EPRI Aging Power Cable Maintenance Guideline 1024044

2012 TECHNICAL REPORT

Aging Power Cable Maintenance Guideline Medium-Voltage Cables, 5 to 35 kV

Aging Power Cable Maintenance Guideline Medium-Voltage Cables, 5 to 35 kV 1024044 Final Report, November 2012

EPRI Project Manager R. Chambers

ELECTRIC POWER RESEARCH INSTITUTE 3420 Hillview Avenue, Palo Alto, California 94304-1338 • PO Box 10412, Palo Alto, California 94303-0813 • USA 800.313.3774 • 650.855.2121 • [email protected] • www.epri.com

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION NAMED BELOW AS AN ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM: (A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR (B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT. REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY ITS TRADE NAME, TRADEMARK, MANUFACTURER, OR OTHERWISE, DOES NOT NECESSARILY CONSTITUTE OR IMPLY ITS ENDORSEMENT, RECOMMENDATION, OR FAVORING BY EPRI. THE FOLLOWING ORGANIZATION PREPARED THIS REPORT: Electric Power Research Institute (EPRI)

NOTE For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or e-mail [email protected]. Electric Power Research Institute, EPRI, and TOGETHERSHAPING THE FUTURE OF ELECTRICITY are registered service marks of the Electric Power Research Institute, Inc. Copyright © 2012 Electric Power Research Institute, Inc. All rights reserved.

ACKNOWLEDGMENTS

The following organization prepared this report: Electric Power Research Institute (EPRI) 1300 West W. T. Harris Blvd. Charlotte, NC 28262 Principal Investigators A. Mantey G. Toman This report describes research that was prepared for and managed by G. Toman and A. Mantey of EPRI and that was documented in the report, Medium-Voltage Cable Aging Management Guide, Revision 1 (1021070). The report was edited by R. Chambers to remove material not applicable to fossil-fueled plants.

This publication is a corporate document that should be cited in the literature in the following manner: Aging Power Cable Maintenance Guideline: Medium-Voltage Cables, 5 to 35 kV. EPRI, Palo Alto, CA: 2012. 1024044. iii

PRODUCT DESCRIPTION

Medium-voltage cables (5- to 35-kV rated cables) have provided reasonable service in power plants. However, there is a concern that cables that have experienced long periods of wet service might degrade and fail in service. Because most plants have had few problems with medium-voltage cables, little on-staff experience with medium-voltage cables exists at most sites. This report provides information that will be of practical use when questions concerning medium-voltage cable longevity in adverse environments and service conditions arise at fossilfired plants. Objectives This report is meant for persons on plant staffs who are responsible for cable system maintenance, operation, and design. This report describes the currently available information concerning the types of cables in service in plants and the best way to address aging and monitoring concerns. Approach This report is unique in that it focuses on the cable types used in the power industry and the conditions that challenge them. The report supports the needs of plant staffs, allowing them to understand the specifics of their cable systems without being confused by the much broader range of cable information associated with power distribution systems that do not apply to the plants. The report was developed by consultants with extensive experience in cable design, manufacturing, installation, troubleshooting, and replacement. They were guided by industry task group meetings that helped maintain focus on the issues facing plant cable personnel. This revision provides additional information and formalizes the report. Results The report provides information on cable system design, cable construction, insulation systems and their aging characteristics, condition assessment, cable installation, and preparation for repair and replacement. The report provides practical information to help choose the correct assessment tests to apply and to help understand critical issues about the selection and installation of cables and terminations that affect cable longevity. Applications, Value, and Use This report provides a wide range of information that can affect the aging of cables and the management of that aging. Information is provided on grounding system designs that, if not understood, could lead to the misapplication of cable types or increase the likelihood of a potentially devastating phase-to-phase fault that could cause failures of transformers and buses

v

in addition to the cable. Guidance is provided on the selection of an appropriate test method, depending on the cable design, and how the cable is expected to degrade with time. Effects of aging of the insulation and metallic shield system must be considered to allow appropriate selection of the test method. Keywords Aging Cable insulation assessment Cable systems Cross-linked polyethylene (XLPE) Ethylene-propylene rubber (EPR) Medium-voltage cable

vi

CONTENTS

1 INTRODUCTION .................................................................................................................... 1-1 1.1

History and Background of Polymer-Insulated Cables ................................................ 1-2

1.2

In-Plant Cables ............................................................................................................ 1-2

1.3

Underground Cables .................................................................................................... 1-3

1.4

Abbreviations and Acronyms ....................................................................................... 1-4

1.5

Terminology ................................................................................................................. 1-5

2 UNDERSTANDING THE DESIGN OF POWER PLANT CABLE SYSTEMS ........................ 2-1 2.1

Shielded and Nonshielded Cables ............................................................................... 2-1

2.2

Grounding Systems, Protection, and Alarms ............................................................... 2-2

2.2.1

Grounding Systems............................................................................................. 2-2

2.2.2

Phase-to-Phase Faults........................................................................................ 2-4

2.2.3

Number of Grounds on a Cable Insulation Shield ............................................... 2-4

2.3

Multiple Cables per Phase and Balanced Magnetic Fields .......................................... 2-6

2.3.1

Configuration ....................................................................................................... 2-7

2.3.2

Eddy Currents ..................................................................................................... 2-8

2.3.3

Circulating Currents ............................................................................................ 2-8

2.4

Protective Relay and Annunciation Alarm Systems ..................................................... 2-9

3 UNDERSTANDING THE PHYSICAL CONDITION OF THE SYSTEM .................................. 3-1 3.1

Conditions in Manholes and Ducts .............................................................................. 3-1

3.1.1

Insulation Deterioration ....................................................................................... 3-1

3.1.2

Pumping and Dryness ......................................................................................... 3-3

3.2

Correcting Adverse Conditions .................................................................................... 3-4

3.2.1

Adverse Dry Conditions ...................................................................................... 3-4

3.2.2

Physical Stress.................................................................................................... 3-4

3.2.3

Vertical Support................................................................................................... 3-5

vii

3.2.4

Adverse Environments ........................................................................................ 3-5

3.2.4.1

Temperature-Related Aging .......................................................................... 3-5

3.2.4.2

Radiation-Related Aging ............................................................................... 3-5

3.2.4.3

High Conductor Temperature from Ohmic Heating ....................................... 3-6

3.2.4.4

High-Resistance Connections ....................................................................... 3-6

3.2.5

Surface Corona and Partial Discharge................................................................ 3-6

4 CABLE DESIGNS .................................................................................................................. 4-1 4.1

Cable Design Summary ............................................................................................... 4-1

4.2

Medium-Voltage Cable Constructions ......................................................................... 4-1

4.2.1

Voltage Rating..................................................................................................... 4-4

4.2.2

Conductors .......................................................................................................... 4-5

4.2.3

Conductor Shield................................................................................................. 4-6

4.2.4

Insulation ............................................................................................................. 4-6

4.2.5

Insulation Shield .................................................................................................. 4-8

4.2.5.1

Semiconducting or High-Permittivity Shield Layer ........................................ 4-8

4.2.5.2

Metallic Shield Layer ..................................................................................... 4-9

4.2.6

Nonshielded Cables ............................................................................................ 4-9

4.2.7

Jacket ................................................................................................................ 4-10

5 SPLICING AND TERMINATING ............................................................................................ 5-1 5.1

Cable Splicing and Terminating Theory ....................................................................... 5-1

5.2

Gradients ..................................................................................................................... 5-1

5.2.1

Electric Fields ...................................................................................................... 5-1

5.2.2

Stress Cones....................................................................................................... 5-3

5.2.3

Voltage Gradient Design ..................................................................................... 5-4

5.3

Splices for Shielded Cables ......................................................................................... 5-5

5.3.1

Cable Preparation for Splices and Terminations................................................. 5-5

5.3.2

Hand-Taped Splices............................................................................................ 5-8

5.3.3

Premolded Splices ............................................................................................ 5-10

5.3.4

Cold-Shrink Splices ........................................................................................... 5-12

5.3.5

Heat-Shrink Splices........................................................................................... 5-13

5.4

Terminations .............................................................................................................. 5-16

5.5

Lugs and Connectors ................................................................................................. 5-18

viii

5.6

Installation Considerations ......................................................................................... 5-18

5.6.1

Connecting the Conductors............................................................................... 5-18

5.6.2

Insulation for Splices ......................................................................................... 5-19

5.6.3

Semiconducting Insulation Shield Materials for Splices and Terminations ....... 5-19

5.6.4

Metallic Insulation Shield for Splices and Terminations .................................... 5-19

5.6.5

Jackets for Splices ............................................................................................ 5-20

5.7

Selection of Splices and Terminations ....................................................................... 5-20

6 FUNDAMENTALS OF CABLE INSULATION SYSTEMS...................................................... 6-1 6.1

Primary Insulations ...................................................................................................... 6-1

6.2

Elastomer Basics ......................................................................................................... 6-1

6.2.1

Cross Linking ...................................................................................................... 6-2

6.2.2

Fillers Used in Rubber Insulations ...................................................................... 6-2

6.2.3

Crystallinity .......................................................................................................... 6-3

6.2.4

Cable Conductor and Insulation Shields ............................................................. 6-3

6.3

Butyl Rubber ................................................................................................................ 6-4

6.3.1

Material Description ............................................................................................ 6-4

6.3.2

Butyl Wire and Cable Insulation .......................................................................... 6-4

6.3.3

Fillers and Other Additives .................................................................................. 6-5

6.4

Ethylene-Propylene Rubber ......................................................................................... 6-5

6.4.1

Material Description ............................................................................................ 6-5

6.4.2

Ethylene-Propylene Rubber Cross Linking ......................................................... 6-6

6.4.3

Fillers for Ethylene-Propylene Rubber Insulation ................................................ 6-7

6.4.4

Compounding (Mixing) of Ethylene-Propylene Rubbers ..................................... 6-7

6.4.5

Shielded Cable Constructions for Medium-Voltage Ethylene-Propylene Rubber Cables for Plant Applications ................................................................. 6-8

6.5

Historical Review of Medium-Voltage Cable Constructions ......................................... 6-9

6.5.1

General ............................................................................................................... 6-9

6.5.2

Ethylene-Propylene Rubber Types ................................................................... 6-10

6.6

6.5.2.1

Black Ethylene-Propylene Rubber .............................................................. 6-11

6.5.2.2

Pink Ethylene-Propylene Rubber ................................................................ 6-12

6.5.2.3

Brown Ethylene-Propylene Rubber ............................................................. 6-14

Cross-Linked Polyethylene ........................................................................................ 6-14

6.6.1

Material Description .......................................................................................... 6-14

ix

7 AGING AND DEGRADATION OF BUTYL, ETHYLENE-PROPYLENE RUBBER, AND CROSS-LINKED POLYETHYLENE CABLES DUE TO ADVERSE ENVIRONMENTS ........... 7-1 7.1

Aging and Degradation of Butyl Rubber ...................................................................... 7-1

7.1.1

General ............................................................................................................... 7-1

7.1.2

Water-Related Degradation of Butyl Rubber....................................................... 7-1

7.1.3

Thermal Degradation of Butyl Rubber................................................................. 7-2

7.1.4

Radiation Degradation of Butyl Rubber............................................................... 7-2

7.1.5

Conclusions......................................................................................................... 7-3

7.2

Aging and Degradation of Ethylene-Propylene Rubber ............................................... 7-3

7.2.1

General ............................................................................................................... 7-3

7.2.2

Water-Related Degradation of Ethylene-Propylene Rubber ............................... 7-4

7.2.3

Thermal Degradation of Ethylene-Propylene Rubber ......................................... 7-6

7.2.4

Radiation Degradation of Ethylene-Propylene Rubber ....................................... 7-9

7.2.5

Conclusions......................................................................................................... 7-9

7.3

Aging and Degradation of Cross-Linked Polyethylene ................................................ 7-9

7.3.1

General ............................................................................................................... 7-9

7.3.2

Water-Related Degradation of Cross-Linked Polyethylene............................... 7-10

7.3.3

Thermal Degradation of Cross-Linked Polyethylene......................................... 7-11

7.3.4

Radiation Degradation of Cross-Linked Polyethylene....................................... 7-11

7.3.5

Conclusions....................................................................................................... 7-11

7.4

Other Degradation Causes ........................................................................................ 7-11

7.4.1

General ............................................................................................................. 7-11

7.4.2

Corona Discharge ............................................................................................. 7-11

7.4.3

Partial Discharge ............................................................................................... 7-12

8 TESTING: MANUFACTURING, INSTALLATION, AND MAINTENANCE OR INSERVICE TESTS ....................................................................................................................... 8-1 8.1

Introduction .................................................................................................................. 8-1

8.2

Purpose of Tests .......................................................................................................... 8-1

8.3

Manufacturing Tests .................................................................................................... 8-2

8.3.1

Standards and Test Methods .............................................................................. 8-2

8.3.1.1

Tests of Special Interest ................................................................................ 8-5

8.3.1.2

Final Electrical Tests for Shielded Cables ..................................................... 8-7

8.3.1.3 Final Electrical Tests for Nonshielded Cables, 2001–5000 Volts Without Metallic Sheath or Armor ................................................................ 8-10 8.3.1.4

x

Qualification Tests for 5 kV–35 kV Shielded Cables ................................... 8-10

8.4

Acceptance Tests ...................................................................................................... 8-12

8.5

Installation Tests ........................................................................................................ 8-12

8.6

Maintenance and In-Service Testing ......................................................................... 8-13

8.6.1

8.6.1.1

Historical Perspective on In-Service Testing ............................................... 8-14

8.6.1.2

Withstand Versus Diagnostic Testing.......................................................... 8-15

8.6.1.3

Global Versus Local Assessment................................................................ 8-15

8.6.1.4

60 Hz Versus Other Frequencies ................................................................ 8-15

8.6.2

Off-Line Diagnostic and Withstand Testing ....................................................... 8-16

8.6.2.1

Dissipation Factor (Tan δ) Testing .............................................................. 8-16

8.6.2.2

Dielectric Spectroscopy ............................................................................... 8-19

8.6.2.3

Off-Line Partial Discharge Measurement .................................................... 8-21

8.6.2.4

AC Withstand Testing.................................................................................. 8-25

8.6.2.5

Tests Under Development........................................................................... 8-27

8.6.3 8.7

Introduction ....................................................................................................... 8-13

On-Line Diagnostics Assessment ..................................................................... 8-27

Applicability of Tests .................................................................................................. 8-31

9 CABLE AGING MANAGEMENT PROCESS ......................................................................... 9-1 9.1

Strategies and Philosophies ........................................................................................ 9-1

9.1.1

9.1.1.1

Run to Failure ................................................................................................ 9-2

9.1.1.2

Diagnostic Testing ......................................................................................... 9-2

9.1.1.3

Withstand Testing.......................................................................................... 9-3

9.1.2 9.2

Test Strategies for Shielded Cable Circuits ........................................................ 9-1

Nonshielded Cable Circuits ................................................................................. 9-4

Prioritization of Cables for Assessment and Testing ................................................... 9-5

9.2.1

Risk Ranking Methodology ................................................................................. 9-6

9.2.1.1

Maintenance Rule and Criticality Screening.................................................. 9-6

9.2.1.2

Insulation Type .............................................................................................. 9-7

9.2.1.3

Significance Factor for Jacket Types or Water-Impervious Designs ............. 9-7

9.2.1.4

Significance Factor for Operating Experience ............................................... 9-7

9.2.1.5

Significance Factor for Diagnostic Test Results ............................................ 9-7

9.2.1.6

Significance Factor for Voltage and Insulating Level .................................... 9-7

9.2.1.7

Significance Factor for Operating Conditions ................................................ 9-7

9.2.1.8

Weighting Factor for Adverse Environment................................................... 9-8

9.2.1.9

Weighting Factor for Current and Amperage Level ....................................... 9-8

xi

10 RESPONSE TO CABLE FAILURES .................................................................................. 10-1 10.1

Electrical Tests of Circuit....................................................................................... 10-1

10.1.1

Logic for Testing to Be Performed ................................................................. 10-1

10.1.2

Logistical Issues for Electrical Testing ........................................................... 10-2

10.2

Fault Location........................................................................................................ 10-2

10.3

Forensics and Failure Assessments ..................................................................... 10-3

10.4

Repair Options ...................................................................................................... 10-6

10.4.1

Replacement of Failed Section ...................................................................... 10-6

10.4.2

Total Replacement ......................................................................................... 10-7

10.4.3

Acceptance Testing ....................................................................................... 10-7

11 REFERENCES ................................................................................................................... 11-1 A MEDIUM-VOLTAGE CABLE FAILURES AND FIELD EXPERIENCE................................. A-1 A.1

Introduction ................................................................................................................. A-1

A.2

Dry Condition Failures ................................................................................................ A-1

A.2.1

Failure of a Cable at the Point of a Shield Crimp ............................................... A-1

A.2.2

Thermal Deterioration of a Butyl Rubber Insulated Cable.................................. A-3

A.2.3

Failure of a Brown Ethylene-Propylene Rubber Insulated Cable Due to Failure of a Zinc Shield Tape ............................................................................. A-5

A.3

Water-Related Failures ............................................................................................... A-8

A.3.1

Failure of a 38-Year-Old Butyl Rubber Cable Due to Water-Induced Degradation........................................................................................................ A-8

A.3.2

Failure of a Wet Okonite Black Ethylene-Propylene Rubber Cable ................. A-16

A.3.3

Failure of a Cross-Linked Polyethylene Insulated Cable from Localized Water Treeing .................................................................................................. A-17

A.4

Failures of Terminations and Splices ........................................................................ A-22

A.4.1 A.5

xii

Failure from Use of an Oversized Termination Sleeve .................................... A-22

Summary of Lessons Learned .................................................................................. A-24

A.5.1

Dry Cable Compression-Related Event ........................................................... A-24

A.5.2

Overheated Butyl Rubber Failure..................................................................... A-24

A.5.3

Failure of Continuity of a Zinc Tape Shield ...................................................... A-24

A.5.4

Water-Related Failure of a Butyl Rubber Cable ............................................... A-25

A.5.5

Failure of an Ethylene-Propylene Rubber Insulated Cable from Long-Term Wetting ............................................................................................................. A-26

A.5.6

Failure of Cross-Linked Polyethylene Insulation from Water Treeing .............. A-26

A.5.7

Failure from Use of an Oversized Molded Termination.................................... A-26

B RESULTS OF NUCLEAR ENERGY INSTITUTE SURVEY AND NUCLEAR REGULATORY COMMISSION GENERIC LETTER 2007-01 .................................................. B-1 B.1

Nuclear Energy Institute Underground Medium-Voltage Cable Survey ...................... B-1

B.1.1

Survey Purpose.................................................................................................. B-1

B.1.2

Survey Scope ..................................................................................................... B-1

B.2

Survey Results Evaluation .......................................................................................... B-2

B.2.1

Contributors........................................................................................................ B-2

B.2.2

Underground Circuit Quantities .......................................................................... B-2

B.2.3

Installed Cable Types......................................................................................... B-2

B.2.4

Shielding ............................................................................................................ B-5

B.2.4.1 B.3

Underground Wet-Duty Failure Assessment ............................................... B-5

Results of Utility Responses to Generic Letter 2007-01 ........................................... B-13

B.3.1

Summary of Results ......................................................................................... B-14

B.3.2

Assessment Methods ....................................................................................... B-16

B.4

Comparison of the Nuclear Energy Institute Survey and Generic Letter 2007-01 Results ....................................................................................................... B-17

C TAN δ DATA FOR ETHYLENE-PROPYLENE RUBBER AND BUTYL RUBBER INSULATED CABLES .............................................................................................................. C-1 C.1

60-Hz Tan δ Data from 1972 Black Okoguard Insulation ........................................... C-1

C.2

Very-Low-Frequency Tan δ Results for Okonite Black Ethylene-Propylene Rubber Insulation ....................................................................................................... C-4

C.3

Tan δ for Anaconda Pink Ethylene-Propylene Rubber UniShield ............................... C-6

C.4

Anaconda Black UniShield Ethylene-Propylene Rubber .......................................... C-11

C.5

Tan δ Results for Butyl Rubber Insulation ................................................................ C-12

C.6

Conclusions Related to Rubber Insulated Cables and Tan δ Results ...................... C-16

C.7

Effects of Mixed Shielded and Nonshielded Segments on Tan δ ............................. C-17

D ADDITIONAL POLYMER MATERIALS INFORMATION ..................................................... D-1 D.1

Fundamentals of Elastomers ...................................................................................... D-1

D.2

Cross Linking .............................................................................................................. D-2

D.3

Butyl Rubber Compositions ........................................................................................ D-3

D.4

Ethylene-Propylene Rubber and Ethylene-Propylene-Diene Monomer...................... D-4

D.5

Dicumyl Peroxide Cross-Linking Agent Byproducts.................................................... D-4

D.6

Ethylene-Propylene Rubber Formulations .................................................................. D-5

D.7

Influence of Clay on Properties of Ethylene-Propylene Rubber.................................. D-7

xiii

E OFF-LINE TESTS THAT ARE UNDER DEVELOPMENT .................................................... E-1 E.1

Introduction ................................................................................................................. E-1

E.2

Isothermal Return Current .......................................................................................... E-1

E.3

Return Voltage ............................................................................................................ E-2

E.4

Oscillating Wave ......................................................................................................... E-3

F INSULATION RESISTANCE TEST MEASUREMENTS: THEIR VALUE AND LIMITATIONS ............................................................................................................................F-1 F.1

Introduction ..................................................................................................................F-1

F.2

Determining Minimum Insulation Resistance Value and Improving Interpretation of Results ....................................................................................................................F-1

F.3

Insulation Resistance of Good Insulation.....................................................................F-3

F.4

Conclusions .................................................................................................................F-4

xiv

LIST OF FIGURES Figure 2-1 Shielded Cable Components .................................................................................... 2-2 Figure 2-2 Single-Point Grounding of a Shield, Showing Voltage of Shield to Ground as a Function of Distance ....................................................................................................... 2-5 Figure 2-3 Shield Losses and Voltages for Single-Conductor Cables ....................................... 2-5 Figure 2-4 Acceptable Configurations: Two Cables per Phase ................................................. 2-7 Figure 2-5 Acceptable Configurations: Three Cables per Phase ............................................... 2-7 Figure 2-6 Acceptable Configurations: Four Cables per Phase ................................................. 2-8 Figure 3-1 Electrostatic Flux Lines and Charged Particles ........................................................ 3-3 Figure 3-2 Example of Manhole Drainage System .................................................................... 3-4 Figure 3-3 White Powder Indicates a Location of Corona Discharge Between a Cable and a Ground Cable in Close Proximity ............................................................................. 3-7 Figure 4-1 Shielded, Single-Conductor, Medium-Voltage Cable Design ................................... 4-2 Figure 4-2 Medium-Voltage Shielded Cable, Water-Impervious Design.................................... 4-3 Figure 4-3 Medium-Voltage Nonshielded Cable Design ............................................................ 4-3 Figure 4-4 UniShield Construction, Shield Wires Embedded in the Semiconducting Outer Jacket ....................................................................................................................... 4-4 Figure 4-5 Conductor Stranding Configuration, Showing Compressed and Compacted Conductor Configurations .................................................................................................. 4-5 Figure 5-1 Equipotential and Flux Lines in a Cable ................................................................... 5-2 Figure 5-2 Electrical Stress Fields, Shield Removed ................................................................. 5-2 Figure 5-3 Termination of an Insulation Shield with a Stress Cone ........................................... 5-3 Figure 5-4 Stress Relief with High Dielectric Constant or High Resistivity Materials ................. 5-4 Figure 5-5 Projection from Semiconducting Layer and Cut into Insulation ................................ 5-5 Figure 5-6 Two Types of Semiconducting Layer Scoring or Stripping Tools ............................. 5-6 Figure 5-7 Measure Carefully to Achieve the Proper Length of Cable ...................................... 5-7 Figure 5-8 Slip Applicable Sleeves over One or Both Cable Ends Before Installing Connector ........................................................................................................................... 5-7 Figure 5-9 Connectors of Several Lengths and Diameters for the Same Conductor Size ......... 5-7 Figure 5-10 Shear-Bolt Connectors for Copper Conductor ........................................................ 5-7 Figure 5-11 Clean the Insulation, Using an Approved Solvent .................................................. 5-8 Figure 5-12 Semiconducting Tape Is Applied over the Connector to Form a Smooth Interface ............................................................................................................................. 5-8 Figure 5-13 Insulating Tape Is Applied Until the Proper Thickness Is Achieved over the Connector ..................................................................................................................... 5-9

xv

Figure 5-14 Another Layer of Semiconducting Tape Is Applied over the Insulating Tape ......... 5-9 Figure 5-15 A Metallic Braid Is Installed, a Ground Strap Is Attached, and the Jacket Tape Is Installed over the Entire Splice Area ................................................................... 5-10 Figure 5-16 Cutaway of a Premolded Splice ........................................................................... 5-11 Figure 5-17 Cold-Shrink Splice, Showing the Direction of the Shrinking Process ................... 5-12 Figure 5-18 Carefully Position the Housing Before Removing the White Core Support .......... 5-12 Figure 5-19 Continue Removing the Core While Holding Its Position on the Cable ................ 5-13 Figure 5-20 High-Permittivity Mastic Material Is Placed over the Connector and Conductor, Without Concern for Smoothness .................................................................. 5-14 Figure 5-21 The Semiconducting Tube Is Slid into Place and Heated Until Properly Shrunk Down to the Cable ............................................................................................... 5-14 Figure 5-22 An Insulation Tube Is Slid into Place and Shrunk Down ...................................... 5-14 Figure 5-23 A Tube That Is Both Insulating on the Inside and Semiconducting on the Outside Is Positioned and Shrunk into Place ................................................................... 5-14 Figure 5-24 Tinned Copper Braid is Wrapped Around the Splice to Replace the Metallic Portion of the Insulation Shield ........................................................................................ 5-14 Figure 5-25 A Ground Strap Spring Is Placed Under One Side of the Cable’s Taped Metallic Insulation Shields ................................................................................................ 5-15 Figure 5-26 The Ground Strap Is Placed Across the Splice and Connected to the Factory Metallic Shield on the Opposite Side of the Splice ........................................................... 5-15 Figure 5-27 An Overall Rejacketing Tube Is Placed Around the Entire Area .......................... 5-15 Figure 5-28 Cold-Shrink Outdoor Termination with Rain Sheds or Skirts ................................ 5-16 Figure 5-29 Dead Break, Premolded Termination ................................................................... 5-17 Figure 5-30 Stub-Type Motor Connection ................................................................................ 5-17 Figure 5-31 In-Line Type Connection ...................................................................................... 5-17 Figure 5-32 Improperly Crimped Connector ............................................................................ 5-19 Figure 6-1 Butyl Rubber Molecule ............................................................................................. 6-4 Figure 6-2 Copolymer of Ethylene and Propylene ..................................................................... 6-6 Figure 6-3 Banbury Mixer Used for Preparing Ethylene-Propylene Rubber Compounds .......... 6-8 Figure 6-4 Various Representations of Crystallinity in Polyethylene ....................................... 6-14 Figure 7-1 Ethylene-Propylene Rubber Cable Failures as Function of Color of the Ethylene-Propylene Rubber Insulations ............................................................................. 7-5 Figure 7-2 Generational Differences in Life Expectancy for Ethylene-Propylene Rubber and Other Insulations ......................................................................................................... 7-6 Figure 7-3 AC Breakdown Strength of Medium-Voltage, Field-Aged Ethylene-Propylene Rubber Insulated Cables at V0 ........................................................................................... 7-8 Figure 7-4 AC Voltage Breakdown Strength of Combined Ethylene-Propylene Rubber Cables Aged in the Laboratory and in Service at Rated Voltage ....................................... 7-8 Figure 7-5 Unfilled and Mineral-Filled Cross-Linked Polyethylene Cable Failures Compared to All Failures .................................................................................................. 7-10 Figure 8-1 Wafer Examination for Voids, Inclusions, and Conductor Shield Protrusions .......... 8-6 Figure 8-2 Hot Oil Test............................................................................................................... 8-6

xvi

Figure 8-3 Vertical Tray Flame Test .......................................................................................... 8-7 Figure 8-4 An Acceptable Partial Discharge Plot ....................................................................... 8-8 Figure 8-5 An Unacceptable Partial Discharge Plot ................................................................... 8-8 Figure 8-6 Partial Discharge Requirement History from AEIC 5 ................................................ 8-9 Figure 8-7 Dry Specimen Design Qualification Tests .............................................................. 8-10 Figure 8-8 Wet Specimen Design Qualification Tests ............................................................. 8-11 Figure 8-9 Derivation of Dissipation Factor (Tan δ) Measurement in Insulation ...................... 8-16 Figure 8-10 Typical Variable-Frequency, Very-Low-Frequency Portable Test Equipment for Performing 0.1-Hz Dissipation Factor (Tan δ) Testing ................................................ 8-17 Figure 8-11 Voltage Dependence of Dissipation Factor for New and Aged Cross-Linked Polyethylene Cable .......................................................................................................... 8-18 Figure 8-12 0.1-Hz Dissipation Factor of Cross-Linked Polyethylene-Insulated Cables.......... 8-18 Figure 8-13 Dielectric Spectroscopy Measurements for Shelf-Aged 15-kV Cross-Linked Polyethylene Cable .......................................................................................................... 8-20 Figure 8-14 Calibration Equipment for Off-Line Partial Discharge Testing .............................. 8-21 Figure 8-15 Test Setup for Performing 60-Hz Partial Discharge Measurements ..................... 8-22 Figure 8-16 Test Profile for Short Duration Off-Line 60-Hz Partial Discharge Testing............. 8-23 Figure 8-17 60-Hz Partial Discharge One-Step Diagnostic Data Capture Approach ............... 8-23 Figure 8-18 Nominal Very-Low-Frequency (0.1-Hz) Sinusoidal Waveform ............................. 8-25 Figure 8-19 Trapezoidal (Bipolar Rectangular) Very-Low-Frequency (0.1-Hz) Waveform ...... 8-26 Figure 8-20 Sensors for On-Line Signal Detection and Data Acquisition System ................... 8-29 Figure 8-21 On-Line Condition Assessment Data Acquisition System .................................... 8-29 Figure 8-22 Influence of Defect Type on Signal Patterns Detected During On-Line Evaluation ........................................................................................................................ 8-30 Figure A-1 External Cable Condition at the Location of the Fault ............................................. A-2 Figure A-2 Condition of the Shield After the Jacket and Burn Hole Were Removed from the Fault ............................................................................................................................ A-2 Figure A-3 Damage to the Conductor from the Fault ................................................................ A-3 Figure A-4 Inner Surface of the Insulation with Severe Indentation from the Semiconducting Tape ....................................................................................................... A-4 Figure A-5 Zinc Shield .............................................................................................................. A-6 Figure A-6 Tracking Path and Failure Location on Interior of Jacket ........................................ A-6 Figure A-7 Tracking Path and Failure Site on Insulation .......................................................... A-7 Figure A-8 Damage to Conductor from Fault ............................................................................ A-7 Figure A-9 Transformer Connection of One of the Cables ....................................................... A-8 Figure A-10 Circuit Breaker Connection of One of the Cables ................................................. A-9 Figure A-11 Cable Configuration in the Manhole Adjacent to the Transformer ........................ A-9 Figure A-12 Location of Failed, Faulted, and At-Risk Cables Within Duct Leading to Building ....................................................................................................................... A-10 Figure A-13 Duct Sealing and Water Leaking from Failed Conductor B-5 ............................. A-11

xvii

Figure A-14 Cables A-5, B-5, and C-5 .................................................................................... A-12 Figure A-15 Close-Up of B-5 Cable Failure ............................................................................ A-12 Figure A-16 Comparison of Conductor Corrosion ................................................................... A-13 Figure A-17 Marked Low-Resistance Channel in the Insulation with Breakdown Hole in Center .......................................................................................................................... A-13 Figure A-18 Micrograph of Insulation Wall at Low-Resistance Channel, Showing Swelling and Fissures ..................................................................................................... A-14 Figure A-19 Temporary, Aboveground Cable System ............................................................ A-15 Figure A-20 Permanent Aerial Structure Routing on Second Unit .......................................... A-15 Figure A-21 Damage to the Copper Shield at the Damage Site ............................................. A-18 Figure A-22 Cross Section of Insulation at the Fault Tube, Showing Carbonization of the Wall of the Fault Tube ..................................................................................................... A-19 Figure A-23 Large Water Tree Viewed Through Hot Oil Bath ................................................ A-20 Figure A-24 Cross Section of the Water Tree Shown in Figure A-23 ..................................... A-20 Figure A-25 Embedded Particle at Base of Water Tree Shown in Figures A-23 and A-24 ..... A-21 Figure A-26 As-Found Condition of Terminations ................................................................... A-22 Figure A-27 Burnthrough of Stress Relief Adaptor ................................................................. A-23 Figure A-28 Burned-Through Insulation Found After Removal of the Stress Relief Adaptor ............................................................................................................................ A-23 Figure B-1 Manufacturers of 5-kV Cable .................................................................................. B-3 Figure B-2 Age Distribution of Units with No Failures ............................................................... B-6 Figure B-3 Age Distribution of Wet Cable Failures for All Insulation Types .............................. B-7 Figure B-4 Age Distribution of Wet Cross-Linked Polyethylene Cable Failures ....................... B-8 Figure B-5 Age Distribution of Wet Ethylene-Propylene Rubber Cable Failures ...................... B-9 Figure B-6 Age Distribution of Butyl Rubber Cable Failures ................................................... B-10 Figure B-7 Age of Wet Cable at Failure Versus Year of Failure ............................................. B-11 Figure B-8 Age at Time of Failure for Wet Ethylene-Propylene Rubber Cables ..................... B-12 Figure B-9 Age at Time of Failure for Wet Cross-Linked Polyethylene Cables ...................... B-13 Figure B-10 Inaccessible Cable Failures by Plant, from Responses to Generic Letter 2007-01 ........................................................................................................................... B-14 Figure B-11 Number of Failures of Wet Cable by Year .......................................................... B-15 Figure B-12 Wet Medium-Voltage Cable Failures Versus Age at Time of Failure .................. B-15 Figure B-13 Age of Cable at Time of Failure By Insulation and Design .................................. B-16 Figure C-1 AC Breakdown Strength Versus 60-Hz Tan δ Results............................................ C-2 Figure C-2 60-Hz Tan δ Results by Test Voltage ..................................................................... C-4 Figure C-3 Very-Low-Frequency Tan δ Result for a 5-kV Okonite Black EthylenePropylene Rubber Cable ................................................................................................... C-5 Figure C-4 Very-Low-Frequency Tan δ Results for 15-kV Okonite Black EthylenePropylene Rubber Cable ................................................................................................... C-6 Figure C-5 60-Hz Tan δ for Anaconda UniShield Pink Ethylene-Propylene Rubber ................ C-7

xviii

Figure C-6 Laboratory-Measured Very-Low-Frequency Tan δ for Anaconda Pink Ethylene-Propylene Rubber UniShield .............................................................................. C-7 Figure C-7 Tan δ Results for Anaconda Pink UniShield Cables (Ordered by 8-kV result) ....... C-8 Figure C-8 Tan δ Results for Anaconda Pink UniShield Cables (Ordered by Difference in 16 kV and 8kV Results) ..................................................................................................... C-9 Figure C-9 Tan δ Results for Anaconda Pink UniShield Cables (Showing Moderate to High Results) ................................................................................................................... C-10 Figure C-10 Tan δ Measurements for Anaconda UniShield Cables Taken One Year Apart ................................................................................................................................ C-11 Figure C-11 Tan δ Measurements for Anaconda UniShield with Black Insulation .................. C-12 Figure C-12 Tan δ Versus Breakdown Strength for Okonex Butyl Rubber ............................. C-13 Figure C-13 Comparison of 60-Hz Tan δ Versus Breakdown Strength of Okonite Butyl Rubber to Black Ethylene-Propylene Rubber ................................................................. C-14 Figure C-14 AC Breakdown Voltage (in V/mil) for Okonite Butyl and Black EthylenePropylene Rubber Insulations ......................................................................................... C-15 Figure C-15 Tan δ Results for Mixed Shielded and Nonshielded Black Okonite EthylenePropylene Rubber Circuits .............................................................................................. C-17 Figure C-16 Tan δ Results for Mixed Shielded and Nonshielded Black Okonite EthylenePropylene Rubber Circuits .............................................................................................. C-18 Figure C-17 Average Tan δ Value Versus Percent of Nonshielded Cable Section ................ C-19 Figure D-1 A Cross-Linked Elastomer ...................................................................................... D-2 Figure D-2 Ethylene-Propylene-Diene Monomer Terpolymer ................................................... D-4 Figure E-1 Isothermal Return Current Plots for Cross-Linked Polyethylene Cable Having Undergone Different Degrees of Aging ............................................................................. E-1

xix

LIST OF TABLES Table 6-1 Typical Components in Medium-Voltage Ethylene-Propylene Rubber Compounds ...................................................................................................................... 6-13 Table 6-2 Percentage of Nuclear Plants with Cross-Linked Polyethylene Insulated Cables by Cable Voltage Rating ...................................................................................... 6-16 Table 8-1 History of Specifications and Testing of Shielded and Nonshielded MediumVoltage Cables ................................................................................................................... 8-3 Table 8-2 Common Production Tests ........................................................................................ 8-4 Table 8-3 Dielectric Constant and Dissipation Factor Acceptance Criteria................................ 8-5 Table 8-4 IEEE Standard 400 Criteria for Assessment for Cross-Linked Polyethylene Insulated Cables .............................................................................................................. 8-17 Table 8-5 Matrix of Applicability of Tests During the Life of a Medium-Voltage Cable ............ 8-31 Table B-1 Originally Installed 5-kV Insulation Types ................................................................ B-3 Table B-2 Distribution of 8-kV Cable Insulation Materials ......................................................... B-4 Table B-3 Distribution of 15-kV Cable Insulation Materials ....................................................... B-4 Table B-4 Distribution of 25-kV to 35-kV Cable Insulation Materials ........................................ B-4 Table B-5 Number of Failures per Plant Reporting Failures ..................................................... B-7 Table B-6 Summary of Assessment and Testing Responses from Generic Letter 2007-01 ........................................................................................................................... B-17 Table D-1 Butyl Rubber Insulation Components ....................................................................... D-3 Table D-2 Butyl Rubber Wire Insulation Components .............................................................. D-3 Table D-3 Typical Black Wire and Cable Insulation Ethylene-Propylene Rubber Compound Components from the 1970s .......................................................................... D-5 Table D-4 Properties of Black Ethylene-Propylene Rubber Compound Shown in Table D-3 .......................................................................................................................... D-5 Table D-5 Medium-Voltage Black Ethylene-Propylene Rubber Insulation Using EthylenePropylene-Diene Monomer ............................................................................................... D-6 Table D-6 Compounds of EPR Using Calcined and Coated Clay ............................................ D-7 Table D-7 Influence of Clay Nature on Ethylene-Propylene Rubber Cable Properties: Calcined Versus Coated Clay ........................................................................................... D-7

xxi

1

INTRODUCTION This report is based on the research performed for the Nuclear Sector and its membership. GenMAC acknowledges their work as comprising the content of this product and its applicability to the Generation Sector. It is our intent to partner where ever possible on this important research to avoid duplication of work or seek to overload limited resources providing this critical information. The medium-voltage cable system (cable rated from 5 kV to 35 kV, for the purpose of this report) is composed of cable, terminations, splices, trays, conduits, ducts, and, in some cases, trenches. Vertical support systems, manholes, fire stops, and water drainage systems also affect the cable. All of these play a role in the longevity of the cable system. This report provides a wide range of information in support of medium-voltage cable aging management. Information is provided on topics such as electrical system grounding practices and operational issues because they affect failure mechanisms, especially in the final stages of failure. For example, some safety-related distribution systems are ungrounded or have highresistance grounds so that, during accident conditions, they could sustain a phase-to-ground fault and continue to operate for some period of time. However, under normal operating conditions, a cable with a suspected or identified ground fault should be de-energized as soon as possible to prevent the fault from converting to a phase-to-phase fault, which is much more damaging. A phase-to-ground fault in an ungrounded or high-resistance ground system is likely to generate heat or increase voltage stress across the adjacent phase insulation so that deterioration of the insulation of a second phase in a relatively short period of time is possible. After the fault converts to a phase-to-phase fault, extremely high currents result, causing severe damage to the cable and connected equipment. Understanding the need for de-energizing cables having a phase-to-ground fault, even though the system is designed to withstand the condition for a short period, is important to limiting damage and stress to the remainder of the system. Topics included in this report related to cable aging management include the following: 

The grounding system design, which determines the way in which ground currents and transient voltages affect cables under fault conditions and affects operating practices at the time of faults



Types of cables and cable designs, including nonshielded and shielded, and the effects of different cable designs on aging and testability



The physical layout of the cable and the various service conditions and stresses imparted to the cable beyond operating voltage and current



Cable design, aging mechanisms, and failure mechanisms



Splice and termination issues that can lead to early failure

1-1

Introduction



Cable insulation materials and changes and improvements that have been made to them since the 1970s



Cable condition assessment tests

This report is an update of the Electric Power Research Institute (EPRI) report Medium Voltage Cable Aging Management Guide (1016689), which was published in 2008 [1]. This update reflects the consensus guidance on aging management of medium-voltage power cables developed in 2009–2010, as well as other developments since the initial report was published. Due to industry attention to submergence of cable, much of the report concerns aging of medium-voltage cables under wet energized conditions. However, cables and terminations can fail under dry conditions if defects exist in the insulation, if damage or installation errors occur, or if the cable is affected by an adverse local environment. Appendix A provides insights into some of the conditions that have resulted in failures under both wet and dry conditions.

1.1

History and Background of Polymer-Insulated Cables

In the early 1960s, high-molecular-weight polyethylene, cross-linked polyethylene (XLPE), and various synthetic rubber insulations (black ethylene-propylene rubber [EPR] and black butyl systems [2, 3]) began to be deployed in medium-voltage distribution networks. Blodgett and Fisher described the development of one such system and provided a brief review of the status of the full range of alternative polymers [2]. With the exception of high-molecular-weight polyethylene, these materials were used in generating stations in the 1960’s. A key feature of these materials was their expected enhanced moisture stability compared with legacy insulations. As a consequence, most distribution utilities adopted high-molecular-weight polyethylene or XLPE due to their low dielectric losses and costs, and then routed such circuits with little concern for moisture. Although these cables have provided good service under dry conditions, none of these materials lived up to their promised moisture stability, and the early designs were superseded by more robust systems. Some utilities followed the distribution industry’s lead and installed XLPE insulated cables. However, most chose to use rubber insulated cables due to their greater flexibility and tighter bending radius, which is important in the tighter confines of a power block. Rubber cables were also expected to be long lived, and they have generally proven to be longer lived than XLPE cables. Over the years, numerous improvements to the design, materials, compounding of insulations, and manufacture of cables have occurred. For example, black EPR was replaced by brown, red, or pink EPR; high-molecular-weight polyethylene was supplanted by XLPE; and XLPE was replaced by tree-retardant XLPE (TR-XLPE). A history of these changes is provided in Section 6, Fundamentals of Cable Insulation Systems.

1.2

In-Plant Cables

In-plant cables are generally located in dry conduits or trays. Dry cables tend to have a long life. When failures occur, they are generally related to an installation error, a manufacturing flaw, or physical or thermal damage. Even under such conditions, failure can take years to decades to occur. Thermal damage can occur from an external source such as an adjacent, uninsulated hot pipe or from ohmic heating on highly loaded cables, especially if there is a significant magnetic or resistive imbalance in the phase cables. 1-2

Introduction

1.3

Underground Cables

Underground routing of power cables has long been desirable for a variety of reasons: the physical protection it provides; the low ambient temperature presented to the cable and the resultant high ampacities; the large natural heat sink provided by the soil and its favorable impact on transient loading capabilities; and the invisibility of the network once installed. Unfortunately, in most parts of the country, installation of cables underground exposes them to some degree of wetting or submergence. The extent and duration of that submergence are influenced by many variables. In systems provided with adequate and well-maintained drainage, short-term submergence consistent with post-storm runoff does little more than wet the surface of the cable, given the slow diffusion of moisture through common jacketing systems. In contrast, directburial installations or poorly sloped and drained raceways will ultimately expose the cables to extended submergence, which provides time for moisture to diffuse through most common polymeric jackets. The earliest applications in the underground network of solid dielectric cables using natural rubber insulation were plagued with moisture-induced problems. The resultant degradation assumed two forms: electrical instability and physical instability. Electrical instability in water was characterized by ever-increasing leakage current versus time, with the risk of localized thermal runaway and failure of the circuit. Physical instability of the insulation system was characterized by leaching of compound constituents or significant swelling when exposed to water. Both modes of physical instability led to the development of large-scale voids and eventual failure. The moisture-related failures had the greatest impact on distribution utilities, because their networks consist of numerous, long underground medium-voltage feeders serving a number of shorter low-voltage circuits. In contrast, underground cable systems typically consist of only a few long medium-voltage circuits. As a consequence, the bulk of the research to date has been directed toward understanding and resolving the moisture-driven issues that were of particular interest to the distribution market. The conclusions drawn from distribution-related research are not always directly transferable to the power industry. Assessing the applicability of distribution research for power applications remains a challenge for both the user and regulatory communities. Some of these issues are addressed here. This report provides a significant amount of information on rubber insulations that are used in conventional and nuclear plants but, until recently, were not often used in distribution systems. The report also provides some insights on design differences between plant cables and those used in distribution systems. Water-related degradation became a concern to the nuclear industry in 2006 and to the United States Nuclear Regulatory Commission (NRC), as indicated by Generic Letter 2007-01, Inaccessible or Underground Power Cable Failures that Disable Accident Mitigation Systems or Cause Plant Transients [4]. Actions taken include implementation of routine inspections of manholes for submergence of cable. The EPRI report Aging Management Program Guidance for Medium-Voltage Cable Systems for Nuclear Power Plants (1020805) [5] recommends limiting cables exposure to long-term wetting by manually or automatically pumping dry manholes, pits, and vaults to preclude long-term wetted or submerged conditions. In addition, it is recommended that cables that are shielded and are known or suspected of being wetted or submerged for longer than a few days at a time be tested to determine whether degradation has occurred. 1-3

Introduction

1.4

Abbreviations and Acronyms

ac

alternating current

ACLT

accelerated cable life test

AEIC

Association of Edison Illuminating Companies

AWG

American Wire Gauge

CPE

chlorinated polyethylene rubber

CSPE

chlorosulfonated polyethylene rubber (Hypalon)

dc

direct current

IEEE

Institute of Electronic and Electrical Engineers

EPRI

Electric Power Research Institute

EPR

ethylene-propylene rubber

Hz

hertz

ICEA

Insulated Cable Engineers Association

IRC

isothermal return current

kcmil

thousands of circular mils, a unit of area for conductor size

kV

kilovolts

LDPE

low-density polyethylene

mil

1/1000th of an inch

NEI

Nuclear Energy Institute

NEMA

National Electrical Manufacturing Association

NRC

Nuclear Regulatory Commission

OW

oscillating wave

pC

picocoulomb

PD

partial discharge

PDEV

partial discharge extinction voltage

PDIV

partial discharge inception voltage

PE

polyethylene

PVC

polyvinyl chloride

RV

return voltage

tan 

tangent  (a loss factor of insulation)

1-4

Introduction

TDR

time domain reflectometer

TR-XLPE

tree-retardant cross-linked polyethylene

V0

phase-to-ground voltage (also referred to as U0 in the literature)

XLPE

cross-linked polyethylene

1.5

Terminology

Hydrophobic. A tendency of a polymer to reject and not to absorb or react with water. Lossy. Referring to insulation having a somewhat higher leakage current. XLPE is a low-loss material, whereas EPR is lossy.

1-5

2

UNDERSTANDING THE DESIGN OF POWER PLANT CABLE SYSTEMS Understanding the design of the cables and the cable system is important because it can influence the rate of degradation and the ability to detect that degradation. This section describes cable shielding, circuit grounding philosophy, cable ground insulation, and cable phase configuration choices that can influence or affect cable aging and testability. Protective relays and annunciation alarms, as well as actions that be taken in response to their actuation, are described.

2.1

Shielded and Nonshielded Cables

Medium-voltage, 5-kV cables can be constructed with or without an insulation shielding system. Factors that influence this decision include the following: 

Depending on the thickness of the insulation in the cable, nonshielded (grounded) cables can continue to function for a limited period with a single phase-to-ground fault should one occur under plant accident conditions. (Single phase-to-ground faults should be de-energized as soon as possible under normal conditions.)



Nonshielded cables are simpler to splice and terminate because there are no insulation shields.



Nonshielded cables have voltage on their exterior surface because the voltage distributes across any insulator, such as air, between the conductor and ground or the conductor and adjacent phases. The voltage at the surface of the cable can be as high as 90% of the conductor voltage.



When nonshielded cables touch or nearly touch grounded cabinets and conduits, high-voltage stress occurs in the air gap, which can lead to corona attach. Due to the high-voltage stress, the air in the gap can break down, causing streams of electrons to impact the cable surface. During each breakdown, the air gap is temporarily shorted, and the voltage is distributed across the insulation. At each discharge, a small increment of damage occurs to the polymer wall of the cable. Over long periods, the wall of the insulation can erode, leading to insulation failure.



Nonshielded cables are more difficult to protect. Relay settings must be less sensitive than for shielded cables. Manual intervention might be required for nonshielded cables to prevent a single-phase fault from developing into a phase-to-phase fault.

2-1

Understanding the Design of Power Plant Cable Systems



With a shielded cable, a multi-grounded metallic shield eliminates any voltage or tracking on the cable’s exterior surface. An additional safety advantage is developed because there is no surface voltage present, and workers can come into close proximity to energized cables, such as in manholes; however, contacting medium-voltage cables, whether shielded or nonshielded, is not recommended while they are energized.



Shielded cables are more complex to splice and terminate.

Two shields exist in a majority of the medium-voltage cables used in power plants today (see Figure 2-1). The shielding material next to the conductor is called the conductor shield. Mediumvoltage cables used in more recent plants have conductor shields. The conductor shield prevents discharges between the conductor and the insulation during operation. The shield that is of concern with regard to shielded and nonshielded cable is the insulation shield, which consists of a semiconducting layer in contact with the outer portion of the insulation, as well as a metallic portion on top of the semiconducting layer. The metallic portion must be grounded in at least one point in every circuit for the shield to contain the voltage stress within the insulation system. If the shield is not grounded, the cable will function just like a nonshielded cable and will have up to 90% of conductor voltage on the cable’s outer surface.

Figure 2-1 Shielded Cable Components

2.2

Grounding Systems, Protection, and Alarms

2.2.1 Grounding Systems Significant differences exist in cable design and applications between medium-voltage cables used in utility residential and industrial customer distribution systems and those used in power plants. These differences are related to the electrical system design practices for power plants, the need for more flexible cables to allow installation within the restricted confines of the power plant, and service continuity requirements. Understanding the designs of cable systems, especially with respect to grounding and fault clearing, is critical to responding properly when a ground is identified but the cable remains in service.

2-2

Understanding the Design of Power Plant Cable Systems

Medium-voltage utility distribution systems in North America are generally served by transformers that are connected in a wye configuration that is solidly grounded. The advantage of this arrangement is that when a single phase-to-ground fault occurs, voltages on the unfaulted phases remain relatively stable and do not subject customers to high phase-to-phase voltages during fault conditions. The disadvantage of this system is that the voltage on the faulted phase drops to near zero even for transient faults and, therefore, interrupts service served from that phase. Medium-voltage power plant auxiliary systems are frequently either connected in a delta configuration or a resistance or impedance is placed in the ground connection. The advantage of these systems is that the voltage on the faulted phase stays high on single-phase faults. This means that motor loads can continue to function and operate the device without interruption and fault currents are restricted during single-phase faults. The importance of such a connection is that a time limit is involved for the circuit to be cleared. There are basic assumptions in these designs concerning how quickly manual action will be taken should a ground fault occur. These assumptions are frequently lost and are not included in system operating instructions. Often, plant personnel believe that the systems are designed to function for long periods with a single-phase fault—which is rarely true. An important consideration is the possibility of the single-phase fault progressing to a two- or three-phase fault that could lead to severe damage to interconnected equipment. Three voltage levels and corresponding insulation thicknesses have been established for medium-voltage cable systems, with corresponding times for a single-phase fault to be cleared [6]. They are included in Association of Edison Illuminating Companies (AEIC) specifications such as CS1-90 [7], which states the following: 100 Percent Level: This insulation level is designated by the normal phase-to-phase system voltage. This is applicable only to systems where the normal voltage between the cable conductor and the insulation shielding tape or metal sheath will not exceed 58 percent of the phase-to-phase voltage. These cables may be applied when the system is provided with relay protection such that the ground faults will be cleared as rapidly as possible, but in any case within one minute. These cables are applicable to the great majority of cable installations that are on grounded neutral systems, and they may be used also on other systems for which the application of cables is acceptable, providing the above clearing requirements are met in completely de-energizing the faulted section. 133 Percent Level: This insulation level corresponds to that formerly designated for ungrounded systems. Cables in this category may be applied when the clearing time requirements for the 100 Percent Level category cannot be met, and yet there is adequate assurance that the faulted section will be de-energized in a time not exceeding one hour. Also they may be used when additional strength over the 100 Percent Level is desirable. 173 Percent Level: Cables of this designation should be applied on systems when the time required to de-energize a grounded section is indefinite. Their use is recommended also for resonant grounded systems.

2-3

Understanding the Design of Power Plant Cable Systems

Faults on power plant cables are expected to clear immediately by automatic means or within a short, finite time. Care must be taken that operators understand that if a cable shows signs of a phase-to-ground fault, such as a ground alarm or popping noises from a manhole or cubicle, the cable is de-energized as rapidly as possible. 2.2.2 Phase-to-Phase Faults The description for the 173% level “when the time required…is indefinite” seems to indicate that if the cable is rated at 173% of the operating voltage there is no reason to clear the fault manually or automatically. This is not true for several reasons. If the fault persists, the heat and arcing from the fault will likely cause additional damage to the adjacent phases. Significant damage can occur rapidly. The other sobering thought is that the failure of the first phase of a cable should lead one to believe that all phases of the cable are deteriorated and hence the other phases might fail soon. If the second failure occurs, the fault current is no longer limited by the resistor or impedance in the ground connection and the current can reach tens of thousands of amperes. Phase-to-phase faults cable faults can lead to weakening or failure of upstream transformers and switchgear. If the transformer does not failure during the event, it can be left susceptible to failure should another surge occur or it can be degraded enough that it fails a short time later. Operators of power plants should take defensive action–that is, de-energize the circuit as soon as reasonably possible if a single phase-to-ground fault condition is suspected to eliminate the possibility of a phase-to-phase fault. 2.2.3 Number of Grounds on a Cable Insulation Shield The metallic portion of the insulation shield must be grounded at least once for each phase of every individual circuit for the shield to perform its function. This ground connection must be of sufficient ampacity to ensure that the protective relaying function is activated in the event of a fault. Two different levels of fault current can be involved in any circuit. If the circuit is fed by a resistance or impedance type of ground connection and is promptly cleared in the event of a single-phase fault, then one #6 copper ground might be adequate. If the circuit is designed to not promptly be cleared by automatic protection, a much larger ground lead (cable shield and grounding connection) must be provided. Single-point grounding is often used on circuits where the conductor is 1000 kcmil (507 mm2) copper or larger. The reason for this is that the second ground on that metallic shield produces a transformer action that causes current to circulate in the shield system. This results in additional I2 R losses that effectively reduce the ampacity of that circuit. Although this is a concern in long transmission circuits, it is often not a problem in most power plant runs. Design calculations (see Figure 2-2), should be performed to determine whether an I2R heating is acceptable when both ends of the insulation shield are grounded. If the heating is too high, only one end should be grounded.

2-4

Understanding the Design of Power Plant Cable Systems

Figure 2-2 Single-Point Grounding of a Shield, Showing Voltage of Shield to Ground as a Function of Distance

However, with single-point grounding, a voltage builds up on the shield as a function of the current in the central conductor and the distance from the grounding point (see Figure 2-3).

Figure 2-3 Shield Losses and Voltages for Single-Conductor Cables [8]

2-5

Understanding the Design of Power Plant Cable Systems

This voltage can be as much as 100 V on a heavily loaded circuit and becomes a concern for the integrity of the jacket as well as a concern for workers that are not aware of the voltages and current that might be present on the shield. When current flows in the conductor of the cable, that current produces electromagnetic flux in the metallic shield, if present, or in any parallel conductor. This becomes a one-turn transformer when the shield is grounded two or more times, because a circuit is formed and current flows when a person touches the cable surface at the ungrounded end of the circuit. The cable circuits that should be considered for single-point grounding are systems with conductors of 1000 kcmil (507 mm2) and larger and anticipated loads of more than 500 A. The EPRI report Power Plant Electrical Reference Series, Volume 4: Wire and Cable (EL-5036) should be reviewed before designing a single-point grounding scheme [8]. In addition, if a discontinuity occurs in the metallic portion of the insulation shield that is grounded at only one end, a discharge site is created that can lead to carbonization and tracking at the point of the break leading to insulation failure. One such example is described in Section A.2.3. If the shield is grounded two or more times or otherwise completes a circuit, the magnetic flux produces a current flow in the shield. The amount of current in the shield is inversely proportional to the resistance of the shield. The voltage on the jacket surface stays at zero if two or more grounds are connected but current is being carried in the shield. This is the reason that it is not safe to cut a shield in half while the cable is energized, because the two open ends become energized. The distance between the grounds has no effect on the magnitude of the current. Whether the grounds are 1 ft or 1000 ft (0.3 m or 305 m) apart, the current is the same—depending on the current in the central conductor and the resistance and impedance of the shield. In the case of multiple cables, the spatial relationship of the cables is also a factor. Figure 2-2 provides equations for calculating shield currents and induced voltages.

2.3

Multiple Cables per Phase and Balanced Magnetic Fields

The large currents that are involved in power plant auxiliary circuits often require two or more cables per phase to carry the load. (For this portion of the report, large current is defined as more than 200 A on a single cable.) Because copper conductors larger than 1000 kcmil (507 mm2) are generally not economically effective, there is a need to properly design and install parallel cables. These rules apply to low-voltage as well as medium-voltage cables. The most obvious fact is that parallel cables must have the same length, conductor size and type, and similar connector types so that the resistance is equal on all cables in that phase. This is extremely important if the parallel cables are quite short—less than 20 ft. To obtain sufficient ampacity, limit voltage drop, and attain load sharing between two or more cables, the position of each cable must be correct. There are three basic areas of concern when designing the installation of single conductor cables: 1) orientation or configuration of the phases, 2) eddy currents, and 3) circulating currents [9, 10].

2-6

Understanding the Design of Power Plant Cable Systems

2.3.1 Configuration When a single-conductor cable carries current, it generates a magnetic field that radiates out and alters the impedance of adjacent cables. If the configuration is not correct, a serious impedance imbalance and, therefore, an imbalance in the current of parallel cables can result. If is the configuration is done correctly, each cable making up one phase has an equal magnetic field acting on it. Parallel cables must have the same conductor size and metal, the same length, the same insulation type, and the same connector type, and they should be larger than #1/0 (~0.08 in2 [~53.5 mm2]). Figures 2-4 through 2-6 show possible configurations for balanced magnetic fields.

Figure 2-4 Acceptable Configurations: Two Cables per Phase

Figure 2-5 Acceptable Configurations: Three Cables per Phase

2-7

Understanding the Design of Power Plant Cable Systems

Figure 2-6 Acceptable Configurations: Four Cables per Phase

2.3.2 Eddy Currents Eddy currents are generated in any ferrous metal (iron or steel) surrounding an individual, singleconductor cable carrying alternating current because of the varying magnetic field. Where a single conductor cable passes through steel plates on enclosures, structural steel, or steel conduits—to name a few—eddy currents will flow. Heat is generated in the surrounding steel. As long as the current is less than about 200 A, the small amount of heat generated is usually not a significant factor. The preferred practice is to use only nonferrous plates—such as aluminum—mounted over one common hole. The single-conductor cables plus the neutral enter through this one plate. No eddy currents are generated by this method because the individual magnetic fields from the three phases cancel one another. This magnetic field cancellation effect also applies where cables pass through structural steel, are run in steel cable trays supported by steel, or are run on steel struts that are supported by steel. As long as the three phases are run through the same steel “window,” the net magnetic field and eddy current are almost zero. When single-conductor cables run through reinforcing bar that might be embedded in concrete floors or walls, there is much less of a problem from heating because any heat that is produced by eddy currents will be dissipated by the large heat sink of the concrete. Direct contact between the cables and rebar should be avoided. 2.3.3 Circulating Currents Circulating currents are generated in a metallic layer of single-conductor cables when the conductor is carrying current—regardless of the type of metal—and when the metallic layer has more than one ground. The magnitude of this circulating current depends on the amount of current in the conductor, the resistance and impedance of the loop formed by this layer, and the ground path. This path could be interlocked armor, shielding tapes, or metallic conduit.

2-8

Understanding the Design of Power Plant Cable Systems

2.4

Protective Relay and Annunciation Alarm Systems

Most medium-voltage cable circuits have protective relaying or annunciation alarm systems or both to alert the operator when a fault occurs on that circuit. Protective relaying might not be designed to automatically trip (de-energize) the circuit, but this is no reason to ignore the cause by simply acknowledging the annunciation alarm. There are many different design factors for power plant circuits and, therefore, there is no single rule regarding the proper action to be taken after an alarm. A generalized protocol should include the following steps: 

Determine whether a fault actually exists and record the current on the phase or phases.



If a phase has a target, do not reclose until troubleshooting determines the cause of the relay operation.



Review the equipment that is served from that circuit.



Evaluate the criticality of repair.



Determine whether there are redundant circuits in operation.



Walk the circuit to see whether the fault has created visible, odor, or sound indications.



Inform supervision of the results of this initial investigation.



De-energize the circuit for additional inspection unless other factors take precedence.



Remember that the initial fault might clear itself by burning away the shield or nearby grounds.



Also remember that a fault that is not de-energized can continue to generate heat and arcing that can damage the adjacent phase insulation and cause a phase-to-phase fault.

If the cable is nonshielded, the likelihood is that the alarm has detected a fault and that reclosing should not be attempted until troubleshooting has confirmed that no fault exists.

2-9

3

UNDERSTANDING THE PHYSICAL CONDITION OF THE SYSTEM The environmental and operating conditions of a cable circuit are a major factor in determining their long-term reliability. Cables operated in benign environmental conditions (dry, low ambient temperature, and low radiation dose levels) and benign operating conditions (lower voltage stress and currents lower than rated) are highly likely to function reliably well beyond 40 years. Cables in benign environments have failed in operation, but for the most part, those failures were attributed to random defects or installation damage rather than an aging problem. This section describes the adverse environments and operating conditions that must be understood by personnel responsible for cable system aging management. Knowing the adverse environments and conditions that affect cable longevity will allow personnel within a plant to identify them and to select appropriate aging management methods for the cable exposed to them.

3.1

Conditions in Manholes and Ducts

Power plant duct and manhole systems that are beneath the earth’s surface can become flooded. If the flooded conditions last for short periods, there will be little effect on most insulation systems. However, if manholes and ducts are prone to long-term flooding, a means of removing the water should be established and, if the cable is testable, periodic testing is recommended. Energized, wet medium-voltage cables are likely to age from electrochemical and electromechanical actions of the water in the polymer matrix. The rate of aging is different for the different insulations, different semiconducting layers, and different vintages of cables. Ethylene-propylene rubber (EPR) and butyl rubber cables produced before 1976–1978 are likely to age more rapidly than cables produced after that time. The aging, even in the more susceptible cables, is slow, with approximately 30 years of operation before the initial age-related failure in the population. Continuous improvements have been made in cross-linked polyethylene (XLPE) cables over the years. However, when they fail, most XLPE cables are replaced with EPR insulated cables. 3.1.1 Insulation Deterioration Historically, cable engineers mainly believed that three inherent factors were related to insulation failures: heat (oxidation), water, and time. Conductor current can cause excessive heating if cables are operated at or above their rated ampacity. The use of larger conductors will reduce ohmic heating, reducing temperature and extending the life of the insulation systems. Ambient environments can be controlled by adding cooling to an area or rerouting the cable. Prolonged immersion in water was recognized to have two major reactions on underground power systems—insulation deterioration and corrosion of 3-1

Understanding the Physical Condition of the System

metal shields and conductors. These conditions can be controlled by design of the cable, constructing the duct system to remain dry, or adding dewatering systems. The length of exposure to the adverse condition will determine the ultimate life of the cable. Short or no exposure to elevated temperature and wetting will have little or no effect on cable life. Extended exposure to elevated temperature or wetting or both will shorten a cable’s life. Natural rubber was the main source of insulation for medium- and low-voltage cables for power plants from 1900 to 1950. Each cable manufacturer had its own formula for rubber compounding. Testing was limited to taking a short sample of a manufacturer’s insulation, weighing it, soaking it in hot water for about one week, and then weighing it again. If the weight of the insulation had increased by less than 10% by the end of that week, the compound was considered adequate for insulation used in wet environments. World War II brought polyethylene (PE) to the cable industry—first for radar cables and, just after the war, to low- and medium-voltage power cables. Using the same weighing method, they found that they could not measure any weight increase. Electrical measurements were incorporated to show any change in properties. These test results seemed to indicate that this new material was unaffected by water. This was not the case, because even a small amount of moisture could deteriorate PE by creating water trees when the cable was energized for a period of time in a wet environment. By the 1950s, researchers determined that moisture could penetrate the insulation wall of a cable if a dc negative polarity was applied to the conductor. Test standards, therefore, specified that dc test sets should be constructed to impose dc negative polarity on the conductor of the cable in a test environment. In the case of ac, the voltage is negative for only one-half cycle, and engineers originally believed that during the positive half cycle, the water would be pushed away from the conductor. However, degradation was still noted at the conductor-to-insulation interface. This condition led to the theory that dielectrophoresis was applicable to water migration in ac cable circuits. Electrophoresis is a term used to describe the movement of charged particles in an electric field. Particles with a positive charge tend to move toward a negative electrode, and negative ions tend to move toward the positive electrode. Dielectrophoresis relates to the movement of an uncharged but polarized particle or molecule in a divergent (ac) field. In the example of a single-conductor electrical cable, the field in the insulation increases as a particle or molecule gets closer to the conductor. An uncharged particle will be polarized at any given point in time, so that it will have a negatively charged dipole with its negative side toward the conductor that is positive at that instant. Because the negative side of this dipole exists in a stronger field than the positive side, the particle will be attracted toward the field of greatest field intensity. In an ac system, as the conductor becomes negatively charged, the polarization process is reversed. This means that the particle is still attracted to the conductor with its higher electric field (see Figure 3-1).

3-2

Understanding the Physical Condition of the System Dielectrophoresis

Figure 3-1 Electrostatic Flux Lines and Charged Particles

The practical effect of dielectrophoresis is that moisture is drawn to the higher dielectric field regions, even in an alternating field. This high stress point is likely to be a small void that is a portion of the initial tree formation. The void that was initially filled with gas now becomes filled with water. Although this does not fully explain the formation of the water tree, it does shed some light on the growth of such trees and the dispersion of moisture in an energized cable. Dielectrophoresis provides a means for water to propagate through the insulation to feed water tree formation, both in the insulation and at the interface between the conductor shield and the insulation. One way to solve the problem of moisture ingress is to use a jacket or impervious metal sheath over the insulation shield system. When there is no voltage drop across the jacket, there is no dielectrophoresis effect. All of this is possible only when the jacket is intact, of course. Putting a jacket over nonshielded conductors does not accomplish this goal because the jacket material does have a voltage drop across it. Even semiconducting layers of insulation do not stop the process because there is some voltage drop across that portion of the cable as well. However, some jacket materials are better than others at preventing water passage through to the insulation. Dyed chlorosulfonated polyethylene (CSPE) jackets allow much more water transfer than do black jackets in which carbon is the colorant. Accordingly, water transfer will occur through some jacket materials without needing dielectrophoresis to cause it. This transfer occurs relatively slowly, on the order of months or more. 3.1.2 Pumping and Dryness One method of keeping the duct and manhole system dry is to install sump pumps. Water from the ducts must drain into the manholes, as depicted in Figure 3-2.

3-3

Understanding the Physical Condition of the System

Figure 3-2 Example of Manhole Drainage System

A well-designed dewatering system does not require a pump in every manhole, but it can be desirable to have a sump pump built into each manhole for occasional water intrusion. Automatic sump pumps might be necessary. They use a float switch or, when the water rises, the entire housing goes up to turn on the pump. Regular inspection and maintenance of sump pumping systems are required; otherwise, they can fail relatively rapidly and allow the cables to be submerged. If pumping is to be performed, whether using automatic sump pumps or periodic inspections with manual pumping, sampling of the water for environmental or radiological conditions might be necessary to determine how to dispose of the water. Pumping oil-contaminated or tritiumcontaminated water into storm drains or into the ground could have serious consequences.

3.2

Correcting Adverse Conditions

3.2.1 Adverse Dry Conditions Although dry cables are expected to have longer lives than wet cables, dry cable failures are possible if adverse conditions exist. These conditions can exist from the time of installation or can occur over the life of the plant. Care should be taken to resolve adverse conditions when they are recognized. A number of dry condition failures are described in Appendix A to provide insights regarding failures that could occur and the related circumstances. Dry failures are often caused by random factors such as manufacturing flaws and installation errors coupled with an adverse environment or condition. The following subsections describe the stressors and their effects. 3.2.2 Physical Stress A number of physical conditions can adversely affect medium-voltage cable life, including overbending, compression, cuts, and gouges. When a shielded cable is bent into too tight a bend radius, the insulation-to-shield interfaces can be disrupted, providing gaps where partial discharge (PD) can occur, which would lead to short life. Permanent compression of the shield and insulation system can cause elevated potential stress in the insulation or disruption of shields. Cuts and gouges, depending on their severity, can disturb shields or even cause elevated potential stress in the insulation. Tension or compression forces on cables due to routing and termination can result in mechanical damage (due to external vibration) or electrical discharge degradation (for nonshielded or single-point grounded cables). Tension can also result in failure at a cable connection point such as a splice or termination. If such conditions are identified, the stress should be resolved and repairs made as necessary, or the cable condition should be monitored through periodical diagnostic testing. At minimum, performing a damage evaluation 3-4

Understanding the Physical Condition of the System

is warranted. Discussions with the manufacturer of the cable might provide insights regarding the importance of the damage and any necessary corrective actions. Compression and damage to cables can occur at the dropouts from trays to local conduits. The cable should be protected from sharp edges of the conduit by a bell or other appropriate fitting, and padding might be necessary on the rungs of the tray where the cable drops out to preclude excessive load on the side of the cable. 3.2.3 Vertical Support Medium-voltage cables with larger conductors are heavy and need appropriate support devices where long vertical drops occur. Improperly supported cable can be crushed at its top support, leading to high electrical stress or disrupted shields at the top of the vertical run. Care must be taken when following the National Electric Code requirements for supporting larger cables. The standard prescribes the same number of supports for any cable larger than 500 kcmil (250 mm2), but if the standard were followed literally, the supports would exceed the support manufacturer’s allowable weight per unit length for cables larger than 500 kcmil (250 mm2). Manufacturer’s literature should be consulted when determining vertical support requirements. If not properly supported, the weight of the cable can also pull on connections at the top of the run, possibly leading to failure. Long vertical runs should be supported by strain relief grips. 3.2.4 Adverse Environments Given that medium-voltage cables run throughout the power plant, localized adverse environments can affect them. These environments can be permanent or the effect of errors or failures in the plant. The most likely adverse environment is elevated temperature and radiant energy conditions that occur when thermal insulation is left off, displaced, or temporarily removed from adjacent high-energy piping. Another possible damaging environmental effect is hot water or steam leaking from a pipe or valve that impinges the cable. These conditions should be corrected as soon as they are observed. 3.2.4.1

Temperature-Related Aging

Many areas and rooms inside the power plant are relatively cool environments, less than 40°C (104°F). However, some areas that contain medium-voltage cables can have temperatures well in excess of 50°C (122°F), which could reduce the life of the cables. These areas must be identified and appropriately managed in accordance with the general guidance provided in this report. 3.2.4.2

Radiation-Related Aging

It is not expected that medium-voltage cables will be subject to radiation levels high enough to cause cable aging in conventional plants. The EPRI has developed reports explaining that higher doses of radiation change the physical properties of the cable. Increased insulation hardness and loss of elongation at a break of the insulation occur after severe aging. Medium-voltage cables identified to be subject to doses greater than 5 Mrd (50 kGy) per 40 years or less should be monitored in accordance with the recommendations in this report.

3-5

Understanding the Physical Condition of the System

3.2.4.3

High Conductor Temperature from Ohmic Heating

Medium-voltage cables can also be affected by long-term high currents due to loading errors or unbalanced magnetic circuits. Such conditions can be compounded at fire stops, where heat transfer is reduced, causing further elevation of temperature within the cable. If conductor temperatures are found to exceed the cable design rating, they must be evaluated, and corrective actions must be taken. To date, the dominant issues relating to ohmic heating have occurred on multiple-conductor cable due to magnetic or resistive imbalances among the individual phase conductors. Studies on heating are available in the Nuclear Energy Institute (NEI) white paper 06-05, “Medium Voltage Underground Cable” [11]. 3.2.4.4

High-Resistance Connections

Improperly made splices and terminations can deteriorate from elevated temperatures due to high-resistance connections. Terminations that are separable and not properly reassembled (as verified by post-maintenance measurement of connection resistance or thermography (if accessible) are also candidates for thermal degradation over time due to high connection resistance. These conditions, if not identified and corrected, will thermally degrade the cable insulation or accessory over time. Identifying and correcting high-resistance connections according to the guidance in this report will help to limit this failure mode. 3.2.5 Surface Corona and Partial Discharge Nonshielded cable can be subject to surface PD (corona) in the small gap adjacent to the location at which the cable touches a grounded metal surface. Corona discharges occur from ionization of the air gap between the cable and the grounded surface. The conductor voltage distributes across the insulation, jacket, and air gap, with a large portion of the voltage across the air. With a high voltage across a small gap, a voltage stress higher than the breakdown stress of the air occurs. The gap discharges and the voltage redistributes across the insulation and jacket so that the discharge is stopped. However, each electron stream causes a small increment of damage to the polymer surface, resulting in erosion of the polymer. Over an extended period, these discharges can erode the surface of the cable’s jacket and continue to slowly reduce the dielectric strength of the insulation system, if not corrected. In such cases, the presence of corona discharge is often indicated by a white powder in the vicinity of the discharge. Corona attack can be identified by visual examination during maintenance when terminations and junction boxes are accessible. An example of corona discharge is shown in Figure 3-3.

3-6

Understanding the Physical Condition of the System

Figure 3-3 White Powder Indicates a Location of Corona Discharge Between a Cable and a Ground Cable in Close Proximity

3-7

4

CABLE DESIGNS This section describes cable constructions that have been installed in typical power plants.

4.1

Cable Design Summary

In general, most plant cable systems are believed to be less susceptible to moisture-related degradation than similar systems in distribution service, due to their low electrical stress levels and the use of rubber insulations, overall jackets, duct bank systems, and well-shielded terminations. Shielding the terminations from lightning strikes removes a significant source of severe voltage surges that can cause failure or initiate electrical treeing in deteriorated cable. Most power plant circuits are terminated inside buildings or sheltered areas so that they will not be exposed to lightning strikes. Medium-voltage cable systems have operating voltages in the lower band of the medium-voltage range, where the electrical stress is lower than in the upper band of the medium-voltage range. This lower voltage stress causes electrical-related degradation to occur more slowly than in cables operating at 35 kV and greater. Much of the literature on cable insulation aging relates to the aging of wetted cable systems that are more prone to electrical failure. The vast literature produced in response to moisture-related degradation in the distribution arena has been assessed as a valuable resource for utilities in understanding and evaluating moisture-related degradation in XLPE insulated cables. It is less applicable to the majority of stations because they have rubber insulation systems. Section 6, Fundamentals of Cable Insulation Systems, describes the differences between XLPE and rubberbased (butyl rubber, ethylene-propylene-diene monomer [EPDM], and EPR) insulation systems.

4.2

Medium-Voltage Cable Constructions

Underground medium-voltage cables at plants are installed as one of three basic cable assembly configurations: 

Individual, insulated single conductors



A twisted combination of the insulated single conductors, known as a triplexed assembly



A jacketed three-conductor cable

In any of these assemblies, the insulated conductors share the same basic construction shown in Figure 4-1, with the exception of the jacketed three-conductor cable, which has an overall jacket over the insulated singles.

4-1

Cable Designs

Figure 4-1 Shielded, Single-Conductor, Medium-Voltage Cable Design

The conductor (A) is typically stranded copper or aluminum, with copper being more common in power plants. The insulation (C) used in conventional and plant medium-voltage cables is EPR or, to a lesser extent, XLPE. Medium-voltage cables in a few early plants had butyl rubber insulation. Shields (B and D), composed of semiconducting polymer in modern designs, help maintain a uniform voltage stress in the insulation. The metallic tape shield (E) provides a continuous drain for the insulation shield and a return path for fault currents. The jacket (F) adds mechanical protection as well as an additional barrier to moisture and external contaminants. In early black EPR and butyl rubber cables, the insulation and conductor shields were formed from helically wrapped carbon black–loaded cotton tape. By the early 1970s, the insulation shield was made of a helically wrapped semiconducting polymer tape. The insulation shields were in tape form to enable installers to differentiate between the black shield and the black insulation. When strippable, extruded black semiconducting polymer shields became common, the industry converted to gray, pink, or brown EPR by eliminating some or all of the carbon black from the insulation. Extruded conductor shields became available for power plant use in the mid 1970s. A modern construction, similar to that shown in Figure 4-1, is the water-impervious design shown in Figure 4-2. The sub-components are virtually the same, but the water-impervious cable is constructed to limit absorption of moisture and external contaminants through additional radial and axial design barriers. One design barrier is that the metallic tape shield (E) has been replaced with a continuous, corrugated tape shield that is copolymer coated and sealed at the overlap. A second design barrier is the use of water-swellable tapes or powders placed in the conductor strands to prevent water migration through the stranded core. These two design changes provide barriers to prevent moisture and external contaminants from penetrating into the layers below. The corrugated metallic shield also should provide lower, more stable shield resistance compared to the helically wrapped copper tape design, which should aid in long-term testability. Other moisture-impervious designs exist, including those that use a fully sealed lead or aluminum sheath instead of a corrugated copper sheath with a glued overlap. For more information on cable design and selection, refer to the EPRI report Plant Support Engineering: Common Medium Voltage Cable Specification for Power Plants (1019159) [12].

4-2

Cable Designs

Figure 4-2 Medium-Voltage Shielded Cable, Water-Impervious Design

Two disadvantages of the water-impervious design are the higher cable cost and the increased difficulty in making field terminations. The cost differential might not be significant in comparison to the overall cost of installation. Using a lead sheath in place of the corrugated copper tape is also possible, but it has limited availability and requires additional care in pulling to preclude damage to the lead sheath. Installation of lead-sheathed cable might not be practical when replacing cables that had tape shields or were nonshielded. The lead-sheathed cable will likely require larger ducts and largerradius bends. Figure 4-3 shows a nonshielded cable design. In this design, there is no insulation shield. This design has been applied in many instances for medium-voltage circuits of 4160 V and lower. The reason is that non-grounded systems, when designed properly, can tolerate a single phase-toground fault while remaining in service for a short period. However, the absence of a shield that confines the voltage stress to the insulation makes testing of the insulation system in a meaningful way quite difficult. Testing is possible in a laboratory by submerging the cable in water and using the water for the ground plane. Submergence of an entire circuit is not practical in a plant and can provide confusing results because the jacket of the cable is in series with the insulation and can affect the test results.

Figure 4-3 Medium-Voltage Nonshielded Cable Design

4-3

Cable Designs

Figure 4-4 shows an example of the UniShield cable design. In this design, the outer polymer layer serves as both jacket and semiconducting insulation shield for the cable. The shield conductor system is composed of six longitudinal, corrugated, neutral wires embedded in the jacket. The design includes a compacted conductor. The result of this design is a smallerdiameter cable. The small-diameter design might be needed when existing ducts are relatively small in diameter because smaller-diameter cable had been previously used. Other designs exist that have small diameters, but they have not been commonly used in power plants to date.

Figure 4-4 UniShield Construction, Shield Wires Embedded in the Semiconducting Outer Jacket

4.2.1 Voltage Rating Typical distribution service feeders are rated 5 kV through 35 kV. Plant auxiliary power distribution system feeder cables are predominantly rated 5 kV; plants constructed after the late 1970s have 8 kV or 15 kV cable, as well. Medium-voltage systems in plants operate at 4.16 kV, 6.9 kV, 12.2 kV, or 13 kV. Plants with 12.2 kV or 13 kV systems generally also have 4.16 kV systems. The combination of lower voltage rating (that is, 5–15 kV) and larger minimum conductor size (see Section 4.2.2, Conductors) means that cable insulations generally operate at the lower end of the range of electrical stresses at which water treeing occurs in XLPE and water-related degradation occurs in EPR. Therefore, such systems are less susceptible to moisture-related degradation, and water-related failures tend to occur later in cable life than has been recognized in distribution systems with smaller conductors and operating voltages of 33 kV and greater. Insulation must be able to withstand the voltage stress experienced during normal operation, as well as voltage surges. Thicknesses of insulation material for a given voltage rating have been established for different materials by the Insulated Cable Engineers Association (ICEA) and AEIC for power plant applications. Commercial and industrial cable applications are governed by Underwriters Laboratories, Inc. and National Electrical Code standards. The voltage ratings are phase-to-phase ac that relates to the maximum phase-to-phase operating voltage being used in the system. The discrete ratings are 5 kV, 8 kV, 15 kV, 25 kV, and 35 kV [8]. 4-4

Cable Designs

When a ground fault occurs on a power system, voltages higher than the cable’s voltage rating can occur. To address this, ICEA and AEIC specify thicker insulation if the overvoltage can last longer than 1 minute. The duration of the overvoltage depends on whether the neutral is solidly grounded or ungrounded. The insulation and resultant thicknesses are classified as 100%, 133%, or 173% insulation levels [8], as follows: 

100% Level. Relay protection normally clears ground faults within 1 minute.



133% Level. The faulted cable de-energizes within 1 hour.



173% Level. The time needed to de-energize the fault is indefinite. This level is recommended for resonant grounded systems.

Even if the 173% level of insulation is used, it does not mean that cables with a ground fault can be left in service. If a ground is suspected (for example, an alarm is received), action should be taken to remove the cable from service as soon as possible. The faulted phase can adversely affect the adjacent phases and lead to a phase-to-phase fault that will have extremely large currents, which could cause damage to the connected equipment. For plant use, some designers opted to use the 133% level as conservatism. 4.2.2 Conductors The conductor (see A in Figure 4-1) is typically stranded copper or aluminum, the former being more common in power plants. The cross-sectional area size, the associated diameter, and number of strands in the conductor are standardized and unaffected by material choice. The conductor size is given in terms of American Wire Gauge (AWG) or kcmil in the United States and in terms of mm2 elsewhere. The outside surface of the conductor can be smoothed to a nearperfect circle by compacting the strands to a compact-round standard, which can result in up to a 3% diameter reduction. The compression also aids in attaining a smoother interface between the strands and the conductor shield. Figure 4-5 shows a comparison of the conductor stranding compression options. Although it reduces flexibility somewhat, the compact-rounding helps to avoid potentially ionizing and discharging air gaps at the inside surface of the insulation, where the voltage gradient is greatest. The smallest diameter stranded conductor configuration results from compacting in which each strand is shaped to allow approximately 9% reduction in diameter. This degree of diameter reduction is used in cables designed for smaller diameter ducts and conduits.

Figure 4-5 Conductor Stranding Configuration, Showing Compressed and Compacted Conductor Configurations

4-5

Cable Designs

Distribution and plant conductor design philosophies differ substantially. Distribution designers frequently use aluminum conductors and typically use solid conductors for aluminum wires up through 2/0 AWG (67.5 mm2) and copper up through 6 AWG (13.28 mm2). Stranded conductors are used for all larger sizes. Although the solid conductor is somewhat stiffer, it does effectively block the flow of water through the interior of the cable. In contrast, plants use stranded copper conductors to allow greater flexibility for installation. Water blocking of stranded conductors can be accomplished through the use of polymer fills, water-swellable powders, or a thixotropic gel filling for the strand interstices. However, water-blocking technologies were not available at the time of construction of most plants, but they are available for replacement cables and new plant use. Historically, most cables were not compacted, with the exception of the UniShield construction, which is designed to be a small diameter for a given voltage rating. 4.2.3 Conductor Shield Air gaps between the conductor and insulation result in high-voltage stresses that cause the gap to periodically discharge. Such discharges can damage the insulation and lead to failure. To eliminate the air gaps between the conductor and insulation, an effective conductor shield or a strand shield (see B in Figure 4-1) is required over the conductor, regardless of whether the insulation itself is shielded. Addition of a semiconducting layer between the insulation and the conductor prevents voltage from building in gaps between the insulation and the conductor by eliminating the gap and causing surface charge on the insulation to be drained to the conductor. The semiconducting layer must be in intimate contact with the inner diameter of the insulation. The conductor shield is typically a thin (~10–20 mil [0.254–0.508 mm]), extruded semiconducting compound that is compatible with the primary insulation. Like the electrodes of a capacitor, the insulation shield on the opposite side of the cable insulation and the conductor shield help to confine the electric field and create symmetrical radial distribution of voltage stress within the dielectric. Due to limits in extrusion technology, helically wrapped carbon black–loaded cotton tapes were used for the conductor shield in medium-voltage cable designs available in the late 1960s until shortly after 1970. With such tapes, stray fibers protruding from the conductor shield tape could become encapsulated in the insulation during extrusion. These protrusions became initiating sites for water-tree growth. The subsequent development of dual-pass and dual-tandem extrusion systems facilitated the use of polymeric conductor shields and the elimination of the inner tape. The Kerite Company’s Permashield design uses an alternative stress reduction technique at the conductor to insulation interface. Instead of using a semiconducting conductor shield, a highpermittivity polymer layer is applied to the conductor and bonded to the insulation. The layer limits electrical stress at the conductor-to-insulation interface. 4.2.4 Insulation The cable’s primary insulation (see C in Figure 4-1) is manufactured of materials that are designed with sufficient dielectric strength to withstand the voltage stress experienced during normal operation, as well as unusual voltage spikes and surges. The insulating material for most medium-voltage plant cables is either XLPE or EPR; however, some early plants have cables with butyl rubber insulation. Although XLPE has additives for fire retardance and processing, it 4-6

Cable Designs

is mostly XLPE polymer, but a copolymer (additional plastic type) is often used. Recent experience within the transmission and distribution utility industry has led to the development of a tree-retardant enhancement of this insulating material. The additives to tree-retardant XLPE (TR-XLPE) do not totally eliminate water trees; rather, they greatly reduce their rate of generation and growth. In contrast to the relatively limited number of variations in XLPE formulations, EPR insulations are comparatively complex compounds that vary substantially among polymer suppliers and cable manufacturers. Different improvements of the EPR materials have evolved and have been distinguished from one another by color, such as black, gray, brown, or pink. The colors alone are not directly related to the way these materials age, but they are indicative of the changes made to the formulations that improved the longevity of cables. Black EPR cables are earlygeneration EPR cables manufactured through the mid to late 1970s. Brown EPR appears to have been resistant to water-enhanced aging effects throughout the period of its use. Pink (or red) EPR cables are the more modern generation, available from the late 1970s through today. A few manufactures changed from black EPR to gray EPR in the late 1970s. The major shift in EPR formulation during that period was the transition from untreated clay to silane-treated clay. The silane treatment caused the EPR to bind more tightly to the clay, and it sealed the clay so that water uptake in the EPR was greatly reduced. A typical dielectric constant for XLPE is ~2.3, whereas that for EPR is distinctly higher, at ~3.2. Thus, EPR has higher dielectric losses per unit length of cable than XLPE, but it is generally more resistant to voltage stress and discharges. The dielectric loss through the EPR insulation drains any charge that could build in the insulation at imperfections and eliminates high localized stresses that result in water-enhanced aging. Because medium-voltage cables in power plants are usually much shorter than those in transmission and distribution applications, electrical system losses are not paramount in the design of station cable systems. Thus, a majority of underground medium-voltage cables at plants have EPR insulation, which has always had an expectation of greater operating life. Rubber insulation systems also have been chosen for in-plant applications due to their flexibility, which is important during installation in the tighter confines of power plant applications. When PE insulation first became available for medium-voltage applications, it was hailed as the cure-all for many of the issues then facing the distribution industry. These PE materials had extremely low losses in a high-stress electrical field, were easier to compound than rubber systems, were lower in cost than either rubber or paper-insulated lead-covered cables, and were quite hydrophobic (tended not to absorb water or moisture), whereas conventional rubber systems were not. Thus, distribution utilities made widespread of various types of PE. generating station designers gave relatively little weight to PE’s low loss characteristics because of the insignificant circuit lengths involved. Those few plant designers who did not choose rubber-insulated systems chose XLPE for its superior mechanical strength and thermal endurance.

4-7

Cable Designs

4.2.5 Insulation Shield 4.2.5.1

Semiconducting or High-Permittivity Shield Layer

Historically, 5-kV rated cables could be purchased with or without an insulation shield. According to NEI 06-05 survey results [11], 30% of the respondents have some nonshielded cables installed. Typically, 8-kV and higher rated underground cables at plants are shielded. When grounded, the shield confines the electric field within the insulation and produces a symmetrical radial distribution of voltage stress within the dielectric, minimizing the potential for surface discharges. In addition, the shielding limits radio-interference generation, allows for individual conductor insulation electrical testing, and if properly grounded, reduces a possible shock hazard to plant personnel. The insulation shield is composed of a semiconducting polymeric insulation shield or screen (see D in Figure 4-2) and an overlying metallic component (see E in Figure 4-2). The Kerite Company offers an alternative shield design, called Permashield, using a high-permittivity shield instead of a semiconducting layer. Kerite uses a Permashield layer as a conductor shield and offers either a semiconducting layer or a Permashield layer for the insulation shield. When both shields are of the Permashield type, the cable is labeled “Double Permashield.” The semiconducting or Permashield layer eliminates air gaps between the primary insulation and the ground plane of the metallic shield that could ionize, discharge, and, in time, degrade the insulation. Earlier cables used cotton tapes, which were ultimately problematic because cotton fibers could enter the insulation during manufacture, leading to high localized stresses in the insulation. Semiconducting tape insulation shields were introduced in the late 1960s and continued to be used into the early 1970s. The use of insulation shield tapes simplified production and ensured that the cable could be readily spliced or terminated. The tapes have printed statements indicating that they must be removed when splicing and are readily discernable from the insulation. However, during the manufacture of this type of cable, the tapes are applied in a separate operation from the extrusion of the insulation. The exposure and handling of the insulation before the tape was applied allowed the possibility of contamination of the interface. Contaminants occurring during manufacture were subsequently identified as the root cause of water trees. These designs were replaced with the higher-reliability, extruded semiconducting shields that are currently available. The extruded semiconducting layers can be either thermoplastic or thermoset materials. As manufacturers switched to extruded semiconducting layers, so-called “dual-pass” extrusion systems were commonly used. Although the extruded insulation shield was a definite improvement over the old tape method, the interface was still exposed to contamination before and during the second pass. When the significance of this contamination was recognized, manufacturers eliminated the exposure through the use of three extruders on a single production line. The so-called “1 + 2” extrusion, in which the conductor shield was applied just upstream of a tandem extruder that applied both the insulation and the insulation shield was introduced in the 1980s. Triple extrusion, in which all three layers—the conductor semiconducting shield, the insulation, and the insulation semiconducting shield—are applied at once, precluding contamination, did not become common until the 1990s.

4-8

Cable Designs

Care must be taken when splicing new cable to old black rubber cables with tape shield designs. In modern cables, the insulation shield layer is easy to identify because it is black as opposed to the gray, brown, or pink insulation. Current splicing crews are familiar with the modern cable designs. However, they are unlikely to have seen woven fabric or polymer tape shields. They must be trained in how to make splices and terminations to these old style cables. If the semiconducting layer is not removed properly, failure is possible immediately on energization or shortly thereafter because the separation between the conductor and the shield will barely withstand operating voltage. 4.2.5.2

Metallic Shield Layer

The second component in an insulation shield system is a metallic layer that allows the system to be grounded. In-plant cables, copper tapes are the most common. Tinning was added to the tape when sulfur-bearing jackets were used to reduce corrosion. Following conventional industrial service guidelines, most plant designers chose a helically applied, 5-mil (0.127-mm) copper tape shield. The copper tape design allows the cable to be more flexible for in-plant installation and was acceptable because large neutral wires are not needed, as plants do not have single-phase loads. Regardless of the metallic shield design, the shield will bleed charge off the semiconducting shield layer and conduct fault current should the insulation fail. An ungrounded (floating) insulation shield performs no useful function. Essentially, the outer surface of the cable will react like a nonshielded cable, and nearly the full operating voltage will be present at the cable surface. Grounding the shield at only one end allows voltage buildup along the surface of the cable jacket that may result in lethal voltages being present at the ungrounded end of the cable. As cables age, light levels of corrosion on helically wrapped copper tape shields can cause significant attenuation of high-frequency signals. Although this does not present any operational concern, the oxidation can insulate the overlaps of the helical tape, forming an inductor that attenuates PD signals, making partial testing impracticable. Alternative shield designs, such as those in UniShield and the linear corrugated tape designs are less susceptible to corrosioninduced attenuation and they allow the use of tests that evaluate high-frequency signals. 4.2.6 Nonshielded Cables Some 5-kV rated rubber-insulated plant cable systems are of a nonshielded design. A few distribution utilities followed this same practice at 5 kV, using rubber and, to a lesser extent, XLPE insulation systems. The lack of a shield at the 5-kV level is not a longevity issue because the relatively thick layer of insulation with respect to operating voltage causes a low-voltage stress in the insulation. However, nonshielded cables are prone to induced-voltage shock hazards and potentially damaging corona surface discharges. In addition, the lack of an insulation shield eliminates the ability to perform meaningful electrical tests.

4-9

Cable Designs

4.2.7 Jacket Jackets protect the power cable’s underlying insulation from mechanical and environmental damage. Jackets are also used to maintain the configuration of multi-conductor cable. Although metallic armor might be used on some medium-voltage constructions, most underground power cables at plants have a nonmetallic jacket (see F in Figure 4-1) as their outer protective sheath (with or without underlying armor). The jacket materials include polyvinyl chloride (PVC), thermoplastic chlorinated polyethylene (CPE), neoprene, and CSPE. Replacement cables might have thermoset CPE or thermoset low-smoke zero-halogen jackets. The transmission and distribution industry also favors a linear low-density polyethylene (LDPE) jacket material for moisture protection, but PE jacket material is not used in plants because it is not flame retardant. Jackets slow moisture intrusion into the underlying cable core. In the case of the UniShield cable, the jacket serves a dual role as the insulation shield and the overall protective sheath. generating stations use flame-retardant designs that could be obtained only through the use of specially modified jacketing systems. When non-flame-retardant cables have been used, flameretardant coverings have been applied to prevention the propagation of fire. Although they are not impervious to water migration, the existence of the polymer jackets on plant cables has impeded the ingress of moisture into the insulation. The presence of this additional diffusion barrier (along with some of the installation attributes described in Section 6, Fundamentals of Cable Insulation Systems) helps explain why moisture-related degradation of medium-voltage cable is just now becoming a concern for plants, although the cables are 25–35 years old.

4-10

5

SPLICING AND TERMINATING This section describes splicing and termination of medium-voltage cables. Splicing of mediumvoltage cable is not a common practice in U.S. power plants, but it must be used in long runs and might be necessary when repairing a failed section of cable.

5.1

Cable Splicing and Terminating Theory

During the installation of medium-voltage cable circuits, connections might be necessary to create long lengths. These connections are referred to as joints or splices. Each circuit must have at least two ends known as terminations. Collectively, joints, splices, and terminations are often referred to as accessories in the literature. As described in Section 4, Cable Designs, there are both shielded and nonshielded mediumvoltage cables. This section concentrates on splicing and terminating cables with insulation shielding (commonly referred to only as shielding or shielded) because they are the dominant medium-voltage cables in plants and are closely related in most aspects. Nonshielded cables are connected the same way as low-voltage cables and the same way as shielded cables, with the exception of the two components of the insulation shielding system—the semiconducting tape and the metallic shield component. The splicing of two pieces of cable can best be visualized as two terminations that are connected together. The most important deviation, from a theoretical view, between splices and terminations is that splices are more nearly extensions of the cable. The splice simply replaces with field components all the various components that were made into a cable at the factory. Instead of two lugs being attached at the center of the splice, a connector is used. At each end of the splice where the cable shielding component has been stopped, electrical stress relief is required, just as it is when terminating.

5.2

Gradients

5.2.1 Electric Fields An electric field in a cable can be visualized with the use of equipotential and flux lines. The equipotential lines represent surfaces of constant potential difference between the two electrodes. The flux lines define the boundaries of dielectric flux between two electrodes. For a shielded, medium-voltage cable, these lines are illustrated in Figure 5-1 [13].

5-1

Splicing and Terminating

Figure 5-1 Equipotential and Flux Lines in a Cable [13]

When the cable is cut so that the shield ends abruptly, the electrical stresses change from being in the semiconducting material to being in the air, as shown in Figure 5-2.

Figure 5-2 Electrical Stress Fields, Shield Removed

To reduce the electrical stress at the end of the cable, the insulation shield is removed for a sufficient distance to provide adequate leakage or creepage distance between the conductor and the shield. The distance depends on the voltage involved as well as the anticipated environmental conditions. The removal of the shield disrupts the coaxial electrode structure of the cable. In most cases, the resulting stresses are high enough that they cause dielectric degradation of the materials at the edge of the shield, unless steps are taken to reduce that stress. The concentration of electric stress is now located at the conductor and end of the insulation shield. The stress lines are the horizontal lines that curve upward at the end of the shield (identified by the arrow in Figure 5-2), and the flux lines are the curved lines that are at right angles to the stress lines. The stress lines are more closely spaced near the conductor, and the flux lines more closely spaced at the end of the shield. These forces are strong enough to actually decompose the factory insulation at that interface and ultimately cause the cable insulation to fail. The stress at the insulation shield remains great because the electrical stress lines converge at the end of the shield. The equipotential lines are closely spaced at the shield edge. If those stresses are not reduced, PD can occur. Electrical stress relief is required in most medium-voltage applications.

5-2

Splicing and Terminating

5.2.2 Stress Cones To produce a termination of acceptable quality for long life, it is necessary to relieve voltage stresses at the edge of the cable insulation shield. The traditional method of doing this was to use a stress cone to control the capacitance in the area of high electrical stress (see Figure 5-3) [14]. Another method of stress control most often used for splicing and terminating in plants is the high-K material described in Section 5.2.3, Voltage Gradient Design.

Figure 5-3 Termination of an Insulation Shield with a Stress Cone [14]

A stress cone increases the spacing from the conductor to the end of the shield, as shown in Figure 5-3. This spreads out the electrical lines of stress as well as providing additional insulation at this high stress area. The ground plane gradually moves away from the conductor and spreads the dielectric field, thus reducing the voltage stress per unit length. The stress relief cone is an extension of the cable insulation. Another way of saying this is that the electrostatic flux lines are not concentrated at the shield edge as they are in Figure 5-2; it follows that the equipotential lines are also spaced farther apart. Stress cones can be taped by hand or premolded. Terminations that are taped achieve this increase in spacing by creating a lapped conical configuration of tape followed by a semiconducting layer that is connected electrically to the insulation shield, as shown in Figure 5-3. Premolded stress cones use the same concepts in the construction. The classic approach to the design of a stress relief cone is to have the initial angle of the cone be a few degrees and take a logarithmic curve throughout its length. This provides the ideal solution, but it was not usually needed for the generous dimensions used in medium-voltage cables. There is such a little difference between a straight slope and a logarithmic curve for medium-voltage cables that, for hand build-ups, a straight slope is acceptable. Premolded designs usually maintain that logarithmic shape. In actual design, the departure angle is in the range of 3° to 7°. The diameter of the cone at its greatest dimension has generally been calculated by adding another insulation thickness to the diameter of the insulated cable at the edge of the shield; therefore, at the maximum diameter of the stress cone, the insulation thickness is twice that of the cable’s insulation. A major disadvantage of such stress cones is that they require much more space between cables than the voltage gradient types that are described in Section 5.2.3, Voltage Gradient Design.

5-3

Splicing and Terminating

5.2.3 Voltage Gradient Design Electrical stress relief can come in different forms. A high-permittivity material (high dielectric constant or high K) can be applied over the cable end, as shown in Figure 5-4 [15]. When materials with different permittivities are subjected to a voltage gradient across their combined thickness, the material with the lower permittivity is subjected to the highest stress. The high K material over the shield maintains the radial voltage gradient in the insulation. The equipotential lines emerge only gradually from the insulation, thus producing a stress gradient, as shown in Figure 5-4. This material can be represented as a long resistor connected electrically to the insulation shield of the cable. By having this long resistor in cylindrical form extending past the shield system of the cable, the electrical stress is distributed along the length of the tube. Stress relief is thus accomplished by using a material with a controlled resistance or capacitance. These are available in cold-shrink, heat-shrink, and hand-taped designs. Other techniques can be used, but the basic concept is to use a material with a high resistance, high dielectric constant, or nonlinear current and voltage characteristics to extend the lines of stress away from the edge of the cable shield.

Figure 5-4 Stress Relief with High Dielectric Constant or High Resistivity Materials

Capacitive-graded materials usually contain particles of silicon carbide or oxides of aluminum, zinc, or iron. Although they are not truly conductive, they become electronic semiconductors and have identical stress relief to that of a stress cone. They do not have a linear E = IR relationship, but rather produce a voltage gradient along their length. One of their useful features is that the diameter is not increased to that of a stress cone. This makes them valuable for use in confined spaces. This voltage gradient does not depend on the IR drop but on an exchange of electrons from particle to particle. Resistive-graded materials contain carbon black, but in proportions that are less than the semiconducting materials used for extruded shields for cable. They also provide a nonlinear voltage gradient along their length. With proper selection of materials and proper compounding, these products can produce almost identical stress relief to that of a stress cone. A termination such as the one in Figure 5-4 obviously will fit in a smaller space than the stress cone design shown in Figure 5-3.

5-4

Splicing and Terminating

5.3

Splices for Shielded Cables

Splices must be electrically as strong as the cables that they join. Insulating materials with different permittivities and dielectric strengths are used in combination, and care must be taken to avoid overstressing the weaker materials [6]. Splices must also dissipate the heat generated in them without creating a hot spot, even though the path for heat generation is usually greater than that of the cable. Applying twice as much insulation, for instance, is not desirable because heat will be held in the splice, and the temperature at the conductor will increase. Splices rated at 5 kV or higher must consider the effect of electrical stresses by maintaining smooth surfaces throughout their design, manufacturing, and installation. Connectors must provide a current path that is as great as the cable conductor and must have as streamlined a contour as practical. Care must be exercised to remove all protruding points and edges because all sharp points produce high electrical stress at their tips (see Figure 5-5) [16]. Each manufacturer has specific instructions and dimensions for their splices and terminations. The removal of the semiconducting shield from the factory insulation at a splice or termination is critical to success. The semiconducting layer must be scored with a straight cut, only part way through the semiconducting layer. Cuts into the insulation wall are not permitted and must be removed by starting the process over. The cut must be circular and not leave any points of semiconducting material projecting into the surface of the insulation. Either of these errors will reduce the service life of the cable system.

Figure 5-5 Projection from Semiconducting Layer and Cut into Insulation

5.3.1 Cable Preparation for Splices and Terminations An important step in the splicing or terminating process is the preparation of the cable. Improperly prepared cable ends provide initiation sites for failures. Over the past 30 years, many newer materials and methods have been introduced for splicing and terminating cable, such as premolded, heat-shrink, and cold-shrink materials. The older materials do not have to be replaced, but when new cables are to be installed, the newer types should be considered because of their advantages.

5-5

Splicing and Terminating

Medium-voltage cables with extruded insulation are prepared for splicing and termination in a similar manner, regardless of the type of insulation. The entire cable must be cut to the proper length with a straight cut. The jacket, metallic shield, and semiconducting layer must then be removed to the proper dimensions to fit the termination being used. Tools are available to score the semiconducting layer for removal (see Figure 5-6). Spiral or longitudinal strips are scored partway through the semiconducting layer but not into the insulation. These narrow strips are then removed one at a time.

Figure 5-6 Two Types of Semiconducting Layer Scoring or Stripping Tools

Cables manufactured in the 1970s occasionally had small bits of semiconducting layer remaining on the surface of the insulation when the strips of the semiconducting layer were removed. These had to be removed by abrading the surface. For these cables, some light sanding might be required to remove the small chunks of the semiconducting material that adhered to the insulation surface. A fine, nonconducting, nonmetallic abrasive paper, such as 400–600 grit, should be used to polish the insulation surface and remove these particles. With present manufacturing technology, there should be no reason to sand or abrade the exposed insulation. Care must be taken to avoid cuts or scrapes, which are places for moisture to be trapped or for elevated electrical stress to occur that can lead to failure. One critical location for cuts is where the semiconducting material is removed. A cut into the insulation at that point is not easily seen. If left there, a cut creates a void that is likely to cause an early failure, especially in XLPE insulated cables. Cuts, scrapes, or gouges cannot be corrected by sanding off the defect or by reducing the insulating wall thickness. Serious errors might require that the cable end be re-cut and preparation restarted. The acceptance criteria for cable end preparation depend on the methods and materials used to construct the splice or termination. Common requirements include a cable insulation surface that is free of contamination, imperfections, and damage. A smooth surface for extruded dielectric insulations minimizes contamination and moisture adhering to the surface. If a rough surface remains, it must be made smooth.

5-6

Splicing and Terminating

Connections to the metallic shield of the cable must not damage the underlying cable components. Cutting into the strands of wire—or, of even greater importance, into a solid conductor—cannot be tolerated. Figures 5-7 through 5-11 illustrate some of the steps involved in cable preparation.

Figure 5-7 Measure Carefully to Achieve the Proper Length of Cable

Figure 5-8 Slip Applicable Sleeves over One or Both Cable Ends Before Installing Connector

Figure 5-9 Connectors of Several Lengths and Diameters for the Same Conductor Size

Figure 5-10 Shear-Bolt Connectors for Copper Conductor

5-7

Splicing and Terminating

Dampen a clean, lint-free cloth with the solvent, start the wipe on the exposed insulation,and wipe toward the connector.

Figure 5-11 Clean the Insulation, Using an Approved Solvent

5.3.2 Hand-Taped Splices Splices were traditionally made in the field, using hand-applied tapes, as shown in Figure 5-12. Tape splices are still used when large variations in conductor sizes must be joined or space limitations preclude using other types of splices. The medium-voltage cable’s insulation in Figure 5-12 is cut to a taper (or penciled) in this hand-taped splice. The connector ends are also tapered. This provides a longer creepage path over the insulation’s surface, as well as allowing the insulating tape to fill the space over the connector and its semiconducting layer without creating a void. The cable must be prepared and the connector installed as shown in Figures 5-12 through 5-15.

Figure 5-12 Semiconducting Tape Is Applied over the Connector to Form a Smooth Interface

5-8

Splicing and Terminating

Figure 5-13 Insulating Tape Is Applied Until the Proper Thickness Is Achieved over the Connector

Figure 5-14 Another Layer of Semiconducting Tape Is Applied over the Insulating Tape

5-9

Splicing and Terminating

Figure 5-15 A Metallic Braid Is Installed, a Ground Strap Is Attached, and the Jacket Tape Is Installed over the Entire Splice Area

Advantages of taped splices are the following: 

They can be used to join cables of different sizes and types.



They can be used in tight confines that require the connection to be bent.

Disadvantages of taped splices are the following: 

They require the most time to complete.



They are the most complex to make.



They require a highly trained and skilled splicer.



They have the least margin for error.

5.3.3 Premolded Splices Premolded splices (see Figure 5-16) are made at a factory and are electrically tested on a mandrel before shipment. The cable ends are prepared and the premolded device is slipped over one end before the connector is installed. The connector is encased in a shielded cavity, so that the splicer need not form a smooth layer over the connector. An overall jacket is required to protect the shields and ground straps from corrosion or mechanical damage.

5-10

Splicing and Terminating

Figure 5-16 Cutaway of a Premolded Splice

Research sponsored by EPRI [17] showed that premolded splices should be selected so that they fit tightly over the cables. Manufacturers show a range of sizes on each splice that the particular molded device will accommodate. When cables are spliced in confined spaces, bending can cause gaps between the premolded splice and looser-fitting cables; these gaps can allow moisture to enter and cause a failure. The solution is to select premolded devices that are on the tighter side of the manufacturer’s size range, even though splicers might not appreciate the extra effort required to assemble the splice. Advantages of premolded splices are the following: 

They require the least time to complete.



They require the least skill to assemble.



They have the greatest margin for error, yet still work.



They are the least complex to make.

Disadvantages of premolded splices are the following: 

Cable sizes must be close to the same on both sides.



The premolded device must be a tight fit, causing it to be hard to push on.



Premolded devices might require stocking several sizes because they are not capable of fitting over a wide range of cable sizes.



Proper connectors and dies must be used. If the crimping process increases the diameter of the flashing of the connector and it becomes too large, the inside of the molded rubber can be damaged when slid into place.

5-11

Splicing and Terminating

5.3.4 Cold-Shrink Splices During the manufacture of a cold-shrink sleeve, the rubber tube is stretched and a helically perforated nylon tube, which holds the rubber tube in the stretched condition, is inserted in the end of the perforated tube and pulled back through the center of the tube. When installing the splice, the cable ends are prepared, the cold-shrink sleeve (or sleeves) is slipped over one end, and the connector is installed. The splicer positions the sleeve in its proper place and then pulls out the nylon strip to allow the sleeve to shrink onto the cable, as shown in Figure 5-17. Figures 5-18 and 5-19 illustrate some of the steps for using cold-shrink splices.

Figure 5-17 Cold-Shrink Splice, Showing the Direction of the Shrinking Process

Figure 5-18 Carefully Position the Housing Before Removing the White Core Support

5-12

Splicing and Terminating

Figure 5-19 Continue Removing the Core While Holding Its Position on the Cable

Advantages of cold-shrink splices are the following: 

They fit over a wide range of cable sizes.



They require less skill to install than other types.



They require less time to complete than other types.



They are the least complex.



They are moderately forgiving.

The disadvantage of cold-shrink splices is that they require time to fully shrink in cold weather. 5.3.5 Heat-Shrink Splices During the manufacture of heat-shrinkable splices, the tube elements are extruded and crosslinked by irradiation. The tubes are then heated, stretched, and held in place while they cool. The tubes remain in their stretched shape until they are reheated, at which time they shrink diametrically to form the splice elements. The material used for these elements is cross-linked polyolefin, which is thermoplastic and re-formable. The splicer prepares the cable ends, as shown in Figure 5-20 through 5-27, and slides all the necessary tubes over the ends of the cable before installing the connector.

5-13

Splicing and Terminating

Figure 5-20 High-Permittivity Mastic Material Is Placed over the Connector and Conductor, Without Concern for Smoothness

Figure 5-21 The Semiconducting Tube Is Slid into Place and Heated Until Properly Shrunk Down to the Cable

Figure 5-22 An Insulation Tube Is Slid into Place and Shrunk Down

Figure 5-23 A Tube That Is Both Insulating on the Inside and Semiconducting on the Outside Is Positioned and Shrunk into Place

Figure 5-24 Tinned Copper Braid is Wrapped Around the Splice to Replace the Metallic Portion of the Insulation Shield

5-14

Splicing and Terminating

Figure 5-25 A Ground Strap Spring Is Placed Under One Side of the Cable’s Taped Metallic Insulation Shields

Figure 5-26 The Ground Strap Is Placed Across the Splice and Connected to the Factory Metallic Shield on the Opposite Side of the Splice

Zipper Type

Tube Type

Figure 5-27 An Overall Rejacketing Tube Is Placed Around the Entire Area

A heat-shrinkable jacket is available as a tube or as a rectangular sheet with a metallic zipper. After the jacket is wrapped around a completed splice, the zipper is closed to form a tube. The tube is positioned and shrunk down to the proper size. These jackets are also used as a repair sleeve that can be installed without cutting the conductor or cable. Advantages of heat-shrink splices are the following: 

They have a wide size range.



They require a moderately skilled splicer.



They are qualified for nuclear Class 1E.

5-15

Splicing and Terminating

Disadvantages of heat-shrink splices are the following: 

They must be shrunk fully to eliminate voids.



They require more time to complete.



They are complex to make, but less complex than a taped splice.



They have a low margin for error, but they are more forgiving than other splices.



They are difficult to heat properly in tight locations, such as near a manhole or cubicle wall, and they might not shrink fully.

5.4

Terminations

Terminations are available in the four types: taped, premolded, cold-shrink, and heat-shrink. They are similar to the splices with respect to installation practices (see Section 5.3, Splices for Shielded Cables). Cold-shrink and heat-shrink type terminations with skirts (see Figure 5-28) can be used indoors, as long as there is sufficient space for the skirts. Having the additional creepage length might be desirable because rain is not available to clean dust and contaminants from terminations in cubicles.

Figure 5-28 Cold-Shrink Outdoor Termination with Rain Sheds or Skirts

The premolded types of terminations (see Figure 5-29) are known as separable connectors that are useful in power plants. They are not exactly the same as those used in the distribution systems because distribution generally uses the load break type. They both consist of a semiconducting shield over the metallic connector system and insulating material, and they have an overall grounded surface. They are used to provide the connection between the cable and the electrical compartment of a transformer, switch, or other device.

5-16

Splicing and Terminating

An advantage of these devices is that, while they are de-energized, they can be taken apart at the motor for testing purposes, if there is sufficient space in their connection box. Taped terminations are destroyed during the disconnection process and require significant time to rebuild.

Figure 5-29 Dead Break, Premolded Termination Courtesy of Elastimold

The insulating portion of the elbow is made of EPDM rubber with an outer covering of similar materials that contain carbon black to make them semiconductive. Other termination types that are used to connect medium-voltage feeder cables to motors, switchgear, or transformers include stub-type (or V-type) and in-line type kits. For stub- or V-type kits, stress-control tubing is installed on shielded cable; compression lugs are installed on each cable and bolted together as shown in Figure 5-30. A variety of shielded cable designs can be accommodated. Sealant and insulating material complete the installation. Because of the many cable sizes that can be involved, these kits do not contain the required hardware.

Figure 5-30 Stub-Type Motor Connection Courtesy of Tyco

For in-line type kits, compression lugs are bolted together in a similar manner as for stub-type kits, but they are connected end-to-end like a normal splice (see Figure 5-31).

Figure 5-31 In-Line Type Connection Courtesy of Tyco

5-17

Splicing and Terminating

5.5

Lugs and Connectors

The electrical connection used to connect the cable in a termination to another electrical device must be considered. This connector, generally called a lug, must be able to carry the rated current of the cable, must provide good mechanical connection to keep from coming loose and creating a poor electrical connection, and must seal water from the cable core. All terminations must keep water out of the conductor strands. Many early connectors were made of a flattened section of tubing that had no actual sealing mechanism, and water could enter along the pressed seams of the tubing. Sealing can be accomplished by filling the space between the insulation cutoff and lug base with a compatible sealant or by purchasing a sealed lug. Connectors in a taped splice must perform the functions of the lugs, but they must have a smooth surface. For medium-voltage cable splices, connectors with tapered edges are recommended. Indents made during compression connections must be smoothed out with semiconducting tape or special stress-relieving mastics. Proper techniques for bolting terminations to a bus or lug are described in the EPRI report Electrical Connector Application Guidelines (1003471) [18].

5.6

Installation Considerations

5.6.1 Connecting the Conductors Cable conductors are generally either copper or aluminum. Copper is a forgiving metal in a splice, and many methods of connecting two copper conductors together are possible, including soldering, compression, welding, and heat fusion. Aluminum conductors are relatively rare in power plant applications; however, when they are used, special connectors and practices are required. Great care must be taken to match the compression tool, die, and connector with one other for aluminum conductors. As conductor sizes approach 1000 kcmil (507 mm2), these concerns must be addressed more completely. One of the facts involved in the larger-size conductors is that, on plant systems, they are the cables that are most prone to extended periods of high-temperature operation. The operation of the connector must be stable throughout load cycling and be capable of carrying the maximum amount of current without causing thermal degradation of the joint. The connector metal should be the same as that of the conductor, when possible. There are situations in which this cannot be done, such as when a copper conductor is to be connected to an aluminum conductor. It is acceptable to use an aluminum connector over a copper conductor, but a copper connector must not be used over an aluminum conductor because, during load cycles, the relative rates of expansion of the two metals causes the aluminum to extrude from the lug and results in a high-resistance connection. The shape of the connection is always of importance if the connection is not in a shielded area such as exist in all premolded splices. To minimize voltage stress at the connection for all those other conditions, special connectors are required for medium- and higher-voltage cables. Tapered shoulders and filled indents are required for these connectors. Semiconducting layers are almost

5-18

Splicing and Terminating

always specified over these connectors. When crimping a connector, the crimping tool must be rotated around the connector about 90° after each crimp to maintain a straight configuration. Parallel crimps result in a banana-shaped connection (see Figure 5-32). The flashing that results on the connector can score the interior of a premolded connector.

Figure 5-32 Improperly Crimped Connector

5.6.2 Insulation for Splices In general, splices should be made using kits, and the vendor’s instructions should be followed explicitly. The wall thickness of the splice covering and its interfaces with the cable insulation must safely withstand the intended electrical stresses. In premolded devices today, the thickness is usually about 150% of the thickness of the factory insulation. The splice insulation for handtaped splices generally uses self-amalgamating tape that is currently made from EPR but was originally made of butyl rubber or PE with thermal properties that matched the cable insulation. Premolded splices are almost always made of EPDM rubber compounds. Heat-shrink splices are made of polyolefin compounds that can be expanded after being cross-linked using irradiation. 5.6.3 Semiconducting Insulation Shield Materials for Splices and Terminations Splice semiconducting shields, like cable shields, perform the task of electrical shielding by having a considerable amount of carbon black (about 30%) compounded into the material to form a semiconducting path. These materials must be compatible with the rest of the cable, as well as having adequate conductance to drain the electrostatically induced voltages, charging currents, and leakage currents. 5.6.4 Metallic Insulation Shield for Splices and Terminations The metallic portion of the insulation shield system must be capable of carrying any fault current across the splice. This is generally done by using two paths. The first is a copper braid or mesh that is placed next to the semiconducting layer. For small amounts of fault current, a solder bead is run over this braid. However, a more robust fault current path is usually accomplished with wires, such as those from the wire shields of a cable.

5-19

Splicing and Terminating

5.6.5 Jackets for Splices Jackets over splices must provide physical strength, seal against entry of moisture into the splice, and resist chemical and other environmental attacks. It is important to use a jacket over the splice when jacketed cables are spliced together because corrosion of metallic neutrals or shields can concentrate at the joint.

5.7

Selection of Splices and Terminations

When making a decision as to the best choice of devices to purchase, the following questions and opportunities should be considered: 

Are the components of the device compatible with the cable being spliced or terminated?



Has the device passed all the tests that were specified, so that it meets the requirements of the electrical system involved?



Are codes applicable in the decision to use the chosen device?



Have all safety requirements involved in the construction, application, and installation of the device been considered?



Will the device meet the mechanical requirements of the installation?



Can the device be assembled with the tooling that is already available, or are special tools required?



Is the splice qualified for its intended environment (such as a harsh environment or wet conditions)?



Has the positioning of the device been considered? Splices are not recommended for installation at bends in the cable.



Are there any existing work practices or procedures that will conflict with the application of this device?



Will the device do the job as well as or better than what is presently used?



Skirted (outdoor) terminations are normally installed in an upright position. Other positions are possible but require special attention.



Environmental conditions are of importance to attain the expected life of any device. Heat can affect the ampacity of the device. Cold can affect the assembly. Eliminating contaminants is critical to preventing unwanted leakage paths within a termination.



Moisture is always the enemy of an underground system and must be controlled in construction and installation.



Consider using separable connectors for ease of disconnecting for future diagnostic testing and troubleshooting.

Careful consideration in selecting terminations and splices is important to ensure long-term, reliable operation and to accommodate future needs, such as periodic diagnostic testing and troubleshooting should a failure occur. 5-20

6

FUNDAMENTALS OF CABLE INSULATION SYSTEMS Cable insulation systems include the primary insulation system (EPR, XLPE, and butyl rubber for nuclear plants) and the conductor and insulation shields. The shields equally distribute voltage stresses within the insulation and prevent PD between the conductor and insulation and between the insulation and the metal insulation shield. This section describes the properties of the insulation and shields and the issues that could affect aging. This section also contains a chronology of improvements made to the insulation and shielding system since the late 1960s. This chronology does not imply that cables manufactured for nuclear plants between 1968 and 1990 will have short lives; rather, it indicates that newer cables should be expected to have lives much longer than those of older-generation cables. More information on insulation materials, formulations, details of chemical processes in manufacture, and aging of insulation polymers is provided in the appendices.

6.1

Primary Insulations

Two types of insulation systems have been used in nuclear plant cables—rubber (elastomeric) and semi crystalline plastics. Ethylene-propylene rubber and butyl rubber are elastomeric, and XLPE is a semi crystalline plastic. Plant designers had a choice to make when determining the type of insulation to use in a power plant. Rubber materials possess more flexibility for installing cable in the confines of a plant and can make installation a bit easier. They can also have longer lives but are generally more costly. XLPE cables are stiffer but have lower electrical losses than rubber-insulated cables, and, at the time of installation in the 1970s, they were thought to have superior capabilities under wet conditions.

6.2

Elastomer Basics

Polymers themselves can be classified as rubbers, plastics, resins, or fibers. Elastomers are polymers that are inherently rubbery and flexible by nature. Elastomers are produced by joining (polymerizing) small molecules (monomers), and converting them into large, long-chain molecules. The process leads to elastomers with high molecular weight. When many of the same monomers (such as ethylene) are joined to others, a homopolymer is formed. If two different monomers such as ethylene and propylene are used, a copolymer such as EPR results. These polymers are called chains—the longer the chain, the higher the molecular weight and the better the properties. Elastomeric chains are not entirely linear and will possess branches. Elastomers, being inherently soft, require inorganic mineral fillers to be useful as insulation. The inorganic fillers improve their strength and make them firmer. Additional details on fundamentals of elastomer technology are provided in Appendix D.1.

6-1

Fundamentals of Cable Insulation Systems

6.2.1 Cross Linking The long elastomer molecules are mixed together, as in a bowl of spaghetti. The “noodles” may have branches, but they are not connected to each other. When the different chains are joined together, this linkage is called the cross link (see Appendix D.2). Cross-linking has numerous beneficial effects. The resulting polymer insulation becomes tougher, resists softening at elevated temperatures, and maintains form stability at elevated temperatures. These changes are particularly important for elastomers such as ethylene copolymers, as compared to homopolymers such as PE because elastomers lack crystallinity, making them soft. Regardless, property improvements induced by cross-linking of the elastomer alone are not adequate for a butyl or EPR cable, and additional additives are needed to create a useful insulation. Cross linking does not improve the electrical properties (dielectric constant or dissipation factor). It only improves strength and form stability of the polymer. 6.2.2 Fillers Used in Rubber Insulations Fillers (inorganic chemicals) are needed in elastomeric insulations to provide structural strength and stability. The nature of the filler additives used in the elastomer blend varies with the elastomer type. A general overview is provided here; specific additives are described with the different polymer types. Typical additives used in an elastomer for wire and cable insulation are the following: 

Inorganic fillers such as clays that have undergone various treatments. This type of filler improves structural integrity of the polymer. Fillers influence stiffness, which in turn influences abrasion [19, 20].



Plasticizers that are used to modify the physical characteristics of the wire coating or the viscosity of the compounded rubber (before extrusion)—often called softeners. Plasticizers cause polymers that are normally rigid to become flexible and stretchable.



Metal oxides that serve as heat and/or moisture stabilizers.



Curing or vulcanization agents. These agents promote the cross linking desired during the curing process.



Co-curing agents—chemicals that facilitate curing.



Antioxidants or antiozonants—chemicals that retard aging.

The inorganic clay component (also referred to as kaolin) is significant, and its nature requires description [21]. Clay is an inorganic mineral composed primarily of aluminum silicate, with trace amounts of other metal oxides and impurities. After mining, it is washed, ground, and cleaned to remove impurities and adjust particle size. Water of hydration is present, and clays are then heat treated (calcined) to reduce the water content from ~14% to ~1%. This thermal treatment takes place at more than 1500°F (815°C). Calcined clays were used in the past to reinforce (strengthen) butyl rubber. Clays can also be treated with functional silane chemicals, which serve to bind the calcined clay with the elastomer. This technology is used with EPR. Experience has demonstrated that as one moves from using hydrous clay to calcined clay to silane-treated clay, the properties of the 6-2

Fundamentals of Cable Insulation Systems

elastomer improve; the tensile strength, modulus, volume resistivity, and breakdown strength all increase. The silane also serves to remove remaining water and provides the capability to link between the polymer and the clay particles. Clay minerals have the ability to adsorb cations on their surfaces [22]; therefore, silane treatments prevent the ions from participating in water treeing. Several different types of silanes can be used. The modification of the clay, a major component of EPR formulation, is a complex process. The term compounding is used to describe the mixing of all the materials in an EPR formulation, and the EPR that is blended in the soon-to-be cable insulation is called a compound. 6.2.3 Crystallinity Crystallinity refers to chain alignment, a tendency of unbranched polyolefins to form structures that are impermeable and exclude impurities [23]. In general, elastomers, including butyl rubber, are not crystalline, but some types of EPR used for wire and cable can possess a low level of crystalline regions as a result of high ethylene-to-propylene comonomer ratios. Branches inhibit crystallinity, as they hinder the tendency of chains to align. To place the subject of crystallinity in perspective, the highest level of crystallinity in an EPR wire or cable insulation material is perhaps 6% to 8%, whereas the level of crystallinity for low-density PE is approximately 50%. In general, the level of crystallinity is low in the EPRs used in medium-voltage cable. Crystallinity plays a bigger role in XLPE insulation (see Section 6.6, Cross-Linked Polyethylene). 6.2.4 Cable Conductor and Insulation Shields Although 5-kV cables might or might not use insulation shields, 8-kV rated cables are generally shielded, and all 15-kV cables are shielded. Modern standards require insulation shields at 8 kV and are tending to require shields at 5 kV. All the medium-voltage cables in use in nuclear plants have conductor shields (shields at the interface between the conductor and the inner wall of the insulation). Different generations of medium-voltage cables have different types of shields. Early generations of nuclear plant cables had cotton tape conductor and insulation shields. Later conductor shields were formed of extruded polymer. Cotton tape insulation shields were supplanted by polymer tape shields, which in turn were supplanted by extruded shields. In a modern cable, an insulation shield consists of an elastomeric (an ethylene copolymer) combined with semiconducting carbon black filler that imparts the semiconducting properties. The role of the cable shield is to provide a controlled stress gradient between the conductor and the insulation. The nature and amount of carbon black influences the conductivity of the shield layer; the carbon black must possess the ability to aggregate into small clusters (termed structure) that provide the conducting path within the elastomeric matrix. The shield must also facilitate a conductor–insulation intermediate interface that is smooth and defect free. The semiconducting shield materials are processed in the same manner as the insulation [23]. The semiconducting insulation shield connects the surface of the insulation to the metallic portion of the shield, so that air gaps do not exist that could ionize at operating voltage and cause PDs that would lead to early insulation failure.

6-3

Fundamentals of Cable Insulation Systems

Elastomeric cables with cable shields are intended to be discharge free, in the sense that they do not exhibit discharges that exceed a specific allowable level on testing—the exact maximum is defined in industry specifications. The requirements have forced lower and lower maximum allowable discharge levels as acceptance criteria became more stringent over the years.

6.3

Butyl Rubber

6.3.1 Material Description Butyl rubber is one of the earliest style cables available for use in nuclear plants, and only a few early plants have this type of insulated cable. Butyl rubber is one of several types of polymers classified as elastomers. All rubber-like materials, as noted, are polymers that are highmolecular-weight compounds consisting of long chains of one or more types of molecules (monomers). Butyl rubber (see Figure 6-1) is a copolymer of isobutylene monomer combined with a small amount (1% to 3%) of isoprene monomer [24].

Figure 6-1 Butyl Rubber Molecule

The isobutylene portion of butyl rubber provides the desired insulation characteristics, and the isoprene functionality is present solely to provide unsaturated sites (double bonds) along the chain for facilitating cross linking. Vulcanization produces chemical links between the loosely coiled polymeric chains; elasticity occurs because the chains can be stretched and the cross links cause them to retract when the stress is released. The major portion of the butyl polymer consists of saturated hydrocarbon chain lengths that impart oxidation, ozone, chemical, and moisture resistance, as well as low gas permeability. A variety of commercial grades of butyl rubber exist. In addition, several curing techniques were used in the past to cure butyl for wire and cable application. 6.3.2 Butyl Wire and Cable Insulation A typical butyl rubber insulation used in the past possessed 0.6% to 1.0% (mol %) unsaturation, an average molecular weight of 350,000, a tensile strength of 2600 psi (17.9 MPa), and an elongation of 700%. This was referred to as Exxon or Enjay Butyl 035, “the most widely used butyl polymer for electrical applications” [25, 26]. However, other grades have been used. The required fillers in the final compound affect both processing and properties. In the past, claims for butyl rubber included “the ozone resistance of butyl rubber coupled with the moisture resistance of its essentially unsaturated hydrocarbon structure finds utility as a high quality electrical insulation” [27]. Butyl rubber has been surpassed by EPR and other, more modern insulations.

6-4

Fundamentals of Cable Insulation Systems

6.3.3 Fillers and Other Additives For butyl rubber, calcined clay was the filler considered to be “of the greatest value in power cable…and applications where the utmost in electrical moisture stability is required” [28]. Other useful inorganic fillers used in butyl rubber are talc (referred to as Mistron vapor talc), calcium carbonate, hydrated alumina, and carbon black. Additional additives were used to bond the carbon black and/or clay with the butyl rubber itself. These old butyl rubbers for wire and cable application did not use the silane-coated clays that are used today in EPR. Indeed, coated clays were developed for EPR when it was determined that the calcined clay that worked well with butyl did not provide the same wet stability when used in EPR [26]. The type of vulcanization used in production of butyl rubber insulation is important with respect to how it ages. Several basic methods have traditionally been used for butyl rubber. Sulfur or organic chemicals that possess sulfur have been used [24]. These agents interact with the unsaturation provided by the isoprene portion of the butyl rubber. Sometimes small amounts of sulfur were used in conjunction with sulfur-containing organic chemicals (referred to as accelerators). Another additive that might be present is called an accelerator activator, a chemical designed to increase the vulcanization rate during manufacture; examples of this chemical are zinc oxide and red lead. The literature suggests that (non-sulfur-based) curing agents were preferred for wire and cable insulation [27]. The stability of the sulfur-based cross links has been less than originally expected. The lack of stability allows thermal degradation to cause the material to soften to the point at which the conductor or shield can migrate through the insulation. Other common cross-linking agents for wire applications are referred to as dioximes and can be used in place of sulfur-containing chemicals. These are converted to dinitrosobenzene, the actual cross-linking agent, which interacts with the unsaturated portion of the butyl rubber molecules. This technology has also been commonly referred to as a quinoid cure, and it is referred to in older literature as being used in wire formulations. In essence, if the system maintains properties (no softening) at elevated temperature, it is considered to be a better cure than sulfur-based cures. Appendix D.3 presents the details of some typical butyl rubber compounds with typical properties. The dielectric strengths for these compounds range from 550 V/mil to 695 V/mil (21.6 kV/mm to 27.3 kV/mm).

6.4

Ethylene-Propylene Rubber

6.4.1 Material Description EPRs are copolymers prepared by polymerization of ethylene and propylene monomers. The ratio of ethylene to propylene can vary widely. This leads to a wide range of commercially available EPR grades. In practice, there is a practical range for the copolymer ratio of ethylene to propylene that is useful for wire and cable insulation. This is the weight percentage from 50:50 to 75:25 of ethylene to propylene. The 50:50 ratio for EPR is completely amorphous, and the 75:25 composition ratio possesses a slight degree of crystallinity.

6-5

Fundamentals of Cable Insulation Systems

Propylene monomer differs from ethylene in that it possesses a -CH3 linkage in place of one -H linkage. When propylene is copolymerized with ethylene, the resulting EPR copolymer is rubbery and soft, can be deformed, and has relatively high elongation. EPRs possess many short chain branches; this is what imparts the rubbery or elastomeric quality. In the absence of the bulkiness provided by the repeating -CH3 units, the PE homopolymer has a tendency to crystallize. That is why the 75:25 ratio for ethylene to propylene is slightly crystalline and the 50:50 ratio is not.

Figure 6-2 Copolymer of Ethylene and Propylene

In addition, it is not necessarily uncommon to use a third monomer in an ethylene-propylene formulation. See Appendix D.4, Ethylene-Propylene Rubber and Ethylene-Propylene-Diene Monomer. 6.4.2 Ethylene-Propylene Rubber Cross Linking As with butyl rubber, EPRs must be cross linked to be useful as insulations, and they must also be compounded with a series of additives, including clay. The cross-linking technology for commercial EPRs involves the use of peroxides. The most common peroxide—dicumyl peroxide—is incorporated into the EPR before the polymer is extruded [29]. At the temperature range of the extrusion process (which melts the pellets and converts the melted pellets into cable insulation), the peroxide is unaffected. However, after the extrusion process is completed, the now-formed cable (insulation with the conductor and shields) is passed through a long tube that is at a higher temperature and pressure. This is referred to as a curing tube or continuous vulcanization tube. Under those conditions, the peroxide decomposes and converts the spaghettilike, mineral-filled polymer into the gelled network. In addition, this method leads to a number of by-products that result from the intended decomposition of the peroxide, which remain in the insulation wall and migrate out over time. Methane is formed and diffuses rapidly. For additional information, see Appendix D.5, Dicumyl Peroxide Cross-Linking Agent Byproducts.

6-6

Fundamentals of Cable Insulation Systems

Cross linking through the use of peroxides imparts improved properties; the resulting elastomer insulation becomes tougher, resists softening at elevated temperatures, and facilitates maintenance of form stability at elevated temperatures. This is particularly important for elastomers (such as ethylene copolymers) as compared to homopolymers such as PE. Cross linking the polymer chains improves strength and structure, but it does not improve the electrical properties such as dielectric constant, dissipation factor, and dielectric strength. 6.4.3 Fillers for Ethylene-Propylene Rubber Insulation EPR must be modified with additives to render it useful as insulation. As with butyl rubber, many additives are used; some are the same, and some differ. Silane-treated clay is a common filler used to impart strength and structure. In addition, depending on the nature of the EPR (the ratio of ethylene to propylene) and the nature of the clay, the amount of clay can vary widely. Fillers can constitute a majority of an EPR compound, but the individual components and their natures vary between suppliers. The filler nature and concentration will vary depending on whether the EPR is black, pink or brown, and can vary even within an EPR formulation that is the same color (from the same or different suppliers). The exact components used in some EPRs are proprietary. The color of any EPR cable insulation is related to the components used by the supplier. Because EPRs are used for many applications beyond wire and cable, scores of compound formulations have been published. Because the individual components influence aging and reliability, the following section focuses on those used for wire and cable. 6.4.4 Compounding (Mixing) of Ethylene-Propylene Rubbers Mixing the required ingredients properly represents a separate technology for elastomers. Such mixing is often performed in a Banbury mixer, which involves a batch process that uses heating to ensure good dispersion while removing volatiles that could cause harmful porosity. The overall Banbury mixing process combines the elastomer with other required ingredients and produces a homogenous blend. The ingredients are first weighed (either at the mixer or in prepackages) and passed along a conveyer belt to the mixer. A multi-step process then ensues. As an example, the unvulcanized rubber is passed through rolls at controlled temperatures, additives are incorporated, the mixed batch is passed through a mill for additional mixing, and more additives can be incorporated. Eventually, sheets are formed [30]. Key points in the overall process include ensuring a controlled temperature range (which differs for different rubber formulations, a specific order of additive entry, controlled mixing times, and proper temperature to prevent premature cross linking. Premature cross linking is called scorch and is a result of either too long a time at elevated temperature or actual temperature being too high (or both). Scorch is a practical concern, if present, and can impact the reliability of a finished cable. All the components for a satisfactory wire and cable insulation are mixed in this or an equivalent manner. A Banbury mixer is shown in Figure 6-3.

6-7

Fundamentals of Cable Insulation Systems

Figure 6-3 Banbury Mixer Used for Preparing Ethylene-Propylene Rubber Compounds

In contrast to batch processing, continuous mixing has been applied using a Buss kneader. This continuous process claims the following advantages: lower temperature levels and precise temperature control during mixing, uniform shearing effect without temperature peaks, shorter residence times of the components in the unit, and short product changeover times. Newer, state-of-the-art elastomer technology uses either mixing method, depending on the cable manufacturer’s mixing capability (or their materials suppliers’ techniques). However, older cables (more than 30 years old) would have compounds blended with a Banbury mixer. 6.4.5 Shielded Cable Constructions for Medium-Voltage Ethylene-Propylene Rubber Cables for Plant Applications The shield accomplishes the following for modern cables [23]: 

It confines the electric field to the interior of the shield region.



It ensures symmetrical radial stress distribution within the insulation.



It eliminates longitudinal and tangential stresses along the insulation surface.



It protects cable from overvoltages by facilitating uniform surge impedance along the cable length.

Conductor shields used for this purpose in the past were composed of semiconducting tapes, helically wrapped over the conductor. Early fibrous tapes, such as cotton, did not provide smoothness at the insulation interface, leading to stress concentrations and insulation damage. Smoothness and cleanliness are critical parameters for shield materials. Tapes were later replaced with extruded shields, and this layer is now extruded directly over the conductor. The geometry of stranded conductors is such that air gaps can exist between the outer surface of the conductor and the inner surface of the insulation. Thus, without a stress control layer, electric fields can cause PDs within these gaps that would harm the insulation. The issue becomes more significant as the voltage rating of the cable increases.

6-8

Fundamentals of Cable Insulation Systems

There are two basic types of conductor shields: conductive and emissive shields. An emissive shield is one that uses a material with a high dielectric constant. However, the most popular type of shield is the extruded, semiconducting conductor shield containing carbon black. Cable types used for nuclear power plants have included 5-kV nonshielded and 5-, 8-, and 15-kV shielded constructions [28]. The tape or extruded shield 5-, 8-, and 15-kV constructions used the following: 

Stranded copper conductors



Extruded strand shields or tape strand shields



EPR insulations of various colors



Extruded insulation shield or tape insulation shields (fiber or polymer)



Helical copper tape shields



Jackets composed of CPE, neoprene, or CSPE

The 5- and 8-kV wire-shielded construction (UniShield) used the following: 

Stranded copper conductor



Extruded strand shield



Pink EPR



Semiconducting CPE jacket with an embedded wire shield

The 5-kV nonshielded constructions used the following: 

Stranded copper conductor



Extruded strand shield



EPR insulation



CPE, neoprene, or CSPE jackets

6.5

Historical Review of Medium-Voltage Cable Constructions

6.5.1 General Butyl rubbers were first commercialized in 1942 for tire inner tubes. Its use for wire and cable followed that, and it was used for nuclear plant (and distribution) cables into the late 1960s and early 1970s. PE and then XLPE replaced butyl for the distribution arena. When EPR was introduced in 1962 and commercial production began in 1963 [30], it quickly replaced butyl rubber for modern day cables for applications in which elastomers were preferred. Extruded thermoplastic shields were introduced after 1965 to avoid the use of tapes. From the 1970s

6-9

Fundamentals of Cable Insulation Systems

onward, a series of improvements were imparted to semi crystalline PE and XLPE systems, whereas EPR changes were less dramatic, but ongoing. Some of the changes to PE and XLPE, although focused primarily in the distribution area, are relevant for nuclear cables. These overall changes included the following: 

Movement from thermoplastic PE to XLPE to improve high-temperature properties



Movement from thermoplastic shields (used for PE) to shields with better high-temperature properties (to match the XLPE); first, deformation-resistant thermoplastic PE shields, and later, extruded cross-linked shields



Improved extrusion processes in which the shields were extruded concurrently with the insulation (to avoid contamination)



Development of TR-XLPE



Use of improved jackets

These changes were mostly directed toward distribution cables, but some affected nuclear cables, as well. One major industry change that took place in the 1980s for XLPE that had little impact on EPR was the shift in cable cross-linking technology from steam curing to dry curing. A major change in the late 1980s–1990s that impacted all shielded cable systems (nuclear and distribution) was the introduction of shield materials that possessed significantly reduced ion components. By that time, ions had been identified as contributing to water treeing and it was known that many harmful ions had been introduced into the cable system from the old carbon blacks used to manufacture the shields. While all these changes were taking place, EPR usage continued to grow steadily, and changes were also occurring for EPR. A major advancement for EPR technology occurred in the mid to late 1970s when a superior grade of EPR that could be extruded on the same type of equipment used for XLPE was provided; others followed shortly. This was one factor that facilitated movement from butyl to EPR. Other EPR changes were more subtle, described in terms of color (change in components of the formulations) rather than content, as the components of EPR formulations were considered proprietary, for the most part. 6.5.2 Ethylene-Propylene Rubber Types ICEA S-97-682-2007 [27] recognizes four classes of EPR: earliest EPR polymers (class I) and second-generation polymers—semi crystalline (class II); 221°F (105°C) rated compounds (class III); and discharge-resistant compounds (class IV). EPRs have often also been classified informally by their colors: black, gray, pink, and brown. The specifications ensure that the cables installed have met some minimum set of industry-accepted criteria and properties. The category breakdown allows one to recognize that 1) changes occurred over the decades, with each change intended to provide improved quality cable, and 2) all EPRs are not alike, independent of the changes made over the years. An EPR compound can contain as many as 10–20 different ingredients. (See Appendix D.6 for some EPR compounds.) Some suppliers of insulation compounds manufacture or have manufactured EPR compounds using commercially available materials that are sold to the different cable manufacturers, who perform the extrusion of the compounds and convert them into cable insulation. In this situation, different cable manufacturers might provide the same 6-10

Fundamentals of Cable Insulation Systems

insulation materials in their constructions, and the difference would be solely in the extrusion technology. However, some manufacturers prefer to mix their own EPR compounds; therefore, the specific components would differ from those of their competitors and be proprietary. In these cases, manufacturers provide their own unique cable systems and apply their own philosophy on how to ensure longevity and reliability. In the past and present, for applications in which elastomers were preferred, EPR cables have been informally classified by the color of the insulation layer. In most cases, these color classifications provide general indications of changes in formulation and capability. However, manufacturers have continued to make changes in formulations and improvements to extrusion practices that occurred well beyond the point at which the color of the polymer was changed. Although it is convenient to classify EPRs by color, the color classification does not provide complete information on the total compound package and the ultimate aging and life characteristics of the material. The following section focuses on the different EPR cables as provided at different times and by different suppliers based on color. Some information on aging is necessarily included; more information on aging of the different cables (butyl, EPR, and XLPE) is presented in Section 7, Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments. 6.5.2.1

Black Ethylene-Propylene Rubber

EPR elastomers were appealing as a butyl rubber replacement for a number of reasons, one of which is that they allowed for peroxide curing, thus avoiding the technology used for butyl rubber. They also opened the door to superior dielectric characteristics. The carbon black– containing EPR compounds were the earliest types and were manufactured from about the mid1960s until about 1971. These early EPRs were black because carbon black was used in butyl rubber formulations, and the practice was continued after butyl was replaced. Carbon black reinforcement was traditionally used, and it is generally accepted that the particle size, aggregate structure, and surface area of carbon black particles are important factors in reinforcing the EPR. It also imparts the necessary high tensile strength, good elongation, and hardness to the EPR. Electrical properties such as volume resistivity, dissipation factor, and dielectric strength, as well as moisture resistance, were enhanced in carbon black–filled EPRs. Carbon black also serves as a light-absorbing additive that inhibits chemical reactions that would lead to insulation degradation for cable sections that were outdoors. The proportion of carbon black to EPR was approximately 1 to 10 (10 parts carbon black to 100 parts EPR. Clay was also used in these EPRs as fillers and strengtheners. The proportion of clay to EPR was 110 to 100 parts. Typical EPR compound constituents are provided in Appendix D.6. The material properties of the black EPR are superior to that of the butyl rubber, in that the water absorption is less than one-tenth that of butyl rubber and the breakdown strength is significantly higher—one compound has 1380 V/mil (54.3 kV/mm), which is more than double that of the butyl rubbers.

6-11

Fundamentals of Cable Insulation Systems

Black EPRs were provided by Anaconda, Okonite, and General Cable. Some failures have occurred before the end of the 40-year design life, especially under wet conditions. NEI 06-05 [11] describes the experience with wet medium-voltage cables at nuclear plants. For example, in a plant that experienced 10 failures of early Okonite black EPR, it was concluded that most of the problem was associated with manufacturing defects (inclusions in the insulation) combined with severe operating conditions. Black EPRs were replaced by pink EPRs for new cables in the mid-1970s, so most of these failures occurred long after the black cable insulation and constructions were no longer being provided. Most of the failures appear to have been related to multiple issues, such as voids and inclusions, or installation damage coupled with wetting and somewhat higher susceptibility to water, rather than just the use of black EPR. Many of the black EPR cables that were installed in the early 1970s remain in service and are in satisfactory condition. 6.5.2.2

Pink Ethylene-Propylene Rubber

Pink EPRs became available in the mid 1970s (carbon black was removed) for EPR cables of greater than 5 kV. One reason noted [31] was to develop contrast between the insulation and the black semiconducting insulation shield. This change made it easier to judge the effectiveness of stripping the insulation shield from the insulation in the field by providing color contrast between the two layers. The change did not affect the properties. Another reason was noted by Zuidema [32], who reports that an Insulated Power Cable Engineers Association (IPCEA) requirement on wet electrical aging in 194°F (90°C) water also led to the change from black to other colors. The pink color is due to the component called red lead (Pb3O4) that was in the black EPR but was not observable until the carbon black was removed. DuPont provided several grades of EPR based on the ethylene to propylene ratio; both amorphous EPRs (~50:50) and slightly crystalline EPRs (~75:25) were available. Nordel 2722 was the semi crystalline grade, and Nordel 1040 was the amorphous grade. Both were compounded with appropriate additives, and the higher ethylene-to-propylene ratio system, after formulation with appropriate basic common additives, could be extruded on equipment used for the more crystalline XLPEs. The semi crystalline ethylene-propylene compound was referred to as Superohm 3728. A significant change was the reduced level of coated clay used. Schulman later commercialized the DuPont technology and provided the compound to cable manufacturers. DuPont eventually ceased supplying the base polymer and, at a later time, Schulman withdrew from the market. Electric Cable Compounds, Inc., then supplied the pink EPR market using base polymer supplied by Uniroyal or Exxon. Independently, and concurrently with these activities, some cable manufacturers developed and commercialized their own carbon-black-free pink EPR formulations. Typical medium-voltage EPR compounds used into the mid 1990s contained coated clays and zinc oxide, as well as typical ingredients common to all elastomers; details are shown in Table 6-1. These components are likely present in many cables installed today.

6-12

Fundamentals of Cable Insulation Systems Table 6-1 Typical Components in Medium-Voltage Ethylene-Propylene Rubber Compounds (Amounts Are Approximate) Amorphous EthylenePropylene Rubber (%)

Semi crystalline EthylenePropylene Rubber (%)

Nordel 1040

39

—

Nordel 2722

0

54

Low-Density Polyethylene (LDPE)

0

2.7

Zinc Oxide

2.0

2.7

Lead Oxide (Red Lead)

2.0

2.7

Silane-Treated Clay (Kaolin)

47

32.4

Vinyl Silane

0.4

0.5

Oil

5.8

0

Wax

1.9

2.7

Antioxidant

0.6

0.8

Dicumyl Peroxide

1.3

1.4

Component

Table 6-1 shows the approximate levels of the various ingredients in typical EPR formulations using different EPR insulation polymers. These compounds were typically used from the 1970s to 1990s. Each component has a specific purpose. Note also the presence of a small amount of LDPE in the semi crystalline EPR compound. The functions of the fillers and additives are the following: 

Zinc oxide improves heat aging.



Lead oxide maximizes wet electric stability.



Silane-treated clay optimizes physical properties. The silane coating on the clay improves interfacial contact with the polymer. This is necessary to prevent the formation of gaps or weak boundary layers that could lead to the formation of voids.



Vinyl silane contributes to improved interfacial contact between the clay and the ethylenepropylene.



Oil is a processing aid.



Wax is a release agent (keeps material from sticking in the extruder).



Dicumyl peroxide is a cross-linking agent.

All elastomers, regardless of the level of crystallinity, require compounding to properly mix the components. The major difference between the noncrystalline and semi crystalline elastomeric compounds is in the level of clay (see Table 6-1). The degree of crystallinity of a semi crystalline EPR is estimated to be in the vicinity of perhaps 5%. Calcined clays that worked so well for

6-13

Fundamentals of Cable Insulation Systems

butyl rubber were inferior for electrical stability in EPRs; this led to the application of silanecoated clays for EPRs. The clay improves the modulus of elasticity and the physical properties, such as tensile strength, tear resistance, or abrasion resistance, of the final product. The influence of silane coating of the clay in improving properties [21] is reviewed in Appendix D.7. 6.5.2.3

Brown Ethylene-Propylene Rubber

EPR cables having a brown insulation are manufactured by Kerite; again, the color is due to the specific components in the insulation formulation. The Kerite design for reliable cables differs from that of other suppliers; their approach for 5-kV cables and greater is to use an insulation composition that is formulated to provide superior discharge (corona) resistance. In other words, cables that use this technology are designed to be discharge resistant, not discharge free as were the pink and black EPRs. This technology [33] has been used by Kerite since the 1970s, and the components are proprietary. Kerite has also used a different shield concept; the stress control layer over the conductor, referred to as Permashield, is not semiconducting and possesses a high dielectric constant. Clearly, the different philosophy leads to differences in 1) ingredients in the formulations, 2) compounding technology, 3) extrusion and curing, and 4) aging response during operation.

6.6

Cross-Linked Polyethylene

6.6.1 Material Description The major difference between elastomers like EPR (or butyl rubber) and XLPE is that XLPE has a tendency to undergo crystallization. This means that the polymer chains have a tendency to align. The alignment, which is limited to unbranched sections or regions, imparts a degree of stiffness and toughness at ambient temperatures that is lacking in elastomers. The nature of the chain alignment can be depicted in several manners as shown in Figure 6-4.

Figure 6-4 Various Representations of Crystallinity in Polyethylene

6-14

Fundamentals of Cable Insulation Systems

The crystalline regions are the parallel lines in all depictions, and the noncrystalline or amorphous regions are represented by the curved lines or folds. The presence of crystallinity means that PE need not be cross-linked to be useful as insulation, but the presence of the cross links imparts slightly improved high-temperature properties and renders the semi crystalline polymer more useful as insulation. Also, the presence of crystalline regions means that it is not necessary to use mineral fillers or many of the other additives required in an elastomer recipe, but occasionally this has been done. From this fundamental perspective, several key points relevant to XLPE are the following: 

The degree of crystallinity can vary, depending on the degree of chain alignment.



Increased crystallinity means increased density for the PE and also a higher melting point.



Because high-density PE has greater crystallinity than LDPE, it is tougher, but it is also more challenging to extrude.



Crystallinity is diminished as the temperature is increased and completely disappears at a temperature related to the density; for LDPE, this is approximately 223°F (106°C).



Any inorganic mineral fillers present will reside only in the amorphous regions.



Cross linking of medium-voltage cables with XLPE insulation is induced by peroxides, the same technology used for EPR elastomers.



Peroxide-induced cross linking of XLPE takes place only in the amorphous regions; crystalline regions are melted during the cross-linking process and re-form after the cable is cooled.



Any impurities—for example, foreign contaminants, ions, or water—are located in the amorphous regions, and moisture transport takes place by movement through those regions.

For additional information, see Electrical Power Cable Engineering [23]. Some cable manufacturers used mineral fillers, such as clay, in their XLPE insulated cables. When filler additives are used, all the issues described in Section 6.1, Primary Insulations, for elastomers apply to this type of XLPE. The difference, as noted, is that the fillers are now inherently located in the amorphous regions of the XLPE (rather than uniformly dispersed throughout the matrix). In principle, mineral filler additives provide a back-up source for imparting toughness, certainly at elevated temperatures at which the crystallinity is reduced or absent. As with EPR, mineral fillers do not improve the electrical properties. Conventional XLPE was provided by many suppliers. Mineral-filled XLPE called Vulkene was provided by General Electric; later, the supplier provided a second grade of Vulkene, called Vulkene II. This technology was later made available by Raychem. As with butyl rubber, mineral-filled XLPE is no longer being provided commercially. Only one plant is known to have Vulkene insulation. Several Vulkene failures occurred under wet conditions at that plant. Table 6-2 lists the percentage of nuclear plants that use XLPE insulated cables, by the voltage rating of the cables.

6-15

Fundamentals of Cable Insulation Systems Table 6-2 Percentage of Nuclear Plants with Cross-Linked Polyethylene Insulated Cables by Cable Voltage Rating

6-16

Voltage Rating of Cable

Percentage of Nuclear Plants

5 kV

31%

8 kV

9%

15 kV

12%

25–35 kV

3%

7

AGING AND DEGRADATION OF BUTYL, ETHYLENEPROPYLENE RUBBER, AND CROSS-LINKED POLYETHYLENE CABLES DUE TO ADVERSE ENVIRONMENTS

7.1

Aging and Degradation of Butyl Rubber

7.1.1 General Butyl rubber will degrade due to the long-term effects of wetting, thermal stress, and radiation. The loss of physical properties during aging, such as changes in tensile strength or elongation, is normally caused by elastomeric chain scission or, to a lesser extent, by additional chain crosslinking. Scission refers to elastomeric chain cleavage, and reversion generally refers to destruction of the cross links rather than elastomer backbone cleavage. External stress-inducing agents include oxygen, ozone, thermal energy (most heavily), ultraviolet light, environmental chemicals, and high-energy radiation. Voltage stress is anticipated to become a factor only as the voltage rating of the cable increases or the insulation system is wetted [34]. Additives to counter the influence of some of these stresses were sometimes incorporated into the elastomer. These additives, such as amines or phenolics, interact with free radicals (electrons) generated by these outside sources, and are referred to as antioxidants or antiozonants. As long as a small percentage of the original antioxidants and antiozonants are available in the polymer, polymer degradation proceeds at a slow pace. When these materials are consumed, degradation can proceed quite rapidly, especially at higher stress levels. 7.1.2 Water-Related Degradation of Butyl Rubber Butyl rubber cables are susceptible to water-related degradation. Operating experience shows that failures have due to water-related degradation have occurred in U.S. power plants after 20–25 years of service. A typical elastomer must pass an electrical stability test, which involves aging the insulated wire in water at elevated temperature with 60 Hz voltage applied and measuring the capacitance and power factor. The moisture sorption characteristics of the common butyl formulations (see Appendix D.3) are 11.1 and 10.9 mg/in2 (1.72 and 1.69 mg/cm2), respectively, after 7 days of wet aging at 185° (85°C) [35, 36]. By today’s standards, these values are high when compared to later (but still old) EPR formulations that absorbed 0.9 mg/in2 (0.14 mg/cm2) (0.1%) after one week of wet aging at 194°F (90°C) [28]. Yet, the butyl rubber values were considered in the 1960s to exhibit “low moisture sorption” [37].

7-1

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

Voltage stress influence was described by Hiranandani [34], who reported on a 30-year-old, nonshielded, 5-kV butyl rubber insulated cable with a neoprene jacket that possessed areas in which the jacket was not bonded to the insulation. The cable had areas that showed discharge activity, but the cable had not failed (before removal and analysis). Surface discharge–induced degradation, which can exist for an extended period, would result from voltage-induced phenomena in the absence of water. Although degradation of butyl rubber is expected to increase the rate of electrical property deterioration, that rate remains low, with electrical degradation taking as long as three or more decades to occur. 7.1.3 Thermal Degradation of Butyl Rubber Thermal stress is an aging issue for butyl rubbers, and aging rates and aging mechanisms differ depending on the type of cure used in manufacturing the insulation. Sulfur-vulcanized butyl rubbers decompose at elevated temperatures; however, purely thermal scission of the sulfur-containing cross links was considered negligible below 320°F (160°C) [35, 36]. Regardless, it has also been noted that sulfur-vulcanized formulations tend to soften during prolonged exposure to elevated temperatures of 300–400°F (150–204°C) [30, p. 257]. The latter temperature is at the lower end of the range for decomposition noted by Conley, and it is likely that the softening is related to the degradation of some sulfur cross links. Fenger [38] reported on a late 1960s vintage (40+ years old) butyl rubber insulated cable (4160 V, 760.5 mm2) that failed due to progressive softening of the insulation due to aging. Softening is an indication of degradation due to reversion. This failure is described in Appendix A. Conant noted that some elastomers undergo loss of strength (reversion) due to degradation as a result of heat aging; the type of accelerator also influences reversion [36, p.116]. Of great significance, quinoid curing systems (C-N cross-link bonds) and peroxide cures used in butyl manufacturing yielded greater heat stability than the C-S-S-C or C-S-C bond [35, p. 246] used in sulfur cures. Degradation, as seen by softening, is clearly relevant to butyl rubber long-term aging, and curing techniques are relevant to overall properties. It is possible that additives incorporated to deter degradation dissipate near end of life. It is also unlikely that all butyl rubbers will behave the same. Some compounds can harden instead of softening due to thermal aging. 7.1.4 Radiation Degradation of Butyl Rubber In nuclear applications, pure polyisobutylene (homo) polymer responds differently to radiation than the filled copolymer that is butyl rubber insulation. More than 40 years ago, the technical literature reviewed the effect of radiation on polyisobutylene (the pure butyl polymer), polyisoprene rubber, and polybutadiene rubber. Polyisobutylene, with no un-saturation, underwent degradation under high energy radiation [39], whereas the other two elastomers underwent cross linking [40]. It is apparent, therefore, that the un-saturation in the latter two hydrocarbon elastomers renders them more susceptible to cross linking. This implies that, for butyl rubber compound (which is a copolymer), cross linking could possibly initially dominate chain cleavage/scission due to the isoprene portion of the molecule, with chain cleavage/scission degradation dominating as the dose (aging time) increases. Therefore, assuming the presence of a small amount of residual un-saturation in the finished and installed butyl rubber insulation, a 7-2

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

rather slight initial improvement in properties of the butyl compound would be expected during radiation aging, ultimately followed by softening during continued aging. Sheirs [41] suggests the opposite with respect to properties, as it states that, with radiation, “butyl rubber crosslinks to become stiffer, lose elongation and turn friable and powdery.” Sheirs refers to gamma radiation, and although some cross-linking is to be anticipated, the presentation makes no mention of dose or dose rate. One would expect other changes to be noticeable before those described by Schiers. It was shown in past years that polyisobutylene cross linking could be enhanced under gamma radiation by use of certain additives [42], but it is not known whether this technology was applied to butyl rubber for cable insulation. In general, few if any butyl rubber insulated medium-voltage cables are expected to be in high radiation zones. The other stresses appear to be more significant at low anticipated radiation doses. Thermal stress is likely to be of the most importance in plant applications. Overall, the direction of physical property changes reported is not unreasonable, because the polyisobutylene portion of a butyl rubber copolymer would be expected to ultimately degrade; however, the radiation doses required for such degradation should far exceed what is seen in practical butyl rubber usage for commercial nuclear power station applications. 7.1.5 Conclusions Butyl rubber was a reasonable choice for general power station applications based on the state of knowledge in the 1960s. However, history has shown that harmful environmental issues, some not foreseeable at the time, influence long-term reliability. These include allowable moisture pickup in butyl rubber cable compounds after extrusion; thermal degradation due to the nature of the cross links in butyl rubber and their influence on thermal stability (not an issue for EPR—see Section 7.2, Aging and Degradation of Ethylene-Propylene Rubber); and possible overall radiation-induced degradation after long aging times. Also, for shielded butyl rubber cables, the issue of contaminants and impurities in the carbon black and shield compounds, discovered in the 1980s (and corrected at that time), also will contribute adversely to the rate of aging and loss of life of butyl cables

7.2

Aging and Degradation of Ethylene-Propylene Rubber

7.2.1 General A number of types of EPR insulations and EPR cable designs exist that are partially identifiable by the color of the EPR (black, gray, pink, and brown). Industry operating experience indicates failures of the cable in certain of the black, gray, and pink EPR types under wetted conditions. Some types of discharge-free pink insulation and all the discharge-resistant insulations have been free of water-related insulation failures. U.S. power industry operating experience indicates that some EPR insulation systems are affected by wetting, but the extent to which an individual manufacturer’s compound is susceptible to water-related degradation differs from one EPR to another. Regardless, failures of EPR cables in wetted environments have occurred and must be addressed as an aging concern.

7-3

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

In some instances, adverse thermal conditions resulted in failures of EPR insulated cables. Identifying those adverse thermal conditions (such as external radiant energy, ohmic heating, or high-resistance connections), determining the effect on the cable and correcting the condition are required to maximize cable reliability. Thermal runaway and embrittlement due to insulation hardening are the effects of long thermal degradation. Thermal runaway occurs when the heat generated in the cable and insulation system are not dissipated by the surrounding environment, causing the insulation resistance of the insulation to decrease. This results in higher leakage currents in the insulation, which, in turn, increases the temperature of the insulation. The effect feeds on itself to the point at which breakdown of the insulation occurs. In general, medium-voltage cables are not exposed to radiation levels that would affect the longterm reliability of EPR cable compounds. However, if medium-voltage cables have been exposed to both higher than normal temperature (122°F [>50°C]) and radiation doses of >5 Mrad (>50 kGy), a loss of elongation and a hardening of the jacket and/or insulation should be expected. Identifying cables in these adverse conditions and verifying the electrical and physical condition is recommended. More information can be found in the EPRI report Aging Management Program Guidance for Medium-Voltage Cable Systems for Nuclear Power Plants (1020805) [5]. 7.2.2 Water-Related Degradation of Ethylene-Propylene Rubber With reference to moisture influence on power plant cables, a 1994 survey, described in the EPRI reports Effects of Moisture on the Life of Power Plant Cables, Part 1: Medium Voltage Cable (TR-103834-P1) and Effects of Moisture on the Life of Power Plant Cables, Part 2: Low Voltage Cable (TR-103834-P2) [43, 44], sought feedback from 50 plants at 24 utilities. The vast majority of the cables were EPR insulated, with a much smaller XLPE population. Only 34 failures in almost 100 plant-years of experience were identified as externally initiated; that is, related to wetting in conjunction with manufacturing defects, damage during installation, or transient surges. A more recent 2003 study, described in the EPRI report Medium Voltage Cables in Nuclear Plant Applications—State of Industry and Condition Monitoring (1003664) [45], touched on this subject. It sought to identify the specific types of medium-voltage cables in service, to assess cable and accessory failure experience, and to outline additional experience with diagnostic tools for condition assessment. Only a few stations (14 units) participated in the survey, and the report (focusing primarily on diagnostics) concluded that insufficient experience with EPR cables existed to understand condition evaluation criteria for cables with EPR insulation. These reports and additional related information are summarized in NEI 06-05 [11]. NEI 06-05 [11] shows that Okonite and Anaconda were main suppliers of the old black EPRs; the newer pink EPRs were also manufactured by Okonite and Anaconda. The UniShield design was also provided using pink EPR. The brown Kerite EPR cables were installed over a long period. (The supplier can change the nature or components of the EPR compounds, but one cannot tell this from the color.) NEI 06-05 also summarizes the plant failures as a function of the age of the cables at failure. The figure from NEI 06-05 is reproduced in Figure 7-1 and indicates that some pink EPRs have failed. Review of the input data for Figure 7-1 indicates that no pink Okonite insulation has failed under wet aging conditions.

7-4

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

Figure 7-1 Ethylene-Propylene Rubber Cable Failures as Function of Color of the Ethylene-Propylene Rubber Insulations [11]

Figure 7-2 provides an indication of onset and expected onset of initial failures associated with water-induced degradation for the various insulations used in medium-voltage cables. Because EPR is the dominant insulation type, Figure 7-2 applies mostly to EPRs. One manufacturer’s insulations (both black and brown discharge-resistant Kerite insulation) has not experienced a water-related failure to date. The black discharge-free insulation was installed in some plants in the early 1970s, with the brown version replacing the black discharge resistant-insulation in the mid 1970s. A significant change in the compounding of EPR insulations occurred in the mid 1970s. Before that time, the clays were heat treated to dry them. In the mid 1970s, the dried clays were also treated with silane, which improved the bonding with the EPR and reduced the water absorption. This change to treated clay greatly reduced the rate of water-related degradation. As a result, the Okonite pink insulation produced since that time has not experienced water-related failures. Accordingly, two of the most dominant cable types in the U.S. nuclear industry where records are more formally maintained (Kerite black and brown EPRs and Okonite pink EPR) have not experienced water-related failures. Failures have been experienced in compact design EPR cables; however, it is not clear at this time whether the failures of these cables is from insulation degradation, design and manufacturing problems, or installation problems.

7-5

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

Figure 7-2 Generational Differences in Life Expectancy for Ethylene-Propylene Rubber and Other Insulations

7.2.3 Thermal Degradation of Ethylene-Propylene Rubber Under most conditions, the thermal life of EPR cables should be long. However, EPR cables that are subjected to adverse thermal conditions such as radiant heat sources (uninsulated steam lines), ohmic heating, or heat generated from high-resistance connections can have shortened lives. The effects of such exposure can result in loss of elongation and hardening of the cable jacket and insulation. Accelerated thermal aging of EPR has been performed to seek to estimate the life of the cables using the Arrhenius relationship; this approach applies the principle that insulation life is inversely exponentially proportional to temperature. Laboratory aging procedures for medium-voltage cable have traditionally been based on methodology developed for XLPE insulated cables. Therefore, it is not uncommon to age cable in pipes (accelerated water treeing test), or in temperature-controlled tanks (accelerated cable life test [ACLT]) with water being present both outside the cable core and within the cable strands. The accelerated water treeing test uses water at approximately 113°F (45°C) at 3V0; the ACLT uses controlled conductor temperatures that can be at ambient or load cycled (daily) to some maximum, such as 140°F, 167°F, or 194°F (60°C, 75°C, or 90°C). The applied stress for the ACLT can be two, three, or four times the rated voltage (V0). Cables can be aged until failure occurs (and the times to failure compared for different conditions), or they can be aged for fixed times, with dielectric strength being determined after those aging times. Retained dielectric

7-6

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

(breakdown) strength is compared to aging conditions [46]. Aging at slightly accelerated voltage stress at ambient temperature (~86°F [~30°C]) has also been performed. Numerous EPR insulated cables of various compound formulations and constructions have been studied and compared using these methods. What is clear from these studies is that pink EPRs do not behave in a manner similar to XLPE or TR-XLPE. For the latter two, the dielectric strength drops continuously on aging (at a slower rate under reduced accelerated conditions, with TR-XLPE having a slower rate than XLPE) with failure occurring only after long aging times. For EPRs, the dielectric strength drops initially, stabilizes, and then remains essentially the same without the cable failing within the timeframe of the test programs. For XLPE, the higher the voltage stress, the more rapid the loss of life; the higher the load cycle temperature, the more rapid the loss of dielectric strength and the more rapid the failure during aging in the water-filled tanks. The EPRs do not follow this trend, and the actual aging conditions are not as significant as for XLPE. Differences in the shields of the EPR cables can also play a significant role in the results obtained. So, although one can obtain an apparent projected life for XLPE and TR-XLPE by using the ACLT, the same cannot be achieved for the pink and brown EPRs. The procedure is not suitable for brown EPRs (see Section 6.5.2.3, Brown Ethylene-Propylene Rubber) and there is little interest in performing such tests on older, black EPRs. Years of study by General Cable Corporation [47–49] have confirmed that ACLT test results for pink EPRs cannot provide absolute predicted service life—there is no correlation between time to failure and predicted service life. It has also been demonstrated that these EPR cables have a lower ac breakdown strength than TR-XLPE cables after aging but that both have long service lives. The ACLT is apparently of greater value for XLPE and TR-XLPE than for pink EPRs. However, ACLT testing of pink EPRs might provide information on 1) whether tested cables comply with industry standards, 2) breakdown strength retention, and 3) relative rankings of different EPRs (or EPR insulation-shield compositions). A comprehensive, comparative study of various EPRs was performed under EPRI sponsorship in which five commercial EPR grade cables of the same configuration, from five different manufacturers, were studied both in the laboratory and in service (underground). Both operating and accelerated voltage stresses were used, with cables being aged at ambient. In service, normal operating conditions were applied; in the laboratory, the cables were aged in tanks at ambient temperature (~86°F [~30°C]) with water in the strands. Normal and elevated voltage stresses were applied in both the laboratory and in service. Katz described details of the tests [50]. Figure 7-3 shows the results from aging the five different pink and brown EPRs (and TR-XLPE) for up to 70 months at rated voltage while in service. Note the differences in initial ac breakdown strengths, as well as the changes on aging; clearly, the EPRs do not respond alike.

7-7

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

Figure 7-3 AC Breakdown Strength of Medium-Voltage, Field-Aged Ethylene-Propylene Rubber Insulated Cables at V0 [51]

Figure 7-4 shows the comparative information from aging the same EPRs [50] in service and in the laboratory at rated voltage (the latter procedure used water in the strands and outside the cable core). Results at elevated voltage stresses applied in both the laboratory and in the field have also been generated [50]. All the EPRs are combined in Figure 7-4, solely for the purpose of easily comparing the in-service and laboratory aging results. The figure shows that water presence alone can accelerate the aging relative to service aging (although the temperature in service was likely different from in the laboratory). Figure 7-4 also shows that, in both the field and laboratory aging, the breakdown strengths remains six to seven times greater than operating voltage, even though saturation with water has caused a significant drop in breakdown strength from the dry condition.

Figure 7-4 AC Voltage Breakdown Strength of Combined Ethylene-Propylene Rubber Cables Aged in the Laboratory and in Service at Rated Voltage

7-8

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

Finally, a legitimate question exists as to whether it is technically appropriate to test all EPR insulated cable under the same accelerated temperature conditions. Kerite cable is much lossier, and as a result, temperature control in tanks is difficult. Furthermore, even if adjustments could be made, this set of accelerated test conditions remains of questionable validity for that significantly different set of cable materials and construction technology. Further information on EPR aging can be found in the literature [43, 50, 52–54]. The earlier studies show that response to aging under water, thermal, and voltage stress yields different results for different EPRs and for different shield types and that application of accelerated thermal and voltage stress conditions must take into account the nature of the EPR compound being evaluated. Valuable information has been obtained from past studies, but the application of this information requires good judgment. 7.2.4 Radiation Degradation of Ethylene-Propylene Rubber Medium-voltage cables in most U.S. power plants are rarely exposed to high levels of radiation. However, if there are specific cases in which medium-voltage cables have been exposed to both higher than normal temperature (122°F [50°C]) and radiation doses of >5 Mrad (>50 kGy), loss of elongation and hardening of the jacket and/or insulation should be expected. Identifying EPR cables in these adverse conditions and verifying the electrical and physical condition of those cables is recommended. 7.2.5 Conclusions The reliability of EPR compounds in general, even the early black EPR compounds, have been quite good. The major reliability challenge to date comes from slow, long-term degradation due to installation-induced or manufacturing-induced defects, often combined with adverse conditions (long-term wetting, thermal, and radiation). To date, research has not indicated when these cables will reach an end of life condition due to normal thermal aging. The main challenge at this time is to identify and correct or manage the adverse environments that have been identified as contributors to the failures evaluated to date.

7.3

Aging and Degradation of Cross-Linked Polyethylene

7.3.1 General NEI 06-05 [11] noted the history of XLPE failures, both unfilled and mineral-filled; 12 XLPE failures occurred at four plants, eight were at one plant that used a “filled XLPE that was used only at that plant.” Figure 7-5 shows failure information for these XLPE cables relative to all cables.

7-9

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

Filled XLPE failures are shown in blue, conventional XLPE failures (“Others”) are shown in red, and all non-XLPE failures are shown in gray. Figure 7-5 Unfilled and Mineral-Filled Cross-Linked Polyethylene Cable Failures Compared to All Failures

7.3.2 Water-Related Degradation of Cross-Linked Polyethylene Wetting of XLPE insulated cables can lead to water trees. However, it is likely that the mineral filler technology in those past years when filled XLPE cables were manufactured and supplied, was more closely related to butyl rubber than to the later EPRs. (This assumption is based on the years of manufacture of these XLPEs and knowledge of how filler technology advanced, as described in Sections 6.3 and 6.3.2). The mineral-filled XLPEs that failed in service were attributed to water treeing (meaning water or voltage as the primary environmental stress). A potential mechanism is inadequate polymer filler interaction, thus leading to weak boundary interfaces and sites for water to enter. Once that occurred, water tree growth would ensue. For non-filled XLPE, which was used in the remainder of the plants having XLPE insulated cables, water treeing occurs at defect sites where voids, inclusions, or contaminants exist from the time of manufacture or installation. The defects allow water to enter the normally hydrophobic material. At these sites, slow electrochemical degradation occurs on a microscopic level. When the polymer’s physical properties degrade from the electrochemical damage, electromechanical forces cause the water to break through the polymer. This microscopic

7-10

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

degradation continues over time and, over decades, the water tree grows to the point at which the insulation becomes sensitive to voltage surges that would either break down the polymer or cause the water tree to convert to an electrical tree, which would grow more rapidly and result in electrical failure of the insulation. 7.3.3 Thermal Degradation of Cross-Linked Polyethylene Although XLPE can be damaged by extended periods of high temperatures, most mediumvoltage cable is not expected to be exposed to damaging temperatures and radiant heating. As with EPR, if XLPE cables are adjacent to uninsulated or improperly insulated high-energy lines, the radiant energy or high ambient temperature could cause deterioration of the XLPE or, possibly, thermal runaway. Inspection of the vicinity of cable trays and conduits containing XLPE insulated medium-voltage cable can readily determine whether hot piping and equipment is in the vicinity of the cable and whether damage has occurred. 7.3.4 Radiation Degradation of Cross-Linked Polyethylene Irradiation of XLPE causes further cross linking and hardening of the polymer. Near the end of the life of the polymer, further cross linking is not possible, chain scission becomes dominant, and the material will powder. The doses required for this to occur far exceed the normal and accident doses in nuclear plants. Historical results indicate that radiation was a less important parameter than water or voltage stress effects. Radiation environments for XLPE cable located outside containment are not severe enough to produce significant aging of medium-voltage XLPE cables. 7.3.5 Conclusions All XLPE insulated cables for power plant application require oversight and attention, with particular emphasis on mineral-filled XLPEs with respect to long-term wetting. Dry XLPE cables should have a long life, provided they are not exposed to high temperatures for extended periods.

7.4

Other Degradation Causes

7.4.1 General Additional degradation modes can affect both shielded and nonshielded cables and could result in cable failures, regardless of insulation type. Corona discharge, surface PD, and internal PD, as well as their effects on the cable, are described in the following subsections. In some cases, visual inspections can identify the degradation (corona discharge damage), whereas in other cases, maintenance or in-service testing can be used to detect PDs. 7.4.2 Corona Discharge Corona discharge is the ionization of air that occurs when an electrical potential high enough to ionize the air is present. Corona can be seen in a darkened room as a blue glow around the conductor at the point of origin. The resultant ozone can degrade the surface of the insulation, and over a long period, can result in insulation damage and even breakdown. Signs of corona discharge damage can be visually observed as a white, bluish white, or pinkish white powder 7-11

Aging and Degradation of Butyl, Ethylene-Propylene Rubber, and Cross-Linked Polyethylene Cables Due to Adverse Environments

that forms adjacent to the discharge site or in a dead air space if an air flow is present. The discharge site is often a grounded surface that is just touching a nonshielded cable or a shielded cable with an ungrounded shield. Periodic visual inspections for powder residue from corona discharges could be performed as a required preventive maintenance task for nonshielded cable in circuit breaker housings and termination boxes and cabinets. These inspections could be done in conjunction with inspections and maintenance of the equipment being fed by the cable. Corona discharge is a subset of PD. Corona discharge occurs only at the surface of the insulation. Section 3.2.5, Surface Corona and Partial Discharge, describes corona discharge and shows a picture of the powder that indicates its presence. 7.4.3 Partial Discharge PD is also the breakdown of solid or fluid electrical insulation that is subject to a high enough voltage stress. PD can occur in voids caused by poor mixing or extruding of the insulation compound, gas spaces in the insulation or between the conductor and the conductor shield, or between the surfaces of an ungrounded or single-point grounded cable that is in close proximity to ground. PDs can also occur in electrical treeing within the insulation. The end result is to reduce the dielectric capability of the insulation system until insulation breakdown occurs. The length of time to insulation breakdown is highly dependent on the location and nature of the defect and can occur quite quickly (days, weeks, months) or take longer (years). PD within the shield and insulation systems cannot be detected visually. PD is difficult to excite in 4 kV cables, and the design of the cables has attempted to eliminate PD (hence the term discharge-free insulation). Modern standards require as-manufactured insulation to have no more than 5 pC of PD at up to four times operating voltage. Discharge-resistant cable (Kerite) has design tests that show that the cable is not susceptible to PD and corona damage after extended exposure to discharge from application of high voltage with respect to ground.

7-12

8

TESTING: MANUFACTURING, INSTALLATION, AND MAINTENANCE OR IN-SERVICE TESTS This section describes the tests performed on cable during design and manufacture, after installation, and during operation. It supports the development and revision of purchase specifications and receipt inspection of the cable. Section 8.6, Maintenance and In-Service Testing, provides a basis for application of the appropriate test protocol for a particular cable type and primary failure mechanism for a particular adverse condition.

8.1

Introduction

Tests are performed on cables in the design stage, during and after manufacture, after installation, and after being in service for some time. The objective of the tests is to verify that the cable meets the requirements of the applicable standards or specifications and to ensure reliable operation. The tests can further be described as follows: 

Manufacturing tests are conducted in connection with the design of a cable, at various stages during the manufacture of the cable, and on the completed cable before shipment.



Acceptance tests are conducted when the material arrives at the users’ facilities.



Installation tests are conducted to verify that the cables, accessories, and associated equipment have been properly installed and can be placed in service.



Maintenance tests are conducted to ensure that the installed cable continues to be suitable for service and has not aged or been damaged to the point at which in-service near-term failure is expected.

8.2

Purpose of Tests

The purposes of manufacturing tests can be divided into several categories. Design tests provide the data that the manufacturer uses to make adjustments in cable materials and design to achieve the desired cable characteristics for the application. These include the design qualification tests described in ICEA and AEIC standards and numerous other tests that manufacturers use to prove their designs. (These qualification tests are design qualification tests; they should not be confused with environmental qualification tests required by IEEE standards 323 and 383.) Manufacturing tests verify that the manufacturing process is under control and will result in finished cable that meets the specified requirements. These tests also minimize scrap by providing information that the product is satisfactory during the manufacturing run such that fewer failures occur during the final tests on the cable.

8-1

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Final tests verify that the product has been properly manufactured and exceeds the requirements of the specifications to which it has been manufactured. From the user’s standpoint, the purpose of manufacturing tests goes beyond verification that specifications have been met. The test data generated form the base from which maintenance tests can be selected and conducted to track aging and determine if the cable characteristics have been deteriorated seriously. A lack of knowledge of the as-manufactured state can result in the selection of maintenance tests that are meaningless, or possibly destructive. For instance, discharge-resistant medium-voltage cables are not deteriorated by PD at operating stresses and are not subjected to PD testing during manufacture. A user unaware of this fact could select PD testing for maintenance purposes. The results would be meaningless and would cause confusion, at best, as well as leading to possible replacement of a cable that was in good condition. Acceptance tests are performed by the user before installing the cable and are performed to partially verify that the relevant specifications have been met and that shipment damage has not occurred. These tests are generally not as stringent as the post-manufacturing tests. Installation tests verify that the installation is correct, damage free, and suitable for placing in service. These tests verify that no gross installation errors that would lead to early failure have occurred. Maintenance testing determines whether the cable is satisfactory for continued service. Two different approaches can be used. Withstand testing represents a go/no-go challenge to the dielectric integrity of the system, whereas diagnostic testing is intended to characterize the condition of the aged dielectric. Although these two approaches can be used independently, they can also be applied in a complementary manner when a higher degree of confidence in the cable circuit is required. These technologies are described in Section 8.6. More exhaustive descriptions of their strengths and weaknesses can be found in IEEE Std. 400-2001 and related subdocuments [51–55, 56]. Although the IEEE standards provide insights into the various test technologies, plant personnel must be aware that most techniques (and the related literature) were developed for the power distribution industry. Although much of the technology readily translates to the power generation industry, the differences in cable construction, preferred dielectric, shield construction, modes of degradation, and cable system architecture can impact or even negate practices that are acceptable in the distribution industry.

8.3

Manufacturing Tests

8.3.1 Standards and Test Methods Manufacturing tests go well beyond industry standards, yet industry standards can be relied on as describing the tests and results that are universally applicable to the cable under consideration. Changes in materials and processes cause industry standards to be works in progress. An accurate understanding of the tests to which all of the cables in a category were subjected and passed in the manufacturing phase can be attained only by knowing the standard involved and the edition in effect at time of manufacture. A review of the specification and common manufacturing testing of shielded and nonshielded medium-voltage cable history of interest to plant operators is presented in Table 8-1. 8-2

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests Table 8-1 History of Specifications and Testing of Shielded and Nonshielded Medium-Voltage Cables Year

Standards Butyl insulated, IPCEA S 19-81 EPR insulated, IPCEA S 19-81

1969

XLPE insulated, IPCEA interim versions of S 66-524 (issued May 71) XLPE insulated, AEIC 5-69 (shielded medium-voltage cables only) It is improbable that AEIC 5-69 was called for by many (if any) plant specifications, as it was generally viewed as an underground distribution specification. Butyl insulated, IPCEA S 19-81

1972

EPR insulated, IPCEA S 19-81 and interim versions of S 68-516 XLPE insulated, IPCEA S 66-524 XLPE insulated, AEIC 5 (shielded medium-voltage cables) EPR insulated, IPCEA interim versions of S 68-516

1975

EPR insulated, AEIC 6 (shielded medium-voltage cables) XLPE insulated, IPCEA S 66-524 XLPE Insulated, AEIC 5 (shielded medium-voltage cables) EPR insulated, IPCEA S 68-516

1979

EPR insulated, AEIC 6 (shielded medium-voltage cables) XLPE insulated, IPCEA S 66-524 XLPE insulated, AEIC 5 (shielded medium-voltage cables) EPR and XLPE (also TR-XLPE) ANSI/ICEA S-94-649, Concentric Neutral Cables Rated 5-46 kV ANSI/ICEA S-97-682, Utility Shielded Cables Rated 5-46 kV

2008

NEMA WC 74/ICEA S 93-639, 5–46 kV Shielded Power Cable for Use in the Transmission and Distribution of Electric Energy NEMA WC 71/ICEA S 96-659, Standard for Non-Shielded Cables Rated 2001–5000 Volts for Use in the Distribution of Electric Energy AEIC CS8, Specification for Extruded Dielectric Shielded Power Cables Rated 5 Through 46 kV

In addition to these standards, U.S. nuclear plants must meet the additional requirements of IEEE Std. 383 for environmental qualification testing [58, 65] and of IEEE Stds. 308, 603, and 1185 [59–61] for installation requirements. To list and describe all of the manufacturing tests for medium-voltage cables would result in a volume beyond the user’s interest. Refer to the appropriate standard (and desired edition) for a summary of all production and qualification tests. Table 8-2 lists the more common production tests found in ICEA S-97-682-2000, “Standard for Utility Shielded Power Cables Rated 5,000– 46,000 Volts” [62].

8-3

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests Table 8-2 Common Production Tests [62] Conductor DC Resistance Diameter Temper Nonmetallic Conductor Shield

Metallic Shields Dimensions Jackets, if Required Unaged and Aged Tensile and Elongation Other Applicable Jacket Tests

Elongation After Aging

Heat Distortion

Volume Resistivity

Heat Shock

Thickness

Cold Bend

Voids, Protrusions, Irregularities

Oil Immersion

Wafer Boil

Radial Resistivity

Spark Test (Nonconducting Layer Only) Insulation

Electrical Tests AC Withstand Test

Unaged and Aged Tensile/Elongation

Partial Discharge

Hot Creep

Jacket Spark Test

Voids and Contaminants

Other Tests

Diameter

Moisture in Conductor

Shrink back (XLPE/TR-XLPE)

Moisture under Jacket

Thickness Nonmetallic Insulation Shield Elongation after Aging Volume Resistivity Thickness Voids and Protrusions Stripping Tension Wafer Boil Diameter

The test methods and the minimum frequency sampling plan are listed in the standard for each test in the table. The word partial is used because additional tests unique to the manufacturer’s process and quality assurance methods are not included. The tests are designed to ensure that the cable received by the end user fully meets the specifications (including customer specifications) involved. The following section concentrates on those tests of greatest interest to power plant operators and those tests that might have an impact on maintenance and field testing. 8-4

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

8.3.1.1

Tests of Special Interest

Incoming material tests are generally not found in the traditional cable standards. Materials purchased from vendors for use in cable manufacture are commonly covered under purchase guides with appropriate standards and test requirements, followed by occasional production plant visits. Materials such as proprietary compounds manufactured in house are subjected to tests typical of compound suppliers as appropriate for the material. Dimensions such as diameters are increasingly being demanded to be met at every point in the cable. Manufacturers are making extensive use of statistical process control methods to ensure this result. Statistical process control has been extended to other test requirements as well. Dielectric constant and dissipation factor testing is conducted in accordance with ICEA T-27-581/NEMA WC-53 [46]. The specification requirements, as called for in ICEA S-97-682-2000 [62], are shown in Table 8-3. Table 8-3 Dielectric Constant and Dissipation Factor Acceptance Criteria Insulation

Type

Properties

Cross-Linked Polyethylene

Tree-Resistant Cross-Linked Polyethylene

EthylenePropylene Rubber, Class I, II, and III

EthylenePropylene Rubber, Class IV*

Dielectric Constant

3.5

3.5

4.0

4.0

Dissipation Factor (%)

0.1

0.5

1.5

2.0 (5–28 kV) 1.5 (>28 kV)

* Discharge resistant

Because these values cover a range of materials, they might not be representative of a specific material in the category. For instance, the dielectric constant most common for TR-XLPE in use is 2.4. Should a manufacturer actually obtain a value of 3.5 for a common TR-XLPE, it would be cause for great concern, investigation, and likely scrapping the cable. Conversely, should a special TR-XLPE consistently have a dielectric constant of 3.5, it would not be a negative indicator of tree resistance and might well be superior. Clearly, plant personnel who wish to know the actual values for the material purchased should require test reports or, at the very least, consult the manufacturer’s literature for the specific product. Discharge resistance tests, in accordance with ASTM D 2275-89 with a 60-Hz test voltage of 21 kV for 250 hours, are conducted on discharge-resistant materials [63]. No failure or surface erosion of the insulation sample is permitted to occur. Although only cables classified as discharge resistant must pass this test, it has been applied to other materials [33]. Void, contaminant, and conductor shield protrusion tests are conducted on a production basis. These are destructive tests performed on samples of production cable that allow the insulation to be physically examined under a microscope. These tests can also be used in forensic examinations of the cable insulation or shielding when a failure occurs in service. For forensic

8-5

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

examinations of XLPE cables, the wafers of the cable are dyed to allow a water treeing examination to be performed. For this type of test, a common method of examination involves “slinkies” cut on a lathe, or simply cutting individual wafers after the conductor has been removed. A typical examination is shown in Figure 8-1.

Figure 8-1 Wafer Examination for Voids, Inclusions, and Conductor Shield Protrusions

Unfortunately, the process is tedious and involves a small sample length relative to the total length of the cable for each examination. For translucent insulations such as unfilled PE, XLPE, and TR-XLPE, the examination includes the entire thickness of the wafer. For opaque materials such as EPR, only the surface of the wafer can be examined. Manufacturers take advantage of the fact that when unfilled XLPE and TR-XLPE are heated sufficiently, they become transparent (virtually clear). With the insulation shield removed and a sample heated (often in oil such as vegetable oil), a trained observer can see contaminants in the insulation and the smoothness of the conductor shield–insulation interface. For forensic assessments, this can also be done using a sample removed from a faulted cable to assess the relative quality of the cable near the fault. This test has been named the hot oil test and is shown in Figure 8-2.

Figure 8-2 Hot Oil Test (Insulation Is Clear in the Oil Bath)

8-6

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Filled-strand or blocked-strand conductors are becoming more popular since water penetration tests on blocked conductors showed their effectiveness. Testing of blocking materials is conducted in accordance with ICEA T-31-610 [64] and demonstrates the water pressure capability of the blocked conductor strand in the completed cable. Flame tests for nuclear plants are called for as required by IEEE 383-2003 [65], which in turn requires the test method of IEEE 1202 [66]. IEEE 383-1974 defined two methods of flame testing—a vertical flame test and a horizontal test using an oil-soaked burlap bag. The revision to the standard requires that the flame test be conducted as defined in IEEE 1202. The test is a vertical tray-type fire test with flames impinging on the cables near the bottom of the tray. Pass– fail criteria involve the extent of burning upward in the tray. A typical tray-type fire test is shown in Figure 8-3.

Figure 8-3 Vertical Tray Flame Test

8.3.1.2

Final Electrical Tests for Shielded Cables

The ac voltage withstand test consists of a 5-minute application of 60-Hz ac voltage to all cable types—XLPE, TR-XLPE, and EPR types I, II, III, and IV—at 200 V/mil (7.9 kV/mm) of nominal insulation thickness. For instance, for a 90-mil (2.3-mm) nominal insulation thickness of a 5-kV rated cable, the test voltage would be 18 kV. The test criteria are the same for ICEA S-93-639 and ICEA S-97-682 [57, 62]. For cables manufactured to AEIC standards before 1979 or only to ICEA standards in that same timeframe, the test voltage was 150 V/mil (5.9 kV/mm) of nominal insulation thickness or 13.5 kV for a 90-mil (2.3-mm) insulation. The dc voltage withstand test consists of a 15-minute application of dc voltage to all cable types at approximately 400 V/mil (15.7 kV/mm) of nominal insulation thickness, but it is optional in ICEA S-93-639 and ICEA S-97-682 [57, 62]. Before and during the 1990s, the test was required. It was made optional (and is not widely used) not because of fear of damage, but because years of experience showed it to be largely ineffective in locating cable defects in modern cables.

8-7

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

PD testing is conducted in accordance with ICEA T-24-380 [67] on all cable types except discharge-resistant cables. The voltage is raised to the maximum ac test voltage. During this time, the voltage at which PD (if any) begins is called the partial discharge inception voltage (PDIV). The voltage is then slowly lowered. The voltage at which the PD disappears is the partial discharge extinction voltage (PDEV). Except for rare occasions when they are equal, the PDEV is always lower than the PDIV. PD is measured in picocoulombs (pC). Accurate measurement requires test room shielding and highly sensitive measuring equipment. Improvements in measurement and cable manufacturing technology have allowed continuous improvement in the PD requirements with time. A graphic plot of PD tests (called an x-y plot) has been common since the 1970s. Figure 8-4 shows a typical passing result for a modern cable in which no PD is measured (5 pC is the accepted lower limit of sensitivity).

Figure 8-4 An Acceptable Partial Discharge Plot

Figure 8-5 shows a failing result. The PDEV does not occur until the cable operating voltage is reached.

Figure 8-5 An Unacceptable Partial Discharge Plot

8-8

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

In 1994, AEIC 5 [68] was revised for XLPE cables with extruded insulation shields to allow no measurable PD in cables manufactured to the specification. Before that time, discharge was allowed, as long as the level did not cross the pass–fail line versus test voltage. The history for PD requirements in AEIC 5 is shown in Figure 8-6.

Figure 8-6 Partial Discharge Requirement History from AEIC 5

Cables not manufactured to AEIC specifications might have met less stringent PD requirements. Plant personnel must consult the proper specification and edition to determine the requirements for the cable of interest. One special case involves cables that used semiconducting tape insulation shields. These were subject to less stringent requirements, even in AEIC specifications. Discharge-resistant cables (Kerite black and brown EPR) are not PD tested in accordance with industry standards. The insulation is not deteriorated by PD at and above operating voltage and, if tested, it would likely show some discharge. Production PD testing techniques date back to the 1950s [33], and the acceptance criteria have become tighter with years of progress. This has resulted in the emphasis for most EPR insulations shifting from greater discharge resistance to constructions demonstrating less measureable PD [33]. Thus, the PD history for older EPR cables is not as great a cause for concern. This reported history seems to be supported by the fact that AEIC 6-73 [69] required the U-bend discharge test as an EPR qualification test, but it was dropped from following editions. Modern EPR insulations are, in general, more discharge resistant than XLPE. Plant personnel considering the use of PD maintenance tests (in-plant field tests) for diagnostic purposes might obtain meaningless results unless the test is appropriate for the cables involved. Insulation resistance is now an optional test appearing as a qualification test in both ICEA S-93639 and ICEA S-97-682 [57, 62]. In the past, this completed cable test, conducted with a dc voltage source of 100–500 volts applied for 60 seconds was used to calculate an insulation resistance constant. The required specification constants were so low as compared to the actual constants for medium voltage XLPE and EPR cables that they were meaningless. A difficulty arose when some field personnel attempted to use these constants to determine a passing insulation resistance for dc high-potential tests. 8-9

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

8.3.1.3 Final Electrical Tests for Nonshielded Cables, 2001–5000 Volts Without Metallic Sheath or Armor The ac voltage wet test is the most common test for nonshielded cables. It requires water immersion for a sufficient time to ensure complete water penetration into the cable reel to provide an electrode to test from the conductor to the made electrode. An ac voltage of 13 kV is applied for five minutes. An alternate dc wet test may be substituted. An equivalent high-potential maintenance or field test is not possible unless a made electrode, such as flooding the ducts or conduits enclosing the cables, is practical. This is seldom the case. 8.3.1.4

Qualification Tests for 5 kV–35 kV Shielded Cables

Before ICEA S-93-639 and ICEA S-97-682, ICEA standards did not specifically call for qualification tests [57, 62]. ICEA standards S-19-81, Rubber Cables; S-68-516, EPR; and S-66-524, XLPE, have been withdrawn [70–72]. AEIC standards 6-73, EPR, and 5-69, XLPE (both of which have been withdrawn) were the first to specifically do so [69, 73]. However, manufacturers conducted many of the tests similar to the AEIC qualification tests in connection with design development, and the test results were published in the manufacturer’s literature. The AEIC qualification tests were mostly tests on dry cable specimens with the exception of the electrical moisture absorption test in AEIC 6-73, EPR [69]. Many of these early tests remain in ICEA S-97-682 [27]. A flowchart of the current standard is shown in Figure 8-7.

Figure 8-7 Dry Specimen Design Qualification Tests

Several manufacturers regularly conducted wet tests on cables, but these were not standardized. AEIC CS-5-79 for XLPE first required a wet qualification test, called the accelerated water/electrochemical treeing test [74]. This was quickly followed in AEIC CS-6 for EPR [69]. The current flowchart found in ICEA S-97-682 is shown in Figure 8-8 [27].

8-10

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Figure 8-8 Wet Specimen Design Qualification Tests [27]

These wet (or treeing) qualification tests had a good history of identifying combinations of insulation and shielding materials in completed, unjacketed, 1/0 AWG (53.5 mm2) aluminum or copper, 100% insulation level, 15 kV cable that resists water treeing. Other qualification tests in the standard include jacket material qualification, continuous vulcanized extrusion qualification, and an “other” qualification test category. ICEA S-93-639 [57] covers many of the cable designs found in ICEA S-97-682 [27]. However, ICEA S-93-639 for medium-voltage cables does not have the same qualification test regimen as ICEA S-97-682. This should not necessarily be a cause for concern, as it will have been qualified by default when the same core is used to make cables to both specifications. The exception is for core designs in ICEA S-93-639 that do not appear in S-97-682 (such as having a semiconducting tape insulation shield). Such designs are not likely to be desired for future needs of plants. ICEA S-93-639 does have several cable features that users may wish to use [57]. This makes it necessary to reference both specifications to cover all of the desired requirements. Low-smoke zero-halogen jackets are growing in popularity. ICEA T-33-655 provides specific tests for acid gas, halogen content, and smoke generation [75]. The importance of requesting and retaining manufacturer’s qualification and test reports concerning the basic cable design and the specific manufacturing run cannot be overemphasized. Industry standards might contain pass–fail criteria that cover a range of materials and cable. In that case, the pass–fail number might not be representative of the cable the user has purchased. Tests reports can be requested, including qualification test data and showing actual test data numbers. These may provide important baseline information for future on-site field testing. During the construction, it was common practice for the consulting firms involved to require technical data from the manufacturer, much of which is retrievable from plant records. That data is often helpful in gaining an understanding of the characteristics of the installed cable. Manufacturers’ literature is an excellent source of test results developed with different cable materials and designs. The literature should be for the cable and time vintage of interest. Some might object that it could have some commercial bias. Although that might be true, the data will at least be representative of the cable and far superior to using generic information or data from an entirely different era.

8-11

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

8.4

Acceptance Tests

Acceptance tests on cables being received by a utility were more common in the era of paperinsulated lead-covered cable than now. A few users extended the practice to extruded dielectric cables. Upon arrival at the plant location, samples would be removed and either inspected on site or sent to an independent laboratory for a partial repeat of several tests conducted in the factory to verify that the cable met the specification requirements. Common tests included dimensions, examination of voids and contaminants in the extruded layers (insulation and shielding), and stripping tension of the insulation shield. Experience showed that manufacturers were doing a good job of meeting the specifications and called into question the expense of conducting the tests. The practice is less common than in the past. In a few cases, the cable ends were exposed and isolated for a withstand test while the cables were on the reel. This overwhelmingly involved a dc high-potential test. Test voltages ranged from the factory test voltage to those used in installation testing. A significant fear was that the short test ends increased the probability of flashover during the test. This could result in traveling voltage waves on the cable, possibly causing damage.

8.5

Installation Tests

Installation tests continue to be used before a cable circuit is placed into service. The cable is new, so aging is not a factor. The main purpose is to verify that the circuit has been properly connected and that the cables and accessories have not been damaged by handling and installation. It might be argued that simply energizing the circuit would constitute a test. Although this is true, should a fault occur upon energizing, the damage involved might be quite extensive, requiring rework and a delay in placing the installation into service. An insulation resistance test is of value to verify that connections have been made properly and that the cable is safe to expose to elevated test voltages, but the insulation resistance test voltage is much too low to identify cable damage and poorly made splices and terminations. The insulation resistance test is acceptable as a pre-test to check connections but is not suitable as a condition monitoring test. Appendix F provides additional information regarding the use of insulation resistance values for cables. The historical installation test of favor has been a dc withstand (high-potential test). Although there are no concerns that the dc voltage from a properly conducted test on a new cable will cause damage, a dc test does not have the capability to identify significant damage short of a near cut-through. However, the dc test has been reported to have some success in identifying poorly made splices and terminations, so the test continues to be used. Instead of dc testing, very-lowfrequency (VLF) tan δ and PD tests, followed by a withstand test, are recommended for acceptance testing. This provides reasonable assurance that no detectable major defect exists and creates a baseline for later diagnostic testing.

8-12

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

8.6

Maintenance and In-Service Testing

8.6.1 Introduction Various in-service tests have been used in recent years to gain an understanding of the state of aged, installed cable systems. However, only a limited set of tests is applicable to the cable types used in plants. This tests include tan δ, dielectric spectroscopy, VLF and 60-Hz withstand, and PD. The selection of the appropriate test depends on the insulation, the type and condition of the insulation shield, and the rating of the cable. The selection also depends on the particular failure mechanism that is of concern. The tests described in this section apply to cables with insulation shields. An insulation shield provides a consistent ground plane that enables electrical testing that indicates the condition of the insulation. These tests do not apply to nonshielded cables. The lack of a ground plane for nonshielded cables has led to a search for techniques that might provide an indication of the state of the insulation. Plant cables that are nonshielded can be located in trays or in conduit. The technique of filling the duct or conduit with water that serves as the ground plane has been studied in the laboratory, but it has not been applied to field testing. One problem with using water as a ground plane is that the jacket of the nonshielded cable is included in the electrical test. The presence of a jacket can result in confusing test results. On-line diagnostics testing (see Section 8.6.3, On-Line Diagnostics Assessment) purports to be suitable as an electrical diagnostic test method for both shielded and nonshielded cable because the diagnostic technology approach is different—it detects the magnetic field resulting from current flow in either the shield or the conductor of a cable. For nonshielded cable testing, special sensors are required to eliminate discharges between the sensor and the cable under test, as well as for safety [76]. For 4-kV circuits, PD-related degradation might not be the dominant failure mechanism due to the low electrical stresses in the insulation system and its interfaces. Accordingly, there might be no signal to detect (that is, no PD or pre-PD) through on-line assessment. Montinari [77] warns that the applicability of either on-line or off-line PD measurement is of questionable value for wetted medium-voltage cables. PD measurement would detect a failure of a wetted cable only after a water tree converts to an electrical tree. In the case of on-line measurements, the need for high signal-to-noise ratio detectors to enhance noise detection can lead to identification of false negatives, whereas external noise detected by the detectors can result in false positives. Montinari states that the value of on-line detection might be as a screen to identify potential problems that should be followed by off-line validation combined with tan δ testing for distributed effects, especially for cables that have been wetted. Thermal and radiation aging (state of the cable) caused by the environment can be assessed by indenter testing, which evaluates compressive modulus. As elastomers age, their physical and mechanical properties change, and the materials generally harden. The indenter is nondestructive. In the test, a small, instrumented probe is pressed against the side of the cable at a constant velocity while force is measured. The modulus is the change in force divided by the change in probe travel. As the material hardens, the modulus increases. Test data are available

8-13

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

for common rubber materials in the EPRI Cable Polymer Aging Database (1011874) [78]. The EPRI report Initial Acceptance Criteria and Data for Assessing Longevity of Low-Voltage Power Cable Insulations and Jackets (1008211) [79] provides a basis for establishing the degree of aging of the polymer by using the indenter and other technologies. The indenter cannot provide any indication of electrically induced degradation but rather would indicate whether the cable had degraded from thermal or radiation exposure. The indenter might be of limited use in evaluating ohmic heating, given that the jacket of the cable is in cooler ambient air. If the ambient temperature is normally cool, hardening of the jacket is a strong indication that ohmic heating is significant. 8.6.1.1

Historical Perspective on In-Service Testing

All cable must pass tests defined in AEIC procedures after manufacturing [80]. Withstand testing has been used as an in-service test method to determine the integrity of an installed, aged cable system. After a significant period of service, plant personnel have been concerned about the condition of aged cable circuits and have sought a simple test to provide information that would reduce or eliminate in-service failures. High-voltage tests have been used to detect gross imperfections or deteriorations in extruded dielectric cables. Appropriate voltage levels and time duration of off-line withstand field testing methods for distribution cables are found in AEIC CS8-07 [80] and IEEE Std. 400 [51, 55, 56]. DC testing of medium-voltage distribution cable systems was the preferred method for evaluating cables in the field until the 1980s. This preference was related to low equipment cost, portability of equipment, and ease of operation. The dc high-potential test, which was an intentionally destructive test for weakened cable, was developed mainly for paper-insulated, lead-covered cables and was later applied to extruded dielectric cables, as well. The dc test experience with paper-insulated, lead-covered cables was quite positive—such experience with paper-insulated, lead-covered cable systems that have been tested in the field with dc for more than 60 years has shown that testing with the recommended dc voltage removes the weak link cables and does not cause deterioration of sound insulation [81]. However, experience has shown that applying dc to extruded polymer cables frequently misses significant degradation and sometimes causes premature failures [82]. Carrying over dc withstand testing from paper-insulated, lead-covered cables to extruded polymer cable without considering that the two types of cables are significantly different in their response to applied stress caused in-service failures of some XLPE cables [83]. In essence, applying dc (in accordance with industry standards) to aged XLPE insulated medium-voltage cable systems might not induce failure in highly aged insulation. If it does not induce a failure, the test itself (due to the application of dc voltage) can result in trapping of space charge in the insulation that causes premature failure after the cable system is placed back in service. Whether failure occurs at the time of dc application is related to the degree of degradation that has taken place during service. The insulation must be at the point of in-service failure for the dc to cause the breakdown, leaving no operating margin. Also, dc testing often does not detect significant flaws in the insulation system and results in the false impression that the tested cable is sound.

8-14

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

As a result of this experience, the industry has followed two paths: 1) seeking a substitute for dc withstand testing that would perform the same role without causing a loss of service life, and 2) seeking alternative diagnostic tools that can estimate future performance again without shortening the life of the tested system. Current industry standards discourage the use of dc high voltage for all types of extruded cable as a means of testing aged insulation; however, it is still recommended for use on paper-insulated, lead-covered cable [51, 55]. 8.6.1.2

Withstand Versus Diagnostic Testing

A withstand test is designed to remove the weak link at the time of testing by causing it to fail at a convenient time for replacement. The cable system is tested while out of service, and an overvoltage is applied by a test set for specific period, such as a half hour. Available tests include 60 Hz, VLF (0.1 Hz), and dc withstand tests. (The dc withstand test was performed historically and is still the recommended test for paper-insulated, lead-covered cables, but it is no longer endorsed in IEEE Std. 400 for testing extruded cables that are more than five years old.) VLF test sets are similar in weight and size to dc test sets, whereas 60 Hz test sets are large and heavy due to the larger charging currents at the higher frequency. Diagnostic tests—PD, tan δ, and dielectric spectroscopy, for the purposes of this report—provide an indication of current condition, allowing inferences about the future performance of the cable system. These tests are intended to be nondestructive and would cause failure only if the cable were deteriorated to the extent that failure in service was imminent. Except for one technology, these tests are performed while the cable is out of service and an overvoltage is applied. The results of these tests are analyzed, and a projection is made whether to maintain the normal test frequency (good condition), test more frequently (further study required), or plan for cable replacement (action required). 8.6.1.3

Global Versus Local Assessment

Local test methods identify the weakest point in the aged cable circuit, whereas global tests determine the overall condition of the aged cable circuit. Local test methods, such as PD and withstand tests, either identify the location of the worst-case damage (PD testing) or cause the site to fail (withstand testing). In the case of one large flaw or a near-through-wall water tree that was near causing failure, withstand testing would cause the degradation to go to failure. If only one location on the cable was near or at partial discharging, PD testing would identify that location. A global test, such as tan δ, is intended to provide guidance on the overall condition of the cable circuit; as cables age, they undergo changes that alter the physiochemical nature of regions both with and without water trees. A global test is intended to evaluate the significance of these broad changes and provide information to estimate future performance of the entire cable length. 8.6.1.4

60 Hz Versus Other Frequencies

Assessment of insulation using different frequencies changes the nature of the test and can result in a different outcome. Direct comparisons between data collected with one frequency versus another may not be useful. However, although the direct comparison of results is not recommended, the diagnostic value of VLF and variable-frequency methods has proven to be as 8-15

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

reliable a tool as the use of 50-Hz or 60-Hz test sets. Considering that there is comparable value in methods, the portability issues for line frequency testing (which can make it impracticable for testing in many locations) makes VLF and variable-frequency testing an attractive choice. 8.6.2 Off-Line Diagnostic and Withstand Testing Guidance for developing a medium-voltage cable system aging management program for power plants is presented in the EPRI report Aging Management Program Guidance for MediumVoltage Cable Systems for Nuclear Power Plants (1020805) [5]. The report describes attributes of cable assessment and testing and provides guidance on those tests most applicable to plant cables. A cable testing regimen should be chosen based on the particular cable design, the adverse environment factors, and operating history. 8.6.2.1

Dissipation Factor (Tan δ) Testing

Tan δ testing is a global test that evaluates the ratio of the resistive current divided by the capacitive current in the insulation layer (see Figure 8-9). In 1981, Bahder et al. were among the first to report on this technique [84]. Later, Bach et al. and Hvidsten et al. reported a correlation between the decreasing ac breakdown level at power frequency and increasing dissipation factor at 0.1 Hz [85, 86]. Baur claims a strong correlation between the off-line 0.1 Hz tan δ value and the amount of water tree damage of the cable insulation [87]. The measurement of the 0.1 Hz tan δ provides a cable-aging assessment method that differentiates between good, defective, and highly deteriorated cable insulation.

Figure 8-9 Derivation of Dissipation Factor (Tan δ) Measurement in Insulation

As with other testing methods, tan δ testing can be performed using line frequency, VLF, or variable-frequency methods. EPRI report 1020805 recommends the VLF method and provides assessment criteria for this method [5]. VLF tan δ measurements are desirable because small portable test sets can be used. VLF tan δ measurements are made at several predetermined voltage level steps, starting at 0.5 V0 and proceeding in 0.5 V0 increments up to a maximum of 2.0 V0, provided no unacceptably high values are encountered. Preprogrammed equipment can be used or test operators can change the voltages at specific times. In a 0.1-Hz test, one cycle takes 10 seconds. Accordingly, the test duration must be long enough to take sufficient measurements to give valid results. Step durations vary from 30 seconds to 5 minutes. The test equipment automatically calculates and records the average and standard deviation of the readings. Although the absolute value of tan δ 8-16

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

provides some indication of aging, the difference in the tan δ readings at 2 V0 and 1 V0, the degree of degradation, is considered a stronger indicator. Table 8-4 provides criteria for XLPE insulation as being good, aged, or highly deteriorated. EPRI report 1020805 provides assessment criteria for EPR and butyl insulations [5]. Table 8-4 IEEE Standard 400 Criteria for Assessment for Cross-Linked Polyethylene Insulated Cables Tan δ at 2V0 10-3

Tan δ Increment at 2 V0 vs. V0 10-3

Assessment

< 1.2

< 0.6

Good cable

≥ 1.2

≥ 0.6

Aged cable

≥ 2.2

≥ 1.0

Highly degraded cable

Figure 8-10 shows a typical portable test device that is suitable for testing the relatively short cables in power plants.

Figure 8-10 Typical Variable-Frequency, Very-Low-Frequency Portable Test Equipment for Performing 0.1-Hz Dissipation Factor (Tan δ) Testing Courtesy of HV Diagnostics, Inc.

8-17

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Figure 8-11 shows the voltage dependence of the dissipation factor at 0.1 Hz for new and service-aged XLPE-insulated cables.

Figure 8-11 Voltage Dependence of Dissipation Factor for New and Aged Cross-Linked Polyethylene Cable

Another way of viewing the data is shown in Figure 8-12. Here the increase of dissipation factor with voltage stress is shown as a bar (the color code is the same as that used in Figure 8-11).

Figure 8-12 0.1-Hz Dissipation Factor of Cross-Linked Polyethylene-Insulated Cables

8-18

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

This test method is global and, therefore, determines the general condition of the cable under test rather than identifying and localizing weakened sites. One drawback is that numerous small water-aging degradation sites can respond in the same manner as one or a few severe (long) water-aging degradation sites, with the latter being the condition of most importance. It is believed that the severe degradation sites will cause the differential measurement between V0 and 2 V0 to be larger, rather than just causing the overall measurement to be larger but stable with increasing voltage. Advantages of 0.1-Hz dissipation factor testing include the following: 

The test identifies the existence of water trees and water-related degradation before the point of conversion to electrical trees.



Test equipment is portable and does not require the use of a van.

Disadvantages of 0.1 Hz dissipation factor testing include the following: 

Because an overvoltage is applied, a failure of a severely aged insulation system is possible during testing.



Because tan δ is an off-line test with elevated voltage, the cables must be disconnected from their loads (such as motors and transformers).



Testing of dissimilar, interconnected cable types, such as XLPE and EPR, will require separation to determine the state of both insulations correctly. In this case, normal EPR results could mask problems in the XLPE segment.

8.6.2.2

Dielectric Spectroscopy

Measurements of capacitance and dielectric losses at power frequency have been used for many years as a method of characterizing cable insulation systems. More recently, measurements over a range of frequencies have been studied. Frequency domain dielectric spectroscopy reveals information about the degree of degradation (water treeing) of aged XLPE cable systems; although cable insulation materials should ideally have an infinite insulation resistance, in practice they demonstrate small conduction currents. As the insulation ages with time, the conduction current tends to increase as a result of aging-induced changes, including oxidation. These changes affect results at different frequencies, depending on the severity of aging. This test method involves measuring dissipation factor over a range of frequencies at several voltage levels. At each voltage level, “swept” frequency measurements are performed at frequencies such as 0.1, 0.2, 0.5, and 1.0 Hz. This was the procedure used in the EPRI report Advanced Diagnostics: Estimation of Life of Extruded Cables (1001727) [88]. The development team for this technology [89] measured in the frequency domain from 0.0002 Hz to 100 Hz (at 20-kV peak).

8-19

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

The data developed have been characterized in several manners in terms of dielectric response of cables without water trees, and several categories for XLPE with water trees XLPE [89], as follows: 

Low-loss linear permittivity. Characterized by an almost frequency-independent capacitance. A linear response across a wide frequency band indicates that little or no aging has occurred.



Voltage-dependent permittivity. Characterized by increases in both capacitance and dissipation factor with increasing voltage, but essentially, the increase is independent of frequency. This response is characteristic of cables in which water tree deterioration is significant, but the trees have not penetrated the insulation wall.



Transition to leakage current. At low voltage levels, the response is similar to the voltagedependent permittivity response, but at higher voltage levels, the dissipation factor losses increase with voltage and higher leakage current occurs at both high and low voltages.

Numerous dielectric spectroscopy curves are presented by Werelius et al. [89]. EPRI report 1001727 [88] refers to responses defined as different types: 

Type 1. A voltage-dependent increase of dissipation factor and capacitance, almost independent of frequency



Type 2. At low voltage levels, a type 1 response, but at higher voltage levels, a nonlinear loss versus frequency characteristic



Type 3. At low voltage levels, a nonlinear type 2 loss versus frequency characteristic

A set of curves for a shelf-aged cable (never energized) is shown in Figure 8-13 [88, 89]. In this example, there is a slight change in dissipation factor with frequency, but change with voltage is essentially constant at any frequency.

Figure 8-13 Dielectric Spectroscopy Measurements for Shelf-Aged (Never Energized) 15-kV CrossLinked Polyethylene Cable

8-20

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

In essence, for unaged extruded cables, capacitance and losses are linear with voltage increases, but higher losses and increasing nonlinear responses are associated with aged cables containing water trees. Further details are provided in the EPRI report Guide for Non-Destructive Diagnosis of Distribution Cable Systems (1001731) [90]. Dielectric spectroscopy is currently offered by at least one testing company. Skilled practitioners are required to interpret the results. 8.6.2.3

Off-Line Partial Discharge Measurement

PD testing can be performed at very low frequency or at 50 or 60 Hz. Testing at 0.1 Hz is described in IEEE 400.2 and 400.3 [56, 91]. As with other test methods, the choice of frequency (line or very low) is mainly logistical, and the portability of VLF test sets make it the method of choice. PD is a breakdown of the gas in a void or gap in cable insulation or at its interface with a shield. It is characterized by a pulse of short duration (perhaps tens of nanoseconds) that travels along the cable in both directions. The pulse is characterized by numerous high frequencies that become attenuated as they travel. By measuring the time that elapsed between the pulse arrival times of the incident (the direct signal from the discharge) and the reflected signal (the reflection of the signal traveling the opposite direction from the discharge site, reflected from the far termination), it is possible to determine the location of PD along the length of the cable. This is referred to as a time domain measurement approach. Because the PD signals are quite small, circuit attenuation can be a significant problem. EPR and butyl rubber insulations cause signal attenuation, and slight corrosion of helical tape shields cause them to act like inductors, causing severe attenuation. A calibration pulse from the test set should be used before performing any PD testing to determine whether a reflection from the far terminal occurs and whether satisfactory results can be expected. Figure 8-14 shows the calibration equipment used before performing PD testing, and Figure 8-15 shows the test setup for performing 60-Hz PD measurements.

Figure 8-14 Calibration Equipment for Off-Line Partial Discharge Testing

8-21

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Figure 8-15 Test Setup for Performing 60-Hz Partial Discharge Measurements

A typical test involves the exposure of the cable system to a brief overvoltage, up to 2V0. During the period of voltage increase, the system looks for the onset of PD. The voltage at which this occurs is known as the partial discharge inception voltage (PDIV). The elevated voltage should be maintained no longer than is required to obtain the necessary partial discharge data. A second and equally important point in the test is the partial discharge extinction voltage (PDEV). In a good cable, the PDIV is greater than the peak test voltage, so that PD is not initiated. In an aged cable, PD, if present, will initiate well above the operating voltage but at less than the peak test voltage. The PDEV should also be well above operating voltage, but it will be significantly less than the PDIV. If the PDIV is too close to the operating voltage and the cable is returned to service, a small voltage surge during operation can start the PD, and the PDEV is likely to be below the operating voltage, so that PD does not stop. Insulation failure could then occur in weeks to months. The level of overvoltage applied and the duration of its application is important. The overvoltage and duration should be as low and as short as possible—but not so low or so short that the test misses the defect. Figure 8-16 shows a voltage test profile for the short duration test, with an overvoltage of 2.5 per unit and a test time of <12 seconds. This is as short as is reasonably possible. Not all testing is performed in such a short period. Figure 8-17 shows the diagnostic data capture approach using reduced dwell time at test voltage.

8-22

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Figure 8-16 Test Profile for Short Duration Off-Line 60-Hz Partial Discharge Testing

Figure 8-17 60-Hz Partial Discharge One-Step Diagnostic Data Capture Approach

A sinusoidal 0.1-Hz test set can also be used to evaluate PD at defect sites. As with 60 Hz, the traveling wave is used to determine the location and magnitude of the discharge and, as with 60 Hz, defect locations in the cable, splice, or termination can be located. The PD acquisition system is coupled to the cable to be tested through a coupling capacitor that detects the highfrequency electromagnetic impulses created by any PD. The PD signature and PD site location can be detected using this method. The VLF sine wave generator is portable. The cable system must be isolated for testing. A typical test will last for less than 10 minutes with peak applied voltage levels of 1V0 to 2V0. Compared to 60-Hz test equipment, the required reactive power for testing a cable at 0.1 Hz decreases by a factor of 600. A reduction in the kilovolt-ampere requirements for the test transformer and a corresponding weight reduction occur; therefore, improved portability makes the 0.1-Hz dissipation factor equipment of practical interest for in situ measurements. In addition, very low frequency leads to more direct and quicker channel growth of electrical trees [86].

8-23

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

If the failure mechanism for a cable issue results in PD, the PD test can determine whether the cable is free from PD at the test voltage, and if not, it will indicate the PDIV, PDEV, and location. In addition, some testers use phase-resolving methodology that provides an indication of where the PD occurs on the ac wave. The phase angle location and density provide insights into whether the discharge is occurring along a surface, at a shield–to-insulation interface, or in the insulation itself. This information helps determine the importance of the PD and whether it will cause failure in a short period. Advantages of the VLF PD test are the following: 

The location of PD activity can be detected and measured.



Cable system insulation condition can be graded as good, defective, or highly deteriorated when the measurement data are compared against historically established cable system PD data that have been developed by the same test method.



Compared to 60-Hz ac PD detection, the VLF test equipment is more compact and transportable (no van is required).



Power requirements are comparable to standard cable fault-locating equipment.



If desired, a combined VLF withstand test and PD test system can be used to monitor the dissipation factor or PD activity, or both, during a 15- to 60-minute withstand test procedure.



Phase-resolution technology allows the type of PD to be identified, which helps in understanding the nature of the failure mechanism.

Potential limitations and disadvantages of this test method are the following: 

There are length limitations (1–2 miles) for accurate measurements. These should not cause a problem for most plant circuits, given their relatively short lengths.



Deteriorated helically wrapped metal tape shield conditions and cable insulation attenuation characteristics can influence the propagation of PD pulses along the cable, thereby influencing detection and locating of PD sites. This could significantly affect detection of PD in wet segments of cable, where wetting would eventually cause corrosion of the tape shield, a common plant shield configuration.



PD data collected with VLF might not be comparable to what is obtained with 60-Hz power frequency data. Because of the slow repetition rate, fewer cycles occur during the test, and fewer PDs are observed per unit period.



Testing of mixed systems (XLPE joined to EPR, for example) can complicate interpretation.



Detecting the existence of water trees during PD testing is unlikely. The string of microvoids connected by oxidized channels that make up a water tree will not discharge under elevated voltage stress when ion-containing water resides in the tree channel. In addition, discharges require an air gap that could not exist if water filled the void.



The cable must be taken out of service for testing to be performed.

8-24

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

8.6.2.4

AC Withstand Testing

AC withstand testing can be performed at line frequency or very low frequency. The test purposely applies an elevated voltage for an extended period, with the aim of failing a weak cable insulation at a time when replacement is more convenient than during periods of high energy cost during the year. VLF is normally used because the test set is more portable. For a cable to be deemed adequate for return to service, the cable circuit under test must withstand a specified voltage applied across the insulation for a specified period of time without breakdown (dielectric failure) of the insulation. A sound cable does not fail. An ac withstand test, as opposed to a dc withstand test, applies voltage stresses to the cable similarly to that of operating voltage. A withstand test of sufficiently high voltage will detect the weakest spot in the cable, no matter where it is located along the cable system length. During a VLF test, an electrical tree is induced at the site of an insulation defect and is forced to penetrate the insulation. The test method and the significance of proper grounding are described in IEEE Std. 400.2-2004 [56]. The test times and voltage stresses have been questioned and are the subject of ongoing study; however, the voltage levels are generally between 2V0 and 3V0 for cables rated between 5 and 35 kV, and the test times range from 15 minutes to one hour. The two main types of 0.1 Hz waves generated—sinusoidal and trapezoidal—are described in the remainder of this section. For further details see IEEE Std. 400.2-2004 [56] and the EPRI report High Voltage, Low Frequency (0.1 Hz) Testing of Power Cable (TR-110813) [92]. The sinusoidal voltage wave is generated by amplitude modulation of a higher-frequency signal transforming up to the required high voltage and demodulation of the signal to produce the 0.1-Hz wave. A schematic of the sinusoidal wave is shown in Figure 8-18.

Figure 8-18 Nominal Very-Low-Frequency (0.1-Hz) Sinusoidal Waveform

8-25

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

The test set voltage is raised to the specified value [56, Table 5] for the specified period. If no failure occurs, the voltage is returned to zero, the cable system and test set are discharged, and the cable system is grounded. If a failure occurs during the test (that is, the test voltage drops) the test set should be turned off to discharge the cable system and the test set. The trapezoidal wave (cosine rectangular) is generated as follows: A dc test set forms the high voltage source with a dc to ac converter (consisting of a high-voltage inductor and a rectifier), changing the dc voltage to the VLF ac signal. The signal has polarity reversals at five-second intervals, which generates a 0.1-Hz bipolar pulse waveform. When testing with a pulse bipolar waveform, the test voltages should be peak voltages rather than rms. A schematic of the trapezoidal wave is shown in Figure 8-19.

Figure 8-19 Trapezoidal (Bipolar Rectangular) Very-Low-Frequency (0.1-Hz) Waveform

Advantages of VLF withstand testing are the following: 

Continuous polarity changes minimize the possibility of space charge buildup within the insulation wall, thus reducing problems related to localized overstress upon return to service.



A defect in the insulation will be detected at a much lower voltage than with a dc highvoltage test.



Mixed systems—from a plant cable perspective, such as circuits composed of a combination of butyl rubber, EPR, and/or XLPE insulated segments—can be tested.



Defective splices and terminations can be identified.



The size of a VLF test set is comparable to that of a dc test set.

8-26

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Disadvantages of VLF withstand testing are the following: 

VLF testing has been reported as working best when locating and eliminating a few defects in essentially good insulation (that is, long water trees); it is unlikely to detect several distributed small defects (multiple short water trees).



The system must out of service during VLF testing.

Additional points requiring consideration and that should not be overlooked when performing VLF withstand testing include the following: 

Information on cable constructions and route maps should be studied in advance,



If a failure does occur, fault location detection is necessary to identify where the circuit has failed.



Because the voltage stress and application time parameters are not (yet) definitive, the results might not be conclusive (that is, degraded insulation might not fail under test or be taken only partway to failure). Work continues on determining the most appropriate test voltage and test duration.

8.6.2.5

Tests Under Development

Several additional diagnostic off-line tests are available, but they are not commonly used and can be considered to be under development. Descriptions of some of the promising tests are provided in Appendix E, Off-Line Tests That Are Under Development. 8.6.3 On-Line Diagnostics Assessment On-line cable assessment is based on analysis of high-frequency signals that are emitted by energized systems; detecting these signals, interpreting their significance, and separating them from noise is the key to the technology. Since the data are collected while the system is energized, the method is considered passive, in the sense that it cannot lead to failures that might occur during diagnostic and withstand testing methodologies [93–95]. This type of test cannot detect the existence of problems that would be excited just above operating voltage that would be indicative of degradation. Most work with this process has been performed on distribution cables operating at 13 kV and higher. Signals result from a displacement current due to internal charge transport in a gas inclusion within a solid dielectric. This current can be observed and measured externally; it has a rapid appearance and rise time and is influenced by conductor, insulation, and shields, as well as by discontinuities. These signals appear at various frequencies [96]. One limitation of this on-line assessment methodology for medium-voltage cables described by Montanari [77] is the inability to detect global degradation such as water treeing in wetted cables. Water-filled water trees do not contain gas and would not emit such signals. If this is the primary cause of degradation for a particular cable circuit, then the off-line tan δ diagnostic methodology would be a better method to detect degradation. For 4-kV rubber insulations, conversion of water-related degradation to electrical treeing (partial discharging) might not occur for submerged cable, because the degradation sites are filled with water, and a gas-filled void does not exist where pre-PD and PD would occur. Failure is more likely to be caused by thermal 8-27

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

runaway at the degradation site. Thermal runaway would occur if the leakage current in a lowresistance zone in the insulation increased to the point at which the heat generated in the insulation could no longer be dissipated by the surrounding medium and excessive insulation temperature occurred. Montanari [77] cautions that false positives could occur due to the dependence on high signal-tonoise ratios required for the sensors used for on-line evaluations. In addition, errors can occur due to the discontinuous nature of PD signals. Extending sampling time or continuous monitoring can reduce these errors, but that is not necessarily a practical solution. The occurrence of false positives can be caused by external noise detected by the sensors. The on-line assessment physically places sensors at readily accessible locations along the cable circuit. For example, for underground testing, access through manholes is used. Data from the signals are transmitted to the home office of the vendor and, after analysis, the condition of the cable system between sensors is determined. The sensors are moved to different locations as testing continues, until the entire desired length of cable is assessed. This approach allows localized regions to be evaluated (segments between sensor placements). Because cables age at different rates along their lengths, independent of age, the technology is able to distinguish between the degrees of degradation in segments of the cable. The technology estimates the condition of all equipment between sensors, such as joints or terminations. High-frequency signals during cable system operation result from current flow in the cable shield and conductor. The magnetic field resulting from the current flow is measured. The sensors are designed to pick up energy from the magnetic field and detect signals in the highfrequency range. The sensors use both inductive and capacitive coupling, which allows signal detection over a wide frequency range. The sensitivity of the inductive coupler depends on the construction and condition of the shield and the size of the cable. This is in contrast to capacitive coupling, in which the signal sensitivity depends primarily on the cable capacitance. The sensors are U-shaped, and various sizes exist to cover commonly used cable configurations. Slight corrosion of helical tape shields will lead to significant attenuation of high-frequency signals along the length of the cable. Signals detected can be due to PD or pre-discharge events. The literature generally refers to all signals detected in this manner as partial discharge detection. Significant signals can result from mechanical damage, foreign impurities, poor splice or connection regions, water intrusion into splices or terminations, or voids. This approach can, therefore, be described as seeking to 1) acquire signal information due to defects, 2) process the signal information to reject noise, and 3) analyze the signals. The U.S. supplier of this technology provides results in terms of the quality of the cable. Technical detail concerning the nature and amplitude of the signals is not provided. Figure 8-20 shows the sensors for the detection system and the data acquisition system. Figure 8-21 shows the on-line condition assessment data acquisition system. Figure 8-22 shows signal patterns and the influence of various defect types.

8-28

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Figure 8-20 Sensors for On-Line Signal Detection and Data Acquisition System

Figure 8-21 On-Line Condition Assessment Data Acquisition System

8-29

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests

Ovals represent sine waves; vertical lines represent signals. Figure 8-22 Influence of Defect Type on Signal Patterns Detected During On-Line Evaluation

If pre-PD or PD is present, the patterns indicate the causes of the signals, allowing personnel to determine their significance relative to future performance. The pulse pattern is more significant in determining relevance to future performance than is the apparent discharge magnitude. The latter, while of interest and value, does not allow distinguishing between defect types that cause the specific signal. After evaluation by the vendor’s technical personnel, the condition of the cable system is placed into categories. The most significant level suggests that immediate action is required, whereas intermediate levels suggest potential retesting after various timeframes. The lowest level represents cables that show little or no aging. In addition to the benefit of detecting defects at operating voltage while the cables are energized, the on-line method purports to provide the following advantages compared to off-line testing methods: 

The type of defect can be identified.



Testing is performed at operating temperatures because cables undergo physiochemical changes as the temperature is increased. Accordingly, on-line assessment evaluates the cables under the conditions of operation, not under conditions following cooldown.



Testing does not require disconnecting and reconnecting of terminations.

Disadvantages of the on-line diagnostic testing are as follows: 

Defects such as water treeing in wetted cables cannot be detected.



Analysis can be performed only by the vendor, so that the technology cannot be transferred to the utility.

8-30

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests



This methodology does not overstress the insulation, so that a gross defect close to breakdown but not producing PD signals might not be detected. Such defects will not be affected by operating voltage but can cause failure if a significant surge voltage occurs.



If the ultimate failure mechanism is thermal runaway and no PD is expected, this methodology will not detect the aging of the cable.

This assessment method attempts to identify extremely small signals, in particular, pre-PD signals. Accordingly, attenuation of the signal can impede the ability to identify adverse conditions. When attenuation is an issue, the sensors must be applied at several spots along the length of the cable to ensure that attenuation does not mask problems. Except for short cables, making measurements at both ends of the cable is highly desirable. Even for cable lengths in the vicinity of common large power plant cables (500 ft [164 m]), testing from one end of the circuit might not fully assess the circuit.

8.7

Applicability of Tests

Table 8-5 provides a list of the more important tests and their applicability at various times in the life of a cable, from manufacture through end of service. The individual tests might have different applicability under special conditions of service or when different failure mechanisms than those listed are of concern. Table 8-5 Matrix of Applicability of Tests During the Life of a Medium-Voltage Cable Manufacturing Test Insulation Spark Gap Test

X

Conductor DC Resistance

X

Jacket Spark Test

X

Partial Discharge

X2

AC Withstand Test

X3

DC Insulation Resistance Test

X4

Acceptance Test

Installation Test

Diagnostic Test 1

Fault Locating And Troubleshooting

X4

X4

X4

X4

X5, 6

X5. 6

DC Withstand Test

X5

60-Hz or Very-LowFrequency Withstand Test

X5, 6

On-Line Partial Discharge / Signal

X

8-31

Testing: Manufacturing, Installation, and Maintenance or In-Service Tests Table 8-5 (continued) Matrix of Applicability of Tests During the Life of a Medium-Voltage Cable Manufacturing Test

Acceptance Test

Installation Test

Diagnostic Test 1

Off-Line Tan Delta

X

X

Off-Line Partial Discharge

X

X7

Fault Locating And Troubleshooting

X

Resistance Bridge (Murray Loop Test)

X8 (Fault Location)

Capacitance Ratio

X9 (Fault Location)

Time Domain Reflectometry

X8 (Fault Location)

Notes: 1. For initial diagnostic testing of aged cables, it is recommended that spare cable, accessories, and installation procedures be available before testing should a replacement be necessary. 2. This test is generally performed between 500 and 5000 Vdc. Its value is in troubleshooting and verifying that the cable under test is capable of withstanding higher test voltages used for PD, tan δ, and high-potential tests. It should not be used as proof of suitability for service. 3. This is a 5-min, 200 V/mil test. 4. This test, described in Section 8.3.1.2, is not performed on Kerite brown EPR cables because the lossy nature of the design makes the test meaningless. An acceptance criterion is <5 pC, but some exceptions are noted in the text. 5. An ac withstand test would be the preferred method, except for paper-insulated, lead-covered cables. VLF requires a smaller test set than 60 Hz testing. DC high-potential testing of new installations is acceptable from the standpoint of not damaging the cable, but it is not recommended for service-aged cables. Test values should be in accordance with IEEE Std. 400. 6. Test values after manufacture are generally 80% or factory test value for acceptance and installation tests and 60% for maintenance tests. 7. PD testing will identify degradation of cables. Test success depends on the quality of the metal shield, so that attenuation due to the high impedance of the shield will not preclude measuring the PD signal. 8. If the cable fault is solidly grounded or can be carbonized sufficiently with a high-potential set so that it is solidly grounded, this test is highly effective in determining the location of a fault. 9. The capacitance ratio test is an effective means of locating a fault for a cable with an open circuit.

8-32

9

CABLE AGING MANAGEMENT PROCESS The cable management process involves obtaining information on the cables installed at the site, determining the adverse environments and service conditions, knowing how those conditions affect the specific cable designs, and applying the best test methods available to determine the degree of degradation. This section describes the diagnostic approaches and the actions to be taken to manage cable aging.

9.1

Strategies and Philosophies

The strategy for management of cable aging can be dictated by the cable circuit design. Strategies and philosophy for shielded and nonshielded cables differ because there are no demonstrated electrical diagnostic test techniques for nonshielded cable that produce meaningful results. In addition, withstand testing of nonshielded cables is not recommended due to the danger of causing damage to otherwise good insulation and the fact that such testing will provide no useful information unless the ground plane can be brought to the surface of the cable, which is not practical in field applications. The EPRI report Aging Management Program Guidance for Medium-Voltage Cable Systems for Nuclear Power Plants (1020805) provides guidance for development of a medium-voltage cable aging management program [5]. Contingency plans should include having replacement cable and termination components, replacement procedures, and a source of personnel skilled in the removal and installation of medium-voltage cables to preclude lengthy outages that can result from the failure of important cable circuits. 9.1.1 Test Strategies for Shielded Cable Circuits Several approaches to aging management exist for shielded cables. EPRI report 1020805 [5] recommends prioritizing medium-voltage cable assessment and testing based on the presence of adverse environmental and service conditions, such as the presence of water at the cable surface, high ambient temperature, high radiation, severe ohmic heating, and high-resistance connections. Cables subject to these conditions should be assessed or tested using an appropriate test method. The appropriate test method is based on the nature of the adverse environment or condition, the cable design, and the expected failure mode. Based on that analysis, performing a diagnostic test such as tan δ measurement, dielectric spectroscopy, or PD testing or an ac withstand test or both might be appropriate. If thermal or radiation damage is suspected, visual inspection can be useful. Infrared thermography can be useful for determining the severity of ohmic heating and high-resistance connections. Visual inspection can also be useful. Cables that are not subjected to adverse conditions are expected to have long lives. Testing of such circuits is not needed in the near term. If an unexpected failure mechanism is found for such circuits or the service duration is extremely long (for example, 50 years), corrective action that includes testing is likely to be necessary. 9-1

Cable Aging Management Process

9.1.1.1

Run to Failure

For shielded cables subject to adverse environments, run to failure is not recommended as an option unless the circuit has no effect on operations or safety, including the considerations contained in 10CFR50.65, Requirements for Monitoring the Effectiveness of Maintenance at Nuclear Plants (generally referred to as the Maintenance Rule) scoping processes [97]. 9.1.1.2

Diagnostic Testing

To reduce the likelihood that a deteriorated section has been only partially driven to failure at the end of the withstand test, the stability of the tan δ measurement during the withstand can be monitored if the test set has a tan δ monitor. If the tan δ measurement varies during the test, a degradation site might have started into its final stages of failure, and increasing the duration of the test might be warranted. If the tan δ is stable, no sites are being driven to failure, and a reasonable period of failure free operation can be expected. The diagnostic tests that are recommended in EPRI report 1020805 [5] include tan δ, dielectric spectroscopy, and off-line PD, depending o the cable type, cable condition, and failure mechanism of interest. These tests are applied at greater than normal line-to-ground voltages but below voltages that would cause a partially aged cable to fail. Any elevated voltage test could cause a highly degraded cable to fail. However, highly degraded cables are at or near the point of in-service failure and should not be in service. Depending on cable characteristics (such as design and shield impedance) and environment (dry or wetted) a properly chosen diagnostic test will provide information on the level of degradation of the cable nondestructively. Care must be taken to understand the tests under consideration and their applicability to the cable design and condition. For example, if aging of cables in wet environments is the concern, then either tan δ or dielectric spectroscopy testing is an appropriate diagnostic method, especially for EPR insulated cables with copper tape shields. Cables that have aged under wet conditions do not emit PD signals during the aging process. If the cable remains wet, there are no gas-filled voids where PD will occur. For 4-kV applications, the voltage levels in the insulation might be too low to excite PD and cause it to continue. In addition, even if PD were to occur, PD testing can be impeded by the naturally higher impedance of EPR and the attenuation of the high-frequency signals resulting from oxidation of the helically wrapped taped shields, causing the PD to be difficult to detect. This attenuation does not affect tan δ and dielectric spectroscopy testing or the operation of the cable. When performing tan δ testing on circuits with segments having different insulation types, such as EPR and XLPE, it is necessary to separate the segments for testing. The normal characteristics of a good EPR section will mask the condition of a bad XLPE segment. Accordingly, the use of separable connectors between dissimilar cable types is recommended to allow separate testing of the segments. For off-site feeds from distribution systems (for example, a 34-kV source) using a distribution type cable with concentric neutrals and multiple splices due to its length, PD testing is appropriate. These types of cables are more likely to exhibit PD either in the cables or the splices, and PD testing could identify deterioration. Light corrosion of the concentric neutral would not cause significant attenuation. 9-2

Cable Aging Management Process

PD testing can also be useful for 13 kV UniShield cables that have distributed shield wires that should not be subject to a high-impedance shield condition if the wires corrode slightly. In addition, PD testing can be useful in identifying long-term aging (after 50–60 years) in dry cable sections of all types of shielded cables. However, such testing is not needed until operating experience indicates that the dry sections of cable are exhibiting signs of aging. In the case of most EPR cable designs, PD might never occur or might occur only when the cables are close to failure. In the latter case, no PD would be detectable for decades in the cable’s life. Detection would be possible only during a short period near the end of the cable’s life. Periodic testing for PD with frequencies on the order of once every six years would have a low probability of usefully detecting PD, because the window between the start of partial discharging in the insulation and the point of failure is a fraction of the period between tests. Detection of PD would be happenstance rather than an expected outcome. If detected, the cable in question would likely need immediate or near-term replacement because the presence of PD in the insulation would indicate that the final stage in the degradation mechanism is in progress. Brown EPR insulated cables are designed for discharge resistance, and some of these cables exhibit levels of PD that would be unacceptable in other cable designs but cause no significant degradation. Use of PD testing on these brown EPR insulated cables can be of little use and can cause needless concern about the longevity of the cable. 9.1.1.3

Withstand Testing

VLF withstand testing in accordance with IEEE 400.2-2004 [56] requires applying a voltage higher than the rated line-to-ground voltage for a recommended period ranging 15 to 30 minutes. (The IEEE ICC Working Group is evaluating longer durations at somewhat lower voltages for their ability to drive significant degradations to failure without causing unnecessary failures.) These test values and durations have the capability to fail a marginal cable in which a localized flaw was near failure at the start of the test. That is, some useful life might have remained before the test, but the cable was almost assuredly degraded before testing and likely to fail in subsequent operation. The concept of this test is that good insulation will easily withstand the high-potential test, whereas degraded cable will purposely fail during the test. Another use of a withstand test is to couple it with a diagnostic test when the diagnostic test indicates that degradation is present. There is some question regarding how long a degraded cable could remain in service without failure. The state of the art in diagnostic tests can identify the need for replacement, but it cannot predict precisely how long a cable will remain functional if it is somewhat degraded. The advantage of applying a withstand test is that it should cause failure under test of cables that are at risk of failure under normal operating conditions in the near term (for example, an operating cycle or two). The decision to use a purposely destructive test is to drive a significant degradation site to failure at a time when the replacement can be addressed in a controlled manner rather than being subjected to the potential for an in-service failure at a time when the power plant’s output is in high demand. A withstand test by nature is a go/no-go test, and although it provides assurance for continued service for some period of time, this test provides no information regarding the amount of cable degradation.

9-3

Cable Aging Management Process

9.1.2 Nonshielded Cable Circuits NEI 06-05 [11] identified that 30% of the <5-kV rated circuits have nonshielded cable. The options for what aging management strategy can be applied for this population of cables is limited because of the lack of useful diagnostic test methods. This is due to the lack of a uniform ground plane for testing, because there is no insulation shield. The EPRI report Aging Management Program Guidance for Medium-Voltage Cable Systems for Nuclear Power Plants (1020805) [5] describes the possibility of creating a uniform ground plane for nonshielded cables by intentionally flooding the cable ducts and verifying that the water was grounded. This would then allow performance of withstand and tan δ diagnostic tests. Although this methodology has been successful in the laboratory, it is not yet been attempted in the field. The following three strategies for aging management of underground, nonshielded cable circuits are possible: 

Commit to the performance of a rigorous forensic assessment if a cable failure occurs to determine whether the failure was related to aging or some other cause and take corrective action as appropriate for like circuits. Consideration of operating experience from other plants and research on like cables would be included in this approach.



Identify a worst-case cable, remove it for laboratory assessment, and base further actions on its condition.



Use a time-based replacement strategy.

The first approach might be the only useful approach until industry operating experience indicates that nonshielded cables of the same type are beginning to experience age-related deterioration. To date, the failure experience for nonshielded cables is equal to or better than that of shielded cable. As plants age, the inability to confirm the condition of nonshielded cables causes increased uncertainty about long-term functionality. Plants using this approach should have contingency plans available, including a ready source of replacement cable and termination components, replacement procedures, and a source of personnel skilled in removal and installation of medium-voltage cables. In addition, there must be a commitment to engage an experienced cable forensics laboratory to perform a detailed forensic analysis to determine whether the failure was age related and whether additional cables in like service should be replaced. As plants with nonshielded cables continue to age, the second and third options should be considered. The second option requires that the cables in the severe environment (wetted, direct buried, and so on) cable runs be reviewed to determine which circuits had the worst-case conditions. For example, cables that were continuously wet, energized, and unloaded, and cables that were continuously wet and highly loaded would be considered worst-case conditions. In the first case, the voltage and water that could produce water-related degradation would be present, but there would be no heating to drive off moisture. In the second case, there might be enough heat to reduce moisture, but also enough to drive temperature-related degradation. Selecting one or more circuits for removal of the wet section for laboratory assessment would be an approach useful to a plant with many underground cables. If the plant had few underground cable circuits, the replacement option might be more useful. Alternatively, if the plant could prove that all underground circuits were dry, then remaining with the first option might be more useful. 9-4

Cable Aging Management Process

An acceptable alternative might be to use industry operating experience of forensic analysis done on similar cables under similar conditions. Forensic analysis performed at another plant on shielded cables with the same insulation design and similar service conditions can also be a suitable source of information for evaluating the condition of a site’s nonshielded cables. Alternatively, the third option, time-based replacement, could be used. A time-based approach for nonshielded cable could be broken down into different strategies as follows: 

Replace all cables.



Prioritize replacement based on various risk factors (safety, business, regulatory).



Prioritize replacement based on environmental risk factors such as thermal, radiation, chemical, or wetted environments (localized adverse environments).



Replace all cables at an arbitrary end-of-life. This strategy could be used in plants with a small population of wetted medium-voltage cables. Prioritized replacements could be used for plants with larger populations of wetted cables. It would be prudent to perform laboratory evaluations of some of the removed cables to determine whether a delayed replacement schedule were possible based on as-found condition of the removed cable.

When replacing a nonshielded system, replacement with a shielded cable design should be considered. However, the design of the system must be considered to ensure that use of a shielded cable is acceptable. In addition, water-impervious cables should be considered as replacements. The use of shielded or water-impervious cables can be precluded by the design of the duct system, pulling restrictions, and cable availability. Water-impervious designs are available that use a corrugated tape shield that is folded over the insulation system and glued shut. Additional information is provided in the EPRI report Plant Support Engineering: Common Medium Voltage Cable Specification for Nuclear Power Plants (1019159) [12]. EPRI will begin submergence qualification for certain cable types based on operating experience and on-going qualification as described in IEEE Std 323-1974 [98], Sections 5.2 and 5.6. A cable that has been submerged for nearly its entire life will be removed from service and subjected to electrical and physical characterization to provide the operating experience portion of the qualification. Cable specimens will then be prepared from the remainder of the cable and will be subjected to submerged, energized accelerated aging in a laboratory to provide the ongoing qualification. The rate of acceleration is expected to reasonably low and require an extended period to complete. The submergence qualification is expected to yield a basis for the expected (qualified) life of the cables so that a basis for continued use or an appropriate replacement interval will be determined. None of these strategies preclude the possibility of a cable failure, so it is highly recommended that plant personnel take action to be prepared should an in-service failure occur.

9.2

Prioritization of Cables for Assessment and Testing

The size and nature of the cable population and the nature of the adverse environments and service conditions must be considered. Cables with the greatest safety and operational importance that are subjected to adverse conditions should be given the highest priority. Earlygeneration cables (those manufactured before 1975 to 1977 having XLPE, butyl, or black EPR insulation) should be given higher priority. Later-generation cable having silane-treated clay or 9-5

Cable Aging Management Process

discharge-resistant insulation (cables produced by Kerite) could have lower priority. Normally energized cables having limiting conditions for operations, such as those connecting startup and auxiliary transformers to busses, should have high priority, as should safety-related cables, especially those that are continuously energized. The prioritization method should include consideration for criticality of the connected components, as defined by INPO AP-913, Equipment Reliability Process Description, [99] and risk significance, as defined by 10CFR50.65 (Maintenance Rule) [97]. Cables that are not Maintenance Rule risk-significant can be given low priority. However, care must be taken that protective relaying and circuit breakers associated with non-risk-significant cables and loads are included in plant maintenance and test programs to ensure that they function when called on should a non-risk-significant cable fail. Prioritization could be further associated with environmental stressors (wetted environment, elevated temperature) and their severity, the cable and accessory design (water impervious, insulation type, and so on), site-specific and industry operating experience, and duty cycle (normally energized and loaded, normally energized and unloaded, periodically energized, or normally de-energized). Weighting factors can be assigned for each of the factors that plant personnel consider important to longevity. The summation of the weighting factors would indicate the relative priority of cables with respect to aging. This prioritization could be used as a method to prioritize cable testing, refurbishment, and replacements. 9.2.1 Risk Ranking Methodology This section provides a method that could be used for prioritization. It is an example, not a requirement. 9.2.1.1

Maintenance Rule and Criticality Screening

Using screening criteria from Maintenance Rule scoping, before applying significance factors and weighting factors, will reduce the number of cables that must be analyzed. The screening criteria identify cable criticality based on the criticality of the source or the load. This is important because a cable failure can result in loss of load and sometimes the source. If a nonsafeguard load, especially one that is noncritical or run to failure, could trip a safeguard feeder circuit that is critical, the cable should be considered critical. Criticality screening should include Maintenance Rule scope and risk-significance classification if criticality scoping does not. Cables that feed or are fed by components with a risk-significant designation should be of equal risk rank as critical components. The highest risk factor would be those that are critical and risk significant, followed by those either critical or risk significant, then noncritical, and then run to failure.

9-6

Cable Aging Management Process

9.2.1.2

Insulation Type

Insulation type significance is based to some degree on operating history and known manufacturing process improvements that have occurred over time. This combination is evident in that XLPE insulation manufactured before 1975 and black EPR manufactured around the same time should have a high significance factor with respect to aging [5]. Butyl rubber and cables having a compact design such as UniShield should be ranked high. Pink, brown, and gray EPRs would be lower-significance insulation types. TR-XLPE cables have not accumulated sufficient operating time but are likely to have significance factors similar to those of pink EPR. 9.2.1.3

Significance Factor for Jacket Types or Water-Impervious Designs

Cables with sealed metal jackets (lead or aluminum) or having a water-impervious design used in newer cables with sealed, corrugated metallic shields are less susceptible to water degradation compared to other types of cables. Most jacket materials provide some barrier to water absorption, but they do not stop water migration. Lead-sheathed cable was used in few plants. Sealed, corrugated metallic shields have just started to be used in some applications starting in about 2005. 9.2.1.4

Significance Factor for Operating Experience

A higher significance factor should be applied to cables of the same type in similar conditions as those that have had a failure or several cable failures in other plants. 9.2.1.5

Significance Factor for Diagnostic Test Results

Cables that have diagnostic test results indicating that they are in an aged but not severely degraded state should be given higher priority (reduction of the period between tests). Cables that have not been tested and are in an adverse environment so that their condition is unknown should be given higher priority than cables that have already been tested and found to be acceptable. Cables that have or have not been tested but are in acceptable dry locations should be valued at the lowest significance but should be assessed to determine whether they are exposed to adverse thermal, radiation, chemical, or oil environments. 9.2.1.6

Significance Factor for Voltage and Insulating Level

Cables with insulation thicknesses greater than that required for the operating voltages will have lower dielectric stresses within their insulation and should be significantly less at risk than cables with the minimum required insulation thickness for the applied voltage. 9.2.1.7

Significance Factor for Operating Conditions

Cables that are wet and normally energized are of higher significance than normally deenergized cables because the conditions could lead to water-related degradation. In dry environments, cables with high currents with respect to their ampacity could have adverse ohmic heating. For wet cables, the most rapid aging is expected to come from being energized with no load or the opposite extreme, energized with high load.

9-7

Cable Aging Management Process

9.2.1.8

Weighting Factor for Adverse Environment

Wetting has been identified as a factor in the highest percentage of operating experience failures for medium-voltage cables and should have the highest weighting factor for adverse environments. Thermal, radiation, ohmic heating, and high-resistance connections should be afforded weight values equally, unless a specific situation is known to be excessive or in combinations (such as a known high-temperature area (>122°F [>50°C]) and high radiation (5 Mrd [50 kGy]). In most cases, adverse environments affect specific portions of a cable run, but in the case of risk ranking, the worst-case condition that a cable circuit experiences should be used to determine the risk factor for the whole circuit. If more than one adverse environment exists, both weighting factors should be applied to the significance factor and the two results should be added together to account for the higher risk. Both electrical testing of the cable to determine the condition of the wet section and inspection of the dry section with the adverse environment or service condition will be necessary. 9.2.1.9

Weighting Factor for Current and Amperage Level

Sizing cables in nuclear power plants generally used derating factors for cables (safety-related and non-safety-related) so that ampacities were reduced by 20% or more. Reducing the ampacity to 80% results in low to moderate conductor temperatures and limited thermal aging. A weighting factor for operating current should be applied if it is significantly greater than 80% of the conductor’s rated ampacity.

9-8

10

RESPONSE TO CABLE FAILURES This section describes the actions and decisions that might be required if a medium-voltage cable fails.

10.1 Electrical Tests of Circuit 10.1.1 Logic for Testing to Be Performed The first decision that must be addressed for the cable is what test should be performed to determine its condition. The most common test to be performed first would almost surely be an insulation resistance test. The value of performing this test is that it might eliminate the need for a more sophisticated test if it indicates that the insulation resistance is low. For medium-voltage cable, insulation resistances that are less than 100 MΩ-1000 ft (30.5 MΩ-km) should be considered inadequate. Insulation resistances as high as 30 MΩ-1000 ft (10 MΩ-km) have been obtained on failed shielded cables because the fault blew out the shield and caused the path between the conductor and shield to have a high resistance. Insulation resistances less than 100 MΩ-1000 ft (30.5 MΩ-km) should be considered inadequate. Insulation resistances as high as 30 MΩ-1000 ft (10 MΩ-km) are direct indications that the insulation has failed or is highly degraded and should not return to service. However, even if the cable has greater than 100 MΩ-1000 ft (30.5 MΩ-km), an insulation resistance test should not solely be relied on for determining serviceability. A thin layer of good insulation in series with a near-through-wall degradation or defect will mask the problem and result in a high insulation resistance. Appendix F provides a basis for cable insulation resistance acceptance criteria. For nonshielded cable, insulation resistance testing might be the only practical means of assessing the cable. For shielded cable, diagnostic testing can be performed to characterize the insulation system. For wetted cables with helically wrapped metal tape shields, the recommended diagnostic test is either tan δ testing or dielectric spectroscopy. These tests grade cables as good, requiring further study, or action required. A good result would allow that phase or all phases to be used further. A requiring further study result would indicate that the phase is not failed but has weakened. An action required result indicates that the phase or phases are either failed or highly deteriorated and should not return to service. With respect to the requiring further study or a borderline action required result, a withstand test can be applied to determine whether the cause is a single severe degradation or several lesser distributed flaws. If the cable passes the withstand test, there is reasonable assurance that the cable will be satisfactory to operate for an interim period.

10-1

Response to Cable Failures

In some cases, acceptance criteria for a specific EPR formulation can be difficult to obtain. When no data are available for a particular cable manufacturer’s insulation, testing of a cable in a dry application is a possible way to establish a baseline for comparison. For example, a dry cable that connects to buses can be relatively easy to test. 10.1.2 Logistical Issues for Electrical Testing The test effort must recognize that lethal voltages and currents exist around energized mediumvoltage cable and cables under test. The pertinent safety practices must be carefully followed. When working in the vicinity of energized medium-voltage cables adjacent to those being evaluated, the possibility of arc flash must be recognized, and appropriate precautions must be taken. Meaningful tests of cables will require separating cables from loads and separating circuits that have multiple conductors per phase. Testing of cables that are paralleled can mask deterioration and would not indicate which of the cables was degraded. Many plants are obtaining their own VLF and tan δ test sets or have contracts in place with testing companies. If a utility has done neither by the time a cable failure occurs, understanding the types of available tests and which companies use them is critical. Each test company has its own test methods. Discussions with a prospective test company should cover the type of cable used at the plant, the parameters of the cable operation, and the suspected failure mode. Otherwise, valuable time can be lost if an inappropriate test method is tried and is not successful. Understanding the test to be applied by the service provider and whether it applies to the installed cable type is critical. Test companies that have tested plant cables, rather than just distribution cables, should be used. Operating experience has identified cables in a wet environment as the major at-risk group of medium-voltage cables. If the cables being investigated are routed underground and are accessible through manholes along the cable run or through underground vaults that might be flooded, pumping will be necessary. At a minimum, these locations must be dry if access to these locations becomes necessary for cable repair. Confined space procedures are necessary when working in manholes and, if other circuits remain energized, arc flash procedures should be used. Another consideration for coordination of testing and replacement work is how to deal with protected equipment that might be in proximity to the cable terminations or have associated cables in the same cable trays. The potential to lose numerous shifts while resolving operational concerns or restrictions related to protected equipment can be minimized or avoided if it is discussed when it is determined that off-line testing or cable replacement is required.

10.2 Fault Location Explicit fault location might or might not be necessary. If the entire length of the circuit must be replaced, there is no need for fault location. If a cable can be cut into two sections—such as the portion with the dry environment and the portion with the wet environment—knowing which section has the fault, such as the wet section, will allow replacement again without the need for explicit fault location. However, when runs are long and replacement of an entire run or section is not practical, fault location methodology is important.

10-2

Response to Cable Failures

If PD testing is used and attenuation of the signal is not an issue, PD test set will provide the location of the fault. If a fault causes a solid grounding of a cable, a time domain reflectometer (TDR) can be used to detect the location of the fault. However, the fault often burns out the insulation shield in its vicinity, leaving a high resistance to ground and limiting the ability to be detected by the TDR. High-voltage test sets are often used to burn the fault in so that it can be detected by a TDR or a fault-locating bridge. A thumper can also be used. A thumper charges a high-voltage capacitor and then discharges it into the cable. The thumper causes an audible sound (a thump) that can be heard from above ground at the location of the fault. However, a thumper will do additional damage at the failure site and sometimes destroy the insulation to the point at which identification of the cause of failure is not possible. Other techniques, such as the use of a Murray bridge or a capacitance bridge, can also be used [6].

10.3 Forensics and Failure Assessments The highest priority during recovery from a cable failure is to identify and repair or replace the damaged cable. However, care must be taken to preserve the failed cable section or component for forensic assessment. Gathering available forensic data and performing a thorough failure assessment will provide valuable information in determining whether the failure was from longterm aging, a manufacturing flaw, or installation damage. The results of the forensic assessment will provide a basis for the extent of condition considerations for the remaining cable population. For these reasons, care should be taken during the troubleshooting and repair to preserve as much information possible so that a proper failure assessment can be made. The single most important piece of information to preserve is the failed cable section or sections. As a minimum, analysis of the failed section of cable and adjacent sections of cable is recommended. A reasonable length of cable on either side of the fault, at least 3–6 ft (1–2 m), is recommended. Longer sections allow additional electrical and physical tests to be performed. Nonfaulted, adjacent cable phases should be preserved and assessed, as well. In addition, a dry section some distance from the fault or a section of cable from the warehouse should be obtained for determining baseline conditions. The sections of cable saved for analysis should not be handled more than necessary or cleaned. Pointing to the fault with a pencil can leave carbon that has nothing to do with the failure. Overhanding can lead to physical disruption of the internal configuration that could mislead the forensics team. Cleaning with solvents could alter the chemistry of the materials and lead to corrosion of conductors that, again, could mislead the forensics team. Digital pictures should be taken during removal of relevant areas of the failure. The segments saved to be analyzed should be wrapped, sealed in plastic, and quarantined with minimal handling to preserve their condition. Any damage that occurs as a result of the removal process should be documented and forwarded to the forensics team for consideration.

10-3

Response to Cable Failures

The forensics laboratory will use the dry or unaged cable to determine the characteristics of a good section of cable for comparison to the damaged section. The section adjacent to the fault will allow the laboratory to determine whether the adjacent phases had similar deterioration or whether the faulted phase had a unique problem. Removal and handling of the specimens should be supervised to ensure that the specimens are properly handled and preserved. Electricians are experienced in installing cables but are often not trained to preserve cables or terminations that have failed. In addition to the cable specimens, the laboratory should be provided with the physical and operational conditions that the cable or cables experienced in service. The following information should be provided to the failure analyst, if possible: 

Cable manufacturer



Cable ratings



Circuit length



Cable routing details (penetrations, bends, cable tray, conduit, direct buried, ducted, and so on)



Number of conductors per phase



Accessories and splices



Time in service



Operating voltage



Operating current



Operating duty cycle (continuous or standby)



Environmental conditions (wetted, flooded, near heat source, and so on)



Operating parameters at the time of failure (plant or system transient, equipment startup, shutdown, change in demand, and so on)



Related operating experience

Forensic analysis of the cable should be performed by a laboratory that specializes in cable analysis. Although many sites have access to corporate laboratories that provide failure analysis, these labs might not have the knowledge or capability to evaluate the physical and electrical properties of the cable specimens. For example, corporate laboratories might have experience only with XLPE insulation used in distribution systems rather than EPR with helically wrapped tape shields used in power plants.

10-4

Response to Cable Failures

The properties of the cable that should be evaluated by the laboratory will vary by the cable type, the manufacturer’s design, and the nature of the failure. The following physical attributes should be measured, if applicable, depending on the design of the failed cable: 

Visual condition of the cable and the failure location



Jacket type, thickness, and condition



Metallic shield type, if applicable (copper tape, corrugated tape, drain wires, and so on) and condition



Insulation shield type and condition



Conductor shield type and condition



Insulation type, thickness, and condition



Conductor type and condition



Tensile strength of jacket and insulation



Elongation of jacket and insulation



Moisture content of jacket, insulation, and conductors



Microscopic examination of insulation for imperfections

In addition to the physical properties, the following electrical properties of the cable should be gathered, as applicable: 

Jacket electrical resistance (ohms/cm)



Semiconducting shield resistance (if shield is different from jacket)



Dissipation factor at 60 Hz or VLF



PD



Insulation resistance (normalized to 1000 foot of cable) at 5000 Vdc (see Appendix F, Insulation Resistance Test Measurements: Their Value and Limitations)



AC voltage breakdown test or dissection of failure areas

Evaluation of local insulation resistance of the insulation between the conductor and the surface of the insulation can be useful in assessing water-related deterioration of rubber insulation. The insulation resistance between the conductor and a probe run over the surface of the insulation is used to detect areas of low resistance through the insulation. Such testing has identified areas of lower insulation resistance that were several orders of magnitude lower than that of the surrounding insulation. This activity showed that several pockets of degradation were distributed along and around the insulation. The insulation consistently failed in these low-resistance pockets when breakdown tests were subsequently performed. Even though the insulation was not uniformly degraded, the identification of several low-resistance pockets indicated that distributed general degradation existed, rather than a unique degradation site. See Section A.3.1, Failure of a 38-Year-Old Butyl Rubber Cable Due to Water-Induced Degradation, for an example of this technique. 10-5

Response to Cable Failures

10.4 Repair Options Repair options differ case by case. In addition to the economic factors (reduced unit output or unit off-line), there might be regulatory issues (environmental qualification; requirements of FERC, NERC; commitments; and so on), environmental issues (extreme cold, hurricanes, flooding, and so on), logistical issues (availability of parts or specialized personnel), considerations of collateral damage to adjacent cable or cables, physical barriers (such as cable stuck in conduits, degraded conduits that prevent pulling new cable, fire stops, or drywell penetrations) that all could play a part in a repair decision. 10.4.1 Replacement of Failed Section Replacement of the failed section, rather than the entire circuit, might be necessary, especially if the failed section is underground rather than in the plant. This option can be used if an appropriate location is available to make a splice. Splicing might be necessary for reasons other than just time. The section of cable that will be retained should be electrically tested to verify its condition. In a few cases, a local repair might be possible; for example, a splice or termination that does not damage the surrounding cable can fail, or a single insulation failure that does not damage the conductor can occur. In such cases, replacing the splice or terminating or reinsulating the failure site might suffice. Many factors can complicate an in situ repair. One complication would be that the environment of the failed section can be a main contributor to initiating the fault. If the cables were in a high temperature or wetted location, for example, a local repair would not be the best option for a permanent correction. It might be acceptable as a temporary1 repair, but rerouting the cable around the adverse environment might be appropriate, as would replacement with a cable that is impervious to the environment. Rerouting the cable in part or whole might be required if the failed cable cannot be physically removed or a section of duct is so badly deteriorated that new cable cannot be pulled through it. Although most plants have been successful in removing cable from ducts for replacement, one plant had a cable jam during removal and had to abandon the duct. No other ducts were available, and the cable system had to be reengineered to allow 4 cables per phase rather than 5. When removing a cable from a duct, having a pull rope on both ends is recommended. If a cable begins to jam during removal, the trailing rope can be used to pull the cable back and possibly extricate the cable from the opposite end. Continuing to pull when a cable jams is likely to tighten the cable in place and prevent removal. Before removing a cable from a duct, the path should be evaluated. If at all possible, any bends in the path should be closest to the pulling end to reduce the force required during removal. Removal forces, especially to start the pull, are likely to be much higher than allowable installation forces. Accumulations of dirt and dried pulling compounds tend to glue the cable in place.

1

Temporary is a relative term. Depending on the environmental stressor, it could take many years, even decades, for the cable to degrade again to the point of failure. However, if the causes of the original failure cannot be eliminated from consideration for the replacement cable, the repair should be considered temporary, and another option must be implemented later to permanently resolve the problem.

10-6

Response to Cable Failures

Fire barriers and containment penetrations can also prevent using the original routing path of replacement cables if disturbing the barriers would create issues for adjacent circuits or affect their ability to continue to perform their design function. Access to the cables can be another issue. Safety concerns for working in a confined location such as a vault or manhole with other energized conductors can become a factor in the repair decision. Corporate safety procedures for working in confined spaces and in areas with the potential for arc flash must be followed. 10.4.2 Total Replacement Depending on the circumstances, the desire might be to replace a complete circuit, especially for safety-related circuits, for which having no splices is desirable. Splices can add a potential failure mode if they are not properly installed, but there is no technical basis for not considering a spliced section of cable acceptable and permanent. Having sufficient cable available to replace the entire length of the longest circuit in the plant is recommended. With sufficient cable on hand, total circuit replacement is still an option. When a failure occurs in service, the time needed to replace an entire circuit can cause an excessively long equipment outage. Replacing a large power cable inside the confines of a plant can take a long time because of limited space for jockeying the cable into trays and the need break and remake fire barriers. For some circuits, especially circuits with several cables per phase, disassembly and reassembly of the tray system might be necessary to remove the old cable and reinstall the new ones. Even though replacement of an entire circuit might be desirable, it might be necessary to replace a segment during a forced outage and then replace the entire circuit during a scheduled outage. 10.4.3 Acceptance Testing The scope and complexity of the postinstallation testing is related to the scope of work performed. Influencing factors are the time required to perform the recommended tests, whether the repair is considered temporary or permanent, the impact of testing on older cable in the circuit that was not replaced, and the value of diagnostic test data if a partial circuit replacement is performed using a different cable type (that is, XLPE cable spliced onto EPR cable). Acceptance and postmaintenance tests should provide reasonable assurance that the repairs have been performed adequately to ensure that the equipment is safe to return to service and capable of performing its design function. In addition, if diagnostic tests are performed as part of this testing, a baseline will be available for comparison to later test results. The following tests should be considered when repairs or replacements have been performed: 

Shield continuity check for all conductors (shielded cable only).



Insulation resistance test to confirm acceptability for performance of a withstand test.



Withstand test of each conductor. The test value should be based on the manufacturer’s recommendations for newly installed cable or in accordance with the guidance provided in IEEE Std. 400.



Diagnostic test of each conductor as a baseline. 10-7

Response to Cable Failures

If replacement of the cable is on the critical path in an outage, diagnostic testing could be deferred to a more convenient time. Aging of a medium-voltage cable takes a long period. Accordingly, baseline testing of the cable a few years into operation would be acceptable. When old and new segments are to be combined, both segments should be tested separately before splicing to verify the adequacy of the segments and then again after splicing to verify the adequacy of the splices and terminations of the completed cable. At minimum, the completed cable should withstand testing to verify that no serious damage has occurred during installation and that no splicing or termination errors have occurred.

10-8

11

REFERENCES 1.

Medium Voltage Cable Aging Management Guide. EPRI, Palo Alto, CA: 2008. 1016689. This product has been superseded by Product 1021070, Medium Voltage Cable Aging Management Guide, Revision 1.

2.

R. B. Blodgett and R. G. Fisher, “A New Corona- and Heat-Resistant Cable Insulation Based on Ethylene Propylene Rubber,” Paper No. 63-162, AIEE Transactions on Power Systems and Apparatus, pp. 980–990 (December 1963).

3.

J. C. Carroll, A. R. Lee and R. B. McKinney, “Design and Evaluation of Butyl-RubberInsulated Power Cables,” AIEE Transactions on Power Systems and Apparatus, Vol. 74, pp. 1204–1214 (December 1955).

4.

Generic Letter 2007-01, “Inaccessible or Underground Power Cable Failures that Disable Accident Mitigation Systems or Cause Plant Transients,” U.S. Nuclear Regulatory Commission, Washington, DC (February 7, 2007).

5.

Aging Management Program Guidance for Medium-Voltage Cable Systems for Nuclear Power Plants. EPRI, Palo Alto, CA: 2010. 1020805.

6.

W. A. Thue, editor. Electrical Power Cable Engineering: Second Edition. New York: Marcel Dekker, Inc. (2003).

7.

AEIC CS1-90, “Specification for Impregnated Paper-Insulated Metallic Sheathed Cable, Solid Type.” 11th ed. Association of Edison Illuminating Companies (1990).

8.

Power Plant Electrical Reference Series, Volume 4: Wire and Cable. EPRI, Palo Alto, CA: 1987. EL-5036-4.

9.

Nexans. “Configuration of Single Conductor Cable to Obtain Equal Load Sharing.” www.nexans.ca/egy/tecdoc/8.htm.

10.

Nexans. “Installation of Single Conductor Cable to Prevent Eddy and Circulating Currents.” www.nexans.ca/egy/tecdoc/7.htm.

11.

NEI 06-05, “Medium Voltage Underground Cable.” Nuclear Energy Institute, Washington, DC (April 2006).

12.

Plant Support Engineering: Common Medium Voltage Cable Specification for Nuclear Power Plants. EPRI, Palo Alto, CA: 2009. 1019159.

13.

T. P. Arnold and C. D. Mercier, editors. Power Cable Manual. Southwire Company (1991).

14.

J. D. Medek, adapted from class notes of “Understanding Power Cable Characteristics and Applications.” University of Wisconsin–Madison (2001).

11-1

References

15. G. Virsberg and P. H. Ware, “A New Termination for Underground Distribution,” IEEE Transactions on Power Apparatus and Systems, Vol. PAS-86, pp. 1129–1135 (1967). 16.

T. A. Balaska, “Jointing of High Voltage, Extruded Dielectric Cables, Basics of Electrical Design and Installation.” IEEE UT&D Conference Record, pp. 318–326 (1974).

17.

Evaluation of Pre-Molded and Field Assembled Joints for Extruded Dielectric Medium Voltage Cables. EPRI, Palo Alto, CA: 2000. 1000045.

18.

Electrical Connections Application Guide. EPRI, Palo Alto, CA: 2002. 1003471.

19.

L. Mullins, “Effects of Filler in Rubber.” In The Chemistry and Physics of Rubber-like Substances: Studies of the Natural Rubber Producers Research Association, edited by L. Bateman, p. 311. Scientek Books (1963).

20.

A. Schallamach, “Abrasion and Tire Wear.” In The Chemistry and Physics of Rubber-Like Substances: Studies of the Natural Rubber Producers Research Association, edited by L. Bateman, p. 374. Scientek Books (1963).

21.

J. D. Hogan and L. A. Berry, “From Dust to Quality Dielectrics—The Formulation of Dielectric Clays,” IEEE Electrical Insulation Magazine, Vol. 21, No. 5. pp. 25-28 (September/October 2005).

22.

T. L. Brown, H. E. LeMay, and B. E. Burstein, Chemistry: The Central Science. 6th ed., p. 876. Prentice Hall Publishing Co. (1994).

23.

B. S. Bernstein, “Electrical Insulating Materials.” In Electrical Power Cable Engineering, edited by W. A. Thue, pp. 65–114. Marcel Dekker, Inc. (2003).

24.

R. Webb, T. Shafer, and A. Tsou, “Butyl Rubber.” In Encyclopedia of Polymer Science and Technology, edited by H. F. Mark, pp. 356–381.Wiley Interscience (2004).

25.

Exxon. “Butyl Rubber.” In Materials for the Electrical Industry, publication SYN-73-1730, pp. 5–16 (1973).

26.

Exxon, “EPR.” In Materials for the Electrical Industry, publication SYN-73-1730, pp. 19–51 (1973).

27.

ICEA S-97-682-2007, “Standard For Utility Shielded Power Cables Rated 5 Through 46 KV” ICEA, Carrolton, GA (2007).

28.

K. Brown, “Performance of Medium Voltage Underground Medium Voltage Cable at United States Nuclear Plants,” IEEE Insulated Conductors Committee Meeting, Scottsdale, Arizona, pp. 886–909 (October/November 2005).

29.

Hercules Chemical Company. Hercules Peroxides, Cross-Linking Agents and Polymerization Catalysts. Bulletin ORC-1031, 2016 (1983).

30. R. L. Zapp and P. Hous, “Butyl and Chlorobutyl Rubber.” In Rubber Technology. 2nd ed. edited by M. Morton, pp. 249–262. Van Nostrand Reinhold Publishing Co. (1973). 31.

11-2

R. Frank, “Black EPR Insulated Medium Voltage Cables,” ICC Minutes, pp. 954-958 (2005).

References

32.

C. Zuidema, “Evolution of EPR Compounds.” ICC Meeting, Education Forum on EPR, pp. 920–953. (Fall 2005).

33.

H. Hu, “Evaluation of Discharge Resistance of Solid Dielectric Power Cable Insulations,” IEEE Electrical Insulation Magazine, Vol. 11, No. 2, pp. 18–22 (July/August 1995).

34.

A. K. Hiranandani, “Condition Assessment of Butyl Insulated Power Cable Used in Generating Stations,” Conference Record of 2004 IEEE International Symposium on Electrical Insulation, pp. 450-453 (2004).

35.

W. C. Smith, “Vulcanization of Butyl, Chlorobutyl and Bromobutyl Rubber.” In Vulcanization of Elastomers, edited by G. J. Allimger and I. J. Srjothun, pp. 230–264. Reinhold Publishing Co. (1964).

36. F. S. Conant, “The Effect of the State of Cure on Vulcanizated Properties.” . 78-125. Vulcanization of Elastomers. In Vulcanization of Elastomers, edited by G. J. Allimger and I. J. Srjothun, pp. 78–125. Reinhold Publishing Co. (1964). 37.

L. A. Dissado and J. C. Fothergill. “Water Treeing Degradation.” In Electrical Degradation and Breakdown in Polymers, IEE Materials and Devices Series 9, G. C. Stevens, ed. United Kingdom: Peter Peregrinius Ltd. (1992).

38.

M. Fenger, ”Failure Analysis—Feedwater Pump Motor Cable: Accurate Assessment for Informed Decision Making,” Kinectrics, Inc., Transmission & Distribution Technologies, publication T&D-CS-12.02.07. http://www.kinectrics.com/en/solutions/CableFailureFeedwaterPump.html.

39. A. Charlesby. Atomic Radiation and Polymers. International Series of Monographs on Radiation of Materials. Pergamon Press (1960). 40.

A. Chapiro. Radiation Chemistry of Polymeric Systems, Sections IX and X. Interscience Publishing, J. Wiley and Sons (1962).

41.

J. Schiers, “Opportunities in Gamma Irradiation of Polymers.” www.steritech.com.au/files/pdf/xlinkapr06.pdf.

42.

G. Odian and B. S. Bernstein, “Radiation Cross-Linking of Polyisobutylene,” Journal of Polymer Science. Polymer Letters, Vol. 2. p. 819 (1964).

43.

Effects of Moisture on the Life of Power Plant Cables, Part 1: Medium Voltage Cable. EPRI, Palo Alto, CA: 1994. TR-103834-P1.

44.

Effects of Moisture on the Life of Power Plant Cables, Part 2: Low Voltage Cable. EPRI, Palo Alto, CA: 1994. TR-103834-P2.

45.

Medium Voltage Cables in Nuclear Plant Applications- State of Industry and Condition Monitoring. EPRI, Palo Alto, CA: 2003. 1003664.

46. ICEA T-27-581/NEMA WC 53, “Standard Test Methods for Extruded Dielectric Power, Control, Instrumentation, and Portable Cables for Test.” NEMA, Roslyn, VA (2005). 47.

E. Walcott, W. S. Temple, and J. T. Smith III. 2005. “Accelerated Wet Testing of MV EPR Insulated Cables,” presented at the Fall 2005 ICC Meeting, Education Forum on EPR (2005). 11-3

References

48.

Advanced Diagnostics: Life Estimation of Extruded Dielectric Cables. EPRI, Palo Alto, CA: 1993. 1001726.

49.

Advanced Diagnostics: Life Estimation of Extruded Dielectric Cables. EPRI, Palo Alto, CA: 2005. 1011499.

50.

C. Katz, “Performance Evaluation of EPR Underground Distribution Cable,” presented at the IEEE Insulated Conductors Committee Meeting (Fall 2005).

51.

IEEE 400-2001, “IEEE Guide for Field Testing and Evaluation of the Insulation of Shielded Power Cable Systems.” IEEE, New York (2001).

52.

In-Service Evaluation of Extruded Dielectric Distribution Cables. EPRI, Palo Alto, CA: 2002. 1009017.

53.

Cable Polymer Aging and Condition Monitoring Research at Sandia National Laboratories under the Nuclear Energy Plant Optimization (NEPO) Program. EPRI, Palo Alto, CA: 2005. 1011873.

54.

Medium Voltage Cables in Nuclear Plant Applications—State of Industry and Condition Monitoring. pp. 28–30. EPRI, Palo Alto, CA: 2003. 1003664.

55.

IEEE 400.1-2007, “IEEE Guide for Field Testing of Laminated Dielectric, Shielded Power Cable Systems Rated 5 kV and Above with High Direct Current Voltage.” IEEE, New York (2007).

56.

IEEE 400.2-2004, “IEEE Guide for Field Testing of Shielded Power Cable Systems Using Very Low Frequency (VLF).” IEEE, New York (2004).

57.

ICEA S-93-639/NEMA WC 74, “5–46 kV Shielded Power Cable for Use in the Transmission and Distribution of Electric Energy.” National Electrical Manufacturers Association, Rosslyn, VA (2006).

58. IEEE 383-1974, “IEEE Standard for Type Test of Class 1E Electric Cables, Field Splices, and Connections for Nuclear Power Generating Stations.” IEEE, New York (1974). 59.

IEEE 308-1978, “IEEE Standard Criteria for Class 1E Power Systems for Nuclear Power Generating Stations.” IEEE, New York (1978).

60. IEEE 603-1998, “IEEE Standard Criteria for Safety Systems for Nuclear Power Generating Stations.” IEEE, New York (1998). 61.

IEEE 1185-1994, “IEEE Guide for Installation Methods for Generating Stations.” IEEE, New York (1994).

62. ICEA S-97-682-2000, “Utility Shielded Power Cables Rated 5,000–46,000 Volts.” ICEA, Carrolton, GA (2000). 63.

ASTM D-2275-89. “Test Method for Voltage Endurance of Solid Electrical Insulating Materials Subjected to Partial Discharges (Corona) on the Surface.” American Society for Testing Materials, Conshohocken, PA.

64.

ICEA T-31-610, “Test Method for Conducting Longitudinal Water Penetration Resistance Tests On Blocked Conductors.” ICEA, Carrolton, GA (2006).

11-4

References

65.

IEEE 383-2003, “IEEE Standard for Type Test of Class 1E Electric Cables and Field Splices for Nuclear Power Generating Stations.” IEEE, New York (2003).

66.

IEEE 1202, “Standard for Flame-Propagation Testing of Wire and Cable.” IEEE, New York (2006).

67.

ICEA T-24-380. “Guide for Partial Discharge Test Procedure.” ICEA, Carrolton, GA (2007).

68.

AEIC 5, “Specifications for Cross-Linked Polyethylene Insulated, Shielded Power Cables Rated 5 through 46 kV.” 10th ed. Association of Edison Illuminating Companies (1994).

69.

AEIC 6, “Specifications for Ethylene Propylene Rubber Insulated Shielded Power Cables Rated 5 through 46 kV.” 1st ed. Association of Edison Illuminating Companies (1973).

70.

ICEA S-19-81, “Rubber Insulated Wire and Cable for the Transmission and Distribution of Electrical Energy.” ICEA, Carrolton, GA (1994). Withdrawn.

71.

ICEA S-68-516/NEMA WC 8-1976, “Ethylene-Propylene-Rubber Insulated Wire and Cable for the Transmission and Distribution of Electrical Energy.” Rev. 6. ICEA, Carrolton, GA (pages dated January 30, 1983). Withdrawn.

72.

ICEA S-66-524/NEMA WC 7, “Cross-Linked Thermosetting Polyethylene-Insulated Wire and Cable for the Transmission and Distribution of Electrical Energy.” Second Edition. ICEA, Carrolton, GA (1982). Withdrawn.

73.

AEIC 5, “Specifications for Polyethylene and Cross-Linked Polyethylene Insulated Shielded Power Cables Rated 2,001–35,000 Volts.” 2nd ed. Association of Edison Illuminating Companies (1969).

74.

AEIC CS5, “Specifications for Thermoplastic and Cross-Linked Polyethylene Insulated Shielded Power Cables Rated 5 through 69 kV.” 6th ed. Association of Edison Illuminating Companies (1979).

75.

ICEA T-33-655, “Guide for Low Smoke, Halogen-Free (LSHF) Polymeric Cable Jackets.” ICEA, Carrolton, GA (1994).

76.

N. Ahmed and N. Srinivas, “In situ Partial Discharge Detection in Power Plant Cables.” Proceedings for the Twenty-Sixth Water Reactor Safety Information Meeting. Vol. 3. U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research, pp. 289-301 (1998).

77.

G. C. Montinari, “On Line Partial Discharge Diagnosis in Power Cables,” 2009 IEEE Electrical Insulation Conference, Montreal, Canada, pp. 210–215 (May 31–June 3, 2009).

78.

Cable Polymer Aging Database, Version 2. EPRI, Palo Alto, CA: 2005. 1011874.

79.

Initial Acceptance Criteria Concepts and Data for Assessing Longevity of Low-Voltage Cable Insulations and Jackets. EPRI, Palo Alto, CA, and U.S. Department of Energy, Washington, DC: 2005. 1008211.

80.

AEIC CS8-07, “Specification for Extruded Dielectric Shielded Power Cables Rated 5 through 46 kV.” 3rd ed. Association of Edison Illuminating Companies (2007). 11-5

References

81. Edison Electric Institute (EEI), “Cable Testing.” In Underground Systems Reference Book. pp. 11-2 to 11-6. New York: Edison Electric Institute (1957). 82. N. Srinivas and B. S. Bernstein, “Effect of DC Testing on Aged XLPE Insulated Cables with Splices,” presented at Jicable 91, Paris, France (June 1991). 83.

Bruce S. Bernstein, “Cable Testing: Can We Do Better?” IEEE Electrical Insulation Magazine, pp. 33-38 (July/August 1994).

84.

G. Bahder, C. Katz, G. S. Eager, E. Leber, S. M. Chalmers, W. H. Jones, and W. H. Mangrum, “Life Expectancy of Cross-Linked Polyethylene Insulated Cables Rated 15 to 35 kV.” IEEE Transactions on Power Apparatus and Systems, Vol. PAS-100, pp. 1581– 1590 (April 1981).

85. R. Bach, W. Kalkner, and H. Oldehoff, “Verlustfactormessungen bei 0.1 Hz an betriebsgealterten PE/VPE-Kabelanlagen,” Elektrizitätswirtschaft, Jg. 92, Heft 17/18, pp. 1076–1080 (1993). 86. S. Hvidsten, E. Ildstad, B. Holmgren, and P. Werelius, “Correlation Between AC Breakdown Strength and Low Frequency Dielectric Loss of Water Tree Aged XLPE Cables.” IEEE Transactions on Power Delivery, Vol. 13, No. 1, p. 40–45 (January 1998). 87.

Testing and Diagnostics of Medium Voltage Cables. BAUR Prüf- und Messtechnik GmbH, Sulz, Austria: 1996. PHG-TD Report 820-002.

88.

Advanced Diagnostics: Estimation of Life of Extruded Cables. EPRI, Palo Alto, CA: 2005. 1001727.

89.

P. Werelius, P. Tharning, R. Eriksson, B. Holmgren, and U. Gafvert, “Dielectric Spectroscopy for Diagnosis of Water Tree Deterioration in XLPE Cables.” IEEE Transactions on Dielecrics and Electrical Insulation, Vol.8, No.1. pp. 27–42 (February 2001).

90.

Guide for Non-Destructive Diagnosis of Extruded Cable Systems. EPRI, Palo Alto, CA: 2005. 1001731.

91.

IEEE 400.3, “IEEE Guide for Partial Discharge Testing of Shielded Power Cable Systems in a Field Environment.” IEEE, New York (1994).

92.

High Voltage, Low Frequency (0.1 Hz) Testing of Power Cable. EPRI, Palo Alto, CA: 1998. TR-110813.

93. N. Ahmed and N. Srinivas, “On-Line Partial Discharge Detection in Cables.” IEEE Transactions on Dielectric and Electric Insulation. Vol. 5, No. 2 (1998). 94.

N. Ahmed and N. Srinivas, “On-Line Partial Discharge Diagnostic System in Power Cable Systems,” presented at the IEEE T&D Conference, Atlanta, GA (2001).

95.

N. Srinivas, B. S. Bernstein, O. Morel, and D. Bogden, “Condition Assessment of Distribution and Transmission Class Voltage Cable Systems,” presented at the IEEE T&D Conference New York (2003).

96.

D. Gross, “Signal Transmission and Calibration of On-Line Partial Discharge Measurements,” presented at the IPCDAM Conference (June 2003).

11-6

References

97.

10CFR50.65, Requirements for Monitoring the Effectiveness of Maintenance at Nuclear Plants. Office of the Federal Register, National Archives and Records Administration, U.S. Government Printing Office, Washington, DC.

98.

IEEE 323-1974/2003, “IEEE Standard for Qualifying Class 1E Equipment for Nuclear Power Generating Stations,” IEEE, New York (2003).

99.

INPO AP-913. “Equipment Reliability Process Description.” Institute of Nuclear Power Operations, Atlanta, GA (November 2001).

100. 10CFR50, Appendix R, Fire Protection Program for Nuclear Power Facilities Operating Prior to January 1, 1979. Office of the Federal Register, National Archives and Records Administration, U.S. Government Printing Office, Washington, DC. 101. Failure Mechanism Assessment of Medium Voltage Ethylene Propylene Rubber Cables. EPRI, Palo Alto, CA: 2007. 1015070. 102. Letter to James H. Riley, Director, Engineering, Nuclear Generation Division, Nuclear Energy Institute, from Michael J. Case, Director, Division of Policy and Rulemaking, Office of Nuclear Reactor Regulation, U.S. NRC; Subject: Response to NEI Letter Dated March 26, 2007, RE: Interpretation of Generic Letter 2007-01, Inaccessible or Underground Power Cable Failures that Disable Accident Mitigation Systems or Cause Plant Transients; dated April 13, 2007. 103. Condition Assessment of Five Cable Sections from [a nuclear plant], privately held report, June 2008. (Data used with permission.) 104. Advanced Diagnostics and Life Estimation of Extruded Dielectric Cable. EPRI, Palo Alto, CA: 2006. 1013085. 105. G. Hoff and H. G. Kranz, “On-site Dielectric Diagnostics of Power Cables Using the Isothermal Relaxation Current Measurements,” presented at the IEEE/PES Winter Meeting, Singapore (2000). 106. E. Gulski, J. J. Smit, P. N. Seitz, “PD Measurements On-Site Using Oscillating Wave Test System,” presented at the IEEE International Symposium of Electrical Insulation, Washington, DC (June 1998). 107. H. Orton. “Diagnosing the Health of Your Underground Cable.” Transmission and Distribution World (June 1, 2002). 108. IEEE 43-2000, “IEEE Recommended Practice for Testing Insulation Resistance of Rotating Machinery,” IEEE, New York (2000).

11-7

A

MEDIUM-VOLTAGE CABLE FAILURES AND FIELD EXPERIENCE

A.1

Introduction

This appendix describes failure and field experience for medium-voltage cables. The purpose is to describe events and the lessons learned from them. These failure and field experience descriptions were provided by utilities; however, the plant and owner’s name are not included. This appendix is divided into five sections: this introduction; failures under dry conditions; failures related to wet conditions; failures of terminations and splices; and a summary of lessons learned. A number of failures under dry conditions are presented here. Dry failures occur less frequently than wet failures (approximately 25% as often). The dry failures occur randomly from a number of different causes and, unlike water-related failures, have no common stress enhancement. Essentially, each dry failure event provides an interesting insight into potential failure mechanisms and issues to avoid.

A.2

Dry Condition Failures

A.2.1 Failure of a Cable at the Point of a Shield Crimp A 19-year-old cable failed in service. The fault was traced to a single phase in a triplexed section located in a tray. The cable was manufactured by Okonite and had an extruded conductor shield, pink ethylene-propylene rubber (EPR) insulation, an extruded insulation shield, a helically wound copper tape shield, and an overall jacket. Figure A-1 shows the external condition at the location of the fault, and Figure A-2 shows the condition of the shield after the jacket was removed.

A-1

Medium-Voltage Cable Failures and Field Experience

Figure A-1 External Cable Condition at the Location of the Fault

Figure A-2 Condition of the Shield After the Jacket and Burn Hole Were Removed from the Fault

A-2

Medium-Voltage Cable Failures and Field Experience

The copper tape shield was found to have three longitudinal depressions around approximately one-third of the circumference of the cable. A small burn hole in the tape shield occurred at the location of the insulation failure. Figure A-3 shows the damage that occurred to the conductor of the cable.

Figure A-3 Damage to the Conductor from the Fault

The exact cause of the failure was not identified. The cable was repaired by removing the damaged insulation, tapering the ends of the insulation, and forming a taped repair equivalent of a splice directly over the conductor. A.2.2 Thermal Deterioration of a Butyl Rubber Insulated Cable The cable that failed in service was 5-kV, 1500-kcmil (760.5 mm2) cable manufactured by General Electric and having a Simplex, Anhydrex-Plastex insulation/jacket system. The cable had been in service approximately 35 years. Segments of the cable were sent to a forensics laboratory for assessment. The cable had a copper conductor, a conductor shield composed of carbon black impregnated cloth tape, butyl rubber insulation, a semiconducting polymer tape insulation shield, and a helically wrapped metal foil shield covered by a polyvinyl chloride (PVC) jacket. When the specimens were disassembled, it was noted that two specimens had experienced more heat damage than the third. The labeling on the exterior surface of the semiconductor polymer tape had turned yellow from its original white, and the main insulation was so soft that the edges of the conductor shield tape were cutting into the insulation (see Figure A-4). The inspection of the specimens determined that there was no evidence of oil or solvents on either the insulation or jacket and that the cause of the softening was not induced by external contaminants.

A-3

Medium-Voltage Cable Failures and Field Experience

Figure A-4 Inner Surface of the Insulation with Severe Indentation from the Semiconducting Tape

To characterize the cause of the softening of the insulation, a number of chemical and physical tests were performed on the insulation and jacket. For the insulation, the tests were the following: 

Fourier transform infrared analysis



Elongation at break



Tensile strength



Gel content



Shore hardness

For the jacket, the tests were the following: 

Solubility



Elongation at break



Tensile strength



Plasticizer content

The Fourier transform infrared analysis confirmed that the insulation material was butyl rubber. The solubility test confirmed that the jacket was PVC. The tensile strength tests of the more thermally damaged insulation specimens ranged from 2.47 to 63.8 psi (0.17 to 0.44 MPa), and the elongation values ranged from 53% to 159%. The less damaged specimen had a tensile strength of approximately 145 psi (1 MPa) and an elongation at break of greater than 250%.

A-4

Medium-Voltage Cable Failures and Field Experience

However, all these values were significantly lower than the Insulated Cable Engineers Association (ICEA) standard minimum tensile strength for new butyl rubber of 597.6 psi (4.12 MPa) and minimum elongation at break of 350%. The gel content of the less damaged specimen was 68%, whereas that of the two more deteriorated specimens ranged from 0% to 13.5%. The expected gel content for unaged butyl rubber insulation should be between 80% to 95%. The Shore “A” hardness for the less damaged cable was 54.4, whereas the other two specimens were too soft to measure. The evaluation estimated an operating temperature of approximately 176°F to 185°F (80°C to 85°C) over its 35-year life. Evaluation of the jacket indicated that the high temperature was not from exterior conditions, because the PVC had not released oily plasticizers or hydrochloric acid, which would have corroded the shield. At 200%, the jacket had twice the minimum required elongation at break and plasticizer levels (24% to 25%) that were within the normal bound (20% to 25%). The conclusion was that the softening of the butyl rubber was due to elevated temperature from ohmic heating of the conductor. The air temperature in the vicinity of the failure was not elevated, and there were no adverse localized thermal environments in the area. Thermal exposure of butyl rubber induces an oxidative chain scission reaction that leads to softening of the rubber. The cause of the ohmic heating was an unbalanced magnetic circuit that induced higher currents on individual conductors in the circuit with several conductors per phase. This condition occurred when conductors were not transposed along the length of the circuit at the time of installation. This led to some individual conductors having low currents with respect to their ampacity, whereas others had high currents. The softening of the material allowed the conductor shield to rise up into the insulation, leading to high localized stress and eventual breakdown of the insulation. The insulation softening that occurred in this case is unique to sulfur-cured butyl rubbers; it would not occur for EPR or cross-linked polyethylene (XLPE) insulation. A.2.3 Failure of a Brown Ethylene-Propylene Rubber Insulated Cable Due to Failure of a Zinc Shield Tape The failure occurred in a 4160-V, 2000-kcmil (1014-mm2) copper cable with brown EPR insulation and a helically wrapped 8-mil (0.2-mm) zinc insulation shield cable. A wrapped cotton tape shield was located between the zinc shield layer and the insulation. The circuit had two cables per phase. The cable was an outdoor cable that was mounted in a tray on the wall of a building. The fault could not be located electrically. The location was finally determined by careful visual observation along the length of the cable. A darkened spot was noted on the building face, which turned out to be the burn mark from the fault.

A-5

Medium-Voltage Cable Failures and Field Experience

The fault occurred at the base of a Kellems grip used to support the cable at a vertical drop in elevation. Dissection of the cable at the faulted area indicated that the zinc shield tape had crystallized and cracked, causing a separation in the shield and resulting in partial discharge (PD) and arcing along the interface between the gap in the shield. Figure A-5 shows a section of good shield, as well as the deteriorated shield. PD and tracking occurred at the gap in the shield around the circumference of the cable, damaging the jacket and the insulation.

Figure A-5 Zinc Shield (Top, Good Condition; Bottom, Crystallized and Burned from Tracking)

Figure A-6 shows the tracking path and the fault site along the interior of the chlorosulfonated polyethylene (CSPE) jacket. The figure shows that the tracking path followed the path of the gap in the shield, as evidenced by the shield marks on the inner surface of the jacket.

Figure A-6 Tracking Path and Failure Location on Interior of Jacket

A-6

Medium-Voltage Cable Failures and Field Experience

Figure A-7 shows the outer surface of the insulation, illustrating the tracking path and the ultimate failure location on the insulation.

Figure A-7 Tracking Path and Failure Site on Insulation

Figure A-8 shows the damage to the conductor of the cable from the fault. It is not clear why zinc was used as the shield in this cable. It is known that a copper production strike occurred in the early 1970s, which might have affected copper cost and availability at the time that this cable was produced. Most plants reporting tape metal shields indicate that copper or tinned copper was used.

Figure A-8 Damage to Conductor from Fault

A-7

Medium-Voltage Cable Failures and Field Experience

A.3

Water-Related Failures

A.3.1 Failure of a 38-Year-Old Butyl Rubber Cable Due to Water-Induced Degradation The final failure of this cable circuit was preceded by periodic, loud popping noises that were heard from a manhole in the underground section of the cable. The loud popping noises were first heard about 2:00 A.M. on the day of the failure and were again heard in the last 30 seconds of operation when the circuit tripped at 2:14 P.M. The cable circuits associated with the event were complex. There were two individual cables connected directly to two separate secondaries on a 13-to-4-kV transformer. Each of the cables connected to buses through circuit breakers, approximately 450 ft (137 m) away. About 120 ft (37 m) of the circuit, starting near the transformer end, was underground, and the remainder of the cable was inside the plant. Each of the 4-kV cable circuits had six conductors per phase, resulting in 36 conductors in the two cables connected to the transformer. The cables were manufactured by Okonite and had Okonex butyl rubber (200 mil [5.1 mm]) insulation. The conductor and insulation shields were carbon black–filled woven fabric. A helically wound copper tape (two layers, with the first separated and the second lapped) covered the conductor shield, and that was covered by a neoprene jacket. At the time of the event, outside temperatures were as low as -13°F (-25°C) at night. Figure A-9 shows the circuit connections on one of the secondary windings of the transformer, and Figure A-10 shows the connections at one of the bus breakers.

Figure A-9 Transformer Connection of One of the Cables

A-8

Medium-Voltage Cable Failures and Field Experience

Figure A-10 Circuit Breaker Connection of One of the Cables

Figure A-11 shows the configuration of the cables in the manhole adjacent to the transformer. An A, B, and C phase passes through each duct to balance the magnetic currents in the underground section of the circuit. The configuration in the manhole converts from a 3 vertical × 4 horizontal matrix to a 4 vertical × 3 horizontal matrix, so that removal of just a few deteriorated cables would require replacing many cables.

Figure A-11 Cable Configuration in the Manhole Adjacent to the Transformer

A-9

Medium-Voltage Cable Failures and Field Experience

Figure A-11 indicates that water could have affected some circuits in the lower ducts, leading to the expectation that cables in the upper ducts would not be affected because of the indication that water had not reached them. As indicated by Figure A-12, however, cables in the upper ducts failed under test or exhibited test results showing significant deterioration. (Boxes in the figure indicate the locations of deteriorated cables.)

Figure A-12 Location of Failed, Faulted, and At-Risk Cables Within Duct Leading to Building

The locations of the deteriorated cables indicated that the water mark in the manhole was not indicative of the locations of deteriorated cable. Moreover, the removal of cables from the manhole would have required removing and repulling nearly every cable in the manhole, a complex and nearly impossible job, given the outdoor temperatures at the time of the event. The plant owner decided to run temporary cables instead of trying to pull new cables into the ducts under adverse temperature conditions and to allow the unit to return to service more quickly. A review of the complexity of the cable layout inside the plant indicated that replacement of the entire lengths of the circuits would require approximately 3 months time and approximately 18,000 ft (5500 m) of cable. The plant had 7500 ft (2300 m) of spare cable on hand. A decision was made to cut the circuits just inside the building wall and to splice temporary cables in place between the transformer and the building wall. Tan δ measurements were used to determine the A-10

Medium-Voltage Cable Failures and Field Experience

cables that had failed and to confirm that the in-building segments of these cables were in acceptable condition. PD testing was attempted, but the corrosion of the shield and the high attenuation of the butyl rubber insulation precluded detecting a reflection of even a high-level calibration signal. This indicated that, even had PD been present, it would not have been detectible from the breaker cubicle test location. The cause of the failures was traced to water-related degradation. The ducts drained toward the building, causing water to leak into the building in winter. The entry point into the building is in an unheated area, allowing ice to form. To eliminate this condition, the ducts had been sealed in the 1990s (see Figure A-13). Figure A-13 also shows water draining from a conductor of cable B-5, the cable that failed during the event.

Figure A-13 Duct Sealing and Water Leaking from Failed Conductor B-5

Figure A-14 shows the as-found condition of cables A-5, B-5, and C-5 when they were removed from the duct. During the original failure of the B-5 cable, the copper shield vaporized, causing the voltage to seek an alternate ground (2 A.M. repetitive flashovers). Subsequently, the jacket failed on the A-5 cable, allowing a flashover path to the shield of the A-5 cable. The 2 P.M. repetitive flashovers are likely to have been to the A phase shield. Fortunately, the A phase insulation withstood the condition, and a phase-to-phase fault did not occur. The original failure of the B-5 cable insulation caused burnout of the copper shield, and the failure propagated to the shield of the A-5 cable. The C-5 cable was not damaged.

A-11

Medium-Voltage Cable Failures and Field Experience

Figure A-14 Cables A-5, B-5, and C-5

Figure A-15 shows a close-up of the damage to the B-5 cable. The copper tape shield has vaporized from the entire vicinity of the fault and is not visible at the edges of the hole in the jacket. The fault evaporated 3/8 in. (1 cm) of the conductor.

Figure A-15 Close-Up of B-5 Cable Failure

A-12

Medium-Voltage Cable Failures and Field Experience

The faulted cable and adjacent phases were sent to a forensics laboratory, where all of the strands of the conductor were found to be corroded in the vicinity of the fault, indicating that the conductor had filled with water (see Figure A-16). B5B is the conductor of the cable that failed. A5 is the conductor that was adjacent; it is slightly corroded. C5 is the adjacent conductor that had good insulation, and B2 is from another cable with good insulation.

Figure A-16 Comparison of Conductor Corrosion

The insulation was found to have a high moisture content, as well. Examination of the insulation showed channels of low insulation resistance between the surface of the insulation and the conductor. The insulation resistances of these areas were 1/1000th of the resistances of the surrounding insulation. These areas were marked, the shield was reformed, and withstand tests were performed. The breakdowns of the insulation occurred at these channels of low insulation resistance (see Figure A-17). The black tape and aluminum foil constitute the shield that was applied to allow breakdown testing after the low-resistance channel was identified.

Figure A-17 Marked Low-Resistance Channel in the Insulation with Breakdown Hole in Center

A-13

Medium-Voltage Cable Failures and Field Experience

Wafers were cut from some of the low-resistance channels and the wafers were heated in water to allow identification of degradation. Swelling and fissures were identified in the low-resistance zones (see Figure A-18).

Figure A-18 Micrograph of Insulation Wall at Low-Resistance Channel, Showing Swelling and Fissures

The conclusion was that the failure was caused by water-induced degradation of the energized cable’s insulation. The corrosion of the entire conductor indicated a long-term presence. The water could have accumulated during construction or through permeation through the cable wall and condensation at the lowest and coldest point on the cable. The failure occurred within a short distance of the point at which the temperature of the cable would have been below freezing for most of the winter. The load on these cables during normal operation is low, so ohmic heating would not have tended to warm the cable and reduce condensation. Temporary cables were run on the surface of the ground to replace the underground section of the cables. Figure A-19 shows the temporary routing, and Figure A-20 shows the aboveground, permanent replacement that was implemented during the next outage of the second unit at the plant.

A-14

Medium-Voltage Cable Failures and Field Experience

Figure A-19 Temporary, Aboveground Cable System

Figure A-20 Permanent Aerial Structure Routing on Second Unit

A-15

Medium-Voltage Cable Failures and Field Experience

The following lessons were learned from this event: 

The failure could have burned into a phase-to-phase fault. It is recommended that a circuit that has a suspected fault, as indicated by loud popping, be removed from service as soon as possible. Had the fault become a phase-to-phase fault, the source transformer could have been damaged.



Having spare cable and necessary splice and termination kits available is critical to a rapid return to service.



Had it been necessary to replace the entire circuit, including the dry sections, a long period would have been required to remove and replace fire stops and to maneuver the large, heavy cable through tightly configured tray systems. Three months was estimated for the total replacement of these circuits.



Sealing ducts on the interior of building walls can cause water to back up into the duct, covering cable in areas that previously were well drained.



Current splicing crews are familiar with modern cable construction in which shield semiconducting materials are easy to distinguish from insulation. Extra training will be necessary when such crews are working with old constructions in which semiconducting materials and the insulation are black.



Significant plant modifications might be required to allow replacement of below-grade circuits. Pulling cables at low temperatures is difficult and could lead to damage of the cable.



Tan δ testing was practical for this cable; however, due to the large attenuation of the corroded shield and butyl rubber, PD testing was not practical. On-line assessment had been performed on this cable. It is likely that the attenuation would have made identification of this degradation difficult using on-line techniques. The configuration of the circuit allowed on-line assessment at only one end, the one most distant from the degradation. Discussions with the developer of the on-line test technique indicated that detection of the problem on this butyl rubber cable from the far end would be unlikely and that testing from both ends would be desirable.



Surface-to-conductor resistance evaluation in the forensic assessment identified lowresistance channels from water-induced degradation. (The 60-Hz tan δ test data for this cable are provided in Appendix C.5, Tan δ for Butyl Rubber Insulation).

A.3.2 Failure of a Wet Okonite Black Ethylene-Propylene Rubber Cable Two separate segments of 32-year-old Okonite black Okoguard EPR insulated 4/0 AWG (107 mm2) cable with a helical copper tape shield and an Okoprene (neoprene) jacket failed. The insulation was 140 mil (3.5 mm) thick, and the jacket was 80 mil (2 mm) thick. The cables had been continuously energized but unloaded for the duration of their operation and had been submerged in brackish water for much of their life. The circuits supplied a transfer switch that fed the starter for a fire protection pump motor. The first circuit failed approximately one week before the other. Both circuits failed with B and C phase overcurrent indications. The circuits had no phase–to-ground alarms (the only circuits in the plant without such alarms) and might have had single phase-to-ground conditions for a period of time before the phase-to-phase failure occurred. When the cables were removed from the ducts, two phases (likely B and C) were found A-16

Medium-Voltage Cable Failures and Field Experience

to have sections of loose or failed jacket, and the third phase (likely A) was found to have its jacket intact. As the cable was being removed from the duct, it was cut into random lengths from 3 to 5 feet in length. The electricians noted that water was dripping from between the strands of the conductors when the sections were cut. The utility’s evaluation concluded that the failures were due to water-related insulation degradation. Up to the point of the failure, the utility had considered insulation resistance testing acceptable for evaluating the condition of cables. However, as a result of the investigation, the utility determined that insulation resistance testing is not a discriminating indicator of condition of medium-voltage cables and that their acceptance criteria of a minimum of 5 MΩ was wholly inadequate. The cables that failed had insulation resistances of 30 MΩ before the event, whereas the other cables had 2 GΩ or higher. At the time of the first event, the alternate feed was found to have low insulation resistance values. The cables were cut in the manhole, the section with the low insulation resistance was replaced, and the cables with 10–18 MΩ insulation resistances were reused. However, 6 hours after reenergization, the reused section B and C phase failed, indicating that the 5 MΩ criterion was unsatisfactory. Sections of the cable were provided to the Electric Power Research Institute (EPRI) for further testing. The 60-Hz tan δ versus breakdown results for this cable are provided in Appendix C.1, 60-Hz Tan δ Data from 1972 Black Okoguard Insulation. The evaluation identified that the neoprene jacket was loose in a number of cases and determined that low insulation resistance channels had developed in the EPR, similar to those described for the butyl rubber cable (see Section A.3.1, Failure of a 38-Year-Old Butyl Rubber Cable Due to Water-Induced Degradation). The full results of the evaluation are provided in the EPRI report Failure Mechanism Assessment of Medium Voltage Ethylene Propylene Rubber Cables (1015070) [101]. A.3.3 Failure of a Cross-Linked Polyethylene Insulated Cable from Localized Water Treeing A 35-year-old XLPE insulated General Cable Corporation 5-kV cable failed in service and was de-energized by the operation of a C phase protective relay. The cable had a 2/0 AWG (67.5 mm2) copper conductor with a nominal 90 mil (2.3 mm) of XLPE insulation. The cable had extruded semiconducting conductor and insulation shields and a helically wrapped copper tape shield with an overall PVC jacket. The cable failed in an underground duct bank. Initially, the fault could not be located. Therefore, the VLF test set was applied to carbonize the fault channel. The fault was then located using a bridge fault detection device. The faulted section of cable, an adjacent nonfaulted phase, and an additional section away from the fault were shipped to a forensics laboratory. Extensive electrical and physical tests were performed on all three sections of the cable. After the faulted section was removed from the failed cable, it and the other sections were subjected to VLF tan δ measurements. The nonfaulted cables had good results. The nonfailed sections of the faulted cable had tan δ results within the requirements of IEEE Std 400-2001; however, a slight rise in tan δ occurred with increasing voltage. The cables all passed a 7-kV, 60-minute VLF high-potential test, during which the tan δ measurement remained stable. The electrical testing indicated that the failure was due to a localized condition. The cables were deemed to have breakdown strengths on the order of at least 3 to 4.75 times the breakdown strength considered by the industry as end of life (200–300 V/mil [7.8–11.8 kV/mm]). A-17

Medium-Voltage Cable Failures and Field Experience

Dimensional checks indicated that all the cable subcomponents were within specification, with the exception of the PVC jacket, which was slightly thinner than expected. The jacket thickness was determined to have no effect on the failure. The section that had failed was evaluated in detail. The fault did not puncture the jacket, and the jacket had to be removed to allow location of the actual fault. Figure A-21 shows the puncture and damage caused by the vaporization of the shield during the fault.

Figure A-21 Damage to the Copper Shield at the Damage Site

Figure A-22 shows the cross section of the insulation at the location of the fault. The fault tube shows the carbonization of the tube from the fault and the fault burn-in process that was used to locate the fault. The inspection of the wafers of insulation at the fault shows no evidence of a large water tree at the failure site. However, the fault did burn away an extensive area of the insulation. No sign of an electrical tree was identified.

A-18

Medium-Voltage Cable Failures and Field Experience

Figure A-22 Cross Section of Insulation at the Fault Tube, Showing Carbonization of the Wall of the Fault Tube

Sections of the insulation surrounding the fault were subjected to a hot oil bath, which turns the insulation clear for visual examination. Figure A-23 shows evidence of a large, single water tree. Figure A-24 shows a cross section of the 61-mil (1.5-mm) water tree after dying.

A-19

Medium-Voltage Cable Failures and Field Experience

Figure A-23 Large Water Tree Viewed Through Hot Oil Bath

Figure A-24 Cross Section of the Water Tree Shown in Figure A-23

A-20

Medium-Voltage Cable Failures and Field Experience

The water tree extends 70% through the wall of the insulation at the site of a small conductor shield protrusion. The protrusion and the contaminant causing it are shown in Figure A-25.

Figure A-25 Embedded Particle at Base of Water Tree Shown in Figures A-23 and A-24

Similar assessments were performed on the other cable segments that had been taken to breakdown during the laboratory testing. The examination at the failure sites showed low to medium density of water trees, with a maximum tree length of 70% to 80% of wall. The overall conclusion was that the most likely cause of failure was a large water tree growing from a conductor shield defect that was present at the time of installation. The results indicate that the insulation system was not highly treed but that a few large trees had occurred at sites of defects from manufacture in the early 1970s, the period before manufacturers made significant efforts to improve the cleanliness of the preparation and extrusion process.

A-21

Medium-Voltage Cable Failures and Field Experience

A.4

Failures of Terminations and Splices

A.4.1 Failure from Use of an Oversized Termination Sleeve After 18 years of service, an Anaconda UniShield cable circuit faulted, and protective relays operated to interrupt the circuit. The cable had a 500-kcmil (253.5-mm2) compact stranded copper conductor, an extruded conductor shield, 0.175-in. (4.45-mm) pink EPR insulation, and an extruded shield and jacket containing shield wires. The failure was traced to the interface between the cable and an Elastimold termination on the B phase. The cable and termination had been in service for 18 years. Figure A-26 shows the as-found condition, including the slightly burned condition of the B phase stress relief adapter. Figure A-27 shows a hole in the stress relief adapter of the termination, and Figure A-28 shows the hole that burned through the insulation beneath the stress relief adapter.

Figure A-26 As-Found Condition of Terminations

A-22

Medium-Voltage Cable Failures and Field Experience

Figure A-27 Burnthrough of Stress Relief Adaptor

Figure A-28 Burned-Through Insulation Found After Removal of the Stress Relief Adaptor

The failed termination and the non-failed C-phase termination were sent to a cable forensics laboratory for evaluation. The laboratory found that the Elastimold adaptor was meant for a cable with a diameter of 1.220 to 1.375 in. (31 to 35 mm) and the cable insulation diameter was 1.16 in. (29.5 mm). It appeared that a molded termination meant for a standard configuration was used on an Anaconda UniShield cable having a compact design with a much smaller diameter. The central part of the adapter should mate tightly with the cable insulation surface. Radial pressure must be exerted on the cable insulation surface by the adapter to quench any PD at the interface of the cable insulation and the stress relief adapter. Instead of the adaptor exerting radial pressure on the cable insulation, it was loose. Accordingly, after the silicone grease dried out, PD eroded both the adapter and the insulation, leading to the failure.

A-23

Medium-Voltage Cable Failures and Field Experience

Although the C-phase termination showed no outward signs of degradation, interior inspection showed that the silicone grease beneath the stress relief assembly was highly discolored and experiencing PD, as well. The lesson learned from this failure is that the Elastimold and other types of molded terminations (as well as cold-shrink and heat-shrink systems) must be sized properly to mate with the cable insulation and shielding system. Similar terminations throughout the plant were assessed to confirm that the correct size stress relief adapter had been used. If this type of problem is suspected, handheld PD detection devices can be passed near the terminations to determine whether PD is present. Repairs were made by replacing 10-ft (3.0-m) sections of the field cable and applying splices and new terminations.

A.5

Summary of Lessons Learned

This section summarizes the lessons learned from these failures. A.5.1 Dry Cable Compression-Related Event Although it is not clear exactly what caused the failure in this case, the event indicates that care must be taken to prevent abusive handling during the installation of cables. No loads should be rested on cables when they are laid out for installation, and no carts or vehicles should traverse them. Care should be taken to ensure that sheaves and pulleys used during pulling provide a smooth arc for the cable to follow and that cables do not run against edges of ducts or trays that could grab or indent their surface. A.5.2 Overheated Butyl Rubber Failure Some butyl rubber medium-voltage insulations will soften when subjected to long-term elevated temperatures from operation or environment. This problem is related to butyl rubbers having sulfur curing processes. Butyl rubber installation systems that remain in service should be checked to determine whether ohmic and ambient heating are well within the ampacity limits of the cable. This is especially important for cables known to be sulfur cured. However, determination of whether a butyl rubber cable was sulfur cured might be difficult at this late date. A.5.3 Failure of Continuity of a Zinc Tape Shield When helically wrapped metal tape shields are used in cable, zinc tapes should be avoided. These can corrode to the point of separation and result in an arcing and tracking path that can lead to insulation failure. The 2005 Nuclear Energy Institute survey results reported in NEI 06-05 indicate that the use of zinc tape shields is rare in the industry [11].

A-24

Medium-Voltage Cable Failures and Field Experience

A.5.4 Water-Related Failure of a Butyl Rubber Cable Although the failure was related to a butyl rubber cable, most of the following lessons learned apply to all cable types. 

The fault did not clear properly and continued to pop (seek ground) over an extended period. The failure could have burned into a phase-to-phase fault. It is recommended that a circuit that has a suspected fault, as indicated by loud popping, be removed from service as soon as possible. Had the fault become a phase-to-phase fault, the source transformer could have been damaged. This fault eventually caused failure of the adjacent phase’s jacket and arced through that phase’s shield.



This concern also applies to cables that have phase-to-ground alarms that do not cause a circuit trip. When such alarms are received, it is recommended that the circuit be removed from service as soon as possible to preclude damage to the adjacent phases and the potential for the condition to become a phase-to-phase fault with high fault currents.



Having spare cable and necessary splice and termination kits available is critical to a rapid return to service. Spare cable alone might not be enough.



Had it been necessary to replace the entire circuit, including the dry sections, a long period would have been required to remove and replace fire stops and to maneuver the large, heavy cable through tightly configured tray systems.. Replacement of the section containing the fault might be the only viable option for return to service in a reasonable time.



Sealing of ducts on the interior of building walls can cause water to back up into the duct, covering cable in areas that previously were well drained.



Current splicing crews are familiar with modern cable construction in which shield semiconducting materials are easy to distinguish from insulation. Extra training will be necessary when such crews are working with old constructions in which semiconducting materials and the insulation are black.



Significant plant modifications might be required to allow replacement of complex belowgrade circuits. Pulling cables at low temperatures is difficult and could lead to damage of the cable.



Tan δ testing was practical for this cable; however, due to the large attenuation of the corroded shield and butyl rubber, PD testing was not practical. On-line assessment had been performed on this cable. It is likely that the attenuation would have made identification of this degradation difficult using on-line techniques. The configuration of the circuit allowed on-line assessment at only one end, the one most distant from the degradation.

A-25

Medium-Voltage Cable Failures and Field Experience

A.5.5 Failure of an Ethylene-Propylene Rubber Insulated Cable from Long-Term Wetting This event provided insights that insulation resistance measurement acceptance criteria that have been considered adequate in the past are unacceptable and should not be used. In this case, two failures occurred in a short period. At the time of the first event, the phases of the cable were found to have low insulation resistance values. The cables were cut in the manhole, the section with the low insulation resistance was replaced, and the cables with 10–18 MΩ insulation resistances, which were thought to be acceptable insulation resistances, were reused. However, 6 hours after reenergization, the reused section B and C phase failed, indicating that the 5 MΩ criterion was unsatisfactory. Tan δ testing of cables with several hundred megohm insulation resistances in the case of the butyl rubber cable event (see Sections A.3.1 and A.5.4) indicated that even these cables were marginal. Acceptable tan δ results occurred only for cables have gigohm insulation resistances, indicating that insulation resistance testing is not a useful indication of insulation condition for medium-voltage cable. Insulation resistance remains a useful tool for troubleshooting, but it should be used with caution for determining serviceability for medium-voltage cables. A.5.6 Failure of Cross-Linked Polyethylene Insulation from Water Treeing This failure indicates that water treeing of XLPE cable is not rampant, even in cables manufactured in the early 1970s, and that the failures will likely occur at locations of minor and relatively infrequent manufacturing flaws. Even with these flaws, the water tree growth can take a significant time to lead to failure. In the case of this cable, it was 35 years. A.5.7 Failure from Use of an Oversized Molded Termination The lesson learned from this failure is that the Elastimold and other types of molded terminations (as well as cold-shrink and heat-shrink systems) must be sized properly to mate with the cable insulation and shielding system. Similar terminations throughout the plant were assessed to confirm that the correct size stress relief adapter had been used. If this type of problem is suspected, handheld PD detection devices can be passed near the terminations to determine whether PD is present.

A-26

B

RESULTS OF NUCLEAR ENERGY INSTITUTE SURVEY AND NUCLEAR REGULATORY COMMISSION GENERIC LETTER 2007-01 This section provides the results from a Nuclear Energy Institute (NEI) survey on underground medium-voltage cables that was performed in 2005 and described in NEI 06-05, “Underground Medium Voltage Cable” [11]. It also summarizes the utility responses to the Nuclear Regulatory Commission (NRC) Generic Letter 2007-01, “Inaccessible or Underground Power Cable Failures That Disable Accident Mitigation Systems or Cause Plant Transients” [4]. Since the time of the survey, a few more medium-voltage cables have failed. These are not included in the data. The results of the Generic Letter response and the NEI survey correlate well, with only one exception, which is described in Section B.3, Results of Utility Responses to Generic Letter 2007-01.

B.1

Nuclear Energy Institute Underground Medium-Voltage Cable Survey

B.1.1 Survey Purpose The purpose of the survey was to identify the medium-voltage cable types in use in underground service, their insulations and shielding systems, and their generational differences. In addition, the survey identified failures that occurred under wet conditions and linked them to the cable types. B.1.2 Survey Scope In response to NRC staff concerns related to wet medium-voltage cable , the NEI formed a Medium-Voltage Underground Cable Task Force to assemble a white paper. The task force was formed of knowledgeable cable personnel from the industry. Shortly after the NEI task force was formed, it was agreed that data from the individual plants were needed before any aging-related failure conclusions could be drawn. Thus, a comprehensive survey was designed in January 2005 to capture the following types of information: 

Extent of medium-voltage cable that is installed underground for safety-related and critical functions



Number of circuits



Rated and applied voltage levels



Cable manufacturer, insulation type, and color



Cable age (based on year installed)

B-1

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01



Cable functions



Cable conductor shield and insulation shield attributes



If failures had occurred, information about the failure root cause and cable replacement

Most survey data were collected in February and March 2005. However, some plants conducting refueling outages submitted data later.

B.2

Survey Results Evaluation

B.2.1 Contributors A total of 81 units (51 plants) provided information to the general survey questions. Of those units, 74 units (47 plants) provided information on originally installed cables, failures, and replacement cables. B.2.2 Underground Circuit Quantities All 81 responding units reported some underground cable applications. Of those, 65 units (80%) reported underground conduits; 76 units (94%) reported underground ducts; 23 units (28%) reported some direct-buried circuits; and 21 units (28%) reported enclosed trenches with cables supported within the trench. The number of circuits per plant was quite variable and appears to relate to vintage. For the 75 units providing data, the total number of circuits identified was 8509. This quantity is for wet and dry applications, underground and in plant. The average number of circuits per unit was 113, with a high of 376 and a low of 4. A total of 77 units provided data on the number of underground circuits; the total number of circuits was 2767. The average number of underground circuits per unit was 36, with a high of 214 and a low of 2 per unit. The ratio of underground circuits to total medium-voltage circuits was 0.32. Of the plants reporting underground circuits, 31% indicated that the circuits were dry, and 69% indicated that the circuits were subject to wetting. B.2.3 Installed Cable Types Most respondents indicated use of 5-kV rated cables in underground applications operating at 4.16 kV. Some units reported more than one type of cable in use. Butyl rubber insulation was in use in the late 1960s and early 1970s; it was used at few sites. The material was replaced by black ethylene-propylene rubber (EPR), cross-linked polyethylene (XLPE), and brown EPR. In the late 1970s to early 1980s, EPR manufacturers achieved improvements in longevity by improving the coating of clay fillers to make them nonhydroscopic. The carbon black was removed, causing the pink pigment to be visible and to indicate the generational difference. EPR is a compound, and each manufacturer’s formulation is somewhat different from others. Some variations in behavior might exist due to formulation and overall cable design differences. Even after the manufacturers changed the insulation color to pink, improvements in compounding and quality of extrusion continued.

B-2

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

Table B-1 lists the 5-kV rated insulation types by number of units reporting the material. Table B-1 Originally Installed 5-kV Insulation Types Insulation

Units

Percent of Reporting Units

Butyl Rubber

4

5

Ethylene-Propylene-Diene Monomer

1

1

Black Ethylene-Propylene Rubber

48

65

Brown Ethylene-Propylene Rubber

20

27

Pink Ethylene-Propylene Rubber

31

42

Cross-Linked Polyethylene

23

31

Figure B-1 shows the distribution of manufacturers of black EPR insulated cables; Okonite and Anaconda were the dominant manufacturers. The brown EPR is manufactured only by Kerite, and the pink EPR was manufactured by either Anaconda (9 units reporting use) or Okonite (21 units reporting use). One unit reported having both Anaconda and Okonite pink EPR.

Figure B-1 Manufacturers of 5-kV Cable

A total of 24 units reported having 8-kV rated cable in use on systems operating at 6.9 kV. Plants with 6.9-kV systems tended to be constructed in the late 1970s and beyond, so that a lower percentage of the plants had black EPR insulations for 8-kV rated cables than for the 5-kV rated population. In addition, 6.9-kV systems were not adopted by most utilities, even in later plants. Some later plants remained with 4.16-kV systems for the bulk of the plant and used 13.8-kV systems to supply large loads such as the circulating water pumps. Table B-2 provides a distribution of the insulation types in use in 8-kV cables.

B-3

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01 Table B-2 Distribution of 8-kV Cable Insulation Materials Insulation

Units

Percent of Reporting Units

Black Ethylene-Propylene Rubber

8

11

Brown Ethylene-Propylene Rubber

2

3

Red Ethylene-Propylene Rubber

20

27

Cross-Linked Polyethylene

7

9

A total of 40 units reported having 15-kV rated cables operating at 13.2-kV to13.8-kV. These cables are used for distribution to large loads and feeds to 13–to-4 kV transformers. Table B-3 provides the distribution of the insulation types in use. Table B-3 Distribution of 15-kV Cable Insulation Materials Insulation

Units

Percent of Reporting Units

Butyl

1

1

Black Ethylene-Propylene Rubber

21

28

Brown Ethylene-Propylene Rubber

9

12

Red Ethylene-Propylene Rubber

26

35

Cross-Linked Polyethylene

9

12

A total of 20 units reported use of cables rated at 25 kV and 35 kV. The 35-kV cables were used in 34.5-kV operating systems associated with off-site power feeds. The 25-kV cables were used in 22-kV circuits between the generator output and auxiliary transformers. Rather than having extruded polymer insulation, some cables are insulated by oil-impregnated paper with an overall sheath of lead. Only a few plants reported using paper-insulated leadcovered cable, and even at those plants, extruded polymer insulation was used on the bulk of the installed cables. Table B-4 provides the distribution of the insulation materials on these cables. Table B-4 Distribution of 25-kV to 35-kV Cable Insulation Materials

B-4

Insulation

Units

Percent of Reporting Units

Black Ethylene-Propylene Rubber

7

9

Brown Ethylene-Propylene Rubber

3

4

Red Ethylene-Propylene Rubber

1

1

Cross-Linked Polyethylene

2

3

Paper-Insulated Lead-Covered

4

5

Unknown

2

3

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

B.2.4 Shielding At greater than 5 kV, cables are typically manufactured with insulation shields; however, 5-kV EPR cables can be manufactured and purchased with or without shields. At 5-kV and greater, XLPE cables have insulation shields. Excluding the general-service cables, 22 units (30%) reported having a total of more than 271 circuits with nonshielded EPR cables. Two plants reported having nonshielded cables, but they did not indicate the quantity. These cables were used in safety, fire protection, operationally important, station black-out, and off-site feed cables. The lack of a shield on the EPR cables is not a reliability issue. Circuits without shields represent an electrical testing issue. Electrical testing at high voltage requires a uniform ground plane. An insulation shield provides such a ground plane. Circuits without a shield would not have a uniform ground plane, and available electrical testing is unlikely to provide useful results. For the cables rated above 5 kV, nearly all the cables were reported to have an insulation shield. A few entries indicated that the respondent did not know; 42 units reported having 5-kV cables with shields. This group with shields included all the XLPE cables and a portion of the EPR cables. Most shields involve a helically wrapped flat copper tape over a semiconducting extruded layer. Earlier-style cables used a cotton or polymer semiconducting tape instead of the extruded layer. Anaconda UniShield cables have 6 neutral wires located in the semiconducting extruded shield and jacket. B.2.4.1

Underground Wet-Duty Failure Assessment

Plants with No Failures Of the 74 units that provided data regarding failure experience, 53 units (72% of reporting units) had no failures of medium-voltage, wet underground cable to date. Figure B-2 shows the age distribution of the plants that have not experienced a failure. This figure shows that 23 of the units with no failures are older than the average age of the fleet and 30 are younger than the average. Nineteen units with no failures are 30–35 years old. (To date, 65% of the reporting units still have not experienced a wet medium-voltage cable failure.)

B-5

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

7 Average Age of Fleet 6

Number of Units

5

4

3

2

1

0 1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

31

33

35

Age of Units with No Failures

Figure B-2 Age Distribution of Units with No Failures

Plants with Failures Of the 74 reporting units, 21 units at 15 plants (28% of reporting units) experienced failures of medium-voltage, wet or possibly wet, underground cable. The 21 units experienced a total of 50 failures in circuits that were safety-related, fire protection, off-site power, station blackout, or operationally important. General-service cables were excluded when sufficient information was provided, because these circuits are not related to safety or power production and are likely to be treated differently with respect to installation and operational practices. Events not related to the wet section of the cable were excluded from the analysis. Cables that were replaced on the basis of test results or as an extended corrective action resulting from failure of a similar circuit were also excluded, because they were replaced before failure. Table B-5 summarizes the number of failures per plant reporting failures. The table indicates that the 6 plants with 3 or more failures account for 72% of the failures.

B-6

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01 Table B-5 Number of Failures per Plant Reporting Failures Failures Per Plant

Plants Reporting Failures

Failures

1

4

4

2

5

10

3

1

3

4

2

8

5

1

5

10

2

20

A parallel evaluation of related data from the Institute of Nuclear Power Operations’ Equipment Performance Information Exchange (EPIX) and Nuclear Plant Reliability Data System (NPRDS) indicates that the failures reported by the 74 units responding to the NEI survey appear to be all the medium-voltage, wet cable failures that occurred before 2006. Figure B-3 shows the age distribution at time of failure for all cable types. The age distribution at time of failure is quite broad, with at little as 5 years and as many as 30 years of service before failure. 7

6

Number of Failures

5

4

3

2

1

0 1

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 Years to Failure

Figure B-3 Age Distribution of Wet Cable Failures for All Insulation Types

B-7

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

Figure B-4 shows the distribution of XLPE cable failures in relation to all failures. The total distribution of failures are shown as gray bars, and the XLPE failures are shown in color. Twelve XLPE failures occurred at 4 plants. Eight of the failures occurred at one plant in a specific type of filled XLPE that was used only at that plant. These 8 failures represent 16% of the total wet underground cable failures. 7

6

Number of Failures

5

4 Filled XLPE Others All Failures

3

2

1

0 1

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 Age at Failure

Figure B-4 Age Distribution of Wet Cross-Linked Polyethylene Cable Failures

Figure B-5 shows the distribution of EPR failures in relation to all failures. Black EPR is the dominant, original insulation in the industry (88% of units reported usage). Pink EPR was installed in units that became commercial in the 1980s and beyond; it has become the dominant replacement insulation. The failure of the pink EPR at five years was determined to be related to a manufacturing flaw, causing a large contaminant at the failure site. The three pink EPR failures at 10 years occurred at one plant; however, the cause of failure was not given.

B-8

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01 7

6

Number of Failures

5

4

Black EPR Red EPR All Failures

3

2

1

0 1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

Age at Failure (years)

Figure B-5 Age Distribution of Wet Ethylene-Propylene Rubber Cable Failures

B-9

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

Figure B-6 shows the distribution of butyl rubber failures in relation to all failures. Although the number of butyl rubber failures appears small, only 4 of the reporting units indicated use of butyl rubber cables. 7

6

Number of Failures

5

4 Butyl Failures All Failures 3

2

1

0 1

3

5

7

9

11

13

15

17

19

Years to Failure

Figure B-6 Age Distribution of Butyl Rubber Cable Failures

B-10

21

23

25

27

29

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

Figure B-7 shows the age of the cables at the time of their failure. The general trend in the age at failure is increasing. The trend in the number of failures at a given age is relatively stable, but it shows a slight increase with time. 35

Age at Failure/Number of Failures

30

25

20

Age at Failure # of Failures Linear (Age at Failure) Linear (# of Failures)

15

10

5

0 1975

1980

1985

1990

1995

2000

2005

2010

Year of Failure

Figure B-7 Age of Wet Cable at Failure Versus Year of Failure

B-11

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

Figure B-8 shows the age distribution at failure versus the year of failure for EPR cables. The trends for both black and pink EPR are parallel with time to failure increasing. The population of black EPR cables is larger than that of pink EPR cables. The first installations of pink EPR cables occurred approximately 10 years after installation of black EPR cables began. 35

Age at Failure/Number of Failures

30

25 Black EPR Pink EPR

20

# Black EPR Failures # Pink EPR Failures Linear (# Pink EPR Failures)

15

Linear (# Black EPR Failures) Linear (Black EPR) Linear (Pink EPR)

10

5

0 1985

1990

1995

2000

2005

2010

Year of Failure

Figure B-8 Age at Time of Failure for Wet Ethylene-Propylene Rubber Cables

B-12

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

Figure B-9 shows the age distribution at failure versus the year of failure for XLPE cables. The six failures through 1985 are of one unique, filled XLPE insulation system at one plant. Failures after that point are a mix of the filled XLPE cables at that one plant and nonfilled XLPE cables at several plants. 30

Age at Failure/Number of Failures

25

20

Age at Failure

15

10

5

0 1975

1980

1985

1990

1995

2000

2005

Year of Failure

Figure B-9 Age at Time of Failure for Wet Cross-Linked Polyethylene Cables

B.3

Results of Utility Responses to Generic Letter 2007-01

In February 2007, the NRC issued Generic Letter 2007-01, “Inaccessible or Underground Power Cable Failures That Disable Accident Mitigation Systems or Cause Plant Transients,” which required utilities to supply failure data on low- and medium-voltage power cables that are inaccessible (for example, located in ducts or direct buried) [4]. The letter also required utilities to supply a description of the condition monitoring and test methods in use for assessing the condition of inaccessible cable. A separate NRC letter [102] indicated that Generic Letter 2007-01 applied to ac power cables rated between 480 V and 15,000 V. Copies of the responses were obtained and evaluated. This section summarizes the results. In some cases, the NRC staff issued requests for additional information; the results of those requests are not included in this report.

B-13

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

B.3.1 Summary of Results Review of the licensees’ responses to the survey indicated that 61 units had no failures on inaccessible medium-voltage cables, 47 units had neither low nor medium-voltage inaccessible cable failures, and 14 units had low-voltage inaccessible cable failures but no medium-voltage cable failures. Figure B-10 summarizes the responses to Generic Letter 2007-01 for plants reporting mediumvoltage cable failures. The only distinctly different results from the NEI 2005 survey were the 43 events in 1993 reported by Browns Ferry. Although the cause was not listed in the response to the NRC, discussions with Tennessee Valley Authority personnel indicated that these events were associated with rodent damage during a particularly cold winter. The rodents were believed to have chewed on the cables and damaged the shield-to-insulation interface, leading to the failure of a large number of cables in the remainder of the year. The data indicated that eight plants had a single cable failure, eight plants had two failures each, three plants had three failures each, four plants had 5 to 7 failures, and one plant had 13 failures. If the 43 rodent-related events are ignored, Browns Ferry would have experienced 7 failures.

Figure B-10 Inaccessible Cable Failures by Plant, from Responses to Generic Letter 2007-01 [4]

The six plants with five or more failures account for 73% of the 121 total failures, and the remaining 19 plants with one to three failures account for 27%. This distribution is heavily skewed by the plant with 43 events associated with physical damage to the cables. If these failures are ignored, 78 failures remain. The six plants with five or more failures account for 56% of the failures. Figure B-11 shows the number of failures by year. The 1993 number is an outlier associated with rodent problem at Browns Ferry. This chart appears to indicate that the number of failures per year is increasing; however, the chart is not normalized to reflect the number of units in service in the United States. B-14

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

Figure B-11 Number of Failures of Wet Cable by Year

Figure B-12 shows the failures with respect to the age of the cable. This chart indicates that the failure rate is relatively constant and shows no peaking with respect to age of the wet cable. The large spike at year 19 is once again associated with the rodent-related events at Browns Ferry in 1993.

Figure B-12 Wet Medium-Voltage Cable Failures Versus Age at Time of Failure

B-15

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01

Figure B-13 provides an analysis of the 68 failures for which insulation type was listed in the response. The data indicate an increasing trend in black EPR failures with length of service and a stable trend with time for XLPE. The data seem to indicate a decreasing failure rate for pink UniShield type cables. The three pink EPR failures seem to be “infant mortality” type failures. Too few data points exist for the brown nonshielded EPR and butyl rubber cables to identify a trend. 5

XLPE Black EPR Pink Uni-Shield Pink EPR Brown Unshielded EPR Butyl Rubber

Number of Failures

4

3

2

1

0 1

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

Age at Failure

Figure B-13 Age of Cable at Time of Failure By Insulation and Design

B.3.2 Assessment Methods Generic Letter 2007-01 requested that utilities describe activities that either assess the condition of the cable or preclude long-term wetting [4]. Table B-6 summarizes the actions associated with assessment and testing of medium-voltage cables. Although one would assume that pumping and dewatering of manholes would occur if manhole inspections identified water, some responses did not explicitly state that dewatering would occur. Accordingly, the responses for inspections and those for pumps and dewatering are stated separately.

B-16

Results of Nuclear Energy Institute Survey and Nuclear Regulatory Commission Generic Letter 2007-01 Table B-6 Summary of Assessment and Testing Responses from Generic Letter 2007-01 Number of Plants Committed

Percentage of 65 Plants

Manhole Inspection

25

38.5

Manhole Pumps and Dewatering

20

30.8

Very-Low-Frequency Tan δ

5

7.7

Very-Low-Frequency High Potential

3

4.6

On-Line Partial Discharge Testing

3

4.6

Power Factor Testing

2

3.1

DC High Potential (Mostly of Motor with Cable; One Cable Separately)

9

13.8

Insulation Resistance Test with Load

47

72.3

Polarization Index with Load

20

30.8

Time-Domain Reflectometry (TDR)

2

3.1

Electromagnetic Interference (with Motor)

1

1.5

GL 2007-01 Medium-Voltage Cable Assessment and Testing Responses

B.4 Comparison of the Nuclear Energy Institute Survey and Generic Letter 2007-01 Results The NEI survey results provided information on the installed and replacement cable population, which was not included in the request for information in Generic Letter 2007-01 [11, 4]. The Generic Letter 2007-01 results included all the U.S. units, whereas the NEI survey results provided failure information for 75 units. The trends in both sets of results are in agreement. The data show a relatively constant failure rate for the overall population of cables. However, when examined individually, the results for black EPR indicate an increase in failure rate with age. Black EPR has the most failures in both sets of data, which is consistent with black EPR being the dominant insulation system in use (65% of responding plants indicated having black EPR in the NEI survey). Although pink EPR is the dominant replacement insulation, the population of black EPR cables still exceeds the size of the population of pink EPR circuits. When the 43 events that occurred at one plant, likely due to rodent damage, are discounted, the results from the two data sets indicate similar numbers of failures and proportions of failures by insulation material type.

B-17

C

TAN δ DATA FOR ETHYLENE-PROPYLENE RUBBER AND BUTYL RUBBER INSULATED CABLES This section provides tan δ data from in-plant and laboratory tests for ethylene-propylene rubber (EPR) and butyl rubber. The EPRI report Aging Management Program Guidance for Medium Voltage Cable Systems for Nuclear Power Plant (1020805) provides assessment criteria based on the data in this section and expert opinion [5]. Because butyl rubber and EPRs are compounds, variations in normal tan δ can be expected between manufacturers’ formulations and between generations. Some variations (~20%) have been noted within a particular manufacturer’s insulation of the same generation. In addition, circuit configurations can affect criteria. One plant that has nonshielded cable for dry segments of cable in series with shielded cables in wet locations has noted much higher tan δ results for normal condition. The following subsections evaluate data from laboratory and field tests to help identify criteria for retaining and replacing EPR and butyl rubber cable based on tan δ results. The approximate range of normal and abnormal results can be identified. The data are provided for specific manufacturers’ formulations and vintages. The data might not apply to other manufacturer’s formulations or vintages of cable.

C.1

60-Hz Tan δ Data from 1972 Black Okoguard Insulation

The EPRI report Failure Mechanism Assessment of Medium Voltage Ethylene Propylene Rubber Cable (1015070) describes the evaluation of 10 segments of an Okonite black Okoguard cable that was removed from service after a failure following 34 years of service under wet conditions [101]. The cable had black EPR insulation. The cable was received by the laboratory in segments ranging from 42 to 70 in. (1.07 to 1.78 m) in length. As part of the assessment, the segments were subjected to 60-Hz tan δ (dissipation factor) testing, followed by ac breakdown tests. Figure C-1 shows the results. The insulation was an average of 170 mils (4.3 mm) thick.

C-1

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

Figure C-1 AC Breakdown Strength Versus 60-Hz Tan δ Results

The line in Figure C-1 is a regression curve obtained by the formula y = c + a*exp(-bx) with R2 (coefficient of determination, equal to 0.83), c = 0.26, a = 56.4, and b = 0.69. The graph demonstrates the predicted breakdown strength of the sample up to 9 × 2.4 kV, based on dissipation factor measured at the cable operating voltage of 2.4 kV. After the field faults were removed from the segments, the lowest laboratory breakdown voltage was four times the operating voltage for the cables. The dry result was obtained by drying a cable segment in a C-2

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

203°F (95°C) circulating air oven for 5 days. The data indicate that, for a 60-Hz measurement, a tan δ of 40 × 10-3 the breakdown strength of the insulation is 4 × V0 (9.6 kV), which cable experts consider extremely weakened insulation (~70 V/mil [2.7 kV/mm]). With a tan δ of 4 × 10-3, the breakdown strength is about 8.6 × V0 (20.6 kV), which cable experts consider highly weakened (147 V/mil [5.8 kV/mm]). Industry experts consider that a value of 200 V/mil (7.9 kV/mm) is indicative of end of life for an XLPE cable. The dry value for the black EPR indicates that the ultimate breakdown voltage for the material was in excess of 55 kV or 394 V/mil (15.5 kV/mm). New cable would likely have a breakdown strength of twice this value. The failure had been cut out of the cable before these tests were performed, and the sections with elevated tan δ and low breakdown voltages were determined to have channels of localized low insulation resistance through the depth of the insulation. The data indicate that there is an inverse correlation between tan δ and breakdown voltage. These tests were performed with a 60-Hz test set because of the short lengths involved. The results are not directly comparable to results from VLF testing, which would likely have values on the order of twice the 60-Hz values. Figure C-2 shows the 60-Hz tan δ results for the same test specimens by voltage step. Segments 2A and 2B resulted when the fault was cut from the original specimen 2. The removal of the fault left two segments with reasonably low tan δ measurements and reasonable stability between voltage steps. Specimens 3 and 4 had high initial tan δ measures at 0.25 V0 and unstable readings at 0.5 V0, preventing measurements at higher voltages. Specimen 5 had somewhat high tan δ at 0.25 V0 and 0.5 V0 and had a step change increase at 0.75 V0, with instability at 1V0. These three specimens were found to have low resistance channels between the conductor and shield where there were small areas having 1/10 or less of the surrounding insulation resistance. When specimen 4 was evaluated, an additional failure was identified. When it was cut out, the remaining segment had tan δ measurements between 5.9 × 10-3 and 5.3 × 10-3 through all the voltage steps. The data for all voltage steps are shown in Figure C-2 because many of the cable subsections were degraded to the point that testing at 1V0 and 2V0 was not possible. Most other data in this report are given in terms of 1V0 and 2V0 and the differential between the two.

C-3

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

90

80

70

60 hz Tan Delta (E-3)

60 0.6 kv 1.2 kV 2.4 kV 3.6 kV 4.8 kV Diff 2Vo-Vo

50

40

30

20

10

0 1

2

3

4

5

6

7

8

9

10

11

Specimen

Figure C-2 60-Hz Tan δ Results by Test Voltage

The data in Figure C-1 show that there is an inverse correlation of tan δ with breakdown strength. Elevated tan δ indicates that the insulation is weakened and susceptible to in-service failure. Although Figure C-2 was developed in an attempt to show that significant increases in tan δ are indicative of deterioration, only specimen 5 tends to show this condition. However, specimens 3 and 4 partially support this concept, in that the tan δ measurement became unstable at the second voltage step.

C.2 Very-Low-Frequency Tan δ Results for Okonite Black EthylenePropylene Rubber Insulation Figures C-3 and C-4 provide VLF tan δ results for cables with Okonite black EPR insulation. The results in Figure C-3 indicate that VLF tan δ results are approximately 1.8 times that for good 60-Hz results, as seen by the dry specimen value (2.8 × 10-3) in Figure C-1. The cables that were tested to obtain Figures C-1 and C-3 are similar 5-kV cables. The cables for Figure C-4 are 15-kV rated cables.

C-4

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables 6

5

Tan Delta (E-3)

4

1.1 kV 3.3 kV Diff 1.1 to 3.3 kV

3

2

1

0 1

2

3

Figure C-3 Very-Low-Frequency Tan δ Result for a 5-kV Okonite Black Ethylene-Propylene Rubber Cable

C-5

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables 12

10

Tan Delta (E-3)

8

8 kV 16 kV Diff 8 - 16 kV

6

4

2

0 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

Figure C-4 Very-Low-Frequency Tan δ Results for 15-kV Okonite Black Ethylene-Propylene Rubber Cable

C.3

Tan δ for Anaconda Pink Ethylene-Propylene Rubber UniShield

The EPRI report Failure Mechanism Assessment of Medium Voltage Ethylene Propylene Rubber Cable (1015070) describes the evaluation of five segments of 15-kV rated Anaconda UniShield cable with pink EPR insulation manufactured in 1979 that was removed from service after a failure following 20 years of service (1984–2004) under wet conditions [101]. After the faulted segment was removed, the remainder was in good condition. Figure C-5 provides 60-Hz tan δ measurements for the cable segments, which ranged from 85 to 233 ft (25.9 to 71 m) in length. Figure C-6 provides the VLF (0.1 Hz) tan δ for the same cables. The VLF tan δ is normally higher than the 60-Hz measurement. In this case, the increase is approximately a factor of 1.8 to 1.9.

C-6

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables 3.5

3

Tan Delta (E-3)

2.5

2 8 kV 16 kV 8-16 kV Diff 1.5

1

0.5

0

A11

2

A2

3

B1

C14

5 C2

Specimen Figure C-5 60-Hz Tan δ for Anaconda UniShield Pink Ethylene-Propylene Rubber 8

7

6

Tan Delta (E-3)

5

8 kV 16 kV 8-16 kV Diff

4

3

2

1

0

A1

1

A2 2

B1 3

C14

C25

Specimen Figure C-6 Laboratory-Measured Very-Low-Frequency Tan δ for Anaconda Pink Ethylene-Propylene Rubber UniShield

C-7

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

Figure C-7 shows field VLF tan δ results for 15-kV Anaconda UniShield cables. This chart shows good through moderate condition results. The results for two dry cables (six individual phases) are shown at the left of the plot. The remainder of the cables were wet for 50% to 57% of their length. The order of this plot is by the value of the initial measurement at 8 kV. The dry results and many of the wet cable results correlate well with the results shown in Figure C-6 for the laboratory tests on the specimens after the damaged sections were removed. (Each cluster of two voltages and a difference represents the test of one cable phase.) 30

25

Tan Delta (E-3)

20

8-kV 16-kV Diff 16-8 kV

15

10

5

0 1

3

5

7

9

11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57

Figure C-7 Tan δ Results for Anaconda Pink UniShield Cables (Ordered by 8-kV result)

C-8

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

Figure C-8 shows the same field VLF tan δ measurements arranged in the order of the difference between the 2V0 (16-kV) and 1V0 (8-kV) measurements. When ordered by this difference, the results tend to also be arranged in the order of most elevated 1V0 to 2 V0 result. (Each cluster represents the test of one cable phase.) 30

25 8-kV 16-kV Diff 16-8 kV

Tan Delta (E-3)

20

15

10

5

0 1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

31

33

35

37

39

41

43

45

47

49

51

53

55

57

Figure C-8 Tan δ Results for Anaconda Pink UniShield Cables (Ordered by Difference in 16 kV and 8kV Results)

C-9

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

Figure C-9 shows the field data in order of the value at 2V0. In this figure, entries with 2V0 values less than 10 × 10-3 have been omitted, and a number of entries have been added for cables having much higher tan δ measurements. Some of these tests with higher measurements were stopped before the 1.5 V0 or 2.0 V0 measurement was taken, because the results indicated significant problems and possible failure during the test. (Each cluster of four voltages and a difference represents the test of one cable phase.) 140

120

4-kV 8-kV 12-kV 16-kV Diff 16-8 kV

100

Tan Delta (E-3)

80

60

40

20

0 1

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

Figure C-9 Tan δ Results for Anaconda Pink UniShield Cables (Showing Moderate to High Results)

Figure C-10 shows tests that occurred approximately one year apart for four Anaconda UniShield cables. The cables were tested in either the winter or the spring. Cable C-3 (arbitrary circuit number for this report) shows acceptable tan δ values with little difference between the measurements taken at 1V0 and 2V0. The A and C phases of cable C-1 show stable and good results from year to year. The B phase has elevated results at 2V0 and a differential that is nearly equal to the measurement at 1V0 in 2007. In 2008, the values for the B phase have dropped somewhat, which could be the effect of temperatures or bulk wetting of the cable at the time of the test. For cable 2, the tests one year apart indicate that the A and B phases have nearly identical behavior. However, the C phase has a more significant change at 2V0 in 2008 than in 2007 and the difference between 1V0 and 2V0 has nearly doubled, indicating a worsening leakage current. Finally, the data for cable C4 show that all three phases have improved in a similar manner from the 2006 tests to the 2007 tests, indicating again that the temperature might have been lower or the circuits might have been drier during the 2007 tests. Given that these cables are showing uniform behavior between the phases, it is likely that the degradation is not worsening significantly and has remained stable. C-10

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables 25

20 8-kV 16-kV Diff 16-8 kV

Tan Delta (E-3)

15

10

5

4-

B2

00 7

6

00 7 C

00 C

4-

A2

00 6

-2

C

4C

00 6

B2 4-

C

00 C

4-

A2

8

00 8

-2

B2

3C

3C

C

6 00

A2

3C

00 8

00 6

-2

B2

3C

3C

C

8 00 -2

A2

3C

00 6

00 8

B2

2C

2C

C

7 00

00 8

00 7

-2

A2

2C

2-

2C

B2

C

8

00 7

00

A2

2C

C

00 8

-2

B2 1-

1C

C

7

00 8

00

A2

1C

C

00 7

-2 1C

C

1C

C

1-

A2

B2

00 7

0

Figure C-10 Tan δ Measurements for Anaconda UniShield Cables Taken One Year Apart

The data shown in Figure C-10 indicate that comparing results between phases and between years for multiple phases is important to determine whether apparent changes are the result of environmental conditions at the time of the test or are actual changes in the condition of the cable insulation.

C.4

Anaconda Black UniShield Ethylene-Propylene Rubber

Figure C-11 shows the available results for VLF tan δ testing of Anaconda UniShield with black EPR insulation. The tan δ for two of the phases had step increases when the voltage was raised from 2 kV to 4 kV, and the tests were stopped at 4 kV. The tan δ measurements for the remainder of the cables under test were reasonably stable with voltage increases. (Each cluster represents the test of one cable phase.)

C-11

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables 18

16

14

Tan Delta (E-3)

12

2 kV

10

4kV 6 kV 8 kV

8

Diff 4 - 8 kV

6

4

2

0 1

2

3

4

5

6

7

8

9

10

11

12

13

14

Figure C-11 Tan δ Measurements for Anaconda UniShield with Black Insulation

C.5

Tan δ Results for Butyl Rubber Insulation

An extensive failure assessment of Okonex 5-kV butyl rubber cables that had failed in service was conducted [103]. The insulation in these specimens was approximately 200 mils thick. The failures were caused by long-term water related degradation. Five sections of cable were assessed. Two had failed in-service or in post-failure testing. Three sections were removed from non-faulted sections of the cables. During the course of the assessment, 60 Hz tan δ tests were performed followed by ac breakdown tests. The faults were removed from the cables and the remaining sections were subjected to tan δ tests and breakdown tests. The efforts continued until the sections that remained had significantly improved breakdown strengths. Figure C-12 shows the relationship of 60 Hz tan δ to breakdown strength.

C-12

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables 1000 800 600 500

60 Hz Dissipation Factor at Cable Operating Voltage, tan (E-3

400 300 200

100 80 60 50 40 30 20

10 8 6 5 4 3 2

1 0

2

4

6

8

10

12

14

16

18

20

AC Breakdown Voltage/Operating Voltage

Figure C-12 Tan δ Versus Breakdown Strength for Okonex Butyl Rubber

C-13

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

Figure C-13 compares the 60 Hz tan δ for black Okonite EPR as shown in Figure C-1 to that for Okonite butyl rubber from Figure C-12. The characteristic shape of the plots is essentially identical for the two materials such that as the breakdown strength drops below 10 times operating voltage, the tan δ value is beginning to increase. For the butyl and black EPR rubber specimens in the assessments, the 10 times operating voltage value is 24,000 Vac. Due to insulation thickness differences the voltage stress for the two materials was significantly different. For the butyl rubber, it was 120 V/mil (4.7 kV/mm) and for the black EPR, it was 141 V/mil (5.5 kV/mm). At a breakdown strength of eight times operating voltage (19,200 Vac), the tan δ for both materials has doubled (2X for EPR, 2.4X for butyl) from its lowest value. As the breakdown strength decreases to four times operating voltage, the tan δ has increased by a factor of 16 for the black EPR and by 24 for the butyl rubber. The comparisons show that the lossier butyl rubber will have characteristically higher 60 Hz tan δ and have a larger change than the black EPR as it deteriorates. 1000

Butyl Black EPR

Tan Delta (x 10-3)

100

10

1 2

4

6

8

10

12

14

16

18

20

22

24

ac Breakdown Voltage/Operating Voltage

Figure C-13 Comparison of 60-Hz Tan δ Versus Breakdown Strength of Okonite Butyl Rubber to Black Ethylene-Propylene Rubber

C-14

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

Figure C-14 shows the same data as Figure C-13 but in terms of volts per mil. The figure shows that tan δ does not change for Okonite black EPR and butyl until the breakdown strength has decreased below 200 to 240 V/mil (7.9 to 9.5 kV/mm). For both materials, tan δ is slightly greater than 10 × 10-3 when the ac breakdown voltage has dropped to 100 V/mil (3.9 kV/mm). When tan δ has increased to 30 × 10-3, the breakdown strength has dropped to 75 V/mil (2.9 kV/mm). 1000

Tan Delta (E-3)

100

Butyl Black EPR

10

1 0

50

100

150

200

250

300

350

400

ac Breakdown (V/mil)

1 V/mil = 0.0394 kV/mm

Figure C-14 AC Breakdown Voltage (in V/mil) for Okonite Butyl and Black Ethylene-Propylene Rubber Insulations

The EPRI report Advanced Diagnostics and Life Estimation of Extruded Dielectric Cables (1001727) also describes partial discharge (PD) testing performed on specimens in the laboratory [88]. These butyl rubber cables had high natural attenuation due to the characteristics of the butyl rubber and had corroded but intact helically wrapped copper tape shields. When the full length of one of the cables with severe water-related degradation was tested, PD could not be detected above background. However, an appreciable PD (50–150 pC, depending on test voltage) could be detected on the full length of the second cable . As both of the cables were cut into shorter sections following breakdowns, PD was identifiable in each section with elevated tan δ. The result of the PD measurements indicates that PD can be present in severely deteriorated butyl rubber cable that is near failure; however, the combination of high attenuation of the butyl rubber and shield corrosion that causes the tape shield to act as an inductor and further attenuate the signal might make detection of PD difficult to impossible. C-15

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

C.6

Conclusions Related to Rubber Insulated Cables and Tan δ Results

IEEE Std 400-2001 provides acceptance criteria for the use of tan δ measurements for XLPE insulated cables [51]. As further field and laboratory data are gathered, that guidance is expected to be updated. This appendix was developed to provide insights regarding the response of various EPR and butyl insulations during tan δ testing and the implications of the results. It is recognized that the data available are limited, but these data do provide valuable insights. Figure C-1 for black EPR and Figure C-12 for butyl rubber indicate that tan δ is inversely proportional to ac breakdown voltage and that, when the breakdown voltage decreases below 10 times operating voltage, tan δ increases exponentially with a drop in breakdown voltage. It is likely that all the rubber insulations behave similarly; however, it should not be assumed that each manufacturer’s black or pink EPR has exactly the same characteristics as that of another manufacturer. It is likely that there will be differences in the initial tan δ and the rate of change as the breakdown voltage increases. The field results indicate that a baseline tan δ can be determined by testing a few cables that are known to be dry. Even after extended periods, the tan δ for such circuits appear to be like new unless there is some non-water-induced flaw in the insulation or an accessory. Figures C-1 and C-12 show 60-Hz tan δ results for black Okonite EPR and black Okonite butyl rubber. Corresponding VLF figures do not exist. However, a comparison of “healthy” results for the materials indicates that a VLF results can be a factor of 1.8 higher than the 60-Hz results. It is not clear that this factor holds true at all degradation values. If it does, it indicates that severely aged black Okonite insulation having ~70 V/mil (2.7 kV/mm) or 4V0 breakdown strength would have a tan δ of approximately 72 × 10-3, as opposed to a healthy tan δ of ~5 × 10-3. A similar factor for black Okonite butyl rubber would indicate an initial VLF value of ~9 × 10-3. The specimens evaluated had a thick layer of insulation (190 mils [4.8 mm]); accordingly, 70 V/mil (2.7 kV/mm) would occur at 5.5V0, which corresponds to a VLF of ~68 × 10-3. Figures C-13 and C-14 show that the 60-Hz tan δ versus ac breakdown strength behavior of EPR and butyl rubber are similar. When the data are given in terms of volts/mil, the degradation graphs overlap as breakdown strength degrades. If the 1.8 factor between 60-Hz and VLF tan δ hold true, then when the VLF tan δ is approximately 20 × 10-3, the ac breakdown strength would be approximately 100 V/mil (3.9 kV/mm). A VLF tan δ of 54 × 10-3 would indicate that the breakdown strength is approximately 75 V/mil (2.9 kV/mm). Although these values might seem to be well within acceptable conditions (for example, on 110-mil [2.8 mm] insulation, the breakdown strength would be 8250 volts at 75 V/mil (2.95 kV/mm) or a factor of 3.4 times 2400 V), cable experts would consider the cables highly degraded. In fact, cable experts consider cables with a breakdown strength of 200 V/mil (7.9 kV/mm) highly degraded. When the breakdown strength has deteriorated to 200 V/mil (7.9 kV/mm) or as little as 75 V/mil (2.95 kV/mm), the insulation system might not be able to withstand switching and lightning surges without an electrical tree forming, leading to failure shortly thereafter. Accordingly, conditions indicating more than doubling of a healthy tan δ indicate that more frequent testing is necessary to confirm that the condition is stable. Values indicating 75 V/mil (2.95 kV/mm) or less indicate that replacement or rejuvenation should be considered.

C-16

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

The laboratory VLF tan δ results for Anaconda pink EPR are consistent with healthy field results. Unfortunately, no complete tan δ versus ac breakdown curve is available for the Anaconda pink EPR. Figure C-8 indicates that increasing tan δ relates to deterioration, as does the difference in tan δ when measured at 1V0 and 2 V0. From Figure C-8, there appears to be a strong tendency of differential increases of tan δ with increasing test voltage for the Anaconda cable. Figure C-9 indicates that, in some cases with a tan δ of above approximately 30 × 10-3, the differential increases with increasing test voltage are so large that testing at 1.5V0 to 2V0 cannot be safely accomplished and that investigation of the need to replace the cable will be necessary.

C.7

Effects of Mixed Shielded and Nonshielded Segments on Tan δ

One plant reports having circuits in which the dry sections have nonshielded cable and underground (wetted) sections have shielded cable. This is an unusual cable system design, and the results might have extremely limited value to plants with nonshielded cable. The nonshielded cables are three-conductor cables with an overall armor. These cables were all black EPR manufactured by Okonite. Figures C-15 and C-16 show the VLF tan δ results for the cables with mixed nonshielded and shielded segments. The test voltages for these two figures were slightly different, requiring the data to be plotted separately. (Each cluster of test voltages represents one phase of a cable.) 90

80

70

1.1 kV 2.2 kV 3.4 kV 4 kV

Tan Delta (E-3)

60

50

40

30

20

10

0 1

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 39

Figure C-15 Tan δ Results for Mixed Shielded and Nonshielded Black Okonite Ethylene-Propylene Rubber Circuits

C-17

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

Figure C-16 Tan δ Results for Mixed Shielded and Nonshielded Black Okonite Ethylene-Propylene Rubber Circuits

In Figure C-15, the cable tests represented by markers 37, 38, and 39 are a special case for two reasons: 1) only 15% of the cable was shielded and 2) each phase had three conductors in parallel. The large amount of nonshielded cable significantly elevates the tan δ. The further elevation of the tan δ results could be caused by having the nonshielded cables between the test set and the shielded sections. The effects of the length of the nonshielded sections on the tan δ is described below. In Figure C-16, the data for a three-phase cable are associated with x scale markers 4, 5, and 6. The data for the A phase of the cable is shown at scale marker 4. This phase had elevated tan δ results and a significant differential between 1V0 and 2 V0. The aged wet section of the cable was replaced with modern Okonite pink EPR. Also in Figure C-16, the C phase (marker 27) of cable having the x scale markers 25, 26, and 27 shows similar degradation. The utility plans to replace the cable based on the tan δ measurements that have been observed.

C-18

Tan δ Data for Ethylene-Propylene Rubber and Butyl Rubber Insulated Cables

Figure C-17 shows a plot of the average tan δ for the mixed circuits versus the percent of nonshielded cable in the circuits. Although the correlation is not perfect, the tan δ is proportionally related to the percent of the nonshielded cable in the circuit. This result is logical, in that the nonshielded segments would have a lower capacitive current than the shielded circuits. At the same time, there would still be leakage currents in the nonshielded circuits with respect to the armor in the three-phase cable segments. The approximate relationship of tan δ to the percent of nonshielded cable is interesting but not accurate enough to allow an adjustment so that an absolute value of tan δ could be used to assess a cable. However, comparison of results between phases and significant increases in tan δ with increasing test voltage obviously can identify deteriorated insulation, even though the circuits have a mix of shielded and nonshielded segments. 70

60

Average Tan Delta (E-3)

50

40 Avg. Tan Delta Linear (Avg. Tan Delta) 30

20

10

0 0

10

20

30

40

50

60

70

80

90

100

% Unshielded

Figure C-17 Average Tan δ Value Versus Percent of Nonshielded Cable Section

The testing of the circuits was by means of connecting the test set to the conductor and to the station ground. It was not necessary to connect directly to the cable shield connection.

C-19

D

ADDITIONAL POLYMER MATERIALS INFORMATION

D.1

Fundamentals of Elastomers

Any polymer, elastomers included, consists of many (branched) chains having a distribution of molecular weights, and so it is common to refer to average molecular weights. One can refer to the number of molecules of each monomer present in the chain or the molecular weight of the chains themselves. In elastomers, these chains are linked together. These long, intertwined polymer chains are important because they give the rubber (after cross linking) the ability to stretch and to return to their original shape. At rest, the long chains have several bends rather than being straight and parallel. Therefore, when insulation is bent or pulled, the insulation can stretch because the long chains can straighten. When the stress is removed, the insulation will return to its original shape Uncured (unvulcanized) elastomers are inherently soft, flexible, and often tacky. If the chains are not cross linked, they will not return to the original shape after relaxation occurs. Therefore, to render them useful, the following two steps must be taken: 

The polymer chains must be extremely long, and different chains must be linked together. This is referred to as cross linking (also referred to as vulcanization or curing). Cross linking is induced after the cable is extruded; it increases the toughness and improves the physical and mechanical properties of the elastomer. Different elastomers use different curing techniques. A cross-linked elastomer is depicted in Figure D-1. The curved lines represent the chains, and the short (almost horizontal and vertical) lines between chains represent the cross links. (See Section D.2, Cross Linking, for details.)



Cross linking alone is not adequate to achieve usefulness. A number of other additives, called fills, must be introduced to the elastomer to attain usefulness. Many of these are inorganic, and one common required ingredient is clay, which imparts toughness. Other ingredients improve mechanical and electrical properties and also facilitate processing and stability. The use of numerous additives requires that sophisticated mixing technology be used to convert the mixture of ingredients into a blend suitable for conversion into cable insulation by extrusion.

D-1

Additional Polymer Materials Information

Figure D-1 A Cross-Linked Elastomer

D.2

Cross Linking

If an elastomer possessed two chains of equal length, the cross linking would lead to an increase in molecular weight; in this case, doubling. Chain length is not a factor in influencing the crosslinking process. As the degree of cross linking increases, more and more chains are incorporated into the cross-linked network. In a practical insulation, a significant majority of the polymer chains are no longer spaghetti-like but are linked together, forming a three-dimensional network. The elastomer’s mechanical properties are dominated by this cross-linked fraction. The crosslinked network is referred to as the gel fraction and the fraction that is not cross linked is referred to as the sol. Practical cable insulations contain a small sol fraction. Insulated Cable Engineers Association and Association of Edison Illuminating Companies specifications require the sol fraction in cross-linked polyethylene (XLPE) to be less than 30% by weight. Because the gel fraction is so important, another key point is noted. In considering the properties that are influenced by cross linking, the length of the chain between the cross links is significant. The closer together the cross links (the small lines in Figure D-1), the lower the length of the chains between the cross links, giving improved mechanical properties to the elastomer.

D-2

Additional Polymer Materials Information

D.3

Butyl Rubber Compositions

The formulations listed in Table D-1 are from the early technical literature; they describe the components present in butyl rubber compounds for wire and cable applications [35, 36]. Table D-1 Butyl Rubber Insulation Components Amount (Relative Weight)

Component Butyl 035 (Enjay)

100

Zinc Oxide

5

Calcined Clay

90

Stearic Acid

0.5-1.5

Catalyst:

See below

Quinone Dioxime Dibenzoate (also referred to as DibenzoGMF)

6

Lead Oxide (Pb3O4)

9

Sulfur

1.5

This formulation had the following properties: 

Tensile strength, 840 psi (0.6 kgf/mm2)



Elongation, 330 %



Moisture absorption, 7 days at 185°F (85°C), 11.1 mg/in2 (1.7 mg/mm2)



Dielectric strength, 550 V/mil (21.7 kV/mm)

A second common compound for wire insulation used a similar polymer-filler composition but a different catalyst system, as shown in Table D-2. Table D-2 Butyl Rubber Wire Insulation Components Component Butyl 035 (Enjay)

Amount (Relative Weight) 100

Zinc Oxide

5

Calcined Clay

90

Stearic Acid Catalyst Quinone Dioxime (also referred to as GMF)

0.5-1.5 See below 1.5

Mercaptobenzothiazole (MBTS)

4

Lead Oxide (Pb3O4)

5

D-3

Additional Polymer Materials Information

The tensile strength for this composition was 930 psi (6.4 MPa); elongation, 290%; moisture sorption, 10.9%; and dielectric strength, 695 V/mil (27.4 kV/mm). The curing agent system plays a significant role in resistance to degradation under thermal stress. The sulfur cure can soften with thermal aging, whereas the lead oxide cure tends to remain stable. Polymers such as phenol-formaldehyde resins can be used as the cross-linking agent for butyl rubber, although no specific notation for wire applications is included in the referenced literature. What is relevant is that each method of inducing cross linking will impart a different nature to the cross-linked regions, different additives in the final rubber, and potentially different responses with aging.

D.4

Ethylene-Propylene Rubber and Ethylene-Propylene-Diene Monomer

Ethylene-propylene rubber (EPR) is a copolymer of ethylene and propylene. Sometimes a third monomer is added to the composition to assist in the cross-linking process; it has no influence in plant operations. This termonomer is polymerized with the ethylene and propylene and is present in rather small amounts (~3-6%). When this third monomer is used, the insulation is referred to as ethylene-propylene-diene monomer (EPDM). The termonomer provides unsaturation (double bonds) and facilitates cross linking by use of sulfur. The term EPDM refers to ethylenepropylene-diene monomer or terpolymer, the latter two referring to the repeating CH2 units in the polymer backbone. Figure D-2 shows the EPDM terpolymer, with ethylene-propylene and 5-ethylidene-2-norbornene as the termonomer. EPDM can be cross linked using sulfur, but this method is not used for modern wire and cable systems.

Figure D-2 Ethylene-Propylene-Diene Monomer Terpolymer

Cable accessories (such as joints and terminations) are made of EPDM, and manufacturers’ literature designates them as EPDM.

D.5

Dicumyl Peroxide Cross-Linking Agent Byproducts

The most common byproducts of the cross-linking process that uses dicumyl peroxide are acetophenone and dimethyl benzyl alcohol (also referred to as cumyl alcohol). Other known byproducts are alpha-methylstyrene and methane gas. All these byproducts are present in small quantities, but they can influence aging and reliability. They diffuse from the cable system over time, with methane diffusing most rapidly. Although the changes due to cross linking do not improve the electrical properties, peroxide cross linking agent byproducts provide superior treeresistant properties to XLPE.

D-4

Additional Polymer Materials Information

D.6

Ethylene-Propylene Rubber Formulations

Exxon [25] has described the process of developing wire and cable insulation formulations based on black EPR. A typical carbon black formulation from the 1970s using silane-coated clay is shown in Table D-3. Table D-3 Typical Black Wire and Cable Insulation Ethylene-Propylene Rubber Compound Components from the 1970s Amount (Relative Weight)

Component EPR Vistalon 404

80

Low-Density Polyethylene

20

Zinc Oxide

5

Coated Clay

110

Carbon Black

10

Flectol H

1.5

Lead Oxide

5

Wax

5

Silane

2

SR-350

1.5

Peroxide

7

Significant properties for this black, filled EPR are shown in Table D-4. Table D-4 Properties of Black Ethylene-Propylene Rubber Compound Shown in Table D-3 Property

Value

Tensile Strength

1210 psi (0.85 kgf/mm2)

Elongation

200%

Elongation Percent Retention After Seven Days at 300°F (149°C)

95%

Moisture Absorption, 194°F (90°C), Seven Days, Percent Weight Increase

0.1% (0.9 mg/in2 [0.14 mg/cm2])

Dielectric Strength

1370 V/mil (53.9 kV/mm)

Dielectric Constant

3.4

Power Factor, Original and One Week, 194°F (90°C) Water

0.005, 0.0044

D-5

Additional Polymer Materials Information

The moisture sorption is greater than an order of magnitude less than that for butyl rubber. Also, the dielectric strength is higher, and the physical properties are excellent. A similar (but not exactly the same) formulation for black power cable insulation from the 1970s was provided by Exxon [26] using Vistalon 2504, an EPDM (see Table D-5). Table D-5 Medium-Voltage Black Ethylene-Propylene Rubber Insulation Using Ethylene-PropyleneDiene Monomer Component

Amount (Relative Weight)

EPDM Vistalon 2504

100

Carbon Black

10

Treated Clay

110

Antioxidant

1.5

Lead Oxide

5

Zinc Oxide

5

Process Oil

15

Paraffin

5

Silane

1

Peroxide

3.5

Formulation components can differ between suppliers. In addition, although the function of the components in any EPR formulation is well known, specific individual components used by individual suppliers can differ from one another.

D-6

Additional Polymer Materials Information

D.7

Influence of Clay on Properties of Ethylene-Propylene Rubber

Table D-6 shows the compounding of two EPRs: one using calcined clay and the other using silane-treated calcined clay, which is more representative of modern cables. Other than the clay types, the polymers are essentially the same. Table D-6 Compounds of EPR Using Calcined and Coated Clay Component

Calcined Clay (%)

Treated Calcined Clay (%)

Vistalon 404

32.7

32.4

Low-Density Polyethylene

8.2

8.1

2

2

Calcined Clay

44.9

44.5

Silane A-172

0

0.8

0.6

0.6

Lead Oxide

2

2

Wax

2

2

SR-350

0.6

0.6

Peroxide

2.8

2.8

4

4

Zinc Oxide

Flectol H

SRF Black

Table D-7 shows the effect of treatment of the clay on EPR properties and the limitations of calcined clay with respect to electrical properties in EPR. The power factor, dielectric constant, and water sorption characteristics all improve in the treated calcined clay system. This indicates that the use of treated calcined clay has been a significant improvement in compounding EPR. Table D-7 Influence of Clay Nature on Ethylene-Propylene Rubber Cable Properties: Calcined Versus Coated Clay Properties

Ethylene-Propylene Rubber with Calcined Clay

Ethylene-Propylene Rubber with Treated Calcined Clay

Power Factor After 194°F (90°C) Water, 2 weeks

0.062

0.0037

Dielectric Constant After 194°F (90°C) Water, 2.5 Weeks

4.09

3.52

0.69 %

0.12 %

Water Sorption After 194°F (90°C), 2 Weeks

Another factor in influencing elastomer properties is clay particle size—the smaller the average size, the greater the polymer–inorganic component interfacial contact and the better the anticipated properties. More even size distributions with smaller-sized particles have been implemented in cable manufacture, as well, to improve electrical and mechanical properties. D-7

E

OFF-LINE TESTS THAT ARE UNDER DEVELOPMENT

E.1

Introduction

Several additional diagnostic tools are being developed but are not commonly used. They are summarized in this section. Additional information is available in the EPRI reports Advanced Diagnostics and Life Estimation of Extruded Dielectric Cables (1013085) and Guide for NonDestructive Diagnosis of Extruded Cable Systems (1001731) [104, 90].

E.2

Isothermal Return Current

The isothermal return current (IRC) test method applies low-level dc to the cable system for a defined time. The treated cable system is then discharged by shorting the cable for a brief period, and then the current from the insulation (also referred to as the relaxation current) is followed and measured. When the discharge current is plotted against time (in seconds), different cable systems provide different relaxation curves and peaks as a function of time elapsed after shorting the system. From the nature of the curves and peaks, the claim is that the state of the cable can be determined and a risk factor provided. IRC plots for several cables subjected to various degrees of aging are shown in Figure E-1 [90, 105]. The shapes of the curves and locations of the peaks change as aging progresses.

Figure E-1 Isothermal Return Current Plots for Cross-Linked Polyethylene Cable Having Undergone Different Degrees of Aging

E-1

Off-Line Tests That Are Under Development

An advantage of this diagnostic method is that the important data are collected after the dc stress has been removed; a disadvantage is that the return currents are extremely small. A typical procedure applies 1000 Vdc for 30 minutes (to allow the current to saturate due to the dc conductivity), grounding occurs for ~20 seconds, and then measurements are made for the next 30 minutes while discharging is taking place. The use of relatively low dc voltage is claimed to be unable to inject charges into the system. This process induces a polarization within the insulation system. The depolarization current is considered to be due to trapped charge release as well as polymer chain motion. However, the currents that are measured are likely not only due to the insulation response to the applied dc, but also to the entire cable construction; for example, shield-insulation interfaces, polymer-filler interaction, impurities such as water (in addition to likely polar water treed regions), and cable length (joints further increase the complexity). The analysis of the data generated requires the use of computer-based network evaluation tools to estimate the aging status of the cable. Information about the cable construction and length can be fed into the database, but not the obviously more subtle information. Regardless, the interpretation is quite complex, and it has been noted that IRC can be considered to be a superposition of different components exhibiting different relaxation times.

E.3

Return Voltage

The return voltage (RV) test applies dc repetitively; after application of the first dc stress, the cable is discharged and then repeatedly recharged and discharged a total of four times. Key information sought includes 1) the peak value of the recovery voltage, 2) the time to reach peak voltage, and 4) the gradient of the recovery voltage curves. The information is fed into a computer program. The RV procedure involves charging for ~15 minutes followed by a two-second discharge; this is again followed by the charge and discharge sequence. The charging time can vary, and the discharge time can vary and be modified by relating it to the charging time. The dc voltage application starts low and is gently increased from 0.5V0 to 1.0V0, 1.5V0, and finally 2.0V0. The test time for one sample can take 1.5 hours. This charge and discharge process induces various polarization processes within the insulation. For example, oxidized regions of aged cables (which result from aging) would possess more dipoles and, therefore, greater potential for changes and then for relaxation after short circuiting. By measuring the return voltage as a function of applied dc voltage stress, it is claimed that a linear response time indicates less aging and a nonlinear response time indicates more aging. Clearly, one can expect different results depending on the degree of aging, charging time, discharging time, and local environment. The procedure measures the open circuit voltage that appears after poling, rather than the shortcircuit current as does IRC. The charging process is, therefore, longer and more repetitive than for the IRC measurement. Because the procedure causes charge to be stored and released, it is to be anticipated that, as cables age, longer times will be required for charges to be released (aged cables are more polar) and easier trapping of charges. Moisture, ionic contaminants, filler-polymer interfaces, and shield adhesion can all influence test results. It is desirable to perform a test concurrent with an unaged cable specimen of the same construction, vintage, and storage history. E-2

Off-Line Tests That Are Under Development

E.4

Oscillating Wave

The oscillating wave (OW) test provides a PD test methodology for locating discharge sites, using a damped oscillating wave as the voltage source. This represents an alternative test voltage waveform that does not require as large a power source as does a power frequency test set. The OW test set consists of a remotely controlled, high-voltage dc supply; a specially designed, solidstate switch; an air-core inductor/voltage divider/PD coupler; and an industrial computer as a control unit. Digital data acquisition and signal storage, as well as analysis and evaluation of the PD signals, take place in the control unit. The OW method [106] charges the test cable using the dc supply; the charging time is of the order of 20 seconds or less (so that potentially harmful space charges do not occur) [E.4]. After this time, service voltage is reached. At this point, the solid-state switch with fast closure time creates a series-resonant circuit between the cable and an air-core inductor. The circuit begins to oscillate, the latter being the resonant frequency with a delay time of 0.1 to 1 second. This produces a few of tens of cycles and, if PD is initiated, the circuit detects pulses that occur during the oscillating wave lifetime. The inductor and capacitance of the cable form a resonant circuit that produces a damped oscillatory sine wave. The frequency range is 50 Hz to 1 kHz. The rate of decay of the oscillation will depend on the resistance of the circuit (that is,, the dissipation factor of the insulation). Typically up to 50 shots are applied at a specified voltage level (it can be varied). From one trace, it is possible to determine the PDIV and PDEV, the pulse magnitude, and the dissipation factor. Although the test set is primarily designed to detect, measure, and locate PDs, it can also measure the dissipation factor; however, its sensitivity is not as high as dissipation factor–specific measuring equipment. A factor to be considered is that the information developed is related to cable length [107]. Details of the technology are described in the EPRI reports Advanced Diagnostics: Life Estimation of Extruded Dielectric Cables (1011499), Guide for Non-Destructive Diagnosis of Extruded Cable Systems (1001731), and especially Advanced Diagnostics and Life Estimation of Extruded Dielectric Cables (1013085) for nuclear issues [49, 90, 104].

E-3

F

INSULATION RESISTANCE TEST MEASUREMENTS: THEIR VALUE AND LIMITATIONS

F.1

Introduction

Very-low-frequency or ac withstand, tan δ, dielectric spectroscopy, and partial discharge (PD) testing are the preferred test methods for extruded polymer testing, depending on the potential problem being evaluated and the nature of the cable design. However, insulation resistance testing is often used to assess cables due to its ease of application and ready availability. Insulation resistance measurement is a valuable troubleshooting tool if used properly, and in some cases, it will be a good indicator of severe degradation of a cable’s insulating system. However, insulation resistance measurements must be used cautiously as a means of understanding age-related degradation. Some plants have adopted inappropriately low values of insulation resistance for acceptance criteria for return-to-service for medium-voltage cable. In addition, this acceptance criterion does not take circuit length into consideration. This section describes why higher values of insulation resistance are appropriate for cable insulation assessment and provides cautions regarding sole dependence on insulation resistance for assessment of medium-voltage cable. For nonshielded cable, insulation resistance is likely to be the only test that is practical to use. Consideration of the cautions provided in this section will improve the value of the test until an improved test is available.

F.2 Determining Minimum Insulation Resistance Value and Improving Interpretation of Results Minimum insulation resistance values exist for cable and motor circuits for the operation and application of elevated voltage tests. Often, insulation resistance testing acceptance criteria for cables have been based on acceptance criteria for rotating machinery contained in IEEE Std. 432000, “IEEE Recommended Practice for Testing Insulation Resistance of Rotating Machinery” [108]. When the original version of this standard was written in 1974, it contained the following formula for minimum acceptance value of insulation resistance for a winding:

IR1min  kV 1

Eq. F-1

Where: IR1min is the recommended minimum insulation resistance, in MΩ, at 104°F (40°C), of the entire three-phase winding. kV

is the rated machine terminal-to-terminal voltage, in rms kV.

F-1

Insulation Resistance Test Measurements: Their Value and Limitations

If individual phases were to be tested separately and guard circuits were used for the phases not under test, the minimum insulation resistance is three times this value. IEEE Std. 43-2000 contains an additional minimum insulation resistance value of 100 MΩ for form-wound windings built after 1970. For separately tested phases, the requirement is 300 MΩ. Although the minimum insulation resistance of Equation F-1 was meant to apply only to wet or deteriorated windings, the industry adopted it for general use and applied it to cables and the combination of motors tested with their field cable. Insulation resistance testing of the motor through the extruded polymer-insulated cable generally will not be a problem with respect to understanding the condition of the motor winding insulation system. As long as the cable has not degraded, the test will be dominated by the motor. However, a number of problems occur when the condition of the cable is evaluated either with the motor or by itself using the criteria of Equation F-1. The minimum insulation resistance of Equation F-1 is applicable to a varnished or asphalt winding that was wet to determine if it was safe to re-energize or to expose to an elevated voltage test. The leakage currents were not necessarily associated with a localized deterioration but rather a generalized leakage over the entire surface of insulation the winding. The 100 MΩ requirement for form-wound windings built after 1970 applies to insulation systems that have epoxy and/or mica tape insulation systems that have better barriers to surface moisture and, accordingly, much higher insulation resistance values, even when damp. Extruded mediumvoltage cable insulation systems do not behave like random-wound varnished or asphalt windings. Wetting the jacket of a shielded cable will have no immediate effect on insulation resistance. Although wetting the jacket of nonshielded cable might provide a return path for the test, it will not cause a low insulation resistance. The criteria of Equation F-1 from IEEE Std. 43 definitely do not apply to extruded polymer insulation. The 100–300 MΩ value for post-1970 form-wound motors comes closer to being a valid minimum. However, the following are also true: 

The insulation resistance of a cable is indirectly proportional to its length. Accordingly, minimum values must be corrected to a specific length, such as 1000 ft or 1 km. If the minimum insulation resistance is 100 MΩ-1000 ft (30.5 MΩ-km), and the circuit length is 250 ft (76 m), the minimum insulation resistance is 400 MΩ. Similarly, if the length is 2000 ft (61 m), the minimum insulation resistance is 50 MΩ.



The introduction to IEEE Std. 43-2000 states “Whereas individual insulation resistance measurements may be of questionable value, the carefully maintained record of periodic measurements, accumulated over months and years of service, is of unquestioned value as a measure of some aspects of the condition of the electrical insulation.” There are many variables associated with insulation resistance testing of motors, including temperature and humidity. The same is true for extruded polymer insulation of cables. Gross trending of the results over time is useful; however, trying to evaluate small changes (factors of 2 to 10 for high values) is inappropriate. General decreasing trends of decades are important.

F-2

Insulation Resistance Test Measurements: Their Value and Limitations



Insulation resistance testing is a global or bulk test. It cannot differentiate between an overall decrease in the insulation resistance in the wall of the polymer and one or two severe, localized, low insulation resistances. In addition, if a near-through-wall defect has a thin layer of good insulation between it and the shield or conductor, the insulation resistance test will not identify it.



Even when evaluating a failed cable, the insulation resistance can be tens of megohms or more because the fault is likely to have blown out the metal shield, causing long surface resistance paths between the exposed conductor and the remainder of metallic shield.

Insulation resistance measurement is a useful troubleshooting tool for evaluation of mediumvoltage cable. For extruded cable, values less than 100 MΩ-1000 ft (30.5 MΩ-km) are a strong indication that a problem exists with the circuit and investigation is necessary. However, the converse is not necessarily true. Reasonable confidence that the circuit is sound exists when values of gigohms or higher are present when adjusted for length, but the test will not necessarily detect voids and partial through-wall damage that could result in ultimate failure under ac operating conditions. Some improvement in value of insulation resistance results are possible for cable testing if the following are considered: 

The insulation resistances of each phase of a cable should be similar to those of the other phases. Decades of difference between phases indicates that a problem might exist with a phase or phases.



Periodic measurements taken over the course of years should be similar. Variations are expected due to differing conditions at the time of test, but the overall value should remain high and relatively consistent.



Gross decreasing trends or major decreasing step changes between periodic measurements are an indication of a problem and should be investigated further. A more sophisticated test is recommended for shielded cable, such as ac withstand, tan δ, dielectric spectroscopy, or PD, as applicable.



The insulation resistance measurement should smoothly increase or remain stable during the period of the test. Instability in the measurement, such as oscillations or continuing decreases in values, are indications of a breakdown phenomenon. Such oscillations might be present in a single-phase cable that has failed and the shield continuity has been lost in the vicinity of the fault. Returning such a cable to service has a high probability of resulting in a phase-tophase fault after a short period of operation.

F-3

Insulation Resistance Test Measurements: Their Value and Limitations

F.3

Insulation Resistance of Good Insulation

Having an insulation resistance of 100 to 300 MΩ-1000 ft (30 to 100 MΩ-km) does not indicate that the cable insulation is in good condition if it just exceeds these values; rather, it indicates that the cables can function and likely withstand operating or test voltage without immediate failure. Cable manufacturing standards, such as Insulated Cable Engineers Association specification S-68-516 [71], require minimum insulation resistances for new cable based on the following formula: IR = 10,000 log10 (D/d) MΩ-1000 ft

Eq. F-2

Where: D

is the diameter over the insulation.

d

is the diameter under the insulation.

To allow a comparison of the new cable requirement to the minimum acceptance criteria, assume a 500-ft (152-m) long cable with a 1/0 AWG (53.5 mm2) conductor and an 80-mil (2-mm) thick 600-V rated insulation. A representative cable of this type would have a 0.49 in. (12.4 mm) diameter. The field acceptance limit would be 1.6 MΩ. The new cable value would equal the following: IR = 10,000 log10 (0.49 in./0.33 in.) MΩ * 1000 ft/500 ft =3440 MΩ If the circuit was 2000 ft (610 m) long, the new cable value would be 860 MΩ. The length of the circuit is a significant consideration related to expected insulation resistance values. A newly installed cable should have a near-factory-value insulation resistance per 1000 ft (305 m), which, in this example, would be 6.8 GΩ. Using 100 to 300 MΩ-1000 ft (30 to 100 MΩ-km) as an acceptance criteria for the continued use of this cable example results in a value that is 23 to 68 times the allowable reduction from the original minimum insulation resistance. Most cables, when manufactured, have insulation resistance values that greatly exceed the minimum manufacturing requirements, making the difference between the asmanufactured to the degraded condition even larger.

F.4

Conclusions

Insulation resistance is a valuable troubleshooting tool. However, it must be used cautiously as a condition monitoring tool. The acceptance criterion of Equation F-1 (1 MΩ plus 1 MΩ per kV of rating) is not adequate for extruded polymer insulation used in medium-voltage cable. A minimum value of 100 MΩ-1000 ft (30 MΩ-km) is recommended to preclude severely degraded cables from being returned to service. Even when this value is used, higher values of insulation resistance might not be indicative of a sound insulation system. The cautions and techniques for improving the value of the results described in Section F.2, Determining Minimum Insulation Resistance Value and Improving Interpretation of Results, should be considered.

F-4

Export Control Restrictions

The Electric Power Research Institute, Inc. (EPRI, www.epri.com)

Access to and use of EPRI Intellectual Property is granted with the spe-

conducts research and development relating to the generation, delivery

cific understanding and requirement that responsibility for ensuring full

and use of electricity for the benefit of the public. An independent,

compliance with all applicable U.S. and foreign export laws and regu-

nonprofit organization, EPRI brings together its scientists and engineers

lations is being undertaken by you and your company. This includes

as well as experts from academia and industry to help address

an obligation to ensure that any individual receiving access hereunder

challenges in electricity, including reliability, efficiency, health, safety and

who is not a U.S. citizen or permanent U.S. resident is permitted access

the environment. EPRI also provides technology, policy and economic

under applicable U.S. and foreign export laws and regulations. In the

analyses to drive long-range research and development planning, and

event you are uncertain whether you or your company may lawfully

supports research in emerging technologies. EPRI’s members represent

obtain access to this EPRI Intellectual Property, you acknowledge that it

approximately 90 percent of the electricity generated and delivered in

is your obligation to consult with your company’s legal counsel to deter-

the United States, and international participation extends to more than

mine whether this access is lawful. Although EPRI may make available

30 countries. EPRI’s principal offices and laboratories are located in

on a case-by-case basis an informal assessment of the applicable U.S.

Palo Alto, Calif.; Charlotte, N.C.; Knoxville, Tenn.; and Lenox, Mass.

export classification for specific EPRI Intellectual Property, you and your company acknowledge that this assessment is solely for informational

Together...Shaping the Future of Electricity

purposes and not for reliance purposes. You and your company acknowledge that it is still the obligation of you and your company to make your own assessment of the applicable U.S. export classification and ensure compliance accordingly. You and your company understand and acknowledge your obligations to make a prompt report to EPRI and the appropriate authorities regarding any access to or use of EPRI Intellectual Property hereunder that may be in violation of applicable U.S. or foreign export laws or regulations.

Program: Generation Maintenance Applications Center

© 2012 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power Research Institute, EPRI, and TOGETHER...SHAPING THE FUTURE OF ELECTRICITY are registered service marks of the Electric Power Research Institute, Inc.

1024044

Electric Power Research Institute 3420 Hillview Avenue, Palo Alto, California 94304-1338 • PO Box 10412, Palo Alto, California 94303-0813 USA 800.313.3774  • 650.855.2121 • [email protected] • www.epri.com