EPRI Phosphat & Caustic Treatment Guidelines

Cycle Chemistry Guidelines for Fossil Plants: Phosphate Continuum and Caustic Treatment

SED R I A L

LICE

N

M AT E

WARNING: Please read the License Agreement on the back cover before removing the Wrapping Material.

Technical Report

Cycle Chemistry Guidelines for Fossil Plants: Phosphate Continuum and Caustic Treatment 1004188

Final Report, January 2004

EPRI Project Manager R. B. Dooley

EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • 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(S) 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. ORGANIZATION(S) THAT PREPARED THIS DOCUMENT EPRI

ORDERING INFORMATION Requests for copies of this report should be directed to EPRI Orders and Conferences, 1355 Willow Way, Suite 278, Concord, CA 94520, (800) 313-3774, press 2 or internally x5379, (925) 609-9169, (925) 609-1310 (fax). Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power Research Institute, Inc. Copyright © 2004 Electric Power Research Institute, Inc. All rights reserved.

CITATIONS This report was prepared by EPRI 3412 Hillview Avenue Palo Alto, California 94304 Authors R. B. Dooley K. Shields A. Aschoff M. Ball A. Bursik M. Gruszkiewicz This report describes research sponsored by EPRI. The report is a corporate document that should be cited in the literature in the following manner: Cycle Chemistry Guidelines for Fossil Plants: Phosphate Continuum and Caustic Treatment, EPRI, Palo Alto, CA: 2004. 1004188.

iii

PRODUCT DESCRIPTION

The purity of water and steam is central to ensuring fossil plant component availability and reliability. These guidelines for drum units provide information on the application of phosphate and caustic boiler water treatments. The guidelines will help operators reduce corrosion and deposition, and thereby achieve significant operation and maintenance cost reductions and greater unit availability. Results & Findings These guidelines include the following new features and control philosophies: •

There are two distinct solid alkali boiler water treatments: phosphate continuum (PC) and caustic treatment (CT). Examples are provided for the low and high levels of phosphate addition: PC(L) and PC(H).

•

The PC is bounded by Na:PO4 molar ratio of 3 and tri-sodium phosphate plus 1 ppm NaOH. It has a minimum phosphate level of 0.2 ppm and a minimum pH of 9 to reduce the possibility for boiler corrosion.

•

There are two distinctly different feedwater all-volatile treatments defined by the oxidation reduction potential (ORP). Feedwater systems having all-ferrous materials and using no reducing agent will operate with AVT(O). Those systems having mixed metallurgy (copper) materials and operating with a reducing agent and low oxygen (<10 ppb) will operate with AVT(R).

•

A separate set of target values and action levels to protect the steam turbine and boiler are included for both PC and CT.

•

Examples of specific and cation conductivity curves have been provided for PC and CT and linked to contaminant (chloride and sulfate) levels in the boiler water.

•

Correction curves and approaches have been developed for the effects of ammonia and carbon dioxide on boiler water pH.

Challenges & Objectives These guidelines have been derived to address the serious corrosion problems that have been experienced in drum units when operating with previous phosphate treatments (equilibrium phosphate treatment, EPT, and congruent phosphate treatment, CPT). The guidelines also provide direction for effective economical control of corrosion and deposition in drum units using solid alkali boiler water treatments (PC and CT).

v

Applications, Values & Use The chemistries in these guidelines can be applied to all drum units needing solid alkali treatments to counteract the effects of contaminants in the cycle. As such, the operator now has three choices for boiler water treatment. The road maps for selection in the guideline will provide the optimization required to ensure boiler corrosion does not occur. EPRI Perspective These PC and CT guidelines will help utilities achieve plant-specific goals in the areas of availability, reliability, and performance. This revision now becomes a part of EPRI’s suite of 10 key fossil plant guidelines that can be employed in every fossil plant. EPRI now has three guidelines for the five fossil plant boiler water treatments and three feedwater treatments – allvolatile treatment (1004187), phosphate continuum and caustic treatment (1004188), and oxygenated treatment (TR-102285). Other guidelines address the selection and optimization of boiler water and feedwater (TR-105040), controlling flow-accelerated corrosion (FAC) (TR108859), startup, shutdown, and layup (TR-107754), chemical cleaning (1003994), condensate polishing (TR-104422), makeup water treatment (TR-113692), and copper in fossil plants (1000457). In the near future, EPRI will revise the guidelines for oxygenated treatment and FAC based on the latest research results from the Boiler and Turbine Steam and Cycle Chemistry Program. Approach EPRI developed an initial skeleton of the PC and CT guidelines to include all pertinent research results. This was used as the basis for a meeting of the EPRI guidelines team. Following this meeting, the team developed a draft document, which was circulated to 75 members of EPRI’s Boiler and Turbine Steam and Cycle Chemistry Target for review and comment. Keywords Power Plant Availability Water Chemistry and Steam Boilers Turbines Corrosion Boiler Tube Failures Phosphate Treatment Caustic Treatment

vi

EPRI Licensed Material

ABSTRACT This Phosphate Continuum/Caustic Treatment (PC/CT) Guideline for Fossil Plants represents a major revision of EPRI’s previous phosphate guideline from 1994, and a combination with a previous EPRI report on the application of Caustic Treatment. This new guideline now represents the solid alkali treatments applicable to fossil plants. The approach developed for PC/CT follows that used in the 2002 revision of EPRI’s AVT Guideline, where the steam and boiler water limits have been decoupled to provide unique operating limits to protect the boiler and the turbine. The boiler water control limits for both PC and CT have been designed so they are below the values which are predicted to meet EPRI’s new steam limits. A phosphate continuum has been developed and is described in Section 1. Although it is a continuum, operators need to recognize that there are extremes at the low (L) and high (H) end of the phosphate range. These have been designated as PC(L) and PC(H), and examples of guideline limits for both have been provided in Section 4. However, to select and optimize the phosphate treatment for each boiler, a road map has been provided in Section 2, which requires, as a minimum, knowledge of the possible contaminants in the boiler, the carryover into the steam, and the level of sodium in steam. Actual derivation of each set of boiler water control curves, especially for cation and specific conductivity, will require reference to Section 3 where the rationale and derivation of each control parameter is extensively described. Caustic treatment has now been developed into full EPRI Guidelines structure in Section 5. A complete set of control curves, normal and action levels has been developed. As with PC, the limits for each parameter have been verified and derived in the rationale of Section 3. These new PC/CT Guidelines provide a complete set of solid alkali guidelines to address the serious corrosion problems currently in the fossil plant industry.

vii


ACKNOWLEDGMENTS The authors of these guidelines R. B. Dooley K. J. Shields A. Aschoff M. Ball A. Bursik M. Gruszkiewicz

EPRI EPRI Consultant Consultant Consultant Oak Ridge National Laboratory

acknowledge the assistance provided by D. Palmer from Oak Ridge National Laboratory in providing guidance on the volatility/partitioning of compounds between boiler water and steam. The draft guideline was reviewed by 75 members of the Boiler and Turbine Steam and Cycle Chemistry Target around the world. The authors particularly acknowledge the contributions from the following: K. Armstrong D. Aspden B. Buecker M. Caravaggio H. Cheng J. Mathews R. Pate D. Reynolds G. Verib

Atco Power ESKOM KCPL Ontario Power Generation CLP Power Hong Kong Ltd. Duke Energy Southern Company Dynegy First Energy

ix


CONTENTS

1

INTRODUCTION TO EPRI’S CYCLE CHEMISTRY PROGRAM, AND TO THE

PHOSPHATE CONTINUUM AND CAUSTIC TREATMENT.....................................................1-1 1.1 THE EPRI CYCLE CHEMISTRY PROGRAM ................................................................1-1 1.1.1

Program Goals and Objectives ...........................................................................1-3

1.1.2

Program Philosophy............................................................................................1-4

1.1.3

Key Cycle Chemistry Guidelines.........................................................................1-5

1.1.4

Program Vision and Future Plans .......................................................................1-6

1.2 RESEARCH SUPPORTING REVISION OF THE CYCLE CHEMISTRY GUIDELINES.........................................................................................................................1-7 1.2.1 (PTZ)

Chemical Environment and Liquid Films in the Phase Transition Zone 1-8

1.2.2

Corrosion Processes in the PTZ of Steam Turbines...........................................1-9

1.2.3 EPRI’s New Steam Limits.....................................................................................1-10 1.2.4

Volatility and Solubility of Impurities in Steam...................................................1-11

1.2.5

Copper Corrosion and Transport in Fossil Cycles ............................................1-13

1.3 FIELD EXPERIENCE SUPPORTING REVISION OF THE PC AND CT GUIDELINES, AND INTRODUCTION TO PC AND CT ......................................................1-14 1.3.1 Phosphate Treatments and Introduction of the Phosphate Continuum Leading to New PC Guidelines.......................................................................................1-14 1.3.2 Introduction to Caustic Treatment and Operating Experiences Leading to New CT Guideline ..........................................................................................................1-19 1.4 HOW TO USE THESE GUIDELINES ..........................................................................1-21 1.5 REFERENCES.............................................................................................................1-24

2

SELECTION AND OPTIMIZATION OF FEEDWATER AND BOILER WATER ................2-1

2.1 CHEMICALLY-INFLUENCED PROBLEMS, AND THE CONTINUUM OF TREATMENTS ......................................................................................................................2-1

xi


2.2 SELECTION AND OPTIMIZATION OF FEEDWATER TREATMENT ...........................2-5 2.2.1 Introduction and Types of Feedwater Treatment....................................................2-5 2.2.2 All-ferrous Feedwater Systems Optimization .........................................................2-7 Optimization of All-Ferrous Feedwater Chemistry .....................................................2-10 Step 1—Review Normal or Current Feedwater Treatment ...................................2-10 Step 2—Monitoring Baseline on Current Feedwater Treatment ...........................2-10 Step 3—Evaluate Reducing Agent Requirements ................................................2-12 Step 4—Monitoring with New Feedwater Treatment ............................................2-13 Steps 5 and 6—Consider Converting to OT..........................................................2-13 Step 7—Continue to Optimize the Feedwater Treatment .....................................2-13 Step 8—Operation and Continuing Monitoring .....................................................2-13 2.2.3 Mixed-Metallurgy Feedwater Systems Optimization ............................................2-14 Optimization of Mixed-Metallurgy Feedwater Chemistry ...........................................2-17 Step 1—Review of Water Chemistry, Operation, and Experience........................2-17 Step 2—Baseline Monitoring.................................................................................2-19 Step 3—Water Chemistry Optimization ................................................................2-20 Step 4—Design and Material Changes.................................................................2-20 Step 5—Operation ................................................................................................2-21 Step 6—Monitoring to Compare with Baseline Data.............................................2-21 Step 7—Normal Operation and Monitoring ...........................................................2-21 Step 8—Continual Check of Chemistry.................................................................2-22 Step 9—Longterm Plans .......................................................................................2-22 2.3 SELECTION AND OPTIMIZATION OF DRUM BOILER WATER TREATMENT .........2-22 Step 1—Review Normal or Current Treatment .....................................................2-24 Step 2—Continue Use of Current Treatment ........................................................2-26 Step 3—Base-Line Monitoring on Current Chemistry ...........................................2-26 Step 4—Initial Evaluation of Boiler Water Treatment Options ..............................2-27 Step 5—Consider Changing to PC(L), AVT or CT ................................................2-27 Step 6—Is There a Condensate Polisher in the Unit Cycle? ................................2-28 Step 7—Convert to AVT or OT .............................................................................2-28 Step 8—Boiler Chemical Clean.............................................................................2-28 Step 10—Need to Determine the Likelihood and Frequency of Cycle Contaminant Events..............................................................................................2-29 Steps 11, 13 and 15—Convert to PC(L), PC(H), or CT ........................................2-29 Steps 9, 12, 14, and 16—Develop Specific Unit Chemistry Guidelines ................2-30

xii


Steps 17 to 19—Monitor to Compare with Baseline Monitoring and to Optimize Treatment Selected................................................................................2-30 Step 20—Normal Operation..................................................................................2-31 2.4 REFERENCES.............................................................................................................2-32

3

PHILOSOPHY FOR GUIDELINE AND RATIONALE FOR SAMPLE POINTS,

ACTION LEVELS AND TARGET VALUES ..............................................................................3-1 3.1 INTRODUCTION............................................................................................................3-1 3.2 BACKGROUND DERIVATION PHILOSOPHY FOR PREVIOUS EPRI GUIDELINES.........................................................................................................................3-3 3.2.1 Illustration and Examples of Boiler Water Limits from Previous EPRI Guidelines.........................................................................................................................3-6 Total Carryover ............................................................................................................3-6 Illustration ....................................................................................................................3-8 3.3 BACKGROUND TO OVERALL DERIVATION OF PHILOSOPHY FOR PHOSPHATE CONTINUUM AND CAUSTIC TREATMENT GUIDELINES ........................3-11 3.3.1 Derivation of Philosophy for Steam and Boiler Water Limits in EPRI’s New Phosphate Continuum and Caustic Treatment Guidelines.............................................3-12 Phosphate Continuum at Low Phosphate Levels (PC(L)) (< 3 ppm) .........................3-15 Phosphate Continuum at High Phosphate Levels (> 3 ppm) (PC(H)) .......................3-20 Caustic Treatment (CT) .............................................................................................3-22 Summary Discussion of Derivation Process for Control Curves for PC and CT........3-27 3.4 RATIONALE FOR SAMPLE POINTS ..........................................................................3-28 3.4.1 Reheat Steam/Superheated Steam......................................................................3-29 3.4.2 Saturated Steam...................................................................................................3-30 3.4.3 Boiler Water..........................................................................................................3-31 3.4.4 Economizer Inlet and Attemperation Water..........................................................3-32 3.4.5 Deaerator Outlet ...................................................................................................3-32 3.4.6 Deaerator Inlet......................................................................................................3-32 3.4.7 Condensate Polisher Effluent (if Applicable) ........................................................3-33 3.4.8 Condensate Pump Discharge...............................................................................3-33 3.4.9 Condenser Leak Detection Trays and/or Hotwell Zones (if Applicable) ...............3-33 3.4.10 Air Removal System Exhaust .............................................................................3-33 3.4.11 Condensate Storage Tank Effluent ....................................................................3-34 3.4.12 Makeup Treatment System Effluent ...................................................................3-34

xiii


3.5 TROUBLESHOOTING, COMMISSIONING AND CORE PARAMETERS ...................3-34 3.6 RATIONALE FOR ACTION LEVELS ...........................................................................3-36 3.7 RATIONALE FOR TARGET VALUES..........................................................................3-37 3.7.1 Steam and Boiler Water .......................................................................................3-37 3.7.2 Feedwater and Condensate .................................................................................3-38 3.8 RATIONALE FOR TARGET VALUES OF INDIVIDUAL PARAMETERS.....................3-39 3.8.1 Sodium .................................................................................................................3-39 Sodium Target Value in Superheated (Reheat/Main) and Saturated Steam .............3-39 Sodium Target Value in Boiler Water.........................................................................3-40 Sodium Target Value in Condensate and Makeup Water..........................................3-45 Sodium Target Value in Condensate Storage Tank Effluent (Aluminum Tanks Only) ..........................................................................................................................3-45 3.8.2 Phosphate Target in Boiler Water ........................................................................3-45 3.8.3 Chloride ................................................................................................................3-47 Chloride Target Value in Steam.................................................................................3-47 Chloride Target Value in Boiler Water .......................................................................3-47 Chloride Target Value in Makeup Water....................................................................3-48 3.8.4 Sulfate ..................................................................................................................3-48 Sulfate Target Value in Steam ...................................................................................3-48 Sulfate Target Value in Boiler Water..........................................................................3-49 Sulfate Target Value in Makeup Water ......................................................................3-49 3.8.5 Silica .....................................................................................................................3-50 Silica Target Value in Steam .....................................................................................3-50 Silica Target Value in Boiler Water ............................................................................3-50 Silica Target Value in Condensate and Makeup Water .............................................3-50 3.8.6 Dissolved Oxygen.................................................................................................3-51 Dissolved Oxygen Target Value in Economizer Inlet.................................................3-51 Dissolved Oxygen Target Value in Drum Boiler Water ..............................................3-51 Dissolved Oxygen Target Value at Deaerator Outlet.................................................3-51 Dissolved Oxygen Target Value at Condensate Pump Discharge ............................3-51 3.8.7 Oxidizing-Reducing Potential (ORP) ....................................................................3-52 3.8.8 Reducing Agents (Hydrazine or Alternates) .........................................................3-52 3.8.9 pH ..........................................................................................................................3-53 pH Control..................................................................................................................3-53 pH Target Values in Boiler Water ..............................................................................3-53

xiv


pH Target Values in Feedwater .................................................................................3-55 3.8.10 Ammonia ............................................................................................................3-55 Ammonia Target Values at the Economizer Inlet.......................................................3-56 Monitoring Ammonia in Boiler Water .........................................................................3-56 3.8.11 Specific Conductivity ..........................................................................................3-56 Specific Conductivity Target Values at Economizer Inlet...........................................3-57 Specific Conductivity in Boiler Water .........................................................................3-57 Specific Conductivity Target Values in Makeup Water ..............................................3-62 3.8.12 Cation Conductivity.............................................................................................3-62 Cation Conductivity Target Values at Economizer Inlet.............................................3-62 Cation Conductivity in Boiler Water ...........................................................................3-62 Cation Conductivity Limit in Steam ............................................................................3-70 3.8.13 Total Organic Carbon (TOC) ..............................................................................3-70 Steam, Condensate, and Makeup TOC Target Values .............................................3-70 3.8.14 Iron and Copper..................................................................................................3-70 Total Iron and Copper Limits......................................................................................3-71 Monitoring Considerations .........................................................................................3-71 3.8.15 Air In-leakage .....................................................................................................3-73 Air In-leakage Limit ....................................................................................................3-75 3.9 GUIDELINES CUSTOMIZATION AND OPTIMIZATION..............................................3-75 3.10 REFERENCES...........................................................................................................3-76

4

CYCLES WITH DRUM BOILERS OPERATING WITH THE PHOSPHATE

CONTINUUM (PC).....................................................................................................................4-1 4.1 INTRODUCTION TO PC................................................................................................4-1 4.2 GUIDANCE AND GUIDELINES FOR PC(L) ..................................................................4-4 4.2.1 Introduction to PC(L) ..............................................................................................4-4 4.2.2 Target Values for PC(L)........................................................................................4-10 4.2.3 Normal Operation for PC(L)..................................................................................4-23 Continual Check of pH and Phosphate Level ............................................................4-24 4.2.4 Operation During Upset Conditions......................................................................4-24 4.2.5 Unit Startup and Shutdown...................................................................................4-26 Cold Startup ...............................................................................................................4-26

xv


Unit Trip with Normal Contaminant Concentrations...................................................4-26 Unit Shutdown ...........................................................................................................4-26 4.3 GUIDANCE AND GUIDELINES FOR PC(H) ...............................................................4-26 4.3.1 Target Values for PC(H) .......................................................................................4-27 4.3.2 Normal Operation for PC(H) .................................................................................4-43 Phosphate Hideout and Its Control ............................................................................4-43 Reaction to Contaminants in Boiler Cycle..................................................................4-44 Startup and Shutdown Operation...............................................................................4-45 Cold Start ..............................................................................................................4-45 Transition to Normal Operation .............................................................................4-45 Hot Startup and the Impact of Hideout.......................................................................4-46 Feedwater and Makeup Control.................................................................................4-46 Condenser Leakage ..................................................................................................4-46 Chemical Feeds—Mixed-Metallurgy Cycles ..............................................................4-46 Chemical Feeds—All-Ferrous Systems .....................................................................4-48 Monitoring and Corrective Actions .............................................................................4-48 4.4 REFERENCES.............................................................................................................4-49

5

CYCLES WITH DRUM BOILERS ON CAUSTIC TREATMENT .......................................5-1

5.1 CAUSTIC TREATMENT GUIDANCE.............................................................................5-1 5.2 GUIDANCE AND GUIDELINES FOR CAUSTIC TREATMENT (CT) ............................5-2 5.2.1 Target Values for CT ..............................................................................................5-2 5.2.2 Target Values for Plants Without Reheat .............................................................5-18 5.3 NORMAL OPERATION FOR DRUM BOILERS ON CAUSTIC TREATMENT.............5-18 5.3.1 Cycle Makeup.......................................................................................................5-18 5.3.2 Condenser Leakage .............................................................................................5-18 5.3.3 Chemical Feeds – Feedwater – Mixed-Metallurgy Cycles ...................................5-18 5.3.4 Chemical Feeds – Feedwater - All-Ferrous Systems ...........................................5-19 5.3.5 Chemical Feeds – Boiler Water............................................................................5-19 5.3.6 Monitoring and Corrective Actions........................................................................5-19 5.4 APPROACHES TO DEVELOPING OPTIMUM CAUSTIC TREATMENT ....................5-20 5.5 REFFERENCES...........................................................................................................5-24

xvi


6

PC AND CT CHEMISTRY CONTROL AND CORRECTIVE ACTIONS ............................6-1

6.1 CHEMISTRY CONTROL................................................................................................6-1 6.2 CORRECTIVE ACTIONS...............................................................................................6-3 6.3 INFLUENCE OF TREATMENT CHEMICAL PURITY ON CYCLE CHEMISTRY.........6-25

A

OXYGEN REMOVAL TECHNIQUES FOR TREATED MAKEUP .................................. A-1

A.1 VACUUM DEAERATION .............................................................................................. A-1 A.2 CATALYTIC OXYGEN REMOVAL ............................................................................... A-3 A.3 MEMBRANE SYSTEMS ............................................................................................... A-4 A.4 MAKEUP WATER STORAGE ...................................................................................... A-7 A.5 References.................................................................................................................. A-10

B

OXIDATION-REDUCTION POTENTIAL (ORP) ............................................................. B-1

B.1 INTRODUCTION .......................................................................................................... B-1 B.2 OVERVIEW OF ELECTOCHEMICAL PHENOMENA INFLUENCING CORROSION PROCESSES IN FOSSIL PLANT UNITS ..................................................... B-2 B.3 OXIDATION-REDUCTION POTENTIAL (ORP) OR REDOX POTENTIAL .................. B-5 B.4 ELECTROCHEMICAL POTENTIAL (ECP)................................................................. B-11 B.5 REFERENCE ELECTRODE CONSIDERATIONS ..................................................... B-13 B.6 MONITORING ORP IN FOSSIL UNITS...................................................................... B-15 Factors Influencing Fossil Plant ORP Readings............................................................ B-16 Interpretation of Fossil Plant ORP Data ........................................................................ B-19 Reference Electrode Calibration and Maintenance Procedures.................................... B-20 Corrosion Potential Probe ............................................................................................. B-21 ORP Probe .................................................................................................................... B-22 B.7 CLOSING COMMENTS ON ORP MONITORING IN FOSSIL PLANTS..................... B-24 B.8 REFERENCES ........................................................................................................... B-25

xvii


C

AIR IN-LEAKAGE MONITORING AND CONTROL ....................................................... C-1

C.1 CYCLE AIR IN-LEAKAGE ............................................................................................ C-1 C.2 ROTAMETERS............................................................................................................. C-1 C.3 MULTISENSOR PROBE .............................................................................................. C-2 C.4 HELIUM AND SULFUR HEXAFLOURIDE METHODS ................................................ C-6 C.5 UTILITY AIR IN-LEAKAGE PROGRAM ....................................................................... C-9 C.6 REFERENCES ........................................................................................................... C-12

D

BENCHMARKING A UTILITY’S CHEMISTRY ORGANIZATION .................................. D-1

D.1 INTRODUCTION .......................................................................................................... D-1 D.2 ASSESSING THE CYCLE CHEMISTRY ORGANIZATION OF A UTILITY ................. D-1

E

CYCLE CHEMISTRY DATA QUALITY ........................................................................... E-1

E.1 INTRODUCTION .......................................................................................................... E-1 E.2 SAMPLING, SAMPLE CONDITIONING AND INSTRUMENTATION ........................... E-2 E.2.1 Factors Affecting Sampling ................................................................................... E-2 E.2.2 Potential Problems in the Design and Operation of Sampling Systems................ E-3 E.2.3 Sample Collection ................................................................................................. E-4 E.2.4 Sample Tubing ...................................................................................................... E-7 Pumps, Valves, and Fittings ....................................................................................... E-7 Sample Water Recovery and Drains........................................................................... E-7 E.2.5 Sample Conditioning ............................................................................................. E-7 Temperature Regulation ............................................................................................. E-8 Flow Rate and Pressure Regulation ........................................................................... E-8 Sample Filters ........................................................................................................... E-10 Sample Panels.......................................................................................................... E-10 E.2.6 Grab Sampling Procedures................................................................................. E-13 Grab Samples ........................................................................................................... E-13 Sample Containers ................................................................................................... E-13

xviii


Sample Collection and Preservation......................................................................... E-14 E.2.7 Corrosion Product Sampling ............................................................................... E-14 E.3 ON-LINE MONITORING TECHNIQUES .................................................................... E-14 E.3.1 Important Considerations for Selection of Proper Monitoring Method ................ E-15 E.3.2 Analyzer Calibration Techniques ........................................................................ E-16 E.3.3 Analyzer Operation and Maintenance................................................................. E-16 E.4 DATA COLLECTION, INTERPRETATION, AND MANAGEMENT............................. E-16 E.4.1 Data Collection.................................................................................................... E-17 E.4.2 Automatic Data Collection and Storage .............................................................. E-17 E.4.3 Manual Storage of Chemistry Data ..................................................................... E-18 E.4.4 Data Analysis and Interpretation ......................................................................... E-18 E.4.5 Roles of Plant Personnel in Water Chemistry Data Collection, Interpretation and Management........................................................................................................... E-18 E.4.6 Expert Systems for Water Chemistry Management ............................................ E-19 E.5 VALIDATION OF CHEMISTRY DATA........................................................................ E-20 E.5.1 Precision, Accuracy, Bias and Drift ..................................................................... E-23 E.5.2 QA/QC for Sampling Systems............................................................................. E-27 E.5.3 QA/QC for On-Line Instruments.......................................................................... E-27 Instrument Calibration and Standardization.............................................................. E-27 Validation of On-Line Monitoring Instruments........................................................... E-28 Synthesizing Standard Samples ............................................................................... E-29 Potential Problems in Preparation of Standards ....................................................... E-30 E.5.4 Charting of Chemistry Data to Track Instrument Performance ........................... E-30 E.5.5 Interlaboratory Assessment of Grab Sample Analysis Methods ......................... E-33 E.5.6 EPRI Experience With Chemistry Data Validation .............................................. E-34 Performance Determination ...................................................................................... E-34 Acceptance Limits..................................................................................................... E-35 Use of Acceptance Limits in QC ............................................................................... E-35 E.6 SUMMARY.................................................................................................................. E-37 E.7 REFERENCES ........................................................................................................... E-39

xix


F

UNIT SHUTDOWN, LAYUP, STARTUP, CYCLING AND PEAKING...............................F-1

F.1 SHUTDOWN ..................................................................................................................F-2 F.2 LAYUP ...........................................................................................................................F-4 F.2.1 Influence of Cycle Metallurgy on Layup .................................................................F-5 All-Ferrous Metallurgy Systems ...................................................................................F-5 Mixed metallurgy systems............................................................................................F-5 F.2.2 Short-term Layup....................................................................................................F-5 F.2.3 Intermediate and Longterm Layup .........................................................................F-6 F.2.4 Maintenance Outage ..............................................................................................F-6 F.3 STARTUP ......................................................................................................................F-6 F.3.1 Corrosion Product Filters........................................................................................F-7 F.3.2 Units With Drum Boilers .........................................................................................F-8 F.4 ROAD MAP FOR SHUTDOWN AND LAYUP..............................................................F-13 F.5 CYCLING AND PEAKING............................................................................................F-17 F.6 REFERENCES.............................................................................................................F-20

G

CALCULATION OF THE pH CORRECTED FOR THE PRESENCE OF

AMMONIA ................................................................................................................................ G-1 G.1 REFERENCES ............................................................................................................. G-7

H

EPRI FOSSIL PLANT CYCLE CHEMISTRY REPORTS, GUIDELINES AND

CONFERENCE PROCEEDINGS ............................................................................................. H-1 CONFERENCE PROCEEDINGS......................................................................................... H-6

xx


LIST OF FIGURES Figure 1-1 Overall Philosophy of EPRI’s Cycle Chemistry Program. CCIP is Cycle Chemistry Improvement Program, BTFR is Boiler Tube Failure Reduction, FAC is Flow-accelerated Corrosion, and CP is Condensate Polishing..........................................1-5 Figure 1-2 Schematic of the Processes in PTZ. The extra arrows pointing in to the “Deposits” indicate that deposition can occur by other methods........................................1-8 Figure 1-3 Model for the Development of Corrosion Processes in the PTZ of Steam Turbines ...........................................................................................................................1-10 Figure 1-4 Partitioning Constants for Common Salts, Oxides and Acids. KD is the partitioning constant. ........................................................................................................1-12 Figure 1-5 pH Dependence of the Cu Release Rates at 95°C (203°F) for Al-Brass, 90Cu10Ni and Admiralty Brass After Treatment in Non-Degassed Feedwater at ORP * OX = +100 mV and –300 mV* Oxidizing, RED = Reducing 1-13 Figure 1-6 Schematic of Operating Ranges of Boiler Water on Equilibrium Phosphate Treatment (EPT), Congruent Phosphate Treatment (CPT) and Phosphate Treatment (PT). It should be noted that these are the old phosphate operating ranges, and are shown here for reference only. They shouldn’t be applied to fossil plants................................................................................................................................1-15 Figure 1-7 Schematic of Operating Range for PC. a) is normal scale, b) is log scale. ............1-17 Figure 2-1 Summary of Possible Boiler Water and Feedwater Treatments for Fossil (1) Plants as a Function of Equipment and Equipment Capability . This is the “Continuum of Treatments”. ...............................................................................................2-5 Figure 2-2 Schematic Representation of Oxide Formed on Ferrous Feedwater Surfaces During Operation with Reducing AVT ................................................................................2-7 Figure 2-3 Change in Oxidizing-Reducing Potential (ORP) and Feedwater Iron Levels (Fe) at the Economizer Inlet when Hydrazine (N2H4) is Gradually Reduced on a (10) 600MW Drum Unit with an All-Ferrous Feedwater System .............................................2-8 Figure 2-4 Schematic Representation of Oxide Formed on Iron-Based Feedwater Surfaces During Operation with Oxidizing AVT and OT.....................................................2-9 Figure 2-5 Road Map for Optimizing Feedwater Treatment for All-Ferrous Feedwater Systems ...........................................................................................................................2-11 Figure 2-6 Surface Images of Admiralty Brass After Treatment at 95°C (203°F), pH 9 and ORP = +100 mV (a,c,e,g) and ORP = -350 mV (b,d,f,h) for 8 (a,b), 24 (c,d), 100 (e,f) and 200 (g,h) hours. The surface oxide under reducing conditions is protective (12) Cu2O. The rougher surface oxide under oxidizing conditions is CuO. ..........................2-16 Figure 2-7 Road Map for Optimizing Feedwater Treatment for Mixed-Metallurgy Systems ....2-18

xxi


Figure 2-8 Road Map for Selecting or Optimizing Boiler Water Treatment for Drum Boilers ..............................................................................................................................2-23 Figure 3-1 Typical Locations of Impurity Ingress, Corrosion and Deposition in a Drum Cycle ..................................................................................................................................3-2 Figure 3-2 Representative Drum Boiler Mechanical Carryover used to Derive Generic Cycle Chemistry Limits; Actual Carryover in Drum Boilers Must be Determined in the Field as There are Significant Variations Due to Drum Design, Firing Rate and Drum Operating Level. .......................................................................................................3-4 Figure 3-3 Distribution Ratios for Common Boiler Water Contaminants (This ray diagram (7,8) was originally developed by N. A. Styrikovich and O. I. Martynova. ) .............................3-5 Figure 3-4 Equilibrium Phosphate Treatment (EPT): Drum Boiler Water Chloride vs. Operating Pressure. Calculated using mechanical and vaporous carryover for a limit of 3 ppb Chloride in Steam. The values were identical for sulfate. It must be noted that these are the old EPT limits derived in the previous Phosphate (1) Treatment Guidelines and should no longer be used. The new PC(L) limits are discussed in Section 4. ......................................................................................................3-9 Figure 3-5 Phosphate Treatment (PT): Drum Boiler Water Chloride vs. Operating Pressure. Calculated using mechanical and vaporous carryover for a limit of 3 ppb Chloride in Steam. It must be noted that these are the old limits derived in the (1) previous Phosphate Treatment Guidelines and should no longer be used. The new PC(H) limits are discussed in Section 4. ..................................................................3-10 Figure 3-6 Partitioning Constants - Neutral Species ................................................................3-13 Figure 3-7 Partitioning Constants - 1:1 Compounds................................................................3-14 Figure 3-8 Partitioning Constants - 1:2 Compounds................................................................3-15 Figure 3-9 Model Predictions for the maximum pressure-dependent concentrations of Sodium in the boiler water. The bottom set of three curves is for conditions typical of PC(L) to ensure less than 2 ppb sodium in steam. The upper set of three curves is for conditions typical of PC(H) to ensure less than 3 ppb sodium in steam. In each set, the top straight line is for mechanical carryover only. The two lower curves are at Na:PO4 = 3 (lower) and TSP + 1 ppm NaOH (higher). It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the control curves in the Guidelines (Section 4). .....................................................................................................................................3-16 Figure 3-10 Model Predictions for the maximum pressure-dependent concentrations of Chloride in the boiler water. The curves are representative of conditions for PC(L) to ensure less than 2 ppb chloride in steam. The top line (1) is for consideration of mechanical carryover only. Line 3 is for Na:PO4 = 3 and includes vaporous carryover. Lines 2 and 4 are for TSP + 1 ppm NaOH and also include vaporous carryover. Line 4 represents the condition where the chloride is limited by the sodium limit. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the control curves in the Guidelines (Section 4)................................................................................................3-17 Figure 3-11 Model Predictions for the maximum pressure-dependent concentrations of Sulfate in the boiler water. The bottom set of three curves is for conditions typical of PC(L) to ensure less than 2 ppb sulfate in steam. The upper set of curves is for conditions typical of PC(H) to ensure less than 3 ppb sulfate in steam. In each set, the top line is for mechanical carryover only. The two lower curves are at Na:PO4 =

xxii


3 (lower) and TSP+ 1 ppm NaOH (higher) and include vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the control curves in the Guidelines (Section 4). .....................................................................................................................................3-18 Figure 3-12 Model Predictions for the maximum pressure-dependent concentrations of Silica in the boiler water to Ensure Less than 10 ppb Silica (as SiO2) in Steam. The curves are representative of PC(L), PC(H) and CT. Curve 1 is for mechanical carryover only. Curve 2 is for mechanical and vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the control curves in the Guidelines (Section 4). .....................................................................................................................................3-19 Figure 3-13 Model Predictions for Boiler Water Chloride as a Function of Pressure. The curves are representative of conditions for PC(H) to ensure less than 3 ppb chloride in steam. The top line (1) is for consideration of mechanical carryover only. Line 3 is for Na:PO4 = 3 and includes vaporous carryover. Lines 2 and 4 are for TSP + 1 ppm NaOH and also include vaporous carryover. Line 4 represents the condition where chloride is limited by the sodium limit, but does include vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the curves in the Guidelines (Section 4)........................................................................................................................3-21 Figure 3-14 Model Predictions for Boiler Water Sodium as a Function of Pressure. The curves are representative of conditions for CT to ensure less than 2 ppb sodium in steam. Curve 1 is for mechanical carryover only. Curve 2 is for mechanical and vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the curves in the Guidelines (Section 5)................................................................................................3-24 Figure 3-15 Model Predictions for Boiler Water Chloride as a Function of Pressure. The curves are representative of the conditions for CT to ensure less than 2 ppb chloride in steam. Curve 1 is for mechanical carryover only. Curves 2 are for mechanical and vaporous carryover with various levels of NaOH (0.5, 1.0 and 2 ppm). Curves 3 are for mechanical and vaporous carryover for AVT conditions at different ammonia levels (pH 9, 9.2, 9.4 and 9.6). It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the curves in the Guidelines (Section 5). ....................................................3-25 Figure 3-16 Model Predictions for Boiler Water Sulfate as a Function of Pressure. The curves are representative of conditions for CT to ensure less than 2 ppb sulfate in steam. Curve 1 is for mechanical carryover only. Curve 2 is for mechanical and vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the curves in the Guidelines (Section 5)................................................................................................3-26 Figure 3-17 Model Predictions for Boiler Water Sodium as a Function of Pressure when Operating with PC(H) to ensure less than 3 ppb sodium in steam at Na:PO4 = 3, and also the contribution from 1 ppm NaOH, if added. The top curve is the same as shown in Figure 3-9. Note: these curves should not be used for boiler water control; they are provided to illustrate the effects of sodium in boiler water. The control curves for PC(H) are provided in Section 4......................................................................3-42 Figure 3-18 Model Predictions for Boiler Water Sodium as a Function of Pressure when Operating with PC(L) to ensure less than 2 ppb sodium in steam at Na:PO4 = 3, and

xxiii


also the contribution from 1 ppm NaOH, if added. The top curve is the same as shown in Figure 3-9. Note: these curves should not be used for boiler water control; they are provided to illustrate the effects of sodium in boiler water. The control curves for PC(L) are provided in Section 4. .....................................................................3-43 Figure 3-19 Specific Conductivity as a Function of Phosphate Level with Varying Na:PO4 ratios and TSP + 1 ppm NaOH. No influence of ammonia is included. It should also be noted that the curve for H3PO4 crresponds to cation conductivity. ..............................3-58 Figure 3-20 Estimates of Specific Conductivity Against Pressure for PC(L). The lower curve reflects the conductivity if only phosphate is present in the boiler water. It is based on the phosphate concentration corresponding to the sodium concentration given in Figure 3-18 to ensure less than 2 ppb sodium in steam. The upper curve reflects the situation with additional chloride and sulfate in the boiler water. Note: these curves must not be used for boiler water control; they are provided here to assist the reader to develop a set of specific conductivity control curves in Section 4 for PC(L)...........................................................................................................................3-59 Figure 3-21 Estimate of Specific Conductivity Against Pressure for PC(H). The lower curve reflects the conductivity if only phosphate is present in the boiler water. It is based on the phosphate concentration corresponding to the sodium concentration given in Figure 3-17 to ensure less than 3 ppb sodium in steam. The upper curve reflects the situation with additional chloride and sulfate in the boiler water. Note: these curves must not be used for boiler water control; they are provided here to assist the reader to develop a set of specific conductivity control curves in Section 4 for PC(H). .........................................................................................................................3-60 Figure 3-22 Estimates of Cation Conductivity Against Pressure for PC(L). The lower curve reflects the conductivity if only phosphate is present in the boiler water. It is based on the phosphate concentration corresponding to the sodium concentration given in Figure 3-18 to ensure less than 2 ppb sodium in steam. The upper curve reflects the situation with additional chloride and sulfate present in the boiler water. Note: these curves must not be used for boiler control; they are provided here to assist the reader to develop a set of cation conductivity control curves in Section 4 for PC(L)...........................................................................................................................3-66 Figure 3-23 Estimates of Cation Conductivity Against Pressure for PC(H). The lower curve reflects the conductivity if only phosphate is present in the boiler water. It is based on the phosphate concentration corresponding to the sodium concentration given in Figure 3-17 to ensure less than 3 ppb sodium in steam. The upper curve reflects the situation with additional chloride and sulfate present in the boiler water. Note: these curves must not be used for boiler control; they are only provided here to assist the reader to develop a set of cation conductivity control curves in Section 4 for PC(H). ......................................................................................................................3-67 Figure 3-24 Cation Conductivity Versus Pressure for PC(L) only taking into account the maximum (normal) levels of chloride and sulfate (1:2 ratio). It must be noted that any contribution from phosphate has been excluded.......................................................3-68 Figure 3-25 Cation Conductivity Versus Pressure for PC(H) only taking into account the maximum (normal) levels of chloride and sulfate (1:2 ratio). It must be noted that any contribution from phosphate has been excluded.......................................................3-69 Figure 3-26 Relationship between ammonia concentration and pH for various carbon dioxide concentrations in feedwater. The lowest curve represents zero CO2. The top curve represents 4 ppm CO2. ...........................................................................................3-74

xxiv


Figure 4-1 Schematic of Operating Range for PC (a is normal scale, b is log scale) ................4-2 Figure 4-2 PC(L): Boiler Water Phosphate vs. Operating Pressure ..........................................4-5 Figure 4-3 Boiler Water pH vs. Phosphate at Various Ammonia Levels. Na/PO4 = 3. The lowest curve is at zero ammonia, the highest curve is at 2 ppm ammonia. .......................4-6 Figure 4-4 Boiler Water pH vs. Phosphate at Various Ammonia Levels. TSP + 1 ppm NaOH. The lowest curve is at zero ammonia. The highest curve is at 2 ppm ammonia. ...........................................................................................................................4-7 Figure 4-5 Relationship Between Corrected Boiler Water pH, Measured Boiler Water pH and Measured Ammonia Concentration. The lines on the figure increase in 0.1 ppm ammonia increments from 0 to 2 ppm. To determine the corrected pH, find the measured pH on the x-axis and the point at which a vertical line intersects the measured ammonia concentration. This point, extended horizontally to the y-axis, indicates the corresponding corrected pH value. ...............................................................4-9 Figure 4-6 Cycle Chemistry Diagram for PC(L) with AVT(O) feedwater. All-ferrous metallurgy in feedwater system. Operating with an oxidizing environment (no reducing agent). ...............................................................................................................4-11 Figure 4-7 Cycle Chemistry Diagram for PC(L) with AVT(R) feedwater. All-ferrous and mixed-metallurgy feedwater systems. Operating with a reducing environment (reducing agent added). ...................................................................................................4-12 Figure 4-8 PC(L): Boiler Water Sodium vs. Operating Pressure .............................................4-15 Figure 4-9 PC(L): Boiler Water pH vs. Operating Pressure. pH due to trisodium phosphate (3.8 to 1.0 ppm phosphate from Figure 4-2)...................................................4-16 Figure 4-10 PC(L): Boiler Water Chloride vs. Operating Pressure ..........................................4-17 Figure 4-11 PC(L): Boiler Water Sulfate vs. Operating Pressure ............................................4-18 Figure 4-12 PC(L): Boiler Water Silica vs. Operating Pressure ...............................................4-19 Figure 4-13 PC(L): Drum Boiler Water Specific Conductivity vs. Operating Pressure showing specific conductivity due to phosphate alone, and due to phosphate plus various amounts of chloride and sulfate. It must be noted that this figure is only an example to illustrate the effects of phosphate, chloride and sulfate. Each operator will need to develop unique curves for each boiler as described in the text of Sections 4.2.2 and 3.8.11. ...............................................................................................4-21 Figure 4-14 PC(L): Drum Boiler Water Cation Conductivity vs. Operating Pressure showing cation conductivity due to phosphate alone, and due to phosphate plus various amounts of chloride and sulfate. It must be noted that this figure is only an example to illustrate the effects of phosphate, chloride and sulfate. Each operator will need to develop unique curves for each boiler as described in the text of Sections 4.2.2 and 3.8.12. ...............................................................................................4-22 Figure 4-15 Cycle Chemistry Diagram for PC(H) with AVT(O) Feedwater. All-ferrous metallurgy in feedwater system. Operating with an oxidizing environment (no reducing agent). ...............................................................................................................4-29 Figure 4-16 Cycle Chemistry Diagram for PC(H) with AVT(R) Feedwater. All-ferrous and mixed-metallurgy feedwater systems. Operating with a reducing environment (reducing agent added). ...................................................................................................4-30 Figure 4-17 PC(H): Boiler Water Sodium vs. Operating Pressure ...........................................4-32 Figure 4-18 PC(H): Boiler Water Phosphate vs. Operating Pressure ......................................4-33

xxv


Figure 4-19 PC(H): Boiler Water pH vs. Operating Pressure. pH due to trisodium phosphate (5.8 to 1.4 ppm from Figure 4-18). .................................................................4-34 Figure 4-20 PC(H): Boiler Water Chloride vs. Operating Pressure..........................................4-35 Figure 4-21 PC(H): Boiler Water Sulfate vs. Operating Pressure ............................................4-36 Figure 4-22 PC(H): Boiler Water Silica vs. Operating Pressure ..............................................4-37 Figure 4-23 PC(H): Drum Boiler Water Specific Conductivity vs. Operating Pressure showing specific conductivity due to phosphate alone, and due to phosphate plus various amounts of chloride and sulfate. It must be noted that this figure is only an example to illustrate the effects of phosphate, chloride and sulfate. Each operator will need to develop unique curves for each boiler as described in the text of Sections 4.3.1 and 3.8.11. ...............................................................................................4-40 Figure 4-24 PC(H): Drum Boiler Water Cation Conductivity vs. Operating Pressure. Showing cation conductivity due to phosphate alone and due to phosphate plus various amounts of chloride and sulfate. It must be noted that this figure is only an example to illustrate the effects of phosphate, chloride and sulfate. Each operator will need to develop unique curves for each boiler as described in the text of Sections 4.3.1 and 3.8.12. ...............................................................................................4-41 Figure 5-1 Cycle Chemistry Diagram for a Drum Unit on Caustic Treatment. All-ferrous and mixed-metallurgy feedwater systems. Operating with a reducing environment (reducing agent added) in the feedwater, AVT(R). ............................................................5-3 Figure 5-2 Cycle Chemistry Diagram for a Drum Unit on Caustic Treatment. All-ferrous feedwater systems. Operating with an oxidizing environment (no reducing agent) in the feedwater, AVT(O). ......................................................................................................5-4 Figure 5-3 Caustic Treatment: Boiler Water pH vs. Operating Pressure. The dashed line indicates the lower limit for pH to ensure that the pH of the boiler water is at least a minimum value to reduce the risk of boiler corrosion. ........................................................5-6 Figure 5-4 Caustic Treatment: Boiler Water Cation Conductivity vs. Operating Pressure.........5-7 Figure 5-5 Caustic Treatment: Boiler Water Specific Conductivity vs. Operating Pressure. The dashed line indicates the lower limit for specific conductivity corresponding to the minimum pH and NaOH concentration shown in Figures 5-3 and 5-6. ..............................................................................................................................5-8 Figure 5-6 Caustic Treatment: Boiler Water Sodium Hydroxide vs. Operating Pressure. The dashed line indicates the lower limit for NaOH in the boiler water. .............................5-9 Figure 5-7 Caustic Treatment: Boiler Water Sodium vs. Operating Pressure..........................5-10 Figure 5-8 Caustic Treatment: HRSG Boiler Water Chloride vs. Operating Pressure .............5-11 Figure 5-9 Caustic Treatment: Boiler Water Sulfate vs. Operating Pressure ..........................5-12 Figure 5-10 Caustic Treatment: Drum Boiler Water Silica vs. Operating Pressure .................5-13 Figure 5-11 Relationship Between Specific Cation Conductivity, Cation Conductivity and pH at 25°C for Boiler Water with Caustic Treatment ........................................................5-16 Figure 5-12 Caustic Treatment: Drum Boiler Water Relationship Between pH and Equivalent Sodium Hydroxide Concentration at 25°C, Showing Effect of Ammonia Concentrations. The lowest curve is at zero ammonia. The highest curve is at 2 ppm ammonia. .................................................................................................................5-17 Figure 5-13 Road Map for Conversion of Plants to Caustic Treatment ...................................5-22

xxvi


Figure 5-14 Road Map to Optimize Plants on Caustic Treatment ...........................................5-23 Figure A-1 Vacuum Degasifier................................................................................................ A-2 Figure A-2 Components of a GTM Unit .................................................................................. A-4 Figure A-3 GTM System Process Flow Diagram (Note: 1 gal = 3.785 litres).......................... A-5 Figure A-4 GTM System Process Flow Diagram (Note: 1 gal = 3.785 litres).......................... A-6 (8)

Figure A-5 A Storage Tank Nitrogen System and Sparging Elements ................................. A-9 Figure B-1 Calculated current-voltage curves for the hydrogen electrode reaction (HER), the oxygen electrode reaction (OER) and the total polarization curve (sum of those for the HER and OER) for a system containing 10 ppb H2 + 5 ppb O2 at 25 °C. Note that the ORP must lie between the equilibrium potentials for the HER and the OER. ...... B-7 e

Figure B-2 Calculated equilibrium potentials for the Hydrogen Electrode Reaction (E her) e and the Oxygen Electrode Reaction (E oer) and the calculated redox potential (ORP) o for the system 10 ppb H2 + X ppb O2 at 25 C. Note that, generally, the ORP approaches the equilibrium potential for the HER at very low oxygen levels, but does not approach the equilibrium potential for the OER at very high oxygen levels. .... B-10 Figure B-3 Plot of calculated redox potential versus molar ratio of O2:H2 for hydrogen concentrations of 2, 5, 10, and 20 ppb H2. The arrow indicates a “redox neutral” system having a molar ratio of 0.5 .................................................................................. B-18 Figure C-1 Multisensor Probe (courtesy of Intek, Inc) ............................................................ C-3 Figure C-2 Multisensor Probe Instrument Schematic .............................................................. C-4 Figure C-3 Flow Schematic for SF6 Analyzer System ............................................................. C-7 Figure C-4 Schematic Diagram of SF6 Sampling System ....................................................... C-8 Figure E-1 Head Cup for Constant Sample Flow (pressure in psi at discharge equals the head, H inches of water, divided by 27.6) ................................................................... E-9 (2)

Figure E-2 Sampling System Configuration Used During RP2712-3 Project ...................... E-12 Figure E-3 Normal Distribution Curve for Random Data........................................................ E-23 Figure E-4 The Relationship Between True Value, Measured Value, and Bias. The bias in the example shown is negative, but a positive bias is also possible. .......................... E-24 Figure E-5 Random and Systematic Errors ........................................................................... E-26 Figure E-6 Laboratory Quality Control Charts........................................................................ E-31 Figure E-7 Cation Conductivity Control Chart, RR%, CC = 0.5 µS/cm .................................. E-32 Figure E-8 Cation Conductivity Control Chart, RSD%, CC = 0.5 µS/cm................................ E-32 Figure F-1 Copper Concentration at Virginia Power’s Chesterfield Unit 6 Before and (3) After a Filter .....................................................................................................................F-7 Figure F-2 Boiler Water Copper Concentration/Drum Boiler Pressure Control Curves (7) Developed at Miami Fort Station ....................................................................................F-11 Figure F-3 Drum Boiler Water Copper vs. Operating Pressure ..............................................F-12 Figure F-4 Road Map to Develop Shutdown and Layup Guidelines Common to Most Units .................................................................................................................................F-14

xxvii


+

Figure G-1 Concentrations of the Aqueous Ammonium Cation (NH4 ) (Continuous Line) and Aqueous Neutral Ammonia Species NH3 (Dashed Line) in Solutions of Pure Ammonia. .......................................................................................................................... G-4 Figure G-2 Relationship Between Corrected Boiler Water pH, Measured Boiler Water pH and Measured Ammonia Concentration, as Determined using Equation G-15. The lines on the figure increase in 0.1 ppm ammonia increments from 0 to 2 ppm.......... G-6

xxviii


LIST OF TABLES Table 1-1 Key Cycle Chemistry Guidelines* .............................................................................1-6 Table 1-2 EPRI’s New Steam Limits........................................................................................1-10 Table 1-3 CEGB Water/Steam Circuit Guidelines for Drum Type Boilers as Currently Amended by One Operator of Caustic Treatment (CT). (Section 5 should be referred to for the latest understanding on pH in the feedwater to control copper corrosion). ........................................................................................................................1-20 Table 1-4 Boiler Water Limits Used by Various Operators of CT for High Heat Flux, High Pressure Drum Boilers (typically 16 MPa (2400 psi)).......................................................1-22 Table 2-1 Percentage of organizations experiencing chemically influenced BTF. Results (4) from survey of 93 organizations in 2001 (Figures in parentheses indicate survey results from 1997). .............................................................................................................2-2 Table 2-2 Key Parameter Limits Relating to Continuum of Treatments.....................................2-4 Table 2-3 Major Unit Transport and Deposition Problem Areas for Units with All-Ferrous and Mixed-Metallurgy Feedwater Systems ........................................................................2-6 Table 2-4 Feedwater Limits for All-Ferrous Systems. Values in parenthesis are easily achieved by the best units..................................................................................................2-8 Table 2-5 Feedwater Limits for Mixed-Metallurgy Systems. The value in parenthesis is easily achieved on the best units with low air in-leakage. ................................................2-15 Table 3-1 EPRI Steam Purity Limit Revisions Based on PTZ Research Findings...................3-11 Table 3-2 EPRI’s Core Monitoring Parameters and/or Minimum Level of Continuous Instruments for All Units Operating on Phosphate Continuum and Caustic Treatments .......................................................................................................................3-29 Table 6-1 Possible Causes of Chemical Excursions .................................................................6-4 Table 6-2 Condensate System – Corrective Actions .................................................................6-5 Table 6-3 Feedwater System – Corrective Actions....................................................................6-6 Table 6-4 Boiler Water - Corrective Actions...............................................................................6-7 Table 6-5 Steam Circuit - Corrective Actions.............................................................................6-8 Table 6-6 Makeup System Malfunction......................................................................................6-9 Table 6-7 Makeup Regenerant - NaOH, H2SO4 or HCl Leaking into the Cycle........................6-10 Table 6-8 Condenser Tube Leak .............................................................................................6-11 Table 6-9 Condensate Polisher Malfunction/Exhausted ..........................................................6-12 Table 6-10 Condensate Polisher – Caustic or Acid Regenerant Leakage...............................6-13 Table 6-11 Excessive or Insufficient Concentration of Oxygen Scavenger .............................6-14 Table 6-12 Corrosion and/or Flow-Accelerated Corrosion in the Preboiler Cycle....................6-15

xxix


Table 6-13 Phosphate Hideout and Return (Phosphate Continuum).......................................6-16 Table 6-14 Low Sodium/Phosphate Ratio and High Phosphate Concentration in Boiler Water (Phosphate Continuum).........................................................................................6-17 Table 6-15 Low Sodium Hydroxide Concentration, Low NaOH:Chloride Ratio, and High Chloride Concentration in Boiler Water (Caustic Treatment) ...........................................6-18 Table 6-16 High Sodium Hydroxide Concentration in Boiler Water (Caustic Treatment) ........6-18 Table 6-17 Phosphate and Sodium Hydroxide Carryover (Phosphate Continuum).................6-19 Table 6-18 High Boiler Carryover and Contaminated Steam (Caustic Treatment) ..................6-20 Table 6-19 Impurity Introduction into Turbine by Attemperating Sprays ..................................6-21 Table 6-20 High Concentrations of Acids and Salts in Steam .................................................6-21 Table 6-21 Air In-leakage to Hotwell........................................................................................6-22 Table 6-22 Colloidal Silica and Iron through the Makeup System ...........................................6-23 Table 6-23 Boiler Layup Problems∆ .........................................................................................6-24 Table B-1 Classification of Corrosion Processes Showing the Potential Conditions for (5) their Occurrence . ........................................................................................................... B-3 Table B-2 Definitions of Various “Potentials” Encountered in Corrosion Science and (5) Engineering, Particularly in Thermal Power Plant Technology ...................................... B-4 Table B-3 To Convert ORP or Corrosion Potential (ECP) Values Measured Using Reference Electrode #1 to Values on Reference Electrode #2 Scale, Add the Indicated Conversion Factor to the Measured Potential. (Source: Reference 22). ........ B-21 Table B-4 Expected ORP Values for Reference Quinone Solutions at pH 4 and pH 7 (Source: Reference 23)................................................................................................... B-23 Table C-1 Examples of a Five-Probe Air In-leakage Measurement System........................... C-5 Table C-2 MSP Probe Indications for Various Probe Positions ............................................... C-5 Table E-1 Summary of Criteria for Sampling ........................................................................... E-6 Table E-2 Working Definitions of QA/QC Terms for Fossil Plant Chemistry Data Validation ........................................................................................................................ E-22 Table E-3 Typical Sampling Errors ........................................................................................ E-27 Table E-4 Continuous Instrument Acceptance Limits, 95% Confidence Interval (Based on Seven Replicates) ...................................................................................................... E-36 (4)

Table F-1 Design features of copper filter ............................................................................F-10 Table F-2 Comparison Of Copper Values in Blowdown (Figures F-2 and F-3) ......................F-13 Table F-3 Maximum Annual Exposure to Contaminant Conditions for Cycling and Peaking Operation ...........................................................................................................F-17

xxx


1 INTRODUCTION TO EPRI’S CYCLE CHEMISTRY PROGRAM, AND TO THE PHOSPHATE CONTINUUM AND CAUSTIC TREATMENT

1.1 THE EPRI CYCLE CHEMISTRY PROGRAM Availability and reliability are of paramount importance to the overall economic performance and profitability of fossil plant unit operations. Industry statistics have demonstrated the negative impacts of improper water chemistry on unit availability and reliability, as a consequence of chemistry-related failures and associated unscheduled outages. Plant assessments have shown how deficient chemistry practices reduce the efficiency and performance of fossil plant components in contact with water and steam. Further, non-optimum chemistry conditions can shorten the useful service life of fossil plant components, requiring that replacement projects begin sooner than normally required. In recognition of these issues, the EPRI Cycle Chemistry Program was established in 1984. Initial efforts and activities addressed the most obvious and apparent needs within the fossil plant industry. The EPRI response to these needs included: •

Improving the recognition and understanding of the impacts on fossil plant equipment caused by deficient chemistry practices.

•

Critically appraising the science of water and steam chemistry, and identifying specific data needs and other deficiencies.

•

Establishing industry guidelines for cycle chemistry in all varieties of fossil plant units.

•

Through open communications, conferences and collaborative research efforts, creating a worldwide network of cycle chemistry specialists, allowing appraisal of the science and technology on a global basis.

•

Preparing publications and other products intended to facilitate technology transfer to fossil plants, designed to simplify application of good chemistry practices.

Over the nearly 20 years the program has been in existence, the understanding of chemistry influenced damage and the effects of deposits on unit performance has increased substantially. Unfortunately, there are still cases where the causes of damage and performance degradation are not properly identified, resulting in situations where the role of chemistry goes unrecognized, or damage and performance losses not involving chemistry are incorrectly determined to be chemistry related.

1-1

EPRI Licensed Material Introduction To EPRI’s Cycle Chemistry Program, and to the Phosphate Continuum and Caustic Treatment

There are also many cases where the optimum cycle chemistry has not been selected and continually validated, or where inadequate instrumentation was responsible for allowing units to operate with gross contamination. Deposits can also impair performance and have been experienced in many areas of the steam-water cycle. Chemistry influenced component damage in fossil plant units is widespread and includes the following mechanisms. •

Condenser tubes: stress corrosion cracking, pitting, condensate grooving.

•

Condenser structure: flow-accelerated corrosion of steam side shell, supports, headers and piping.

•

Deaerators: flow-accelerated corrosion, pitting, corrosion fatigue, and stress corrosion cracking.

•

Feedwater heaters and associated piping: general corrosion and pitting, corrosion fatigue, flow-accelerated corrosion, stress corrosion cracking, and deposits.

•

Economizer tubes: pitting, flow-accelerated corrosion and corrosion fatigue.

•

Boiler tubes: hydrogen damage, acid phosphate corrosion, caustic gouging, corrosion fatigue, pitting, and deposit induced overheating.

•

Superheaters and reheaters: pitting corrosion, stress corrosion cracking and corrosion fatigue.

•

Turbines: corrosion fatigue, erosion and corrosion, stress corrosion cracking, crevice corrosion, pitting, and deposits (reducing efficiency and capacity).

It should be noted that some of these damage mechanisms were unknown at the inception of the program. Others were not readily distinguished from superficially similar damage mechanisms, including some that are not influenced by chemistry. In other cases, the extent of components that were vulnerable to the damage mechanism was not fully appreciated. Today, there is a very good understanding of damage mechanisms, including the influence of chemistry on many of them. Permanent solutions, based on identification of the responsible root cause and initiation of action to “kill the mechanism” are generally available. Deficient chemistry is either a root cause or significant influencing factor in all chemistry influenced damage mechanisms. (1) Initial interim chemistry guidelines were issued in 1986. Subsequent research findings, field experience with the interim guidelines, and worldwide cycle chemistry practices justified updates and revision activities. As a result, individual guidelines for phosphate, all-volatile, and oxygenated treatments were issued in the 1990s.(2-4) Additionally, a document describing (5) favorable international experience with caustic treatment of drum boilers was published.

This report represents the second in a series of “third generation” EPRI cycle chemistry guidelines for fossil plants which will be published between 2002 and 2004. The first of these revisions was for all-volatile treatment.(6)

1-2


This guideline introduced a couple of very important new concepts into the world of fossil plant cycle chemistry: Decoupling of the steam and boiler limits, while providing unique protection for the steam turbine and the boiler. Clear distinction between AVT(O) and AVT(R) for oxidizing and reducing conditions for all-ferrous and mixed-metallurgy feedwater systems respectively, and Use of ORP as a core parameter for controlling feedwater chemistry with mixed-metallurgy feedwater systems operating on AVT(R). This second guideline is for the solid alkali phosphate and caustic treatments. As such it introduces the concept of Phosphate Continuum (PC), which replaces the two treatments (EPT (2) and PT) in the previous phosphate guideline. In 2004, the OT Guideline will be revised. 1.1.1 Program Goals and Objectives The overall objectives of the program are to provide guidelines, technology and training materials, which together will assist in avoiding the major damage and failure mechanisms in fossil plants. By implementation of improved water chemistry, the following goals, which have been set for the EPRI program, are attainable by virtually all fossil plant units: •

No boiler tube failures influenced by cycle chemistry

•

No turbine problems involving the cycle chemistry, specifically:

•

•

•

•

–

no corrosion fatigue in low pressure turbine components

–

no stress corrosion cracking in disks

–

minimum deposits (no availability losses or performance concerns)

Optimized feedwater treatment to: –

eliminate serious flow-accelerated corrosion failures

–

minimize iron and copper transport (each to less than 2 ppb in the feedwater)

Operational guidelines for all unit designs and operating conditions –

selected to protect boiler and turbine

–

customized for each unit

Simple and reliable chemistry instrumentation and control –

minimum (“core”) levels of instrumentation for all units and treatments

–

continual chemistry surveillance, control and alarms for all units

Optimized procedures for unit shutdown and layup

1-3


•

•

•

Eliminate unneeded chemical cleanings –

appraise need to clean

–

establish effective approaches and procedures

Optimum managerial approach and support for cycle chemistry –

training of staff

–

benchmarking assessments of plant chemistry programs

–

value and risk-based management tools for assessment of cycle chemistry improvements

Cost effective cycle chemistry programs

There are already a number of world class utility organizations that enjoy the benefits of operating without chemistry-related boiler and turbine failures, with minimal rates of corrosion product transport, requiring few (if any) chemical cleanings, etc. Many others are working with EPRI to improve their chemistry programs and making measurable progress, with commensurate changes in unit availability and performance. 1.1.2 Program Philosophy The overall philosophy of EPRI’s Cycle Chemistry Program for Fossil Plants is shown in Figure 1-1. Various projects, including state-of-knowledge assessments, technology appraisals, research and development programs, and relevant non-technical investigations are performed to improve the overall understanding of the science of water treatment technology and how to optimally apply it to working fossil plant units.

1-4


Figure 1-1 Overall Philosophy of EPRI’s Cycle Chemistry Program. CCIP is Cycle Chemistry Improvement Program, BTFR is Boiler Tube Failure Reduction, FAC is Flow-accelerated Corrosion, and CP is Condensate Polishing.

The results of these projects serve as critical input to development and products for use by plant personnel. As indicated in bold in the figure, the main products consist of cycle chemistry guidelines, a cycle chemistry advisor (ChemExpert), and various training programs that ensure proper technology transfer to plant staff for optimal understanding and application. By following this approach, it has been possible to continually refine the understanding of the underlying science while also making appropriate changes in the products utilized by plant personnel. 1.1.3 Key Cycle Chemistry Guidelines In all, there are 10 essential cycle chemistry guideline documents that should be available for use by all utility personnel responsible for fossil plant cycle chemistry. Included are two operating guidelines(4,6) as well as this current document for PC and CT, four selection, process and (7-10) (11-13) Table 1-1 indicates the subject transient guidelines, and three cycle support guidelines. matter of these guideline publications, the year of publication, and the year in which publication of revised and updated guidelines is planned.

1-5

EPRI Licensed Material Introduction To EPRI’s Cycle Chemistry Program, and to the Phosphate Continuum and Caustic Treatment Table 1-1 Key Cycle Chemistry Guidelines* Guideline Type/Subject

Year Published

Planned Updates

Operating Guidelines •All-Volatile Treatment (AVT)

2002

•Oxygenated Treatment (OT)

1994

•Phosphate Continuum (PC) and Caustic Treatment (CT)

2003

2004

Selection, Process and Transient Guidelines •Selection and Optimization

1994

+

•Flow-Accelerated Corrosion

1997

2004

•Cycling/Startup/Shutdown/Layup

1998

2004/2005

•Control of Copper in Fossil Plants

2000

Cycle Support Guidelines •Makeup (Revision 1)

1999

•Chemical Cleaning (Revision 2)

2001

•Condensate Polishing

1996

2005

*See Appendix G for further information on these and other publications. + This report will be removed from circulation after publication of the revised OT Guidelines in 2004. The selection and optimization process in now included in each EPRI Guideline (see for example Section 2).

As shown in Figure 1-1, the chemistry guideline documents are integral to the content of the training programs and other tools developed for operations, maintenance, technical, and management personnel. The updated guidelines are the initial conduit through which the findings of research and development efforts are transferred to the plants. However, the important new concepts introduced in the guidelines are subsequently integrated into training program materials and future versions of the expert system code (ChemExpert). 1.1.4 Program Vision and Future Plans Utilities desiring optimum benefits from these and future cycle chemistry guidelines publications will derive the best results as follows: •

1-6

Perform initial benchmarking assessments of existing cycle chemistry and boiler tube failure reduction programs to establish worldwide rankings for each unit and to identify areas of deficiency. (EPRI’s approach to Cycle Chemistry Benchmarking is included as Appendix D.)


•

Organizations desiring optimized chemistry should arrange for Boiler Tube Failure Reduction/Cycle Chemistry Improvement Program (BTFR/CCIP) training to familiarize staff with: a) the controllable aspects of the key cycle chemistry program guidelines, b) the importance of formalized, management supported, BTFR/CCIP Programs, and c) the importance of establishing customized chemistry treatment programs based on the guidelines. Use of a Corporate Mandate document provides the needed support from management.

•

Use EPRI ChemExpert to provide minute-by-minute surveillance of the chemistry, early warning of chemistry excursions, direction to response actions which will minimize or prevent damage to equipment. This program will also eventually allow assessment of the risks of continued operation with out-of-specification chemistry until normal chemistry can be restored, as well as the value of program improvements

•

Perform follow-up benchmarking assessments to track progress in improving plant chemistry programs and identify any areas where the staff’s understanding of the guidelines would benefit from review of training materials

Feedback obtained from utility personnel working with these products is an important means by which EPRI’s Cycle Chemistry Program is continually upgraded.

1.2 RESEARCH SUPPORTING REVISION OF THE CYCLE CHEMISTRY GUIDELINES EPRI research conducted over the last ten years has resolved prior knowledge deficiencies in several critical areas. These include: •

Chemical environment and liquid films in the Phase Transition Zone (PTZ) region of the low pressure steam turbine

•

Corrosion processes in the PTZ of steam turbines

•

Volatility and solubility of impurities in steam and water

•

Copper corrosion and transport in fossil plant steam-water cycles

The influence of research in these areas on the development of guidelines follows in subsequent sections. Sections 2 and 3 provide further details on how the research findings influence selection of treatments and the rationale for these new PC and CT chemistry guidelines.

1-7


1.2.1 Chemical Environment and Liquid Films in the Phase Transition Zone (PTZ) The risk of chemistry related turbine damage is greatest within the phase transition zone (PTZ), where corrosion of low pressure (LP) blades and disks results in substantial availability losses with commensurate cost impacts on the affected fossil generating units. EPRI has sponsored a series of research projects intended to understand the environment in the PTZ in detail, to improve the understanding of the corrosion damage processes in the PTZ and the effects of the (14-18) chemistry on these processes. As a result of 10 years of research and international collaborative work with 23 organizations, Figure 1-2 schematically shows the important processes that take place in the PTZ. The impurities, oxides and ions in the superheated steam act as centers for the heterogeneous nucleation of the first drops of moisture sometimes called the “first” or “early” condensate. These drops concentrate the impurities (for chloride and sulfate up to 200 times) and are now known to be electrically charged. However, it is very important to note that there is no oxygen within these droplets, even for units operating on oxygenated treatment (up to 400 ppb oxygen). The droplets can impinge on the turbine surfaces (blades and disks) and give rise to liquid films on the surfaces. The properties of these liquid films have been extensively studied in EPRI research.(17-19) The concentration of impurities in the liquid film is at least ten times higher than in the droplets and the pH can drop down to below 7. These liquid films are important because they provide the dynamic environment for the PTZ corrosion mechanisms. It has also been shown that the liquid films have a potential and a high conductivity. As the liquid films flow off the blade surfaces they generate relatively large droplets compared with the early condensate droplets. Both are charged as they flow into the exhaust hood at an approximate moisture level of around 8%.

Processes in PTZ Charged Condensate

Steam and Impurities, Oxides, Ions

Volatility and Partitioning

Nucleation of Moisture

Deposits Charged Liquid Films

E x h a u s t

Figure 1-2 Schematic of the Processes in PTZ. The extra arrows pointing in to the “Deposits” indicate that deposition can occur by other methods.

1-8


The EPRI research has also determined that the unit cycle chemistry has a major effect on the properties of the liquid films. Parallel work has also addressed deposition of salts, oxides and impurities onto the PTZ surfaces. Some of the main conclusions from this large body of research, which directly influences the development of the revised EPRI AVT Guidelines include: •

It appears that less deposition and thus more reliable operation in the steam turbine PTZ will be accomplished by reducing the current steam limits for Na, Cl, and SO4 to below 2 ppb, although there is significant transport and deposition of steam impurities in units with concentrations of sodium and chloride in steam of less than 0.1 ppb.

•

The concentration of impurities by drying of liquid films and moisture droplets on surfaces, where the surface temperature is above the steam saturation temperature can be significant. Liquid films have been detected on blade surfaces above the saturation line.

•

Pitting can be initiated in relatively pure steam on typical turbine materials in cycling units that do not apply any shutdown protection to the turbine.

•

The cycle chemistry plays a major influence on the chemical composition of the early condensate and of the liquid film on blades.

•

Low volatility salts, volatile and film-forming amines in steam affect the condensation process, particularly at lower steam expansion rates such as those in turbines.

The relation of these conclusions to cycle chemistry treatment and steam purity is clear, and these “third generation” chemistry guidelines have included these results in deriving some of the revised limits. 1.2.2 Corrosion Processes in the PTZ of Steam Turbines These extensive new findings on the environment and liquid films in the PTZ have led to a new model of how corrosion processes (corrosion fatigue (CF) and stress corrosion cracking (SCC)) take place. Figure 1-3 shows the model. When the turbine is operating, the chemical environment in susceptible locations in the PTZ consists of dynamic liquid films and deposition of salts, oxides and impurities. There is no oxygen in the liquid films. However, when the turbine shuts down, most organizations do not provide any protective environment, and so the deposits become moist once the surfaces cool down, and cause passivity breakdown and pits to form. Repetition of the operating/shutdown environments eventually leads to microcracks. Only when the turbine is operating is the loading (cyclic or steady state) sufficient to drive the microcracks into corrosion fatigue or stress corrosion cracks. Only when the unit is operating are liquid films present, which provide the environment for the cracks to propagate. Thus the main ramifications for the development of these new guidelines are the importance of higher purity steam (reduction of the limits for sodium, chloride and sulfate) and of protecting the turbine during shutdown.

1-9


Loading

Chemistry

Ultimately, research results in this area will allow development of a computer model of the LP turbine that can be used to appraise the potential for corrosion fatigue and stress corrosion cracking. This model will include predictive capabilities to assess the time to failure once precursor conditions (pits, crevices and microcracks) are determined to be present. The improved model will eventually be integrated into other EPRI software such as BLADE and ChemExpert.

Turbine Operation

Shutdown

Deposits and Liquid Films (No O2)

Deposits, Oxygen and Moisture

Pitting Crevices

Pitting Microcracks

Cyclic

CF

Steady State

SCC

No Loading

Figure 1-3 Model for the Development of Corrosion Processes in the PTZ of Steam Turbines

1.2.3 EPRI’s New Steam Limits One of the overall results of all the research on the PTZ is the derivation of a new set of EPRI steam limits. These are shown in Table 1-2. Table 1-2 EPRI’s New Steam Limits OT

AVT

PC(L)

CT

PC(H)

Na, Cl, SO4 (ppb)

2

2

2

2

3

Cation Conductivity (µS/cm)

0.15

0.2

0.2

0.2

0.3

1-10


1.2.4 Volatility and Solubility of Impurities in Steam More than a decade of research(20-26) has shown that volatility predictions based on the ray diagram are not accurate in defining volatile carryover in drum boilers of fossil plant units. In some cases, the values suggested by the ray diagram are off by as much as two orders of magnitude. Volatility experiments conducted to reevaluate impurities of interest resulted in a much better understanding of this subject and the complexities involved in properly applying this science to fossil plants. Most of the results from 10 years of EPRI research are shown in Figure 1-4. Key findings of the volatility experiments and subsequent data evaluations may be summarized as follows. •

Impurities may carry over as neutral salts, as 1:1 ionic compounds or as 1:2 ionic compounds. Neutral, 1:1 and 1:2 compounds cannot be compared directly and separate diagrams for each of these groups are shown in Section 3.

•

Chloride exhibits relatively high volatility as hydrochloric acid (HCl), minor volatility as ammonium chloride (NH4Cl), and only minimal volatility as sodium chloride (NaCl).

•

Sodium hydroxide (NaOH) has a volatility comparable to that of NaCl and introduction of these species to the steam path from the boiler is strongly dependent on mechanical carryover rather than volatile carryover.

•

Under oxidizing conditions, volatile transport of sulfur species to the steam occurs mainly as sulfuric acid (H2SO4), however, under reducing conditions, volatile carryover of sulfur dioxide, (SO2), a neutral compound, can introduce sulfur to the steam. Also, carryover of ammonium sulfate (NH4HSO4) will be present in steam at high pH under AVT chemistry conditions.

•

Phosphoric acid (H3PO4) is considered a neutral compound as it does not ionize appreciably at boiler operating temperatures, yet volatile carryover was determined to be quite low. Transport of phosphates to the steam is essentially the result of mechanical carryover. Trisodium phosphate (Na3PO4), is, for practical purposes, non-volatile.

•

Organic acids such as formic acid (HCOOH) and acetic acid (CH3COOH) are neutral compounds under fossil plant operating conditions and are very volatile, however, any organic salts in steam would exhibit lower volatility and could lead to a reduction of liquid film pH.

•

Volatile carryover of copper from boiler water to the steam occurs mainly as cupric hydroxide (Cu(OH)2) and cuprous hydroxide (CuOH); both species exhibit very high volatility across the whole range of temperatures and pressures experienced in a fossil plant from startup to full operation.

•

Silica volatilization characteristics are quite consistent with the findings of other researchers.

•

Carryover predictions for operating fossil plant units with drum boilers require use of a thermodynamic model to make an appraisal of the boiler water components as a system; a model has been developed that allows prediction of saturated steam chemistry and the chemistry of the early condensate. Component volatility is pH dependent and mechanical carryover rate has a significant impact on steam purity.

1-11


Figure 1-4 Partitioning Constants for Common Salts, Oxides and Acids. KD is the partitioning constant.

This model was used in developing the AVT Guidelines(6) to validate the derivation of boiler water limits. It is used in a similar way in this PC and CT Guideline as discussed in Section 3. The final version of the code will be incorporated within the EPRI ChemExpert software. It is envisioned that doing so will allow utilities to customize their boiler water limits and action levels for individual drum boiler units and to predict changes in steam purity and early condensate chemistry with variations in the boiler water chemistry.

1-12


1.2.5 Copper Corrosion and Transport in Fossil Cycles Under Program Copper, EPRI sought to establish an improved understanding of copper corrosion, transport and deposition in fossil plant cycles, leading to new guidelines for control of mixed-metallurgy cycles.(10,25-32) It has been shown that the corrosivity of individual copper alloys is dependent primarily on oxidation-reduction potential (ORP), and also on pH, as shown in Figure 1-5. Ramifications here for the revised guidelines are that the minimum in corrosion for copper alloys occurs in the pH range 9.0-9.3 under reducing conditions. Also, the solubility of cupric oxide (CuO) in steam, about 1-2 ppb under fossil plant conditions, was found to be generally consistent with prior research, while the solubility of cuprous oxide (Cu2O), about 23.5 ppb, was determined to be somewhat greater than predicted by other investigators. In boiler water, the copper oxides exhibit increasing solubility as a function of temperature up to around o o 200 C (392 F). At typical boiler operating temperatures however, the solubilities appear to be independent of temperature, about 2 ppb for CuO, and about 6 ppb for Cu2O. Boiler pH (ammonia concentration) has considerable influence on copper oxide solubility at high temperatures. These results have been used to develop revised boiler water copper limits. Because of the solubility of the copper oxides in boiler water and the high volatility of the oxides to steam, it is of paramount importance that reducing conditions exist during any shutdown periods to minimize the transport of copper to steam.

Figure 1-5 pH Dependence of the Cu Release Rates at 95°C (203°F) for Al-Brass, 90Cu-10Ni and Admiralty Brass After Treatment in Non-Degassed Feedwater at ORP +100 mV and –300 mV* * OX = Oxidizing, RED = Reducing

1-13


1.3 FIELD EXPERIENCE SUPPORTING REVISION OF THE PC AND CT GUIDELINES, AND INTRODUCTION TO PC AND CT This current guideline is a derivative and combination of the previous 1994 EPRI Guideline on Phosphate Treatment(2) and a report on Caustic Treatment(5). This version now introduces the Phosphate Continuum (PC), which together with the consolidation of the previous caustic report, provides all the guidance necessary for solid alkali treatments in fossil plants. This subsection provides a short historical perspective on phosphate and caustic treatments, and introduces the philosophy behind the new PC. 1.3.1 Phosphate Treatments and Introduction of the Phosphate Continuum Leading to New PC Guidelines The use of phosphate chemicals for internal boiler water treatment is more than 70 years old, and is currently used in about 65% of drum units worldwide. Coordinated pH-phosphate control was introduced in 1942 to protect boiler tubes from “caustic embrittlement” as well as the effects of condenser in-leakage of water hardness contaminants. The treatment was designed to preserve magnetite and provide protection against “caustic under-deposit” corrosion attack and hydrogen damage from “under-deposit acidic chloride” attack. The use of coordinated treatment led to a number of failures believed, at the time, to be caustic gouging and, as a result, to the use of congruent phosphate treatment (CPT) with an operating range below the sodium-to-phosphate molar ratio curve of 2.6 (Figure 1-6). However, the move to CPT was also not without challenges. Many utilities experienced phosphate hideout with increasing load and pressure, (a decrease in phosphate and an increase in pH), and hideout return when the unit load decreased (an increase of phosphate occurring with a pH depression). The depression of pH on startup to below 7 or 8 is now known to exacerbate boiler tube failures, particularly those occurring by corrosion fatigue. More often, problems, notably acid phosphate corrosion (APC), arise as a result of the use of the mono- and di-sodium phosphate in a vain effort to “chase” hideout so as to maintain the control point within the CPT range. An international survey conducted in 1994 showed that over 90% of drum units in the U.S. operating with congruent phosphate treatment experienced hideout, and more than 60% had corrosion consequences.

1-14


Figure 1-6 Schematic of Operating Ranges of Boiler Water on Equilibrium Phosphate Treatment (EPT), Congruent Phosphate Treatment (CPT) and Phosphate Treatment (PT). It should be noted that these are the old phosphate operating ranges, and are shown here for reference only. They shouldn’t be applied to fossil plants.

In the early 1990s, EPRI introduced a new guideline for phosphate treatment(2). These 1994 Guidelines introduced equilibrium phosphate treatment (EPT) and phosphate treatment (PT) after much EPRI and independent research into corrosion mechanisms (acid phosphate corrosion, APC), and after good operating experiences in Russia, Germany and on Great Lakes cooling waters. The ranges are shown in Figure 1-6, and although the Guidelines suggested that only trisodium phosphate and NaOH should be used, Figure 1-6 shows the lower bound of the EPT/PT regions at Na:PO4 molar ratio of 2.8. This was based on research conducted by Tremaine, which showed that maricite (the corrosion product of APC) would not form above higher molar ratios than 2.8 at boiler operating temperatures of 360°C (680°F)(33). This guideline brought enormous success to the world of phosphate treatments in conventional fossil plants. It is not possible to form maricite (NaFePO4) with only trisodium phosphate additions. A more recent international survey in 2003 found that as an increasing number of operators had moved to EPT and PT, only about 40% of boilers now experience hideout and that less than 20% have evidence of acid phosphate corrosion.

1-15


Despite the two phosphate treatments, EPT and PT, successfully addressing acid phosphate corrosion, there have been an increasing number of operational uncertainties and some other corrosion problems: •

The improper use of low level phosphate treatments (such as EPT) have been associated with an increase of hydrogen damage failures in both conventional plants and HRSGs(34). There have been over 30 fully documented cases. This results from the wrong choice of boiler water control curves for chloride, which allow too high a level of chloride in the boiler water. Hydrogen damage has been increasing over the last 10 years. Basically operators did not make a clear distinction between the extremes of EPT and PT in their ability to “neutralize” contaminant.

•

The name of EPT also caused lots of confusion because there isn’t really any “equilibrium”, and although EPRI published a road-map to derive the “equilibrium” level (minimum level that a boiler can maintain without hideout), very few organizations used this procedure. But many high pressure boilers continue to operate at about 0.2 ppm of phosphate or lower.

•

Most operators were not clear about the distinct differences between EPT and PT and thus exactly where to operate. Figure 1-6 shows an artificial boundary between the two treatments at about 2.4 ppm phosphate.

•

Most operators did not measure and/or know the percentage of mechanical carryover on their drum units.

•

Many operators did not take into account the ammonia levels when operating at low phosphate levels or alternatively made the wrong connection. Similarly operators did not account for sodium associated with contaminant anions when these were present.

In summary, the 1994 Guidelines were very successful in addressing the main corrosion concern of acid phosphate corrosion, but the application of the guideline was often unsuccessful in addressing control and overall boiler corrosion protection (hydrogen damage). These concerns have led to the need for a new control technology and operating philosophy for phosphate treatments, which could not simply be addressed by changing a few things in the EPT/PT envelope. The new guidelines now encompass a Phosphate Continuum (PC). The operating ranges are shown in normal and log plots in Figure 1-7.

1-16


(a)

(b) Figure 1-7 Schematic of Operating Range for PC. a) is normal scale, b) is log scale.

1-17


The new PC treatment keeps the overriding criteria that drove the 1994 Guidelines, but also adds new features to address the concerns delineated above. This combination leads to a treatment which: •

Will be applicable to a wide range of boilers (pressures) and contaminant levels

•

Has a wide range from low phosphate to around 10 ppm to assist in controlling hideout

•

Requires only the addition of TSP and NaOH to ensure that acid phosphate corrosion will not occur and is thus encompassed by Na:PO4 = 3 and TSP + 1 ppm NaOH

•

Covers the possibility of contaminant control at the low phosphate end by having a minimum of 0.2 ppm phosphate. This will prevent operations using < 0.2 ppm where there is essentially no protection from contamination.

•

Covers the high phosphate end for low pressure boilers and in cases where contaminants are likely/probable. (See Section 2.3 on selection and optimization of boiler water treatments.)

•

Continues the use of boiler water contaminant control curves (Section 4)

Phosphate continuum is a much better terminology as phosphate treatments are indeed a true continuum of treatments, as follows: •

Provides boiler protection from very low pressures (e.g. LP HRSG evaporator circuits at 60/70 psi) to very high conventional plants operating at drum pressures above 2600 psi (18 MPa) and to HP HRSG evaporator circuits.

•

Is a continuum of phosphate treatments from 0.2 ppm to 10 ppm, with the minimum being to prevent excessively low phosphates being used which provide little resistance to contaminant ingress.

•

Extends from the low phosphate levels (< 3 ppm) where the pH is moderated by ammonia, to the high phosphate levels, where ammonia has a smaller effect.

•

The higher phosphate levels have a much greater ability to counteract contaminant ingress.

•

The opposite ends of the PC have much different abilities to partition sodium into steam (Section 3).

Although PC is a continuum as described above, it also has to be recognized that there are two extremes of low (L) and high (H) phosphate levels. In Section 4 these two regimes have been called PC(L) and PC(H) with a very loose demarcation between the L and H variants at around 3 ppm phosphate. However, this demarcation is not marked on the PC schematic (Figure 1-7) to prevent organizations polarizing between the two treatments in the way that was done for EPT and PT. The emphasis must be on recognizing that there are these two extremes (and intermediate conditions) and they will behave differently as indicated by the last set of bullets above. This very important point is discussed again extensively in the Rationale (Section 3) and in the Guidelines (Section 4).

1-18


In summary, the 1994 EPRI Phosphate Guidelines made a tremendous contribution to the industry. However, clearly operators/chemists have experienced numerous problems in developing reliable operating ranges, and there were some serious corrosion problems as discussed in the next Section. Thus the time has come to make the previous phosphate treatments obsolescent. The introduction of PC has done that. It keeps the best features of the old, but also addresses the deficiencies of the old. 1.3.2 Introduction to Caustic Treatment and Operating Experiences Leading to New CT Guideline Caustic treatment has been used successfully for many years in coal and oil fired drum units as an alternative to phosphate treatment, particularly where difficulties with phosphate hideout, and monitoring and control have been experienced. In a survey conducted in the early 1990s for (5) EPRI , at least 50,000 MWe of plant was being successfully operated with caustic treatment (CT), mostly in the United Kingdom, Ireland, Hong Kong and South Africa. The EPRI survey included details of operating experience, including problems and how they were overcome. Guidance on conditioning using sodium hydroxide was developed into an EPRI interim (5) guidelines . The EPRI Interim Cycle Chemistry Guidelines for Combined Cycle Heat Recovery (35) Steam Generators published in 1998 included caustic treatment as a method of conditioning the boiler water for drum type units. More recently, caustic treatment has started to be used on “conventional” fossil fired drum units in the USA. High concentrations of sodium hydroxide and sodium phosphate were used in many countries in the 1950s for conditioning the boilers of low pressure plant. However, when these methods of conditioning were applied to new high pressure plant in the 1960s, there were frequent failures due to on-load corrosion in the form of caustic gouging or hydrogen damage due to acid attack. There was a great deal of confusion at the time, which led to the use of sodium hydroxide being abandoned in the USA in favor of AVT and various forms of phosphate treatment. While the use of high or “uncontrolled” concentrations of sodium hydroxide can certainly cause caustic gouging, it is now known that many examples of what were thought to be caustic gouging, were in fact acid phosphate corrosion, due to phosphate hideout and dosing acidic forms of sodium phosphate. In other countries, particularly the UK, the use of sodium hydroxide was pursued and refined, so that the problems experienced with phosphate treatment, particularly those associated with hideout, could be avoided. The addition of low concentrations of sodium hydroxide to the boiler water used in caustic treatment significantly increases the amount of chloride that can be tolerated without corrosion, compared with AVT. This is important when dealing with condenser leaks, particularly when cooling is with brackish or sea water. It is still important to limit chloride ingress with CT and ensure that there is sufficient sodium hydroxide present to prevent potentially acidic conditions, while at the same time limiting the amount present, to prevent caustic gouging in the boiler and excessive carryover into the steam, resulting in damage to the turbine.

1-19


The former Central Electricity Generating Board (CEGB) in England developed guidelines for feedwater, boiler water and steam for once-through and drum type boilers over a period of several decades. The limits for feedwater were the same regardless of operating pressure, but those for boiler water and steam reduced as the pressure increased. AVT was recommended where high purity boiler water could be maintained and CT was recommended when this was not the case, for example when there were condenser leaks. CT was also recommended when the condensers were cooled with sea water, since in the event of condenser leaks, the sodium hydroxide would help to prevent acidic conditions forming in the boiler water. Like the phosphate treatments, the pH of the boiler water needs to be maintained at a sufficiently high value with CT to minimize the risk of corrosion in the boiler. To avoid excessively high concentration of sodium hydroxide, the concentration can be linked to the chloride concentration, this being the most corrosive contaminant present. A semi-empirical relationship was introduced in the 1970s, namely NaOH = 2.5 x chloride. Initially, this was based on laboratory tests and practical operating experience, but was later substantiated by pilot plant tests. A certain minimum amount of sodium hydroxide is required to achieve the required pH. In the CEGB Guidelines, the minimum amount of NaOH and the permitted chloride concentrations reduce with increasing pressure. The CEGB Guidelines worked well for low heat flux coal fired plants, but examples of caustic gouging occurred in high heat flux zones of oil fired boilers and it became necessary to reduce (36) the limits for NaOH and chloride in these plants. The 1985 CEGB Guidelines are still used as the basis for chemical operation by all large operators in the UK, albeit with some reductions to the limits. The CEGB Guidelines for CT, as currently amended by one operator, are given in Table 1-3. Table 1-3 CEGB Water/Steam Circuit Guidelines for Drum Type Boilers as Currently Amended by One Operator of Caustic Treatment (CT). (Section 5 should be referred to for the latest understanding on pH in the feedwater to control copper corrosion). Fuel Pressure (MPa) [psi]

Coal

Coal

Coal

Oil

6 [900]

10 [1500]

16 [2400]

16 [2400]

Feedwater Ammonia (ppb) pH at 25°C Boiler Water Chloride (ppm) Equivalent NaOH Minimum NaOH (ppm) Silica (ppm) Steam Sodium (ppb) Silica (ppb)

1-20

500 or 1000 when copper alloys absent 8.8 – 9.2 or 8.8 –9.4 when copper absent

3.5 2.5 × Cl 5 5

2.5 2.5 × Cl 5 1.5

1.2 2.5 × Cl 2 0.2

0.3 2.5 × Cl 0.5 0.2

20 20

10 20

2 20

2 20


CT is used in countries other than the UK, including Germany, Ireland, Hong Kong and South Africa. Early examples of guidelines and operating experience were included in the EPRI(5) report. Recent examples of CT operation are given below. The Guidelines used by Scottish Power(37) for their 16 MPa (2400 psi) plants are similar to the CEGB Guidelines(36). They have condensate polishing and readily achieve the much lower limits set for high heat flux oil fired plant, despite being a coal fired plant. Oil and gas fired 16 MPa (2400 psi) plants operated by Scottish and Southern(38) also, basically, use the CEGB Guidelines. (39) In England, PowerGen use the CEGB Guidelines, suitably modified to include the use of lower NaOH and chloride limits for intermediate pressure oil fired plants, with lower limits for sodium (40) in steam for all intermediate and high pressure plants. Innogy (formerly National Power) also use guidelines based on the CEGB Guidelines, adopting the lower limits given for high heat flux plants for all their plants, where possible. Although it is accepted that lower limits for boiler water recommended for high heat flux boilers and the lower limits for sodium in steam are desirable for all plants, it is also accepted that this is not always possible, and less restrictive limits may have to be adopted in some cases. The limit for chloride in the boiler water adopted by the operators in England and Scotland is (41) typically 0.3 ppm chloride for 16 MPa (2400 psi) high heat flux plants. In Ireland , South (42) (43) Africa and Hong Kong the limits adopted for chloride are about half this value, because this (44) is readily achieved, as most plants have condensate polishing. The Coronado plant in the USA (3) typically operates with chloride concentrations similar to the old EPRI AVT limits , but is allowed to operate up to 0.2 ppm chloride, before operating pressure is reduced. The main boiler water limits used by various operators of CT are given in Table 1-4.

1.4 HOW TO USE THESE GUIDELINES This introductory section of the guidelines has presented the EPRI Cycle Chemistry Program objectives and achievable goals, and has shown how recent research findings provided the direction for the revisions to the new AVT guidelines(6). It has also included the field experience with phosphate and caustic treatments, and outlined the introduction of the new PC. These chemistry guidelines are applicable to fossil units with drum boilers operating with PC and CT. As with prior versions, the guidelines represent the best practices for those units that should utilize PC and CT. This in no way implies that the guidelines, as presented, can be immediately applied to individual units. Additional guidance is provided to assist users of the guidelines to develop an optimized cycle chemistry program with customized limits and action levels. Further, it is important to understand that PC and CT are not acceptable treatment for all units, and that the suitability and applicability of the cycle chemistry treatment approaches to a given unit will change over time. As described in Section 2, it is important to verify the suitability of PC or CT for use in a specific unit prior to initial operation with PC or CT and whenever significant changes in unit design and operation occur. In the case of existing units operated on a particular treatment for more than five years, users of these guidelines are strongly urged to use the road maps and supporting discussions of Section 2 to verify that the present program is still optimum for the unit in question because the latest research findings have influenced the applicability of the available feedwater and boiler water treatment options.

1-21

EPRI Licensed Material Introduction To EPRI’s Cycle Chemistry Program, and to the Phosphate Continuum and Caustic Treatment Table 1-4 Boiler Water Limits Used by Various Operators of CT for High Heat Flux, High Pressure Drum Boilers (typically 16 MPa (2400 psi)) Organization

Year

Chloride (ppm Cl)

CEGB

1985

0.3

1988/85

VGB


Sodium Hydroxide (ppm NaOH) 0.5–0.75

50+

0.8

ESB (Moneypoint)

1992

0.12

1.5

0.8 max

ESKOM

1993

0.15

3

0.87–2.6

EPRI (proposed)(5)

1995

0.4

8

1.0–1.5

EPRI (HRSG)(35)

1998

0.35

6.5

0.9–1.6

PowerGen

Current

0.3

Innogy

Current

0.3 (target)

Scottish Power

Current

(0.33* equiv)

4

Scottish and Southern

Current

(0.37* equiv)

4.5

0.5 or 2.5×Cl

ESB (Moneypoint)

Current

0.12

1.5

0.8 max

ESKOM

Current

0.15

1.5

1.1–2.0

CLP Power Hong Kong Ltd

Current

0.2

3.0

1.8–2.2

Coronado

Current

0.2

Note:

0.50–0.75

0.2–0.8

*chloride equivalent to cation conductivity allows for phosphate if added

+

Section 3 of the guidelines presents and discusses the new rationale used to develop these guidelines to establish appropriate target values and action levels. This section is most important (a “must read”) as it will be noted that the approach to be followed is completely new and different from previous EPRI guidelines, and differs with the system metallurgy. In Section 3, there are many figures, which have been included to illustrate how the guidelines normal and action levels have been derived in Sections 4 and 5. It is very important to note that none of these figures should be used to control the chemistry of fossil plants. The Section 3 presentation also defines the preferred sample points, chemical feed points and required on-line instrumentation for all situations to which PC or CT is the desired cycle chemistry.

1-22


Sections 4 and 5 cover the revised PC and CT guidelines for drum boiler units, respectively. As in prior guidelines, cycle diagrams are used to summarize the sample collection and chemical addition point, the basic (generic and uncustomized) chemistry target values and action levels, and the monitoring requirements. As explained in each of these sections, the instrumentation requirements now differ depending on whether an oxidizing or reducing feedwater chemistry is employed (AVT(O) or (AVT(R)). Guidance is given on how the guidelines should be used to establish unit-specific target values and action levels. Also presented are discussions of best practices for unit shutdown, layup, startup and cycling service as well as makeup water requirements and the role of chemical cleaning in optimized PC and CT programs. Section 6 addresses other aspects of optimized chemistry control programs. Primary attention is focused on prompt identification of chemistry excursions by plant personnel, accurate analysis of the cause or causes of the deviations from normal chemistry limits, and provision of corrective actions consistent with this analysis. In addition, this section considers the purity of treatment chemicals to be applied to units operated on PC or CT. Sections 1-6 are followed by eight appendices, providing further details on important topics related to PC or CT usage. Appendix A describes ways to remove dissolved oxygen from treated makeup water, which is very beneficial when filling units for startup. Techniques for monitoring feedwater ORP are presented in Appendix B. Condenser air in-leakage monitoring and control, necessary to ensure attainment of target values for dissolved oxygen in condensate, are presented in Appendix C. Appendix D provides information on the EPRI approach to cycle chemistry benchmarking. The importance of proper sampling and analysis practices as they relate to chemistry program management is reviewed in Appendix E. The contributions of staff training to the overall success of the program are also covered. Appendix F covers unit shutdown, layup, startup, cycling and peaking. Appendix G covers the calculation of the corrected pH in the boiler water to take account of the presence of ammonia. As indicated in Table 1-1, there are 10 key cycle chemistry guideline documents that all personnel within the organizations supporting the EPRI Cycle Chemistry Program should have. However, these publications represent only a small part of the products created since program inception in 1984. Appendix H provides a listing of all currently available guidelines, technical reports, and conference proceedings.

1-23


1.5 REFERENCES 1. Interim Consensus Guidelines on Fossil Plant Water Chemistry, EPRI, Palo Alto, CA: June 1986. CS-4629. 2. Cycle Chemistry Guidelines for Fossil Plants: Phosphate Treatment for Drum Units, EPRI, Palo Alto, CA: December 1994. TR-103665. 3. Cycle Chemistry Guidelines for Fossil Plants: All Volatile Treatment, EPRI, Palo Alto, CA: April 1996. TR-105041. 4. Cycle Chemistry Guidelines for Fossil Plants: Oxygenated Treatment, EPRI, Palo Alto, CA: December 1994. TR-102285. 5. Sodium Hydroxide for Conditioning the Boiler Water of Drum-Type Boilers, EPRI, Palo Alto, CA: January 1995. TR-104007. 6. Cycle Chemistry Guidelines for Fossil Plants: All-Volatile Treatment, EPRI, Palo Also, CA: November 2002. 1004187. 7. Selection and Optimization of Boiler Water and Feedwater Treatments for Fossil Plants, EPRI, Palo Alto, CA: March 1997. TR-105040. 8. Guidelines for Controlling Flow-Accelerated Corrosion in Fossil Plants, EPRI, Palo Alto, CA: November 1997. TR-108859. 9. Cycling, Startup, Shutdown and Lay-up Fossil Plant Cycle Chemistry Guidelines for Operators and Chemists, EPRI, Palo Alto, CA: August 1998. TR-107754. 10. Guidelines for Copper in Fossil Plants, EPRI, Palo Alto, CA: November 2000. 1000457. 11. Revised Guidelines for Makeup Water Treatment, EPRI, Palo Alto, CA: October 1999. TR-113692. 12. Guidelines for Chemical Cleaning of Conventional Fossil Plant Equipment, EPRI, Palo Alto, CA: November 2001. 1003994. 13. Condensate Polishing Guidelines, EPRI, Palo Alto, CA: September 1996. TR-104422. 14. Steam, Chemistry and Corrosion in the Phase Transition Zone of Steam Turbines, Volume 1: Key Results, Summary, and Interpretation, EPRI, Palo Alto, CA: February 1999. TR-108184-V1. 15. Steam, Chemistry and Corrosion in the Phase Transition Zone of Steam Turbines, Volume 2: Part 1: Individual Contributions of Participants, EPRI, Palo Alto, CA: February 1999. TR-108184-V2P1.

1-24


16. Steam, Chemistry and Corrosion in the Phase Transition Zone of Steam Turbines, Volume 2: Part 2: Individual Contributions of Participants, EPRI, Palo Alto, CA: February 1999. TR-108184-V2P1. 17. Turbine Steam, Chemistry and Corrosion: Generation of Early Liquid Films in Turbines, EPRI, Palo Alto, CA: September 1999. TR-113090. 18. Turbine Steam, Chemistry and Corrosion: Experimental Turbine Tests, EPRI, Palo Alto, CA: September 1997. TR-108185. 19. Corrosion of Low Pressure Steam Turbine Components, EPRI, Palo Alto, CA: November 2000. 1000557. 20. Behavior of Ammonium Salts in Steam Cycles, EPRI, Palo Alto, CA: December 1993. TR-102377. 21. Assessment of the Ray Diagram, EPRI, Palo Alto, CA: August 1996. TR-106017. 22. Volatility of Aqueous Sodium Hydroxide, Bisulfate and Sulfate, EPRI, Palo Alto, CA: February 1999. TR-105801. 23. Vapor-Liquid Partitioning of Sulfuric Acid and Ammonium Sulfate, EPRI, Palo Alto, CA: February 1999. TR-112359. 24. Volatility of Aqueous Acetic Acid, Formic Acid, and Sodium Acetate, EPRI, Palo Alto, CA: July 2000. TR-113089. 25. Behavior of Aqueous Electrolytes in Steam Cycles: The Solubility and Volatility of Cupric Oxide, EPRI, Palo Alto, CA: November 2000. 1000455. 26. The Volatility of Impurities in Steam/Water Cycles, EPRI, Palo Alto, CA: September 2001. 1001042. 27. State-of-Knowledge of Copper in Fossil Plant Cycles, EPRI, Palo Alto, CA: September 1997. TR-108460. 28. Corrosion of Cu-Ni-Zn Alloys in Water-Ammonia Power Plant Environments: Development of High Temperature Potential-pH (Pourbaix) Diagrams, EPRI, Palo Alto, CA: November 1999. TR-113697. 29. Copper Alloy Corrosion in High Purity Feedwater, EPRI, Palo Alto, CA: November 2000. 1000456. 30. Influence of Water Chemistry on Copper Alloy Corrosion in High Purity Feedwater, EPRI, Palo Alto, CA: October 2001. 1004586.

1-25


31. Copper Alloy Corrosion in High Purity Feedwater, EPRI, Palo Alto, CA: May 2003. 1007391. 32. Influence of Water Chemistry on Copper Alloy Corrosion in High Purity Feedwater, EPRI, Palo Alto, CA: September 2003. 1007612. 33. P.R. Tremaine, L.G.S. Gray, B. Wiwchar, P. Taylor and J. Stodola, “Phosphate Interactions th with Metal Oxides Under High Performance Boiler Hide-Out Conditions,” 54 International Water Conference, Pittsburgh, PA. IWC-93-35. 34. R.B. Dooley and W.P. McNaughton, Appropriate Controls for Phosphate Boiler Water Treatments to Avoid Acid Phosphate Corrosion and Hydrogen Damage, Power Plant Chemistry, 2001, 3(3), pp. 127-134. 35. Interim Cycle Chemistry Guidelines for Combined Cycle Heat Recovery Steam Generators (HRSGs), EPRI, Palo Alto, CA: November 1998. TR-110051. 36. CEGB, GOM72, Part 2, Chemical Control of the Steam Water Circuit of Drum Type Boilers, CEGB, England, Issue 6, September 1985. 37. A. Berg, Scottish Power, Personal Communication to M. Ball, December 2002. 38. K. Frazer, Scottish Southern, Personal Communication to M. Ball, December 2002. 39. R.R. Harries, PowerGen, Personal Communication to M. Ball, November 2002. 40. G. Bignold, Innogy. Personal Communication to M. Ball, October 2002. 41. F. McCarthy, J. Bane and G. O’Connor, “Oxygenated Treatment in a 300MW Drum Type th Boiler”, Proceedings: 6 International Conference on Cycle Chemistry in Fossil Plants, EPRI, Palo Alto, CA: April 2001. 1001363. 42. F. Wessels, T. Shabalala and R. Cramer, ESKOM, Personal Communication to B. Dooley, October 2002. 43. H. Cheng, CLP Power Hong Kong Ltd., Personal Communication to B. Dooley, October 2002. 44. T.W. Robertson, Salt River Project, Personal Communication to B. Dooley, October 2002.

1-26


2 SELECTION AND OPTIMIZATION OF FEEDWATER AND BOILER WATER

2.1 CHEMICALLY-INFLUENCED PROBLEMS, AND THE CONTINUUM OF TREATMENTS The unit chemistry selection and continuing optimization of feedwater and boiler water treatments are critically important to the longterm performance of a fossil plant. Often the chemistry for both is decided during initial unit design, and then “fine-tuned” during commissioning or early operation. Historically, the science of chemical treatments has changed over the last 20 years of the EPRI Cycle Chemistry Program, and many organizations have indeed changed treatments. Some examples of this include: •

a gradual progression from coordinated phosphate to congruent phosphate to trisodium phosphate treatments (EPT and PT)

•

conversion from a phosphate treatment to AVT

•

conversion from AVT to a phosphate treatment

•

conversion from AVT to OT, and

•

for all-ferrous feedwater systems, conversion from a reducing AVT environment (using reducing agents) to either an oxidizing AVT environment (without reducing agents) or OT.

However, it is very rare that an organization conducts a comprehensive monitoring campaign when such changes are made. Most often it is accomplished quickly, and often without a chemical clean, which is especially important if a boiler water treatment is changed. Also, and often more importantly, is the fact that as the unit gets older, it becomes more prone to contaminant ingress from increasing condenser leaks and air in-leakage. The majority of organizations do not continually (at least on a six months frequency) check whether the treatment is still optimized or even adequate for the changed conditions. Most often, when asked, organizations indicate “we operate with this boiler water (or feedwater) treatment because we always operated this way.” This situation is not satisfactory. Within this overall scenario is included the fact that very often organizations do not change contaminant limits when boiler water chemistry is changed. The classic example in the mid to late 1990s was the change from congruent phosphate treatment (CPT) to equilibrium phosphate (1,2) treatment. This change from CPT to EPT was desperately needed by the industry in the early 1990s to prevent phosphate hideout, hideout return, and acid phosphate corrosion. The publication of the 1994 EPRI Phosphate Guidelines(3) has reversed these trends. 2-1

EPRI Licensed Material Selection and Optimization of Feedwater and Boiler Water

However, on changing from CPT to EPT, there also needs to be a change of boiler water control limits, especially for chloride. For example for a 2500 psi (17.2 MPa) boiler, the chloride limit with EPT was very different to that for PT. Thus it is not surprising that there has been a drastic increase around the world in hydrogen damage tube failures in EPT units if the boiler water limits had not been chosen correctly. Overall this is part of a very disturbing trend of an increasing number of chemically influenced failures. Table 2-1 indicates the percentage of organizations experiencing cycle chemistry influenced BTF. Here it has to be said that not only those organizations with the poorest BTF statistics and poorest cycle chemistry performance have experienced hydrogen damage; in some cases it has occurred in organizations that have been previously benchmarked in the “World Class” category. Table 2-1 Percentage of organizations experiencing chemically influenced BTF. Results from survey of 93 organizations in 2001(4) (Figures in parentheses indicate survey results from 1997). Organizations having Chemically Influenced BTF

81%

(61%)

•

Hydrogen Damage

57%

(37%)

•

Acid Phosphate Corrosion

25%

(17%)

•

Corrosion Fatigue

45%

(43%)

•

Pitting

40%

(7%)

•

Stress Corrosion Cracking (SCC)

28%

(18%)

•

Caustic Gouging

11%

(11%)

Since 2000, EPRI has benchmarked a large number of organizations based on five key fossil plant cycle chemistry factors.(1) The benchmarking process is provided in Appendix D. The process provides a ranking for a unit/plant/organization from “World Class” to “Below Average.” World Class essentially means that a unit has no chemically influenced failures or performance losses, has a full complement of cycle chemistry instruments (see Table 3-2) which are alarmed for operators, has very few chemical cleans based on optimum choice of feedwater treatment, and has minimum cycle losses (increased makeup addition) due to blowdown. EPRI also conducts a boiler tube failure reduction program/cycle chemistry improvement program (BTFRP/CCIP), and over the period 1999 to 2002 has provided initial training and review to over 50 organizations worldwide. Some interim results from these two activities of benchmarking and BTFRP/CCIP have recently been collated into an assessment of the (1) importance of cycle chemistry on fossil plant performance. It is clear from this assessment that the most reliable and best performing units in the world operate on OT or AVT(O), have condensate polishing, all-ferrous feedwater systems, do not use reducing agents, have tight condensers, and excellent air in-leakage control. These organizations also take time to ensure that the boiler and feedwater treatments match the equipment and the equipment capability.

2-2


It is, of course, appreciated that while all units to be built in the future should adhere to these guiding principles to achieve performance requirements, not all currently operating units can instantly change to this spectrum. This again emphasizes the need for continuing and continual assessment of a unit’s cycle chemistry, which is the main focus of this section. There are four possible choices for drum boiler water treatment, as discussed in Section 2.3: AVT, OT, PC, and CT. There are three possible choices for feedwater treatment, as discussed in Section 2.2: AVT(O), AVT(R) and OT. The chosen treatments need to match the unit, unit materials (particularly in the feedwater system), cooling water, and possible contaminant ingress. The recent performance results mentioned above have led to a diagram, which simplifies the process of selection, or at least acts as a first-cut analysis. EPRI calls this the “Continuum of Treatments.” Figure 2-1 attempts to provide an overview of the Continuum. The width of the wedge or funnel, in which the continuum of treatments is located, is proportional to the possible level of contamination in the cycle, and to the ability of any of the five boiler water treatments to neutralize or buffer any contaminants. As the point of the funnel is approached the level of possible contaminants becomes less and the more likely the plant will have very good feedwater control. Such a plant will have a condensate polisher and/or a very tight condenser. As the open end of the funnel is approached, plants might have no condensate polisher, seawater or high TDS water from a cooling lake or a cooling tower. The better the air in-leakage, the nearer to the point will the plant be positioned. The vertical dotted line represents an arbitrary distinction between units with all-ferrous and mixed-metallurgy feedwater systems. Clearly there is a relationship with the boiler water treatments: OT and AVT(O) are only applicable to units with all-ferrous feedwater systems, but AVT(R), PC, and CT are applicable to both all-ferrous and mixedmetallurgy. In fact in a number of cases around the world, a hybrid version of CT is used on units effectively operating with OT where the operator adds a small amount of NaOH to the boiler water. For the Phosphate Continuum, the approximate range of operation has been inserted into Figure 2-1. As already discussed in Section 1-3, there is a need to consider both ends of the PC, and to accomodate this, both PC(L) and PC(H) have been shown. Clearly PC(L) has much less ability to buffer contamination, whereas PC(H) has the best ability of all the boiler water treatments. In terms of cycle purity, moving towards the point of the funnel requires lower feedwater contaminants. For example (Table 2-2), the continuum for cation conductivity at the economizer inlet might be: for OT (< 0.15µS/cm), for AVT(O) and AVT(R) (< 0.2 µS/cm), and with CT (< 0.2 µS/cm). For PC it will depend on whether the treatment is at the low end (< 3 ppm phosphate) or at the high end (> 3 ppm phosphate). These are termed PC(L) and PC(H) with possible cation conductivity limits of < 0.2 µS/cm and < 0.3 µS/cm, respectively. Table 2-2 also provides an example for boiler water at a drum pressure of 2500 psi (17.2 MPa) with a mechanical carryover of 0.18% of the normal control values.

2-3

EPRI Licensed Material Selection and Optimization of Feedwater and Boiler Water Table 2-2 Key Parameter Limits Relating to Continuum of Treatments (It should be noted that the boiler water values are only provided as an example at a drum pressure of 2500 psi (17.2 MPa) for a drum with a mechanical carryover of 0.18%. Normal control values for all pressures and carryovers will need to be obtained from the appropriate EPRI Guidelines).

Treatment

OT

AVT

Parameter

CT

PC PC(L)

PC(H)

Feedwater and Steam Cation Conductivity µS/cm

0.15

0.2

0.2

0.3

0.2

Steam, Na, Cl, SO4 ppb

2

2

2

3

2

Cation Conductivity µS/cm

1.5

1.5

10(1)(4)

25(1)(4)

9(4)

Cl ppb

50(3)

50(2)

150(4)

600(4)

300(4)

SO4 ppb

100(3)

100(2)

300(4)

1200(4)

600(4)

SiO2 ppb

180

180

180

180

180

Boiler Water 2500 psi (17.2 MPa) 0.18% Mechanical Carryover

Notes:

1. 2. 3. 4.

For Na:PO4 molar ratio of 3 New AVT Guideline(17) New OT Guideline to be published in 2004 Sections 4 and 5 of this Guideline

In terms of performance, the best results occur using OT and AVT(O), whereas operating with PC and CT there are operational problems such as hideout and hideout return. There clearly has to be benefit in an organization changing the chemistry treatments and maybe the unit equipment (adding a polisher, changing out copper feedwater heaters). This is the focus of EPRI’s program on assessing the value of cycle chemistry.(5) The primary purpose of this section of the new Guidelines is to provide the needed advice, guidance, and road maps for selection and optimization of boiler water and feedwater. The process can, and should, be used each and every time a change in treatment is contemplated, and also used for continual assessment of boiler water and feedwater. (6) This process, which was developed by EPRI in 1997, is even more important today because of the deteriorating statistics of chemically influenced failures and performance losses (Table 2-1). The 1997 document will be removed from circulation after publication of the revised OT Guideline in 2004. Each guideline now contains a selection and optimization section (see the next Section 2.2).

2-4


Figure 2-1 Summary of Possible Boiler Water and Feedwater Treatments for Fossil Plants as a Function of Equipment and Equipment Capability(1). This is the “Continuum of Treatments”.

2.2 SELECTION AND OPTIMIZATION OF FEEDWATER TREATMENT 2.2.1 Introduction and Types of Feedwater Treatment For all-ferrous feedwater systems (no copper alloys in the feedwater, but there can be copperbased condenser tubing), the generation and transport of corrosion products (magnetite, hematite and ferric oxide hydrate) occurs mainly due to corrosion and flow-accelerated corrosion of low pressure and high pressure feedwater heaters, deaerators, economizer inlet tubing and piping, feedwater piping and drain lines. For mixed-metallurgy systems, the generation and transport of corrosion products (cupric and cuprous oxide) occurs mainly due to corrosion of any low pressure and high pressure feedwater heater tubes manufactured from copper alloys. Some possible consequences of these corrosion problems are outlined in Table 2-3.

2-5

EPRI Licensed Material Selection and Optimization of Feedwater and Boiler Water Table 2-3 Major Unit Transport and Deposition Problem Areas for Units with All-Ferrous and MixedMetallurgy Feedwater Systems All-Ferrous Systems •

•

Mixed-Metallurgy Systems

Waterwall pressure drop due to ripple magnetite deposits on waterwalls of oncethrough units

• HP turbine copper deposits • Primary superheater copper deposits • Copper deposits in HP heaters

Flow-accelerated corrosion

• Pump strainer/screen copper deposits All Feedwater Systems •

Boiler deposits and increased boiler pressure drop

•

At least five boiler tube failure mechanisms

•

Frequent need to chemical clean

•

Boiler feedpump fouling

•

Orifice fouling

Thus feedwater chemistry is critical to the overall reliability of fossil plants. Corrosion products are generated here, flow around the cycle, deposit in various areas acting as initiating centers for most of the major failure mechanisms, and often need removing by chemical cleaning. There are three distinctly different feedwater treatments: •

Reducing all-volatile, AVT(R), which uses ammonia and a reducing agent. Here the oxidation-reduction potential, ORP, should be reducing and may be in the range –300 to – 350mV [Ag/AgCl/sat, KCl], which is necessary to protect mixed-metallurgy feedwater systems.

•

Oxidizing all-volatile, AVT(O), where the reducing agent has been eliminated. Here the ORP will be around 0 mV but could be slightly positive or negative.

•

Oxygenated treatment (OT) where oxygen and ammonia are used. Here the ORP will be around +100 to +150mV.

In each case, the ORP value is simply an indication that the ORP needs to be reducing or oxidizing. Appendix B should be consulted to understand how the ORP is a function of many variables. Most operators can easily meet the feedwater requirements for cation conductivity. However, it is clear that this is not sufficient by itself, and major efforts need to be applied to the generation and transport of corrosion products. Less than 2 ppb of iron and/or copper is very achievable by the best units operating on AVT(R). For units operating with AVT(O) the iron levels should be around 1 ppb, and for units operating with OT iron levels should be around 0.5 ppb.

2-6


2.2.2 All-ferrous Feedwater Systems Optimization All of the three feedwater treatments mentioned above in Section 2.2 are possible for use in allferrous feedwater systems. The basis of either of the AVT treatments is an elevated pH in all cycle streams. The common alkalizing agent is ammonia. Originally, and up to the late 1980s, the ammonia dosing was always combined with a reducing agent feed, such as hydrazine. This treatment is now termed AVT(R), which indicates that the oxygen level at the condensate pump discharge (CPD) is low enough (< 10 ppb) (minimum air in-leakage) that a reducing agent can be added to the cycle to produce a reducing environment with ORP < 0 mV. Figure 2-2 illustrates the oxide formation (magnetite, Fe3O4) which will be formed on all the ferrous surfaces throughout the feedwater system. The dissolution of Fe3O4 into the feedwater flow is dependent on the ORP. The more reducing is the feedwater the greater is the dissolution and thus the higher is the amount of iron corrosion products measured at the economizer inlet. Flowaccelerated corrosion (FAC) occurs by exactly the same mechanism, which is accelerated at (7-9) locations where the flow hydrodynamics are elevated. Under reducing conditions that produce FAC or normal corrosion, organizations are not able to meet the guideline requirements of less than 2 ppb iron in the final feedwater at the economizer inlet. Investigations performed since the late 1980s, and hundreds of unit operating experiences have (10) indicated that eliminating the reducing agent feed minimizes the corrosion product generation. This treatment is now termed AVT(O). It also requires the air in-leakage be minimized to produce oxygen levels at the CPD of less than 10 ppb. Figure 2-3 shows a very typical example in a 600 MW drum unit with an all-ferrous feedwater system. As the reducing agent (hydrazine) was reduced over a period of 90 days, the ORP increased from about –350mV to around +80mV in the oxidizing range, and the iron levels reduced markedly. There was no change in the feedwater oxygen level.

Figure 2-2 Schematic Representation of Oxide Formed on Ferrous Feedwater Surfaces During Operation with Reducing AVT

2-7


Figure 2-3 Change in Oxidizing-Reducing Potential (ORP) and Feedwater Iron Levels (Fe) at the Economizer Inlet when Hydrazine (N2H4) is Gradually Reduced on a 600MW Drum Unit with an All-Ferrous Feedwater System(10).

The feedwater key parameters for AVT(R), AVT(O) and OT are summarized in Table 2-4. Table 2-4 Feedwater Limits for All-Ferrous Systems. Values in parenthesis are easily achieved by the best units. Parameter

AVT(O)

AVT(R)

OT

pH

9.2–9.6

9.2–9.6

D 9–9.6 O 8–8.5

< 0.2

< 0.2

< 0.15

Fe (ppb) at EI

< 2 (<1)

<2

< 2 (0.5)

Cu (ppb) at EI

<2

<2

Oxygen (ppb) at EI

< 10

< 5 (< 2)

D 30–50 O 30–150

Oxygen (ppb) at CPD

< 10

< 10

< 10

No

Yes

No

Not needed

–300 to –350

Not needed


Reducing Agent ORP (mV) at DAI*

Notes: EI - economizer inlet, CPD - condensate pump discharge, DAI - deaerator inlet, D - drum unit, O - Once-through unit * - Reference to Appendix B.

2-8


While an elevated pH is the basis of the two AVT treatments, oxygenated treatment (OT) uses oxygenated high purity feedwater to minimize the corrosion and FAC in the feedwater systems. Elevated pH is not a necessity with OT. The addition of oxygen is the preferred choice. The feedwater key parameters for OT are shown in Table 2-4. For the application of oxygenated treatment in units with once-through and drum boilers with allferrous feedwater systems, there is one indispensable prerequisite that the cation conductivity must be less than 0.15 µS/cm (at 25°C). The basis for the success of OT depends on the formation of a layer of ferric oxide hydrate (FeOOH/ferric oxide (Fe2O3)) on the surface of the magnetite oxide layer and within the pores. This is illustrated in Figure 2-4. In this case the FeOOH has a much lower solubility than Fe3O4 in feedwater, so the dissolution of surface oxide layers and the measurement of iron corrosion products will be minimal. FAC should also be eliminated on feedwater surfaces that have the red FeOOH/Fe2O3 appearance. It should be also noted that the use of AVT(O) also produces red surface layers of FeOOH/Fe2O3 throughout the feedwater systems. However, the lower levels of oxygen, maintained by good air in-leakage control, are usually not high enough to passivate 100% of the feedwater (particularly the deaerator) surfaces. Thus the lowest levels of iron transport cannot be expected; typically iron levels for optimized AVT(O) feedwater treatment will be around 1 ppb.

Figure 2-4 Schematic Representation of Oxide Formed on Iron-Based Feedwater Surfaces During Operation with Oxidizing AVT and OT

2-9


Optimization of All-Ferrous Feedwater Chemistry Figure 2-5 shows a road map for optimizing the feedwater treatment in all-ferrous systems. The primary purposes of this important activity are to minimize corrosion product transport, eliminate any possibility for FAC, and thus to reduce the accumulation of corrosion product deposition on the boiler waterwalls. The methodology described here is equally applicable for both drum and once-through units with all-ferrous feedwater systems. Step 1—Review Normal or Current Feedwater Treatment

This step involves a review of the current feedwater treatment, which is probably AVT(R) with a reducing agent. If there are no problems such as indicated in Table 2-3 then continue to use the current treatment. Such a review would indicate that the operating experience has been good, that minimal chemical control problems have been experienced, that no BTF in the waterwalls relating to waterside problems have occurred in the last five years, that no turbine deposition or blade failure problems have occurred, and that the feedwater is operating in the optimum fashion with minimum levels of feedwater corrosion products (less than 2 ppb at the economizer inlet). In such cases of good experience, no changes need to be made. However, it is suggested that the road map is reviewed as there may be considerable economic savings to be gained by converting to AVT(O) or OT, and it should be remembered that FAC is always possible with reducing feedwater chemistry. Step 2—Monitoring Baseline on Current Feedwater Treatment

This step involves a complete base-line monitoring to quantify the current chemical parameters and, in Step 3, to determine whether continued use of a reducing agent and AVT(R) or a change to AVT(O) should be contemplated. This program would utilize the installed chemistry monitoring system, supplemented by monitoring those parameters in the cycle chemistry diagrams under “troubleshooting and commissioning.” The monitoring program should pay particular attention to the adequacy of the makeup and chemical feed systems, condenser tightness, air in-leakage, and corrosion product transport.

2-10


Figure 2-5 Road Map for Optimizing Feedwater Treatment for All-Ferrous Feedwater Systems

2-11


This monitoring involves taking a “thumb-print” of the unit under “typical operating conditions” to identify under controlled conditions exactly how the unit chemistry is behaving. It may involve a review of the operating chemistry logs, but this often is not satisfactory and it is preferable to undertake a monitoring campaign. Before this campaign is initiated, it is important to review the utility’s chemistry monitoring capability and reliability. This should include Quality Assurance (QA) and Quality Control (QC) of existing and normally utilized analytical chemistry monitoring and analysis methodology and equipment (see Appendix E). The monitoring campaign should include: •

Varying Operating Conditions—base load, startup and shutdown

•

Steam Chemistry—cation conductivity, sodium, chloride, silica and sulfate

•

Feedwater Chemistry—cation conductivity, chloride, corrosion products (Fe, Cu), oxygen, pH, oxidizing-reducing potential (ORP)

•

Operation of Condensate Polishers (if included)

If this step indicates a low level of feedwater corrosion product transport (such as Fe <2 ppb) and acceptable feedwater purity from a dissolved solids standpoint with the control chemistry meeting the guideline values provided in Sections 4 and 5, then an organization might consider to continue with the current AVT(R) feedwater chemistry. Further confirmation of optimum feedwater chemistry will be that waterwall deposition rates have historically been much less than 2 2 1 mg/cm /1000 hours (about 1 g/ft /1000 hours), that the interval between chemical cleans has historically been better than every 10 years, and that there has been no indication of FAC in the feedwater, feedwater steamside, deaerator, or feedwater heater drain lines. Step 3—Evaluate Reducing Agent Requirements

This step is a subset of Step 2 and should involve a series of tests to minimize the generation and transport of feedwater corrosion products. Many utilities with all-ferrous systems have found that, with proper air in-leakage control (dissolved oxygen at the CPD of less than 10 ppb), the reducing agent (such as hydrazine) can be (10) eliminated without jeopardizing chemistry control on the unit and moving to AVT(O). Thus in Step 3, a series of tests could be performed to evaluate the need for a reducing agent and, if needed, to determine the proper reducing agent level. The tests should utilize the monitoring system instrumentation (used in Step 2) while reducing or eliminating the reducing agent dosage. Particular note should be made of dissolved oxygen levels and the level of corrosion products generated during each test. Reference can be made to Figure 2-3, which illustrates the results from a typical test. Table 2-4 should be referenced for typical parameter ranges. Particular attention should be paid to the fact that there can be a long time between changes in the feedwater chemistry and corrosion product transport, so careful planning is needed for accurate extended tests. For instance, two months after the test outlined in Figure 2-3, the iron levels were below 2 ppb and generally approaching 1 ppb.

2-12


Step 4—Monitoring with New Feedwater Treatment

Step 4 involves a period of normal operation with the new feedwater treatment, which occasionally requires repetition of monitoring in Step 2 to confirm that running with reduced or zero reducing agent provides the optimum feedwater treatment. This might involve a reduced monitoring effort which just looks at feedwater oxygen and corrosion products at the economizer inlet, in parallel with the cycle “core” parameters (Table 3-2), which should be normally continuously recorded and alarmed. In units with all-ferrous feedwater systems, but with copper based condenser tubing, there is a possibility of copper corrosion particularly in the air removal section of the condenser. If increased copper transport is measured at the CPD, then considerations should be given to a reduction in cycle pH from 9.2–9.6 to 9.0–9.3 as needed to reduce copper levels to those present during baseline monitoring. Of course such an increase in copper dissolution will not be a result of the AVT(R) to AVT (O) change, but the pH reduction was probably needed when operating in the AVT(R) mode. Steps 5 and 6—Consider Converting to OT

Once the baseline monitoring (Step 2) and a period of normal operation has been undertaken then the question can be raised about whether the unit could run on oxygenated treatment (Step 5). The reader is referenced to the OT Guideline,(11) which provides step-by-step guidance on whether a unit is suitable for OT and how to convert units to OT (Step 6) in a sequential fashion. The transition check list in the OT Guidelines(11) should be referenced. It should be noted that conversion to OT will take a number of weeks (more for drum units) and must include the heater drains. Step 7—Continue to Optimize the Feedwater Treatment

This step continues the efforts described in Steps 3 and 4. Step 8—Operation and Continuing Monitoring

There are now three treatments that can be used for the feedwater: • • •

AVT(R), AVT(O), and OT

The feedwater treatment used on each unit should be continually checked to ensure it is always the optimum treatment. As well as the core parameters at the key cycle points, it is now necessary to ensure that the operation of the deaerator vents and heater vents are operated in a manner which provides minimum levels of corrosion products in all parts of the cycle.

2-13


2.2.3 Mixed-Metallurgy Feedwater Systems Optimization AVT(R) feedwater treatment is regarded as the only suitable treatment for mixed-metallurgy (12,13) and the EPRI Copper Guidelines(14) the feedwater. Based on recent EPRI research results basis of AVT(R) is an elevated pH (9.0–9.3) (see Figure 1-5), minimum air in-leakage to ensure that the dissolved oxygen at the CPD is less than 10 ppb, and the addition of a reducing agent to the feedwater to ensure a reducing ORP. Maximum effort must be given to ensure the reducing environment during operation and shutdown. Appropriate shutdown and layup is discussed in Appendix F. (14) The research in Program Copper has clearly indicated that cuprous oxide (Cu2O) is the more protective oxide of choice for feedwater surfaces. This oxide is formed when the feedwater environment is reducing, with an optimum ORP between –350mV and –300mV. Thus there is a need to monitor ORP (versus Ag/AgCl, sat. KCl) at the deaerator inlet (to protect any LP and HP feedwater heaters with copper alloy tubing). On units without a deaerator then ORP can be monitored at the economizer inlet. Details on monitoring ORP and the necessary calibration procedures are included in Appendix B.

Controlling air in-leakage is of paramount importance in reaching the optimum ORP value. Experience has indicated for over 15 years that if the air in-leakage is kept below 1 scfm/100MW, the level of oxygen within the condensate (at the condensate pump discharge, CPD) can be kept below 10 ppb. Detailed procedures for monitoring air in-leakage are included in Appendix C. It is now clear that the previous oxygen limit of 20 ppb at the CPD within the original EPRI guidelines, in combination with a reducing agent addition based on the oxygen level (three times the oxygen concentration) would not always produce a reducing environment in the LP feedwater. The preferred approach is now to focus effort on air in-leakage to reduce oxygen in the condensate, and to add only sufficient reducing agent to reach and maintain the optimum ORP at the deaerator inlet. Under reducing conditions (ORP –350mV), the surface oxide is cuprous oxide (Cu2O) and its formation results in a compact even layer on the surface. (Figure 2-6, photos b, d, f and h), whereas under oxidizing conditions (ORP +100mV) the surface oxide contains increasing amounts of cupric oxide (CuO) (Figure 2-6, photos a, c, e and g). The copper release rate under oxidizing conditions can be up to 30 times higher than under the reducing conditions, because the surface roughness increases markedly with increasing ORP. The importance of maintaining a reducing environment during operation cannot be overemphasized. The cuprous oxide surface starts to change to cupric oxide within ten hours once the environment becomes oxidizing. Also, when a unit is shut down, if the ORP is kept reducing at an ORP of –350mV the surface oxide (Cu2O) will remain on the surface. However, during a non-protected shutdown, when the ORP is allowed to become oxidizing, the surface oxide layers start to change quickly (within 10 hours) from Cu2O to CuO again with a marked increase in surface roughness and transported copper on startup. Similar results have been obtained for all the copper alloys (admiralty brass, aluminum brass, Cu-Ni alloys, and Monel). The feedwater key parameters for AVT(R) for mixed-metallurgy feedwater systems are summarized in Table 2-5.

2-14

EPRI Licensed Material Selection and Optimization of Feedwater and Boiler Water Table 2-5 Feedwater Limits for Mixed-Metallurgy Systems. The value in parenthesis is easily achieved on the best units with low air in-leakage. Parameter

AVT(R)

pH (25°C, 77°F) at EI

9.0–9.3

Cation Conductivity (µS/cm) (25°C, 77°F)

< 0.2

Fe, ppb at EI

<2

Cu, ppb at EI

<2

Oxygen (ppb) at EI

< 5 (< 2)

Oxygen (ppb) at CPD

< 10

Reducing Agent

Consistent with a Reducing ORP

ORP (mV) at DAI (25°C, 77°F)*

–300 to –350

Notes: EI - economizer inlet; CPD - condensate pump discharge; DAI - deaerator inlet. Values in parentheses indicate best achievable. * Reference should be made to Appendix B to review how a reducing ORP is very dependent on a number of variables.

The most common alkalizing agent is ammonia. The most common and most satisfactory reducing agent is hydrazine. Any alternates (amines and new organic water treatment chemicals) must be evaluated in a monitoring campaign as indicated below under “Optimization” prior to continuous use (Figure 2-7). They are not preferred for mixed-metallurgy systems because they break down in the cycle and their use has been found to directly increase the transport of copper corrosion products.

2-15


a)

c)

b)

d)

e)

f)

g)

h)

Figure 2-6 Surface Images of Admiralty Brass After Treatment at 95°C (203°F), pH 9 and ORP = +100 mV (a,c,e,g) and ORP = -350 mV (b,d,f,h) for 8 (a,b), 24 (c,d), 100 (e,f) and 200 (g,h) hours. The surface oxide under reducing conditions is protective Cu2O. The rougher surface oxide under oxidizing conditions is CuO.(12)

2-16


In summary, for plants with mixed-metallurgy feedwater systems, the approach should involve: •

first, getting the ORP in the optimum range (–350 to –300 mV) at the deaerator inlet (DAI) by addressing air in-leakage, reducing agent addition and oxygen levels,

•

second, optimizing the feedwater pH,

•

third, tracking the feedwater copper (and iron) levels, and

•

fourth, optimizing the complete feedwater treatment to minimize copper (< 2 ppb) and iron (< 2 ppb).

The importance of proper shutdown and layup have already been discussed, and the necessary procedures are further discussed in Appendix F. Sampling and analysis procedures are covered in Appendix E. Optimization of Mixed-Metallurgy Feedwater Chemistry Figure 2-7 shows a road map for optimizing the feedwater treatment in mixed-metallurgy feedwater systems. The primary purposes of this important activity are to minimize corrosion product transport, and to reduce the accumulation of copper deposits throughout the boiler (waterwalls and primary superheater) and the HP turbine. All of the steps included on this Road (14) Map are covered in detail in EPRI’s Copper Guideline. Step 1—Review of Water Chemistry, Operation, and Experience

Water Chemistry. The first part of this review is verification of proper sampling and chemistry monitoring practices to ensure that valid data are collected; in particular, for monitoring of iron and copper in feedwater boiler water and steam, and determination of drum boiler carryover. Any deficiencies must be corrected before Step 2 is initiated. The chemistry data should be compared to customized limits based on the EPRI guideline limits shown in Sections 4 and 5, and a judgement should be made whether they are acceptable or not. Review of Operation. The type of operation, layup, and startup procedures should be reviewed for their relation to cycle chemical transport, particularly with regards to oxidation of copper alloys and boiler carryover. For cycling and peaking units, the influence of operation on air inleakage, layup protection, and transient chemistry should be reviewed. Experience. The maintenance reports need to be reviewed and personnel interviewed to evaluate the results of nondestructive testing and inspections, and the residual life of major components influenced by water and steam chemistry. Specifically, the status of condenser and feedwater heater, boiler, and superheater tubing and problems with flow-accelerated corrosion of carbon steel in the feedwater cycle and in the economizer will be used in the decision making about design and material changes.

2-17


Figure 2-7 Road Map for Optimizing Feedwater Treatment for Mixed-Metallurgy Systems

If there are no problems such as those indicated in Table 2-3, then continue to use the current treatment. Such a review would indicate that the operating experience has been good, that minimal chemical control problems have been experienced, that minimal BTF in the waterwalls relating to waterside problems and deposition of feedwater corrosion products have occurred in the last five years, that there have been no turbine deposition problems or blade failures, and that the feedwater is operating in the optimum fashion with minimum levels of feedwater corrosion products (<2 ppb Fe and <2 ppb Cu at the economizer inlet). In such cases of good experience, no changes need to be made. However it is suggested that the road map is reviewed as there may be considerable economic savings to be gained by optimizing AVT(R). 2-18


Step 2—Baseline Monitoring

Baseline monitoring is a step which determines cycle chemical characteristics such as boiler carryover, oxygen control, and transport of corrosion products. During the baseline monitoring with verified sampling and analytical procedures, compliance with specifications of key water and steam chemistry parameters is quantified. If problems exist, this monitoring should also determine where and during what type of operation any corrosion and water chemistry problems occur. For example, which feedwater heater generates copper alloy corrosion products, tube side vs. shell side, during startup and/or normal operation, etc. This program should use the installed chemistry monitoring system, supplemented by monitoring those parameters in the cycle chemistry diagrams under “troubleshooting and commissioning.” The monitoring program should pay particular attention to the adequacy of the makeup and chemical feed systems, condenser tightness and air in-leakage, and corrosion product transport. This monitoring involves taking a “thumb-print” of the unit under “typical operating conditions” to identify under controlled conditions exactly how the unit chemistry is behaving. It may involve a review of the operating chemistry logs, but this often is not satisfactory and it is preferable to adopt a monitoring campaign. Before this campaign is initiated, it is important to review the utility’s chemistry monitoring capability and reliability (as in Step 1). This should include Quality Assurance (QA) and Quality Control (QC) of existing and normally utilized analytical chemistry monitoring and analysis methodology and equipment (see Appendix E). The monitoring campaign should include: •

Varying Operating Conditions—base load, startup and shutdown

•

Steam Chemistry—cation conductivity, sodium, chloride, silica and sulfate

•

Feedwater Chemistry—cation conductivity, chloride, corrosion products (Fe, Cu), oxygen, pH, oxidizing-reducing potential (ORP)

•

Operation of Condensate Polishers (if included)

During the tests for optimum pH and ammonia concentrations (9.0–9.3) for mixed metallurgy systems, the pH measurements should be made with on-line equipment designed for pH measurement in high purity water at carefully controlled sample temperatures (25°C±0.1). Ammonia concentrations should be related to specific conductance to be sure that no appreciable carbon dioxide is present. If this step indicates a low level of feedwater corrosion product transport (such as Fe <2 ppb and Cu <2 ppb) and acceptable feedwater purity from a dissolved solids standpoint, with the control chemistry meeting the guideline values provided in Sections 4 and 5, then an organization might consider to continue with the current chemistry. Further points of confirmation of optimum 2 feedwater chemistry will be that waterwall deposition rates are much less than 1 mg/cm /1000 hours (about 1 g/ft2/1000 hours), that the interval between chemical cleans has historically been better than every 10 years, and there has been no indication of FAC in the heater drain lines.

2-19


Step 3—Water Chemistry Optimization

If water chemistry, corrosion and scale problems are identified in the review part of the optimization process, cycle chemistry optimization can possibly improve the situation. This step needs to involve operation, startup, and layup. The feedwater chemistry optimization usually consists of reducing the ingress of impurities, such as air in-leakage and condenser leakage, minimizing ammonia concentration, and determining the optimum reducing agent concentration that will minimize copper corrosion while controlling corrosion and flow-accelerated corrosion of carbon and low alloy steels. In high pressure drum boiler and once-through cycles, there is little room for application of alternate reducing agent and volatile amines because of the problems with their thermal decomposition and generation of organic acids and CO2. Layup chemistry and chemical dosing and disposal of the layup solutions must also be considered in this step. See Appendix F. Thus in this step, a series of tests should be performed to determine the optimum reducing agent levels. The tests should use the monitoring system instrumentation (Step 2) while lowering the reducing agent dosage. Particular note should be made of dissolved oxygen levels, quantities of corrosion products, and the oxidizing/reducing potential (ORP). The latter is particularly important as a reducing environment is required for the copper containing alloys. The choice of reducing agent is also involved here. This includes safety as well as optimized reaction rates and concern for the alternate behavior of any thermal decomposition products and their effects in other parts of the cycle. Although the surface oxide on copper can start to change within ten hours after losing a reducing environment, attention should be paid to the fact that there can be a long time (maybe six months) between changes in feedwater chemistry (ORP) and measured changes in corrosion product transport at the economizer inlet. Thus careful planning is needed for accurate extended tests. Step 4—Design and Material Changes

After a number of years of operation with copper feedwater heaters, with perhaps non-optimum chemistry such as high air in-leakage, high oxygen and oxidizing ORP, many organizations are realizing that the heaters should be replaced. The most common replacement material is stainless (14) steel, but all material options are thoroughly discussed in the EPRI Copper Guideline. Such changes are likely to need detailed economic analysis. EPRI has also developed the tools to assist in these analyses.(5)

2-20


Step 5—Operation

Operational changes can be implemented which are not very costly and can reduce water and steam chemistry problems, such as generation and transport of copper alloy and carbon steel corrosion products. Other changes can be applied to layup practices and to system fill and startup. Examples of operational changes include more frequent and longer use of condensate polishers, and filling with deoxygenated makeup water (Appendix A). Again many of these (14) solutions are discussed in the Copper Guidelines. Each suggested operational change should be tested with the monitoring/instrumentation as in Steps 2 and 3. Step 6—Monitoring to Compare with Baseline Data

This step is aimed at verifying the effectiveness of the implemented solutions to water chemistry and corrosion problems. This water chemistry monitoring and determination of the new cycle chemistry transport characteristics should cover base load, load variations, startup and shutdown, and it should include all the parameters in the water and steam chemistry guidelines. Special instrumentation and sampling could be used to ensure that the corrective measures are effective. There should also be a longterm monitoring plan which includes equipment inspections and nondestructive testing. Step 7—Normal Operation and Monitoring

When Step 6 results are satisfactory, indicating trouble-free water chemistry and operation, then normal operational monitoring resumes. However, there should be periodic reviews of water chemistry and equipment scale and corrosion problems, and chemical transport studies should be performed every two to three years to ensure that equipment aging and other changes do not result in problems. The water chemistry and operating manuals should be updated to reflect the changes resulting from the optimization. If the feedwater copper level at the economizer inlet exceeds the guideline value (< 2 ppb) during normal operation, then attention has to be drawn to maintaining reducing chemistry. ORP in the range of –300 to –350mV has to be maintained not only during the steady state operation, but also during load changes, cycling and peaking operation, and startup and shutdown. Here careful control of air in-leakage, optimized reducing agent dosing into the feedwater upstream of the LP heaters to ensure a reducing ORP at the deaerator inlet, and feedwater oxygen of less than 5 ppb are particularly important.

2-21


Step 8—Continual Check of Chemistry

The continual check on the optimized treatment (Step 8) and the core parameters will determine whether fine tuning is necessary (Repeat of Steps 3 and 6). Step 9—Longterm Plans

Based on the effectiveness of the solutions of water chemistry and corrosion problems implemented in a specific unit and on the equipment life predictions, i.e., feedwater heater and condenser tube deterioration, then longterm plans can be formulated. This can include planning for ultimate replacement of copper alloys, retrofit of condensate polishers, nitrogen sparging and blanketing of the storage tanks used to fill the system during startup (Appendix A). Again all these are discussed in detail in the Copper Guidelines.(14)

2.3 SELECTION AND OPTIMIZATION OF DRUM BOILER WATER TREATMENT It is assumed that the boiler will be supplied with feedwater which has already been optimized according to one of the previous two sub-sections. The previous discussion on the Continuum of Treatments in Section 2.1 indicated that the ability of a unit to produce feedwater of a certain purity has a direct influence on the choice of boiler water treatment (OT, AVT, PC, CT), and that the choice largely reflects the tolerance to ingress of impurities, cooling media, out of specification operation, type of boiler, maximum heat flux, and the amount of blowdown required. As also discussed in Section 2.1, the initial choice of boiler water treatment and its continual assessment has a direct influence on the performance, availability and reliability of the unit. Figure 2-8 provides a road map of the steps that an organization should take to determine the optimum boiler water treatment for each specific unit.

2-22


Figure 2-8 Road Map for Selecting or Optimizing Boiler Water Treatment for Drum Boilers

2-23


Step 1—Review Normal or Current Treatment

Here the first part of this step is to determine if the current boiler water treatment is adequate or whether the unit should be converted to another treatment. This can be performed by answering some of the following questions: •

Historically has the feedwater cation conductivity most often been below 0.2 µS/cm? This might indicate that serious condenser leakage over a period of time has not been of concern. The answer here is very important for assessing the approach in Step 10.

•

Can the boiler water meet the cation conductivity requirements?

•

Does the boiler have a consistent carryover problem? Has it been measured within the last six months?

•

Does the unit have a condensate polisher?

•

Does the unit have a satisfactory monitoring system with the ability to monitor the essential “core parameters” as indicated in the cycle chemistry diagrams of Sections 4 and 5, and in Table 3-2? If this assessment is being conducted for drum boilers on AVT then the EPRI (17) AVT guideline should be referenced for the appropriate core level of instruments .

The answers to these questions will be needed in the other steps of the road map. The second part of this step will be accomplished by a complete review of the current and past experience with the current treatment system (phosphate, CT or AVT) to determine if an alternate treatment should be considered. Note that, for boiler water, AVT may be either AVT(O) or AVT(R), as determined during efforts to select and optimize the feedwater treatment. In many cases, the operating experience will be good, and indicate that only minimal chemical control problems have been experienced, that no BTF in the waterwalls relating to waterside problems have occurred in the last five years, that no turbine deposition or blade failure problems have occurred, and that the feedwater is operating in the optimum fashion with minimum levels of feedwater corrosion products (Fe <2 ppb, Cu <2 ppb). In such cases of good experience, maybe no changes need to be made. However, it is suggested that the road map be reviewed as there may be considerable economic savings to be gained. In other cases, some problems may have occurred or still be occurring, which necessitate a more thorough review and optimization. Information is provided in this step for units with boilers currently operating with a phosphate treatment or CT, and for units operating on AVT. For Units with Drum Boilers Currently Operating with a Phosphate Treatment or CT. The initial review involves operating experience with the current phosphate treatment and in particular, determination of whether phosphate hideout and return occurs, and whether excessive sodium carryover occurs. Controlling carryover with CT is particularly important to minimize damage to the steam circuits.

2-24


Do phosphate hideout and return occur, together with pH instability during transients, shutdown and startup? Another question is “has the boiler experienced any Boiler Tube Failures (BTF) in the last five years, and have any of these been water-side related?” Here the relevant failure mechanisms are (15) hydrogen damage, caustic gouging, acid phosphate corrosion, or corrosion fatigue. The last is important for AVT units as well, as it has been shown that pH depressions, which can occur during shutdown, load reductions, and during the early startup period can exacerbate the corrosion fatigue process. Another question is whether any deposits have been observed in the turbine. Particularly relevant is the deposition of copper-based materials in the HP turbine, or of sodium compounds in the HP and IP turbines; either of which may have resulted in serious capacity and efficiency losses. If phosphate deposits have been found, then boiler drum level control and boiler carryover problems might exist, which need to be addressed prior to optimizing the current treatment or converting to PC or CT. A review of condenser in-leakage and any makeup water contaminant excursions should also be made. If the answer to any or some of these questions is positive, implying that the current phosphate or caustic treatment is not providing adequate chemical control, then the other road map steps should be followed. For Units with Drum Boilers Currently Operating on AVT. There are also a large number of drum units operating on AVT and it is advisable to review the operating experience to date and to answer the following questions. The AVT may be set up incorrectly and/or the boiler may need a solid alkali treatment (PC or CT). Alternatively the boiler may be operating very successfully on AVT and may be suitable for conversion to oxygenated treatment (OT). •

Has the unit experienced any of the problems relating to the generation, transport and deposition of feedwater corrosion products (Table 2-3)?

•

Does the boiler water experience pH depressions during the shutdown and early startup periods which require the addition of either phosphate or caustic?

•

Are chemicals other than trisodium phosphate or sodium hydroxide added to the boiler water for startup or in response to contaminant ingress?

•

Does the feedwater cation conductivity often and significantly rise above 0.2 µS/cm?

•

The questions on BTF, turbine problems and condenser and makeup water contaminants in the section above for units on phosphate and caustic treatment are also applicable here.

If the answer to any or some of these questions is positive, implying that the current AVT is not providing optimum chemical control, then the other road map steps should be followed starting with Step 3. Alternatively if the answers are negative then the unit may be suitable for conversion to OT starting at Step 7.

2-25


Step 2—Continue Use of Current Treatment

If the review in Step 1 indicates no problem or no economic benefit in changing or optimizing the current boiler water treatment, then it is appropriate to continue using the current treatment. Step 3—Base-Line Monitoring on Current Chemistry

If the review in Step 1 indicates that consideration should be given to converting the boiler to another treatment or that the current treatment requires optimizing, then Step 3 involves a complete base-line monitoring program to quantify the chemical parameters for the particular unit in question. This program would use the installed chemistry monitoring system, supplemented by monitoring those parameters in the cycle chemistry diagrams under “troubleshooting and commissioning.” The monitoring program should pay particular attention to the adequacy of the makeup and chemical feed systems, condenser tightness, air ingress, and corrosion product transport. This monitoring involves taking a “thumb-print” of the unit under “typical operating conditions” to identify, under controlled conditions, exactly how the unit chemistry is behaving. It may involve a review of the operating chemistry logs, but this often is not satisfactory and it is preferable to adopt a monitoring campaign. Before this campaign is initiated, it is important to review the utility’s chemistry monitoring capability and reliability. This should include Quality Assurance (QA) and Quality Control (QC) of existing and normally utilized analytical chemistry monitoring and analysis methodology and equipment (see Appendix E). The monitoring campaign should include: •

Varying Operating Conditions—base load, startup, shutdown, over-pressure, etc.

•

Boiler Water Chemistry with Blowdown Open and Closed—sodium, phosphate, free hydroxide, ammonia, chloride, sulfate, silica, pH, cation conductivity, Fe, Cu, carryover.

•

Phosphate Balance and Hideout.

•

Steam Chemistry—cation conductivity, sodium (as a measure of both salts and free hydroxide), phosphate, chloride, silica, sulfate, and boiler carryover of these chemicals.

•

Feedwater Chemistry—cation conductivity, specific conductivity, corrosion products (Fe, Cu), oxygen, pH, oxidizing-reducing potential (ORP) in mixed-metallurgy units.

•

Operation of Condensate Polishers.

Mechanical carryover (moisture) and chemical volatility in the steam must be measured by monitoring total carryover. Concern must also be given to any contaminants that may be contributed to the steam by feedwater utilized for steam temperature control (attemperation). The feedwater treatment should be reviewed as discussed in the previous two subsections. The outcome from this step should be a quantification of those problems recognized in Step 1.

2-26


Step 4—Initial Evaluation of Boiler Water Treatment Options

For boilers currently operating with a phosphate treatment (EPT, PT, or even CPT) the basic questions to be answered here are whether classic phosphate hideout is occurring, whether this is causing any boiler tube corrosion problems, or whether any BTF have occurred. The information collected in Steps 1 and 2 should answer these questions. Hideout could be indicated by pH and phosphate changes during the shutdown or load reduction period, and corrosion may be indicated by an increase of boiler water iron, for boilers operating with CPT. However, it should be noted that units, which only have tri-sodium phosphate injected, can experience a pH elevation together with a phosphate elevation during the shutdown or load reduction period. Hideout can also occur on boilers operating with CT, but it is usually less pronounced than those operating with PC. So for boilers operating on any previous phosphate treatment (EPT, PT or CPT in Figure 1-6), then at this stage in the roadmap two routes can be taken (A or B) based on the analysis in Steps 1–3. Operators/Chemists that should initially consider Route A include those with boilers operating: •

On EPT with contaminant levels exceeding the EPT boiler control curves (see for instance Figure 3-4).

•

On CPT experiencing phosphate hideout and/or adding mono- or di-sodium phosphate to the boiler.

•

On PT with no problems other than maybe phosphate hideout.

Operators/Chemists that should initially consider Route B include those with boilers operating: •

On EPT with contaminant levels not generally exceeding the EPT boiler control curves.

For boilers operating on AVT or CT there are also two cases: a) Where the current AVT or CT is not adequate (Steps 1 and 2) the utility should initially consider changing to an alternate treatment and starting this process at Step 5. b) Where the current AVT or CT is operating satisfactorily without any serious problems, but a utility recognizes that the treatment could be optimized or the AVT converted to OT. Here a utility should start at Step 7 for AVT and at Step 12 for CT. Step 5—Consider Changing to PC(L), AVT or CT

Here the basic direction is determined by the answers to Steps 6 and 10. Most mistakes in choosing, optimizing and changing a boiler water treatment result from not considering these two Steps comprehensively.(1,2) It will directly increase the risk of under-deposit corrosion and BTF.

2-27


Step 6—Is There a Condensate Polisher in the Unit Cycle?

Prior to making a decision on converting to PC, AVT, or CT it is necessary to consider the possible contaminant ingress into the cycle and whether a condensate polisher is available or not. If there is no condensate polisher in the cycle, then it is necessary to review the types of possible contaminant ingress and the contamination events of the unit (Step 10). Alternatively it is preferable to justify the inclusion/addition of a condensate polisher in the cycle. Here, help is (5) available through the cost/benefit approaches developed recently by EPRI. If a condensate polisher is included in the cycle, then the logical conversion route is to AVT or OT (Step 7). However, some utilities might feel that a phosphate boiler chemistry offers a little more buffering capacity and thus could choose the PC(L) route (Step 11). In this case a utility might also want to have a condensate polisher in the cycle in order to meet the strict feedwater specifications. In both cases it is advisable to consider chemically cleaning the boiler prior to conversion (Step 8). This will be particularly important for changes from AVT or OT. Step 7—Convert to AVT or OT

If AVT is selected on this route then the new AVT Guidelines(17) provide the advice, guidance and detailed steps to be followed. Once the AVT chemistry has been selected or it has been decided to optimize the current AVT, then a monitoring campaign should be conducted (Step 17), which can provide direct comparison with the base-line monitoring conducted in Step 3. It is important to note that the optimization of the feedwater chemistry with AVT is very important and particular attention should be given to the choice of AVT(R) or AVT(O) (Section 2.2). An alternative route is for the unit to be converted to OT, and then the OT guidelines provide the (11) necessary detailed guidance. Step 8—Boiler Chemical Clean

It is usually preferable to chemically clean the boiler immediately before conversion to a new treatment. This becomes almost a necessity if considering changing from an older phosphate treatment (such as CPT) to PC or AVT. Those organizations planning on changing from EPT to PC(L) with no history of tube failures should base the chemical cleaning decision on the normal criteria based on deposit levels. The EPRI Chemical Cleaning Guidelines(16) provide the relevant information.

2-28


Step 10—Need to Determine the Likelihood and Frequency of Cycle Contaminant Events

This most important step is the one most often not considered when changing treatments. Very often plants have continued to operate with the older EPT with very low levels of phosphate (0.1–0.2 ppm), or AVT, either with no condensate polisher and/or seawater cooling, or where, as the unit has become older, condenser leaks occur more frequently. In any of these cases there is clearly a need to move to a more resistant boiler water treatment, at least until the condenser can be retubed and/or a condensate polisher is retrofitted to the unit. This step helps in deciding whether to convert to PC(L) or PC(H) or CT, or whether to maintain and/or optimize the current treatment. Here there is a need to look at the past operating chemical records (Step 1) and gauge the seriousness of any contaminant ingress. One way of doing this is to determine how closely the unit has met the guideline limits for feedwater, and the amount of time, in the past 2–3 years, that the unit has been out of normal target limits. For example, plant personnel might want to consider the cumulative hours equivalent to Level 1 for economizer inlet cation conductivity or boiler water chloride as being acceptable and controllable contamination. Here the boiler water control curves in Sections 4 and 5 will need to be referenced and compared with operating practices. If the contaminant episodes have been very infrequent then the organization might want to consider operating with PC(L) (Step 11). If, however, the contamination is greater, then the organization will need to consider converting to PC(H) (Step 15) or CT (Step 13); in these cases there is a need to develop operating procedures which have closer adherence to the guideline limits. Steps 11, 13 and 15—Convert to PC(L), PC(H), or CT

The necessary detailed guidance to convert a unit to PC(L) and PC(H) is included in Section 4. If a chemist/operator followed Route A in Step 4 and/or the unit is susceptible to frequent feedwater contamination (Step 10), then the utility will want to convert and optimize to PC(H) or CT for the unit. The PC(H) treatment is markedly different from the older CPT and PT, and will be used to allow greater flexibility in phosphate control. It essentially optimizes the chemistry by eliminating the need to chase phosphate hideout and ensures that correct additive chemicals (only trisodium phosphate and sodium hydroxide) are added to the unit. With PC(H) it is also very important to consider changes or potential changes to the unit operation, which could change the heat flux patterns on the walls, or to the thermal hydraulic conditions within the walls. Such situations might include: •

change of burner configuration or operation

•

change of fuel sources

•

changes to overpressure operation

These types of situations could cause hideout to be initiated; thus there is a need to ensure that any hideout is recovered normally, by monitoring following any such changes to unit or boiler operation. If hideout does occur with PC(H), there is no danger of it resulting in acid phosphate corrosion because no mono- or di-sodium phosphate will be added to the boiler.(2) 2-29


Section 5 on the use of caustic treatment, using sodium hydroxide for conditioning the boiler water of drum boilers, covers operating experience, boiler and steam chemistry, corrosion, sampling, on-line monitoring and conversion of plant to caustic treatment. This section should be studied when considering CT. CT avoids the complications with monitoring, chemical control and the hideout that occur with phosphate treatments. However, the concentration of chloride in the boiler water needs to be carefully controlled to prevent hydrogen damage due to acid chloride. (This is also true of any of the phosphate treatments and, particularly for AVT). The concentration of sodium hydroxide should normally be maintained between about 1 ppm and 1.5 ppm, and must not be allowed to exceed 3 ppm, to reduce the risk of caustic gouging in the boiler, and carryover, leading to deposition and a potential risk of corrosion in the steam circuit. The pressure dependence of NaOH concentration for CT is shown in Figure 5-6. It is most important that no boiler drum level control or boiler carryover problems exist. It is important in these steps and in Step 12 to ensure that the high pressure boilers are not constrained by steam purity. Of particular concern is for sodium. Steps 9, 12, 14, and 16—Develop Specific Unit Chemistry Guidelines

By this stage, a utility should have decided on their optimum boiler water treatment. The importance of this step can’t be overemphasized as survey results have indicated that, in the past, the development of unit specific guidelines was performed very infrequently to follow the suggestions in all EPRI Cycle Chemistry Guidelines. Most utilities simply use the target values and action levels and do not customize the guidance into specific unit cycle chemistry guidelines. The results of this step should be unique normal target and action levels for each specific unit. Steps 17 to 19—Monitor to Compare with Baseline Monitoring and to Optimize Treatment Selected

These are very important steps which link with Steps 9, 12, 14 and 16. They are essentially a repeat of Step 3. The same matrix of operating conditions and parameters should be looked at. For the PC and CT treatments, particular emphasis should be directed to monitoring the free hydroxide in the boiler water and sodium in the steam to ensure that any excessive levels are recognized early and will not lead to caustic gouging BTF or to turbine problems. Sections 3.8.1, and 4 and 5 should be extensively referenced during this step. With all treatments, it is essential to monitor and compare with the baseline data (Step 17). If the conditions achieved (Step 18) are in any way unsatisfactory, it is necessary to proceed via the “check loop” and reconsider the method of chemical treatment (Step 19). Such instances may be an increasing number of condenser leaks, or a marked increase in air in-leakage, or a level of sodium in steam which exceeds the EPRI steam limits (Sections 4 and 5). Only when satisfactory chemical operation has been obtained, can “normal operation” (Step 20) be considered to have been achieved. 2-30


Step 20—Normal Operation

At this stage, the unit should be operating with the optimum treatment. For the PC treatments this should involve no phosphate hideout, or at least understandable hideout which is fully under control through unit specific guidelines. There should be no cases where chemical/operating personnel add the wrong chemicals. For all treatments, the utility should have developed a set of guidelines for the chemical control of the drum unit. There should be at least a minimum level of instrumentation on the unit, and the key monitoring points and parameters should be alarmed in the control room. All operating/chemical personnel should know exactly what to do in the case of a contaminant ingress. Contaminant episodes need to be dealt with promptly. The source of contamination needs to be removed as soon as possible, pressure restrictions applied if maximum contaminant levels are exceeded, and boiler water chemistry adjusted as needed to get back into the control limits established for the unit. As mentioned previously for feedwater treatments, it is important to continually re-evaluate the appropriate use of boiler water treatments in individual fossil units. A general review of treatment applicability should be made at six month intervals. Follow-up monitoring is desirable if the general review raises any concerns.

2-31


2.4 REFERENCES 1. R.B. Dooley, “The Relationship Between Cycle Chemistry and Performance of Fossil Plants, Power Plant Chemistry, 2002, 4(6), pp. 320-327. 2. R.B. Dooley and W.P. McNaughton, Appropriate Controls for Phosphate Boiler Water Treatments to Avoid Acid Phosphate Corrosion and Hydrogen Damage, Power Plant Chemistry, 3(3), pp. 127-134: 2001 3. Cycle Chemistry Guidelines for Fossil Plants: Phosphate Treatment for Drum Units, EPRI, Palo Alto, CA: December 1994. TR-103665. 4. International Conference on BTF and HTF, and Inspections. Phoenix, Arizona, November 2001, EPRI, Palo Alto, CA: 2002. 1007347. 5. Justifying Cycle Chemistry Upgrades to Improve Availability, Performance, and Profitability, EPRI, Palo Alto, CA: June 2001. 1001557. 6. Selection and Optimization of Boiler Water and Feedwater Treatments for Fossil Plants, EPRI, Palo Alto, CA: March 1997. TR-105040. 7. B. Chexal, J. Horowitz, R.B. Dooley, et al., Flow-Accelerated Corrosion in Power Plants, EPRI, Palo Alto, CA: 1998. TR-106611-R1. 8. R.B. Dooley and B. Chexal, Flow-Accelerated Corrosion, Paper 347, Corrosion 99, 26–28 April 1999. Also published in International Journal of Pressure Vessel and Piping, 77, pp. 85–90: 2001. 9. Guidelines for Controlling Flow-Accelerated Corrosion in Fossil Plants, EPRI, Palo Alto, CA: November 1997. TR-108859. 10. R.B. Dooley, J. Mathews, R. Pate and J. Taylor, “Optimum Chemistry for ‘All-Ferrous’ th Feedwater Systems: Why Use an Oxygen Scavenger,” Proc. 55 International Water Conference, Pittsburgh, PA, Oct. 1–Nov. 2, 1994. 11. Cycle Chemistry Guidelines for Fossil Plants: Oxygenated Treatment, EPRI, Palo Alto, CA: December 1994. TR-102285. 12. Copper Alloy Corrosion in High Purity Feedwater, EPRI, Palo Alto, CA: November 2000. 1000456. 13. Influence of Water Chemistry on Copper Alloy Corrosion in High Purity Feedwater, EPRI, Palo Alto, CA: October 2001. 1004586. 14. Guidelines for Copper in Fossil Plants, EPRI, Palo Alto, CA: November 2000. 1000457.

2-32


15. R.B. Dooley and W.P. McNaughton, Boiler Tube Failures: Theory and Practice, EPRI, Palo Alto, CA: 1996. TR-105261. 16. Guidelines for Chemical Cleaning of Conventional Fossil Plant Equipment, EPRI, Palo Alto, CA: November 2001. 1003994. 17. Cycle Chemistry Guidelines for Fossil Plants: All-Volatile Treatment, EPRI, Palo Alto, CA: November 2002. 1004187.

2-33


3 PHILOSOPHY FOR GUIDELINE AND RATIONALE FOR SAMPLE POINTS, ACTION LEVELS AND TARGET VALUES

3.1 INTRODUCTION These guidelines are designed to assist organizations in improving the operating efficiency and in reducing the losses in boiler and turbine availability influenced by the cycle chemistry by accomplishing the following objectives: •

selecting and optimizing the boiler water and feedwater (Section 2),

•

minimizing the ingress of corrodents,

•

minimizing corrodent concentrations, and

•

reducing the generation and transport of corrosion products.

The rationale of the guidelines in achieving these objectives is based upon an understanding of the effects of impurities on cycle component materials under stress and heat transfer conditions, and recently developed knowledge of chemical transport from the sources of the impurities throughout the water and steam cycle. Figure 3-1 identifies sources of impurity ingress, corrosion and deposition areas around a typical drum boiler cycle. These guidelines suggest target values and action levels for drum boiler units for all key cycle contaminants as a continuum over a broad range of operating pressures. Each of the solid alkali treatments applicable to drum boilers is addressed including those employing trisodium phosphate and (as needed) sodium hydroxide, namely the Phosphate Continuum (PC) (covered in Section 4) and those utilizing only sodium hydroxide, known as Caustic Treatment (CT, covered in Section 5). The guidelines summarize all sample points, monitoring and control parameters, target values and action levels; and feature a consistent, methodical rationale that relates to all sample points, target values and action levels for each parameter to the component(s) most affected by that parameter.

3-1

EPRI Licensed Material Philosophy for Guideline and Rationale for Sample Points, Action Levels and Target Values

Figure 3-1 Typical Locations of Impurity Ingress, Corrosion and Deposition in a Drum Cycle

The guidelines include target values and action levels for significant contaminants at critical sample points. In addition, the criteria for establishing the target values include: •

the optimal steam and water chemistry for protecting cycle components (boiler, turbine, and pre-boiler components) from deposition and corrosion, based on EPRI’s research;

•

limits of detection based on the capability of currently available on-line analyzers and grab sample analysis methods;

•

achievability with current operating units and water treatment; and

•

practicality, based on overall economic benefits without imposing undue burdens on operations.

This solid Alkali Treatment Guideline is a revision of the original Phosphate Treatment Guidelines(1), and of a report on worldwide experience with sodium hydroxide treatment(2). It is based on the results of over 10 years of EPRI research and as such there are a number of key changes and philosophies to the guideline values and the approach. Thus the next two subsections have been added to outline the differences between the previous approach (Section 3.2) and the current approach (Section 3.3).

3-2


3.2 BACKGROUND DERIVATION PHILOSOPHY FOR PREVIOUS EPRI GUIDELINES Immediately, it must be emphasized that this subsection on the philosophy used to derive guideline limits in all previous EPRI Guidelines(1-5), is only provided here for historic perspective and for comparison with the new approach, which was used to develop the revised AVT Guidelines(6) and will be used for these PC and CT Guidelines as outlined in Section 3.3. Users of these combined PC and CT Guidelines are advised that the material presented in Section 3.2 should not be adapted or applied to fossil plant units. During development of the initial (interim consensus) EPRI Cycle Chemistry Guidelines(4), a review of the sensitivities of cycle components to water steam purity determined that limiting the concentrations of ionic contaminants throughout the cycle to levels consistent with the levels in the steam tolerable by the turbine, adequately protected the boiler and other cycle components. This philosophy formed the basis for the derivation of the initial specific guidelines for Allvolatile Treatment (AVT)(3), Phosphate(1) and Oxygenated Treatments(5). This rationale for fossil cycle water and steam chemistry control required that overall cycle chemistry must satisfy the turbine requirements for steam purity. The basic rule was that the concentration of a molecular impurity in superheated steam should not exceed its solubility anywhere in the turbine. The solubility was considered to be lowest just before the saturation line. The target values for sodium and chloride were derived from the limits for hydroxide and chloride in the sodium or ammonium forms depending on the treatment. The process to derive boiler water contaminant limits for sodium, chloride, sulfate and silica started with the allowed turbine steam composition (based on the solubility limits). By considering mechanical and vaporous carryover, the allowable impurity concentrations in boiler water could be determined. Both mechanical (Figure 3-2) and vaporous (Figure 3-3) carryover are pressure dependent and, therefore, the target values subsequently derived for boiler water ionic contaminants were also pressure dependent. Figure 3-2 is an upper-limit for mechanical carryover. A safety factor of 2 was included in Figure 3-2 to provide a conservative estimate of the amount of mechanical carryover. It was stressed as being very important for the user to determine this limit for each drum unit as actual unit mechanical carryover is unit specific. Figure 3-3 is commonly referred to as the “ray diagram”. This was originally developed by (7,8) Styrikovich and Martynova in the early 1960s. It has been used in all previous EPRI guidelines except the recently published AVT Guidelines(6), to calculate the vaporous carryover. Smooth, continuous relationships of boiler water target values as a function of pressure were developed for drum boilers. This approach, which was originally developed by EPRI in the (4) Interim Consensus Guidelines for pulverized coal-,oil-, and gas-fired units , also allows units having the capability of changing load via sliding pressure operation to benefit from relaxed boiler water chemistry limits when reducing load by reducing pressure. Target values for feedwater chemistry in drum boilers were made comparable to those for steam chemistry to minimize the cost of blowdown and also maximize the availability of blowdown as a corrective action. This rationale gave operators the maximum amount of response time to take corrective action when upsets occur and is practical because of the significant cost of the heat lost in blowdown.

3-3


Feedwater chemistry target values also reflected the need to control feedwater corrosion by optimizing the dissolved oxygen concentrations and pH. It is now clear that ORP is the critical parameter in the feedwater. 4.8

6.2

7.6

9.0

Drum Pressure (MPa) 10.3 11.7 13.1 14.5

15.8

17.2

18.6 19.6

0.3

Mechanical carryover (%)

0.2

0.1 0.09 0.08 (Note: This curve includes a safety factor of 2)

0.07 600 700

900

1100 1300 1500 1700 1900 2100 2300 2500 2700 2850 Drum Pressure (psig)

Figure 3-2 Representative Drum Boiler Mechanical Carryover used to Derive Generic Cycle Chemistry Limits; Actual Carryover in Drum Boilers Must be Determined in the Field as There are Significant Variations Due to Drum Design, Firing Rate and Drum Operating Level.

3-4


200

400

600

1000

1200

1600

2000

2200

2800 2600

3000

3200

Drum Pressure (psia)

10-0

Fe3O4 Al2O3

10-2

B2O3

NH

4 Cl

, (N

NiO

H

4) 2

SO

CuO

4

10-3

SiO2 10-4 Cu2O

CaSO4 MgO

CaCl2

10-5

Na 4

O4

2S

PO Na 3

Na

10-6

BaO

LiCl

Cl

Concentration In Vapor/Concentration in Liquid

10-1

NaOH

10-7 22

1

20

2

16

12 8 6 Drum Pressure (MPa)

4

3

6 10 20 60 Density of Liquid/Density of Steam

2

1

100

Figure 3-3 Distribution Ratios for Common Boiler Water Contaminants (This ray diagram was originally developed by N. A. Styrikovich and O. I. Martynova.(7,8))

3-5


It was suggested that the boiler water target values could be recalculated and made to fit individual unit characteristics where the steam purity limits could be modified (for example, in higher backpressure turbines) and where the mechanical carryover had been measured and validated, and differed from the values in Figure 3-2. Consideration of high heat fluxes in drum boilers needed to be included when recalculating boiler water target values. Reducing the target values for boiler water impurities by as much as 50% was recommended for units known to have ”high duty” boilers (such as some oil-fired boilers); future work planned by EPRI in a project called “Program Boiler Corrosion” will assess this need and other aspects of boiler water chemistry, surface deposits, and their effects on corrosion activity. The higher tolerance of non-reheat turbines to impurities in steam was acknowledged in the previous guidelines and this was reflected in the establishment of two sets of guidelines for reheat and nonreheat drum boiler units. 3.2.1 Illustration and Examples of Boiler Water Limits from Previous EPRI Guidelines The following example is provided only to illustrate the historic perspective delineated in Section 3.2. It should not be adapted or applied to fossil plant units. This approach was used in (1) previous EPRI Cycle Chemistry Guidelines including the 1994 Phosphate Treatment Guideline ; (2) this derivation approach was not followed in the earlier report on sodium hydroxide treatment or in the recently issued AVT Guideline(6). Total Carryover The allowable impurity concentrations in boiler water were calculated from those of the steam and the amounts of mechanical and vaporous carryover. Carryover is defined as the ratio of concentration of a chemical specie in the saturated steam exiting the boiler drum to the concentration in the boiler water (usually measured in the blowdown). Total carryover (T) is defined as: T=M+V

(eq. 3-1)

where M is mechanical carryover (with steam moisture) and V is the vaporous carryover. Total carryover is usually measured by analysis of sodium:

T=

3-6

Na in saturated steam Na in blowdown

(eq. 3-2)


Mechanical carryover is a measure of the separation efficiency of the boiler. It can be calculated as: M = T – V*

(eq. 3-3)

where V* is a vaporous carryover of the prevalent sodium salt (either trisodium phosphate or sodium hydroxide depending on the solid alkali treatment in use) identified in a particular boiler at the boiler operating pressure. Volatile carryover of chemical species associated with unknown molecules, Vx, is then calculated as: Vx = Tx – M where Tx = and

(eq. 3-4)

x in saturated steam x in blowdown

M is from eq. 3-3

The following equation can be derived from a mass balance around the boiler and can be used to calculate the boiler water concentrations from the allowable steam concentrations, distribution ratios, and mechanical carryover rate: CB ≤

CS M + k d (1 − M )

(eq. 3-5)

where CB = boiler water target value (ppb) CS = reheat steam target value (ppb) M = mechanical carryover kd = fractional distribution ratio It was strongly suggested in all previous EPRI guidelines that this equation should be used to modify the target values for boiler water chemistry when validated data on a specific plant’s mechanical carryover rates were collected during commissioning and periodically thereafter and found to be substantially different from those in Figure 3-2. The equation does not include the effects of attemperation water, which will be negligible when the economizer inlet chemistry is maintained at a value equal to or better than the reheat steam chemistry as recommended in the guidelines.

3-7


Illustration The following example illustrates the procedure for calculating the boiler water target values. Example: Calculate the boiler water sulfate target values and action levels for a plant, using mechanical carryover rate data from Figure 3-2. This plant has a 2500 psi (17.2 MPa) drum boiler with reheat operated on PT. Solution: Calculate the target value and action level for operation at full load, 2500 psi (17.2 MPa) at which the mechanical carryover from Figure 3-2 is 0.00185.

1. Target value of sulfate in steam is 3 ppb based on solubility. 2. The distribution ratio of sodium sulfate at 2500 psi (17.2 MPa) from Figure 3-3 is 2.3 x 10-6 (0.0000023). 3. The target value of sulfate in the boiler water using equation 3-5: CB =

3 ppb 0.00185 + 0.0000023 (1 − 0.00185)

= 1600 ppb (1.6 ppm) Action Level 2 was derived by multiplying the target value by a factor of 2. Action Level 3 was derived by multiplying the Action Level 2 by a factor of 2.

4. Target values at 2500 psi (17.2 MPa): Normal: Action Level 1: Action Level 2: Action Level 3:

≤ 1.6 ppm > 1.6 ppm ≤ 3.2 ppm > 3.2 ppm ≤ 6.4 ppm > 6.4 ppm

A similar process was used for other boiler water contaminants. By using sodium chloride as the volatile specie for chloride and its distribution ratio from the ray diagram, the PT boiler water target values for chloride were the same as for sulfate based on the sodium sulfate salt. Figures 34 and 3-5 show the boiler water chloride target values and action levels as a function of operating pressure for EPT and PT chemistry. These curves were derived as described above using the generic mechanical carryover curve (Figure 3-2). (1) The 1994 EPRI Phosphate Treatment Guidelines derived a family of control curves for EPT and PT. The curves for EPT (such as Figure 3-4 for chloride) assumed that chlorides and sulfates would be present as the more volatile (as predicted by the ray diagram) ammonium salts. Thus, limits for these parameters in boiler water on EPT were much lower than for units operated on PT where the volatility of chloride and sulfate was primarily considered as their respective sodium salts (as shown in Figure 3-5 for chloride). In fact the EPT limits were identical to the

3-8


limits derived for AVT chemistry(3). The report on CT(2) did not apply the derivation procedure used for the Guidelines for other chemistries. Drum Pressure (MPa) 3.0

4.8

6.2

7.6

9.0

10.3

11.7

13.1

14.5

15.8

17.2

18.6 19.6

2500

2700 2850

2.0 1.5 1.0 0.9 0.8 0.7 0.6 0.5 0.4

Chloride (ppm Cl)

0.3

Action level 3

0.2 Action level 2 Action level 1

0.10 0.09 0.08 0.07 0.06

Normal

0.05 0.04 0.03

0.02

0.01 600 700

900

1100

1300

1500 1700 1900 Pressure (psia)

2100

2300

Figure 3-4 Equilibrium Phosphate Treatment (EPT): Drum Boiler Water Chloride vs. Operating Pressure. Calculated using mechanical and vaporous carryover for a limit of 3 ppb Chloride in Steam. The values were identical for sulfate. It must be noted that these are the old EPT limits derived in the previous Phosphate Treatment Guidelines(1) and should no longer be used. The new PC(L) limits are discussed in Section 4.

3-9


Figure 3-5 Phosphate Treatment (PT): Drum Boiler Water Chloride vs. Operating Pressure. Calculated using mechanical and vaporous carryover for a limit of 3 ppb Chloride in Steam. It must be noted that these are the old limits derived in the previous Phosphate Treatment Guidelines(1) and should no longer be used. The new PC(H) limits are discussed in Section 4.

3-10


3.3 BACKGROUND TO OVERALL DERIVATION OF PHILOSOPHY FOR PHOSPHATE CONTINUUM AND CAUSTIC TREATMENT GUIDELINES As discussed in Section 1.3.1, much of the philosophy used in the previous EPRI Guidelines is still very much applicable. However, based on operating experiences, availability problems (such as boiler tube failures) and over 10 years of EPRI research, a number of new concepts and philosophies are included in this revision. The new guidelines retain the concept of smooth continuous relationships of drum boiler water contaminant values. This may allow units to have the capability, during changing load (reducing pressure), of benefiting from relaxed boiler water limits. EPRI has conducted much work over the last 13 years to understand the phenomena of nucleation, early condensate and liquid films in the phase transition zone (PTZ) of the low (9-13) Deposition studies were pressure steam turbine, as outlined in Sections 1.2.1 and 1.2.2. conducted by 23 members of an International Collaboration using a converging-diverging nozzle to reproduce conditions in the PTZ. The results of this work clearly indicated that to minimize deposition, the steam limits for sodium, chloride and sulfate should be reduced from the levels indicated in the earlier Guidelines; Table 3-1 summarizes the changes in steam purity limits resulting from this work, which have now been incorporated into these revised guidelines, as well as those used to revise the AVT Guidelines(6) and that will be used to revise the OT Guidelines in 2004. This table is in full agreement with the overall concept of the Continuum of Treatments discussed in Section 2.1 and shown in Figure 2-1. Table 3-1 EPRI Steam Purity Limit Revisions Based on PTZ Research Findings Cycle Chemistry

Steam Purity Parameter OT/AVT

PC(L)

CT

PC(H)

Na, Cl, SO4, ppb

2/2

2

2

3

Cation Conductivity, µS/cm

0.15/0.2

0.2

0.2

0.3

Visits to hundreds of plants and benchmarking of the chemistry programs in use has indicated that very few utilities around the world measure and/or know the percentage of mechanical carryover on their drum units. This observation, combined with the inaccuracies of the ray diagram, has led to questions about the continued use of the approach followed in the previous Guidelines of deriving boiler water limits from steam solubilities using mechanical and vaporous (ray diagram) carryover. Thus EPRI initiated a detailed study of partitioning of impurities, salts and oxides into steam, which has now led to a validation of the boiler water limits (see Section 1.2.4). This is discussed in the next section.

3-11


3.3.1 Derivation of Philosophy for Steam and Boiler Water Limits in EPRI’s New Phosphate Continuum and Caustic Treatment Guidelines As indicated in Section 3.2, the concepts of carryover were included in the derivation of boiler water limits from the steam limits in all previous EPRI Guidelines prior to 2002. In order to apply these concepts to a variety of individual systems with widely differing chemical and physical environments (e.g. water chemistry, boiling and condensation temperatures), it is necessary to quantify the extent of partitioning of various solutes between water and steam under phase change conditions (boiling and condensation). Jonas pointed out discrepancies between partitioning data obtained from operating plants and the values predicted from the ray diagram.(14) For almost 50 years the ray diagram has been used to express vaporous carryover and used for boiler water limit derivation. The plant measurements of distribution coefficients for chloride and sulfate were higher by orders of magnitude than those given by the ray diagram for chloride and sulfate. This observation led to the conclusion that both the experimental and theoretical basis for the ray diagram should be investigated in some detail, to include those solutes for which no experimental partitioning data were available, and to develop a more rigorous treatment of solute partitioning between liquid water and steam. This new model for solute carryover could then be applied to the wide range of operating conditions and water chemistry found in fossil plants worldwide. This initiated a series of EPRI sponsored research projects to rigorously measure experimentally the partitioning constants for most of the important compounds, which can be present in boiler water and steam. Over the last ten years, work has addressed the following compounds and related species: NH4Cl, NH3, HCl, NH4HSO4, Na2SO4, NaHSO4, NaOH, H2SO4, organic acids (acetic and formic), Cu(OH)2, H3PO4, NaH2PO4, and Na2HPO4. The results have been reported by EPRI.(15-21) The results of this new thermodynamic approach and the partitioning equations have been summarized in a recent report(21); the overall results were shown in Figure 1-4 and discussed in Section 1.2.4. The partitioning constants (i.e. the concentration in the steam over the activities of the constituent ions in the water) can be illustrated on a partitioning diagram by plotting the logarithm of the partitioning constant against reciprocal absolute temperature (in Kelvin). Separate curves have been derived for each compound and related species: NH4Cl, NH3, HCl, NH4HSO4, Na2SO4, NaHSO4, NaOH, H2SO4, organic acids, Cu(OH)2, H3PO4, NaH2PO4 and Na2HPO4. Some species, NH3, SO2, organic acids, Cu(OH)2 and H3PO4 are neutral, others are 1:1 ionic compounds, NH4Cl, HCl, NaOH, NaCl, NaHSO4, NH4HSO4 and NaH2PO4, while Na2SO4 and Na2HPO4 are 1:2 ionic compounds. The partitioning constants for the neutral species are shown in Figure 3-6, the 1:1 compounds in Figure 3-7 and the 1:2 species in Figure 3-8. Sulfuric acid can be classed as a 1:1 compound, + because under the conditions of the experiments, it only ionizes one stage to H and HSO4 . The relative volatilities of each type of species can be compared directly on each graph, but because of the different ways in which 1:1 or 1:2 electrolytes dissociate, neutral compounds can not be compared with 1:1 compounds, or 1:1 compounds with 1:2 compounds. However, it is clear from the graphs that the volatilities of all the ionic solutes increase with temperature.

3-12


20

Tc

300

200

150°C

NH3 (N)

Partitioning Constants

SO2 (N) CH3COOH (N)

10

HCOOH (N)

Log KD

Cu(OH)2 (N)

0

H3PO4 (N)

-10

-20

1.6

1.8

2.0

2.2

2.4

2.6

1000 K/T

Figure 3-6 Partitioning Constants - Neutral Species

3-13

EPRI Licensed Material Philosophy for Guideline and Rationale for Sample Points, Action Levels and Target Values Tc

300

200

150°C

20

Partitioning Constants 10

HCI (1:1) Log KD

0

NH4CI (1:1)

NaCI (1:1) H2SO4 (1:1)

-10

NaOH (1:1) CH3COONa (1:1)

NaH2PO4 (1:1) NH4HSO4 (1:1)

-20

NaHSO4 (1:1)

1.6

1.8

2.0

2.2

2.4

2.6

1000 K/T

Figure 3-7 Partitioning Constants - 1:1 Compounds

Because the data in Figures 3-6 to 3-8 cannot be used simply as with the ray diagram to calculate vaporous carryover, all of this data has been incorporated into a computer code(21). This code allows predictions to be made of the composition of the boiler water, steam and early condensate. Some initial examples were provided where the steam and early condensate composition was predicted from the boiler water composition. The model can predict vaporous carryover alone and with various degrees of mechanical carryover. Equilibrium conditions are assumed to apply all around the water/steam circuit. The model can be used as an excellent check on the validity of any derived boiler water limits. Starting with the new steam limits (Table 3-1) of 2 ppb of sodium, chloride and sulfate for PC(L) and CT, 3 ppb for PC(H), and 10 ppb for silica, the model was applied in the “backwards” direction to calculate the corresponding boiler water values across a range of pH and mechanical carryover under various PC(L), PC(H) and CT regimes to assess the impact of variations in pH and other factors. The curves presented in the following discussions of PC(L), PC(H) and CT represent the most conservative and restrictive limits within the range of normal chemistry conditions allowed for each treatment. 3-14


20

Tc

300

200

150°C

Partitioning Constants 10

Log KD

0

-10 Na2HPO4 (1:2)

-20

Na2SO4 (1:2)

1.6

1.8

2.0

2.2

2.4

2.6

1000 K/T

Figure 3-8 Partitioning Constants - 1:2 Compounds

Phosphate Continuum at Low Phosphate Levels (PC(L)) (< 3 ppm) Figures 3-9 to 3-12 show the model predictions for the maximum pressure-dependent concentrations in the boiler water of sodium, chloride, sulfate and silica values to ensure less than 2 ppb of sodium, chloride and sulfate, and 10 ppb of silica in steam. These figures show the effect of assuming mechanical carryover in agreement with Figure 3-2. These are the upper lines in Figures 3-9 to 3-12. The added effect of vaporous carryover predicted with the thermodynamic model as shown by the lower lines in each figure.

3-15


Figure 3-9 Model Predictions for the maximum pressure-dependent concentrations of Sodium in the boiler water. The bottom set of three curves is for conditions typical of PC(L) to ensure less than 2 ppb sodium in steam. The upper set of three curves is for conditions typical of PC(H) to ensure less than 3 ppb sodium in steam. In each set, the top straight line is for mechanical carryover only. The two lower curves are at Na:PO4 = 3 (lower) and TSP + 1 ppm NaOH (higher). It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the control curves in the Guidelines (Section 4).

3-16


Figure 3-10 Model Predictions for the maximum pressure-dependent concentrations of Chloride in the boiler water. The curves are representative of conditions for PC(L) to ensure less than 2 ppb chloride in steam. The top line (1) is for consideration of mechanical carryover only. Line 3 is for Na:PO4 = 3 and includes vaporous carryover. Lines 2 and 4 are for TSP + 1 ppm NaOH and also include vaporous carryover. Line 4 represents the condition where the chloride is limited by the sodium limit. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the control curves in the Guidelines (Section 4).

3-17


Figure 3-11 Model Predictions for the maximum pressure-dependent concentrations of Sulfate in the boiler water. The bottom set of three curves is for conditions typical of PC(L) to ensure less than 2 ppb sulfate in steam. The upper set of curves is for conditions typical of PC(H) to ensure less than 3 ppb sulfate in steam. In each set, the top line is for mechanical carryover only. The two lower curves are at Na:PO4 = 3 (lower) and TSP+ 1 ppm NaOH (higher) and include vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the control curves in the Guidelines (Section 4).

3-18


Figure 3-12 Model Predictions for the maximum pressure-dependent concentrations of Silica in the boiler water to Ensure Less than 10 ppb Silica (as SiO2) in Steam. The curves are representative of PC(L), PC(H) and CT. Curve 1 is for mechanical carryover only. Curve 2 is for mechanical and vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the control curves in the Guidelines (Section 4).

3-19


The trends for sodium, chloride and sulfate with PC(L) indicated in Figures 3-9 to 3-11 are quite similar to those derived for control of boiler water sodium, chloride and sulfate on PC(H); differences in relative limit values predicted by the model are nearly in the same proportion as the allowable steam purity limits for the two chemistry treatments. With respect to sodium (Figure 3-9), the major feature is that mechanical carryover is the decisive factor. Thus there is a need for each operator/chemist to correctly maintain and operate the separation devices and know the total carryover characteristic for each drum unit. Sodium control in the boiler water is extremely important when using PC(L) as a significant fraction of the sodium in the steam could be present as sodium hydroxide. The figure also shows that volatility starts to influence the sodium limit only at about 2500 psi (17.2 MPa). The limits of the thermodynamic model are, due to the constant carryover contribution, more restrictive at lower pressures while at higher pressures the reduced volatility predicted by the model in combination with the lower allowable steam purity limit (2 ppb) results in less restrictive boiler water limits than are predicted using the ray diagram and higher steam purity limit (3 ppb) used to derive the earlier EPT Guidelines. For chloride, Figure 3-10 illustrates that volatility is one of the influencing factors on steam purity at pressures above about 900 psi (6.2 MPa) and becomes progressively more significant as pressure increases. Allowable boiler water chloride levels predicted by the thermodynamic model are considerably higher than would be allowed if volatile carryover is calculated from the ray diagram. For sulfate (Figure 3-11), mechanical carryover appears to be the exclusive influencing parameter at least up to about 2700 psi (18.6 MPa). Allowable boiler water sulfate levels predicted by the thermodynamic model are considerably higher than would be allowed if volatile carryover is calculated from the ray diagram. Figure 3-12 confirms that vaporous carryover is the major contributor of silica in steam. Also, the silica data (Curve 2) show very good agreement with limits derived using the old methodology. Phosphate Continuum at High Phosphate Levels (> 3 ppm) (PC(H)) Figures 3-9, 3-13, 3-11 and 3-12 show the model predictions of boiler water sodium, chloride, sulfate and silica values to ensure 3 ppb of sodium, chloride and sulfate, and 10 ppb of silica in steam. These figures show the effect of assuming mechanical carryover in agreement with Figure 3-2. These are the upper lines in each figure. The added effect of vaporous carryover predicted with the thermodynamic model is shown by the lower lines in each figure.

3-20


Figure 3-13 Model Predictions for Boiler Water Chloride as a Function of Pressure. The curves are representative of conditions for PC(H) to ensure less than 3 ppb chloride in steam. The top line (1) is for consideration of mechanical carryover only. Line 3 is for Na:PO4 = 3 and includes vaporous carryover. Lines 2 and 4 are for TSP + 1 ppm NaOH and also include vaporous carryover. Line 4 represents the condition where chloride is limited by the sodium limit, but does include vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the curves in the Guidelines (Section 4).

3-21


With respect to sodium (Figure 3-9), the major feature, as for PC(L), is that mechanical carryover is the decisive factor. Thus there is a need for each operator/chemist to correctly maintain and operate the separation devices and know the total carryover characteristic for each drum unit. As is the case with PC(L), sodium control in the boiler water is very important when using PC(H) as a portion of the sodium in the steam likely will be present as sodium hydroxide if there is any free sodium hydroxide in the boiler water. The figure also shows that volatility starts to influence the sodium limit only at about 2500 psi (17.2 MPa). The limits of the thermodynamic model are, due to the constant carryover contribution, more restrictive at lower pressures while at higher pressures the reduced volatility predicted by the model in combination with the lower allowable steam purity limit (3 ppb) results in less restrictive boiler water limits than are predicted using the ray diagram and higher steam purity limit (5 ppb) used to derive the earlier PT Guidelines. For chloride, Figure 3-13 illustrates that volatility is one of the influencing factors on steam purity at pressures above about 500 psi (10.3 MPa) and becomes progressively more significant as pressure increases. Allowable boiler water chloride levels predicted by the thermodynamic model are remarkably similar to those calculated using volatile carryover from the ray diagram. For sulfate (Figure 3-11), mechanical carryover appears to be the exclusive influencing parameter at least up to about 2600 psi (17.9 MPa). Allowable boiler water sulfate levels predicted by the thermodynamic model are higher than those calculated using volatile carryover from the ray diagram. Figure 3-12 confirms that vaporous carryover is the major contributor of silica in steam; the control curves for PC(L) and PC(H) predicted with the thermodynamic model are virtually identical. Caustic Treatment (CT) Figures 3-14 through 3-16, and Figure 3-12, respectively, show the model predictions of boiler water sodium, chloride, sulfate and silica values to ensure 2 ppb of sodium, chloride and sulfate, and 10 ppb of silica in steam. These figures show the effect of assuming mechanical carryover in agreement with Figure 3-2. These are the upper lines in each figure. The added effect of vaporous carryover predicted with the thermodynamic model is shown by the lower lines in each figure. To minimize the risk of corrosion and deposition in the turbine with CT, the limits for sodium, chloride and sulfate in steam have all been set at 2 ppb, and those for silica at 10 ppb. The limits for these constituents in the boiler water for CT, calculated using the thermodynamic model, Figures 3-14 to 3-16 and Figure 3-12, respectively, are similar to those derived for PC(L), shown in Figures 3-9 to 3-12. This is because the limits for sodium, chloride, sulfate and silica in steam are the same for both CT and PC(L). Note how the presence of a solid alkali in the boiler water, in this case NaOH, reduces the vaporous carryover of chloride in the steam, compared with AVT (Figure 3-15).

3-22


With respect to sodium (Figure 3-14), the major feature is that mechanical carryover is the decisive factor. Thus there is a need for each operator/chemist to correctly maintain and operate the separation devices, and to know the carryover characteristics for each drum. Sodium control in the boiler water is crucial when using CT as nearly all sodium present in the steam is likely to be present as sodium hydroxide; it is for this reason that an upper operating limit of 2500 psi (17.2 MPa) at the drum is imposed on CT. The figure also shows that volatility starts to influence the sodium limit only at about 2500 psi (17.2 MPa). For chloride, Figure 3-15 illustrates that volatility is one of the influencing factors at pressures above about 1500 psi (10.3 MPa) and becomes progressively more significant as pressure increases. Also, evaluations showed that chloride volatility increases as levels of sodium hydroxide decreased, converging with predictions for AVT chemistry. For sulfate (Figure 3-16), mechanical carryover appears to be the exclusive influencing parameter at least up to about 2500 psi (17.2 MPa). The curves for CT shown in Figures 3-14 to 3-16 extend to 2850 psi (19.6 MPa) so that they can be compared to the corresponding curves for PC. However, it must be noted that CT is only proven up to 2500 psi (17.2 MPa), and thus, the control curves given in Section 5 only extend to 2500 psi (17.2 MPa). As with PC(L) and PC(H), Figure 3-12 confirms that vaporous carryover is the major contributor of silica in steam.

3-23


Figure 3-14 Model Predictions for Boiler Water Sodium as a Function of Pressure. The curves are representative of conditions for CT to ensure less than 2 ppb sodium in steam. Curve 1 is for mechanical carryover only. Curve 2 is for mechanical and vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the curves in the Guidelines (Section 5).

3-24


Figure 3-15 Model Predictions for Boiler Water Chloride as a Function of Pressure. The curves are representative of the conditions for CT to ensure less than 2 ppb chloride in steam. Curve 1 is for mechanical carryover only. Curves 2 are for mechanical and vaporous carryover with various levels of NaOH (0.5, 1.0 and 2 ppm). Curves 3 are for mechanical and vaporous carryover for AVT conditions at different ammonia levels (pH 9, 9.2, 9.4 and 9.6). It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the curves in the Guidelines (Section 5).

3-25


Figure 3-16 Model Predictions for Boiler Water Sulfate as a Function of Pressure. The curves are representative of conditions for CT to ensure less than 2 ppb sulfate in steam. Curve 1 is for mechanical carryover only. Curve 2 is for mechanical and vaporous carryover. It should be noted that the curves in this figure are not to be used as boiler control limits. The curves are only to verify and illustrate the curves in the Guidelines (Section 5).

3-26


Summary Discussion of Derivation Process for Control Curves for PC and CT The boiler water limits for sodium, chloride, and sulfate predicted by the thermodynamic model for PC(L), PC(H), and CT are not intended for use as control diagrams in fossil plant units. The new boiler water control curves for sodium, chloride and sulfate are provided in Sections 4 (PC) and 5 (CT). The limits for sodium are pressure dependent with steam purity concerns becoming more likely at high operating pressures. The limits for chloride and sulfate in boiler water (Sections 4 and 5) established to prevent corrosion damage are lower than those derived using the thermodynamic model and the steam limits. Thus, the limits for chloride and sulfate in the boiler water are defined by the need to prevent initiation of corrosion in the boiler, rather than the need for limiting the concentration in steam. In developing these guidelines it was recognized that not all plants would be able to routinely monitor boiler water sodium and that even fewer would be equipped to monitor chloride and sulfate for chemistry control purposes. At those plants where these parameters can be checked, the new control curves of Sections 4 and 5 can be applied and potentially customized if reliable drum carryover data are available. In cases where these parameters are not routinely monitored, the curves are still useful for problem evaluation and assessment of possible corrosion damage. (21) Thus in summary, the results of the partitioning work , Figures 3-6 to 3-8, and the predictions from the volatility model (Figures 3-9 to 3-16) indicate that it is possible to decouple the derivation of boiler water limits from steam limits as was used initially in EPRI’s revised AVT Guidelines(6). It should be noticed that Figures 3-10 and 3-11 (for PC(L)), 3-13 and 3-11 (for PC(H)) and 3-15 and 3-16 (for CT) all clearly show that there is usually quite a difference between the boiler water curves for chloride and sulfate based on the approach employing volatilites indicated on the ray diagram (Figure 3-3) and mechanical carryover from Figure 3-2 used in earlier EPRI Guidelines, and the boiler water curves presented in Figures 3-9 to 3-16, derived using the thermodynamic model and the mechanical carryover data from Figure 3-2. Thus the new steam limits of 2 ppb of Na, Cl and SO4, for PC(L) and CT, 3 ppb for PC(H) and 10 ppb of SiO2 determined from EPRI’s PTZ research should not be limiting in providing boiler water corrosion protection for either PC or CT. The only exception may be for sodium with PC, and this will require careful monitoring as indicated in Section 3.8.1 Also the volatility model predictions indicate much higher levels of boiler water chloride and sulfate are required in most cases to raise the steam values above the new limits, and that these limits are generally greater than would be required to protect the boiler against corrosion. Thus the development of boiler water cation conductivity curves has been used to establish boiler water limits that provide boiler waterwall corrosion protection. This is thoroughly discussed in Section 3.8.12 for both PC and CT. Thus these new solid alkali guidelines will have steam limits to protect the turbine based on PTZ research findings (Table 3-1), and boiler water limits to protect the boiler (Sections 4 and 5); the boiler water limits will not, under normal operating circumstances, allow excedance of the steam limits.

The curves for silica obtained from the thermodynamic model are similar to those already in use and will therefore be used for boiler water silica limits in these revised guidelines.

3-27


3.4 RATIONALE FOR SAMPLE POINTS Recommended sample points were developed based on the optimal cycle chemistry requirements, and the rationale for each sample point is discussed in this section. The parameters to be monitored fit into two categories, as shown on the cycle chemistry diagrams in Sections 4 (PC) and 5 (CT): a) those parameters which all fossil plants should have for optimum chemistry control (core parameters) and b) those parameters which are regarded as diagnostic for troubleshooting or commissioning. See Table 3-2 for a listing of core instruments for units operating on alkali solid treatments. The core parameters are considered the minimum level of surveillance needed for all units. In general, use of on-line analyzers for continuous analysis of chemistry is preferred. However, it is recognized that certain parameters such as iron, copper, sulfate, total organic carbon and chlorides (usually) will require the procurement and analysis of grab samples. Primary chemistry control for these units is through the use of cation and specific conductivity. Conductivity instrumentation is reliable, relatively maintenance free and provides an early indication of cycle contamination and increased carryover. Specific conductivity also provides an accurate indication of ammonia levels in the feedwater. For units on PC, conductivity needs to be used in conjunction with pH and phosphate measurements. Target levels have been established for sodium, chloride, silica, sulfate and organics in steam. Limits for these parameters were set to minimize the risk of deposition and corrosion in turbines, while ensuring that the limits provide adequate corrosion protection to the boiler. In drum units, it is important to avoid levels of mechanical and vaporous carryover, which will exceed these target values. The routine measurement of carryover (every six months) is considered of such importance that it is included as a core parameter.

3-28

EPRI Licensed Material Philosophy for Guideline and Rationale for Sample Points, Action Levels and Target Values Table 3-2 EPRI’s Core Monitoring Parameters and/or Minimum Level of Continuous Instruments for All Units Operating on Phosphate Continuum and Caustic Treatments

Oxidizing AVT (AVT(O)) Feedwater and PC or CT in the Boiler Water (Drum units without a reducing agent and with an all-ferrous feedwater system) Cation conductivity Specific conductivity pH Dissolved oxygen Sodium Phosphate Air in-leakage Total Carryover

CPD, CPO or EI, RH (or MS), Blowdown Makeup, Blowdown Blowdown CPD, EI CPD, CPO or EI, Blowdown, RH (or MS) Blowdown (PC only)

Reducing AVT (AVT(R)) Feedwater and PC or CT in the Boiler Water (Drum units with a reducing agent and with a mixed-metallurgy or all-ferrous feedwater system) Cation conductivity Specific conductivity pH Dissolved oxygen Sodium Phosphate ORP Air in-leakage Total Carryover

CPD, CPO or EI, RH (or MS), Blowdown Makeup, Blowdown Blowdown CPD, EI CPD, CPO or EI, Blowdown, RH (or MS) Blowdown (PC only) DAI

Using the approach in Section 3.3, contaminant control curves were developed for drum units. These curves are presented in Sections 4 and 5. By confirming these values with mechanical and vaporous carryover (partitioning) modeling it has been possible to stress the importance of steam carryover and boiler drum pressure to provide the concentration of the contaminant of interest. 3.4.1 Reheat Steam/Superheated Steam Factors affecting steam chemistry, other than the mechanical and vaporous carryover from the boiler, include the following:

•

contamination of steam in superheaters and reheaters by attemperation water, and

•

precipitation of impurities as deposits in the superheater, reheater, and in the turbine due to changes in steam temperature and pressure.

Monitoring of key contaminants at this sample point indicates the actual impurity levels in the steam and indicates whether the turbine blades are protected against deposition and/or corrosion. This monitoring also verifies compliance with the turbine manufacturer’s guarantee condition. 3-29


Should this sample point not be available, the steam chemistry may be calculated from the chemistries of the saturated steam and feedwater, accounting for any impurity ingress associated with the superheater and reheater attemperating water (feedwater). Steam sampling requires special techniques, which are described in an EPRI Report(23) and in Appendix E. 3.4.2 Saturated Steam Monitoring of this sample point provides verification of compliance with the boiler manufacturer’s performance guarantee for steam purity, which may apply only to the saturated steam. The sample source may be from one steam offtake or all steam offtake tubes from the drum as long as a representative sample is taken. This sample point also serves as a vital diagnostic tool to monitor the total carryover (a core parameter) of impurities into the steam. The saturated steam chemistry data and resulting carryover figures can also be related to the performance of the steam drum moisture separator devices. Excessive carryover generally indicates either poor moisture separator performance or operation with a higher than intended water level in the steam drum. Mechanical carryover has the greatest impact upon steam chemistry at boiler drum operating pressures below 2500 psi (17.2 MPa); for most species, particularly for sodium and sulfate, mechanical carryover is the major component of the total carryover. For silica and copper, however, volatile carryover is very significant, even at lower pressures. As indicated previously, a knowledge of carryover is considered so important, that it is now a core parameter (Table 3-2), which should be routinely checked (every six months). As discussed in Section 3.3.1, EPRI research on volatility has shown that, for sodium, chloride and sulfate, carryover of these constituents is almost entirely attributable to mechanical carryover. Volatile carryover only becomes significant at the higher operating pressures. Thus, it becomes clear that the steam purity with respect to these constituents is dependent on factors such as the integrity of steam separator devices and drum level control. Also, at higher pressures, the boiler water chemistry must be carefully controlled to prevent contamination of the steam. The control curves presented in Sections 4 and 5 define boiler water conditions that will protect both the boiler and turbine. However, the turbine protection aspect assumes that boiler carryover is effectively controlled. As in earlier chemistry guidelines, industry mechanical carryover criteria were used and a safety factor was applied; see Figure 3-2. Vaporous carryover was determined with the EPRI thermodynamic model developed as a result of earlier volatility studies (see Section 3.3.1). To ensure that the control curves of Sections 4 and 5 are providing turbine protection it is imperative to periodically assess boiler carryover in every drum-type unit. This involves making a routine check of total carryover at six month intervals. Total carryover is defined as the fraction of a given constituent in the saturated steam, expressed as a percentage of the same constituent in the boiler water. For routine chemistry surveillance of total carryover, the determination should be made using sodium as the selected constituent. Total carryover typically varies with boiler load and pressure. Thus it is necessary to evaluate the total carryover under a range of operating conditions, preferably with the boiler chemistry in normal operating limits.

3-30


Sodium monitoring in boiler water and steam may be performed with either on-line analyzers or by laboratory analysis of grab samples. For units that meet the instrumentation criteria of Table 3-2, it will be possible to make total carryover assessments with the installed analyzers, providing that the saturated steam sample can be directed to the sodium analyzer normally used to monitor reheat steam or main steam. (Increased sodium readings at these sample points are a possible indication of increased carryover from the boiler water and this may be evaluated by checking sodium in the saturated steam.) Once the carryover has been calculated, then the value of the sodium in boiler water for PC(L) or PC(H) should be compared with Figure 3-9 (i.e., should be less than the values shown in Figure 3-9 to ensure that the sodium value in steam is not exceeded). The value of sodium in steam for CT should be compared with Figure 3-14. These curves provide the overriding criteria to protect the turbine. The boiler water sodium curves for PC(L) (Figure 4-8), PC(H) (Figure 4-17), and CT (Figure 5-7) in Sections 4 and 5 have been designed to be the same as or lower than the sodium values in Figures 3-9 and 3-14, respectively. 3.4.3 Boiler Water This sample point monitors drum boiler water chemistry to minimize deposition and corrosion in the boiler tubes. This sample point allows control of boiler water chemistry through blowdown and chemical feed, and is a primary control point for saturated steam purity. Boiler drum samples (either blowdown or downcomer) can be used for boiler water analysis. With some boiler designs, the blowdown may not provide a representative boiler water sample, and the downcomer sample should be used. The downcomer samples should also be used for cycling units and for layup chemistry control. Samples from the downcomer will be diluted with the feedwater, and this will reflect a lower concentration of the various chemical species when compared to blowdown samples. This effect may be considerable, depending on boiler design. Therefore, limits derived from downcomer samples must reflect this dilution effect when compared to limits given in these guidelines, which are derived for blowdown samples. Once a satisfactory data base on total carryover is established, it may be possible to customize the boiler limits from what is indicated in the generic control curves presented in Sections 4 and 5. If warranted, relaxation of the boiler water sodium and phosphate (for PC) limits may provide added boiler protection, while still protecting the turbine, so long as the sodium values in Figure 3-9 for PC(L) and PC(H), and Figure 3-14 for CT, are not exceeded. Thus the need to periodically appraise total carryover remains, regardless of whether or not the boiler limits are customized. On the other hand, tightening of the boiler water limits due to steam purity concerns (exceeding the maximum allowable boiler water sodium values predicted to ensure satisfactory steam purity as indicated in Figures 3-9 and 3-14), will compromise the boiler protection aspect of the Section 4 and 5 control curves. If total carryover testing indicates that the limits for boiler water are impacting steam purity, it is imperative that actions be initiated as soon as possible to identify and correct the root cause responsible for the high carryover rate.

3-31


3.4.4 Economizer Inlet and Attemperation Water This sample point allows the direct measurement of the total contaminant ingress to the boiler and to the steam via the attemperation water. In new units, it also permits the determination of whether the feedwater entering the boiler meets the feedwater chemistry limitations required by the boiler manufacturers. This sample point is also used to monitor oxygen and feedwater corrosion product transport, and serves as a sampling point for control of ammonia feed. However, it does not permit the evaluation of flow-accelerated corrosion (FAC) in the economizer header or tubes. 3.4.5 Deaerator Outlet This sample point (in conjunction with the deaerator inlet point) permits an evaluation of how the deaerator performs in removing dissolved oxygen from the feedwater, especially during periods when the condensate oxygen levels are excessive, such as during startups and periods of high makeup with air-saturated water, or when the air in-leakage is high and occurring at locations that do not allow the oxygen to be removed in the condenser. 3.4.6 Deaerator Inlet In most fossil units, the significant reduction of feedwater dissolved oxygen occurs in the deaerator. In these guidelines a distinction is made as to whether the feedwater is being operated in a reducing or oxidizing mode (AVT(R) or AVT(O)). As explained in the guideline document on copper(24), it is considered mandatory that mixed-metallurgy cycles be operated in a reducing regime to minimize copper alloy corrosion, transport and subsequent deposition. This requires the feed of a reducing agent such as hydrazine, in addition to minimizing oxygen ingress to the cycle. To properly determine that a reducing atmosphere is being provided, the measurement of oxidizing-reducing potential (ORP) at the deaerator inlet has been designated a core parameter. Although the reaction between reducing agents and oxygen increases at high temperatures, the reactivity of dissolved oxygen with metal is greater and predominates in the high pressure heater trains. Thus the sample point should be used as a control point for reducing agent (hydrazine) feed at the condensate polisher outlet or condensate pump discharge. All-ferrous systems can be operated in either a reducing or oxidizing mode. As previously noted, an increasing number of power plants with all-ferrous systems are finding that the elimination of the reducing agent (hydrazine) leads to superior results relative to iron transport while, at the same time, minimizing chemical feed costs. This sample point also monitors deaerator oxygen removal performance by comparison to results for samples collected at the deaerator outlet.

3-32


3.4.7 Condensate Polisher Effluent (if Applicable) This sample point is required to determine the effectiveness of the condensate polishers and to determine their need for regeneration. This sample point also permits the evaluation of resin particle “throw” from the condensate polishers. Measurements of sodium and cation conductivity at this point can substitute for the measurement of these parameters at the economizer inlet. 3.4.8 Condensate Pump Discharge This sample point monitors the following:

•

Presence and magnitude of contaminants introduced by condenser leakage.

•

Presence and magnitude of contaminants introduced by the makeup treatment system.

•

Amount of oxygen entering the feedwater train and thus, indirectly, the air in-leakage.

•

Corrosion products from heater drains returned to the condenser.

•

In extreme cases, carryover of contaminants and treatment chemicals in steam.

For plants without condensate polishers, this point monitors the chemistry at the start of the feedwater system. A major change was made in the new AVT Guidelines(6), which involved a tightening of the dissolved oxygen normal target to 10 ppb. This is important for both mixedmetallurgy and all-ferrous feedwater systems and is included in these PC/CT Guidelines. 3.4.9 Condenser Leak Detection Trays and/or Hotwell Zones (if Applicable) This sample point, if provided and available, may allow condenser tube leaks to be detected earlier than by the sample point located at the condensate pump discharge. This aids in timely determination of needed corrective actions. Monitoring at these leak detection tray sample points allows determination of the tube sheet on which the leak is occurring. Quick detection and repair of leaks minimizes the ingress of impurities into the heat cycle. Each individual condenser hotwell section may also be monitored. 3.4.10 Air Removal System Exhaust Condensate oxygen as well as carbon dioxide levels in condensate and low pressure feedwater are a direct function of cycle air in-leakage activity. Air in-leakage in excess of that which can be removed by the air removal system design will result in increased oxygen and carbon dioxide levels, which may cause an increase in corrosion-product generation. The monitoring of air inleakage is considered so important that it has been designated a core parameter. Those organizations that have active air in-leakage programs can easily maintain the 10 ppb oxygen limit at the condensate pump discharge. Minimization of air in-leakage will also prevent corrosion of the condenser shell. A further discussion on measurement of air in-leakage can be found in Appendix C.

3-33


3.4.11 Condensate Storage Tank Effluent Monitoring the effluent from the storage tanks indicates the quality of the available makeup to the condensate. Also, certain water-quality conditions in aluminum condensate storage tanks can lead to aluminum corrosion, resulting in difficult-to-remove aluminum corrosion-product deposits in the boiler and turbine. Serious consideration should be given to protecting the condensate storage tank from oxygen ingress; this is especially important for units that are subject to frequent startups. Refer to Appendix A for further discussion of oxygen control. 3.4.12 Makeup Treatment System Effluent This sample point monitors the performance of the makeup treating system. The achievable makeup water purity is dictated by the design of the makeup water treatment equipment and, may also depend on the makeup rate. Target values have been recommended for the makeup treatment system based on current state-of-the-art equipment capabilities. These guidelines should be used as a performance guarantee for any new makeup treatment systems. Serious consideration should be given to providing oxygen removal equipment and procedures for the makeup effluent and/or the condensate storage tank. Additionally the condensate storage tank containing deoxygenated water should be provided with protection against air ingress. Air ingress not only increases oxygen levels but also introduces carbon dioxide to the cycle. A full discussion on the subject of oxygen removal is given in Appendix A. Oxygen removal for the makeup effluent is of particular importance for mixed-metallurgy systems where a reducing atmosphere is essential for the control of copper corrosion. Use of deoxygenated water to fill equipment prior to startup is also recognized as good practice.

3.5 TROUBLESHOOTING, COMMISSIONING AND CORE PARAMETERS In subsequent cycle diagrams, (Sections 4 and 5) which graphically depict the chemistry guidelines, those parameters at sample points that should be used for routine chemistry monitoring and control are designated by an asterisk(*). These are the EPRI core parameters shown in Table 3-2. Those parameters utilized as supplements for troubleshooting or plant commissioning are designated by “T”. It is suggested that the plant staff develop a specific cycle chemistry diagram for each of their units which only contains the EPRI core parameters, plus any others that, based on customization and optimization activities, are of importance to the individual unit in question. Such a diagram may be more useful for posting in the control room and chemistry laboratory than a diagram displaying both categories of parameters. Water chemistry commissioning should be a part of the total commissioning effort during which a new unit or a retrofitted unit is transferred from the supplier to the operator. A monitoring campaign should similarly be conducted when a major chemistry change is planned for the feedwater or boiler water.

3-34


The main objectives of the commissioning or monitoring are:

•

to prevent equipment damage due to malfunction of water chemistry related equipment.

•

to determine the chemical transport characteristics of the cycle, make final selection of water treatment approach, and establish water and steam chemistry limits.

•

to assess major sources of impurities and corrosion.

•

to quantify total carryover. Note that this is a core parameter, which should be determined at regular intervals (every six months). However, it should be determined as soon as possible if there is a sudden change in steam purity. It should also be measured following each major outage where inspection and repair of the drum has been conducted. Carryover should be measured under a range of operating conditions.

There have been too many cases where unit commissioning was not performed at all or was performed insufficiently, resulting in major equipment damage within weeks or months of the initial operation. There have also been too many instances where a monitoring campaign was not initiated when the feedwater or boiler water treatment was changed. Section 2 delineates clearly the steps that are needed in either case. Typical problems that can be avoided by proper commissioning and monitoring include:

•

high carryover leading to superheater, reheater, and turbine deposits and superheater and reheater overheating failures.

•

dirt and debris left in the boiler or elsewhere in the system resulting in buildup of deposits or foreign object damage.

•

destruction of passive films on tube surfaces, and

•

flow-accelerated corrosion (FAC).

The water chemistry-related systems and functions that should be tested during commissioning and monitoring include:

•

boiler carryover,

•

chemical addition equipment,

•

sampling and analytical equipment,

•

deaeration (condenser, deaerator),

•

effects of reducing agents (particularly for mixed-metallurgy units),

•

iron and copper oxide generation and transport (determination of the points where the corrosion occurs),

•

phosphate levels and hideout return characteristics,

•

blowdown and cycles of concentration,

•

makeup system and associated regeneration equipment,

•

pretreatment system, 3-35


•

condensate polishers and associated regeneration equipment,

•

condensate and feedwater storage systems,

•

cleanliness and passivation effectiveness after preoperational chemical cleaning of cycle components, and

•

purity of returned condensate in cogeneration units.

The importance of monitoring/commissioning for fossil units on PC and CT cannot be over emphasized. The development of optimum alkali solid treatment of boiler water, and associated feedwater treatments for drum units, provided in the road maps in Section 2, depends critically on continual assessment and unit performance.

3.6 RATIONALE FOR ACTION LEVELS Action levels for control parameters at critical sample points have been retained from the previous EPRI Guidelines(1-5) to provide guidance on the relative need for corrective action in the event of cycle chemistry excursions. These action levels may be revised based on a periodic review of experience. Four action levels were established based on the following criteria:

•

Normal. Values are consistent with longterm system reliability. A safety margin has been provided to avoid concentration of contaminants at surfaces and under deposits.

•

Action Level 1. There is a potential for the accumulation of contaminants and corrosion. Return values to normal levels within 1 week.

•

Action Level 2. The accumulation of impurities and corrosion will occur. Return values to normal levels within 24 hours.

•

Action Level 3. Experience indicates that rapid corrosion could occur, which can be avoided by shutdown of the unit within 4 hours.

•

Immediate Shutdown. There is clear evidence of rapid boiler tube damage by low boiler water pH. Immediate shutdown of the unit is required to avoid such damage.

If the problem is not corrected within the time allotted to each action level, the next higher action level takes effect. In drum boilers having the capability of sliding pressure operation, one way to respond to an action level condition is to reduce boiler pressure. The target values for Action Levels 1, 2, and 3 were established to allow and provide guidance during transient or abnormal operating conditions. The following criteria were used for most of the parameters:

•

Action Level 1 Value: 2 x [Normal Target Value]

•

Action Level 2 Value: 2 x [Action Level 1 Value]

•

Action Level 3 Value: > [Action Level 2 Value]

•

Immediate Shutdown: Boiler water pH (drum units) indicates a serious acidic contamination and the unit should be shutdown in an orderly fashion.

3-36


Operation at the lowest achievable impurity levels, with the shortest and least frequent excursions, is the most desirable mode of operation. The probability of corrosion damage and deposit formation grows significantly with increasing impurity concentration, and the damage is proportional to the exposure time. When corrosive impurities concentrate on metal surfaces, the breakdown of passivity (damage to the protective oxide layer) occurs within a few minutes to several hours. (Reheater, superheater, and boiler tubes have been destroyed within days when the impurity concentrations have been high, in some cases as a result of improper chemical cleaning operations or seawater ingress.) During excursions, when the impurity concentration increases sharply, the accumulation of impurities in the cycle (turbine deposits, boiler scale, superheater deposits) can occur quickly, but the subsequent cleanup or return to good chemistry may take weeks. Therefore, yearly time accumulations have also been established as a guide. Another useful approach to damage accumulation control is to set an annual goal for all key parameters of “(contaminant concentration) x (days).” Basically, this approach involves setting a limit on the absolute amount of contaminant transport to a component per year. Contaminant concentrations are then monitored for fixed lengths of time, multiplied by the respective sampling periods, and totaled over the year. The yearly concentration-days calculated in this way can then be compared with the set limit. This approach has been used to rate unit chemistry, compare performance of multiple units, and, for some organizations, serve as an input for assessing the need for boiler chemical cleaning. However, this approach should not be used for accomodating preventable excursions, which amount to using up most of the allowance at one time.

3.7 RATIONALE FOR TARGET VALUES Target values are designated for chemistry parameters at various sample points around the cycle. Organizations that periodically and properly assess their units to determine the most suitable chemistries will generally be more successful at meeting target values and, in so doing, minimize the risk of chemistry-related damage and performance problems. The reader is referred to Section 2 for further discussion on selection and optimization of both boiler water and feedwater. 3.7.1 Steam and Boiler Water Discussion on the rationale and philosophy for boiler water and steam limits is found in Sections 3.3 and 3.3.1. The limits for chloride and sulfate in boiler water in Sections 4 and 5 are derived from the curves for cation conductivity. To ensure adequate protection of the steam turbine (steam limits in Table 3-1), these boiler water chloride and sulfate limits must be lower than the curves in Figure 3-10 and 3-11 for PC(L), Figures 3-13 and 3-11 for PC(H), and Figures 3-15 and 3-16 for CT. It is known that chloride represents a greater risk than sulfate in initiating corrosion, so rather than equal concentrations of chloride and sulfate, the concentration of sulfate has been taken as twice that of chloride. This lower ratio of chloride also makes some allowance for the fact that the cation conductivity includes other impurities. The limits for sodium in boiler water are based on limiting the excess of sodium over chloride and sulfate, and phosphate for PC, and thus, the amount of free sodium hydroxide that could be present in the boiler water. 3-37


A controlled excess of free NaOH is required for CT. The limits for silica are based on the need to limit the amount of silica carried over into the steam, based on partitioning data and the thermodynamic model. Concern related to possible “acid phosphate corrosion” and “caustic” under-deposit corrosion, boiler tube damage mechanisms which do not involve contamination by impurities but instead involve improper control of the boiler water chemistry, are thoroughly addressed by the operating guidelines presented in Sections 4 and 5 of these guidelines. The PC treatments are designed to avoid any possibility of a boiler experiencing acid phosphate corrosion by only allowing additions of trisodium phosphate and NaOH. Each section includes appropriate control limits that prevent the formation of conditions needed to initiate the relevant under-deposit corrosion damage mechanisms. 3.7.2 Feedwater and Condensate As discussed in detail in Section 2, there are three possible feedwater treatments: A. Reducing. Ammonia plus a reducing agent is added after the condensate pump or condensate polisher (if applicable). This treatment must be used for mixed-metallurgy systems but is also used for some all-ferrous systems with poor feedwater cation conductivity (> 0.3 µS/cm). This is AVT(R). B. Oxidizing. Ammonia is added after the condensate pump or condensate polisher (if applicable). Addition of a reducing agent is not applicable. This is the feedwater of choice for all-ferrous systems, especially in those units that generally have good feedwater purity (< 0.3 µS/cm) and/or may not consistently be able to meet the requirements for OT. This is AVT(O). C. Oxygenated. Ammonia plus oxygen is added. This is the feedwater of choice for all-ferrous systems with a condensate polisher or with the ability to maintain feedwater cation conductivity of less than 0.15 µS/cm. This is OT. This feedwater treatment has been used on a number of drum units operating with the CT boiler water treatment. In conventional fossil units, the main emphasis for feedwater must be to minimize the generation of feedwater corrosion products as these act as initiating centers for many BTF, and result in the need to chemically clean. The secondary emphasis is to minimize FAC. For mixed-metallurgy units, iron and copper levels of less than 2 ppb are also achievable if the feedwater is optimized according to Section 2. This requires careful monitoring of air in-leakage and the addition of a reducing agent after the condensate pump. Controlling ORP at the deaerator inlet (a core parameter) can ensure that the low pressure feedwater systems will run under (1-5) reducing conditions (ORP < 0mV). Previous EPRI Cycle Chemistry Guidelines suggested controlling oxygen at the CPD at less than 20 ppb, and loosely related this with air in-leakage of less than 1 scfm/100 MW. However, the importance of maintaining a reducing environment in the feedwater for mixed-metallurgy systems now suggests that the oxygen at the CPD should be less than 10 ppb, independent of the air in-leakage rate. The previous limit of < 1 scfm/100 MW is, however, still a good operating target, particularly for condenser performance purposes. 3-38


Recent research in Program Copper(24) has indicated that the optimum pH for mixed-metallurgy systems is between 9 and 9.3 under reducing conditions. This new CPD limit for oxygen also applies to all-ferrous systems to allow full benefit of the oxidizing environment control. All organizations with all-ferrous systems should be striving to meet less than 2 ppb of iron at the economizer inlet with AVT(R), at around 1 ppb with AVT(O) and less than 1 ppb on OT. These levels are easily achievable if the processes described in Section 2 are followed.

3.8 RATIONALE FOR TARGET VALUES OF INDIVIDUAL PARAMETERS Based on the general rationale discussed earlier, target values were developed for each parameter at each sample point. Specific rationales for these target values are discussed below. While the cycle diagrams in Sections 4 and 5 for drum units only show reheat units, the limits for nonreheat limits can generally be taken as twice those of reheat units. As discussed in Section 5.2.2 for CT, care must be taken in adopting this relaxation in non-reheat units. For PC similarly, monitoring needs to be conducted to ensure that the steam limits are not exceeded. 3.8.1 Sodium Sodium is a core monitoring parameter in steam, boiler water (blowdown or downcomer), condensate and feedwater. Monitoring of sodium is necessary because sodium compounds such as sodium hydroxide and sodium chloride are two major corrodents of concern within the turbine, especially in relation to stress corrosion cracking of some turbine blades and stainless steel superheater tubes. Uncontrolled and/or chronically high sodium hydroxide concentrations have been noted to cause under-deposit corrosion damage and failures of boiler tubes. Sodium Target Value in Superheated (Reheat/Main) and Saturated Steam The thermodynamic model indicates that the reheat/main steam sodium target is probably the least conservative limit in these guidelines (see Figures 3-9 and 3-14 for Na in Section 3.3.1). The new limits of 3 ppb for PC(H) units and 2 ppb Na for PC(L) and CT units do not guarantee the absence of NaOH deposition on surfaces within the LP turbine. Thus, lower levels of sodium should be maintained if possible. Saturated steam (and boiler water) sodium should be monitored to determine total carryover (a core parameter).

3-39


Sodium Target Value in Boiler Water Boiler water sodium target values for PC (Section 4) and CT (Section 5) are based on several factors. These include:

•

Attainment of suitable steam purity to protect the turbine, provided drum carryover is not excessive (In practice this means that sodium in the boiler water must not exceed the values provided for PC and CT in Sections 4 and 5. It should be remembered that these have been set below the maximum allowable levels predicted in Figures 3-9 for PC(L) and PC(H) and 3-14 for CT.)

•

Conformance to treatment operating ranges designated for PC(L), PC(H), and CT

•

The presence of reasonable and allowable levels of chloride and sulfate in the boiler water; contaminant levels may be monitored indirectly via conductivity when direct measurement capabilities are unavailable.

•

For PC(L) and PC(H), proper control of free sodium hydroxide to allowable levels.

With routine assessment of total carryover to ensure that it is within expected values, meeting the boiler water sodium target also ensures attainment of acceptable sodium levels in saturated steam. Meeting the boiler water sodium target values also provides protection of the boiler from contaminant ingress (which can lead to hydrogen damage) and minimizes the risk of underdeposit corrosion by caustic gouging in CT. Organizations applying the PC guidelines of Section 4 (Figure 4-8 for PC(L) and Figure 4-17 for PC(H)) to fossil units should keep in mind that the sodium measured in boiler water represents the total sodium, not just sodium associated with phosphate. It includes any sodium that is associated with anionic impurities such as chloride, sulfate, and carbonate plus any free hydroxide present as a result of caustic additions to the boiler, makeup system or condensate polisher malfunctions, use of impure sodium hydroxide or possibly, in the case of those units with highly alkaline cooling waters, condenser leaks. Thus it is inappropriate to use the “raw” sodium and phosphate data to calculate a sodium to phosphate molar ratio in the boiler water. An approach sometimes considered is to monitor sodium hydroxide levels in the boiler water and use this information, along with the measured phosphate concentration to assess the sodium to phosphate ratio. However, available procedures for sodium hydroxide analysis reportedly are not (34,35) completely reliable when applied to boiler water samples containing phosphate. Determination of the sodium to phosphate ratio, and of the presence or absence of free NaOH and thus the adherence to limits for PC requires that calculations be made, using measured boiler water phosphate concentration and pH, corrected for the influence of ammonia at standard conditions. Publications by Verib provide a good assessment of these relationships and how to (35,36) Boiler water sodium measurements are also necessary for manage them in fossil plants. routine control purposes, or for correction of out of specification chemistry in boiler water. Guidance on control of PC is given in Sections 4 and 6.

3-40


Boiler water sodium target values were derived from sodium limits for the reheat/superheated steam for PC(L), PC(H) and CT. The values for PC(H) are consistent with conventional treatment practice for lower pressure boilers but are lower than commonly maintained in practice for higher pressure boilers. These lower values have been included not only in consideration of steam purity but also to reduce problems with phosphate hideout during load changes. Lower sodium, and consequently, lower phosphate, have been found to reduce the severity of hideout. Although this practice gives less buffering capacity for contaminants introduced by condenser tube leakage, adherence to the guidelines throughout the rest of the cycle will diminish the need for excess buffering capacity. Phosphate levels consistent with the sodium target values in the guidelines will provide adequate capacity when all other boiler target values are being met. Target values for PC(L) provide less boiler protection than those for PC(H), however, they both require and allow for better steam purity control, which is crucial in high pressure drum-type units. Operation with free sodium hydroxide in the boiler water increases the boiler protection but this sodium can be introduced to the steam with mechanical carryover. The contribution of 1 ppm of free sodium hydroxide in boiler water as a function of the pressure is shown in Figures 317 and 3-18. This is allowed with PC(H) and PC(L) when included in the total allowable sodium permissible to ensure suitable steam purity under PC(H) and PC(L) chemistries. Thus the importance of monitoring sodium in the boiler water to avoid steam purity excursions becomes apparent; it is especially critical at higher pressures.

3-41


Figure 3-17 Model Predictions for Boiler Water Sodium as a Function of Pressure when Operating with PC(H) to ensure less than 3 ppb sodium in steam at Na:PO4 = 3, and also the contribution from 1 ppm NaOH, if added. The top curve is the same as shown in Figure 3-9. Note: these curves should not be used for boiler water control; they are provided to illustrate the effects of sodium in boiler water. The control curves for PC(H) are provided in Section 4.

3-42


Figure 3-18 Model Predictions for Boiler Water Sodium as a Function of Pressure when Operating with PC(L) to ensure less than 2 ppb sodium in steam at Na:PO4 = 3, and also the contribution from 1 ppm NaOH, if added. The top curve is the same as shown in Figure 3-9. Note: these curves should not be used for boiler water control; they are provided to illustrate the effects of sodium in boiler water. The control curves for PC(L) are provided in Section 4.

3-43


A minimum phosphate concentration for boiler water of 0.2 ppm as phosphate has been established for PC(L) in recognition of the limited neutralization and pH buffering capabilities of this treatment. An upper limit for phosphate for PC(H) is around 10 ppm as PO4, however, steam purity concerns and hideout characteristics will effectively constrain the practical upper limit of PC(H) at higher boiler operating pressures. Further discussion of the boiler water phosphate target values for PC is provided in Section 3.8.2. Target values for CT have been selected to provide better boiler protection than is possible with PC(L) but not as good as with PC(H). As with PC(L), the associated steam purity requirements for CT are more stringent than for PC(H). As discussed in Section 5, this allows CT to be used in many high pressure drum-type units. Key aspects of the rationale for CT, presented in Section 3.3, and reflected in the operating guidelines presented in Section 5, may be summarized as follows.

•

Steam purity must be controlled to ensure that concentrations of sodium, chloride and sulfate are less than 2 ppb.

•

Total carryover monitoring (at least every six months) is routinely performed to ensure the steam purity remains satisfactory.

•

A pressure-dependent maximum allowable free sodium hydroxide concentration in boiler water (Figure 5-6) is required to minimize the risk of underdeposit corrosion by caustic gouging; this also helps to protect the turbine.

•

Routine control of boiler water chemistry is based on monitoring of the pH, sodium, and conductivity (cation and/or specific).

•

Results for pH and sodium provide an indication of the treatment control while sodium also is useful in steam purity control; the conductivity readings are used to identify contaminant levels and must be controlled to protect the boiler.

•

Although no absolute minimum free caustic level is indicated, the intent is to maintain the sodium hydroxide concentration at least 2.5 times the chloride concentration. (It should be kept in mind that the degree of boiler protection lessens as the caustic concentration decreases; complete loss of free caustic in the boiler water results in the same degree of boiler protection, essentially none, afforded by AVT.)

•

At low sodium hydroxide levels the pH is influenced by ammonia entering the boiler with the feedwater; see Figure 5-12. The allowable corrected pH is dependent on the measured sodium hydroxide concentration and should never be less than 9.0.

•

In Section 5, lower limits for pH, specific conductivity and sodium hydroxide concentration are shown as dashed lines in order to provide guidance on the minimum pH. The area between the dashed line and Action Level 1 should be taken as the normal operating zone.

Sodium analysis of units on CT can also be useful in identifying and responding to incidents of caustic hideout. Although less common than sodium phosphate hideout, hideout of caustic is occasionally observed in units that operate on CT. Section 5 characterizes caustic hideout and provides guidance on how to deal with it.

3-44


For PC and CT, higher treatment levels are permitted by the action levels during excursions but are limited in duration since these higher levels only buffer boiler water pH and will not deter the carryover of corrodents such as chlorides and sulfates into the turbine when their boiler water concentrations exceed target values. Sodium Target Value in Condensate and Makeup Water Low sodium levels are required for condensate and makeup water for the following reasons:

•

To achieve recommended steam purity under normal operation without routinely relying on boiler blowdown capability (which remains available to control transients) and maintain the chemistry within guidelines at all times.

•

To allow maximum flexibility during excursions.

Limits on sodium concentration in the condensate can be more relaxed for units with condensate polishers than for units without polishers, however, it is not advisable to operate for extended periods with contaminated condensate, even in units with polishers. Monitoring of condensate polisher effluent (where applicable) for sodium is useful in evaluating polisher performance. Sodium Target Value in Condensate Storage Tank Effluent (Aluminum Tanks Only) Corrosion of aluminum condensate storage tanks can occur in the presence of free sodium hydroxide. Aluminum corrosion products form difficult-to-remove boiler tube and turbine deposits. Free sodium hydroxide may be present in the condensate storage tank whenever sodium levels become excessive. The major sources of sodium in the condensate are from the makeup treatment system effluent and from condenser leakage. The sodium level in condensate storage tank effluent should be comparable to that in makeup treatment system effluent. 3.8.2 Phosphate Target in Boiler Water Monitoring of boiler water phosphate when using PC is necessary for the following reasons:

•

The sodium-to-phosphate molar ratio determines the control point relative to the pH/PO4 curves (as illustrated in Figure 1-7), and

•

Acid phosphate corrosion is a concern if the boiler water is operated with low sodium-tophosphate ratios.

Corrosive conditions may result from the concentration of either low ratio phosphate treatment chemicals or excessive free NaOH beneath porous magnetite and other boiler wall deposits, and in steam blanketed areas. This may lead to tube damage by corrosion and failures in boilers with accumulated waterside (porous) deposits.

3-45


Careful consideration should be given to avoid the feed of excessive amounts of treatment chemicals in PC(H) units. As discussed previously this may result in hideout, but will not result in corrosion if only trisodium phosphate is added as the phosphate addition. Studies have now shown that the hideout behavior under acceptably “clean” conditions is very different than when excessive deposits are present. The deposits are essentially magnetite but can also contain other corrosion products from different pre-boiler construction materials. Key aspects of the rationale for PC, presented in Sections 1 and 3.3, and reflected in the operating guidelines presented in Section 4, may be summarized as follows.

•

High pressure boilers will generally need to be operated at the lower end of the PC, normally in the region designated PC(L). While this treatment offers a limited boiler protection capability, it must also ensure that steam of satisfactory purity is generated for expansion in the turbine.

•

Low pressure boilers can usually be operated at the higher end of the PC, in the region designated PC(H); this treatment offers additional neutralization and buffering capacity to protect the boiler and also has less stringent steam purity criteria that are considered sufficient to protect the turbine.

•

Trisodium phosphate is the only phosphate-based chemical used with PC; sodium hydroxide may be added as needed to control the pH while maintaining a free sodium hydroxide residual of up to 1 ppm as NaOH. Attention must be given to Section 3.8.1, and particularly to Figures 3-17 and 3-18 to ensure compliance with steam sodium limits.

•

Phosphate hideout is possible with either PC(H) or PC(L); chasing of hideout with repetitious additions of chemicals is not required and should be discouraged.

•

A minimum allowable boiler water phosphate concentration of 0.2 ppm as PO4 has been established, in recognition of the need for a minimum limit on boiler protection in the event of contaminant ingress; maintaining some free sodium hydroxide is especially desirable when phosphate concentrations are < 1 ppm as PO4. The minimum boiler water pH is 9.0. If for some reason a high pressure boiler can only operate with phosphate levels below 0.2 ppm, then it is suggested that the boiler water treatment is converted to CT or AVT. Section 2.3 provides the necessary guidance.

•

As discussed in Section 1, PC should be regarded as having the two extremes of PC(L) and PC(H), which behave and react in very different ways. A loose transition between these two extremes has been suggested at about 3 ppm PO4. The maximum allowable phosphate concentration for PC(H) is 10 ppm as PO4, however, not all boilers will be able to operate over the entire range for PC(H), due to steam purity constraints (sodium as indicated in Figure 4-17), or excessive hideout and return of phosphate.

•

Routine control of boiler water chemistry is based on monitoring of the pH, phosphate, sodium, and conductivity (cation and/or specific; see examples in Figures 4-13 and 4-14 for PC(L), and Figures 4-23 and 4-24 for PC(H)).

•

Results for pH and phosphate, provide an indication of the treatment control, while sodium is useful in steam purity control and as a general indicator of contamination; the conductivity readings are used to identify and track changes in contaminant levels and must be controlled to protect the boiler.

3-46


•

At the low phosphate levels associated with PC(L), the pH is strongly influenced by ammonia entering the boiler with the feedwater; see Figures 4-3 and 4-4; as indicated in Figure 4-4, the influence of ammonia on pH is less pronounced when the boiler water contains free sodium hydroxide. The pH correction may be determined directly from the measured pH and ammonia values using Figure 4-5. The ammonia effect is also discussed in Section 3.8.10.

•

The influence of ammonia on pH is less significant when the boiler water phosphate concentration is above about 3 ppm as PO4; thus the effect of ammonia on boiler water pH is generally not of as great a concern when employing PC(H).

3.8.3 Chloride Monitoring of chloride is necessary, especially for troubleshooting, because chloride contributes to the following:

•

corrosion fatigue, stress corrosion cracking, and pitting in LP turbines;

•

corrosion, hydrogen damage, and pitting in boilers; and

•

synergistic effect on drum boiler waterwall corrosion in the presence of oxygen and copper oxide

Chloride Target Value in Steam (9-12) Based on the turbine deposition studies and the solubility of sodium chloride in superheated steam, a limit of 2 ppb was chosen for units on PC(L) and CT, and 3 ppb for units on PC(H). Compliance with this limit should result in avoidance of damage to the boiler and turbine, and help to minimize deposition in the LP turbine.

Chloride Target Value in Boiler Water Boiler water chloride target values were derived to provide boiler corrosion protection (Figure 410 for PC(L), Figure 4-20 for PC(H), and Figure 5-8 for CT). They were derived from curves for boiler water cation conductivity and chloride for CT as described in Sections 3.3 and 3.3.1. These values will protect both the boiler and the turbine. Chloride represents a greater risk than sulfate in initiating corrosion, so the concentration of chloride has been taken as half that of sulfate. This lower ratio of chloride also makes some allowance for the fact that the cation conductivity also includes other impurities. (21) As was discussed in The Volatility of Impurities in Water/Steam Cycles, chloride may undergo volatile transport to the steam as both HCl and NH4Cl, the ratio of which depends largely on the pH of the liquid. Whereas the dominant species for NH4Cl solutions in both the high and low temperature liquid phase are NH4+ and Cl , the species transported to the steam are predominantly HCl and NH3. At low ammonia concentrations, the major source of chloride in the steam is partitioning of HCl from the liquid. On the other hand, the primary contribution of HCl partitioning to the chloride content in the steam does not imply that the subsequent appearance of chloride, in early or first condensate, will be as HCl.

3-47


For the volatility and speciation in ammonium chloride solutions, it is important to remember that the solute transports to the steam as mixtures of HCl, NH3 and NH4Cl, the relative proportions being dependent on temperature, total chloride, solution pH and presence of other cations (e.g. Na+). The difference in pH and the increasing volatility of HCl, accounts for the increased importance of chloride as HCl at high temperatures. Chloride Target Value in Makeup Water The makeup water target value of 3 ppb is necessary for the following reasons:

•

to achieve recommended steam purity under normal operation without depending on boiler blowdown, leaving the full blowdown capability available to control transients and maintain the chemistry within guidelines at all times, and

•

to allow maximum flexibility during excursions.

This target value should be modified if necessary to be consistent with the capabilities of the makeup treatment equipment design. 3.8.4 Sulfate Monitoring of sulfate is necessary for troubleshooting purposes for the following reasons:

•

sodium sulfate in combination with chloride contributes to turbine corrosion,

•

sulfate causes off-load corrosion of steam-side components,

•

acid sulfate can cause boiler and turbine corrosion, and

•

sulfate can deposit in reheaters and subsequently lead to pitting during non-protected shutdown periods.

Grab sample analysis is normally used to assess sulfate, supplemented by indirect on-line continuous analysis by cation conductivity. Sulfate Target Value in Steam (9-12) Based on the turbine deposition studies, and that the solubility of sodium sulfate in superheated steam is estimated to be similar to that of sodium chloride, and acid sulfate is as corrosive as acid chloride, the same limits were chosen for sulfate in steam (2 ppb for PC(L) and CT, 3 ppb for PC(H)) as those for chloride.

3-48


Sulfate Target Value in Boiler Water Boiler water sulfate target values were derived to provide boiler corrosion protection (Figure 4-11 for PC(L), Figure 4-21 for PC(H), and Figure 5-9 for CT). They were derived from curves for boiler water cation conductivity as described in Sections 3.3 and 3.3.1. These values will protect both the boiler and the turbine. Sulfate represents a lesser risk than chloride in initiating corrosion, so the concentration of sulfate has been taken as twice that of chloride. (21) As was discussed in The Volatility of Impurities in Water/Steam Cycles , the situation with carryover of sulfate is more complex than for chloride because sulfuric acid is divalent and can dissociate in two steps, forming bisulfates, as well as sulfates. It is also less volatile than hydrochloric acid (Figure 3-7).

Although H2SO4 can be considered as a 1:1 compound or a 1:2 compound, at high temperatures, it behaves more as a 1:1 compound. Sulfuric acid was found to be the dominant form of sulfur (VI) species in steam under most conditions and is the principal transporting medium for the (21) vaporous carryover of sulfate in steam . Reduced sulfur species (SO2) were observed in all vapor samples taken over sulfuric acid solutions. The partitioning coefficients for SO2 obtained from the sulfuric acid volatility data, showed that with increasing pH, the hydrolysis of the weaker sulfurous acid becomes significant and the level of sulfite decreases in the vapor phase. Sulfur dioxide is a neutral specie and is very volatile (Figure 3-6). Reduced sulfur species may provide an important route for the transport of sulfate and sulfite to the steam. Thus the carryover of sulfate should be significantly reduced for plants with oxygenated treatment by eliminating SO2 transport. It is also interesting to compare the figures for chloride and sulfate in Section 3.3.1, which show the predictions from the thermodynamic partitioning code. Higher values of sulfate in boiler water are required than for chloride in boiler water to generate 2 ppb in steam. Sulfate Target Value in Makeup Water The makeup water sulfate target value of 3 ppb is necessary for the following reasons:

•

to achieve recommended steam purity under normal operation without relying upon boiler blowdown, leaving the full blowdown capability available to control transients and maintain the chemistry within guidelines at all times, and

•

to allow maximum flexibility during excursions.

This target value should be modified, if necessary, to be consistent with the capabilities of the makeup treating equipment.

3-49


3.8.5 Silica Monitoring of silica is necessary for the following reasons:

•

precipitation of silica forms silicate deposits on the turbine that are not soluble in water and are very difficult to remove;

•

silicate deposits cause losses in turbine capacity and efficiency; and

•

significant amounts of silica can enter the condensate/feedwater undetected in the nonreactive colloidal form through the makeup, causing boiler water and steam target values to be exceeded in spite of apparently good-quality makeup.

Silica Target Value in Steam Based on the solubility of silica in superheated steam, a limit of 10 ppb was chosen for PC and CT units. Compliance with this target value should limit deposition of silica in the turbine. Silica Target Value in Boiler Water Vaporous carryover is the major contributor to silica in steam as illustrated by the partitioning predictions of silica (Figure 3-12 in Section 3.3.1). Also Figure 3-3 shows the distribution ratio, or concentration in the steam to concentration in the boiler water, of silica as a function of pressure. Boiler water silica target values were derived to keep silica below the steam limit (Figure 4-12 for PC(L), Figure 4-22 for PC(H), and Figure 5-10 for CT). High boiler water pH reduces silica carryover. However, this effect is insignificant at the pH levels maintained in high pressure boilers on PC and CT and, for that reason, is not considered in these guidelines. Operation at higher boiler water pH values will cause these guidelines to be slightly conservative. The limits for silica are based on the need to limit the amount of silica carried over into the steam and are based on partitioning data and the thermodynamic code. Silica Target Value in Condensate and Makeup Water The target value of 10 ppb silica in condensate is necessary for these reasons:

•

To achieve recommended steam purity under normal operation without relying on boiler blowdown, leaving the full blowdown capability available to control transients and maintain the chemistry within guidelines at all times, and

•

To allow maximum flexibility during excursions.

This target value should be modified if necessary to be consistent with the capabilities of the makeup treating equipment design.

3-50


3.8.6 Dissolved Oxygen Dissolved oxygen is a core monitoring parameter in condensate and feedwater. Monitoring of oxygen is necessary for the following reasons:

•

Copper alloy corrosion in condensate and feedwater systems is controlled by ORP. The value of ORP is influenced by the levels of oxygen and reducing agent in the feedwater and by pH.

•

In all-ferrous feedwater systems, very low levels of oxygen (<1 ppb) in conjunction with high levels of reducing agent (hydrazine or alternative) result in very strong reducing (26) conditions. Under such conditions, the feedwater reacts with the ferrous alloys in the feedwater system to generate corrosion products. These corrosion products are subsequently transported to the boiler, which leads to a) increased chemical cleaning requirements, and b) increased risk of boiler waterwall tube failures by under-deposit corrosion.

•

Single phase flow-accelerated corrosion of carbon steel components is accelerated by overfeeding reducing agents and stopped by eliminating them and/or adding oxygen at concentrations above 20 ppb in high purity feedwater. Overfeeding of reducing agents is, therefore, not recommended and some minimal amount of oxygen in high purity water is (26) beneficial in reducing generation of iron oxides.

•

Many plants with all-ferrous feedwater systems operate without the use of reducing agents, and have significantly reduced iron transport.

Dissolved Oxygen Target Value in Economizer Inlet This dissolved oxygen target provides an indication of the oxygen control for reducing, oxidizing and oxygenated feedwater, and is thus important for controlling feedwater material corrosion in all-ferrous and mixed-metallurgy feedwater systems. Dissolved Oxygen Target Value in Drum Boiler Water This dissolved oxygen target is provided as an indication of potential boiler waterwall corrosion. Its level is determined by the oxygen at the economizer inlet and the boiler recirculation ratio. Dissolved Oxygen Target Value at Deaerator Outlet Monitoring of oxygen at the deaerator outlet serves as a check for the effectiveness of deaeration. Dissolved Oxygen Target Value at Condensate Pump Discharge This dissolved oxygen target value (10 ppb) is provided to minimize total corrosion in mixedmetallurgy low pressure feedwater heaters and to provide an indication of air in-leakage into the cycle for all feedwater systems.

3-51


3.8.7 Oxidizing-Reducing Potential (ORP) ORP is a core parameter for systems having mixed-metallurgy. To minimize copper transport in such systems, it is essential to provide a reducing atmosphere in the feedwater. Measurement of reducing agent concentration at the deaerator inlet does not necessarily guarantee that a reducing condition has been attained. An ORP reading of -300 to -350 mV at the deaerator inlet ensures a reducing condition which will minimize copper corrosion and transport. However, operators are not encouraged to continually add a reducing agent to the feedwater in the understanding that these ORP readings must be achieved, since experience has shown that this approach is ineffective unless the dissolved oxygen is maintained at less than 10 ppb. Appendix B should be referenced for further understanding of ORP readings. Reducing ORP levels are very dependent on a number of variables, and it will not be possible, in all systems, to reach ORP readings of –300 to –350 mV. ORP is not a core parameter for all-ferrous systems operated in an oxidizing mode. However, the use of ORP in this situation will facilitate assessing and fine tuning feedwater chemistry control. A positive ORP value will ensure an oxidizing condition. A complete discussion of ORP measurement and interpretation can be found in Appendix B. 3.8.8 Reducing Agents (Hydrazine or Alternates) Monitoring of the reducing agent and ORP at the deaerator inlet can provide feedback control of the feed rate. The reducing agent concentration in conjunction with the oxygen level at this location should be consistent with the ORP value (Section 3.8.7). ORP is a much more useful measurement than either oxygen or reducing agent. For all-ferrous feedwater systems, there is evidence linking higher reducing agent concentrations and low oxygen levels at the economizer inlet (<1 ppb) with flow-accelerated corrosion and (26) increased transport of feedwater corrosion products (see Section 2). Thus, it is necessary to advise that for feedwater heater trains constructed in all-ferrous materials, that good operating practice requires that the need for reducing agent addition be confirmed by monitoring of the feedwater corrosion products, ORP and oxygen levels at the economizer inlet. Many operators have found that additions of reducing agent in these all-ferrous systems make very little difference to the level of oxygen, but produce a major change in the oxidizing/reducing potential of the feedwater. Thus many have been able to eliminate their use, and oxidizing AVT, AVT(O), is the feedwater treatment of choice for all-ferrous systems. This AVT(O) feedwater treatment can be used for units where the boiler is treated with PC or CT. Cycles with copper alloy tubes in the feedwater system require lower oxygen levels and higher reducing agent levels for AVT(R) than cycles with all-ferrous components. The ORP at the deaerator inlet in these mixed-metallurgy feedwater systems must be determined and coordinated with reducing agent dosage and a low air in-leakage level to ensure reducing conditions exist (ORP of -300 to -350 mV). These levels of reducing ORP are not always achieved because ORP is very dependent on a number of variables (see Appendix B).

3-52


3.8.9 pH Boiler water pH is a core monitoring parameter. Monitoring of pH is necessary for the following reasons:

•

corrosion of cycle materials in contact with feedwater and boiler water is a function of pH, and

•

alkaline pH values increase the stability of the magnetite film and reduce magnetite solubility in water.

pH Control It is important to keep in mind that ammonia treatment of the feedwater (alkaline AVT (R) or AVT (O)) is intended to reduce general corrosion and FAC in the pre-boiler system. However, ammonia (or ammonium hydroxide) provides minimal, if any, neutralizing effect in the boiler water. Ammonium hydroxide is essentially fully associated at boiler operating temperatures. It, therefore, does not materially contribute to neutralizing possible acidic feedwater contaminants that may be available to concentrate beneath porous oxide waterwall tube deposits. Alkali solids (trisodium phosphate and/or sodium hydroxide) are added to the boiler water to neutralize any acidic contaminants that may be present. However, at typical sampling conditions, ammonia in boiler water samples is partially dissociated and therefore affects the pH. For this reason extreme care should be taken in the measurement of pH from the standpoint of both sample temperature (25°C, 77°F) and proper calibration of electrodes in an on-line measurement system. The effect of ammonia in the sample on the measured boiler water pH is discussed in the next section. pH Target Values in Boiler Water The measured pH in conditioned boiler water samples will be influenced by ammonia. This effect is most pronounced under PC(L) and under CT when levels of sodium hydroxide are less than one ppm. For effective surveillance and control of PC and CT, a correction for the influence of ammonia should be made. By measuring the boiler water pH and ammonia concentration, it is possible to calculate the corrected pH value, by manipulation of the relationship between the concentration of hydrogen and hydroxide ions in the water; the following approach has been described by Verib that is (35) useful in defining the basic relationships involved : pH = -log[H]

(eq. 3-6)

pH + pOH = 14

(eq. 3-7)

Concentrations are expressed in moles per liter. Be rearranging, it is possible to express the OH concentration as a function of the pH: -14 -pH [OH] = 10 /10

(eq. 3-8)

3-53


To this point, all terms relate to initial measurement of the pH in the boiler water. The following equations may be applied to calculate the corrected pH by subtracting the influence of the measured ammonia concentration, also expressed in moles per liter: [OH]c = [OH]m – [NH3]m

(eq. 3-9)

pOHc = -log[OH]c

(eq. 3-10)

pHc = 14 – pOHc

(eq. 3-11)

In this series of equations, [OH]m is the value calculated in equation 3-8 using the measured pH. Terms with the subscript c denote corrected values, as determined by calculcation. The corrected pH, if determined by this methodology is, at best, a rough estimate, since not all of the ammonia measured in the boiler water sample will dissociate to ammonium and hydroxide ions. Complete dissociation of ammonia is assumed by direct substitution of the measured ammonia concentration in Equation 3-9. The actual degree of ammonia hydrolysis is incomplete and variable in conditioned samples, influenced by both sample temperature and concentrations of ammonia and alkali treatments present. These influences are considered in the EPRI thermodynamic model used to develop Figures 4-3 and 4-4 presented in Section 4 for PC and Figure 5-12 in Section 5 for CT. While Figure 5-12 may be readily used to determine the influence of ammonia on pH when employing CT, the situation is more complicated in the case of PC, where both phosphate and sodium hydroxide may be present, resulting in a spectrum of sodium to phosphate ratios intermediate to the relationships shown in Figures 4-3 (only trisodium phosphate) and Figure 4-4 (trisodium phosphate plus 1 ppm of caustic.) To facilitate correction of pH for the effects of ammonia, Figure 4-5 has been developed, which can be used for PC(L), PC(H) and CT. The derivation process is described in detail in Appendix G. The corrected pH, that is, the pH that would be produced by the chemical treatment in the absence of the ammonia can be determined as a function of the measured pH value at various measured ammonia concentrations and independent of the actual measured phosphate concentration. However, to use the figure for diagnostic purposes the boiler water phosphate concentrations should be verified to be within the range (0.2–10 ppm) required when operating with PC. Proper use of Figure 4-5 is addressed in Subsection 4.2.1, Introduction to PC(L). However, the figure is applicable to the entire PC working range and thus may also be used with PC(H) if desired. Comparison of pH correction results obtained with Figure 5-12 for CT with those of Figure 4-5 shows satisfactory agreement. This result was anticipated, since dissociation of trisodium phosphate to disodium phosphate and sodium hydroxide is nearly complete within the pH range of interest at standard sample temperature (25°C (77°F). However, derivation of a corrected pH equation of CT (based only on sodium hydroxide and ammonia) was performed (as discussed further in Appendix G) to confirm this consistency. Thus under normal treatment conditions, and with accurate measurement of boiler water pH and ammonia, the ammonia specific correction curves shown in Figure 4-5 are applicable to operation with CT as well as PC.

3-54


When managing boiler water chemistry, it must also be kept in mind that errors in measurement of the boiler water parameters needed to determine corrected pH may be significant and will thus affect the results obtained when making the pH correction. General guidance on chemistry data quality is included in Appendix E. Immediate shutdown is recommended for low drum boiler water pH of less than 8 (as measured) and descending. Immediate shutdown is recommended since the pH will generally be much lower locally, which could lead to severe localized attack on waterwall tubes depending on the local conditions, particularly if the waterwall tubes are dirty and boiler chemical cleaning is due. pH Target Values in Feedwater The corrosion rate of carbon steel reaches minimum values over the pH range of 9.2 to 9.6. The optimal pH for protection of copper has been researched extensively over the last 5 years and is dependent upon temperature and ORP. Thus systems having mixed-metallurgy should now be able to minimize the corrosion rate of both carbon steel and the copper alloy by maintaining the condensate/feedwater pH between 9.0 and 9.3 under reducing conditions. The feedwater pH can be affected by carbon dioxide which enters with air in-leakage (Figure 3-26). For units with condensate polishers in H-OH form, polisher run lengths decrease significantly at higher pH values. Operation at the lower end of the feedwater pH range may be necessary to maintain reasonable condensate polisher run lengths and acceptable effluent quality when using hydrogen form cation resins. Operation in the ammonia form will maximize run lengths at the expense of somewhat higher sodium leakage (normally within guidelines for sodium). 3.8.10 Ammonia Monitoring of ammonia can be used to supplement the direct measurement of feedwater pH and/or specific conductivity for the control of the ammonia feed rate. It also influences the pH of samples from boilers operating on PC and CT (Section 3.8.9). Curves displaying these effects, as determined by the thermodynamic model, are presented and discussed in Sections 4 and 5 (Figures 4-3 to 4-5 for PC, and Figure 5-12 for CT). For plants containing copper alloys in the condensate/feedwater systems, monitoring of ammonia is also important to minimize ammonia attack on copper alloys (particularly when oxygen and carbon dioxide are present), and particularly in the air removal section of the condenser.

3-55


Ammonia Target Values at the Economizer Inlet The ammonia concentration in the final feedwater should be consistent with the pH requirements. The pH of the feedwater is affected by carbon dioxide as illustrated in Figure 3-26. Monitoring Ammonia in Boiler Water As described earlier, ammonia in the boiler water will influence the pH. This effect becomes more pronounced when the levels of trisodium phosphate and/or sodium hydroxide in the boiler water are low. Figures 4-3 to 4-5 for PC and 5-12 for CT illustrate the effect of ammonia on pH as predicted by the thermodynamic model. In reality, the measured pH is also influenced by any other dissolved solids (chlorides, sulfates, organics, etc.) present in the boiler water; however, in the absence of significant contamination (as indicated by chloride and sulfate values in boiler water), this effect does not interfere with routine monitoring and control. For routine cycle chemistry control purposes, it is quite satisfactory to estimate the corrected pH using relationships shown in Figures 4-3 and 4-4 for PC, and Figure 5-12 for CT. Figure 4-5 allows determination of the required pH correction using the results of analysis for measured (uncorrected) pH and ammonia in the boiler water. The correction will be less precise when excessive levels of contaminants are present but such conditions will result in pH changes and increased conductivity (cation and/or specific) readings and thus require operator response. However, by using the conductivity to detect contamination of the boiler, it is envisioned that better protection of the boiler will be possible. It should be recognized that ammonia levels in downcomer samples will tend to be somewhat higher than those measured in the blowdown. Thus the former sample point will provide a more realistic indication of bulk water chemistry in generating tubes. However, the latter point is satisfactory for routine monitoring and control of PC and CT. 3.8.11 Specific Conductivity Specific conductivity is an inexpensive, reliable, on-line method for monitoring the overall level of contamination, treatment chemicals and their trends. It is a core monitoring parameter for treated makeup and boiler water (blowdown). Since the feed of ammonia is relatively high compared to contaminant levels normally present, feedwater specific conductivity values reflect ammonia concentration and mask the levels of impurities present.

3-56


Specific Conductivity Target Values at Economizer Inlet Specific conductivity at the economizer inlet is an indication of the ammonia level and hence pH, during normal operation, since other ionic impurity levels are relatively low. The specific conductivity target value should, therefore, be consistent with the ammonia level. Where applicable, specific conductivity of condensate polisher effluent may be monitored (prior to the chemical injection point) to assess performance. Specific Conductivity in Boiler Water The measurement of specific conductivity in the blowdown or downcomer is a core monitoring parameter and is a very useful diagnostic parameter for PC. It provides a reliable on-line indication of boiler water purity and includes the specific conductivity of phosphates, chlorides and sulfates as well as ammonia and caustic (if used). Utilities can rely on this measurement in conjunction with cation conductivity (discussed later) to initiate control measures to mitigate the undesirable effects associated with increasing levels of contamination. This measurement also provides an indirect method of assessing ammonia and caustic concentration (if used). It should be kept in mind that samples from the downcomers will be diluted with feedwater to a level dependent on the boiler circulation ratio. Care should be taken to ensure optimum sample flow rate and temperature control. The specific conductivity control curves presented as examples in Sections 4 and 5 are designed to protect the boiler. They are based on experience of limiting the concentrations of impurities in the boiler water to minimize corrosion. They consider the boiler water chloride and sulfate levels, together with phosphate, ammonia and caustic (if used) as a function of pressure (see the examples in Figure 4-13 for PC(L), Figure 4-23 for PC(H), and Figure 5-5 for CT). Specific conductivity control curves have been constructed for CT (Figure 5-5). It is much more difficult to construct similar curves for PC. Figure 3-19 shows the model predictions for specific conductivity as a function of phosphate for solutions of phosphoric acid and sodium phosphates with varying Na:PO4 ratios, including trisodium phosphate plus 1 ppb NaOH. These curves are (27) essentially in agreement with those published by Bosselmann and Brunner . Specific conductivity is strongly influenced by the phosphate concentration, which increases by a factor of 50 over the entire PC operating range. Thus it becomes very difficult to produce generic control curves that are applicable to all or even most fossil units. Figures 3-20 and 3-21 show the results of estimates for boiler water specific conductivity under PC(L) and PC(H) conditions. Each figure includes two curves; the lower curves indicate conductivity attributable only to the phosphate, while the upper curve reflects the conductivity associated with an additional maximum allowable levels of chloride and sulfate in the boiler water.

3-57


It should be noted that the curves presented in these figures correspond to selected phosphate concentrations at sodium to phosphate molar ratios of 3.0 with no free sodium hydroxide present; the allowable specific conductivity limits suggested by these figures would change with variations in treatment chemical (including ammonia) inventories and be influenced by contaminant inventories actually present in the boiler water. Also, steam purity could become an issue in boilers at the higher operating pressures unless the mechanical carryover rate is lower than that used to establish the guidelines.

Figure 3-19 Specific Conductivity as a Function of Phosphate Level with Varying Na:PO4 ratios and TSP + 1 ppm NaOH. No influence of ammonia is included. It should also be noted that the curve for H3PO4 crresponds to cation conductivity.

3-58


Figure 3-20 Estimates of Specific Conductivity Against Pressure for PC(L). The lower curve reflects the conductivity if only phosphate is present in the boiler water. It is based on the phosphate concentration corresponding to the sodium concentration given in Figure 3-18 to ensure less than 2 ppb sodium in steam. The upper curve reflects the situation with additional chloride and sulfate in the boiler water. Note: these curves must not be used for boiler water control; they are provided here to assist the reader to develop a set of specific conductivity control curves in Section 4 for PC(L).

3-59


Figure 3-21 Estimate of Specific Conductivity Against Pressure for PC(H). The lower curve reflects the conductivity if only phosphate is present in the boiler water. It is based on the phosphate concentration corresponding to the sodium concentration given in Figure 3-17 to ensure less than 3 ppb sodium in steam. The upper curve reflects the situation with additional chloride and sulfate in the boiler water. Note: these curves must not be used for boiler water control; they are provided here to assist the reader to develop a set of specific conductivity control curves in Section 4 for PC(H).

3-60


Personnel responsible for operation of fossil units on PC(L) and PC(H) are encouraged to monitor boiler water specific conductivity and examine the correlation this parameter displays with respect to pH, phosphate and sodium levels. Over time, this should lead to a better understanding of “normal” readings for each individual boiler. Once this is known, samples of boiler water should be tested for chloride and sulfate to verify that the accepted normal specific conductivity readings also ensure satisfactory control of these parameters within the target values indicated in Section 4. Specific conductivity will increase as a result of several factors, such as addition of phosphate or caustic to the boiler water or increases in the feedwater ammonia dosing rate. Increasing specific conductivity that is not related to chemical feed activity is most likely due to contamination entering the boiler with the feedwater; phosphate concentrations will begin to fall if the contamination includes any hardness. In most instances, the contamination can be readily identified and responded to if the cycle chemistry program includes the appropriate instrumentation and operator response procedures. However, in many fossil units, extremely low level contamination is a chronic problem that can lead to concentration of impurities in the boiler water to levels capable of initiating corrosion damage. A longterm trend of increase in the conductivity readings may be the only indication that a problem exists until the damage produces failures. As suggested by Figures 3-20 and 3-21, allowable specific conductivity increases associated with contamination are fairly low and become increasingly restrictive at increased operating pressures. Organizations seeking to establish unit specific limits for boiler specific conductivity are encouraged to dose the boiler to the maximum allowable phosphate and free sodium hydroxide levels; this should be done carefully and with the blowdown closed or at the normal minimum setting since overfeed will require opening the blowdown, which will impact subsequent sampling and analysis activities. Once this treatment level is established the conductivity should be checked and samples collected for analysis of chlorides and sulfates. The effect of treatment on steam purity will also need to be checked at this time. By using this data and the information presented in this section, it will then be possible to establish site specific limits that will require blowing down the boiler before contaminants increase to levels that put the boiler at risk. As continual reliance on blowdown is costly and limits flexibility in response to more serious chemistry excursions, personnel responsible for the unit are advised to identify and correct sources of chronic contamination at the earliest possible opportunity. Examples of boiler water specific conductivity control curves for PC are provided in Section 4 (Figure 4-13 for PC(L) and Figure 4-23 for PC(H)). As with other parameters, it is very important to customize the specific conductivity limits for each boiler as just outlined. This is included in the road map to optimize PC and CT in Section 2. In summary, neither specific or cation conductivity give unambiguous results for PC, and thus there is a need to also check pH and phosphate. It is difficult to separate the individual contributions of phosphate and impurity anions on conductivity with PC. It is much easier with CT since there is no contribution from phosphate. With CT, specific and cation conductivity can give a good indication of pH, free NaOH (Figure 5-11) and anionic contamination.

3-61


Specific Conductivity Target Values in Makeup Water The specific conductivity target value is based on the sum of conductivities contributed by the individual ions at the recommended effluent quality and by carbon dioxide. 3.8.12 Cation Conductivity Cation conductivity is a core monitoring parameter in steam, boiler water (blowdown or downcomer) condensate and feedwater, and can be used to indirectly assess levels of chloride and sulfate for corrosion avoidance purposes. It serves as an excellent diagnostic tool as the cations are removed via cation exchange and the H+ ion has a markedly higher equivalent conductance than all other cations. Target values of cation conductivity at the condensate pump discharge can be more relaxed for units with polishers than for those units without polishers. The reason for this relaxation is the removal of possible condensate contaminants in condensate polishers. Cation conductivity measurements do not detect contamination by alkali hydroxides (NaOH and ammonia) and only imperfectly of contamination by alkaline reacting contaminants. In high pressure units, if mixtures of amines or organic reducing agents are used for feedwater treatment, the thermal decomposition products will increase the cation conductivity. Cation Conductivity Target Values at Economizer Inlet The economizer inlet cation conductivity target values reflect the sum of cation conductivities contributed by individual anions and some carbon dioxide that may exist from air in-leakage, or as a result of the thermal decomposition of chemical additions such as alternate amines or organic reducing agents. Measurement of cation conductivity at the condensate polisher effluent (if applicable) can substitute for cation conductivity measurements at the economizer inlet. Cation Conductivity in Boiler Water In order to help operating personnel avoid boiler damage caused by impurities, it is recognized that controls based on monitoring of boiler water cation and specific conductivity are needed for routine control of PC and CT. This need was also recognized and followed during development of interim chemistry guidelines for heat recovery steam generators (HRSGs) in combined cycle (22) plants . The derivations need to consider the thermodynamic model predictions (to prevent carryover) and the experience of fossil plants that monitor cation conductivity (to avoid boiler corrosion damage by chloride and sulfate compounds).

3-62


AVT limits on maximum allowable boiler water impurity concentrations to minimize corrosion in the boiler are based on longstanding fossil plant experience with boiler water cation conductivity around the world(6). It is possible to calculate the concentrations of chloride and sulfate from cation conductivity, provided the proportions of these and the other anions present are known. However, these proportions vary from plant to plant and operating conditions, thereby requiring that the proportion be assumed. As it is known that chloride represents a greater risk than sulfate in initiating corrosion, the concentration of sulfate has been taken as twice that of chloride when calculating the limits from cation conductivity. Note that the 1:2 concentration ratio for chloride:sulfate corresponds to a 40/60 stoichiometric ratio. Taking less than 50% ratio for chloride also makes some allowance for the fact that the cation conductivity (6) includes other impurities. This approach was used in EPRI’s new AVT Guidelines . In order to develop cation conductivity curves for the high and low phosphate extremes of PC, it is necessary to consider the effect of the phosphate on the measured cation conductivity value; hydroxides have no effect on this parameter. It is also necessary to assess the level of contamination that the alkali solid in the boiler water is capable of neutralizing and/or buffering. The cation conductivity control curves for PC(L) and PC(H), and CT determine these effects and then overlay the cation conductivity associated with the allowable (un-neutralized and/or unbuffered) contaminants (based on and as used with AVT) to define the boiler water limit values. The sodium boiler water control curves presented in Sections 4 and 5 are based on maximum permissible levels of chloride and sulfate in the boiler water assuming they are present as sodium salts. They also account for additional sodium associated with allowable levels of any phosphate and free hydroxide in the boiler water. For PC this allowable concentration is 1 ppm sodium hydroxide as NaOH while for caustic treatment it ranges from 1-3 ppm sodium hydroxide as NaOH depending on the pressure. Effects of other contaminants that may be present (such as carbon dioxide and organics) on boiler water cation conductivity are also considered. In units where boiler water sodium is not routinely monitored, there is risk of caustic corrosion if free sodium hydroxide is present in the boiler water at higher than intended levels. This is one reason why sodium is a core parameter (Table 3-1). Boiler water pH limits are set to avoid excessive free hydroxide conditions but this parameter is affected by ammonium hydroxide for all of the alkali solid treatments. Like ammonium hydroxide, sodium hydroxide affects the specific conductivity of the boiler water but not the cation conductivity. Thus, for indirect, conductivity-based monitoring and control of free sodium hydroxide in boiler water in units that do not monitor sodium directly, it may desirable to monitor the specific conductivity as discussed in Section 3.8.11. Thus, the measurement of cation conductivity in the blowdown or downcomer provides a reliable on-line indication of the level of contamination entering the boiler and concentrating by recirculation. For this reason, plant operators can rely on this measurement to initiate control measures to mitigate the undesirable effects associated with increasing levels of contamination. Corrective action, typically blowing down and/or pressure reduction, is needed when increased contaminant levels are observed even if there is no apparent change in boiler water pH or chemical treatment levels.

3-63


It should be kept in mind that samples from the downcomers will be diluted by feedwater to a level dependent on the boiler recirculation ratio. Care should be taken to ensure optimum sample flow rate and temperature control. The boiler cation conductivity control curves, which are discussed in Sections 4 and 5, and are designed to protect the boiler from corrosion. They are currently based on experience of limiting the concentrations of impurities in the boiler water to minimize corrosion as well as the effects of the alkali solid treatments in use. They indirectly consider the boiler water chloride and sulfate levels as a function of pressure together with a number of other contributors such as organic acids, carbonates, etc. For PC, phosphate and contaminant anions all affect the cation conductivity, making it difficult to assess the individual contributions without further data (pH and phosphate). With CT, cation conductivity provides a direct indication of anionic contamination. Cation conductivity control curves have been constructed for CT (Figure 5-4). It is much more difficult to construct similar curves for PC since the measured cation conductivity is strongly influenced by the phosphate concentration, which increases by a factor of 50 over the entire PC operating range. Thus it becomes very difficult for EPRI to produce generic control curves that are applicable to all or even most fossil units. Figures 3-22 and 3-23 show the results of estimates for boiler water cation conductivity under PC(L) and PC(H) conditions. Each figure includes two curves; the lower curves indicate the cation conductivity attributable only to the phosphate, while the upper curves reflect the conductivity associated with the same level of phosphate and an additional maximum allowable level of chloride and sulfate in the boiler water. It should be noted that the curves presented in these figures correspond to a selected phosphate concentration at sodium to phosphate molar ratios of 3.0; the allowable cation conductivity limits suggested by these figures would change with variations in treatment chemical inventories (although cation conductivity is not affected by sodium hydroxide or ammonia) and be influenced by contaminant inventories actually present in the boiler water. Also, steam purity could become an issue in boilers at the higher operating pressures unless the mechanical carryover rate is lower than that used to establish the guidelines. At present, cation conductivity is often not measured in the boiler water of fossil units that utilize phosphate treatments. As was indicated earlier for specific conductivity, personnel responsible for operation of fossil units on PC(L) and PC(H) are encouraged to monitor boiler water cation conductivity and examine the correlation this parameter displays with respect to pH, phosphate and sodium levels. Over time, this should lead to a better understanding of “normal” readings for each individual boiler. Once this is known, samples of boiler water should be tested for chloride and sulfate to verify that the accepted normal cation conductivity readings also ensure satisfactory control of these parameters within the target values indicated in Section 4.

3-64


Cation conductivity will increase as a result of addition of phosphate to the boiler water. Increasing cation conductivity that is not related to chemical feed activity is most likely due to contamination entering the boiler with the feedwater; phosphate concentrations will begin to fall if the contamination includes any hardness. In most instances, the contamination can be readily identified and responded to if the cycle chemistry program includes the appropriate instrumentation and operator response procedures. However, in many fossil units, extremely low level contamination is a serious problem that can lead to concentration of impurities in the boiler water to levels capable of initiating corrosion damage. A longterm trend of increase in the conductivity readings may be the only indication that a problem exists until the damage produces failures. Figures 3-24 and 3-25 show the cation conductivity values versus pressure for PC(L) and PC(H) only taking into account the maximum (normal) levels of chloride and sulfate (1:2 ratio). No contribution of phosphate is included. Organizations seeking to establish unit limits for boiler cation conductivity are encouraged to dose the boiler to the maximum allowable phosphate concentration; this should be done carefully and with the blowdown closed or at the normal minimum setting since overfeed will require opening the blowdown, which will impact subsequent sampling and analysis activities. Once this treatment level is established the cation conductivity should be checked and samples collected for analysis of chlorides and sulfates. The effect of treatment on steam purity will also need to be checked at this time. By using this data and the information presented in this section it will then be possible to establish site specific limits that will require blowing down the boiler and/or reduction of pressure before contaminants increase to levels that put the boiler at risk. As continual reliance on blowdown is costly and limits flexibility in response to more serious chemistry excursions, personnel responsible for the unit are advised to identify and correct sources of chronic contamination at the earliest possible opportunity. Examples of boiler water cation conductivity control curves for PC are provided in Section 4 (Figure 4-14 for PC(L) and Figure 4-24 for PC(H)). Calculation of boiler water cation conductivity limits for PC(L) and PC(H) at a drum pressure of 2500 psi (17.2 MPa) and 0.18% carryover has been performed and reported to be around 10 and (28) 25 µS/cm, respectively ; similar values are indicated in Figures 4-14 and 4-24. It will be noted that these values reflect considerably different capacities to cope with boiler contamination. The cation conductivity value for PC(L) (10 µS/cm) at a normal chloride limit of 150 ppb is about the same as for CT (see Figure 5-4) (about 9 µS/cm) at a normal chloride limit of 300 ppb, while the value for PC(H) is more than double that permissible with CT. The relative levels of boiler protection available with these treatments is reflected in both the continuum of treatments (Figure 2-1) and the boiler water treatment selection road map (Figure 2-8) presented and discussed in Section 2. As with other parameters, it is very important to establish customized cation conductivity limits for each boiler as just outlined. This is included in the road map to optimize PC and CT in Section 2.

3-65


Figure 3-22 Estimates of Cation Conductivity Against Pressure for PC(L). The lower curve reflects the conductivity if only phosphate is present in the boiler water. It is based on the phosphate concentration corresponding to the sodium concentration given in Figure 3-18 to ensure less than 2 ppb sodium in steam. The upper curve reflects the situation with additional chloride and sulfate present in the boiler water. Note: these curves must not be used for boiler control; they are provided here to assist the reader to develop a set of cation conductivity control curves in Section 4 for PC(L).

3-66


Figure 3-23 Estimates of Cation Conductivity Against Pressure for PC(H). The lower curve reflects the conductivity if only phosphate is present in the boiler water. It is based on the phosphate concentration corresponding to the sodium concentration given in Figure 3-17 to ensure less than 3 ppb sodium in steam. The upper curve reflects the situation with additional chloride and sulfate present in the boiler water. Note: these curves must not be used for boiler control; they are only provided here to assist the reader to develop a set of cation conductivity control curves in Section 4 for PC(H).

3-67


Figure 3-24 Cation Conductivity Versus Pressure for PC(L) only taking into account the maximum (normal) levels of chloride and sulfate (1:2 ratio). It must be noted that any contribution from phosphate has been excluded. Note: These curves must not be used for boiler control; they are provided here only to assist the reader to develop a set of cation conductivity control curves in Section 4 for PC(L).

3-68


Figure 3-25 Cation Conductivity Versus Pressure for PC(H) only taking into account the maximum (normal) levels of chloride and sulfate (1:2 ratio). It must be noted that any contribution from phosphate has been excluded. Note: These curves must not be used for boiler control; they are provided here only to assist the reader to develop a set of cation conductivity control curves in Section 4 for PC(H).

3-69


Cation Conductivity Limit in Steam Continuous measurement of cation conductivity at this location will provide a reliable indication of the presence of harmful salts and acids that are known to cause turbine corrosion. Elevated steam cation conductivities can be caused by an increased carryover and/or decomposition of alternate amines and organic reducing agents. 3.8.13 Total Organic Carbon (TOC) Organics may cause foaming in the boiler and high mechanical carryover, and may also affect the boiler and turbine corrosion, depending on the specific species or its thermal decomposition products that are present. Organics can form organic acids and carbon dioxide from decomposition in the boiler and typically are detected by an elevation of cation conductivity. Styrene sulfonic acids released from the oxidation of strong acid cation resin in condensate polishers constitute a source of interference for sulfate by IC; they can potentially contribute as is, or as their pyrolysis products, to corrosion product generation, release and transport. Organics introduced with the makeup water may release chloride or bromide upon decomposition depending on the disinfection agent used for pretreatment. Steam, Condensate, and Makeup TOC Target Values The target values are practical, achievable values based on the limited amount of data available for plants using ammonia and hydrazine. Plants using other, carbon-based amines for the control of pH and/or as reducing agent may not be able to achieve these target values. Since the contribution of the decomposition products of all the various carbon-based amines to cycle corrosion has not yet been determined, the target values have not been modified for units using these chemicals. 3.8.14 Iron and Copper Corrosion products form deposits on waterwall tubes under high heat flux conditions, which can trap and concentrate impurities, contributing to tube failures by caustic gouging, acid phosphate, hydrogen damage, and other forms of localized corrosion. Copper oxides act as oxidizers and sources of oxygen. Dissolved copper transported from the feedwater system deposits in the high pressure turbine, causing loss of turbine capacity and (24) efficiency due to pressure and flow changes. Chemical or mechanical cleaning is required to remove these copper deposits.(24,25) In boilers, co-deposition of copper and iron oxides on boiler tubes complicates boiler chemical cleaning. It should be recognized that sampling (analyzing) for copper may be difficult because of its deposition (up to 60-70% observed in superheated steam) (24) in the sampling lines when pressure and temperature are reduced.

3-70


For mixed-metallurgy systems it is necessary to maintain low levels of dissolved oxygen (< 10 ppb) in the feedwater and utilize a reducing agent in order to maintain reducing conditions to minimize copper transport. ORP monitoring must be used to ensure that a reducing condition is maintained at all times. ORP is a core parameter for mixed-metallurgy systems. A reducing environment (-300 to -350 mV) at the deaerator inlet is necessary to minimize copper transport. Total Iron and Copper Limits The total iron and copper limits are based on the minimum values practically achievable and on the solubility of iron in water and steam, and of copper and its oxides in superheated steam (to avoid deposition in the turbine). Meeting these limits requires that the feedwater chemistry be periodically and properly selected and optimized. Monitoring Considerations Although not core monitoring parameters, iron and copper are extremely important diagnostic parameters. Any routine monitoring or more rigorous monitoring campaign conducted to assess the performance of feedwater treatment program must include monitoring of iron and, in the case of mixed metallurgy cycles, copper. In general, use of on-line analyzers for monitoring metals transport is not extensive in fossil units. An on-line monitoring system that employs x-ray (29) fluorescence is available but costly and application has been confined mainly to nuclear units . On-line colorimetric analyzers have been developed and the general experience has been that they are difficult to maintain. Also, colorimetric methods are generally incapable of measuring iron and copper at the concentrations corresponding to target values for these metals in feedwater. Devices known as integrated corrosion product samplers provide for collection of a composite grab sample over a user-selected time interval. If desired, nonfilterable (soluble) and filterable (suspended) metals may be collected and analyzed separately. Methods of analysis typically used for iron and copper may be applied to samples collected with these devices. Commercially available samplers have been used extensively in both fossil and nuclear power plants; some organizations have designed and constructed their own samplers. On-line integrated sampling devices allow for concentration of filterable species on a 0.22 or 0.45 µm membrane, and for concentration of non-filterable colloidal and soluble species on cation (and, if needed, anion) ion exchange membranes. The accumulated waterborne particulates can then be analyzed in one of two ways. In the most common approach, filters are removed and the solids are digested by appropriate procedures for subsequent analysis by flame atomic absorption, graphite furnace atomic absorption, or atomic emission spectroscopy. Integrated sampling devices require access to a representative water sample, flowing at rates high enough (up to 2 liters or 0.53 gallons per minute) to avoid corrosion product particle dropout. Significant pressure drops in sample coolers should be minimized for the same reason.

3-71


The sample flow is split so that water is delivered continuously to the integrated sampling device at a constant flow rate in the range of about 50 to 200 cm3/min (0.013-0.053 gpm), and the remainder flows to the drain. The filtration unit holds a 0.45 µm membrane to collect the particulate matter and, beneath that, one or more cation ion exchange membranes are positioned to collect colloidal and soluble species passing through the 0.45 µm filter. A constant sample flow rate is maintained through filter and ion exchange membranes by incorporating a pressure regulator downstream of the filtration unit. The volume of water passing through the integrated sampling device is measured with a sample flow totalizer before the sample flows to the drain. For laboratory analyses of iron and copper, the filter membrane (carrying the particulates) and cation resin impregnated membranes (carrying the soluble metals) are digested separately, typically in hydrochloric-nitric acid mixtures to produce the integrated and solubilized samples needed for reliable chemical analysis. A concentration factor of 1000 or more is typical using this approach. Thus, these samples can be used to analyze separately for particulate loading and for dissolved corrosion products (iron and/or copper) in the original sample. The following noncolorimetric analysis methods may be used for analysis of iron copper in water.

•

Atomic Absorption Spectroscopy (AAS) (flame/furnace)

•

Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES)

•

Ion Chromatography (IC)

•

X-Ray Fluorescence (XRF)

•

Capillary Ion Analysis (CIA)

•

Capillary Electrophoresis (CE)

Of these, AAS (furnace) has been most widely used for analysis of iron and copper at the low levels normally present in condensate, feedwater and steam. IC also offers the potential to monitor these metals at low ppb levels. ICP-AES and AAS (flame) are better suited to levels of corrosion products which may be present in the boiler water. (Use of ICP/mass spectrometry (ICP/MS) permits detection of copper at sub ppb levels.) XRF is potentially applicable to low level analysis if the sample is concentrated prior to analysis. CIA and CE are ion analysis methods which may eventually be utilized for low level metals analysis. Each of these noncolorimetric techniques involves use of instrumentation not available at most fossil plants. AA, IC and ICP-AES are most likely to be present, especially when used for environmental compliance. All of these instruments are costly and should only be operated by (30) (31,32) are available for trained technicians or chemists. Standard methods and ASTM procedures analysis of copper by AA and ICP. Methods for the other techniques have been developed by the manufacturers.

3-72


There are numerous colorimetric methods available for analysis of soluble iron and copper in water. For iron, there are procedures based on phenanthroline compounds such as orthophenanthroline and bathophenanthroline, and specific triazine compounds. Colorimetric procedures for copper include methods employing compounds such a neocuproine, bathocuproine, bicinchoninate, and porphyrin. Some of the methods were at one time recognized by ASMT(31,32) but may no longer be supported. Some may be recognized as standard methods(30). Some of the methods in use have been developed by and/or are supported by manufacturers of (33) testing apparatus and reagents . Users considering application of any of these colorimetric methods for metals analysis are advised that they are not generally well suited to analysis of grab samples of feedwater, since limits of detection for these methods are usually well above target values for iron and copper. These methods may be of some benefit when testing for metals during startup or when analyzing metals concentrated by integration corrosion product sampler devices. Users should determine if the limits of detection, precision and accuracy, interferences and other characteristics of these methods are suitable for the intended application. 3.8.15 Air In-leakage Air in-leakage is the primary source of oxygen and carbon dioxide in the condensate and feedwater system; it is now designated as a core monitoring parameter (Table 3-2). Oxygen and carbon dioxide can promote corrosion, and carbon dioxide can degrade condensate polisher performance. High air in-leakage rates can also cause corrosion damage to the condenser shell and can result in reduced condenser vacuum, thereby reducing the efficiency of the cycle. Carbon dioxide also can have an effect on the pH of feedwater. The relationship between ammonia concentration and pH, and the effect of carbon dioxide on feedwater are shown in Figure 3-26.

3-73


Figure 3-26 Relationship between ammonia concentration and pH for various carbon dioxide concentrations in feedwater. The lowest curve represents zero CO2. The top curve represents 4 ppm CO2.

3-74


Air In-leakage Limit Because of the importance of air in-leakage for maintaining good cycle chemistry for all feedwater systems and materials, it is essential that measurements of this parameter be made using the most modern technology. The past practice of relying on rotameters to measure air inleakage is not considered acceptable today. Reference is made to Appendix C for a complete discussion of this topic. Also it should be noted that the emphasis in these revised PC and CT Guidelines has been placed on the 10 ppb oxygen limit at the condensate pump discharge. This change was made because of the uncertainty as to the location of any air in-leakage (above or below the water line).

3.9 GUIDELINES CUSTOMIZATION AND OPTIMIZATION Ideally, cycle chemistry control customization and optimization will start in the design phase. The actual cycle characteristics will be determined during commissioning, which would lead to customization of these guidelines. Selection, customization and optimization of the cycle water and steam chemistry includes the following steps:

•

selection of unit-specific boiler water and feedwater treatments,

•

determining optimum unit-specific control limits and alarm settings,

•

selection of unit-specific cycle chemistry process/control equipment,

•

development of operating guidelines (Episodes of contamination ingress must be dealt with quickly. The source of contamination should be eliminated and any repairs made as soon as possible. Procedures should detail what is to be done in a clear and concise way. General guidance is available in Section 6.),

•

determining sampling and instrumentation requirements,

•

commissioning,

•

specification of layup practices, and

•

determining optimum chemical cleaning criteria.

3-75


3.10 REFERENCES 1.

Cycle Chemistry Guidelines for Fossil Plants: Phosphate Treatment for Drum Units. EPRI, Palo Alto, CA: December 1994. TR-103665. (out of print)

2.

Sodium Hydroxide for Conditioning the Boiler Water of Drum-Type Boilers. EPRI, Palo Alto, CA: January 1995. TR-104007.

3.

Cycle Chemistry Guidelines for Fossil Plants: All-Volatile Treatment. EPRI, Palo Alto, CA: April 1996. TR-105041. (now out of print)

4.

Interim Consensus Guidelines on Fossil Plant Cycle Chemistry. EPRI, Palo Alto, CA: June 1986. CS-4629. (Now out of print.)

5.

Cycle Chemistry Guidelines for Fossil Plants: Oxygenated Treatment. EPRI, Palo Alto, CA: December 1994. TR-102285.

6.

Cycle Chemistry Guidelines for Fossil Plants: All-Volatile Treatment, EPRI, Palo Alto, CA: November 2002. 1004187.

7.

N.A. Styrikovich and O.I. Martynova, “Contaminations of the Steam in Boiling Reactors from Solutions of Water Impurities.” Soviet Atomic Energy, Vol. 15, 1963, pp. 917-921.

8.

O.I. Martynova, “Transport and Concentration Processes of Steam and Water Impurities in Steam Generating Systems,” in Water and Steam—Their Properties and Current Industrial Applications, pp. 547-562. Ed. by J. Straub and K. Scheffler for International Association for the Properties of Steam, Pergamon Press, New York, 1980.

9.

Steam, Chemistry, and Corrosion in the Phase Transition Zone of Steam Turbines, Volumes 1 and 2. EPRI, Palo Alto, CA: February 1999. TR-108184.

10.

Turbine Steam, Chemistry and Corrosion. EPRI, Palo Alto, CA: September 1997. TR108185.

11.

Turbine Steam Chemistry and Corrosion: Electrochemistry in LP Turbines. EPRI, Palo Alto, CA: September 2001. 1006283.

12.

Proceedings: Workshop on Corrosion of Steam Turbine Blading and Disks in the Phase Transition Zone. EPRI, Palo Alto, CA: November 1998. TR-111340.

13.

Steam Turbine Efficiency and Corrosion: Effects of Surface Finish, Deposits, and Moisture. EPRI, Palo Alto, CA: October 2001. 1003997.

14.

O. Jonas, “Transport of Ionic Impurities in Fossil and PWR Cycles—New Observations.” International Water Conference, Pittsburgh, PA. IWC-81-40. 1981.

3-76


15.

Behavior of Ammonium Salts in Steam Cycles. EPRI, Palo Alto, CA: December 1993. TR-102377.

16.

Volatility of Aqueous Sodium Hydroxide, Bisulfate and Sulfate. EPRI, Palo Alto, CA: February 1999. TR-105801.

17.

Vapor-Liquid Partitioning of Sulfuric Acid and Ammonium Sulfate. EPRI, Palo Alto, CA: February 1999. TR-112359.

18.

Volatility of Aqueous Acetic Acid, Formic Acid and Sodium Acetate. EPRI, Palo Alto, CA: July 2000. TR-113089.

19.

Behavior of Aqueous Electrolytes in Steam Cycles: The Solubility and Volatility of Cupric Oxide. EPRI, Palo Alto, CA: November 2000. 1000455.

20.

Vapor-Liquid Partitioning of Phosphoric Acid and Sodium Phosphates, EPRI, Palo Alto, CA: 2003. 1007291.

21.

The Volatility of Impurities in Water/Steam Cycles. EPRI, Palo Alto, CA: 2001. 1001042.

22.

Interim Cycle Chemistry Guidelines for Combined Cycle Heat Recovery Steam Generators (HRSGs), EPRI, Palo Alto, CA: November 1998. TR-110051.

23.

Development of a Steam Sampling System. EPRI, Palo Alto, CA: 1991. TR-100196.

24.

Guidelines for Copper in Fossil Plants. EPRI, Palo Alto, CA: 2000. 1000457.

25.

Guidelines for Chemical Cleaning of Conventional Fossil Plant Equipment, EPRI, Palo Alto, CA: November 2001. 1003994.

26.

B. Dooley, J. Mathews, R. Pate and J. Taylor, “Optimum Chemistry for ‘All-Ferrous’ Feedwater Systems: Why Use an Oxygen Scavenger?” IWC Paper-94-53. International Water Conference, Pittsburgh, PA, October 31/November 2, 1994.

27.

W. Bosselmann and R. Brunner, Conductivity of Phosphate Solutions, Kraftwerkstechnik, 58, February 1979, pp. 175-178.

28.

R.B. Dooley and K.J. Shields, “Cycle Chemistry for Conventional Fossil Plants and Combined Cycle/HRSGs”, Seventh International Conference on Cycle Chemistry in Fossil Plants, Houston, Texas, June 2003. EPRI, Palo Alto, CA: 2003. 1009194, December 2003.

29.

Reference Manual for On-Line Monitoring of Water Chemistry and Corrosion: 1998 Update, EPRI, Palo Alto, CA: April 1999. TR-112024.

3-77


30.

th Standard Methods for the Examination of Water and Wastewater, 20 Edition, American Public Health Association, American Waterworks Asociation and Water Environmental Federation, Washington, DC: 1998. (Part 3000, Metals)

31.

ASTM D 1068-95, “Standard Test Methods for Iron in Water”, ASTM, Philadelphia, PA: 1998. (Vol. 11.01).

32.

ASTM D 1688-95, “Standard Test Methods for Copper in Water,” ASTM, Philadelphia, PA: 1998. (Vol. 11.01)

33.

rd Water Analysis Handbook, 3 Edition, Hach Company, Loveland, CO: 1997. (Methods 8506 and 8143)

34.

D. Goldstrom and T. Robertson. “Low Phosphate-Low Sodium Hydroxide Treatment in 2600 psig Boilers at Salt River Project’s Coronado Station”, presented at the 50th International Water Conference, Pittsburgh, PA. IWC-89-46. 1989.

35.

G.J. Verib. “Sodium to Phosphate Ratios”, presented at the ASME Research Committee on Power Plant & Environmental Chemistry, San Diego, CA. March 2001.

36.

G.J. Verib. “History and Strategy of Phosphate Chemistry Control at Ohio Edison th Company”, presented at the 57 International Water Conference, Pittsburgh, PA. IWC96-23. 1996.

3-78


4 CYCLES WITH DRUM BOILERS OPERATING WITH THE PHOSPHATE CONTINUUM (PC)

4.1 INTRODUCTION TO PC A brief overview of the development of various phosphate treatments was provided in Section 1.3.1. There it was discussed how the Phosphate Continuum had grown out of the treatments in the previous EPRI Guidelines. The remarkable successes of the previous treatments (EPT and PT) and the remaining concerns were delineated. The main reasons for the development of the Phosphate Continuum (Figure 4-1) were very common operating deficiencies. It was indicated that very few organizations attempted to determine an equilibrium (minimum) level of phosphate for their boilers, and most often the level of phosphate used in high pressure boilers was very low (< 0.2 ppm). Additionally there has been the problem of corrosion (hydrogen damage) in cases of contaminant ingress, in conjunction with the application of the incorrect contaminant curves and ignoring of the ammonia/pH correction when determining the boiler water pH and the sodiumto-phosphate ratio. The PC treatment keeps the overriding criteria from the 1994 Phosphate Guidelines and also adds some new features to address the concerns. This combination leads to the PC treatment which:

• • • •

•

• •

Is applicable to a wide range of boilers (pressure) and contaminant levels. Has a wide range from low phosphate to around 10 ppm to assist in controlling hideout. Requires only the addition of TSP and NaOH, to ensure that acid phosphate corrosion will not occur, and is thus encompassed by Na:PO4 = 3 and TSP + 1 ppm NaOH (Figure 4-1). Covers the possibility of contaminant control at the low phosphate end by having a minimum of 0.2 ppm phosphate. This will prevent operation using less than 0.2 ppm, where there is essentially no protection from contamination. Together with the proper choice of contaminant control curves, this will prevent the onset of hydrogen damage. Ensures that a minimum boiler water pH (corrected for ammonia) of 9.0 is maintained to provide a reasonable level of boiler corrosion protection in the event that acid forming contaminant ingress is experienced; with low trisodium phosphate dosing rates some free sodium hydroxide must be present to maintain the pH at or above 9.0. Covers the high phosphate end for low pressure boilers and in cases where contaminants are likely/possible in high pressure boilers. (See Section 2.3 on Selection and Optimization of Boiler Water Treatments.) Continues the use of boiler water contaminant and pH control curves (Sections 4.2 and 4.3).

4-1

EPRI Licensed Material Cycles with Drum Boilers Operating with the Phosphate Continuum (PC)

(a)

(b) Figure 4-1 Schematic of Operating Range for PC (a is normal scale, b is log scale)

4-2


The PC treatment keeps the overriding criteria from the 1994 Phosphate Guidelines and also adds some new features to address the concerns. This combination leads to the PC treatment which:

•

Is applicable to a wide range of boilers (pressure) and contaminant levels.

•

Has a wide range from low phosphate to around 10 ppm to assist in controlling hideout.

•

Requires only the addition of TSP and NaOH to ensure that acid phosphate corrosion will not occur and is thus encompassed by Na:PO4 = 3 and TSP + 1 ppm NaOH (Figure 4-1)..

•

Covers the possibility of contaminant control at the low phosphate end by having a minimum of 0.2 ppm phosphate. This will prevent operation using less than 0.2 ppm, where there is essentially little protection from contamination. Together with the proper choice of contaminant control curves, this will prevent the onset of hydrogen damage.

•

Ensure that a minimum boiler water pH (corrected for ammonia) of 9.0 is maintained to provide a reasonable level of boiler corrosion protection in the event that acid forming contaminant ingress is experienced; with low trisodium phosphate dosing rates some free sodium hydroxide must be present to maintain the pH at or above 9.0.

•

Covers the high phosphate end for low pressure boilers and in cases where contaminants are likely/possible in high pressure boilers. (See Section 2.3 on Selection and Optimization of Boiler Water Treatments.)

•

Continues the use of boiler water contaminant and pH control curves (Sections 4.2 and 4.3).

Although PC is a continuum as shown in Figure 4-1, it needs to be recognized that there are two very different extremes of low (L) and high (H) phosphate levels. In the following two subsections (4.2 and 4.3) these regimes are PC(L) and PC(H) with a very loose demarcation around 3 ppm phosphate. To determine which of these treatments is applicable to a boiler, an operator will need to be continually cognizant of the following two main factors:

•

possible boiler contaminant levels, and

•

total carryover, particularly carryover of sodium into steam, as the sodium boiler water limits are the least conservative in this PC Guideline.

The process of selection of either PC(L) or PC(H) has been described in Section 2.

4-3


4.2 GUIDANCE AND GUIDELINES FOR PC(L) 4.2.1 Introduction to PC(L) For any phosphate treatment operating with low levels of phosphate within the Phosphate Continuum, there is a risk of corrosion, and thus for these PC guidelines it has been decided to provide Guidance and Guidelines for two extremes of PC—PC(L) and PC(H). The PC treatments at lower phosphate levels are discussed in this section, and those at higher phosphate levels are discussed in Section 4.3. There is not any distinct boundary between PC(L) and PC(H) within the PC. Obviously, the operator must be cognizant of the possible contaminant ingress frequency and quantity, and use this information in selecting and optimizing the exact phosphate treatment within the PC. Pressure-dependent normal target values and action levels for the phosphate content in the boiler water when applying PC(L) are depicted in Figure 4-2. In the phosphate region shown, ammonia in the boiler water has a marked influence on the pH particularly when operating at the Na-toPO4 ratio = 3 without any sodium hydroxide excess (Figure 4-3). Even when operating with trisodium phosphate + 1 ppm NaOH at phosphate contents less than 3 ppm (Figure 4-4), the effect of ammonia may not be neglected.

4-4


Figure 4-2 PC(L): Boiler Water Phosphate vs. Operating Pressure

4-5


Figure 4-3 Boiler Water pH vs. Phosphate at Various Ammonia Levels. Na/PO4 = 3. The lowest curve is at zero ammonia, the highest curve is at 2 ppm ammonia.

4-6


Figure 4-4 Boiler Water pH vs. Phosphate at Various Ammonia Levels. TSP + 1 ppm NaOH. The lowest curve is at zero ammonia. The highest curve is at 2 ppm ammonia.

4-7


Figure 4-5 may be used to determine the corrected pH, which is the pH attributable to the presence of phosphate and/or sodium hydroxide treatments, and to eliminate the influence of ammonia. While also applicable to PC(H) and even to CT, use of Figure 4-5 is most beneficial when employing PC(L). For routine surveillance and control of boiler water chemistry, the figure provides a reliable indication of the corrected pH. Users of Figure 4-5 should be cognizant of the following circumstances, which could lead to incorrect and misleading results:

•

Errors in measured data: the relationships shown in the figure assume that the measured pH and ammonia data are accurate. Any measurement errors will naturally influence the pH correction factor determined by the figure and the resulting corrected pH. Data validity is an issue of general concern and information pertinent to data quality are provided in Appendix E.

•

Correction assessment using Figure 4-5 is appropriate only when the boiler water contains at least the requisite minimum level of treatment chemical and this should be verified prior to making the pH correction. Under circumstances of confirmed loss of treatment chemicals, determining the reason for reduced treatment chemical inventory and restoration to normal levels is the more important issue.

Thus, when correctly applying PC(L) (using the ammonia/pH correction), i.e., when operating correctly with Na:PO4 ratio ≥ 3, establishing an equilibrium phosphate concentration is superfluous and should not be undertaken. There are three boiler water parameters that have to be monitored and only two that will be used for routine operation and optimization of PC(L): Ammonia

for making the pH/ammonia correction.

pH

as the primary boiler protection parameter; with PC(L) the measured pH must be corrected for the influence of ammonia;

PO4

as the secondary boiler protection parameter. The actual level of the phosphate concentration depends on the actual cycle contamination.

As is discussed in the next Sub-Section, cation conductivity should be used as the main control of contaminants in the boiler water for PC(L) in conjunction with pH and phosphate. This will enable effective diagnosis of any operational problems. Any phosphate treatment, even PC(L), is necessarily accompanied with phosphate hideout and hideout return, but when trisodium phosphate is the only phosphate added to the boiler, any hideout will not result in corrosion. For this reason, trials in high pressure boilers to establish the equilibrium concentration with nearly zero (or negligible) hideout will result in very low phosphate concentrations being applied. This, in connection with neglecting the necessary ammonia/pH correction, has led to operation with an actual Na:PO4 ratio < 3, often very much less than 3. This incorrect phosphate control together with a small ingress of contaminants has resulted in serious boiler tube failures due to hydrogen damage. This important feature of PC(L) is discussed further in Sections 3.8.2, 3.8.9 and 3.8.10.

4-8


Figure 4-5 Relationship Between Corrected Boiler Water pH, Measured Boiler Water pH and Measured Ammonia Concentration. The lines on the figure increase in 0.1 ppm ammonia increments from 0 to 2 ppm. To determine the corrected pH, find the measured pH on the x-axis and the point at which a vertical line intersects the measured ammonia concentration. This point, extended horizontally to the y-axis, indicates the corresponding corrected pH value.

4-9


Cycle chemistry target values and action levels are presented in this section for each monitoring point for plants having drum boilers operating on PC(L). A target, or normal value, and as many as three action levels are given for each parameter at each monitoring point. This section is intended to be used by operating staff and management personnel. 4.2.2 Target Values for PC(L) The sample points, monitoring parameters, target values, and action levels were developed for those plants that have reheat. To customize these values for plants without reheat, the reader is referenced to Section 3.8. The following figures show the target values and action levels for each sample point and monitoring parameter:

•

Figure 4-6 for all-ferrous systems operating with oxidizing conditions, AVT(O), (but not on OT), and with no reducing agent.

•

Figure 4-7 for mixed-metallurgy and all-ferrous systems operating under reducing conditions, AVT(R). These systems use a reducing agent and provide (for mixed-metallurgy systems) the reducing conditions necessary to minimize copper transport.

Figure 4-7 features the use of ORP instrumentation at the deaerator inlet to ensure that reducing conditions are maintained at all times. ORP is not specified as a core parameter for Figure 4.6 (oxidizing regime). Figures 4-6 and 4-7 list target values (N) and from one to three action levels for a compendium of sampling points around the cycle. The steam values are more restrictive than featured in previous Guidelines in order to provide enhanced turbine protection (Section 1.2.3) (steam limits of 2 ppb sodium, chloride and sulfate versus 3 ppb in previous editions). The limit for silica in the steam remains the same at 10 ppb. Research has indicated that, for mixed-metallurgy systems (with strict air in-leakage and oxygen control, the addition of a reducing agent, and with reducing conditions), the pH range at the economizer inlet can be increased from 8.8 to 9.2 (previously) to 9.0 to 9.3. This increase in pH will reduce the transport of iron in the cycle, while minimizing the copper bearing materials from corrosion. It must be recognized that the ORP needs to remain in the reducing regime for these mixed-metallurgy systems. Simply changing the pH will not suffice if the air in-leakage remains high.

4-10


Figure 4-6 Cycle Chemistry Diagram for PC(L) with AVT(O) feedwater. All-ferrous metallurgy in feedwater system. Operating with an oxidizing environment (no reducing agent).

4-11


Figure 4-7 Cycle Chemistry Diagram for PC(L) with AVT(R) feedwater. All-ferrous and mixed-metallurgy feedwater systems. Operating with a reducing environment (reducing agent added).

4-12


There are a number of features to keep in mind when adapting Figures 4-6 and 4-7 to an individual plant. Particular emphasis has been placed in the development of these new guideline cycle diagrams to make control of a unit as simple as possible, and to minimize repetitive grab sampling.

•

The Core Parameters (Table 3-2) are indicated by an asterisk (*). These instruments should be monitored continuously and on-line. These have a “C” for “continuous” in the sample column.

•

There are a number of other parameters, which have also been designated by a “C” as they provide useful confirmation of unit condition, but are not regarded as “Core” instruments. These parameters include silica and sodium in makeup, specific conductivity at the economizer inlet, and boiler water sodium.

•

There are a number of other parameters on these diagrams, which have been designated in this guideline as “T” for troubleshooting. It is only necessary to monitor these if one of the continuous “Core” instruments provides an indication that the parameter is out of its “normal” guideline value. In most cases, this will be accomplished by grab sample; however, in some cases an organization might have decided to install a continuous analyzer. In the case of iron and copper at the economizer inlet, it should be recognized that in a well operated plant it should not be necessary to monitor these parameters more frequently than once or twice a year; and then only to confirm the feedwater regime remains optimized.

Other points of interest on Figures 4-6 and 4-7 are:

•

Dissolved oxygen at the CPD should be controlled to less than 10 ppb. This is a much better control for the feedwater than air in-leakage into the condenser and is critical to be able to control copper corrosion in the feedwater for mixed-metallurgy feedwater systems, as well as oxidizing environments for all-ferrous feedwater systems.

•

A limit for copper at the economizer inlet remains on Figure 4-6 for those units with copper based condenser tubing.

•

Reducing agent (hydrazine preferred) feed in Figure 4-7 is a requirement, whereas in Figure 4-6 it is not required nor is it necessary if the unit is to be operated with AVT(O). For both AVT(R) and AVT(O) rigorous air in-leakage control must be maintained to ensure oxygen levels at the CPD of less than 10 ppb.

As discussed in Section 3 (Table 3-2), a new suite of core parameters has been developed for Figures 4-6 and 4-7. The following core parameters are applicable for PC(L):

•

Makeup treatment system effluent specific conductivity is required to detect malfunction of the makeup treatment system.

•

Sodium and cation conductivity at the condensate pump discharge provide cycle contamination control from condenser leaks and makeup malfunctions.

•

Dissolved oxygen at the condensate pump discharge ensures the optimum feedwater potential can be controlled. It also provides the important indication of air in-leakage into the cycle.

4-13


•

Deaerator inlet ORP (Figure 4-7) ensures reducing conditions are maintained to minimize copper transport. Reference to Appendix B will illustrate how reducing ORP levels can vary dependent on a number of variables and does not always need to be in the range of -300 to 350mV to be reducing.

•

Economizer inlet sodium (or condensate polisher effluent sodium) on units with a condensate polisher reveals condensate polisher malfunction and guards against cycle contamination.

•

Economizer inlet dissolved oxygen and cation conductivity.

•

Boiler water pH (corrected for ammonia) and cation conductivity. Cation conductivity is the main control for boiler water contaminants, and thus, corrosion.

•

Boiler water sodium and phosphate are necessary to ensure proper feed of trisodium phosphate and caustic (to 1.0 ppm if used). Boiler water specific conductivity is a further check on boiler water purity and also reflects the concentration of ammonia being used, chlorides, sulfate and caustic (to 1.0 ppm if used).

•

Saturated steam carryover is required to be determined on a routine/ periodic basis (every six months or after an outage or when drum work is conducted) to prevent excessive contamination of the turbine caused by boiler drum internal separation failure, foaming, high boiler water levels, etc.

•

Air removal system exhaust air in-leakage is required to prevent serious corrosion via excess cycle dissolved oxygen and carbon dioxide.

The core parameters shown in Figures 4-6 and 4-7 are the minimum required for routine chemistry monitoring and provide basic input data for the EPRI computer program “ChemExpert”. The basic requirement for every unit is that a minimum level of instruments should be installed that can uniquely identify any problem on the unit. A series of curves have been developed which present smooth, continuous relationships of boiler water target values as a function of boiler pressure. These curves have been designed to provide boiler corrosion protection, but as indicated in Section 3.3.1 they have been validated to ensure they are below the boiler water contaminant values which would cause the steam values to be exceeded. Thus these curves provide protection to the boiler and the turbine. The following boiler water limits are presented:

•

sodium versus operating pressure (Figure 4-8)

•

pH versus operating pressure (Figure 4-9)

•

phosphate versus operating pressure (Figure 4-2)

•

chloride versus operating pressure (Figure 4-10)

•

sulfate versus operating pressure (Figure 4-11)

•

silica versus operating pressure (Figure 4-12)

4-14


Figure 4-8 PC(L): Boiler Water Sodium vs. Operating Pressure

4-15


Figure 4-9 PC(L): Boiler Water pH vs. Operating Pressure. pH due to trisodium phosphate (3.8 to 1.0 ppm phosphate from Figure 4-2)

4-16


Figure 4-10 PC(L): Boiler Water Chloride vs. Operating Pressure

4-17


Figure 4-11 PC(L): Boiler Water Sulfate vs. Operating Pressure

4-18


Figure 4-12 PC(L): Boiler Water Silica vs. Operating Pressure

4-19


Copies of these figures, modified to reflect specific unit characteristics by monitoring (see Section 2), should be included in the plant operating procedures and prominently displayed in the control room, water and steam sample room, and the chemistry laboratory. The specific and, in particular, cation conductivity are very important core parameters when operating on PC(L). Extensive discussion took place on these parameters in Sections 3.8.11 and 3.8.12. It was illustrated how these conductivities depend on the level of phosphate and any contaminants (chloride and sulfate). The effect of phosphate by itself, and of chloride/sulfate was shown in a series of figures. Those organizations wishing to establish specific and cation conductivity control curves are referenced to the suggested procedures in Sections 3.8.11 and 3.8.12. Copies of these figures, modified to reflect specific unit characteristics by monitoring (see Section 2), should then be included in the plant operating procedures and prominently displayed in the control room, water and steam sample room, and the chemistry laboratory. As an example, two figures are included in this PC(L) guideline section to illustrate the sensitivity of specific and cation conductivity to the balance between phosphate and chloride/sulfate (Figures 4-13 and 4-14). Both of these figures show the specific and cation conductivity due only to phosphate, and also due to additional amounts of chloride and sulfate. To assist operators in using these two figures or developing a set of curves for their boiler, a few words are provided on how these curves were derived. For specific conductivity (Figure 4-13). The contribution of phosphate at the normal limits shown in Figure 4-2 has been converted into specific conductivity using the relationship shown in Figure 3-19 for Na3PO4. The contributions of chloride and sulfate at the normal limits shown in Figures 4-10 and 4-11 have been converted to specific conductivity. The top dashed curve in Figure 4-13 represents the “envelope” corresponding to the upper curve in Figure 3-20. There are no contributions in Figure 4-13 from NaOH or ammonia. For cation conductivity (Figure 4-14). The contribution of phosphate at the normal limits shown in Figure 4-2 has been converted into cation conductivity using the relationship shown in Figure 3-19 for H3PO4. The contributions of chloride and sulfate at the normal limits in Figures 4-10 and 4-11 have been converted to cation conductivity. The top dashed curve in Figure 4-14 corresponds to the upper curve in Figure 3-22.

4-20


Figure 4-13 PC(L): Drum Boiler Water Specific Conductivity vs. Operating Pressure showing specific conductivity due to phosphate alone, and due to phosphate plus various amounts of chloride and sulfate. It must be noted that this figure is only an example to illustrate the effects of phosphate, chloride and sulfate. Each operator will need to develop unique curves for each boiler as described in the text of Sections 4.2.2 and 3.8.11.

4-21


Figure 4-14 PC(L): Drum Boiler Water Cation Conductivity vs. Operating Pressure showing cation conductivity due to phosphate alone, and due to phosphate plus various amounts of chloride and sulfate. It must be noted that this figure is only an example to illustrate the effects of phosphate, chloride and sulfate. Each operator will need to develop unique curves for each boiler as described in the text of Sections 4.2.2 and 3.8.12.

4-22


If control curves such as illustrated in Figures 4-13 and 4-14 are derived, then specific and cation conductivity can be used as the main control for contaminants in the boiler water for PC(L), but they should be utilized together with other monitored parameters (pH and phosphate) to diagnose effectively various potential operational problems. For example, from a baseline steady load and uniform boiler water specific and cation conductivity, a sudden or continuing change in either or both of these parameters can be the basis for determining the source of the potential problem and of initiating the appropriate corrective action. The four case studies in the PC(H) section (Section 4.3.1) illustrate the diagnostic indications. Boiler water pH must be kept within a narrow alkaline range to maintain the protective layer of magnetite on the internal boiler water tube surfaces. Contaminants such as chlorides, sulfates and silica must be maintained at reasonable low concentrations to prevent corrosion or excessive deposits within the boiler tubes or downstream steam circuits. In PC(L), boiler water alkalinity is maintained with an adequate concentration of a blend of sodium hydroxide and trisodium phosphate. The rationale for this treatment is described in the introduction to this section and in Section 1.3. 4.2.3 Normal Operation for PC(L) Once the optimum phosphate level based on the risk of possible contamination has been determined for the boiler, then the unique operating limits for the unit should be determined. The cycle chemistry diagrams and contaminant curves, that are provided in Figure 4-2 and Figures 46 to 4-12, should be regarded purely as a guide, and that need to be customized for each unit. Here reference should be made to the roadmap in Section 2. The importance of this procedure cannot be overemphasized: it is very important to measure the carryover from the boiler drum into the steam for each unit and to use this information to develop specific unit operating chemical guidelines. It is of particular concern with PC(L), operating with a very small level of free hydroxide, that particular attention is paid to the steam sodium levels. During normal operation, only trisodium phosphate and sodium hydroxide should be added to the boiler water. Under PC(L), it is necessary to maintain some free sodium hydroxide in the boiler water to maintain the ammonia corrected pH above 9.0, particularly if the normal phosphate concentration is under 1 ppm. In units where sodium leakage from makeup systems and polishers is well controlled, there will be generally a need for sodium hydroxide dosing to avoid boiler corrosion if the boiler operates with low concentrations of phosphate. The use of the boiler blowdown should be minimized. If this operation can be maintained then the addition of chemicals to the unit will be minimized. If blowdown is used for some reason (e.g., for reducing the contaminant concentration in the boiler water), additional dosing of trisodium phosphate and sodium hydroxide is required to keep the boiler water pH in the optimum region and to ensure that the sodium-to-phosphate ratio actually is ≥ 3.

4-23


Continual Check of pH and Phosphate Level The boiler water should be continuously monitored during normal operation for pH and phosphate, and particularly during periods of rapid load reduction. Since the chemical requirements of a unit can change over time, significant shifts of the phosphate concentration can and do occur. Changes in the phosphate concentration may also be caused by contaminant ingress into the boiler. It is also important during these periodic checks to ensure that the sodium levels in the boiler water are consistent with the pressure/concentration curves for sodium (Figure 4-8) to prevent possible damage to the turbine from mechanical and vaporous carryover of sodium compounds. The free NaOH should be restricted to a range of 0.2 to 1.0 ppm. Section 3.8.1 should be referred to for the effects of sodium. The operator needs to be cognizant of a) the effects of ammonia in the PC(L) treatment, and b) the protection of the boiler from corrosion through cation conductivity (Sections 3.8.12 and 4.2.2 and associated figures). The effects of ammonia and appropriate correction can be conducted using Figure 4-5. 4.2.4 Operation During Upset Conditions The following discussion is concerned with PC(L) during non-normal conditions; specifically what modifications to PC(L) are necessary during these contamination episodes and how should an operator react? When contamination occurs, the boiler blowdown should be opened and the source of contamination isolated quickly to minimize chemical upset. In parallel with boiler blowdown, sodium hydroxide dosing is required for ensuring the correct pH level in the boiler water. The cation conductivity should be referenced to a unit specific version of Figure 4-14, and if necessary, the concentrations of chloride and sulfate in the boiler can be determined using an ion chromatograph. If the concentrations exceed limits (such as those given in Figures 4-10 and 411) the pressure of the unit should be dropped to stay within the sliding chloride and sulfate pressure limits adopted for the plant. In order to provide chemical protection during high contamination episodes, the phosphate concentration can be adequately increased. Using sodium hydroxide, the pH should be maintained within the normal limits of 9.0 to 9.6 adjusted for the presence of ammonia or to the unit specific phosphate/pH diagram corrected for ammonia. With PC(L), a measured pH reading of less than 9.0 requires action, regardless of the ammonia concentration, since the corrected pH (using Figure 4-5) will clearly be less than 9.0 under such circumstances. Boiler water conductivity readings should also be taken to determine if the low pH is due simply to loss of treatment chemical (typically via blowdown and possibly involving malfunction or failure of the chemical feed equipment) or involves contamination. Boiler conductivity and other plant instrumentation readings should also be checked to see if they indicate the possibility of an ongoing contamination event.

4-24


In the absence of boiler water contamination, the degree of corrosion protection afforded by lower than desired levels of treatment (and commensurate low corrected pH values) may be completely satisfactory, however, restoration of normal chemistry is required as quickly as possible to ensure that the boiler will be protected in the event contamination is experienced. With PC(L), if the absence of contamination is confirmed and the phosphate residual is within the customized limits of the unit, addition of caustic is needed to raise the pH. On the other hand, addition of trisodium phosphate should be made first if (again, in the confirmed absence of contamination) the phosphate level is less than the minimum allowable concentration; a subsequent addition of caustic would be needed if the trisodium phosphate addition does not restore the corrected pH value to normal limits. The more serious situation is the one involving the presence of boiler contamination, typically confirmed by increased boiler water conductivity readings. Operation of PC(L) without pH correction, and without surveillance and control of the conductivity, can give a false sense of security with respect to corrosion protection. When the corrected pH and treatment chemical levels are within expected ranges for control, the boiler water can provide corrosion protection as indicated in the impurity and conductivity control curves (Figures 4-10, 4-11, 4-13 and 4-14); corrosion protection is compromised at higher boiler water impurity levels. However, when the corrected pH is less than normal, this means that the levels of treatment chemicals are also lower than needed to provide corrosion protection at the levels indicated in the impurity and conductivity control curves; in the event of complete loss of chemical treatment, the boiler corrosion protection would be the same as for all-volatile treatment. If the measured pH is rapidly falling and at risk of reaching the immediate shutdown level of 8, there is clearly no point in making the pH correction until and unless the boiler chemistry can be stabilized quickly and without the need for immediate shutdown. For example, it may be possible to keep the unit in service by reducing pressure and/or load and restoring program treatment levels, in which case correction of the measured pH for ammonia should be resumed. Loss of the ammonia feed system could conceivably lead to a rapid yet relatively minor reduction in the measured boiler water pH. However, the corrected pH should remain at an acceptable level providing the cycle is free of significant contamination and levels of phosphate and/or sodium hydroxide treatments are in expected ranges for control. Under such circumstances, the boiler corrosion protection is maintained but the ammonia feed system must be returned to service as soon as possible to control the feedwater pH at the optimum level needed to minimize corrosion of metals from the waterside surfaces of preboiler components and piping and subsequent transport of these metals to the boiler. If at any time the boiler blowdown pH falls to below 8, the unit should be taken off line and drained. In the case of a longterm operation with markedly increased contamination, chemical cleaning should be considered. After the source of contamination has been isolated and the concentration of the contaminants in the boiler blowdown is within limits, the unit can be converted back to standard PC(L) chemistry.

4-25


4.2.5 Unit Startup and Shutdown Cold Startup In order to give added protection against contaminant ingress during the initial stages of operation, the pH and the phosphate concentration may be adjusted to the maximum concentration permitted by this guideline. The phosphate level may even go into the PC(H) region. The pH should be maintained at the upper pH limit, adjusted for ammonia. During startup, this adjustment will likely be significant if the use of excessive reducing agent is necessitated by high feedwater oxygen levels. Failure to account for positive bias of the pH due to the presence of ammonia will lead to inadequate chemistry; the boiler will actually operate below the PC(L) pH range. This will increase the risk of boiler tube corrosion and failures. The unit should be converted to standard chemistry when the first main burner is brought on line assuming contamination does not exceed the pressure/contaminant curves. Heavy boiler blowdown can be maintained throughout the startup until contaminant levels are well within limits at which time the boiler blowdown should be isolated. The sodium hydroxide dosing has to ensure that the necessary pH is maintained even in periods with an increased blowdown. The boiler blowdown remains isolated during normal operation except in periods of abnormal contamination. Unit Trip with Normal Contaminant Concentrations The pH and the phosphate concentration should be monitored closely when a unit trips. A sudden release of phosphate (hideout return) may indicate that the phosphate concentration used during normal operation is too high. As a rule, any phosphate hideout return causes a pH increase when only trisodium phosphate is added to boiler. Unit Shutdown During shutdown, the pH may also increase as the result of the phosphate hideout return. The pH should be maintained within the selected control range even during shutdown. If necessary, sodium hydroxide can be added to drive the pH into the acceptable range to protect the boiler.

4.3 GUIDANCE AND GUIDELINES FOR PC(H) Cycle chemistry target values and action levels are presented in this section for each monitoring point for plants having drum boilers on PC(H). A target, or normal value, and as many as three action levels are given for each parameter at each monitoring point. This section is intended to be used by operating staff and management personnel.

4-26


4.3.1 Target Values for PC(H) Cycle chemistry diagrams have been developed for reheat drum units operating on PC(H). To customize these values for plants without reheat, the reader is referenced to Section 3.8. Two separate diagrams have been provided:

•

Figure 4-15 for all-ferrous systems operating with oxidizing conditions, AVT(O), (but not on oxygenated treatment) and with no reducing agent feed, which minimizes iron transport in the cycle.

•

Figure 4-16 for mixed-metallurgy and all-ferrous systems operating under reducing conditions, AVT(R). These systems use a reducing agent and provide (for mixed-metallurgy systems) the reducing conditions necessary to minimize copper transport.

Figure 4-16 features the use of ORP instrumentation at the deaerator inlet to ensure that reducing conditions are maintained at all times. ORP is not specified as a core parameter for Figure 4-15 (oxidizing regime). Figures 4-15 and 4-16 list target values (N) and from one to three action levels for a compendium of sampling points around the cycle. The steam values are more restrictive than featured in previous Guidelines, in order to maintain enhanced turbine protection (Section 1.2.3), with steam limits of 3 ppb sodium, versus 5 ppb in previous editions. The limit for silica in the steam remains the same at 10 ppb, as does chloride and sulfate at 3 ppb each. Research has indicated that, for mixed-metallurgy systems (with strict air in-leakage and oxygen control, the addition of a reducing agent, and with reducing conditions), the pH range at the economizer inlet can be increased from 8.8 to 9.2 (previously) to 9.0 to 9.3. This increase in pH will reduce the transport of iron in the cycle, while minimizing corrosion of the copper bearing materials. It needs to be recognized that the ORP needs to remain in the reducing regime for these mixed-metallurgy systems. Simply changing the pH will not suffice if the air in-leakage remains high. There are a number of features to keep in mind when adapting Figures 4-15 and 4-16 to an individual plant. Particular emphasis has been placed in the development of these new guideline cycle diagrams to make control of a unit as simple as possible, and to minimize repetitive grab sampling.

•

The Core Parameters (Table 3-2) are indicated by an asterisk (*). These instruments should be monitored continuously and on-line. These have a “C” for “continuous” in the sample column.

•

There are a number of other parameters, which have also been designated by a “C” as they provide useful confirmation of unit condition, but are not regarded as “Core” instruments. These parameters include: silica and sodium in makeup, specific conductivity at the economizer inlet, and boiler water sodium.

4-27


•

There are a number of other parameters on these diagrams, which have been designated in this guideline as “T” for troubleshooting. It is only necessary to monitor these if one of the continuous “Core” instruments provides an indication that the parameter is out of its “normal” guideline value. In most cases, this will be accomplished by grab sample; however, in some cases an organization might have decided to install a continuous analyzer. In the case of iron and copper at the economizer inlet, it should be recognized that in a well operated plant it should not be necessary to monitor these parameters more frequently than once or twice a year; and then only to confirm the feedwater regime remains optimized.

Other points of interest on Figures 4-15 and 4-16 are:

•

Dissolved oxygen at the CPD should be controlled to less than 10 ppb. This is a much better control for the feedwater than air in-leakage into the condenser, and is critical to be able to control copper corrosion in the feedwater for mixed-metallurgy feedwater systems, as well as oxidizing environments for all-ferrous feedwater systems.

•

A limit for copper at the economizer inlet remains on Figure 4-15 for those units with copper based condenser tubing.

•

Reducing agent (hydrazine preferred) feed in Figure 4-16 is a requirement, whereas in Figure 4-15 it is not required nor is it necessary if the unit is to be operated with AVT(O). For both AVT(R) and AVT(O) rigorous air in-leakage control must be maintained to ensure oxygen levels at the CPD of less than 10 ppb.

As discussed in Section 3 (Table 3-2), a new suite of core parameters has been developed for Figures 4-15 and 4-16. The following core parameters are applicable for PC(H):

•


•

Sodium and cation conductivity at the condensate pump discharge provide cycle contamination control from condenser leaks and makeup malfunctions.

•

Dissolved oxygen at the condensate pump discharge ensures the optimum feedwater potential can be controlled. It also provides the important indication of air in-leakage into the cycle.

•

Deaerator inlet ORP (Figure 4-16) ensures reducing conditions are maintained to minimize copper transport. Reference to Appendix B will provide an understanding of how reducing ORP is dependent on a number of variables and does not always need to be in the range of 300 to -350 mV to be reducing.

•

Economizer inlet sodium (or condensate polisher effluent sodium) on units with a condensate polisher reveals condensate polisher malfunction and guards against cycle contamination.

•

Economizer inlet dissolved oxygen and cation conductivity.

•

Boiler water pH and cation conductivity. Cation conductivity is the main control for boiler water contaminants, and thus, corrosion.

•

Boiler water sodium and phosphate are necessary to ensure proper feed of trisodium phosphate and caustic (to 1.0 ppm if used). Boiler water specific conductivity is a further check on boiler water purity and also reflects the concentration of ammonia being used, chlorides, sulfate and caustic, (to 1.0 ppm if used).

4-28


Figure 4-15 Cycle Chemistry Diagram for PC(H) with AVT(O) Feedwater. All-ferrous metallurgy in feedwater system. Operating with an oxidizing environment (no reducing agent).

4-29


Figure 4-16 Cycle Chemistry Diagram for PC(H) with AVT(R) Feedwater. All-ferrous and mixed-metallurgy feedwater systems. Operating with a reducing environment (reducing agent added).

4-30


•

Saturated steam carryover is required to be determined on a routine/ periodic basis (every six months or after an outage or when drum work is conducted) to prevent excessive contamination of the turbine caused by boiler drum internal separation failure, foaming, high boiler water levels, etc.

•

Air removal system exhaust air in-leakage is required to prevent serious corrosion via excess cycle dissolved oxygen and carbon dioxide.

With PC(H), the influence of ammonia in the boiler water is generally less significant than it is for PC(L), as indicated in Figures 4-3 and 4-4. The corrected pH may be determined using Figure 4-5. The core parameters shown in Figures 4-15 and 4-16 are the minimum required for routine chemistry monitoring and provide basic input data for the EPRI computer program ChemExpert. The basic requirement for every unit is that a minimum level of instruments should be installed that can uniquely identify any problem on the unit. A series of curves have been developed which present smooth, continuous relationships of boiler water target values as a function of boiler pressure. These curves have been designed to provide boiler corrosion protection, but as indicated in Section 3.3.1 they have been validated to ensure they are below the boiler water contaminant values which would cause the steam values to be exceeded. Thus these curves provide protection to the boiler and the turbine. The following boiler water limits are presented:

•

sodium versus operating pressure (Figure 4-17)

•

phosphate versus operating pressure (Figure 4-18)

•


•

chloride versus operating pressure (Figure 4-20)

•

sulfate versus operating pressure (Figure 4-21)

•

silica versus operating pressure (Figure 4-22)

4-31


Figure 4-17 PC(H): Boiler Water Sodium vs. Operating Pressure

4-32


Figure 4-18 PC(H): Boiler Water Phosphate vs. Operating Pressure

4-33


Figure 4-19 PC(H): Boiler Water pH vs. Operating Pressure. pH due to trisodium phosphate (5.8 to 1.4 ppm from Figure 4-18).

4-34


Figure 4-20 PC(H): Boiler Water Chloride vs. Operating Pressure

4-35


Figure 4-21 PC(H): Boiler Water Sulfate vs. Operating Pressure

4-36


Figure 4-22 PC(H): Boiler Water Silica vs. Operating Pressure

4-37


The sodium vs. pressure curve, Figure 4-17 is based on the mechanical carryover curve presented in Figure 3-2 plus the vaporous carryover from the research conducted by EPRI and discussed in Section 3.3.1. As indicated in Figure 3-9, vaporous carryover remains unimportant until pressures reach 2500 psi and above. Even at these high pressures, mechanical carryover is the dominant factor in determining sodium in the steam. This research finding emphasizes the necessity of having an accurate measurement of total carryover from each boiler drum to prevent excessive carryover of sodium to the steam The mechanical carryover curve, Figure 3-2, contains a safety factor of 2 to provide conservatism. Thus, the sodium curve (Figure 4-17) illustrates that relatively low boiler water sodium levels must be maintained in the boiler to prevent sodium steam limits from being exceeded. If a particular operating boiler has mechanical carryover less than indicated on Figure 3-2, sodium levels in the boiler may be increased above those in Figure 4-17 without exceeding sodium steam limits. Conversely, if actual mechanical carryover is greater than shown on Figure 3-2, sodium limits in the boiler must be reduced to maintain proper sodium steam limits. These considerations make it mandatory that total carryover be measured in each boiler with some frequency (twice per year at least). The feed of pure trisodium phosphate will, obviously, result in a molar ratio of sodium phosphate of 3.0. However, contaminants, such as sodium chloride and sodium sulfate, as well as the feed of up to 1.0 ppm of caustic will elevate the sodium values in the boiler water resulting in sodium to phosphate ratios greater than 3.0. The sodium vs. pressure curve, Figure 4-17, represents the maximum sodium in the boiler water to maintain a steam limit of 3 ppb, irrespective of whether the sodium is from trisodium phosphate, sodium chloride, sodium sulfate or caustic. As discussed later in this section, the curves for specific and cation conductivity reflect sodium from these other sources. These conductivity curves should be used in conjunction with measurements of sodium in the boiler water to ensure that the sodium levels in Figure 4-17 are not exceeded. The chloride and sulfate curves, Figures 4-20 and 4-21, are based on protecting the boiler from corrosion. These curves were developed from extensive experience with operating boilers on (2) phosphate treatment (PT) and caustic treatment (CT) which is thoroughly discussed in Section 5. Sulfate was recognized as a lesser risk than chloride in initiating corrosion in boilers, so the concentration of sulfate has been taken as twice that of chloride. These curves have also been validated to ensure they are below the boiler water contaminant values, which would cause the steam values to be exceeded. (Compare to Figures 3-11 and 3-13.) Thus these curves provide protection to the boiler and turbine. The specific and, in particular, cation conductivity are very important parameters when operating with PC(H). Conductivity methods utilize simple and reliable instrumentation and are, therefore, the preferred methods of evaluating and controlling boiler water chemistry. Conductivity measurements should, of course, be supplemented by measurements of other control parameters such as pH, sodium and phosphate to ensure proper control of the boiler at all times. Extensive discussion took place on these parameters in Sections 3.8.11 and 3.8.12. It was illustrated how these conductivities depend on the level of phosphate and any contaminants (chloride and sulfate). The effect of phosphate by itself, and of chloride/sulfate was shown in a series of figures. 4-38


Those organizations wishing to establish specific and cation conductivity control curves are referenced to the suggested procedures in Sections 3.8.11 and 3.8.12. Copies of these figures, modified to reflect specific unit characteristics by monitoring (Section 2) should be then included in the plant operating procedures and prominently displayed in the control room, water and steam sample room, and the chemistry laboratory. As an example, two figures are included in this PC(H) guideline section to illustrate the sensitivity of specific and cation conductivity to the balance between phosphate and chloride/sulfate (Figures 4-23 and 4-24). Both of these figures show specific and cation conductivity due only to phosphate, and also due to additional amounts of chloride and sulfate. To assist operators in using these two figures or developing a set of curves for their boiler, a few words are provided on how these curves were derived. For specific conductivity (Figure 4-23). The contribution of phosphate at the normal limits shown in Figure 4-18 has been converted into specific conductivity using the relationship shown in Figure 3-19 for Na3PO4. The contributions of chloride and sulfate at the normal limits shown in Figures 4-20 and 4-21 have been converted to specific conductivity. The top dashed curve in Figure 4-23 represents the “envelope” corresponding to the upper curve in Figure 3-21. There are no contributions in Figure 4-23 from NaOH or ammonia. For cation conductivity (Figure 4-24). The contribution of phosphate at the normal limits shown in Figure 4-18 has been converted into cation conductivity using the relationship shown in Figure 3-19 for H3PO4. The contributions of chloride and sulfate at the normal limits in Figures 4-20 and 4-21 have been converted to cation conductivity. The top dashed curve in Figure 4-24 corresponds to the upper curve in Figure 3-23.

4-39


Figure 4-23 PC(H): Drum Boiler Water Specific Conductivity vs. Operating Pressure showing specific conductivity due to phosphate alone, and due to phosphate plus various amounts of chloride and sulfate. It must be noted that this figure is only an example to illustrate the effects of phosphate, chloride and sulfate. Each operator will need to develop unique curves for each boiler as described in the text of Sections 4.3.1 and 3.8.11.

4-40


Figure 4-24 PC(H): Drum Boiler Water Cation Conductivity vs. Operating Pressure. Showing cation conductivity due to phosphate alone and due to phosphate plus various amounts of chloride and sulfate. It must be noted that this figure is only an example to illustrate the effects of phosphate, chloride and sulfate. Each operator will need to develop unique curves for each boiler as described in the text of Sections 4.3.1 and 3.8.12.

4-41


If control curves such as illustrated in Figures 4-23 and 4-24 are derived, then specific and cation conductivity can be used as the main control for contaminants in the boiler water for PC(H), but they should be utilized together with other monitored parameters (pH and phosphate) to diagnose effectively various potential operational problems. For example, from a baseline steady load and uniform boiler water specific and cation conductivity, a sudden or continuing change in either or both of these parameters can be the basis for determining the source of the potential problem and of initiating the appropriate corrective action. The following four case studies illustrate the diagnostic indications. Case I: Boiler Water Specific Conductivity Rises While Cation Conductivity Remains Steady A. Ammonia feed increasing

•

pH at economizer inlet increases

•

Boiler water pH increases

B. Sodium hydroxide feed increasing

•


C. Sodium hydroxide cycle contamination

•


•

Makeup specific conductivity and condensate pump discharge sodium increase—makeup system malfunction

•

Condensate polisher effluent sodium and economizer inlet pH increase

Case II: Boiler Water Specific Conductivity Falls While Cation Conductivity Remains Steady A. and B. Same as Case I, except parameters fall Case III: Boiler Water Specific Conductivity and Cation Conductivity Both Rise A. Phosphate feed increasing

•

Boiler water pH and phosphate increases

B. Cycle contamination

•

Makeup specific conductivity increases—makeup system malfunction

•

Condensate pump discharge cation conductivity and sodium increases—condenser leak and/or makeup system malfunction

•

Condensate polisher cation conductivity increases—polisher malfunction

•

Economizer inlet cation conductivity increases—may reflect any of the other bullets in B.

4-42


•

Boiler water pH drops—acid contamination—makeup and/or polisher malfunction (economizer inlet pH drops), brackish water intrusion

Case IV: Boiler water specific and cation conductivity both fall A. Phosphate feed decreasing

•

Boiler water pH and phosphate drop

A varying unit load may add a complication to these diagnostic procedures related to possible chemical hideout. In this situation, additional monitoring may be required to verify the underlying root causes for chemistry parameter deviations from their normal target values. Copies of Figure 4-15 through 4-22 and the customized versions of Figure 4-23 and 4-24, should be included in the plant operating procedures and prominently displayed in the control room, water and steam sample room, and chemistry laboratory. It may be more convenient to develop a separate figure, based on Figures 4-15 and 4-16, which only includes the core parameters for a specific unit. 4.3.2 Normal Operation for PC(H) Phosphate Hideout and Its Control With PC(H), the phosphate during normal operation will probably be at a level that the boiler cannot maintain without hideout, especially for high pressure boilers. Hideout will be expected during initial startup on PC(H). It is important to be patient during this period and not to chase phosphate while attempting to maintain pH-phosphate control relationships. As indicated in Figure 4-1, the control range (pH-phosphate) for PC has been enlarged over that for phosphate treatment (PT) from the previous EPRI Guidelines, and there is an allowance for up to one ppm of free sodium hydroxide (NaOH) over a given range of phosphate not to exceed that dictated by boiler pressure. No additions of mono- and di-sodium phosphate should be made to correct for any observed hideout return. Controlled additions of trisodium phosphate can be made to increase phosphate concentration levels materially affecting pH. Minor sodium hydroxide additions to give boiler water concentration of always less than one ppm (“free” sodium hydroxide beyond the Na/PO4 ratio of 3.0) can be utilized to increase pH if it becomes necessary, but should be avoided, if possible. It is important to have knowledge about the typical amounts of chemicals that can be added to the boiler. Normally, hydrated trisodium phosphate (Na3PO4 · 12H2O) is utilized because of its better solubility as compared with the anhydrous Na3PO4 material. Any additions should be made based on the targeted sodium to phosphate ratio. The phosphate concentration is dependent on what the boiler will accept at the operating pressure utilized.

4-43


A modern natural circulation boiler of about 500 MW contains around 600,000 lbs of water at normal operating pressure. For example, the total quantity of TSP utilized for a Na/PO4 of 3.0 to achieve a boiler water phosphate level of 2 ppm would be 978 g or 2.13 lbs. Careful records of the additions should be kept. With the blowdown valve normally closed, the only additions required should be for average carryover and sample losses (hopefully less than 0.3%). It is advisable to use good technical grade chemicals as specified by the American Chemical Society. Any sodium hydroxide utilized should be of a grade specified by ASME PTC-31 (for regeneration of condensate polishing resins). The latter (as crystalline material) should be stored in a “dry” atmosphere as it readily absorbs moisture from humid environments. Since operation should normally be with the blowdown valve closed or nearly closed, careful control of the inventory of chemicals added must be maintained to ensure that additions are not being made to correct for phosphate hideout. Blowdown should be used only for contaminant control. Reaction to Contaminants in Boiler Cycle Condenser cooling water in-leakage or makeup water contamination can and will occur. The reaction to this is dependent on the kind of contamination. Three examples could be:

•

chloride or sulfate excursions due to condenser in-leakage;

•

caustic or acid contamination from makeup (or condensate polisher if utilized) regenerant chemical leakage or

•

silica from extraneous sources

The reaction in all cases should be to increase blowdown, necessitating boiler water treatment chemical addition to compensate for the loss, and in some cases variation of additions to correct for such things as pH depression (caustic addition). Such caustic injection should never result in the boiler water caustic concentration exceeding 1 ppm.

4-44


Startup and Shutdown Operation During startup the effect of chemistry control impacts on the potential for internal boiler deposits and related corrosion. Focus should be directed towards the identification and correction of the root causes of chemical control problems as follows: Cold Start

The ramp to full pressure is dependent on boiler design and boiler/turbine stress limitations. Frequent analysis of the boiler water with emphasis on suspended solids and silica is recommended. Suspended solids is a measure of the magnetite inventory and the source of deposition and subsequent under-deposit corrosion. Silica is monitored out of concern for turbine deposition.

•

Adjust boiler chemical feed to take advantage of tolerance for higher alkalinity at lower pressure.

•

Adjust phosphate feed to support anticipated boiler blowdown rate.

•

The use of conductivity measurements will provide an indirect assessment of phosphate concentrations as previously described.

•

Anticipate boiler startup and begin pumping chemicals so that target treatment is in place when boiler is fired.

•

Maximize boiler blowdown and perform bottom blow (if present) during startup when solids are at their highest concentration.

•

With pressure increases above 1500 psi the phosphate residual should be allowed to slide back to a concentration consistent with the guidelines values for PO43– versus pressure (Figures 4-15 and 4-16).

Transition to Normal Operation

•

Continue boiler blowdown reduction in 50% increments.

•

Adjust phosphate chemical feed to support measured or estimated blowdown rate.

•

Increase trisodium phosphate (TSP) to increase alkalinity. Mono-sodium phosphate (MSP) should not be added even though it is recognized that many operators have used MSP in the past to lower pH. In the case that the pH is too high, then the appropriate action is to increase the blowdown rate.

•

Be precise in measuring chemical feed dose.

•

Monitor chemical feedpump performance.

•

Now is the time to conserve boiler blowdown since excess suspended solids have been removed, chemical feed is stable and consistent. Boiler blowdown should be zero during normal operation.

4-45


Hot Startup and the Impact of Hideout Avoid making changes to boiler chemistry in reaction to chemistry change from a boiler trip (such as depressed or elevated pH). This chemistry change is normal and is usually the result of “hideout” phenomena. Careful characterization of the trip induced chemistry change is necessary, and it is important to avoid “chasing your tail.” Additional discussions of shutdown, layup, startup, cycling and peaking can be found in Appendix F. Feedwater and Makeup Control Makeup water to the cycle is an important source of oxygen, carbon dioxide and chemical contamination, thus affecting adversely the levels of iron and copper contamination. Older makeup systems may not provide the necessary quality of water, which generally requires a final polishing stage, usually a mixed bed demineralizer. Contaminants such as chlorides, sulfates, caustic and acids can emanate from makeup treatment systems. Such contamination may cause an increase in copper pickup in the cycle. Serious consideration must be given to removing oxygen from the makeup and condensate water and storing the makeup water unvented to the atmosphere if copper and iron are to be controlled to minimal levels. Oxygen removal can be achieved in a number of ways. This topic is fully discussed in Appendix A. Condenser Leakage Condenser leakage will introduce chemical contamination and oxygen, which will adversely impact the control of iron and copper in the cycle as well as having other deleterious effects. Condenser leakage must be continuously monitored by sodium and cation conductivity. Chemical Feeds—Mixed-Metallurgy Cycles Maintaining a reducing environment in the feedwater is essential if feedwater copper corrosion transport is to be controlled to an acceptable level. The availability of a deaerating condenser and deaerating heater will have a positive effect on oxygen control. However, it will still be necessary to feed a reducing agent to ensure a negative ORP at the most desirable level of –350 mV at the deaerator inlet. It is important to reference Appendix B to understand how reducing ORP levels are dependent on a number of variables. The most desirable and recommended reducing agent is hydrazine. The reducing agent should be fed at the condensate pump discharge, or, if polishers are installed, at the outlet of the polishers. It is not satisfactory to add the reducing agent after the deaerator.

4-46


It is not an acceptable or effective practice to increase the reducing agent to counteract increasing levels of air in-leakage. In the LP feedwater system (up to the deaerator) the reaction rate between a reducing agent and oxygen is very low. The optimum practice is to monitor ORP at the LP heater outlet (deaerator inlet) to ensure that reducing conditions are maintained. If air inleakage is allowed to be excessive with oxygen levels at the condensate pump discharge greater than 10 ppb, then it is not possible to add sufficient reducing agent to make the low pressure feedwater reducing (at deaerator inlet). A number of alternate reducing agents have been introduced in the past several years and are in use at some power stations. The main concern with the use of these alternate chemicals is their thermal decomposition ( breakdown) products, which usually include organic acids and carbon dioxide. The use of these chemicals must be carefully considered and the breakdown products carefully evaluated before use in the cycle is allowed. Cation conductivity typically will be elevated above the normal target value of 0.3 µS/cm around the cycle when using alternate treatments. The cycle pH should be controlled in the range of 9.0 to 9.3 using ammonia. Ammonia should be fed at the same location as with hydrazine. Separate lines are preferred. Reagent grade chemicals are recommended to avoid any unknown contamination from chemical additions (see Section 6.3 for further discussion). A number of alkaline organic chemicals have been utilized to control pH. The same admonishments apply to these chemicals as to hydrazine and alternatives, as previously discussed. However, it should be remembered that the overriding philosophy of the EPRI guidelines is that organic amines and organic reducing agents are not needed, required or recommended in the fossil plant cycle. The following target values at the deaerator inlet and the economizer inlet are required for proper copper control:

•

Reducing agent (hydrazine or alternate) - to maintain an ORP of between -300 and -350mV at the deaerator inlet (use minimum amounts of reducing agent to achieve the desired ORP) (see Appendix B for further discussion on ORP and how reducing ORP levels depend on a number of variables).

•

Oxygen- deaerator inlet ≤ 10ppb, economizer inlet ≤ 5ppb

•

pH- 9.0 to 9.3

•

Cation conductivity- ≤ 0.3 µ S/cm

Maintenance of these parameters at the above levels should produce the desirable feedwater copper level of 2 ppb or less.

4-47


Chemical Feeds—All-Ferrous Systems While it is not required or recommended (assuming adequate oxygen and air in-leakage control) to maintain a reducing atmosphere, oxygen and carbon dioxide should still be rigidly controlled by reducing air in-leakage and by minimizing dissolved oxygen in the makeup water. Many operators have found that reducing agent feed is not required if these practices are followed. By eliminating reducing agent feed while also minimizing air in-leakage and dissolved oxygen in the makeup, it has been found in numerous instances that iron transport was effectively minimized, and levels of around 1 ppb at the economizer inlet are easily achievable. All-ferrous systems can operate at the higher pH levels of 9.2 to 9.6, further minimizing iron transport. Reagent grade chemicals are recommended to avoid any unknown contamination from chemical additions. Monitoring and Corrective Actions A comprehensive monitoring program should be established, based upon the use of continuous monitors where available, supplemented by grab samples, if necessary as outlined previously. Comprehensive monitoring protocols are given in several EPRI Guidelines such as the Cycling, Startup, Shutdown and Layup Fossil Plant Cycle Chemistry Guidelines for Operators and (1) Chemists. The cycle diagrams in that publication are based upon the applicable boiler water chemistry in use (phosphate, AVT or caustic). An important new parameter for copper control in mixed-metallurgy cycles in this PC and CT Guidelines, is ORP. The ORP must be in the reducing range to control copper adequately, as previously described. It is especially important that ORP is monitored in the LP feedwater at the deaerator inlet, and not at the economizer inlet. Conditions could be reducing in the HP feedwater (at the economizer inlet), but oxidizing in the LP feedwater due to excessive air inleakage or oxygen saturated makeup additions. Many utilities have reported problems with the use of ORP sensors and variable ORP readings. It is important that these instruments are calibrated and subjected to frequent QA. Please see Appendix B for comprehensive details. Deviations from established target values must be vigorously investigated and brought rapidly under control if copper and iron are to be minimized in the cycle.

4-48


4.4 REFERENCES 1. Cycling, Startup, Shutdown and Lay-up Fossil Plant Cycle Chemistry Guidelines for Operators and Chemists, EPRI, Palo Alto, CA: August 1998. TR-107754. 2. Cycle Chemistry Guidelines for Fossil Plants: Phosphate Treatment for Drum Units, EPRI, Palo Alto, CA: December 1994. TR-103665.

4-49


5 CYCLES WITH DRUM BOILERS ON CAUSTIC TREATMENT

5.1 CAUSTIC TREATMENT GUIDANCE As outlined in Section 1.3.2, caustic treatment has been used successfully for many years in coal and oil fired drum units as an alternative to phosphate treatment, particularly where difficulties with phosphate hideout, and control and monitoring have been experienced. The general operating experience with CT is limited to 2500 psi (17.2 MPa). (1) As with the recently revised AVT Guidelines and Section 4 of this report on PC, it is necessary to consider boiler limits to protect the boiler, as well as steam limits to protect the turbine. The limits on boiler water impurity concentrations to minimize corrosion in the boiler are based on the chloride concentration; chloride being the most corrosive constituent normally found in boiler water. In previous guidelines(2-4), the same limits were used for sulfate and chloride. However, it is known that chloride represents a greater risk than sulfate in initiating corrosion, so (1) in the most recent guidelines for AVT , the concentrations of sulfate were taken as twice those of chloride. Note that the 1:2 concentration ratio for chloride to sulfate corresponds to a 40/60 stoichiometric ratio. Taking a lower ratio for chloride makes allowance for the fact that the cation conductivity includes other impurities.

Compared with AVT, the small amounts of sodium hydroxide added with CT, increase the amount of chloride that can be tolerated, without initiating boiler corrosion. However, it is still important to limit the chloride ingress and ensure that there is sufficient sodium hydroxide to prevent potentially acidic conditions forming, whilst preventing excessive carryover into the steam. Experience has shown that reducing the limits for chloride with increasing pressure reduces the risk of initiating corrosion. Sodium hydroxide is required to raise the pH of the boiler water to the desired range and additional sodium hydroxide is added to take account of the chloride present, using a semi-empirical relationship NaOH = 2.5 × Cl. It is well understood that uncontrolled concentrations of sodium hydroxide may lead to caustic gouging in the boiler, dryout is particularly dangerous. Excess carryover into the steam may result in damage to the turbine and austenitic components in the steam circuit. Although far less common than phosphate hideout, it is still possible for caustic to hideout. If this occurs, the reasons for it should be investigated, since it might indicate dryout.

5-1

EPRI Licensed Material Cycles with Drum Boilers on Caustic Treatment

5.2 GUIDANCE AND GUIDELINES FOR CAUSTIC TREATMENT (CT) 5.2.1 Target Values for CT A cycle chemistry diagram has been developed for drum units with all-ferrous and mixed Fe-Cu metallurgy feedwater heaters operating with reducing conditions in the feedwater, AVT(R), and is provided in Figure 5-1. Reducing conditions (ORP of feedwater at deaerator inlet -300 to -350 mV) and <5 ppb oxygen at the economizer inlet, with strict air in-leakage, oxygen control and the addition of a reducing agent, must be used for mixed Fe-Cu metallurgy feedwater heaters. However, non-reducing or mildly oxidizing conditions, with up to 10 ppb oxygen at the economizer inlet, have been used successfully with CT for some plants with all-ferrous feedwater heaters. The cycle diagram for these AVT(O) feedwater conditions is shown in Figure 5-2. There are a number of features to bear in mind when adapting Figures 5-1 and 5-2 to individual plant situations:

•

The core parameters are indicated with an asterisk (*).The instruments should be on-line and monitored continuously. This is shown by “C” in the sample column.

•

There are a number of other parameters, which also have the designation “C”, since they provide useful confirmation of the unit condition.

•

There are a number of parameters designated “T” for troubleshooting. They can be used if other parameters are out of their normal guideline values.

Other points of interest include:

•

The dissolved oxygen at the condensate pump discharge should be <10 ppb.

•

There is a limit for copper at the economizer inlet for units with copper alloys.

•

A reducing agent (hydrazine preferred) is required for units operating under reducing conditions (essential when copper alloys are present in the feedwater heaters).

•


•

Sodium and cation conductivity at the condensate pump discharge provide cycle contamination control in the event of condenser leaks and makeup malfunctions.

•

Dissolved oxygen control at the condensate pump discharge ensures optimum feedwater redox potential (ORP) and provides an indication of air in-leakage.

•

Deaerator inlet ORP ensures reducing conditions to minimize copper transport when operating with AVT(R).

•

Economizer inlet sodium (or condensate polisher effluent sodium) guards against polisher malfunction and cycle contamination.

•

Boiler water cation conductivity is the main control for boiler water contaminants.

5-2


Figure 5-1 Cycle Chemistry Diagram for a Drum Unit on Caustic Treatment. All-ferrous and mixed-metallurgy feedwater systems. Operating with a reducing environment (reducing agent added) in the feedwater, AVT(R).

5-3


Figure 5-2 Cycle Chemistry Diagram for a Drum Unit on Caustic Treatment. All-ferrous feedwater systems. Operating with an oxidizing environment (no reducing agent) in the feedwater, AVT(O).

5-4


•

Boiler water pH, specific and cation conductivity are the main controls for ensuring appropriate conditions for dealing with contaminants and minimizing corrosion.

•

Saturated steam sodium monitoring is required to monitor carryover, identify drum internal failures and to prevent contamination of the steam to the turbine.

A series of curves have been developed which give smooth, continuous relationships for boiler water target values as a function of pressure up to 17.2 MPa (2500 psi), the current limit of operating experience for caustic treatment. These curves are designed to provide boiler protection and ensure steam limits of 2 ppb sodium, chloride and sulfate, and 10 ppb silica. The following boiler water limits are presented:

•


•

cation conductivity versus operating pressure (Figure 5-4)

•

specific conductivity versus operating pressure (Figure 5-5)

•

sodium hydroxide concentration versus operating pressure (Figure 5-6)

•

sodium concentration versus operating pressure (Figure 5-7)

•

chloride concentration versus operating pressure (Figure 5-8)

•

sulfate concentration versus operating pressure (Figure 5-9)

•

silica concentration versus operating pressure (Figure 5-10)

5-5


Figure 5-3 Caustic Treatment: Boiler Water pH vs. Operating Pressure. The dashed line indicates the lower limit for pH to ensure that the pH of the boiler water is at least a minimum value to reduce the risk of boiler corrosion.

5-6


Figure 5-4 Caustic Treatment: Boiler Water Cation Conductivity vs. Operating Pressure

5-7


Figure 5-5 Caustic Treatment: Boiler Water Specific Conductivity vs. Operating Pressure. The dashed line indicates the lower limit for specific conductivity corresponding to the minimum pH and NaOH concentration shown in Figures 5-3 and 5-6.

5-8


Figure 5-6 Caustic Treatment: Boiler Water Sodium Hydroxide vs. Operating Pressure. The dashed line indicates the lower limit for NaOH in the boiler water.

5-9


Figure 5-7 Caustic Treatment: Boiler Water Sodium vs. Operating Pressure

5-10


Figure 5-8 Caustic Treatment: HRSG Boiler Water Chloride vs. Operating Pressure

5-11


Figure 5-9 Caustic Treatment: Boiler Water Sulfate vs. Operating Pressure

5-12


Figure 5-10 Caustic Treatment: Drum Boiler Water Silica vs. Operating Pressure

5-13


The boiler water control limits for chloride are based on 0.3 ppm for 17.2 MPa (2500 psi) plants, as used by various operators of caustic treatment (see Table 1-4). Although higher chloride limits have been used for low heat flux plants, this increases the risk of caustic gouging in high heat flux areas, and makes it unlikely that 2 ppb limits for sodium, chloride and sulfate in steam would be achieved. Table 1-3 shows how the limits for chloride in the boiler water at 16 MPa (2400 psi) can be increased by factors of 2 at 10 MPa (1500 psi) and 3 at 6 MPa (900 psi), respectively, while still minimizing the risk of initiating corrosion. The chloride concentrations of 0.3 ppm at 17.2 MPa (2500 psi), 0.6 ppm at 10 MPa (1500 psi) and 0.9 ppm at 6 MPa (900 psi) have been used to construct the boiler water control curves for chloride (Figure 5-8). The control curves for sulfate (Figure 5-9) are twice those of chloride. Both sets of curves are below those calculated using the EPRI predictive model for 2 ppb chloride and 2 ppb sulfate in steam (see Figures 3-15 and 3-16). A cation conductivity curve has been calculated from the chloride and sulfate curves and is shown in Figure 5-4. With caustic treatment a certain amount of sodium hydroxide is required to achieve the required pH. In addition, there should be sufficient sodium hydroxide to combat the chloride. NaOH = 2.5 x chloride is commonly used. Curves for the minimum and maximum concentrations of sodium hydroxide (Figure 5-6) can be obtained by allowing a factor of two, between maximum and minimum, whilst still encompassing the NaOH = 2.5 x chloride factor. The maximum NaOH concentrations correspond to 1.0 ppm at 17.2 MPa (2500 psi), 1.8 ppm at 10 MPa (1500 psi) and 2.5 ppm at 6 MPa (900 psi). pH (Figure 5-3) and sodium (Figure 5-7) curves can be calculated from the NaOH concentrations. The sodium curves also allow for at least half the chloride and sulfate being associated with sodium. The curve for sodium is just below that calculated using the predictive model for 2 ppb sodium in steam (see Figure 3-14), demonstrating that higher concentrations of chloride, sulfate and sodium hydroxide should not be present, if the 2 ppb sodium limit in steam is to be obtained. Curves for specific conductivity (Figure 5-5) can be calculated from the sodium hydroxide, chloride and sulfate curves. The silica curves (Figure 5-10) are similar to the silica curves for PC(H) and PC(L). Curves have also been prepared showing the relationships between specific conductivity, cation conductivity and pH (Figure 5-11), and pH and equivalent sodium hydroxide concentration (Figure 5-12). Figure 5-12 clearly shows that, when applying CT, particularly at low NaOH levels, the measured pH readings in the boiler water will be affected by any ammonia present and should therefore be corrected. This may be accomplished by using the information in Figure 5-12. Figure 4-5 is also applicable for making the correction with CT, providing the measured pH and ammonia data are accurate and the boiler water sodium hydroxide concentration is 0.4 ppm (equivalent to pH 9.0) or greater. Specific and cation conductivities of the boiler water should be regarded as the main control parameters for drum boilers operating on caustic treatment. Sodium in steam should be regarded as the main protection against excessive impurities in steam.

5-14


Maintaining the boiler cation conductivity within the limits will provide boiler corrosion protection and includes the effects of chloride and sulfate. It should be noted that chloride and sulfate are designated as “T” parameters in the cycle chemistry diagram. Curves have been provided for chloride and sulfate (Figures 5-8 and 5-9). The chloride curves are designed to minimize the risk of initiating corrosion. Sulfate represents a lesser risk than chloride in initiating corrosion, so the concentration of sulfate has been taken as twice that of chloride. Both sets of curves are provided so that an organization can check chloride and sulfate levels, in cases where the cation conductivity limits are exceeded. The chloride curves can be used directly, if there is a continuous chloride analyzer. This is the preferred method, particularly where sea water is used for cooling the condensers and should be designated as a “C” parameter in the cycle chemistry diagram. Monitoring specific conductivity, in conjunction with the cation conductivity, and pH can be used as an indication of the sodium hydroxide concentration. Figure 5-11 shows the relationship between specific conductivity and pH in samples cooled to 25°C. Figure 5-12 shows the relationship between sodium hydroxide and pH at 25°C. The curves for sodium in boiler water (Figure 5-7) are based on the sodium hydroxide required to achieve the pH, plus an allowance for sodium associated with the chloride and sulfate. Maintaining operation within the boiler water sodium limits should prevent excessive carryover of sodium, chloride and sulfate into the steam. However, sodium in steam should be monitored online. The curves for silica (Figure 5-10) are based on the need to limit the amount of silica carried over into the steam and are derived from partitioning data. Silica is also designated a “T” parameter. Figures 5-3 to 5-10 contain curves for Action Levels 1, 2 and 3 above the normal operating zone. Figures 5-3, 5-5 and 5-6 show curves for pH, specific conductivity and sodium hydroxide concentration, respectively. The dashed lines in Figure 5-3, 5-5 and 5-6 indicate the lower limits for pH, specific conductivity and sodium hydroxide. The area between the dashed line and Action level 1 should be regarded as the normal operating zone. Figures 5-1 to 5-8, modified if necessary to reflect specific unit characteristics and Figures 5-11 and 5-12, could be included in the plant operating procedures and displayed in the control room, steam sample room and chemistry laboratory.

5-15


Figure 5-11 Relationship Between Specific Cation Conductivity, Cation Conductivity and pH at 25°C for Boiler Water with Caustic Treatment

5-16


Figure 5-12 Caustic Treatment: Drum Boiler Water Relationship Between pH and Equivalent Sodium Hydroxide Concentration at 25°C, Showing Effect of Ammonia Concentrations. The lowest curve is at zero ammonia. The highest curve is at 2 ppm ammonia.

5-17


5.2.2 Target Values for Plants Without Reheat Less restrictive target and action levels have been applied to plants without reheat in previous EPRI Guidelines(1-5). However, increasing the boiler water limits for cation conductivity, chloride, sulfate and sodium hydroxide concentrations by a factor of two, would require the limit for sodium in steam to be increased accordingly, unless it could be demonstrated that mechanical carryover was at an unusually low level in the units concerned (see Section 5.2). Whether this is possible or desirable would depend on local conditions and decisions.

5.3 NORMAL OPERATION FOR DRUM BOILERS ON CAUSTIC TREATMENT 5.3.1 Cycle Makeup Makeup water is an important source of oxygen, carbon dioxide and chemical contamination and can adversely affect the levels of iron, copper and other contaminants in the feedwater and boiler water. 5.3.2 Condenser Leakage Condenser leakage introduces chemical contamination to the feedwater and boiler water and must be continuously monitored by sodium and cation conductivity. Sufficient sodium hydroxide should be present in the boiler water to combat chloride. 5.3.3 Chemical Feeds – Feedwater – Mixed-Metallurgy Cycles Maintaining a reducing environment in the feedwater is essential if copper transport is to be maintained at an acceptable level. It will be necessary to feed a reducing agent (hydrazine is preferred) to ensure a negative ORP; the desirable level is -300 to -350 mV at the deaerator inlet, however, Appendix B should be referenced to understand how the reducing ORP is dependent on a number of variables. The reducing agent should be added at the condensate pump discharge, or, if polishers are fitted, at the outlet of the polishers. It is not satisfactory to add the reducing agent after the deaerator. The reducing agent should not be increased to above 30 ppb to counteract air in-leakage, the source of the leakage should be identified and fixed. The feedwater pH should be controlled in the range 9.0 to 9.3 using ammonia. The ammonia should be fed at the same location as the hydrazine, preferably via a separate dosing line. The relationship between ammonia concentration and pH, and the effects of carbon dioxide on feedwater are shown in Figure 3-26. Oxygen at the deaerator inlet should be <10 ppb and at the economizer inlet <5 ppb. The cation conductivity should be <0.2 µS/cm.

5-18


5.3.4 Chemical Feeds – Feedwater - All-Ferrous Systems Feedwater dosing is similar to that described above for mixed-metallurgy cycles. Control of air ingress is desirable to minimize oxygen and carbon dioxide in the makeup and feedwater. Some operators have found it unnecessary to add reducing agents to the feedwater, provided relatively low oxygen concentrations can be achieved, e.g., <10 ppb at the deaerator outlet. All-ferrous systems can be operated at higher feedwater pH levels of 9.2 to 9.6, further reducing iron transport. 5.3.5 Chemical Feeds – Boiler Water Sodium hydroxide is added to the boiler water at the boiler drum to maintain the pH in the range 9.1 to 9.7, depending on the operating pressure and the amount of sodium hydroxide required to combat the chloride present (see Figures 5-3 and 5-6). As with PC, the presence of ammonia contributes to the pH of the boiler water (see Figures 4-3 and 4-4). However, ammonia has a poor high temperature alkalizing potential and it is important to ensure that adequate solid alkalizing agent is present in the boiler water. It is good practice to occasionally confirm that sufficient sodium hydroxide is present by measuring the ammonia and sodium hydroxide in grab samples of boiler water. If necessary, correct the pH for the ammonia present using Figures 4-5 or 5-12. 5.3.6 Monitoring and Corrective Actions A comprehensive monitoring program should be established based on continuous monitors where possible, supplemented by grab samples. Monitoring protocols are given in several EPRI Guidelines, including the Cycling, Startup, Shutdown and Layup Fossil Plant Cycle Chemistry (6) Guidelines for Operators and Chemists . Many utilities have reported problems with the use of ORP sensors and variable ORP readings. It is important to note that a reducing ORP will not always be in the range –300 to –350 mV. Appendix B gives details of ORP measurement, and on how the ORP reading is dependent on a number of variables. The sodium hydroxide concentrations of boiler water can be measured directly on grab samples by titration or inferred from specific and cation conductivity (see Figure 5-11). The chloride concentrations of boiler water can be measured directly from online chloride analyzers, if fitted (preferred), or on grab samples, or inferred from the online cation conductivity measurements (Figure 5-4). Further information on chemical control and corrective actions are given in Section 6.

5-19


5.4 APPROACHES TO DEVELOPING OPTIMUM CAUSTIC TREATMENT Guidance is given below on how to develop and optimize caustic treatment for drum boilers, when AVT has been found to be inadequate for controlling corrosion, or when a phosphate treatment has been prone to hideout, or been difficult to monitor and control. A general road map for converting a unit to caustic treatment is given in Figure 5-13 and information on optimizing caustic treatment is given in Figure 5-14. Note that general operating experience of CT is limited to 17.2 MPa (2500 psi). Before using or commencing conversion to caustic treatment, it is important to be sure that dryout does not occur in the evaporator tubes. When converting from a phosphate treatment, it is important to ensure that any hidden out phosphate returns to the boiler water and can be controlled by blowdown. Carefully monitoring sodium in steam, while still operating on phosphate, gives a good indication of whether excessive carryover of boiler water impurities in steam is occurring. If this is the case, the boiler may not be suitable for caustic treatment. The following stages apply to optimizing boilers on caustic treatment and converting boilers currently operating on AVT or a phosphate treatment: a) Install additional equipment for dosing sodium hydroxide into the boiler water. b) Install additional sampling, sodium, chloride and conductivity monitoring equipment for monitoring the boiler water and steam. c) Undertake a carefully monitored trial by (re)commissioning the boiler on AVT and then adding a low concentration of sodium hydroxide to the boiler water, e.g. quarter to half of that recommended in Figure 5-6. d) Observe the concentrations and stabilities of the chloride and alkalinity in the boiler water, look for loss of sodium hydroxide due to hideout or dryout, and observe the concentration of sodium in steam, under various operating conditions, e.g. startup, steady load, load changes. If satisfactory values are obtained: e) Increase the concentration of sodium hydroxide in the boiler water to the values given in Figure 5-6. f) Repeat d) above. If satisfactory values are not obtained: g) Return to AVT, or add the lower concentration of sodium hydroxide as in c) above and continue the trial.

5-20


If the sodium hydroxide used for dosing is lost from the boiler, this is serious and it may be necessary to stop adding NaOH. As a minimum it is necessary to: h) Monitor hideout during load changes and carryover into steam. If excessive hideout or loss from the boiler is occurring, this is also very serious and it may be necessary to stop adding NaOH. It may be necessary to: 24 i) In very special cases a trial using a radioactive trader (Na OH) can be used to determine the extent and location of the hideout.

If excessive carryover into the steam is occurring: j) Inspect the steam separation equipment for damage. k) It may be necessary to modify the method of boiler operation (e.g. boiler water level control) or reduce the limits for impurities and the amount of sodium hydroxide added to the boiler water. l) Review the results and decide on the further course of action. If necessary: m) Change to AVT or PC.

5-21


Figure 5-13 Road Map for Conversion of Plants to Caustic Treatment

5-22


Figure 5-14 Road Map to Optimize Plants on Caustic Treatment

5-23


5.5 REFFERENCES 1.

Cycle Chemistry Guidelines for Fossil Plants: All-Volatile Treatment, Electric Power Research Institute, Palo Alto, CA., 1004187, November 2002.

2.

Interim Cycle Chemistry Guidelines for Combined Cycle Heat Recovery Steam Generators (HRSG’s), Electric Power Research Institute, Palo Alto, CA., TR-110051, November 1998.

3.

Cycle Chemistry Guidelines for Fossil Plants: All-Volatile Treatment, Electric Power Research Institute, Palo Alto, CA., TR-105041, April 1996.

4.

Cycle Chemistry Guidelines for Fossil Plants: Phosphate Treatment, Electric Power Research Institute, Palo Alto, CA., TR-103665, December 1994.

5.

Interim Consensus Guidelines on Fossil Plant Water Chemistry, Electric Power Research Institute, Palo Alto, CA., CS-4629, June 1986.

6.

Cycling, Startup, Shutdown and Layup Fossil Plant Cycle Chemistry Guidelines for Operators and Chemists, Electric Power Research Institute, Palo Alto, CA., TR-107754, August 1998.

5-24


6 PC AND CT CHEMISTRY CONTROL AND CORRECTIVE ACTIONS

6.1 CHEMISTRY CONTROL The main purposes of chemistry control are to minimize the risk of deposition and corrosion in the boiler and turbine. The EPRI approach in these revised guidelines for the phosphate continuum (PC) and caustic treatment (CT), (outlined in Sections 4 and 5, respectively) involves setting limits to protect both the boiler and turbine. The boiler limits are compatible with limits set to protect the turbine. Ammonia, with or without the addition of a reducing agent, is used to condition the feedwater and increase the pH. The solid alkalis used, trisodium phosphate plus sodium hydroxide for PC, and sodium hydroxide for CT, can be only used with drum type boilers and not once-through boilers. Operating pressure, type of firing (coal, oil or gas) and whether the unit is fitted with all-ferrous or mixed metallurgy feedwater heaters and condensate polishing have important bearings on feedwater and boiler water treatment. Higher concentrations of impurities can be tolerated with drum type boilers, than for oncethrough boilers, because, within limits, these can be allowed to accumulate in the boiler water, without impacting on steam purity, and can then be removed from the system by blowdown. As heat fluxes and operating pressures increase, so does the concentration factor between the bulk water and surfaces of the boiler tubes. Unless the concentration of impurities is reduced accordingly, increased corrosion and deposition will result. With drum type boilers, the amount of boiler water impurities and conditioning chemicals carried over into the steam depends on both vaporous and mechanical carryover. Vaporous carryover used to be calculated in the earlier EPRI guidelines from the “ray” diagram, but this has considerable limitations. A much improved method is now available based on the EPRI sponsored work on volatility discussed in Section 3. Mechanical carryover is based on the graph developed by EPRI (Figure 3-2), based on information supplied by boiler manufactures. Under PC and CT conditions, vaporous carryover usually represents only a relatively small fraction of the total carryover (mechanical carryover plus vaporous carryover). It is recommended that total carryover, which is now a core monitoring parameter, be measured on individual units under various operating conditions.

6-1

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions

Traditionally, ammonia and hydrazine, added to increase the pH and “remove” residual oxygen, have been used as the method of conditioning the feedwater. Oxygenated treatment (OT), where a small amount of oxygen is added to the feedwater instead of a reducing agent, is now frequently used for once-through boilers. OT is also being used for drum type boilers. Higher purity feedwater is required for OT and for this reason such units should be equipped with condensate polishing plants. The experience gained with OT demonstrates that the traditional strict limits of low oxygen are not required for the feedwater, provided that there are no copper containing alloys in the feedwater heating systems. Indeed, the presence of traces of oxygen in the feedwater reduces the risk of flow-accelerated corrosion in all-ferrous feedwater heaters. Where the concentrations of residual oxygen are not excessive, it is also possible to condition the feedwater using only ammonia, to increase the pH, without adding a reducing agent. This can also have the benefit of reducing the risk of flow-accelerated corrosion in the feedwater system and connecting pipework. Mixed-metallurgy feedwater heating systems are unsuitable for use with OT and ammonia without reducing agent. Low feedwater oxygen concentrations must be maintained by the use of a reducing agent (hydrazine) in order to minimize the dissolution of copper and carryover and deposition in the superheater and HP turbine. All-volatile treatment (AVT) can be used for drum type boilers with no further addition of conditioning chemicals, but being volatile, the alkalizing agent is rapidly lost from the boiler water into the steam and very little remains as a buffer against the ingress of impurities. Also, volatile additives are hardly dissociated at boiler water temperatures and, therefore, do not provide much alkaline pH correction under these conditions. The addition of solid alkalis to the boiler water, trisodium phosphate and/or sodium hydroxide with PC and CT, allows operation with higher concentrations of impurities, without corrosion, than is possible with AVT, because of the improved pH buffering capacity of trisodium phosphate and sodium hydroxide towards potentially acidic impurities. Sodium hydroxide has a lower buffering capacity than phosphate and the aim is to match the amount with the concentration of impurities present. Such additions can be increased as a temporary protective measure to protect the boiler during excursions in water purity, for example due to condenser leaks. Even with PC, it may be necessary to add some sodium hydroxide to control the pH. However, great care is needed to control the buildup of impurities and alkalizing agents to prevent damage to the boiler and excessive carryover into the steam. Even under apparently similar operating and water chemistry conditions, some plants are far less tolerant than others and are more susceptible to corrosion and failures. Condenser leaks are often the main source of impurities and, therefore, their incidence and nature of the cooling water has an important bearing on the plant chemistry and corrosion. Similarly, some plants are much more susceptible than others to carryover of impurities into steam. As indicated earlier, total carryover is now a core parameter for units with drum boilers. Regular measurements of carryover and the results, together with the actual history of the plant should to be taken into consideration when implementing the guidelines.

6-2


Makeup water can also be a source of contamination, as can off-site sources of condensate and contamination during storage. Condensate polishers are valuable in improving the quality of condensate. Makeup plant and condensate plant malfunctions can also be sources of contamination. Target and Action Levels have been developed for makeup water quality, condensate, feedwater, boiler water and steam. Information on sampling points, frequency of sampling, core and diagnostic parameters, limits and action levels are given in earlier sections of the guidelines. These should be adhered to and used as the basis of chemical control. It is essential that a full QA/QC program is in place for recording, evaluating and storing the data (see Appendix E).

6.2 CORRECTIVE ACTIONS A series of corrective actions has been developed for PC and CT. Although comprehensive, they are generic and are not meant to be all-inclusive or universally applicable. Users of the Guidelines are encouraged to use this information in developing customized unit-specific corrective actions. Two approaches are included, first from key sampling points around the water/steam cycle (as used in earlier versions of the guidelines) and then by looking at different types of problems, the approach used in EPRI ChemExpert Cycle Chemistry Advisor for Fossil Plants program. Possible causes of chemistry excursions for key points in the condensate, feedwater, boiler water and steam circuit are described in Table 6-1. This table can be used to help identify the sources of ingress of impurities. Corrective actions for key points in the condensate system, feedwater system, boiler water and the steam circuit are given in Tables 6-2 to 6-5. Advice on cycle chemistry problems, diagnosis, actions to be taken and the consequences of not taking the actions have been developed within the EPRI ChemExpert training program. Information on the most common problems for phosphate continuum (PC) and caustic treatment CT) has been abstracted from ChemExpert. These include: makeup system malfunctions, condenser leaks, condensate polisher malfunctions, excessive or insufficient reducing agent, corrosion and flow-accelerated corrosion (FAC) in the preboiler, phosphate hideout and return, and sodium:phosphate ratios in boiler water (PC), and the concentrations of chloride and sodium hydroxide (CT), carryover of boiler water impurities into steam and attemperation spray, air inleakage to the hotwell, colloidal silica and iron through the makeup systems, and boiler layup problems. The results are given in Tables 6-6 to 6-23. Not all plants have condensate polishers, deaerators and mixed-metallurgy feedwater heaters, as given in the tables. Further examples of problems, diagnosis, actions to be taken and the consequences of not taking action are contained within the ChemExpert program.

6-3

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-1 Possible Causes of Chemical Excursions Phosphate or Caustic Feed System

Ammonia Feed System

Insufficient Boiler Blowdown

Condenser Leakage

Condensate Polisher

Makeup Demineralizer

1. Sodium

X

X

X

X

2. Chloride

X

X

X

X

3. Sulfate

X

X

X

X

4. Silica

X

X

X

X

X

X

X

X

X

X

X

X

X

X

Parameter

5. pH 6. Cat/Spec Conductivity

X

7. Oxygen 8. TOC 9. Na:PO4 Ratio 10. Free Hydroxide 11. Ammonia 12. Phosphate

X

Air Inleakage

Deaerator

Feedwater Oxygen

Feedwater pH

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X X

13. Iron

X

X

X

X

14. Copper

X

X

X

X

15. ORP

X

X

6-4

Polisher Resin Leakage

X

X X

Reducing Agent Feed System

X

X

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-2 Condensate System – Corrective Actions Parameter

Corrective Action Condenser Leak Detection Trays

Sodium

Locate and repair 1. Leaking tube or joint. 2. Waterbox flange to shell leaks. Condensate Pump Discharge

Sodium*

1. Identify source of contaminant by checking makeup treatment system effluent; condenser tube leakage.

Cation Conductivity* Condensate Polisher Effluent

Cation Conductivity*

1. Identify source of contaminant by checking condenser tube leakage; makeup treatment system effluent; exhausted polisher resins; fouled or damaged resins.

Silica

2. Regenerate polishers.

Sodium*

3. Clean or replace resins, if necessary.

* Core parameters

6-5

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-3 Feedwater System – Corrective Actions Parameter

Corrective Action Deaerator Inlet

Dissolved Oxygen Reducing Agent ORP

1. Identify source of dissolved oxygen by checking condenser and feedwater train for air in-leakage. 2. If reducing agent dosing reduced, it may be necessary to increase. 3. If increasing reducing agent feed, check for excessive increase of ammonia and pH resulting from decomposition of reducing agent. Deaerator Outlet

Dissolved Oxygen

1. Check deaerator performance. 2. Check deaerator vent control. 3. Check deaerator inlet oxygen. Economizer Inlet

pH

1. Check ammonia and adjust feed, as necessary. 2. Identify source of contaminants by checking condensate polisher quality; condenser for tube and air leakage.

Ammonia

1. Check pH and adjust chemical feed, as necessary.

Specific Conductivity Cation Conductivity*

1. Check ammonia in feedwater (specific conductivity) 2. Identify source of contaminants by checking makeup treatment system effluent quality; condensate polisher effluent quality; condenser for tube and air in-leakage.

Iron Copper

1. Identify cause by checking pH of feedwater; dissolved oxygen in condensate; reducing agent dose; condenser for air in-leakage; FAC, iron and copper pickup.

Reducing Agent Dissolved Oxygen*

1. Identify source of oxygen, optimize in condensate and feedwater. 2. If reducing agent dose reduced, increase to control oxygen. 3. If increasing reducing agent feed, check ammonia and pH increase resulting from decomposition of reducing agent.

*Core parameters

6-6

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-4 Boiler Water - Corrective Actions Parameter

Corrective Actions Boiler Water or Blowdown

Sodium Chloride Sulfate Silica Cation Conductivity*

1. Increase blowdown. 2. Check phosphate, sodium to phosphate ratio, free sodium hydroxide, adjust chemical feed as appropriate. 3. Identify source of contamination by checking makeup treatment system effluent quality; condensate polisher effluent quality; condenser tube leakage.

pH*

1. Check phosphate, sodium to phosphate ratio, free sodium hydroxide, adjust chemical feed within limits allowed. 2. Check economizer inlet pH and adjust ammonia feed within limits allowed. 3. Identify source of contamination by checking makeup treatment system effluent quality; condensate polisher effluent quality; condenser tube leakage. 4. Increase blowdown. 5. Shut-down if pH drops below limits.

Phosphate (for PC)

1. Check pH, sodium, free sodium hydroxide, sodium to phosphate ratio, adjust chemical feed within allowed limits. 2. Check for hideout. 3. Do not attempt to “chase” hideout by dosing excess phosphate. 4. Increase blowdown, if appropriate.

Sodium to Phosphate Ratio (for PC)

1. Check pH, sodium, free sodium hydroxide, sodium to phosphate ratio, adjust chemical feed within allowed limits. 2. Check for hideout 3. Do not add acidic phosphates (NaH2PO4 or Na2HPO4).

Free Sodium Hydroxide (for PC)

1. Check pH, sodium, free sodium hydroxide, adjust chemical feed within allowed limits. Avoid over-dosing. 2. Monitor sodium in steam. 3. Increase blowdown, if appropriate.

*Core parameters

6-7

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-5 Steam Circuit - Corrective Actions Parameter

Corrective Actions Saturated, Superheated and Reheated Steam

Sodium* Chloride Sulfate Silica Total Organic Carbon

1. Closely monitor sodium in steam. 2. Increase blowdown, if necessary. 3. Identify source of contaminants by checking makeup treatment system effluent quality; condensate polisher effluent quality; condenser tube leakage; steam attemperation, flowrates, water quality, and boiler drum internals. TOC useful for troubleshooting. Possible ingress of furnace gasses (e.g., sulfate in reheaters).

Cation Conductivity*

1. Increase blowdown, if necessary. 2. Identify contributing parameters by checking individual anions and total organic carbon in steam.

*Core parameters

6-8

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-6 Makeup System Malfunction Diagnosis 1. Check specific conductivity* and silica of the makeup treatment system effluent. 2. Check condensate pump discharge sodium* and cation conductivity*; if the increase coincides with the makeup addition, it confirms problem. 3. Check condensate storage tank conductivity, sodium and sulfate. 4. If no condensate polishers: check relationship pH - NH3 conductivity in feedwater. 5. Check sodium, cation conductivity* and pH* in boiler water. 6. Check chemical feed for impurities.

Actions to be Taken

Consequences+

1. For Action Levels 1 and 2, use regenerated makeup train, drain condensate storage tank, as needed. 2. Reduce load immediately, do not increase boiler blowdown until ingress of impurity is stopped. 3. If Action Level 3 of any parameter occurs: orderly shutdown, drain condensate storage tank, deaerator tank and boiler, as needed. 4. Check contamination of the superheater, reheater and turbine. Chemically clean and water wash, as needed. 5. Find root cause and correct the situation.

1. Increased boiler corrosion. 2. Buildup of silica and other deposits in the turbine and possibly other cycle components.

+

Consequences if not corrected *Core parameters

6-9

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-7 Makeup Regenerant - NaOH, H2SO4 or HCl Leaking into the Cycle Diagnosis

Actions to be Taken

Consequences+

Sodium Hydroxide Leaking into the Cycle 1. Check specific conductivity* & pH of makeup plant effluent. 2. Check condensate pump discharge sodium* and cation conductivity*. If the increase coincides with the makeup addition to the condenser, it is a contamination problem. 3. Check condensate storage tank conductivity, sodium and sulfate or chloride. 4. Units without condensate polishers: check the relationship of feedwater conductivity – pH – NH3. 5. Check sodium, cation conductivity* and pH* in boiler water blowdown.

1. Action Level 3: reduce load immediately, do not increase blowdown; switch to regenerated makeup train and drain condensate storage tank, as needed. 2. Action 2 level: reduce load immediately. 3. For orderly shutdown, drain condensate storage tank, deaerator storage tank and boiler, as needed. 4. Check contamination of the superheater, reheater and turbine. Chemically clean and water wash, as needed. 5. Find root cause and correct the situation.

1. Possible stress corrosion cracking in the turbine and austenitic stainless steel sections of the superheater and reheater. 2. Possible caustic gouging in the boiler generating tubes. 3. Possible stress corrosion of cracking of stainless steel feedwater heater tubes.

Sulfuric or Hydrochloric Acid Leaking into the Cycle As 1, 2, 3, 4 and 5 above. 6. Check operation of regeneration of ion exchange in makeup plant.

+


6-10

As 2, 3, 4 and 5 above.

1. Cycle corrosion and contamination. 2. General corrosion and FAC of carbon steel in the preboiler cycle. 3. Hydrogen damage of boiler tubes. 4. Pitting of boiler steam generating tubes, superheater and reheater lower bends. 5. Pitting and corrosion fatigue in turbine.

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-8 Condenser Tube Leak Diagnosis 1. Compare sodium* and cation conductivity* of condensate pump discharge and hotwell. All increased is confirmation. 2. Polisher effluent conductivity drifting up is confirmation. 3. Blowdown sodium, cation conductivity drifting up, pH drifts down = confirmation.+ ∆ 4. Check chemical feeds for impurities.

Actions to be Taken

Consequences+

1. Boiler water at Action Level 3 immediate shutdown, drain & flush boiler, check deaerator storage tank, and condenser hotwell water, drain and flush. 2. Parameters within Action Level 2 – immediately reduce load, isolate one condenser water box, find leaks and plug. Feed emergency Na3PO4 or NaOH to control boiler pH. 3. Maintain lower load and keep condensate polisher effluent within limits. 4. Increase boiler blowdown.

1. Increased corrosion of the preboiler cycle. 2. Scale formation and hydrogen damage of the boiler generating tubes. 3. Stress corrosion cracking of austenitic stainless steel superheater, reheater and turbine components. 4. Turbine deposits, pitting and corrosion cracking, if problem persists.

+

Consequences if not corrected *Core parameters + With alkaline cooling water, boiler water pH may drift up ∆ As indicated, pH depression is symptomatic of contaminant ingress due to cooling water ingress via condenser tube leaks in the vast majority of plants. However, there are a few plants where the cooling water is alkaline, contains relatively low levels of chloride and sulfate and may not exhibit this tendency.

6-11

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-9 Condensate Polisher Malfunction/Exhausted Diagnosis

Actions to be Taken

Consequences+

1. Compare condensate pump discharge conductivity*, sodium*, condensate polisher effluent cation conductivity* and silica to these parameters in the effluent. 2. Compare the in and out ratio for iron. 3. Analyze effluent samples from individual polishers for resin fines. 4. Check chemical feeds for impurities. 5. Check purity (Cl-) of NaOH regenerant.

1. Take malfunctioning polishers out of service and regenerate. 2. Determine whether problem is incomplete backwashing of acid regenerant, resin leakage or polisher exhaustion. 3. Take resin sample and restore or replace the resin, as required. 4. Check the regeneration procedure and correct, as required. 5. Check condensate polisher strainers and fix.

1. Caustic or acid corrosion of boiler steam generating tubes and turbine. 2. Pitting in reheater due to acid sulfate deposition. 3. Silica and other deposits in the turbine.

+


6-12

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-10 Condensate Polisher – Caustic or Acid Regenerant Leakage Diagnosis

Actions to be Taken

Consequences+

Leakage of Caustic Regenerant 1. Compare feedwater sodium to condensate and makeup. 2. Check sodium* in the combined and individual polisher effluents & in steam. 3. Increasing trend in boiler & feedwater pH shows problem. 4. Check feedwater pH - NH3conductivity relationship; high pH indicates problem. 5. Check chemical feeds for impurities.

1. Reduce load and boiler pressure. 2. Check boiler carryover and steam sodium: if high, wash the turbine within two days. 3. Fix the polisher regeneration system.

1. Caustic cracking of austenitic stainless steels in heaters, superheater, reheater and turbine. 2. Caustic gouging in the boiler in the high heat flux zones.

Leakage of Sulfuric or Hydrochloric Acid Regenerant 1. Compare feedwater, condensate, makeup and steam cation conductivities. 2. Analyze feedwater, blowdown, condensate and polisher effluent samples for sulfate or chloride. 3. Analyze the conductivity – pH – NH3 relationships. 4. Monitor SO4 or Cl in steam and monitor pH* in boiler. 5. Check chemical feeds for impurities.

1. Reduce boiler pressure and unit load. 2. If boiler water sulfate, chloride or cation conductivity reached Action Level 2, go to orderly shutdown, drain the boiler, fill with ammoniated deaerated water with reducing agent, analyze after 3 hours, repeat if necessary. 3. If the cation conductivity and sulfate or chloride in the boiler water was only Action Level 1, feed emergency trisodium phosphate or NaOH. 4. Take boiler tube samples, clean boiler as needed.

1. Increased corrosion and FAC of the preboiler cycle and economizer. 2. Pitting and hydrogen damage in boiler generating tubes. 3. Pitting, corrosion fatigue and stress corrosion cracking in the turbine. 4. Generation of large quantities of corrosion products and their deposition, dirty boiler.

+


6-13

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-11 Excessive or Insufficient Concentration of Oxygen Scavenger Diagnosis

Actions to be Taken

Consequences+

Excessive Reducing Agent Leading to High Iron in Feedwater 1. Check historical data for feedwater iron and copper concentration vs reducing agent. 2. Analyze grab samples for iron and copper in feedwater.

1. Optimize the concentration of reducing agent in relation to iron and copper concentration in final feedwater. 2. Minimize air in-leakage. 3. Use condensate polishers as much as possible.

1. Increased rate of corrosion and FAC in the preboiler cycle and economizer. 2. Scale buildup in boiler steam generating tubes. 3. Scale buildup on boiler orifices (assisted circulation boilers) and on feedwater control valves. 4. Reduced output. 5. If orifice plugging, danger of long or even short-term overheat and high pumping costs for circulating pumps.

Insufficient Reducing Agent Leading to High Copper in Feedwater 1. Check reducing agent and oxygen concentrations around the cycle. 2. Check the reducing agent addition system (day tank, pump, etc). 3. Measure ORP and copper in feedwater.

+

Consequences if not corrected

6-14

1. Increase reducing agent addition rate to achieve the optimum concentration in feedwater, as determined by ORP being reducing (–300 to – 350 mV). 2. If air in-leakage rate is higher than recommended, reduce to the normal level. Confirm by ensuring oxygen at CPD is <10 ppb.

1. Increase rate of corrosion of the copper in the feedwater. 2. Deposit buildup on the boiler waterwalls. 3. Buildup on boiler orifices. 4. Possible deposition on HP turbine and reduced output.

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-12 Corrosion and/or Flow-Accelerated Corrosion in the Preboiler Cycle Diagnosis

Actions to be Taken

Consequences+

1. Optimize reducing agent concentration, oxygen and pH. 2. Operate condensate polishers as much as possible. 3. Check the condensate polishers resins for iron fouling and clean the fouling, if needed. 4. Ensure preboiler system cleanliness and proper preoperational cleaning of feedwater during cold starts. 5. Fill the deaerator storage tank and boiler with deaerated water, install condensate storage tank nitrogen sparging and blanketing or other deaeration, if needed.

1. Scale formation in the boiler generating tubes. 2. Increased boiler carryover due to suspended solids, possibly leading to deposit buildup in the turbine. 3. Reduced life and possible safety issue of corroding components. 4. Possibility of perforation and rupture of piping or feedwater tubing. 5. Plant safety issues.

As 1 and 2 above. 3. Minimize air in-leakage by eliminating sources. 4. Optimize layup procedures for copper bearing components, e.g. condenser & feedwater heaters. 5. Find root cause of copper alloy corrosion and plan actions. 6. Replace copper alloys in high pressure feed heaters asap. 7. Operate with reduced boiler pressure when Cu concentration in boiler water is above 50 ppb.

As 1 above, plus superheater. As 2 above. 3. Copper deposits in the HP turbine and reduced MW generating capacity and efficiency.

Carbon Steel 1. Analyze iron in feedwater during normal operation and startups: high during cold startups show layup corrosion. 2. Inspect shell condenser, shell and tube sides of feedwater heaters (particularly tube inlets), audit FAC, audit of feedwater & wet steam systems. 3. Check the concentrations of oxygen, reducing agent and pH, and their latest changes in condensate and final feedwater. Increased concentration of reducing agent and reduced pH can be the cause of the problem. Copper Alloys As 1 above for Cu, Ni and Zn. 2. Inspect copper alloy tubing used in condensers and feedwater heaters. 3. Check with maintenance for use of copper-bearing thread lubricants. 4. Verify the concentration of reducing agent, oxygen, ammonia and pH in condensate and final feedwater. 5. Check air in-leakage and its recent trends. +


6-15

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-13 Phosphate Hideout and Return (Phosphate Continuum) Diagnosis 1. Take duplicate samples of boiler water and analyze for Na, PO4, pH and conductivity. 2. Review Na/PO4 molar ratios for past and current operation, compare for boiler pressure and load increases and decreases. 3. Estimate sodium* and phosphate* mass balance around the boiler. +

Actions to be Taken 1. Reduce load. 2. Increase blowdown to reduce sodium phosphate concentration in the boiler water and, at the same time, bring pH within recommended limits by chemical additions.

Consequences+ 1. Severe hideout can lead to corrosion of waterwall tubes. 2. High concentration of sodium hydroxide can lead to increased boiler carryover.

Consequences if not corrected *Core parameter Information note: Moderate phosphate hideout does not lead to boiler problems in most boilers. It is prudent to perform a mass balance, test for corrosion and inspect waterwall tubes in the high heat flux and low mass flow regions.

6-16

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-14 Low Sodium/Phosphate Ratio and High Phosphate Concentration in Boiler Water (Phosphate Continuum) Diagnosis

Actions to be Taken

Consequences+

Low Sodium/Phosphate Ratio in Boiler Water 1. Take duplicate samples of boiler water and analyze for Na, PO4, pH and conductivity. 2. Review Na:PO4 molar ratios for past and current operation, compare for boiler pressure and load increases and decreases. 3. Estimate sodium* and phosphate* mass balance around the boiler.

1. Reduce load. 2. Increase blowdown to reduce sodium phosphate concentration in the boiler water and, at the same time, bring pH within the recommended limits by chemical additions.

1. Na:PO4 ratio less than 3.0 can lead to corrosion of boiler waterwall tubes. 2. Occasional low ratio does not lead to boiler problems in most boilers.

High Phosphate Concentration in Boiler Water As 1 above. 2. Check the phosphate injection system.

As 1 above. 2. Increase blowdown to reduce sodium phosphate concentration in the boiler water.

1. Concentration of sodium phosphate in waterwall tubes in high heat flux zones and horizontal tubes. 2. Phosphate hideout and possible waterwall tube corrosion. 3. Increased sodium phosphate in steam and deposition in reheater and turbine leading to overheat failures of reheater tubes and MW loss.

+

Consequences if not corrected *Core parameter

6-17

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-15 Low Sodium Hydroxide Concentration, Low NaOH:Chloride Ratio, and High Chloride Concentration in Boiler Water (Caustic Treatment) Diagnosis

Actions to be Taken

Consequences+

1. Check boiler water pH*, conductivity*, chloride and sodium* concentrations. 2. Check NaOH pumping rate. 3. Check blowdown rate. 4. Check cation conductivity* and acid regenerant in the condensate polisher effluent.

1. The most critical is reduced pH to Action Level 2 and 3. If pH is low, reduce load immediately. 2. Increase the chemical addition for NaOH. 3. If all parameters, except Na, pH and conductivity in the boiler are within normal limits, reduce boiler blowdown.

1. With sea water or other acidic water ingress, there may be hydrogen damage, pitting and corrosion fatigue of boiler tubes. 2. Volatile acids in steam can cause pitting, stress corrosion and corrosion fatigue of LP turbine blades and disks.

+

Consequences if not corrected * Core parameter

Table 6-16 High Sodium Hydroxide Concentration in Boiler Water (Caustic Treatment) Diagnosis 1. Check boiler water pH*, conductivity*, sodium* and OH. 2. Check NaOH chemical addition rate. 3. Check blowdown. 4. Check pH*, conductivity* and sodium* in condensate, makeup effluent and storage tanks. 5. Check pH, conductivity and sodium in condensate polisher effluent.

+


6-18

Actions to be Taken 1. If makeup and condensate polisher effluent OK (no high sodium or acid), increase blowdown and reduce NaOH feed rate. 2. Reduce load.

Consequences+ 1. Caustic gouging corrosion of boiler tubes in the areas of the high local heat flux and steam blanketing. 2. Foaming and increased boiler carryover, particularly if organics present. 3. Sticking turbine valves and possibility of turbine destructive overspeed, if deposits are built up in the valve bushing. 4. If elevated NaOH in steam persists, possibility of stress corrosion cracking of LP blade attachments, in the superheater and elsewhere.

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-17 Phosphate and Sodium Hydroxide Carryover (Phosphate Continuum) Diagnosis

Actions to be Taken

Consequences+

Phosphate Carryover 1. Take duplicate samples of boiler water, saturated and superheated steam and condensate; analyze for Na*, PO4, pH and conductivity. 2. Review chemical charts, particularly pH, conductivity* and sodium* in steam. 3. Check with control room operators about boiler pressure, drum level control and use boiler circulation pumps (where applicable). 4. Check with control room operators and turbine engineers about the turbine efficiency and generating capacity and IP turbine exhaust pressure trends. 5. Check the makeup and polisher effluent sodium.

1. Reduce load. 2. Reduce boiler water phosphate concentration to the minimum normal limit. 3. Where boiler separation and drum level control are causing the problem, maintain the reduced pressure and load until the problem can be fixed. 4. Inspect superheater and reheater for deposits. 5. Wash turbine as needed (without disassembly at reduced rpm).

1. Deposition of sodium phosphate in the reheater and turbine leading to overheat failures of reheater tubes. 2. Loss of turbine MW and efficiency. 3. Sticking turbine valves.

As 1 to 5 above.

1. Deposition of sodium hydroxide in the reheater and turbine leading to overheat failures of reheater tubes. 2. Austenitic steels can crack within hours. 3. Deposits and overheat failures of reheater and superheater tubes. 4 and 5 as 2 and 3 above.

Sodium Hydroxide Carryover As 1 to 5 above.

+


6-19


Table 6-18 High Boiler Carryover and Contaminated Steam (Caustic Treatment) Diagnosis

Actions to be Taken

Consequences+

1. Reduce load and boiler pressure. 2. Eliminate sources of high hydroxide alkalinity, suspended solids and organics, as needed. 3. Verify correct operation of condensate polishers and the regeneration. 4. Inspect superheater and reheater for deposits. 5. Wash turbine as needed (without disassembly at reduced speed). 6. Inspect boiler drum internals and level control. 7. May need to repeat carryover tests.

1. Deposition and corrosion in the turbine. 2. Possibility of scale formation in the superheater and reheater and short or longterm failures. 3. Sticking turbine valves.

1. Reduce load. 2. Adjust drum level control as needed. 3. Decrease sodium concentration in the boiler water. 4. Analyze boiler water and feedwater for organic contamination and eliminate the source.

1. Stress corrosion cracking of turbine, superheater and reheater. 2. Possibility of turbine efficiency and MW loss due to deposit buildup. 3. Sticking turbine valves and possibility of destructive overspeed.

High Boiler Carryover 1. Check with operators about drum level, boiler pressure and changes of boiler pressure and load. 2. Analyze samples of feedwater and boiler water and for alkalinity, TOC and suspended solids 3. Check makeup system and polisher sodium. 4. Compare current carryover with commissioning measurements at same load. 5. If the problem persists, inspect the turbine. 6. Check % carryover* (total) and compare with baseline. Contaminated Steam 1. Check boiler water pH*, conductivity*, chloride and sodium* concentrations. 2. Check NaOH pumping rate. 3. Check blowdown rate. 4. Check drum level control in the control room and on the drum sight glass. 5. Check sodium* in condensate and in polisher effluent. +


6-20

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-19 Impurity Introduction into Turbine by Attemperating Sprays

Diagnosis 1. Analyze steam before and after attemperation. 2. Discuss with Operations the use of attemperating sprays. 3. Check feedwater sodium, cation conductivity and copper, and latest trends.

+

Actions to be Taken

Consequences

1. Stop ingress of impurities into feedwater (condensate polishers, condenser leaks, makeup). 2. Minimize the use of attemperation, if the problem persists 3. Measure metal temperatures.

1. Deposit buildup and corrosion in the turbine. 2. Deposition of salts in the superheater and reheater leading to tube failures from pitting during shutdown. 3. High copper in feedwater could lead to deposits in superheater & HP turbine.

+


Table 6-20 High Concentrations of Acids and Salts in Steam Diagnosis

Actions to be Taken

Consequences+

1. Check cation conductivities* of steam and condensate and their trends. 2. Analyze the troubleshooting parameters (chloride, sulfate, etc) in steam and condensate. 3. Check performance of condensate polishers. Check for condenser leaks. 4. Measure air in-leakage. 5. Check use of attemperating water for superheater and reheater attemperation.

1. Reduce load. 2. Correct drum level control, if needed. 3. Find and correct the source of acid impurities. 4. Reduce air in-leakage.

1. Pitting and corrosion fatigue in the LP turbine. 2. Flow-accelerated corrosion in the wet steam piping.

+


6-21

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-21 Air In-leakage to Hotwell Actions to be Taken

Consequences+

1. Using inert gas leak detector, find locations of air in-leakage. 2. Activate additional air ejectors or vacuum pumps, as needed, to keep condensate oxygen & cation conductivity within normal values. 3. Typical air in-leakage locations are LP turbine glands, condenser/ LP turbine base joint, LP turbine rupture disks, No 1 and 2 feedwater heater safety valves and side glasses.

1. Fast exhaustion and release of impurities from condensate polishers, if in operation. 2. FAC and increased corrosion of copper based alloys in feedwater. 3. Increased iron and copper scale buildup in the boiler steam generating tubes. 4. FAC in the economizer and LP turbine. 5. Increased propensity for oxygen pitting of carbon steel components during shutdown.

Diagnosis Air In-leakage above the Hotwell Line 1. Check air eject of vacuum exhaust air flow and its history. 2. Check condensate pump discharge oxygen* and cation conductivity* and their trends for different recent loads. 3. Where air in-leakage is high and oxygen is low, it is an indication of air in-leakage above the water line.

Air In-leakage below the Hotwell Line As 1 and 2 as above. 3. Where air in-leakage is high and oxygen is high, it is an indication of air in-leakage below the water line. 4. Condensate pump discharge oxygen* and cation conductivity* increased or drifting up is confirmation. +


6-22

As 1 and 2 as above. 3. Frequent air in-leakage locations are suction of condensate pump, various hotwell connections, improperly designed makeup addition, pits and holes at the bottom of the condenser hotwell.

As 1, 2, 3, 4 and 5 as above.


Table 6-22 Colloidal Silica and Iron through the Makeup System Diagnosis

Actions to be Taken

Consequences+

Colloidal Silica 1. Compare silica (iron) in makeup effluent, condensate storage tank, condensate polisher in and out, blowdown & saturated or hot reheat steam. Feedwater silica (iron) higher than polisher effluent or condensate pump discharge indicates colloidal silica (iron). 2. Analyze grab samples for total organic carbon at the makeup system effluent and condensate storage tank. High TOC may be due to presence of organic colloids containing silica (iron). 3. Filter makeup system and condensate storage tank samples and analyze filtrates. 4. Audit the pre-treatment and makeup treatment system and measure variation of the source water iron, silica and organics.

1. Correct the situation leading to the pickup of colloids at the influent to makeup system (raw water). 2. Check and correct the pretreatment system. 3. Modify the makeup system. 4. Operate with increased blowdown as needed.

1. Scale buildup in the boiler and turbine, lower turbine output and efficiency.

As 1, 2, 3 and 4 as above.

1. Increased rate of scale buildup in the boiler. 2. Condensate polisher resin fouling. 3. Foaming, increased boiler carryover and turbine deposits.

Colloidal Iron As 1, 2, 3 and 4 as above.

+


6-23

EPRI Licensed Material PC and CT Chemistry Control and Corrective Actions Table 6-23 Boiler Layup Problems∆ Diagnosis

Actions to be Taken

Consequences+

Lost Nitrogen Blanket 1. Check nitrogen pressure and supply system.

1. Restore nitrogen supply. 2. Fix nitrogen leaks as needed.

1. Boiler corrosion. 2. Corrosion-generated iron and copper oxides will precipitate and redeposit in the boiler. Some will be carried into the superheater and turbine during startup. 3. Possible effects on turbine MW generating capacity and efficiency. 4. Iron and copper fouling of condensate polishers.

1. Restore pH by injecting ammonia and circulating boiler water. 2. If there is a high concentration of suspended oxides in the boiler water, drain and refill the boiler.

1. Dissolution and subsequent redeposition of iron and copper oxides in the boiler. 2. Carryover of oxides into the superheater and turbine during startup. As 3 and 4 above.

1. Restore nitrogen supply. 2. Fix nitrogen leaks as needed.

As 1, 2, 3 and 4 above (for lost nitrogen blanket).

Low Boiler Water pH 1. Take duplicate sample of boiler water, measure pH* and conductivity*.

High Boiler Water Oxygen 1. Check nitrogen pressure and supply system. +

Consequences if not corrected Also see Appendix F *Core parameter

∆

6-24


6.3 INFLUENCE OF TREATMENT CHEMICAL PURITY ON CYCLE CHEMISTRY It should be clearly understood that the EPRI Chemistry Guidelines for PC and CT and the tables on chemistry excursion troubleshooting and corrective actions generally assume that the treatment chemicals applied to the cycle are not a significant source of impurities that would influence parameters such as cation conductivity, sodium, chloride, sulfate, silica and total organic carbon. However, this is not always the case. Fortunately, the frequency of cycle chemistry excursions traceable to variations in chemical purity is low; they are much more likely to be problematic in drum boilers that do not feed solids based treatments than in boilers operating on PC or CT. Only very rarely do chemical purity deviations result in significant and apparent excursions from normal chemistry in the steamwater cycle. In most cases, only very minor changes in cation conductivity and perhaps pH will be observed. Such changes are often and usually correctly associated with variations in other unit-specific conditions such as condenser air in-leakage rates, makeup water consumption rates, makeup water organics content, or contamination attributed to (sometimes intermittent) condenser leaks. Therefore, cases may exist where treatment chemical purity contributes to the cycle chemistry problem but is not even considered as a possible cause for the out of specification readings. Whenever an investigation of a water chemistry excursion using the tables presented earlier in this section (or, preferably, versions of these generic response guidelines customized to individual units in the fleet) fails to conclusively identify the responsible root cause of the excursion, the purity of the treatment chemicals should be evaluated. This may be accomplished in a number of ways. These include:

•

Collection of a sample of each treatment chemical in use for laboratory analysis-this approach is obviously effective but may require an unacceptable amount of time, depending on the analytical capabilities of the organization.

•

Analysis of condensate (or feedwater) samples collected on each side of the chemical injection point(s)-this approach is not likely to be beneficial except in cases of fairly significant excursions that have no other apparent cause.

•

Isolation of the suspect chemical feed system(s) for a brief period and observing the impact on cycle chemistry-this approach is most effective when the unit has satisfactory levels of reliable on-line instrumentation, however, other chemistry parameters such as pH and ORP will also be affected by this action.

•

Isolation of the suspect chemical feed system(s) and simultaneous startup of backup feed system(s) with treatment chemicals of satisfactory purity-this is clearly a preferred practice in units that provide redundant feed systems.

6-25


In cases where sampling of the treatment chemical for analysis is desired, consideration should also be given to dilution water purity. This means that in units where the treatment chemicals are diluted on-site prior to injection to the cycle, samples of diluted and undiluted chemical solution should be collected. The last approach indicated can most easily be followed if a new supply of the chemical is applied to the unit before all of the existing supply has been consumed. As stated previously, cycle chemistry deviations caused by impurities in the treatment chemicals are generally infrequent events of very minor magnitude. Therefore, the possibility exists that extended and undetected unit operation with low levels of impurities present in the treatment chemicals applied could ultimately result in or contribute to chemistry related damage in the boiler and turbine. Traditionally, organizations have avoided such concerns through development of chemical purchase specifications and through testing of chemicals received from suppliers to verify compliance with specifications. However, in today’s competitive business climate, this approach is all too commonly minimized or eliminated. In those organizations that purchase treatment chemicals on the open market and emphasize price rather than purity, the importance of operating a unit with a new chemical supply and evaluating the chemistry prior to consuming all of the previous supply should be readily apparent. Ultimate responsibility for development of optimum treatment chemical specifications applicable to individual fossil plants must remain with the organization. One reason for this is that there are differences in chemical manufacturing, processing, dilution, packaging and labeling practices around the world and there is no single chemical purity specification that can adequately encompass the variety of product strengths and grades that may be available at any given location. Also, the organization’s chemical purity specification usually becomes part of a legal agreement between the organization and the chemical supplier. Again, there are clearly no general product specifications that will be universally applicable to all conceivable business arrangements for provision of the needed chemical treatments. In developing individual treatment product specifications, the ultimate goal is to ensure that normal usage of the treatment chemical will not result in deviations from any of the plant specific chemistry limits. This is applicable to normal operating conditions as well as chemistry limits for any other operating conditions, such as wet layup, for which the chemical is to be used. This goal is, in general, readily attainable providing that only fully deionized water is used to dilute the active chemical. However, it is most reliably attained when ammonia is used for feedwater pH control and hydrazine (when not prohibited and for reducing AVT in cycles with copper) is used as the reducing agent. Use of neutralizing amines for pH control and of alternative reducing agents, whether due to site prohibitions on hydrazine usage or for other reasons, makes it more difficult to attain guideline chemistry values, particularly cation conductivity, by virtue of the fact that these treatment chemicals are subject to thermal degradation, resulting in formation of carbon dioxide and/or organic acids. Special care must therefore be taken when establishing customized chemistry limits and action levels for fossil units that are treated with these chemicals. Also, it should be recognized that the effects these treatments have on the cycle chemistry tend to make it more difficult to identify the responsible cause of minor deviations from normal chemistry.

6-26


A OXYGEN REMOVAL TECHNIQUES FOR TREATED MAKEUP

Previous Sections have described the importance of maintaining strict control to prevent dissolved oxygen and carbon dioxide from entering the heat cycle in order to minimize iron and copper corrosion and transport. It was noted that 40% of utilities surveyed by EPRI put aerated water directly into the cycle, while 66% put makeup into storage tanks which are vented to atmosphere.(1) It is obvious that these types of operating procedures are counter- productive if corrosion is to be fully controlled. This appendix describes methods by which makeup water can be deoxygenated (and some of these methods also remove carbon dioxide) and stored in such a way as to prevent further aeration. Another important source of ingress of oxygen and carbon dioxide is air in-leakage via negative pressure in the condenser and low end of the heat cycle. This subject is addressed in Appendix C.

A.1 VACUUM DEAERATION Vacuum deaeration is a process, which removes both oxygen and carbon dioxide. A typical vacuum degasifier, which accomplishes this task, is shown in Figure A-1. This equipment consists of a packed tower with a vacuum pulled from beneath the packing and with water sprayed over the top of the packing. The water trickles down through the packing, which provides a large surface area for removal of oxygen and carbon dioxide. The vacuum induces vaporization of water and produces an upward flow of water vapor together with oxygen and carbon dioxide. The water is collected in the bottom of the tower and is pumped from there to the next stage of the system.

A-1

EPRI Licensed Material Oxygen Removal Techniques for Treated Makeup

Water inlet

Access and packing installation opening

Packing material

Packing removal connection Vapor removed by vacuum pump

Degasified water storage

Figure A-1 Vacuum Degasifier

The process of vacuum degasification is based on Henry’s Law, which states that the concentration of an ideal gas in solution is proportional to its vapor pressure (or partial pressure). Therefore, if the pressure over a solution is reduced by a vacuum, dissolved gases must leave the water and enter the vapor phase to establish a new equilibrium. Increasing the vacuum increases the degree of gas removal. The vacuum supply to the degasifier can either be steam jet ejectors or vacuum pumps. The degasifier is typically located preceding the first anion unit of a demineralizer train in order to minimize the ionic loading to the anion unit by removing carbon dioxide. Forced draft decarbonators effectively remove carbon dioxide but saturate the water with oxygen. Typical removal levels for a vacuum degasifier are: carbon dioxide to 5 ppb and oxygen to 10-20 ppb. Increasing the packing height in the tower and increasing vacuum will result in higher gas removal. The vacuum degasifier is quite tall, typically in the range of 25-40 feet (7.6–12.2 m). One reason for this is to provide room for the packing material and the other is to provide sufficient suction head for the transfer pump. Because of its height, vacuum degasifiers may be difficult to backfit into an existing power plant arrangement. This equipment requires considerable maintenance in order to maintain constant vacuum conditions. In addition, the vacuum equipment is relatively energy intensive to operate.

A-2

EPRI Licensed Material Oxygen Removal Techniques for Treated Makeup (2) For a typical 300 gpm vacuum degasifier unit, the following costs are estimated in year 2003 dollars:

Capital Cost: $135,000 Installation Cost: $70,000 to $215,000 depending on location and complexity of installation. Operating and Maintenance (O&M) Cost: 3 to 5% of capital cost per year for normal operations.

A.2 CATALYTIC OXYGEN REMOVAL The use of hydrazine or other reducing agents to remove oxygen is commonly practiced at many power plants and is usually fed at the condensate pump discharge or the outlet of the condensate polisher (if one is installed). The reaction between hydrazine and oxygen can be expressed as follows: N2 H4 + O2 → N2 + 2H2O

(eq. A-1)

The reaction is stoichiometric, requiring equal amounts of hydrazine and oxygen for the reaction to proceed to completion. At ambient temperatures this reaction proceeds very slowly. However, in the presence of a catalyst, such as activated carbon, the reaction proceeds rapidly. Therefore, this technology can and is being used to remove oxygen from makeup water. The equipment consists of a vessel containing activated carbon to which hydrazine is fed using a positive (3) displacement chemical pump. The process is patented and is said to achieve makeup oxygen concentrations less than 10 ppb. Overall costs for the catalytic process are about $1.25 to $1.75 per 1000 gallons depending on system flow rates. This catalytic process can treat the water either going to storage or coming from storage. However, when large quantities of water are required, such as during startups, it may be impractical to provide equipment of sufficient size to meet such demands when processing water from storage. One disadvantage of this catalytic process is that carbon dioxide is not removed. Therefore, when used in demineralizer systems treating waters high in bicarbonate alkalinity, operating costs can be expected to be substantially higher than when using a vacuum degasifier, for example. However, when the catalytic systems are used in conjunction with a membrane system, both carbon dioxide and oxygen can be simultaneously removed (see discussion which follows under membrane systems). Another disadvantage associated with the carbon bed system is that an ion exchange bed is (3) required to remove excess hydrazine and carbon leachables from the effluent stream .

A-3


A.3 MEMBRANE SYSTEMS Membrane systems are presently utilized today for many applications, including reverse osmosis, ultrafiltration, microfiltration, and electrodeionization. Another membrane process, which is achieving popularity in water treatment applications, is the use of microporous, hydrophobic membranes for the transfer of gases to and from aqueous streams. Gas transfer membranes (GTM’s) for the removal of oxygen and carbon dioxide from makeup water are constructed of an array or bundle of hollow fiber hydrophobic polypropylene membranes which are gas permeable and water impermeable. The hollow fiber configuration, with the gas phase inside the fiber and the liquid phase outside, provides high surface area for maximum efficiency for gas removal processes. One such design encloses the hollow fiber array within a pressure vessel, which includes an internal baffle to promote turbulent flow and mass transfer. An example of this equipment is shown in Figure A-2. Sweep gas

Vacuum and/or sweep gas Distribution tube

Baffle

Cartridge

Aqueous stream

Aqueous stream Hollow fiber membrane

Collection tube

Housing

Figure A-2 Components of a GTM Unit

By altering the partial pressure of a gas at the air/water interface, a driving force is created which results in the mass transfer of a gas into or out of solution. There are three ways to alter the partial pressure of a gas in a GTM system:

•

Apply a vacuum to the inside of the hollow fiber

•

Use a “sweep” gas, usually nitrogen, which effectively lowers the partial pressure of other gases such as oxygen and carbon dioxide

•

Use both a vacuum and a “sweep” gas

A-4


The effectiveness of this membrane technology is governed by several factors:

•

Water flow through the system

•

Water temperature

•

Amount of vacuum applied

•

Volume and purity of the sweep gas

With a suitably designed system, oxygen levels of less than 1 ppb in the effluent water stream are said to be achievable. An example of a GTM system used to remove oxygen from boiler feed (4) water at a petroleum refinery is shown in Figure A-3. In this case, two parallel trains of GTM units are provided, each having 3-10 inch (7.6–25.4 cm) membrane contactors connected in series. The nitrogen sweep gas is 99.9% pure, and flows at a rate of 1.5-2.0 scfm. Vacuum requirements are ~ 25 acfm at –28 inches of mercury. The 200 gpm (760 litres/min) of product water contains less than 15 ppb of dissolved oxygen. An example of a hybrid system which combines both GTM’s for carbon dioxide and a catalytic (5) oxygen removal system is shown in Figure A-4. This is an outsourced plant in use at a nuclear facility. 99.9% N2 Reverse osmosis

1 Municipal water 150 - 390 gpm

Concentrate

2

3

Product water 200 gpm

GTM Skid #1 Vacuum

Ion exchange softening

Activated carbon

5µ depth filter

99.9% N2 Reverse osmosis 1

Concentrate

2

3

GTM Skid #2 Vacuum

Figure A-3 GTM System Process Flow Diagram (Note: 1 gal = 3.785 litres)

A-5


Anthracite filters Well water

Reverse osmosis Storage tank and pump

Hybrid Deoxygenation System GTM skid Carbon catalyst Sweep air Vacuum pump Hydrazine

RO reject

Mobile Ion Exchange Polisher

Up to 300 gpm <5 ppb D.O. Figure A-4 GTM System Process Flow Diagram (Note: 1 gal = 3.785 litres)

This GTM system consists of three skid-mounted contactors in parallel, a vacuum pump, valving to allow an air sweep, and instrumentation to monitor vacuum level and air sweep volume. The membrane system is primarily for carbon dioxide removal, but also removes some oxygen. (The use of an air sweep rather than nitrogen probably limits the extent to which oxygen can be removed in this part of the system). Hydrazine injection occurs at the GTM effluent using a small, diaphragm positive displacement pump. Any remaining oxygen in the RO permeate water reacts with the hydrazine as it passes through two activated carbon vessels connected in parallel. The water then enters a three-bed ion exchange system for further polishing. This system consists of a strong acid cation, strong base anion and a mixed bed unit. This hybrid system produces up to 300 gpm of purified water at less than 5 ppb of dissolved oxygen. It is significant to note that the deoxygenated water is stored in a tank with a nitrogen blanket. The tank effluent is tested daily for dissolved oxygen, which consistently measures below 5 ppb. A southwestern electric utility plant chose to utilize gas transfer membrane technology (GTM) to reduce their dissolved oxygen levels to below 20 ppb prior to their demineralized water storage (6) tank . Their system consists of conventional ion exchange plus GTM to produce 500 gpm of demineralized and deoxygenated water to storage. A unique feature of this system is that the GTM is located after the strong acid cation and before the strong base anion unit. Thus, the GTM is acting as a replacement for a forced draft decarbonator to remove the carbon dioxide generated in the strong acid cation unit. Locating the GTM at this point appreciably reduces the load on the anion unit while also removing oxygen.

A-6


These GTM systems can be skid mounted and installed easily in a minimum of space. One such system(7) consists of two 1x1x1 (3 in series) arrays on two skids which produce a combined effluent of 200 gpm at 15 ppb (or less) of dissolved oxygen. The two skids each occupy less than 370 cubic feet of space. Overall costs for the GTM process are about $0.65 to $1.25 per 1000 gallons (3785 litres) depending on the degree of oxygen removal required and the systems flow rate.

A.4 MAKEUP WATER STORAGE The results of oxygen removal by any of the above are completely negated if the makeup water is stored in tanks, which are vented to atmosphere. There are several ways that tanks can be equipped to prevent air ingress:

•

Plastic or rubber diaphragms

•

Floating covers

•

Nitrogen positive pressure purging

The use of plastic or rubber diaphragms to protect condensate storage tanks from oxygen ingress has been very limited in the electrical generating industry. This is the result of relatively high installed costs, high maintenance costs and the fear (perhaps justly so) that the diaphragm might fail causing a major disruption in power plant operations. Also, diaphragms would be difficult, and possibly impossible to install in existing tanks. Floating covers have been in use, especially in petroleum refinery tanks, for many years. Their purpose in this case is to prevent undue evaporation of volatile liquids. Here again, installation of floating covers has not been popular at electrical generating plants for some of the same reasons as given above for diaphragms. Backfit of floating covers on existing condensate tanks, while perhaps possible, would certainly require considerable engineering and would undoubtedly be quite expensive. Probably the most feasible means to prevent oxygen ingress to condensate storage tanks is to provide a nitrogen blanketing system. Such a system is relatively easy to install on existing tanks and inexpensive to operate, given the low cost of nitrogen and the modest amount of nitrogen required to operate the system. (8) One such system is shown in Figure A-5. This system has a nitrogen bottom sparger (A), inlet pipe sparger (B), nitrogen supply and controls (D), oxygen or conductivity analyzer (F), tank pressure control (C), and an overflow pipe (E). The sparging itself is through sintered stainless steel sparging elements. Two pressure regulators control the nitrogen pressure to the spargers; one for the bottom sparger and one for the inlet pipe sparger. This is followed by a passive critical orifice flow element. This element prevents the nitrogen flow from exceeding the capacity of the tank pressure control siphon, which vents the excess pressure.

A-7


By having a sparger in the inlet water supply to the tank, any residual dissolved oxygen in that water should be effectively removed. The sparger in the bottom of the tank will prevent the ingress of air. The following costs are estimated to be associated with the system shown in Figure A-5(8) in year 2003 dollars: Capital Cost: $22,000 to $32,000 Installation Cost: $5,500 to $11,000 Nitrogen Consumption Costs: $110 per month assuming storage tank inflows and outflows are not excessive. The system in Figure A-5 would most probably use cylinder nitrogen, because of the relatively low consumption. The purity of cylinder (or cryogenic) nitrogen is 99.995%. Other, simpler nitrogen blanketing systems merely provide a slight positive pressure in the storage tank by using a small continuous flow of nitrogen applied at the top of the tank. This design, as compared to the above, probably requires a larger consumption of nitrogen and lacks the obvious advantage of the purge system for incoming water to the tank. While removal of oxygen from makeup may appear costly at first glance, it is essential to the proper control of iron and especially copper in the feedwater cycle. In the overall picture, the use of deaerated makeup water will produce savings, since iron and copper corrosion and transport will be reduced. An innovative use for GTM systems involves minimizing oxygen ingress to an existing (9) demineralized water storage tank at a nuclear power generating facility . In this installation the inlet water to the tank has been previously deoxygenated to less than 10 ppb. The tank is equipped with a bladder to miinimize oxygen ingress. However, without continuous oxygen removal, storage tank levels would gradually reach 1-2 ppm. A GTM system was installed on a constantly recirculating “kidney loop” in a full-flow 100 gpm pilot plant test. Dissolved oxygen levels in the storage tank were then consistently maintained to less than 10 ppb. The conclusion reached from the full flow pilot plant test was that a recirculating GTM system is a viable alternative to installing a nitrogen blanket, floating roof or diaphragm for maintaining low levels of dissolved oxygen in a storage tank.

A-8


Sparging element

C

E Top View Pipe support Sparging element

Inlet

B A

Outlet

Orifice F Regulator Option

D

Figure A-5 A Storage Tank Nitrogen System and Sparging Elements(8)

A-9


A.5 References 1. “Survey of 63 Utilities”. Fifth International Conference on Fossil Plant Cycle Chemistry. EPRI Proceedings. Edited by B. Dooley and J. Matthews TR-108459. December 1997. 2. S. Dominick, U.S. Filter Corporation, Personal Communication to A. Aschoff, August 2000. 3. R. C. Dickerson, W. S. Miller. U. S. Patent 4,556,492, “Deoxygenation Process”. December 3, 1985. 4. P. H. Kinghorn and W. E. Haas, “Low Level Deoxygenation of Boiler Makeup Water Using Gas Transfer Membranes”. Proceedings of the International Water Conference, p. 527, October 18-20, 1999. 5. S. B. Gorry, W. E. Haas, and J. W. Mahaffee, “Effective Deoxygenation by a Process Combining Gas Transfer Membranes with Catalytic Oxygen Reduction”. Proceedings of the International Water Conference, p. 535, October 18-20, 1999. 6. R. M. Mode, W. E. Haas and W. Greene, “Dissolved Oxygen Reduction by Membrane Technology to Reduce Feed Water Corrosion”. Proceedings of the International Water Conference, p. 406, October 20-24, 2002. 7. Ecolochem, Inc. Application Bulletin, December 5, 1999. 8. O. Jonas, Jonas Inc., Personal Communication to A. Aschoff, April, 2000. 9. J. Helmrich, W. E. Haas and J. E. Staton, “Maintaining Storage Tank Dissolved Oxygen Levels Utilizing Gas Transfer Membranes”. Proceedings of the International Water Conference, p. 207, October 22-26, 2000.

A-10


B OXIDATION-REDUCTION POTENTIAL (ORP)

B.1 INTRODUCTION Research investigating the electrochemistry of copper alloys under feedwater heater conditions suggests that maintenance of a stable cuprous oxide layer requires tight control of oxidationreduction potential (ORP), also referred to as redox potential, plus dissolved oxygen and pH. (As explained later, ORP, as measured in fossil plants, reflects a balance between the concentrations of dissolved oxygen and reducing agent in the feedwater, although the value measured in a given unit is also dependent on other factors.) Results of these research activities are summarized in a guideline document for copper (1). The practical implications of the copper research are reflected in the new target values for mixed metallurgy cycles indicated in this (2) report and in an earlier guideline report for all-volatile treatment (AVT) . Further, electrochemical corrosion experiments have pointed out that, during unit shutdowns, elemental copper and cuprous oxide will be converted to cupric oxide in a matter of hours, unless reducing chemistry is maintained; research assessments of various copper alloys used in fossil plants has (3,4) indicated that cuprous oxide is most stable at ORP values of around –350 to –300 mV . Effective implementation of reducing AVT feedwater chemistry in drum type fossil units where the boiler water is treated with phosphate or caustic therefore requires that on-line surveillance of ORP be practiced and that action be taken as needed in response to deviations from target values for this parameter in the cycle. EPRI has, for many years, been interested in real time measurements of corrosion rates in generating plants and considerable effort has been devoted to development of analyzers for this purpose. Commercially available monitors are available for surveillance of the electrochemistry of the operating environment of interest. Field experience with these devices is most extensive in the nuclear power industry, with both in situ and sidestream analyzers in use, but generally limited in the fossil industry to ORP analysis of conditioned samples. Most of the early field experience in fossil plants has been with ORP surveillance of final feedwater at the economizer inlet. However, findings of EPRI’s Program Copper clearly indicate the deaerator inlet to be the preferred monitoring point for ORP due to the difficulty of maintaining reducing conditions in (1) the low pressure heaters .

B-1

EPRI Licensed Material Oxidation-Reduction Potential (ORP)

Fossil plant experience with ORP measurement has shown the benefits of using ORP analysis as part of the control strategy to minimize corrosion and transport of copper-based feedwater heater materials. This experience also indicates that confusion exists, with respect to the physical meaning of ORP measurements, interpretation of ORP data and its relation to other chemistry data and, in some cases, how to set up an ORP monitoring capability. This appendix, adapted from a technical paper (5) prepared by a team of electrochemistry, materials and corrosion experts, provides the information fossil plant personnel need to be aware of when establishing ORP monitoring capabilities in the mixed metallurgy units for which they are responsible.

B.2 OVERVIEW OF ELECTOCHEMICAL PHENOMENA INFLUENCING CORROSION PROCESSES IN FOSSIL PLANT UNITS Over the past twenty years, researchers, power plant operators and chemists alike have come to realize that all forms of corrosion damage throughout the heat transport circuits of thermal power plants (fossil and nuclear), including general corrosion, pitting, stress corrosion cracking, corrosion fatigue, and crevice corrosion, are primarily electrochemical phenomena that involve (6) charge transfer at interfaces . As with any charge transfer process, the rate at which charge transfer occurs (the current) depends upon the voltage applied across the system. Faraday’s Law provides the link between the current density across the interface, due to the oxidation of the substrate metal or alloy, and the rate of mass loss (the “corrosion rate”). This link is the Faraday Constant, equal to 96,487 coulombs per gram equivalent – which means that 1 gram equivalent of metal will be lost when 96,487 coulombs of charge pass across the metal/corrodent interface. In electrochemical and corrosion systems, the voltage of a corroding metal with respect to a suitable reference is referred to as the corrosion potential (Ecorr). Other commonly used names for Ecorr are open circuit potential (OCP) and electrochemical corrosion potential (ECP). In addition, there exist various critical potentials above or below which certain types of corrosion occur (Table B-1). Thus, general corrosion, pitting corrosion, stress corrosion cracking, corrosion fatigue, flow-accelerated corrosion, and crevice corrosion occur only at potentials that are more c c , E cfc , E FAC , and E cc , positive than the corresponding critical values (referred to as E gc , E cp E scc respectively) at a rate that depends strongly upon the (positive) over potential, which is defined as = E – Ec, where E is the applied or spontaneous potential (e.g., Ecorr or ECP). Because of this close relationship between the potential and the occurrence of various forms of corrosion, potential control represents a powerful strategy for minimizing or preventing deleterious forms of corrosion in power plant heat transport circuits (in fact, for all corrosion systems) (6). For example, if the corrosion potential (Ecorr or ECP) is displaced to a value that is more negative than E gc , general corrosion can be prevented. This is the technical basis for cathodic protection, as extensively practiced in the pipeline and marine industries; in this case, the potential of the metal (now no longer the open circuit or corrosion potential) is displaced by coupling the system to a more active metal (e.g., Zn, Al, Mg) or by passing a direct current through the environment (e.g. water, soil) to the metal from an anode installed especially for this purpose. Frequently, two or more corrosion processes, each characterized by a critical potential, occur on the same substrate under the same conditions.

B-2


In a sense, this is always the case, because all engineering alloys suffer general corrosion, but most systems also display simultaneously or under slightly different potential conditions, one or more forms of localized corrosion. This is an important issue, because the blind application of potential control techniques may simply substitute one problem for another. Table B-1 Classification of Corrosion Processes Showing the Potential Conditions for their Occurrence (5). Form of Corrosion

Critical Potential

Condition for Spontaneous Corrosion and Morphology of Attack

General Corrosion

E gc

Ecorr > E gc . Uniform attack with point-to-point variation in metal loss less than 20% of the mean loss. Aspect ratio (depth/width of local corroded area) < 1.0.

Pitting Corrosion

E cp

Ecorr > E pc Localized attack in the absence of a mechanical load, with KI < KIscc. Aspect ratio (depth/width of local corroded area) > 1.0.

Stress Corrosion Cracking

c E scc

c Ecorr > Escc Localized attack in the presence

of a constant mechanical load, with KI > KIscc. Aspect ratio (depth/width of local corroded area) > 1.0. Corrosion Fatigue

E cfc

Ecorr > Ecfc . Localized attack in the presence of a time varying mechanical load, with ∆KI > ∆KIscc. Aspect ratio (depth/width of local corroded area) > 1.0.

Flow –accelerated Corrosion

c E FAC

c Ecorr > EFAC . Localized attack, generally of

low aspect ratio (L/D < 1.0) and influenced by fluid flow. Crevice Corrosion

E cc

c Ecorr > Ecc . Localized attack within a high

aspect ratio crevice (L/D > 5) in the absence of a mechanical load. Hydrogen Induced Corrosion

E Hc

Ecorr > EHc . Localized attack in the form of hydrogen induced fracture (HIF) or blistering (HIB), due to the ingress of hydrogen into the substrate.

B-3


Unfortunately, confusing terminology has evolved with regard to the term “potential”, with the result that the term is often incorrectly used or is taken to mean different things in different fields of power plant technology. Suitable working definitions of various potentials that are reasonably consistent with other fields of electrochemistry and corrosion science are summarized in Table B- 2. The potential of a metal exposed to a condensed, ionically conducting medium measured against a suitable reference electrode in the same system is referred to in the broad sense as the “electrochemical potential”. It should be noted that this terminology is in unfortunate conflict with similar terminology used in electrochemical thermodynamics in referring to the chemical potential that has been modified to include electrical work. The problem has been compounded in recent years by researchers and workers in the nuclear power industry referring to the corrosion potential as an “electrochemical (5) corrosion potential (ECP)”, particularly in studies dealing with BWRs . Another example is the use of the term “oxidation/reduction potential” (ORP) in fossil fuel power plant technology to describe what is in reality the redox potential of the system. The term “redox potential” has a specific, well-defined meaning in electrochemistry that is synonymous with ORP, as used in fossil plant technology. Table B-2 Definitions of Various “Potentials” Encountered in Corrosion Science and Engineering, Particularly in Thermal Power Plant Technology (5) Term

Common Usage in Power Plant Technology

Electrochemical Definition

Potential

Potential measured against a reference electrode.

Reversible work done in bringing a unit positive charge from a field free point at infinity to the point of interest.

Electrochemical Potential

Corrosion potential.

Potential of an indicator electrode measured against a reference electrode

Equilibrium Potential

Equilibrium potential

Potential of an electrochemical cell at equilibrium.

Standard Hydrogen Electrode Potential

Zero on potential scale in electrochemical thermodynamics. (seldom used).

Potential of a hypothetical hydrogen electrode (H2/H+) having unit activity of H+ and unit fugacity of H2. The potential of the SHE (standard hydrogen electrode) is arbitrarily assigned a value of 0 V at 25 oC/77 oF. The potential at any other temperature depends upon the standard state convention employed for the partial molal Gibbs energy.

Reference Potential

B-4

Potential on the standard hydrogen electrode scale adopted by a reference electrode at equilibrium that is invariant or variant in a known manner to the independent variables

EPRI Licensed Material Oxidation-Reduction Potential (ORP) Mixed Potential

Not commonly used.

Potential of a metal or semiconductor observed in a non-equilibrium system containing two or more charge transfer reactions, measured against a suitable reference electrode using a voltmeter having infinite input impedance.

Redox Potential

Oxidation-Reduction Potential (ORP)

Potential observed at an inert metal electrode sensing redox reactions with infinite exchange current densities. When more than one redox reaction is present, the redox potential is a particular example of a mixed potential.

Corrosion Potential

Open circuit potential of a corroding metal.

Potential of a metal or semiconductor that is itself participating directly in the charge transfer reaction, measured against a suitable reference electrode using a voltmeter having infinite input impedance. A particular example of a mixed potential.

“Open Circuit” Potential

Corrosion potential. Often mistakenly used to mean ORP.

Potential measured with respect to a reference electrode using a voltmeter having infinite input impedance.

Critical Potential.

Not commonly used.

Calculated or measured potential that marks the upper or lower boundary of a particular phenomenon, measured against a suitable reference electrode using a voltmeter having infinite input impedance

B.3 OXIDATION-REDUCTION POTENTIAL (ORP) OR REDOX POTENTIAL ORP has been used extensively to define operating ranges for the feedwater of fossil plants(7). Two particular examples are in mixed-metallurgy systems, where it is essential to keep the ORP in the reducing range, and in all-ferrous systems where optimum performance is achieved with oxidizing environments. Thus, the redox potential, which is commonly referred to as the ORP by personnel in fossil plants (see Table B-2), is becoming recognized as an important parameter in characterizing the chemical state of the heat transport circuit. It must be emphasized again that the term “redox potential” is already firmly established in electrochemistry, corrosion science, and geochemistry for the same quantity. It is also important to recognize that the ORP is not synonymous with the corrosion potential (Ecorr, OCP, or ECP), which is measured on the structural substrate of interest. Indeed, the redox potential and the Ecorr can differ by hundreds of millivolts and can respond quite differently to changes in system conditions, depending upon the nature of the corrosion reaction (which does not occur on platinum in the measurement of ORP).

B-5


As noted earlier, ORP is defined as the potential that is measured at an inert indicator electrode (e.g., platinum) in contact with a solution containing electrochemically active, redox species(6). Generally, the system is not at equilibrium and, furthermore, the redox species generally are not 2+ 3+ conjugately related (for example, if the solution contained Fe and Fe , the species are referred 3+ to as being “conjugate”, because they are related by the charge transfer reaction, Fe + e 2+ Fe ). Indeed, complex aqueous systems generally contain many conjugate pairs. These include, 2+ 3+ 1+ 2+ for example, in the case of mixed metallurgy fossil unit feedwater systems, Fe /Fe , Cu /Cu , + H2/H , and O2/H2O in a partially aerated aqueous solution contained within carbon steel and cuprous materials (pipe or tube), with H2 being produced by corrosion, for example. The Nernst equation (discussed further later) is often used incorrectly to calculate ORP in complex systems containing multiple conjugate redox couples; however, the use of the Nernst equation is possible only if a single conjugate couple is present, because, only in this case, can the system be in electrochemical equilibrium. Even then, the Nernst equation can be used only if the redox reaction is reversible and the system is in equilibrium. Reactions that are notoriously irreversible or quasi-reversible, for which the use of the Nernst equation is excluded, include the oxygen electrode reaction (OER). If more than one conjugate couple is present, the system is not (8) in equilibrium, and the potential is calculated as a mixed potential . The basis of this issue is examined below. In Figure B-1, is plotted the partial and total currents for two independent redox couples in the system, each characterized by an equilibrium potential ( E1e and E 2e , respectively) and an exchange current density ( i0,1 and i0, 2 , respectively). The curves were generated using the general Butler-Volmer equation for a redox reaction, R ↔ O + ne − , as (8):

i=

e (E−E

e

) / ba e

− e −( E −E

e

) / bc

1 e ( E − E ) / ba e −( E − E + − i0 il , f il , r

e

) / bc

(eq. B-1)

where ba and bc are the anodic (forward direction) and cathodic (reverse direction) inverse Tafel constants for the redox reaction, and il , f and il ,r are the corresponding mass transfer limiting currents. In fossil power _plant systems, Subscript “1” might refer to the hydrazine electrode reaction, N 2 + 4 H + + 4e ⇔ N 2 H 4 , and Subscript “2” to the oxygen electrode reaction (OER), O2 + 4 H + + 4e − ⇔ 2 H 2 O , but it is stressed that the reactions could be any of a large number of possible charge transfer processes involving redox species, depending upon the composition of the system. Equation B-1 is strictly valid for i0 << il,f, il,r., but it is a very good approximation if i0 is no greater than about 0.1 times the limiting current densities. This is normally the case in corrosion and redox systems.

B-6


Abs(Current Density, A.cm-2)

1e-5 o T = 25 C [H2] = 10 ppb

HER OER Total Current

1e-6 1e-7

[O2] = 5 ppb pH = 7 d = 10 cm v = 5 cm/s e E her = -0.351 Vshe e E oer = 0.757 Vshe Eredox = 0.139 Vshe

1e-8 1e-9 1e-10

Eredox (ORP)

1e-11 1e-12 1e-13 -1.0

E

E

e her

-0.5

0.0

0.5

1.0

e oer

1.5

Potential, Vshe Figure B-1 Calculated current-voltage curves for the hydrogen electrode reaction (HER), the oxygen electrode reaction (OER) and the total polarization curve (sum of those for the HER and OER) for a system containing 10 ppb H2 + 5 ppb O2 at 25 °C. Note that the ORP must lie between the equilibrium potentials for the HER and the OER.

B-7


A more accurate approach is to express the partial reactions for the forward and reverse directions as:

if =

e ( E−E

e

) / ba

(eq. B-2)

e

1 e ( E − E ) / ba + i0 il , f

and

ir = −

e −( E −E

e

) / bc

1 e −( E − E − i0 il , r

e

) / bc

(eq. B-3)

with

i = i f + ir

(eq. B-4)

Note that if and ir are positive and negative quantities, respectively, while the exchange current density, i0, is a positive quantity. Thus, the total current density for two redox reactions in the system becomes:

i = (il , f + il ,r )1 + (il , f + il ,r ) 2

(eq. B-5)

where subscripts “1” and “2” signify the two reactions. Solution of Equation B-5 for E at i = 0 (8) yields the redox potential (ORP) . Before proceeding further, it is necessary to express the parameters in Equations (2) and (3) in terms of specie concentrations. Thus, the exchange current densities for the two conjugate reactions can be expressed as: C i0 = i  R0  CR 0 0

  

α

 CO  0 C  O

   

β

(eq. B-6)

where i00 is the standard exchange current density, C R0 and CO0 are the concentrations of the reactant and product, respectively, in the standard state (assigned values of unity), and α and β are constants that are related to the reaction orders, which can be determined experimentally (common) or by examining viable reaction mechanisms (less common), and the stoichiometric coefficients for the reaction.

B-8


The limiting current densities are most conveniently expressed as dimensionless correlations between the Sherwood number and the Reynolds and Schmidt numbers for a particular flow geometry/regime in a specified solution (9). The dimensionless correlations are of the form: Sh = A Re a Sc b

(eq. B-7)

where Sh is the Sherwood number ( Sh = kd / D ), A is a geometry-dependent constant, Re is the Reynolds number ( Re = Vd /ν ), Sc is the Schmidt number ( Sc = ν / D ), k is the mass transfer coefficient, ν is the kinematic viscosity (dynamic viscosity divided by the density), D is the diffusivity of the species of interest, V is the bulk fluid velocity, d is the characteristic dimension, and a and b are geometry-dependent constants that also depend upon the flow regime (laminar, turbulent, or transitionary). The mass transfer coefficient is related to the mass transport-limited current density, iL, by: k = i L / nFC b

(eq. B-8)

where Cb is the bulk concentration of the species under consideration, n is the number of electrons involved in the oxidation or reduction reaction, and F is Faraday’s constant. Note that Re, Sc, and Sh are dimensionless quantities and that correlations of the type in Equation (B-7) have been developed for many systems. For example, the correlation for pipe flow under (10) turbulent conditions is given as :

Sh = 0.0165 Re 0.68 Sc 0.33

(eq. B-9)

which is valid over the range 8000 < Re < 200,000. Thus, it can be seen (Figure 1) that, in the case shown, the two conjugate couples have different equilibrium potentials, with the consequence that the couple that has the most negative equilibrium potential moves in the oxidation direction (R1 O1 + n2e-), while that with the R2). most positive equilibrium potential moves in the reduction sense (O2 + n2e Accordingly, a local, internal current flows between the two couples, with the partial anodic and cathodic currents being equal, resulting in a gradual decrease in the concentrations of R1 and O2 and to a gradual increase in the concentrations of O1 and R2. The potential at which the two (8) partial currents are equal is known as the redox potential . It should be pointed out that when multiple conjugate pairs are present in the system, the redox potential becomes a very complex function of the system composition and that the redox potential is not simply related to the equilibrium potential of any given conjugate couple. As the species concentrations change, due to the reactions indicated above, the equilibrium potentials move towards one another, as can be confirmed by examining the Nernst equations for each couple. This “fuel cell” effect will continue until the equilibrium potentials coincide, at which point the system attains a state of equilibrium. Note, however, that if the environment is very large compared with the indicator electrode surface area, or if the components are continually renewed, (e.g., by being supplied from an external source), both of which are normally the case in fossil plant feedwater, the concentrations of the electro-active components are effectively constant and hence so are the equilibrium potentials and the redox potential. B-9


Thus, in fossil plants, corrosion processes, once initiated frequently proceed until failures are experienced unless the electrochemical environment is changed. An important “rule-of-thumb” in analyzing ORP data is that the redox potential lies closest to the equilibrium potential for the “fastest” conjugate pair. It is sufficient to note at this juncture that the “speed” of a charge transfer reaction is measured by the exchange current density and the Tafel constants, with the faster reactions having larger values for the exchange current density and smaller values for the Tafel constants. If the components of one conjugate pair are at sufficiently high concentration, the redox potential may approximate the equilibrium potential of that couple. This is illustrated in Figure B-2 at low [O2] values, where the measured redox potential, as would be detected by a platinum indicator electrode, approaches the hydrogen (5) equilibrium potential . Note that because the oxygen electrode reaction (OER) is a very slow reaction, even on platinum, the ORP does not approach the equilibrium potential for the OER, even at oxygen concentrations as high as 10 ppm.

1.0

Eeoer

Potential, Vshe

0.8 0.6

o T = 25 C [H2] = 10 ppb

0.4

[O2] = X ppb

0.2

pH = 7 d = 10 cm v = 5 cm/s

0.0

Oxidizing ORP

Reducing

-0.2

Eeher

-0.4 -0.6 1e-4

1e-3

1e-2

1e-1

1e+0

1e+1

1e+2

1e+3

1e+4

1e+5

[O2], ppb Figure B-2 Calculated equilibrium potentials for the Hydrogen Electrode Reaction (Eeher) and the Oxygen Electrode Reaction (Eeoer) and the calculated redox potential (ORP) for the system 10 ppb H2 + X ppb O2 at 25 oC. Note that, generally, the ORP approaches the equilibrium potential for the HER at very low oxygen levels, but does not approach the equilibrium potential for the OER at very high oxygen levels.

B-10


Upon increasing [O2], and hence upon increasing the ratio [O2]/[H2], the potential is observed to shift sharply in the positive direction as oxygen exerts an increasingly strong effect over the potential of the electrode. This “sigmoid” dependence of the potential on the ratio of redox species is highly characteristic of redox systems and is a feature that strongly differentiates redox systems from equilibrium systems, where the potential varies linearly with the logarithm of the ratio of the electro-active species concentrations within a single conjugate system at equilibrium (i.e., as described by the Nernst equation). Figure B-2 provides a clear interpretation of the meaning of ORP as measured in a fossil plant feedwater system. Thus, the ORP reflects the balance between various conjugate redox systems, such that the resulting potential (ORP) is the potential at which the total interfacial current is zero (8,11) . Since the partial current for any given reaction (e.g., R1 → O1 + n1e) is a function (commonly linear) of the concentration of the reactant (in this case, R1), the ORP reflects the composition of the system with respect to electro-active species. Of prime importance, however, is the impact that these species have upon corrosion. While the ORP (which is measured using a platinum indicator electrode) must not be confused with the corrosion potential, the value of the ORP is a useful indicator of the corrosivity of the environment.

B.4 ELECTROCHEMICAL POTENTIAL (ECP) In the case of a corrosion potential measurement, the metal electrode may be the power plant component itself (such as a feedwater pipe or heat exchanger tube) or it may be a separate piece of metal (a “probe”) having a chemical composition similar to the component of interest and immersed in the environment of interest. When a metal, like iron, is corroding in an environment, such as water containing dissolved oxygen, the primary oxidation reaction is irreversible and heterogeneous:

Fe → Fe 2+ + 2e −

(eq. B-1)

This reaction, together with the primary reduction reaction,

1 / 2O2 + H 2 O + 2e − → 2OH −

(eq. B-11)

controls the measured corrosion potential. Theoretically, each corrosion product is thermodynamically stable only in a specific corrosion potential range. A change in the corrosion potential, for whatever reason, may induce a change in the identity of the corrosion product; for 2+ example from Fe to Fe3O4 to Fe2O3 as the corrosion potential is increased, and may change the reaction mechanism. As noted above, when the measuring electrode is platinum or some other noble, non-corroding metal, the potential measured is the ORP (redox potential). If the system contained only a single conjugate redox reaction, the primary oxidation reaction, occurring at the metal/environment interface, is simply the reverse of the reduction reaction, as noted above.

B-11


For instance, the partial oxidation reaction corresponding to Reaction B-11 would be:

2OH − → H 2 O + 1 / 2O2 + 2e −

(eq. B-12)

However, the rates of the forward and backward reactions of the oxygen electrode reaction (OER) are usually quite small, so that the oxygen electrode usually displays poor reversibility, even on platinum, and the redox potential seldom agrees with the equilibrium potential calculated from thermodynamics. On the other hand, the hydrogen electrode reaction is “fast” and the calculated and measured potentials are normally in excellent agreement, as indicated in Figure B-2 at very low oxygen concentrations with reducing agent present. When the substrate also oxidizes (corrodes), the picture is altered significantly in that the current due to that oxidation reaction must be added to the total current flowing across the interface. In the mathematical sense, this is equivalent to adding another redox reaction to the system, in which the reduced conjugate species is the metal substrate. Thus, charge balance at the interface can be expressed as: N

icorr ( E ) + ∑ iO / R , j ( E ) = 0

(eq. B-13)

j =1

where icorr ( E ) is the current density for the oxidation of the metal substrate and iO / R , j ( E ) is the current density for the jth redox couple, as given by Equation (1). The “E” contained in parentheses emphasizes that the currents are functions of the potential. Indeed, solving Equation B-13 for E yields the desired corrosion potential, commonly referred to by power industry personnel as the electrochemical potential (ECP). Equation B-13 forms the basis of the Mixed Potential Model (MPM), which has been used extensively and successfully for many years to calculate corrosion potential (ECP) values in (12) power plant coolant circuits (particularly in nuclear units) . This work, directed towards understanding and control of corrosion processes that effect nuclear plant components, in particular steam generators, has resulted in considerable interest in and application of ECP techniques. ECP measurement and the MPM, while certainly applicable in principle to fossil cycles, have not been evaluated or applied extensively to the operating environment which exists in these units. However, the value and benefits of the MPM and ECP monitoring to fossil units may be considered under EPRI’s Program Boiler Corrosion activities. In summary, while ORP and corrosion potential are measured in a similar way, the former reflects the oxidative/reductive power of the environment, while the latter reflects the corrosion reaction occurring on the metal surface. The corrosion potential normally (but not always) reflects the ORP, in that an increase in ORP will normally be accompanied by an increase in the corrosion potential, but ORP and corrosion potential seldom have the same value. ORP measurement is a useful tool for control of the feedwater chemistry in fossil units with copper alloy heaters. ECP has no clear role within the context of Cycle Chemistry Guidelines at this time but could become important in the future as work under Program Boiler Corrosion proceeds. B-12


B.5 REFERENCE ELECTRODE CONSIDERATIONS This discussion of ORP would be incomplete without consideration of reference electrodes, because it is in the choice of the reference electrode that the non-expert most commonly errs. Only potential differences can be measured with electrochemical methods; the role of the reference electrode is to provide a stable, system-invariant potential with which the potential of the indicator electrode (i.e., the electrode under study) can be placed on a rational scale. This rational scale preferably is the Standard Hydrogen Electrode (SHE) scale, or is a scale that is directly transferable to the SHE scale, so that the measured data may be related to appropriate thermodynamic data bases (e.g., Pourbaix diagrams). However, in some cases, only a stable potential is required and, hence, it is possible to employ a pseudo reference electrode in the measurement. Generally speaking, pseudo reference electrodes are non-equilibrium systems, so that a convenient way of classifying reference electrodes is whether or not they are equilibrium systems. It must be emphasized that the classification is sometimes artificial and arbitrary, in that even an equilibrium reference electrode may have a contribution to the measured potential from non-equilibrium processes, such as thermal diffusion and isothermal diffusion across a (13, 14) . liquid junction It is important to note that, in fossil plants, the ORP is usually reported as a voltage versus that of a Ag/AgCl (sat. KCl) reference electrode. The silver/silver chloride reference electrode is used because it is easy to prepare and maintain, it is exceptionally stable, and it is well understood (14, 15) . Because of the importance of this reference electrode, a brief thermodynamically discussion of the thermodynamics is given below. The potential determining equilibrium of the Ag/AgCl, KCl (aq) system is written as:

AgCl + e − ⇔ Ag + Cl −

(eq. B-14)

which yields an equilibrium potential ( E e ), according to the Nernst equation, of: o E e = E Ag / AgCl −

2.303RT log(aCl − ) F

(eq. B-15)

o where E Ag / AgCl is the standard potential and a Cl − is the activity of chloride ion in the internal o solution of the electrode. The standard potential is given as a function of temperature (t, C) as: o E Ag / AgCl =

0.23732-5.28194x10-4t-2.440055x10-6t2

(eq. B-16)

where R is the gas constant (R = 8.31441 J/K.mol), F is Faraday’s constant (F = 96,487 C/equiv), and T is the temperature in Kelvin (t° C + 273.15).

B-13


“Activity” is often referred to as being the “thermodynamic concentration” (as opposed to the “real concentration”) and is defined as: aCl − = γ Cl − (mCl − / mClo − )

(eq. B-17)

where γ Cl − is the activity coefficient (a dimensionless quantity), mCl − is the molal chloride ion concentration, and mClo − is the molal concentration of chloride in the standard state ( mClo − = 1 mol/kg). The activity coefficient may be measured experimentally (e.g. from solubility or isopiestic studies) or may be calculated using various models for electrolyte solutions (14). At any rate, γ Cl − is a function of the ionic strength of the solution, which, for a solution of a 1:1 salt such as KCl, is equal to the concentration. Examination of Equation B-15 reveals that the only independent variable is the activity of chloride. Accordingly, if this quantity is not controlled, considerable error may be incurred in using a silver/silver chloride reference electrode. The errors commonly arise from the fact that the electrode contains a liquid junction for electrolytic communication between the electrode internal environment and the external (test) environment. Flow of solution through this junction, commonly as a result of pressure changes in the environment, can result in considerable changes in chloride concentration in the internal environment and hence in significant errors in the reference potential. This problem is effectively overcome by using a saturated solution of a chloride salt (e.g., KCl) + + + in the inner compartment. Thus, for a saturated solution of MCl (commonly, M • Li , Na , K ), in which solid MCl is in equilibrium with the solution, the equilibrium solution process may be described as:

MCl ( s ) ⇔ M + + Cl −

(eq. B-18)

with the equilibrium condition being written as:

K sp = a M + aCl − = (γ M + m M + )(γ Cl − mCl − ) = (γ M + γ Cl − )m M + mCl −

(eq. B-19)

We now define the mean molal activity coefficient as γ ± = (γ M + γ Cl − )1 / 2 , whence: K sp = (γ ± m M + )(γ ± mCl − ) = aCl2 −

(eq. B-20)

1/ 2 , and substitution of this (because m M + = mCl − for the solution of a 1:1 salt). Thus, aCl − = K sp relationship into Equation (15) yields the Nernst equation as:

o E e = E Ag / AgCl −

B-14

2.303RT log10 ( K sp ) 2F

(eq. B-21)


However, Ksp is a system constant that is not a function of composition (but does depend on temperature and pressure) and hence no independent variable now exists in the Nernst equation. Accordingly, for a given temperature and pressure, Ee is fixed unequivocally, provided that solid KCl remains in the inner reference electrode compartments. At the saturation vapor pressure of pure water, the temperature dependence of the equilibrium potential for the Ag/AgCl/KCl(sat) on the SHE scale is given by: 2 3 Esssce=0.3835-2.1251e-4*T-1.4894e-6*T +2.1927e-10*T

(eq. B-22)

where T is in degrees Kelvin. An important advantage of using KCl over other alkali chlorides (e.g., LiCl, NaCl, CsCl, RbCl) as the internal electrolyte is that the transference numbers of K+ and Cl- are nearly equal at all practical temperatures, resulting in near suppression of the isothermal liquid junction potential (14) (another source of error), particularly if the KCl concentration is very high (e.g., saturated) . Another electrode that is sometimes used as a reference electrode for ORP measurements is the saturated calomel electrode (SCE), which is electrochemically described as (Pt)Hg/Hg2Cl2/KCl (aq). The thermodynamics of this system are identical to that developed above for the silver/silver chloride system, except that the standard potential refers to the half-cell reaction Hg2Cl2 + 2e = 2Hg + 2Cl . The same advantage accrues in this reference electrode in using a saturated internal solution as it does in the Ag/AgCl system, in which case the equilibrium potential for the SCE is given by: 2 3 Esce=0.3420+2.02242e-5*T-1.232835e-6*T +5.925097e-11*T

(eq. B-23)

on the SHE scale.

B.6 MONITORING ORP IN FOSSIL UNITS As indicated earlier, commercially manufactured monitors are available for use in surveillance of either ORP or ECP. Field experience with these monitors is quite extensive in the nuclear power industry, with both in situ and sidestream analyzers in use. In contrast, plant experience in fossil o plant units is limited, generally to ORP analysis of conditioned samples at 25°C (77 F). Most of the early field experience in fossil plants involved ORP monitoring of final feedwater at the economizer inlet for both reducing and oxidizing AVT and for oxygenated treatment. However, as a result of EPRI research, it is now known that ORP monitoring and control is essential (and designated as a core parameter) for minimizing copper corrosion and transport in the feedwater of units with mixed metallurgy feedwater systems and that reducing AVT is the treatment of (1, 2) choice for these units .

B-15


When making corrosion potential and ORP measurements, it is important to use a voltmeter with an input impedance that is sufficiently high that current does not flow through the reference electrode. Current flow would cause the reference electrode to “polarize” and give inaccurate readings. This is particularly important in test environments (such as high purity water) that have 12 very low conductivities. An electrometer with an input impedance exceeding 10 ohms is usually adequate in these low conductivity solutions, but higher input impedances may be necessary in some cases. The selection of a higher input impedance is often accompanied by an increased sensitivity to AC pick-up and “noisy” measurements. Where the electrometer allows selection of the input impedance, the optimum setting can be achieved by starting at a very high input impedance, then reducing the input impedance stepwise until the corrosion potential reading begins to drift or change in an unexpected fashion, indicating polarization of the electrodes. At this juncture, the input impedance should be increased immediately to a somewhat higher value so as to eliminate the polarization. Factors Influencing Fossil Plant ORP Readings Two important reactions that influence the ORP of a fossil power plant environment can be written as follows: + O2 + 4H + 4e ↔ 2H2O, E eoer =1.228 - 0.0591pH + 0.0148log(pO2)

2H+ + 2e- ↔ H2, E eher =0.000 - 0.059 1pH - 0.0296log(pH2)

(eq. B-24) (eq. B-25)

where E eoer and E eher are the equilibrium potentials for the oxygen electrode reaction and the hydrogen electrode reaction, respectively, at 25°C, stated with respect to a standard hydrogen electrode (SHE); and pO2 and pH2 are, respectively, the partial pressures of oxygen and hydrogen in the environment. The two-directional arrows in Equations B-24 and B-25 indicate the reversibility of the reactions, but it is emphasized that the OER seldom, if ever, displays the (16) equilibrium potential given by Equation B-24 . As previously discussed, the ORP lies closest to the equilibrium potential for the principal redox couple in the system (that with the greatest exchange current density), with the thermodynamically mandated relationship E eher < ORP < E eoer . So, it is clear that environments containing dissolved oxygen have a higher ORP than those containing hydrogen at all pH values. Furthermore, while the equilibrium potentials decrease by about 59 mV for each increase of 1 pH o unit at 25 °C (77 F), the ORP may change at quite different rates, depending upon the pH dependencies of the exchange current densities for the various redox reactions in the system, as shown in Equation B-6. The ORP also decreases as the partial pressure of oxygen decreases and the partial pressure of hydrogen increases, although not necessarily as indicated by the equilibrium Equations B-24 and B-25, because the ORP is, fundamentally, a kinetic property of the system.

B-16


Nevertheless, the ORP may be used to control or monitor the oxidizing power of fossil plant feedwater on oxygenated treatment and to control flow-accelerated corrosion, or, most importantly, to assure the desired reducing environment in the feedwater of mixed-metallurgy systems, which can only be accomplished by keeping the oxygen level low (<10ppb) and adding a reducing agent.. Designation of a feedwater system being “oxidizing” or “reducing” is somewhat subjective, because the use of those terms requires careful definition of the metallurgy in the circuit and in particular the susceptibility of the metallurgy to various forms of corrosion. Thus, for a feedwater system that contains copper alloys, the recommended practice (1) is to operate under reducing conditions by adjusting the water chemistry to yield a redox potential of –300 mV [Ag/AgCl/KCl(sat)] or –101 mV (SHE), or lower. (Note that the conversion from the Ag/AgCl/KCl(sat) scale to the SHE scale at ambient temperature is affected by adding 199 mV to the former – also see Table B-3). This is normally done, by adding a reducing agent, such as hydrazine, to the feedwater, which contains a low level of oxygen (<10 ppb). Assuming that hydrazine acts electrochemically like hydrogen, the molar ratio required to achieve this condition -3 is calculated to be about 10 , based on the hydrogen equivalence of hydrazine of 2 (N2H4 N2 + 2H2) (Figure B-3). In this figure, the calculated redox potential is plotted versus the molar ratio of oxygen to hydrogen for four different hydrogen concentrations. The redox potential describes a sigmoid variation with the molar ratio, and is predicted to be insensitive to the hydrogen concentration except under reducing conditions, where the hydrogen electrode reaction dominates the redox behavior. One important feature displayed by these calculations is the high sensitivity of the redox potential to small amounts of oxygen in the system. Thus, noting that the equilibrium potential for the hydrogen electrode reaction is of the order of –0.39 Vshe, the addition of 0.0001 mole of oxygen per mole of hydrogen at very low molar ratios has the effect of shifting the redox potential in the positive direction by 200 mV. This feature has been observed (8) experimentally in simulated BWR environments .

B-17


0.4 0.3

2ppb H2

Redox Potential Vshe

5ppb H2 0.2

10ppb H2 20ppb H2

0.1 0.0 -0.1

T = 25 oC d = 10 cm v = 5 cm/s

-0.2 -0.3 -0.4 1e-5

1e-4

1e-3

1e-2

1e-1

1e+0

1e+1

1e+2

Molar Ratio (mO2/mH2) Figure B-3 Plot of calculated redox potential versus molar ratio of O2:H2 for hydrogen concentrations of 2, 5, 10, and 20 ppb H2. The arrow indicates a “redox neutral” system having a molar ratio of 0.5

In any event, achieving a redox potential as low as –0.3 Vshe requires very low oxygen levels (Figure B-3), and the use of a reducing agent. On the other hand, oxidizing conditions are now (17) recommended for all-ferrous feedwater systems , primarily to produce a dissolution-resistant Fe(III) oxide (e.g., γ-Fe2O3) in the outer layer of the passive film. This requires the redox potential, as measured at ambient temperature, to be 0.2 Vshe or greater (Figure B-3). At lower potentials, the steel is covered by a magnetite (Fe3O4) barrier layer and a magnetite outer layer, which is susceptible to both oxidative and reductive dissolution, depending upon the exact value of the redox potential. It should be noted that it is also possible to define “oxidizing” and “reducing” conditions in terms of the molar ratio of the oxidizing and reducing species in the coolant, as shown in Figure B-3. A “redox neutral” system has a molar ratio of 0.5, as indicated by the arrow and corresponding to the complete consumption of both species should reaction occur between them (2H2 + O2 2H2O). Because of the differences in molecular weight between oxygen and hydrogen, the weight-based (ppb) concentration ratio (ppbO2:ppbH2) is 8.

B-18


Finally, it is important to note that the practical classification of systems being “oxidizing” or “reducing” for a given molar ratio is sensitive to temperature, because of the effects of temperature on the kinetics of the redox reactions. Accordingly, great care must be exercised when using redox potentials measured at ambient temperature to infer the state of the system at temperatures that are several hundred degrees higher. It is for this reason that redox and corrosion potentials need to be obtained for the high temperature regions of interest, either by direct measurement or by correcting the ambient temperature values to elevated temperatures by (8) using a mixed potential model of the type that was employed for constructing the figures in this paper and of the kind that has been used extensively for calculating ECP in BWR coolant circuits (6-8, 12) . Returning now to Figure B-3 and noting that to convert a potential reading from the SHE scale to the Ag/AgCl, KCl(sat) scale it is necessary to subtract 200 mV; accordingly, the redox neutral potential (RNP) defined above is about 0 V [Ag/AgCl, KCl(sat)]. However, in this region, the ORP is very sensitive to small changes in the balance between reducing and oxidizing species and the potential may swing one way or the other, depending upon the monitoring system design. Interpretation of Fossil Plant ORP Data The implication for fossil units may be stated as follows: If the oxygen concentration in the feedwater is low (<10 ppb) and a reducing agent (e.g., hydrazine) is added to the cycle then the ORP [Ag/AgCl, KCl(sat)] will be somewhere around –200 to –300 mV, and the reading clearly indicates that the fluid is reducing. Then, with the other extreme of no reducing agent but adding oxygen, the ORP will be around 100 mV (300 mV SHE) or higher. The feedwater is then clearly oxidizing. However, if only the reducing agent is eliminated with no additional oxygen, then the ORP will be somewhere between –100 mV and +50 mV [Ag/AgCl, KCl(sat)] depending on the oxygen and conditions as above. Under these circumstances, it is difficult to specify categorically the redox state of the system, because the coolant is poorly redox buffered and even trace amounts of redox species (including corrosion products), often below the limits of detection, can cause the ORP to deviate from the redox neutral state. Furthermore, the redox neutral potential (RNP) is known to depend upon the relative exchange current densities of the OER and the HER and it is possible that, for any given sensor, the RNP may be negative, even though the system is actually oxidizing in nature. It is for this reason that the redox sensor should be internally calibrated on known plant coolant conditions, if the opportunity arises. In principle, fitting the MPM to the plant data and calculating the potential for the appropriate mole ratio of oxidizing to reducing species will then yield the RNP. As might be expected, the ORP is similarly influenced by the presence of other redox reducing agents, such as hydrazine, carbohydrazide, and hydroquinone, and ORP measurements are often (18, 19) used to ensure that hydrazine levels are maintained at correct levels . Further details of how and why ORP and corrosion potential measurements are made in power plants can be found elsewhere (20,21).

B-19


Reference Electrode Calibration and Maintenance Procedures As previously mentioned, all ECP measurements, including corrosion potential and ORP, utilize a reference electrode that must be maintained and periodically checked for proper operation. It is common practice to check reference electrode potentials against the potentials of other nominally identical reference electrodes. All measurements should agree within one millivolt and any reference electrode not meeting this standard should be withdrawn from service. Some reference electrodes can be rejuvenated and brought back into service simply by replacing the internal electrolyte (see below), but some reference electrodes are not designed for rejuvenation and must be discarded. The general care that should be given to reference electrodes is discussed at length (15) by Ives and Janz , but their monograph is designed primarily for research scientists. Nevertheless, it is an excellent source of information on reference electrodes in general and on the Ag/AgCl, KCl(sat) system in particular. From this source and others, it is possible to formulate a set of “general rules” for the care and calibration of reference electrodes, as follows:

•

Reference electrodes should never be allowed to dry out. They should be kept immersed in saturated KCl solution at all times between applications.

•

The KCl internal solution must be maintained in the saturated state by ensuring that solid KCl is present at all times.

•

A reference electrode should never be polarized by an external power source.

•

Reference electrodes should be maintained in groups, with all electrodes being connected to a common bus and all being maintained in the same saturated KCl electrolyte. Thus, if any difference exists between two electrodes, shorting of the leads will cause the two to discharge to a common (correct) voltage.

Ideally, all reference electrodes should be calibrated periodically, because they tend to change with time. The apparatus needed to make such a calibration using a hydrogen electrode as the (22) standard is described in ASTM G5-87 . The rate that a reference electrode degrades depends partly on the type of service it sees and the contaminants it picks up from the service environment. For instance, a silver/silver chloride (Ag/AgCl) electrode, commonly used for both corrosion potential and ORP measurements at temperatures up to 300 °C (572 °F), is susceptible to irreversible damage if contaminated with sulfide. The formation of silver sulfide on the electrode surface causes a permanent change to occur that cannot be fixed by recalibration. Calibration may also be impossible if the silver chloride component has been reduced to silver by hydrogen that has diffused through the reference electrode internal electrolyte from the bulk environment. Dilution of the internal electrolyte (usually potassium chloride, KCl) with the bulk environment may also occur, especially at high temperatures (e.g. above 200 °C or 392 °F) and when there is a pressure differential between the inside and outside of the reference electrode chamber, as noted previously in this paper. Dilution of the KCl will cause the reference potential to drift from the initial value, so periodic replacement of this solution is recommended in electrode designs that allow this.

B-20


Corrosion Potential Probe If the reference electrode has been properly maintained and calibrated, nothing further can be done to calibrate the corrosion probe, because there is no “correct value” or standard corrosion potential available for calibration purposes. When reporting corrosion potentials, the reference electrode should always be cited. For instance, if a potential measurement, E, is made using a Ag/AgCl reference electrode that contains an internal 3.0 M KCl solution, the potential would be written as E volt (Ag/AgCl, 3.0 M KCl). Standard conversion tables are available that relate all reference electrode potentials to the standard hydrogen electrode (SHE) scale. Consequently, the measured potential can be converted easily by the reader to an equivalent value measured using a different reference electrode. Some conversion factors are shown in Table 3 for reference electrodes in common use (22) and for temperatures in the range 20-30°C (68-86°F) . For instance, an ECP of +500 mV measured at 20°C (68°F) versus an SCE can be converted to an ECP versus an SHE, by adding 247 mV, to yield a potential of 747 mV(SHE). Table B-3 To Convert ORP or Corrosion Potential (ECP) Values Measured Using Reference Electrode #1 to Values on Reference Electrode #2 Scale, Add the Indicated Conversion Factor to the Measured Potential. (Source: Reference 22). Add the Conversion Factor Below to Convert To Potential (mV) Versus Reference Electrode #2*

Convert From Potential (mV) Versus Reference Electrode #1* (Listed Below)

To SCE

To Ag/AgCl (3M KCl)

To Ag/AgCl (sat. KCl)

To SHE

20°C 68°F

25°C 77°F

30°C 86°F

20°C 68°F

25°C 77°F

30°C 86°F

20°C 68°F

25°C 77°F

30°C 86°F

20°C 68°F

25°C 77°F

30°C 86°F

From SCE

—

—

—

+34

+35

+36

+45

+45

+45

+247

+244

+241

From Ag/AgCl (3M KCl)

-34

-35

-36

—

—

—

+11

+10

+9

+213

+209

+205

From Ag/AgCl (sat. KCl)

-45

-45

-45

-11

-10

-9

—

—

—

+202

+199

+196

From SHE

-247

-244

-241

-213

-209

-205

-202

-199

-196

—

—

—

* Ag/AgCl = silver chloride; SCE = saturated calomel electrode; SHE = standard hydrogen electrode. Note: the presence of liquid junction potentials may result in the listed conversion factors being in error by 1 or 2 mV.

B-21


ORP Probe As noted earlier, the great majority of ORP probes used in fossil power plants use one type of reference electrode: the Ag/AgCl reference electrode. Also, fortunately, the internal electrolyte in the reference electrode, which influences the electrode potential, is typically a solution of potassium chloride having a concentration in the range 3M to saturated. An increase in the chloride concentration over this range would decrease the reference electrode potential by no more than 11 mV for a temperature in the range 20°C to 30°C (68 to 86 °F); see Table B-3. Since an error of 11 mV would be considered small for most practical ORP measurements, knowledge of the precise KCl concentration is usually not an issue at near-ambient temperatures. Corrections for KCl concentration may be necessary if the ORP is measured at higher temperatures. Nevertheless, confusion has arisen when ORP values measured versus a Ag/AgCl reference electrode are inadvertently compared with others measured versus an SCE or converted to the SHE. (Note that there is no such thing, physically, as a “Standard Hydrogen Electrode”. Instead, the SHE scale is simply a hypothetical scale that was developed in thermodynamics to permit various electrodes to be compared against one another). Consequently, it is essential that the reference electrode used for the measurement is always specified along with the measured ORP values [e.g., ORP = 100 mV (Ag/AgCl, KCl(sat))]. Again, when comparing ORP measurements, they should be referred to the same reference electrode scale. As described above, the conversion factors shown in Table B-3 can be used to convert from one reference electrode scale to another. The instrument used to measure ORP may be calibrated using standard solutions that have known ORP values. For instance, one possible set of standard ORP reference solutions are based on pH 4 and pH 7 buffer solutions, similar to those used for pH meter calibration. However, for ORP standards, quinone and hydroquinone of specified concentrations are added to the buffer solutions, giving rise to a reversible redox couple. Hydrogen ions participate in the reaction between the quinone and hydroquinone, creating a pH dependent equilibrium: + C6H4O2 + 2H + 2e ↔ C6H6O2

Quinone (Q)

(eq. B-26)

Hydroquinone (HQ)

with

E eQ / HQ = E oQ / HQ - [2.303RT /(2F)] × log(aHQ/aQ) - [2.303RT / F] × pH (eq. B-27) When the activity of quinone (aQ) is equal to the activity of hydroquinone (aHQ), the second term on the right hand side of Equation B-27 drops out, so that the electrode potential, E, depends only on pH. In practice, it is common to prepare the ORP reference solutions with “quinhydrone”, which is formulated to provide an equimolar solution of quinone and hydroquinone.

B-22


The quinhydrone solutions are prepared by dissolving 10 grams of quinhydrone in one liter of pH 4 or pH 7 buffer solution (24). Quinhydrone is not very soluble, so only a small amount will dissolve in the buffer solution, changing it to an amber color. However, it is important that excess quinhydrone be used, so that solid crystals are always present. The quinhydrone powder poses a moderate health risk, causing irritation of the lungs with prolonged exposure to the dust. The calibration solutions are fairly innocuous unless ingested in large amounts. For the pH 4 and 7 quinone solutions, ORP values of about 263 mV and 86 mV, respectively, o can be expected at 25°C (77 F) when measured against a Ag/AgCl, KCl(sat) reference electrode (21, 24, 25) . These values are consistent with a change of 59.0 mV per pH unit, which in turn is consistent with the calculated value of 2.303 RT/F (= 59.2 mV at 25 °C or 77 °F) in Equation B27. ORP values for the standard solutions at other temperatures and for other reference (24) electrodes are provided in ASTM D1498-93 and are reproduced in Table B-4. Table B-4 Expected ORP Values for Reference Quinone Solutions at pH 4 and pH 7 (Source: Reference 23). Reference Electrode*

ORP Value (mV) pH 4 Buffer Solution

pH 7 Buffer Solution

20°C (68°F)

25°C (77°F)

30°C (86°F)

20°C (68°F)

25°C (77°F)

30°C (86°F)

Ag/AgCl

268

263

258

92

86

79

SCE

223

218

213

47

41

34

SHE

470

462

454

295

285

275

*Ag/AgCl = silver/silver chloride (sat. KCl) SCE = saturated calomel electrode SHE = standard hydrogen electrode.

The quinhydrone standards are easily made, but they are stable for only about 8 hours and are not expected to yield highly reproducible ORP values at longer times. Nevertheless, ORP values in freshly prepared solutions are expected to be within 10 mV of the values listed in Table B-4, and an ORP value measured with a probe in one of these standards can usually be considered acceptable if it lies within 30 mV of the listed value. If the electrode does not respond as expected or has a slow response, it should be cleaned and the calibration procedure repeated. Removal of oily or organic deposits can be achieved by washing with a detergent or, if necessary, methanol or isopropyl alcohol. For removal of inorganic contaminants or mineral deposits, the electrode should be soaked in 10% nitric acid for 10 minutes. Alternatively, ASTM (24) D1498-93 suggests the use of warm (70 °C or 158 °F) aqua regia for about one minute . As a last resort, the platinum surface can be polished with a 600 grit wet-dry emery cloth or a 1-3 micron alumina polishing powder to remove any stubborn coatings or particulates. The electrode will need to soak in tap water or purified water for at least 30 minutes after performing any of these cleaning procedures.

B-23


A single point calibration (or “standardization”) like this is often considered adequate, because ORP meters, unlike pH meters, do not (yet) have “slope calibration” to allow adjustment of the response to a specified change in ORP. Nevertheless, it is usually good practice to verify that the ORP probe (platinum/reference electrode combination) is operating in a predictable fashion by measuring the ORP in a second standard solution. If the electrode potential does not change by the expected amount, the electrode should be cleaned and re-calibrated, as described above. Other ORP calibration solutions are available based on the ferrous/ferric cyanide [ Fe(CN ) 36− / Fe(CN ) 64− ] or ferrous/ferric sulfate equilibrium reactions, but the quinhydrone standards are more widely used. Also, it has been suggested that a saturated sodium sulfite (26) solution could be used to verify instrument response to negative ORP readings .

B.7 CLOSING COMMENTS ON ORP MONITORING IN FOSSIL PLANTS ORP is identical to the redox potential that is used extensively in electrochemistry and corrosion science. ORP should not be confused with corrosion potential, although the ORP value does provide a useful indicator of the corrosivity of the environment. The ORP reflects the oxidizing/reducing power of the environment, while the corrosion potential reflects the corrosion reaction occurring on the metal surface. The corrosion potential generally reflects the ORP but they do not have the same value. Thus ORP can be an excellent tool to control fossil plant feedwater chemistry, corrosion and flow-accelerated corrosion. However, care has to be taken in interpreting the values measured as ORP changes with pH, partial pressure of oxygen in the solution, mass transport properties, flow rates and the materials in the cycle. Thus measured ORP values cannot be directly compared from unit to unit, and similar units containing similar amounts of oxygen and/or reducing agent will not necessarily read the same ORP, because of differences in measuring parameters. Also, the voltage at which conditions change from “reducing” to “oxidizing” (redox neutral potential) will not always be the same. Thus ORP should be used to control the oxidizing power of the feedwater in all-ferrous systems, or, and more importantly, the reducing power of the feedwater in mixed-metallurgy feedwater systems. The future direction should be to develop the technology to measure ORP in-situ in feedwater, and to extend the mixed-potential model for use at the elevated temperatures, where components are actually in contact with high temperature water, such as in the boiler.

B-24


B.8 REFERENCES 1. Guidelines for Copper in Fossil Plants, EPRI, Palo Alto, CA: 2000. 1000457. 2. Cycle Chemistry Guidelines for Fossil Plants: All-Volatile Treatment: Revision 1, EPRI, Palo Alto, CA: 2002. 1004187. 3. Copper Alloy Corrosion in High Purity Feedwater, EPRI, Palo Alto, CA: 2002. 1000456. 4. Influence of Water Chemistry on Copper Alloy Corrosion in High Purity Feedwater, EPRI, Palo Alto, CA: 2002. 1004586. 5. B. Dooley, D. Macdonald, and B. C. Syrett, “ORP – The Real Story for Fossil Plants”, PowerPlant Chemistry, vol. 5, No. 1, pp. 5-15 (Jan. 2003). 6. D. D. Macdonald, I. Balachov, and G. Engelhardt, PowerPlant Chemistry, 1(1), 9 (1999). 7. R. B. Dooley and D. E. Hubbard, Latest Developments in Fossil Plant Cycle Chemistry. Power Plant Chemistry, 2000, 2(7), pp 389-395. 8. D. D. Macdonald, A. C. Scott, and P. Wentrcek. “Redox Potential Measurements in High Temperature Aqueous Systems”. J. Electrochem., Soc., 128, 250-257 (1981). 9.

D. D. Macdonald and L. B. Kriksunov, “Flow Rate Dependence of Localized Corrosion in Thermal Power Plant Materials”. Adv. Electrochem. Sci. Eng., 5, 125-193 (1998).

10. J. R. Selman and C. W. Tobias, Adv. Chem. Eng., 10, 211 (1978). 11. T. Tsuruta and D. D. Macdonald. “Measurement of pH and Redox Potential in Boric Acid/Lithium Hydroxide Buffer Solutions at Elevated Temperatures”. J. Electrochem. Soc., 128, 1199-1203 (1981). 12. D. D. Macdonald, “Viability of Hydrogen Water Chemistry for Protecting In-Vessel Components of Boiling Water Reactors”. Corrosion, 48(3), 194-205 (1992). 13. D. D. Macdonald, A. C. Scott, and P. R. Wentrcek. “External Reference Electrodes for Use in High Temperature Aqueous Systems”. J. Electrochem. Soc., 126, 908-911 (1979). 14. D. D. Macdonald, “Reference Electrodes for High Temperature Aqueous Systems - A Review and Assessment”. Corrosion, 34(3), 75-84 (1978). 15. D. J. G. Ives and G. J. Janz, Reference Electrodes, Theory and Practice”, Academic Press, 1961. 16. J. P. Hoare, “The Electrochemistry of Oxygen”, Interscience, NY, 1968. 17. R. B. Dooley, The Relationship between Cycle Chemistry and Performance of Fossil Plants. Power Plant Chemistry, 2002, 4(6), pp 320-327.

B-25


18. S.J. Shulder. “Using ORP Measurements to Control Hydrazine Feed.” Presented at the 1999 Southwest Chemistry Workshop, Steamboat Springs, Colorado, August 11-13, 1999. 19. S.J. Shulder, M.A. Janick, E.C. Gwin and S.T. Filer. “Measuring Oxidation-Reduction Potential (ORP) and Its Use In Controlling Oxygen Scavenger Injection.” EPRI. Presented at the Sixth International Conference on Cycle Chemistry in Fossil Plants, Columbus, Ohio, June 27-29, 2000. 20. S. Filer, “Power Plant Chemistry Measurement Advancements: Oxidation Reduction Potential”, Australia Power Station Chemistry Conference, New Castle, New South Wales, Australia (March 1998). 21. S. Filer, A.S. Tenney III, D. Murray and S.J. Shulder, “Power Plant ORP Measurements in High Purity Water”, NUS International Chemistry On-Line Process Instrumentation Seminar, Clearwater Beach, FL (November 1997). 22. ASTM G5-87, “Standard Reference Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements”, ASTM, Philadelphia, PA (1987). 23. B.C. Syrett and W.C. Micheletti, Reference Manual for On-Line Monitoring of Water Chemistry and Corrosion: 1998 Update. EPRI, Palo Alto, California. TR-112024, April 1999. 24. ASTM D1498-93, “Standard Practice for Oxidation-Reduction Potential of Water”, ASTM, Philadelphia, PA (1993). 25. L. McPherson, Chemical Engineering, p. 143-145, March 1994. 26. Personal Communication with S.J. Shulder, Constellation Power Source Generation Inc, August 2000.

B-26


C AIR IN-LEAKAGE MONITORING AND CONTROL

C.1 CYCLE AIR IN-LEAKAGE A recent EPRI publication, Condenser In-Leakage Guideline(6), gives comprehensive details on how to conduct a monitoring program associated with air in-leakage using tracer gases. This document would be a valuable reference for those contemplating the purchase and use of this technology. Program Copper research has clearly indicated that higher levels of air in-leakage into the cycle, which cause the ORP to change from reducing to oxidizing, will change the surface oxide layers from Cu2O to non-protective CuO. Allowing the air in-leakage to continue for more than 8 hours will lead to increasing conversion to CuO and increased levels of copper in the feedwater. Constant vigilance must be exercised to prevent air in-leakage to the cycle from those equipment elements under vacuum at any or at all times. Particular problem areas include pump seals, LP turbine shaft seals, valve bonnets, threaded joints, and especially the expansion joint between the turbine and condenser. Monitoring the amount of air in-leakage (flow to the vacuum pumps) is essential for proper control of oxygen / carbon dioxide. Such monitoring will determine when an exhaustive effort must be made to find and fix the source of air leakage. The Heat Exchange Institute recommends that air in-leakage be restricted to no more than 1.0 scfm per 100 MW of generating capacity. This same value has been included in all EPRI Guidelines to date.

C.2 ROTAMETERS Excess air in-leakage not only affects iron and copper corrosion and subsequent deposition, but can also increase turbine backpressure, which directly reduces electrical generating capacity. A complete analysis of the effects of air in-leakage on condenser efficiency was made (2, 3) using a “multisensor” probe. This multisensor probe and its use in measuring air in-leakage is further described later in this Section. Typically, utilities employ a rotameter at the exhaust of the steam jet air ejector or mechanical vacuum pump to quantify air in-leakage. This was the standard approach at the time plants were designed and constructed. Manual readings were planned to be taken, normally once a day, however, current manpower often restricts this. Air in-leakage can increase and remain at potentially damaging levels for hours or possibly days before the excess in-leakage is discovered, since turbine backpressure may not have increased significantly or noticeably.

C-1

EPRI Licensed Material Air In-Leakage Monitoring and Control

Under certain conditions, rotameters provide adequate flow rate measurements of air in-leakage. However, many factors can affect the accuracy of the readings so obtained, such as debris, exhauster pulsations, water carryover, etc. Rotameters also have limited range, generally 10:1. Therefore, most problematic air in-leakages cause the rotameter to be pegged against the high level stop. Rotameters receive essentially very little maintenance and are checked very infrequently at many plants. In these situations, the operator has no idea of the extent of the air in-leakage, given the fact that the turbine backpressure rise may appear to be rather small and insignificant. In an attempt to mitigate these inherent problems with rotameters, a series of instruments were developed which featured continuous readouts of air in-leakage. These earlier developments had some serious drawbacks, such as not being able to distinguish between water vapor and air; and, operating under the principle that the condenser exhaust gases were always at saturation, when in fact, they were not.

C.3 MULTISENSOR PROBE Subsequent developments have produced instrumentation, which apparently solved these earlier limitations and is available commercially at the present time. One such instrument features a multisensor probe (MSP) which is installed in the exhaust line from the condenser. This device is depicted in Figure C-1. This instrument has the ability to directly measure the mass properties of gas in the condenser vacuum line, as well as computed parameters based on thermodynamics. The instrument combines a gas analyzer with a total mass flow meter. The instrument utilizes a velocitymeasuring device, and other facilities for measuring temperature, pressure, and relative saturation. By including the measurement of pipe diameter, the instrument can separate the effects of non-condensable gas from water vapor and to compute individual and combined properties of the gas mixture. As a result, the following parameters of the gas flowing in the exhauster suction line are either directly measured or are determinable:

•

Air in-leakage

•

Water vapor density

•

Volumetric flow

•

Air density

•

Total mass flow

•

Partial pressure of water

•

Water vapor mass flow

•

Partial pressure of air

•

Water vapor/air mass ratio (relative saturation)

•

Total pressure

•

Temperature

C-2


Pressure port

Temp. reference

RS sensor port Flow sensor

Figure C-1 Multisensor Probe (courtesy of Intek, Inc)

Air in-leakage, as measured by the MSP(2) is the sum of all leak sources entering the subatmospheric pressure system and subsequently flowing through the exhauster line upstream of the sensor location. The MSP, however, does not measure air entering the low-pressure system beyond that (downstream), such as at valves, input connections, and shaft seals. The MSP does measure the exhauster capacity, which is useful to gauge the effectiveness of the exhausters to remove the gas mixture from the condenser.

C-3


Evaluation and analysis of condenser system performance and measurement of in-leakage is best accomplished with the use of multiple MSP probes such as shown in Figure C-2. Each probe is identified by a number that is used to categorize performance information and to help pinpoint possible leak locations since air in-leakage is removed primarily through the air removal section closest to the leak. As illustrated in Table C-1 and Table C-2, a self validating arrangement can be established that will allow the accuracy of all probe locations to be identified.

3

2

4

1

5

Condenser Figure C-2 Multisensor Probe Instrument Schematic

C-4

Exhauster 2

Exhauster 1

EPRI Licensed Material Air In-Leakage Monitoring and Control Table C-1 Examples of a Five-Probe Air In-leakage Measurement System Leak Location

Probe Indications of Air In-Leakage (SCFM) and Plant Dissolved Oxygen (DO)

Below water line, left side of condenser

Much higher than normal DO, 1>2 1 + 2 = 3 = 4 + 5, 4=5

Above water line, right side of condenser

Slightly higher or normal DO, 2 > 1, 1 + 2 = 3 = 4 + 5, 4=5

Small leak or faulty exhauster down stream of probe 4

Slightly higher or normal DO, 1 = 2, 1 + 2 = 3 = 4 + 5, 4<5

Large leak of failed exhauster down stream of probe 4

Slightly higher or normal DO, 1 = 2, 1 + 2 = 3 < (4 + 5), 5 > 4 (back flow at 4)

Center joint seal, LP bearing seal, or other central location

Slightly higher or normal DO, 1 = 2, 1+2=3=4+5

Table C-2 MSP Probe Indications for Various Probe Positions Leak Location

Probe Indications of Air In-Leakage (SCFM)

Normal tight system

1 = 2.5 SCFM 2 = 2.5 SCFM 3 = 5 SCFM 4 = 0 SCFM (exhauster 2 not in service) 5 = 5 SCFM

Abnormal, need to locate and fix leaks (central joint seal)

1 = 15 SCFM 2 = 18 SCFM 3 = 33 SCFM 4 = 16.5 SCFM 5 = 16.5 SCFM

Abnormal, need to locate leaks and fix Nash pump shaft seal in exhauster 2

1 = 20 SCFM 2 = 20 SCFM 3 = 40 SCFM 4 = 20 SCFM (Differences due to reversed flow 5 = 60 SCFM through shaft seal being sensed)

C-5


C.4 HELIUM AND SULFUR HEXAFLOURIDE METHODS The use of the MSP is certainly helpful in determining the extent of air in-leakage and providing guidance relative to where, generally, leaks may be occurring. After this assessment, it will be necessary to further determine the exact location of the leak(s). The two methods typically used for this purpose are helium leak detection and sulfur hexafluoride leak detection. Both use a sophisticated analyzer in conjunction with these tracer gases. Air in-leakage testing is done by drawing a sample of condenser off-gas from the exhauster discharge while releasing the tracer gas (helium or sulfur hexafluoride) at those areas within the vacuum boundary where leaks may be expected to occur. Since the condenser is under vacuum, the tracer gas migrates into the condenser shell, and from there it is removed by the air-removal system. A leak is identified when an off-gas sample containing the tracer gas passes through the analyzing equipment and is detected. Helium was the first tracer gas to be successfully utilized in determining air leaks. The testing is accomplished by injection of helium at various locations (within the vacuum boundaries) and detection of helium in the condenser off-gases by use of helium mass spectrometry. This system is still in use today for detecting air in-leakage. There are some limitations to the use of this technique, however, particularly as regards sensitivity of response, which is particularly important in the detection and evaluation of very small leaks. The fundamental advantage of sulfur hexafluoride (SF6) is that it can be detected in very low concentrations, as low as 0.1 ppb. By contrast, the lowest detectable concentration of helium is 1 ppm. SF6 analysis is about 10-100 times more sensitive than helium detection. The sensor technology for SF6 is based on the principle of electron capture(4). Figure C-3 is a general view of the SF6 analyzer while Figure C-4 is a schematic diagram showing how the offgas sample passes through the analyzer. To remove oxygen from the gas sample, hydrogen gas is introduced. This mixture enters a catalytic reactor in which the hydrogen and oxygen react to form water.

C-6


Moisture outlet

Sample flowmeter

Hydrogen metering valve

Pump

Catalyst assembly

Hydrogen flowmeter

Water trap

Dessicant dryer

Sample inlet

Exhaust flowmeter

Solenoid valve

Air metering valve

7mm filter

Exhaust

Electron capture cell

Pressure regulator

Hydrogen supply

Figure C-3 Flow Schematic for SF6 Analyzer System

C-7


Sample inlet

Dessicant tower

Copper cooling coil

Water trap Air mover To exhaust system

SF6 analyzer

Water discharge Strip chart recorder Figure C-4 Schematic Diagram of SF6 Sampling System

C-8

Exhaust

Hydrogen tank


SF6 can generally be used wherever helium can be used, although the converse may not always be true and there are situations in which use of helium is preferred. The following guidelines may be applied:

•

If unit air in-leakage is in excess of 10 CFM, either gas may be used. If less, SF6 is preferred.

•

If the unit is running at greater than 20% turbine power, either gas may be used. If the unit has no turbine power and the leak is so bad that the unit can not be brought up to any turbine power level, standard procedure would dictate the use of helium.

•

The search for the cause of air leakage below the steam space requires the use of SF6 as the standard procedure.

•

Inspections of units of less than 50MW capacity should always use helium.

The determination of air in-leakage using tracer gases requires a great deal of experience and a considerable outlay of capital if the necessary equipment is to be purchased. There are, however, consulting firms specializing in the evaluation of air in-leakage. These firms have their own equipment and are well versed in this technology. A typical study is estimated to cost between (5) $3500 and $5000 in today’s market .

C.5 UTILITY AIR IN-LEAKAGE PROGRAM One large utility has developed a formal program to control air in-leakage at each of their power plants (1). The program starts with a comprehensive written procedure which is continuously updated as new information is forthcoming and new experiences noted. The program and its revisions are controlled by central staff and implemented at each plant by the plant’s performance engineering group. Formal training is given to each participant in the program. The program is based on a daily monitoring and recording of air in-leakage at the vacuum pump or steam jet air ejector exhaust. The instrument presently in use for this purpose is a thermal gas mass flow meter, costing in the range of $2200 to 3200 each. Air in-leakage at the utilities’ various plants is normally in the range of 2–10 scfm/ 100MW, depending upon unit size, configuration, etc. In addition, close attention is paid to unit backpressure, an increase of which could indicate increased air in-leakage. Continuous readings of dissolved oxygen at the condensate pump discharge are carefully reviewed as another indication of increased air inleakage. (6,8) (Other plants utilize the MSP System which incorporate a thermal gas mass flow meter and other sensors as previously described. Its reported advantages over any exhauster discharge measurements are: freedom from inadvertent high output spikes or oscillations; measurement of exhauster capacity degradation caused by shaft seal leaks, wear or regulation; measurement of the water vapor to air mass ratio which triggers the onset of increased condenser pressure caused by air in-leakage and finally, high resolution measurement of air in-leakage and other noncondensables being removed from the condenser.)

C-9


At the first indication of increased air in-leakage (scfm readings above the established base, increased backpressure, higher dissolved oxygen), a leakage test is initiated. An important part of the program is the establishment of a checklist, which delineates all of the various sampling points and their location at each plant elevation. For example, the turbine room sample points are broken down by directions- north, south, east and west relative to the turbine. The equipment used for air in-leakage detection is a helium mass spectrometer. The detector is located at the vacuum pump or steam ejector exhaust. A unique feature of the program is the use of a wireless video camera, which constantly focuses on the readings of the air in-leakage detection unit. The results (leakage or no leakage) are transmitted to the operator via a remote wireless receiver. Thus, only one person is required to conduct a test, minimizing manpower. The detection instrument has an internal vacuum pump and requires a dry sample for proper operation. The inlet sample tubing is passed through a cooling coil embedded in ice, followed by two water traps in series prior to entering the instrument. Tests start at the top of the unit, working down, since helium, being lighter than air, will rise. If testing started at the bottom of the unit, the rising helium would be sucked into leaks further up, severely hampering interpretation of results. Most tests are conducted at medium, or preferably, at full load, to maximize condenser vacuum and, subsequently, potential air in-leakage. At times, however, it is necessary to test at low loads to determine leakage at those areas of the low end of the cycle that are under pressure at high loads but under vacuum at low loads. An example would be the low pressure feedwater heaters. Some examples of sampling points which are emphasized in the test procedure are condenser and condensate pump expansion joints, various flanges and pipe fittings, turbine horizontal shell flanges, turbine rupture diaphragms, turbine shaft seals, various turbine penetrations, pipe plugs, valve packings, etc. Helium is sprayed at each test area at low pressure, about 10 psi, for only about 2-3 seconds to prevent too much helium from escaping, which could result in false readings and misinterpretations. It is necessary to note any extraneous air currents at a particular sampling location to prevent helium from “spreading out” and causing false leak indications. Generally, to save time, a group of adjacent test points are sprayed with helium at the same time. If there are no leaks found, the next group of test points is sprayed. However, if leaks are recorded, further tests are required to isolate the offending location. A record of leak location is made together with the relative intensity of the leak. Another practice, which is followed on a periodic basis, about once per year, is to flood the condenser with water to above the tube sheet and to look for water leaks from possible cracks in the condenser shell, attached piping welds or valve packings. Before performing this operation it will be necessary to determine explicitly whether the structural supports provided are sufficient to hold the extra weight of water within the condenser.

C-10


Results of all testing are submitted to the maintenance staff at the plant for necessary repairs, replacements, etc. This program has been very effective in reducing oxygen and carbon dioxide from entering the cycle. Corrosion and corrosion transport have been reduced substantially since the program has been initiated. Another utility experience highlighting the economic value of finding and repairing air leaks was (7) reported at a recent EPRI Conference . Recent developments in understanding the shell side dynamics of condensers(8) have shown that condenser defects cause air blanketing, affecting the condenser’s effective heat transfer coefficient. These regions are subcooled and have a significant concentration of noncondensable gases with high partial pressure. Therefore, they can contribute greatly to dissolved gases and corrosion even though air in-leakage is low.

C-11


C.6 REFERENCES 1. D. W. Reynolds, Dynegy Midwest Generation, Personal Communication to A. Aschoff, April 2002 2. J. W. Harpster, Intek Inc., “On Understanding Mechanisms that Control Dissolved Oxygen in Condenser Condensate.” Proceeding of the 21st Annual Electric Utility Workshop, University of Illinois, Champaign, IL, May 8-10, 2001. 3. W. H. Stroman and J. W. Harpster “Continuous Monitoring for Condenser Air In-Leakage” Presented at the 2002 Spring Meeting, ASME Research Committee on Power Plant and Environmental Chemistry, Charleston, SC, March 11-13, 2002 4. La Porte and R. E. Putman, “ The Practical Application of SF6 and Helium for Condenser Tube and Air In-leakage Detection”. Proceedings of the 19th Annual Electric Utility Chemistry Workshop, University of Illinois, May 11-13, 1999. 5. La Porte, Conco Services Corp., Personal Communication to A. Aschoff, April 2000. 6. Condenser In-Leakage Guideline. Electric Power Research Institute, Palo Alto,CA.: TR-112819, 2000. 7. W.A. Kaufman, “Experiences with Steam Cycle Air Ingress at Alliant Energy Fossil Plants: Case Studies Documenting the Effects of Air In-Leakage on Steam Plant Operations”, presented at the EPRI Condenser Technology Seminar, EPRI, Palo Alto, CA: September 2002. 1004116. 8. J.W. Harpster, “An Overview of a Comprehensive Study of Condenser Shell Side Dynamics”, EPRI Seminar, San Antonio, TX, Sept. 10–12, 2002.

C-12


D BENCHMARKING A UTILITY’S CHEMISTRY ORGANIZATION

D.1 INTRODUCTION Utility chemists often ask how good or bad they are in terms of their cycle chemistry operation and organization, and where do they rank with other utilities for a utility of their size. If such questions were answerable then they would be able to use the information as the first step in justifying new cycle chemistry equipment, or improvements to their current ability. Insufficient quantitative data exist to rank utilities in terms of chemically influenced availability loss, reliability and performance degradation. For example, the NERC/GADS database has not continued to collect boiler tube failure (BTF) mechanistic data since the finish of EPRI’s Boiler Tube Failure Reduction Program (BTFRP) in 1991. An initial attempt was made with EPRI’s Cycle Chemistry Improvement Program (CCIP) to measure a utility’s improving cycle chemistry by monitoring a series of direct and indirect cost indices. This was very successfully used by the CCIP participating utilities. Currently most utilities require short payback periods with associated accurate savings to justify cycle chemistry equipment and improvements. Much thought has been given to this topic and the current assessment methodology will provide an initial assessment for rating a utility’s chemistry organization. Participation and survey by a large number of utilities over the next year will solidify the rating. It could also be used by individual utilities to monitor improvements.

D.2 ASSESSING THE CYCLE CHEMISTRY ORGANIZATION OF A UTILITY The attached form is a self-assessment. It involves a series of “results oriented”, “cycle chemistry influenced” factors. Each of the five factors is capable of being assessed quantitatively, and, as a whole, represent the key performance and availability indicators (results) by which the chemistry organization should be judged. Each factor relates to the Goals of the EPRI Cycle Chemistry Program. It is suggested that a utility make the initial assessment with data from the last five years. Improvements/changes could then be assessed on an annual basis. The methodology will work best when applied to a single unit/plant, but can be also applied to a utility system.

D-1

EPRI Licensed Material Benchmarking a Utility’s Chemistry Organization

A. Chemically influenced Boiler Tube Failures (BTF) What percentage of the total numbers of BTF over the last five years, that have resulted in forced outage or unit unavailability, have been influenced by the cycle chemistry. A supplementary listing of those BTF influenced by the chemistry is provided for water-touched and steamtouched tubing. For further information on BTF Mechanisms, see “Boiler Tube Failures: Theory and Practice”. B. Dooley and W. McNaughton. TR-105261. B. Chemically influenced turbine problems This is in two parts. The first questions whether any LP blade/disk cracking has occurred in the last five years: this would involve either failure or identification by NDE during overhaul that cracking has occurred in the common corrosion fatigue or stress corrosion cracking areas. Has pitting been observed on LP blades The second part involves whether copper deposition has been experienced in the HP turbine or whether excessive corrosive deposits in the LP turbine have been observed, which have required shot blasting or rinsing. For further information on chemically influenced turbine mechanisms, see “Steam Turbine Damage: Theory and Practice”. T. McCloskey, B. Dooley and W. McNaughton. TR-108943, 1999. C. Cycle Chemistry Instrumentation Control This has two parts. The first involves a comparison of plant instrumentation with EPRI’s core level of instruments which should be installed in every plant. A supplementary list is provided, which indicates 10 key instruments/parameters (or 11 if a polisher is installed) for drum boilers, and 9 for once-through units. The second part involves whether these key control parameters are alarmed in the control room. A supplementary question might be whether the operators know how to react to an alarm, but this was not included as the answer would be subjective and difficult to quantify. D. Chemical Cleaning This is in two parts relating to the two types of feedwater metallurgy system possible. The need to chemically clean boilers and the frequency are key indicators of whether the feedwater and boiler water treatments are optimized. E. High Purity Water Preparation and Use This factor deals with the production of makeup water and its use. Because the cost of makeup production varies with supply (the range of cost is not known) and geography, an indicator based on the actual cost could not be used, although the question is asked. The addition of makeup to the cycle is dependent on the type of boiler (once-through or drum). The makeup rate is a good indicator of whether chemical control is optimized, particularly for drum boilers.

D-2


Assessment of Cycle Chemistry Organization in a Utility

•

Size of thermal plant capacity ❏ ❏ ❏ ❏

> 30,000 MW 20-30,000 MW 10-20,000 MW < 10,000 MW Factor

Weighting 3

A.

Points

Total

Chemically influenced BTF (What is percentage of total number of BTF that is cycle chemistry influenced. See attached for relevant BTF) ❏ ❏ ❏ ❏ ❏

0% < 5% 5-15% 15-30% > 30%

0 1 2 3 4

Sub-total (Points x Weighting) 3

B.

Chemical influenced turbine problems ❏ ❏

LP blade/disk cracking, or pitting No LP blade/disk cracking, or pitting

2 0


❏

❏

Copper deposition in HP reducing capacity or efficiency, or excessive deposits in LP turbine Neither HP or LP deposition

2 0

Sub-total (Points x Weighting)

D-3


Weighting 2

Factor C.

Points

Total

Cycle Chemistry Instrumentation and Control What is percentage of EPRI core level of instruments (see attached for core level) ❏ ❏ ❏ ❏

100 % 90-99 % 70-89 % < 70%

0 1 2 3


Do you have these key control parameters alarmed in control room ❏ ❏ ❏

Yes. All alarmed Yes. Some alarmed No. None alarmed

0 1 2

Sub-total (Points x Weighting) D. 1

Chemical Cleaning Frequency of cleaning boilers with mixed-metallurgy feedwater heaters ❏ ❏ ❏

< 5 years 5-10 years > 10 years

3 2 1


Frequency of cleaning boilers with all-ferrous feedwater heaters ❏ ❏ ❏

Sub-total (Points x Weighting) Factor

Weighting E. D-4

< 5 years 5-10 years > 10 years

High Purity Water Preparation and Use

3 2 1 Points

Total


1

Do you know the cost of makeup water ($/1,000 gals) for each plant/unit ❏ ❏

Yes (what is it? No

$/1000 gals)

0 1

Sub-total (Points x Weighting) Addition of makeup to the cycle. For drum units, is addition generally ❏ ❏ ❏

1

<0.3 % 0.3 – 0.5 % >0.5 %

0 1 2

For once-through units, is addition generally ❏ ❏

<0.2 % >0.2 %

0 1

Sub-total (Points x Weighting) Total

Rating System Excellent/World Class Very good Good Average Below Average

<5 6-10 11 –15 16–30 >30

D-5

EPRI Licensed Material Benchmarking a Utility’s Chemistry Organization Supplementary Information for Factor A BTF Mechanisms in Water-touched Tubes that are Influenced by Cycle Chemistry Mechanism

Nature of Chemistry Influence

Hydrogen damage

Excessive feedwater corrosion products form excessive deposits and combine with a source of acidic contamination.

Caustic gouging

Excessive feedwater corrosion products form deposits and combine with a source of caustic.

Acid phosphate corrosion

Excessive feedwater corrosion products form deposits and combine with a source of phosphate

Chemical cleaning damage

Excessive deposits in waterwalls lead to chemical cleaning; process errors lead to tube damage.

Corrosion fatigue

Poor water chemistry, shutdown or layup practices, and improper chemical cleaning worsen contribution of the environment to causing damage.

Supercritical waterwall cracking and overheating

Excessive internal deposits lead to increased tube metal temperatures; exacerbates mechanism.

Fireside corrosion

Excessive internal deposits lead to increased tube metal temperatures; exacerbates mechanism.

Short-term overheating

Plugging of waterwall orifices by feedwater corrosion products.

Erosion/corrosion of economizer inlet headers

Attack by reducing feedwater conditions.

Pitting (economizer)

Stagnant, oxygenated water formed during shutdown.

BTF Mechanisms in Steam-Touched Tubes that are Influenced by Cycle Chemistry Longterm overheating (creep)

If caused by restricted steam flow as a result of contaminant deposits, debris, etc.

Short-term overheating

Blockage from improper chemical cleaning (of SH/RH or waterwalls).

Stress corrosion cracking

Variety of bad environment influences, most directly related to chemistry control and practices.

Pitting (RH loops)

Carryover of Na2SO4 or poor shutdown practices allowing for oxygenated, stagnant condensate.

Chemical cleaning damage

Poor chemical cleaning practice.

For further information on BTF Mechanisms, see Boiler Tube Failures: Theory and Practice. B. Dooley and W. McNaughton. TR-105261.

D-6


Supplementary Information for Factor C “Core” Monitoring Parameters (Minimum level of instruments for all plants/units) (all instruments should be on-line and continuously monitored) Parameters

Measurement Locations

Cation Conductivity

CP Discharge

Cation Conductivity

Polisher Outlet and Economizer Inlet

Cation Conductivity1

Blowdown or Downcomer

Cation Conductivity

Hot Reheat Steam or Main Steam

Dissolved Oxygen

CP Discharge Economizer Inlet

pH (Drum Boilers)

Blowdown or Downcomer

Sodium

CP Discharge

Sodium

Polisher Outlet or Economizer Inlet

Sodium

Hot Reheat Steam or Main Steam

Note 1: Drum boilers on AVT and OT.

D-7


E CYCLE CHEMISTRY DATA QUALITY

E.1 INTRODUCTION Effective fossil plant cycle chemistry programs consist of several components. These include:

•

Properly selected feedwater and boiler water treatment approaches

•

Establishment of optimized and customized target values and action levels

•

Provision of a suite of on-line analyzers consistent with the selected treatment approaches

•

Implementation of operating practices and procedures that allow for maintaining desired chemistries under all operating conditions

•

Development of operating response procedures that enable plant personnel to promptly identify deviations from normal chemistry and to initiate appropriate corrective actions

These aspects have been addressed in detail as they apply to fossil units with drum type boilers operating on phosphate and caustic treatments. In this section, other important aspects of cycle chemistry surveillance and control are discussed, including the following.

•

Sampling and sample conditioning capabilities

•

Monitoring methods

•

Data acquisition, analysis and management practices

•

Quality assurance and control requirements, including on-line analyzer chemistry data validation practices

•

Plant staff capabilities and responsibilities

These important aspects of cycle chemistry programs are addressed in ensuing subsections. However, it should be noted that this appendix represents a summarized account of best available practices in these areas. Users of these guidelines seeking additional details are strongly encouraged to consult the source documents on which this presentation is based.

E-1

EPRI Licensed Material Cycle Chemistry Data Quality

E.2 SAMPLING, SAMPLE CONDITIONING AND INSTRUMENTATION This subsection addresses the important considerations involved with proper sampling, on-line instrumentation and analysis of grab samples. Readers should also refer to Sections 4 (phosphate continuum) and 5 (caustic treatment), which delineate the types of analyzers to be used. The following discussion is provided as a summary of these topics. A comprehensive discourse on (1-6) these and related topics can be found in other EPRI publications. E.2.1 Factors Affecting Sampling Sampling is an important aspect of water chemistry control within the power plant. A meticulously performed analysis is of no value if a bad sample is used. Obtaining a good sample involves careful consideration of the design and operation of the sampling system. Many factors affect how well a sample represents the medium from which it is drawn. The sample point must be selected so the sample is representative of the bulk solution, particularly when suspended material is present. Also, if chemicals are injected or other streams are introduced into the sampled medium, the sample collection point should be placed far enough downstream to ensure a completely mixed sample. Assuming turbulent flow, locating the sample collection point an equivalent length of 25 pipe diameters downstream of the chemical injection point is considered acceptable. An equivalent length of 50 pipe diameters is recommended for laminar flow. The material and design of the sampling apparatus also play an important role in the collection of representative samples. The materials chosen for nozzles, tubing, pumps, and valves must be corrosion-resistant and compatible with the sample conditions and the analyses performed. Austenitic stainless steels (types 304, 316, and 347) are most commonly used. In addition, it is essential to select the most effective piece of equipment for carrying out a particular task. For example, needle valves have been used for reducing pressure in sample conditioning systems. However, at high temperatures, needle valves have been found to cause the thermal dissociation of water because energy is released within the small volume of the valve. Alternate methods of pressure reduction (discussed later) for high-temperature/pressure locations are available. These methods are most effective in obtaining an unaltered and thus a more representative sample. The sample must be rapidly cooled, preferably as close to the sampling point as possible, to preserve sample integrity.

E-2


The frequency with which a grab sample is taken is another factor affecting its representativeness. Samples should be taken at a frequency high enough to sufficiently monitor critical conditions that could reach intolerable limits. Also, the sampling frequency should be influenced by periods of load changes. For example, during startup, thermal transients release a variety of contaminants throughout the cycle that should be closely monitored. Also, the increase in the makeup requirement during these load swings introduces a greater risk of obtaining high oxygen levels unless provision is made to remove oxygen as part of the treatment (Appendix A). At such times, oxygen should be very closely monitored. E.2.2 Potential Problems in the Design and Operation of Sampling Systems It is important to be aware of the potential problems in the design and operation of sampling systems. The causes of sample degradation and contamination within the tubing are numerous. One factor contributing to contamination is “crud bursts” caused by sudden increases in velocity. In addition, if the velocity of sample flow is not high enough and if suspended particles are present, deposition may occur. Chemical reactions within the tubing may also contribute to sample degradation. Some of the chemical reactions that can occur include

•

decomposition of hydrazine (or proprietary reducing agents) and/or organic contaminants;

•

reaction of oxygen with hydrazine (or proprietary reducing agents), corrosion products, and/or base metals;

•

dissociation of water; and

•

adsorption of ionic impurities onto corrosion-product (magnetite) particles and deposits.

Another consideration is that in any sample system, there is an interchange of contaminants (such as iron, nickel, and chromium) between the flowing sample and the sample line surfaces. Eventually, an equilibrium state is reached. Whenever the sample is not in equilibrium with the surface, the sample will be changed from its original state. The time needed to reach equilibrium decreases with decreased surface area and increased sample velocity. If insufficient sample velocity, less than approximately 6 feet/second (1.8 m/second) is maintained, then the equilibrium process has been known to take up to 1 month to achieve. Sample line blockage, caused by the accumulation of deposits or large particles caught in smalldiameter tubing, is another source of potential problems. Lengthy sample lines or low sample flow increase the probability of sample line blockage and can cause unacceptable time lags between sample collection and analysis. Proper conditioning of samples is another essential consideration in the design and operation of sampling systems. Continuous analyzers require precise flow and temperature regulation to perform accurately.

E-3


Another source of potential problems in a sampling system is the sample cooling equipment. Avoiding cooling water shell-side deposits and corrosion and maintaining a constant sample flow rate through the cooling devices are important considerations. There are also many possibilities for the degradation of a grab sample during collection and storage. Such potentials for sample degradation are especially critical for pH, conductivity, dissolved oxygen, and hydrazine analyses. Special preparation of grab samples and/or sample containers may be required, depending upon the type of analysis being performed. Chemicals are often added to the container before the sample is added to prevent sample degradation. Also, collection methods for samples to be analyzed for pH, conductivity, dissolved oxygen, ammonia, hydrazine, or organics must exclude any contact between the sample and the air. In addition, storage methods and holding times of samples from collection to analysis require special consideration to avoid degradation of samples. Applicable procedures for collection, preparation of sample containers, and sample storage and preservation are given in ASTM D 3370-95 (26) “Standard Practices for Sampling Water” , and in ASTM D4453-02, “Standard Practice for Handling of Ultra-Pure Water Samples”(27). E.2.3 Sample Collection Obtaining a representative sample of superheated steam requires special sampling techniques. The goals are to avoid deposition of suspended and dissolved solids from the superheated steam, avoid effects caused by pipe surfaces, and withdraw representative amounts of all phases (steam vapor, solid particles, and water droplets). Sampling of steam and water for on-line and grab sample chemical analysis can represent a major source of errors. Changes of up to several orders-of-magnitude in concentration of dissolved and suspended impurities have been observed when sample withdrawal, cooling, and/or transport are not properly performed. Recognizing this problem, EPRI established research project RP2712-8 to define priorities for sampling points and monitored parameters. By analyzing collected samples from a supercritical unit, the project team found that particulates (magnetite) in steam can transport significant amounts of contaminants. On the basis of their findings, the team designed a new isokinetic sampling nozzle, which can collect particulates in superheated steam and impurity-rich water droplets in saturated steam. The team evaluated the sampling system using the new nozzle at four utility plants. Investigations demonstrated that the preferred sampling location for fossil cycles with reheat is (3) the hot reheat steam pipe. This report describes the following in detail:

•

the optimum location of a sampling nozzle in a long vertical section of a steam pipe

•

a single-port tapered nozzle, now called the “EPRI Nozzle”

•

the method of nozzle attachment of a steam pipe and nondestructive inspection methods

•

mechanical design requirements for vortex shedding and resonant frequency

•

the location of depressurizing valves and placement of sample coolers as close to the valves a possible

E-4


In all, evaluations showed that the “EPRI Nozzle” provides a representative total sample of the impurities involved. The nozzle extracts isokinetically a homogeneous sample from a flow region removed from the pipe inner surface and representative of the average flow velocity of the sample fluid. This arrangement results in the withdrawal of the sample with representative concentrations of dissolved, suspended, and volatile constituents. Isokinetic nozzles are equally applicable for liquid samples (condensate), especially when sampling for particulates, such as magnetite or copper oxides. An appropriate nozzle is selected for the desired sample flow and typical conditions of the sample fluid. When the flow rate (velocity through the pipe) changes, sample flow should be adjusted to maintain isokinetic sampling. A summary of sampling criteria is depicted in Table E-1.

E-5

EPRI Licensed Material Cycle Chemistry Data Quality Table E-1 Summary of Criteria for Sampling Sample

Criteria

Pressurized Condensate and Feedwater with Representative Sampling of Particulates

• Isokinetic sampling velocity • Velocity of 5-6 ft/s (1.5-1.8m/sec) • Primary sample cooler at the sample source • Isokinetic sample probe(3)

Boiler Water from Blowdown Lines or Circulation Pump Manifold

• Velocity of 5-6 ft/s (1.5-1.8m/sec) • Primary sample cooler at the sample source

Boiler Water from a Downcomer or Other Pipeline

• Isokinetic sampling velocity • Velocity of 5-6 ft/s (1.5-1.8 m/sec) if consistent with isokinetic requirements • Primary sample cooler at the sample source • Isokinetic sample probe(3)

Saturated Steam from Steam Leads

• Isokinetic sampling velocity • EPRI single port nozzle located 0.75 to 2 steam lead diameters from the steam lead inlet • Sufficient nozzles should be used to provide sampling of a representative fraction of all steam leads along the drum • Primary sample cooler at the sample source

from Large Diameter Line

• Isokinetic sample probe(3)

from Dry Pan Sampler

• Velocity of 5-6 ft/s (1.5-1.8 m/sec)

Superheated Steam

• Isokinetic EPRI sample probe(3) • Primary sampler cooler at the sample source

Condensate at Subatmospheric Pressure (requires a special pump for this service, not available on many fossil units)

• Velocity of 5-6 ft/s (1.5-1.8 m/sec) • Large diameter sample nozzles, net positive suction head actual (NPSHA), greater than net positive suction head required (NPSHR) at all conditions

E-6


E.2.4 Sample Tubing All components of the system in contact with the sample after it leaves the boiler must be constructed of corrosion-resistant materials such as type 304, 316, or 347 austenitic stainless steels (or their low-carbon analogs). The sample tubing should be shaped so that sharp bends, dips, and low points are avoided, thus preventing crud from collecting. Routing should be planned to protect samples from exposure to extreme temperatures, and also to protect plant personnel from high-temperature sampling lines. Steam and water velocities in the sample lines must be high enough to maintain necessary turbulence, preventing deposition in the piping and providing a well-mixed sample. A constant water velocity of more than 6 ft/s (1.8 m/sec) is desirable. In addition, line lengths must be as short as possible to reduce lag times and changes in sample composition. Tube diameters should be small enough to maintain adequate velocities without excessive conditioning or waste (2) requirements. For the EPRI RP2712-3 monitoring program, 1/4-inch (6.4 mm) OD type 316 stainless steel tubing was used throughout and assembled using compression-type fittings. (Wall thickness should be compatible with pressure and temperature at the sample point.) This selection proved to be a good one, as no problems were encountered with pluggage or leakage at any of the four plants monitored and sample flow rates were consistently in the turbulent range. Pumps, Valves, and Fittings All components of the sampling system must be noncontaminating and able to withstand the mechanical and thermal stresses (pressure/temperature) involved. Both diaphragm pumps with priming devices and centrifugal pumps with high rates of recycle may be used for extracting samples from sources operated at subatmospheric pressures. Valves and fittings should be maintained leak-free. Sample Water Recovery and Drains Sample flows should be adjusted to achieve the desired sample velocity, with any excess bypassed to the drain or to the condenser hotwell, or, if appropriate, a sample collection tank. A typical sampling system handles a substantial volume of condensate-quality water each day. If this water is not reused, a significant increase in makeup requirements could result. Uncontaminated samples should be recycled where feasible, while contaminated samples should be directed to waste via the drain. E.2.5 Sample Conditioning This subsection describes general requirements for sample conditioning. Specific requirements of continuous analyzers must also be satisfied. Sample conditioning includes temperature, pressure, and flow rate regulation as well as filtration, the latter applied as required to ensure proper performance of individual analyzers.

E-7


Temperature Regulation Temperature reduction should always be carried out before pressure reduction to prevent flashing, which results in deposit formation. Roughing coolers, located as close to the sample point as possible, are used to cool the sample to approximately 100°F (38°C). Isothermal baths, placed as close to the analyzers and grab sample points as possible, should be used to cool the sample and maintain it at 77°F±1°F (25°C±0.5°C). This eliminates the necessity of making temperature corrections for pH and conductivity, thus improving the accuracy of these measurements. Isothermal baths are available from several vendors specializing in the custom manufacture of sampling panels and related equipment. Flow Rate and Pressure Regulation Most continuous analyzers are sensitive to sample flow rate, and the appropriate manufacturer’s specifications must be followed. Usually, on-line instruments are designed to include flow regulating devices to serve as a check for those devices located upstream, but these on-line devices must not be relied upon. A minimum recommended flow rate for grab samples is 300 mL/min (4.8 gal/hr). However, the actual flow rate should be governed by the rate required to maintain adequate and constant sample line velocities. The design velocity can be maintained by routing excess flow either to waste or the condenser hotwell, as appropriate. Pressure regulating devices must be designed to maintain constant flow to the analyzers during normal, startup, or changing load conditions. Of the many pressure-regulating devices that exist, available types that have been used successfully are described:

•

needle valves

•

rod-in-capillary devices

•

constant head devices and back pressure regulators

Needle valves are sufficient for reducing sample pressures from up to about 500 psig (3.45 MPa). However, at pressures nearing the critical pressure, needle valves have been found to cause the dissociation of water. This is an undesirable situation, causing a rise in the hydrogen and oxygen content of the water. Also, at higher pressures, needle valves are subject to wiredrawing. Rod-in-capillary devices are available in two designs: the fixed rod-in-capillary device and the variable rod-in-tube capillary device. Fixed rod-in-capillary devices have a tendency to plug, and are difficult to clean. A variable rod-in-tube device is effective over a larger pressure range than a fixed rod-in-capillary device because it provides an adjustable pressure drop. In addition, the variable rod device is more easily cleaned.

E-8


Most on-line analyzers require inlet pressure to be 5-20 psig (34-138 kPag) for proper operation. The outputs from most instruments are also flow sensitive, so flow rate should be maintained as constant as possible. These flow and pressure requirements can be met by using a back pressure regulating device or a head cup. A head cup is a relatively simple device (see Figure E-1) and has been used successfully to supply pH and conductivity cells. Immediately upstream of the analyzer, a vertical leg is used to develop an 8 to 10 foot (2.4 to 3.0 meter) head. A trough for collecting overflow is placed at the top of this vertical leg. An overflow is required at all sample conditions to avoid atmospheric sample contamination. They are very reliable but they suffer a major disadvantage: they provide only a very low inlet pressure so usually they cannot be used with analyzers requiring a 5-20 psig (34-138 kPag) sample pressure. Consequently, a back pressure regulator is normally preferred over a head cup when the analyzer requires constant flow at pressure. Use of an orifice installed in a sample flow line is not recommended for pressure and flow control because the orifice size cannot be modified easily to allow flow rate adjustments, and the orifice size changes as it wears, and therefore would have to be replaced frequently to maintain the desired flow rate and pressure, even under baseload operating conditions.

Figure E-1 Head Cup for Constant Sample Flow (pressure in psi at discharge equals the head, H inches of water, divided by 27.6)

E-9


Sample Filters Filters are placed before the pressure and flow regulators to prevent fouling of analyzers in those sample lines not being tested for turbidity or used as grab samples. Closer attention is needed for filters used in cycling plants; because of the heavy accumulation of scale and corrosion products, these filters require frequent maintenance. For some analyzers, filters are included in the internal design. Some analyzers also require the use of micron filters immediately upstream. Manufacturer’s specifications should be checked for special requirements concerning filters. (3) Because of the interactions of soluble impurities with suspended metal oxides , filtered samples may not represent total concentrations. The metal oxides retained on the filter may either adsorb or desorb trace levels of soluble impurities or change the sample composition or its properties, i.e. temperature, pH, etc. Sample Panels In fossil plants, it is quite common to direct all sample lines from one or more units to a common location, often within or adjacent to the laboratory. Most sample panels include the following components:

• • • • • • •

pressure reducers relief valves constant temperature baths valves to control flow rate sample pressure, temperature, and flow rate indicators analyzer cells and elements sample sink for grab samples

A recorder-analyzer panel is typically provided as a separate unit. Sample panels are usually centrally located for convenience. In addition, the surrounding environment is more easily controlled when a central location is chosen for the panels.

E-10


The air in the sample room should be filtered. The room should have a positive pressure to maintain low dust levels. In addition, relative humidity should be between 50% and 70% and temperatures between 65°-78°F (18-26°C). Adequate electrical shielding is required, and refrigeration units should be located outside the sample room, whenever possible, to keep vibrations to a minimum. Local satellite panels should be considered for new plants and for existing plants requiring supplemental instrumentation. Local sample panels provide advantages over a centrally located panel such as shorter sample lines and troubleshooting capabilities. However, these panels must have a central display. Some of the major problems concerning sample panels are the difficulties encountered during servicing and adjustment. Piping, tubing, and fittings should be arranged so instruments or devices may be removed, serviced, or calibrated without disturbing adjacent piping or tubing. The sampling system configuration used in the RP2712-3 monitoring project is shown in (2) Figure E-2 . This configuration proved to be quite effective, and generally caused few maintenance problems during the project’s four monitoring periods.

E-11


Figure E-2 Sampling System Configuration Used During RP2712-3 Project(2)

E-12


E.2.6 Grab Sampling Procedures This section discusses only general requirements for grab sampling procedures. Analysis methods and standards should be checked for any special requirements pertaining to grab sampling procedures. Grab Samples Grab samples serve many purposes within chemical control programs, including:

•

verification of results obtained from on-line monitors

•

a means to measure other chemical parameters not measurable with on-line instrumentation

•

troubleshooting upsets in cycle chemistry

•

optimization of treatments

•

determination of mechanical carryover in drum boilers

Grab samples can be taken at local sampling sites using portable conditioning units. More often, however, grab samples are taken centrally. Sample Containers The selection of appropriate (noncontaminating) sample containers and closures, in addition to the container cleaning procedures, depends upon the parameter being tested and the method of analysis. Before taking samples, the sample container should be properly labeled. This procedure helps to maintain meaningful records. A typical sample label should have the space for and include the following information:

•

sample number

•

date and time of sampling

•

source of sample

•

point of sampling

•

load and mode of operation

•

type and quantity of preservative added

•

results of analysis performed

•

name of sample collector

Provision of complete information is essential where off-site analysis of grab samples is required.

E-13


Sample Collection and Preservation Sample points that are regularly monitored should flow continuously. Local grab sampling points used for troubleshooting should be flushed at least four hours before collection of the sample. After the sample has been allowed to flow at a steady rate for this period, the sample containers and closures should be rinsed with the sample medium and the sample should be taken. The container should be filled completely, leaving no trapped air. The necessary measures should be taken to ensure sample preservation. Special procedures, often involving the addition of a stabilizing chemical, are suggested in various ASTM standards and should be followed where applicable. E.2.7 Corrosion Product Sampling Optimization of feedwater treatment involves corrosion product monitoring campaigns (Section 2). Unfortunately, centralized sampling facilities do not support collection of representative grab samples of metal oxide particulates. Lengthy sample lines have detrimental effects on any suspended corrosion products which may be present. Deposition of these materials in the line leads to either flow restriction or plugging. In some instances, deposited corrosion products may become reentrained in the flowing sample; this phenomenon is sometimes referred to as “crud bursts”. Effective surveillance of corrosion products generally requires a continuously flowing sample. Temperature and pressure reductions should be performed as close to the source as practical to minimize sample line length. This approach is desirable whether the intent is to collect grab samples for laboratory analysis or to beneficially utilize integrated corrosion product sampler devices. For iron and copper transport characterizations, it is very desirable to minimize the length of the sample line and position primary coolers as close to the extraction point as possible. As indicated earlier in Table E-1, isokinetic sampling is recommended and is especially important when monitoring metals in the steam/water cycle. Failure to maintain isokinetic conditions introduces bias, since suspended solids (and moisture in saturated steam) will not be collected in quantities consistent with their levels at the sample source. Additional information and guidance regarding sample extraction and conditioning are available (2,3,5,7) Also, utility personnel should consider a recent report in other EPRI publications. (publication pending) from the American Society of Mechanical Engineers. This document, entitled, Steam and Water Sampling, Conditioning and Analysis in the Power Cycle, was (8) prepared by the Performance Test Codes Committee (PTC 19.11).

E.3 ON-LINE MONITORING TECHNIQUES Water chemistry monitoring is essential for all fossil units to achieve optimum performance from their chemistry programs. This section provides an overview of these topics but further information is available in other EPRI publications(2,5,7).

E-14


E.3.1 Important Considerations for Selection of Proper Monitoring Method Selecting the appropriate monitoring method is vital for obtaining reliable results. Method selection is based on a variety of factors; primarily, the choice must meet the required sensitivity and precision. The sensitivity, precision, and bias of some methods are not suitable for monitoring conformance with these guidelines. Variations in accuracy may mask process fluctuations and erroneously indicate conformance or nonconformance; therefore, the choice of methodology must also meet the required accuracy. Appropriate concentration levels and compensation for possible interferences must be considered before selecting a monitoring method. The selection of a methodology also requires the evaluation of available equipment, time limitations, and the cost of analysis. The experience of the analyst and the amount of quality assurance necessary are also critical in choosing a method. If possible, there should be a trial period for analyzers, and other users’ experiences with the analyzer should be investigated. In addition to using the proper procedure, proper sampling techniques must be followed to obtain reliable results. The analyzers should be relatively simple and easy to calibrate, maintain, and check. Fewer and less complicated components increase an analyzer’s dependability. Certain characteristics should be considered in analyzer selection, including the following.

•

The fluid path should be simple and use a minimum amount of small-bore plastic tubing.

•

Solid-state electronics should be utilized

•

Auxiliary equipment, switches, and alarms should be external

•

A reasonably fast response time should be provided

•

The analyzer should have simple purge and flush requirements for startup and shutdown

•

Maintenance and calibration should be easy and infrequently required

•

The analyzer should be able to resist fouling and scaling

•

The analyzer should have simple sample conditioning requirements

•

A comparison of the various analyzer-specific designs available should be made

•

The analyzer should be able to compensate for variations in sample flow and temperature

A very desirable feature is that the instrument sample manifold provides a means by which independently prepared standard samples may be introduced continuously, wherever desired, for the normal sample flow. Periodic checks of samples represent the surest means of determining the precision and bias of continuous instrumentation. The monitoring and control system must furnish the laboratory with a representative sample, provide dependable feedback of measurements, and automatically protect important system constituents. The control system must also provide local grab sample lines in appropriate locations to evaluate specific pieces of equipment and to sample corrosion products.

E-15


Selection of the appropriate equipment is essential in the control of water chemistry, which is a mandatory requirement for longterm operation of all power plants. E.3.2 Analyzer Calibration Techniques Calibration of monitoring equipment is essential for accurate and precise analysis. The frequency and method of calibration will vary among different types of analyzers; therefore, specific vendor instructions should be followed as closely as possible. Calibration frequency may also depend on the concentration of the particular parameter, the accuracy requirements of the analysis, and the stability of the instrument used for the measurement. All types of analyzers will lose accuracy when measuring zero or very low levels of concentrations over extended periods of time, in which case the analyzer should be calibrated more frequently. A calibration frequency should be established and rigorously followed for each individual instrument. Another aspect of calibration accuracy is the reliability of the standards used for interlaboratory and intralaboratory comparisons. Chemical quality and the care exercised in preparation of standard solutions greatly affect chemical calibrations. E.3.3 Analyzer Operation and Maintenance Sufficient priority must be given to proper operation and maintenance to ensure reliable results. The analyzer is only one part of the system and all parts must be functioning properly to guarantee reliable operation. Manufacturers usually supply specific operating and maintenance instructions for their equipment. Failure to utilize this information will result in unsatisfactory performance and equipment downtime. A preventive maintenance schedule should be established initially and then strictly followed until sufficient performance data exist to warrant modification of the schedule. Such programs vary with respect to the type of analyzer and the nature of its environment.

E.4 DATA COLLECTION, INTERPRETATION, AND MANAGEMENT A data management program is essential to collect and analyze the voluminous information collected via a modern sampling and instrumentation system. Data should be recorded promptly in a format easily interpreted and retrievable. Data must be interpreted promptly by comparison to clearly established and understood criteria. The interpretation must occur in the early stages of data analysis in order to assess the need for any corrective action, which, if necessary, must be rapidly communicated to personnel who can take appropriate corrective action. Interpretation of both short-term (hours to days) and longterm (weeks to months) trends is important.

E-16


Topics covered herein include:

•

data collection

•

data storage and retrieval

•

data analysis and interpretation

•

computerized systems

E.4.1 Data Collection Standardized log sheets for the purpose of record keeping should be developed for the various sample points throughout the plant. Such log sheets should be developed to record the calibration and maintenance of laboratory equipment and continuous analyzers to ensure the regularly scheduled preventive maintenance is being performed in compliance with the maintenance schedule. These records are a necessity in identifying sources of inconsistencies in data. Information recorded on data log sheets should include the following:

•

sample point location

•

date and time of sampling

•

limits or action levels

•

sample temperatures

•

plant load

•

comments

Comments should include information such as the status of the plant at the time of data collection, action levels reached, and corrective actions taken. Examples of successfully used data sheets are provided in EPRI’s Cycle Chemistry Improvement Program.(4) Data from continuous analyzers may be collected manually by periodically logging values from local indicators, or by interpreting strip charts if the plant does not have a data acquisition system. Individual values logged from local indicators will generally be more accurate than values read from strip charts; however, local indicators do not provide the thorough summary of parameter fluctuations that the strip charts do. E.4.2 Automatic Data Collection and Storage Continuous analyzers can be configured to send information directly to data acquisition systems, personal computers, or control room computers. Automatic data collection provides the same information as a strip chart but with greater precision and eliminates both the time necessary to evaluate strip charts and the element of human error.

E-17


E.4.3 Manual Storage of Chemistry Data Manual data storage and retrieval should include accessible files of completed data sheets, daily chemistry report forms, trending plots, and action level notifications. Each data sheet should be reviewed by plant personnel responsible for chemistry to ensure the water chemistry is within normal ranges for control and that any necessary corrective actions have been taken. A computer-based system is more efficient for storing and managing larger quantities of data. These data can be entered manually or recorded directly from a data acquisition system. Computer storage of data allows past data to be easily retrieved and, depending on the system, also provides the capability to manipulate and analyze present and past data interactively. E.4.4 Data Analysis and Interpretation Data analysis and interpretation provide for the identification of adverse trends and problems in system chemistry so corrective actions can be taken. This section will introduce the means by which data analyses and interpretations are accomplished. The “eyeball” method, as the phrase infers, is a process for viewing of data as it is being collected to detect “out-of-spec” values and inconsistencies among the data. Graphical plotting makes use of a visual, easily interpreted format that is easy to use, low in cost, and emphasizes the order of observation with respect to time. Putting information into graphic form is recommended for all measured parameters so slow changes can be identified in a timely manner. This procedure should be done so as to result in a simple trend graph or a control chart. A trend graph is a plot of sample parameters versus time and should have a duration of at least 30 measurements to be useful in identifying a meaningful trend. A control chart is equivalent to a trend chart, with the addition of a zero deviation point and upper/lower control points, providing specific action limits. Such capabilities, as well as generation of mass balances and histograms, are standard features of most modern, computerbased data management programs. E.4.5 Roles of Plant Personnel in Water Chemistry Data Collection, Interpretation and Management Plant personnel (typically technicians or operators) are responsible for the collection and analysis of samples, completion of data sheets, construction of graphical charts, and calibration of analyzing devices. The initial responsibility for chemistry program management lies with these individuals. They must be trained to recognize not only out-of-specification conditions, but also trends that appear to be abnormal. When either of these conditions occurs, they must be instructed, by procedure and training, to notify the appropriate personnel for initiation of corrective actions.

E-18


Supervisory personnel are responsible for a more detailed review of chemistry and trends. They must be able to explain recorded occurrences and make suggestions for corrective actions to be taken and then report their findings to management. It is the responsibility of the operating personnel to be aware of sudden changes in operating parameters and to interpret and take corrective actions when conditions needing immediate attention arise. The operators must be adequately trained so they are able to perform the described duties. E.4.6 Expert Systems for Water Chemistry Management An EPRI project was initiated for the development and demonstration of an expert system to gather and analyze plant chemistry data in such a way as to result in significant conclusions relative to the need for corrective actions, as well as what those corrective actions should be. The resulting software package, known as EPRI ChemExpert, offers a variety of capabilities and features of considerable value to fossil plants(9). These include:

•

Real time displays

•

– Cycle arrangement – Sample points and chemical feed points – On-line analyzers (location, current readings and status) Management of chemistry data

•

– On-line analyzer readings (direct input) – Grab sample data (manual input) – Operating parameters (via interface with plant data acquisition systems or distributed control systems) – Reports and graphics – Data storage Data analysis and interpretation – Chemistry problem identification (over 30 problems) – Corrective action recommendations – Diagnostic capabilities based on EPRI chemistry guidelines and experience of experts in the field of fossil plant water chemistry – Tracking of on-line analyzer performance – Calculation of mechanical carryover and unit specific boiler water chemistry limits (units with drum boilers)

E-19


•

Personnel training

– – – –

Definitions of chemistry terms Key chemistry concepts Cycle chemistry tutorials Chemistry control simulation

The ChemExpert software package currently consists of four separate programs. A customization program allows the user to define unit characteristics and cycle chemistry. Either two or three action levels may be specified by the user. A diagnostic program provides verification of analytical data, warning of chemistry readings outside of limits (including problem identification and recommendation of actions that should be taken), determination of pressure and carryover dependent limits for impurities in the boiler water of drum type units, and data storage, plotting and report preparation. A cycle description and maintenance program may be used as a database of unit equipment characteristics and for planning and scheduling of operation, maintenance and calibration functions related to the chemistry program. The training program is designed for beneficial use by plant operators and technical personnel responsible for cycle chemistry.

E.5 VALIDATION OF CHEMISTRY DATA Verification of chemistry data is generally required prior to initiation of corrective actions, not only to determine that action is, in fact, required, but also to assure that the contemplated actions are appropriate for the chemistry conditions which exist in the unit. As many chemistry conditions require rapid recognition, evaluation and response, it is imperative that those plant personnel present at the onset of possible chemistry excursions be capable of providing the verification function. At facilities that neither have chemistry specialists on site at all times nor utilize an expert system, it is therefore crucial that operators or technicians normally present on a “24/7” basis be trained to take the steps needed to verify chemistry data and proceed with corrective actions that may be required. Conversely, responsibility for chemistry data validation measures, as preferably outlined in the plant Quality Assurance (QA) and Quality Control (QC) program, must remain the responsibility of technical specialists. The QA/QC procedures and other data validation functions should serve to demonstrate the statistical significance of the chemistry data, thereby minimizing any doubts or concerns that actions are needed or may be inappropriate when the verification process is followed and confirms the existence of a chemistry excursion. Chemistry data validation activities should therefore, be used proactively for best results. Plant cultures which apply them reactively (only or mainly after possible chemistry excursions or after chemistry-related damage has been experienced) or have no formal data verification programs are clearly at a disadvantage. Those facilities in the latter category often operate with a false sense of security when they provide suitable levels of chemistry surveillance capabilities and data verification procedures but fail to address data verification requirements.

E-20


The needs for data validation are by no means unique to the fossil electric industry. QA/QC programs are applied in innumerable applications, many of which have far more rigorous requirements than those of fossil plants. Important and obvious examples include medical technology, the aerospace and aeronautics industry, pharmaceuticals production, computer components and communications technology, defense applications, food processing, security systems, and nuclear power generation, to name but a few. As a result, the language of QA and QC features an extensive vocabulary, which often, unfortunately, constrains its application to areas where such action is elective, rather than mandatory. Fortunately, application of effective validation techniques to fossil plants requires familiarity with a relatively small number of terms and concepts, as summarized in the working definitions of Table E-2. Recognition of the importance of cycle chemistry data validation and quality has resulted in considerable effort on the parts of utility organizations, consultants, and instrument manufacturers to evaluate requirements of fossil plants and to establish manageable programs capable of meeting these (10-12) needs . Ensuing discussions provide guidance on this important subject.

E-21

EPRI Licensed Material Cycle Chemistry Data Quality Table E-2 Working Definitions of QA/QC Terms for Fossil Plant Chemistry Data Validation Accuracy:

The agreement between the amount of a component measured by the test method and the amount actually present. Accuracy includes both precision and bias of the method.

Precision:

The reproducibility of a method when repeated on a homogeneous sample under controlled conditions, regardless of whether or not the observed values are widely displaced from the true value as a result of systematic or constant errors (also sometimes referred to “reproducibility”).

Bias:

Note.

Both accuracy and precision can vary as a function of the concentration being measured. They can also vary due to other factors such as temperature, concentrations of interferences, etc. Therefore, statements of precision and accuracy should be expressed in terms of the concentration over the range of concentrations to which they apply. If precision and accuracy are stated in percent, the range over which the relative precision or accuracy applies should be explicitly stated (e.g. “a relative accuracy of ± 10% if the amount present is achieved over a span of 0.01 to 10 mg/L”).

St

Interinstrument precision—the standard deviation of the measured values of a specific parameter obtained by multiple continuous instruments (of the same type) at the same value in homogeneous samples of the same composition (matrix). Also, interlaboratory precision when referring to grab sample analysis by multiple laboratories on samples of the same matrix as defined by ASTM D 2777-86.

So

Intrainstrument precision—the pooled standard deviation of the measured values of a given parameter obtained by multiple continuous instruments (of the same type) on repeated measurement of the same specified value of the parameter in homogeneous samples of the same sample composition (matrix) when each measurement follows an interval of time where the instrument was equilibriated to a value of the parameter that was significantly different from that specified. Also, the interlaboratory or single operator precision, as defined in ASTM 2777-86, when referring to grab sample analysis by multiple laboratories on samples of the same matrix.

The persistent positive or negative deviation of the method average value from the assumed or accepted true value. Note:

Both precision and bias vary as a function of the concentration being measured. They can also vary because of factors such as temperature, concentrations of interferences, etc. Therefore, statements of precision and bias should be related to the concentrations to which they apply.

Sensitivity:

The minimum observable difference in very small amounts of a component in a sample. Also, sometimes totally incorrectly referred to as “detection limit” or “detectability.”

Range:

The span of amounts or concentrations over which a component can be measured monotonically in a sample. Also sometimes referred to as “span.” Usually, the range covers the concentrations for which the method or instrument is actually calibrated.

Response Time:

The period of time required for an analyzer to properly measure the components in a sample from the time the sample enters the analyzer. Also known as “analyzer dead time.” Note:

Response time should be quantified by referencing the time necessary to provide a certain percentage response to a step change in concentration. Preferably the response time for a 95% response should be provided; however, a statement of response time in terms of a somewhat lesser response (such as a 90% response) may provide sufficient information.

Sampling Error:

The deviation between the true concentration, pH, conductivity, etc. in the stream prior to sampling and the concentration in the sample presented to the analyzer; or contained, in the sample container prior to preservation or transportation for analysis.

Analytical Interference:

Analytical errors caused by interference from some chemical compounds in the mixture being analyzed. Interferences should be listed for all water treatment chemicals used and for their decomposition products.

E-22


E.5.1 Precision, Accuracy, Bias and Drift The American Society of Testing and Materials (ASTM) defines accuracy as “a measure of the degree of conformity of a single test result generated by a specific procedure to the assumed or accepted true value and includes both precision and bias”.(18) In other words, it is a measure of exactness or correctness of a value and is affected by both random sources of error (precision) and systematic error (bias). ASTM defines precision as “the degree of agreement of repeated measurements of the same property, expressed in terms of dispersion of test results about the arithmetical mean result, obtained by repetitive testing of a homogeneous sample under specified conditions. The precision of a test method is expressed quantitatively as the standard deviation computed from (13) the results of a series of controlled determinations”. In other words, precision is a measure of the reproducibility or random uncertainty of the data set. In the absence of systematic uncertainty, data should follow a random distribution pattern. This pattern is characterized by the normal distribution curve shown in Figure E-3.

Figure E-3 Normal Distribution Curve for Random Data

E-23


A measurement of precision is given by the standard deviation (Sx) of the data set. Standard deviation can be used to express the distribution (or spread) of the data around the average value of the data set (X ). In many cases, it is convenient to express the standard deviation relative to the concentration as a relative standard deviation (RSD), sometimes expressed as the percent relative standard deviation (%RSD):

RSD = [S X / X ]

(eq. E-1)

%RSD = [S X / X ]× 100%

(eq. E-2)

and

In this way, precision can be compared at varying concentrations and %RSD can be used to quantify changing precision as a function of concentration. Understanding %RSD can provide insight into instrument detection limits and instrument linearity. The %RSD tends to be very large for very low analyte concentrations. For example, for concentrations at the decision limit (LC), %RSD ≈ 60%; at the detection criteria (LD), %RSD ≈ 30%; at the determination limit (LQ), (14) %RSD ≈ 10 (as specified by Currie ). ASTM defines bias as “the persistent positive or negative deviation of the average value of a test method from the assumed or accepted true value”.(15) If the data are randomly distributed, accuracy expresses the deviation from the expected value as shown in Figure E-4. Bias is called recovery when expressed as a percentage of the true value. Thus, accuracy may be expressed as a combination of recovery and relative standard deviation at a given true value.

Figure E-4 The Relationship Between True Value, Measured Value, and Bias. The bias in the example shown is negative, but a positive bias is also possible.

E-24


In some cases, it is convenient to express accuracy requirements in terms of percent deviation from the target value (i.e., from the true value):

% Deviation = [X / T arg et ]× 100%

(eq. E-3)

Many measurement methods are more accurate at the high end of the range covered, leading to percent deviation values that change over the dynamic range of the measuring instrument. For convenience, average values may be cited for specific ranges of the parameter measured, e.g., ±20% for 1-100 ppb and ±10% for 100 ppb - 1 ppm. In some cases, a lower limit of accuracy may also be cited, e.g., ±0.5 ppb or ±10%, whichever is larger. Clearly, using percent deviation limits in this is imperfect and does not address the overall capability of the instrument. A more scientific approach utilizes the Student’s “t” test.(16) This approach measures accuracy as a function of the deviation from the expected value at a specified confidence coefficient, usually 95%. The test assesses the instrument’s accuracy based on observed instrument behavior and is most appropriate for instruments that are used for diagnostic purposes. The Student’s “t” test mathematically determines whether a statistically significant bias exists at a specified confidence coefficient. In this way it can be said conclusively that a statistically significant bias does or does not exist. Several commercially available programs can be used to calculate statistics (such as percent deviation or the Student’s “t” test), make data comparisons, and archive information. Selection of a specific program will depend upon a number of factors, including a user’s data acquisition capabilities and data management needs. Special testing can be performed to verify accuracy and precision of instrumentation or grab sample analyses over various working ranges based on the needs of the user. The precision and accuracy of a measurement may be affected by many factors, such as temperature, concentration of reagents, interferences, availability of traceable calibration standards at the concentration expected in the samples, calibration procedure, and sample size, as well as variations that develop from one instrument, operator, or laboratory to another. Consequently, accuracy and precision statements should include the range of conditions over which they apply. While the accuracy of an instrument can be determined immediately when it is placed into service, drift develops over time. Drift may be defined as the change in bias in a specified period caused by changes in the electronic zero of the instrumentation but, more particularly, by deterioration of the sensor. The American National Standards Institute (ANSI) defines drift in ANSI Standard C85.1-1963 as “an undesired change in output over a period of time, which (17) change is unrelated to the input, environment, or load.” Drift is an important example of systematic errors (Figure E-5), which must be minimized to ensure data quality. Instrument manufacturers do not always provide precision, accuracy, bias and drift information. Such information must therefore be determined by the user.

E-25


Figure E-5 Random and Systematic Errors

E-26


E.5.2 QA/QC for Sampling Systems The effect of sampling errors on chemistry data are potentially significant and should be minimized through proper equipment selection. As shown in Table E-3, there are many sources of sampling errors. Total sampling error can span several orders-of-magnitude, much larger than typical analytical errors. Except for the error caused by improper withdrawal (collection) of the sample, other errors can be determined by analyzing known samples before and after the sampling system at otherwise identical sample conditions (temperature, flow, chemical species). Table E-3 Typical Sampling Errors • Sample Withdrawal

— not representing the sampled stream — not representing all phases (solid, liquid, gas)

• Deposition in sample tubing and other components • Deposit release • Chemical reactions in sample line or cooler • Deposition in the pressure reducing device • Dissociation of water in the pressure reducing device • Leakage of cooling water into the sample in the cooler • Corrosion of the sampling system—generation of corrosion products

Sample withdrawal error can be determined by using alternate sampling taps and nozzles or a different sampling location. Experiments have shown that for superheated and wet steam sampling, an isokinetic sampling nozzle is needed, correctly located in the steam pipe, such as (3) described in EPRI TR-100196, Development of a Steam Sampling System. The same type of nozzle is needed for feedwater when sampling for suspended oxides. (Also see E.2.3 and E.2.7) In sampling boiler water from the steam drum, the sample could be mixed with the incoming feedwater or separated moisture and thus be different from the water in the boiler tubes. E.5.3 QA/QC for On-Line Instruments Instrument Calibration and Standardization Establishment of effective calibration and standardization procedures are the starting point of the QA/QC program. Since manufacturer’s procedures vary widely in their ability to ensure traceable accuracy, it is advisable to independently calibrate the instrument components, especially the sensing elements, after the manufacturers’ procedures have been followed. Independent calibration is particularly advised for temperature sensing elements in conductivity, pH, and other instruments where automatic temperature compensation is an integral part of the instrument measurement process. E-27


For high-purity water, conductivity instrument cell calibration by comparison with a standard reference conductivity cell and measurement bridge (without temperature compensation and on the same sample flow) is recommended initially and possibly at various later intervals, depending on the fouling characteristics of the particular sample stream. For pH, calibration is best accomplished with flowing standards using two concentrations, as discussed below, one above and one below the desired measurement range, following completion of the manufacturers’ initial procedures. Validation of On-Line Monitoring Instruments At least four approaches to validation of on-line monitoring instruments are available: the reference sample procedure, the line sample procedure, instrument comparison, and the standard (18) injection method. The first two approaches are described in ASTM D3764 , a standard practice that was developed for validation of process stream analyzers in petroleum or petrochemical product steams, but which is equally applicable to power plant water streams. The last two methods are perhaps better suited to calibration of instruments but they have been used on occasion for validation purposes. The reference sample procedure involves introduction of reference standard samples into the analyzer and subsequent comparison of the measured values with the reference values. For instance, to validate a pH monitoring device, the plant water stream could be diverted and replaced with a standard buffer solution flowing for a period long enough to allow the pH measurement. This approach is particularly useful for validating new and replacement analyzers to ensure they have been installed and commissioned in accordance with the manufacturer’s recommendations. The line sample procedure makes use of a plant water stream of interest and involves taking at least seven grab samples for laboratory analyses and comparing these results with those obtained from the on-line analyzer at the times of sampling. The data are evaluated statistically (Student’s “t” test) to validate the analyzer. Suitable statistical software packages are available from some instrument suppliers to help perform this task. The line sample procedure is used more as an operational check on an analyzer that has been previously validated, and is useful in establishing the ongoing precision and bias of an analyzer in service. The third approach involves comparing the results obtained from the on-line analyzer (A) with those obtained from a second on-line analyzer (B) exposed simultaneously to the same water stream. The latter may be a high quality, portable unit that is first validated and calibrated in the laboratory, then attached to the plant water stream system when needed to validate or calibrate analyzer A. Measurements on analyzer B are assumed to be correct and the accuracy of analyzer A can be assessed by comparing measurements on the two instruments. This method is useful for validating and verifying analyzers that measure very low analyte concentrations where the grab sampling process may introduce significant error. However, in such cases, analyzer B must first be calibrated using a water having analyte constituent concentrations near those needing measurement in the plant water stream. Clearly, the value of this method is very dependent on the precision and lack of drift in instrument B during the validation period.

E-28


The standard injection method can be considered a variation of the reference sample procedure in that it involves the introduction of standard solutions into the analyzer. However, the standard injection method does this by injecting the standard solution at a controlled rate into the plant water stream, also flowing at a known rate, such that the concentration of the resulting diluted standard solution can be calculated accurately. The method is otherwise similar to the reference sample procedure. In high-purity water systems, it has the great advantage of presenting the standard sample without contamination and permitting measurement with similar matrix and flow. A general protocol for establishing the validity of on-line water analysis instruments is provided (19) (20) in ASTM D3864. More specific protocols, detailed in EPRI Report GS-7556 , were followed in a collaborative validation program at four power plant sites. As with the grab sample methods, on-line instruments must be validated by collecting a minimum of six data sets at each concentration level. For the four-plant collaborative study, this minimum requirement was satisfied for all of the methods except three: on-line ion chromatography, total organic carbon, and dissolved oxygen. Nevertheless, these latter three methods should be considered for use, although treated with appropriate caution. The successful validation of on-line instrumentation requires: proper installation of water and steam sample lines; close control of sample temperature, pressure, and flow rate; preparation of standard samples; and accurate analysis of grab samples. These requirements are discussed in more detail below. Synthesizing Standard Samples Validation of the measurement methods can be accomplished using standard solutions (samples), over a range of well defined concentrations, produced in a continuous flow. Grab samples can be taken from this flow or the flowing standard solution can be directed to the on-line monitoring equipment under test. A method of synthesizing flowing standard solutions was described in (2) EPRI Report GS-7556 Volume 1. It involves the continuous production of very high quality reagent water, flowing at a constant, continuously measured rate, into which one or more concentrated standard solutions are injected continuously with metering pumps. The very high quality reagent water is prepared from steam condensate which is passed through a de-oxygenation exchange bed, a carbon bed, a mixed bed and a 0.2-micron filter. If source water of lower quality than steam condensate (e.g. makeup storage tank water) is used, additional purification steps may be needed. To avoid precipitation of the chemicals in the concentrated standards, Cl, SO4, and organic acids are prepared in an acid solution, and Na, HCO3, N2H4, NH3, and SiO2 are prepared in an alkaline solution; oxygen is injected as air-saturated (STP) pure water. Standard concentrations are calculated from the continuously logged standard solution flow, the metering pump flow, and the assayed concentration of the concentrated standard solutions. The metering pump flow is calculated from calibration curves, established from the rate of change of weight of the vessels containing the concentrated standards, as measured by an electronic analytical balance. Sample synthesizers that are capable of delivering sample streams of known chemical composition to several analyzers are also commercially available. E-29


Potential Problems in Preparation of Standards Standards containing ammonia and organic acids (acetate, formate, etc.) may decompose or change in concentration before use unless care is taken to avoid this. Changes in composition can be prevented or minimized by using the standards immediately after preparation, or refrigerating them to 4°C during storage. The refrigerated samples should be allowed to come to room temperature just before use not only to avoid effects of temperature on the accuracy of the metering pump calibration curves, but also to avoid condensation on the outside of the standard solution tanks which, during creation of the calibration curves, would cause a weighing error. E.5.4 Charting of Chemistry Data to Track Instrument Performance Whatever the interval and concentration selected for the QC tests on each instrument, including those used for analysis of grab samples, the results should be recorded on a control chart. General examples are presented in Figure E-6. As an example using data from an on-line cation conductivity instrument, charts such as those show in Figures E-7 and E-8 record the RRs and RSDos obtained from each month’s QC test. The upper and lower acceptance limits for RR and RSDo are shown along with the past test results and the most recent results from the example QC test, those for month 12. Again, the use of relative recovery and relative standard deviation simplifies charting as well as the necessary calculations since it is likely there will be small differences in the true concentration used in each QC test episode. The most useful feature of control charts is for the detection of trends in instrument performance that may require alternation of the test frequency or the instrument maintenance and calibration procedures. A trend of decreasing cation conductivity RR, as in Figure E-7, suggests a greater degree of exhaustion of the resin cartridge at the time of the tests. The latter could be the result of increased sample flow rate, increased amine level in the sample, decreasing quality of the fresh cartridge, or change in the conductivity cell constant. A review of the record of specific conductivity and pH indicates amine level to be relatively unchanged over the last several QC tests. Thus, the RR trend, when coupled with the observed increase in RSDo, noted in Figure E8, suggests that the first priority for potential corrective action would be the careful review of the resin cartridge maintenance procedures.

E-30


Figure E-6 Laboratory Quality Control Charts

E-31


Figure E-7 Cation Conductivity Control Chart, RR%, CC = 0.5 µS/cm

Figure E-8 Cation Conductivity Control Chart, RSD%, CC = 0.5 µS/cm

E-32


E.5.5 Interlaboratory Assessment of Grab Sample Analysis Methods For individual plants, the requirements for grab sample analysis include use of appropriate sample withdrawal and conditioning equipment, use of appropriate sample collection techniques and preservation methods and charting of data. As these requirements have been covered previously, results of interlaboratory assessment of available grab analysis methods are summarized. Five methods of analyzing grab samples for trace substances in high purity water were validated in an inter-laboratory program, designed in accordance with ASTM D2777-85.(21) Up to 20 participating laboratories developed precision and recovery estimates, limits of detection, and limits of quantification for the following methods:

•

Proposed Method of Test for Low-Level Sodium in High Purity Water by Graphite Furnace Atomic Absorption Spectrophotometry (ASTM Draft 3)

•

Proposed Method of Test for Low-Level Calcium in High Purity Water by Graphite Furnace Atomic Absorption Spectrophotometry (ASTM Draft 1)

•

Proposed Method of Test for low-Level Magnesium in High Purity Water by Graphite Furnace Atomic Absorption Spectrophotometry (ASTM Draft 1)

•

Standard Test Methods for Trace Anions in High Purity Water by Ion Chromatography (ASTM D5542(22), Method A: Chloride, Phosphate, and Sulfate)

•

Standard Test Methods for Trace Anions in High Purity Water by Ion Chromatography (ASTM D5542(22), Method B: Fluoride, Acetate, and Formate)

ASTM practice calls for a minimum of six data sets at each concentration level in order to derive valid estimates of performance. Because insufficient data were generated in the study for the following grab sample methods, they could not be properly validated but, with appropriate caution, they should be considered for future use:

•

Proposed Method of Test for Trace Cations in High Purity Water by Ion Chromatography (ASTM Draft 1)

•

Standard Test Method for Low-Level Total Silica in High Purity Water by Flameless Atomic Absorption Spectroscopy (ASTM D4517-85(23))

•

Standard Test Method for Low-Level Total, Organic, and Inorganic Carbon in High Purity Water by Ultraviolet (UV) or Persulfate Oxidation, or Both, and Infrared Detection (ASTM D4779(24))

(25) The results of this validation study were reported in 1988 and presented in detail in Appendix E of EPRI Report GS-7556 Volume 1. In this program, grab samples were collected in special sample bottles from flowing synthetic standards (prepared as described above), then refrigerated until they were analyzed.

E-33


E.5.6 EPRI Experience With Chemistry Data Validation The EPRI RP2712-3 project provided many useful insights with respect to validation of chemistry data collected from operating fossil plants.(2,20) For all continuous instruments, a satisfactory QC program must include the periodic use of a flowing standard sample that is presented to the instrument under the same conditions that exist for the samples from the cycle. This procedure is especially important for instruments on highpurity sample streams where grab samples are easily contaminated. Continuous pH instruments on high-purity water are best calibrated, as well as validated, by passing through the flow cell, standards of composition similar to those of the samples to be measured, as described in ASTM D 5128. For power plant operation where ammonia is used for pH control in the feedwater and steam, high-purity water with 100-500 µg/L NH3 makes suitable standards. Where the use of a flowing standard is not feasible, a neutral ionic buffer should be added to the grab sample before measurement of the sample pH with an independent reference instrument. In neither case, use of flowing standards or grab sample comparison, should the pH cell be standardized with concentrated buffer solutions at any time other than the initial installation of new electrodes. Where grab samples collected in parallel with the continuous instrument must be used for quality control, a minimum of seven replicate samples must be collected and analyzed under steady-state conditions, one for each of the seven required separate instrument data points, at a concentration of interest (see following section). The methods of analysis of the grab samples also must have appropriate quality control, preferably using traceable standards run “blind” by the analysts, on the basis of one standard per day or one per 10 samples analyzed, whichever is greater. (2,20) During the EPRI RP2712-3 project , no determination was made of the optimum frequency at which each type of continuous instrument should be validated. However, the two validation runs at each site during RP2712-3 were separated by an approximately two-week interval, which was sufficient for some instruments to show substantial change in bias (measured concentration minus true). The bias for the continuous SiO2 analyzer, for example, increased from 8 to 12 µg/L over that time. Specific conductance, the most stable and precise of the continuous instruments, showed at a level of 2.0 µS/cm, an increase in bias from 0.14 µS/cm to 0.42 µ/S/cm. Therefore, a conclusion can be drawn that the frequency of validation at a plant should be initially set at no less than once per month. This validation should be performed at a single concentration selected to be in the middle of the normally encountered operation range, until sufficient validation data is gathered for each instrument to warrant changing the frequency. More test levels as well as more frequent tests may be required in some circumstances and with certain instruments.

Performance Determination Seven replicate presentations at the same desired test level are necessary to develop a satisfactory estimate of intrainstrument (single instrument) standard deviation (So) for comparison with established precision standards for a particular instrument at that level.

E-34


The presentation of each of the seven flowing standard replicates must be preceded by a reasonably short period of stable instrument reading on the regular sample source (or other source) where the concentration is different from that of the flowing standard. This interlude should be of such length that >90% of the expected reading is attained. Seven replicates are required at a particular concentration in order to develop meaningful acceptance limits by which to determine if the instrument measurements are in control—that they perform with the expected precision and accuracy. The large number of replicates required for statistically useful limits led to the above recommendation that only one test level be selected for the initial QC program for each instrument. When using only one test concentration for the QC test for each instrument, it is important that the same test level be employed for the QC test each month, or other chosen interval, since the acceptance limits vary considerably with test concentration for some instruments while remaining nearly constant for others. Acceptance Limits The standard deviation and mean of the seven replicate instrument measurements can be compared to the expected precision and mean previously determined for the particular instrument being evaluated. Since the flowing standard method may not produce exactly the same numerical results when performed at different times, (whether at the same plant or at different plants), a useful procedure is to express both the standard deviation and the mean in terms relative to the true value of the standard used. Thus, the relative recovery (RR) is the measured mean concentration of the seven replicates divided by the true concentration (T) expressed in percent, X/T x 100. Similarly, the intrainstrument relative standard deviation (RSDo) is the standard deviation of the seven replicates divided by the mean concentration expressed in percent, So/X x 100. Both the RR and RSDo determined from the QC test can then be compared with values shown in Table E-4 for the nearest true concentration, since both RR and RSDo can be considered constant for a reasonable range on either side of the selected value. The calculated value of RR actually obtained for a validation should not be greater than the upper value nor less than the lower shown in Table E-4 for RR at the selected true value. Similarly, the calculated RSDo should not exceed the upper nor be less than the lower value shown for RSDo in Table E-4. Use of Acceptance Limits in QC A QC test of a particular cation conductivity (cc) instrument, for example, would start with presenting the instrument with a flowing standard sample, with a desired target cc of say 0.5 µS/cm. The standard would be presented seven times in alternate succession with a nonstandard sample with a different cc, say 0.25 or 1.0 µS/cm. Sufficient time would be allowed for the reading to stabilize. The true conductance of the flowing standard was found to be say 0.55 µS/cm based on the analysis of the concentrate, and the dilution ratio used to produce the flowing standard. If the mean of the seven replicate cc results, X, was 0.34 µS/cm, the RR would be X/T x 100, or 62%. Reference to Table E-4 shows cc = 0.5 µS/cm is the closest to the 0.55 µS/cm standard. The corresponding RR shows an upper acceptance limit of 130% and a lower limit of 59%. Therefore, it was concluded the instrument bias was within control.

E-35


Table E-4 Continuous Instrument Acceptance Limits, 95% Confidence Interval (Based on Seven Replicates) True Concentration

Specific cond. (µS/cm)

Cation cond. (µS/cm)

Na (µg/L)

NH3 (µg/L)

Relative Recovery (1%)

Intrainstrument Relative Standard Deviation (%)

Lower

Upper

Lower

Upper

1

79

119

3.7

7.4

2

89

109

3.6

7.4

5

93

105

3.6

7.4

10

94

104

3.6

7.4

0.1

79

115

6.5

14.3

0.2

70

121

4.4

9.7

0.5

59

130

3.1

6.8

1.0

53

135

2.6

5.8

1

5

372

5.8

11.2

2

1

243

4.7

9.0

5

57

153

3.7

7.1

10

74

124

3.3

6.4

100

61

145

7.5

18.7

200

86

110

7.6

18.8

500

92

97

7.6

18.9

1000

93

94

7.6

18.9

Likewise, assume the seven replicate cc’s yield a So of 0.22, and an RSDo of 6.5% ((0.22/0.34) x 100). Examination of Table E-4 for cc = 0.5 µS/cm shows that the RSDo should not be less than 3.1% nor greater than 6.8% for the instrument to qualify as producing acceptable results. Since both the RR and the RSDo are within the acceptance limits, the instrument is operating in control. Where continuous water quality instruments of a given type can be supplied simultaneously with an alternative high-quality water supply, useful information can be gained on multipleinstrument precision. Multiple-instrument precision has been shown to be generally smaller than true interinstrument precision but greater than instrainstrument precision. Thus, the RSDs obtained from alternative water supply tests normally can be expected to exceed slightly the range of acceptance limits in Table E-4, which are based on intrainstrument precision. No information can be obtained from alternative water supply tests, however, on relative recovery unless the composition of the alternative water supply can be established independently and with sufficient accuracy.

E-36


E.6 SUMMARY A quality assurance (QA) program for cycle chemistry data is essential if comparisons are to be made between measurements at different points within the cycle, between the same point in the cycle in different plants, and between specific points in the cycle and the plant’s operating chemistry limits and action levels. The QA program must include quality control (QC) procedures for both grab sample and continuous instrument measurements. To follow the instrument manufacturers’ calibration procedures is generally not sufficient, as these procedures seldom employ traceable standards. More importantly, they do not determine reproducibility (intra-instrument standard deviation) or recovery so that instrument performance may be documented and trends in performance readily observed. Periodic validation of measurement procedures, especially continuous instrument performance may be documented and trends in performance readily observed. Periodic validation of measurement procedures, especially continuous instrument procedures, by making replicate measurements on a continuously flowing standard generates a record than can be compared readily with suggested acceptance limits as derived from the EPRI RP2712-3 program. Charting the record of validation results allows early detection and correction of problems with instrument accuracy. Close control of cycle chemistry requires accurate chemistry measurements. Analytical quality control must be incorporated into laboratory procedures if reliable results are to be obtained. A quality control program provides for the detection/correction of problems and statistical verification of performance. The objective of such a program is to reduce measurement errors to a minimum so data can be used reliably to monitor and control cycle chemistry. Unacceptable analytical variability arises from many sources. These include:

•

incorrect analytical procedures

•

calibration errors

•

poor technique in following analytical procedures

•

degradation of reagents utilized in analyses

•

improper sampling procedures, contamination of the sample during collection and storage, reaction of the sample with the sample container

•

chemical reactions during sample storage

•

organic growth of bacteria and algae in the sample container

Quality control programs must provide guidelines for calibration and standardization of all laboratory equipment and reagents; and, they must establish procedures for estimating precision, bias, and sensitivity for the analytical methods performed. Limitations in the availability of equipment, number of samples, or time and experience of personnel can affect the reliability of the data obtained. Therefore, quality assurance procedures must be developed specifically in consideration of each plant’s needs and abilities.

E-37


Systematic and random errors must be controlled to the maximum practical extent before evaluating precision and accuracy. Random error is present in any set of analyses in which repeated measurements are taken on the same sample. Random errors are the scatter of measurements about the true value. Systematic errors are referred to as bias and result from a tendency for the data to be greater or smaller than the true value. Reduction in systematic errors can be accomplished through proper analytical procedures, traceable calculations, optimum personnel performance, and proper purity of reagents used in all analyses. Reagent water is a major source of systematic error; therefore, care should be used in its preparation to avoid contamination. Various methods of evaluation of performance are available. A convenient way of monitoring the precision and accuracy of a method is to construct control charts. Control charts are used to evaluate the quality of the measurement process. Individual, multiple sample, and X-R control charts can be generated. Individual and multiple control charts operate similarly; however, multiple control charts span a range of concentrations, which avoids potential bias. The X-R control chart incorporates changes in the mean, X, and changes in the standard deviation, S, thereby making it the most useful of the three charts available. Confidence in measurements in strengthened by inter-laboratory as well as intra-laboratory comparisons. Inter-laboratory programs should be used to establish agreed-upon requirements and laboratory procedures necessary to achieve desired quality standards. An independent facility should then inspect laboratory equipment and evaluate performance. Intra-laboratory programs identify the variability of measurement obtained when a single laboratory measures a sample repeatedly. System audits also provide valuable quality assurance information. A system audit is an investigation of the interaction of all laboratory procedures. The audit should be performed at appropriate intervals to guarantee that each aspect of the QC program is operating efficiently. A performance audit of an individual instrument is based on the outcome of the analyses of a “blind” standard sample, either on an inter-laboratory or intra-laboratory level. A quality control program must be implemented for all analyses and must be evaluated on a regular basis to be effective. A formal quality assurance program, with written instructions and formal documentation of its implementation, is required. Plant chemistry manuals should also reflect quality control provisions. Analytical procedures and analytical evaluation methods must be established to detect quality control problems early and correct them as soon as possible. When properly executed, a quality assurance program will provide measurements of the desired precision and bias, together with the associated documentation. A final point that must be made on cycle chemistry data quality is that in far too many fossil (28) plants this important issue has been given a very low priority . Changes in the industry to reduce staff size and, in many cases, experience levels have often resulted in situations where routine chemistry monitoring and control is performed by inexperienced personnel with no relevant education or training in chemistry. For organizations operating under the new paradigm, procedures and other measures designed to ensure chemistry data quality should be regarded as a high priority issue; implementation of suitable quality control and assurance provisions are essential to day to day operation of fossil units in which high levels of availability and reliability are required. E-38


E.7 REFERENCES 1.

Guideline Manual on Instrumentation and Control for Fossil Plant Cycle Chemistry. EPRI, Palo Alto, CA: April 1987. CS-5164.

2.

Monitoring Cycle Water Chemistry in Fossil Plants, Volume 3: Conclusions and Recommendations. EPRI, Palo Alto, CA: October 1991. GS-7556-V3.

3.

Development of a Steam Sampling System. EPRI, Palo Alto, CA, December 1991. TR100196.

4.

Cycle Chemistry Corrosion and Deposition: Correction, Prevention and Control. EPRI, Palo Alto, CA: December 1993. TR-103038.

5.

Reference Manual for On-Line Monitoring of Water Chemistry and Corrosion: 1998 Update, EPRI, Palo Alto, CA: 1999. TR-112024.

6.

Guidelines for Copper in Fossil Plants, EPRI, Palo Alto, CA: 2000. 1000457.

7.

Guideline Manual on Instrumentation and Control for Fossil Plant Cycle Chemistry, EPRI, Palo Alto, CA: 1987. CS-5164.

8.

American Society of Mechanical Engineers, Performance Test Codes Committee, PTC 19.11. “Steam and Water Sampling, Conditioning and Analysis in the Power Cycle”. PTC 19.11-1997.

9.

O. Jonas, L. Machemer and B. Dooley. “EPRI ChemExpert: Cycle Chemistry Advisor for Fossil Plants”. Sixth International Conference on Cycle Chemistry in Fossil Plants, EPRI, Palo Alto, CA: 2001. 1001363.

10.

M. Stoffel. “Quality Assurance in Power Cycle Monitoring”. Sixth International Conference on Cycle Chemistry in Fossil Plants, EPRI, Palo Alto, CA: 2001. 1001363.

11.

D. Meils. “On-Line Instrumentation QA/QC Experience at TVA’s Kingston Power Station”. Fifth International Conference on Cycle Chemistry in Fossil Plants, EPRI, Palo Alto, CA: 1997. TR-108459.

12.

D. Meils. “Chemistry On-Line Process Instrumentation QA/QC Practices: Are You Getting What You Paid For?” ASME Research Committee on Power Plant & Environmental Chemistry Meeting, September, 2001.

13.

ASTM D2777-96, “Standard Practice for Determination of Precision and Bias of Applicable Methods of Committee D19 on Water”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.01 (Water I), ASTM International, Philadelphia, PA. (Updated November 1998)

E-39


14.

Lloyd A. Currie, “Limits for Quantitative Detection and Quantitative Determination”, Analytical Chemistry, Vol. 40, No. 3, March 1968.

15.

ASTM D2777-96, “Standard Practice for Determination of Precision and Bias of Applicable Methods of Committee D-19 on Water”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.01 (Water I), ASTM International, Philadelphia, PA. (Updated November 1998)

16.

Standard Mathematical Tables and Formulae. CRC Press, Cleveland, OH 1982.

17.

American National Standard C85.1-1963 [Addendum A (1966); Addendum B (1972); standard now withdrawn], Terminology for Automatic Control. American National Standards Institute, NY, NY.

18.

ASTM D3764-92, “Standard Practice for Validation of Process Stream Analyzers”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.01 (Water I), ASTM International, Philadelphia, PA: March 1992. (Updated April 2001)

19.

ASTM D3864-96, “Standard Guide for Continual On-Line Monitoring Systems for Water Analysis”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.01 (Water I), ASTM International, Philadelphia, PA: January 1996.

20.

Monitoring Cycle Water Chemistry in Fossil Plants, Vol. 1 Monitoring Results, by A.F. Aschoff, D.M. Sopocy, D.T. Eglar, O. Jonas, J.K. Rice, C.C. Stauffer, and W.E. Allmon. EPRI, Palo Alto, CA: October 1991, GS-7556, Volume 1.

21.

ASTM D2777-85 , “Standard Practice for Determination of the Precision and Bias of Applicable Methods of Committee D-19 on Water”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.01 (Water I), ASTM International, Philadelphia, PA. (Updated November 1998)

22.

ASTM D5542-94, “Standard Test Methods for Trace Anions in High Purity Water by Ion Chromatography”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.01 (Water I), ASTM International, Philadelphia, PA: November 1994.

23.

ASTM D4517-85 (1994)e1, “Standard Test Method for Low-Level Total Silica in High Purity Water by Flameless Atomic Absorption Spectroscopy”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.02 (Water II), ASTM International ,Philadelphia, PA: January 1985.

24.

ASTM D4779-93, “Standard Test Method for Total, Organic, and Inorganic Carbon in High Purity Water by Ultraviolet (UV) or Persulfate Oxidation, or Both, and Infrared Detection”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.02 (Water II), ASTM International, Philadelphia, PA: November 1993.

E-40


25.

J.K. Rice, G.T. Upperman, A.F. Aschoff, R.B. Dooley, “Interlaboratory Performance of Methods of Analysis for Trace Substances in High Purity Water”, Ultra Pure Water, Vol. 5, No. 5, p. 36 (July/Aug. 1988).

26.

ASTM D3370-95, “Standard Practices for Sampling Water from Closed Conduits”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.01 (Water I), ASTM International, Philadelphia PA: November 1995.

27.

ASTM-D4453, “Standard Practices for Handling of Ultra-Pure Water Samples”, Annual Book of ASTM Standards: Water and Environmental Technology, Volume 11.01 (Water I), ASTM International, Philadelphia PA: August 2002.

28.

D.E. Meils, Power Plant Chemistry QA/QC, a Practical Application, Seventh International Conference on Cycle Chemistry in Fossil Plants, EPRI, Palo Alto, CA: 2004. 1009194.

E-41


F UNIT SHUTDOWN, LAYUP, STARTUP, CYCLING AND PEAKING

Initial efforts by EPRI to establish cycle chemistry guidelines for fossil units emphasized normal unit operations, including baseload operation, load cycling, and peaking service. Implementation of these guidelines in fossil plants around the world led to significant improvements for many organizations. At the same time, however, it also became apparent that the chemistry must be properly monitored and controlled during shutdown and layup, and that doing so was the only way in which organizations applying the guidelines could attain all of the cycle chemistry program goals (Section 1). EPRI benchmarking activities further reflect the relationship between effective shutdown, layup and startup procedures and the ranking of individual cycle chemistry programs. In the absence of proper shutdown and layup procedures, the cycle is vulnerable to contamination and corrosion activity. For example, during off-load conditions, deposited impurities may have been released into the circuit or have entered from outside, e.g. condenser leaks, carbon dioxide (CO2) with air inleakage. Therefore, it is important to consider the effects on the feedwater and steam cation conductivity, silica, sodium, chloride and sulfate concentrations during startup. If there is a leak in the reheaters, CO2 and, more importantly, flyash and sulfur dioxide, (SO2), may be drawn in, resulting in acidic condensed steam in the reheaters. There is some evidence that silica is washed off turbines during shutdown and startup. Oxygen concentrations in the feedwater require special consideration. Restriction will need to be applied to this parameter, but some relaxation from the normal EPRI guideline values should be possible until steady state operating conditions are achieved. During unit startup, the effect of chemistry control impacts on the potential for fluid-side deposits, corrosion and impeding of heat transfer (furnace tube metal overheating). Focus should be directed towards the identification and correction of the root cause of the chemical control problem.

F-1

EPRI Licensed Material Unit Shutdown, Layup, Startup, Cycling and Peaking

F.1 SHUTDOWN The shutdown periods should be viewed as a continuum of the good practices used during normal operation. For mixed-metallurgy systems, it is necessary to maintain reducing conditions in the condensate and feedwater. Having a negative ORP near the optimum value of -300 to -350 mV best ensures this and will maintain the protective cuprous oxide (Cu2O) layer on the surface of copper-based feedwater heater tubes. The shutdown period is a prelude to the layup of the unit. Layup procedures are determined by the duration of the layup, whether short or longer term. The use of nitrogen blanketing the shell side of feedwater heaters to protect feedwater heaters is presently practiced by only 13% of the (1) recently surveyed utilities . Nitrogen blanketing serves an essential purpose in preventing oxygen ingress during short or longer-term layups. Nitrogen blanketing is an especially significant topic, since it plays an important role in the control of copper in fossil fired power plants. In order to produce an effective nitrogen blanket for oxygen exclusion from electricity generating units, comprising boilers, turbines, heaters, and all interconnecting systems, both the appropriate nitrogen supply/delivery system and the appropriate application procedures are required. The final key to success is to follow the established procedures implicitly at the time the application is needed. Nitrogen supply systems need to be large enough to sufficiently supply nitrogen to steam systems while pressure is collapsing during cooling in order to prevent any sub-atmospheric condition from occurring; and, to preclude any air ingress into the systems being blanketed. If cryogenic storage systems are used, they should be of about twice to three times the gaseous volume of all the equipment that is being protected per application. The vaporizing equipment must be sized to keep pace with the evacuating demand produced by the cooling and draining of the various pieces of equipment. Membrane systems are also commercially available to generate nitrogen on-site. These units are typically skid-mounted and can easily be moved from place to place on the site. These units are available in a wide variety of sizes to accommodate most requirements of nitrogen blanketing. The units use compressed air as an inlet stream. Nitrogen purity ranges to 99.9%, and is a function of the air pressure to the unit. As the purity requirements increase, the capacity to generate nitrogen decreases. The nitrogen which is produced is very dry, with a dew point of –90 F. The units are simple to operate with no moving parts. Membrane systems may be feasible at a cost savings over cryogenic systems. However, due consideration must be given to final purity of the product gas, volume of gas production and the operating power requirements and associated auxiliary power costs. There is little information in the literature relative to the required purity of nitrogen gas to be used for blanketing. Oxygen is the largest impurity in commercial or on-site generated nitrogen. This oxygen has a great propensity for dissolving in water, particularly deoxygenated water, which will enhance corrosion to metal surfaces.

F-2


Based upon considerable experience with nitrogen blanketing systems, the suggested minimum purity for nitrogen used for blanketing purposes is 99.5%. This agrees well with the practice of feedwater heater manufacturers which use 99.6% purity nitrogen when shipping their heaters. Nitrogen supply piping must be connected separately to all circuits and components that are normally isolated by shut-off, stop or stop/check valves during normal shutdown. Exception to this can be made if non-routine valve positioning is explicitly stated in the blanketing procedures and is accomplished during the application process. This is to ensure that all parts of all circuits can be blanketed with nitrogen. All nitrogen piping connections must be per code requirements of the component or circuit involved. Connections can be made permanently, with welded connections and appropriate isolation and check valves; or, using the quick connect type that can be disconnected to prevent blanketing by accident if the unit is going to be inspected instead of blanketed. The valves on the permanent type connections can be of the manual type or (preferably) can be motor-operated in order to automate the blanketing process. These valves can be wired into the control system so as to admit nitrogen as soon as the specified pressure set point for that component or circuit is reached. If the system is manual, care must be taken to monitor the pressures and to valve-in the nitrogen supply to that component at the appropriate time in order to prevent sub-atmospheric pressure from occurring, which would nullify the effectiveness of the nitrogen blanket. Shutdown procedures that previously had been established for the unit will need to be reviewed and revised as necessary to facilitate application of the nitrogen blanket. The objective is always to replace the collapsing steam with nitrogen to prevent air ingress into the system, including the condenser. As the turbine comes off line, the condenser and reheater need to be flooded with nitrogen to achieve positive pressure as fast as possible before shutting down the air removal system. Care must be taken not to violate any turbine cooling requirements and to not over-pressure the condenser, which may cause the rupture discs to blow. The extraction system, which is frequently neglected, needs as much attention as any other system. The heater shells and drains need appropriate blanketing. The deaerating heater and its associated storage tank can be pegged with nitrogen as the steam pressure collapses. The superheater vents and drains should be open and nitrogen admitted at such pressure as necessary to effect a purge prior to isolation. The boiler should be drained under nitrogen pressure if draining is desired; or, capped under a continuous nitrogen pressure if it is to remain full. The boiler can be “flash drained” (drained while above boiling temperature) to produce a dry condition, all the while under nitrogen pressure, if such a condition is desired.

F-3


If the condensate, feedwater and associated heaters are to be drained, this should be accomplished under positive nitrogen pressure. A slight positive pressure should be maintained on the isolated system once the drain is accomplished. If these systems are to remain full of condensate, a positive nitrogen pressure must exist at all times to prevent any air ingress and to replace any water that may have leaked from seals, flanges, etc. If the unit being blanketed contains any non-drainable components or circuits, these must be identified and an appropriate nitrogen supply made available to them in order to have an effective nitrogen blanket. Once all systems are drained and blanketed, all vents and drains should be closed. Then, a slightly positive nitrogen pressure of 2-5 psig (13.8–34.5 kPa) should be maintained on all components. In the case of the main condenser, once the steam seals are off and the air removal systems are stopped, a positive flow of nitrogen should continue to the condenser. If a slight pressure can be maintained without excessive nitrogen consumption, this procedure is feasible. Again, care must be taken not to blow the rupture disks. A continuous small flow through the condenser and out the turbine seals is totally sufficient to keep air out of that area in almost any situation. Cost effective and physically effective nitrogen blanketing of steam electric generating systems can be accomplished. There are certain prerequisites that cannot be circumvented if effective blanketing is to be the desired result. The nitrogen supply system has to be capable of delivering the appropriate volume and purity in the required time frame. The nitrogen application points must be located in all necessary components and circuits. The nitrogen blanket application procedures, as well as unit shutdown procedures must address all of the necessary conditions and situations. Finally, the performance of the process must be exacting in order to be successful to accomplish the ultimate objective, which is to exclude oxygen from the equipment during downtime and lay-up.

F.2 LAYUP The following guidelines will give desirable results for iron and copper control during short and longer-term layups. Additional information can be found in EPRI’s Startup, Shutdown and Layup Guidelines.(3)

F-4


F.2.1 Influence of Cycle Metallurgy on Layup All-Ferrous Metallurgy Systems Adequate layup practices for layup of idle equipment fall into two general categories: the wet and the dry procedures. Wet layup requires filling of most of the system with an alkaline solution (ammonia) and preventing air ingress by pressurization with nitrogen. In units operated with a feedwater ORP of >0, reducing agents (hydrazine or alternates) should not be used for wet layup. Dry layup requires drainage while hot, and removal of all water followed by pressurization with a moisture-free inert gas or by use of dehumidified air to maintain a low-moisture environment. Use of wet layup solutions in units with all-ferrous feedwater systems on AVT(O) or OT should generally be restricted to short term protection. Extended wet layup without a reducing agent (as is optimum in these cycles for short term protection) could lead to pitting attack unless the equipment is reliably nitrogen capped at all times and there is a means of circulating the solution. In view of these concerns, dry layup methods are clearly preferred for units with all-ferrous feedwater metallurgy and AVT(O) or OT feedwater chemistry. Conversely, if the normal feedwater chemistry is AVT(R) extended wet layup solutions containing a suitable reducing agent can be used effectively. Mixed metallurgy systems Adequate layup practices for layup of idle equipment fall into two general categories: the wet and the dry procedures. Wet layup requires filling of most of the system with an alkaline reducing solution (ammonia and hydrazine) and preventing air ingress by pressurization with nitrogen. It is emphasized that the reducing conditions have to be maintained not only during normal operation but also in all other situations (shutdown, layup, startup). Dry layup requires drainage while hot, and removal of all water following by pressurization with a moisture-free inert gas or by use of dehumidified air to maintain a low-moisture environment. F.2.2 Short-term Layup Short-term layup presumes that the unit will be required to operate again within a relatively short time frame. Therefore, no major changes are required from normal operating conditions, with the exception being that the unit must be protected from air ingress. The condenser vacuum and turbine seals should be maintained. The deaerator and heater shells are inerted with nitrogen, and the feedwater chemistry is maintained as before shutdown.

F-5


F.2.3 Intermediate and Longterm Layup The two methods, which are available for intermediate and longterm layup, are the wet and the dry procedures. In the dry procedure, all components of the condensate and feedwater systems must be drained hot. Dry air or nitrogen is circulated through the equipment to remove residual moisture and to prevent subsequent corrosion of the components of the cycle. The preferred wet procedure, only suitable for units with feedwater on AVT(R), involves filling the various components with deoxygenated demineralized water containing ammonia and hydrazine. Reducing agents should not be used for layup of units with feedwater on AVT(O) or OT. For units containing copper alloys, the pH of the feedwater should be limited to 9.0 to 9.3; for all-ferrous systems 9.2 to 9.6. Hydrazine is held to 40 to 50 ppm. The pH in mixedmetallurgy systems should not be allowed to rise, because increasing pH increases the solubility of copper oxides. A separate chemical purging system may be required for this purpose. Wet layup methods generally should not be used for intermediate or longterm layup of units with feedwater on AVT(O) or OT feedwater treatment. Dry layup is preferred. Nitrogen blanketing is preferred for all wet layups (Figure F-4). As the turbine spins down on shutdown, nitrogen is added to the condenser, which also purges the reheater system. A continuous purge of nitrogen is necessary to account for leakage through the turbine steam seals. Nitrogen is added to the deaerator and storage tank while the system is still hot. A small flow of nitrogen is provided to purge this system. A nitrogen cap is maintained on the shell side of the feedwater heaters. F.2.4 Maintenance Outage Only equipment requiring maintenance should be drained (wet storage only), and nitrogen (if used) purged with air to provide an environment suitable for entry of personnel. Since nitrogen does not support human life, it is extremely important that it be completely displaced by air. The atmosphere within the equipment to be maintained should be tested with suitable test equipment to ensure the equipment is safe for personnel entry.

F.3 STARTUP Again it is necessary to prevent oxygen ingress, if iron and copper are to be controlled to acceptable levels. The use of deaerated makeup will appreciably reduce oxygen levels during startup (see Appendix A). Air in-leakage must be controlled, as additional parts of the heat cycle are under vacuum during low load operation (see Appendix C). The use of auxiliary steam to peg the deaerator (if available) will enhance oxygen removal at this point in the cycle. Detection and repair of condenser leaks during startup is critical to prevent ingress of harmful contaminants such as chlorides and sulfates.

F-6


The use of a cleanup loop, in conjunction with condensate polishers and/or filters (if available) was considered in the 1980s to be helpful in reducing contaminants during startup, however they are quite expensive to back-fit; and, therefore are not in widespread usage today. There have been some back-fit installations made for those units which have been converted from base loaded to cyclic operation, and, therefore are expected to have frequent startups. Even without cleanup loops, condensate polishers and/or condensate filters can markedly decrease startup times. F.3.1 Corrosion Product Filters A recent development, condensate filters for copper(2) shows promise for markedly reducing copper in feedwater. Results from the application of such a unit are shown in Figure F-1. Field testing of this copper-reducing filter, as reported, indicates that copper can be reduced in feedwater by more than 95%, from 600 ppb to less than 20 ppb. Iron concentrations are also said to be reduced. 1200

Copper Concentration (ppb)

1000

800

Inlet copper concentration

600

400

Outlet copper concentration

200

0 0

200

400

600

800

1000

1200

1400

1600

Run Time (minutes) Figure F-1 Copper Concentration at Virginia Power’s Chesterfield Unit 6 Before and After a Filter(3)

F-7


The copper filter is a design derived from filters already being applied in condensate filtration for particulate iron removal. The filter is a cartridge type utilizing a proprietary melt blown filter media that has a unique affinity for copper. It has no leachables, different from other plastic filter media of similar construction. Maximum operating temperature is 180°F (82°C), and maximum (4) differential pressure is 50 psi. Other design features are shown in Table F-1 . After the initial tests at Chesterfield Station, the copper filter was fitted with an improved cartridge design, and retested at Chesterfield with improved results(5). Additional tests were performed at Pacificorp’s Naughton Plant using the improved cartridge design. As a result of the satisfactory test results, a full set of copper reducing elements were ordered for installation at (7) Naughton . A primary limitation of condensate filters is that they cannot remove iron and copper which enter the feedwater. (Condensate polishers exhibit comparable filtration capabilities and limitations.) Development of filter elements capable of working in the feedwater environment would increase the usefulness of their devices. However, it should also be kept in mind that the need for corrosion product removal is minimized when the chemistry is controlled to minimize corrosion of materials in the cycle. F.3.2 Units With Drum Boilers In general and to the extent practiced, it may be desirable to discard heater drains during initial startup operations (first 2-4 hours after turbine synchronization) to reduce the amount of iron and copper entering the boiler. Initial startup depends on the layup procedures followed prior to startup. If the layup was shortterm, and the unit filled with water treated for typical normal operations, startup can proceed without delay. If the feedwater system was full of water treated for longer-term layup (pH 9.0 to 9.6, hydrazine 40-50ppm) some water may have to be dumped to reduce hydrazine and ammonia to more appropriate values. If the unit was laid up dry, the feedwater system for mixed-metallurgy systems must be filled with deaerated water from the makeup system treated with ammonia to pH 9.0-9.3, and hydrazine to achieve an ORP in the range -300 to -350mV, prior to startup. For all-ferrous systems, pH would be 9.2 to 9.6. During startup, the boiler feedwater should be closely monitored. The following limits for the boiler feedwater have been used prior to firing the boiler:

•

Oxygen < 100 ppb

•

Iron < 100 ppb

•

Copper < 10 ppb

F-8


By following the guidelines and procedures in Sections 4 and 5 of this document, it is expected that these limits can (and should) be reduced. Attainment of these prefiring limits will be more difficult in units not equipped with condensate polishers and / or filtration equipment. Best control will require close compliance with the guidelines for unit shutdown and layup. However, for mixed-metallurgy systems the Program (6) Copper research has indicated that transition of protective cuprous oxide will occur to nonprotective cupric oxide in a matter of hours in oxygenated water. Thus it is important to quickly reach ORP levels between –350 and –300 mV for these systems.

F-9

EPRI Licensed Material Unit Shutdown, Layup, Startup, Cycling and Peaking Table F-1 Design features of copper filter(4) Sizes Filter Diameter (in/mm)

Filter Length (in/mm)

Max. Flow (gpm/lpm)

6 / 152.4

40 / 1016

83 / 321

6 / 152.4

60 / 1524

130 / 492

Housings Number of Filters

Aqueous Rated Flow (gpm/lpm)

1

130

2

260

3

390

4

520

7

910

12

1560

19

2470

After firing the boiler, monitoring of the condensate and feedwater must be continued, both for chemistry parameters and to ensure that temperature ramps are maintained according to equipment manufacturer’s specifications. Boiler water chemistry concentration levels vs. pressure curves (see Sections 4 and 5) can be utilized to control sodium, silica, chlorides and sulfates to provide protection to the steam turbine. Boiler pressure should remain at reduced levels such that these contaminant values, as shown on the curves, are not exceeded before pressure can be increased to the next level. Maximum use of blowdown and condensate polishing (if available) will minimize startup times. Several organizations with mixed-metallurgy systems have established boiler water “holds” for copper during startups, similar to the philosophy established very much earlier for “holds” on (7) silica boiler water levels during startups. One such startup curve is shown in Figure F-2 . The first version was established in 1987, but was modified in 1993 in an effort to shorten unit startup times. The revised curve for 1993 reduced startup times from 48-60 hours (1987) to about 12 hours.

F-10


Figure F-2 Boiler Water Copper Concentration/Drum Boiler Pressure Control Curves Developed at Miami Fort Station(7)

Figure F-3 shows a boiler water copper control curve derived for a 2 ppb copper limit in steam, (6) which was developed for the EPRI Copper Guidelines. Since it is difficult to quantify analytically the individual species (cupric or cuprous) in the boiler water, it was suggested that total copper be used for control purposes. The curve shown in Figure F-3 is based on the volatility of cuprous oxide, and could be used to provide an approximation of the total copper level in the boiler water which should minimize the copper in steam. The startup curves in Figure F-2 are similar to the total copper curve of Figure F-3 in the higher operating pressure ranges, as shown in Table F-2. Of course it should be recognized that use of a curve implies that the good control philosophies for boiler water, which ensure reducing conditions are maintained during a shutdown, have not been adopted. A key example may include a maintenance outage to work or internally inspect the boiler. On the other hand, if reducing conditions are rigorously maintained then such a figure should not be needed. In this regard, it can not be emphasized more strongly that, according to the Program Copper research (Sections 1.2.4 and 3.3.1), the copper oxides will partition into (6) steam from boiler water at high levels even during the early startup. Applying these guidelines to control the chemistry under all conditions is the best approach to avoidance of copper transport problems.

F-11


Drum Pressure (MPa) 4.8

6.2

7.6

9.0

10.3

11.7

13.1

14.5

15.8

17.2

18.6 19.6

2000

1000 800 600 500 400 Total Copper (ppb) in Boiler Water

300 200

100 80 60 50 40 30 20 Normal 10 8 6 5 4 3 600 700

900

1100 1300 1500 1700 1900 2100 2300 2500 2700 2850

Drum Pressure (psia) Based on 2 ppb copper in steam Figure F-3 Drum Boiler Water Copper vs. Operating Pressure

F-12

EPRI Licensed Material Unit Shutdown, Layup, Startup, Cycling and Peaking Table F-2 Comparison Of Copper Values in Blowdown (Figures F-2 and F-3) PRESSURE psi (MPa)

FIGURE F-2 (ppb)

FIGURE F-3 (ppb)

1993 DATA(1)

1987 DATA

1950 (13.4)

275

110

85

2000 (13.8)

160

100

75

2100 (14.5)

140

75

65

2200 (15.1)

100

70

50

2300 (15.8)

75

50

45

2400 (16.5)

60

45

35

2500 (17.2)

40

40

30

2600 (37.7)

25

25

25

During startup, any chemical excursions must be dealt with quickly and effectively to return the unit expeditiously to normal operating values. Target values at the deaerator inlet and the economizer inlet are given in Figures 4-1 and 4-2.

F.4 ROAD MAP FOR SHUTDOWN AND LAYUP A road map to develop shutdown and layup procedures is shown in Figure F-4(3). Because of variations in design, some generating units may require deviations to Figure F-4, or may require customization to adapt to a particular utility’s needs. Figure F-4 is divided into nine steps, which are further described as follows: Step 1 - Short-Term Layup

Short-term layup presumes that the unit will be required to operate within a relatively short timeframe. In consideration of this, no major changes are required from normal operating conditions, with the exception being that the unit must be protected from air ingress. The condenser vacuum and turbine seals are maintained; the deaerator, heater shells and boiler are inerted with nitrogen or steam; and, the feedwater chemistry is maintained according to the requirements of the treatment philosophy employed.

F-13

EPRI Licensed Material Unit Shutdown, Layup, Startup, Cycling and Peaking Step 2

Step 1

Short term shutdown

No

Intermediate and long term

Yes

Inert the deaerator and heater shells

Yes Step 6

Step 5

Low O2 scavenger

Traditional Yes

Break reheater vacuum with Nitrogen purge

Inert the boiler with automatic system

No

Dry layup

Yes Step 3

Add 5-10 ppm Hydrazine when boiler pressure decays to 200 psi (1.4 MPa)

Fill feedwater system with 200 Hydrazine* 10 ppm Ammonia**

Add Nitrogen cap when boiler pressure decays to 5 psi (0.03 MPa)

Establish boiler Nitrogen cap of 5 psi (0.03 MPa)

Maintain feed water without change

No Step 4

No

Yes

Fill boiler with 10 ppm Ammonia; and up to 200 ppm Hydrazine*

Drain condenser under Nitrogen

Maintain chemical limits per guidelines

Drain system to remove all water

Yes

Drain system to remove all water

Wet layup

Yes

Evacuate reheater with condenser vacuum

Maintain condenser vacuum and turbine seals

Dry air

No

Pressurize with N2 all wetted parts

Maintain small Nitrogen flow through condenser, turbine and deaerator Notes: *No Hydrazine for AVT(O) or OT units **Limit pH to 9.0 to 9.3 if units have copper alloys in cycle; maintain Hydrazine at 40-50 ppm

Back fill superheater 200 ppm Hydrazine* 10 ppm ammonia. Nitrogen cap Step 8

Very long term storage Yes

Isolate reheater

Backfill reheater and superheater with 200 ppm Hydrazine* 10 ppm Ammonia Nitrogen cap

Step 7

Add Nitrogen to condenser while turbine spins down. Maintain slow N2 flow.

Add Nitrogen to deaerator and storage tank while still hot. Maintain slow N2 flow

Step 9

For maintenance: purge with air all N2 from equipment to be maintained. Test to ensure safe environment.

Maintain Nitrogen cap on shell side of feedwater heaters

Figure F-4 Road Map to Develop Shutdown and Layup Guidelines Common to Most Units Notes: * No hydrazine for AVT(O) or OT units. ** Limit pH to 9.0 to 9.3 if units have copper alloys in cycle; maintain hydrazine at 40–50 ppm

F-14


Step 2 - Intermediate and Longterm Layup Common to Dry and Wet Layup

Intermediate and longterm layup require additional steps to be taken to prevent corrosion during intermediate periods of layup, such as for maintenance and indeterminate cycling or peaking requirements; and during longterm layup for indefinite periods of time. Certain procedures are common regardless of whether the units are to be laid up dry or wet. There is a danger of offload corrosion (pitting) if there are salts present. Consideration needs to be given to whether the reheater supports can take the weight if the reheater is to be filled with water. During shutdown, the turbine, condenser (steam side) and reheater are generally considered together because, unless special facilities are incorporated, there is no practical way to isolate them. With special facilities incorporated, the reheater can be isolated from the turbine and may be stored wet (see Step 8). The turbine, however, can only be laid up dry. The condenser may be flooded and laid up wet, but several factors limit the feasibility of this procedure:

•

The condenser must be supported from the bottom to handle the extra weight of a flooded condenser.

•

The expansion joints between the turbine and condenser are not designed to support the weight of a flooded condenser.

•

A flooded condenser tends to pull the turbine bearings out of alignment.

•

Water in the condenser may cause moisture vapor to enter the turbine causing corrosion.

For these reasons, the steam side of the condenser is normally laid up dry. For Step 2, during shutdown, the reheater is evacuated by utilizing the vacuum in the condenser. The vacuum is then broken using nitrogen pressure. The condenser is drained under nitrogen. If the turbine steam is supplied from a header system, all valves must be tight to prevent moisture entrance into the turbine. Step 3 - Dry Air Layup

The dry air layup procedure requires that all components of the system be drained. There are several methods of maintaining dry air flow through the equipment. The advantages and disadvantages of dry air layup and the equipment required have been previously described in other EPRI documents(3,6). Step 4 - Dry Layup with Nitrogen

Nitrogen can be used for blanketing equipment, which is drained but not completely dry, or for blanketing equipment either filled with water or not, to prevent air ingress. Step 4 is similar to Step 3, except nitrogen is used for a positive pressure on all components rather than a dry air purge. A small continuous purge of nitrogen is required, however, to protect the turbine and deaerator storage tank.

F-15


Step 5 - Wet Layup: Traditional Method (Boiler and Feedwater Heaters)

For those units having copper alloy condenser tubes and/or feedwater heaters, feedwater pH should be limited to 9.0 to 9.3 by reducing ammonia. Hydrazine is held at 40-50 ppm. For these units, it is most important to maintain reducing conditions (ORP < 0 mV and preferably to -350 mV) to prevent increased attack of the copper alloys associated with a change of the surface oxide layers from cuprous to cupric oxide. A nitrogen cap of 5 psi (0.03 MPa) is maintained on the boiler. Step 6 - Wet Layup: Low Oxygen Scavenger Method (Boiler and Feedwater Heaters)

Many alternatives to the traditional wet layup method have been used successfully by markedly reducing the level of hydrazine. These procedures may not require boiler draining prior to startup, thereby protecting the environment and not requiring special disposal techniques or permits. In one procedure, 5 to 10 ppm of hydrazine equivalent is injected into the boiler when the boiler pressure decays to 200 psi (1.4 MPa) (typically 3 days). The chemical injection is made using the normal chemical feed system. Natural boiler circulation at 200 psi (1.4 MPa) is sufficient to mix adequately the chemicals in the boiler water. No other changes need be made to the boiler or to the feedwater chemistry. When the boiler pressure decays to 5 psi (0.03 MPa) (typically 7 days) a nitrogen cap is applied to the boiler. Upon return to service, the boiler may be fired without draining the layup solution. Also, since it typically take 3 days before the boiler pressure decays to 200 psi (1.4 MPa), weekend outages normally do not require a chemical injection treatment. Step 7 - Wet Layup (Balance of Cycle)

As the turbine spins down, nitrogen is added to the condenser, which also purges the reheater system. A continuous purge of nitrogen is necessary to account for leakage through the turbine steam seals. Nitrogen is added to the deaerator and storage tank while this system is still hot. A small flow of nitrogen is provided to purge this system. A nitrogen cap is maintained on the shell side of the feedwater heaters. The superheater is back filled with a solution of up to 200 ppm of hydrazine and 10 ppm ammonia and a nitrogen cap maintained. Step 8 - Very Long Storage

For very long storage periods, the reheater is isolated from the turbine, and back filled with a solution of up to 200 ppm hydrazine and 10 ppm ammonia, and capped with nitrogen. The superheater is treated similarly.

F-16


Step 9 - Maintenance

Only equipment requiring maintenance should be drained (wet storage only), and nitrogen (if used) purged with air to provide an environment suitable for entry of personnel. Since nitrogen does not support human life, it is extremely important that nitrogen is completely displaced by air. The atmosphere within the equipment to be maintained should be tested with suitable test equipment to ensure the equipment is safe for personnel entry. Experience indicates that for mixed-metallurgy systems unit startups in the past have been a major source of copper transport to the boiler, with concentrations as high as 10 ppm during cold starts. By following the preceding startup guidelines, especially for oxygen control and air inleakage, copper transport can be minimized during these periods.

F.5 CYCLING AND PEAKING Cycling is a load following operation. The unit load fluctuates with system demand, with the unit synchronized at very low loads during low-demand periods. A typical load variation for cycling units might range from 30% to 100% of design capacity. Peaking is a form of cycling in which the unit is operated only during peak power demand periods. At off-peak hours the unit is on hot standby, depending upon the estimated time between restarts. Two-shift operation is typical of peaking units, which generally furnish power for the morning and evening high demand hours. For guidance, Table F-3 gives suggested cumulative hours per year of exposure to contaminant conditions for cycling operations. Table F-3 Maximum Annual Exposure to Contaminant Conditions for Cycling and Peaking Operation

TARGETS

HOURS PER YEAR

NORMAL

-

ACTION LEVEL 1

672

(28 days)

ACTION LEVEL 2

96

(4 days)

ACTION LEVEL 3

16

IMMEDIATE SHUTDOWN

2

F-17


Organizations attempting to apply the criteria to peaking units are challenged by the chemistry variations associated with shutdown and startup periods, especially if no provision is made for equipment layup. For “chemistry accounting” purposes, it is generally accepted that accrued times in action level conditions should only be counted while the unit is in service at or above minimum load at which the unit is normally operated to supply power to the system. However, this approach does not lessen the value and importance of optimizing cycle chemistry under all operating conditions. Some concerns relative to cycling and operation are given below. An * indicates the items affected by water and steam chemistry:

•

increased boiler component cyclic stress with loss of unit life*

•

increased turbine rotor cyclic stress with loss of rotor life*

•

increased thermal stress on turbine rotor, steam chests, valves, and inner casing,* leading to corrosion fatigue

•

increased solid particle erosion of HP and IP turbine blading from superheater and reheater tube exfoliation

•

turbine vibration during startup/shutdown*

•

possibility of furnace implosion/explosion during startup/shutdown

•

steam requirements for maintaining condenser vacuum

•

the need for improved boiler control to minimize unit restart and reloading times*

•

acid dew point condensation- corrosion and plugging of air preheaters and back end ductwork

•

flame scanner monitoring problems requiring much attention

•

burner turndown ratio/flame stability at low loads- feeder turndown and mill response times during load ramps too slow

•

increased frequency of chemical cleaning and more monitoring and control of steam and water chemistry needed*

•

poor back end equipment performance at low load conditions

•

increased possibility of mill fires and explosions during frequent starting and stopping of a unit

•

distortion and internal rubbing of the boiler feed pump, if pump is not fully warmed up at startup

F-18


Some of the possible solutions to these problems are given below. Many of these solutions require considerable expenditures. An * denotes possible effects on cycle water chemistry:

•

incorporate steam bypass systems*

•

incorporate full arc admission

•

add additional controls and monitoring equipment*

•

incorporate variable pressure operation*

•

use two half-sized boilers

•

install turning gears on ID fans

•

incorporate a condensate polishing system*

•

incorporate better turbine seals

•

install smaller coal mills for low load operation

Additional detailed information on this topic can be found in the EPRI’s Cycling, Startup, and Layup Guidelines.(3) and in a recent ASME publication(8).

F-19


F.6 REFERENCES 1.

“Survey of 63 Utilities”. Fifth International Conference on Fossil Plant Cycle Chemistry. EPRI Proceedings. Edited by B. Dooley and J. Matthews TR-108459. December 1997.

2.

K. Shimmoller, “Plant O&M: Taking the Long View”. Power Engineering, February 2000.

3.

Cycling, Startup, and Lay-up Fossil Plant Cycle Chemistry Guidelines for Operators and Chemists. EPRI, Palo Alto, CA.: August 1998. TR-107754.

4.

R. Friis, Pall Corporation, Personal Communication to A. Aschoff, August 2000.

5.

K. Laird, D. Beard, E. Conners, J. Canil, R. Stoyell, Jr., and R. Gieger, “Copper th Reduction in Power Plants,” presented at the 8 Annual Southwest Chemistry Workshop, August 1999.

6.

Guidelines for Copper in Fossil Plants. EPRI, Palo Alto, CA: 2000. 1000457.

7.

M.L. Hoffman, “Concerted Utility Efforts Tames Copper Deposition,” Power, June 1994.

8.

Consensus for the Layup of Boilers, Turbines, Turbine Condensers and Auxiliary Equipment”, ASME, New York, NY: 2002. CRTD-Vol. 66.

F-20


G CALCULATION OF THE pH CORRECTED FOR THE PRESENCE OF AMMONIA

The sodium to phosphate ratio is defined as: R = [Na]/[PO4]

(eq. G-1)

where [Na] = [NaOH] + [NaPO4 ] is the molar concentration of sodium balanced by hydroxyl and phosphate anions, and [PO4] is the total molar concentration of phosphate. Note that the sodium balanced by other anions, such as chloride and sulfate, is not counted in [Na]. If the small concentration of the hydrogen ion is neglected, ammonium remains the only other cation balancing the hydroxyl anions, so that [NaOH] = [OH−] – [NH4+], and R can be written: [OH - ] - [NH 4 ] + [Na PO4 ] R = [PO 4 ]

(eq. G-2)

The determination of [NaPO4], the amount of sodium necessary to balance the phosphate anions, requires the knowledge of the speciation of phosphate taking into account all three dissociation constants and all four species resulting from the three-step dissociation of orthophosphoric acid: 23H3PO4, H2PO4 , HPO4 , and PO4 . Such a calculation, arithmetically rather cumbersome, is necessary to obtain a general solution valid in the entire range of pH from 0 to 14. However, it can be shown that in the range of pH relevant for phosphate treatment chemistry in samples of drum boiler water cooled to ambient temperature, practically all the phosphate is present as the 2HPO4 anion. This fortunate simplification is due to large differences between successive dissociation constants of the tri-basic acid (4 to 5 orders of magnitude). Therefore, it can be assumed that solid trisodium phosphate dissociates according to the reaction: Na3PO4

→ (2 Na+ + HPO42- ) + ( Na+ + OH−) .

(eq. G-3)

The effect of reaction 3 is a contribution of one mole of NaOH by each mole of phosphate present in the solution. In other words, 1 ppm additional phosphate as the trisodium salt contributes the same increase in alkalinity as 0.421 ppm additional NaOH. It should be emphasized, however, that this is only true for samples of boiler water at room temperature, when pH is not much less than 9 or greater than 10.

G-1

EPRI Licensed Material Calculation of the pH Corrected for the Presence of Ammonia

Using equation G-3 the amount of sodium associated with phosphate as [NaPO4 ] = 2 [PO4] can immediately be identified, and the ratio R (eq. G-2) can be expressed as a function of pH by the following equation: R =

10 pH − pK w - [NH +4 ] + 2 [PO4] [PO 4 ]

(eq. G-4)

+ -14 where Kw = [H ][OH−] = 10 (pKw = 14) at 25 °C (77 °F) is the ion product of water.

The last step remaining before obtaining the final relationship between R and pH is to determine the concentration of the ammonium cation. It may be tempting to assume that ammonia is completely dissociated, and each mole of ammonia contributes one mole each of hydroxyl and ammonium ions. Unfortunately, the concentration of the ammonium ion varies with the total concentration of ammonia and with pH, and so it has to be calculated using the dissociation constant of ammonia as discussed below. Since ammonia is a weak base, the equilibrium constant of the reaction: NH3 + H2O

NH4+ + OH–

(eq. G-5)

proceeding from the left to the right side is relatively small; at 25 °C (77 °F):

Kd

≈

[ NH 4+ ][OH − ] = 1.8 ⋅ 10 -5 , [ NH 3 ]

pK d = 4.74

(eq. G-6)

The small value of K d does not mean that ammonia is always far from completely dissociated + and that the concentration of the ammonium cation NH4 is always small compared to the total concentration of ammonia. In fact, at sufficiently low concentration the dissociation is indeed practically complete. However, quantitatively it turns out that “sufficiently low concentration” means much less than the ammonia levels relevant for drum boilers of power plant steam generators. As a rule ammonia is partly dissociated in boiler water samples at ambient temperature.

G-2


To correctly calculate the actual dissociation degree of ammonia in dilute solutions as a function of concentration, it is necessary to solve the system of four equations:

[ NH 4+ ][OH − ] [ NH 3 ] [H + ][OH − ] =

=

Kd

Kw

(eq. G-7)

(eq. G-8)

[ NH 4+ ] + [ NH 3 ] = m 0

(eq. G-9)

[H + ] + [NH 4+ ] = [OH − ]

(eq. G-10)

with four unknown concentrations in square brackets. The solution of this system leads to a third degree polynomial. If it is assumed that the concentration of hydroxyl ions coming from selfdissociation of water is very small in comparison to those coming from dissociation of ammonia, + then [OH−] = [NH4 ] and the system can be reduced to two equations (eq. G-7 and G-9) leading to a second degree polynomial. This simplified solution, most often presented in textbooks, is accurate if the dilution is not extreme. Figure G-1 plots the concentrations of the aqueous ammonium ion and ammonia as a function of total ammonia obtained as a solution of equations G-7 through G-10. It shows that at about 0.6 ppm ammonia the degree of dissociation is about 50%. The enlargement of the dilute region shows that at concentrations greater than about 50 ppb (0.05 ppm) the dissociation of ammonia can not be considered complete even for approximate practical considerations. Thus, the assumption that each mole of ammonia contributes one mole of the OH ion to the solution, and so is equivalent to one mole of NaOH, is incorrect, and in general may lead to significant errors in the estimation of the pH correction for ammonia.

G-3


Figure G-1 Concentrations of the Aqueous Ammonium Cation (NH4+) (Continuous Line) and Aqueous Neutral Ammonia Species NH3 (Dashed Line) in Solutions of Pure Ammonia.

It should also be kept in mind that Figure G-1 represents solutions of pure ammonia in water. Additional alkalinity due to phosphate and excess NaOH will reverse the dissociation (eq. G-5) and further decrease the degree of dissociation. Using equations G-7 and G-9 the concentration of the ammonium cation as a function of pH can now be calculated, by first substituting [OH-] = 10pH – pKw (equation G-8) . The result is G-4


[ NH 4+ ] =

K d m NH 3 , 10 pH − pKw + K d

(eq. G-11)

and the final form of equation G-4 expressing the Na/PO4 ratio R as a function of pH, phosphate molality mPO4, and ammonia molality mNH3 is:

10 pH − pK w R ( pH , m PO 4 , m NH 3 ) =

10

K m NH 3 + Kd m PO 4

d pH − pK w

+ 2 mPO 4 (eq. G-12)

where pH is the measured value, with ammonia. Note that essentially the same reasoning leading to the above expression was presented by Verib.G1 Of course, without ammonia, equation G-12 is reduced to:

R ( pH 0 , m PO 4 ) =

10 pH 0 − pK w m PO 4

+ 2,

(eq. G-13)

yielding, after solving for pH0 without ammonia (the goal of the calculation),

pH 0 ( R, m PO 4 ) = log[( R − 2) m PO 4 ] + pK w .

(eq. G-14)

Since the sodium to phosphate ratio remains the same with or without ammonia, the equation G12 can be substituted into equation G-14 to obtain the following expression for pH0 without ammonia as a function the measured pH with ammonia and the concentration of ammonia, mNH3,   K d m NH 3  + pK w , pH 0 ( pH, m NH 3 ) = log 10 pH − pK w pH − pK w 10 + K d  

(eq. G-15)

and the following expression for the correction ∆pH = pH – pH0 to be subtracted from the measured value of pH:   K d m NH 3  − pK w ∆pH ( pH , m NH 3 ) = pH − log 10 pH − pK w pH − pK w 10 + K d  

(eq. G-16)

The molar concentration of ammonia mNH3 can be substituted by the mass fraction A expressed as ppm using mNH3 = 0.001 A / MNH3 = A / 17031. For the purpose of boiler water chemistry surveillance, equation G-15 may be used to calculate corrected pH from the measured boiler water pH and ammonia concentration. A graphical presentation of equation G-15 at various ammonia concentrations is shown in Figure G-2. A slightly simplified version of Figure G-2, which allows plant personnel to easily determine the corrected boiler pH, is also shown in Section 4 as Figure 4-5.

G-5


Note that equations G-15 and G-16 do not depend on the concentration of phosphate, and so are valid for CT, PC(L), and PC(H). The only approximations used in the derivation of these equations and Figure G-2 are the omission of the hydrogen ions in equation G-2 and the assumption that reaction G-3 describes the speciation of phosphate, as discussed previously. It is also worth noting that the sodium-to-phosphate ratio R used in this discussion can vary continuously from zero to infinity. In practice the amounts of sodium above R = 3 are expressed as ppm NaOH, ppm NaOH = 1000 MNaOH mPO4 (R – 3)

(eq. G-17)

or  K d m NH 3 ppm NaOH ( pH , mPO 4 , m NH 3 ) = 1000 M NaOH 10 pH − pK w − pH − pK w + Kd 10 

 − mPO 4  

(eq. G-18)

where 1000 MNaOH = 39997 kg/mol, and the molal concentration of phosphate mPO4 can be substituted by the ppm mass fraction P using mPO4 = 0.001 P / MPO4 = P / 94971.

Figure G-2 Relationship Between Corrected Boiler Water pH, Measured Boiler Water pH and Measured Ammonia Concentration, as Determined using Equation G-15. The lines on the figure increase in 0.1 ppm ammonia increments from 0 to 2 ppm.

G-6


G.1 REFERENCES 1. G.J. Verib. “Sodium to Phosphate Ratios” Power Plant Chemistry 4(11), 687-690, 2002.

G-7


H EPRI FOSSIL PLANT CYCLE CHEMISTRY REPORTS, GUIDELINES AND CONFERENCE PROCEEDINGS

1.

INSTRUMENTATION AND CONTROL GUIDELINES CS-5164, April 1987. (Replaced by item 28 Reference Manual, April 1999)

2.

STEAM SAMPLING TR-100196, December 1991

3.

MONITORING CYCLE CHEMISTRY IN FOSSIL PLANTS GS-7556, Vol.1 Monitoring Results. October 1991 Vol.2 International Cycle Chemistry. Dec. 1992 Vol.3 Conclusions & Recommendations. Oct. 1991

4.

MEASUREMENT OF pH and POTENTIAL IN SUPERCRITICAL WATER. TR-102277. Vol.1 Development of Sensors. April 1993 Vol.2 Long Term Testing and Measurement. Sept. 1993

5.

BEHAVIOR of AMMONIUM SALTS IN STEAM CYCLES. TR-102377. December 1993

6.

CYCLE CHEMISTRY CORROSION and DEPOSITION: CORRECTION, PREVENTION AND CONTROL. TR-103038. December 1993

7.

IN-SITU MEASUREMENT of PARTICULATES BY ACOUSTIC DETECTION TR-103627. February 1994

8.

DEVELOPMENT of OXYGEN and pH SENSORS for AQUEOUS SYSTEMS TR-103400. April 1994

9.

BEHAVIOUR OF SODIUM PHOSPHATES UNDER BOILER CONDITIONS TR-102431. June 1994

10. TURBINE STEAM, CHEMISTRY, AND CORROSION TR-103738, August 1994 11. CYCLE CHEMISTRY GUIDELINES FOR FOSSIL PLANTS: PHOSPHATE TREATMENT FOR DRUM UNITS TR-103665. December 1994 (To be replaced by this PC/CT Guideline in December 2003) H-1

EPRI Licensed Material EPRI Fossil Plant Cycle Chemistry Reports, Guidelines and Conference Proceedings

12. CYCLE CHEMISTRY GUIDELINES FOR FOSSIL PLANTS: OXYGENATED TREATMENT TR-102285. December 1994 13. SODIUM HYDROXIDE FOR CONDITIONING THE BOILER WATER OF DRUM TYPE BOILERS TR-104007. December 1994 (To be replaced by this PC/CT Guideline in December 2003) 14. INVESTIGATION AND OPERATING EXPERIENCE of RUSSIAN HIGH PRESSURE DRUM BOILERS USING PHOSPHATE TREATMENT TR-104008. January 1995 (not published) 15. ASSESSMENT OF THE RAY DIAGRAM. TR-106017. August 1996 16. CONDENSATE POLISHING GUIDELINES TR-104422. September 1996 17. SELECTION and OPTIMIZATION OF BOILER WATER AND FEEDWATER for FOSSIL PLANTS TR-105040. March 1997 18. CYCLE CHEMISTRY IMPROVEMENT PROGRAM (CCIP) TR-106371. April 1997 19. TURBINE STEAM, CHEMISTRY and CORROSION: EXPERIMENTAL TURBINE TESTS TR-108185, September 1997 20. STATE of KNOWLEDGE of COPPER in FOSSIL PLANT CYCLES TR-108460, September 1997 21. MOISTURE NUCLEATION in STEAM TURBINES TR-108942, October 1997 22. GUIDELINES for CONTROLLING FLOW-ACCELERATED CORROSION IN FOSSIL PLANTS TR-108859, November 1997 23. FLOW-ACCELERATED CORROSION IN POWER PLANTS EPRI BOOK. TR-106611. Revision 1. July 1998.

H-2


24. CYCLING, STARTUP, SHUTDOWN, AND LAYUP FOSSIL PLANT CYCLE CHEMISTRY GUIDELINES FOR OPERATORS AND CHEMISTS. TR-107754. August 1998. 25. STEAM, CHEMISTRY, and CORROSION in the PHASE TRANSITION ZONE OF STEAM TURBINES (Two Volumes) TR-108184, February 1999 26. VOLATILITY of AQUEOUS SODIUM HYDROXIDE, BISULFATE and SULFATE. TR-105801, February 1999 27. VAPOR-LIQUID PARTITIONING OF SULFURIC ACID and AMMONIUM SULFATE TR-112359, February 1999 28. REFERENCE MANUAL for ON-LINE MONITORING of WATER CHEMISTRY AND CORROSION. 1998 Update TR-112024, APRIL 1999 29. SODIUM PHOSPHATE HIDEOUT MECHANISMS TR-112137, May 1999 30. TURBINE STEAM, CHEMISTRY and CORROSION TR-113090, September 1999 31. INVESTIGATION of ELECTROPHYSICAL EFFECTS in the TURBINE EXHAUST upon STEAM FLOW AND POWER OUTPUT TR-113091, September 1999 32. REVISED GUIDELINES for MAKEUP WATER TREATMENT TR-113692, October 1999 33. CORROSION OF Cu-Ni-Zn ALLOYS IN WATER-AMMONIA POWER PLANT ENVIRONMENTS TR-113697, November 1999 34. THE ROLE of HETEROGENEOUS NUCLEATION on the STEAM CONDENSATION PROCESS TR-113524, February 2000 35. VOLATILITY of AQUEOUS ACETIC ACID, FORMIC ACID, and SODIUM ACETATE TR-113089, July 2000

H-3


36. CORROSION of LOW PRESSURE STEAM TURBINE COMPONENTS 1000557, November 2000 37. GUIDELINES for COPPER in FOSSIL PLANTS 1000457, November 2000 38. BEHAVIOR of AQUEOUS ELECTROLYTES in STEAM CYCLES Solubility and Volatility of CuO. 1000455, November 2000 39. COPPER ALLOY CORROSION in HIGH PURITY FEEDWATER 1000456, November 2000 40. RADIAL FLOW CONDENSATE POLISHING, RADIAL ION EXCHANGE SIMULATION STUDIES 1001409, March 2001 41. JUSTIFYING CYCLE CHEMISTRY UPGRADES to IMPROVE AVAILABILITY, PERFORMANCE and PROFITABILITY 1001557, June 2001 42. GUIDELINES for OFF-SITE REGENERATION of CONDENSATE POLISHING RESINS 1001502, June 2001 43. CONDENSATE POLISHING GUIDELINES: AMMONIUM FORM OPERATION 1004322, September 2001 44. TURBINE STEAM CHEMISTRY and CORROSION Electrochemistry in LP Turbines. 1006283, September 2001 45. The VOLATILITY OF IMPURITIES in WATER/STEAM CYCLES 1001042, September 2001 46. ELECTROSTATIC CHARGE AND ITS INFLUENCE ON THE CONDENSATION OF STEAM IN A TURBINE 1001332, September 2001 47. INFLUENCE OF WATER CHEMISTRY ON COPPER ALLOY CORROSION IN HIGH PURITY WATER 1004586, October 2001

H-4


48. STEAM TURBINE EFFICIENCY AND CORROSION: EFFECTS OF SURFACE FINISH, DEPOSITS AND MOISTURE 1003997, October 2001 49. GUIDELINES FOR CHEMICAL CLEANING OF CONVENTIONAL FOSSIL PLANT EQUIPMENT 1003994, November 2001 50. BEHAVIOUR OF AQUEOUS ELECTROLYTES IN STEAM CYCLES: THE SOLUBILITY AND VOLATILITY OF COPPER (I) AND COPPER (II) OXIDES 1003993, December 2001 51. VALUING CYCLE CHEMISTRY IN FOSSIL PLANTS 1004641, November 2002 52. CYCLE CHEMISTRY GUIDELINES FOR FOSSIL PLANTS: ALL-VOLATILE TREATMENT REVISION 1. 1004187. November 2002 53. RADIAL FLOW CONDENSATE POLISHING, RADIAL ION EXCHANGE PHYSICAL MODEL EXPERIMENTS 1004542, December 2002 54. STATE OF KNOWLEDGE ON DEPOSITION, PART I: PARAMETERS INFLUENCING DEPOSITION IN FOSSIL PLANT UNITS, 1004194, December 2002 55. VAPOR-LIQUID PARTITIONING OF PHOSPHORIC ACID AND SODIUM PHOSPHATES 1007291, May 2003

H-5


CONFERENCE PROCEEDINGS 1.

1985 Water Chemistry Symposium. CS-4950, Dec 1986

2.

Second Cycle Chemistry Conference. GS-6166, Jan. 1989

3.

International Conference on Cycle Chemistry. TR-100195. Dec 1991

4.

Interaction of Iron-Based Materials with Water and Steam. TR-102101, May 1993

5.

Fourth International Conference on Fossil Plant Cycle Chemistry TR-104502. December 1994

6.

Interaction of Non Iron-Based Materials with Water and Steam. TR-108236, July 1997

7.

Fifth International Conference on Fossil Plant Cycle Chemistry TR-108459. December 1997

8.

Workshop on Corrosion of Steam Turbine Blading and Disks in the Phase Transition Zone. TR-111340. November 1998

9.

Steam Chemistry: Interaction of Chemical Species with Water, Steam, and Materials during Evaporation, Superheating, and Condensation. TR-114837. August 2000

10.

Sixth International Conference on Fossil Plant Cycle Chemistry. 1001363. April 2001

11.

Seventh International Conference on Fossil Plant Cycle Chemistry. 1009194. December 2003.

H-6

Program: Boiler and Turbine Steam and Cycle Chemistry

SINGLE USER LICENSE AGREEMENT THIS IS A LEGALLY BINDING AGREEMENT BETWEEN YOU AND THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI). PLEASE READ IT CAREFULLY BEFORE REMOVING THE WRAPPING MATERIAL. BY OPENING THIS SEALED PACKAGE YOU ARE AGREEING TO THE TERMS OF THIS AGREEMENT. IF YOU DO NOT AGREE TO THE TERMS OF THIS AGREEMENT,PROMPTLY RETURN THE UNOPENED PACKAGE TO EPRI AND THE PURCHASE PRICE WILL BE REFUNDED. 1. GRANT OF LICENSE EPRI grants you the nonexclusive and nontransferable right during the term of this agreement to use this package only for your own benefit and the benefit of your organization.This means that the following may use this package: (I) your company (at any site owned or operated by your company); (II) its subsidiaries or other related entities; and (III) a consultant to your company or related entities, if the consultant has entered into a contract agreeing not to disclose the package outside of its organization or to use the package for its own benefit or the benefit of any party other than your company. This shrink-wrap license agreement is subordinate to the terms of the Master Utility License Agreement between most U.S.EPRI member utilities and EPRI.Any EPRI member utility that does not have a Master Utility License Agreement may get one on request.

About EPRI EPRI creates science and technology solutions for the global energy and energy services industry. U.S. electric utilities established the Electric Power Research Institute in 1973 as a nonprofit research consortium for the benefit of utility members, their customers, and society. Now known simply as EPRI, the company provides a wide range of innovative products and services to more than 1000 energyrelated organizations in 40 countries. EPRI’s multidisciplinary team of scientists and engineers draws on a worldwide network of technical and business expertise to help solve today’s toughest energy and environmental problems. EPRI. Electrify the World

2. COPYRIGHT This package, including the information contained in it, is either licensed to EPRI or owned by EPRI and is protected by United States and international copyright laws.You may not, without the prior written permission of EPRI, reproduce, translate or modify this package, in any form, in whole or in part, or prepare any derivative work based on this package. 3. RESTRICTIONS You may not rent, lease, license, disclose or give this package to any person or organization, or use the information contained in this package, for the benefit of any third party or for any purpose other than as specified above unless such use is with the prior written permission of EPRI.You agree to take all reasonable steps to prevent unauthorized disclosure or use of this package. Except as specified above, this agreement does not grant you any right to patents, copyrights, trade secrets, trade names, trademarks or any other intellectual property, rights or licenses in respect of this package. 4.TERM AND TERMINATION This license and this agreement are effective until terminated.You may terminate them at any time by destroying this package. EPRI has the right to terminate the license and this agreement immediately if you fail to comply with any term or condition of this agreement. Upon any termination you may destroy this package, but all obligations of nondisclosure will remain in effect. 5. DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES NEITHER EPRI,ANY MEMBER OF EPRI,ANY COSPONSOR, NOR ANY PERSON OR ORGANIZATION 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 PACKAGE, 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 PACKAGE 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 PACKAGE OR ANY INFORMATION, APPARATUS, METHOD, PROCESS OR SIMILAR ITEM DISCLOSED IN THIS PACKAGE. 6. EXPORT The laws and regulations of the United States restrict the export and re-export of any portion of this package, and you agree not to export or re-export this package or any related technical data in any form without the appropriate United States and foreign government approvals. 7. CHOICE OF LAW This agreement will be governed by the laws of the State of California as applied to transactions taking place entirely in California between California residents. 8. INTEGRATION You have read and understand this agreement, and acknowledge that it is the final, complete and exclusive agreement between you and EPRI concerning its subject matter, superseding any prior related understanding or agreement. No waiver, variation or different terms of this agreement will be enforceable against EPRI unless EPRI gives its prior written consent, signed by an officer of EPRI.

© 2004 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power Research Institute, Inc. Printed on recycled paper in the United States of America 1004188

EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA 800.313.3774 • 650.855.2121 • [email protected] • www.epri.com

EPRI Phosphat & Caustic Treatment Guidelines

Recommend Documents