORESONANCE IN THREE-PHASE DISTRIBUTION SYSTEMS Figure 6.3 shows a typical situation in which ferroresonance can occur. One three-phase transformer is fed through a cable circuit from an overhead line, or from a pad-mounted switching enclosure on an underground feeder. The transformer primary windings are connected in delta, and load is not connected to the secondary during switching. The fuses providing fault protection to the transformer and cable circuit are located at the cable riser pole or switching enclosure, whatever the situation may be. A permanent connection is made between the transformer primary terminals and the cable circuit. When the unloaded three-phase transformer and cable circuit, as in Figure 6.3, are energized or de-energized with single-pole switches, and just one or two switches are closed, a series LC circuit, similar to that in Figure 6.2, is established, where the inductance is nonlinear. The capacitance is from the primary cable on the open phase(s), and the nonlinear inductance is due to the transformer exciting impedance(s). If the values of L and C are in a specific range, ferroresonance can occur, producing overvoltages from both phase-to-phase and phase-to-ground on the open phases. In Figure 6.3, only the phase A switch is closed, so overvoltages can appear on phases B and C. These overvoltages can persist as long as one or two primary phases remain open. After all three phases are closed or opened to eliminate the single-phase condition, ferroresonance is not possible. When ferroresonance occurs, the transformer may be very noisy because of magnetostriction in the core. The sound emitted by the transformer frequently is described as rattling, whining, or

2 4 6 – Se c t io n 6

6 (Walling, 1992) with grounded loud humming. However, wye-wye transformers on fivewhen low-level overvoltages Ferroresonance is legged cores have shown that, occur across the transformer not a high-current when sufficient capacitance is windings during ferroresopresent to create ferroresonance, or the voltages are phenomenon, but nance, the overvoltage was less than rated, the transhigh overvoltages present in virtually every formers may not emit any may be present. switching event. From case to unusual noises. case, the maximum overvoltIf the overvoltages do not age magnitudes varied within cause an insulation failure or a range. short circuit, the currents usuThe preceding discussion assumed that load ally will not activate overcurrent protective dewas not connected to the secondary side of the vices. Consequently, the overvoltages on the transformer in Figure 6.3. If sufficient resistive open phases persist until all three phases are eiload (reasonably balanced) is connected to the ther connected to, or disconnected from, the secondary, ferroresonant overvoltages will not source. However, if the overvoltages cause the occur. However, most system operators will not failure of cable insulation, transformer insulaintentionally switch a cable circuit and connected tion, or surge arresters on the open phases, the transformer with consumer load connected to currents may operate overcurrent devices. If a the secondary because doing so makes a singlecable insulation failure occurred on open phases phase condition that may cause harmful overB or C in Figure 6.3 when just the phase A fused voltages if insufficient load is connected. switch was closed, the fault current would not blow the fuse in phase A. The operator at the switch location might not be aware of the insuEQUIPMENT AFFECTED BY lation failure. But, when the fused switch in the FERRORESONANT OVERVOLTAGES faulted phase is subsequently closed, high curThe overvoltages produced by ferroresonance rent blows the fuse in the faulted phase. can cause insulation in major equipment to fail. When the transformer primary windings are As early as 1954, the literature mentioned the ungrounded as in Figure 6.3, and the cable cirfailure of reclosers and surge arresters in 24.9-kV cuit is a specified length, a defined switching rural systems from higher than normal 60-Hz operation may produce ferroresonant overvoltovervoltages (Crann and Flickinger, 1954). When ages some of the time and, at other times, overcable-fed transformers are energized or de-enervoltages will not occur. That is, there is a finite gized by switching from a riser pole (as in Figprobability that ferroresonant overvoltages will ure 6.3), transformer insulation failures are occur when the single-pole switches energize or numerous, especially in the early days of underde-energize the circuit and transformer (Young, ground distribution when some pad-mounted Schmid, and Fergestad, 1968). Factors affecting three-phase distribution transformers employed the probability are the point on the voltage the T-T winding connections. Tests with T-T wave at which the switch is operated, the residcable-fed transformers produced transient peak ual flux in the core of the transformer, the initial voltages as high as nine times normal peak voltcharge on the cable capacitance at the time of age during ferroresonance (Young, Schmid, and switching, the cable circuit length, the size of Fergestad, 1968). the transformer, and the switching sequence. Other components damaged by ferroresonant The ferroresonant overvoltage probabilities reovervoltages are cables and elbow connectors. ported by Young, Schmid, and Fergestad (1968) Overvoltages have caused corona in separable are the probabilities of obtaining the ferroresoinsulated connectors used in 34.5-kV undernant mode rather than the normal mode. ground systems (Locke, 1978). Earlier investigations (Smith, Swanson, and Surge arresters applied on the distribution Borst, 1975) and more recent investigations system are either gapped SiC arresters, gapless


6 B

47”

4 7”

A

C

88”

49” 58 ”

” 52

N

Conductor Heights Above Ground A = 300” B = 317” C = 300” N = 267”

CL of Pole

FIGURE 6.4: Conductor Spacings for an Overhead Line on an Eight-Foot Crossarm.

φB CAB

MOV units, or gapped MOV units. In general, a gapped arrester of a given duty cycle voltage rating can withstand a higher ferroresonant overvoltage than a gapless arrester can, provided the peak voltage does not exceed the gap sparkover voltage. The effect of ferroresonant overvoltages on gapless MOV arresters is much less than would be presumed by examining the standard temporary overvoltage (TOV) curves (Walling et al., 1992). The standard TOV curves are developed using stiff 60-Hz sources, but the ferroresonant circuit is weak compared to the load imposed by the MOV arrester in its conductive state. This means that the arrester can hold down the voltage, usually without drawing a large amount of current. However, depending on the heat transfer characteristics of a given MOV arrester design, the arrester may eventually overheat. IMPACT OF CIRCUIT CONSTRUCTION One of the parameters that determines if ferroresonance occurs with single-pole switching of a circuit with an unloaded transformer is the circuit capacitance. The equivalent capacitances of overhead lines are much less, by at least a factor of ten, than the phase-to-ground capacitance of an underground distribution cable of equal length. Because of the higher capacitance, ferroresonance is more likely in underground systems than in overhead systems.

CBC

φA

φC CAC

CAG

CBG

Circuit and transformer capacitances are important in ferroresonance equations.

CCG

Ground

FIGURE 6.5: Equivalent Capacitance Network for an Overhead Multigrounded Neutral Line.

Capacitances of Overhead Lines An overhead line, consisting of three phase conductors and a multigrounded neutral conductor, as shown in Figure 6.4, is represented by six equivalent capacitors as shown in Figure 6.5. There is an equivalent capacitance between each pair of phase conductors and from each phase conductor to ground. The neutral conductor does

2 4 8 – Se c t io n 6

6 not appear in this representation because, capacitively, it is at the same potential as the ground. TABLE 6.1: Values for Equivalent Capacitances of an Overhead Line With 4/0 ACSR Phase Conductors and a 1/0 ACSR Neutral Conductor. Phase-to-Ground Capacitances (microfarads/mile)

Phase-to-Phase Capacitances (microfarads/mile)

CAG = 0.0090

CAB = 0.0033

CBG = 0.0081

CBC = 0.0032

CCG = 0.0092

CAC = 0.0016

Insulation (220 Mils)

Insulation Shield

D

d Phase Conductor Conductor Shield

Neutral Wire

FIGURE 6.6: Cross Section of a Multiwire Concentric Neutral Cable.

Equation 6.4: Shielded Cable Capacitance. C=

0.03886 K µFarads/mile log D/d

where: C = Capacitance, in microfarads/mile D = Diameter over the insulation, in inches d = Diameter over the conductor shield, in inches K = Dielectric constant of the insulation (For HMWPE and XLPE insulation, K is about 2.3. For EPR insulation, K is about 3.0.)

For symmetrical three-phase distribution lines, the phase-to-phase equivalent capacitances are about 0.002 microfarads per mile, and the phaseto-ground capacitances are about 0.010 microfarads per mile (Hopkinson, 1965). These are easy-to-remember, rule-of-thumb values. For a line with 4/0 ACSR phase conductors and a 1/0 ACSR neutral conductor, with the conductor heights and spacings in Figure 6.4, values for the equivalent capacitances are given in Table 6.1. The three phase-to-ground capacitances are not equal because of the unsymmetrical conductor configuration. Also, the phase-to-phase capacitances are not the same, with the capacitance from A to C being the smallest as these two phase conductors are the farthest apart. Capacitances of Cable Circuits With single-conductor shielded cable, all capacitance is from phase to ground; there is no phaseto-phase capacitance. Figure 6.6 shows a cross section of a concentric neutral cable. The separation between the two “plates” making the capacitor—the conductor shield and the insulation shield—is the thickness of the insulation, which is less than one-half inch. In contrast, the conductor separation in an overhead line is several feet or more, so the cable capacitance is much larger than any of the equivalent capacitances of an overhead line. Although the calculation of the equivalent capacitances for the overhead line is rather involved, the calculation of the shielded cable capacitance is straightforward. The capacitance is found with Equation 6.4. In Equation 6.4, the logarithm is to the base 10. Tables 6.2, 6.3, 6.4, and 6.5 list the capacitance of cables with nominal insulation thicknesses of 175, 220, 260, and 345 mils in sizes up through 1,000 kcmil. Each table gives the conductor size, diameter over the insulation, the capacitance for cables with HMWPE and XLPE insulation, and the cable charging in kVA per mile for a three-phase circuit operating at the indicated phase-to-phase voltage. For cables with EPR insulation with the same nominal diameter over the insulation and the same insulation thickness, the capacitances are approximately 1.3 times those in the tables.


6 TABLE 6.2: Representative Capacitance and Three-Phase Charging for XLPE Insulated Cables With 175 Mils Insulation.* Conductor Size (AWG or kcmil)

Nominal O.D. of Insulation (inches)

Capacitance (µFarads/mile)

Three-Phase Charging 12.47 kV (kVAC/mile)

2

0.655

0.269

15.8

1

0.690

0.291

17.0

1/0

0.725

0.312

18.3

2/0

0.765

0.336

19.7

3/0

0.855

0.391

22.9

4/0

0.915

0.427

25.0

250

0.965

0.457

26.8

300

1.020

0.490

28.7

350

1.065

0.517

30.3

400

1.110

0.543

31.9

500

1.195

0.594

34.8

600

1.275

0.641

37.6

750

1.375

0.701

41.1

1,000

1.520

0.786

46.1

@

* Dielectric constant of 2.3. For EPR cables with same nominal O.D., multiply capacitance and charging values in table by 1.3. Note. 175 mil insulation no longer allowed by RUS

TABLE 6.3: Representative Capacitance and Three-Phase Charging or XLPE Insulated Cables With 220 Mils Insulation.* Conductor Size (AWG or kcmil)




2

0.745

0.230

13.5

1

0.780

0.248

14.5

1/0

0.810

0.263

15.4

2/0

0.850

0.282

16.6

3/0

0.940

0.326

19.1

@

4/0

1.005

0.357

21.0

250

1.050

0.379

22.2

300

1.105

0.405

23.8

350

1.155

0.429

25.2

400

1.200

0.451

26.4

500

1.280

0.489

28.6

600

1.360

0.527

30.9

750

1.465

0.576

33.8

1,000

1.610

0.645

37.8

* Dielectric constant of 2.3. For EPR cables with same nominal O.D., multiply capacitance and charging values in table by 1.3.

2 5 0 – Se c t io n 6

6 TABLE 6.4: Representative Capacitance and Three-Phase Charging for XLPE Insulated Cables With 260 Mils Insulation.* Conductor Size (AWG or kcmil)




@

1

0.902

0.240

28.0

1/0

0.942

0.256

30.0

2/0

0.986

0.275

32.1

3/0

1.036

0.295

34.6

4/0

1.092

0.318

37.3

250

1.158

0.345

40.4

300

1.210

0.366

42.9

350

1.261

0.387

45.3

400

1.306

0.405

47.4

500

1.389

0.439

51.4

600

1.443

0.461

53.9

750

1.578

0.515

60.3

1,000

1.767

0.590

69.1


TABLE 6.5: Representative Capacitance and Three-Phase Charging for XLPE Insulated Cables With 345 Mils Insulation.* Conductor Size (AWG or kcmil)



Three-Phase Charging Three-Phase Charging 34.5 kV (kVAC/mile)

1/0

1.070

0.199

46.6

2/0

1.110

0.212

49.7

95.1

3/0

1.200

0.241

56.4

108.1

@ 24.94 kV (kVAC/mile) @

89.3

4/0

1.265

0.261

61.2

117.1

250

1.310

0.275

64.5

123.4

300

1.365

0.292

68.5

131.0

350

1.415

0.308

72.2

138.2

400

1.460

0.322

75.4

144.5

500

1.540

0.346

81.2

155.3

600

1.620

0.371

87.0

166.5

750

1.720

0.401

94.1

179.9

1,000

1.870

0.447

104.8

200.6


The capacitance values in the tables assume that the diameter over the conductor shield, d, is the diameter over the insulation, D, minus twice the insulation thickness. The diameter over the insulation of a given size cable will

vary somewhat from manufacturer to manufacturer, with the values in the second column taken from one manufacturer’s handbook. Cable size has a major effect on capacitance and, consequently, the likelihood of ferroresonance. The


6 TABLE 6.6: Phase-to-Ground Capacitance of Three-Phase GroundedWye Capacitor Banks.* Nominal Three-Phase Rating (kVA)

Capacitance in Microfarads 12.47 kV System 24.9 kV System

150

2.94

0.75

300

5.88

1.53

450

8.83

2.21

600

11.77

2.94

900

17.65

4.41

* Capacitance values based on 115 percent of nominal kVA rating.

capacitance values in the tables are used in application criteria for calculating the maximum cable length that can be switched without exceeding 1.25-pu ferroresonant overvoltages. Capacitance of Capacitor Banks Capacitor banks on a three-phase primary circuit being switched with a transformer may cause ferroresonance because the capacitor acts like a long cable circuit. Table 6.6 gives the capacitances of three-phase grounded-wye capacitor banks installed in 12.47- and 24.9-kV systems. The capacitance values are based on capacitors rated either 7.2 or 14.4 kV, with the capacitance calculated with the assumption that the capacitor may deliver up to 115 percent of nominal kVAR at rated voltage. Even the smallest capacitor banks on a three-phase circuit look like at least a mile of shielded cable.

is particularly true of banks with ungrounded primaries (e.g., floating wye, delta, or open delta). It is also true of grounded-wye/groundedwye pad-mounted transformers using five-legged cores at 24.9 and 34.5 kV. While there are many inherent capacitances internal to a transformer, the relevance of each depends on the primary winding connection. For a grounded-wye primary, the net capacitance between primary winding layers is the major contributor to phase-ground capacitance in transformers with the primary winding wound outside of the secondary (SP construction). Minor contributions are made by the capacitance between the outer primary winding layer to the core and tank. In transformers with the secondary wound on both the inside and the outside of the primary (SPS construction), the capacitance between the outer layer of the primary winding and the first layer of the outer half of the secondary winding is also a major contributor. It should be noted that there is no simple means to measure the equivalent phase-to-ground capacitance of a grounded-wye winding and manufacturers’ design data are needed to calculate this parameter. The equivalent phase-to-ground winding capacitances of a number of groundedwye pad-mounted transformers have been calculated; the average trend versus rated line-to-line primary winding voltage (in kV) and rated kVA has been reduced to the empirical calculation in Equation 6.5 (Walling, 1992). Equation 6.5

Capacitance of Transformer Windings Transformer windings have an inherent capacitance to ground. This capacitance adds to that provided by underground cables, overhead lines, or capacitor banks and it contributes to creating a ferroresonant circuit. For a given transformer, a maximum capacitance to ground can be left connected to an open phase during single-phase switching without risking excessive ferroresonant overvoltage. The transformer capacitance, thus, directly reduces the allowable amount of capacitance that can be provided by the connected cable, overhead line, or capacitor bank. In some cases, the transformer capacitance alone is sufficient to create ferroresonance. This

CXFMR =

0.000469 × (kVA)0.4 (kVA)0.25

For transformers with ungrounded primary connections (e.g., floating wye, delta, or opendelta), the winding capacitances contributing to ferroresonance are not the same as just described for grounded-wye primaries and the empirical equation does not necessarily apply. The layerto-layer capacitance does not contribute to the phase-to-ground capacitance, for example. In a transformer with SPS construction, the capacitance between both (a) the outer layer of the primary and the first layer of the outer secondary

2 5 2 – Se c t io n 6

6 and (b) the innermost primary winding layer and the last layer of the inner secondary add to the phase-to-ground capacitance and, thus, have an effect on the likelihood of ferroresonance.

Ferroresonance When Switching at the Primary Terminals of Overhead and Underground Transformer Banks

Because the winding in transformers with ungrounded primaries does not shunt the phaseto-ground capacitances, the total capacitance can be easily measured.

The primary windings of single-phase distribuFerroresonance can occur with certain transtion transformers in banks in the overhead sysformer connections and switching operations in tem, and in the underground system, may or may overhead distribution systems when the switchnot be grounded by connection to the multiing is done at the primary terminals of the transgrounded neutral conductor of the primary system. former bank (two or three single-phase transformers). Similarly, ferroresonance can occur in underground systems with the same transformer GROUNDED PRIMARY WINDINGS connections when the switching is done at the When the primary windings of the single-phase primary terminals of the bank. The capacitance transformers in the bank are grounded, either forming the ferroresonant circuit in these cases the grounded-wye or the open-wye connection is the inherent capacitance of the transformer is employed. With grounded primary windings, windings as discussed in the previous subsection. whether the secondary is in wye or delta, ferroresBefore widespread loss evaluation, ferroresoonance is impossible during single-pole switchnant overvoltages seldom occurred on five- and ing at the primary terminals of the transformer 15-kV class overhead distribution systems when bank, at any primary voltage level, whether load the switching was done at the primary terminals is or is not connected to the secondary side of of the bank (Stoelting, 1966). Ferroresonance did the bank. But if power factor capacitors, conoccur during this era in 24.9- and 34.5-kV overnected in either delta or floating-wye, are aphead systems when energizing or de-energizing— plied on the secondary side of the otherwise unwith single-pole switches located at the primary loaded bank, ferroresonance is possible. terminals—small transformer banks connected When opening and closing switches at the floating-wye on the primary and delta on the primary terminals of transformer banks (with secondary. Recent tests performed on a bank of two or three single-phase transformers) with the modern low-loss 13.8-kV (line-to-line) transgrounded-wye or the open-wye primary windformers, with the primaries connected in floatings, whether the primary circuits are overhead ing-wye, have shown very seor underground, disconnect vere ferroresonant overvoltsecondary capacitors before All switching of ages. Thus, it is no longer safe the switching is performed if to consider 15-kV class transthe capacitors are connected transformers or formers immune to ferroresoin delta or ungrounded wye. transformer banks nance when switching at the terminals of banks with unUNGROUNDED PRIMARY with ungrounded grounded primary windings. WINDINGS primary connections This subsection considers In the past, when the primary situations in which the switchwindings of the single-phase should be considered ing is done at the primary tertransformers in the bank were as having the potential minals of transformer banks not grounded in 12.47-kV and for ferroresonance. made from single-phase translower voltage systems, and formers when no primary cirload was not connected to the cuits are connected to the secondary system, ferroresoopen terminals. The lessons learned when nance and harmful overvoltages generally did switching transformer banks in overhead systems not occur when single-pole switching was perapply equally well to underground systems. formed at the primary terminals (Ferguson,


6 1968). One can no longer transformers in a 34.5-kV assume this to always be true system resulted in steady-state Temporary when more modern loss-evalvoltages to ground on the neutral-point grounds uated distribution transformers open phases of about 2.2 times are used. Connections in normal (Pennsylvania Electric installed for switching which the windings are Company, 1964). More recent must always be ungrounded are delta, openfull-scale tests of modern lowremoved for normal delta, and floating-wye. If load loss 25-kVA transformers with is connected to the secondary silicon-steel cores, banked in a operation. when the switching is perfloating-wye/delta connection formed at the primary termiat 25 kV, yielded peak overnals, and the load is voltages in excess of four per reasonably balanced, overvoltages also will not unit (Walling, 1991). Whether the primary system occur. is overhead or underground, overvoltages can When 24.9-kV and higher voltages were adoptoccur during single-pole switching at the primary ed for overhead distribution systems, ferroresoterminals of small floating-wye/delta banks in 15-, nance did occur when switching at the primary 25-, and 35-kV class systems. terminals of small floating-wye/delta transformer The neutral point of the primary windings banks without load on the secondary. Tests by should be temporarily connected to the neutral the RUS showed that phase-to-ground overvoltconductor of the primary system to prevent ages as high as 2.5 pu occurred on the open priovervoltages during the single-pole switching at mary phase when switching banks made from the primary terminals of the smaller floating10- and 25-kVA transformers in 24.9-kV systems wye/delta banks. This connection is represented (Crann and Flickinger, 1954). Full-scale tests by in Figure 6.7 by the closing of switch SW1 in the one utility with three 15-kVA units in a 34.5-kV connection between the neutral of the bank and floating-wye/delta bank resulted in phase-tothe neutral conductor of the primary system. Ferground overvoltages of five per unit (Shultz, roresonance occurs during single-phase condi1964). Tests by another utility with three 50-kVA tions, and it doesn’t matter if the cutouts are being closed to energize the bank or opened to de-energize the bank. Thus, the switch in the neutral is closed before the three fused cutouts Multigrounded Neutral Primary Feeder are closed to energize the bank, and the neutral φA switch is also closed before the three fused Neutral Conductor cutouts are opened to de-energize the bank. φB After the cutouts are closed to energize the bank, neutral switch SW1 must be opened. If the φC switch remains closed, the transformer bank acts Fused Cutouts as a ground source for the primary feeder under φA φB φC normal conditions, tending to balance the load Service Surge Arresters Switch on the three primary phases. Furthermore, if an open phase occurred on the primary feeder between the substation and the location of the grounded-wye/delta transformer bank, the bank Pole-Top Transformer Bank would supply the load on the open primary SW1 phase beyond the open point. This condition may produce loadings where the fuses for the grounded-wye bank do not provide overload FIGURE 6.7: Floating-Wye/Delta Transformer Bank with Fused Cutouts protection. The transformers can fail thermally at Primary Terminals. before the fused cutouts operate to relieve the

2 5 4 – Se c t io n 6

6

Ferroresonance with Cable-Fed, Three-Phase Transformers with Delta or Ungrounded-Wye Connected Primary Windings

overload. Of even more importance is the fact that the backfeed condition from the groundedwye bank will be hazardous to personnel working on the lines. Temporarily grounding the neutral of the floating-wye/delta bank prevents ferroresonance during planned single-pole switching. This grounding will not prevent ferroresonance if a phase opens in the primary supply lines when the bank and lines are unloaded or lightly loaded, as the neutral will not be grounded. The floating-wye/delta bank is employed to supply three- and four-wire delta secondaries. An alternative to temporarily grounding the neutral of the floating-wye/delta bank, to prevent ferroresonance during switching at the primary terminals, is to select transformer connections that can be grounded yet do not act as a ground source for the primary system. The open-wye/ open-delta transformer bank satisfies these criteria for service to three- and four-wire delta loads in 25- and 35-kV class systems. However, although this connection prevents ferroresonance for single-pole switching, it may create higher voltage unbalance in the secondary system than

the floating-wye/delta bank; if the two transformers in the open-wye/open-delta bank are inadvertently connected to the same primary phase, the secondary phase-to-phase voltage across the missing leg will be two times normal. As shown in Figure 6.7, if load is connected to the secondary of the floating-wye/delta bank during switching and the load is reasonably balanced, overvoltages will not occur from phase to neutral on the bank side of the open cutouts on the primary side regardless of primary system voltage. However, if the load connected to the secondary is badly unbalanced or connected across just one phase, the phase-to-neutral voltage at the open primary terminal can be as high as 2.65 pu. This is not due to a nonlinear resonance, but to voltage feedback through the secondary load (Gasal, 1986). Such occurrences have caused the failure of gapped SiC and MOV surge arresters connected to the terminals of the bank. If a floating-wye/delta bank is to be switched at its primary terminals with load connected to the secondary, regardless of primary voltage, the load should be reasonably balanced.

When 15-kV class voltages were selected for overhead multigrounded neutral systems, the ungrounded-wye or delta connections frequently were employed for primary windings in distribution transformer banks and in three-phase transformers. These connections had been used successfully in lower voltage primary systems. In the 1950s and ’60s, when pad-mounted and submersible transformers were first produced for UD systems, many of them also had the delta or ungrounded-wye primary windings that had been applied successfully in overhead systems. Early UD systems often consisted of a threephase transformer fed through a cable circuit from an overhead line, with fused cutouts at the riser pole to energize and de-energize the cable and connected transformer. This arrangement is shown in Figure 6.8. During single-pole switching at the riser pole, transformers with ungrounded primary windings sometimes failed, or emitted unusual noises, and externally gapped surge arresters “spat” across the external gap.

Sometimes the arresters failed. The cause of these problems was ferroresonance. For the system configuration of Figure 6.8, Figure 6.9 shows the measured voltage waveforms and current into terminal H2 of the transformer when the cable and transformer are energized with single-pole switches. These waveforms are from tests on a 150-kVA delta/ grounded-wye transformer bank fed through a cable circuit with a phase-to-ground capacitance of 0.1 microfarads per phase. Terminal H1 is energized first by closing the switch in phase A, followed 12 cycles later by closing the switch in phase B to energize terminal H2. During this 12cycle interval, the voltages from H2 to ground and H3 to ground are as high as three per unit, and the transient response has not fully decayed. From the waveforms, the ferroresonant mode response in the first 12 cycles is at fundamental frequency. After the switch in phase B is closed, the transient voltage from terminal H3 to ground approaches four per unit. Although not


6 Surge Arresters

Shielded Cable Circuit

Transformer Switch

L∆

Fused Cutouts φA

Pad-Mounted Transformer kVAT kV IE% X1

H1 XC

No Load

H2

φC

X2 H3

XC

X3

φB XC Cable Capacitance

FIGURE 6.8: Three-Phase Cable-Fed Transformer with a Delta-Connected Primary Winding.

The factors with the greatest effect on the likelihood of overvoltages on the open phases during the switching of a cable circuit and a connected transformer are the following:

2 pu

3 pu H3V

4 pu H2V

H2 Energized

• Transformer kVA size (kVAT in Figure 6.8), • Primary voltage level (kV in Figure 6.8), • Phase-to-ground capacitance (cable length) of the circuit being switched with the transformer, and • Transformer exciting current at rated voltage (IE% in Figure 6.8).

1 pu

Measures that can limit the voltage on the open phases to 1.25 pu are as follows:

H1V H1 Energized

H2 I

FIGURE 6.9: Voltage and Current Waveforms During Ferroresonance with a 150-kVA Delta/Grounded-Wye Bank.

shown, closing the switch in phase C terminates ferroresonance and eliminates the overvoltages. Other full-scale tests show that, when cable-fed transformers have the delta or ungrounded-wye primary windings, steady-state overvoltages as high as four per unit can occur during ferroresonance (Young, Schmid, and Fergestad, 1968).

• Grounding, through a resistor, the neutral of the wye-connected primary windings (secondary connected in delta). This approach has not been widely accepted because of cost, complexity, and the fact that it cannot be used with transformers having the delta-connected primary winding. • Connecting resistive load to the secondary side of the three-phase transformer during single-pole switching. Most users rejected reliance on secondary load as they did not want to intentionally single-phase their customers. Also, if the load was not large enough, high overvoltages could occur and damage the customer’s load.

2 5 6 – Se c t io n 6

6 • Applying three-pole switches at the riser pole. Many users rejected three-pole switches because of their high cost in comparison to fused cutouts. • Performing the single-pole switching at the primary terminals of the three-phase transformer, with no cable connected to the de-energized primary terminals. The de-energized primary terminals of the transformer are those that are not connected directly to the source system, but can have voltage on them because of coupling through the transformer, or because of load on the secondary side of the transformer. • Limiting the length of the cable circuit being switched with the transformer. A 1.25-pu voltage level is used for establishing ferroresonance criteria for the maximum cable length that can be switched with a connected transformer, as temporary overvoltages at this level should not be harmful to the system or equipment. Overvoltages of this and higher magnitudes occur during ground faults. Gapped SiC arresters can withstand 1.25-pu temporary overvoltages without trouble. Distribution class and riser pole MOV surge arresters can withstand temporary overvoltages of 1.25 times system line-to-neutral voltage (about 1.0 times arrester duty-cycle voltage rating) for one to two hours or more, depending on the specific design and manufacturer. The 1.25-pu voltage level is significantly below the applied voltage test given to transformers with ungrounded primary windings and the induced voltage test given to transformers with a grounded primary winding. Similarly, the insulation of other equipment such as cables, cable terminators, splices, separable connectors, and fused cutouts can withstand the 1.25-pu line-to-neutral voltages from ferroresonance, and commonly occurring ground faults. MAXIMUM ALLOWED CABLE LENGTHS TO LIMIT OPEN-PHASE VOLTAGES TO 1.25 PU The maximum cable lengths with a connected transformer that can be switched with singlepole switches so that voltages do not exceed 1.25 pu were found using either full-scale tests (Young, Schmid, and Fergestad, 1968) or TNA

simulations. Normally, the transformer is at the end of the cable circuit as in Figure 6.8, but it may be connected at any point along the cable. The results from TNA simulations are more conservative than those obtained from the full-scale tests performed in the same era. Therefore, the conventional criteria for maximum allowed cable lengths have been based on the TNA studies for transformers with both the delta and ungrounded-wye connected primary windings (Hopkinson, 1967, 1968). The significance of no-load loss to ferroresonance susceptibility was not understood at the time the TNA work was performed because there was good correlation between rated-voltage exciting current and ferroresonance susceptibility. More recent investigations on grounded-wye transformers indicate that exciting current at rated voltage does not accurately reflect ferroresonance susceptibility, but no-load loss does (Walling et al., 1992). Computer simulation of delta-wye transformers tends to indicate that the same is true for transformers with ungrounded primaries (Walling, 1992). Full-scale testing with modern transformers has yet to be performed to determine if the previous approach to ferroresonance guidelines, based on rated exciting current, is valid for low-loss units. In the absence of a verified new approach, the conventional approach based on the TNA investigations of the late 1960s is used in this subsection. When the phase-to-ground voltages on the open phases are limited to 1.25 pu, the voltage across the windings, either delta- or wye-connected, will be less than 1.1 times winding rated voltage. Such overvoltage will not damage transformers as they can continuously withstand across their windings, at no load, 110 percent of winding rated voltage. Application Criteria For a three-phase unloaded transformer with the delta-connected primary windings fed through a cable circuit (as in Figure 6.8), the voltages to ground on the open phases during single-pole switching will not exceed 1.25 pu if the inequality of Equation 6.6 is satisfied. This inequality is expressed in terms of more readily available transformer and system parameters by the inequality of Equation 6.7.


6 It should be noted that Equation 6.7 ignores the capacitance contribution provided by the transformer windings. This transformer capacitance parameter is not readily available to utilities but it can be very important because it may equal the capacitance of 50 or 100 feet of cable,

Equation 6.6 XC ≥ 40 XM

depending on transformer voltage and kVA rating plus the manufacturer’s design practices. From the inequality of Equation 6.7, it can be seen that ferroresonant overvoltages above 1.25 pu are more likely to occur with smaller transformers and longer cable circuits, and at the higher primary voltage levels. The term involving primary voltage, kV, is squared, so doubling the primary voltage reduces the term on the left side of the inequality by a factor of four.

Ferroresonance is more likely with small transformers, higher primary voltages, and longer cable circuits.

where: XC = Phase-to-ground capacitive reactance of one phase of the cable circuit, in ohms. This capacitive reactance is identified in Figure 6.8. XM = Equivalent exciting reactance of the transformer, in ohms. This is equal to the line-to-line rated voltage of the primary winding in volts divided by the rated voltage exciting current in amperes, divided by √3.

Equation 6.7 kVATIE% ≥ 0.286 kV2CµF/ML∆ where: kVAT = Nameplate kVA rating of the three-phase transformer = Exciting current of the transformer at rated voltage in percentIE% age of rated current kV = Rated phase-to-phase voltage of the transformer primary winding in kV. This is the voltage on the nameplate. CµF/M = Capacitance of the shielded single-conductor cable in microfarads per mile. This capacitance is found from Equation 6.4. Representative values are given in Tables 6.2, 6.3, and 6.4. = The total length of the cable being switched with the transL∆ former with delta-connected primary winding. If the cable extends beyond the transformer, this is the total length of the cable being switched.

When the transformer parameters and cable capacitance are known, the maximum length (L∆max) of cable circuit that can be switched with the transformer with the delta-connected primary winding, so that the voltages do not exceed 1.25 pu, is given by Equation 6.8. This equation includes the correction for transformer winding capacitance. Use of Equation 6.8 is illustrated by the following example and data: kVAT IE% kV CµF/M

CXFMR

= = = =

1,500 kVA 1.0% 12.47 kV 0.312 µF/M (for 1/0 phase conductor with 175 mils of XLPE insulation per Table 6.2) = 0.006 µF

Placing these values in Equation 6.8 gives the following: Equation 6.8 L∆max = 3.5kVATIE% CXFMR × 5,280 L∆max = – kV2CµF/M CµF/M where: CXFMR = The equivalent phase-to-ground winding capacitance of the transformer in µF

3.5 × 1,500 × 1.0 0.006 × 5,280 – 12.472 × 0.312 0.312 = 6.7 feet

2 5 8 – Se c t io n 6

6 This example reveals that, even with larger kVA transformers with the delta primary winding, the allowed length of cable to limit the overvoltage to 1.25 pu is impractically short, 6.7 feet, at 12.47 kV. If the primary voltage was 24.9 kV, the winding capacitance correction would exceed the other term and this equation would yield a negative critical cable length. The voltage could, therefore, exceed 1.25 pu if the transformer was switched single phase, solely as a result of internal transformer capacitances. For three-phase transformers with the ungrounded-wye primary winding, the voltages on the open phases during remote single-pole switching through a shielded cable will not exceed 1.25 pu if the ratio of XC to XM is greater than 30 (Hopkinson, 1968). From this, it is concluded that the maximum length of cable circuit, identified as LYmax, and connected transformer with the floating wye-connected primary winding that can be switched so that the voltages do not exceed 1.25 pu is given by Equation 6.9. Equation 6.9 LYmax =

4.7kVATIE% CXFMR × 5,280 – feet kV2CµF/M CµF/M

All terms in Equation 6.9 are the same as in Equation 6.8. From the constant terms, the allowable cable lengths with the ungrounded-wye primary winding are, at most, 34 percent greater than those allowed with the delta primary winding. However, with the ungrounded-wye primary windings, the lengths that limit voltages to 1.25 pu are so short that practical applications usually cannot be made. In light of the low loss levels and small exciting currents in modern distribution transformers, there is little value in specifying maximum cable lengths for transformers with ungrounded primary windings. In all cases, prudence dictates that these transformers be switched only at the primary terminals. Phase-to-ground winding capacitance varies with rating and also greatly with transformer design practices. For those units in which unacceptable overvoltages result from single-phase switching at the terminals (such as

by elbows), three-phase ganged switches must be provided or other means used to control the overvoltages. SWITCHING (OPERATING) PROCEDURES TO PREVENT FERRORESONANCE If transformers with the delta or ungroundedwye primary windings must be used in the UD system, there are two options for preventing ferroresonant overvoltages above 1.25 pu: either use three-pole switches or do single-pole switching at the transformer terminals. The latter is effective only when the internal capacitance of the transformer is less than the critical capacitance. This tends to be true for higher loss transformers in 15-kV class systems and with the larger transformers in the 25-kV class systems. It is not recommended for any size transformer in 35-kV class systems. Reasons sometimes given for the requirement of the delta or ungrounded-wye primary winding connections, rather than the grounded-wye primary connections, especially for the larger three-phase distribution transformers, are as follows: • To isolate the primary and secondary systems so that the fundamental frequency component of the unbalanced load current, the third harmonic load current, and its odd multiples (9th, 15th, 21st harmonic . . .) do not flow in the neutral conductor of the primary system, and • To isolate the primary and secondary systems so that ground relays on the primary system do not see ground faults on the secondary system. Single-Pole Switches When only single-pole switches are available, ferroresonant overvoltages above 1.25 pu can be prevented by switching procedures whereby unloaded transformers and cable circuit are not switched as a single entity. This is accomplished by first, energizing the cable circuit with all transformers disconnected and second, connecting the transformers to the energized cable circuit with switching devices at the primary terminals of the transformer. When the transformer is connected to the cable circuit, no cable is connected to the de-energized primary terminals.


6 the time between closing of the first pole and This procedure is effective at voltage levels of closing of the third pole, referred to as pole 15 kV and below, as long as the transformers span, must not be too long, or high overvoltages are not low loss. However, at the 24.9-kV voltwill build up. With pole spans of one cycle (16.6 age level, or in the case of low-loss transformmilliseconds) or less, harmful overvoltages will ers, single-pole switching at the transformer ternot develop. Similarly, the switch pole span minals to energize just the transformer may remust be one cycle or less on opening. sult in high overvoltages as a result of the interA three-pole switch installed in the threenal capacitances of the transformer. Single-pole phase transformers with the delta or unswitching of transformers with ungrounded prigrounded-wye primary mary windings is not recomwindings prevents overvoltmended in 34.5-kV systems. ages for switching at the priSwitching at the transformer Do not do single-pole mary terminals to energize primary terminals can be done switching of just the transformer. This with load-break elbow conswitch is closed to energize nectors, load-break fusing detransformers with the transformer after the cable vices internal to the transungrounded primary circuit is energized from a reformer, or switches internal to windings at 34.5-kV. mote location, with either sinthe transformer. This is illusgle- or three-pole switches. trated with an example using Similarly, the three-pole switch the radially fed transformer in at the transformer is opened Figure 6.8. before the cable circuit is de-energized at the To energize the cable and connected transremote location. former, do the following: STEP 1: Open the switching devices at the trans-

former, disconnecting the transformer from the cable. STEP 2: Close the single-pole switches at the source end of the cable circuit to energize just the cable circuit. STEP 3: Close the switching devices at the transformer to energize the transformer. If the cooperative desires to energize transformers from a remote location, for operator safety, this switching procedure is unacceptable. To de-energize the cable circuit and transformer, as in Figure 6.8, just the opposite procedure is used. First, the switching devices at the transformer are opened, de-energizing the transformer. Second, the single-pole switching devices at the source end of the cable circuit are opened to de-energize just the cable. Three-Pole Switches Three-pole switching of the cable circuit and connected transformer prevents ferroresonance for either energizing or de-energizing operations. For the three-pole switch to be effective,

Temporary Grounding of the Neutral If the three-phase transformer has the ungrounded-wye primary windings with the neutral of the wye available external to the case, overvoltages do not occur if the neutral is grounded during single-pole switching of the cable circuit and connected transformer, or during switching of just the transformer. When the primary winding of the transformer is rated EY volts, where E is the phase-to-phase rated voltage of the primary, the neutral of the wye is not brought outside the tank. If the primary winding is rated 3 EY/E volts, where E is the rated phase-to-neutral voltage of the primary windings, the neutral of the wye-connected primary is brought out. If the installation is made from three single-phase transformers, the neutral of the primary windings is available. Temporary neutral grounding is effective at any primary voltage level. The neutral point should be grounded only during the switching. If left permanently grounded, the transformer may be damaged thermally if single-phasing occurs on the primary system on the source side of the grounded-wye/delta bank.

2 6 0 – Se c t io n 6

6 Ferroresonance with Cable-Fed Three-Phase Transformers with Grounded-Wye Primary Winding and Five-Legged Core

For transformers with ungrounded primary windings, the cable lengths that allow singlepole switching are so short that practical systems usually cannot be operated. Single-pole switching at the transformer primary terminals to energize or de-energize just the transformer (without cable connected to the de-energized primary terminals) usually prevents objectionable overvoltages in 12.47-kV systems, but is not effective in 24.9- and 34.5-kV systems or where lower loss transformers are applied. And installation of three-pole switching devices is expensive compared with installation of single-pole switching devices. As a result of these limitations, alternative transformer designs allowing single-pole switching without objectionable overvoltages were sought by the utility industry. Some papers on ferroresonance written in the 1960s suggested that the grounded-wye primary winding connections would prevent overvoltages during single-pole switching of a cable circuit and a connected transformer. Operating experience and tests showed that the effectiveness of the grounded-wye primary windings in preventing overvoltages depended on whether the transformer was assembled on a five-legged core or used triplex construction. From tests, the following was learned: • When the three-phase transformer employs triplex construction with grounded primary

H1

H2

H3

FIGURE 6.10: Five-Legged Wound-Type Core with Grounded-Wye Primary Windings.

windings, and power factor capacitors connected in delta or ungrounded-wye are not connected to the secondary, ferroresonance and overvoltages will not occur for singlepole switching of the cable circuit and transformer or for single-pole switching at the primary terminals. • When the transformer has a five-legged core and grounded-wye primary windings, overvoltages can occur when a cable circuit and connected transformer are switched with single-pole switches. In general, much longer lengths of cable and connected transformer can be switched when the transformer has grounded-wye primary windings rather than ungrounded primary windings without exceeding 1.25-pu voltage. • When the transformer has a five-legged core and grounded-wye primary windings, overvoltages can occur during single-pole switching at the primary terminals of the lower-kVA, lowerloss transformers used in 24.9- and 34.5-kV systems. These overvoltages occur due to the internal capacitances of the transformer. The performance of the five-legged core transformer with grounded-wye connected primary windings is discussed in the following subsection. Grounded-wye primary windings are used with grounded-wye secondary windings or ungrounded-wye secondary windings, depending on the type of service. Delta-connected secondary windings should not be used with grounded-wye primary windings. CORE CONFIGURATION Most three-phase distribution transformers with grounded-wye primary windings are constructed on a five-legged, wound-type core. Transformers with grounded-wye primary windings cannot be constructed on a three-legged core as common unbalances in the primary system give severe heating in the transformer tank. Figure 6.10 illustrates the configuration of the five-legged, wound-type core. The core assembly is made from four individual wound-type cores. The two inner core loops have the same mean length and the two outer core loops have the same mean length, but the mean length of


6 an inner core loop is longer than the mean length of an outer core loop. In Figure 6.10, only the primary winding for each phase is depicted. The secondary winding for each phase is wound concentric to the primary winding. Disregarding the effects of transformer winding capacitances, the magnetic circuit of the fivelegged core transformer in Figure 6.10 shows that, when rated voltage is applied between the line terminal of any one winding and ground, with all other windings open circuited, voltage appears between the line terminal and ground of the open windings. Similarly, if voltages from two phases of a three-phase system are applied between the line terminal and ground of any two windings, voltage appears between the line terminal and ground of the open winding. Tests run in the 1970s on transformers rated 12.47/7.2 kV showed that the rms value and the peak value of the voltage from the open terminal to ground did not exceed winding rated rms and peak voltages, respectively (Smith, Swanson, and Borst, 1975). The voltage appearing on the open phases is not sinusoidal as is the applied voltage because of the nonlinear characteristics of the core loops. Tests run more recently with low-loss transformers rated 12.47/7.2 kV also show that the rms value of the voltages from the open terminals to ground does not exceed winding rated rms voltage, but the peak value is about two percent above rated peak voltage as a result of harmonics (Millet, Mairs, and Stuehm, 1990). Thus, for practical purposes, switching at the primary terminals of the grounded-wye primary five-legged core transformers prevents overvoltages in systems with voltages of 15 kV and below. However, test results made available in the summer of 1992 show that the five-legged core transformer internal capacitances can react with the magnetic circuits to produce overvoltages when the switching is done at the primary terminals of the smaller low-loss transformers used in 24.9- and 34.5-kV systems (Walling et al., 1992). From tests on transformers applied in 24.9-kV systems, the peak line-to-ground voltage was 1.29 pu; for transformers applied in 34.5-kV systems, the peak line-to-ground voltage was

1.47 pu. These transformers had lower loss levels than did the transformers used in prior tests in the early 1970s. The voltages appearing at the open-circuited terminals demonstrate that magnetic coupling exists between the phases of the five-legged core transformer. This magnetic coupling between phases, in conjunction with the capacitance to ground (neutral) of the primary cable connected to the open phases, or the internal capacitance of the transformer if high enough, produces a series/parallel LC circuit in which overvoltages are possible. Most published information on the performance of the five-legged core grounded-wye/ grounded-wye transformer is based on tests and TNA simulations done in the early 1970s, before losses were evaluated by most utilities. The significance of the age of this information is that the losses of transformers on which the application criteria in this section are based are higher than the losses of many newer transformers. The material on ferroresonance with five-legged core grounded-wye transformers in this section and, in particular, the criteria for allowed cable length to limit the overvoltages to 1.25 pu, is based on the published literature (Smith, Swanson, and Borst, 1975; Millet, Mairs, and Stuehm, 1990) and personal experience. However, recent tests with newer transformers having lower core losses show that the currently accepted application criteria for allowed cable lengths need to be modified to take into account the lower losses. In fact, core loss is a better indicator of the critical capacitance than is exciting current as used in the older guidelines (Walling et al., 1992). The maximum peak voltage during ferroresonance with the five-legged core grounded-wye transformer is 2.1 pu, based on tests in the 1970s. Tests on the lower loss five-legged core grounded-wye transformers of more recent design show that the sustained voltages during ferroresonance are as high as 2.4 pu (Walling et al., 1992). In comparison, the maximum steady-state overvoltages possible with the delta or ungrounded-wye windings are 4 pu (Young, Schmid, and Fergestad, 1968).

2 6 2 – Se c t io n 6

6 Shielded Cable Circuit

Surge Arresters

Transformer Pad-Mounted Transformer Five-Legged Core Switch

LGY

Fused Cutouts φA

H1

XC

φC

X1 No Load

XC

φB

H2

H3

X3

X2

XC Cable Capacitance

FIGURE 6.11: Three-Phase Cable-Fed Transformer with a GroundedWye Primary Winding on a Five-Legged Core.

__1.02 pu __1.09 pu

1 Cyc. at 60 Hz

1 Cyc. at 60 Hz

(a)

__.71 pu __1.41 pu

1 Cyc. at 60 Hz (b)

1 Cyc. at 60 Hz

__2.0 pu

1 Cyc. at 60 Hz (c) __.77 pu

__1.04 pu 1 Cyc. at 60 Hz

(d)

1 Cyc. at 60 Hz

FIGURE 6.12: Open-Phase Voltage Waveforms with Five-Legged Core, Grounded-Wye Transformers.

MAXIMUM ALLOWED CABLE LENGTHS TO LIMIT OPEN-PHASE VOLTAGES TO 1.25 PU Figure 6.11 shows a five-legged core transformer fed through a cable circuit. The length of the circuit is designated as LGY. Although the transformer is connected to the end of the cable, the system response is the same irrespective of where the transformer is located along its length. The reason for this is that the voltage drop through the series impedance of the cable circuit is negligible during ferroresonance. The type of response and peak value of the overvoltages on the open phases for single-pole switching are affected to a great extent by the distance between the switches and transformer. Roughly, when switching is performed at the primary terminals of the five-legged core transformer (no cable connected to the de-energized terminals), the voltages to ground on the open phases are at a minimum. As the cable length being switched with the transformer increases, the voltage increases, reaching a maximum of about two to 2.5 pu. The distance at which this maximum occurs depends primarily on the kVA rating of the transformer, transformer core loss level, primary voltage, and cable capacitance. Whether one or two phases are connected to the source also affects the responses. Generally, for a given cable length, the overvoltages are higher when just one phase is open. Figure 6.12 shows examples of the steady-state, phase-to-ground voltages on the open phases during single-pole switching, based on full-scale tests with 150-, 225-, and 500-kVA transformers (Smith, Swanson, and Borst, 1975). The response represented by the two voltage waveforms in Figure 6.12(a) is cyclical at fundamental frequency and symmetrical, as only odd harmonics are present. In Figure 6.12(b), the voltage waveforms also are cyclical at fundamental frequency, but they are not symmetrical about the time axis because of the presence of even harmonics. Nonharmonic responses also occur during ferroresonance with the five-legged core transformers, as illustrated in Figure 6.12(c). Here the waveform never repeats itself and there are no identifiable cyclical patterns. These responses produced the maximum voltage of 2.1 pu, and it is during these types of responses that the transformer


6 can be very noisy because of magnetostriction. Twenty- and 30-Hz subharmonic steady-state responses also occur with the five-legged core transformer, as shown in Figure 6.12(d). Work completed in 1992 has re-examined the ferroresonance susceptibility of grounded wye/wye transformers using a five-legged core (Walling, 1992). This investigation concluded that core loss is the key parameter defining the critical capacitance creating ferroresonant overvoltage. Previous guidelines using exciting current as a basis have been generally satisfactory because core losses and exciting current have historically correlated. New low-loss designs and wider variations in design flux density have shown the pitfalls of exciting-current-based guidelines. One such shortcoming, for example, is that the measured exciting current on a transformer can be dominated by the winding capacitance. The older guidelines yield a longer critical cable length for a transformer that has a high measured exciting current because of a high internal capacitance. This internal capacitance adds to that of the cable, however, and the actual critical cable length should be shorter than for that of a transformer with a smaller exciting current that is more inductive. Equation 6.10 kV2Ct ≤ 0.00493 Pnl where: kV = Rated phase-to-phase voltage of the transformer primary winding in kV; the voltage on the nameplate Ct = Total capacitance in µF connected to the open phase, including both cable and internal transformer capacitance Pnl = Three-phase, no-load loss of the transformer at rated excitation in watts

Equation 6.11 LGY =

1 CµF/M

26.0

Pnl kVA0.4 – 2.48 kV2 kV0.25

where: LGY = Length of cable in feet with connected transformer having the grounded-wye primary winding and five-legged core

The guidelines in this subsection are based on more recent research. They apply to a single transformer, without secondary load, connected to cable circuit as shown in Figure 6.11. Application Criteria For the voltage to ground to be limited to 1.25 pu during single-pole switching of the cable circuit and connected transformer, the inequality of Equation 6.10 must be satisfied (Walling, 1992). From the inequality of Equation 6.10, overvoltages above 1.25 pu are more likely to occur with the smaller or more efficient transformers at the higher primary voltage levels and with longer cable circuits. The maximum length of cable circuit, LGYmax, that can be switched with the transformer so that the voltages do not exceed 1.25 pu is given by Equation 6.11. This equation is the inequality of Equation 6.10 combined with an approximate empirical relationship between transformer rating and internal capacitance and solved for LGY. If the primary cable extends beyond the transformer in Figure 6.11, but does not serve any other transformers, LGY is the total length of cable being switched with the unloaded transformer. Transformer no-load loss values can vary widely for transformers of a given kVA rating. This wide variation occurs because the loss evaluation factors used in transformer procurement by various utilities vary widely, and the core losses of various manufacturers’ designs also vary, even when bid to the same loss evaluation specification. Where possible, the actual no-load loss should be used in Equation 6.11 but this is not always feasible when standard practices are developed. Equation 6.12 provides an approximate empirical relationship between transformer kVA rating and no-load losses, reflecting the fact that the percentage of losses tends to decrease for larger transformers (Walling, 1992). This loss relationship is also used in the guidelines of Tables 6.6, 6.7, and 6.8 provided later in this section.

Equation 6.12 Pnl = kVA [4.54 – 1.13 Log10 (kVA)]

2 6 4 – Se c t io n 6

6 EXAMPLE 6.1: Maximum Lengths of Cable Circuit Possible. kVA

= 150 kVA

= Nameplate kVA rating of the three-phase transformer with the five-legged core kV = 12.47 kV = Rated phase-to-phase voltage of the transformer primary winding in kV CµF/M = 0.312 µF/M = Capacitance of the shielded cable circuit in microfarads per mile Placing these values into Equations 6.12 and 6.11 and solving the equations gives the following. From Equation 6.12: Pnl = 150 [4.54 – 1.13 Log109 (150)] = 312.2 watts From Equation 6.11: LGY =

1 312.2 1500.4 26.0 – 2.48 12.470.25 0.312 12.472 LGY = 136 feet For transformers that have small kVA ratings and, consequently, low no-load loss wattage, highly efficient transformers, or transformers with high primary voltage ratings, Equation 6.11 can yield a negative maximum cable length. Of course, a negative cable length is physically meaningless except that it indicates that the transformer internal capacitance is likely to be large enough that sustained voltages exceeding 1.25 pu can occur for single-phase switching at the transformer terminals. Example 6.1 illustrates the use of Equations 6.11 and 6.12. In comparison, with the delta-connected primary winding, all other parameters being the same, the allowed length of cable found from Equation 6.8 is 5.4 feet even if the transformer is assumed to have no capacitance. The maximum cable length calculated above (136 feet) is sufficiently long to permit single-pole switching in many practical applications in which a single transformer is fed radially from an overhead line or from a switching compartment in a UD system. But, if the primary voltage level is 24.9 kV, and everything else remains the same, Equations 6.12 and 6.11 give a maximum allowable cable length of 24 feet. Lengths this short will not allow single-pole switching for any practical application.

Application Data Tables—Maximum Cable Lengths With Equation 6.11, and using Equation 6.12 to approximate typical core losses, the length of cable that can be switched with a five-legged core transformer with grounded-wye primary winding is easily calculated. From this, application data tables can be prepared from typical data. At the 12.47-kV primary voltage level, in many cases the allowed cable lengths permit remote single-pole switching of radially fed transformers. But in 24.9-kV systems, the allowed cable lengths with the smaller and medium-size transformers are so short that, for most practical situations, single-pole switching must be performed at the primary terminals of the transformers. For 34.5-kV systems, as well as low-loss 24.9-kV transformers, the internal transformer capacitances of the smaller kVA-rated transformers are sufficient to create ferroresonant overvoltages in excess of 1.25 pu even when switched at the terminals without connected cable. 12.47-kV Systems Table 6.7 lists the maximum cable lengths that can be energized or de-energized with the transformer (unloaded) in a 12.47-kV system if the voltages are not to exceed 1.25 pu during singlepole switching. Values are given for transformers fed by cables of three different sizes, and typical core loss values are assumed for each kVA rating. For a transformer with greater no-load loss than assumed here, the maximum cable length will be longer. Likewise, a more efficient transformer will have a shorter maximum length. The effect of loss variations do not make an exactly proportional effect on maximum length because of the winding capacitance term (the second term on the right-hand side of Equation 6.11). Note the effect of cable size on allowed lengths. 24.9-kV Systems Table 6.8 lists the maximum cable lengths that can be switched with the transformer in a 24.9kV system if the voltages on the open phases are not to exceed 1.25 pu. If the cable extends beyond the transformer, but serves only one transformer, the total length of cable being switched should be limited to the value given in the table. The maximum allowed cable lengths to limit


6 TABLE 6.7: Maximum Allowed Cable Lengths in 12.47-kV Systems to Limit Open-Phase Voltages to 1.25 PU.

Transformer Nameplate (kVA)

Assumed No-Load Loss (w)

Maximum Cable Length in Feet for the Indicated Cable Size Grounded-Wye Primary #2 1/0 4/0

75

182

100

87

64

112.5

250

144

126

93

150

312

184

161

119

225

423

257

225

166

300

522

323

283

208

500

745

473

413

305

750

968

623

545

401

1,000

1,150

745

652

480

1,500

1,426

930

813

599

2,000

1,619

1,057

924

681

2,500

1,750

1,141

998

735

Note. Cable capacitances of #2, 1/0, and 4/0 cables are 0.230, 0.263, and 0.357 µFarads/ mile, based on 220 mils of TR-XLPE insulation. Allowed cable lengths are longer with 260-mil insulation because of lower capacitance.

TABLE 6.8: Maximum Allowed Cable Lengths in 24.9-kV Systems to Limit Open-Phase Voltages to 1.25 PU.



Maximum Cable Length in Feet for the Indicated Cable Size Grounded-Wye Primary 1/0 4/0

75

182

5

4

112.5

250

12

10

150

312

19

15

225

423

31

25

300

522

42

34

500

745

69

56

750

968

96

77

1,000

1,150

118

95

1,500

1,426

151

122

2,000

1,619

172

139

2,500

1,750

185

149

Note. Cable capacitances of 1/0 and 4/0 cables are 0.256 and 0.318 µFarads/mile, based on 260 mils of TR-XLPE insulation.

the voltage to 1.25 pu with the small and medium kVA five-legged core grounded-wye transformers are very short. Single-pole switching of the cable circuit and connected transformer cannot be performed in many practical systems. For the larger kVA transformers, the allowed cable lengths are sufficiently long that single-pole switching can be performed. 34.5-kV Systems Table 6.9 gives the maximum cable length with a connected transformer that can be energized or de-energized with single-pole switches in a 34.5kV system if the voltages on the open phases are not to exceed 1.25 per unit. For the smaller kVA transformers, overvoltages are likely with singlepole switching at the transformer terminals even without connected cable. The allowed cable lengths are very short even with the larger kVA transformers, virtually excluding the use of single-pole switches other than for switching at the primary terminals of the larger transformers without cable connected to the de-energized terminals. SWITCHING (OPERATING) PROCEDURES TO PREVENT FERRORESONANCE If the cable lengths are longer than those listed in the tables or calculated from Equation 6.11, there are two options for preventing overvoltages above 1.25 pu. The first is to use only three-pole, gang-operated switches when energizing or deenergizing a cable circuit and connected transformer. The second is to do single-pole switching at the primary terminals of the transformer so that, when the transformer is being energized or de-energized, no cable is connected to the deenergized primary terminals. Although this approach is effective for any size transformer at 15-kV and below, it may not limit the overvoltages to 1.25 pu with the smaller kVA, low-loss 24.9- and 34.5-kV transformers. To limit the voltages in these cases, install a three-pole switching device in the transformer. Single-Pole Switches For situations in which overvoltages above 1.25 pu will not occur from single-pole switching at the primary terminals of the five-legged core

2 6 6 – Se c t io n 6

6 TABLE 6.9: Maximum Allowed Cable Lengths in 34.5-kV Systems to Limit Open-Phase Voltages to 1.25 PU.



Maximum Cable Length in Feet for the Indicated Cable Size Grounded-Wye Primary 1/0 4/0

75

182

112.5

250

150

312

225

423

1

1

300

522

5

4

500

745

16

13

750

968

26

21

1,000

1,150

35

28

1,500

1,426

47

38

2,000

1,619

71

54

2,500

1,750

58

47

Do not switch single-phase, even at terminals

grounded-wye transformers (power factor capacitors are not connected to the secondary), switching procedures exist that allow the energization of the smaller kVA transformers connected to, or at the end of, long cable circuits. Such procedures will be illustrated with the arrangement in Figure 6.11, where the cable circuit does not extend beyond the transformer.

Three-Pole Switches Three-pole switches allow the energization and de-energization of the circuit and unloaded transformer with grounded-wye primary without ferroresonant overvoltages if the switch pole span does not exceed one cycle. Furthermore, the three-pole switch will prevent ferroresonant overvoltages even if power factor capacitors are connected to the secondary side of the transformer. Three-pole switches in the lower kVA, lowloss 24.9- and 34.5-kV transformers also will prevent overvoltages above 1.25 pu that otherwise can occur with single-pole switching at the primary terminals.

With grounded-wye primary windings and fivelegged core construction, there are limitations on allowed cable lengths, especially at the 24.9and 34.5-kV primary voltage levels, when energizing and de-energizing cable circuits and connected transformers with single-pole switches. Switching procedures exist that allow single-pole switching without producing overvoltages above 1.25 pu with the five-legged core grounded-wye transformer, excluding the lower loss, lower kVA 24.9- and 34.5-kV transformers. But these procedures may be difficult to implement with multiple transformers on a circuit.

FIVE-LEGGED CORE, GROUNDED-WYE TRANSFORMER TANK HEATING The five-legged core transformer with groundedwye primary windings can experience severe tank heating during certain unbalances in the system. Although the five-legged core prevents tank heating for most unbalances, some unbalances have caused transformer fires. Figure 6.13 illustrates how this happens. With a solid ground fault on phase A of the shielded cable circuit, the riser-pole fuse in the faulted phase blows. The voltage from terminal H1 to ground at the transformer is zero, with

Note. Cable capacitances of 1/0 and 4/0 cables are 0.256 and 0.318 µFarads/mile, based on 260 mils of TR-XLPE insulation.

Ferroresonance with Cable-Fed, Three-Phase Transformers with Grounded-Wye Primary Windings and Triplex Construction

To energize the cable and transformer, open the single-pole switching devices at the transformer; of course, the switching devices for the cable circuit are open. Then close the singlepole switches for the cable circuit to energize just the cable. Finally, close the single-pole switches at the transformer to energize the unloaded transformer. To de-energize the transformer and cable circuit in Figure 6.11, use the opposite procedure. Specifically, open the single-pole device at the transformer. Then open the single-pole switches to de-energize the cable. Although the preceding example is for the simple case of a radially fed transformer at the end of the cable circuit, it can be adapted to radial systems with more than one transformer and to multiple cable segments. This is discussed later in this section.


6 The five-legged core prevents tank heating in transformers with the grounded-wye primary Pad-Mounted Transformer windings during phase-to-ground faults whether Fused Cutouts Five-Legged Core Shielded Cable Circuit or not single-pole overcurrent devices are inφA stalled in the primary feeder. All faults on circuits H1 X1 No XC made with concentric neutral cable will be from Load one or more phases to ground. Consequently, φC when all segments of the primary feeder beH2 X3 X2 H3 XC tween the substation and transformer are made φB with concentric neutral cable, tank heating in XC five-legged core transformers is unlikely. Cable Capacitance Line-to-ground faults, with only the overcurCable Shield and rent devices in the faulted phases opening, Concentric Neutral would have caused excessive tank heating if the Riser Pole transformer in the preceding example had been Line Fuses constructed on a three-legged core. If a solid ungrounded phase-to-phase fault occurs from phases B to C in Figure 6.13 on the overhead line and the fuse in only one of the two faulted phases blows, say phase B, then φA φB φC Single-Phase Reclosers phase C voltage is applied to terminals H2 and H3 of the transformer and phase A voltage is applied to terminal H1. That is, the same voltage is applied to two of the primary phases downstream FIGURE 6.13: Overhead System Supplying a Cable-Fed, Grounded-Wye from the fuses and to two of the high voltage Transformer on a Five-Legged Core. terminals of the transformer. A solid ungrounded fault can occur in an overhead approximately rated voltage line if an insulator breaks at an applied from terminal H2 to angle pole and the phase conFive-legged core ground and from terminal H3 ductor is pulled across one of transformers with to ground. The current in the the other two phase conducB and C phase fuses at the tors. Manufacturing tolerances grounded-wye riser pole is due to the load and/or different preloadings primary windings can on the transformer and a small are reasons only one of the experience severe tank component fed back to the two line fuses blows. ground fault on phase A. ConAn ungrounded phase-toheating during certain sequently, the fuses in phases phase fault also happens when system unbalances. B and C do not blow and the line crews jumper two phases transformer remains energized together to bring temporary until switching is manually service to single-phase conperformed. Because the transsumers following a fault. These former is constructed on a five-legged core, tank conditions impress approximately 58 percent heating does not occur. Similarly, if the solid zero-sequence voltage across the primary windfault to ground (concentric neutral) involves two ings of the transformer, induce currents into the of the single-conductor cables in Figure 6.13, the tank, and cause tank heating unless the transfuses at the riser pole supplying the two faulted former is quickly de-energized. Because the fivephases blow and one phase of the transformer legged core transformer is not symmetrical, the remains energized until switching is manually time to produce high temperatures is a function performed, but tank heating does not occur. of which two terminals are fed from the same Multigrounded Neutral

Overhead MGN Feeder

Surge Arresters

2 6 8 – Se c t io n 6

6 depending on the connection of the secondary primary phase, the induction level at rated voltload. In Figure 6.13, assume a fuse opens at the age, and other design parameters. Regardless, riser pole and sufficient load is connected to the tanks have heated the oil above the flash point, secondary such that the secondary load deterwhich caused the bushings to leak oil, which mines the voltage appearing at the transformer caused a fire. terminals. That is, the primary cable is so short Ungrounded three-phase solid faults on overthat its capacitance is not a factor. If all sechead feeders, with only two of the three singleondary load is constant impedance connected pole overcurrent devices opening, energize all from phase to neutral, tank heating will not three phases downstream from the overcurrent occur. The voltage on the open phase will coldevices at the same voltage. Then the same voltlapse to zero, just as for a ground fault on the age is applied to all three HV terminals of the primary cable. If all secondary load is constant five-legged core transformer. Because the unimpedance connected from phase to phase, and grounded three-phase fault applies 100 percent balanced, the voltages impressed on the transzero-sequence voltage, the tank currents are former have a zero-sequence component of 50 higher and heating occurs in a shorter time than percent and tank heating can occur. If all load is for the ungrounded phase-to-phase fault. During three-phase induction motors that maintain a these conditions, the currents in the transformer speed such that the motors’ negative-sequence fuses usually are not high enough to blow the impedance is less than one-half the motors’ posfuse unless a short circuit develops. itive-sequence impedance, tank heating will not The I2R losses in the tank from the tank curoccur. In an actual system, the total secondary rents are the main source of heating. The tank load is connected from both phase to phase and heating raises the transformer oil temperature phase to neutral and may not be balanced, and because there is an inward heat flow. A few a portion is induction motor; thus, it is difficult users have applied eutectic fuse links inside the to predict whether tank heating will occur with five-legged core transformer to sense oil temperjust an open phase. Regardless, tank heating inature and de-energize the transformer before the cidents have occurred for an open primary oil reaches the flash point, the bushings leak, or phase in the absence of a fault. a fire starts. The effectiveness of the eutectic fuse links is not documented in the literature. However, several utilities have indicated the euTRIPLEX TRANSFORMER CORE tectic fuse links prevented severe tank heating CONFIGURATION… that otherwise would have occurred with fuse Three-phase distribution transformers with links that do not respond to oil temperature. triplex construction have three single-phase If only three-pole overcurrent devices are in core-coil assemblies inside a common tank, as the primary circuits (see Figure 6.13), tank heatillustrated in Figure 6.14. When the primary ing with a five-legged core transformer will not windings are connected in grounded-wye and occur from ungrounded phasethe secondary windings are to-phase faults because all connected in either groundedthree phases are de-energized. wye or ungrounded-wye, Triplex transformers Or, if ungrounded phase-tothere is no magnetic coupling with grounded-wye phase faults are impossible bebetween phases of the transcause concentric neutral cable former. There is no possibility primary windings are is used for all primary circuits of tank heating for unbalances not susceptible to from the station to all transwhere two or three terminals ferroresonance. formers, tank heating will not of the transformer are eneroccur, even when single-pole gized at, or below, rated voltovercurrent devices are used. age from the same primary However, if a primary phase phase. Also, they are not susopens in the absence of a fault, the five-legged ceptible to ferroresonance during single-pole core transformer may experience tank heating, switching, regardless of the primary circuit


6 H1

length or voltage, if it is made with single-conductor shielded cables and ungrounded capacitors are not connected to the secondary system at the time of switching.

the secondary during single-pole switching, voltage appears on the open primary phases. The voltage is due to the phase-to-phase connected load on the secondary applying voltage to the secondary (LV) terminals corresponding to the open primary phases. The magnitude of the voltage to ground on the open primary phase is determined primarily by the ratio of the phaseto-ground load to the phase-to-phase load on the secondary side, the magnitude and power factor of the phase-to-phase secondary load, the transformer leakage impedance, and the phase-to-ground capacitive reactance of the primary feeder. The phase-to-ground voltage on the open phase almost always is less than nominal, although voltages five to 10 percent above nominal phase-to-ground voltage are theoretically possible.

…WITHOUT SECONDARY POWER FACTOR CORRECTION CAPACITORS Figure 6.15 shows a triplex transformer with grounded-wye primary and grounded-wye secondary windings fed through single conductor shielded cables. With no capacitive coupling between phases of the primary cable circuit, with no magnetic coupling between phases of the transformer, and with no load on the secondary, single-pole switching does not cause ferroresonance or overvoltages, regardless of the length of the primary cable circuit. If lagging power factor load is connected to

…WITH SECONDARY POWER FACTOR CORRECTION CAPACITORS If capacitors are connected to the secondary side of the triplex-core transformer having the grounded-wye primary windings and groundedwye secondary windings, single-pole switching remote from the transformer or at the transformer, with no other load connected to the secondary, may cause ferroresonance. Whether it does depends on the connections of the secondary capacitors. If the capacitors are connected in groundedwye (from phase to neutral) on the secondary

H2

H3

FIGURE 6.14: Triplex-Type Wound Core with Grounded-Wye Primary Windings.

Surge Arresters Fused Cutouts

Shielded Cable Circuit

Transformer Switch

Pad-Mounted Transformer Triplex Core

L

φA

H1

X1

φC φB

No Load H3

H2

X3

X2

Cable Capacitance

FIGURE 6.15: Cable-Fed, Triplex-Core Transformer with Grounded-Wye Primary Windings.

2 7 0 – Se c t io n 6

6

Ferroresonance in Underground Feeders Having More Than One Transformer

system ahead of the service disconnect switch in Figure 6.15, ferroresonance and overvoltages will not occur for single-pole switching on the primary side, either at the primary terminals of the triplex-core transformer or remote from the transformer. If the capacitors are connected in delta or ungrounded-wye on the secondary system ahead of the service disconnect switch in Figure 6.15, ferroresonance and overvoltages can occur for single-pole switching on the primary side, either at the primary terminals of the triplex-core transformer or remote from the transformer. Most capacitors for application in low-voltage secondary systems are connected in delta. Thus, when single-pole switching is performed on the primary

side of the triplex-core transformer, secondary capacitors should be disconnected. In UD systems with single-pole switching on the primary, three-phase transformers supplying four-wire wye services, as well as three-wire delta services, should have grounded-wye primary windings. The secondary is connected in grounded-wye or ungrounded-wye. With triplex construction, the possibility of tank heating and ferroresonance, which can occur with fivelegged core transformers, is virtually eliminated. Also, there is no need to develop special switching procedures to prevent ferroresonance. Triplex transformers are, however, inherently heavier and may cost more than five-legged, wound-core designs.

When there is more than one three-phase transformer on a cable circuit when single-pole switching is performed, the total length of cable being switched should be limited so that the voltage to ground does not exceed 1.25 pu. DiPietro and Hopkinson (1976) studied this situation. Their investigations were performed on the TNA with transformers having the deltaconnected HV windings. They concluded that the criterion for limiting the voltage to ground on the open phase to 1.25 pu was the same as if there were just one transformer on the circuit, provided an equivalent capacitive reactance and

an equivalent magnetizing reactance were found. This method can be extended to five-legged core, grounded-wye primary transformers, using the no-load-loss-based approach presented in this section. APPLICATION CRITERIA Transformers with delta or ungrounded-wye primary windings are not recommended in UD systems that use single-pole switching. If triplex-core transformers are used, ferroresonance is not a concern and operators can design and switch the system without developing complex procedures.

Fused Single-Pole Switches

L1(C1)

L2(C2)

SW1

Lj(Cj)

LS(CS) Ti kVAi PNLi

T1 kVA1 PNL1

TN-1 kVAN-1 PNLN-1

TN kVAN PNLN

L3(C3) Symbols Lj - Length of section j in feet Cj - Capacitance of section j in µf/mile

T2 kVA2 PNL2

FIGURE 6.16: Circuit with “S” Cable Sections and “N” Five-Legged Core Grounded-Wye Primary Transformers.


6 Equation 6.13 [C1µF/ML1 + C2µF/ML2 + CjµF/MLj + CSµF/MLS] + 2.476 [kVA10.4 + kVA20.4 + kVA0.4 + kVAN0.4 ] ≤ kV0.25 26 [Pnl1 + Pnl2 + Pnli + PnlN] kV2 where: CjµF/M = Capacitance of cable section “j” in microfarads per mile = Length of cable section “j” in feet Lj S = Number of cable sections in the system during the switching operation kVAi = Nameplate kVA rating of three-phase transformer “i” that is connected to the circuit being switched = No-load loss in watts of three-phase transformer “i” that is Pnli connected to the circuit being switched kV = Rated phase-to-phase voltage in kV of the primary windings of the transformers on the circuit; all transformers are assumed to have the same rated voltage N = Number of transformers connected to the cable circuit during the switching

Fused SinglePole Switches

The multiple transformer criterion given is for cable circuits with transformers having the grounded-wye primary windings and constructed on a five-legged core. With reference to the system in Figure 6.16, which has “S” three-phase cable sections and “N” three-phase transformers, the voltage to ground during single-pole switching at SW1 will not exceed 1.25 pu if the inequality of Equation 6.13 is satisfied. When Equation 6.13 is applied, the number of three-phase cable sections, “S,” and the number of three-phase transformers, “N,” on the circuit need not be the same. Each cable section “j” can have a different length and capacitance pu of length, and each transformer “i” can have different noload loss and kVA rating. There are no restrictions in the topology of the circuit to which Equation 6.13 applies. However, it assumes all transformers are three-phase, five-legged core units. Application of Equation 6.13 is demonstrated with the three-phase system in Figure 6.17 and Example 6.2, assuming the system phase-tophase voltage is 12.47 kV. Table 6.10 lists the transformer and cable data for the system.

3-Way Junction

L1(C1)

L2(C2)

L3(C3)

L4(C4) N.C.

T1 kVA1 PNL1

T2 kVA2 PNL2

L5(C5)

Symbols - Normally closed separable connector - Normally opened separable connector Lj - Length of section j in feet Cj - Capacitance of section j in µf/mile

T4 kVA4 PNL4

FIGURE 6.17: Circuit Configuration for Switching Example 6.2.

T3 kVA3 PNL3

N.O.

2 7 2 – Se c t io n 6

6 TABLE 6.10. Transformer and Cable Data for the System of Figure 6.17. Transformer Data

Cable Circuit Data

Number

Rating (kVA)

No-Load Loss (w)

Section Number

Size (AWG)

T1

500

745

1

4/0

330

0.427

T2

225

850

2

4/0

280

0.427

T3

300

810

3

4/0

350

0.427

T4

75

182

4

4/0

1,500

0.427

5

2

130

0.269

Length (feet)

Capacitance (µF/mile)

Note. Based on 175 mil TR-XLPE Insulation.

EXAMPLE 6.2: Energizing Multiple-Transformer System with Single-Pole Switches. In Figure 6.17, transformers T1, T2, and T3 are loop-feed units with two HV bushings per phase, and transformer T4 is a radial-fed unit supplied from the three-way junction. The normally open point of the loop is at transformer T3. To determine if the entire system can be energized with the singlepole switches at the source end of cable section 1, assuming that load is not connected to the transformers, place the data in Table 6.10 into Equation 6.13 as follows:

[(330 × 0.427) + (280 × 0.427) + (350 × 0.427) + (1,500 × 0.427) + (130 × 0.269)] + 2.476 (5000.4 + 2250.4 + 3000.4 + 750.4) ≤ 0.167(745 + 850 + 810 + 182) or 12.470.25 1,133 > 432 As 1,133 is not less than 432, single-pole switching at the source end of cable section 1 causes phase-to-ground voltages above 1.25 pu. The data in Table 6.10 show that cable section 4 is quite long, 1,500 feet, which suggests that disconnecting cable section 4 from transformer T2 may enable energizing transformers T1, T2, and T4 with single-pole switching at the source end of cable section 1. Assuming that cable section 4 is disconnected from transformer T2, placing the data into Equation 6.13 gives the following:

[(330 × 0.427) + (280 × 0.427) + (350 × 0.427) + (130 × 0.269)] + 2.476 (5000.4 + 2250.4 + 750.4) ≤ 0.167(745 + 850 + 182) or 12.470.25 480 > 297 As 480 is still more than 297, this example illustrates that energizing practical multitransformer loop circuits on a single-pole switching basis often cannot be performed without creating ferroresonant overvoltages in excess of 1.25 pu, even on 12.47-kV circuits.

If all transformers in Figure 6.17 employed triplex construction (grounded-wye primary windings), the entire system could be energized by closing the single-pole switches at the source end of cable section 1, and voltages above 1.25 pu would not occur. Triplex transformers greatly simplify operating procedures, reduce the time to energize or de-energize a circuit with multiple transformers, and prevent ferroresonance if a conductor or jumper opens at light load. SWITCHING (OPERATING) PROCEDURES TO PREVENT VOLTAGES ABOVE 1.25 PU When transformers have groundedwye primary windings (five-legged core), procedures can be developed that allow single-pole switching of the cable circuit and connected transformers without producing voltages above 1.25 pu. This is possible because switching at the primary terminals without cable connected to the de-energized primary terminals and without capacitors on the secondary does not produce voltages above 1.25 pu (excluding the lower loss, lower kVA units in 724.9- and 34.5-kV systems).


6 Also, if only three-pole switching is performed to energize the cable circuit and connected transformer, overvoltages will not occur. Single-Pole Switches Figure 6.17 shows numerous possibilities for energizing or de-energizing a system with singlepole switches when the transformers are loop feed and load-break separable connectors are used at each transformer, junction, and switching compartment. With Equation 6.13, switching procedures can be developed for taking a circuit out of service and then restoring it, so that the voltages do not exceed 1.25 pu. For radially fed

Summary of Techniques for Preventing Ferroresonance in Underground Systems

There are two options for preventing ferroresonance under all conditions that can exist in the UD system if all capacitor banks are connected in grounded wye. First, use only three-pole switches to energize and de-energize cable circuits and their connected transformers if any of the transformers have primary windings connected in either delta, floating-wye, or grounded-wye windings. The second option, which enables singlepole switching of the cable circuit and connected transformer, uses only grounded-wye primary windings and triplex construction for three-phase transformers, or else uses three single-phase transformers with grounded-wye primary connection. This second option also prevents ferroresonance should a jumper or conductor open under light load conditions. The second option is the recommended approach for new systems and additions. Also, cable-fed transformers with openwye/open-delta connections are not susceptible to ferroresonance during single-pole switching. For conditions other than those defined above, the possibility of ferroresonance always exists. However, design and operating procedures that limit the voltages on the open phases to 1.25 pu during single-pole switching are available. These are summarized in the following subsections. DISTRIBUTION SYSTEM DESIGN Primary Cable Circuit Length Ferroresonant overvoltages can be limited to 1.25 pu, when designing the system, by limiting the length of the primary cable circuit that can

transformers, tables can be developed giving the maximum length of cable that can be switched with a transformer of a given size. Ferroresonance always should be considered when you are switching in UD systems. Three-Pole Switches Three-pole switches at switching enclosures and in three-phase loop-feed transformers with the grounded-wye primary windings will allow the greatest flexibility for energizing and de-energizing circuits and connected transformers. However, having three-pole switches at each loop-feed transformer may be difficult to justify economically.

be switched with the transformer. When the transformers have ungrounded primary windings, limiting the length is almost always not practical. Furthermore, single-pole switching at the primary terminals of the transformer with ungrounded primary windings can produce voltages to ground above 1.25 pu. When the transformers have grounded-wye primary windings and are constructed on a fivelegged core, the cable lengths that can be energized or de-energized with single-pole switches without producing voltages above 1.25 pu are long enough that systems can be designed and operated at the 12.47-kV primary level for many situations. But at the 24.9-kV voltage level, and especially at the 34.5-kV voltage level, the cable lengths are short. With the lower loss, lower kVA transformers used in 24.9- and 34.5-kV systems, overvoltages above 1.25 pu occur when switching at the terminals of the transformer. Three-Pole Switches Another system design option for controlling overvoltages is to use only three-pole switches at locations where single-pole switching of the cable circuit and connected transformer(s) will produce voltages above 1.25 pu. If the transformers have ungrounded primary windings, a large number of three-pole switches will be required. When the five-legged core transformers have grounded-wye primary windings, the number of three-pole switches required in the 15-kV class systems may be small because of the relatively

2 7 4 – Se c t io n 6

6 above 1.25 pu may occur for single-pole switching at the primary terminals of the transformer. To recap the switching procedures for limiting overvoltages during single-pole switching, consider a system represented by Figure 6.18. Assuming that energizing cable section 1 (SEC 1) and the transformer with single-pole switches at location SW1 produces voltages above 1.25 pu, either of the following two switching procedures could be used. First, assume the single-pole switches at locations SW1, SW2, and SW3 are open but the switches are closed at location SW4. This represents the situation in which the transformer is loop feed with load-break elbow connectors, or internal loop-feed switches, but does not have a field-operable disconnect between the loop-feed bus of the transformer and the primary winding. To energize the system, do the following:

long lengths of cable and connected transformers that can be switched with single-pole switches. But at the 24.9- and 34.5-kV voltage levels, the lengths of cable that can be switched with fivelegged core, grounded-wye primary transformers are much shorter, by factors of approximately four and nine, respectively. DISTRIBUTION SYSTEM OPERATION (SWITCHING PROCEDURES) With existing distribution systems, switching procedures can be developed that limit overvoltages during single-pole switching operations to 1.25 pu when the transformers have groundedwye primary windings and a five-legged core. Implementation of these switching procedures requires that switching devices—such as loadbreak elbow connectors, fused or solid disconnects, or internal under-oil switches—are located at the transformer primary terminals. They can be used in 15-kV and lower voltage systems having five legged-core transformers with grounded-wye primary windings because switching at the primary terminals, without connected cable, does not result in objectionable overvoltages. They can also be used in 24.9- and 34.5-kV systems if it is recognized that 1.5 pu overvoltages may occur when switching the lower loss, lower kVA grounded-wye transformers at their terminals. However, implementing these procedures in the field may be difficult, especially when restoring service during and after severe storms. These switching procedures generally are not practicable in systems with delta or ungroundedwye primary windings because overvoltages

1. Close the switches at location SW1 to energize cable section 1 (SEC 1). 2. Close the switches at location SW2 to energize the transformer. 3. Close the switches at location SW3 to energize cable section 2 (SEC 2) up to the normally open point. If the switches at SW3 were closed before the single-pole switches at SW2 were closed, cable would be connected to the de-energized primary terminals of the transformer, and overvoltages could occur for single-pole switching at location SW2. Second, assume the single-pole switches are open at all four locations shown in Figure 6.18.

Single-Pole Switching Devices SW1

SW2

SW3

N.O.

SEC 1

SEC 2

SW4

Single-Pole Switches or Fused Disconnect Devices

Five-Legged Core

FIGURE 6.18: Single-Line Diagram of a Portion of a UD System.


6 This represents the situation in which the transformer is loop feed with internal single-pole switching devices for connecting the transformer primary windings to the internal loop-feed bus. To energize the transformer and cables, close the single-pole switches at locations SW1 and SW2. The order of closing is not significant from a ferroresonance standpoint because the switches at location SW4 are open. Then the transformer windings are energized by closing the singlepole switching devices at location SW4, either before or after the switches at location SW3 are closed to energize cable section 2 (SEC 2). In developing switching procedures to prevent or limit ferroresonant overvoltages, the cooperative should consider whether it matters if liquid-filled transformers are energized with switching devices at their primary terminals or at a remote location.

that prevent overvoltages above 1.25 pu usually can be developed. Implementation of these procedures may be difficult under practical conditions. For the basic types of services supplied by three-phase transformers, or banks of singlephase transformers, the preferred winding connections in the UD system, from a ferroresonance standpoint, are defined below. Four-Wire Wye Services Figure 6.1 shows the two most common connections for supply of four-wire wye services. Delta/grounded-wye connections should be avoided in cable-fed transformers unless only three-pole switching devices are used. Grounded-wye/grounded-wye connections should be used in transformers supplying fourwire wye services in UD systems. Triplex construction of three-phase transformers, or use of three single-phase transformers, prevents ferroresonance and eliminates the possibility of tank heating that can occur with the five-legged core transformer. Triplex construction is recommended for three-phase units. Use of five-legged core, three-phase transformers with groundedwye primary windings usually prevents voltages above 1.25 pu for switching at the primary terminals and with reasonable lengths of cable connected to the primary terminals.

SELECTION OF DISTRIBUTION TRANSFORMER CONNECTIONS Transformer connections in UD systems affect the likelihood of ferroresonance during singlepole switching of cable circuits and connected transformers (or just one transformer). In general, delta and ungrounded-wye connected primary windings should not be used for cable-fed transformers in 15-, 25-, and 35-kV class UD systems, unless only three-pole switches are used. For transformers in UD sysFour-Wire and Three-Wire tems, grounded-wye primary Delta Services windings are preferred. With Figure 6.1 shows transformer Transformer triplex-core three-phase transwinding connections for supconnections in UD formers and banks of singleplying the 240/120-volt, fourphase transformers with wire and 240-volt, three-wire systems affect the grounded-wye primary winddelta services. Delta/delta, likelihood of ings, ferroresonance will not floating-wye/delta, and openferroresonance occur during single-pole delta/open-delta connections switching of cables and conshould be avoided if singleduring single-pole nected transformers. With the pole switching of cable circuit switching. five-legged core, voltages and connected transformers is above 1.25 pu will occur if contemplated. These conneccable lengths are too long, or tions are acceptable only if if switching is done at the prithree-pole switches are used mary terminals of the lower loss, lower kVA for all switching operations. 24.9- and 34.5-kV transformers. When the cable Open-wye/open-delta connections prevent lengths are greater than the length allowed to ferroresonant overvoltages during single-pole limit voltages to 1.25 pu, switching procedures switching of cable circuits and connected

2 7 6 – Se c t io n 6

6 Another option is to provide two separate transformer. However, these connections are not service voltages. The 240-volt, three-wire load is symmetrical and are a source of voltage unbalsupplied from a triplex-core ance. Intentional oversizing of grounded-wye/grounded-wye the transformers in this configtransformer with a secondary uration will minimize voltage Use grounded-wye rated 240Y/138 volts. The neuunbalance on the secondary primary windings tral point of the wye-connectside. As long as the conneced secondary windings may tions do not cause objectionand triplex cores be grounded or floating. The able voltage unbalance, and in three-phase 120/240-volt, single-phase the possibility of energizing load is supplied from a singleboth high-voltage terminals transformers to avoid phase transformer with its from the same primary phase ferroresonance. primary connected from phase is minimal, these are the recto neutral. The experience of ommended connection for the cooperative with the opencable-fed transformers. Otherwye/open-delta connection in overhead systems wise, delta/delta or floating-wye/delta conneccan serve as a benchmark in determining the tions must be used with appropriate installation acceptability of the connections for cable-fed of three-pole, gang-operated switches and opertransformers. ating procedures to prevent ferroresonance.


Ferroresonance in UD systems is a complex phenomenon. The probability of its occurring and the severity of the associated overvoltages are a function of many parameters. If the following recommendations are observed in the design and operation of the system and in the selection of transformer connections, problems caused by ferroresonance will be minimized. If only grounded-wye primary windings and triplex cores are used in three-phase transformers, ferroresonance during single phasing is virtually impossible. It is not necessary to develop special switching or operating procedures, use three-pole switches, or limit cable length as may be required with five-legged core transformers with groundedwye primary windings. Triplex construction of three-phase transformers with grounded-wye primary windings prevents tank heating. 1. For service to four-wire wye loads from 12.47-, 24.9-, and 34.5-kV UD systems, use grounded-wye/grounded-wye winding connections. If the three-phase, cable-fed transformer is constructed on a five-legged core, there are limits on the length of cable with a connected transformer that can be energized or de-energized with single-pole switches so

that the overvoltages are limited to 1.25 pu. Do not exceed these cable lengths. If the physical location of the equipment makes it impossible to limit the length of cable, develop switching procedures whereby the cable circuit can be energized or de-energized with the transformer(s) disconnected from the cable circuit. Switching at the primary terminals of the lower loss, lower kVA grounded-wye primary five-legged core transformers in 24.9- and 34.5-kV systems may produce voltages above 1.25 pu. If the three-phase transformer has triplex construction, there are no limits on cable length during the single-pole switching of the cable circuit and transformer with grounded-wye primary windings. Triplex core transformers will not experience tank heating as is possible with five-legged core transformers with grounded-wye primary windings. When purchasing three-phase transformers with grounded-wye/grounded-wye winding connections, always consider both triplex and five-legged core designs, especially in the lower kVA sizes. Always purchase triplex designs if their evaluated cost (includes first cost, cost of losses, etc.) is less


6 than or equal to that of the five-legged core unit. Because the triplex designs simplify switching and operating procedures and eliminate the possibility of tank heating, these benefits should be evaluated in the purchasing decision. Installation of three single-phase transformers, from a ferroresonance and tankheating standpoint, offers the same advantages as do the triplex design transformers. 2. For service to three-wire ungrounded (delta) loads from the 12.47-, 24.9-, and 34.5-kV UD systems, use the grounded-wye/floating-wye winding connection. This connection may also be used to supply the corner-grounded delta secondary system by grounding one of the secondary phase conductors. If the threephase, cable-fed transformer is constructed on a five-legged core, there are limits on the length of cable with a connected transformer that can be energized or de-energized with single-pole switches so that the overvoltages are limited to 1.25 pu. Do not exceed these cable lengths. If the physical location of the equipment makes it impossible to limit the length of cable, develop switching procedures whereby the cable circuit can be energized or de-energized with the transformer(s) disconnected from the cable circuit. If the threephase cable-fed transformer has triplex construction, there are no limits on cable length during the single-pole switching of the cable circuit and connected transformer with grounded-wye primary winding. When purchasing three-phase transformers with grounded-wye/floating-wye winding connections, always consider both triplex and five-legged core designs, especially in the lower kVA sizes. Always purchase triplex designs if their evaluated cost (including first cost, cost of losses, etc.) is less than or equal to that of the five-legged core unit. Three-phase transformers for this application either should have the neutral of the primary windings connected to the transformer tank or else should have the primary neutral brought out through a separate insulated bushing. When the neutral of the highvoltage winding is brought out through an insulated bushing, the neutral should

always be grounded before the transformer is energized. The neutral point of the low-voltage windings rated 480Y/277 volts, in either case, should be brought out through an insulated bushing so that the transformer can serve either a four-wire grounded wye secondary, a three-wire delta (ungrounded) secondary system, or a corner-grounded secondary system. When the transformer serves the ungrounded or corner-grounded systems, the secondary neutral bushing (terminal) should be insulated by the cooperative to avoid unintentional grounding. When the low-voltage windings are rated 240Y/138 volts to supply a 240-volt ungrounded or corner-grounded system, the neutral may or may not be brought out, depending on the preference of the user. However, if the secondary neutral is not brought out on an insulated bushing, it must be floated (isolated) within the tank. 3. For service to 240/120-volt four-wire delta loads in the 12.47-kV UD system employing single-pole switching of cable circuit and connected transformers, use open-wye/opendelta connections. An alternative is to provide two separate services. The 240-volt, three-wire service is supplied from a triplexcore transformer or three single-phase transformers with the primary and secondary windings connected in grounded wye. The secondary winding of the three-phase unit must be rated 240Y/138 volts. The singlephase 120/240-volt service is supplied from a single-phase transformer with its primary winding connected from phase to neutral. Some utilities discourage new applications for the 240/120-volt, four-wire delta services. Instead, they promote four-wire wye service at 208Y/120 volts. This type of service allows use of grounded-wye/grounded-wye connections with triplex construction. If only three-pole switching is used to energize or de-energize the cable circuit and connected transformers, delta, open-delta, or floating-wye connections may be used for the primary windings, provided three-pole switches are also installed at each transformer

2 7 8 – Se c t io n 6

6 to connect and disconnect the transformer from the cable. The advantage of the closeddelta connection is that the maximum possible voltage unbalance, under worst-case conditions, is lower than with the openwye/open-delta connections. 4. For service to 240/120-volt, four-wire delta loads in 24.9- and 34.5-kV systems, use open-wye/open-delta connections. An alternative is to provide two separate services as described in recommendation three, or else promote the 208Y/120-volt, four-wire wye service over the four-wire delta service with grounded-wye windings. If only three-pole switching is used to energize and de-energize the cable circuit and connected transformers, and only three-pole switches are used to connect the transformer to the cable circuit, delta, open-delta, or floating-wye connections may be used for the primary windings. 5. Consumer load connected to the cable-fed transformer should not be used or relied on to prevent ferroresonance during single-pole switching on the primary side of the distribution transformer. If the load is too small, it will not prevent overvoltages, yet may be damaged by the resultant overvoltages. With floating-wye/delta transformer connections, badly unbalanced secondary load will prevent ferroresonance but cause high overvoltages by a different mechanism during single phasing. 6. Delta or ungrounded-wye primary windings can be used with three-phase transformers or banks of single-phase transformers in 24.9- and 34.5-kV UD systems without the

possibility of ferroresonance only if threepole switches are used to energize and deenergize cable circuits and their connected transformers and three-pole switches are used at the HV terminals to connect the transformer to the cable circuit. Single phasing, caused by conductor or jumper opening, may result in ferroresonance under light load conditions. Cable-fed transformers with delta or ungrounded-wye primary windings are not recommended for use in new UD systems. 7. Delta or ungrounded-wye primary winding connections should not be used with cablefed, three-phase transformers or banks of single-phase transformers in 12.47-, 24.9-, and 34.5-kV UD systems when single-pole switching of cable circuits and connected transformers will be performed. The only exception to this recommendation is if the transformer primary windings are connected ungrounded-wye and provisions are made at the transformer to temporarily ground the neutral during single-pole switching operations. 8. Delta or ungrounded-wye connected primary windings should not be used for three-phase transformers or banks of three single-phase transformers even when single-pole switching will be done only at the primary terminals of the transformer. The only exception to this recommendation is if the transformer primary windings are connected ungrounded-wye and provisions are made at the transformer to temporarily ground the neutral during single-pole switching operations.


6 References

Anderson, P.H. Analysis of Faulted Power Systems. Ames, Iowa: Iowa State University Press, 1982. Crann, L.B., and R.B. Flickinger. “Overvoltages on 14.4/24.9kV Rural Distribution Systems.” AIEE Transactions (Power Apparatus and Systems) 73, part III (October 1954): 1208–1212. DiPietro, J., and R.H. Hopkinson. “Ferroresonance on Underground Feeders Having Several Transformers.” Southeastern Electric Exchange Engineering and Operating Meeting, New Orleans, La., April 26–27, 1976. Feldman, J.M., and A.M. Hopkin. “A Simple Nonlinear Analysis of the Single-Phase Ferroresonant Circuit.” Paper C 74 233-3, IEEE PES Winter Meeting, New York, N.Y., January 27, 1974. Ferguson, J.S. “A Practical Look at Ferroresonance.” Missouri Valley Electric Association Engineering Conference, Kansas City, Mo., April 17–19, 1968. Gasal, J. “Prevent Overvoltage Failure of Arresters.” Electrical World (July 1986): 47. Germany, N., S. Mastero, and J. Vroman. “Review of Ferroresonance Phenomena in HighVoltage Power System and Presentation of a Voltage Transformer Model for Predetermining Them.” CIGRE Paper 33-18, August 21–29, 1974. Hendrickson, P.E., I.B. Johnson, and N.R. Schultz. “Abnormal Voltage Conditions Produced by Open Conductors on Three-Phase Circuits Using Shunt Capacitors.” AIEE Transactions 72, part III (1953): 1183–1193. Hopkinson, R.H. “Ferroresonance During SinglePhase Switching of Three-Phase Distribution Transformer Banks.” IEEE Transactions on Power Apparatus and Systems PAS84 (April 1965): 289–293, discussion June 1965, 514–517. Hopkinson, R.H. “Ferroresonant Overvoltage Control Based on TNA Tests on Three-Phase Delta-Wye Transformer Banks.” IEEE Transactions on Power Apparatus and Systems PAS86, no. 10 (October 1967): 1258–1265.

Hopkinson, R.H. “Ferroresonant Overvoltage Control Based on TNA Tests on Three-Phase Wye-Delta Transformer Banks.” IEEE Transactions on Power Apparatus and Systems PAS87, no. 2 (February 1968): 352–361. Locke, P. “Check Your Ferroresonance Concepts at 34 kV.” Transmission and Distribution (April 1978): 3239. Millet, R.D., D.D. Mairs, and D.L. Stuehm. “The Assessment and Mitigation Study of Ferroresonance on Grounded-Wye/Grounded-Wye ThreePhase Pad-Mounted Transformers.” Final Report: NRECA Energy Research Division, January 1990. Pennsylvania Electric Company. “Field Investigation of Ferroresonance on 20/34.5-kV Distribution Three-Phase Transformer Banks.” PENELEC, October 14, 1964. Rudenberg, R. Transient Performance of Electric Power Systems. Cambridge, Mass.: The MIT Press, May 1970. Schultz, R.A. “Ferroresonance in Distribution Transformer Banks on 19.8/34.5 kV Systems.” Rocky Mountain Electric League Spring Conference, Boulder, Colo., April 21, 1964. Smith, D.R., S.R. Swanson, and J.D. Borst. “Overvoltages with Remotely Switched Cable-Fed Grounded Wye-Wye Transformers.” IEEE Transactions on Power Apparatus and Systems PAS94, no. 5 (September/October 1975): 1843–1853. Stoelting, H.O. “A Practical Approach to Ferroresonance as Established by Tests.” Pacific Coast Electric Association Engineering and Operating Meeting, San Francisco, Calif., March 4, 1966. Walling, R.A. “Ferroresonance in Today’s Distribution Systems.” Presentation to the Western Underground Committee, Palo Alto, Calif., May 2, 1991. Walling, R.A. “Ferroresonance Guidelines for Modern Transformer Applications.” Final Report to the Distribution Systems Testing, Application, and Research (DSTAR) Consortium, July 1992.

2 8 0 – Se c t io n 6

6 Walling, R.A., K.D. Barker, T.M. Compton, and L.E. Zimmerman. “Ferroresonant Overvoltages in Grounded Wye-Wye Pad-Mounted Transformers with Low-Loss Silicon-Steel Cores.” Presentation at the IEEE 1992 Summer Power Meeting.

Young, F.S., R.L. Schmid, and P.I. Fergestad. “A Laboratory Investigation of Ferroresonance in Cable-Connected Transformers.” IEEE Transactions on Power Apparatus and Systems PAS-87, no. 5 (May 1968): 1240–1248.

Cathodic Protection Require m e n t s – 2 8 1

7 In This Section:

Cathodic Protection Requirements

Special Note Introduction What to Protect Where to Protect Types of Cathodic Protection Systems Amount of Cathodic Protection

Cathodic Protection Design with Galvanic Anodes Cathodic Protection Installation and Follow-Up Calculation of Resistance to Ground Summary and Recommendations

Special Note

With the 2006 transition to jacketed mediumvoltage distribution-class cables, cathodic protection is not generally needed in today’s applications. With the older BCN cables (now not RUS accepted), cathodic protection was a necessary

precaution to avoid corrosion on the exposed bare concentric neutrals in certain soil types. This section is being left in this manual as a historical reference for those situations in which BCN cables are still in operation.

Introduction

Cathodic protection is an effective and economical means for avoiding underground corrosion in electrical grounding to ensure safe and reliable operation of the electric system. Cathodic protection is protection of the neutral, ground electrodes, and other metal in contact with soil through the use of sacrificial anodes or rectifiers and impressed-current anodes. Cathodic protection has become a necessity for electric utilities because of the broad shift to underground construction and the use of nonconducting materials. In the past, the electric neutral and ground wires were connected to buried steel piping, conduit, tanks, wells, and anchors at many locations. Copper grounds and copper wires in soil received cathodic protection at the expense of buried steel. The large extent

of buried steel, together with surface films on the copper, caused the resulting corrosion of steel to be so slow that it was generally ignored. Now, copper-jacketed ground rods and copper wires may be the only earth contact for safety and electrical protection. With no steel connected, the copper is vulnerable to corrosion because of variations in the soil and from ac voltages present on the neutral. Corrosion of copper in these circumstances can result in loss of electrical protection, property damage, and hazards to operating crews and the public. This section explains, step by step, how to design and install cathodic protection with sacrificial anodes, and how to recognize where such protection will be the most important.

2 8 2 – Se c t io n 7

7 What to Protect

THE ELECTRIC NEUTRAL AND GROUNDS The first requirement is to protect the electric neutral and grounds. The necessity for effective grounding and continuity of the neutral return conductor should be obvious. Cathodic protection is a cost-effective means for avoiding problems in these areas.

OTHER BURIED, GROUNDED METAL Buried steel conduit, anchors, pipes, and well casings are subject to corrosion when they are connected to the common neutral, particularly when the grounding is with copper materials. Cathodic protection of the neutral and grounding system is needed to avoid or control such corrosion.

Where to Protect

Consider cathodic protection at the time of construction where any of the following apply:

are vulnerable to accelerated corrosion because of the effects of dissimilar metals. • At the ends of copper-grounded cable routes in very high-resistivity soils, where ac voltages on the neutral may cause accelerated corrosion of buried copper (Zastrow, 1981). • Near cathodically protected pipelines and in the vicinity of rectifiers that supply dc for cathodic protection; also, near dc-powered railways and mining operations. Control of corrosion from external dc sources may require special measures in addition to installation of cathodic protection (Zastrow, 1979).

• In new residential subdivisions with nonmetallic sewer lines, water lines, gas lines, and copper-grounded electric facilities, where the only buried metal connected to the neutral is copper. Copper may corrode rapidly in these situations because of the mixing of soils during regrading and the absence of the cathodic protection usually provided by buried steel. • Along other routes with widely variable soil conditions that may result from differences in terrain, soil moisture, drainage, and the presence of contaminants such as ashes, coal, dumped refuse, or drainage from barnyards or irrigated fields. • At services from copper-grounded electric circuits where steel pipes, tanks, or well casings

Pole Line

Electron Flow

Anchor (Anode)

Copper Grounds (Cathodes)

Copper Grounds (Cathodes)

Copper Grounds Steel Pipe (Anode)

FIGURE 7.1: Dissimilar Metal Effects Between Buried Metals Connected to the Neutral of an Electric Distribution Line.

Irrigation Well (Anode)

To understand underground corrosion and corrosion-control measures, one must recognize that the electric neutral and ground connections behave as a dc circuit and must be treated as such. The electric neutral, ground electrodes, and other buried metal components connected to them act as a huge galvanic cell. The more noble buried metal surfaces, usually copper, become cathodes and are protected against corrosion. The less noble metals, usually iron and steel, become anodes and are corroded (see Figure 7.1). The electric grounding system may be in an area of widely varying soil resistivity (see Figure 7.2). Shaded areas on the map represent locations of low-resistivity, corrosive soils. Metals in corrosive soil become anodes and corrode, whereas the metals in less-corrosive soil are protected against corrosion. See Section 5 for a detailed discussion of soil electrical resistivity. When both copper and steel are present in variable soils, as at a connection between an underground cable and pole line (see Figure 7.2), the steel anchor in corrosive soil becomes the anode and corrodes, whereas the copper in less corrosive soil is protected against corrosion. If


7 Electron Flow

B

Electron Flow

B

B

A

(A) Corrosive soils (B) Less corrosive soils

FIGURE 7.2: Electric System Map Shaded to Show Corrosive Soil Locations. only buried copper is present, as may be true with underground cables, the copper in corrosive soils sacrifices itself to protect the copper in less corrosive soils.

TABLE 7.1: Typical DC Potentials in Soil. Material

Potential, Volts*

Zinc

-1.1

Iron

-0.6 to -0.7

Copper

0 to -0.1

Carbon

+0.2

* To a copper-copper sulfate half cell

THE SIGNIFICANCE OF DC POTENTIALS Underground corrosion occurs because of differences between dc potentials of the buried metals, either because of dissimilar metals or because of differences in soil. Typical dc potentials of some common metals and of carbon are shown in Table 7.1. The higher (more negative) the dc potential, the more likely the metal is to corrode when connected to other buried metals. Potentials such as shown in Table 7.1 are measured with a high-resistance voltmeter and copper-copper sulfate half cell (see Figure 7.3). If two buried metals are connected, the one with a higher negative potential is corroded while the other is protected (see Figure 7.4). A single buried metal, such as copper, corrodes in varying soils as shown in Figure 7.5. SOIL RESISTIVITY Include soil resistivity measurements as part of a preconstruction survey along each proposed underground cable route. At the same time, record the locations of pipeline crossings and other possible dc sources that may cause cathodic protection interference. Soil resistivity is measured with a four-terminal ground test instrument, with four equally spaced probes placed in a straight line (see Figure 7.6). If measurements are made in the vicinity of a BCN cable or buried pipe, the probes should be off to one side and at right angles to the buried metal.

Detail of Half Cell To Voltmeter

Voltmeter

Copper Rod

Copper Sulfate Solution Excess Crystals Porous Plug

Soil

Half Cell

Soil Metal

FIGURE 7.3: Measurement of Potential to a Copper-Copper Sulfate Half Cell.

Soil resistivity measurements are essential for success in any corrosion control effort.

SOIL AND TERRAIN FEATURES The appearance of soil and the nature of the terrain often reveal locations of corrosive soils as well as soils not likely to be corrosive. Swamps,

2 8 4 – Se c t io n 7

7 –0.05V

–0.06V

C o p p e r

Soil

Copper

EXAMPLE 7.1: Measuring Earth Resistivity.

Electron Flow

Iron

Soil

I r o n

Protected

+ Ions

In Figure 7.6, if A = 5.2 feet, multiply the meter reading by 10 to find earth resistivity in ohm-m. For 5.2-foot spacing, if the meter reads R = 2.4 ohms, soil resistivity is 24 ohm-m. For 10.4-foot spacing, multiply the meter reading by 20.

Corroding

Additional information about soil resistivity measurements and grounding is given in Section 5 of this manual.

FIGURE 7.4: Dissimilar Metal Effects Between Copper and Steel.

streams, and poorly drained areas indicate severely corrosive soils. Well-drained areas and presence of carbonates (lime) usually indicate locations of no significant corrosion. The appearance of soils at cable depth may be significant for the following reasons.

Soil BareNeutral Cable

Ions Anodic (Corroding) Area

Electron Flow

Ions Cathodic (Protected) Area

Arrows represent the flow of electrons in connecting wires and movement of positive ions in the soil. To show “conventional flow” (movement of positive charge), reverse the arrows that represent electron flow.

FIGURE 7.5: Dissimilar Soil Effects on Buried Copper Wires.

A A = Distance between probes

P1

P2

C1

C2

A

• Red clay is only mildly corrosive to buried steel. Red signifies the presence of iron oxide, indicating the presence of oxygen, which helps form passive protective films on iron or steel. • Blue or gray clay, sometimes mottled with white, is severely corrosive to both copper and steel. This usually dense clay is deficient in oxygen and associated with poorly drained soils. • White alkali on the surface in dry areas, or low locations including marshes where drainage is poor, represent corrosive soils. These may be good locations for sacrificial anodes.

Recollections of underground crews about soil types may be valuable.

A Cable

FIGURE 7.6: Measurement of Earth Resistivity with a Four-Terminal Ground Tester.

CORROSION EXPERIENCE Make use of maintenance and replacement records and recollections of underground crews to identify the areas of most probable corrosion. Note the locations and ages of components (cable neutrals, ground wires, ground rods, and anchor assemblies) as well as their condition at the time of observed deterioration or failure.


7 Types of Cathodic Protection Systems

Cathodic protection may be provided by sacrificial anodes of magnesium or zinc, or by rectifiers and impressed-current anodes. RECTIFIER SYSTEMS Rectifier systems are used where there is a need for higher output voltages and/or currents than galvanic anodes can provide. Rectifiers use an ac source to apply a negative potential to the protected structure and return the dc to earth by means of one or more impressed-current anodes. Rectifier systems are more exacting in their requirements for design and regular attention to maintenance. A rectifier system may either prevent or cause corrosion problems, depending on design and the physical location of the anode or anode bed. Adjustable rectifiers, along with impressedcurrent anodes, are used for protecting large underground structures such as pipelines, storage tanks, and wells in oil fields. They may be the most economical means for protecting grounding systems at generating stations, substations, and major industrial facilities. These rectifier systems should be designed and installed by individuals

–0.4V

–1.1V

–0.8V

Electrons

Electron Flow

+ Ions Copper Iron (Protected) (Corroding)

Positive Ions Zinc

Copper (Protected)

Iron

Zinc (Corroding)

FIGURE 7.7: Potentials of a Copper-Steel Couple Before and After Connecting a Zinc Anode.

of proven competence who have experience with such installations. Special attention must be given to the location of anodes and adjustment of the rectifiers to avoid serious damage to the grounding system or other nearby facilities. Small constant-current rectifier units are used to provide more current output than is available from sacrificial anodes. They are usually installed at pad-mounted equipment where the anodes can be buried at minimum cost with a minimum of digging. They have been suggested for retrofitting along existing BCN underground cables. Results have been mixed in terms of reliability and service life, particularly in soils with very high resistivities. Use of these units should be limited at first in order to gain experience. SACRIFICIAL ANODE SYSTEMS Sacrificial anodes are widely used for cathodic protection on electric distribution facilities for reasons of cost and the minimum maintenance required. The balance of this section will be addressed to cathodic protection by means of sacrificial anodes. Sacrificial anodes make use of dissimilar metal effects to protect buried metals against corrosion. For example, if steel is corroding because of a connection to copper (see Figure 7.4), a zinc anode can be added to provide protection (see Figure 7.7). The potentials in Figure 7.7 show the effect of surface films, which have the effect of reducing the amount of current required for cathodic protection. The potentials of the individual metals in soil for copper, iron, and zinc are 0, -0.6, and -1.1 volts, respectively (see Table 7.1). Note: When they are connected together, the potential that results is more negative than the average of the individual metal potentials. For the copper-iron couple, the resulting potential is -0.4 instead of -0.3 volt. The potential of the copper-iron-zinc combination is -0.8 volt, even though the average of the three is -0.57 volt. The difference is due to films that usually form on cathodic surfaces.

2 8 6 – Se c t io n 7

7 The cathodic protection should cause enough current flow to make the dc neutral potential sufficiently negative to prevent corrosion. The potential selected for providing protection is important, as the cost of cathodic protection increases directly with the shift in dc potential to be achieved. Experience in the cooperative’s service area is the best guide for deciding on potentials that are effective yet practical. Steel or iron in soil is usually regarded as protected against corrosion at potentials of -0.85 volt or more negative. In most soils, however, steel anchor assemblies and ground rods are lasting more than 25 years at potentials such as -0.5 to -0.6 volt. Buried copper is generally free of corrosion at potentials of -0.1 to -0.25 volt, but may corrode at more negative potentials in the presence of ac voltages (Zastrow, 1981). Examples drawn from RUS experience are given in Table 7.2. With these potentials as a guide, cathodic protection is designed to maintain dc neutral potentials equal to or more negative

Amount of Cathodic Protection

TABLE 7.2: Suggested DC Potentials for Cathodic Protection.* Conditions

Potential (volts dc)**

Along jacketed cables*** In noncorrosive soils

-0.7

In corrosive soil areas

-0.85

At locations of grounded steel wells, tanks, conduit

-0.85

Along BCN cables In noncorrosive soils

-0.3


-0.4 to -0.7

At locations of grounded steel wells, tanks, conduit

-0.85

At cable terminal poles In noncorrosive soils

-0.6


-0.85

At connections to extensive copper-grounded facilities

-0.4

* From long-term personal experience on electric systems financed by RUS. ** Volts to a copper-copper sulfate half cell. ***There is a lack of experience with protection of cables with semiconducting jackets. Use the same values as for cables with insulating jackets, if these levels are practical; otherwise, try, as a minimum, to achieve those indicated for BCN cables.

than those shown in Table 7.2. More negative potentials provide a greater margin of protection; less negative potentials increase the probability of underground corrosion problems. The potentials in Table 7.2 are intended to provide a starting point until experience is gained in selecting potentials to provide the desired degree of protection at an acceptable cost. Experience in the service area, interpreted in light of “as found” dc potentials, should be helpful for deciding on the ones to be used for cathodic protection design. Noncorrosive soils are defined as those in which steel ground rods and steel anchor assemblies, pipes, and wells connected to a copper-grounded neutral lasted for 20 years or more without significant losses resulting from underground corrosion. Anchor assemblies bonded to pole line neutrals would not have experienced difficulty before underground construction, with first failures after 20 years or more. No underground corrosion of copper would have been noticed before the installation of BCN underground cables. Corrosive soils are those in which significant numbers of anchor rods bonded to a copper grounded neutral failed within 15 years after installation, or in which significant corrosion of buried copper has been experienced. At locations of steel wells, tanks, pipes, and conduit, the design should avoid any objectionable flow of dc in service neutrals. At connections to extensive copper-grounded facilities, the dc potentials are strongly influenced by “as-found” conditions. To achieve a more negative cable neutral potential at such locations, the owner may need to provide cathodic protection for the “foreign” grounding system. Or, as an alternative, the owner should install anodes at locations of grounds along the new cable so that, after a mile or so, the desired potentials will be achieved. Cathodic protection designs, and these specifications, are only approximate because of wide variations in soil properties, variable effects of polarization films, and uncertainties about the characteristics and extent of buried metal structures connected to the neutral. Even so, procedures such as these are necessary to avoid the high cost of ineffective installations and wasted, improperly located anodes.


7 Cathodic Protection Design with Galvanic Anodes

A cathodic protection system is actually a dc circuit (see Figure 7.8). The anode is the voltage source, ground connections (metals buried in soil) are the load, and the neutral conductor and ground wires provide the connections to the load. The current return path is through the soil. (See Figures 7.4, 7.5, and 7.7.) The object of cathodic protection is to shift the dc potential of the neutral to a sufficiently negative value to control or stop corrosion. Cathodic protection design requires at least five steps: STEP 1: Calculate the neutral resistance to

ground.

Cables in Conduit, Large Power Users, and Connections to Other Facilities—only addi-

tional factors to be considered are included. JACKETED CABLES AND OVERHEAD POLE LINES A jacketed underground cable, with an insulating jacket over the neutral wires, is similar to a pole line with regard to grounding and cathodic protection design. In both, most of the grounding is by means of driven ground rods along the line and on consumers’ premises. Additional grounding is provided by other buried metal (conduit, pipes, wells, tanks, and pole anchor assemblies) connected to the common neutral and in contact with the soil.

STEP 2: Decide on the shift in dc neutral

potential that will be necessary for adequate control of corrosion. STEP 3: Calculate the anode output current required. STEP 4: Select the anode types, sizes, and numbers. STEP 5: Decide on approximate locations for anodes along the cable. These steps will now be described in detail for Jacketed Cables and Overhead Pole Lines. For the following other types of situations— Protection of Bare Concentric Neutral Cables, Cables with Conducting (Semicon) Jackets,

STEP 1: Calculate the Neutral Resistance and Conductance to Ground The discussions that follow refer to conductance (the reciprocal of resistance) instead of resistance to avoid the cumbersome formulas that are necessary for finding the equivalent of resistances in parallel. Conductances of individual grounds in parallel can be combined by simple addition or multiplication. Note: There is a great difference between the value of resistance or conductance to ground per unit length (per mile or kft) of neutral and the value for the complete neutral, which usually has a resistance to ground of a fraction of one ohm.

Rw

Ea

En

Ra

Rn

Ea = Open circuit anode potential, the dc potential the anode would assume if not connected to anything else En = DC neutral potential Rn = Resistance-to-earth of the neutral and grounds Ra = Anode resistance Rw = Resistance of the anode lead

FIGURE 7.8: Equivalent Circuit for a Galvanic Anode Connected to the Electric Neutral.

2 8 8 – Se c t io n 7

7 EXAMPLE 7.2: Calculating the Neutral Conductance to Ground Per 1,000 Feet of Cable. TABLE 7.3: Calculated Resistance and Conductance to Ground of Individual Ground Rods as Related to Soil Resistivity.

Ground Rod Size

5

Soil Resistivity (ohm-m*) 10 15 20 100 500 (Resistance, ohms, and conductance, siemens)

2,500

5/8 in. x 8 ft. Resistance

1.995

3.990

5.985

7.981

39.90

199.5

997.6

Conductance

0.5013

0.2506

0.1671

0.1253

0.0251

0.00501

0.00100

Resistance

1.936

3.871

5.807

7.743

38.71

193.6

967.9

Conductance

0.5168

0.258

3.1722

0.1291

0.0258

0.00517

0.00133

Resistance

1.607

3.214

4.821

6.427

32.14

160.7

803.4

Conductance

0.6223

0.3111

0.2074

0.1556

0.0311

0.00623

0.001245

3/4 in. x 8 ft.

3/4 in. x 10 ft.

* For resistivity in ohm-cm, multiply by 100.

Estimate the number of ground rods per 1,000 feet (kft) along the line and on consumers’ premises. If soil resistivities vary substantially, indicate separately the numbers of ground rods in the lower or higher resistivity areas. (Note that this calculation is not precise!) If pole line anchors are included, assume that each is equivalent to half a ground rod. Determine the conductivity to ground from the numbers of driven grounds and the information in Table 7.3. Equations for calculating the values given in Table 7.3 are explained later in the subsection, Calculation of Resistance to Ground. Assume that this is an underground cable with an insulating jacket over the neutral wires. There are, on average, 15 driven grounds per mile, including those on consumers’ premises. One-third of them are in soils with resistivities of approximately 20 ohm-m, and two-thirds are in highresistivity soils, 500 ohm-m and higher. Calculate the conductance to ground per mile of cable: • If grounding is with 5/8-in. × 8-ft rods, and • If grounding is with 3/4-in. × 8-ft rods.

The conductance per eight-foot ground rod (Table 7.3) is as follows: In 20 ohm-m soil, 5/8-in. rods = 0.125 siemens 3/4-in. rods = 0.129 siemens In 500 ohm-m soil, 5/8-in. rods = 0.0050 siemens 3/4-in. rods = 0.0052 siemens Conductance to ground per mile of cable neutral: For 5/8-in. rods, 5 rods in 20 ohm-m soil: 5 × 0.125 = 0.625 siemens 10 rods in 500 ohm-m soil: 10 × 0.005 = 0.050 siemens Sum: 0.675 siemens For 3/4-in. rods, 5 rods in 20 ohm-m soil: 5 × 0.129 = 0.645 siemens 10 rods in 500 ohm-m soil: 10 × 0.0052 = 0.052 siemens Sum: 0.697 siemens


7 STEP 2: Determine the Shift in Potential Required To determine the dc shift in neutral potential to be achieved by cathodic protection, select the neutral potential needed for adequate protection (Table 7.1 or 7.2). The selection will usually be

EXAMPLE 7.3: Determining Required Shift in Potential. Assume, for this example, that the selected neutral potential is -0.7 volt, to minimize probable corrosion of buried steel connected to the neutrals along the cable route. The shift in potential required is the difference between -0.7 volt and the potential the neutral would have without cathodic protection. The potential of a neutral without cathodic protection is determined by ground rods and other buried, connected metals. Potentials that are likely, with no cathodic protection, are -0.1 volt if all grounding is with copper and -0.6 volt if grounding is with steel disregarding the short-term effects of galvanizing (Table 7.4). TABLE 7.4: Potentials to a Copper-Copper Sulfate Half Cell. Buried Metal or Material Zinc or new galvanized steel Old steel or iron Copper

Typical DC Potential (volts)

STEP 3: Calculate the Anode Output Current Required Calculate the anode output current required from Ohm’s law, I = V × G, where V is the shift in potential to be achieved and G is the neutral conductance (reciprocal of resistance), in siemens, to ground. See Example 7.4. STEP 4: Select Anode Types, Sizes, and Numbers Decide on the anode types and sizes needed. The selections of anodes and their locations are likely to determine the effectiveness—and indeed, the success or failure—of the cathodic protection installation. The selections will depend on soil resistivities at anode locations and on the anode characteristics that determine output, service life, and installed cost. Table 7.5 provides information about standard sizes and types of anodes, as follows:

–1.1 –0.5 to –0.6 0 to –0.1

Carbon (in insulation shield or jacket)

a compromise between an “ideal” level of protection desired and cost of the cathodic protection. See Example 7.3.

+0.2

The shift in neutral potential needed to achieve a neutral potential of –0.7 volt is –0.1 volt for steel and –0.6 volt for copper.

EXAMPLE 7.4: Calculating Required Anode Output Current.

• Column (a), anode weight and the backfill package size, for calculating resistance. • Column (c), anode resistance in soil with the resistivity shown in column (b). • Column (d), the current output and estimated life when protecting the neutral at each of four potentials shown. The driving potential for each output calculation is equal to the solution potential indicated for each anode material less the structure potential. The estimated lives in Table 7.5 are based on ampere years’ output of magnesium and zinc anodes as follows:

With copper-jacketed ground rods, Magnesium anodes:

I = 0.6 × 0.675 = 0.405 A (405 milliamperes [mA]) per mile With steel ground rods, I = 0.1 × 0.697 = 0.0697 A (70 mA) per mile

17- and 20-lb. sizes, 1.0 ampere year 32-lb. size, 2.0 ampere years 48- and 50-lb. sizes, 3.0 ampere years Zinc anodes:

30-lb. size, 1.2 ampere years 60-lb. size, 2.4 ampere years

2 9 0 – Se c t io n 7

7 TABLE 7.5: Sacrificial Anode Resistance, Output Current, and Estimated Life.

(a) Anode Nominal Weight (lb.)

Horizontal Anodes at 6-Foot Depth (c) (d) Current Output and Estimated Life for Structure Potential (volts) Soil Resistivity Anode Resistance –0.3 –0.5 –0.7 –0.85 (ohm-m**) (ohms) (mA) (yrs) (mA) (yrs) (mA) (yrs) (mA) (yrs) (b)*

Package Size (in.)

Standard Magnesium (Solution Potential = –1.55V) 17

32

50

22 × 7

26 × 8.5

22 × 10

2

9

139

7

117

9

94

11

77

13

50

23

54

19

46

22

37

28

30

33

100

46

27

37

23

!

18

!

15

!

20

7.7

162

12

136

15

110

18

91

22

50

19

66

30

55

36

45

!

37

!

100

38

33

60

28

!

22

!

18

!

20

7.1

176

17

147

20

120

25

99

31

50

18

69

43

58

!

47

!

39

!

High-Potential Magnesium (Solution Potential = 1.73V) 17

20

48

38 × 6

64 × 5

34 × 8

50

19

75

13

65

15

54

19

46

22

100

39

37

32

32

23

26

38

23

43

200

78

18

!

16

!

13

!

11

!

100

31

46

22

40

25

33

30

28

36

200

62

23

44

20

!

17

!

14

!

500

154

9

!

8

!

7

!

6

!

20

7.2

199

15

171

18

143

21

122

25

50

18

79

38

68

44

57

!

49

!

Zinc (Solution Potential = –1.10V) 30

60

66 × 6

66 × 6

20

5.7

140

8.5

105

11

70

17

44

27

30

8.6

93

13

70

29

47

26

29

41

50

14

57

21

43

28

29

41

18

!

10

2.9

275

8.7

207

12

138

17

86

28

20

5.7

140

17

105

23

70

34

44

!

30

8.6

93

26

70

34

47

!

20

!

! = Not meaningful; exceeds 45 years. * At anode depth. ** To express in ohm-cm, multiply by 100.


7 EXAMPLE 7.5: Selecting Anode Types, Sizes, and Numbers. See Table 7.5, in the columns for a 0.7-volt structure potential, to find the anodes that might be selected, calculated current outputs, and estimated lives: In 20 ohm-m soil:

Standard magnesium, 32-lb., 110 mA, 18 years High-potential magnesium, 48-lb., 143 mA, 21 years Zinc, 30-lb., 70 mA, 17 years Zinc, 60-lb., 70 mA, 34 years

In 500 ohm-m soil: High-potential magnesium, 20-lb., 7 mA, the anode life will exceed 45 years. For copper-grounded cable, the 405 mA required for each mile can be provided with four 32-lb. standard magnesium anodes (440 mA), or three 48-lb. high-potential magnesium anodes (429 mA), or six zinc anodes (420 mA), if they can be located in 20 ohm-m soil. For steel-grounded cable, the 70 mA required can be provided with one zinc anode per mile, installed in 20 ohm-m soil. Prior experience with steel-grounded pole lines, if available, may show that little cathodic protection is needed for cables with insulating jackets and galvanized steel ground rods.

TABLE 7.6. Conductance to Ground of BCNs with Effective Diameters as Indicated.*

1.00

Effective Diameter (inches) 2.00 4.00 8.00 (siemens per 1,000 feet of cable)

12.00

5

23.03

24.03

25.12

26.32

27.07

10

11.51

12.01

12.56

13.16

13.54

25

4.61

4.81

5.02

5.26

5.41

50

2.30

2.40

2.51

2.63

2.71

100

1.15

1.20

1.26

1.32

1.35

150

0.768

0.801

0.837

0.877

0.902

500

0.230

0.240

0.251

0.263

0.271

1,500

0.077

0.080

0.084

0.088

0.090

7,500

0.015

0.016

0.017

0.018

0.018

Soil Resistivity (ohm-m)

* For method of calculating, see information later in this section.

STEP 5: Decide on Approximate Anode Locations Along the Route Decide on approximate anode locations along the route, and also the kinds of anodes, using soil resistivity profiles obtained during preconstruction surveys. Take maximum advantage of the lowest

soil resistivity locations available, but use discretion regarding maximum distances between anodes. Give special attention to locations where the newly installed cable is connected to other facilities (particularly copper-grounded stations), old BCN cables, and loads such as irrigation wells with unusual amounts of grounded metal surrounded by irrigated soil areas. PROTECTION OF BARE CONCENTRIC NEUTRAL CABLES Most of the grounding of BCN cable is by direct contact with the concentric neutral wires. Additional grounding is by means of driven ground rods and by other buried metal connected to the neutral and ground wires. Neutral Conductance to Ground Table 7.6 shows calculated conductances to ground of BCN neutrals with effective diameters as indicated. The conductances are inversely proportional to soil resistivity, so values in soils with other resistivities can be determined without the need for detailed calculations. The effective diameter in Table 7.6 is the diameter of the individual cable, or of the group of cables in a multiphase circuit. If cables are in a flat or irregular configuration, an estimate should be made of an equivalent circle that would enclose the conductor cross sections. Soil resistivity is the most important variable of all, as is shown in Table 7.6. Yet soil resistivity data are subject to considerable error because of practical limitations of field surveys, seasonal temperature variations, and changes in soil moisture. How accurate can the cooperative afford to be? Do the following: 1. Obtain the best soil resistivity data at cable depth that time and resources allow. Particularly, find the approximate boundaries of lowest and highest resistivity areas if variations are in the range of five to one or greater. Learn to identify the extremes from the nature of terrain, soil appearance, and vegetation. The rate of progress in field surveys is slow at first but will increase with experience. 2. To calculate the neutral conductance to ground if there are wide variations in soil resistivity, do the following:

2 9 2 – Se c t io n 7

7 EXAMPLE 7.6: Estimating Neutral Conductance to Ground of BCN Cable. Assume a single-phase one-inch-diameter cable installed with one-third of its length in soils with resistivities on the order of 20 ohm-m and two-thirds of its length in soil with resistivities of 400 to 1,000 ohm-m (500 ohm-m average). From Table 7.6, conductance to ground of one-inch-diameter cable neutrals in 20 ohm-m soil is 5.75 (half of 11.51) siemens per kft, and for cables in 500 ohm-m soil, 0.23 siemens per kft. Following the above proportions, a mile of BCN cable will have 1,760 ft (1.76 kft) of its length in 20 ohm-m soils and 3.52 kft in the 400 to 1,000 ohm-m soils. The neutral conductance to ground per mile of cable is as follows: In 20 ohm-m soil, In 500 ohm-m soil,

5.75 × 1.76 = 10.12 siemens 0.23 × 3.52 = 0.81 siemens Total per mile = 10.93 siemens

Conductance to ground of driven rods along the cable, based on 15 per mile, is as follows: For copper rods, 5/8 in. × 8 ft, In 20 ohm-m soil, In 500 ohm-m soil,

5 × 0.125 = 0.625 siemens 10 × 0.0052 = 0.050 siemens Sum = 0.675 siemens

For steel rods, 3/4 in. × 8 ft, In 20 ohm-m soil, In 500 ohm-m soil,

5 × 0.129 = 0.645 siemens 10 × 0.0052 = 0.052 siemens Sum = 0.697 siemens

Note that more than 90 percent of the conductance to ground—and, therefore, most of the need for cathodic protection—is in the 20 ohm-m soil.

EXAMPLE 7.7: Determining Required Shift in Neutral Potential. For this example, assume that the selected potential is -0.5 volts to a copper-copper sulfate half cell. As -0.1 volts is the probable potential of copper without cathodic protection (Table 7.4), the shift in potential needed is -0.4 volts for copper and zero for steel rods.

Equation 7.1 I=E×G where: I = Current, in amperes E = DC potential shift, in volts G = Conductance, in siemens

a. Estimate the proportion of cable in each resistivity range as suggested in Example 7.6. b. Calculate the conductance separately for cables in each resistivity range and add them together to obtain the total. 3. Locate anodes of suitable size and type in the lowest resistivity soil locations available at reasonable intervals and in numbers sufficient to provide the total output needed. 4. Recognize as normal the wide seasonal variations in soil resistivity that follow variations in temperature and soil moisture. A sacrificial anode system is largely self-adjusting, with anode outputs and current requirements going up and down together. Concerning “accuracy,” recognize that results from a first installation are likely to miss the design objective, in terms of potential shift achieved, by a considerable margin. Reliability of design estimates will improve with experience. Shift in Neutral Potential Required From Table 7.1 or 7.2, select the potential that should be achieved by cathodic protection. The selected potential might be -0.2 or -0.3 volts to a copper-copper sulfate half cell to protect copper concentric neutral wires and grounds, or it might be -0.7 volts (or even -0.85 volts in corrosive soils) to protect anchor rods and other buried steel connected to the neutral. The level of protection selected makes a great difference in the cost of cathodic protection, particularly where grounding is mainly by means of copper in contact with the soil. Anode Output Current and Anodes Required The required anode output current per mile of cable, to achieve the desired potentials, is given by Equation 7.1. See Example 7.8. CABLES WITH CONDUCTING (SEMICON) JACKETS Cables with a conducting (semiconducting) jacket over the concentric neutral wires provide conductivity to surrounding earth through the jacket in addition to that provided by the metallic grounds. Additional cathodic projection capacity will be needed to accommodate the conductivity effect of the jacket. The amount of additional cathodic


7 EXAMPLE 7.8: Determining Output Current and Anodes Required. Using the assumptions of Examples 7.6 and 7.7, determine the anode output current and anodes required. With copper-jacketed ground rods, the potential shift required is –0.4 volts for BCN wires and the ground rods. Conductance to ground: BCN wires = 10.93 siemens, as in Example 7.6 Ground rods = 0.68 siemens Sum = 11.61 siemens Output current required for –0.4 volt shift: I = 11.61 × 0.4 = 4.644 A (4,644 mA) per mile With steel ground rods, the potential shift required is –0.5 volts for the BCN wires and zero for steel ground rods:

BCN cables must not be installed in direct-buried, nonmetallic conduit. Cables in conduit are much more vulnerable to underground corrosion than are cables in soil. Variations in the environment are extreme, going from soil to mixed air and soil, into a humid atmosphere and possible water and then back into soil. Conduit made of nonconducting material presents an insulating barrier around the cable so that cathodic protection from sources outside cannot reach the cables inside. Steel conduit, on the other hand, tends to sacrifice itself to protect bare copper neutral wires inside. If BCN cables must be inside nonmetallic conduit, a length of zinc ribbon anode should be pulled in with the cables and the core wire connected to the cable neutrals at a splice at one or both ends. For new construction, cables with insulating jackets should be installed inside conduit.

Output current required: I = 10.93 × 0.4 = 4.372 A (4,372 mA) per mile See Table 7.5, in the columns for a –0.5 volt structure potential, to find the anodes that might be selected, calculated current outputs, and estimated lives: In 20 ohm-m soil:

Standard magnesium, 50-lb., 147 mA, 20 years High-potential magnesium, 48-lb., 171 mA, 18 years Zinc, 60-lb., 105 mA, 23 years

In 500 ohm-m soil:

High-potential magnesium, 20-lb., 8 mA, the life of the anode will exceed 45 years.

If all anodes can be installed in 20 ohm-m soil, the numbers needed are as follows: For 4,644 mA per mile: 32 per mile of 50-lb. standard magnesium (4,704 mA), or 27 of 48-lb. high-potential magnesium (4,617 mA), or 44 of 60-lb. zinc (4,620 mA). For 4,372 mA per mile: 30 per mile of 50-lb. standard magnesium (4,410 mA), or 26 of 48-lb. high-potential magnesium (4,446 mA), or 42 of 60-lb. zinc (4,410 mA).

protection capacity will be dependent on the volume resistivity of the jacket and the neutral configuration. It also may depend on the presence of ac voltages (Zastrow, 1981). CABLES IN CONDUIT Virtually all cooperatives now use jacketed cable. However, if the cooperative still uses BCN cables,

LARGE POWER USERS The grounds at a large power load, such as an industrial grounding system or a center-pivot irrigation well, may present a resistance to ground that is low compared with that of the electric neutral at that location. Such a ground has a dc potential that is not readily changed unless the owner installs cathodic protection in addition to that along the electric line. Two alternative approaches to consider are as follows: 1. Encourage the owner to install cathodic protection. 2. Install additional anodes at grounding locations nearby. The anodes provide a zone of protection for individual grounds and anchor assemblies, some protection to the consumerowned grounds, and, over a distance along the line, bring the neutral to the desired potential. CONNECTIONS TO OTHER FACILITIES Install additional cathodic protection where a new cable is connected to an existing BCN cable or a copper-grounded substation. Follow procedures as for large power loads. Or install cathodic protection for the existing station or other facility at the time the new cable is installed. (See information later in this section.)

2 9 4 – Se c t io n 7

7 Cathodic Protection Installation and Follow-Up

Making a diligent effort at cathodic protection design, as discussed in the previous subsection, will be wasted effort and expense without the same diligent effort in the actual installation of the cathodic protection. Particular attention should be paid to the following: • Where the anodes are installed along the route, • The location of the individual cathodic protection installations relative to the protected equipment, • How the cathodic protection should be installed, and • How the cathodic protection is connected to the equipment.

existing equipment, the cathodic protection should be further adjusted to those locations as long as they are also near existing corrosion areas and/or lower soil resistivity areas. • Banking Anodes. A final adjustment may be made to the cathodic protection design to locate the greatest amount of cathodic protection at those locations that meet the above criteria. This adjustment, of course, needs to be tempered by the fact that better overall protection may be provided by cathodic protection distributed along the route rather than lumped together at one or more locations.

After cathodic protection locations are determined along a particular route, the same effort In addition, determining the long-term performust continue relative to positioning the camance of the cathodic protection requires a thodic protection with respect to the protected means to monitor the performance. equipment (such as cable) and how it is installed. After the preliminary cathodic protection reAs discussed in the cathodic protection design quirements have been determined and the apsubsection, the anode output is dependent on propriate spacing calculated, the practical asthe soil resistivity (the resistance between the pects of locating this cathodic protection come anode and protected equipment) and the anode into play. Spacing the cathodic lead length. Consequently, the protection equally along the position of cathodic protection route without regard to soil is a compromise between Consider soil conditions, existing equipment these two elements. In addicondition, equipment locations, and so on should be tion, practical considerations avoided at all costs. Cathodic concerning easements and the locations, and so protection installed at equal expense to trench in cathodic on when spacing distances along the route is protection conductors have to wasteful and expensive and be considered. cathodic protection. will generally not provide the On the one hand, installing best protection. Cathodic proanodes far from the cable will tection should be adjusted in protect the greatest length of accordance with the following (in priority order): cable. On the other hand, the protection level is reduced and the existing easements and expense • Known Corrosion Locations. If the existing may not permit installation of cathodic protection a greater distance away from the cable. As a pracequipment is corroding at a particular locatical matter, cathodic protection should be located tion, it is a good assumption that the cathodic protection will corrode (i.e., protect) at the 10 to 50 feet from the protected equipment with 25 feet as a practical compromise (see Figure 7.9). same location. Anodes can be installed either vertically in au• Soil Resistivity. As discussed in the previous gured holes with a shallow trench to the prosubsection, the cathodic protection output is greater in lower soil resistivity areas. Consetected equipment, or horizontally in a trench. Either method will provide similar protection requently, reasonable effort should be made to locate the cathodic protection in these areas. sults. The method used is generally dependent on the equipment available and personal preference. • Existing Equipment. Because above-ground connections can be completed more easily at In either case, the anode should be installed


7 Roadway

Ditch Cable Anode

Possible Lead Routes 10-Ft Minimum (25 Ft Preferred)

FIGURE 7.9: Anode Positioning.

Hose Clamp

Concentric Neutrals

Hose ClampType Connector Neoprene Cushion

Anode Lead

Cable Tinned Copper Strap

Cable

FIGURE 7.10. Anode Connector.

Threaded Post 0.01-ohm Shunt 1 2 3

Leads to Equipment/Cable

Anode Lead

FIGURE 7.11. Test Station Connector.

with its entire length below (one-foot minimum) the protected equipment, which will put the anode generally in more moist soil and will give maximum output. A few guidelines relative to the actual installation should be observed:

• Do not use the anode lead to install the anode. A disconnection may render the anode useless. • To improve anode performance, distribute the anode backfill to surround the anode. • In horizontal installations, turn the anode parallel to the cable or with the anode lead connection furthest from the cable. Doing so will reduce the possibility of the anode corrosion disconnecting itself, although this is unlikely. Anode connections to the neutral are often a compromise. In all cases, the connection to the neutral should be with the best connection (compression, if possible). It is relatively easy to complete a compression connection at above-ground locations and when completing the connection on newly installed cable. Existing cable that has been in the ground a number of years makes the use of compression connections difficult, if not impossible. In these cases, the use of RUS-approved connectors should be considered (see Figure 7.10). In all connections below ground, the concentric neutrals should be thoroughly cleaned and the connection sealed to the extent possible to reduce exposure to soil moisture. In the years after the cathodic protection is installed, it will be necessary to determine whether the cathodic protection is still operating and providing the necessary protection levels. Determination of continuing anode effectiveness is facilitated by the installation of test stations along the route. It is not necessary to install test stations at all cathodic protection locations. Test stations should be installed to provide a representative sample of the cathodic protection. For example, if a number of cathodic protection locations are in similar soil along the route, only one or two test stations are necessary. This is based on the assumption that whatever happens at each location is similar. Of course, test stations adjacent to or inside existing equipment are preferred. There are many commercially available test stations in either freestanding or flush-mount models. Freestanding models are much easier to find in rough terrain but may not be aesthetically pleasing if installed in someone’s front yard. It is recommended that the anode and neutral be connected through a 0.1-ohm shunt resistor at test stations to facilitate testing without disconnection in the future (see Figure 7.11).

2 9 6 – Se c t io n 7

7 The calculation of resistance to ground for purposes of cathodic protection is different from the calculation of ground resistance for purposes of lightning protection as discussed elsewhere in this publication. A search of literature on calculations of resistance to ground reveals that all seem to come from one primary source. This calculation method applies to the resistance of ground rods, buried pipes and conductors, and

Calculation of Resistance to Ground

Equation 7.2 R= where: R ρ L a s

= = = = =

ρ 4L 4L s s2 s4 In + In – 2 + – 2 + ... 4πL a s 2L 16L 512L4

resistance, in ohms soil resistivity, in ohm-cm (ohm-m × 100) half the length, in cm (the wire length is 2L cm) wire radius, in cm twice the depth, in cm (the depth is s/2)

Equation 7.3 R= where: R ρ L a s

= = = = =

ρ 4L ρ L2 2L4 1 – 2 + 4 ... In – 1 + 3s 5s 4πL a 4πs

resistance, in ohms soil resistivity, in ohm-cm (ohm-m × 100) half the length, in cm (the wire length is 2L cm) wire radius, in cm twice the depth, in cm (the depth is s/2)

Equation 7.4 R= where: R ρ L a

= = = =

ρ 4L In – 1 2πL a

resistance, in ohms soil resistivity, in ohm-cm (ohm-m × 100) length of the rod, in cm radius of rod, in cm

anodes for cathodic protection. The source is a work published in 1936 by H.B. Dwight entitled Calculation of Resistances to Ground. Dwight gives equations for resistance to ground of a long-buried cylinder (wire, pipe, or BCN cable), a short buried cylinder (anode), and a vertical rod or cylinder. The resistance of a long horizontal wire or cylinder buried in soil is given by Equation 7.2. (The conductances to ground of BCN cables, Table 7.6, are calculated by use of Equation 7.2.) See Equation 5.13 for an equivalent expression using more conventional measurements. The resistance of a short wire (or cylinder) at a depth greater than its length is given by Equation 7.3. (The resistances of sacrificial anodes, Table 7.5, are calculated by use of Equation 7.3.) The resistance of a vertical rod is given by Equation 7.4. (The resistance and conductance to ground of driven ground rods, Table 7.3, are calculated by use of Equation 7.4.)



1. Cathodic protection is now a necessity, no longer an option, because of the broad shift to underground construction and the use of nonconducting materials underground. 2. Without cathodic protection, serious property damage and electric shock hazards may occur. 3. Buried copper as well as steel is vulnerable to corrosion. 4. Soil resistivities must be known. Without soil resistivity data, efforts to control corrosion are likely to fail. Soil resistivities must be measured as part of the preconstruction survey. 5. The electric neutral and grounding system must be treated as a dc circuit; design by “rules of thumb” and assumptions must be avoided. 6. The steps for cathodic protection in electric grounding differ from the practices now

7.

8. 9.

10. 11.

widely used because of differences between underground pipelines, from which the practices evolved, and electrical grounding systems. Anodes must be placed at the lowest soil resistivity locations available. Otherwise, many may be ineffective. Rectifiers must be used with great care until experience has been gained. Follow-up measurements to monitor the effectiveness of the cathodic protection should be planned. A schedule must be established for monitoring cathodic protection. Follow-up measurements should be made after construction to monitor performance of the cathodic protection.


Direct-Buried System De s i g n – 2 9 9

8

Direct-Buried System Design

In This Section:

Trench Construction Considerations

Depth of Burial

Trench Design Components

Joint-Occupancy Trenches

Trench Layout/Routing Considerations


Most cooperatives find that direct-buried electrical distribution systems are the most cost-effective method to use for serving residential, commercial, and some industrial consumers. Material costs are generally less expensive—and labor costs are far less—than for conduit systems, in most cases. However, the benefits of a direct-buried system can be realized only if the cooperative’s engineer uses sound judgment in the following: • • • •

Route selection, Coordination with other utilities, Trench construction details, Designs based on method of installation (e.g., trenching, plowing, directional boring), • Roadway/railway standards, and • Terrain considerations.

Trench Construction Considerations

RUS Bulletin 1728F-D806 (U.G. Distribution Specifications) and the 2007 NESC describe the basic standards of trench construction, principally in terms of depth, width, and cable separations. (Later in this chapter, burial depths are described in detail.) However, many other factors affecting trench design must supplement the national standards to truly accomplish an effective design. First of all, direct-buried systems

Many of these factors are standard and generic for all parts of the country, and the 2007 National Electrical Safety Code (NESC) (ANSI/IEEE C2) mandates many specific requirements that are to be followed without exception. However, a good working knowledge of local considerations is also a necessary factor. Soil conditions, climate, other utility practices, U.S. Department of Transportation issues, local railway system policies, and so on must be thoroughly understood and applied. Over time, working relations with local developers and building contractors come into play in establishing methods to use (or not use) in designing direct-buried systems.

represent a style of construction that is minimally protected from dig-ins by other utilities or other outside agents. The following questions should be asked when making the decision to use direct-buried systems: • Is there a true likelihood of dig-ins from other utilities (or others) in the location of the project?

3 0 0 – Se c t io n 8

8 • Will the installation of trench warning tape(s) protect cables and effectively avoid dig-ins? • Is additional protection—such as conduit, added depth, concrete barrier, or concreteencased duct bank—needed? • Are animals (rodents) an issue? • Are soil conditions (rock) an issue? • Is future cable maintenance an issue, considering landscaping or other surface treatments? Second, what construction equipment is available for installation that may alter the design of the trench construction? The most common trench installation is done with conventional chain-type trenchers that provide a trench six to eight inches wide. Knowing the answers to the following questions is critical in specifying proper trench units: • Is the available equipment appropriate to provide for effective tamping/soil compaction/installation of trench warning tape given local soil conditions? • What backfilling methods are available for such a narrow excavation?

Trench Design Components

in detail in Book II). However, the economic use of plowing needs to be determined on the basis of the following, which may affect trench design: • Are cables to be installed at multiple depths (e.g., primary at 42 to 48 inches/secondary at 30 to 36 inches)? • Are flexible (HDPE) conduits to be plowed in, with or without cables? • Are sufficient amounts of cables required for installation to justify the cost of plowing equipment/operation? • Are terrain/accessibility/soil types at the job site appropriate for the size of plow required? • Are other utilities planning joint-use participation? • Is the distance between transformers, pedestals, and switching cabinets such that productivity can be realized? • Will different cable configurations be used sequentially? • Are there other existing utilities or other subsurface features that must be crossed or paralleled?

If rocky soil is a consistent issue, the use of a backhoe (12 to 18 inches wide) might be more appropriate to ensure good physical examination of the trench floor for possible installation of clean soil/sand bedding below and above the cable(s). Backhoe trenches allow better access for proper backfilling and compaction if soil conditions or surface treatment requirements are a concern. An alternative form of trenching is the plow, which is a major labor-saving device. There are two types of commonly used plowing methods: the static plow and the vibratory plow (discussed

Another alternative form of direct-buried trenching is the directional boring method, which again should be justified regarding the following:

In addition to the basic requirements called out by RUS specifications and the NESC, the cooperative’s engineer needs to evaluate several other design considerations as a part of a successful installation. The following is a list of both material and labor factors that play into a successful design.

installed 12 inches below final grade during the backfilling/compaction process. If the top surface is paved (asphalt or concrete), it is recommended the tape depth be lowered to 18 inches to avoid disruption by maintenance/replacement of the paved surface treatment. Trench warning tape is typically 6 inches wide, and can be supplied foil-backed for trench-locating purposes. Consistent with industry standards for utility location services, warning tape for electric systems

TRENCH WARNING TAPE Many cooperatives install trench warning tape to assist in preventing dig-ins. Typically, tapes are

• Additional cost of installation (for reasons of existing landscaping, pavement, etc.), • Soil conditions, • Warning tape that cannot be installed, • Depth of burial, • Ampacity requirements of cables, • Multiple cables at varying depths, and • Quantity of cables to be installed.


8 Proctor density standards are based on the maximum compactibility of a given soil in laboratory conditions. A standard of 90 percent to 95 percent density is typically very difficult to achieve and definitely requires mechanical tamping. As a general rule, undisturbed soils naturally occur at 80 percent to 85 percent Proctor density, to put all this into perspective. If the particular project requires a certain level of advanced compaction, 90 percent to 95 percent Proctor density should be specified, and reference should be made to American Association of State Highway and Transportation Officials Designation T-99 and ASTM Designation D-698. Compaction testing is relatively simple to perform and many local testing companies provide these services at nominal cost. FIGURE 8.1: Typical Trench Warning Tape. Source: Electromark Industries, 2004.

is usually red with “DANGER—ELECTRIC LINES BURIED BELOW” wording. Though not required by the NESC, trench warning tape can be an effective tool in preventing dig-ins. Some states require the use of warning tapes as a part of their utility-locating programs. BACKFILL/COMPACTION The cooperative engineer should be knowledgeable about local soil conditions and should understand what is required for successful backfill/compaction of native soils. Sandy or loamy coastal soils compact differently than do stiff clays. Locations with rocky conditions require additional care to ensure clean backfill above and below cable systems to avoid jacket or insulation punctures or other cable damage. The cooperative engineer should also be aware of the site specifics regarding whether the cables are to be trenched on consumers’ premises or on roadway rights-of-way. Many state departments of transportation specify minimum compaction standards, typically 90 to 95 percent Proctor density levels. It is not merely a rating of restoring the disturbed earth to assume settling no more than 10 percent to five percent, respectively.

CABLE COMPACTION BEDS In most trench backfill/compaction specifications, it is typical to call for a minimum bedding of clean backfill four inches below and above the directburied cables to prevent insulation or jacket puncture from rocks (the 2007 NESC specifies four inches of tamped backfill in rocky soils, Section 352A). The NESC further specifies no machine compacting within six inches of the cable. If finding clean backfill (or screening rocky or unsuitable backfill) is not cost-effective, many utilities elect to import sand for this purpose. It should be remembered, however, that the thermal conductivity of sand is often much lower than typical native soils. This lower thermal conductivity can de-rate cable ampacities and should be examined closely on substation circuit exits, bulk feeder cables, or cables expected to be loaded heavily. RISER POLE DESIGN The riser pole is one part of the underground system that must be carefully considered during the design process. Not only is the riser often the limiting factor for cable circuit ampacity, but the physical arrangement of the cable circuit must also be carefully considered. Cable U-guards must be used with caution. Gaps between the pole surface and the guard can pose an opportunity for public access to unprotected cable surfaces. Conduit risers generally provide a more satisfactory

3 0 2 – Se c t io n 8

8 avoid a dig-in, but it also must be recognized that the barrier might have to be removed for cable repairs. Some utilities use flowable-fill, a light-duty concrete mix that uses fly ash rather than gravel as its aggregate and sets up at around 400 to 500 psi breaking strength. This 400 to 500 psi rating appears equivalent to concrete when exposed, but can be removed easily with a standard backhoe bucket. Most ready-mixed concrete plants offer flowable-fill at 75 to 80 percent of the cost of normal concrete mixes. Most state departments of transportation approve this mix on rights-of-way. Consideration should be given to requesting that the ready-mixed concrete (or flowable-fill mixes) be tinted with red dye for added recognition as an electric cable barrier. Trench warning tape added above the dye-tinted concrete may also reduce the probability of dig-ins. FIGURE 8.2: Cable Route Marker. Electro-Mark “DoMark” Style Mfg., 2005.

installation in most cases, even if vents have to be installed to obtain adequate circuit ampacity. Conduits (or U-guards) near traffic ways should be placed in a position with minimum exposure to traffic. Adequate cable support must be provided at the top of the conduit and supported bends should be installed at the bottom. Conduit supports must be of a design that will prevent unaided climbing by the public. Many other aspects of riser design are covered elsewhere in this publication. Also see Section 36 of the National Electrical Safety Code. CONCRETE PROTECTION BARRIERS On certain critical bulk feeder installations, or for high-voltage cable installations, consideration should be given to pouring a three- to four-inch thick (nonreinforced) layer of concrete 12 to 18 inches below grade to act as a protective barrier from dig-ins. Typically, these barriers should be used selectively and only for specific instances in which circuit continuity is critical. Concrete mix should not exceed 2,000 to 2,500 psi because it should provide enough protection to

TRENCH MARKERS On underground substation circuit exits, underground bulk feeder lines, or underground transmission lines, the engineer should consider the installation of cable route markers to denote critical cable routes. The most effective route marker is a plastic pedestal-type marker that extends 24 to 36 inches out of the trench and generally lists contact information, along with the color red and “DANGER—ELECTRIC LINES BURIED BELOW” wording. Route markers typically are specified to be installed every 100 to 200 feet, at road intersections, other utility crossings, and angles or changes in direction, recognizing terrain and likelihood of damage/vandalism in light of the use of the land traversed. In rural areas, the normal spacings of these markers over straightline trench routes can be lengthened to every 1/8 to 1/4 mile. Neither the 2007 NESC nor RUS require cable route markers, as the philosophy of the use of these devices is a combination of the following: • Marking the route for cable protection/dig-in avoidance, and • Being a good “utility neighbor” by notifying other utilities of a critical system.


8 The consuming public tends to regard cable route markers as an eyesore, which can lead to vandalism. Without mandates by industry stan-

Trench Layout/ Routing Considerations

The successful layout/design of an underground electrical distribution system depends, to a large degree, on the effectiveness and workability of the routing selected for the site conditions. Routing should be selected from point to point in the straightest path to provide a logical geographical layout, both for initial construction ease and for future troubleshooting, repairs, and cable location efforts. The 2007 NESC specifies this in 351.A.2, and implies cables should be located so they will be subjected to minimal future disturbance. Bends and turns should respect equipment capabilities, and, more important, respect minimum cable bending radii. Cable routes should be selected to avoid natural detriments such as swamps, steep slopes, streams, bad or corrosive soils, mud, or unstable soils that could shift, causing cable damage (2007 NESC 320.A.2). Following are considerations about other physical entities along the trench route. 1. Cable routes along roadways (longitudinally) should be in the shoulder area far enough away to avoid undermining the road surface and to avoid disturbance from road surface maintenance. (Many state departments of transportation provide minimum separations and added burial depths.) 2. Bridges require additional separation, both for cable protection and to ensure trenching operations do not undermine bridge supports. 3. Railway systems require special attention, as the 2007 NESC calls for 50-inch minimum burial depth below the top of the rails (36inch depth if the rail system is a trolley car line), as per NESC 320.A.5.a. Many railway systems require additional burial depths for crossings and have many restrictions for longitudinal routes. Most railway companies mandate steel conduit or casings for electric circuit installations—rather than bare, directburied cables—often over the entire expanse

dards or regulations, cooperatives find maintenance of route markers ongoing.

4.

5.

6.

7.

of the railroad right-of-way. Cables may be permitted parallel to tracks on their rights-ofway, but pedestals, junction boxes, or manholes generally are not allowed. Direct-buried cables should be not closer to in-ground swimming pools (or auxiliary pool equipment) than five feet, as per NESC 351.C.1. If this separation is not possible, conduit must be added. Other utilities should be recognized when routes are selected to avoid crossing conflicts and to provide each utility with the ability to maintain its lines in the future. Typically, 12 inches is the established minimum separation between electric lines and other utilities, including telephone, CATV, water, sewer, gas, and steam lines. However, each utility can require additional separation by mutual agreements. Many telephone and CATV utilities require greater separation distances of five to 15 feet to provide for maintenance of lines. Steam lines require additional separation to avoid problems with heat dissipation, which reduces cable ampacities. Cables closer to steam lines than 18 to 24 inches will require thermal insulation material between the two systems. A utility crossing or installation close by needs to reflect the need to not undermine either utility with the initial trench installation or in future maintenance excavations. Cables should not extend under buildings but, if required, must have mechanical protection (conduit) and should be done in a manner that avoids foundation settlement and does not damage cable systems (NESC 351.C.2.). Cables should be installed below the seasonal frost line in an area, if possible, to avoid mechanical shear on cables resulting from freezing and thawing creating contraction and expansion forces on cables.

3 0 4 – Se c t io n 8

8 The NESC specifies the minimum earth cover reSome existing soil conditions, such as solid or quired over direct-buried power cables. This layered rock, prevent cable burial at the required minimum cover is the distance from the top of depth. The 2007 NESC 352.D.2.b allows lesser the cable to the earth surface. burial depths if supplemental Table 8.1 shows the 2007 protection is provided. This NESC requirements. supplemental protection must Burial depths To achieve the minimum protect the cable from damage cover, make the trench depth resulting from normal activity should exceed the three to four inches greater. at the earth surface. Conduit, NESC minimums. The deeper trench depth alor concrete-encased conduit, lows for a three- to four-inch is the typical method of supsoil bedding under the cable plemental cable protection. and for the cable diameter. Figure 8.3 illustrates minimum allowable burial depths for various caCLEARANCE FROM OTHER UTILITIES bles. Another requirement for direct-buried power caThe cooperative’s engineer should typically bles is separation from other buried utilities, inspecify burial depths in excess of the NESC mincluding the following: imums. Typical trench depths are 30 to 36 inches for secondary cable and 36 to 42 inches • Sewers, for primary systems. These depths allow some • Fuel lines, margin for installation error and minor surface • Natural gas lines, changes after completion. It must be recognized • Water lines, that the cooperative should make special al• Telephone lines, and lowances for areas where the surface may be • CATV lines. lowered later. For example, in rural areas, particularly in The 2007 NESC requires a minimum radial areas subject to cultivation, consideration should separation of 12 inches from other utilities. Ocbe given to using burial depths of 42 to 48 casionally, terrain or available easements prevent inches for primary cables, and 36 to 42 inches a 12-inch radial separation. For these instances, for secondary cables, to accommodate all types the cooperative and a telephone or cable utility of farm machinery. This added depth helps to may agree to use random separation. NESC Secminimize dig-ins and future shallow cable issues tion 354 contains extensive special requirements resulting from change in grade as a result of for both electric and communication utilities farming practices. It also provides more room using random separation. However, separation for future cable installations, particularly in areas from other utilities must not be less than 12 where rights-of-way are narrow or congested inches. (See NESC Sections 353 and 354.) with multiple utilities. These minimum separations should provide enough space for either utility to work on its underground lines without damaging the other utilTABLE 8.1: Minimum Cover Requirements. Adapted from the 2007 ity’s lines. If this is not possible with a 12-inch NESC, Table 352-1. separation, then a greater separation is necesOperating Voltage Minimum Cover sary. The NESC also recommends 12 inches of vertical separation at crossings of different un0–150 volts phase-to-ground (streetlight cable ONLY)* 18 in. derground facilities. The cooperative and the 0–600 volts phase-to-phase 24 in. other utility can agree to a lesser separation if the following apply: 601–50,000 volts phase-to-phase 30 in.

Depth of Burial

50,001 volts and over phase-to-phase

42 in.

* Area or streetlight cables only if conflicts with other underground facilities exist

• There is no harmful interaction between systems, and


8 D

D

4” 4”

2”

2” W

W

UR2 (D × W) Trenching Unit One Cable or Cable Assembly

UR2–1 (D × W) Trenching Unit Multiple Power Cables Primary, Secondary, or Service

LEGEND Sand or Clean Soil Compacted Backfill Unless Otherwise Specified Undisturbed Earth

NOTES: 1. Depth (D) and width (W) are specified in description of units. 2. Depths specified are to finished grade. 3. Over-excavate trenches as necessary to allow for (a) sand bedding or (b) loose sandy soils or (c) where more than one cable will be installed in trench and laying first cable may cause trench damage and reduction in depth. 4. Sand bedding is not part of these units and will be specified as needed. 5. Backfilling is part of all trenching units, including joint-use trenches. 6. Optional warning tape is recommended to be placed above the installed cable.

D

P

4”

12” Minimum

T

2” W UR2–2 (D × W) Trenching Unit Power and Telephone Cable TRENCHES FOR DIRECT BURIAL CABLES

2000

FIGURE 8.3: Burial Depth Requirements. Adapted from RUS Bulletin 1728F-806.

UR2 TO

UR2–2

3 0 6 – Se c t io n 8

8 natural hazards are areas subject to erosion. Preferably, the cooperative engineer avoids these areas during the project layout phase. However, if cable must be installed in these Steam lines with only the required 12-inch areas, then the burial depth must be increased. separation can lead to thermal damage of the If the area has moderate to severe erosion, the underground power cable. Steam lines create cooperative engineer may consider supplemenhigh ambient earth temperatures that signifital protection, such as installation in Schedule 40 cantly decrease the ampacity rating of the cable. PVC conduit or encasement in concrete. To avoid these problems, the cooperative engiThe potential forces of man are also a factor neer must route power cable outside the effecto consider in the layout of the system. This is tive thermal range of a steam line. If adequate particularly true if the cable circuit is being inseparation is not feasible, a thermal barrier must stalled in the proximity of existing or future be placed between the facilities. water, sewer, or gas lines. Other utilities have the potential to not only cause substantial disOTHER CONSIDERATIONS ruption to electric service when they fail, but reAnother aspect to consider when choosing a pair of the other utilities will often require burial depth is grade change. The 2007 NESC excavation under emergency conditions. This in352.D.2.c requires the minimum cover requirecreases the chances for accidental dig-ins where ment to be met at the time of installation and at there is reduced separation. all times afterward. The best way to meet this Another consideration for placing cable requirement is to wait until final grade before indeeper is protection from random dig-ins. If the stalling any cable. The cooperative engineer area is congested with other underground utilimust also anticipate grade changes that occur afties, then the chances of cable ter final grade. For example, a damage by these other utilities developer will state that the increase. These chances insubdivision is at final grade Dig-ins are the crease even more when the before individual driveways major cause of faults power cable is installed before are constructed. If the cable is on the underground other utilities. This is particuplaced along the front proplarly true with secondary caerty lines, then it needs to be secondary system. bles, because their required buried deeper in anticipation burial depth is only 30 to 36 of an earth cut for driveway inches. An industry survey reand sidewalk construction. If ports that dig-ins are the major cause of faults the cooperative accommodates the developer by on the underground secondary system. Increasinstalling facilities before final grade is estabing the depth of burial, especially on secondary lished, the cooperative incurs a substantial addicable, can help reduce dig-ins. tional burden during construction and becomes A final consideration is clearance from underdependent on the developer to make final adground structures. According to the 2007 NESC justments to match plans. 351.C.2, underground power cable should not The forces of nature are also factors in deterbe installed directly under building or storage mining proper burial depth. In some areas, the tank foundations. If a cable must be placed befrost level reaches the cable burial depth. The neath a structure, the structure must have adeearth movement caused by frost formation can quate support to prevent a harmful load transfer move the buried cable and nearby objects. In to the cable. these areas, very close attention should be given All these requirements mean that the design to clean bedding material near the cable during engineer must be thoroughly familiar with site backfilling. This, along with firm tamping, will conditions and intended uses before establishing minimize the opportunity for the freeze-thaw a cable route and choosing an appropriate burial cycle to move the cable against stones. Other • Each utility can access its facilities without damaging the other’s facilities.


8 depth. Information on the route and depth must be communicated clearly to construction personnel. Any route changes required by field conditions must be clearly recorded by construction personnel so this information can be included in

Joint-Occupancy Trenches

permanent project maps. Failure of the designconstruction team to follow a proper route at a proper depth will likely increase the number of consumer outages and require future relocation of the underground lines.

interfere with access and be more susceptible to The NESC recognizes two types of joint trench: accidental damage by other deliberate separation and ranutility crews. dom separation. A deliberateThe NESC defines random separation joint trench reDeliberate-separation separation as any common quires a minimum of 12 inches trench arrangement in which of separation between the difjoint trenches require the cables have fewer than 12 ferent utilities. This separation a minimum separation inches of radial separation. can be horizontal or vertical of 12 inches. This type of joint trench is reand is illustrated in Figure 8.4. strictive; only certain utilities Maintaining the 12-inch sepacan place their facilities with ration allows electric utilities random separation. The NESC to share a trench with the folallows random separation of different electric lowing utilities: power cables. For example, the cooperative can place primary and secondary • Telephone, voltage cables in the same • CATV, trench at the same depth with • Gas (not generally Only certain utilities no horizontal separation. The recommended), NESC also allows the random • Water, and can share randomseparation of some power and • Sewer (not generally separation joint communication cables if cerrecommended). tain requirements are met. trenches. Table 8.2 summarizes the Trench sharing with storm types of power cables that can or sanitary sewers is generally be in random separation with not practical because of the telephone and cable television cables. The table size disparity in the facilities. In addition, sewer also lists the requirements that the cooperative is lines are often excavated for replacement or to responsible for according to the 2007 NESC. clear obstructions. In these cases, the cables will

3 0 8 – Se c t io n 8

8 W LEGEND Bedding Sand or Clean Soil Compacted Backfill Unless Otherwise Specified Undisturbed Earth D 4” 12” Minimum

S

W

T

2” UR2–3 (D × W) Service or Secondary and Telephone D

W P

S

12” Minimum

4” T

2” UR2–5 (D × W) Primary, Secondary, and Telephone

D S

12” Minimum

4” T NOTES:

12” P 2” UR2–4 (D × W) Primary and Secondary or Telephone

1. Depth (D) and width (W) are specified in description of units. 2. Depths specified are to finished grade. 3. Over-excavate trenches as necessary to allow for (a) sand bedding or (b) loose and sandy soils or (c) where more than one cable will be installed in trench and laying first cable may cause trench damage and reduction in depth. 4. Sand bedding is not part of these units and will be specified as needed. 5. Backfilling is part of all trenching units, including joint-use trenches.

TRENCHES FOR DIRECT BURIAL CABLES

2000

FIGURE 8.4: Joint Trench Use. Adapted from RUS Bulletin 1728F-806.

UR2–3 TO

UR2–5


8 TABLE 8.2: Requirements for Random-Lay Joint Trench. Adapted from 2007 NESC Section 354. Type of Power Cable 600-V Insulated Cable

Operating Voltage 240/120 V, 1Ø 240 V, 3Ø delta 208/120 V, 3Ø grounded-wye 480/277 V, 3Ø grounded-wye

Requirements None [NOTE: 480-V or 600-V, 3Ø, delta systems cannot be in random-lay with communication cables.]

15-, 25-, or 35- kV Bare Concentric Neutral Cable or Semiconducting Jacketed Cable

4,160/2,400 V, 1Ø or 3Ø 12,470/7,200 V, 1Ø or 3Ø 24,940/14,400 V, 1Ø or 3Ø 34,500/19,900 V, 1Ø or 3Ø

• Ground conductor must be in continuous contact with earth. Short sections of conduit for crossing under roads are allowed if neutral is continuous in conduit. Long sections of conduit require installation of a separate ground conductor that is in contact with the earth and close to the cable. • Ground conductor must be adequate to withstand available fault conditions. • When faulted, the cable will be promptly de-energized. • Ground conductor and communication cable shield or sheath must be bonded at 1,000-foot intervals (maximum spacing). • Concentric neutral must be corrosion-resistant material. • Semiconducting jacket must have a radial resistivity of 100 ohm-m or less.

15-, 25-, or 35- kV Jacketed Concentric Neutral Cable • Direct Buried or • Installation in Nonmetallic Conduit

4,160/2,400 V, 1Ø or 3Ø 12,470/7,200 V, 1Ø or 3Ø 24,940/14,400 V, 1Ø or 3Ø 34,500/19,900 V, 1Ø or 3Ø

• Copper concentric conductor must be effectively grounded. • Minimum conductance of concentric neutral must equal one-half conductance of phase conductor. • Ground conductor must be adequate to withstand available fault conditions. • Minimum of eight ground rods per mile, not including grounds at individual services. • Prompt de-energization of a faulted conductor. • Ground conductor and communication cable shield or sheath must be bonded at 1,000-foot intervals (maximum spacing).


1. Typical trench depths are 30 to 36 inches for secondary cable and 42 to 48 inches for primary cable (36 and 48 inches should be strongly considered in many rural areas). 2. The most common method for installing cable is trenching in suburban areas and plowing in rural areas. Selecting the appropriate equipment depends on soil type, trenching depth, terrain, and trenching distance. 3. For trenches in rocky soils, cable should be placed on a four-inch (minimum) bedding of select backfill, then covered with four inches of select backfill. Covering should continue with native clean backfill and compaction. Compaction should be made to 95 percent Proctor density where settlement control is important. Avoid mechanical compaction within six inches of a cable.

4. Cable plowing does not open a trench and eliminates the need to backfill. Pull plowing is suitable for installing flexible conduit or cable in conduit. Chute plowing should be used to install cable without conduit. 5. Some locations, such as existing subdivisions where cable replacement is considered, may require directional boring or horizontal directional drilling. 6. There are two types of joint trench. Deliberate-separation joint trench requires a minimum separation of 12 inches. Randomseparation joint trench is very restrictive. 7. Conduit should be used wherever additional cable protection is required or where the deferral of future excavation costs will justify the additional initial expense.


Conduit System De s i g n – 3 1 1

9 In This Section:

Conduit System Design

Conduit System Design and Installation


Cable Pulling

Conduit System Design

In some situations, conduit systems may offer many substantial advantages to electric utilities. Although conduit-enclosed cable installations have a higher initial cost, the lifetime advantages may make them the preferable installation in a variety of circumstances. Major advantages of conduit installation include the following: • Improved cable protection from dig-ins, • Ability to add cables along the route without additional excavation, • Ability to replace cable without excavation, and • Better use of utility easement for multiple circuits. The disadvantages when compared with direct burial are the following: • Higher initial cost, and • Lower ampacity for a given cable size. CONDUIT SYSTEM TYPES Distribution system conduit installations generally fall into one of three categories: 1. Direct buried, 2. Concrete encased, and 3. Concrete encased with manholes (or splice boxes).

Direct-buried conduit is simply installed in a trench and conventional backfill techniques are used. This approach generally requires a stronger conduit (Schedule 40 or better), but it has the lowest initial cost of all conduit systems. Typical direct-buried conduit installations are street crossings, single-circuit runs, and exits for small substations. Direct-buried conduit is particularly suitable where only minimal mechanical protection is needed and low cost is important. The next level in the hierarchy of conduit systems is a concrete-encased duct bank. This is generally used where multiple circuits are installed, or will be installed, along a route. Encased conduit is also advisable where additional mechanical protection from dig-ins is needed. As with direct-buried conduits, the length of runs is limited by cable-pulling criteria. However, in loose soils, the encased conduit will be more stable under high cable-pulling tensions. In fact, encased conduit is advised for longer runs, particularly if bends are involved. Where conduit runs are longer than allowable cable-pulling lengths, or where access for laterals or taps is needed, manholes and/or splice boxes must be installed for access and splicing. These allow intermediate pull points that will lead to lower cable-pulling tensions. Manholes and/or splice boxes can also be strategically located to eliminate bends or sharp angles in conduit runs.

3 1 2 – Se c t io n 9

9 Judicious location of manholes will also lower conductor in its own steel conduit, without a pulling tensions and yield a more convenient inneutral return) passing through a steel conduit stallation. Longer duct runs may also need to be will produce heat as a result of eddy currents in tapped for intermediate service locations. Where the conduit. When this condition is encountered, such service points are either a present or a funonmagnetic conduits must be used. ture requirement, a manhole installation will Metallic conduit other than galvanized Schedsimplify access to the cable circuits and produce ule 40 (or Schedule 80) should never be used on a more flexible system. a utility system. Types not recommended include Typically, manholes are full-size, below-grade electrical metallic tubing (EMT) and intermediate enclosures that allow personnel to enter and metal conduit. Such conduits are lighter in work within, standing erect, usually with six to weight, have less secure couplings, and are eight feet of head room. Splice boxes are usuoften more susceptible to corrosion. The only ally surface mounted, not intended for persongeneral exception is Schedule 40 aluminum connel entrance, and usually only three to five feet duit, which might be used in some locations deep. Much smaller in size, splice boxes are where atmospheric corrosion is a concern and worked from the surface and earth contact can be avoided. generally involve only one to Where corrosive conditions two feeders. exist, Type 304L or Type 316 Metallic conduit In summary, conduit sysstainless steel conduits might tems are recommended wherbe considered. other than galvanized ever additional cable Plastic conduits are now Schedule 40 (or protection is required or the available to utilities in a wide Schedule 80) should deferral of future excavation variety of sizes and materials. costs will justify the additional The predominant material is not generally be used initial expense. Manhole/duct PVC and will be used in all on a utility system. bank systems are particularly examples. Other nonmetallic advantageous where long, materials include acrylonitrilecontinuous runs of multiple butadiene-styrene (ABS) plascircuit underground are expected over the life of tic, HDPE, and fiberglass-reinforced epoxy (FRE). the project. ABS conduit is similar to PVC, but it has higher wall thickness to compensate for a lower material strength. In addition, the solvent cement CONDUIT TYPES welds on ABS and PVC are chemically different, Over the years, the electric utility industry has so neither will give satisfactory results on the used a wide variety of conduit types. Today, the other conduit. Therefore, ABS and PVC should main conduits used on UD systems are steel and never be mixed on a project. It is also advisable plastic (predominantly PVC and HDPE). Each of that a cooperative not mix these conduit materithese materials is offered in several configurations. als on its system as different solvent cements are The steel conduit that utilities use almost exrequired and defective joints will be produced if clusively is galvanized Schedule 40. This type is the products are accidentally interchanged. mostly used where extra mechanical protection HDPE conduit is much more flexible than PVC is needed. Examples are riser poles or some diconduit and generally comes on reels, rather rect-buried conduit applications. Although steel than precut straight lengths. HDPE conduit is conduit generally provides better protection particularly useful in directional bored applicathan does a similar-size PVC conduit during a tions, where the flexible nature of the conduit is dig-in, steel does have its disadvantages. It is an advantage. The conduit comes in smooth harder to bend, susceptible to corrosion, and wall, ribbed wall, and corrugated wall, dependoften more expensive. Furthermore, galvanized ing on the application, and also comes in multisteel conduit is magnetic. This means that heavy ple colors and markings. Reels can be provided unbalanced currents (such as a single-phase

Conduit System De s i gn – 3 1 3

9 with continuous lengths up to 4,000 feet, (e.g., required special skills and care during installatwo-inch inside diameter [I.D.]. The one caution, tion. With the advent of economical PVC and and possibly only drawback, to HDPE conduit is concerns about asbestos content, asbestos cethat it tends to have a coiled “memory” and canment conduit is no longer installed. not be allowed to curve in the trench floor bePlastic conduit is the most commonly used fore compaction. Eliminating these curves can electrical duct material. Therefore, engineering be very difficult when the coiled HDPE conduit and construction personnel need a working is being installed in an open trench. If the conknowledge of the plastic conduit types comduit is not installed in a controlled, straight line, monly used on electric systems. the additional curved bends, while seemingly inTable 9.1 gives the specific classifications of significant during conduit installation, greatly plastic conduit. amplify cable-pulling tensions. One important determination for a conduit FRE conduit is a specialty item generally used application is whether it will be used strictly unin applications more typically associated with derground or if it may have above-ground instalsteel conduit because FRE conduit can have lations. As solar radiation affects most plastics, higher strength than even Schedule 80 PVC. It only those conduits classified for above-ground can also be installed with longer unsupported use may be applied in sunlight. Classifications of spans without excessive long-term sag. FRE condirect burial (DB) and encased burial (EB) mean duit, like PVC, has high corrosion resistance. that all above-ground exposure must be These characteristics make FRE conduit particuavoided. Conduit types classified “above ground” larly attractive for duct lines suspended beneath may be used in either location. bridges and for riser installations. Table 9.2 shows the dimensions of these variThroughout the history of electric utilities, a ous conduit configurations. Table 9.3 compares variety of other conduit matethe relative strengths of these rials have been used. These conduit types in the four-inch range from early treated wood nominal size. Table 9.4 lists Never use DB or conduits to terra cotta tile the impact strength of the variducts. Some of the other mateous sizes of PVC conduit. For EB conduit rials most commonly used similar information on FRE above ground. since the 1950s include fiber, conduit, refer to the specific concrete, and asbestos ceFRE manufacturer. ment. Fiber duct, often known by the trade name Orangeburg, was made of Direct-Buried Conduit Design and Installation molded wood fiber impregnated with an asThe most common type of conduit application is phaltic compound. Fiber duct was used mainly the direct-buried system. Here, the conduit conin concrete encasement but, even then, it would taining the electrical cable is placed into the eventually absorb moisture and deteriorate. Conground without additional encasement. When crete duct was generally installed as a directthe installation is made by trenching, the conduit buried conduit or multiple-tile duct. This is placed on a smooth trench bottom before material was naturally very heavy and did not backfill is placed. When the trench is being pregain wide acceptance with electric utilities. pared, the bottom must be leveled to provide Asbestos cement duct, otherwise known as even support to the conduit. Rock outcroppings Transite™, gained wide acceptance in concretemust be cushioned with a layer of clean, comencased conduit duct banks. It also saw use as pacted fill to avoid high-pressure points on the a direct-buried conduit. Major advantages inconduit when backfill is placed. cluded a smooth interior surface and very high The initial backfill layer should be tamped on flame resistance. However, this material was exthe sides of the conduit to develop sidewall suptremely hard and brittle. It was also extremely port. This support is important to provide stabilinflexible. The combination of these characteristics ity during the pulling process and to resist

3 1 4 – Se c t io n 9

9 TABLE 9.1: Classifications of Plastic Conduit. Specification Conduit Designation

Material

NEMA

Other

Description and Application

EB-20

PVC

TC-6

ASTM F512, UL651A

Encased burial

EB-35

PVC

TC-8

ASTM F512

Encased burial, extra strength

DB-60

PVC

TC-6

ASTM F512

Direct burial

DB-120

PVC

TC-8

ASTM F512

Direct burial, extra strength

Schedule 40, Type II

PE

—

—

Schedule 40, Type III

PVC

TC-2

UL 651

Normal duty, above ground

Schedule 80, Type IV

PVC

TC-2

UL 651

Heavy duty, above ground

HDPE, Smooth-Wall

HDPE

TC-7

ASTM D3035 ASTM D2239 ASTM D2160

Normal duty, direct burial

HDPE, Ribbed

HDPE

TC-7



HDPE, Corrugated

HDPE

TC-7




TABLE 9.2: PVC Duct Dimensions—Minimum Wall Thickness. TC-6 EB-20 DB-60

TC-8 EB-35 DB-120 Schedule 40 Schedule 80

2”

0.060

0.060

0.060

0.077

0.154

0.218

0.154

3”

0.061

0.092

0.076

0.118

0.216

0.300

0.216

4”

0.082

0.121

0.100

0.154

0.237

0.337

0.237

5”

0.103

0.152

0.126

0.191

0.258

0.375

0.258

6”

0.125

0.182

0.152

0.227

0.280

0.432

0.280

Conduit Size Minimum Inside Diameter

HDPE-40

TABLE 9.3: Comparison of Characteristics for Four-Inch Diameter PVC Duct.

Characteristics Collapse pressure, psi

TC-6 PVC EB-20 DB-60

TC-8 PVC EB-35 DB-120 Schedule 40 Schedule 80

HDPE-40

6.7

17.0

9.2

36.6

108.7

326.5

96.0

Impact resistance, ft-lb.

25.0

60.0

40.0

80.0

220.0

310.0

N/A

Weight, lb./100 feet

92.0

127.0

109.0

158.0

234.0

310.0

136.2

Pipe stiffness, lb./in.*

20.0

60.0

35.0

120.0

461.2

117.8

N/A

* Specifically, pounds per inch deflection at five percent change in internal diameter


9 TABLE 9.4: PVC Duct—Impact Strength (Foot-Pounds). TC-6

TC-8


EB-20

DB-60

EB-35

DB-120

2”

20

20

20

25

190

300

3”

20

40

30

50

220

525

4”

25

60

40

80

220

525

5”

30

85

55

110

220

525

6”

40

120

75

150

220

525

Schedule 40 Schedule 80

be accomplished successfully in moderate temcrushing when full vertical backfill pressure is peratures with proper equipment. With conduit applied later. Crushing forces can also be resizes larger than two inches or duced by not tamping above during cool weather, make the conduit until an adequate sure the conduit installation is thickness of backfill has been Curves in coilable straight and does not have placed. The thickness of backbends caused by conduit fill required will depend on conduit can greatly “memory” as it is removed the force applied by the tamp, increase cable-pulling from the reel. Curves caused the surface area of the tamp, by this phenomenon can drasand the soil characteristics. tensions. tically increase cable-pulling Type DB conduit is more sustensions. This condition should ceptible to crushing than also be avoided if a coilable Schedule 40 (or 80) conduit conduit, with or without a cable, is being inbecause its wall is thinner. If rocky soils are usustalled by conventional trenching. ally present or clean backfill cannot be ensured, In some cases, short runs of straight, jointed Schedule 40 conduit should be used instead of conduit sections may be installed by the pull-in Type DB for direct-buried installations. Failure plow method. In these cases, it is extremely imto follow these guidelines will lead to conduit portant that all joints are properly made and with blockages or reduced inside dimensions. In cured before pulling begins; otherwise, the these cases, the conduit will be unusable or may joints can separate. damage cable during installation. If multiple conduits for electric circuits are being installed in the same trench, a minimum Concrete-Encased Duct Design and Installation of three inches of clearance must be provided Concrete-encased duct banks are generally used between the conduits. This clearance not only where multiple circuits are required along conwill allow for proper backfill placement and gested routes or where extra physical protection tamping but also will improve heat dissipation. for cables is warranted. Installations of this type If the decision is made to use close spacing of require careful site investigation and advance conduits for individual enclosure of large, highplanning, particularly because of the size of the capacity cables (this is not recommended), these duct line and the need to keep it straight and should be Class DB-120 or Schedule 40 to withproperly graded. Unexpected conflicts with unstand the point pressures created by conduit-toderground obstructions can cause major problems conduit contact. as a large multiple-conduit, concrete-encased Where conduit is installed by plowing, a coilduct line is being installed. able polyethylene product will usually be used. This type of installation begins with an open Plowing of coilable polyethylene conduit may trench. The trench must be wide enough to

3 1 6 – Se c t io n 9

9 2" Min.

3" Min.

3-Way Direct-Buried

3-Way Concrete-Encased



* Designates Circuit Location Undesirable for Loaded Circuit


FIGURE 9.1: Typical Duct Configurations.

provide proper duct spacing and side clearances. However, the trench must not be made too wide, as the trench wall will generally be used as a form for the concrete encasement. Extra trench width will lead to the need for unnecessary concrete. Trench design should also recognize the desirability of well-drained conduits. This requires sloping of the conduit to a point, such as a manhole, where water can be removed from the system. Each section of conduit can be sloped in a single direction or it can be drained toward each end. While the overall slope of the trench is important, the conduit must be graded so there are no local pockets that can accumulate water.

Establishing a well-drained conduit system will facilitate cable installation and removal as well as improving the cable operating environment. The trench bottom should be adequately compacted where conduit support spacers will be installed. Loose material in the trench bottom should be removed or compacted so the duct bank will have proper support at all points. For electrical distribution duct lines, three inches of concrete cover should be provided at both the top and bottom extremities of the bank. A minimum of two inches of horizontal concrete cover should be provided between the outside ducts and the trench wall. For proper heat dissipation, three inches of clearance should be provided between conduits (see Figure 9.1). A simple way to maintain these dimensions both vertically and horizontally is by using conduit spacers. Concrete installed around duct banks must be properly placed to fill all voids, provide optimum heat transfer, and properly protect cables. Concrete that has small aggregate, generally one-half-inch or less, which will readily flow between ducts must be used. A concrete slump of seven to eight inches must also be specified to allow reasonable flow. The slump value is a measure of how much fluid is in the concrete. A higher slump value means the concrete contains more fluid and is “wetter.” A slump value higher than eight inches is not recommended as the high fluid content will make it more difficult to hold the conduits in place during the pouring process. Concrete strength should be specified in the range of 1,500 to 2,500 psi. Although standard ready-mix concretes are often delivered with strengths of 3,000 to 4,500 psi, there is no need for the stronger and more expensive mix in unreinforced duct bank encasements. The concrete supplier should be consulted beforehand in order to get the proper concrete mix at the job site. He can then design an economical mix that will meet the special needs of duct bank construction. It is also vitally important that vibration be used during the pouring process. Vibration will facilitate the flow of concrete and minimize voids. During the actual pouring process, three important requirements should always be verified:


9 • Hold-down of the duct and spacers, • Control of the concrete flow, and • Prevention of duct collapse. Precast concrete weights, often called “suitcases,” should be applied to the top of the conduit before pouring. These weights will keep the conduit from floating when it is surrounded with wet concrete. As noted above, the slump value should be fewer than eight inches to provide friction that will help keep conduits in place. The amount of weight required will depend on the following factors: • • • • •

Number of conduits, Size of conduits, Slump of concrete, Use of vibration, and Other anchoring methods used.

The amount of weight needed to keep conduit from floating is determined by the buoyancy of the empty conduits in wet concrete, given that the concrete has a unit weight of about 150 lb. per cubic foot. To counteract the maximum possible buoyancy of a six-way, five-inch duct bank, a 150-lb.weight is needed on each foot of duct line while the concrete is being poured. This example shows how important it is to keep concrete slump as low as practicable and to use only as much vibration as needed to achieve good concrete flow. As an alternative to the precast concrete suitcases, some manufacturers of duct spacers have provisions for driving hold-down rods through the spacers to avoid conduit floating. This method, coupled with nonmetallic bands or straps, can also be used to prevent conduit floating. Caution should be used with hold-down rods in soft soil and larger duct bank configurations. Softer soils offer lower pullout resistance, thus requiring longer hold-down rods driven deeper. When pouring concrete encasement, the distance that the wet concrete falls into the trench should be minimized. If the ready-mix delivery truck chute cannot be placed near the top of the ducts, a hopper (funnel) and hose arrangement can be attached to minimize the free-fall distance. If the fall distance is too great, there will be two adverse results:

1. The aggregate will tend to segregate out of the concrete mix, thereby producing porous (honeycomb) sections in the encasements, and 2. An excessive free-fall distance may disrupt the conduit configuration or break joints. The use of a splash board will also help direct the concrete flow into the trench and should be used to prevent concrete flow against unsupported trench walls. Otherwise, loose dirt will be embedded in the wet concrete, causing voids. All the above conditions must be avoided for a satisfactory duct installation. When large duct banks (above six-way) are being installed, the collapse pressure rating of the particular conduit should be computed with the expected compressive force on the bottom conduit during pouring. To make this calculation, the depth of the lowest conduit in feet is multiplied by 1.03 psi/ft of depth to find the compressive force (in psi) on the lowest conduit. For instance, if a three-layer duct bank has a layer of five-inch conduit with three inches of top cover and three inches between layers, the lowest conduit is about 25 inches (2.1 feet) deep. This produces a compressive force of approximately 2.2 psi. Table 9.5 shows that the compressive strength of five-inch type EB-20 conduit is 5.9 psi. Therefore, this application of EB-20 is satisfactory. If the compressive force had approached the conduit rating of 5.9 psi, a higher grade of conduit such as EB-35 would be required. Crews must be instructed that conduits, particularly the thinner-wall Type EB varieties, should not be walked on to avoid cracking, collapse, or deformation of the conduit. GENERAL CONDUIT SYSTEM LAYOUT The first problem in engineering an underground conduit system is determining the loads to be served by the system, both now and in the future. Present and future requirements must also be determined for circuits traversing the design area to serve loads in other areas. These requirements, coupled with the characteristics of the design area, will determine the type of underground system to be installed.

3 1 8 – Se c t io n 9

9 TABLE 9.5: PVC Duct Collapse Pressure (PSI). TC-6

TC-8


EB-20

DB-60

EB-35

DB-120

2"

11.2

11.2

11.2

26.6

117.1

595.8

3"

6.6

15.2

8.2

34.0

181.3

487.3

4"

6.7

17.0

9.2

36.6

108.7

326.5

5"

5.9

18.9

10.3

38.2

75.5

235.6

6"

6.1

19.6

11.2

38.0

57.0

212.5

For example, if the problem is substation exit circuits in an open rural area with a wide area for circuit exits and good soil conditions, the obvious answer may be direct-buried circuits. These have lower initial cost and better thermal performance than either direct-buried conduit or concrete-encased conduit systems. However, if only a limited space is available for installing electric circuits, construction conditions are difficult, or several circuit additions are expected over the life of the facility, a conduit system is probably the proper answer for lowest long-term costs. Of course, cable installation forces and damage probability must also be considered when choosing the final conduit configuration. Regardless of whether direct-buried conduit or encased conduit is chosen, the total system must be designed in light of present and future loads and the circuits required to serve these loads. Therefore, the first step is to define loads within the design area and determine the transformer locations required to provide service. Then distribution circuits will be designed to provide primary voltage to all transformers, generally with an open-loop configuration. If a transformer location will provide service through radial secondary circuits, these secondary circuits must be designed. This situation is often encountered in congested areas such as shopping centers. After all circuits are designed for local service, any underground circuits that will pass through the design area should be considered. In the case of sites near substations, the through-circuits may be the only factor considered. After the circuit design is complete, the conduit system configuration should be designed to

Schedule 40 Schedule 80

accommodate the circuits in each location. Standard conduit configurations as shown in Figure 9.1 should be used to simplify construction. Duct bank configurations of greater than three conduits are generally installed using concrete encasement. Encased duct will also be needed where cable-pulling tensions are high or bends are in the conduit. Inner conduits should not be used for heavily loaded cable circuits as heat dissipation is much better for peripheral locations. Section 4 discusses further the thermal performance of cables in a conduit system. If both primary and secondary circuits are located in the same duct run, it is generally preferable to plan on secondary circuits being located in the upper ducts, particularly if the secondaries are serving a load at an intermediate point in the duct run. Turning of the upper ducts is simpler and allows the lower conduits to continue straight with the main conduit run. Other considerations are the greater mechanical protection afforded lower conduits and the simplified manhole internal arrangement. See Figure 9.2. Small features of duct bank design that are often overlooked are provision for area lighting circuits and future electric utility communication circuits. Both of these uses generally require small conduits (two-inch minimum diameter is recommended) and are easy to initially install. Street and area lighting conduits should be located in the upper corners of the duct bank (see Figure 9.3). Lighting conduits should be looped to facilitate multiple light locations to be served by a single circuit. Communication conduits are for the installation of circuits owned and maintained by the


9 To Area Lighting F

F

F

Manhole

A

B

E

Secondary Duct

Transformer Location

D

Secondary Duct

C

D

D

Manhole

Primary Duct

B E

F

C

Primary & Secondary Ducts

A

cooperative only. Such circuits might be used for alarm, control, or metering associated with electric distribution. Section 32 of the NESC has information on the allowable location of communication ducts and circuits. If other conditions allow, the preferred location of communication circuits is the top center conduit. Making provisions for these circuits is recommended, especially in areas where concrete-encased ducts are installed or high load density exists.

To Area Lighting

D

To Remote Loads

To Local Loads

FIGURE 9.2: Typical Duct Line and Manhole Arrangement.

7

8

9

4

5

6

4

5

6

4

5

6

1

2

3

1

2

3

1

2

3

Section A-A

Section B-B

Section C-C

Duct Size Use 1 5" Primary Loop 2 5" Spare 3 5" Primary Loop 4 2" Area Lighting 5 5" Spare 6 2" Utility Communications

Duct Size Use 1 5" Primary Loop 2 5" Spare 3 5" Primary Loop 4 5" Secondary 5 5" Secondary 6 5" Secondary 7 2" Area Lighting 8 2" Utility Communications 9 2" Area Lighting

Duct Size Use 1 5" Primary Loop 2 5" Spare 3 5" Primary Loop 4 2" Area Lighting 5 2" Utility Communications 6 2" Area Lighting

1

2

Section D-D Duct Size Use 1 5" Secondary 2 5" Secondary 3 5" Secondary

3

4

5

6

1

2

3

Section E-E Duct Size Use 1 5" Primary Loop 2 5" Spare 3 5" Primary Loop 4 5" Secondary 5 5" Secondary 6 5" Secondary

1

2

Section F-F Duct Size Use 1 2" Area Lighting 2 2" Area Lighting

FIGURE 9.3: Typical Arrangements for System in Figure 9.2.

DETERMINATION OF CONDUIT SIZES FOR CABLE INSTALLATION An important step in analyzing the duct bank system is to examine the size and number of cables required and to select the appropriate conduit size. For the analysis of conduit fill, the National Electrical Code (NEC) is an excellent source that has been tried and tested countless times. Although the NEC does not legally bind cooperatives, it is still an excellent standard and application guide on conduit fill. Tables 9.12 through 9.15 list the minimum size of conduit necessary to accommodate certain numbers and sizes of underground power and secondary cables. The tables are based on the maximum fill requirements of the NEC, which are 53 percent maximum fill for one cable in a conduit, 31 percent maximum fill for two cables in a conduit, and 40 percent maximum fill for three or more cables in a conduit. The trade sizes, inside diameters, and maximum areas of fill for various sizes of conduit are shown in Table 9.6. The cables shown in Tables 9.7 through 9.11 all have ICEA Class B concentric stranded conductors, unless an “S” indicating solid conductor appears beside the conductor AWG size. The listed cables have standard thickness of conductor shield, insulation, insulation shield, and jacket, and standard numbers and size of concentric neutral wires in accordance with ICEA specifications. If other than standard specifications are used for cable, or if other than stranded or solid conductor is used, the overall cross-sectional area of 15-, 25-, or 34.5-kV power cable can be calculated with Equation 9.1. For 600-volt secondary cable, the overall cross-sectional area can be calculated with Equation 9.2.

3 2 0 – Se c t io n 9

9 TABLE 9.6: Conduit Fill.

Trade Size (in.)

Inside Diameter (in.)

Area (sq. in.)

1 Cable Area x 53% (sq. in.)

2 Cables Area x 31% (sq. in.)

3 Cables Area x 40% (sq. in.)

2

2.067

3.36

1.78

1.04

1.34

2 1/2

2.469

4.79

2.54

1.48

1.92

3

3.068

7.39

3.92

2.29

2.96

3 1/2

3.548

9.89

5.24

3.06

3.95

4

4.026

12.73

6.75

3.95

5.09

5

5.047

20.01

10.60

6.20

8.00

6

6.065

28.89

15.31

8.96

11.56

TABLE 9.7: Conductor Shield Thickness.

Equation 9.1: Shielded Concentric Neutral Cable Diameters.

Conductor Size (AWG or MCM)

Diameter = C + 2CS + A + 2I + 0.030 + 2IS + 2N + 2J where: C = Diameter of the conductor CS = Thickness of the conductor shield (see Table 9.7) A = Addition Factor: • 0.010 inches for 25-kV and 34.5-kV cables with conductor larger than #4/0 • 0.000 inches for all other cable constructions I = Insulation wall thickness IS = Insulation shield thickness (See Table 9.8) N = Thickness of concentric neutral wires J = Thickness of outer jacket: • 0.080 inches for conductors through 1.50 inches over the concentric neutral • 0.120 inches for conductors larger than 1.50 inches over the concentric neutral

Conductor Shield (in.)

#8–#4/0

0.012

25–550

0.016

551–1,000

0.020

1,001 and larger

0.024

TABLE 9.8: Insulation Shield Thickness. Diameter Over Insulation*

Insulation Shield (in.)

0–1.000

0.060

1.001–1.500

0.075

1.501–2.000

0.090

2.001 and greater

0.105

* Diameter over insulation = C + 2CS + A + 2I

TABLE 9.9: Concentric Neutral Thickness—Aluminum Cables. Full Neutral

1/3 Neutral

ALUMINUM Conductor (AWG or MCM)

Neutral Wire Size (AWG)

Thickness (in.)

Through #1/0

Through 350

#14

0.0641

#2/0, #3/0

500–750

#12

0.0808

#4/0, 250

1,000

#10

0.1019

350

1,250–1,500

#9

0.1144

N/A

Over 1,500

#8

0.1285

N/A = not applicable


9 TABLE 9.10: Concentric Neutral Thickness-Copper Cables. Full Neutral

1/3 Neutral

COPPER Conductor (AWG or MCM)

Neutral Wire Size (AWG)

Thickness (in.)

Through #2/0

Through 250

#14

0.0641

#1–1/0

350

#12

0.0808

2/0–3/0

500–650

#10

0.1019

4/0

750

#9

0.1144

N/A

1,000–2,000

#8

0.1285

Equation 9.2: Unshielded Cable Diameter. Diameter = C + 2I where: C = Diameter of the conductor I = Insulation wall thickness

TABLE 9.11: Secondary Cable Insulation Thickness. Insulation Thickness (in.) Conductor Size (AWG or MCM)

Regular*

Ruggedized*

#4–#2

0.060

0.075

#1–#4/0

0.080

0.100

225–500

0.095

0.130

600–1,000

0.110

0.145

* Regular insulation consists of one layer of low-density polyethylene. Ruggedized design consists of two layers of equal thickness bonded together: an inner layer of low-density polyethylene and an outer layer of high-density polyethylene. Various manufacturers use different combinations of layers and layer thickness to achieve ruggedized designs. Verify actual diameters with actual cables being used.

CONDUIT/CABLE TABLES Tables 9.12 through 9.15 have been developed on the basis of the aforementioned requirements of the current 2002 NEC and Equations 9.1 and 9.2 for determining the diameters of required conduits for typical primary and secondary conductors. The sizes of conduit recommended are suggested based on an average of many manufacturers of cables, reflecting the fact that actual diameters of cables vary greatly.

Selection of the appropriate conduit size must also consider the following factors: • • • • •

Future cable capacity increases, Pulling tension considerations, Heating and ventilation (see Section 4), Cost considerations, and Standardization of conduit sizes.

3 2 2 – Se c t io n 9

9 TABLE 9.12: 220-Mil Primary Cable. Minimum Size of Conduit Necessary to Accommodate Primary Underground Power Cable: 15-kV Cable – 220-Mil Insulation Wall, Concentric Neutral Construction Minimum Conduit Size (Inches) for Numbers of Primary Cables, Based on Neutral Construction Conductor

1 Cable per Conduit

2 Cables per Conduit


AWG or MCM

Full

1/3

Full

1/3

Full

1/3

2S*

2

2

3

3

3 1/2

3 1/2

2

2

2

3

3

3 1/2

3 1/2

1S*

2

2

3

3

3 1/2

3 1/2

1

2

2

3 1/2

3

3 1/2

3 1/2

1/0S*

2

2

3 1/2

3

3 1/2

3 1/2

1/0

2

2

3 1/2

3 1/2

3 1/2

3 1/2

2/0

2

2

3 1/2

3 1/2

4

4

3/0

2

2

4

3 1/2

4

4

4/0

2 1/2

2

4

4

5

4

250

2 1/2

2 1/2

5

4

5

5

350

2 1/2

2 1/2

5

5

5

5

500

3

5

6

750

3

6

6

3 1/2

6

6

1,000

* S = Solid Conductor Note. Table 9.12 is based on NEC requirements. Maximum conduit fill is 53 percent for one cable, 31 percent for two cables, and 40 percent for three cables in a conduit. Unless noted, conductors are concentric stranded. If different conductors, such as compressed or compacted, are used, see Equation 9.1 for method of calculating. Outside diameters are based on ICEA Publication ANSI/ICEA S-94-649-2000.


9 TABLE 9.13: 260-Mil Primary Cable. Minimum Size of Conduit Necessary to Accommodate Primary Underground Power Cable: 25-kV Cable—260-Mil Insulation Wall, Concentric Neutral Construction Minimum Conduit Size (Inches) for Numbers of Primary Cables, Based on Neutral Construction Conductor

1 Cable per Conduit



AWG or MCM

Full

1/3

Full

1/3

Full

1/3

1S*

2

2

3

3

3 1/2

3 1/2

1

2

2

3 1/2

3 1/2

3 1/2

3 1/2

1/0S*

2

2

3 1/2

3 1/2

3 1/2

3 1/2

1/0

2

2

3 1/2

3 1/2

3 1/2

3 1/2

2/0

2

2

3 1/2

3 1/2

4

3 1/2

3/0

2

2

4

3 1/2

4

4

4/0

2 1/2

2

4

4

5

4

250

2 1/2

2

4

4

5

5

350

2 1/2

2 1/2

5

5

5

5

500

3

5

5

750

3

6

6

3 1/2

6

#

1,000

* S = Solid Conductor # Indicates that a 6-inch conduit is not of sufficient size to accommodate three cables of this size without exceeding the maximum fill requirement. Note. Table 9.13 is based on NEC requirements. Maximum conduit fill is 53 percent for one cable, 31 percent for two cables, and 40 percent for three cables in a conduit. Unless noted, conductors are concentric stranded. If different conductors, such as compressed or compacted, are used, see Equation 9.1 for method of calculating. Outside diameters are based on ICEA Publication ANSI/ICEA S-94-649-2000.

3 2 4 – Se c t io n 9

9 TABLE 9.14: 345-Mil Primary Cable. Minimum Size of Conduit Necessary to Accommodate Primary Underground Power Cable: 34.5-kV Cable—345-Mil Insulation Wall Minimum Conduit Size (Inches) for Numbers of Primary Cables, Based on Neutral Construction Conductor

1 Cable per Conduit



AWG or MCM

Full

1/3

Full

1/3

Full

1/3

1S*

2

2

4

4

4

4

1

2

2

4

4

4

4

1/0S*

2 1/2

2 1/2

4

4

5

5

1/0

2 1/2

2 1/2

4

4

5

4

2/0

2 1/2

2 1/2

5

5

5

5

3/0

2 1/2

2 1/2

5

5

5

5

4/0

2 1/2

2 1/2

5

5

5

5

250

3

2 1/2

5

5

6

5

350

3

3

6

5

6

6

500

3

6

6

750

3 1/2

6

#

1,000

3 1/2

#

#

* S = Solid Conductor # Indicates that a 6-inch conduit is not of sufficient size to accommodate two (or three) cables of this size without exceeding the maximum fill requirement. Note: Table 9.14 is based on NEC requirements. Maximum conduit fill is 53 percent for one cable, 31 percent for two cables, and 40 percent for three cables in a conduit. Unless noted, conductors are concentric stranded. If different conductors, such as compressed or compacted, are used, see Equation 9.1 for method of calculating. Outside diameters are based on ICEA Publication ANSI/ICEA S-94-649-2000.

MANHOLE TYPES In the design or selection of manholes to use with a conduit system, there are many factors to consider. The enclosure must be structurally adequate to withstand the loads in the selected location. The manhole must also provide reasonable access to the conduit system with room to pull and splice cables. It should also provide space and facilities to properly mount cables and still allow access and working room. The location of manholes is one of the first things to carefully consider when designing the duct system. Previous statements have emphasized the location of manholes relative to load centers or based on cable-pulling limits. Accessibility and

safety of manhole location is also a major consideration. If possible, manhole entrances should be located outside the paved roadway to minimize the hazard to workers and inconvenience to the public while the manhole is open. The roadway within street intersections should be particularly avoided, if possible. However, if a duct line is installed in a roadway and there is an intersection with another duct line, doing so may be impossible. The location of existing water, sewer, storm drain, and communication lines will strongly influence the location of electric facilities and may force location within the roadway. See Figure 9.4 for examples of preferred manhole and duct line locations.


9 TABLE 9.15: Conduit Fill—Secondary Cable. Minimum Size of Conduit Necessary to Accommodate 600-Volt Secondary Underground Power Cablel Minimum Conduit Size (Inches) for Numbers of Secondary Cables, Based on Insulation Construction Conductor AWG or MCM

1 Cable per Conduit




Regular

Ruggedized

Regular

Ruggedized

Regular

Ruggedized

Regular

Ruggedized

4S*

2

2

2

2

2

2

2

2

4

2

2

2

2

2

2

2

2

2S*

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

1S*

2

2

2

2

2

2

2

2

1

2

2

2

2

2

2

2

2

1/0S*

2

2

2

2

2

2

2

2

1/0

2

2

2

2

2

2

2

2

2/0

2

2

2

2

2

2

2

2

3/0

2

2

2

2

2

2

2

2 1/2

4/0

2

2

2

2

2

2

2 1/2

2 1/2

250

2

2

2

2 1/2

2 1/2

2 1/2

2 1/2

3

350

2

2

2 1/2

2 1/2

2 1/2

3

3

3

500

2

2

3

3

3

3

3 1/2

3 1/2

* S = Solid Conductor Regular = Normal insulation Ruggedized = Ruggedized, two-layer insulation. See Table 9.11 for more information. Note. Table 9.15 is based on NEC requirements. Maximum conduit fill is 53 percent for one cable, 31 percent for two cables, and 40 percent for three cables in a conduit. Unless noted, conductors are concentric stranded. If different conductors, e.g., compressed or compacted, are used, see Equation 9.2 for method of calculating. Outside diameters are based on ICEA Publication ANSI/ICEA S-94-649-2000.

Access to the manhole is provided through a ring and cover assembly. Covers are usually round. Noncircular covers should not be used without due consideration of the fact that such covers can fall into the manhole. The minimum clear dimension for a cover is 26 inches. However, on electric manholes, the clear opening should be a minimum of 30 inches to not only ease access but also provide more room to maneuver cables during the pulling process. Covers of 36-inch diameter (or more) allow even more room; however, crews often find that the added weight of these units is a concern. In addition to having adequate cover size,

the chimney between the top of the manhole and the cover should have walls without protrusions that could injure personnel or damage cable. Manholes may be divided into three general categories. The first type is used for locations in straight conduit runs where access is desired mainly for cable pulling. The second type is for locations where duct lines may intersect at an angle near 45° or where a more narrow intersection manhole is needed. The third type is designed to accommodate the intersection of two major duct banks. These patterns are illustrated in Figure 9.5. Each has advantages in the location of cables.

3 2 6 – Se c t io n 9

Tap Ducts to Load or Transformer Location

Main Ducts

Sidewalk

Street

Sidewalk

9 Main Ducts

Preferred Duct and Manhole Location

Street

Main Ducts

Main Ducts

Undesirable Duct and Manhole Location Street

(a) Tap Manhole

(b) Straight-Line Manhole

Main Ducts

Main Ducts

Sidewalk

Street

Sidewalk

Main Ducts

Main Ducts

(c) Intersection Manhole

FIGURE 9.4: Preferred Location of Duct Lines in Roadways.

In the past, some electric utilities, particularly those in urban areas, located special switching and/or fusing equipment in manholes. However, this practice should be avoided because of the congestion and safety problems it causes. In addition, all switches used in manholes must be suitable for remote operation from outside the manhole. All these factors make installation of pad-mounted switchgear much more practical. Here, equipment is more accessible for maintenance and operation. Moreover, personnel are working at ground level with less restricted exit paths in case of equipment problems. MANHOLE/PULL BOX CONSTRUCTION Manholes and pull boxes must be designed to sustain all expected loads that may be imposed on the structure. The manhole or pull box must be capable of withstanding vertical and horizontal live loads, dead loads, equipment loads, impact loads, loads caused by water table or frost, shear, and bending moments. Careful consideration

FIGURE 9.5: Typical Manhole Configurations.

should be given to the location of the structure and loads that may be encountered. In roadway areas, heavy trucks may subject the structure to extreme live loads and forceful impacts. In industrial areas, large cranes may travel over or near the manhole and create extreme point loads when their outriggers are extended. Live load requirements should be increased by 30 percent to account for impact forces. The following publications should be referenced when load requirements are analyzed: • Federal Specification RR-F-621D; • AASHTO Standard Specifications for Highway Bridges, 1983; • National Electrical Safety Code, Section 32; and • RUS Bulletin 1753F-151. Another facet of manhole/vault design is allowance for uplift on the structure when the surrounding soil is saturated. If the manhole is well


9

FIGURE 9.6: Rectangular Manhole Construction Details.

3 2 8 – Se c t io n 9

9

FIGURE 9.7: Rectangular Manhole Installation Details.

drained, the net buoyancy should be calculated. This force will have to be balanced by the effective weight of the overburden and soil shear acting on the walls by saturated soil. Failure to take this into account could result in a manhole floating out of the ground when the soil becomes saturated. This is not only extremely detrimental to system reliability it is also quite disturbing to the public. Figures 9.6, 9.7, 9.8, and 9.9 represent two styles of manhole—rectangular and octagonal— with details indicating depth of install, personnel entrance duct interface, and other construction details. Typically, precast manhole manufacturers can supply these type of manholes and accessories that comply with industry standards.

Wall Thickness/Concrete Strength/ Reinforcing Steel When precast manholes and pull boxes are used, the manufacturer will select the proper concrete strength, rebar type/spacing, and wall thickness based on the loading requirements supplied by the purchaser. Most manufacturers use Grade 60 reinforcing steel and 4,500 psi concrete and design to current ASTM and ACI standards. The purchaser should require the manufacturer to furnish load certifications sealed by a licensed professional engineer. Design guidelines for site-built manholes can be found in Specifications and Drawings for Conduit and Manhole Construction, RUS Bulletin 1753F-151.


9

FIGURE 9.8: Octagonal Manhole Construction Details.

3 3 0 – Se c t io n 9

9

FIGURE 9.9: Octagonal Manhole Installation Details.

The purchaser should require the precast manhole manufacturer to provide grounding provisions so the reinforcing steel within the structure walls can be connected to the system neutral and grounding electrodes. Conduit Entrances/Knockouts Many precast manhole manufacturers provide knockout panels in each wall to accommodate a wide variety of conduit and duct bank sizes. Knockout panels are typically three inches thick, void of rebar, and allow for quick tie-in of duct banks and conduits.

Personnel/Equipment Entrances Access openings in manholes should be large enough for workers to enter the manhole on a ladder and to lower equipment needed for cable pulling, splicing, and testing. Manhole openings should be free of obstructions that would prevent the worker from safely and quickly exiting the manhole. Large manholes may have more than one entrance for convenience. Personnel access openings should not be located directly over cables or equipment. Manhole covers should be at least 30 inches in diameter and designed so they cannot fall into the


9 manhole and harm personnel or equipment. Personnel access openings should be located where safe access can be provided and outside of pedestrian traffic areas when possible. Manhole covers should have markings to identify the type of utility and ownership. Sufficient means through weight, design, or location should be employed to prevent access by the public and unqualified persons. Sump Pit/Drain Lines Manholes should have adequate drainage to keep them dry and free of standing water. In some locations, a sump pit is designed in the manhole floor. The sump pit should be excavated two feet below the manhole floor and filled with small stones to allow water to seep out of the manhole. In high water table areas, the sump pit in the floor would be ineffective. Of course, if the natural water table is higher than the manhole floor, there will be a flow of groundwater into the manhole with the potential for undermining of the structure. In some cases, a drain in the manhole floor may be connected to a local storm sewer if there is no chance of the storm sewer backing into the manhole during periods of high flow. However, the manhole drain must never be connected to a sanitary sewer since there is a danger of sewer gas entering the electric manhole. When a manhole drain is provided, the manhole floor should be sloped to direct all accumulated water to the drain. Also, the exterior walls of the manhole or pull box may need to be waterproofed in areas with a high water table to minimize seepage through the walls. In extreme cases, manholes in an area routinely subject to flooding may have automatic sump pumps to remove water. Pulling Irons and Pulling Eyes Where pulling appurtenances are furnished, they should be installed with a safety factor of two (2.0), based on the expected load. Pulling irons can be supplied in galvanized steel or plastic coated steel. Pulling irons are used for lifting the roof and floor panels. Pulling eyes and inserts are available for lifting wall panels. The precast manufacturer can size the pulling eyes and pulling irons on the basis of the load specifications supplied by the purchaser.

Joint Sealants Joints in precast manholes are typically designed to be self-aligning during the assembly process. Some precast manholes also have cast-in-concrete weld plates that, when welded together, prevent shifting of manhole sections and create a rigid assembly. Asphaltic butyl compounds are usually used in the joints to provide a water seal that is resistant to temperature changes, shock, shrinkage, and mild chemicals. Sealant should comply with Federal Specification SS-S-210A and AASHTO M-198B. Ring and Cover Assemblies—Specifications Ring and cover assemblies should comply with the loading requirements of Federal Specification RR-F-621D. When specifying ring and covers, the interchangeability of new manhole lids with existing manhole lids should be considered. If possible, a standard for lid diameter, thickness, lettering, and so on should be enforced. When future pavement overlays are expected, the vertical seat thickness of the manhole cover and available cast iron riser ring sizes must be considered. Bolted covers can eliminate manhole cover blow-off caused by rising waters in low lying areas. Manhole covers with watertight gaskets will keep surface water from flowing into the manhole. The utility name and the type of utility occupying the manhole should be specified on the manhole cover. See NESC 323J. Manhole Racks/Cable Supports Cable racks and supports must be installed in manholes to support cables at joints and keep cables from lying on the manhole floor. The NESC requires all cables to be at least three inches off the manhole floor. Usually cable racks are placed six inches from the ends of cable joints and every three feet around the manhole for general cable support. Cable racks are usually fastened to the manhole walls with expansion bolts, power-installed studs, or threaded inserts. Metal cable rack/hook assemblies are available with plastic coatings to reduce corrosion and come in a variety of sizes. Metal cable racks should be bonded to the system neutral for safety. Cable racks are also available in fiberglass or other nonconductive materials that avoid corrosion and grounding concerns. Cable racks should be in-

3 3 2 – Se c t io n 9

9 Cable Pulling

stalled with provisions to allow for cable expansion and contraction in long duct runs during load cycles. The use of sliding cradle insulators and generous radii on exit bends will prevent damage from abrasion during load cycles.

Waterproofing Electrical manholes may be “painted” on exterior walls with an asphalt-base, waterproof sealer to prevent water from seeping into the manhole. In addition, sealed covers, mentioned above, can be used to reduce runoff from the road surface.

Many utilities are encountering situations in which the best installation is cable installed in conduit. This type of installation may be a single, directburied conduit or a major, concrete-encased, multiple duct bank. Regardless of configuration, the cable must be installed in the conduit without incurring mechanical damage that will impair electrical performance or cable longevity. Therefore, analysis of the cable-pulling problem must be performed during the design process. The main limiting factors in cable pulling are tension and sidewall bearing pressure (SWBP). Tension must be limited to avoid overstressing the metallic central conductor of the cable. It is assumed that the central conductor carries all tensile forces and these forces must be kept well below the conductor yield point. The origin of cable tension is friction between the outer surface of a cable and the inner surface of the conduit. The force the cable exerts on the conduit wall and the coefficient of friction between the two surfaces governs the amount of friction. Sidewall bearing pressure is the force applied perpendicularly to the outer surface of the cable when it is being pulled through a bend or sheave. Excessive sidewall pressure will distort cable components, particularly the outer jacket and the insulation shield. In some cases, concentric neutral wires or the tape shield may be pushed into the semiconducting insulation shield. More extreme cases may damage the insulating layer or its semiconducting shield. All these conditions involve severe damage to the cable structure and decreased cable life. The overall cable-pulling problem may be considered as a combination of factors:

• Allowable cable tension.

• • • •

Cable-conduit friction, Cable weight, Conduit bends, Sidewall bearing pressure, and

Each of these will be discussed separately to develop a comprehensive analysis of the cablepulling problem. CABLE-CONDUIT FRICTION The most important factor in cable pulling is the friction that exists between the outer surface of the cable and the inner surface of the conduit. This friction force for a horizontal pull is classically expressed as shown in Equation 9.3.

Equation 9.3. T = W × WC × f × l where: T = Tension, in lb. W = Weight of cable, per unit of length, in lb. per ft WC = Weight correction factor (where required) f = Coefficient of friction l = Length of cable, in feet

This relationship is illustrated in Figure 9.10. Equation 9.3 shows that, if there is no friction between the cable and the conduit (f = 0), there is no tension in the cable as it is being pulled. Friction between the cable and conduit is a very complex phenomenon. Contributing factors include the following: • Surface roughness of both cable jacket and conduit, • Deformation characteristics of jacket and conduit materials as shear develops at interface, and • Other materials, either dirt or lubrication, present at interface.


9 calculate the force required to keep a cable moving in the conduit. Most pulling calculations emphasize the dynamic coefficient because the cable should be continually in motion during the pulling process. This condition is the one most commonly encountered. However, tensions resulting from the static coefficient should always be considered because unforeseen circumstances may stop a pull at any point and the cable must be restarted. The static coefficient of friction is used to calculate the tension required to initially move a cable that is in a conduit. Static friction coefficients are always higher than the dynamic coefficients for similar materials. The ratio of these values with conventional lubricants has been observed between 1.25 and 1.85. Tensions resulting from the static coefficient should always be considered because unforeseen circumstances may stop a pull at any point and the cable must be restarted. The presence or absence of other materials at the cable-conduit interface is also a major contributing factor to the tensions actually experienced during the pulling process. This factor can be either positive or negative. Most dirt is a strong negative factor because granules can

Direction of Pull

Inner Conduit Wall

Cable Weight (w)

T1

Tension T2 Friction (f)

Friction (f)

Friction (f)

T2 = T1 + (f × w)

FIGURE 9.10: Cable/Conduit Friction and Pulling Tension.

In light of the complex nature of cable-conduit friction, empirical tests are the only reasonable way to predict friction factors. The variability of these same factors also means that the apparent coefficient of friction experienced during pulling may also differ from values measured with the same materials under similar circumstances. However, the materials comprising the cable jacket and the conduit constitute the best factors for beginning characterization of the friction coefficient. Typical values of this coefficient are found in Table 9.16. Table 9.16 gives dynamic friction coefficients. The dynamic coefficient of friction is used to

TABLE 9.16: Recommended Dynamic Friction Coefficients for Straight Pulls and Bends Using Soap/Water or Polymer Lubricants. Soap/Water Lubricants Straight Pulls One Cable Three Cables Bends (SWBP @75°C @75°C > 150 lb./ft)

Static Friction

Dynamic Friction

0.15

0.14

0.11

0.45

0.15

0.14

0.11

0.50

0.60

0.30

0.17

0.12

Concentric Neutral

0.40

N/A

Not Recommended

N/A

N/A

XLPE

0.60

0.65

0.25

0.15

0.14

PE

0.50

0.55

0.25

0.11

0.11

PVC

0.65

0.70

0.30

0.15

0.15

Concentric Neutral

0.50

N/A

Not Recommended

N/A

N/A

Duct Material

Cable Jacket

PVC

XLPE

0.40

0.60

PE

0.40

PVC

Steel

Polymer Lubricant (SWBP = 200 lb./ft)

Note. Use straight-pull values for bends with sidewall bearing pressure (SWBP)<150 lb./foot. For other conduit or cable jackets, see AEIC G5-90; EPRI EL-3333, Vol. 2; or lubricant manufacturer test data. Polymer lubricant is American Polywater J. Others may vary. N/A = Not Applicable

3 3 4 – Se c t io n 9

9 partially embed themselves in lubricant works will depend either the cable jacket or the not only on a low internal Use the static conduit wall, or both. When shear value but also on how coefficient of friction this happens, simple sliding well the film is maintained, friction is changed to a process especially under the pressure in cable-pulling that scores the jacket and conof bends. calculations. duit surfaces. This process not The technology of cable luonly damages the surfaces but bricants is continually advancalso produces much higher ing. Product types include friction forces. Therefore, one petroleum jelly, soap-andNo amount of basic requirement is to have water mixtures, and polymer all conduit clean before mixtures. Selection of a cable lubricant can pulling begins. lubricant should consider the compensate for even A beneficial effect can be expected coefficient of fricachieved by introducing and tion, long-term compatibility small amounts of dirt maintaining a lubricant bewith the cable jacket, flammain a conduit. tween the surfaces of the cable bility, and adhesion of cable and conduit. Lubricants proto conduit after the lubricant duce a film of slippery material dries. The best combination of between the surfaces. The film will have very properties currently offered seems to be by polylow internal shear values compared with the mer-based lubricants. However, this is subject to bare friction of the surfaces and, therefore, the change as lubricant technology advances. cable-pulling tension will be less. How well the CABLE WEIGHT Equation 9.3 shows that pulling tension is diEquation 9.4.A Equation 9.4.B rectly proportional to cable weight. For simple cases of straight runs with a single cable, this is 3 Cables Cradled 3 Cables Triangular generally true. Weight is a consideration be2 2 –0.5 4 d d cause, in simple cases, it is the major contributWC = 1 + WC = 1– 3 (D – d) (D – d) ing element to the force exerted on the conduit by the cable. where: Wc = Weight correction factor However, in the practical case of multiple conductor runs, other considerations warrant the D = Conduit inside diameter, in inches use of a weight correction factor, WC. This factor d = Cable outside diameter, in inches depends on the configuration of cables in conduit. For a single cable in conduit, WC = 1.0. However, if multiple cables are present in a conduit, the values will be shown by the following Clearance = C two equations. Equation 9.4 shows the values for a cradled configuration as well as the values for a triangular configuration. Figure 9.11 illustrates the cradled and triangular cable configurations. For pulling calculations, the total weight of all cables in the conduit must be considered. For example, if three cables are being pulled in one Three Cables—Triangular Three Cables—Cradled conduit and each weighs 1.5 pounds per foot, the resulting cable weight (W) will be 4.5 lb./ft. FIGURE 9.11: Cable Configurations in Conduit. Cable weight is given in general data sheets.


9 However, for critical cable-pulling calculations, the specific cable manufacturer should be consulted for confirmation or more exact information. SLOPED PULLS Equation 9.5 applies only to straight horizontal cable pulls. For sloped conduit installations, different tension equations apply. Equation 9.5.A applies for upward slopes, and Equation 9.5.B for downward slopes. These equations reflect the additional tension needed to pull the weight of the cable upward and the reduced tension resulting from the assistance of gravity when the cable is pulled downward. The path of the cable is assumed to be a straight line. In the case of a vertical pull, the value of θ in Equations 9.5.A and 9.5.B is 90°. The expressions in parentheses in the equations then become +1 for upward vertical pulls and -1 for downward vertical pulls.

Equation 9.5.A

Equation 9.5.B

T = l × W × (f × WC × cosθ + sinθ)

T = l × W × (f × WC × cosθ – sinθ)

where:

T = l = W = f = WC = θ =

Tension Length of cable Weight of cable, per unit of length Coefficient of friction Weight correction factor (where required) The angle of the slope measured from the horizontal, in degrees

TABLE 9.17: Inside Bend Radius for 90° Schedule 40 Conduits.

Duct Size

Duct I.D. (Inches)

18"

Bend Centerline Radius 24" 30" 36"

48"

2

2.067

1.41'

1.91'

2.41'

2.91'

3.91'

4

4.026

—

1.82'

2.33'

2.83'

3.83'

5

5.047

—

—

2.29'

2.79'

3.79'

6

6.065

—

—

—

2.75'

3.75'

CONDUIT BENDS Cable pulling would be much simpler if all conduit runs were straight. However, bends are necessary and, in many cases, cause very large increases in cable-pulling tensions. Bends also cause compressive stresses on the cable as it is pulled through the bend. This force is referred to as sidewall bearing pressure (SWBP) and can damage the cable insulation and shielding systems. Regardless of the length of a cable run or type of cable, the sum of all bends in a run should never exceed 270°. For example, that could be two 90° vertical bends and one 90° horizontal bend. Gradual changes in conduit direction that do not have a sweep bend fitting or an elbow should also be included in observing this limit. In most cases, the calculated pulling tension for power cables will be exceeded before the 270° limit is reached. The effect of conduit bends not only is a function of the angle turned but also strongly depends on the inside radius of the conduit bend. Table 9.17 gives the inside radius for typical bends available for PVC and steel conduit. When planning conduit installations, the engineer must consider the minimum bending radius of cables. This minimum cable bending radius is determined by the cable construction and is generally given as a multiple of the cable outside diameter. If the minimum cable bending radius requirement is not met, the shielding and insulation systems of the cable may be damaged even if pulling tension and SWBP are low. Conduit bends must never be installed with an inside radius less than the minimum bending radius of the largest cable anticipated for installation in that location. The minimum bending radius for most commonly encountered shielded distribution cables is 12 times the outside diameter. For example, a 350-kcmil shielded power cable with an outside diameter of 1.6 inches should always have a bending radius of more than 12 times its outside diameter, or a minimum bending radius of 19.2 inches. Information in Table 9.17 shows that this cable should not be installed in a two-inch conduit bend with a radius of 18 inches. It would also be close to limits if it was installed in either a twoor four-inch conduit with a 24-inch sweep radius.

3 3 6 – Se c t io n 9

9 As underground conduits generally do not have the space limitations encountered by inbuilding wiring, the largest practical radius should be used on all bends. Most utilities standardize on 36-inch radius bends for duct banks. If more than one conduit size is being used in a duct bank, all sizes must have the same radius. The amount of increase in pulling tension that results from pulling around conduit bends can be calculated from Equation 9.5.C. Equation 9.5.C is a simplified equation that should be adequate for most practical cases. The amount of error is not likely to exceed approximately 15 lb. of tension for each bend.

Equation 9.5.C T2 = T1×ef × WC × θ where: T2 = T1 = e = f = WC = θ =

Tension at exit from bend Tension at entrance to bend 2.71828 Coefficient of friction Weight correction factor Angle of bend, in radians (1 radian = 57.296 degrees)

Tension Before Bend T1

ide Ins

Ra

Co of R) ( s diu

B uit nd

d en

Direction of Pull

SWBP =

Inner Conduit Wall

T2 (lb.) Inside Radius of Conduit Bend

Tension After Bend T2 SWBP results from cable tension in bends and creates additional pulling tension.

FIGURE 9.12: Sidewall Bearing Pressure.

Equation 9.6 SWBP =

T2 (single cable) R

Equation 9.7 SWBP =

(3WC – 2)T2 (three cables cradled) 3R

SIDEWALL BEARING PRESSURE SWBP is a measure of the compressive force applied perpendicular to a cable surface by the inner wall of a conduit bend (see Figure 9.12). The SWBP depends on the tension in the cable at that point and the radius of the bend. Equation 9.6 defines SWBP for the simple case of a single cable in a conduit bend. Equation 9.7 defines the SWBP for three cables in a cradled configuration. If the cables assume a triangular configuration, Equation 9.8 applies. In equations 9.6 through 9.8, T2 is the cable tension at the exit of the bend, R is the inside radius of the bend in feet, and WC is the applicable weight correction factor for the particular cable configuration. See Equations 9.4 and 9.5. SWBP will distort the cable cross-section. If SWBP is excessive, it will permanently damage the cable structure. This damage may be a crushed metallic shield, indentation of the semiconducting shield, or mechanical failure of the insulation. All these conditions will lead to shortened cable life, if not immediate failure. The allowable limits of SWBP for various cable types are given in Table 9.18. This information was developed under Electric Power Research Institute (EPRI) Project EL-3333 in 1984. These values are higher than previously recommended by cable manufacturers

Equation 9.8 SWBP =

WCT2 (three cables triangular) 2R


9 but represent the result of an dependent on the bend conextensive cable installation figuration, the soil characterisIf SWBP is excessive, testing program. tics, and the degree of permanent damage One cautionary note is that compaction. High-tension the SWBP values in Table 9.18 cable pulls should never be to the cable structure are for concrete-encased duct attempted through bends in will occur. or rigid steel conduit that is direct-buried conduit. well supported in bends. It Another concern with unmay be necessary to provide supported bends in directconcrete encasement in the buried conduit runs is the immediate area of a bend in possibility of burn- through by High-tension cable order to achieve adequate pulling cables or ropes. Burnsupport. If cable pulls apthrough results from frictional pulls should never proaching the limits of Table heat build up at the inner surbe attempted through 9.18 are attempted through diface of the conduit bend. This bends in direct-buried rect-buried conduit, the bends condition is much less likely may collapse or move with when steel pulling lines are conduit. disastrous results. Not only used instead of ropes. Burnwill the duct rupture, but the through will always result in cable will be cut and jammed. extensive cable damage or a Limiting SWBP values cannot be given for dijammed conduit. Burn-through is not a problem rect-buried conduit because they are highly where rigid steel bends are used and it is significantly reduced by the use of fiberglass reinforced epoxy (FRE) bends. TABLE 9.18: Recommended Maximum Sidewall Bearing Pressures. Source: EPRI EL-3333 (1984). JAM RATIO Cable Construction Type

Maximum SWBP (lb./ft)

XLPE Insulation - 600V Cable

1,200

• • • •

1,200 750 2,000

PE and XLPE insulation, concentric wire shield: Without jacket, single conductor Without jacket, three conductors With encapsulating jacket

PE and XLPE insulation, LC shield, LDPE jacket

1,500

PE, XLPE, EPR insulation, concentric wire or tape shield, LDPE and PVC sleeved jackets

2,000

Note. LDPE = low-density polyethylene; MDPE = medium-density polyethylene.

Equation 9.9 J=

D d

where J = Jam ratio D = Conduit inside diameter d = Outside diameter of single cable

When three single cables are pulled in parallel in a conduit, wedging action may develop in bends. This is caused by cables changing from a triangular configuration to a cradled configuration as they are pulled through the bend. This change in configuration will force the two outer cables further apart. If the conduit diameter is too small to accommodate this wider configuration, the cables will become jammed in the bend. The jam ratio (J) must be checked to predict this phenomenon. The jam ratio is simply the ratio of the cable diameter to the conduit inside diameter given in Equation 9.9. Either measure or refer to the manufacturer’s literature for the cable outside diameter. The inside diameter of typical conduit sizes may be found in Table 9.17. When the jam ratio is calculated, the probable cable configuration in the conduit can be determined. A listing of probable configurations is given in Table 9.19. (See Figure 9.11.) Experience has shown that cable jamming is most likely between J = 2.5 and J = 3.0. This is

3 3 8 – Se c t io n 9

9 particularly true if the SWBP in a bend exceeds CLEARANCE FACTOR 1,000 lb./ft. Therefore, this combination of conThe clearance between the upper cable and the ditions must be avoided. The most obvious top of the conduit should always be checked to method to prevent jamming is to always use a determine the size conduit required for a given conduit with an inside diameter at least three cable configuration. This clearance (C) is illustimes the outside diameter of the cable being trated in Figure 9.11. For a single cable in a conpulled. Doing so is often not practical, especially duit, the clearance is obviously the difference with high-ampacity distribution class cables. between the inside diameter of the conduit and Therefore, the jam ratio must be calculated, and the outside diameter of the cable. If the allowthe range of J = 2.5 to J = 3.0 should be avoided. able variation in cable diameter of five percent is As mentioned earlier, cable jamming does taken into account, the expected clearance is exoccur in bends. It is generally pressed by Equation 9.10. not a problem in straight pulls. When three cables are The cable will assume the conpulled into a conduit, they asClearance must be figuration indicated by Table sume either a triangular or a 9.19 and remain in that configcradled configuration. Howmaintained or the uration throughout the pull. In ever, because clearance is a cable will jam in addition, cables that are problem only as the cable the conduit. triplexed (twisted) before enapproaches the maximum tering the conduit will tend to conduit capacity and the camaintain the triangular configbles are always triangular uration through the bends. Bewhen the conduit fill is high, cause they will not change to the cradled clearance needs to be calculated only for the configuration, jamming will be avoided. triangular configuration. Equation 9.11 gives clearance under this condition. After calculating the expected clearance, make TABLE 9.19: Cable Configuration for Various sure that it is at least 0.5 inches or greater. This Jam Ratios. much clearance is needed to allow for possible conduit variations. If the upper cable contacts Jam Ratio Range Cable Configuration the top of the conduit, the cable will jam. J<2.4

Triangular

2.4–2.6

More likely triangular

2.6–2.8

Either triangular or cradled

2.8–3.0

More likely cradled

J>3.0

Cradled

Equation 9.10 C = D – 1.05d

Equation 9.11 C=

D 1.05d – 1.434d + 0.5(D – 1.05d) 1– 2 D – 1.05d

2 0.5

CABLE-PULLING EYE TENSION LIMITS The preceding discussions have covered many of the factors affecting the expected tension. The objective of the calculations has been to avoid overstressing cable during the pulling operation. The practical limit for cable tension is based not only on cable tensile stress but also on stresses in the connection point of the pulling wire (or rope) and the cable. Table 9.20 shows the allowable tension based on various pulling eye types and the conductor in the cable being pulled. From Table 9.20, it is determined that a 350-kcmil aluminum cable with an aluminum compression pulling eye can accommodate a tension of up to 2,800 lb. (350,000 cm × 0.008 lb./cm). If three cables are pulled in a single conduit, allowances must be made for unequal sharing of the total tension. It is generally assumed that, on


9 three-conductor pulls, only two of the cables actually develop tension. Therefore, the total tension of the three cables must be divided by two to establish the expected tension in each cable. This assumption must be applied to pulling eye load calculations. If cable is pulled with basket-type grips instead of pulling eyes, lower tension limits apply because of the mechanical stresses that are

TABLE 9.20: Recommended Maximum Pulling Tension Stress for Pulling Eyes on Copper and Aluminum Conductors. Source: EPRI EL-3333 (1984).

Conductor

Aluminum Compression (lb./cmil)

Solder Filled (lb./cmil)

Epoxy Filled (lb./cmil)

Copper (annealed)

0.011

0.013

—

Aluminum-solid (1/2 to full hard)

0.006

N/A

0.008

Aluminum-stranded (3/4 and full hard)

0.008

N/A

0.011

N/A = not applicable

imposed on the cable insulation, shield, and jacket by the grips. Table 9.21 gives the limits for pulling single and multiple cables with various grip arrangements. Cable-pulling grips must be used carefully and in strict accordance with the grip and cable manufacturers’ instructions. Grips must be attached with provisions to maintain grip compression even if the tension drops to zero. Cable under grips and for a distance of at least two feet beyond the end of the grip must be cut off and discarded. In addition, basket-type grips must be of a type specifically designed for pulling insulated cable because of the characteristics of the insulated cable and its behavior under the compressive forces developed by the basket grips. Grips generally used for cable support at riser poles are not satisfactory for cable pulling. Special split-basket cable grips are sometimes connected to cables in intermediate manholes to pull slack in these locations. Because of the construction of these devices and the fact that cable under the grips cannot be cut out, tension on these grips should be limited to 1,000 lb. If

TABLE 9.21. Recommended Maximum Pulling Tension Limits for Basket-Type Pulling Grips.* Source: EPRI EL-3333 (1984). Single Cable (lb.)

Three Cables in One Grip (lb.)

Three Cables (One Grip per Cable) (lb.)

XLPE insulation—600-V cable

2,000

2,000

4,000

EPR and Neoprene—600-V cable

2,000

2,000

4,000

PE and XLPE insulation, concentric wire shield, with and without encapsulating jacket—all voltages

10,000

5,000

20,000

PE and XLPE insulation, L.C. shield, LDPE jacket—15-, 25-, and 35-kV cable

8,000

4,000

16,000

PE and XLPE insulation, concentric wire or tape shield, LDPE and PVC sleeved jackets—all voltages

10,000

5,000

20,000

EPR insulation, concentric wire or tape shield, LDPE and PVC sleeved jackets— all voltages

10,000

10,000

20,000

XLPE insulation, copper wire or ribbon shield, MDPE sleeved jacket—all voltages

18,000

9,000

36,000

Cable Construction Type

* Conductor tensions must not exceed values calculated from Table 9.20.

3 4 0 – Se c t io n 9

9 higher levels of tension are expected, the cable manufacturer should be contacted for recommendations before this work is attempted. CABLE-PULLING CALCULATION SEQUENCE As described above, several factors determine cable tensions during the pulling process. In addition, many of these factors influence not only the pulling tension but also other factors in the pulling process. For example, the friction encountered in the first straight section of conduit will mean higher entering tension in the first bend. This tension on the cable entering the bend will contribute to higher SWBP in the bend and a higher tension contribution from the bend. Therefore, tension calculations must be organized to follow the same sequence as the actual field pull. Prior organization of essential data will also enhance the calculation process. The following sequence of steps will yield the greatest efficiency in the calculation process. STEP 1: Determine cable characteristics:

STEP 2:

STEP 3:

STEP 4:

STEP 5:

STEP 6:

• Weight • Diameter • Outer jacket material. Determine duct characteristics: • Material • Diameter. Determine friction factors for a given combination of cable outer jacket material and conduit composition. Consider the lubrication to be used and the effect of higher SWBP in bends. Calculate the jam ratio, clearance factor, and weight correction factor. If three cables are being pulled in one conduit, determine whether the cable configuration will be triangular or cradled. Determine cable limits including maximum allowable tension, maximum allowable SWBP, and minimum bending radius. Calculate the tension in each conduit section progressing from reel end to winch end. Determine if tensions and SWBP limits are met in each section. Consider the limitations of the cablepulling equipment. Use the ending

tension (T2 ) in each section as the beginning tension (T1 ) in each succeeding section. In most actual conduit runs, especially those with more than one bend, calculating pulling tensions in each direction will be beneficial. Often it will be found that factors such as conduit slope or the combination of bend locations will give lower pulling tensions if a particular direction of pull is used. In simple cases, such as a straight downhill pull or a pull with a single bend near one end, the best pulling direction may be obvious. However, as conditions grow more complex, examination of both pull directions will be beneficial. In addition, field access conditions may force installation crews to strongly consider the alternative pulling direction, and prior calculation of the expected tensions will simplify their decision-making process. Under any conditions, cable installation crews must be instructed to always use the specific direction of pull shown on project drawings unless prior consideration of the alternative direction has shown it to be acceptable. See Appendix J for several example calculations for selected cable pulls that outline the methodology to be used in evaluating the cable-pulling limitations. CABLE-PULLING SOFTWARE Although the methods described in this manual and referenced documents will allow calculation of expected pulling tensions, extensive pulling calculations can be simplified by using one of the calculation programs designed for use on a personal computer. Spreadsheet calculation programs might also be used, but the relatively low cost of currently available cable-pulling programs makes them the preferred approach. The engineer using a computer-based pullingtension program must recognize that a comprehensive program will require detailed knowledge of all cable and conduit parameters as well as access limitations. A pulling-tension program is a tool to simplify the calculation process, not a substitute for knowledge of expected field conditions. Data must still be accumulated and the pulling process organized in a logical fashion before calculations begin.


9 other values critical to accurate pulling calculaThe use of cable-pulling software is recomtions. Default values provided in software packmended on complex pulls or those that may ages may not be completely applicable to your closely approach allowable tension limits. One particular materials or pulling configuration. such software program is Cable Pulling Assistant Software allows the convenient examination (CPA) that was developed by General Electric of multiple actions and enCompany as part of its Distribsures consistent calculation acution Systems Testing, Applicuracy if data are entered cations, and Research (DSTAR) A cable pullingcorrectly. The better software program. This is available tension program is packages also use more exact through Cooperative Research equations for pulling tensions, Network (CRN). Another popa tool to simplify the especially in more complex ular program, “Pull Planner calculation process, conduit arrangements. This ex2000 for Windows,” has been actness should yield greater developed by American Polynot a substitute for accuracy with the same effort water Corporation. knowledge of expected on the user’s part. However, Regardless of the source of field conditions. even with the most advanced the software, it is incumbent software, the results are only on the user to confirm the valas accurate as the data that are ues used for coefficient-of-fricentered; good judgment must tion, allowable sidewall still be used in applying the results. bearing pressures, allowable cable tension, and


1. Conduit should be used wherever additional cable protection is required or the deferral of future excavation costs will justify the additional initial expense. 2. The sum of all bends in a cable run should never exceed 270°. 3. High-tension cable pulls should never be attempted through unsecured bends in direct-buried conduit. 4. The types of cable circuits to be installed will determine the conduit system design. 5. Manholes and/or splice boxes will provide convenient access points to conduit systems and are useful for cable pulling and splicing.

6. Cable pulling should be planned as part of the conduit system design process. 7. Limits on cable tension and SWBP should be observed to avoid cable damage. 8. The expected pulling direction and all field conditions must be considered in cable-pulling calculations. Detailed cablepulling instructions must be provided to field crews. 9. Cable under basket-type pulling grips must always be discarded.


Joints, Elbows, and Termin a t i o n s – 3 4 3

10 In This Section:

Joints, Elbows, and Terminations

Application of Joints, Elbows, and Terminations Joints, Elbows, and Terminations for 200-Ampere Primary Circuits Joints, Elbows, and Terminations for 600-Ampere Primary Circuits Joints and Terminations for Secondary Circuits Summary and Recommendations

Application of Joints, Elbows, and Terminations

On new and existing underground distribution systems, it is routinely necessary to join cables to provide continuous lengths and reconnect cables where a service failure has occurred. This joining or reconnection is accomplished by joints; typical designs are given in this section for primary and secondary cables. ANSI and IEEE standards use the word joint as standard terminology in lieu of the word splice, but the two terms mean the same thing. This section is written using the ANSI/IEEE terminology. Every cable circuit must have at least two permanent terminations, one at each end of the circuit. In secondary circuits, a termination is required for mechanical support of the cable, connection to equipment, and physical protection of the insulation and conductor, particularly against entry of water into the conductor interstices. In primary circuits, in addition, the termination must reduce the radial and longitudinal electrical stress between the conductor and ground. This reduction is accomplished by means of stress cones or other stress-control devices. In primary circuits, elbows are used at transformers, junctions, and switches to terminate a section of the cable circuit. With this type of cable termination, the cable can be disconnected

and reconnected to the apparatus without disturbing the cable structure. Typical elbows in common use and their application are described in this section. Also within cable circuits, separable joints and elbow connectors are sometimes used to join cables, provide branch circuits, and make separable connections at apparatus for additional circuits. The types and applications of separable connectors are discussed. A typical primary UD circuit begins at a substation or from an overhead line where there is a transition to an underground cable. The primary cable extends down a structure and into the ground to the first piece of equipment or apparatus, or it may have a joint to extend the cable run. A typical secondary circuit begins at the distribution transformer (or pedestal), runs underground, and emerges at the meter base located on the consumer’s premises. Under RUS standards, joints, elbows, and terminations are designed to be installed on approved primary cables rated for 15-, 25-, or 35-kV service. At present, industry standards provide for two classes of primary cable accessories. These general classes have continuous current ratings of 200 and 600 amperes. The actual circuit

3 4 4 – Se c t io n 1 0

10

Joints, Elbows, and Terminations for 200-Ampere Primary Circuits

current, however, will be governed by the circuit conductor size and circuit breaker rating. If the circuit has a small conductor size, such as No. 2 or 1/0 AWG, a 200-ampere component can be used. Normally, a 200-ampere component is used because it is more economical and is available with load-break capabilities. If the circuit will carry close to 200 amperes, it may be preferable to use a 600-ampere component because the larger component will run cooler, which is desirable, and because future loads

may go above 200 amperes. Typically, 200-ampere components can be used on phase conductors up to 4/0 stranded AWG. Also, 600-ampere components can be used on cables 4/0 stranded AWG and larger. Components rated for 600 amperes should always be used when the circuit exceeds 200 amperes. For secondary 600-volt circuits, joints and terminations are designed for the same current rating as that of the cables to which they are attached.

CABLE JOINTS Design Features of Joints RUS-approved joints may be premolded, cold shrink, or heat shrink. Taped joints are not approved. When two cables are joined, the length of the insulation between the conductor and insulation shields must be made greater than the radial thickness of the insulation because the insulation surface will not support voltage stress.

Therefore, the path between the conductor and the insulation shields is increased by the removal of the insulation shields and outer coverings for a specified distance. When this is done, the leakage path between the conductor and insulation shields is greater; however, the voltage gradient increases abruptly along the insulation surface at the edge of the insulation shield (see Figure 10.1). Figure 10.2 represents a premolded

Connector Conductor

Insulation Equipotential Lines

Equipotential Lines 0 10 20 30 40 50 60 70 80 90 100

100

80

Joint Housing Shield

50

15 30

Insulation Shield

Conductor Cable Insulation Shown by equipotential lines at the end of the removed insulation shield. Values shown are a percentage of total voltage.

FIGURE 10.1: Voltage Stress Concentration.

Insulation Shield

Equipotential values are a percentage of total voltage.

FIGURE 10.2: Voltage Stress Distribution in a Typical Premolded Joint Housing.


10 joint housing cross-section showing the voltage stress distribution along the body of the joint insulation. This reduces the voltage gradient across the insulating components and increases the resistance of the joint to failure. Before molded or heat-shrink joint designs, joints consisted of hand-wrapped layers of insulating tapes to a predetermined contour which was commonly referred to as a pencil. For electrical integrity and watertightness, these joints were highly dependent on the skill of the splicer. A shielding tape was wrapped over the conductor connector smoothly with no creases or sharp projections. The cable insulation was tapered (pencilled) to allow the insulating tapes to be wrapped to a smooth contour with minimum air gaps. The tape for the insulation shielding was wrapped over the insulation with minimum creases. The concentric neutral was spliced with a copper jumper wire. It is easy to see how difficult it was to fabricate a good, long-lived cable joint, particularly in a splicing pit in the rain! RUS does not currently approve taped joints, either temporary or permanent. Premolded Permanent Straight Joints A straight joint is used when a direct-buried cable is repaired because of a fault or dig-in. For cable installed in conduit, faulted cable usually is

7 5

4

6

3

2 1. 2. 3. 4.

1

Molded Insulation Insulation Shield Crimped Connector Conductor Shield

5. Extended Cable Entrance 6. Grounding Eye 7. Spliced Concentric Neutral

Cable jacket and covering over the neutral not shown.

FIGURE 10.3: Premolded Permanent Straight Joint for Primary Cables. Source: Elastimold Corporation, a division of Thomas & Betts Company.

replaced with new cable, and joints are used in pulling vaults to join sequential runs of cable in duct. Sometimes a straight joint is used when the length of a direct buried cable run exceeds the lengths of replacement cable available. However, the use of joints for such purposes is not desirable. When a straight joint is used for repair, if there is no slack in the cable, it is sometimes necessary to use a cable stub with a joint on each end. However, there are special premolded joints which will allow the repair of localized faults that have limited conductor and insulation damage. Premolded joints are preferred over other styles of joints because the critical voltage stress control features are fabricated under controlled factory conditions and less is left to the skill of the splicer and a favorable splicing environment. With premolded joints, the joint body and all other joint components are closely sized for the cable to be joined and a standard tubular compressed conductor connector is used. Voltage stress control over the spliced conductor is achieved by a built-in layer of conducting rubber that extends onto the cable insulation. The cable insulation shields are bridged by the molded outer shielding layer of the joint housing, which thus places this shield at ground potential. Moisture entry into the joint-cable interface is prevented by a joint housing with a tight interference fit between the cable insulation and the insulation shield and the use of a special silicone assembly lubricant. The elastic properties of the joint housing materials are such that a constant pressure is maintained against the cable, creating a watertight seal. For this reason, it is mandatory to select the joint kit that provides a tight fit. There are even cable joints available that will accommodate two slightly different sizes of cable. Use of such kits might reduce the need to stock an individual kit for each of the sizes of cable on a cooperative system. A typical premolded joint is shown in Figure 10.3. A one-piece housing is normally used. However, in cases of large conductor sizes and limited available space, it may be necessary to use a joint in which the housing is made in two sections. This is less desirable than a joint with a one-piece housing because it provides an additional path for water entry where the housing parts fit together.

3 4 6 – Se c t io n 1 0

10

STEP 1: Position Premolded Housing and Sleeves.

STEP 2: Lock Premolded Housing.

STEP 3: Roll Rubber Sleeves Over Cable Jacket.

FIGURE 10.4: Jacket Replacement Assembly (Method C). Source: Elastimold Corporation, a division of Thomas & Betts Company.

Protective Covering Over Neutral Current RUS specifications require a covering (or jacket) over the cable neutral. The spliced neutral must also be protected from the environment by one of four methods: 1. Method A—A wrap-around heat-shrink polymeric sleeve, 2. Method B—A tubular heat-shrink polymeric sleeve, 3. Method C—A tubular cold-shrink polymeric sleeve, and 4. Method D—A prefabricated assembly.

installation is greater than for the premolded joint, particularly for the application of the heat required to shrink the various components. Premolded Permanent Wye Joints Although somewhat outdated and infrequently used, the premolded wye joint is used to connect a branch circuit. These type joints are very much permanent and, as such, have limited usefulness because of the lack of sectionalizing flexibility and trouble-shooting opportunities that a three-way connection provides. These joints are constructed with an inner metallic bus in the form of a wye. A typical molded wye joint is shown in Figure 10.5. The purpose of the grounding eye is to ground the joint housing to the neutral. The purpose of the test point is to test if the circuit is energized.

Heat-Shrink and Cold-Shrink Straight Joints Heat-shrink and cold-shrink permanent joints consist of a crimped conductor connector over which is placed a succession of wrapped, stresscontrol, and insulating layers. The heat-shrink Separable Molded Joints version is reduced over the cable with a torch; Separable molded straight, wye, and tee joints the cold-shrink version is applied by pulling out are sometimes used for making temporary conan inner coiled expander barrier. After the connections. Typically, separable molded joints are centric neutral is spliced and taped, an overall made up of a series of two, three, or more “Tjacket repair sleeve is shrunk down over the body” elbow terminations that use interconnectjoint and neutral to waterproof the joint. ing plugs to mate one T-body This type of joint is someelbow to another with insulattimes employed rather than a ing caps on the elbow ends, premolded joint. Some users Do not use separable once assembled. These joints prefer this type because a molded joints in typically must be assembled given joint kit is applicable to direct-buried cable with a torque wrench or spea wide range of conductor cial spanner wrench to tighten sizes. On the other hand, the applications. the interconnection plugs. skill and time required for


10 8

2

9 3

5

4 1. 2. 3. 4. 5.

6

7

Wye-Shaped Metallic Bus Bar Stress Relief Adapter Premolded Housing Crimped Conductor Connector Holding Collar

1 6. 7. 8. 9.

Conductor Connector Shield Molded Shielding Insert Grounding Eye Test Point

Spliced neutrals not shown.

FIGURE 10.5. Premolded Permanent Wye Joint for Primary Cables. Source: Elastimold Corporation, a division of Thomas & Betts Company.

They sometimes are used to subdivide UD circuits and are installed in boxes in handholes, in cabinets above ground, or in vaults. If separable molded joints are to be used for permanent service, they should be used only in non-directburied locations—that is, in manholes, vaults, junction boxes, switching cabinets, and so on— where they can be mounted to take all tension and mechanical stress off the components.

Advantages of heat- and cold-shrink joints are as follows:

Advantages and Disadvantages of Straight Premolded and Heat- and Cold-Shrink Primary Joints Premolded and heat- and cold-shrink primary joints have various advantages and disadvantages. Advantages of premolded joints are as follows:

• Open flame hazards for heat-shrink units, • Greater skill required for installation, and • Longer length on some designs.

• Built-in electrical stress control, insulation, and shielding in a factory-made joint housing, • Minimization of voids and contaminants in the insulation, • Factory pretesting of joint housing, • Fewer installation steps, • Shorter installation time, • Convenient circuit modifications, and • Shorter total length. A disadvantage of premolded joints is that they are sized to a particular cable diameter range.

• Wider range of cable diameters with one kit, and • Smaller overall diameter. Disadvantages of heat-shrink and cold-shrink joints are the following:

The selection of joint types for general use on the cooperative system must weigh all factors involved in the particular situations where the joints will be used. ELBOWS Application Elbows are used to terminate primary cables at transformers and switches. Elbows are also used at junction boxes where taps and line extensions are made to existing cable systems. Two basic types of elbows are permitted by RUS. One type is called dead-break (formerly non-load-break) because it must be engaged or disengaged while the circuit is de-energized. The

3 4 8 – Se c t io n 1 0

10 second type, which can be engaged or disengaged from an energized circuit, is called load-break by virtue of built-in arc-quenching elements.

11

5

9 1. 2. 3. 4. 5.

10 1

6. 7. 8. 9. 10. 11.

2 8

4 3

Elbow Housing Housing Shielding Voltage Sress Relief Inner Shield Insert Interference Fit Between Cable Insulation Elbow Grounding Eye Cable Entrance Test Point Hot-Stick Eye Conductor Connector Male Conductor Contact

5 6

7

Neutral and jacket not shown. (Note: 200-ampere units are not RUS approved.)

FIGURE 10.6: Dead-Break Elbow for Primary Cables. Source: Elastimold Corporation, a division of Thomas & Betts Company.

1b

3

2

5

4

1a

6 12 11 7

13

1a. 1b. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

8

10

Arc Follower Conductor Contact Elbow Housing Locking Ring Conductor Connector Hot-Stick Eye Indentification Band Test Point Voltage Sress Relief Cable Entrance Grounding Eye Inner Shield Insert Housing Shielding Interference Fits Between Cable Insulation Surface and Insulation of Elbow and Between Adapter and Elbow Insulation

9

Concentric neutral and jacket not shown.

FIGURE 10.7: Load-Break Elbow for Primary Cables. Source: Elastimold Corporation, a division of Thomas & Betts Company.

Dead-Break Elbows Dead-break elbows can be used for 200- and 600-ampere systems. RUS, however, does not approve 200-ampere dead-break elbows, but does approve 600-ampere dead-break elbows. As with a molded joint, the end of the cable must be prepared for insertion into the elbow. The elbow contains a conductor-shielding layer, insulation, and an insulation-shielding layer. A special connector is crimped onto the conductor; this connector fits to a contact inside the elbow. A typical dead-break elbow is shown in Figure 10.6. Elbows with a test point are often used to enable the operator to determine if the circuit is energized. The elbows also have a grounding eye for grounding the housing. They contain a built-in voltage relief stress cone that fits tightly over the insulation, thus reducing the voltage stress at the end of the cable insulation shield. Manufacturers’ instructions must be followed carefully when preparing the cable for insertion into the elbow. Load-Break Elbows Load-break elbows are used in 200-ampere circuits only. For 600-ampere circuits, it is necessary to use dead-break elbows or some other means for opening the circuit. The load-break elbow connects the primary cable to apparatus such as transformers, switches, and junction boxes. A typical load-break elbow is shown in Figure 10.7. The conductor contact area contains a locking ring to prevent the elbow from being quickly dislodged when the load is interrupted. The housing of the elbow is constructed differently from the housing of the dead-break elbow so as to extinguish an arc during removal, thus interrupting the primary circuit. Elbows at Junctions Both 600-ampere dead-break and 200-ampere load-break elbows are used at junctions that are used for sectionalizing, looping, tapping, and jointing. Junctions are installed in handholes, apparatus cabinets for transformers, switches, and so on, or pedestals above ground.


10 Junction

Feed-Through Insert

Bushing Well

Bushing Insert

Insulating Cap

Parking Bushing

Load-Break Type Elbow

Grounding Elbow

Feed-Through Bushing Insert

FIGURE 10.8: Typical 200-Ampere Elbow Accessories. Source: Elastimold Corporation, a division of Thomas & Betts Company.

Retaining Clip Sealing Tube Solderless Ground Clamp Accessory Sealant

FIGURE 10.9: Heat-Shrink Jacket Seal at Elbow. Source: Raychem Corporation.

Elbow Accessories Typical accessories used with 200-ampere loadbreak elbows are shown in Figure 10.8. A bushing insert, mounted on the apparatus, is used between the elbow and apparatus to connect the elbow to a transformer, switchgear, and other devices. A feed-through insert is used between two elbows to feed a cable circuit past a piece of apparatus; the feed-through insert is attached to the apparatus. A parking bushing is an insulated bushing that isolates and dead-ends a cable terminated in an elbow. These accessories can be used to convert a radial-feed transformer to loop feed. Insulating caps are used for dead-ending or sealing off a bushing insert, feed-throughs, and junctions. They will also waterseal open bushings. PREPARATION OF CABLES FOR USE WITH ELBOWS Preparation of Cable The end of the cable to be inserted into the elbow must be cut to a length that will allow convenient operation of the elbow during switching. As with all cable preparation, cleanliness is extremely important during the elbow installation process. In addition to these general steps, follow the elbow manufacturers’ recommendations for cutback dimensions. Sealing of Cable Jacket at Entrance to Elbow It is recommended to seal the cable jacket to the elbow to prevent moisture entry, especially in areas of high humidity. If elbows are used underground, such as in handholes or manholes, it is mandatory to place a waterproof seal over the cable jacket and elbow to prevent moisture from entering the elbow. Two types of seals can be used: the cold-shrink seal and the heat-shrink seal. Figure 10.9 shows a typical heat-shrink seal in place. Follow manufacturers’ recommendations when a seal is used. Electrical Ratings of Elbows Table 10.1 gives the electrical ratings of elbows for primary cables. Production tests are performed before shipment. Design tests are performed by the manufacturer in order to qualify for the ANSI/ IEEE Standard 386 rating. The current ratings of 200 and 600 amperes for elbows are indicated in Table 10.1.

3 5 0 – Se c t io n 1 0

10 TABLE 10.1: Electrical Rating of Elbows. Source: ANSI/IEEE Standard 386. 200-Ampere Elbows

600-Ampere Elbows

Voltage Class

15 kV

25 kV

35 kV

15 kV

25 kV

35 kV

(A) Production Tests (a) Minimum Corona Level, kV, rms (b) AC Withstand, 1 min., kV (c) Test Point Voltage

11 34 *

19 40 *

26 50 *

11 34 *

19 40 *

26 50 *

200** 10

200** 10

200** 10

600** 25

600** 25

600** 25

300

300

300

900

900

900

*** **** 53

*** **** 78

*** **** 103

*** **** 53

*** **** 78

*** **** 103

(B) Design Tests (a) Continuous Current Operation (b) Short Time Current, 0.17 sec., Amperes x 1,000 (c) 8-Hour Overload Current, Amperes (d) Switching Current (e) Fault Closure (f) DC Withstand, 15 min., kV *

A test voltage is applied to the conductor system of the elbow. The response of a suitable sensing device on the elbow test point shall indicate an energized condition.

**

The ratings are for the following service conditions: • In air, including exposure to direct sunlight, • Buried in earth, • Intermittently or continuously submerged in water at a depth not exceeding six feet, • Environmental temperature of I –40°F to +140°F (for dead-break elbows) I –4°F to +149°F (for load-break elbows), and • Altitude not exceeding 6,000 feet.

*** Applicable to 200-ampere load-break elbows only: The elbow will withstand 10 complete switching operations at rated voltage and rated current without arcing to ground or impairment of its ability to meet other requirements of the specification. **** Applicable to 200-ampere load-break elbows only: The test is to verify that the elbow is capable of closing on the short time currents of (B)(b) above after the switching current test.

Elbow Connectors An elbow connector is an elbow in the form of a tee for use on dead-break 200- and 600-ampere circuits. Elbow connectors can be used for taps and joints to sectionalize, loop, tap, and join cables. Modular arrangements are used mainly in pedestals or in apparatus cabinets and handholes. CABLE TERMINATIONS Primary cable terminations are devices used to make the transition from air-insulated conductor systems, such as overhead lines, to solid-dielectric insulation systems, such as an underground distribution system. The termination controls the

electrical stress at the end of the primary cable and seals the cable end from water entry. Terminations are, thus, used to connect primary cables to overhead lines, switchgear, or other equipment that is air-insulated. Terminations for primary cables incorporate a stress cone to control the voltage stress at the end of the cable insulation shield. They also are designed to prevent water entry into the cable and, on some types of terminations, to provide mechanical support for the cable. Types of Terminations RUS specifications permit the following types of cable terminations:


10 • • • •

for outdoor use are sometimes employed indoors where the environment is not clean. Outdoor terminals also contain a seal at the end of the conductor to prevent moisture entry.

Premolded, Porcelain, Heat shrink, and Cold shrink.

Premolded The type of cable termination shown in Figure 10.10 is a premolded slip-on stress cone for indoor use. For outdoor use, because of airborne contamination and wet conditions, the creepage path between the conductor and ground must be increased. This is accomplished with integral or separately stacked premolded skirts as shown in Figures 10.11 and 10.12. Terminations designed

Porcelain A typical porcelain terminal for outdoor use is shown in Figure 10.13. These terminals incorporate an inner molded rubber stress cone. The outer surface of porcelain terminations are easier to clean and have higher tracking resistance. For locations with heavy airborne pollution and wet conditions, this type is preferred.

1 1 2

1 2 2

3

3

3

4

4

5

4

6

5

6 7 5

7 6

8 1. 2. 3. 4. 5. 6.

Cable Insulation Interference Fit Stress Cone Insulation Stress Relief Shielding Internal Step for Correct Positioning Grounding Eye


FIGURE 10.10: Premolded Indoor Termination (Slip-on Stress Cone) for Primary Cables. Source: Elastimold Corporation, a division of Thomas & Betts Company.

1. 2. 3. 4. 5. 6.

Contact Connector Molded Rubber Cap Water Seal Retainer Washer Insulator Cable Insulation Interference Fit Between Cable Insulation and Insulator 7. Stress Relief Shielding 8. Ground Connection

1. 2. 3. 4. 5. 6. 7.



FIGURE 10.11: Premolded Integral Indoor/Outdoor Termination for Primary Cables. Source: Elastimold Corporation, a division of Thomas & Betts Company.

Contact Connector Molded Rubber Cap Water Seal Premolded Rubber Skirts Cable Insulation Ground Connection Clamp Grounding Eye Stress Cone

FIGURE 10.12: Premolded Modular Indoor/Outdoor Termination with Separate Skirts for Primary Cables. Source: Elastimold Corporation, a division of Thomas & Betts Company.

3 5 2 – Se c t io n 1 0

10 Corrosion-Resistant Brass Top Cap and Threaded Stud Connector Silicon Tape Seal

+

Silicon Insulator

Or

Aerial Lug

Hi-K Stress Relief Tube

Flat Pad Lug

Or CorrosionResistant Brass Top Cap with Eye Bolt

Silicon Grease Semi-Con Tape

O-Ring Moisture Seal

Ground Strap Assembly Mastic Seal

Marker Tape Spring-Loaded Elastomer System Arc-Resistant Porcelain Insulator

CorrosionResistant Aluminum

Stress Cone Integral Cable Ground & Moisture Seal

1. Install ground strap assembly and seal with mastic

FIGURE 10.13: Porcelain Indoor/Outdoor Terminal for Primary Cables. Source: Joslyn Manufacturing Company.

2. Position terminator over cable

3. Remove core

4. Seal top with rubber tape

FIGURE 10.14: Cold-Shrink Indoor/Outdoor Termination for Primary Cables. Source: 3M Electric Products Division.

Cold Shrink and Heat Shrink Figure 10.14 shows a typical cold-shrink terminal. A heat-shrink termination is similar in construction except that some components are wrapped on and others heat shrunk. For ease of installation, the premolded or porcelain types are preferred. Outdoor Terminal Corrosion Protection An important consideration in the construction of outdoor terminals is the corrosion resistance of the exposed metallic parts. Galvanic corrosion results when two dissimilar metals are connected

in the presence of an electrolyte. In sunlight, and dissolved in moisture, pollutants produced by motor vehicles and coal-burning plants become corrosive acids. Airborne saltwater from the sea or from winter roads is also very corrosive. In these environments, terminal hardware is best made of silicon bronze rather than plain aluminum. Areas that are dry and have little airborne pollution may have terminal hardware made of aluminum or aluminum plated with tin. Table 10.2 gives the relative corrosion resistance of metal combinations for use in outdoor terminals.


10 TABLE 10.2: Relative Corrosion Resistance of Metal Combinations for Outdoor Terminations.

Joints, Elbows, and Terminations for 600-Ampere Primary Circuits

Corrosion Resistance Rating

Aerial Cable Conductor

Exposed Termination Material

Aerial Connector Material

Best

Copper

Copper or Bronze

Bronze

Very Good

Copper

Copper or Bronze

Tinned Bronze

Very Good

Aluminum

Tinned Aluminum

Tinned Aluminum

Good

Aluminum

Copper or Bronze

Tinned Aluminum

Poor

Copper

Tinned Aluminum

Tinned Aluminum

Poor

Aluminum

Copper or Bronze

Bronze

CABLE JOINTS The designs, types, and construction of 600-ampere joints are similar to those of 200-ampere joints except the former are larger in physical size because they accommodate larger conductors. Some manufacturers provide cable adapters allowing 600-ampere joints to be used on smaller cables of 200-ampere circuits with the advantage that a smaller joint inventory can be maintained.

FIGURE 10.15: Stick-Operable, Non-Loadbreak Elbow Applied to Pad-Mounted Switchgear. Source: Elastimold Corporation, a division of Thomas & Betts Company, 2008.

ELBOWS Elbows for 600-ampere circuits are basically the same as for 200-ampere circuits, except for physical size. Most manufacturers of 200-ampere elbows cannot accept cables larger than 4/0 AWG, which is typically the point at which 600-ampere elbows are needed. Load-break elbows are not used on 600-ampere circuits because it is not practical to interrupt a high-current arc. These type elbows are not currently being manufactured. Generally, 600-ampere elbows are used on high-current apparatus bushings—such as large pad-mounted transformers (particularly at voltages of 4,160/2,400 volts)—and 600-ampere sectionalizing switches. Currently, several manufacturers of 600-ampere class pad-mounted and vaultmounted sectionalizing switches offer 600-ampere load-break, group-operated disconnect switches, equipped with 600-ampere threaded-stud bushings to accommodate dead-break 600-ampere elbows. Protected positions out of these devices are typically 200-ampere cables, using power fuses, vacuum interrupters, or SF6 interrupters. As these devices typically include load-break switching on the 600-ampere positions, elbows for these positions are dead-break and must be installed (or removed) de-energized. In recent years, several manufacturers have introduced product lines of stick-operable, deadbreak 600-ampere elbows, as the use of these type switches and larger transformers has expanded. Initially, and in the foreseeable near future, these stick-operable, dead-break elbows and accessories

3 5 4 – Se c t io n 1 0

10 will be quite expensive, compared with the fixed dead-break counterparts. Some of the stick-operable, dead-break devices require specialized tools and, quite often, specialized training for safe operation. Figure 10.15 shows a style of stick-operable, dead-break elbows, with the noted accessories available as options. As an alternative to the high cost of some sectionalizing switches, manufacturers now offer padmounted junction boxes that can be equipped with multipoint, insulated 600-ampere bushing terminals where two-, three-, or four-way modules can be provided to the cables together in multiple directions. With the use of stick-operable, dead-break elbows, this “junction box” becomes a 600-ampere sectionalizing switch, at a greatly reduced cost. However, the following limitations and cautions must be recognized with this alternative: • The additional cost of stick-operable, dead-break elbows, • The additional stocking requirements for elbows and accessories, • Specialized training for operating personnel, and • Limited dead-break switching operations.

8

7

6

5

4 3 2 1

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Cable Adapter Stress Relief Crimped Connector Grounding Eye Test Point Test Point Cap Inner Shield Insert Standard Bushing Shape Dead-End Plug Elbow Connecting Plug Reducing Tap Well Plug 200-Ampere Reducing Plug

Accessories: 9

10

11

12

Cable connects at bottom of elbow.

FIGURE 10.16: Dead-Break 600-Ampere Elbow Connector and Accessories for Primary Cables. Source: Elastimold Corporation, a division of Thomas & Betts Company.

In addition, 600-ampere cables are typically larger and, as a result, stiffer to handle with “operating” sticks. Their larger size also requires additional room to train cables and the need for extra space for slack cable. Elbow Connectors Elbow, or T-Body, connectors for 600-ampere circuits are used in applications similar to those for 200-ampere circuits. A typical elbow connector is shown in Figure 10.16. Special plugs (shown in Figure 10.16) are used to dead-end one side of the connector, to connect to another connector that terminates another circuit, and to allow the connector to be used on a 200-ampere circuit. Several applications of the use of 600- and 200ampere elbows deserve noting here, as follows: • When a 600-ampere elbow is terminated on a 600-ampere apparatus bushing, the use of a 600-ampere to 200-ampere load-break tapreducing plug provides a location to (1) extend a 200-ampere cable to serve additional load, or (2) install an elbow-type, metal-oxide lightning arrester. Using this tap-reducing plug and elbow arrestor also provides an excellent location to verify phasing (with the arresters removed) with conventional phasing sticks. • Multiple 600-ampere elbows can be “spliced” together with a 600-ampere “connector plug” that basically couples elbow-to-elbow to form a modular joint that can be separated later. As noted previously, these separable (modular) joints should never be direct buried and should be mounted in manholes, vaults, and so on to take all weight and mechanical stress off the elbows and connector plugs. • Elbow “dead-end plugs” provide a test point when used with certain voltage testers to determine if a cable is indeed energized. • Elbows provide a temporary grounding point. Note: Most 600-ampere elbows and elbow accessories use fairly large-diameter threaded studs, conductive hex nut, and single-hole compression terminals for current-carrying capability up to the full 600-ampere rating. This single point connection’s integrity is critical to the safe and stable operation of the circuit. Therefore, it is imperative


10 that all 600-ampere devices be tightened securely and supported securely to guarantee full current-carrying capacity. Many of these elbows

Joints and Terminations for Secondary Circuits

CABLE JOINTS For new installations of secondary circuits, it is not generally necessary to use joints. In some cases, however, joints are used when a secondary circuit is damaged. Four basic types of joints are in use (see Figures 10.17 to 10.19):

and accessories require the use of a torque wrench and many require special installation tools and wrenches to assure proper connectivity.

Insulate underground secondary cable terminations at transformers.

FIGURE 10.17: Housing Assembly Joint for Secondary Cables. Source: Blackburn, Thomas & Betts Company.

FIGURE 10.18: Cold-Shrink Joint for Secondary Cables. Source: 3M Electric Products Division.

FIGURE 10.19: Heat-Shrink Joint for Secondary Cables. Source: Raychem Corporation.

• • • •

Housing assembly, Cold shrink, Heat shrink, and Rubber sleeve.

The housing assembly joint is the simplest to use. It consists of a molded housing and rubber end caps that are placed on the cables before the conductor is spliced. The molded housing does not contain a voltage stress cone, as is the case for primary cables, because of the low voltage stress. The function of the joint is to prevent water entry and corrosion of aluminum conductors. The cold-shrink joint contains an expanded sleeve over a removable spiral core. The core is removed, allowing the sleeve to shrink down over the spliced conductor. No heat is required. The heat-shrink joint has an adhesive-lined sleeve that is shrunk down over the spliced conductor using a source of heat such as a torch. The rubber sleeve joint is not shrunk down but relies on an interference fit when it is slid over sealant strips that are wrapped over the cables adjacent to the ends of the conductor.

TERMINATIONS AT TRANSFORMERS RUS requires that terminations for underground secondary cables at transformers be insulated in their dead-front designs. However, stress relief at terminations is not required because the voltage stress is low. The insulation is molded onto some terminations when they are manufactured. Cable terminations at transformers mounted in enclosures, as well as underground transformers, must be moisture-sealed and insulated. The covering of the terminations serves to prevent corrosion, especially of aluminum conductors, and is a safety measure in case of accidental contact by a worker. Hand-applied tapes to cover the bushings and busses of the transformer are not permitted by RUS.

3 5 6 – Se c t io n 1 0

10 There are numerous types of terminations for secondary cables. The most popular type of termination for a single line is the sealed stud termination in which a sealing cap covers the bolted terminal to the transformer. An insulated bus is used for multiple terminations; the insulation is supplied by a rubber boot that covers the bus and the cable terminations at the bus. An alternative to this type is the housing and sleeve assembly type. In this case, the housing protects the cable termination from moisture and prevents accidental personal contact. Typical examples of these terminations are shown in Figures 10.20, 10.21, and 10.22. All of these terminations, when carefully selected for the cable size, conductor composition, and bushing configuration are satisfactory in most environments.

Cap for Porcelain Bushings

Threaded Stud Secondary Bushing

Bushing Neck

O-Rings in Sealing Cap Seal Around Bushing Neck and Connector

FIGURE 10.20: Sealed Stud Termination for Secondary Cables. Source: Blackburn, Thomas & Betts Company.

Transformer Bushing Metallic Bus

“BOOT” Transformer Bushing

FIGURE 10.21: Bus and Rubber Cover Termination for Secondary Cables. Source: Blackburn, Thomas & Betts Company.

FIGURE 10.22: Housing and Sleeve Assembly Termination for Secondary Cables. Source: Blackburn, Thomas & Betts Company.



1. Use factory-made joints, elbows, terminations, and elbow connectors in 200- and 600-ampere primary circuits. 2. Use factory-made joints and terminations in secondary circuits. Taped joints and terminations are discouraged and are to be used only in an emergency, as a temporary device. 3. Avoid joints for primary circuits in new installations. When required for long lines, branch circuits, and so on, premolded joints are preferred; those with one molded housing rather than a split housing are most acceptable because of less likelihood of water entry. Permanent joints, in which the conductors are joined by a crimped connector, are preferred over separable joints. The latter should not be buried directly in the ground, but installed in handholes, boxes, and cabinets. Use them to connect circuits that are likely to be changed at an early date. 4. Use load-break elbows in 200-ampere circuits; use dead-break elbows in 600-ampere circuits. The most popular use of elbows is

to terminate short lengths of cables at transformers within the circuit; other uses are to terminate cables at apparatuses such as switches and junctions. 5. Use T-body elbow connectors, which are separable devices, at apparatuses such as switches, junction boxes, and transformers to connect branch circuits and other equipment such as grounding cables. They are for dead-break 600-ampere circuits only. 6. Permanently connect the ends of a primary cable run or circuit to a cable termination, which provides voltage stress relief between the conductor and ground and prevents entry of moisture into the cable conductor. Premolded terminations, consisting of polymeric materials, are the most popular type. They are subject to surface tracking, however, and, in areas where contamination from the environment is likely, porcelain terminations are preferred. 7. Carefully select terminations to prevent corrosion at the cable termination and to provide good mechanical support for cables.


Cable Te s t in g – 3 5 9

11 In This Section:

Cable Testing

Reasons For and Benefits of Cable Testing by the User Primary Cable Tests by the User Secondary Cable Tests by the User Tests by the Cable Manufacturer Summary and Recommendations

Reasons for and Benefits of Cable Testing by the User

Although power cables are subjected to extensive testing by the manufacturer to ensure high quality and suitability for the intended service, accidental damage can occur during shipment, storage, or installation. Most material tests are performed on a sampling basis that, in some cases, allow imperfections in sections of the cable to go undetected. RUS cable specifications are stringent and require that manufacturers conduct the tests recommended in this section and many more to demonstrate cable quality. Despite efforts to prevent damage during shipment and installation, and despite extensive testing by manufacturers, sometimes the installed cables contain defects that result in premature service failures. Many users, therefore, conduct their own quality and acceptance tests on new cable installations. There are numerous benefits to the cooperative that checks new cable quality and conducts

tests on new cable installations. The most important is improvement in service reliability. Experience demonstrates that, in most cases, damage to new cable during handling and installation causes service failures within two to three years. Failures require expensive repairs and consumer complaints and are costly to the cooperative. Replacement of cable often results in expensive landscape restoration or loss of service. When cable dimensions are outside the range of industry specification requirements, premolded joints and terminals will not fit properly. They may be too tight or too loose to apply to the cable. The latter may cause separation from the cable or water entry during service and shorten cable or accessory life, again requiring costly repairs and creating consumer dissatisfaction. Consequently, it is important that cable dimensions and concentricity be checked by the cooperative before the cable is installed.

Primary Cable Tests by the User

TESTING OF NEW PRIMARY CABLES Because the manufacturer is required to conduct many production and final quality assurance tests in accordance with RUS specifications, it is

likely that cables are of high quality. Therefore, it is not necessary for the user to conduct a large number of tests. It is recommended that the user, whether by himself or by a third-party testing

3 6 0 – Se c t io n 1 1

11 facility, make at least the following tests on newly received cables: • Dimensional conformance, including insulation thickness and concentricity and cable diameters, before cable installation; • Microscope examination for voids, contaminants, and protrusions; • Insulation shield strip test; and • High-voltage proof test of critical cables before placing in service (e.g., substation circuit exits). Additional quality assurance is achieved if a cable engineer is sent to the factory to witness tests or if the cooperative contracts with a recognized consultant to perform this service. The cooperative also is advised to conduct the hot silicone oil test on TR-XLPE cables to check for protrusions from shields, skips in the shields, voids in the insulation, and other irregularities on samples of newly received cables before installation. This sample test also can be contracted with an independent laboratory. Typically, short cable samples (approximately two feet long) are sent to a testing laboratory for examination and testing conducted at low cost and with a quick turnaround. As a general rule, it is recommended that cooperatives require tests to be performed on samples from the first and last reels of orders of fewer than 50,000 feet, with one extra sample for each additional 50,000 feet of cable. It is recommended that cooperatives notify their suppliers in advance that they will be sample testing. Further, they should establish responsibilities and procedures in case of a failure, such as the following: “Any evidence of noncompliance with the enclosed specifications shall be justification for: 1. Further testing at manufacturer’s expense (each shipping reel), 2. Rejection of the tested reel and possibly the reels preceding and following in the manufacturing process, and 3. Rejection of the entire order, depending on the severity and frequency of noncompliance.”

MEASUREMENT OF PRIMARY CABLE DIMENSIONS The measurement of the diameters and concentricity of cable and of selected cable components is an easy, straightforward way to check key parameters for compliance to RUS/ICEA specifications. This check also is important to ensure that the cable properly fits premolded commercial joints and terminations that are generally made by a company other than the cable manufacturer. Cable diameters and tolerances for primary cables are given in ICEA Specification S-94-649 and referred to in RUS Bulletin 1728F-U1 cable specification. Dimensions for the most popular cables are listed in Tables 11.1 (concentric stranding) and 11.2 (compressed stranding). The diameter over the conductor should be measured with a diameter tape or other suitable instrument readable to at least 0.001 inch (1.0 mil). When the diameter of a stranded conductor is determined by a micrometer or caliper, it should be measured around the circumference of the conductor perpendicular to the axis of the conductor and on the extension of a line through the center of the conductor and through the center of two wires in the outer layer that are 180° apart. The average of three measurements is taken as the diameter. The measurement of the insulation thickness should be made with a caliper after the extruded conductor and semiconducting insulation shields are removed from the cable core. If the conductor shield is bonded to the insulation, the thickness of the insulation can best be measured with a microscope. The average thickness of insulation must not be less than 220 mils for 15-kV cables, 260 mils for 25-kV cables, or 345 mils for 34.5-kV cables made to current RUS specifications. The measurement of the thickness of the insulation shield should be measured with a micrometer, caliper, or other suitable instrument readable to at least 0.001 inch (1.0 mil) after removing the shield from the cable. The diameter over the insulation shield should be measured with a diameter tape, micrometer, or other suitable instrument readable to at least 0.001 inch (1.0 mil). The measured values should be within the minimum and maximum values calculated in Tables 11.1 and 11.2.


11 TABLE 11.1: Dimensions for Primary Cables to ICEA Specification S-94-649-2000 with Concentric Neutral (Concentric Stranding). Conductor (Aluminum or Copper) Nominal AWG or Diameter kcmil (in.)

15-kV Cable (220 mils) Diameter (mils) Over Insulation Min. Max.

25-kV Cable (260 mils) Diameter (mils)

Over Shielding Min. Max.

Over Insulation Min. Max.





2 Solid

0.258

700

790

760

890

—

—

—

—

—

—

—

—

2

0.292

735

825

795

925

—

—

—

—

—

—

—

—

1 Solid

0.289

735

820

795

920

805

895

865

995

—

—

—

—

1

0.332

775

865

835

965

845

935

905

1,035

—

—

—

—

1/0 Solid

0.325

770

855

830

955

840

930

900

1,030

1,010

1,110

1,090

1,230

1/0

0.373

815

905

875

1,005

885

980

945

1,080

1,055

1,155

1,135

1,275

2/0

0.418

865

950

925

1,050

935

1,025

995

1,125

1,105

1,200

1,185

1,320

3/0

0.470

915

1,000

975

1,100

985

1,075

1,045

1,175

1,155

1,255

1,235

1,375

4/0

0.528

970

1,060

1,030

1,160

1,040

1,135

1,120

1,255

1,210

1,310

1,290

1,430

250

0.575

1,025

1,115

1,105

1,235

1,095

1,190

1,175

1,310

1,265

1,370

1,345

1,490

350

0.681

1,135

1,220

1,215

1,340

1,205

1,295

1,285

1,415

1,375

1,475

1,455

1,595

500

0.813

1,265

1,355

1,345

1,475

1,335

1,430

1,415

1,550

1,505

1,605

1,615

1,755

600

0.893

1,355

1,445

1,435

1,565

1,425

1,520

1,505

1,640

1,595

1,695

1,705

1,845

700

0.964

1,425

1,515

1,505

1,635

1,495

1,590

1,575

1,710

1,665

1,765

1,775

1,915

750

0.998

1,460

1,550

1,540

1,670

1,530

1,625

1,640

1,775

1,700

1,800

1,810

1,950

800

1.031

1,490

1,580

1,570

1,700

1,560

1,655

1,670

1,805

1,730

1,835

1,840

1,985

900

1.094

1,555

1,645

1,665

1,795

1,625

1,720

1,735

1,870

1,795

1,895

1,905

2,045

1,000

1.152

1,610

1,705

1,720

1,855

1,680

1,775

1,790

1,925

1,850

1,955

1,960

2,105

3 6 2 – Se c t io n 1 1

11 TABLE 11.2: Dimensions for Primary Cables to ICEA Specification S-94-649-2000 with Concentric Neutral. (Compressed Stranding). Conductor (Aluminum or Copper) Nominal AWG or Diameter kcmil (in.)

15-kV Cable (220 mils) Diameter (mils) Over Insulation Min. Max.








2

0.283

725

815

785

915

—

—

—

—

—

—

—

—

1

0.322

765

855

825

955

835

925

895

1,025

—

—

—

—

1/0

0.362

805

895

865

995

875

965

935

1,065

1,045

1,145

1,125

1,265

2/0

0.405

850

935

910

1,035

920

1,010

980

1,110

1,090

1,190

1,170

1,310

3/0

0.456

900

985

960

1,085

970

1,060

1,030

1,160

1,140

1,240

1,220

1,360

4/0

0.512

955

1,045

1,015

1,145

1,025

1,115

1,105

1,235

1,195

1,295

1,275

1,415

250

0.558

1,010

1,100

1,090

1,220

1,080

1,175

1,160

1,295

1,250

1,350

1,330

1,470

350

0.660

1,115

1,200

1,195

1,320

1,185

1,275

1,265

1,395

1,355

1,455

1,435

1,575

500

0.789

1,240

1,330

1,320

1,450

1,310

1,405

1,390

1,525

1,480

1,580

1,560

1,700

600

0.866

1,325

1,415

1,405

1,535

1,395

1,490

1,475

1,610

1,565

1,670

1,675

1,820

700

0.935

1,395

1,485

1,475

1,605

1,465

1,560

1,545

1,680

1,635

1,740

1,745

1,890

750

0.968

1,430

1,520

1,510

1,640

1,500

1,595

1,580

1,715

1,670

1,770

1,780

1,920

800

1.000

1,460

1,550

1,540

1,670

1,530

1,625

1,640

1,775

1,700

1,805

1,810

1,955

900

1.061

1,520

1,610

1,630

1,760

1,590

1,685

1,700

1,835

1,760

1,865

1,870

2,015

1,000

1.117

1,575

1,670

1,685

1,820

1,645

1,740

1,755

1,890

1,815

1,920

1,925

2,070

CALCULATION OF DIAMETERS OF PRIMARY CABLES The allowable diameter and tolerances for 15-, 25-, and 34.5-kV RUS specification cables (Table 11.1) can be calculated as indicated in Table 11.3. The nominal diameter over the cable insulation shield can be calculated by adding the appropriate nominal adder for extruded insulation shield

shown in Table 11.4 to the nominal diameter over the insulation. Calculated cable diameters for primary cables with conductor sizes from #2 AWG through 1,000 kcmil are given in Tables 11.1 and 11.2. The diameter over the cable insulation or over the insulation shield may be measured with a diameter tape or it may be calculated.


11 TABLE 11.3: Cable Diameter Tolerances. Conductor Size (AWG or kcmil)

Equation 11.1

Diameters Over Insulation (mils) Minimum Nominal Maximum

2–4/0

C + 30 + A + 2T

Add 30

Add 60

250–500

C + 40 + A + 2T

Add 30

Add 60

600–1,000

C + 50 + A + 2T

Add 30

Add 60

where: C = Conductor diameter The second term (30, 40, or 50) is twice the extruded conductor shield thickness. A = Adder (10 mils) for 25-kV cable; do not use for 15-kV cable. T = Minimum average insulation thickness (RUS Bulletin 1728F-U1). 1 mil = 0.001 in. Note.

If a conductive tape and a subsequent extruded shield have been applied over the conductor, the minimum diameter over the insulation must be calculated by Equation 11.1.

Dmin = C + 80 + A + 2T where: Dmin = Minimum diameter over insulation

Table 11.4. Adders for Extruded Insulation Shield (Mils) to Obtain Nominal Diameter Over Insulation Shield of Cable. Calculated Minimum Diameter Over Insulation (inches)

Extruded Insulation Shield Adders (mils) Minimum Nominal Maximum

0–1.000

50

100

150

1.001–1.500

70

120

170

1.501–2.000

110

160

210

1 mil = 0.001 in.

EXAMPLE 11.1: Diameter Calculation. Calculate minimum, nominal, and maximum diameter over cable insulation for 1/0 AWG concentric (Class B) stranded conductor with extruded conductor and insulation shields, 220 mils insulation thickness, with jacket, 15 kV. C = Conductor diameter For the conductor shield extruded over the conductor (see Table 11.3) A 2T = Insulation Thickness (2 × 220) Calculated diameter over the cable insulation

= 373 mils = 30 mils = 0 mils = 440 mils = 843 mils

Round up

= 845 mils

Nominal diameter over insulation (add 30 mils per Table 11.3) Maximum diameter over insulation (add 60 mils per Table 11.3)

= 875 mils = 905 mils

Calculate minimum, nominal, and maximum diameter over cable insulation shield. Nominal diameter over cable insulation = 875 mils Minimum adder of insulation shield (Table 11.4) = 50 mils Minimum total diameter over insulation shield = 925 mils Nominal diameter over cable insulation Nominal adder of insulation shield (Table 11.4) Nominal total diameter over insulation shield

= 875 mils = 100 mils = 975 mils

Nominal diameter over cable insulation Maximum adder of insulation shield (Table 11.4) Maximum total diameter over insulation shield

= 875 mils = 150 mils = 1,025 mils

3 6 4 – Se c t io n 1 1

11 HOT SILICONE OIL TEST FOR XLPE AND TR-XLPE PRIMARY CABLES This test is a quick and easy way to check for the cleanliness of the insulation, presence of voids in the insulation, the smoothness of the interface between the insulation and the conductor shield, and presence of skips in the extruded shields. It is possible to perform this test because the polyethylene insulation becomes transparent when raised to high temperature. The test is not useful for EPR cables because that insulation does not become transparent. Suspend a sample of cable, approximately 10 inches long, with the jacket, neutral, and insulation shield removed, in a clear glass container filled with clean silicone oil that has been heated to about 180°C (356°F). A typical test setup is shown in Figure 11.1. Look for contaminants. Also look for protrusions of the conductor shield into the insulation and for skips in the conductor shield. If any are found, do not install

the cable. Contact the cable manufacturer or a cable consultant for further advice. The determination of the actual size of any observed protrusion must be done by cutting the insulation into wafers and examining them with a microscope or an optical comparator. The description of imperfections and their limitations as to size allowable in RUS cable specifications are discussed fully in ICEA Specification S-94-649. INSULATION SHIELD STRIPPING TEST FOR PRIMARY CABLES This test is performed to demonstrate that the insulation shield can be removed from the insulation by normal workmanship and to demonstrate that no conducting material is left on the surface of the insulation upon removal of the shield. STEP 1. Use a cable sample approximately 15

inches long. STEP 2. Cut the semiconducting shield longi-

tudinally and vertically down to the insulation. STEP 3. Make a second, similar cut at 1/2-inch

separation from the first cut. STEP 4. Construct a suitable measuring device

and use it to measure the tension required to pull away the 1/2-inch-wide strip of insulation shield from the cable. See the arrangement in Figure 11.2. STEP 5. Attach the measuring device by remov-

ing approximately two inches of the 1/2-inch strip of each end of the cable by pulling it away at a 90° angle from the cable. STEP 6. Measure the pulling tension in pounds

FIGURE 11.1: Setup for Hot Silicone Oil Test. Two cable samples in a hot silicone oil bath show the transparency of polyethylene insulation. The oil temperature is approximately 180°C (356°F).

by increasing the force on the strip until the strip separates from the insulation at a pulling speed of approximately 1/2 inch per second. Make the test at ambient temperature, one test at each end of the cable in opposite directions and 180° apart. The minimum allowable tension


11 is six pounds and the maximum allowable tension is 18 lb. for TR-XLPE, and three pounds minimum and 18 lb. maximum for EPR insulation. No conducting material that is not readily removable may be left on the insulation and the insulation must not be damaged. A local job shop should be able to fabricate the measuring device, or an outside consultant can advise on where to obtain a device or how to construct it. The stripping test is an ICEA specification and, therefore, an RUS specification test requirement; consequently, cable manufacturers possess this measuring device. Further details of this test, plus additional information related to this test, are given in ICEA specifications.

FIGURE 11.2: Setup for Insulation Shield Stripping Test. Tension is measured as a 1/2-inch-wide strip of insulation shield is removed from an XLPE cable.

HIGH-VOLTAGE PROOF TEST FOR PRIMARY CABLES The high-voltage (dc) test is an important acceptance test made on primary cable before the cable is placed in regular service. This test is conducted with the cable joints and either temporary or regular terminations connected to the cable. All other devices (e.g., lighting arresters, transformers) are disconnected from the cable except the regular grounding devices, which remain connected to the cable as used in service. The proof test has the advantage of indicating the condition of the insulation under high-voltage stress conditions. A high-voltage proof test is recommended for large installations, for important feeder cables, or where continuity of the power is of paramount importance (e.g., substation circuit exits). The proof test may not be necessary for small installations of less importance when testing of every piece of cable becomes quite time consuming. Test equipment manufacturers provide lightweight portable dc test equipment along with complete instructions on how to perform proof tests on cable systems.

Preparation for High-Voltage Proof Test The following preparations for the high-voltage proof test for primary cables are recommended: • Read the test equipment manufacturers’ testing recommendations, when available, before performing this test. • Read IEEE Standard 400 covering safety procedures and dc testing techniques. See the references for the sources of these publications. • Operate the test equipment to become familiar with the instruments and how to interpret their readings. • Decide what level of test voltage to use and the time of voltage application. The recommended values are given in Table 11.5. It also is desirable to obtain the cable manufacturer’s recommendation for these test voltages. • Keep personnel out of the area in which the test is to be performed and the areas at both ends of the cables being tested.

3 6 6 – Se c t io n 1 1

11 • Erect barriers between the test area and its surroundings and display signs warning of high-voltage testing. Commercial signs are available for this purpose. • Determine test values for cable terminations, joints, etc., which may have limits on testing voltages. • Always have at least two people present during the test. Setup for Proof Test of Primary Cables The following test setup is recommended when conducting the proof test: • When testing the cable make sure all other equipment—such as surge protective devices, transformers, etc.—is disconnected. • Make sure the cable terminals are clean, dry, and free of sharp points. Sharp points cause corona and flashovers that can be eliminated by covering with commercial electrical putty such as 3M Scotchfil™ or clear plastic bags. Elbow terminations need to be packed in insulated packing stands, or covered with clear plastic bags. It is desirable to conduct the test with the cable terminals installed. If the test is performed before connecting the terminals, remove the insulation shield for about 18 inches and clean the exposed insulation. Although it is usually not necessary, a commercial prefabricated terminal having a voltage rating at least as high as that of the cable also can be applied in accordance with the manufacturer’s instructions. • Check the circuit with a 500- or 1,000-volt megohmmeter to make sure that there are no obvious problems before starting the proof test. Allow a clearance between cable ends and surrounding objects of at least one inch per 10 kV. • Check the high-voltage test set by suspending its high-voltage lead off the ground with a dry plastic cord. Turn the voltage up to the highest value to be used during the test. The micro-ammeter should read close to zero. Conducting the Proof Test After making the proper preparations for the high-voltage dc proof test described above, follow these steps:

STEP 1. Connect the high-voltage test set to the

conductor of the cable under test. STEP 2. Switch on the test set. Raise the voltage

from zero to the test voltage selected from the applicable AEIC specification value indicated in Table 11.5. Raise the voltage slowly, so as to reach the desired level in one to one-and-one-half minutes. STEP 3. Maintain the voltage for the preselected time. For acceptance of new cable not yet placed in service, the AEIC specification recommended time is five minutes during installation or 15 minutes after installation. Use the time indicated in Table 11.5 or the time recommended by the cable manufacturer. The five-minute time of application of voltage is specified during cable installation because temporary terminations are often used during the test. These terminations may have corona discharge or high leakage currents that can cause damage to the cable ends. Most cable damage caused during shipment or installation can be detected during the shorter five-minute test. In addition, less space charge will build up in the cable during the shorter time test. STEP 4. Reduce voltage and switch off the test set. If a flashover occurs during the test, turn the test set off immediately. A flashover tells you that an insulation breakdown has occurred, a termination has flashed over, or the test set has failed. If external flashover occurred during the testing, check to see if the correct voltage was applied. Clean the cable terminations and reposition the test leads if they were too close or became separated from the line. If flashover occurred inside the test set, check to be sure you supplied the correct voltage for the correct time. Internal flashovers may indicate problems with the equipment. STEP 5. With the test set voltmeter indicating zero, discharge the test set to ground. Use an approved discharge stick with


11 proper resistance followed with a grounding stick to drain the charge from the tested cable. When the voltage drops below about 1 kV, connect the grounding stick to the cable terminal and keep the cable grounded for a period of time equal to four times the length of the dc test. Draining the charge to ground on tested cables is a vital safety procedure. If the cable withstood the voltage for the recommended time period, and if the leakage current reached a steady-state condition, the insulation on the cable system—including the joints and terminations—is suitable for service. If the test voltage caused flashover before the end of the time period or if the test set overcurrent circuit tripped, the insulation of the system is not satisfactory to place in regular service. If the

TABLE 11.5: DC Proof-Test Voltages (Conductor to Ground) for Primary Cables. Cable Rated Voltage φ – φ (kV)

Period of Test

DC Voltage (kV)

Time of Application (minutes)

Insulation Thickness — 220 mils 15

During installation

60

5

15

After installation, before service

64

15

15

During first 5 service years

52

5

15

After 5 service years

32

5


During installation

75

5

25


80

15

25


65

5

25


40

5


During installation

93

5

35


100

15

35


81

5

35


50

5

leakage current continued to gradually increase in any step of the test, lengthen the time of that step until it can be determined the leakage current will stabilize. It may be necessary to remove the cable terminations and repeat the high-voltage proof test to determine if the cable under test, rather than the terminations, has failed. If it is desired to repair the cable and if the exact location of the failure is not known, use fault location equipment to locate the fault. Most high-dc-voltage proof testers contain a fault-locating device (thumper) that may be used to pinpoint the exact location of a cable fault. When this thumping equipment is used, the fault is found by listening for a thumping sound, at the cable failure, while walking along the ground above the cable. A pick-up coil and earphones can be employed to facilitate hearing the thumping. Manufacturers of proof test equipment and thumping equipment provide detailed instructions on the operation of their equipment and its use to locate cable faults. Obtain these instructions and follow them when locating cable faults. HIGH-VOLTAGE STEP TEST FOR PRIMARY CABLES The step-voltage test is a variation of the proof test, and the same preparations and procedures as with the proof test are used. The step-voltage test is particularly applicable to cable circuits because if the cable, joint, or termination has an incipient fault, a flashover will occur at a voltage below the failure point. This will reduce the likelihood of damaging good cable. When the step voltage is used, it is desirable to use voltage steps of the same magnitude, divided in equal times between zero and the maximum test voltage level. For example, a 15-minute test involves five three-minute steps increased in equal amounts at each step. As the cable current at each higher voltage increases, and then decreases to a steady value in less than one minute, this will allow time to read the current (or to show unreadably low values). The test set microammeter should be readable to 0.1 microampere. At each step, the current for satisfactory cable should be steady and should be recorded. If, at any step, the current begins to

3 6 8 – Se c t io n 1 1

11 increase rapidly, this indicates impending failure on the cable system. The test set should be turned off and left off for a period of time equal

to four times the length of the dc test to ground the system. A high and/or unstable leakage current is most likely due to contamination on the terminations, too small a clearance between energized and grounded components, or a defect in the test equipment. Check and reclean the terminations, recheck all clearances, and perform a leakage current test on the dc equipment by raising the voltage with the output open circuited. After performing these steps, repeat the step test on the cable system. If high or unstable leakage current is still observed, the problem is probably with the cable system and the proof test voltage must be applied for the required time to determine the location of the defect. TYPICAL EQUIPMENT FOR HIGH-VOLTAGE DC TESTS The portable, high-voltage dc test set for conducting the proof and step-voltage tests should have the maximum test voltage (usually negative polarity), a means of increasing the voltage continuously or in small steps, satisfactory output voltage stability, the output voltage filtered to provide pure dc voltage, and 0.1 microampere resolution. Commercial test equipment usually complies with these requirements. A typical commercial test set is shown in Figure 11.3. A discharge resistor used to discharge the cable after test should have a resistance of not less than 10,000 ohms per kV of test voltage. Commercial discharge resistors are designed to withstand the full test voltage without flashovers and to withstand the discharge energy without overheating. They have an insulating hook stick and a flexible conductor to connect the resistor across the cable terminal and ground. A typical commercial discharge and grounding stick is shown in Figure 11.3.

FIGURE 11.3: Typical High-Voltage DC Test Set with Cable Grounding Probe.

PROOF TEST PRECAUTIONS WHEN SPLICING TO SERVICE AGED CABLES When new primary cable is spliced to existing service-aged PE, TR-XLPE, or EPR cables, special precautions must be observed when conducting a high-voltage dc proof test. The existing in-ser-


11

Secondary Cable Tests by the User

vice cables most likely contain water trees in the insulation as a result of moisture penetration from the environment. The jacket slows, but does not eliminate, this penetration. Water trees lower the breakdown voltage of the cable. In severe cases of water treeing, the dc test voltage may be high enough to further damage adjacent cable and cause it to fail in service prematurely. This precaution need not be taken when splicing to paper-oil insulated cable as this type of cable is not damaged by dc testing. This characteristic of PE, TR-XLPE, and EPR is recognized by the ICEA specifications and is reflected in the recommended dc test voltages. In

Table 11.5 note that the recommended dc test voltages are reduced during the first five years of service. They are further reduced when the cable has been in service for more than five years. It is highly recommended that the reduced values be used if the dc tests are performed. The above precautions apply to all maintenance testing of PE, TR-XLPE, and EPR cable, both jacketed and unjacketed. It should be noted that RUS-specification cables have thicker insulation than do ICEA-specification cables. The thicker insulation retards but does not eliminate water tree formation. Therefore, the lower test levels given in Table 11.5 should be used.

TESTING OF NEW SECONDARY CABLES As for primary cables, the manufacturer is required to conduct many production and quality assurance tests on secondary cables. Because 600-volt cables operate at a low voltage stress, they are much less likely to fail (from electrical stress) in service than are primary cables. Consequently, the cooperative is advised to make only the following tests on newly received cables before installation:

MEASUREMENT OF SECONDARY CABLE DIMENSIONS The most important test for secondary cables before installation is the measurement of the insulation thickness. The insulation thickness should be in accordance with RUS Specification for 600Volt Underground Power Cable (RUS Bulletin 1728F-U2) and is given in Table 11.6 for the applicable conductor size. The thickness of the insulation (or composite insulation) should be measured with a caliper, steel ruler, or micrometer. The average thickness must be taken as one-half of the difference between the mean of the maximum and minimum diameters over the insulation (or composite insulation) at one point and the average diameter over the conductor or any separator measured at the same point. The minimum of the insulation (or composite insulation) must be taken as the difference between (1) a measurement made over the conductor or any separator plus the thinnest insulation (or composite insulation) wall, and (2) the diameter over the conductor or any separator. The first measurement must be made after slicing off the opposite side of the insulation (or composite insulation). The thickness of any separator between the conductor and the covering must not be included in the thickness of the insulation.

• Insulation thickness, and • Concentricity of insulation. Some users conduct a high-voltage proof test of installed secondary cable before placing the cable in service. This practice is not recommended unless it is believed that the cable was damaged during installation. TABLE 11.6: Insulation Thickness of Secondary Cables. Standard XLPE 1 Layer Average (mils)*

Ruggedized XLPE 2 Layer Average (mils)*

4–2 AWG

60

60

1–4/0 AWG

80

80

225–500 kcmil

95

95

Conductor Size

*The minimum thickness will not be less than 90% of the average.

3 7 0 – Se c t io n 1 1

11 PROOF TEST OF SECONDARY CABLES It is not customary or generally necessary for a cooperative to proof-test underground secondary power cable that complies with RUS specifications, including those that also are marked as complying with the Underwriters Laboratories Type USE cable, before placing it in service. If there is reason to believe that the cable has been damaged during installation, an insulation tester can be used to check the cable. It is best to conduct the test when the ground or conduit is wet because there is no integral ground plane as with a primary cable. If the cable complies with the requirements of Table 11.6, it is not likely that a failure will occur when the proof test is performed unless substantial damage occurs during handling or installation. In a proof test, the cable should be removed from the circuit. It is not necessary to disconnect joints and terminals from the cable. The voltage should be about 3,000 volts ac applied for one minute after being increased from zero over a period of 60 seconds. AEIC specifications do not cover 600-volt power cable as is the case with

Tests by the Cable Manufacturer

Various types of tests are performed by the manufacturer when it is making primary cables to comply with RUS specifications. The manufacturer is required to continually conduct many electrical and physical tests on TR-XLPE and EPR insulated cables. The testing may be separated into two categories: qualification tests and production tests on samples and on a full reel of cable. MANUFACTURER QUALIFICATION TESTS ON PRIMARY CABLES These tests are intended to demonstrate the capability of the manufacturer to furnish high-quality cable with the performance characteristics suitable for RUS member systems. Before the manufacturer and the cable design is accepted by RUS, certified test data on a particular design must be submitted to RUS showing compliance with the RUS specifications, which include applicable requirements of ICEA Specifications S94-649. If requested by the purchaser, the manufacturer is required to furnish a certified

primary cables. ICEA and Underwriters Laboratories specifications do not suggest voltage tests after installation for secondary cables. The proof-test equipment used for this test is generally a portable insulation tester (Hipot) test set. A 1,000-volt megohmmeter also may be used for this test. In this case, satisfactory cable should have an insulation resistance not less than one megohm. Equipment manufacturers’ instructions also should be followed when making a proof test or an insulation resistance test. INSULATION RESISTANCE TEST OF SECONDARY CABLES Some users routinely measure the insulation resistance of new cable before placing it in service. This test is not advised for primary cables because many factors influence the IR reading and results are meaningful in only a few cases. For secondary cables, the measurement of insulation resistance is useful, as indicated previously, as a means of determining whether cable has been severely damaged during installation.

copy of the qualification test data of the cable being purchased. If a manufacturer changes the insulation or the semiconducting conductor or insulation shields, that cable with the new components must also be qualified to RUS specification. Each combination of overall jacket material, jacket application method, concentric neutral design, insulation shield type, insulation type, and conductor size range must be subjected to selected qualification tests to prove adequate performance before acceptance by RUS. The qualification tests ensure that the cable design represents a high-quality, state-of-the-art product. MANUFACTURER PRODUCTION TESTS ON PRIMARY CABLES These tests are conducted on a sampling basis during production to ensure that cable performance is equivalent to that of the cable that received qualification approval, to ensure compliance with the applicable ICEA specification, and to detect any manufacturing defects. Numerous


11 tests on the insulation, the semiconducting shields, and mechanical properties of the cables are conducted during cable fabrication to ensure quality. Partial discharge and voltage withstand tests are conducted on full reel lengths of completed cables. The requirements for the voltage withstand tests are given in Table 11.7. The dc voltage test values are higher than those recommended in Table 11.5 for testing installed new cables. MANUFACTURER PRODUCTION TESTS ON SECONDARY CABLES The cable manufacturer is required by RUS to continually conduct tests on 600-volt XLPE and ruggedized composite XLPE-insulated cables. These tests are fewer and less stringent than those required by RUS for primary cables. The testing may be separated into three categories:

TABLE 11.7: Manufacturers’ Voltage Withstand Tests on Completed Cable. Cable Rated Voltage φ – φ (kV)

Minimum Nominal Insulation Thickness (mils)

5-Minute AC Test Voltage (kV)*

15-Minute DC Test Voltage (kV)*

15

220

44

80

25

260

52*

95*

35

345

69

125

35

420

84

155

*Withstand voltages are based on insulation thickness.

TABLE 11.8: Manufacturers’ Voltage Tests on Cables Rated Zer0 to 600 Volts. Conductor Size (AWG or kcmil)

AC Test Voltage (kV)

DC Test Voltage (kV)

AC Spark Test DC Spark Test Voltage (kV) Voltage (kV)

4–2

5.5

16.5

15.0

21.0

1–4/0

7.0

21.0

17.5

28.0

225–500

8.0

24.0

20.0

33.5

Note: The manufacturer is required to conduct at least one of the above voltage tests in accordance with NEMA WC-7/ICEA S-66-524 (XLPE) or ICEA P-81-570 (Ruggedized Extruded Insulation).

1. RUS acceptance and listing of cable tests, 2. Routine production tests, and 3. Completed cable tests. As part of the RUS acceptance and listing of cable, the manufacturer must include certified test data demonstrating compliance with RUS Specification for 600-Volt Underground Power Cable (RUS Bulletin 1728F-U2). The manufacturer must conduct routine production tests required by NEMA Standard WC-7/ICEA 5-66-524 or ICEA P-81-570. Manufacturers’ voltage tests on cable rated 600 volts are given in Table 11.8. MANUFACTURER’S CERTIFIED TEST REPORTS The cable user is urged to specify copies of certified test reports (CTRs) for primary cables at the time of ordering. The manufacturer will furnish certified copies of the qualification test results representative of the cable being purchased and of the actual production test values. These should be compared with the cable specifications. When cable is not shipped directly from the manufacturer but through a local distributor, the latter may provide typical performance test data for the type of cable being used. These reports provide useful information in the following circumstances: • When changes or additions are made on the cable system so that similar cable can be ordered, • When problems arise with the performance of the cable, and • When selecting joints and terminals for the system. As with primary cables, the manufacturer must submit certified test data demonstrating compliance with the applicable secondary cable specifications. For secondary cables, these specifications are RUS Specification for 600-Volt Underground Power Cable (Bulletin 1728F-U2) and ANSI/ICEA S-66-524 for TR-XLPE cables and ICEA P-81-570.

3 7 2 – Se c t io n 1 1


Cable testing should be performed to ensure conformance to specifications before installation and to ensure that the cable was not damaged during installation. 1. Conduct dimensional conformance, hot silicone oil, and insulation stripping tests before installation to identify defects that cause early failures. 2. Conduct a high-voltage dc acceptance test when installing a large quantity of the same type of cable at a given site or important feeder cables, or when continuity of the power supply is of paramount importance. Contract with an outside laboratory or use in-house equipment to conduct the hot silicone oil and the insulation shield stripping tests. Apparatus for the latter test may be purchased on the outside. 3. Use a diameter tape, caliper, or micrometer or equivalent device that can measure to 0.001 inch (1.0 mil) to measure dimensional conformance.

4. Use commercial high-voltage dc test equipment. Manufacturers’ tests ensure a high-quality product at the time of cable shipment. They also show that the cable design and materials have the capability of operating satisfactorily in service. Cable specifications ensure that the manufacturer fabricates the cable to be suitable for use with standard joints and terminations and complies with the cooperative’s order for the cable. Cable specifications provide a means to document test results in the form of a written and certified test report. • Cooperatives that do not have the capability or equipment to conduct cable tests should make arrangements for an outside testing facility. • In all cases, the manufacturer’s certified test reports (CTRs) must be obtained.

Calculations for Reliability St u d i e s – 3 7 3

A In This Appendix:

Reliability Index

Calculations for Reliability Studies

Reliability Index

Calculation of Reliability

Acceptability Criteria

Importance of Sectionalizing

The reliability index that is probably quoted most often in the literature is the average service availability index (ASAI). This index is defined as the ratio of total consumer-hours of available service to total consumer-hours demanded. From a particular consumer’s point of view, this index could be viewed as the ratio of total hours of available service per year to the number 8,760, which is the total number of hours in a year. Simple mathematical formulas relate the ASAI index to the total hours interrupted per year. The total-hours index is also called the System Average Interruption Duration Index (SAIDI). The formulas relating ASAI and SAIDI are shown in Equation A.1.

For a sample application of Equation A.1, assume a feeder has experienced outage times amounting to 3.5 consumer-hours per consumer per year (SAIDI). The corresponding ASAI figure is as follows:

ASAI =

8,760 – 3.5 = 0.9996 8,760

The ASAI number is interpreted to mean that a typical consumer served from the feeder can expect electric service to be present 99.96 percent of the time. The second part of Equation A.1 can be used to calculate the SAIDI from the ASAI:

Equation A.1 ASAI =

8,760 – SAIDI 8,760

SAIDI = 8,760 × (1 – ASAI)

SAIDI = 8,760 × (1 – 0.9996) = 3.5 The total-hours index (SAIDI) will be used in this analysis because it is less abstract than the ASAI.

3 7 4 – Ap p e n d i x A

A Acceptability Criteria

Acceptable reliability criteria are defined in RUS Bulletin 161-1. Table A.1 summarizes the guidelines of that bulletin.

TABLE A.1: Acceptable Outage Hours Per Year Per Consumer. Type of Area

Maximum Acceptable Outage Hours Per Year

Urban

1.0

Rural, Near Urban Areas

2.0

Remote Rural

5.0

Calculation of Reliability

Although loop-feed UD designs are recommended and normally used, these designs are certainly not equivalent to automatically transferred dualfeed designs. Transfer to the alternate feed in a typical loop-feed UD design requires human intervention after an outage. Consequently, the loopfeed design does nothing to reduce the frequency of outages. The advantage of the loop-feed

Although cooperatives are traditionally associated with rural areas, UD systems installed by cooperatives are likely to be in or near urban areas. It is thus reasonable to design UD systems to meet the acceptable outage time criterion of one hour per year. Furthermore, the RUS guidelines represent the total outage time to an individual consumer and include the outage time of the overhead system supplying the UD system. Therefore, the design outage time of the UD system itself should be much less than one hour per year.

design is in reducing the duration of cable-failure outages; this is reviewed in Section 1. Reliability analysis of a UD system is performed in the same way as for any radial distribution system. The critical system components delivering the power are considered to be in series from a reliability point of view, which means that failure of any such component causes an

Overhead Line

Surge Arrester Cable Termination


Elbow Terminator

Secondary Connections

Elbow Surge Arrester

Secondary/Service Cable Primary Cable

Primary Cable

FIGURE A.1: Components Affecting Outage Rate to the Consumer.

Calculations for Reliability St u d i e s – 3 7 5

A outage at the point under study. Therefore, the outage rates associated with the individual components may be added arithmetically to determine the outage rate expected for the consumer. This direct addition rule also applies to the total hours of outage per year for the consumer as calculated from component total outage hours. Calculation of the expected reliability of service to a single-phase consumer on a UD system proceeds as follows: STEP 1. Identify the critical components. For UD

systems, these are the supply to the riser structure, the terminator at the riser, the total cable length energized from the riser, the elbow terminators, any lightning arresters connected to the primary system section, the pad-mounted

Importance of Sectionalizing

An outage at an individual point on a UD system occurs when an overcurrent device operates. Therefore, the calculation of reliability must encompass all the UD cable length and other vulnerable components that will make any overcurrent device operate and interrupt service to the point under study. This collection of components that might fail is referred to as the total “exposure” to outages that is associated with a particular study point. This situation implies that reducing the total exposure will reduce the outage rate and hours of outage at a point on the system. Limiting the cable length and other components protected by each overcurrent device reduces exposure. Reducing exposure can be accomplished in two ways: 1. By increasing the number of circuits used to serve an area from a supply point, and 2. By installing coordinated overcurrent devices at several selected points on a large circuit. The first method is more effective, but also more costly. The second method yields mixed effectiveness, because consumers located far from the supply point do not benefit from exposure reductions as much as do consumers near the supply point.

transformer, the secondary connections, and the secondary cables. STEP 2. Determine the outage rates and restoration times for each component. It will normally be necessary to pool the experience of many utilities to get the large statistical base needed to achieve accuracy for these parameters. STEP 3. Determine total outage hours per year for each component by multiplying the component outage rate by the component detection and restoration time. STEP 4. Sum the component outage rates to get the expected outage rate at the consumer. STEP 5. Sum the component outage hours to get the expected total outage hours for the consumer.

The concept of dividing the area to be served into many overcurrent-device sections to improve reliability is known as sectionalizing. Sectionalizing is often the most cost-effective way to improve reliability. However, if there are serious reliability problems with critical system components, sectionalizing alone may not achieve acceptable reliability. As an example of the effectiveness of sectionalizing, consider a single-phase UD area consisting of four cable runs in the configuration illustrated in Figure A.2. The area is supplied from the west over a 3,000-foot cable run ending in a junction enclosure from which three additional 3,000-foot cable runs feed north, east, and south. Each cable run, including the supply run, serves 20 pad-mounted transformers by feed-through bushings. This configuration requires 41 elbow terminations per cable run, including the elbows in the junction enclosure and on the arrester at the last transformer. The following outage rates might be typical for the primary UD components:

Cable = 0.0020 failures/kft/year Elbow = 0.0001 failures/year

3 7 6 – Ap p e n d i x A

A If no sectionalizing is installed on the UD system, the outage rate expected for each consumer on the system is calculated as shown in Equation A.2.

Equation A.2 Cable Outage Rate = 12 kft × 0.0020 failures/kft/year = 0.0240 outages/year Elbow Outage Rate = 164 elbows × 0.0001 failures/year = 0.0164 outages/year Outage Rate for Each Consumer = 0.0240 + 0.0164 = 0.0404 outages/year

N

Open Point

20 Transformers

Riser

20 Transformers

20 Transformers Sectionalizing Enclosure

This outage rate represents the primary UD system only. Each consumer will also be exposed to outages on transformers, secondaries, and the overhead primary system that supplies the UD area. The consumer outage rate caused by UD primary facilities can be reduced by using a sectionalizing enclosure instead of a junction enclosure at the center of the UD system. The sectionalizing enclosure will provide fuse protection for the cable runs to the north, east, and south. With the sectionalizing devices in service, the primary UD outage rate to consumers on the west cable run will become one-quarter of the previous cable because three-quarters of the failures will be cleared by sectionalizing fuses in the central enclosure. For consumers on the north, east, and south cable runs, the outage rate will be one-half the previous value. These consumers will be without service whenever a failure occurs in their own sectionalized area or in the supplying run from the west. The resulting consumer outage rates from primary UD failures with the sectionalizing devices in service are as follows:

West Cable Area

0.0101 outages/year

North, East, or South Cable Area

0.0202 outages/year

Open Point

The outage rate is 0.0404 per year for all consumers without the sectionalizing cabinet. Open Point

FIGURE A.2: Sectionalized UD Area.

20 Transformers

Transformer and Secondary Voltage Dro p – 3 7 7

B In This Appendix:

Transformer and Secondary Voltage Drop

Voltage Flicker

Secondary voltage drop consists of two components: 1. The transformer voltage drop, and 2. The secondary and service voltage drop to the point of delivery. The total drop allowed in RUS Bulletin 169-4 is four volts on a 120-volt base, or 3.33 percent (see Table B.1). Closer to the substation, a maximum of six volts (five percent) of combined drop is allowed. This threshold recognizes that the full eight-volt drop allowed on the primary system probably will not occur at locations close to the substation. However, the engineer needs

TABLE B.1: Allowable Voltage Drop on a 120-Volt Base. Maximum Drop (Volts)

Percentage Drop

Substation regulated bus (output) to last distribution transformer (primary)

8

6.67

Distribution transformer (primary) to service delivery connection to consumer’s wiring (meter or entrance switch)

4

3.33

4 6

3.33 5.00

Utility service delivery point (meter or entrance switch) to consumer’s utilization terminal (outlet): • Loads including lighting • Loads without lighting

to ensure that voltage level at the point of delivery to the consumer are consistent with levels outlined in RUS Bulletin 169-4 and/or ANSI Standard C84.1. Several different types of service configurations may be present at the utilization voltage level. The following are the most common: • • • •

Three-phase, four-wire, wye; Three-phase, four-wire, delta; Three-phase, three-wire, wye or delta; Two phases and neutral from three-phase, four-wire, wye system; • Single-phase, three-wire; and • Single-phase, two-wire. To further complicate the situation, all but the last listed above may be subject to unbalanced load conditions. Each type of service, along with the extent of load unbalance for each, creates a unique situation requiring customized techniques for calculating voltage drop. In many cases, the engineer can overcome the complications involved with the many configurations by identifying a worst-case single-phase situation that is embedded in a multiphase situation. An engineer who is skilled in both singlephase and balanced three-phase voltage drop calculations can, thus, usually find the worstcase voltage drop involved in a complicated, unbalanced situation.

3 7 8 – Ap p e n d i x B

B The simplest service voltage drop problem is also the most common: single-phase, 240-volt, three-wire service, with the load balanced between the two 120-volt legs. It is reasonable to assume a balanced load on this type of service as the larger appliances are almost always connected line to line. The balanced 120-volt load cancels out the neutral current, so the only impedances that need to be considered are the transformer impedance and the ungrounded

Equation B.1 For resistive circuit with 100% power factor load: VDROP = IR where: VDROP = Voltage drop, in volts I = Current flowing in circuit, in amperes R = Supply circuit resistance, in ohms

Equation B.2 VDROP = IDR + IQX where: VDROP ID R IQ X

= = = = =

Voltage drop, in volts Real component of current, in amperes Supply circuit resistance, in ohms Reactive component of current, in amperes Supply circuit reactance, in ohms

Equation B.3 ID = Icosθ IQ = Isinθ where: ID I cosθ IQ θ

= = = = =

Real component of current, in amperes Measured load current, in amperes Power factor in decimal form Reactive component of current, in amperes Power angle, the arc cosine of the power factor

conductor impedance on one leg. This procedure will permit direct calculation of voltage drop on a 120-volt base. The same calculating procedure used for one leg of a balanced single-phase 120/240-volt system can be used for one phase of a balanced 208/120-volt, three-phase, four-wire, wye system. As 208/120-volt services are also very common, skill in performing the basic single-phase voltage drop calculation is a valuable tool for an engineer. Voltage drop on a purely resistive circuit serving a 100 percent power factor load is very simple to calculate from Ohm’s law, as shown in Equation B.1. In actual cases, however, the supply circuit is not purely resistive and the load is somewhat less than 100 percent power factor. This combination of both the circuit and the load current having an inductive component produces a vector load current interacting with a supply circuit impedance, itself a vector quantity. Their product, IZ, is also a vector quantity that will be somewhat out of phase with the source voltage. The voltage drop under these circumstances is the in-phase component of IZ. Fortunately, it is not necessary to perform an exact calculation using complex arithmetic to get a sufficiently accurate voltage drop for UD transformer and secondary configurations. Equation B.2 produces a very close approximation to the voltage drop for virtually all situations. The components of current needed for Equation B.2 are determined from the load current and load power factor by using Equation B.3. The R and X components of the supply circuit impedance, Z, are found separately for the transformer and the secondary/service cables. After these values are found for each part of the circuit, the separate R values are totaled and the separate X values are totaled to get the R and X values to use in Equation B.2. In the case of the transformer, the impedance is given on data sheets and nameplates as “%IZ.” Equation B.4 can be used to convert this impedance to ohms as seen by the secondary circuit. After the transformer impedance, Z, in ohms is found from Equation B.4, the next step is to find the resistive component, R, of this impedance. An estimate of the winding losses of the


B Equation B.4 ZT = where: ZT E %IZ kVA

= = = =

E2(%IZ/100) (kVA)(1,000)

Impedance per leg or phase of transformer, in ohms Voltage of leg or phase, in volts (120 volts, in most cases) Transformer percentage impedance Transformer kVA rating per leg or phase (one-half the rating of single-phase transformers or one-third the rating of three-phase wye transformers)

transformer is usually available from data gathered when purchasing transformers, and the R value can be directly calculated from winding losses by using Equation B.5. It is important that only winding losses should be used in Equation B.5. Transformers also experience core losses, but this loss component does not affect the transformer R used in voltage drop calculations. After the transformer’s Z and R are determined, the transformer reactance, X, can be calculated with Equation B.6. Equation B.6

Equation B.5 RT = where: RT E W

kVA

XT = ZT2 – RT2

E2W (kVA ×1,000)2

= Transformer resistance, in ohms = Voltage of leg or phase, in volts (120 volts, in most cases) = Transformer winding losses per leg or phase, in watts (one-half of the winding losses of single-phase transformers or one-third of the winding losses of threephase transformers) = Transformer kVA rating per leg or phase

where: XT = Transformer reactance, in ohms ZT = Transformer impedance, in ohms RT = Transformer resistance, in ohms Example B.1 illustrates a voltage drop calculation for a consumer service immediately adjacent to a transformer. Consideration of secondary and service cable impedances begins after this example.

EXAMPLE B.1: Transformer Voltage Drop Calculation. Determine the transformer voltage drop for a 93-ampere load at 85% power factor served immediately adjacent to a 25-kVA, 120/240-volt single-phase transformer. The transformer has 3% impedance and 280 watts of winding losses. The current components are obtained from Equation B.3:

ID = I cos θ = (93)(0.85) = 79 amperes θ = arc cosine (0.85) = 31.8° IQ = I sin θ = (93)(sin 31.8°) = 49 amperes

The transformer’s Z is obtained from Equation B.4: ZT =

E2(%IZ/100) (120)2(0.03) = = 0.03456Ω (kVA)(1,000) 12,000

The transformer’s X is obtained from Equation B.6: XT = ZT2 – RT2 = (0.03456)2 – (0.01290)2 = 0.03206Ω The values needed by Equation B.2 are now available, so that equation can be used to calculate the voltage drop: VDROP = IDRT + IQXT = (79)(0.01290) + (49)(0.03206) = 2.59 volts

The transformer’s R is obtained from Equation B.5: RT =

E2W (kVA × 1,000)2

=

(120)2(0.03) (12,000)2

= 0.01290Ω

This drop of 2.59 volts is on a 120-volt base. Refer to Table B.1 to see that the amount of drop is within guidelines.

3 8 0 – Ap p e n d i x B

B In most actual situations, the consumer’s service entrance is not immediately adjacent to the transformer. Therefore, it is necessary to include secondary and service cable impedance in the voltage drop calculation. This cable impedance consists of resistive (R) ohms and reactive (X)

ohms that are respectively added to the transformer RT and XT before Equation B.2 is used. The resistance of secondary cables can be found from standard references. Table B.2, taken from the Aluminum Electrical Conductor Handbook, is an excellent compilation. For voltage

TABLE B.2: Resistance of Class B Concentric-Strand Aluminum Cable with Thermosetting and Thermoplastic Insulation for Secondary Distribution Voltages (to 1 kV) at Various Temperatures and Typical Conditions of Installation (Ohms per 1,000 feet). Adapted from the Aluminum Electrical Conductor Handbook (1989) 60 Hz ac at 60°C Multiconductor One Single Cable or 2 Conductor in or 3 Single Air, Buried, or Conductors in in Nonmetallic One Metallic Conduit Conduit

60 Hz ac at 75°C

60 Hz ac at 90°C

Multiconductor One Single Cable or 2 Conductor in or 3 Single Air, Buried, or in Conductors in dc at Nonmetallic One Metallic 75°C* Conduit Conduit

Multiconductor One Single Cable or 2 Conductor in or 3 Single Air, Buried, or in Conductors in dc at Nonmetallic One Metallic 90°C* Conduit Conduit

Class B (AWG or kcmil)

dc at 60°C*

6

0.7650

0.7650

0.7650

0.8080

0.8080

0.8080

0.8480

0.8480

0.8480

4

0.4830

0.4830

0.4830

0.5070

0.5070

0.5070

0.5330

0.5330

0.5330

3

0.3820

0.3820

0.3820

0.4020

0.4020

0.4020

0.4220

0.4220

0.4220

2

0.3030

0.3030

0.3030

0.3190

0.3190

0.3190

0.3350

0.3350

0.3350

1

0.2400

0.2400

0.2400

0.2530

0.2530

0.2530

0.2660

0.2660

0.2660

1/0

0.1910

0.1910

0.1910

0.2010

0.2010

0.2010

0.2110

0.2110

0.2110

2/0

0.1510

0.1510

0.1510

0.1590

0.1590

0.1590

0.1670

0.1670

0.1670

3/0

0.1190

0.1190

0.1200

0.1260

0.1260

0.1270

0.1320

0.1320

0.1330

4/0

0.0953

0.0954

0.0963

0.1010

0.1010

0.1020

0.1050

0.1060

0.1070

250

0.0806

0.0808

0.0822

0.0847

0.0850

0.0865

0.0890

0.0892

0.0908

300

0.0672

0.0674

0.0686

0.0706

0.0708

0.0720

0.0741

0.0744

0.0756

350

0.0575

0.0578

0.0593

0.0605

0.0608

0.0623

0.0635

0.0638

0.0654

400

0.0504

0.0507

0.0525

0.0530

0.0533

0.0552

0.0556

0.0560

0.0580

500

0.0403

0.0406

0.0428

0.0424

0.0427

0.0450

0.0445

0.0448

0.0472

600

0.0336

0.0340

0.0370

0.0353

0.0357

0.0381

0.0370

0.0374

0.0400

700

0.0288

0.0292

0.0320

0.0303

0.0307

0.0337

0.0318

0.0322

0.0353

750

0.0269

0.0273

0.0302

0.0282

0.0288

0.0317

0.0296

0.0302

0.0333

1,000

0.0201

0.0207

0.0239

0.0212

0.0218

0.0253

0.0222

0.0228

0.0265

1,250

0.0162

0.0176

0.0215

0.0169

0.0177

0.0216

0.0178

0.0186

0.0228

1,500

0.0135

0.0143

0.0184

0.0141

0.0150

0.0193

0.0148

0.0158

0.0203

1,750

0.0115

0.0124

0.0168

0.0121

0.0131

0.0177

0.0127

0.0137

0.0186

2,000

0.0101

0.0111

0.0158

0.0106

0.0117

0.0166

0.0111

0.0122

0.0173

* Calculated from ICEA resistance tables for Class B stranding and corrected for temperature. Note. The metallic conduit is assumed to be steel. If aluminum is used, the effective resistance is about the same as for single conductor in nonmetallic conduit to 4/0 size and, for larger sizes, is in the range of 1/2% to 2% more than the resistance of the conductor in nonmetallic conduit and, hence, of little significance except in critical cases.

Transformer and Secondary Voltag e Dro p – 3 8 1

B Equation B.7 A

2πƒ s 0.0153 + 0.1404log10 1,000 r

where: X = Inductive reactance to neutral of one conductor, in ohms per 1,000 feet s = Spacing between centers of conductors, in inches r = Radius of the metal portion of the conductor, in inches, including strand shielding, if any f = Frequency, in Hertz (it is convenient to use 377 for 2π × 60)

A

s=A

Right Angle Triangle

A

s = 1.122A

A

Symmetrical Flat

s = 1.26A

B A

drop studies, the resistance at 60°C should be used unless the conductors are being loaded to very near their thermal limits, which is not usually the case. The reactance of secondary cable is composed of inductance and capacitance. However, the effect of shunt capacitance can be ignored in secondary voltage calculations because of its negligible effect on the results. The inductive reactance can be calculated with the following equations and tables. Equation B.7 determines the inductive reactance of one line conductor. The distances (assumed average effective) for various conductor arrangements are shown in Figure B.1. Table B.3 is a table of corrections for Equation B.7. In Table B.3, the term sector refers to a single conductor in which the strands are arranged approximately as a 120° section of a circle as opposed to a round conductor. This conductor configuration is not usually encountered in contemporary UD systems. The designation single conductor refers to one of several single conductors of a single circuit that lie loosely together in one conduit, not bound together or closely adjacent on a support. The increase for random lay in this instance is the result of unequal spacing of the conductors in the conduit. Table B.4 gives conductor diameters (r = D/2) and outside diameters for XLPE insulation as defined in the footnote.

Equilateral Triangle

A

X=

C Unequal

3

s = (A × B × C)

FIGURE B.1: Distance for Various Conductor Arrangements.

3 8 2 – Ap p e n d i x B

B TABLE B.3: Corrections for Multiconductor Cables. Adapted from the Aluminum Electrical Conductor Handbook (1989).

Conductor Size (kcmil, up to) 250 300 350 400 500 600 700 750

Nonmagnetic Binder Magnetic Binder Round Sector Round Sector Multiplying Factor 1.000 0.975 1.149 1.230 1.000 0.970 1.146 1.225 1.000 0.965 1.140 1.220 1.000 0.960 1.134 1.216 1.000 0.950 1.122 1.203 1.000 0.940 1.111 1.199 1.000 0.930 1.100 1.191 1.000 0.925 1.095 1.186

Single Conductors in Conduit Nonmagnetic: Increase 20% for random lay Magnetic: Increase 50% for magnetic effect and random lay

Multiple Conductor Cables in Conduit Nonmagnetic: No correction Magnetic: Use value for round conductors with magnetic binder

TABLE B.4: Comparison of Conductor Diameter and Approximate Cable Outside Diameter of Typical Single, Class B Concentric-Strand Aluminum Cables. Voltages are ac line-to-line with grounded neutral* except as stated. Adapted from the Aluminum Electrical Conductor Handbook (1989).

Size (AWG Conductor or kcmil) Diameter (in.) 6 0.184 4 0.232 2 0.292 1 0.332 1/0 0.373 2/0 0.418 3/0 0.470 4/0 0.528 250 0.575 350 0.681 500 0.813 750 0.998 1,000 1.152 1,250 1.289 1,500 1.412 1,750 1.526 2,000 1.632

Approximate Outside Diameter of Cable Thermosetting or Thermoplastic Insulation (Inches) Nonshielded Fully Shielded 600 V 1 kV 5 kV** 5 kV 15 kV 25 kV 35 kV 46 kV 0.32 0.34 0.62 0.74 0.37 0.39 0.67 0.79 0.43 0.45 0.73 0.88 1.16 1.16 0.51 0.53 0.77 0.92 1.20 1.68 0.55 0.57 0.85 0.96 1.24 1.72 1.45 0.60 0.62 0.89 1.00 1.29 1.77 1.50 0.65 0.67 0.95 1.06 1.34 1.83 1.55 1.82 0.71 0.73 1.01 1.11 1.40 1.92 1.61 1.87 0.79 0.81 1.08 1.20 1.44 1.96 1.65 1.92 0.90 0.92 1.18 1.31 1.56 2.06 1.80 2.03 1.03 1.05 1.32 1.44 1.75 2.17 1.97 2.24 1.25 1.27 1.50 1.63 1.93 2.38 2.14 2.34 1.40 1.42 1.73 1.85 2.09 2.56 2.30 2.50 1.58 1.60 1.91 2.02 2.26 2.73 1.70 1.72 2.04 2.13 2.38 2.96 1.82 1.84 2.15 2.22 2.49 3.07 1.92 1.94 2.29 2.36 2.61 3.13

* For voltages through 5 kV, the diameters also apply if the neutral is ungrounded. For cables above 5 kV with ungrounded neutral or cables at 133% insulation level, consult manufacturer’s lists. ** The 5-kV nonshielded cable, as well as all shielded cables, has strand shielding. The listed overall diameters of 600-volt cables are from Column 4 of Table 5 of the NEC (1981) and are fairly representative of Type THW and triple-rated RHW/RHH/USE unjacketed cable with XLPE insulation; the values are increased by 0.02 in. for 1 kV. The values in the other columns correspond closely with those listed in ICEA No. S-94-649-2000, when increased to allow for jackets. By omitting the jacket, sometimes a lead sheath may be included without increase of diameter. These diameters do not apply to cable with metallic armor. Although the listed values are generally suitable for preliminary studies, important calculations should be made by using the actual diameter of the selected cable.


B EXAMPLE B.2: Secondary Cable Resistance and Reactance. Find the resistance and reactance per 1,000 feet for each conductor of a 250-kcmil, three-conductor, 600-volt cable, aluminum, concentric stranded, 0.79-inch diameter in nonmagnetic conduit. The conductors are bound with tape or twisted to maintain the conductors as an equilateral triangle (triplexed). The spacing between conductors is equal to the outside diameter of a single conductor. Determine the resistance per 1,000 feet of each conductor. From Table B.2, the ac resistance at 60°C is as follows:

R= 0.0808Ω/1,000 feet Calculate the reactance per 1,000 feet of each conductor. From Equation B.7:

X=

where:

377 s 0.0153 + 0.1404log10 1,000 r

s = 0.79 inches r = Diameter ÷ 2 = 0.575 ÷ 2 = 0.2875, from Table B.4

X=

377 0.79 0.0153 + 0.1404log10 = 0.029Ω/1,000 feet 1,000 0.2875

From Table B.3, Corrections for Multiconductor Cables, no random-lay correction is necessary. If this cable was in a magnetic conduit, the correction factor would be 1.149 and the reactance would be as follows: 0.029 × 1.149 = 0.033Ω/1,000 feet

Example B.2 illustrates the methods for calculating secondary cable resistance and reactance. The results of Example B.2 show that the conductor resistance (0.0808 Ω/1,000 feet) is nearly three times larger than the reactance (0.029 Ω/1,000 feet) for 250-kcmil aluminum conductors. For smaller conductors, the resistance increases by a larger factor than the reactance, so the disparity between the two is even greater. Examining Equation B.2, the basic voltage drop equation, reveals that the voltage drop depends more on resistance than reactance for power factors above 71 percent. At 71 percent power factor, ID (real component of current) begins to be greater than IQ (reactive component of current). It can be concluded from these observations

that accurately estimating the conductor reactance is not as important as accurately estimating the resistance for normal load power factors. Therefore, for most studies, the engineer may get the conductor reactance directly from a table rather than spend time calculating the reactance from Equation B.7. Table B.5, taken from the Aluminum Electrical Conductor Handbook (1989), may be used to quickly estimate conductor reactance. This table assumes random lay of conductors, so the values tabulated need to be divided by 1.2 if tightly bound cables are being used. Example B.3 is a continuation of Example B.1 to illustrate the combined effect of transformer and cable impedances on voltage drop.

3 8 4 – Ap p e n d i x B

B TABLE B.5: 60 Hz Reactance of Conductors in the Same Conduit (Ohms per 1,000 feet). Source: Aluminum Electrical Conductor Handbook (1989).

Wire Size (AWG or kcmil)

60

Nonmagnetic Conduit (Aluminum)

Magnetic Conduit (Steel)

Conductor Covering Thickness (Insulation + Cover) (mils)

Conductor Covering Thickness (Insulation + Cover) (mils)

80

95

110

125

140

150

170

190

60

80

95

6

0.0404 0.0430 0.0455

0.0505 0.0537 0.0568

4

0.0386 0.0402 0.0424

0.0475 0.0503 0.0530

2

0.0359 0.0379 0.0398

0.0449 0.0473 0.0497

110

125

140

155

1

0.0367 0.0384 0.0400 0.0415 0.0430 0.0443

0.0458 0.0480 0.0500 0.0519 0.0538 0.0554

1/0

0.0357 0.0373 0.0387 0.0402 0.0416 0.0428

0.0446 0.0466 0.0484 0.0502 0.0520 0.0535

2/0

0.0348 0.0363 0.0376 0.0389 0.0402 0.0414

0.0435 0.0453 0.0470 0.0487 0.0503 0.0517

3/0

0.0339 0.0353 0.0366 0.0378 0.0390 0.0401

0.0424 0.0442 0.0459 0.0473 0.0488 0.0501

4/0

0.0332 0.0344 0.0356 0.0367 0.0378 0.0388

0.0415 0.0431 0.0445 0.0459 0.0473 0.0486

170

190

250

0.0338 0.0349 0.0360 0.0370 0.0380 0.0390 0.0399

0.0423 0.0436 0.0450 0.0453 0.0475 0.0487 0.0499

300

0.0333 0.0342 0.0353 0.0363 0.0372 0.0381 0.0390

0.0416 0.0428 0.0441 0.0453 0.0464 0.0475 0.0482

350

0.0328 0.0337 0.0347 0.0356 0.0364 0.0373 0.0382

0.0410 0.0421 0.0433 0.0445 0.0456 0.0467 0.0477

400

0.0324 0.0333 0.0342 0.0351 0.0359 0.0367 0.0375

0.0405 0.0416 0.0427 0.0439 0.0449 0.0459 0.0469

500

0.0318 0.0326 0.0334 0.0343 0.0350 0.0358 0.0365

0.0397 0.0407 0.0418 0.0428 0.0438 0.0447 0.0457

600

0.0321 0.0329 0.0336 0.0343 0.0350 0.0357

0.0401 0.0411 0.0420 0.0429 0.0438 0.0447

700

0.0317 0.0324 0.0331 0.0338 0.0345 0.0351

0.0397 0.0405 0.0414 0.0422 0.0431 0.0439

750

0.0315 0.0322 0.0329 0.0335 0.0342 0.0349

0.0394 0.0403 0.0411 0.0419 0.0428 0.0436

The above tabular values include a 20% adjustment for random lay of single conductors in a nonmagnetic conduit and a 50% adjustment for random-lay and magnetic effect in steel conduit. If the conductors are part of a multiconductor cable with fixed spacing, multiply the tabular values in the left-hand section by 0.833. For the right-hand section in such a case, multiply the adjusted left-hand section values by the magnetic-binder adjustment factors shown in Table B.3. Thus, for a triplexed 250-kcmil cable with minimum 155-mil insulation thickness of each conductor, the reactance when in nonmagnetic conduit is 0.0380 × 0.0833 = 0.0316 ohms per 1,000 ft., and when in magnetic circuit is 0.0316 × 1.149 = 0.0363 ohms per 1,000 ft.


B EXAMPLE B.3: Complete Secondary Voltage Drop Calculation. Determine the total transformer and cable voltage drop that will occur if the consumer service of Example B.1 is served over 130 feet of AWG No. 1/0 aluminum secondary triplexed UD cable with 80-mil insulation thickness. The transformer and load current conditions are the same as given in Example B.1. From Table B.2, the resistance of AWG No. 1/0 aluminum conductor is 0.191Ω/1,000 feet at 60°C. For the actual length of 130 feet, the resistance is as follows:

R=

130 (0.191) = 0.02483Ω 1,000

The reactance is obtained from Table B.5, divided by 1.2 to adjust for close spacing, and prorated for the 130-foot actual distance:

X=

130 0.0357 = 0.00387Ω 1,000 1.2

Next, the total of transformer and cable resistances and reactances is calculated (see Example B.1 for transformer values): RTOT = RT + R = 0.01290 + 0.02483 = 0.03773Ω XTOT = XT + X = 0.03206 + 0.00387 = 0.03593Ω Equation B.2 can now be used to calculate the total secondary voltage drop (see Example B.1 for the determination of the ID and IQ values): VTOT-DROP = IDRTOT + IQ XTOT = (79)(0.03773) + (49)(0.03593) = 4.74 volts Comparing this result with the guidelines of Table B.1 shows the 4.74-volt drop is excessive. However, if the location is fairly close to the substation, the 4.74-volt drop is acceptable because it is less than the 6-volt limit applicable under that circumstance.

Voltage Flicker

Secondary flicker usually is caused by an inrush of current into consumer equipment. This inrush is usually associated with motor starting and can be five to six times the normal full-load-rated amperes of the motor. Although motors are the most common cause of inrush, other electrical equipment such as welders, arc furnaces, or large blocks of electric heat can also cause problems. The problem with the sudden current increase is that the secondary system (transformer and conductors) must carry this momentary current with its accompanying voltage drop. The

voltage flicker is calculated either on a 120-volt base or as a percentage of nominal voltage. Allowable levels of voltage dip or flicker are very subjective. At low levels, some voltage dips go unnoticed. At slightly higher levels, the consumer becomes aware of the voltage dips, but the magnitude and frequency are tolerable. However, as the magnitude or frequency of the voltage dip increases, the dips become annoying. The word frequency used in flicker evaluation is a reference to how often the voltage dips occur, such as three per hour or four per day.

3 8 6 – Ap p e n d i x B

B greater amount of flicker is to be tolerated to control the costs of correction to a lower level. However, it is important to understand that areas exist where special problems preclude strict adherence to this interpretation of how the flicker guidelines are to be applied in actual cases. The impedance of the primary system ahead of the transformer is sometimes a significant contributor to the total voltage dip during a large secondary current inrush. The value of this impedance is available from primary fault current calculations performed in conjunction with sectionalizing studies. A complete analysis of the issue is beyond the scope of this appendix. However, Example B.4 illustrates the method for translating primary impedance values to the secondary. The general method for calculating the magnitude of voltage flicker is the same as previously demonstrated for voltage drops caused by load current.

Figure B.2 is a chart for evaluating the magnitude of permissible voltage flicker. This chart appears in RUS Bulletin 160-1 and in ANSI standards. The limits are based on expected consumer annoyance levels, which is usually the concern. However, if the voltage dip is allowed to become very severe, equipment operation may be impaired or the motor that is causing the voltage dip may not maintain sufficient terminal voltage to start. For services to individual residential consumers in UD developments, the cooperative should try to limit voltage flicker to the level shown in Figure B.2 marked “Flicker Limits for Installations Serving Many Consumers.” The higher flicker level, marked “Revised Flicker Limits for Installations Serving Few Consumers,” should be applied only in cases in which the cooperative has discussed the flicker problems with the involved consumers and both parties have agreed that a

10

10 Voltage Flicker Limits Revised 120V Basis

9

9

Revised Threshold of Objection 8

8 7

Flicker Limits for Installations Serving Few Consumers

6

6

5

5

4

4

3

3

2

2 Flicker Limits for Installations Serving Many Consumers

1

1 Revised Threshold of Perception

0

4

8 12 Per Day

1

2

5

10

20 30

1

2

Per Hour

FIGURE B.2: Permissible Voltage Flicker Limits.

5

10

Per Minute

20 30

1

2

5

10

Per Second

20 30

0

Volts Change

Volts Change

7


B EXAMPLE B.4. Voltage Flicker Calculation. The transformer and service arrangement described in Example B.3 is located where the primary system line-toground bolted fault current is 750 amperes. The primary line-to-neutral voltage is 7.2 kV, and the fault X/R ratio at the sample location is 1.0. Estimate the voltage dip at the service when an 18-ampere, 230-volt, air-conditioning compressor is starting. The transformer and secondary cable impedances are the same as calculated in Example B.3. The following procedure is used to estimate the additional impedance reflected from the primary system: ZP =

Primary Voltage 7,200 volts = 9.6Ω = Fault Current 750 amperes

When reflected through the transformer to the secondary side, this impedance is reduced by the square of the transformer turns ratio:

ZS = ZP

ES 2 240 2 = 9.6 = 0.01067Ω EP 7,200

The impedances calculated in Example B.3 are for one leg only of the 120/240-volt single-phase service. The impedance, ZS, above includes both legs and must be divided by two to put it on the 120-volt base used in Example B.3. ZS 0.01067Ω = = 0.00534Ω 2 2 A system X/R ratio of 1.0 means that the system R and X are equal, and that each is equal to Z/√2. Therefore, RS X 0.00534Ω = S= = 0.00378Ω 2 2 2

The total supply resistance and reactance at the point of the service are found by adding the reflected primary values to the previous totals from Example B.3. RTOT = 0.03773 + 0.00378 = 0.04151Ω XTOT = 0.03593 + 0.00378 = 0.03971Ω Because of the great variety of air-conditioning equipment that exists, it is often difficult to obtain the starting amperes and starting power factor for the equipment involved in a particular application. A conservative estimate is to use a starting current of seven times the full-load running current. Starting power factors for single-phase motors also vary widely. A reasonable estimate is 80 percent if no specific information is known. In light of these guidelines, the following values are estimated for ID and IQ for this example: ID = (7)(18)(0.8) = 101 amperes IQ = (7)(18)(0.6) = 76 amperes Equation B.2 can now be used to estimate the voltage dip. VDIP = IDR + IQX = (101)(0.04151) + (76)(0.03971) = 4.19 + 3.02 = 7.21 volts The expected voltage dip of 7.21 volts on a 120-volt base is below the threshold of objection for 10 starts per hour (see Figure B.2), and it is within the stated guideline for two starts per hour for installations serving few consumers. However, it is substantially above the allowable limits for residential consumers. It can be concluded that the arrangement is marginally acceptable if consumer agreement is obtained.


Sample Specification UGC2 for 600-Volt Secondary Underground Power C a b l e – 3 8 9

C

Sample Specification UGC2 for 600-Volt Secondary Underground Power Cable

Scope General Specifications Referenced Specifications Conductor Insulation

6. 7. 8. 9.

Table of Contents

1. 2. 3. 4. 5.

Abbreviations

ANSI ASTM AWG HDPE ICEA IEEE KCMIL LDPE MDPE NEMA NESC NRECA PE REA RUS USE XLPE

1. Scope

This document is to provide a sample specification for 600-volt single-conductor and multiconductor secondary underground power cable.

Tests Miscellaneous Markings Multiconductor Cable Assemblies

American National Standards Institute American Society for Testing and Materials American Wire Gauge High Density Polyethylene Insulated Cable Engineers Association, Inc. Institute of Electrical and Electronics Engineers, Inc. Thousand Circular Mil Low Density Polyethylene Medium Density Polyethylene National Electrical Manufacturers Association National Electrical Safety Code National Rural Electric Cooperative Association Polyethylene Rural Electrification Administration Rural Utility Services Underground Service Entrance Cable (UL Approved) Cross-Linked Polyethylene

The NRECA Transmission and Distribution Underground Subcommittee prepared this Sample Specification UGC2.

3 9 0 – Ap p e n d i x C

C 2. General Specifications

a. This specification covers the construction, mechanical, and electrical requirements for single- and multiconductor 600-volt cables with standard cross-linked polyethylene and ruggedized composite cross-linked polyethylene extruded insulations. The cables shall be suitable for use in direct-burial installation in ducts, conduits, or air in wet or dry locations. b. Cable insulation shall be cross-linked polyethylene or ruggedized cross-linked polyethylene as specified by purchaser. c. Conductor sizes No. 6 AWG and larger for copper and aluminum are included.

d. Where provisions of Sample Specification UGC2 conflict with the presently approved REA Bulletin U-2 or its successor document, RUS may require conditional approval. e. Cable insulation shall be capable of operating continuously in both wet and dry locations at a conductor temperature of 90°C under normal and 130°C under emergency operating conditions. The cable shall have an allowable conductor temperature of 250°C under short-circuit conditions.

3. Referenced Specifications

a. Cable shall be in compliance, as noted, with the latest revisions of the following industry standards:

• ASTM B 787, “Specification for 19 Wire Combination Unilay-Stranded Copper Conductors for Subsequent Insulation” • ASTM B 800, “Specification for 8000 Series Aluminum Alloy Wire for Electrical Purposes—Annealed and Intermediate Tempers” • ASTM B 801, “Specification for Concentric-Lay-Stranded Conductors of 8000 Series Aluminum Alloy for Subsequent Covering or Insulation” • ASTM B 901, “Specifications for Compressed Round Stranded Aluminum Conductors Using Single Input Wire Construction” • ASTM B 902, “Specifications for Compressed Round Stranded Copper Conductors Using Single Input Wire Construction” • ASTM D 1248, “Specification for Polyethylene Plastics Molding and Extrusion Materials” • ASTM D 1693, “Test Method for Environmental Stress-Cracking of Ethylene Plastics”

• ICEA S-81-570, “Standard for 600 Volt Rated Cable of Ruggedized Design for Direct Burial Installations as Single Conductors or Assemblies of Single Conductors” • ICEA S-105-692, “Standard for 600-Volt Single Layer Thermoset Insulated Utility Underground Distribution Cable” • ASTM B 3, “Specification for Soft or Annealed Copper Wire” • ASTM B 8, “Specification for ConcentricLay-Stranded Copper Conductors, Hard, Medium-Hard, or Soft” • ASTM B 230, “Specification for Aluminum 1350-H19 Wire for Electrical Purposes” • ASTM B 231, “Specification for ConcentricLay-Stranded Aluminum 1350 Conductors” • ASTM B 400, “Specification for Compact Round Concentric-Lay-Stranded Aluminum 1350 Conductors” • ASTM B 496, “Specification for Compact Round Concentric-Lay-Stranded Copper Conductors” • ASTM B 609, “Specification for Aluminum 1350 Round Wire, Annealed and Intermediate Tempers, for Electrical Purposes” • ASTM B 784, “Specification for Modified Concentric Lay Stranded Copper Conductors for Use in Insulated Electrical Cables” • ASTM B 786, “Specification for 19 Wire Combination Unilay-Stranded Aluminum 1350 Conductors for Subsequent Insulation”

b. Availability of Publications (1) Copies of the American National Standards Institute/Insulated Cable Engineers Association, Inc. (ANSI/ICEA), publications can be obtained from Global Engineering Documents for a fee at the address indicated below: IHS Global Engineering Documents 15 Inverness Way East Englewood, CO 80112 Telephone: 877.413.5184 E Mail: [email protected] Web Site: global.ihs.com


C (2) Copies of American Society for Testing and Materials (ASTM) publications referenced in this specification can be obtained from ASTM for a fee at the address indicated below: ASTM 100 Barr Harbor Drive West Conshohocken, PA 19428-2959 Telephone: 610.832.9585 Web Site: astm.org

4. Conductor

a. Conductors shall be copper or aluminum as specified by the purchaser. b. Copper conductors shall be Class B stranded annealed copper in accordance with ASTM B-3 or ASTM B-8. c. Aluminum conductors shall be one of the following: (1) Solid EC grade aluminum; either soft, half-hard, or three-quarter hard in accordance with ASTM Specification B-230 or ASTM Specification B-609. (2) Class B stranded EC grade aluminum; either three-quarter hard or hard-drawn in accordance with ASTM Specification B-230 or ASTM Specification B-231, or Series 8000 aluminum alloy in accordance with ASTM B 800.

5. Insulation

(3) Copies of the National Electrical Safety Code (NESC) can be obtained from IEEE for a fee at the address indicated below: Institute of Electrical and Electronics Engineers, Inc. (IEEE) IEEE Service Center 445 Hoes Lane Piscataway, NJ 08855 Telephone: 800.678.4333 Web Site: shop.ieee.org/ieeestore

(3) Compact-round stranded in accordance with ASTM Specification B-400. Combination unilay stranded aluminum phase conductors shall conform to ASTM B 786. Combination unilay stranded Series 8000 aluminum alloy conductors shall comply with ASTM B 800. (4) Single-input-wire (SIW) stranded and compressed in accordance with ASTM B 901 and ANSI/ICEA S-105-692 and ANSI/ICEA S-81-570. SIW stranded Series 8000 aluminum alloy conductor shall comply with ASTM B 800. d. The solid conductor and center strand of stranded conductors shall be indented with the manufacturer’s name and year of manufacture at regular intervals with no more than 12 inches (0.3 m) between repetitions.

a. Insulation shall be one of the following as specified by the purchaser: (1) Ruggedized Construction I

I

Ruggedized composite cross-linked polyethylene insulation composed of either: (a) an inner layer of black low- or medium-density polyethylene and an outer layer of black high-density polyethylene firmly bonded together, or (b) a single layer of black crosslinked high density polyethylene. The composite insulation shall comply with the physical and electrical properties as indicated in ICEA Pub-

I

I

lication S-81-570, and the finished cable shall meet the mechanical abuse requirements specified in Section 6.3. The physical, electrical, and mechanical abuse requirements shall be tested in accordance with ICEA Publication S-81-570. The nominal composite insulation wall thickness and each layer forming the composite insulation wall (LDPE or MDPE inner layer and HDPE outer layer) shall have a nominal thickness as shown in Table C.1. The composite wall minimum thickness shall not be less than 90 percent of these values.

3 9 2 – Ap p e n d i x C

C TABLE C.1: Nominal Composite Insulation Layer Thickness (Ruggedized)

Conductor Size (AWG or kcmil)

Composite or Single-Layer Insulation Wall Thickness (mils)

Inner Layer (mils)

Outer Layer (mils)

6-2

60

30.0

30.0

1-4/0

80

40.0

40.0

*250-500

*95

*47.5

*47.5

501-1,000

110

55.0

55.0

Nominal Thickness of Each Layer for Composite Insulation Construction

* Or 80 mil (40/40 mil) thickness, as specified by user.

(2) Alternative: Non-Ruggedized Construction I

6. Tests

Standard cross-linked polyethylene insulation that meets the requirements of ICEA Publication S-105692. The nominal thickness of insulation shall not be less than shown in Table C.2.

a. Qualification Tests. As part of a request for RUS consideration for acceptance and listing, the manufacturer shall submit certified test data results to RUS that detail full compliance with ICEA S-81-570 for ruggedized cable design. (1) Test results shall confirm compliance with each of the material tests, production sampling tests, tests on completed cable, and qualification tests included in ICEA S-81-570.

7. Miscellaneous

a. Watertight seals shall be applied to all cable ends to prevent the entrance of moisture during transit or storage. Each end of the cable shall be firmly and properly secured to the reel. b. Cable shall be placed on shipping reels suitable for protecting it from damage during shipment and handling. Reels shall be provided with a suitable covering to help identify shipping damage to the cable.

TABLE C.2: Nominal Insulation Thickness (Non-Ruggedized) Conductor Size (AWG or kcmil)

Insulation Wall Thickness (mils)

6-2

60

1-500

80

The minimum thickness shall not be less than 90 percent of these values. b. If requested by the purchaser, the cable shall meet the requirements of and labeled as complying with Underwriters Laboratories standards for Type USE 600-volt singleconductor cable. c. Mechanical Abuse Requirements (Ruggedized Design Only) (1) Tests are to be performed on completed cable, with No. 1/0 AWG conductor, in accordance with ICEA Publication S-81-570.

b. Production tests shall be performed in accordance with ICEA S-81-570 and ICEA 692. If requested by the purchaser, a certified copy of the results of all production tests performed in accordance to this section shall be furnished on all orders. For all orders in excess of 10,000 feet, the manufacturer shall furnish results of production tests unless the purchaser waives this requirement through written instructions. c. For Type USE cable, the manufacturer shall perform all production tests required by Underwriters Laboratories. d. Frequency of sample tests shall be in accordance with ICEA T-26-465/NEMA WC53.

c. A durable label shall be securely attached to each reel of cable. The label shall indicate the purchaser’s name and address, purchase order number, cable description, reel number, feet of cable on the reel, tare and gross weight of the reel, and beginning and ending sequential footage numbers.


C 8. Markings

All cable provided under this specification shall have suitable markings on the outer surface of the jacket at sequential intervals not exceeding two feet (0.61 m). The legend shall indicate the name of the manufacturer, conductor size and

material, type and thickness of insulation, voltage rating, and year of manufacture. There shall be no more than six inches (0.15 m) of unmarked spacing between legend sequences.

9. Multiconductor Cable Assemblies

a. Cable shall be furnished in multiconductor assemblies if specified by the purchaser. Such assemblies shall consist of two (duplex), three (triplex), or four (quadraplex) single-conductor cables which individually meet all requirements of this specification. b. The cable assembly shall meet all requirements of ICEA S-105-692 or ICEA S-81-570 as applicable. c. A reduced neutral conductor (if specified) shall be no more than two standard sizes smaller than the phase conductor and no smaller than #2 AWG aluminum or #4 AWG copper or, where the phase conductors are smaller than #2 AWG aluminum or #4 AWG copper, the neutral conductor shall be the same size as the phase conductor.

d. A neutral conductor shall be clearly identified in each assembly. Neutral identification shall be in the form of three extruded weather resistant yellow stripes 120° apart. Each stripe shall cover a minimum of 20 percent of the neutral outer circumference. Stripes shall also be durable under conditions typically found in direct burial installations. (1) A solid yellow neutral insulation shall be supplied if specified by the purchaser. e. Multiconductor assemblies shall be furnished with a lay not exceeding sixty (60) times the diameter of an individual cable. f. Only one cable within a multiconductor assembly shall have sequential footage markings.


Checklist for Information Require m en t s – 3 9 5

D

Checklist for Information Requirements

Project Information Checklist (1) Delivery is: (2) Service voltage needed:

1φ

3φ

120/240 volts 208/120 volts, grounded wye 480/277 volts, grounded wye 240 volts, delta 480 volts, delta Other ______________________________________________________

(3) Connected loads Residential

kW

pf

Diversity

Heating* Cooling* Water heater Range/oven Miscellaneous LRA = __________ Amperes

*If available, record the locked rotor amps (LRA) of the largest compressor: Commercial

kW or Hp

pf

Diversity

Heating Cooling Lighting Base load (receptacles, small motors) Water heater Process machinery Large motor loads Size (horsepower) of largest motor started across line:

_________________

Number of times started per day:

_________________

3 9 6 – Ap p e n d i x D

D (4) Project Schedule Developer/Contractor Schedule

Planned Date

Actual Date

Temporary power service needed Permanent power service needed Roadways cut Property pins (front and back) in place and lot numbers displayed Final grade established Roads paved Sidewalks and curbing installed Start Date Utility Schedule

Water/sewer

Planned

Completion Date Actual

Planned

Actual

Telephone

Cable television Gas

Power

(5) Copies of Pertinent Plans Plans

Date Received

a. Subdivision Plat b. Grading Plan c. Utility Installation Plans

Water/sewer/surface drainage Telephone

Cable television Traffic control

Streetlight circuits

(6) Consumer-Owned Underground Facilities:

Water line Sewer line Septic tank and drain lines Satellite dish cable Irrigation system Electric lines Other ___________________________________________________________________________

Specification UGC1 – 3 9 7

E

Sample Specification for 15-, 25-, and 35-KV Primary Underground Medium Voltage Concentric Neutral Cable (Specification UGC1) Purpose General Specifications Referenced Specifications Conductor Conductor Shield (Stress Control Layer) Insulation

7. 8. 9. 10. 11. 12.

Insulation Shielding Concentric Neutral Conductor Overall Outer Jacket Dimensional Tolerances Tests Miscellaneous

Table of Contents

1. 2. 3. 4. 5. 6.

Abbreviations

ac ANSI ASTM AWG EPR ICEA LDPE LLDPE RUS TR-XLPE XLPE

1. Purpose

a. This document is to provide a sample specification for the purchase of medium voltage 15-, 25-, and 35-kV single-phase and multiphase medium voltage underground power cable. The NRECA Transmission and Distribution Underground Subcommittee prepared

this Sample Specification UGC1. The requirements of this specification are generally consistent with RUS’s proposed changes to REA Bulletin 50-70 (U-1). When accepted, the new RUS document will be Bulletin 1728F-U1.

2. General Specifications

a. This specification details recommended requirements for 15-, 25-, and 35-kV power cables for use on 12.5/7.2-kV (15-kV rated), 24.9/14.4-kV (25-kV rated), and 34.5/19.9-kV

(35-kV rated) underground distribution systems with multigrounded neutral. Cable complying with this specification shall consist of a single solid or strand-filled conductor which is

Alternating Current American National Standards Institute American Society for Testing and Materials American Wire Gauge Ethylene Propylene Rubber Insulated Cable Engineers Association, Inc. Low Density Polyethylene Linear Low Density Polyethylene U.S. Department of Agriculture Rural Development, Electric Program Tree Retardant Cross-Linked Polyethylene Cross-Linked Polyethylene

3 9 8 – Ap p e n d i x E

E insulated with tree-retardant cross-linked polyethylene (TR-XLPE) or ethylene propylene rubber (EPR), with concentrically wound copper neutral conductors covered by a nonconducting or semiconducting jacket. b. The cable may be used in singlephase and multiphase circuits. c. Acceptable conductor sizes are: No. 2 AWG (33.6 mm2) through 1,000 kcmil (507 mm2) for 15-kV cable, No. 1 AWG (42.4 mm2) through 1,000 kcmil (507 mm2) for 25-kV,

3. Referenced Specifications

a. The following specifications/standards are considered pertinent to this sample specification: • ANSI/ICEA S-94-649, “Standard for Concentric Neutral Cables Rated 5,000–46,000 Volts” • ANSI/IEEE C2, “National Electrical Safety Code” • ICEA S-97-682, “Utility Shielded Power Cables Rated 5 Through 46 kV” • ICEA T-31-610, “Guide for Conducting a Longitudinal Water Penetration Resistance Test for Sealed Conductor” • ICEA T-32-645, “Guide for Establishing Compatibility of Sealed Conductor Filler Compounds with Conductor Stress Control Materials” • ASTM B 3, “Specification for Soft or Annealed Copper Wire” • ASTM B 8, “Specification for ConcentricLay-Stranded Copper Conductors, Hard, Medium-Hard, or Soft” • ASTM B 230, “Specification for Aluminum 1350-H19 Wire for Electrical Purposes” • ASTM B 231, “Specification for ConcentricLay-Stranded Aluminum 1350 Conductors” • ASTM B 400, “Specification for Compact Round Concentric-Lay-Stranded Aluminum 1350 Conductors” • ASTM B 496, “Specification for Compact Round Concentric-Lay-Stranded Copper Conductors” • ASTM B 609, “Specification for Aluminum 1350 Round Wire, Annealed and Intermediate Tempers, for Electrical Purposes”

and 1/0 (53.5 mm2) through 1,000 kcmil (507 mm2) for 35-kV cable. d. Except where provisions therein conflict with the requirements of this specification, the cable shall meet all applicable provisions of ANSI/ICEA S-94-649. e. Where provisions of this specification conflict with the presently approved REA Bulletin 50-70 (U1) or its successor document (1728F-U1), RUS may require conditional approval.

• ASTM B 786, “Specification for 19 Wire Combination Unilay-Stranded Aluminum 1350 Conductors for Subsequent Insulation” • ASTM B 787, “Specification for 19 Wire Combination Unilay-Stranded Copper Conductors for Subsequent Insulation” • ASTM B 835, “Specification for Compact Round Stranded Copper Conductors Using Single Input Wire Construction” • ASTM B 836, “Specification for Compact Round Stranded Aluminum Conductors Using Single Input Wire Construction” • ASTM B 901, “Specifications for Compressed Round Stranded Aluminum Conductors Using Single Input Wire Construction” • ASTM B 902, “Specifications for Compressed Round Stranded Copper Conductors Using Single Input Wire Construction” • ASTM D 412, “Test Methods for Vulcanized Rubber and Thermoplastic Rubbers and Thermoplastic Elastomers-Tension” • ASTM D 746, “Test Method for Brittleness Temperature of Plastics and Elastomers by Impact” • ASTM D 1248, “Specification for Polyethylene Plastics Molding and Extrusion Materials” • ASTM D 1693, “Test Method for Environmental Stress-Cracking of Ethylene Plastics” • ASTM D 2275, “Test Method for Voltage Endurance of Solid Electrical Insulating Materials Subjected to Partial Discharges (Corona) on the Surface” • ASTM D 2765, “Test Methods for Determination of Gel Content and Swell Ratio of Cross-Linked Ethylene Plastics”


E • ASTM D 3349, “Test Method for Absorption Coefficient of Ethylene Polymer Material Pigmented with Carbon Black” • ASTM D 4496, “Test Method for DC Resistance or Conductance of Moderately Conductive Materials” • ASTM E 96, “Test Methods for Water Vapor Transmission of Materials” b. Availability of Publications (1) Copies of the American National Standards Institute/Insulated Cable Engineers Association, Inc. (ANSI/ICEA) S-94-649 publication can be obtained from IHS for a fee at the address indicated below: IHS 15 Inverness Way East Englewood, CO 80112 Telephone: 303.397.7956 or 877.413.5187 Fax: 303.397.2740 E-Mail: [email protected] Web Site: global.ihs.com

4. Conductor

a. Central phase conductors shall be copper or aluminum as specified by the purchaser within the limits of section 2.c. b. Central copper phase conductors shall be annealed copper in accordance with ASTM B3. Concentric-lay-stranded phase conductors shall conform to ASTM B 8 for Class B stranding. Compact round concentric-lay-stranded phase conductors shall conform to ASTM B 496. Combination unilay stranded phase conductors shall conform to ASTM B 787. If not specified otherwise by the purchaser, stranded phase conductors shall be Class B compressed strand. c. Central aluminum phase conductors shall be one of the following: (1) Solid: Aluminum 1350, H14 or H24, H16 or H26, in accordance with ASTM B 609. (2) Stranded: Aluminum 1350, H14 or H24, H16 or H26, in accordance with ASTM B 609. Concentric-lay-stranded (includes compressed) phase conductors shall conform to ASTM B 231 for Class B

(2) Copies of American Society for Testing and Materials (ASTM) publications referenced in this specification can be obtained from ASTM for a fee at the address indicated below: ASTM 100 Barr Harbor Drive West Conshohocken, PA 19428-2959 Telephone: 610.832.9585 Web Site: astm.org (3) Copies of the National Electrical Safety Code (NESC) can be obtained from IEEE for a fee at the address indicated below: Institute of Electrical and Electronics Engineers, Inc. (IEEE) IEEE Service Center 445 Hoes Lane Piscataway, NJ 08854 Telephone: 800.678.4333 Web Site: shop.ieee.org/ieeestore

stranding. Compact round concentric-laystranded phase conductors shall conform to ASTM B 400. Combination unilay stranded aluminum phase conductors shall conform to ASTM B 786. If not specified otherwise by the purchaser, stranded phase conductors shall be class B compressed strand. d. The interstices between the strands of stranded conductors shall be filled with a material designed to prevent the longitudinal migration of water that might enter the conductor. This material shall be compatible with the conductor and conductor shield materials. The outer surfaces of the strands that form the outer layer of the stranded conductor shall be free of the strand fill material. Compatibility of the strand fill material with the conductor shield shall be tested and shall be in compliance with ICEA T-32-645. Water penetration shall be tested and shall be in compliance with ICEA T-31-610.

4 0 0 – Ap p e n d i x E

E e. The center strand of stranded conductors shall be indented with the manufacturer’s name and year of manufacture at regular

5. Conductor Shield (Stress Control Layer)

6. Insulation

a. A non-conducting (for discharge-resistant EPR) or semiconducting shield (stress control layer) meeting the applicable requirements of ANSI/ICEA S-94-649 shall be extruded around the central conductor. b. The minimum thickness at any point shall be in accordance with ANSI/ICEA S-94-649 except minimum thickness requirements shall also be met at all points. See Table E.1. c. The conductor shield shall have a temperature rating equal to, or higher than, that of the insulation. d. The void and protrusion limits on the conductor shield shall be in compliance with the ANSI/ICEA S-94-649.

a. The insulation shall conform to the requirements of ANSI/ICEA publication S-94-649 and may either be tree retardant cross-linked polyethylene (TR-XLPE) or ethylene propylene rubber (EPR), as specified by the purchaser.

intervals with no more than 12 inches (0.3 m) between repetitions.

TABLE E.1: Extruded Conductor Shield Thickness. Conductor Size

Extruded Shield Thickness Minimum Point Mils mm

AWG or kcmil

mm2

8–4/0

8.37–107

12

0.30

212–550

107–279

16

0.41

551–1,000

279–507

20

0.51

b. The thickness of insulation shall be as shown in Table E.2. c. The contamination, void, and protrusion limits on the insulation shall be in compliance with the ANSI/ICEA S-94-649.

TABLE E.2: Nominal, Minimum, and Maximum Insulation Thickness.

7. Insulation Shielding

Cable Rated Voltage

Nominal Thickness

Minimum Thickness

Maximum Thickness

15 kV

220 mils (5.59 mm)

210 mils (5.33 mm)

250 mils (6.35 mm)

25 kV

260 mils (6.60 mm)

245 mils (6.22 mm)

290 mils (7.37 mm)

35 kV

345 mils (8.76 mm)

330 mils (8.38 mm)

375 mils (9.53 mm)

a. A semiconducting thermosetting polymeric layer meeting the requirements of ANSI/ ICEA S-94-649 shall be extruded tightly over the insulation to serve as an electrostatic shield and protective covering. The shield compound shall be compatible with, but not necessarily the same material composition as, that of the insulation (e.g., copolymer shield may be used with EPR insulation). A semi-conducting thermoplastic layer meeting

the requirements of ANSI/ICEA S-94-649 will be allowable on discharge-resistant EPR cable. b. The thickness of the extruded insulation shield and the concentric neutral indent shall be in accordance with ANSI/ICEA S-94-649. See Table E.3. c. The shield shall be applied such that all conducting material can be easily removed without the need for externally applied heat. Stripping tension values shall be six through


E 18 pounds (2.72 through 8.16 kg) for EPR discharge-free cable and for TR-XLPE. Discharge-resistant cables shall have strip tension value of zero through 18 pounds (zero through 8.16 kg).

d. The void and protrusion limits on the insulation shield shall be in compliance with the ANSI/ICEA S-94-649.

TABLE E.3: Insulation Shield Thickness for Cables with Wire Neutral. Calculated Minimum Diameter Over the Insulation

Insulation Shield Thickness Minimum Point Maximum Point mils mm mils mm

Maximum Concentric Neutral Indent mils mm

inches

mm

0–1.000

0–25.40

30

0.76

60

1.52

15

0.38

1.001–1.500

25.43–38.10

40

1.02

75

1.91

15

0.38

1.501–2.000

38.13–50.80

55

1.40

90

2.29

20

0.51

55

1.40

105

2.67

20

0.51

2.001 and larger 50.83 and larger

8. Concentric Neutral Conductor

a. A concentric neutral conductor shall consist of annealed round, uncoated copper wires in accordance with ASTM B 3 and shall be spirally wound over the insulation shield with uniform and equal spacing between wires. The concentric neutral wires shall remain in continuous intimate contact with the extruded insulation shield. Full neutral is required for single phase and 1/3 neutral for three phase applications unless otherwise specified. The minimum wire size for the concentric neutral is 16 AWG (1.32 mm2).

b. When a flat strap neutral is specified by the purchaser, the neutral shall consist of copper straps applied concentrically over the insulation shield with uniform and equal spacing between straps and shall remain in intimate contact with the underlying extruded insulation shield. The straps shall not have sharp edges. The thickness of the flat straps shall be not less than 20 mils (0.5 mm).

9. Overall Outer Jacket

a. An electrically nonconducting or semi-conducting outer jacket shall be applied directly over the concentric neutral conductors.

thermoplastic polyethylene (LDPE, LLDPE) compound meeting the requirements of ANSI/ICEA S-94-649, and ASTM D 1248 for Type I, Class C, Category 4 or 5, Grade J3 or Type II before application to the cable. Polyvinyl chloride (PVC) or chlorinated polyethylene (CPE) jackets are not acceptable. (3) Semi-conducting jackets shall have a radial resistivity not exceeding 100 ohmmeters and a maximum water vapor transmission rate of 2 g/m2/24 hours at 38°C (100°F) and 96 percent relative humidity in accordance with ASTM E 96.

(1) The jacket material shall be an extrudedto-fill jacket that fills the area between the concentric neutral wires and covers the wires to the proper thickness. The jacket shall be free stripping. The jacket shall have three red stripes longitudinally extruded into the jacket surface 120° apart as per ANSI/ICEA S-94-649. (2) Nonconducting jackets shall consist of low density, linear low density, or black

4 0 2 – Ap p e n d i x E

E TABLE E.4: Extruded-to-Fill Jacket Thickness. Calculated Minimum Diameter Over the Concentric Neutral

Insulation Shield Thickness Minimum Point Maximum Point mils mm mils mm

inches

mm

0–1.500

0–38.10

45

1.14

80

2.03

1.501 and larger

38.13 and larger

70

1.78

120

3.05

10. Dimensional Tolerances

Cables conforming to this specification shall have all dimensional tolerances meeting the requirements of ANSI/ICEA S-94-649.

11. Tests

a. Qualification Tests. As part of a request for RUS consideration for acceptance and listing, the manufacturer shall submit certified test data results to RUS that detail full compliance with ANSI/ICEA S-94-649 for each cable design. (1) Test results shall confirm compliance with each of the material tests, production sampling tests, tests on completed cable, and qualification tests included in ANSI/ICEA S-94-649. (2) The testing procedure and frequency of each test shall be in accordance with ANSI/ICEA S-94-649. (3) Certified test data results shall be submitted to RUS for any test, which is designated by ANSI/ICEA S-94-649 as being “For Engineering Information Only,” or any similar designation. b. Partial Discharge Tests. Manufacturers shall demonstrate that their cable complies with paragraph 11.b. (1) or 11.b. (2) of this specification. (1) Each shipping length of completed cable shall be tested and have certified test data results available indicating compliance with the partial discharge test requirements in ANSI/ICEA S-94-649.

b. The minimum thickness of the jacket over metallic neutral wires or straps shall comply with the thickness specified in ANSI/ICEA S-94-649. See Table E.4.

(2) Manufacturers shall test production samples and have available certified test data results indicating compliance with ASTM D 2275 for discharge resistance as specified in the ANSI/ICEA S-94-649. Samples of insulated cable shall be prepared by either removing the overlying extruded insulation shield material, or using insulated cable before the extruded insulation shield material is applied. The sample shall be mounted as described in ASTM D 2275 and shall be subjected to a voltage stress of 250 volts per mil of nominal insulation thickness. The sample shall support this voltage stress, and not show evidence of degradation on the surface of the insulation for a minimum test duration of 100 hours. The test shall be performed at least once on each 50,000 feet (15,240 m) of cable produced, or major fraction thereof, or at least once per insulation extruder run. c. Jacket Tests. Tests described in this section shall be performed on cable jackets from the same production sample as in section 11.b of this specification. (1) A Cold Bend Test shall be performed in accordance with the applicable provisions of the ANSI/ICEA S-94-649. The test temperature shall be -35°C (-31°F).


E The sample shall show no cracks visible to the normal, unaided eye at the conclusion of the test. The test shall be performed at least once on each 50,000 feet (15,240 m) of cable produced, or major fraction thereof, or at least once per jacket extruder run. (2) A Spark Test shall be performed on nonconducting jacketed cable in accordance with ANSI/ICEA S-94-649 on 100 percent of the completed cable prior to its being wound on shipping reels. The test voltage shall be 4.5 kV ac for cable

12. Miscellaneous

a. All cable provided under this specification shall have suitable markings on the outer surface of the jacket at sequential intervals not exceeding two feet (0.61 m). The label shall indicate the name of the manufacturer, conductor size, type and thickness of insulation, center conductor material, voltage rating, year of manufacture, and jacket type. There shall be no more than six inches (0.15 m) of unmarked spacing between text label sequences. The jacket shall be marked with the symbol required by Rule 350G of the National Electrical Safety Code and the purchaser shall specify any markings required by local safety codes. This is in addition to extruded red stripes required in paragraph 9.a. (1) of this specification.

diameters <1.5 inches and 7.0 kV for cable diameters >1.5 inches., and shall be applied between an electrode at the outer surface of the nonconducting jacket and the concentric neutral for not less than 0.15 second. d. Frequency of sample tests shall be in accordance with ANSI/ICEA S-94-649. e. If requested by the purchaser, a certified copy of the results of all tests performed in accordance to this section shall be furnished on all orders.

b. Watertight seals shall be applied to all cable ends to prevent the entrance of moisture during transit or storage. Each end of the cable shall be firmly and properly secured to the reel. c. Cable shall be placed on shipping reels suitable for protecting it from damage during shipment and handling. After the cable is wound on the reel, it shall be covered with a suitable covering to help provide physical protection to the cable. d. A durable label shall be securely attached to each reel of cable. The label shall indicate the purchaser’s name and address, purchase order number, cable description, reel number, feet of cable on the reel, tare and gross weight of the reel, and beginning and ending sequential footage numbers.

4 0 4 – Ap p e n d i x E

E Underground Cable Specification ATTACHMENT “A” Cooperative Name: _________________________

Contact: __________________________________

Phone #: ____________

E-Mail: ___________________________________

Fax #:______________

Conductor Material:

Aluminum

Copper

Conductor Type:

Solid

Stranded

Conductor Size:__________________________________________________________________________ Voltage Rating:

15 kV

25 kV

35 kV

Conductor Shield Compound(s): ___________________________________________________________ Insulation Type:

EPR

TR-XLPE

Alternate Insulation Thickness:

Either

Min.________

Nominal _________

Max. _________

Alternate Insulation Compound(s): _________________________________________________________ Insulation Shield Compound(s): ____________________________________________________________ Neutral Design:

Full

1/3

1/6

1/8

1/12

Outer Jacket Type:

Semi-Conducting:

Yes _______________

No________________

Reel Type:

Returnable:

Yes _______________

No________________

Wood Lagging Required:

Yes _______________

No________________

Maximum Reel Size (inches):

Width _____________

Diameter __________

Maximum Loaded Reel Weight (pounds): ____________________________________________________ Shipping Address:________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________

Shipping Method:

Flanges parallel with trailer centerline Flanges perpendicular to trailer centerline

Note: Axis of arbor holes must be horizontal

Additional Comments: ____________________________________________________________________ ____________________________________________________________________ Signature: _______________________________________________________________________________

Allowable Short Circuit Currents for Solid Dielectric Insulated C a b l e s – 4 0 5

F

Allowable Short Circuit Currents for Solid Dielectric Insulated Cables

Figures F.1 through F.8 show the allowable short circuit current duration for common configurations of solid dielectric cable. Figures F.1, F.2, F.5, and F.6 assume the prefault conductor temperature is 75°C for cables with thermoplastic insulation. Figures F.3, F.4, F.7, and F.8 assume a prefault conductor temperature of 90°C for cables with thermoset insulation. Figures F.1, F.2, F.3, and F.4 use an upper temperature limit that, if exceeded, would cause immediate permanent damage to the cable insulation. Figures F.5, F.6, F.7, and F.8 use the upper temperature limit equal to the emergency rating of the insulation that, if exceeded, adds incrementally to a loss of useful life of the cable. It is recommended that Figures F.5 through F.8 be used in the selection of overcurrent protection system for cables. Usually, this will not pose a problem in the overall coordination scheme.

Figure

Time-Current Characteristic

F.1

PE/HMWPE Insulation, Aluminum Conductor, 150°C Final

F.2

PE/HMWPE Insulation, Copper Conductor, 150°C Final

F.3

TR-XLPE/EPR Insulation, Aluminum Conductor, 250°C Final

F.4

TR-XLPE/EPR Insulation, Copper Conductor, 250°C Final

F.5

PE/HMWPE Insulation, Aluminum Conductor, 90°C Final

F.6

PE/HMWPE Insulation, Copper Conductor, 90°C Final

F.7

TR-XLPE/EPR Insulation, Aluminum Conductor, 130°C Final

F.8

TR-XLPE/EPR Insulation, Copper Conductor, 130°C Final

4 0 6 – Ap p e n d i x F

F 60 50

3,600 3,000

40

2,400

30

1,800

.2

12

.1 .09 .08 .07 .06 .05

6.0 5.4 4.8 4.2 3.6 3.0

.04

2.4

.03

1.8

.02

1.2

.01

.6


18

50,000

.3

40,000

24

30,000

.4

20,000

60 54 48 42 36 30

6,000 7,000 8,000 9,000 10,000

1 .9 .8 .7 .6 .5

5,000

120

4,000

2

3,000

180

2,000

3

600 700 800 900 1,000

240

500

4

400

600 540 480 420 360 300

300

10 9 8 7 6 5

200

1,200

100

Time (Seconds)

750

500

350

250 4/0

3/0

2/0

1/0

#2

#1

20

Current (Amperes)

FIGURE F.1: Aluminum Conductor/Thermoplastic Insulation (PE/HMWPE). Allowable Short Circuit Currents Based on 75°C Initial Conductor Temperature and 150°C Final Temperature.


F 60 50

3,600 3,000

40

2,400

30

1,800

.2

12

.1 .09 .08 .07 .06 .05

6.0 5.4 4.8 4.2 3.6 3.0

.04

2.4

.03

1.8

.02

1.2

.01

.6

Current (Amperes)

FIGURE F.2: Copper Conductor/Thermoplastic Insulation (PE/HMWPE). Allowable Short Circuit Currents Based on 75°C Initial Conductor Temperature and 150°C Final Temperature.


18

50,000

.3

40,000

24

30,000

.4

20,000

60 54 48 42 36 30

6,000 7,000 8,000 9,000 10,000

1 .9 .8 .7 .6 .5

5,000

120

4,000

2

3,000

180

2,000

3

600 700 800 900 1,000

240

500

4

400

600 540 480 420 360 300

300

10 9 8 7 6 5

200

1,200

100

Time (Seconds)

750

500

350

250 4/0

3/0

2/0

1/0

#1

#2

20

4 0 8 – Ap p e n d i x F

F 60 50

3,600 3,000

40

2,400

30

1,800

.2

12

.1 .09 .08 .07 .06 .05

6.0 5.4 4.8 4.2 3.6 3.0

.04

2.4

.03

1.8

.02

1.2

.01

.6


18

50,000

.3

40,000

24

30,000

.4

20,000

60 54 48 42 36 30

6,000 7,000 8,000 9,000 10,000

1 .9 .8 .7 .6 .5

5,000

120

4,000

2

3,000

180

2,000

3

600 700 800 900 1,000

240

500

4

400

600 540 480 420 360 300

300

10 9 8 7 6 5

200

1,200

100

Time (Seconds)

750

500

350

250 4/0

3/0

2/0

1/0

#2

#1

20

Current (Amperes)

FIGURE F.3: Aluminum Conductor/Thermoset Insulation (TR-XLPE/EPR). Allowable Short Circuit Currents Based on 90°C Initial Conductor Temperature and 250°C Final Conductor Temperature.


F 60 50

3,600 3,000

40

2,400

30

1,800

.2

12

.1 .09 .08 .07 .06 .05

6.0 5.4 4.8 4.2 3.6 3.0

.04

2.4

.03

1.8

.02

1.2

.01

.6

Current (Amperes)

FIGURE F.4: Copper Conductor/Thermoset Insulation (TR-XLPE/EPR). Allowable Short Circuit Currents for 90°C Rated Insulation Based on 90°C Initial Conductor Temperature and 250°C Final Conductor Temperature.


18

50,000

.3

40,000

24

30,000

.4

20,000

60 54 48 42 36 30

6,000 7,000 8,000 9,000 10,000

1 .9 .8 .7 .6 .5

5,000

120

4,000

2

3,000

180

2,000

3

600 700 800 900 1,000

240

500

4

400

600 540 480 420 360 300

300

10 9 8 7 6 5

200

1,200

100

Time (Seconds)

750

500

350

250 4/0

3/0

2/0

1/0

#1

#2

20

4 1 0 – Ap p e n d i x F

F 60 50

3,600 3,000

40

2,400

30

1,800

.2

12

.1 .09 .08 .07 .06 .05

6.0 5.4 4.8 4.2 3.6 3.0

.04

2.4

.03

1.8

.02

1.2

.01

.6


18

50,000

.3

40,000

24

30,000

.4

20,000

60 54 48 42 36 30

6,000 7,000 8,000 9,000 10,000

1 .9 .8 .7 .6 .5

5,000

120

4,000

2

3,000

180

2,000

3

600 700 800 900 1,000

240

500

4

400

600 540 480 420 360 300

300

10 9 8 7 6 5

200

1,200

100

Time (Seconds)

750

500

350

250 4/0

3/0

1/0

2/0

#1

#2

20

Current (Amperes)

FIGURE F.5: Aluminum Conductor/Thermoplastic Insulation (PE/HMWPE). Allowable Short Circuit Currents for Conductor to Not Exceed Insulation Emergency Operating Temperature Rating Based on 75°C Initial Conductor Temperature and 90°C Final Conductor Temperature.


F 60 50

3,600 3,000

40

2,400

30

1,800

.2

12

.1 .09 .08 .07 .06 .05

6.0 5.4 4.8 4.2 3.6 3.0

.04

2.4

.03

1.8

.02

1.2

.01

.6


18

50,000

.3

40,000

24

30,000

.4

20,000

60 54 48 42 36 30

6,000 7,000 8,000 9,000 10,000

1 .9 .8 .7 .6 .5

5,000

120

4,000

2

3,000

180

2,000

3

600 700 800 900 1,000

240

500

4

400

600 540 480 420 360 300

300

10 9 8 7 6 5

200

1,200

100

Time (Seconds)

750

500

350

250 4/0

2/0

3/0

1/0

#1

#2

20

Current (Amperes)

FIGURE F.6. Copper Conductor/Thermoplastic Insulation (PE/HMWPE). Allowable Short Circuit Currents for Conductor to Not Exceed Insulation Emergency Operating Temperature Rating Based on 75°C Initial Conductor Temperature and 90°C Final Conductor Temperature.

4 1 2 – Ap p e n d i x F

F 60 50

3,600 3,000

40

2,400

30

1,800

.2

12

.1 .09 .08 .07 .06 .05

6.0 5.4 4.8 4.2 3.6 3.0

.04

2.4

.03

1.8

.02

1.2

.01

.6


18

50,000

.3

40,000

24

30,000

.4

20,000

60 54 48 42 36 30

6,000 7,000 8,000 9,000 10,000

1 .9 .8 .7 .6 .5

5,000

120

4,000

2

3,000

180

2,000

3

600 700 800 900 1,000

240

500

4

400

600 540 480 420 360 300

300

10 9 8 7 6 5

200

1,200

100

Time (Seconds)

750

500

350

250 4/0

3/0

2/0

1/0

#1

#2

20

Current (Amperes)

FIGURE F.7: Aluminum Conductor/Thermoset Insulation (TR-XLPE/EPR). Allowable Short Circuit Currents for Conductor to Not Exceed Insulation Emergency Operating Temperature Rating Based on 90°C Initial Conductor Temperature and 130°C Final Conductor Temperature.


F 60 50

3,600 3,000

40

2,400

30

1,800

.2

12

.1 .09 .08 .07 .06 .05

6.0 5.4 4.8 4.2 3.6 3.0

.04

2.4

.03

1.8

.02

1.2

.01

.6


18

50,000

.3

40,000

24

30,000

.4

20,000

60 54 48 42 36 30

6,000 7,000 8,000 9,000 10,000

1 .9 .8 .7 .6 .5

5,000

120

4,000

2

3,000

180

2,000

3

600 700 800 900 1,000

240

500

4

400

600 540 480 420 360 300

300

10 9 8 7 6 5

200

1,200

100

Time (Seconds)

750

500

350

250 4/0

3/0

2/0

1/0

#1

#2

20

Current (Amperes)

FIGURE F.8. Copper Conductor/Thermoset Insulation (TR-XLPE/EPR). Allowable Short Circuit Currents for Conductor to Not Exceed Insulation Emergency Operating Temperature Rating Based on 90°C Initial Conductor Temperature and 130°C Final Conductor Temperature.


Ampacity Ta b l e s – 4 1 5

G

Ampacity Tables

comprehensive set of cable ampacity calculations, the limiting ampacity for soil interface temperature criteria is not readily available from this Standard. It is suggested that both references be used to determine the ampacity limits for the circumstances under evaluation. A copy of Figure 4.10 is included at the end of this appendix for reference.

Cable ampacity values shown in Appendix G are based on calculations provided by Okonite Company for the 1992 edition of this manual. They have been retained because values are included for cases where the soil interface temperature between the cable and the soil, or between the conduit and the soil, govern. While IEEE Standard 835-1994 does provide a more

TABLE G.1: Configuration No. 1—15-kV Copper. 75% Load Factor

100% Load Factor

Soil Interface Temp. Amps at 60ºC

Amps


Conductor Size

Amps

4/0 (1/3 neutral)

390

69ºC

350

350

74ºC

308

350 (1/3 neutral)

506

72ºC

446

450

76ºC

385

500 (1/3 neutral)

603

74ºC

525

532

77ºC

445

750 (1/3 neutral)

689

75ºC

580

602

79ºC

495

1,000 (1/6 neutral)

804

76ºC

675

700

79ºC

575

TABLE G.2: Configuration No. 1—15-kV Aluminum. 75% Load Factor Conductor Size

Amps

100% Load Factor


Amps


1/0 (full neutral)

207

67ºC

192

187

71ºC

168

4/0 (1/3 neutral)

308

69ºC

278

276

73ºC

241

350 (1/3 neutral)

406

71ºC

364

362

75ºC

313

500 (1/3 neutral)

488

73ºC

426

432

77ºC

367

750 (1/3 neutral)

593

74ºC

512

521

78ºC

436

1,000 (1/6 neutral)

698

75ºC

596

609

78ºC

504

4 1 6 – Ap p e n d i x G

G TABLE G.3: Configuration No. 1—25-kV Copper. 75% Load Factor

100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

384

65ºC

361

346

70ºC

310

350 (1/3 neutral)

499

68ºC

454

445

73ºC

391

500 (1/3 neutral)

591

70ºC

528

522

74ºC

446

750 (1/3 neutral)

688

72ºC

602

602

76ºC

511

1,000 (1/6 neutral)

806

73ºC

698

703

77ºC

584


Amps

100% Load Factor


Amps


1/0 (full neutral)

203

63ºC

197

185

67ºC

171

4/0 (1/3 neutral)

302

65ºC

287

272

70ºC

247

350 (1/3 neutral)

399

67ºC

373

357

72ºC

318

500 (1/3 neutral)

481

69ºC

435

427

73ºC

375

750 (1/3 neutral)

587

70ºC

526

517

75ºC

442

1,000 (1/6 neutral)

692

71ºC

619

606

75ºC

517


100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

413

68ºC

376

364

73ºC

314

350 (1/3 neutral)

505

72ºC

442

438

77ºC

368

500 (1/3 neutral)

570

76ºC

483

489

79ºC

401

750 (1/3 neutral)

654

78ºC

535

557

81ºC

444

1,000 (1/6 neutral)

714

78ºC

558

606

82ºC

482


Amps

100% Load Factor


Amps


1/0 (full neutral)

231

63ºC

223

206

68ºC

188

4/0 (1/3 neutral)

340

66ºC

317

301

71ºC

267

350 (1/3 neutral)

430

70ºC

385

376

75ºC

324

500 (1/3 neutral)

499

73ºC

433

431

77ºC

363

750 (1/3 neutral)

578

76ºC

485

494

80ºC

401

1,000 (1/6 neutral)

666

76ºC

565

570

79ºC

468



100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

404

64ºC

388

358

70ºC

323

350 (1/3 neutral)

500

69ºC

454

437

74ºC

379

500 (1/3 neutral)

562

71ºC

496

487

76ºC

413

750 (1/3 neutral)

648

74ºC

554

556

78ºC

455

1,000 (1/6 neutral)

719

76ºC

606

613

80ºC

494


Amps

100% Load Factor


Amps


1/0 (full neutral)

222

59ºC

222

200

65ºC

191

4/0 (1/3 neutral)

328

62ºC

320

293

68ºC

269

350 (1/3 neutral)

421

66ºC

393

371

71ºC

329

500 (1/3 neutral)

486

68ºC

445

425

73ºC

370

750 (1/3 neutral)

569

72ºC

498

491

76ºC

413

1,000 (1/6 neutral)

666

73ºC

577

572

77ºC

480

EXAMPLE G.1: Ampacity Reduction for Direct-Buried Versus Conduit Encasement for Flat-Spaced Installation. TABLE G.9: Configuration No. 2, 3" Type DB Conduit—15-kV Aluminum. 75% Load Factor

100% Load Factor

Conductor Size

Amps

Soil Interface Temp.

Amps at 60ºC

Amps


Amps at 60ºC

750 (1/3 neutral)

467

55ºC

467

420

62ºC

408

TABLE G.10: Configuration No. 2, 3.5" Type DB Conduit—25-kV Aluminum. 75% Load Factor Conductor Size

Amps


750 (1/3 neutral)

479

55ºC

100% Load Factor Amps at 60ºC

Amps


Amps at 60ºC

479

430

61ºC

424

Tables G.9 and G.10 show the effect of encasing the single cables of Configuration No. 2 in Type DB conduit instead of direct burying the cables. Using conduit decreases ampacity by 19.2% for 15-kV cable and 15.82% for 25-kV cable.

4 1 8 – Ap p e n d i x G


100% Load Factor

Conductor Size

Amps


Amps


4/0 (1/3 neutral)

307

46ºC

290

51ºC

350 (1/3 neutral)

407

47ºC

376

53ºC

500 (1/3 neutral)

474

49ºC

436

55ºC

750 (1/3 neutral)

557

49ºC

508

56ºC

1,000 (1/6 neutral)

649

50ºC

590

57ºC


100% Load Factor

Amps


Amps


1/0 (full neutral)

162

45ºC

152

50ºC

4/0 (1/3 neutral)

242

46ºC

226

52ºC

350 (1/3 neutral)

326

47ºC

302

53ºC

500 (1/3 neutral)

392

48ºC

361

55ºC

750 (1/3 neutral)

484

49ºC

441

56ºC

1,000 (1/6 neutral)

568

50ºC

516

57ºC


100% Load Factor

Conductor Size

Amps


Amps


4/0 (1/3 neutral)

315

45ºC

293

51ºC

350 (1/3 neutral)

408

47ºC

377

54ºC

500 (1/3 neutral)

488

48ºC

447

55ºC

750 (1/3 neutral)

563

50ºC

513

57ºC

1,000 (1/6 neutral)

658

51ºC

597

57ºC


100% Load Factor

Amps


Amps


1/0 (full neutral)

169

44ºC

158

50ºC

4/0 (1/3 neutral)

249

45ºC

231

51ºC

350 (1/3 neutral)

327

47ºC

302

53ºC

500 (1/3 neutral)

401

47ºC

368

54ºC

750 (1/3 neutral)

485

49ºC

443

56ºC

1,000 (1/6 neutral)

570

50ºC

519

57ºC



100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

357

73ºC

312

313

77ºC

266

350 (1/3 neutral)

463

75ºC

400

402

79ºC

337

500 (1/3 neutral)

537

77ºC

454

463

80ºC

379

750 (1/3 neutral)

616

78ºC

512

526

81ºC

423

1,000 (1/6 neutral)

717

79ºC

582

610

82ºC

484


Amps

100% Load Factor


Amps


1/0 (full neutral)

191

70ºC

170

169

74ºC

146

4/0 (1/3 neutral)

281

73ºC

246

247

77ºC

210

350 (1/3 neutral)

369

74ºC

320

321

78ºC

270

500 (1/3 neutral)

441

76ºC

378

381

80ºC

314

750 (1/3 neutral)

532

77ºC

446

456

80ºC

374

1,000 (1/6 neutral)

624

78ºC

516

532

81ºC

431


100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

352

69ºC

318

309

74ºC

270

350 (1/3 neutral)

454

72ºC

396

395

76ºC

335

500 (1/3 neutral)

534

74ºC

460

461

78ºC

385

750 (1/3 neutral)

618

76ºC

518

527

80ºC

434

1,000 (1/6 neutral)

719

76ºC

604

613

80ºC

498


Amps

100% Load Factor


Amps


1/0 (full neutral)

188

67ºC

174

167

72ºC

149

4/0 (1/3 neutral)

277

69ºC

250

244

74ºC

213

350 (1/3 neutral)

364

71ºC

324

318

76ºC

273

500 (1/3 neutral)

436

73ºC

381

378

77ºC

317

750 (1/3 neutral)

528

74ºC

456

454

78ºC

380

1,000 (1/6 neutral)

620

75ºC

534

531

79ºC

439

4 2 0 – Ap p e n d i x G


100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

290

51ºC

290

264

57ºC

264

350 (1/3 neutral)

382

52ºC

382

345

59ºC

345

500 (1/3 neutral)

443

54ºC

443

398

61ºC

395

750 (1/3 neutral)

516

55ºC

516

459

62ºC

447

1,000 (1/6 neutral)

583

54ºC

583

532

63ºC

509


Amps

100% Load Factor


Amps


1/0 (full neutral)

154

49ºC

154

142

55ºC

142

4/0 (1/3 neutral)

229

51ºC

229

209

57ºC

209

350 (1/3 neutral)

306

52ºC

306

277

58ºC

277

500 (1/3 neutral)

367

53ºC

367

330

60ºC

330

750 (1/3 neutral)

449

55ºC

449

400

62ºC

392

1,000 (1/6 neutral)

525

56ºC

525

466

63ºC

451


100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

297

50ºC

297

270

57ºC

270

350 (1/3 neutral)

383

53ºC

383

345

59ºC

345

500 (1/3 neutral)

454

54ºC

454

406

61ºC

403

750 (1/3 neutral)

521

56ºC

521

463

63ºC

451

1,000 (1/6 neutral)

607

56ºC

607

538

64ºC

515


Amps

100% Load Factor


Amps


1/0 (full neutral)

160

49ºC

160

146

55ºC

146

4/0 (1/3 neutral)

234

50ºC

234

213

57ºC

213

350 (1/3 neutral)

307

52ºC

307

277

59ºC

277

500 (1/3 neutral)

374

53ºC

374

335

60ºC

335

750 (1/3 neutral)

450

55ºC

450

401

62ºC

393

1,000 (1/6 neutral)

527

56ºC

527

468

63ºC

453



100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

240

63ºC

233

209

70ºC

189

350 (1/3 neutral)

312

65ºC

296

268

71ºC

238

500 (1/3 neutral)

358

66ºC

333

306

73ºC

268

750 (1/3 neutral)

410

68ºC

375

348

74ºC

298

1,000 (1/6 neutral)

474

69ºC

429

400

75ºC

342


Amps

100% Load Factor


Amps


1/0 (full neutral)

130

61ºC

129

114

68ºC

106

4/0 (1/3 neutral)

191

63ºC

186

166

69ºC

151

350 (1/3 neutral)

251

64ºC

238

216

71ºC

194

500 (1/3 neutral)

298

66ºC

278

255

72ºC

224

750 (1/3 neutral)

358

68ºC

328

304

74ºC

264

1,000 (1/6 neutral)

416

68ºC

381

352

75ºC

303


100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

246

63ºC

235

213

70ºC

192

350 (1/3 neutral)

312

65ºC

296

268

72ºC

238

500 (1/3 neutral)

364

66ºC

339

310

73ºC

272

750 (1/3 neutral)

414

68ºC

378

350

74ºC

300

1,000 (1/6 neutral)

479

69ºC

430

405

75ºC

346


Amps

100% Load Factor


Amps


1/0 (full neutral)

133

61ºC

132

116

68ºC

107

4/0 (1/3 neutral)

194

63ºC

187

167

70ºC

152

350 (1/3 neutral)

251

65ºC

238

216

71ºC

192

500 (1/3 neutral)

302

66ºC

283

257

72ºC

226

750 (1/3 neutral)

359

68ºC

329

304

74ºC

262

1,000 (1/6 neutral)

418

68ºC

380

354

75ºC

304

4 2 2 – Ap p e n d i x G

G EXAMPLE G.2: Increase in Ampacity for Duct Bank Installation When Type EB Conduit is Used Versus Schedule 40. TABLE G.27: Configuration No. 6, 6" Type EB Conduit—15-kV Aluminum. 75% Load Factor Conductor Size

Amps


750 (1/3 neutral)

364

69ºC

100% Load Factor Amps at 60ºC

Amps


Amps at 60ºC

331

307

75ºC

262

TABLE G.28: Configuration No. 6, 6" Type EB Conduit—25-kV Aluminum. 75% Load Factor

100% Load Factor

Conductor Size

Amps


Amps at 60ºC

Amps


Amps at 60ºC

750 (1/3 neutral)

365

69ºC

332

308

75ºC

263

Tables G.27 and G.28 show the effect of using Type EB conduit instead of Schedule 40 for the concrete duct bank installation shown in Configuration No. 6. Using the thinner walled conduit gives an increase in ampacity of approximately 1.67% (see Configuration No. 6 of the ampacity tables and Figure 4.10).


100% Load Factor


Amps


Conductor Size

Amps

4/0 (1/3 neutral)

228

66ºC

215

196

72ºC

173

350 (1/3 neutral)

294

68ºC

270

250

74ºC

216

500 (1/3 neutral)

337

69ºC

303

285

75ºC

242

750 (1/3 neutral)

384

71ºC

342

322

76ºC

270

1,000 (1/6 neutral)

443

72ºC

390

371

77ºC

310


Amps

100% Load Factor


Amps


1/0 (full neutral)

124

64ºC

119

107

70ºC

96

4/0 (1/3 neutral)

180

66ºC

170

155

72ºC

138

350 (1/3 neutral)

236

67ºC

219

201

74ºC

176

500 (1/3 neutral)

280

69ºC

255

237

75ºC

205

750 (1/3 neutral)

335

70ºC

299

282

76ºC

238

1,000 (1/6 neutral)

389

71ºC

345

326

77ºC

275



100% Load Factor



Conductor Size

Amps

Amps

4/0 (1/3 neutral)

232

66ºC

218

198

72ºC

175

350 (1/3 neutral)

294

68ºC

270

250

74ºC

215

500 (1/3 neutral)

342

69ºC

307

288

75ºC

245

750 (1/3 neutral)

387

71ºC

341

324

77ºC

272

1,000 (1/6 neutral)

448

72ºC

390

374

77ºC

310


Amps

100% Load Factor



Amps

1/0 (full neutral)

126

64ºC

121

109

71ºC

97

4/0 (1/3 neutral)

183

66ºC

172

156

72ºC

138

350 (1/3 neutral)

236

67ºC

219

201

74ºC

175

500 (1/3 neutral)

283

69ºC

258

239

75ºC

206

750 (1/3 neutral)

336

70ºC

300

282

76ºC

238

1,000 (1/6 neutral)

391

71ºC

346

327

77ºC

276

B 7.5"

Configuration 4

36"

36"

36"

A

C 18"

7.5"

Configuration 7

5"

30"

30"

Configuration 6

36"

5" 19"

18"

7.5"

Configuration 5

Configuration 3

36"

Configuration 2

7.5”

7.5"

7.5"

19"

26.5"

19" × 19" Duct Bank

19" × 26.5" Duct Bank

FIGURE 4.10: Three-Phase Cable Installation Configurations.

19"

Configuration 1


Industry Specifica t i o n s – 4 2 5

H

Industry Specifications

1. RUS Bulletin 1728C-100 List of Materials Acceptable for Use on Systems of RUS Electrification Borrowers 2. IEEE Standard 404 Standard for Extruded and Laminated Dielectric Shielded Cable Joints Rated 2,500 V to 500,000 V 3. IEEE Standard 592 Exposed Semiconducting Shields on High-Voltage Cable Joints and Separable Insulated Connectors 4. IEEE Standard 48 Standard Test Procedures and Requirements for Alternating-Current Cable Terminations 2.5 kV through 765 kV 5. NEMA/ANSI C 119.0 Sealed Insulated Underground Connector Systems Rated 600 Volts 6. RUS Bulletin 1728F-U1 RUS Specification for 15-kV, 25-kV, and 35-kV Primary Underground Power Cable 7. RUS Bulletin 50-6 (D-806) Specifications and Drawings for Underground Electric Distribution 8. ICEA S-81-570 Standard for 600-Volt-Rated Cables of Ruggedized Design for Direct Burial Installation as Single Conductors or Assemblies of Single Conductors 9. ICEA S-105-692 600-Volt Single Layer Thermoset Insulated Utility Underground Distribution Cables

10. ANSI/ICEA S-94-649 Concentric Neutral Cables Rated 5 through 46 kV 11. ANSI/ICEA S-97-682 Utility Shielded Power Cables Rated 5 through 46 kV 12. IEEE Standard 495 Guide for Testing Faulted Circuit Indicators 13. ANSI/IEEE C57.91 IEEE Guide for Loading Mineral-OilImmersed Overhead and Pad-Mounted Distribution Transformers Rated 500 kVA and Less with 65°C or 55°C Average Winding Rise 14. BSR/IEEE C57.12.00 Standard General Requirements for LiquidImmersed Distribution, Power, and Regulating Transformers 15. IEEE Standard C12.26 Pad-Mounted Compartmental-Type, Self-Cooled, Three-Phase Distribution Transformers for Use with Separable Insulated High-Voltage Connectors (34,500 Grd Y/19,920 Volts and Below, 2,500 kVA and Smaller) 16. IEEE Standard 835 Standard Power Cable Ampacity Tables 17. NEMA MG 1-12.35 Locked-Rotor Current of 3-Phase 60-Hz Small and Medium Squirrel-Cage Induction Motors Rated at 230 Volts 18. IEEE Standard C62.11 Standard for Metal-Oxide Surge Arresters for AC Power Circuits (>1 kV)

4 2 6 – Ap p e n d i x H

H 19. IEEE Standard C62.1 Standard for Gapped Silicon-Carbide Surge Arresters for AC Power Circuits 20. IEEE Standard 81 Guide for Measuring Earth Resistivity, Ground Impedance, and Earth Surface Potentials of a Ground System 21. IEEE Standard 80 Guide for Safety in AC Substation Grounding 22. RUS Bulletin 169-4

Voltage Levels on Rural Distribution Systems 23. IEEE Standard 386 Standard for Separable Insulated Connector Systems for Power Distribution Systems Above 600 V 24. REA Bulletin U2 February 1975. Out of print but may be available by special request to RUS Electric staff, Washington, D.C.

Component Manufac t u re r s – 4 2 7

I

Component Manufacturers

TABLE I.1: Cable Installation Equipment Manufacturers.

Manufacturer

Trenchers

Guided Backhoes Cable Plow Boring Tools

Piercing Tools

TrackHydraulic Mounted Trench Auger-Type Pipe Pusher Cable Plows Compactors Boring Tools

Am. Augers

X

Ditch Witch

X

Cleveland Trencher Co.

X

X

X

X

X

Holladay Constr. Co.

X

Pow-r-Devices

X

StraightLine Mfg. Co.

X

UtilX Corp. Vermeer Mfg. Co.

X X

X

X

X

X

X

American Augers, Inc. 135 US Route 42 P.O. Box 814 West Salem, OH 44287 800.324.4930 www.american-augers.com

Holladay Construction Co., Inc. 5419 Hickory Ridge Road Spotsylvania, VA 22553 540.582.2700 www.holladayconstco.com

UtilX Corp. P.O. Box 97009 Kent, WA 98064-9709 800.252.0556 www.utilx.com

The Charles Machine Works, Inc. (Manufacturer of Ditch Witch Equipment) P.O. Box 66 Perry, OK 73077-0066 800.654.6481 www.ditchwitch.com

Pow-r-Devices, Inc. 5940 Goodrich Road Clarence Center, NY 14032-0245 800.344.6653 www.powrdevices.com

Vermeer Manufacturing Co. 1210 Vermeer Road East P.O. Box 200 Pella, IA 50219 641.628.3141 www.vermeer.com

Cleveland Trencher Co. 1755 West Market Street Akron, OH 44313 330.869.2800 www.cleveland-trencher.com

StraightLine Manufacturing, Inc. (Finco Inc.) 1816 East Wasp Road Hutchinson, KS 67501 800.654.3484 www.straightlinehdd.com

4 2 8 – Ap p e n d i x I

I TABLE I.2: Cable Installation Equipment Manufacturers.

Joints— Primary Circuits Manufacturer

Elbows— Primary Circuits

Terminations—Primary Circuits

Joints— Secondary Circuits

Terminations— Secondary Circuits

Heat Elbow Heat Cold Above All Above All Premolded Shrink Premolded Connectors Premolded Porcelain Shrink Shrink Ground Locations Ground Locations

3M Elec. Prod.

X

X

X

X

X

Amp Inc.

X

X

X

X

Burndy Corp.

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

Cooper Power Systems - RPE

X

X

X

X

Elastimold, Thomas & Betts

X

X

X

X

X

Fargo Mfg. Co. G&W Electric Co.

X

Homac Mfg. Co. Kearney Raychem Corp.

X

X

X

Reliable Elec. Prod.

X

X

X

X

X

X

X

X

3M Electric Products Division 6801 River Place Road Austin, TX 78726 800.245.3573 www.mmm.com

Cooper Power Systems Components & Protective Equipment 1045 Hickory Street Pewaukee, WI 53072 www.cooperpower.com

G&W Electric Co., Inc. 3500 West 127th Street Blue Island, IL 60406 708.388.5010 www.gwelec.com

Raychem Corp. Tyco Electronics Corp. 300 Constitution Drive Menlo Park, CA 94025-1164 650.361.3333 www.raychem.com

AMP Inc. (Tyco) P.O. Box 3608 Harrisburg, PA 17105 www.amp.com

Elastimold, Thomas & Betts 8155 T&B Blvd. Memphis, TN 38125 888.862.3289 www.tnb.com

Homac Mfg. Co. 12 Southland Road Ormond Beach, FL 32174 386.673.5025 www.homac.com

Reliable Electrical Products MacLean Power Systems 1000 Allanson Road Mundelein, IL 60060 847.566.0010 www.maclean-fogg.com

Burndy Corporation 825 Old Trail Road Etters, PA 17319 800.346.4175 www.fciconnect.com

Fargo Mfg. Co. (Hubbell) 210 N. Allen Street Centralia, MO 65240 www.hubbellpowersystems.com

Kearney (Cooper Power Systems) 1319 Lincoln Avenue Waukesha, WI 53186 262.524.3300 www.cooperpower.com

Component Manufac t u re r s – 4 2 9

I TABLE I.3: Manufacturers of Joint, Elbow, and Termination Accessories and Kits. Elbow Accessories

Elbow & Termination Sealing Kits

Grounding Kits

Plugs

Adapters

Inserts

Cable Jacket Restoration Kits

3M Electric Products

X

X

X

X

X

X

Cooper Power Systems

X

X

X

Elastimold, Thomas & Betts

X

X

X

X

X

X

Homac Mfg.

X

X

X

Raychem

X

X

X

Manufacturer

TABLE I.4. Partial Listing of Cable Testing Equipment Suppliers. Sources of DC Proof Test Equipment Associated Research, Inc. 13860 W. Laurel Drive Lake Forest, IL 60045 800.858.8378 www.asresearch.com

Biddle Instruments 510 Township Line Road Blue Bell, PA 19422 866.586.3872 www.avobiddle.com

Hipotronics, Inc. 1620 Route 22 P.O. Box 414 Brewster, NY 10509 845.279.8091 www.hipotronics.com

The Von Corporation 1038 Lomb Avenue, S.W. P.O. Box 110096 Birmingham, AL 35211 205.788.2437 www.voncorp.com


Cable-Pulling Exa m pl e s – 4 3 1

J

Cable-Pulling Examples

EXAMPLE J.1: Cable-Pulling Example 1: Maximum Straight-Pull Distance for Three 25-kV Cables Installed in Five-Inch PVC Conduit. Determine the maximum straight-pull distance for three 25-kV cables installed in five-inch PVC conduit. The following specific data apply in this situation: Cable Type: Cable Jacket: Cable Size: Cable Weight (Each):

Aluminum, EPR, 25 kV, with 1/3 neutral Polyethylene 350 kcmil 1.64 lb./ft

Cable Outside Diameter: 1.83 in. Conduit Inside Diameter: 5.047 in. Pulling Lubricant: Soap and water Type of Attachment: Aluminum compression

The solution sequence follows the procedure given in the main text under Cable-Pulling Calculation Sequence. STEPS 1 and 2. Determine the cable and conduit data. These are readily determined from the data above. For a three-cable pull, cable weight is three times 1.64 lb./ft, yielding 4.92 lb./ft. STEP 3. Determine the friction factors. Table 9.16 gives a friction factor of 0.45 for three polyethylene jacketed cables installed in PVC conduit in a straight-pull situation. STEP 4(a). Calculate the jam ratio. The jam ratio is calculated from Equation 9.9. J=

D 5.047 inches = = 2.76 d 1.83 inches

This jam ratio value would be risky if bends were present in the conduit. However, the value of 2.76 is acceptable for straight pulls. STEP 4(b). Calculate the clearance factor. According to Equation 9.11, the clearance factor is 1.13 inches. This value is well above the 0.5-inch minimum acceptable value. STEP 4(c). Calculate the weight correction factor. Before calculating the weight correction factor, determine if the three cables will take on a triangular or a cradled configuration.

Table 9.19 shows that either configuration is likely when the jam ratio is 2.76. Therefore, a cradled configuration should be assumed because it leads to more conservative results. The weight correction factor can now be calculated from Equation 9.4: Wc = 1 +

2 4 1.83 × = 1.43 3 5.047 – 1.83

STEP 5(a). Determine the maximum allowable cable tension. Table 9.20 gives a value of 0.008 lb./cmil maximum pulling tension when an aluminum compression eye is used on stranded aluminum cable. The maximum tension for one 350-kcmil cable is: 350,000 cmil × 0.008 lb./cmil = 2,800 lb. When three cables are being pulled, the maximum allowable tension is determined by doubling the tension maximum of one cable, as it is assumed that the total load will be shared by two of the three cables. Therefore, the maximum allowable tension for the sample situation is as follows: 2 × 2,800 lb. = 5,600 lb. Continued

4 3 2 – Ap p e n d i x J

J EXAMPLE J.1: Cable-Pulling Example 1: Maximum Straight-Pull Distance for Three 25-kV Cables Installed in Five-Inch PVC Conduit. (cont.) STEP 5(b). Determine the maximum allowable sidewall bearing pressure (SWBP). Consult Table 9.18 for this purpose. The maximum allowable SWBP is 2,000 lb./ft for jacketed EPR cable. However, SWBP is not a concern for straight pulls. STEP 6(a). Calculate the pulling tension per foot. Pulling tension for straight pulls is calculated from Equation 9.3:

STEP 6(b). Determine the maximum straight-pull distance. Assume 50 lb. of tension exists from the cable reel to the conduit entrance. Since the maximum allowable tension is 5,600 lb., a tension of 5,600 lb. less 50 lb. (5,550 lb.) is allowed for pulling tension. Therefore, the maximum pulling distance is calculated as follows: IMAX =

5,550 lb. = 1,753 ft 3.166 lb./ft

T = W × WC × f ×l T(1 foot) = 4.92 lb./ft × 1.43 × 0.45 × 1 ft T(1 foot) = 3.166 lb.

EXAMPLE J.2: Cable-Pulling Example 2: Feasibility of Pulling Three 25-kV Cables into a Six-Inch PVC Conduit. Determine the feasibility of pulling three cables of the same type described in Example J.1 into an installation of six-inch PVC conduit consisting of the following sections: A. 100-foot horizontal straight pull beginning at a manhole location B. 22-1/2° bend and beginning of upward slope C. 500-foot upward 1:20 slope

D. 90° bend at base of riser pole E. 30-foot vertical section at riser pole

The cable-reel end is to be at the manhole and the pulling end at the riser. STEPS 1 and 2. Determine the cable and conduit data. Cable data, lubricant, and type of grip are the same as given in Example J.1. For Example J.2, six-inch conduit of 6.065 inches inside diameter is to be used. STEP 3. Determine the friction factors. For three polyethylenejacketed cables installed in PVC conduit, Table 9.16 gives a friction factor of 0.45 for straight pulls and 0.15 for pulls through bends where SWBP exceeds 150 lb./ft. STEP 4(a). Calculate the jam ratio. The jam ratio is calculated from Equation 9.9: D J= = d

6.065 inches 1.83 inches

= 3.31

This jam ratio value is acceptable. STEP 4(b). Calculate the clearance factor. The clearance factor is calculated from Equation 9.11 and is found to be 2.2 inches. This greatly exceeds the 0.5-inch required minimum.

STEP 4(c). Calculate the weight correction factor. Table 9.19 reveals that the cables will take on a cradled configuration for the jam ratio of 3.31 calculated above. Therefore, Equation 9.4 is used to calculate the weight correction factor: Wc = 1 +

2 4 1.83 × = 1.25 3 6.065 – 1.83

STEP 5(a). Determine the maximum allowable cable tension. This tension limit calculation is identical to that found in Example J.1. The maximum allowable pulling tension is 5,600 lb. STEP 5(b). Determine the maximum allowable SWBP. Table 9.18 is consulted for this purpose, and the maximum allowable SWBP is found to be 2,000 lb./ft for jacketed EPR cable. STEP 6(a). Calculate the tension for the 100-foot horizontal straight pull beginning at the manhole location. The tension calculated by Equation 9.3 is added to the entering tension from the reel, which is assumed to be 50 lb. T2 = T1 + W × WC × f ×l T2 = 50 + (4.92)(1.25)(0.45)(100) = 327 lb. Continued

Cable-Pulling Exa m pl e s – 4 3 3

J EXAMPLE J.2: Cable-Pulling Example 2: Feasibility of Pulling Three 25-kV Cables into a Six-Inch PVC Conduit. (cont.) STEP 6(b). Calculate the tension for the 22-1/2° bend. Equation 9.3.C applies for calculating the tension increase resulting from pulling around conduit bends. SWBP is not expected to exceed 150 lb./ft for this bend, so the friction factor of 0.45 is used. The angle of the bend must be stated in radians, and 22-1/2° is 22.2 × 0.01745 radians/degree = 0.3927 radians. T1 is the tension result from the previous steps. T2 = T1ef × WC × φ T2 = (327)e(0.45)(1.25)(0.3927) = 406 lb. When a bend is involved, SWBP must also be calculated. Equation 9.7 applies in the present case. The typical value for R, the inside radius of a bend, for six-inch conduit is 2.75 feet. SWBP =

(3WC – 2)T2 (3.75 – 2)(408) = = 87 lb./ft 3R (3)(2.75)

This result is far less than the 2,000 lb./ft limit and is fully acceptable. As expected, the value is also less than 150 lb./ft, confirming the use of 0.45 as the correct friction factor. STEP 6(c). Calculate the tension for the 500-foot upward 1:20 slope. Equation 9.5.A applies in this case. As a preliminary step, the slope angle, θ, is calculated from the 1:20 slope ratio: θ = Tan–1

1 = 2.86º 20

Equation 9.5.A is then applied as an increment to the tension T1 of 408 lb. from the previous step. T2 = T2 + lW(fWC cosθ + sinθ) T2 = 408 + (500)(4.92)[(0.45)(1.25)(0.99875) + 0 .05] T2 = 408 + (2,460)(0.6118) = 408 + 1,505 = 1,913 lb.

STEP 6(d). Calculate the tension for the 90° bend at the base of the riser pole. At this location, cable tension has increased to the point that SWBP is expected to exceed 150 lb./ft. Therefore, the friction factor of 0.15 for bends may be used (see Table 9.16). The 90° angle of bend converts to 1.5708 radians. Equation 9.3.C is used. T2 = (1,913)e(0.15)(1.25)(1.5708) = 2,568 lb. The calculation of SWBP is also required. The radius of the bend is assumed to be 2.75 feet, and Equation 9.7 applies. SWBP =

(3WC – 2)T2 (3.75 – 2)2,568 = = 545 lb./ft 3R (3)(2.75)

The result is well within the 2,000 lb./ft allowed maximum. It is also above the 150 lb./ft value necessary to allow use of 0.15 as the friction factor. STEP 6(e). Calculate the tension for the 30-foot vertical section at the riser pole. A vertical rise is equivalent to an upward slope of 90°, so Equation 9.5.A applies. The evaluations cos 90° = 0 and sin 90° = 1 yield the following simplified form of the tension equation: T2 = T1 + lW = 2,568 + (30)(4.92) T2 = 2,568 + 148 = 2,716 lb. This result for final tension is less than half the 5,600-lb. maximum allowed, so the proposed cable-pulling operation is feasible.


Abbrevi a t i o n s – 4 3 5

abbrevia t i on s Ω-m

Ohm-meters (Ohm-m), a unit of measure for volume resistivity Microhenry, one-millionth of a Henry Micrometer, one-millionth of a meter Microsecond, one-millionth of a second Percentage Impedance

HDPE HMWPE Hp HV HVAC

IACS

AWG

Acrylonitrile-Butadiene-Styrene Plastic Alternating Current (sometimes AC) Association of Edison Illuminating Companies American National Standards Institute Average Service Availability Index American Society for Testing and Materials American Wire Gauge

BCN BIL

Bare Concentric Neutral Basic Impulse Insulation Level

CATV cm cmil Cont. CPE CRN

Cable Access Television Centimeter Circular Mil Continuous Chlorinated Polyethylene NRECA’s Cooperative Research Network Certified Test Results Chemical symbol for Copper Continuous Vulcanizing Chopped Wave Withstand

µH µm µs %Z ABS ac AEIC ANSI ASAI ASTM

CTR Cu CV CWW

Hz

High-Density Polyethylene High-Molecular-Weight Polyethylene Horsepower High Voltage Heating, Ventilation, and Air Conditioning Hertz

International Annealed Copper Standard ICEA Insulated Cable Engineers Association, Inc. I.D. Inside Diameter IEEE Institute of Electrical and Electronics Engineers in2 Square Inches IR Any Product of Current (I) Times Resistance (R) IZSURGE Current Times Surge Impedance JCN

Jacketed Concentric Neutral

kA kcmil

kV kVA kV/ft kW

Kiloamperes Thousand Circular Mil, wire size commonly used for multiple stranded conductors over 4/0 AWG in size (formerly MCM) Kilofoot (1,000 feet) Kips Per Square Inch (Thousands of Pounds Per Square Inch) Kilovolt (1,000 Volts) Kilovolt Amperes Kilovolts Per Foot Kilowatt

L lb. L.C. LDPE LLDPE

Inductance Pound(s) Longitudinally Corrugated Low-Density Polyethylene Linear Low-Density Polyethylene

mA

Milliampere, One-Thousandth of an Ampere See Kcmil Maximum Continuous Operating Voltage Medium-Density Polyethylene Multigrounded Neutral Metal Oxide Varistor, a type of surge arrester Square Millimeters

kft ksi

DB dc di/dt

Direct Burial (conduit classification) Direct Current (sometimes DC) Change in Current with Time (usually expressed as kA/µs)

EB EC EMT EPR EPRI EVA

Encased Burial (conduit classification) Electrical Conductor (grade of aluminum) Electrical Metallic Tubing Ethylene Propylene Rubber Electric Power Research Institute Ethyl Vinyl Acetate

FCI FOW FRE

Faulted-Circuit Indicator Front-of-Wave Fiberglass-Reinforced Epoxy

MDPE MGN MOV

H

Henry, a unit of inductance

mm2

MCM MCOV

4 3 6 – Ab b re v i a t i o ns

a b b re v i ations mm3 MVA

Cubic Millimeters Megavolt Amperes

NEC NEMA

National Electrical Code National Electrical Manufacturers Association National Electrical Safety Code NanoFarad (Billionth of a Farad) Nominal National Rural Electric Cooperative Association

SAIDI SF6

NESC nF Nom. NRECA

O.D. Outside Diameter Ohm-m Ohm-meters, a unit of measure for

SiC SIW SLG SR SWBP Sym.

System Average Interruption Duration Index Sulfur Hexafluoride, a synthetic gas used to insulate high-voltage equipment and serve as an interrupting medium in switchgear, one of six types of greenhouse gases to be curbed under the Kyoto Protocol Silicon Carbide, used in valve arresters Single Input Wire Single Line-to-Ground (Fault) State Road Sidewall Bearing Pressure Symmetrical

volume resistivity PE ppm psi psia psig pu PVC

Polyethylene Parts Per Million Pounds Per Square Inch Pounds Per Square Inch Absolute Pounds Per Square Inch Gauge Per Unit Polyvinyl Chloride

REA rms ROW RTU RUS

Rural Electrification Administration Root Mean Square Right of Way Remote Terminal Unit Rural Utility Services, U.S. Department of Agriculture Rural Development— Electric Program (formerly REA)

TNA Transient Network Analyzer TOV Temporary Overvoltages TR-XLPE Tree Retardant Cross-Linked

Polyethylene UD

Underground Distribution

V

Volt

W

Watt

XLPE X/R

Cross-Linked Polyethylene Reactance/Resistance (Ratio)

ZnO ZSURGE

Zinc Oxide Surge Impedance

www.crn.coop NRECA MEMBERS ONLY

Underground Distribution System Design Guide

Recommend Documents