ML20125E566
| ML20125E566 | |
| Person / Time | |
|---|---|
| Site: | Haddam Neck File:Connecticut Yankee Atomic Power Co icon.png |
| Issue date: | 03/31/1992 |
| From: | ELECTRIC POWER RESEARCH INSTITUTE |
| To: | |
| Shared Package | |
| ML19303F122 | List: |
| References | |
| EPRI-NP-6864-L, EPRI-NP-6864-L-R01, EPRI-NP-6864-L-R1, NUDOCS 9212170124 | |
| Download: ML20125E566 (215) | |
Text
_ - _ _ _ _ _ _ _ _ _ - - - _ - _ - - _ - _ - - - - - - - - - - - - - - - - - - - - - - - - . - - - - -
Keywords: EPRI NP-6864-L I
Nuclear steam generators Revision 1 PWR Projects S404-12.15,18,19.
Electric Power in-servico inspection 21, 24, 32. 37, 90,-92. 93 Researen Institute Tubes Final Report Nonproprietary Version March 1992 1
PWR Steam Generator Tube l
~
Repair Limits: Technical Support M Document for Expansion Zone PWSCC in Roll Transitions (Revision 1)
Nonproprietary 'fersion I '
i Prepared by Electne Power Researen Institute Palo Alto. California D AD K a 0213 P FDR
NOTICd:
For further information on licensing terms and conditions for this report, contact the EPR1 Office of Commercial Development. (415) 855-2974.
Electric Power Research Institute and EPRI are registered service marks of Electric Power Research Institute, Inc.
Copyright @ 1990 and 1991 Electric Power Research Institute, Inc. All rights reserved.
NOTICE This report was prepared by the Electric Power Research Institute, Inc. (EPRI).
Neither EPRI, members of EPRI, nor any person acting on behalf of any of them:
(a) makes any warranty, express or implied, with respect to the use of any information, apparatus, method, or process disclosed in this report or that such use may not infringe privately owned rights; or (b) assumes any liabilities with respect to the use of, or for damages resulting from the use of, any information, apparatus, method, or process disclosed in this report.
ii
COMMITTEE FOR ALTERNATE REPAIR LIMITS FOR E7. PWSCC Rev.1 D. A. Steininger, Chairman R. Pearson S. B wn" NORTHERN STATES POWER ELECTRIC POWER RESEARCH INSTITUTE J. Smith ROCHESTER GAS & ELECTRIC G. Airey NUCLEAR ELECTRIC C. Smoker WISCONSIN PUBLIC SERVICE F. Anderson NORTHEAST UTILITIES J.Jensen AMERICAN ELECTRIC POWER J. Blomgren COMMONWEALTH EDISON P. Anderson ABB-COMBUSTION ENGINEERING G. Bollini SPANISH UTILITIES R. Gamble"
^ '
J. Engstrom SWEDISH STATE POWER BOARD T. Gerber" STRUCTURAL INTEGRITY ASSOC., INC.
J. Frick SOUTH CAROLIN A ELECTRIC & GAS J. Gorman" DOMINION ENGINEERING, INC.
P. Hernalsteen" LABORELEC H. Houscrman
. R. Vollmer ZETEC, INC.
ELECTRICk E FRANCE G. Kammerdeiner I' #8' Y .,
DUQUESNE LIGHT COMPANY \VESTINGHb SF E ECTRIC CORP.
G. Kent PORTLAND GENERAL ELECTRIC \VILCOX BABCOCK D. Maves .
DWE NWER ALLEN NUCLEA ASSOCIATES B. Miller SOUTHERN NUCLEAR OPERATING CO.
Y.dditional participants for document's first issue: S. Green (EPRI), C. Welty, Jr. (EPRI),
M. Browne (SCE&G), D. Goetcheus (TVA), S. Leshnoff (GPU), B. Dow (EMSI).
" Document principal authors.
iii
n d
ABSTRACT .,
Primary water stress corrosion cracking (PWSCC) of steam generator tubes has been diagnosed in the tubesheet expansion zone (EZ) of many PWR steam generators.
When existing tube repair limits based on crack depth are applied (such as plug at 2 40% through wall), many tubes may require unnecessary repair. Allowing tubes with axial PWSCC to remain in service can be justified based on a combination of inspection by rotating pancake coil (RPC), a repair limit based on crack length rather than crack depth, a limit on the number of cracked tubes remaining in service, and a reduced primary to secondary allowable leak rate. This report provides the technical support for a length-based tube repair limit for steam gener- <
ator tubes in U.S. PWR power plants. In this epproach, axial crack length is used as the measure of tube integrity and tube leakage potential, and operation with known through-wall cracks is permitted. An additional repair limit to preclude significant circumferential cracks in combination with axial cracks and modified repair limits for partial-depth rolled tubes are included. This document has been prepared by a '
committee of U.S. and foreign industry participants who are experts on the technical and licensing issues associated with development and implementation of steam generator tube repair limits. The document represents the cornmittee's recommended approach, and presents information for use by utilities as a reference or supplement to site-specific analyses for developing revised tube repair limits associated with PWSCC in roll transition expansion zones.
'4 iv m
CONTENTS Section Page 1 INTRODUCTION AND BACKGROUND 1-1 1.1 Overview 1-1 1.2 Compliance With General Design Criteria 1-7 1.3 Background 1-10 L4 Current International Practices 1-13 1.5 References 1-15 2 CRACKED TUDE BURST DATA 2-1 2.1 Integrity Criteria 2-1 2.2 Tube Rupture 2-1 2.3 References 2-11 3 CRACK LENGTH MEASUREMENT 3-1 3.1 Background 3-1 3.2 Rotating Pancake Coil (RPC) 3-1 3.3 Demonstrated Crack Length Measurement Accuracy 3-6 __
3.4 References 3-16 . _ . .
4 ALLOWABLE CRACK LENGTH 4-1 4.1 Approach 4-1 4.2 Tube Rupture Equation 4-4 4.3 Loads and Safety Factors 4-6 4.4 Material Properties and Tube Dimensions 4-6 4.5 Correction for Tubesheet Constraint 4-7 4.6 Allowance for Crack Growth 4-11 4.7 Allowance for Crack Length Measurement Uncertainty 4-13 4.8 Allowable Crack Length Sample Calculation 4-16 4.9 References 4-18 5 LEAK RATE CONSIDERATIONS 5-1 5.1 Leak Rate Under Normal Loads 5-1 5.2 Leak Rate at Postulated Faulted Loads 5-4 5.3 References 5-8 v
, COMBINATION AXIAL AND CIRCUMFERENTIAL CRACKS 6-1 6.1 Introduction 6-1 6.2 Detection of Circumferential Cracks Between Axial Cracks 6-2 6.3 Pressure Capacity of Tubes With Combination Cracks 6-2 6.4 Additional Repair Limits to Preclude Significant 6-3 Combination Cracks.
7 REPAIR LIMITS FOR THE EXPANDED REGION OF PARTIAL- 7-1 DEPTH ROLL TUBES 7.1 Tube Burst and Allowable Crack Length 71 7.2 Leak Rate Consideration 7-2 7.3 Summary of Repair Limits for Partial-Depth Rolled Tubes. 7-2 7.4 References 7-3 8
SUMMARY
8-1 8.1 Overview 8-1 8.2 Comparison With European Practices 8-3 APPENDlX A STEAM GENERATOR DEGRADATION-SPECIFIC A-1 MANAGEMENT A.1 Steam Generator Management A-1 A.2 Steam Generator Management Strategy A-2 A.3 Defense In Depth A-2 A.4 EZ PWSCC Degradation Management A-4 A.5 References A-7 APPENDIX B BACKGROUND OF PWSCC B-1 B.1 Causes of Primary Water Stress Corrosion Cracking B-1 B.2 Plant Experience B-2 B.3 Impact of Roll Transition PWSCC B-6 B.4 Remedial Measures for PWSCC in Expanded Areas B-9 B.5 References B-12 APPENDIX C NDE OF ROLL TRANSITIONS - INDUSTRY EXPERIENCE C-1 C.1 Tube Pull Data C-1 C.2 Specific Details C-1 APPENDIX D BURST DATA FOR PART-THROUGH-W.ALL AND D-1 THROUGH-WALL FLAWS IN STEAM GENERATOR TUBES APPENDIX E LEAK RATE DATA FOR STEAM GENERATOR TUBES E-1 WITH SCC IN THE EXPANSION ZONE ROLL TRANSITION REGION vi
APPENDIX F CRACK GROWTH DATA F-1 F.1 Introduction F-1 F.2 Kiss Roll Rolls F-1 F.2 Standard Roll Data F-5 F.3 References F-6 APPENDIX G ANALYSIS hiETHODS FOR LEAK RATE CALCULATION G-1 G.1 Introduction G-1 G.2 Base Deterministic hiodel G-1 G.3 hiodel Validation G-5 G.4 Probabilistic hiethodology G-7 G.5 Computer Program G-8 G.6 References G-15 APPENDIX H COhiBIN ATION AXIAL AND CIRCUhfFERENTIAL H-1 CRACK DETECTION CAPABILITY UTILIZING RPC EDDY-CURRENT PROBES H.1 Introduction H-1 H.2 NDE Experiments H-1 H.3 Results H 13 APPENDIX I PRESSURE TESTS OF STEAhi GENERATOR TUBE WITH I-1 COhiBINATION AXIAL AND CIRCUhiFERENTIAL CRACKS I.1 Introduction I-1 1.2 LABORELEC Data I-1 L3 TECNATOhi Data I-2 1.4 Westinghouse Data I-2 ..
I.5 References I-3 _ - - _
ILLUSTRATIONS Eigure Page 1-1 Schematic of Roll Transition Axial Crack 1-2 1-2 EZ-PWSCC Axial Crack Configurations Acceptable for Continued 1-3 Service Per the Guidance in This Document 1-3 EZ-PWSCC Axial Crack Configurations That Should be Repaired 1-4 Per the Guidance in This Document 1-4 Examples of EZ PWSCC Crack Configurations With Distinct 1-6 Circumferential Crack Components for Which This Document Does Not Apply and Therefore Should be Repaired 2-1 Belgian (BELGATOM) Axial Defect Burst Test Data and Correlation 2-4 2-2 French (EdF/Framatome) Axial Defect Burst Test Data and 2-5 Correlation 2-3 Westinghouse Axial Burst Test Data and Correlation 2-7 2-4 Comparison of Several Tube Burst Test Correlations Along With 2-8 Lower Bound Tube Rupture Equation 2-5 Comparison of Burst Curve for Cracks With Experimental Data 2-10 3-1 Typical Rotating Pancake Coil 3-2 3-2 RPC Data Acquisition / Analysis 3-4 3-3 Zetec SM-15" Analysis Display 3-5 3-4 Crack Length Prediction - Composite Industry Experience 3-8 3-5 Crack Length Prediction - Standardized Residual Plot 3-9 3-6 Crack Length Prediction - Histogram of Residuals 3-10 viii
3-7 Crack Length Prediction - Electricit6 de France 3-11 3-8 Crack Length Prediction - Laborelec 3-12 3-9 Crack Length Prediction - Swedish State Power Board 3-14 3-10 Crack Length Prediction - Spanish Utilities 3-15 4-1 Illustration of Allowable Crack Lengths for Axial and Inclined 4-2 Axial Cracks 4-2 Allowable EZ PWSCC Axial Crack Length Calculation Steps 4-3 4-3 Tube Rupture Equation in Terms of Dimensionless Burst Pressure 4-5 and Critical Crack Length 4-4 Allowance for Tubesheet Reinforcement 4-11 4-5 Estimate of EZ PWSCC Average Crack Growth Rate Curve (Based on 4-13 Kiss-Roll Data) 4-6 Definition of Crack Length Measurement Correction, aNDE. 4-15 5-1 Leak Rate Versus Nominal or Average Axial Crack Length for 5-2 Steam Generator Tubes With Corrosion Cracks in the Expansion Zone Roll Transition Region, (Data From Appendix E)
A-1 Degradation-Specific Management Decision Tree A-8 A-2 EZ PWSCC Degradation Management Decision Tree A-9 B-1 Standard and " Kiss" Roll Transitions With Typical PWSCC B C-1 Ringhals 4 - Tube R22-C97. RPC Isometric - High and Low C-2 Sensitivity. Indication Just Below Roll Expansion C-2 Ringhals 4 - Tube R22-C97. Comparison of Composite Test Results C-3 C-3 Dampierre 1 - Tube L12-C43. Comparison of RPC and Destructive C-5 Test Results C-4 Doel 2 Tube R13-C30. Comparison of RPC and Destructive C-6 Test Results
~.;
ix
s t
C-5 Doel 3 Tubes R27-C52 and R23-C53. Actual Shape of Through-Wall C-7 Cracks and Comparison With RPC Length Measurement C-6 Ringhals 4 - Tube R47-C50. Axial anct Circumferential RPC C-8 Indications C-7 Ringhals 4 - Tube R47-C50. Axial and Circumferential Cracks C-9 C-8 Dampierre 1 - Tube L8-C48. Circumferential Cracks in Kiss Roll C-11 Transition Comparison of RPC and Destructive Test Results C-9 Ringhals 3 - Tube R18-C79. Short ID Axial Cracking C-12 ,
C-10 Ringhals 3 - Tube R18-C79. Composite Test Results C-13 C-11 Ringhals 4 - Tube R17-CSI. Multiple RPC Indications C 14 C-12 Ringhals 4 - Tube R17-C81. Composite Test Results C-15 F-1 Average Crack Growth Rates Based on LABORELEC Repeated F-4 Measurements of Kiss and Standard Roll Transitions with EZ PWSCC Indications. -
G-1 Plastic Component of the Crack Leakage Area G-3 G-2 Flow Discharge Coefficient Assumptions G-4 ,
G-3 Leakage Model Validation G-5 . .
G-4 Input Data G-10 G-5 Distribution of Crack Lengths G-11 G-6 Distribution of Actual (End of Cycle) Crticial Crack Lengths G-11 G-7 Average Crack Opening Width as a Function of Crack Length G-12 G-8 Distribution of Crack Opening Widths (Illustrated for a G-12 10 mm-long Crack)
G-9 Contribution of Each Crack Length Category to the Total Leak Rate G-13 G-10 Distribution of Tube Leak Rates G-13 G-11 Distribution of Steam Generator Leak Rate G-14 X
H-1 Typical Test Sample Geometry (Expanded Tube With U-Shaped H-2 Notch at Roll)
H-2 Pancake Coil - Diameter, Sample # Z-9300 H-3 H-3 Pancake Coil - Diameter (Shielded), Sample # Z-9300 H-4 H-4 Pancake Coil - Diameter, Sample # Z-9299 H-5 H5 Pancake Coil - Diameter (Shielded), Sample # Z-9299 H-6 H-6 Pancake Coil - Diameter, Sample # Z-9289 H-7 H-7 Pancake Coil - Diameter (Shielded), Sample # Z-9289 H-8 H-8 Pancake Coil - Diameter, Sample # Z-9281 H-9 H-9 Pancake Coil - Diameter (Shielded), Sample # Z-9281 H-10 H-10 Pancake Coil - Diameter, Sample # Z-9280 H-11 H-11 Pancake Coil - Diameter (Shielded), Sample # Z-9280 H-12 1-1 LABORELEC Combination Notch Geometries I-7 I-2 LABORELEC Through-Wall Angle Defect Pressure Tests I-8 I-3 LABORELEC Through-Wall "L" Shaped Defect Pressure Tests I-9 l-4 Spanish Combination Defect Pressure Test - Specimen Configuration I-10 1-5 TECNATOM Part-Through-Wall Defect Pressure Tests I-11 S 1-6 Westinghouse Through-Wall "L" Shaped Defect Pressure Tests I-12 I
I-7 Westinghouse Modified Through-Wall and Part-Through-Wall I-13 "L" Shaped Defect Pressure Tests I-8 Sketches of Axial and Circumferential Cracks in Westinghouse I-14 RP301-9 Specimens I-9 Westinghouse Stress Corrosion Crack Pressure Tests I-15 xi
TABLES Table Page 1-1 Full-Depth Rolled Plants With PWSCC 1 12 7 4-1 Alloy 600 Strength Properties, Mill-Annealed Tubing 4-8 4-2 Alloy 600 Strength Properties, Mill-Annealed Tubing 4-9 4-3 Estimate of E7. PWSCC Average Crack Growth Rates 4-12 (Based on Kiss-Roll Data) 4-4 Sample Calculations of Allowable Crack Length for 4-17 Tubes 5-1 Example Crack Distribution from RPC Inspection for Calculation 5-7 of Leak Rate at Faulted Load for Tubes 5-2 Estimated Leak Rates for Postulated Faulted Load for Two 5-7 Example Flaw Distributions 8-1 Comparison of Length-Based Repair Limits for Axial EZ PWSCC 8-4 Roll Transition Cracks A-1 Definitions A-10 A-2 Management Options A-11 1
B-1 B-5 l Reported Extent of EZ PWSCC at Plants With 210% Tubes Affected B-2 Status of PWSCC at Part-Depth-Rolled Plants in USA B-7 B-3 Status of PWSCC at Full-Depth-Rolled Plants in USA B-8 C-1 Listing of Tubes Removed for PWSCC C-17 D-1 NRC Phase II Burst Tests, Tubes With Axial SCC D-2 XIi
- ~ - ----.____ _ _ _ _ _ ,_ _ _
l i
D-2 CE S/G Tube Burst Data, AxialIGA Cracks Simulating D3 Those in St. Lucie 1 l D-3 Westinghouse S/G Tube Burst Data _ EDM Notches in D-4 ,
Roll Transition Region. Axial Through-Wall Cracks l i
E-1 CE S/G Tube Leakage Test Results. Axial Roll Transition Cracks, E-2 600*F E2 \W tinghouse Leak Rate Tests. Axial Roll Transition Cracks E-3 at 200*F l E-3 In Service Leak Rates - Axial Roll Transition Cracks E4 l F-1 LABORELEC Analysis of EZ PWSCC Crack Growth Data From F-2 I Doel-3 and Tihange 2 F-2 Crack Growth Data From Belgian Plants With Kiss Rolls F-3 H-1 Test Sample Matrix H 1 H-2 Summary of Pancake Coil Data H-13
.i I-1 LABORELEC Combined Axial and Circumferential Crack I .
Pressure Test Data l I2 TECNATOM Combined Axial and Circumferential Crack I5
+
Pressure Test Data I-3 Westinghouse Combined Axial and Circumferential Craw I6 Pressure Test Data l
xiii
-...- - :, u -.- .-
- ... .a .. . - . . ~ . . . ..= .
NOMF.NCLATURE a = axial crack length I = reference crack length als = correction for tubesheet constraint Ar- = allowance for crack growth awx = allowance for NDE uncertainty A = allowable crack length b = interc(pt of regression line c = cirr imferential crack length cs = circumferential spacing C3 = Gow stress coefficient S = leakage area E = elastic modulus p = Ouid density k = slope of regression line K = discharge coefficient A = dimensionless crack length m = stress magnification (bulging) factor 11 = Poisson's ratio P = tube (differential) pressure li = dimensionless pressure Per = critical pressure Q = volumetric Bow rate R = tube mean radius r = correlation coefficient t = tube wall thickness Sy = yield strength Su = ultimate strength ,
s = standard deviation o = hoop (circumferential) stress 5 = material Dow stress x = corrected reference crack length Y = NDE measured crack length yid = lower bound y xiv
i i
EXECUTIVE
SUMMARY
Primary water stress corrosion cracking (PWSCC) of alloy 600 steam generator tubes ;
has been diagnosed in the tubesheet expansion zor. * (EZ) roll transition of steam generator tubes in many PWR power plants. When existing tube repair (plugging /
sleeving)!!mits based on crack depth are applied, many tubes may require repair which is annece.,sary from either a safety or reliability standpoint. Allowing tubes with PL/ SCC ta remain in service can be justified based on a combination of enhanced in st rvice inspection, a repair limit based on crack length rather than crack depth, a limit on the number of cracked tubes remaining in service, and a ;
reduced prirr.ary-to-secondary allowable leak rate.
This report provides the technical support for a length-based repair limit for axial PWSCC in roll transition expansion zones of steam generator tubes in U.S. PWR power plants. In this approach, crack length is used as the measure of tube integrity and tube leakage potential, and operation with cracks that may be through-wall is ,
permitted. ,
The combination of remedial measures, inspection, an acceptable length based repair limit, and a reduced allowable leak rate, along with several repair options for cracked tubes, provides a series of plant-specific alternatises that can be used to develop the most cost effective means to maintain safety and acceptable reliability for steam generators experiencing PWSCC. The criteria described in this report are ,
in addition to previous work that established degradation specific evaluation criteria (e.g., P ', F', wastage, etc.) that provided significant operational benefit while maintaining adequate safety margins. ,
This document has been prepared by a committee of U.S. and foreign industry participants who are cxperts on the technical and licensing issues associated with "
development and implementation of steam generator tube repair limits. The documeni represents the_ committee's recommended approach, and presents ,
irdormation for use by utilities as a reference or supplement to site-specific analyses-for developing revised tube repair limits associated with axial PWSCC in roll transition expansion zones.
A length based repair limit approach is currently being used by the Belgian and Swedish utilities. A repair limit relying on the combination of a leak-before-break
- (LDB) argument and a length-based limit is being employed by Electricits de France 3
- Expansion zone is that region bf th 3 tube from the primary face of the tubesheet up so and including the final roll transition.
S-1
\
- . . . _ . . . _ . _ . . . _ . . . . _ . , _ , . . , , . , _ _ , , _ _ . ~ , ,_ _. _ - - , . . , . . , , . , _ . , , - _ , , - . . . . . .
(EdF) and a Spanish utility, it is believed that a similar length based limit approach is viable for those U.S. power plants that now have EZ PWSCC.
The length-based repair limit described in .his report has been developed for tubes with full- and partial-depth roll expansion zones. For tubes with full-depth roll expansion zones, crack length is used to determine that adequate leakage limits are maintained at faulted load and there is acceptable margin against tube rupture. For_
tubes with partial depth roll expansion zones, confinement of the cracked section of the tube within the tubesheet precludes tube burst at pressure less than that of an unflawed tube and the repair limit is based on maintaining acceptable leakage limits at faulted load for the distribution of cracked tubes.
Field experience indicates that circumferential cracks or combinations of circumferential and axial cracks occasionally occur in tube expansion zones. To provide added assurance that combined axial and circumferential cracks will not reduce margin against tube rupture or increase leakage at faulted load beyond allowable limits, an additional repair limit has been defined. This limit requires the spacing between adjacent axial crack indications to be or more.
This document applies to axial or inclined axial cracks where the axial extent of the inclined crack is greater than the circumferential extent. This document is not applicable for NDE indications evaluated to be distinct circumferential cracks. Tubes with identified distinct circumferential cracks should be repaired.
The elements of the proposed U. S. approach include:
Application of EZ PWSCC remedies to limit the number of cracks ,
expected in service.
- At each scheduled inspection outage, performance of rotating pancak.e coil (RPC) eddy-current inspections of 100% of the roll transitions of tube expansion zones in regions of the steam generator (i.e., the hot leg) where the tube expansion zones are susceptible to PWSCC.
Repair of tubes with axial cracks longer than a conservatively established repair limit which includes the following elements:
--Use of a tube rupture curve bounded by published tube burst test correlations, o --Application of USNRC Regulatory Guide 1.121 safety factors.
-Use of lower bound tube material properties.
-Correction for tubesheet constraint.
1
--Allowance for crack growth between inspections. l
-Allowance for eddy-current crack length measurement uncertainty. l i
--Restriction on spacing between adjacent axial cracks.
L..__,_.._._._____._.___.__,.__.____._____.____.._-_
t Implementation of the proposed U.S. approach is expected to require the following utility actions:
- Computation of the plant specific length based repair limit using the i
information in this report supplemented by available plant-specific data on material properties, crack growth rates and NDE uncertainties. .
- Limiting allowable primary to secondary leak rate limit to gpm to provide assurance against abnormal leakage and tube rupture at normal and faulted loads.
- Calculation of the potentialleak rate expected during postulated accident loads from the cracked tubes that remain in service.
- Calculation of maximum allowable site-specific leak rate at postulated faulted load to ensure that 10CFR100 dose limits are maintained.
Implementation of these elements constitutes a defense-in depth approach that was developed to ensure adequate levels of safety and compliance with applicable General Design Criteria in 10CFR50. The inspection scope and procedures, crack length and distribution limits, and the leak rate limits developed for tubes with EZ PWSCC ensure adequate margins against failure and excessive leakage and meet the requirements specified in the applicable General Design Criteria.
Section 1 of this document summarizes the overall approach, need, and justification for a length based repair limit for EZ PWSCC. Section 2 provides a discussion of data and criteria that can be used to predict and justify tube burst conditions for EZ PWSCC. Section 3 describes the NDE capability and develops the approach for dealing with inspection uncertainty. Section 4 provides a comprehensive discussion of allowable crack length. Key topics include analyses with safety factors, loads (accidents and normal operation), material properties, crack growth and NDE uncertainty considerations. Section 5 provides a discussion of when/how a leak-before-break argument applies to the recommended U.S. degradation-specific management strategy, and the calculation of the leak rate resulting from postulated accident loads and the cracked tubes that remain in service following an inspection (i.e., tubes with crack lengths below the repair limit) . Section 6 defines an addi-tional requirement to ensure that adequate margins for tube rupture and leakage at faulted load are maintained in the unlikely event that combined through-wall axial and circumferential flaws develop simultaneously in service. Repair limits for tubes with partial-depth roll expansion zones are defined in section 7. Section 8 provides a summary of the recommended length based repair limit approach.
Appendices A through I provide additional background and supporting material, and'are appropriately referenced in the body of the document.
.S-3 _;
l l ,
Section 1 INTRODUCTION AND BACKGROUND 1.1 OVERVIEW This report documents the justification for a repair limit for axial primary water strens corrosion cracking (PWSCC) that initiates at the tube inside diameter (ID)in the tubesheet expansion zone * (EZ) roll transition of steam generator tubes in U.S.
PWR power plants. In this approach, crack length is used as the measure of tube integrity and tube leakage potential, and operation with cracks that may be through-wall is permitted.
Figure l-1 shows a schematic of a tube and tubesheet along with important geometric parameters for EZ PWSCC roll transition cracks. The tube mean radius and thickness are designated as R and t, respectively. The axial crack length, a,is the extent of the longest crack in the axial direction. As indicated in figure 1-1, there are three possible locations of the EZ roll transition relative to the top of the tubesheet.
In case A, the roll transition is at the top of the tubesheet and is typical of a full-depth roll tube. In case B, the roll transition is below the top of the tubesheet. This is the case of a partial depth roll tube or of a tube which has been "under" rolled. In tha third case shown in figure 1-1, case C, the roll transition is above the top of the tubesheet. This would be typical of a tube which had been "over" rolled.
Repair limits have been developed for axial cracks at or above the tubesheet (figure 1-1, cases A and C) and for axial cracks in partial-depth rolled tubes (figure 1-1, case B). For cracks above the top of the tubesheet, limited leakage at faulted load and acceptable crack length are the bases for the repair limit, while limited leakage at faulted load is the basis for the repair limit for partial-depth rolled tubes.
This document can be used to evaluate single or multiple part-through-wall or through-wall axial and inclined axial cracks with configurations shown in figures 1-2 and 1-3.
- Expansion zone is that region of the tube from the primary face of the tubesheet up to and including the final roll transition.
1-1
i .-
)
i T
Figure 1-1. Schematic of EZ Roll Transition Axial Cracks 1-2.
t i
t 7
i i
t i
o i
i
~
f Figure 1-2. EZ-PWSCC axial crack configurations acceptable for continued service per the guidance in this docurnent.
1-3
_ . . . . _ - _ . - . .-.--.--;.-._- . . . _ .,_._.;_.u..._..
. . _ . . - _ u _ a _ ._ _ .. _ . _ - - n - . . .
_, .. .m .
. _. _ _ . . _ _.~ . _ . _ . . . ..__. - _ ____ _ _ -. _. _ . _ _ . . . _ . . _ . . _ . . - . . ._
t J
1
+
1 t
'i I
i
-t a
s 5
Figure 1-3. EZ-PWSCC axial crack configurations 'that should be <
repaired per_ the guidance in this document.
4 14
-y.. % s., ---a, , - ;,,...m.m..,, ..,,-.v . . . . , , . , , ,y - c-r=-,nt<.- ,,,: , , , . , , , -
,m y.. --weg,,m, . , , , ,. ,,,y-., ..yg,7,---,m--
- This document does not apply to inclined cracks where the circumferential extent is larger than the axial extent or to NDE indications evaluated to be distinct circum-ferential cracks or cracks with significant circumferential components, e.g., "L", "T",
"U", or "Z" shaped cracks, examples of which are shown in figure 1-4. Tubes with identified distinct circumferential cracks should be repaired.
This length based repair limit approach is based on tube burst, leak rate, PWSCC i growth data, and demonstrated crack length measurement capabilities using rotating pancake coil (RPC) equipment and procedures. Together, these support an alternative to the eddy current depth-based repair limit used by most U.S. PWRs, and can be used to support site-specific submittals to the U.S. NRC for revision of -
tube repair limits.
Application of the repair limits developed in this document requires at each scheduled inspection outage, performance of rotating pancake coil (RPC) eddy-current inspections of 100% of the roll transitions of tube expansion zones in regions of the steam generator where the tube expansion zones are susceptible to PWSCC (i.e., the hot leg).
'A leak rate analysis should be performed for cracked tubes that do not require repair ,
to determine if the dose rate from leakage at faulted load will remain a small _
fraction of the 10CFR100 limits. If results from the accident leak analysis indicate an excessive leak rate, tubes should be repaired selectively so that the dose rate is a
. small fraction of 10CFR100 limits.
To provide assurance against abnormalleakage and tube rupture at normal and faulted loads, a leak rate of has been established as the allowable primary to secondary leak rate limit during normal operation.
Implementation of this document satisfies the following general requirements:
- Complies with the General Design Criteria.
- Avoids conditions that lead to exceeding a conservative leakage limit of
- Provides adequate margin against tube rupture under normal operating and postulated accident loads (e.g., steam line break).
- . Avolds excessive leakage under postulated accidentl oads.
1-5
..-.....- - _ -_ _ . u - - - . _ - - . - _ - _ - - - _ _ - - - . - . -
Figure 1-4. Examples of EZ-PWSCC crack configurations with distinct .
circumferential crack components for which this document does not apply and therefore should be repaired.
16
- 1.2 COMPLIANCE WITH GENERAL DESIGN CRITERIA [
The repair limit for axial PWSCC in tube expansion zones has been developed to '
ensure compliance with the applicable General Design Criteria (GDC) in Part 50 of Title 10 of the Code of Federal Regulations (10CFR50). The GDC were reviewed and j lt was concluded that GDC 14,15,30,31 and 32 are applicable to the development of l repair limits for axial PWSCC occurring in steam generator tubes. The remainder of paragraph 1.2 summarizes the bases for compliance with the applicable GDC.
1.2.1.GDC 14 i
GDC 14 requires the reactor coolant pressure boundary (RCPB) to be designed, fabricated, crected and tested so as to have an extremely low probability of abnormal leakage, of rapidly propagating to failure, and of gross rupture.
R.G.1.121, "Dases for Removing Degraded PWR Tubes from Service," provides 1 explicit and implied safety margins for tube loading. R.G.1.121 explicitly states that tube loading should have a safety factor of 3.0 under normal operating conditions.
The regulatory guide further states that the margins of safety against tube rupture under postulated conditions should be consistent with the margin of safety deter = ;
mined by the stress limits specified in Section III of the ASME Boller and Pressure Vessel Code. The repair limits discussed in this report are shown to meet all of the above acceptance criteria.
Following the implementation of the EZ PWSCC tube repair criterion, stearn ,
generator tube integrity is maintained both by eddy-current inspection and by measuring steam generator primary to secondary leakage. To further help to ensure adequate margins of safety are maintained during service, in-service inspections are performed for 100% of the affected regions during each refueling outage. Any tubes found to have flaws laiger than those necessary to maintain acceptable matgins against tube rupture (including consideration of additional flaw growth during service) will be repaired. Service experience indicates leakage levels at plants with through-wall PWSCC in the expansion zone roll transition can be expected to remain at vcry low levels and range from 5 to 35 gpd (1). - A maximum leak rate of has been established for normal operation to maintain leakage within acceptable levels during all plant conditions. This level of leakage helps to ensure that the dosage contribution from tube leakage will be limited to a small fraction of 10CFR100 dose guideline values in the event of a steam generator tube rupture or steam line break event.
1.2.2 GDC15 f
GDC 15 requires the reactor coolant system and associated auxiliary, control, and protection systems to be designed with sufficient margin to ensure the design -
margins of the RCPB are not exceeded during any condition of normal operation, including anticipated operating occurrences.
1-7
i As the steam generator tubing represents a large portion of the total primary system pressure boundary, to help to ensure that steam generator tube integrity is main-tained during normal operation, including anticipated operational occurrences, a i i
factor of safety of three margin on normal pressure loads is used in the alternate repair criterion to define the maximum axial crack length allowed to rem e n in .
service. In addition, a maximum leak rate of has been establishet to enhance the likelihood of leak before break.
1.2.3 GDfl0 ;
GDC 30 requires that components which are part of the reactor coolant pressure boundary shall be designed, fabricated, erected, and tested to the highest quality standards practical. Also, means shall be provided for detecting and, to the extent practical, identifying the location of the source of reactor coolant leakage.
Following implementation of the steam generator tube repair limit, use of the rotating pancake coil (RPC) eddy-current probe is an example of how the reactor coolant pressure boundary will be tested to state-of the art quality standards. The ;
eddy-current sample size is 100% of the affected area.
During reactor operation the secondary side of the steam generator will be monitored for radioactivity to detect leaks from cracks in steam generator tubes. If leakage exceeding 0.1 gpm is detected, the unit will be shutdown and the steam generator tubes will be inspected to determine the source of the leakage.
1.2.4 fiDC 31 GDC 31 requires the reactor coolant pressure boundary be designed with sufficient margin to ensure that when stressed under operating, maintenance, testing, and
, postulated accident conditions (1) the boundary behaves in a nonbrittle manner and (2) the probability of rapidly propagating fracture is minimized. The design shall- ,
reflect consideration of service temperatures and other conditions of the boundary material under operating, maintenance, testing and postulated accident conditions .
and the uncertainties in determining (1) material properties, (2) irradiation effects on material properties, (3) residual, steady state and transient stresses, and (4) size of flaws.
To ensure that the tubes behave in a nonbrittle manner and the probability of--
rapidly propagating fracture is minimized during operating, maintenance, testmg and postulated accident conditions, the tubes cre manufactured from ductile-materials and conservative margins are applied to normal, maintenance, testing and postulated accident loads. These margins have been confirmed from.
experiments with tu'oes experiencing EZ PWSCC. This testing has been accredited >
for use in the development of the tube repair criterion.
1 P
- t v- , e- , , ,ww -w,. ,
,,w, ,- -- -~~~---,,e--~n m, --r - sy m .,n. a -~.-e -
The margins have been determined considering the temperature and pressures at normal and postulated accident loads, and the uncertainties associated with material properties, stresses, flaw size, flaw length measurement error, and in service crack growth. Conservative values and margins for these variables have been used to ,
account for these uncertainties. Again as noted above, upon implementation of the criterion, the margins are in compliance with those specified in Regulatory Guide 1.121, the ASME Code, and the individual plant Final Safety Analysis Report. ;
Where plant specific measured values for these variables are unavailable, conservative values have been used.
t 1.2.5 GDC 32 GDC 32 requires that components which are part of the reactor coolant pressure I
boundary be designed to permit inspection and testing of important areas and features to assess their structural and leak tight integrity.
The presence of multiple, axial, (potential) through-wall cracks in the EZ region does not adversely affect the ability of an eddy-current analyst to interpret the eddy _ t current signal and categorize the condition of the tube. Eddy-current inspections '
using the RPC probe will be performed for 100% of the tube expansion roll transitions in regions of the steam generator where the tube expansion zones are susceptible to PWSCC. Performing these inspections will provide assurance that the .
PWSCC is within limits such that the safety margins used in the structural integrity evaluation are maintained during service conditions and steam generator primary to secondary leak rates during normal and postulated accident condition loads remain within required limits.
1.2.6 GDC Summ2rg +
Section 2 of this report defines the burst pressure for tubes with PWSCC in roll transition expansion zones, while sections 4,6 and 7 define the allowable crack length, including consideration of material properties and in service crack growth, that will ensure adequate margins against rapidly propagating failure and nonbrittle fracture. The information in sections 2 and 4 provide the bases for compliance with GDC 14,15, and 31.
Section 3 of this report describes the examination methods, including provisions for inspection error, that will be used to detect and size PWSCC in tube expansion zones. Implementing the methods described in section 3 provides bases for compliance with various requirements in GDC 14,15,30,31, and 32.
Sections 5,6 and 7 describe the leak rate and crack distribution limits that will .
ensure adequate margins against abnormal leakage during normal and postulated accident conditions. Implementing the leak rate and crack length limits as described
'in section 5 will ensure compliance with GDC 14,15,30, and 32.
1-9 L
1,3 BACKGROUND 1.3.1 Steam Generator Tube Degrastalign Experience shows that steam generator tubes may be susceptible to degradation from a variety of mechanisms. As degradation progresses, the affected tubes (or tube segments) are repaired based on in-service inspection (ISI) results, or when primary to secondary leakage exceeds a pre-established limit during power operation. Defec-tive segments are repaired either by taking the entire tube out of service, plugging, or by installing internal sleeves in the area of local degradation. If degradation progresses and more segments are repaired, core cooling requirements ultimately may dictate that either the plant be derated or that steam generators be replaced.
Guidelines for evaluating steam generator tube integrity are contained in Regulatory Guide 1.83, (Revision 1, July 1975), " Inservice Inspection of Pressurized Water Reactor Steam Generator Tubes" Q), and Regulatory Guide 1.121, (Draft for Review and Comment, July 1976), " Bases for Plugging Degraded PWR Steam Generator Tube," Q).
In the United States, the current application of the depth based guidelines for steam-generator tube defect management can be broadly characterized as follows:
- 1. Assume each degradation form is of equal concern;
- 2. Determine the allowable degradation size based on part through-wall crack depth regardless of defect length, volume of materialloss or cause of the degradation:
- 3. Establish a shutdown leak rate to help ensure that a degradation that penetrates through wall during an operating cycle will leak rather than break;
- 4. Assume that the combined application of known remedial measures, the inherent leak-before-break nature of alloy 600 for many degradation-types and the mandated inspection protocol provide assurance that the consequences of tubing damage will be acceptable.
While the depth-based approach has proved acceptable,[and experience shows that there have been few tube ruptures, and that the consequences of those ruptures are ,
acceptable (4)] in certain cases it has led to excessive and unnecessary tube plugging (specifically in the case of small volume degradation such as isolated pits and primary-water-initiated strees corrosion cracks). In such cases, a more degradation-specific approach has been developed to provide significant benefit to affected' plants while still maintaining acceptable margins. Several examples of these degradation-specific criteria include :
1-10
e
- 1. P' and P Criteria - permit through wall cracks to remain in service within the steam generator tubesheet region (5 #). A more detailed discussion of the P'/F* criteria is provided in Appendix A.
- 2. Tube Support Plate Primary Water Stress Corrosion Cracki.DS-proposed a justification for through-wall axially oriented cracks to remain in service at the support plate elevation with denting present @.
- 3. Pittine Degradation - justifies a 647o of wall thickness repair limit for '
tubes experiencing pitting QQ,11).
allowable wall loss of 827o; implementing Reg. Guide 1.121 considerations the corresponding tube repair limit was conservatively limited to 517e wall thickness (11.1)).
-5. Wastage justifies a 477c of wall thickness repair limit for steam generator tube thinning (E1E). _
In each of these examples, a degradation-specific limit has been developed that satisfies the intent of Reg. Guide 1.121. The justification for degradation-specific ,
criteria integrates concern for structural capability and leakage of the degraded tubing with nondestructive examination accuracy, tube crack tolerance, and leak detection capability. Appendix A provides a more detailed discussion of the generic :
degradation-specific approach to developing tube repair limits.
1.3.2 EZ PWSCC PWSCC has been observed in steam generators with mill-annealed alloy 600 tubing.
It can be caused by denting at tube support plates or top of the tubesheet, and/or processes associated with tubesheet expansion zone or inner row U-bend fabrication.
PWSCC due to denting is controlled by eliminating denting through improved -
secondary water chemistry, and inner row U-bend cracking is controlled by local in situ stress relief or by preventively plugging susceptible tubes. Thus, the current -
area of concern is EZ PWSCC which is the focus of this document.
Many plants have experienced EZ PWSCC. Appendix B provides tables. _
summarizing EZ PWSCC experience at plants with partial-depth and full-depth rolls. Plants with full-depth roll expansion that have experienced EZ PWSCC are -
L :. listed in Table 1-1.
1-11
Table 1-1 FULL-DEPTH ROLLED PLANTS WITH EZ PWSCC *
(Ref.16)
' Limited to plants with mill-annealed tubing.
" Table B-3 includes more recent data.
1-12
i l
Experience shows that EZ PWSCC cracks are generally short, axially oriented and :
confined to the expansion zone, and may be through-wall; the cracks generally have ;
been found in tubes in the steam generator hot leg Pulled tube and RPC data from l standard roll plants show that cracks grow slowly,if at all, after they have grown through the high residual stress field in the roll transition zone. Laboratory testing and some field experience indicates that circumferential cracks, or axial cracks with j circumferential components, may develop as well. In the field, expansion zone crack leakage has been low and has increased gradually over time, and there are no cases of sudden tube rupture with expansion zone cracks.
1.4 CURRENTINTERNATIONAL PRACTICES 1.4.1 EdF Practices for EZ PWSCC EdF is not governed by USNRC regulations and, as such, has developed an alternate approach through interactions with the French regulatory authorities. Specifically, the EdF approach uses a leak before risk-of bmak (LBRB) argument, characterized as follows: (a) uses analysis to show leak before-break for most axial cracks; (b)in
. regions for which LBRB could be violated (i.e. sludge pile area) completes 100% RPC inspection at each annual outage; (c) utilizes sensitive on line leak detection capa-bility (steam line N16 monitors); (d) applies plant-specific Technical Specification leakage limits; (e) uses sensitive off-line helium leak location methods; and (f) applies systematic bobbin coil sampling and limited RPC follow-up on bobbin coil signa s that are indicative of "long" cracks. Tubes are repaired: (a)if leak tests indicate "long" cracks, (b)lf IU'C indicates cracks that could result in reaching the Technical Specification leakage limit during the next operating cycle, (c)if RPC inspection locates any crack in the sludge zone), or (d) if any circumferential crack is indicated and not otherwise dispositioned.
The number of tubes inspected at the outset with the bobbin coil varies from inspection to inspection based on previous plant-specific damage history-generally 12.5% or 25% on one or three generators.
1.4.2 Ilclgian Practices for EZ PWSCC The Belgian utility (Electronucleaire) is required to conform to USNRC Regulatory Guides. However, their proposed use of a modified version of Regulatory Guide 1.121 to develop an alternate tube repair limit has been reviewed and accepted by the Belgian safety authorities. Essentially, they use at temperature material properties for ultimate and yield stresses and develop an approach using safety margins on crack length rather than pressure (load) for analysis, supplemented with field data for axial crack growth rate, and with testing to take credit for tubesheet reinforce-ment. Using this approach, ~ hey have an axial crack length plugging limit of tubing. This is coupled with a 100% inspection of the tubesheet region with RPC technology at each refueling outage. +
i-13
1
- 1.4.3 Spanhb Practices for EZ PWECC i
The repair limits at one Spanish utility are pnmarily based on EdFs leak-before-risk-of break (LBRB) method. However, limits are also placed on crack lent;th, as determined by inspections, foi operational reliability reasons, i.e , to protect against i forced outage 3 For this reason, axial cracks in roll transition zone of tubes with normal roll geometries require repair if the crack length above the top of the 4
tubesheet CITS) is or more. One hundred percent of steam generator tubesheet areas raust be inspected each refueling outage, although bobbin coil methods are acceptable. All anomalous and crack like indications require RPC 2
inspection. Allowable leak rates for the different units were selected by the safety ,
authorities based on evaluations of the possible interactions of various defect mechanisms (e.g., OD IG5CC) with PWSCC, and based on concerns that actual leak rates through individual cracks may be lower than predicted.
1.4.4 Swedish Practices for EZ PWSCC 1 The Swedish utility initially followed an axial crack-depth-based 50% through-wall .
tube repair limit. Shortly after the PWSCC problem was detected and determined to be s;gnificant at the lead plant with Model D3 steam generators, Ringhals 3, the U.S.
-developed P' criteria were adopted. These criteria allow axial cracks to remain in service if tubes or the tube support structure prevent tube pull out even if the tube severs at the indication location.
An additional change was also adopted soon after discovery of EZ PWSCC at Ringhals. This change allows axial cracks of any length to remain in place if they were located below the top of the tubesheet. This position is justified on the basis that axial cracks do not affect pull out strength, or significantl.y impact bkage.
Circumferential degradations in the P' regions, and all indications above the top of the tubesheet, were still governed by the 50% through-wall rule.
In 1988 and 1989, new criteria were developed and adopted using Framatome assistance. These new criteria are based on a degradation identification and monitoring approach. The criteria are based on a critical crack size under accident conditions of based on tube minimum properties and worst case ,
dimensions. To protect against growth of cracks to beyond this value, allowable crack length as determined during inspection is less than the critical size to allow for degradation growth and length measurement uncertainty.
o 1
1-14 0
1.5 REFERENCES
- 1. P. Hernalsteen. Steam Generatm PWSCC In Service Leakage Experience in ikigium. LABORELEC Report hf 00-90-022/R-PH, hiay 1990.
- 2. U.S. Nuclear Regulatory Commission Regulatory Guide 1.83. ' Inservice Inspection of Dressurizeu Water Reactor Steam Generator Tubes."
Revision 1, July 1975.
- 3. U.S. Nuclear Regulatory Commission Regulatory Guide 1.121. " Bases for Plugging Degraded PWR Steam Generator Tubes." Draft for Comment, August 1976.
- 4. NRC Integrated Program for the Resolution of Unresolved Safety issues A-3, A 4, and A 5 Regarding Steam Generator Tube Integrity. NUREG-0844, September 1988.
- 5. Tubesheet Region Plugginn Criterion for Full Depth Hardroll Expanded Tubes. WCAP-10950, September 1985.
- 6. Tubesheet Recion Fluggi.ng Criterion. WCAP-11225, July 1986.
- 7. Tubesheet Region Plugging Criterion. WCAP 11225, Rev.1, October 1986,
- 8. Docket Nos. 50-369 and 50-370. Issuance of Amendment No. 59 to Facility Operating License NPF-9 and Amendment No.40 to Facility Operating License NPF-17 for the hicGuire Nuclear Station, Units 1 and 2, August 1986.
- 9. North Anna Unit 1 Steam Generato Tube Integrity Safety Evaluation.
WCAP-11311, October 1986.
- 10. Docket No. 50-286. Request for Amendment to Facility Operating License DPR-64. " Proposed Changes to Technical Specifications Regarding Steam Generator Tube Plugging Limit." J. P. Bayne (New York Power Authority) to W. R. Denton (US NRC), September 1984.
- 11. Docket No. 50-286. Issuance of Amendment No. 55 to Facility Operating License DPR 64 for Indian Point Nuclear Generating Unit No. 3, hiay 1985.
- 12. Docket No. 50-255. Facility Operating License DPR-20. "1983/1984 Steam Generator Evaluation and Repair Report, Palisades Plant." D. J. VandeWalle (Consumers Power Company) to D. H. Crutchfield & US NRC), April 1984.
- 13. Docket 50-225, LS05-84-06-015. "1983/1984 Steam Generator Inspection, Re:
Palisades Plant." D. J. Crutchfield (US NRC) To D. J. VandeWalle (Consumers Power Company) June 1984.
1-15
- 14. Docket No.50-261. Facility Operating License No. DPR 23. " Supplemental :
Request for License Amendment for Cycle 9 Operation." P. W. Howe j
.(Carolina Power and Ligh0 to S. A.Varga (US NRC), July 1982.
- 13. Docket No. 50 261, Issuance of Amendment No. 71 to Facility Operating License No. DPR-23 for the H. B. Robinson Steam Generator Electric Plant, Unit No. 2, July 1982. ,
- 16. " Steam Generator Progress Report, Revision 7." September 1991. (Available from EPRI Steam Generator Reliability Project staff.)
r t
M i
5 Y
?
~ 1-16
~a, ,-,,,,w,-.. -- -- -:-.. , . - - - - . , , , ~ , . , - , , , , . , - - . , ~ - . n-,,
l Section 2 CRACKED TUDE UURST DATA 2.1 INTEGRITY CRITERIA To ensure compliance with the General Design Criteria of 10CFR50 for tubes with EZ PWSCC, several criteria are used to assess the integrity of steam generator tubes and develop a length based acceptance criterion for axial cracks in the roll transition _
region. These include: (1) a crack length limit (allowable crack length) which assures margins against structural tube failure, (2) enhanced NDE of 100% of the tubes that may be affected by EZ PWSCC to provide a high level of confidence that
- tubes exceeding the allowable crack length are repaired,(3) controlling the distribution of potential leaking tubes to preclude exceeding the site boundary dose limits during postulated faulted loads, and (4) reduced allowable leak rate from 0.35 to 0.1 gpm to enhance leak before-break.
The remainder of this section identifies conditions and bases that can be used to determine the axial crack length corresponding to tube rupture. Allowable crack lengths are developed in section 4, while crack distribution end leak rate limits are discussed in section 5. Restrictions on spacing between axial cracks based on the potential for the presence of combined axial and circumferential cracks are developed in section 6. Repair criteria for cracks that exist in partial-depth rolled tubes (see figure 11B) are presented in section 7.
2.2 TUBE RUPTURE 3.2.1 Axial Through-Wall Cracks Tube rupture is the condition of a tube segment with a crack at the maximum pressure it can support under actual steam generator conditions. Rupture occurs by significant tube deformation followed by crack tip extension.
The rupture bohavior of alloy 600 steam generator tubes with axial cracks has been studied by a mber of investigators (1-4). These studies have generally involved tests with tubes containing electrodischarge machine (EDM) slits to simulate cracks, with a patch or oladder in the tube to prevent leakage from the through-wall slit while the tube is pressurized until rupture occurs. A limited number of tests have been conducted on cracked tubes removed from service , These tests show that the rupture behavior of tubes with in-service cracking is comparable to tubes with EDM slits.
2-1
= .. ______ _ ____
i In addition to reporting experimental results, investigators have developed burst behavior relationships and correlated them with experimental results. These relationships are in terms of tube parameters that control burst behavior and can be used for predicting tube burst performance. l'arameters used in the correlations are tube and crack dimensions, material properties and load. t in this section, available relationships for predicting steam generator tube burst behavior with axial defects are reviewed and compared. A bounding relationship is then developed.
The Belgian agency, BELGATOM, has conducted an experimental investigation of '
burst behavior of alloy 600 tubes with axial defects. Early work was concentrated on-a single heat of tube (1). Later work included tests with different heats and tube diameters. These tests confirmed the earlier burst behavior ;
correlation (1). ,
BELGATOM has correlated the burst test results with a limit load equation which, for a flat plate,is written as o= (2-1) where o is the applied stress and o is the material flow stress. 4 The critical length of an axial crack in a tube can be found by considering the stress magnification factor, m, associated with the bulging of the tube. In this case, e
o= (2 2)' 4 Different forms of the bulging factor have been proposed and used. BELGATOM found that the bulging factor expression due to Erdogan (5) best fit their data. This equation is ,
m= (2-3)
{
I where a is the through-wall crack length, R is the tube mean radius, and t is the tube wall thickness. Further, BELGATOM used a flow stress of
- o. = (2-4) and expressed the applled stress as a function of the differential pressure i
o=
-(2-5) i 2-2
1 where Su is the ultimate strength, Sy is the yield strength, and P is the differential pressure. BELGATOM's initial correlation with bunt test data is shown in figure 21.
Later work with additional data supports this correlation (1).
Additional burst tests have been carried out by Framatome and EdF (2) on alloy 600 and alloy 690 tubing with nominal diametert of . These burst tests have been performed with various types of degradation located in the straight
- portion of tubes away from supports, in the roll transition, in the vicinity of supports and in U bends.
These tests demonstrate that the burst behavior of tubes with through wall cracks can be correlated with the expressions similar to those used by BELGATOM with slightly different correlation parameters. These include for the flow stress 0= (2-6) and for the bulging factor, m= (2-7) :
Figure 2-2 shows the French tube burst behavior correlation plotted with through wall burst data (2).
Like other investigators, the French have pointed out that the tube hoop stress due to internal pressure controls the burst behavior of tubes with axial cracks, and that axial stresses are relatively unimportant (2). In addition, their experiments confirm that the burst pressure for a tube with a crack near the top of the tubesheet is higher than that of a tube with the same defect remote from supports.
Westinghouse (3) has also conducted burst tests with both tubes. Further, they have correlated the burst behavior of tubes with through wall axial cracks with the following two dimensionless terms A= (2-8)-
E" '(2-9)-
Figure 2-3 shows normalized burst data correlated with a ninth order polynomial fit shown below.
F=
(2-10) 2-3 ,
^
- r. u,--- - - , , . , .. e 4.m-~-w. .-.,nn-, ...e.,. e,.,,,..,.mn,,. -r. - . ,.. -.,--..e-, L - . , , . ,,,--,,.n , a+. . . . . . . - .a..e, . . , . .=- .n..'.,
Figure 2-1. Belgian (BELGATOM) Axial-
- Defect Burst Test Data and Correlation (1) 2-4:
t ---
i i
i I
h L
+
t t
t 1
5
.i ,
e t
}
I-i
.t Figure 2-2. French (EdF/Framatome) Axial Defect Burst Test Data and Correlation (1)'
2 .1
. - . . . ....~,.,_..,_m... . . _ . _ _ , , - , , , _ _ _ , , , . . . - , . . , , , _ , , , . . _ . , ~ , , _ , _ . , . . ~ . . . . , , , , , . . - . . - . . . - - . . . . . . .
The British (.4) have recently reported the results of burst test data for tubes with through-wall axial corrosion cracks and compared their results with their propri-etary tube burst and leak code TUBELEAK. TUBELEAK is reportedly based on a correlation with the burst test results of other researchers, and predictions with this code are indistinguishable from the Westinghouse correlation shown in figure 2-3.
The Belgian, French, Westinghouse and British correlations for the burst behavior of through-wall axial cracks are shown plotted together in figure 2-4. As indicated, the Westinghouse and British correlations are nearly indistinguishable. Both these correlations fall below the Belgian and French correlations.
To develop a single tube rupture equation, a curve which bounds the four correlations shown in figure 2-4 was developed. This curve is in terms of the dimensionless parameters II and A and it is also shown in figure 2-4. The equation for the curve, valid for , is li =
(2-11)
A=
This relationship is used to determine the through wall critical crack lengths at postulated faulted load and for calculation of reference crack lengths used in calculation of allowable creack length in section 4.
Tube burst tests have been conducted with either a patch or bladder to prevent -
leakage through the test specimen crack until tube rupture. Although efforts have ,
been made to minimize the effect of the crack sealing system on tube burst behavior, this can not be assured.
Recently, LABORELEC conducted burst tests with - tubing .
( wall thickness at room temperature) using a flow source of 100 gpm and without a crack sealing system. In these tests, leakage occurs as the tube internal pressure is increased to tube rupture. Tests with tubes containing long axial cracks were conducted. The tube burst pressures from these tests are plotted in figure 2-4. These data suggest that tube rupture equation as derived here ,
is quite conse,rvative.
i 2-6
, . - , . - , . - . . ....:-..---.-...-.--.-.. --.._.-..w--.-,- -.
t Figure 2-3, Westinghouse Axial Defect Burst Test-
- Data and Correlation Q) ,
27
-.oy-=r,-e,+=w_ .e e ,' , ES - w -c. - w!-"-, - -
y
i l
1 i
f e
. L L,
ri U
i 7
Figure 2-4. Com; -ison of Several Tube-Burst Test
- Correlations Alont with Lower Bound Tube
- m . Rupture Equation 2-8" .
4 1
-s., * *-
, , . m , , - , - - , . , , e .- ~ s.-
2.2.2 Multiple Axial Tluough-Wall Cracks The effect of multiple axial cracks on steam generator tube burst behavior has been studied by a number of investigators (6-2). Westinghouse conducted burst tests with tubes containing multiple axial cracks and concluded that the longest crack controls the burst behavior and the presence of multiple cracks does not have a deleterious effect (table D-1 of appendix D). Framatome conducted similar tests with tubes containing multiple artificial defects and reached a similar conclusion (6 a 8).
Framatome concluded that burst pressure was not influenced by the presence of up to cracks located in the same tube section.
Finally, EdF nas performed burst test with tubes removed from service which contained up to service cracks (9). These tests confirm the conclusion that burst behavior is controlled by the longest crack and is not influenced by the presence of _
multiple cracks.
2.2.3 Axial Part-Through-wall Cracks Because part-through-wall cracks may also exist in the roll transition region, data were collected to determine if the critical through-wall crack lengths obtained from Eq. 2-11 would Mund the critical part-through-wall crack burst lengths. Figure 2-5 presents a comparison of part-through-wall crack burst data with the through-wall burst curve obtained using Eq. 2-11. The individual data points shown in figure 2-5 and their s. es a listed in tables D-2 through D-3 of appendix D. Also included for referer 'ata points for the through wall cracks listed in table D-1 of appendiv Figure 2-5 , nat the mrough-wall crack burst curve for a tube prov: ic Tservative bound to the data for both part-through-wall and through-wat d cracks in steam generator tubes (the tube burst curve is _
lower). This information demonstrates that a length based allowable crack criterion is acceptable for predicting burst conditions for both through. wall and part through-wall cracks in the roll transition region of steam generator tubes.
2-9
,. + + <-m...-.a :u u. , pa . a ....c - -u. v. +m.. + no..a .- , . - . . 2 a. ..a.- , - - -. . + + +
i
,3
}
.g
~
.1 4
L
}:
s Figure 2-5. Comparison of Burst Curve for Cracks With
. Experimental Data (Data From Appendix D)-
2-10.
- _ . . . ~. _ . _ .. _
.__-2. -__:.___ _ _ - . .
P
2.3 REFERENCES
. 1. G. Frederick, P.- Hernalsteen, and D. Dobbeni. Belgian Approach to Steam Generator Tube Plugging for Primary Water Stress Corrosion Cracking.
Palo Alto, Calif.: Electric Power Research Institute, March 1990. NP-6626-SD.
- 2. B. Cochet and B. Flesch. " Application of the Leak-Before-Break Concept to
- Steam Generator Tubes." 10th SMIRT Conference, Volume G, August 1989.
- 3. J. A. Begley and J. L. Houtman. Inconel Alloy 600 Tubine - Material Burst and Strennth Properties. WCAP-12522, January 1990.
- 4. P. McIntyre. " Procedures for Specimen Preparation, Leak Rate Measurement and Burst Testing of PWR Steat., Generator Tubing Containing Axial Through-Wall Cracks and Preliminary Res alts for Tests on Fatigue Precracked Specimens." CEGB Draft Memorandum, March 1988.
- 5. F. Erdogan. " Ductile Fracture Theories for Pressure Riser Pipes and Containers." International Tournal of Pcessure Vessels and Piping, vol. 4,.
pp. 253-283,1976.
- 6. B. Cochet. Assessment of the Integrity of Steam Generator Tubes - Burst Test Results - Validation of Rupture Criteria (Framatome Data). Palo Alto, Calif.:
Electric Power Research Institute, June 1991. NP-6865-L, vol.1
- 7. B. Cochet. PWR Steam Generator Tube Fluggine Criteria - Leak-Before-Break Analysis for Primary Water Stress Corrosion Crackine Near the Tubesheet (Framatome Data). Palo Alto, Calif.: Electric Power Research Institute, June 1991. NP-686S-L, vol. 2.
- 8. B. Flesch and B. Cochet. " Leak Before Break in Steam Generator Tubes."
International Tournal of Pressure Vessels and Pipine,1989,
- 9. P. Berger, F. Cattant, P. Calle, and J. P. Hutin. Characteristics of Roll Transition Cracks in Steam Generator Tubes. EdF Report.
2-11 i:
Section ?
CRACK LENGTH MEASUREMENT
3.1 BACKGROUND
Nondestructive evaluation (NDE) is a critical element in implementing a degradation specific tube repair approach, and the inspection method must provide a reliable measurement of critical crack dimensions. Additionally, the inspection process can provide crack growth rate data which is required to establish allowable crack length (see paragraph 4.6).
Conventional bobbin coil technology used for routine steam generator examination has limitations in inspecting roll expansions. Because of the change in tube diameter, this region is the source of a large extraneous signal which can mask signals associated with expansion zone cracking. Other limitations of the bobbin coil include the inability to determine flaw type (volumetric or crack-like), to estimate the number of cracks present or to _ determine crack orientation and length. These limitations can be overcome by using rotating pancake coil (RPC) eddy current.
technology which is the recommended method for inspecting the roll expansion zone when using a length based repair limit. Other inspection methods could be-used if they can be appropriately qualified.
This section discusses industry experience in measuring crack length using RPC technology. Crack length measurement uncertainty, an important consideration in
- establishing the allowable crack length, is discussed in paragraph 4.7.
3.2 ROTATING PANCAKE COIL (RPC) 3.2.1 Data Acquisitinn The development and use of RPC eddy-cunant technology in conjunction with a stable probe delivery system provides an important field diagnostic tool for'-
? measuring EZ PWSCC crack lengths.
RPC tube inspection is accomplished using a surface-riding. coil which is rotated l around the tube axis. Figure 3-1 shows the probe body with details of the coil shown L in end-view in the enlarged schematic. The coilis spring-loaded to maintain contact with the tube inner surface as it moves through the expansion transition region. Lift-off variations between the coil and the tube surface caused by a change -
in tube ' diameter due to the roll expansion are significantly reduced when compared with the bobbin coil.
l.
3-1
z Figure 3-1. Typical Rotating Pancake Coil
- 3-2 ,. l 1
d I
- - - - _ _ _ __-__.__m. __. .._. _ ..____ .__ _ _ _ _. ,_
q j
' As' the probe coil is translated ~and rotated through the tube it describes a helical .
- pathi shown in figure 3-2(a). _ A linear discontinuity within the tube will be scanned once during each rotation of the probe. The coil output voltage from a given rotation is used to generate a line scan which represents signal amplitude as a function of coil position around the tube circumference (see figure 3-2(b)). Image -
- formation (two-dimensional cylindrical coordinate system) is accomplished by
- plotting a series of consecutive line scans with the line-scan generation syn-chronized with probe rotation. This allows for the reconstruction of an image in perspective format as shown in figure 3-2(c). Crack presence is determined by recognizing linear features present in the reconstructed image, and orientation is inferred by noting the direction of the major axis of the image, q Essential acquisition test variables include coil diameter and coil excitation frequency. Multiple-frequency eddy current instrumentation is utilized so that the coil is driven over a range of frequencies in order to independently optimize
-detection, characterization and measurement. Coil diameter which is typically in-
- the range of , along with driving frequency, determine detection arid resolution capability. Based on pulled-tube experience, a lower bound estimate -
on through-wall crack length detection capability for axial or circumferential cracking is approximately Multiple through wall axial cracks must generally be separated by approximately one coil diameter before the-
- separate cracks are readily distinguishable (see section 6).
3.2.2 RPC Data Analysis Crack length measurement is discussed with the aid of figure 3-3 which shows the display screen following automatic analysis of a selected tubescan. In the display --
shown, all measurements are in millimeters. The numbered indications represent crack sequences by axial location of the upper crack tip. The compass orientation shown on the left of the screen denotes the circumferential_ position of the crack.
The' vertical dotted line passing through +00 at the bottom of the screen shows the position of the top of tubesheet which is automatically detected-using a low fre =
- quency applied to the test coll. The measurement shown in the upper right portion of the screen (labeled TTS) is the measured distance from the tube end to the top of f the tubesheet. This absolute dimension is recorded to provide a reference for future inspections.
4 3-3 a . . ~ _ _ :
.. .. . . _ _ . _ . . _ _ . _ . _ _ _ _ . . _ _ _ . . _ _ . _ _ _ _ . _ _ _ _ _ . __._.....___.__...__m--_.___
Y J
p u
v c: ,
y-h.
a
-W a
I-h r --:
f i
f i
i l
4 1 r
I 4
i
'.f h
i ,
'A d'
i 7.
I n
Figure 3-2.: RPC Data Acquisition / Analysis 4
?
{ '.
3-4--
p-4 f
4 . ., , ..
- - w ewe - -
- - , .-n..< ~a-az --es, , . a,-- - - ,a -
a d
A M
%w#
b 2w Q
.2 m
>s 9
c rn T
e d
Vh U
b N-m U-w u:
3 i
A list of detected cracks appears below the tubesheet thickness number. CK# is the sequence number of the crack beginning with the uppermost crack tip. Start is the location of the crack's upper tip with respect to the top of the tubesheet, LEN is the length of the indication. DEG is the circumferential position. P/T is the ratio of the highest signal peak within a given crack indication to the signal threshold value used for detection. The signal displayed as a horizontal chart at the very top of the screen represents the lift off component of the test coil signal caused by the roll transition present in the tubing. The signal displayed as a horizontal chart at the very bottom of the screen represents the signal offset caused by the tubesheet inter-face used to automatically locate this axial position. To the right of that chart is a smaller chart showing the signal used to locate the tube end with the differential bobbin coil positioned on the rotating shaft.
Crack length estimates are accomplished by counting the number of line scans for a given crack. The axial length is then the coil pitch multiplied by the total number of hits or line scans. For crack lengths much greater than the coil diameter, the apparent eddy-current length will, in general, be longer than the true length because -
of the finite coil diameter. Thus, in practice, the total number of line scans is adjusted downward to compensate for this effect. The actual adjustment value is a function of the particular eddy-current procedure used for measuring crack length.
3.3 DEMONSTRATED CRACK LENGTH MEASUREMENT ACCURACY The ability to reliably measure crack length in the roll expansion region of steam generator tubing is an important consideration. This has been assessed using tubes removed from operating plant, , detailed listing of which is provided in appendix C. Results from these studes have compared eddy-current RPC predic-tions of crack length with results detecnined destructively from metallographic examinations and, in some instances, witn dye penetrant estimates of crack length.
This section focuses on crack length measurement accuracy using this industrywide pulled tube database. Linear regression equations relating true crack length (the crack length measured on the inner surface of pulled tubes) and eddy current measured crack length are derived for an overall industry composite data set and for individual organization's specific data sets. The industry composite data set used to describe how the experimental data is used in order to arrive at a length measurement uncertainty for use in establishing a repair limit.
The impact of length measurement uncertainty on the repair limit is discussed in
, paragraph 4.7, with the uncertainty used as one element in the calculation of allowable crack length. In general, crack length measurement uncertainty will be a function of eddy current instrumentation, probe type, calibration procedures and analysis methods. Accordingly, the uncertainty statistics presented in this section may not necessarily be the appropriate basis for establishing a repair limit for a specific plant. As an alternative, specific in-service inspection vendor acquisition and analysis procedures may be qualified / demonstrated using pulled tubes or other L appropriate approaches.
3-6 l
l -- ~ . -- . . . -
l 3.3.1 Industry Composite Data Set A scatter-plot showing RPC-measured eddy-current crack length versus true crack length for a composite data set comprised of 152 points based on tube pulls from France, Belgium, Sweden and Spain (2, a)is shown in figure 3-4. The data set is described by the regression equation y = where y is the NDE measured crack length and x is the true crack length with both y and x measured in mm. The correlation coefficient is r = (Paragraphs 3.3.2 - 3.3.5 below discuss the individual data sets that are included in the composite.)
In establishing measurement uncertainty for repair limit calculations, the regression equation is used as a transfer function to first relate true length to an apparent or _
measured eddy current length. For the composite data set, y = Ata particular y value, which in this case is the measured eddy current length equiva-lent to the maximum allowable length, confidence limits are then calculated using standard statistical methods. The confidence limit is in general a function of the spread or dispersion of the data about the least-squares fit line. This spread can be seen by plotting the residual error--which is the difference between actual and measured lengths--for all the observations or experimental data points that com-prise the data set. A plot of the residual error--in mm-for the composite data set is shown in figure 3-5. As shown in the figure, the mean error or deviation is zero (as measured about the regression line); this is a result of the least-squares fit procedure.
The measurement error is treated globally, e.g., independent of length in order to arrive at a single distribution from which error statistics are derived. A histogram of this error-in mm-is shown in figure 3-6. A typical measure of data dispersion is the standard deviation, s, which for the composite data set is As discussed in paragraph 4.7, the industry composite plot regression equation and the standard deviation are used to calculate NDE error used in the allowable crack length calculation.
3.3.2 Electricite de France Data The EdF data set (2__) used in estimating crack length is shown in figure 3-7 and includes data points derived from pulled tubes. These data are described by the regression equation y = The data are strongly correlated with a correlation coefficient of r = .
3.3.3 LABORELEC Data Figure 3-8 shows data (3) from three tubes (32 data points) removed from two Belgian plants (Doel 2 and 3). These data are described by the regression equation y= ; the correlation coefficient is r = ,
3-7
.. . _ . _ _ . _ . _ _ _ _ _ . _ _ _ _ . . . _ . . _ _ _ . . _ . _ _ _ _ _ . _ _ . . _ . . . . ~ . . . . _ . . . . . _ .
,f \~
's o
t
.: h i
t I
(
8 1
4 i
Figure 3-4. Crack Length Prediction - Composite Industry Experience t I
f e
' .m ,
t r
f ll '
i 4
- i. 3-8:
,_ ,~. . .. . -
s .. ._ _ _ _ . . _ _ - . . . , - _ . . _ _ . _ _ _ _ . . . . . . _ . _ . . _.
e
- 5 Figure 3-5. Crack Length Prediction - Standardized Residual Plot i ...
+
4 I
I-3-9
, , . . ., , : a. ;, .. ; . . . = . . .. . . .
a bha D %\.4,.&-H4 a.ma-- -L-~ sc nA*A ,2,L-- -%m-Ms3A**N- 5~E, *J*A b 'FA -5 AB*L-' e>& 4 A *3 J- - M 4-'
m Q L 4
b m
N' U
.T .
Q' a
M .
~
h A i b
. g C
ene -
0
.C-O
-.U.-
/
- -].
w .
b
. esip -
=
.. v -
-M 6-
,U
- ,,1~
_w.
3 '
. ;7
. m i'
1-'
l
' 3-10
. . . _._ - -d L' - - - , ~. , ,-..,_ . , . , , . _,; . __a,_,, , , , , , . ,
l Figure 3-7. Crack Length Prediction - Electricit6 de France 3-11
.l
~
- l i
b t.
Figure 3-8. Crack Length Prediction'- LABORELEC 7
s i
3 . .;
l 3.3.4 Swedish State Power Board Data The Swedish State Power Board, in conjunction with Zetec, recently conducted extensive RPC repeatability and accuracy studies on estimating crack length using data derived from sixteen cracks on four pulled tubes (4). The average error in predicting crack length was quoted as being conservative. A regression analysis for the four tubes (28 data points) removed from Ringhals 3 and 4 is shown in figure 3-9. These data are described by the regression equation y = ; the data set correlation coefficient is r =
3.3.5 Scanish Data A scatter plot for data (nine cracks) from three tubes removed from Spanish units (E) _.
is shown in figure 310. These data are described by the regression equation y = with a correlation coefficient of r = .
3-13.
. . . . . . -. . . - . . - , . ~ . - .. , . _ - . . . - . .
--i m
.- j
.i t
L I
t Figure 3-8. Crack Length Prediction - Swedish State Power Board e l
(
x h
I_-
3-14
.- . ~ , _ _ c- _ . . _ . . ,
~
l
-i e
d
?
m
.,4
-._~
Figure 3-10. Crack Length Prediction - Spanish Utilities :
e l 3-15 r-
- - , .-..., , . . . . . ~ . . - , .,, - . . - , . .
. , . , ~
i __.
q
3.4 REFERENCES
- 1. H. Houserman et al. " Development and Field Use of a Tubesheet Scanner System for Primary-Side Crack Detection and-Measurement." EPRI Eighth Annual Steam Generator NDE Workshop, August 1989.
- 2. F. Cattant. "Mesure de la Longueur des Fissures Longitudinales de Corrosion Sous Contrainte en Milieu Primaire Situses dans la Zone de fin de Dudgeonnage Mdcanique des Tubes." Electricit6 de France Report D.5004 /CTT/RB.90.52, May 1990.
- 3. D. Dobbeni. " Implementation of a Fast and Accurate RPC Method for a 100%
Inspection of Roll Transitions." EPRI Workshop on PWSCC Remedial Measures, Clearwater Beach, Florida, February 1989.
- 4. B. Wilson. Qualification of New MRPC Equipment for Crack Length Measurement in Roll Transition Zone in Ringhals 3 and 4 Steam Generators.
Swedish State Power Board Report, November 1989.
- 5. L. Gonzalo. Results of the Pulled-Out Tubes from Steam Generators of Spanish Nuclear Power Plants. Tecnatom Report Cl-21-01, November 1989.
l 1
-f 3-16
.~ . _ __ - . . _
Section 4 ALLOWABLE CRACK LENGTH 4.1 APPROACH This section presents methods and criteria for calculating the allowable axial length of EZ PWSCC roll transition cracks in recirculating steam generator tubes made of -
alloy 600. The allowable axial crack length is the length of the largest axially oriented crack that can be left in service until the tube is next inspected. This length .
i includes margins on load, consideration of variable material properties, a correction for tubesheet constraint, an allowance for crack growth during the time period
~ between inspections and an allowance for crack length measurement uncertainty.
The methods described in this section apply to axial cracks and inclined axial cracks that are located near or above the top of the tubesheet (figure 1-1, cases A and C).
This document can be used for inclined axial cracks when the circumferential extent of the inclined crack is equal to or less than the axial extent of the inclined crack.
The applicable crack conditions are illustrated in figure 4-1. This document cannot be used for evaluation of NDE indications evaluated to be distinct circumferential cracks. Tubes with identified distinct circumferential cracks should be repaired.
For tubes having multiple axial cracks, additional conditions on spacing between cracks apply as described in section 6. Evaluation of cracks in partial-depth rolled tubes (figure 1-1, case B) is described in section 7.
The steps for calculating the allowable axial crack length are shown schematically _in figure 4-2. The first step is to define the tube rupture equation which describes the relationship between crack length, tube dimensions, loads, and material properties.
With appropriate input parameters, including safety factors, the tube rupture equation is used to calculate a reference crack length,5. Three corrections to the reference crack length 5 are needed. These are a correction for tubesheet constraint
. (aTS), an allowance for crack growth until the tube is next inspected (aCG) and an allowance for crack length measurement uncertainty (aNDE). The allowable crack length, A,is found by adding the correction for tubesheet constraint and subtracting the allowances for crack growth and crack length measurement uncertainty from the reference _ crack length,3.
I; 4-1.
o N.1
~
'b1 i
1 l
,\
i Case A: Axial Crack -
r Case B: Inclined Axial Crack _
i I
Figure _4-1. Illustration of Allowable Crack Lengths for
,-- . Axial and Inclined Axial Cracks
~4-2
+ W GCh-FmO 9- c e '%- F % ree-dr-* *- **MWE- P 45 W ? fr
. .- . . . . - - . .. . ,. .. ... . ... ~. ..- . . _ -. .
- 4 5
1; 4
Figure 4-2. Allowable EZ PWSCC Axial Crack Length Calculation Steps-
. :- ' 4-3 r=,
,e - ,.. u - -
4 4.2 TUBE RUPTURE EQUATION In section 2, major tube burst studies were reviewed and summarized in addition, .
the tube rupture equations proposed and used by the different investigators were presented. The equations were plotted using a common format, and a curve which:
bounds these relationships was developed. It is this curve _which is used here as the-tube rupture equation (a description of these controlling equations and the resulting burst curve are repeated below--Eqs. 4-1 through 4-3 and figure 4-3).
Key parameters for the tube rupture equation are dimensionless pressure, P and dimensionless crack length, A.
P=
(4-1) where P is the tube differential pressure, Su and S y are the material ultimate and yield strength, respectively. R and t are the tube mean radius and thickness as shown in figure 4-1.
A= (4-2) where 5 is the reference crack length.
For an axial crack in the free span of a straight tube away from the tubesheet, the steam generator tube burst correlations can be bounded by the equation:
Or (4-3)
A=
This equation is shown plotted in figure 4-3 and is valid for Eqs. (4-1) through (4-3) or Eqs. (4-1) and (4-2) in combination with figure 4-3 can be used to solve for the burst pressure given an axial crack length or the critical axial crack length given the tube differential pressure.
-1 l
l I
l 4-4
5 L
t 3
J l
4 i.
- Figure 4-3. Tube Rupture Equation in Terms of. .
- Dimensionless Burst Pressure and Critical Crack Length-f
]
4 4
_. l
4.3 LOADS AND SAFETY FACTORS Tube rupture depends on hoop membrane stress. In addition, for straight tube sections, hoop membrane stress is almost entirely due to the differential pressure between the inside and outside of the tube q). For normal operating conditions, tube differential pressure is the difference between the primary system pressure and the secondary system pressure. Maximum differential pressure for faulted condi-tions is generally associated with the design basis accident of a main steam line break (MSLB) or feedwater line break (FWLB). In these accidents, the primary system pressure could be higher than normal operating conditions with secondary system pressure effectively ambient.
USNRC Regulatory Guide 1.121 (3) recommends that when establishing tube plugging limits based on depth, that safety factors be applied to the tube load.
Recommended safety factors are:
- 3 under normal service conditions.
- A value consistent with the limits set by the ASME Code,Section III, paragraph NB-3225 for accident conditions. In compliance with NB-3225, this factor is taken as 12 for accident conditions (emergency and faulted conditions, service levels C and D).
In calculating the reference crack length,it is recommended that safety factors of 3 be applied to normal operating differential pressures and a factor of 15 be applied to accident differential pressures. The smaller of the crack lengths calculated with these two pressure terms is the reference crack length, a.
4.4 MATERIAL PROPERTIES AND TUBE DIMENSIONS The material property controlling tube rupture is the flow stress, i As discussed in section 2, the flow stress is a simple function of the sum of the material yield and-i ultimate strengths. The tube rupture equation is in terms of the tube material yield j plus the ultimate strength (Sy + Su).
I Three sources of tube yield and ultimate strength values are suitable for calculation of allowable axial crack lengths. If properties of the actual tube have been measured, l
these should be used for that tube. Further,if these properties were measured at room temperature, they should be adjusted for steam generator operating tempera-tures. Operating temperature yield strength should be taken as .
of room tem-perature yield strength while ultimate strength at temperature should be- of room temperature ultimate strength (4).
4-6
7 h
- ^= seconld possibi_lity is to use 95%/95% lower bound estimate
- of yield and ultimate <
- ' strengths for the different heats'of material found in a steam generator. This' _ ___
- eliminates the need to identify the heat of each individual tube. If lower bound room temperature yield and ultimate strength properties are used, the properties should be corrected for temperature as is the case where individual tube properties are used.- .
A third source of steam generator tube mechanical properties is the lower botmd
-estimate based on Westinghouse testing of multiple heats of material (4). If steam _
generator specific data are not available, these properties can be utilized.
Westinghouse mean and 95%/95% lower bound estimates of-mechanical properties for mill-annealed tubing are given in tables 4-1 and 4-2.
- In the absence of an identified wall thinning mec hanism, tii s recommen ed d that nominal tube dimensions be used for calculating the reference axial crack lengths.
For tubes, these dimensions are R = -and t= For tubes, R = and t= .
4.5 CORRECTION FOR TUBESHEET CONSTRAINT" The tube rupture equation described in paragraph 4.2 describes.the burst behavior for an axial flaw in a straight section of tube away from the tubesheet or support ,
plates. With appropriate safety factors and tube mechanical properties, the tube -
rupture equation yields the reference flaw length,5.
The presence of the tubesheet near EZ PWSCC influences the tube burst behavior. :
For a given flaw length, the tubesheet presence results in higher burst pressurei for a-given tube pressure, the presence of a tubesheet results in larger critical flaw sizes.
. The effect of tubesheet constraint decreases as the distance from the crack tip t.o the tube-to-tubesheet contact point increases.
- Limit above which 95% of data would be expected with 95% confidence.
"Altemate tube repair limits have been proposed which would modify the substance of this section and the basic methodology presented in the body of the report. For example, an alternate criteria for all indications could consist of only comparing the crack length above the top of the tubesheet to an allowable crack length based on an associated, smaller tubesheet constraint factor than that
- recommended in this document. The smaller tubesheet constraint factor would need to be demonstrated -
- by burst testing of tubes with EDM slots or SCC that extends below the top of a simulated tubesheet at
- least, for example, as far as the length of the exposed crack. NDE crror and crack growth rates would'
- also need to be addressed.
4 _
a y
Table 4-1
[- ALLOY 600_ STRENGTH PROPERTIES MILL-ANNEALED _ TUBING Q) .
TEMPERATURE, F Sample Size of Population, N Tolerance Factor, K .
Yield Strength, Sp, kai Sample Mean,i Standard Deviation, s -
957o/95% Lower Tolerance Limit, X-Ks Ultimate Strength, Su, ksi Sample Mean, R Standard Deviation, s 957e/95% Lower Tolerance Limit,I-Ks Sv+Su, ksi
' Sample Mean,I
~ Standard Deviation, s 95%/95% Lower Tolerance Limit,I-Ks 9-4-8
l Table 4 2 ALLOY 600 STRENGTH PROPERTIES MILL-ANNEALED TUBING (O TEMPFJMILTil F Sample Size of Population
Tolerance Factor, K Yield Strength, Sy, ksi Samnie Mean, X Stanitard Deviation, s 957e/95% Lower Tolerance Limit, X-Ks
~ Ultimate Strength, Su, ksi Sample Mean,i Standard Deviation, s 95%/95% Lowe- Tolerance Limit, E Ks Sv4 Su,k !
Sa.nple Mean, i Standard Deviation, s 957c/95% Lower Tolerance Limit, E-Ks 4-9
To take credit for tubesheet constraint, the crack must be in the vicinity of the tube-to tubesheet contact point. In the calculation of the allowab!e crack length,it is assumed that the reference axial crack effectively spans the roll transition.' In this case, the lower end of the crack will be at the bottom of the roll near the contact point.
If during in service inspection it is noted that a crack indication does not extend to l the bottom of the roll transition: l
- The tubesheet constraint term should be neglected (assumed to be zero) ;
in the calculation of the allowable crack length, or
- The distance from the lower end of the crack indication to the bottom of the roll transition should be added to the indicated crack length and this length compared to the allowable crack length calculated with tubesheet constraint.
The effect of tubesheet constraint has been experimentally studied and an empirical relationship describing the strengthening effect developed. The French _(1) and Spanisn (fi) have accounted for tubesheet constraint by modifying the bulging factor in the limit load equation based on the length of crack outside the tubesheet and an effective tube thickness which accounts for the interaction with the tubesheet.
The Belgians (D developed an empirical fit to their tubesheet constraint data using '
an exponential fall off of reinforcement with crack length. This fit is shown in '
figure 4-4.
A tubesheet constraint correction for tubes was derived from the tubesheet constraint relationship developed by BELGATOM (D.
This correction is designated as aTS and is the crack length to be added to the reference crack length to account for tuocsheet constraint.
For tubes:
- aTS = (4-4)-
- aTs =
or ; aTS =
For tubes:
- aTs = (4-5)
- aTS a or ; aTS =
'This is a reasonable assumption given the size of the reference crack length, on the order of 4-10
-i
,+- .-,..,.%,, -._,n.,_,['_.um., ... . . _ _ , , ,_,.U.,... m,., .r., J_y .... , , m. ,,. d . m., .- , , , , , , . , . - . . . . - 5,,._,,
b These relationships are also shown in figure 4-4.
I I
t I
Figure 4-4. Allowance for Tubesheet Reinforcement m
4.6 ALLOWANCE FOR CRACK GROWTli As discussed in section 3, inspection with the RPC eddy-current probe will identify EZ PWSCC and provide information on location, orientation and length. The .
reported length will subsequently be compared with the allowable crack length, and
'if longer than the allowable, the tube .will be removed from service. ,
There are two corrections to the reported crack length that need to be considered.-
These are a correction for crack growth that is expected to occur until the time the tube is next inspected and a correction for the crack length measurement uncer- .
~ tainty. These corrections could be handled by adding an incremental crack lengtl' for each correction to the indicated crack length, or alternatively, subtracting the incremental crack lengths from the reference crack length in the allowable crack _ . 1 length calculation. It is this latter approach which has been used by others and is used here. The allowance for crack growth is discussed below and the allowance for
- crack length measurement uncertainty is discussed in paragraph 4.7. .
The allowance for crack growth is developed by estimating EZ PWSCC crack growth rate and multiplying the rate by the time interval until the next planned inspection.
Further, EZ PWSCC crack growth rates can be estimated from sequential measure-
= ment of EZ PWSCC cracks in service. Individual growth rate data points obtained in-this manner are given by" 4-11
. .u , , , , ~ . . ., , , . . , - -, , _n ,- - ,. ,, , ~ . , -i,-
I a2 - ai a tirne (4-6) i t
where a2 is crack length at time 2, a1 is crack length at time 1 and a time is the '
operating time interval between inspections performed at time 1 and time 2.
t Appendix F discusses available data from which EZ PWSCC crack growth rates can t be estimated. It is noted that only recently has limited data for standard roll tran-sitions become available although more cxtensive data for kiss rolls has been avail-able for some time. Consequently, crack growth rates for EZ PWSCC cracks have been estimated from kiss roll transition data. ;
Crack growth data for kiss roll transitions is abundant because of the Belgian practice of performing RPC examinations of 100% of the roll transitions at each annual outage. Because of the large number of data points from several plants for several years, the data provide a good basis for establishing crack growth rates for plugging criteria development. In addition,it is considered that growth rates measured for kiss rolls are conservative for standard rolls. - This is based on the information that in_dicates that crack growth rates in kiss roll plants may be higher than in standard j roll plants particularly for short crack lengths (see appendix F). l Using a fitted and smoothed curve to the average values of the Belgian kiss roll data presented in Appendix F results in the values given in table 4-3 and shown in figure 4-S. Note that growth rates are in terms of effective full-power years (EFPY).
Table 4-3 ESTIMATE OF EZ PWSCC AVERAGE CRACK GROWTH RATES (BASED ON KISS ROLL DATA)
Initial Lencth Averace Growth Rate l
t' l'
1 4-12 L
I
...-,,_m ..,...__.,_,__._...__,...~,,_,,_,,._,,,_.__.,__.m..,._ . _ _. #. 4 -
i Figure 4-5. Estimate of EZ PWSCC Average Crack Growth Rate Curve (Based on Kiss Roll Data)
Until plant-specific data and a detailed statistical analysis (which includes estimates of crack length measurement error associated with the particular inspection methods ured to sequentially measure crack length) become available, it is recom-mended that the avcrage crack rate information given in table 4-3 and shown in figure 4-5 be used for calculating the allowance for crack growth. This recommen-dation is justified as kiss roll crack growth rates may be higher than rates in standard roll transition zones and the apparent variability in crack growth rates is likely dominated by crack length measurement uncertainty which is accounted for separately in the allowable crack length calculation.
4.7 ALLOWANCE FOR CRACK LENGTH MEASUREMENT UNCERTAINTY Eddy-current measurement of EZ PWSCC crack length is discussed in section 3.
Crack length measurement capability is presented in terms of a sizing model where the sizing model is a regression equation developed from NDE performance data i 4-13
relating true crack length for cracks of various sizes to an NDE measured dimension obtained using the applicable NDE equipment, procedure and analysis method.
These data are generally developed from tube pulls and/or round robin exercises.
The regression line is defined by the equation y= (4-7) where y is the NDE measured value of the crack length, x is the true 1. D. crack length (or corrected reference crack length),
b is the intercept of the regression line with the y axis, k is the slope of the line defined by the regression equation, and s is the standard deviation or data scatter about the regression.
The sizing model regression line is showr
- matica"v in figure 4-6.
From the standpoint of efficiency in tube L n ; . n.p4 .3y those tubes that could fail unacceptably, the ideal sizing moi. m ~ f.te a slope, k, of 1.0 and a y axis intercept, b, of 0, with no data scatter; how, rer, from the standpoint of reliability, i.e., assuring that all such tubes be removed, ,r is only necessary to know what the slope, intercept and data scatter are. These are then accounted for in the determination of the allowance for NDE uncertainty, discussed below.
As is the case with the allowance for crack growth between inspections, the allowance for crack length measurement uncertainty, an will be one of the terms subtracted from the reference crack length to obtain the allowable crack length.
Further, as the regression formula shows, the eddy current measured crack length depends on the true crack length. Consequently, the first step in calculation of an is to calculate the allowable crack length without an allowance for crack length measurement uncertainty to use as the true crack length in Eq.4-7.
This crack size is the reference crack size,5, plus the correction for tubesheet constraint, alS, minus the allowance for crack growth,acc X= (4 8)
Referring to figure 4-6, for a given true crack length x (which in our case is the corrected reference crack length), the regression equation can be used to calculate the expected NDE crack length, y, and the lower bound NDE crack length, yt.B. The systematic NDE error is the difference between the true crack length and the expected NDE crack length, x-y. The random error is the difference between the expected NDE crack length and the 'ower bound NDE crack length, Taking the standard deviation, s, as the 95%/95% confidence interval, 4-14
._ .._ _ . . . _ _ . - _ . _ . _ _ - . _ _ _ - _ . _ ~ _ _ _ _ . . .
1
)
1 I
i Figure 4-6. Definition of Crack Length
- Measurement Correction asur _
The NDE crack length measurement uncertainty is the sum of the systematic error,
, and the random error, anor =
" (4-9)
Calculated in this way, the lower bound NDE crack length is, by definition, the allowable crack length, A, since A=
(4-10).
l n
l' 4-15
4.8 ALLOWAULE CRACK LENGTil SAMPLE CALCULATION Calculation of allowable EZ PWSCC crack length follows the process shown in figure 4 2.
The first step is to calculate a reference crack length, H. Information needed to establish reference cracks length includes steam generator tube nominal dimen-sions, tube mechanical properties and normal operating and faulted condition tube primary-to secondary differential pressures. Applying a safety factor of 3 to normal operating differential pressure and a factor of V2 to faulted condition differential pressure, a reference crack length is calculated from Eqs. 4-1,4-2, and 4 3, or from Eqs. 4-1 and 4-2 and figure 4-3. The smaller of the calculated crack length from the two differential pressure terms is utilized as the reference crack length. Generally, normal operating differential pressure will be the controlling load.
As an example of this calculation, consider a tube with nominal dimensions of R = and t = . Further, assume that mechanical properties for neither the particular tube nor the steam generator are available and that industry lower bound properties (provided in paragraph 4.4) are used. In this case, Sy+ Su = Assuming that the normal operating differential pressure is and faulted load dif-ferential pressure is , the controlling load will be three times the normal operating differential pressure. Using this load, the reference crack length (3) from Eqs. 41 through 4-3 or figure 4-3 is Correction for tubesheet constraint (aTS)is found from Eq.4-S or figure 4-4. For this exampie of a tube with 3 equal to , aTs is In the absence of plant specific crack growth data, allowance for crack growth (aCG)is found from table 4-3 or figure 4-5. Assuming that it will be one effective full-power year until the tube is next inspected and utilizing the result that K + aTs =
,acc = . The allowable crack length without an allowance for crack length measurement uncertain' ,
X* =
(4-11)
For this example, the composite crack length measurement error statistics presented in Section 3 are used for calculating an allowance for crack length measurement uncertainty (ayot). Substituting the value of allowable crack length calculated above in equation (4-11) for x into the regression equation presented in paragraph 3.3.1 gives y=
4-16
As polm'd out in paragraph 3.3.1, the random error in measuring crack length is characterized by the standard deviation of , Thus, the The NDE crack length measurement uncertainty is from Eq. 4-9, the sum of the systematic error (x y) which for the allowable crack length is equal to plus the random error or aNDE" Thus the allowable crack length, A,is the reference crack length, a, plus the correction for tubesheet constrain, aTs , minus allowances for crack growth, acc, and crack length measurement uncertainty, a wDt,. For the tube example-A= (4-12) j The results of this sample calculation for a tube are provided in table 4-4. A similar sample calculation for a tube is included in this table as well. These results demonstrate how the calculations should be '
performed. --Plant-specific calculations may be carried out using the information available in th(s report supplemented by available plant-specific data.
Tabic 4-4 SAMPLE CALCUATIONS OF ALLOWABLE CRACK LENGTH FOR TUBESt R :
t ,
Sy+ Su Nominal AP Faulted AP a
ars acc
-aNDE A
'The results are not representative of a particular steam generator nor >
should they be considered typical. ,
p 4-17 R
p l . _, J.--, .,,,_...,J.,._,._,
, . , . . . _ _ . . - . , _ _.,,..J.,- ,__a.
j
4.9 REFERENCES
- 1. G. Frederick, P. Hernalsteen, and D. Dobbeni. Belgian Approach to Stea_rn '
Generator Tube Pluccine for Primary Whter Stress Corrosion Crackinc.
~
I Palo Alto, Calif.: Electric Power Research Institute, hiarch 1990. NP-6526-SD. l
- 2. P. DeRosa et al. Evaluation of Steam Generator Tube. Tube Tubesheet and Divider Plate Under Combined LOCA Plus SSB Conditions. WCAP-7832, April 1978.
- 3. U.S. Nuclear Regulatory Commission Regulatory Guide 1.121. " Bases for Plugging Degraded PWR Steam Generator Tubes." Draft for Comment, August 1976.
- 4. J. A. Begley and J. L. Houtman. Inconel Alloy 600 Tubinn-Material Burst and Strength Properties. WCAP-12522, January 1990.
- 5. B. Cochet and B. Flesch. " Application of the Leak-Before-Break Concept to Steam Generator Tubes." 10th SMIRT Conference, Volume G, August 1989.
- 6. C. Cueto-Felgueroso and J. M. Garcia Alonso. Criteria for Plugginn of SG Tubes in Model D3 Steam Generators Based on the Leak-Before-Break Criterion Applicable to Stress Corrosion Crackinn. Tecnatom Report: EPRI-CT-D3, July 1989.
4-18
...a-... .
l Section 5 LEAK RATE CONSIDERATIONS 5.1 LEAK RATE UNDER NORMAL LOADS 5.1.1 1&LkJ1 ale.3 To evaluate the potential for leak-befote-break (LUB), axial through-wall crack length is correlated with leak rate for normal operating pressure.
A number of laboratory leak tests with primary water stress corrosion cracks have been conducted at normal operating pressure using tubes with axial corrosion cracks in the tube expansion zones, and the results are presented in figure 5-1. Also shown in figure 5-1 are data representing in-service leaks from tubes with axial corrosion cracks in the expansion zone. The leak rate data and the data sources are listed in appendix E. The indicated crack lengths are the reported nominal values or the average of the reported ID and OD crack lengths.
The data in figure 5-1 include experiments conducted on several heats of tubes at pressures rangirig from . No effort was made to normalize these data in a manner similar to that shown in figure 2 4.
The data in figure 5-1 indicate a wide range of leak rates at crack lengths of interest in this work. This range may be due individually or in combination to effects such as tube sheet reinforcement, the presence of unbroken ligaments along the crack face, location within the sludge pile, and/or crud buildup at the crack. Additionalin-service experience indicates that even some relatively long cracks (12 mm or longer) in the roll transition region may not leak at all during normal operation (D.
To assess the potential for LBB behavior over the leak rate range indicated in figure 5-1, two curves are presented. One is the best fit mean curve (using the laboratory data and assuming a relationship in the form Q ), while the other represents the mean leak rate divided by 10 at any specified crack length. The lower curve encompasses most of the data and represents tubes where the leak rate may be restricted due to the presence of unbroken ligaments along the crack, crud build-up, etc. Both curves are considered in the following LBB assessment.
5-1
Figure 5-1. Leak rate versus nominal or average axial crack length for steam generator tubes with corrosion cracks in the expansion zone roll
- transition region, (data from appendix E).
.5-2
5.L2 Leak Befomlhnk To establish the allowable leak rate at normal operation that would preclude failure at faulted load, the critical crack length at faulted pressure first must be determined from Eq. 2-11 (or figure 2-4). This crack length is then used to enter figure 5-1 and determine a leak rate at normal operation. If this leak rate is greater than the plant allowable leak rate limit then leak before break at faulted load is assured.
As an example, consider e typical tube where mean radius, R, equals and wall thickness is . A conservative value for the sum of the yield and ultimate stresses at temperature for this size tube has been determined as (2) (see table 4 2). Substituting these values into Eq. 2-11 and including the tubesheet constraint correction described in paragraph 4.5, the critical crack size at a faulted pressure of is . Using similar procedures and a value of for the sum of ultimate and yield stress (2)
(see table 4-1), the critical crack size for a tube is .
Using average material properties, critical flaw sizcs at a faulted pressure of are and for tubes, respectively.
If the conservative assumption is made that the leak rate at normal operation is produced by a single crack, then critical crack sizes determined from Eq. 2-11 can be used in figure 51 to define the allowable leak rate at normal operation for a corrosion crack in any region of the tube bundle.
As seen from the mean curve in figure 51, crack lengths of and
) correspond to leak rates of about .
respectively. For these conditions, the typical Technical Specification limit of 0.35 gpm generally ensures LBB at faulted load.
However, when the leak rate at normal operation is determined from the lower curve, which may be representative of some in service leak rates of cracks in a roll transition, the indicated allowable leak rate at normal operation ranges from about
. for a long crack to for an long crack. This leak rate range is significantly less than typical Technical Specification limits. To impose such leak rate limits from the lower edge of the leak rate data may cause undue restrictions on plant operation and result in unnecessary plant outages, radiation exposure, and cost of repair. In addition, because some-or longer cracks may not leak at normal operation (1) it is not feasible to ensure LBB for all tubes by reducing the leak rate limit.
While it is not feasible to have a leak rate limit that ensures LBB for all tubes, leaks may occur during service (even from cracks in regions other than the roll-transition). This possibility requires an assessment to be made of an appropriate leak rate limit. To determine this limit consider the potential for LBB at normal i
5-3
operating pressure. Eq. 2-11 gives a critical through-wall crack length at normal operation of about for both the tubing.
Using the curve of figure 5-1, the leak rate at normal operation for a long through wall crack is about . To ensure margin exists for LBB at normal operation when an in-service leak is detected, a leak rate limit of is recom-mended. Using a leak rate as the leak rate limit assures a crack length of using the curve in figure 5-1. The minimum burst capability would exceed for either size tube for this crack length. In addition, the on line burst data described in paragraph 2.2.1 indicates that the Belgian burst curve is more representative of in-service burst capability than the lower bound curve represented by Eq. 2-11. On the basis of the Belgian burst curve, the burst capability of the tubing with a long crack is at least . LBB at postulated accident pressure levels therefore is essentially satisfied.
To provide a high level of confidence that tubes that may not exhibit LBB behavior at faulted load are removed from service, RPC inspection will be performed for 100% of the tube roll transitions of expansion zones in regions of the steam generator (i.e.,- the hot leg) that are affected by EZ PWSCC. Tubes with crack lengths greater than a conservative allowable length (see section 4) will be repaired.
5.2 LEAK RATE AT POSTULATED FAULTED LOADS 5.2.1 Ilghgtntnd The Standard Technical Specifications and many plant Technical Specifications limit the allowable primary to secondary leakage through all steam generators not isolated from the reactor coolant system (RCS). The allowable limits are based on the following considerations:
- The total steam generator tube leakage limit of for all steam generators not isolated from the RCS ensures that the dosage contribution from the tube leakage will be limited to a small fraction of 10 CFR 100 limits in the event of either a steam generator tube rupture or steam line break event. The limit is consistent with the assumptions used in the analysis of these accidents.
- The limitations on the specific activity of the primary coolant in the plant technical specification ensure that the resulting doses at the site boundary will not exceed an appropriately small fraction of 10CFR100 limits following a steam generator tube rupture in conjunction with an assumed steady state primary to-secondary leakage rate of .
l
- The leakage limit per steam generator is intended to assure that steam generator tube integrity is maintained 5-4
in the event of a main steam line rupture or under LOCA conditions.
Permitting operation with leakage in excess of this limit increases the potential that steam generators may be vulnerable to tube rupture during a postulated steam line break event.
Because current tubing plugging criteria are aimed at precluding penetration of the tube wall, significant leakage is not expccted even at faulted load. However, because the objective of this work is to justify the presence of through wall and/or deep part through-wall cracks, it is recommended that the leakage at faulted load be assessed.
This assessment would identify the leak rate that must be tolerated when deter-mining if the site boundary dose remains an acceptably small fraction of the 10CFIU00 limits during postulated accidents. This leak rate is site-specific and, among other factors, will depend on the number and length of the cracks allowed to remain in the steam generator tubes subsequent to inspection.
The accidents that should be considered are those that assume that primary to secondary leakage is combined with secondary steam release to the environment (e.g.,
steam line break, locked reactor coolant pump rotor, SG tube rupture, control rod ejection, loss of load / loss of off-site power). For many plants, the limiting event with regard to primary to secondary leakage will be steam line break. For some plants, the locked rotor event may be more limiting depending upon the number of assumed fuel failures and the assumed iodine partition coefficient. To assure that the most limiting condition is addressed relative to primary to secondary leakage, all accidents which combine leakage with steam release should be considered.
This section outlines a method that can be used to assess the leak rate during a postulated steam line break (SLB) for any specified crack distribution (number and length of cracks). Several examples are evaluated to illustrate its application and obtain an estimate of the range of leak rates that can be expected for various EZ PWSCC distributions.
5.2.2 I eak Rate Calculations The method used to predict leak rate uses clastic plastic fracture mechanics analysis to estimate the crack opening for pressure loading. Thermal-hydraulic analysis is used to predict the mass flow through the crack opening. The analysis includes consideration of the crack shape, crack growth, and resistance to flow through the crack opening. To eliminate unnecessary conservatism, statistical methods are used to model ranges in the variables that affect leak rate, and to compute the total leak rate at faulted load for any specified crack distribution.
The distributed variables include material strength, crack <"itribution, crack growth, tube wall thickness, and ECT sizing error. The statistical analyses are performed using Monte Carlo sampling techniques. A detailed description of the analysis method is presented in appendix G.
55
~
-t There are several uncertainties associated with the analysis methods, including the flaw shape factor (i.e., the ratio of the ID crack length to the OD crack length) and the flow discharge coefficient (the resistance to flow through the crack). To obtain an accurate assessment of these variables,it is necessary to bench mark the leak rate ,
calculational methods with experimental data. Currently, there are limited experi-
, mental data that can be used to define leak rate at faulted pressure and bench mark -
computational procedures; however, experiments are now being performed to obtain a correlation between axial crack length, faulted pressure and leak rate (see appendix G). When this work is complete, the revised data can be used to support plant specific analyses. ,
t To estimate the leak rate prior to obtaining data to bench mark the computational ,
procedure, sensitivity studies were performed for two example crack distributions.
These distributions were based on available plant inspection results and are con-sidered typical of distributions that may result from EZ PWSCC. The dis'oributions are presented in table 51, and represent a steam generator returning to service subsequent to inspection where all tubes having cracks greater than an assumed allowable length of have been removed from service. Each distribution has '
cracks; the average crack length is about for distribution A and about for distribution B. The tubes are assumed to have a wall thickness, Leak rate calculations with the LABOLEAK program (2,1), were performed for each i distribution using a best-estimate flaw shape parameter ( from destructive -
examination of pulled tubes) and two curves of the flow discharge coefficient, which are believed to approximate the actual field behavior (see curves 1 and 2 in figure 1 G-2 of appendix G). The results, presented in table 5-2, indicate that a distribution of cracks where the mean crack length is about (distribution A) will have a leak rate of about for curve 2.
Table 5-2 also indicates that the leak rate at faulted load associated with distribution B in table 5-1 is higher than normally assumed in the plant safety evaluation. ' To assess the effect on leak rate as a function of the largest crack length allowed to remain in service, distribution B in table 5-1 was modified by reducing the plugging limit from . The leak rate calculations were repeated and the computed leak rate was reduced from - , a factor of less than These results indicate that for distributions with approximately cracks where the average crack length is about , the leakage at postulated faulted load will be in excess of that normally assumed in the plant safety evaluation.
Attempts to reduce the leakage to within assumed limits by reducing the plugging limit may not be practical, and site-specific calculations may be necessary to demon-strate that the site boundary dose associated with faulted load remains within accept -
able limits. It should be noted that assuming the crack distribution is unchanged,-
the leak rate will be proportional to the number of cracks present in the steam generator; i.e., cracks will result in of the leak rates shown in table 5-2.
5-6
-en s ,,-Am.w.~ .,,s.,-m u,, mA.- ,e,-,- -- A n m ,,,.,,,rc
i Table 5-1 EXAMPLE CRACK DISTRIBUTIONS FROM RPC INSPECTIONS FOR CALCULATION !
OF LEAK RATE AT FAULTED LOAD'FOR TUBES i
Axial Crack I ength' Number of Cracks" Distribution A Distribution B t
i
' Assumed repair limit for this calculation is .
" Total number is 1000 cracks.
Table 5-2 ESTIMATED LEAK RATES FOR POSTULATED FAULTED LOAD FOR TWO EXAMPLE FLAW DISTRIBUTIONS ,
Leak Rate -
Flow Discharge Coefficient" .
Repair Limit Distribution Curve 1 Curve 2 ,
A.
B B
bak rates are based on the crack lengths in table 51 adjusted for in-service crack growth to the end of the upcoming operating cycle.
" Refers to curves 1 and 2 as described in figure G-1 of appendix G.
- "This calculation is detailed in appendix G (figures G-5 through G-11).
5 i
5.3 REFERENCES
- 1. P. Hernalsteen. Steam Generator PWSCC In Service Leakage Experience in Belgium. LABORELEC Report hiOO 90 022/R PH, hiay 1990.
- 2. J. A. Begley, and J. L. Houtman. Inconel Alloy 600 Tubing - hiaterial Burst and Strength Properties. WCAP-12522, January 1990.
- 3. P. liernalsteen. LABOLEAK: Iaborelee Leaktate Probabilistic Assessment Program for PWSCC Crack , ' team Generator Tube Roll Transitions - User's hianual. LABORELEC Rep 400 90 023/R PH, hiay 1990.
- 4. P. Hernalsteen. LABOLEAK: Laborelec Leakrate Probabillsjic Assessment Program for PWSCC Cracks in Steam Generator Tube Roll Transitions -
12escription and Validation. LABORELEC Report h100 90-024/R-PH, hiay 1990.
7 5-8
p i i
Section 6 COMUINATION AX1AL AND CIRCUMFERENTIAL CRACKS
6.1 INTRODUCTION
-l l
While this degradation-specific tube repair document addresses axial PWSCC in _
tube roll transition zones, field experience indicates that circumferential cracks and occasionally combinations of circumferential and axial cracks can occur. :
Pulled tube experience (see appendix B) from European plants indicates that combination axial and circumferential cracks can occur in the EZ roll transition.
In several instances, the circumferential cracks were through-wall, although not t always in locations where interaction with axial cracks would have occurred. In '
almost all cases where there were through-wall circumferential cracks, the cracks were detected by IU'c inspections even though the inspections were not being
- performed specifically to detect circumferential flaws.
i The field experience in the United States with combination axial and circum-ferential cracking indicates significantly fewer confirmed indications of combined cracking. There have been several Instances where through-wall circumferential cracks have been confirmed in U.S. plants. In one case, the through wall crack was in a tube that also contained an axial crack, while in the other instance, there were no axial cracks present. The through-wall circumferential cracks were detected by RPC and confirmed when tubes were pulled. There has been one instance where a tube contained an axial crack in combination with a part through-wall circum-ferential crack.
While more tubes may develop combined axial and circumferential cracks in the future,it is expected that combined cracking will not be widespread or may take significant time to develop.
When distinct circumferential cracks are identified during in service inspection, the repair limits in this document do not apply. In these cases, the tube should be repaired.
To ensure that if circumferential cracks should develop in combination with axial.
PWSCC, structurally significant circumferential cracks will be readily detected by i
NDE and that undetected circumferential cracks will be small and would not leak :
significantly under faulted conditions, an additional repair limit is defined. Because
. there are few incidents of confirmed combined circumferential and axial cracking that can be used to evaluate in-service circumferential crack growth rates, the added '
6-1
+ y-+r-+--<rww .- - - ,e r m g , -s. , . , , -+- . , -y +,,.. - m g- q -
w -
w e ert we s- r r + r---ae-- -- vm gr *rm*w -v-*#
I repair limit was developed using postulated final crack sizes. These include through-wall axial cracxs in combination with circumferential cracks. The detection capability and pressure capacity of these postulated combination cracks were measured in laboratory tests. The repair limit resulting from these considerations requires a minimum spacing between axial cracks that are greater than the allowable axiallength defined in section 4. Tubes with axial cracks longer than the allowable length requirements of section 4 must be repaired.
~
The ability to detect circumferential cracks in combination with axial cracks is discussed in section 6.2 and appendix H. Tests which measured the pressure capacity of tubes with combination cracks are reviewed in section 6.3 and appendix 1. The additional repair limit to ensure that postulated combination cracks will not be structurally significant is described in section 6.4.
6.2 DETECTION OF CIRCUMFERENTI AL CRACKS BETWEEN AXI AL CRACKS A series of laboratory experiments were conducted to assess the ability m the IU'C eddy-current inspection to detect circumferential cracks in the presence of axial cracks (see appendix H). These experiments utilized tubes with combination axial and circumferential EDM notch configurations. Inspections utilizing different coil diameters were carried out and the detection of part-through wall circumferential components in the presence of through wall axial components was evaluated.
These experiments show that the conventional diameter unshielded pancake coil IU'C can readily detect a through-wall circumferential defect in the presence of two through wall axial defects spaced or more apart.
With a smaller coil diameter, circumferential defects can be reliably detected between axial defects closer than 6.3 PRESSURE CAPACITY OFTUBES WITH COMBINATION CRACKS l A number of pressure tests have been conducted with steam generator tubes containing combination axial and circumferential defects (see appendixI). These tests involved pressurizing a section of steam generator tube containing either machined notches or stress corrosion cracks. For through-wall degradations, a crack sealing system was used to maintain tube pressure. Tubes were pressurized and the maximum pressure that the tube could support was recorded. Generally, maximum .
pressure was achieved when part through crack ligaments tore or the crack scaling system was lost. This would result in a tube leak. With a limited flow capacity pressurization system, higher pressures could not be applied.
These tests show that with the axial component of the combination crack through-wall and with length equal to approximately the allowable length defined in section 4, combination cracks, with part-through-wall circumferential components less than through-wall, support maximum pressures greater than times normal operating differential pressure. In fact, most combination crack configura-6-2
tions supported pressures above the burst curve which characterizes the behavior of axial cracks alone in the absence of tubesheet constraint. Even completely through-wall "U" shaped cracks with axial componer s approximately equal to the ,
allowable length will support pressures grei than MSLB pressures. Finally, the test show that stress corrosion combination uacks have significantly higher pressure ,
capacity than EDM notches with similar lengths. ,
6.4 ADDITIONAL REPAIR LIMIT TO PRECLUDE SIGNIFICANT COMBINATION CRACKS While plant experience suggests that combination axial and circumferential cracks are unlikely, for conservatism, it will be assumed that small circumferential cracks can exist together with through-wall axial cracks.
3
.To assure that structurally significant circumferential cracks are readily detected by ;
NDE and that undetected circumferential cracks in combination with axial cracks are small and would not cause a significant leak under faulted conditions, axial cracks
_ longer than the allowable length defined in section 4 and whose circum-ferential separation
- is less than should be removed from service.
The spacing requirement of between axial cracks need not be applied to axial cracks with lengths less than By requiring that axial cracks with length greater than be spaced further than -
or more apart to remain in service, one can be assured that circum-
' ferential cracks through wall or deeper will be detected by standard RPC methods. In the absence of significant vibration loads, circumferential cracks .
through wall and completely around the tube circumference will not compromise ;
structural safety margins.
By limiting the length of axial cracks spaced more closely than to
, one is assured that even a through-wall circumferential crack connecting the two through wall axial cracks will not leak significantly at MSLB pressures and, based on limited data for real cracks, likely support much higher pressures.
A smaller crack spacing limit can be used if justified by qualification tests. This qualification should demonstrate that circumferential cracks ranging from to of wall thickness can be detected reliably when located between adjacent axial through-wall cracks of length approximately equal to A and separated by the proposed spacing limit. ,
"Circumferential separation is defined as the shortest circumferential distance tetween axial cracks as measured on any circumferential plane.
V 6-3 i
, , _ . - g - - . - - - , . . , , . , . - , --
, - - . - . . . , , , = - , - . . , -
--m,,, . -, , , -
Section 7 REPAIR LIMITS FOR Tile EXPANDED RFGION OF PARTI AL DEPTil ROLLED TUDES This section presents repair limits for PWSCC in the expanded region, including the transition, of partial-depth rolled tubes (see figure 1-10).
7.1 TUDE UURST AND ALLOWAULE CRACK LENGTl!
Experimental results (D have been performed to determine the burst pressure for tubes having outside diameter initiated axial cracks that are contained within a support with relatively small annular distances. The results from these experiments show that tube burst below the burst pressure for an unflawed tube is precluded by the constraint of the tube radial displacement when the cracked section of the tube remains within the tubesheet and the diametral gap is less than approximately Further experiments have been performed using combined axial and circumferen-tial cracks, where the axial crack was above the tubesheet and circumferential cracks were either above or below the tubesheet (1). The results from these experiments indicate that there is no influence of circumferential cracking on the axial crack burst curve when the circumferential cracks are located below the top of the tubesheet. Consequently, the structural integrity of partial depth rolled tubes is defined by axial separation and tube pull out (which depends on the extent of circumferential cracking) rather than burst from the presence of axial defects.
Experimental data Q) indicate that the extent of circumferential cracking needed to produce axial separation of the tube is very large (i.e., approximately a through-wall crack extending around the circumference, or a part-through-wall crack of the wall thickness and around the circumference). These crack sizes are readily detectable by standard NDE methods for tube inspection.
Because circumferential cracks can be difficult to detect by standard NDE methods when they are located between closely spaced axial cracks (see section 6.2), it is assumed for partial depth rolled tubes that a circumferential crack exists between two adjacent axial cracks less than apart (see section 6.2). An additional conservative assumption is made that the postulated circumferential cracks are through-wall Circumferential cracks need not be postulated between adjacent axial cracks separated by or more. Circumferential cracks need not be assumed between axial flaws less than the allowable defined in 7-1
i section 4. Tubes having NDE indications evaluated to be distinct circumferential cracks cannot be evaluated using this document and should be repaired.
The maximum allowable circumferential crack extent should be conservatively determined using a continuous through-wall flaw and the R.G.1.121 margins. The tube is acceptable for continued service if the allowable circumferential extent is greater than the extent required to be postulated by the axial spacing indicated by the NDE results. The postulated circumferential extent is determined by the sum of the circumferential distances between adjacent axial cracks that have spacings less than 10 mm (0.4 in.).
7.2 LEAK RATE CONSIDERATION Leakage from cracks within the tubesheet will be restricted at both normal and faulted loads compared to cracks located in the free span of the tube. Although the leakage is expected to be smaller compared to the leakage from a crack in the free span, no analytical methodology or laboratory testing is available to quantify the level of leakage expected. Therefore, it is suggested that the leak rate at faulted loads for partial-depth rolled tubes can be computed conservatively using the methods defined in section 5 for cracks above the tubesheet. The leak rate calculated from each tube need not exceed the maximum leak rate possible through the annulus between the tube and tubesheet.
7.3 SUMM ARY OF REPAIR LIMITS FOR PARTIAL-DEPTH ROLLED TUBES The following repair limits are applicable for plants with partial-depth rolled tubes:
- 1. The cracks must be located within the defined expansion zone, which includes the expanded portion of the tube, the transition between the expanded and unexpanded portions of the tube, and a short distance (e.g., ) of the unexpanded portion of the tube above the transition.
- 2. The length of the axial and inclined axial cracks are only restricted by leakage considerations. Leakage shall be calculated based on the requirements in section 5. To be considered an inclined axial crack, the axial extent of an inclincd crack must be equal to or greater than the circumferential extent of the crack. Indications evaluated to be distinct circumferential cracks cannot be evaluated using this document.
- 3. Circumferential cracks should be postulated to be present between the identified axial cracks due to nondestructive detection limita-tions. The extent of cracking retained in service should be limited such that the postulated circumferential crack extent in the presence of axial cracks does not exceed a portion of the tube circumference 7-2
that would provide adequate margin against axial separation of the l tube. This provides assurance that axial separation of the tube can not occur. Circumferential cracks need not be postulated between .
axial cracks that are more than apart or whose i lengths are less than the allowable defined in section 4. J 7,4 REFERENCES
- 1. T. A. Pitterle et al. Steam Generator Tubing Outside Diameter Stress Corrosion Cracking at Tube Support Plates Data 13ase for Alternate Repair Limi_ti. i I
Palo Alto, Calif.: Electric Power Research Institute. NP-7480 L (to be issued in '
December 1991).
- 2. Evaluation of the impact of the Pressuie of Circumferential Orientation Fissures in the Transition Zone of the Expanded Portion of Steam Generator-Tubes. Tecnatom Report, December 1990. ;
3, Northeast Utilities Tube Specimen 13urst Tests. Westinghouse Report WCAP-12916, May 1991.
(
7-3 .
A
--. ...-.p-+,_--. , ww w. ,.w--...v .$ .vm._e____,_-m.._,__. ..,,'__, i.v_s,.._i,_.m -,_ . _-.,._m,- -__ _, m -.i.-.,-. *_. - ,. . . . ..e.__+. _ - . .., + -. . = :
l Section 8
SUMMARY
8.1 OVERVIEW This document provides guid .ce and the supporting technical bases which may be used by a utility to justify leaving small, axially oriented, EZ PWSCC roll transition -
cracks in service. These cracks are typically through-wall, short, and are expected to grow slowly,if at all, with continued service.
Current technical specifications require repairing a tube if the ECT measured crack depth exceeds a specified limit, typically 40% of the wall thickness. Use of this tube repair limit for EZ PWSCC cracks would result in removal of many tubes from service which neither challenge steam generator tube burst capability nor l'ak a significant amount.
The justification for leaving some EZ PWSCC in service is based on a safety-related defense in depth approach. Elements in this defense include:
- 1. Performing enhanced eddy-current inspection (RPC) of oftube roll transition expansion zones in regions of the steam generator (i.e.,
the hot leg) where the tube expansion zones are susceptible to PWSCC.
- 2. Repair of tubes with axial defects longer than a conservatively established repair limit.
- 3. Limiting total number of cracked tubes in service based on crack size distribution to limit leakage during postulated accidents.
- 4. Implementing reduced primary-to-secondary leak rate limit to supplement leak-before-break detection.
The effectiveness of elements 1 and 2 is due to the ability to reliably detect and measure EZ PWSCC cracks, and the ability to establish allowable axial crack length (tube repair limits) which preclude tube rupture during normal operation or accident loading conditions. Elements 3 and 4 provide additional assurance that small EZ PWSCC left in service will not be a significant safety concern.
8-1
i i
Inspection of tube expansion zones with the rotating pancake coil (RPC) probe is discussed in section 3. RPC eddy current inspection methods will both detect and size EZ PWSCC axial cracks. Further, based on a comparison between measured and '
actual crack lengths, crack length measurement uncertainty can be calculated.
Guidance for calculation of an allowable EZ PWSCC axial crack length or tube repair limit is developed in section 4. Allowable crack length is based on a tube rupture curve which bounds published tube burst test cornelations, Reg. Guide 1.121 safety-factors applied to tube loads, lower bound tube material properties, a correction for tubesheet constraint, an allowance for crack growth between inspections and an allowance for crack length measurement uncertainty. Although some data are provided and sample calculations of allowable crack sizes are included in section 4, this was done to show how the calculation is rnade. It is expected that a utility user will perform plant-specific calculations using parameters and data applicable to the pc ;icular plant.
As proposed, EZ PWSCC tube repair criteria will leave short, possibly through-wall, 1 axial cracks in service. The leak before-break (LBB) nature of EZ PWSCC roll transition axial cracks is discussed in paragraph 5.1. Based on tube burst test and leak rate data,it is shown that many but not all EZ PWSCC cracks which could burst under faulted loading conditions would leak a detectable amount during normal operation. Although lowering of the allowable primary-to-secondary leak rate limit would further enhance LBB, due to the variable nature of tube leak rate data, LBB cannot be guaranteed. To provide a high level of confidence that tubes that may not exhibit LDB behavior at faulted load are removed from service, NDE will be performed on of the tube roll transition expansion zones in regions of the steam generator where the tube expansion zones are susceptible to PWSCC. Tubes with crack lengths greater than a conservatively calculated allowable crack length will be repaired. In addition, estimates of expected leak rate during faulted loading '
conditions are needed. Calculation of leak rates for cracks left in service also is discussed in paragraph S.2. These calculations depend on the distribution of cracks in a generator (number and size) as well as other hydraulic parameters. The number and size of cracks allowed to remain in service will be limited to ensure that the site boundary dose will not exceed 10CFR100 limits for postulated accident conditions. ,
P Section 6 defines an additional requirement to ensure that adequate margins for tube rupture and leakage at faulted load are maintained in the unlikely event that combined through-wall axial and circumferential flaws develop simultaneously in -
service. Repair limits for tubes with partial depth roll expansion zones are defined in section 7.
The combination of the inspection scope and procedure, repair limits, crack distribution limits, and leak rate limits ensure that conservative margins exist against failure and abnormal or excessive leakage at normal and faulted loads.
Using these elements to evaluate acceptance for return to service of tubes with.
8-2 l
1
.,.--.,_ , _.~.._,, _ . - - _ _ _ , _ _ . . - . . . _ _ . _ _ _ _ _ _ _ _ . . _ _ . . _ . . . _ . _ . .. . _
EZ PWSCC will provide levels of structural and leakage integrity that are in compliance with GDC 14,15,30,31, and 32.
l 8.2 COMPARISON WITH EUROPEAN PRACTICES l The repair limits that result from the European practices, together with pe tinent input assumptions, are summarized in table 8-1, European practices are disi ussed in section 1. The EPRI recommended sample calculation values from section 4 are also shown. The main conclusions drawn from review of table 8-1 and from the review of European practices and experience, are as follows: l i
- Alternate repair criteria have been in effect for up to about eight years in France and Belgium. These criteria have allowed the continued operation of many thousands of steam generator tubes with short, through-wall axial cracks. This has resulted in substantial economic-savings and reduced personnel exposure. More importantly, no tube ruptures or other safety problems have resulted.
- A variety of approaches have been used to develop alternate repair criteria to ensure that they provMe necessary margins of safety. All of the approaches have resulted in the conclusion that operation with short axial through-wall cracks at roll transitions is acceptable from both safety and operational standpoints.
- Two main approaches have been used in development and application of alternate repair criteria. One approach has relied on leak-before- ;
break (leak-before-risk-of-break in French terminology), and the other on expanded inspections of the roll transition area. While both approaches have attractive features, these guidelines emphasize the inspection-based approach. However, leak-before-break is considered and provides useful additional assurance of safety.
8-3
. . _ _ _ ~ _- -
=, -. . . ~ , . - - - . . ..-_ _- _ . , _ - .
Tab'c 8-1 COMPARISON OF LENGTii-BASED REPAIR LIMITS FOR AXIAL EZ PWSCC ROLL TRANSITION CRACKS j
Parameter S Approach inspection Method '
Leak Rate, gym i Crstical Length (at accident conditiore)
- In Free Span, mm a Within tubesheet, mm Controlling Condition for AllowaMe Length Allowable Length (before subtraction of margim)
- In Free Span, mm
- Within tubesheet, mm Margin for
- NDE uncertainty, mm
- Crack growth, mm (we tr>te 4)
Alkswable Length, mm ;
(after subtraction of margins)
Assumrtions Differential Pressure
- Normal Operation, psi a Steam Une Break, psi Tensile + Yicki, min., k*i L
Tube Coometry l
Mechanical Propeitics Notes: 1. Belgian uses most limiting of: (D bestetimate case with safety factor, and (2) worst caw parameters with no safety fators_ Only wont caw is irnhnled ;
since it is centrolling. ,
' 2i French perform RPCof sludge pile areas and perform Rit of Irin:m and bobtsn indicatm>ns. '
- 3. A!!owable crack length above the top of tubesheet.
- 4. Crack growth for one year .
'l
- 5. The EPRI criterion includes an additional requirement on the spacing between adjacent asial cracks and alternative criteria kar partialslepth tolk4 plants whkh
[
are not summarized here.
Appendix A
'3 TEAM GENERATOR DEGRADATION-SPECIFIC MANAGEMENT A STEAM GENERATOR MANAGEMENT The goal of steam generator management is to develop tools to optimize steam generator operation relative to safety, reliability, and cost-effective maintenance.
This section presents a management approach that can be appiied by utilities to deal _
with EZ PWSCC, which is one element of an overall steam generator management plan.
The strategy supported by this document uses the relationship between degradation, remedial measures, leakage during opc:ation and inspection as the basis for changes to steam generator tube repair limits for EZ PWSCC. The strategy maintains a defense-in-depth concept to assure that steam generator tube leakage is minimized and that safety issues are adequately addressed. The approach is utility-speci'ic and will require application of the techniques outliaed in other portions of this document to the situation and the condition of the plant under consideration.
Brordly state.. , the specific advantages of this stategy are:
- 1. Avoiding premature / unnecessary tube repair .
- 2. Optimizing steam generator availability _
- 3. Applying the ALARA principle
- 4. Maintaining flexibility for long-term repair options
- 5. Maximizing the available heat transfer area
- 6. Optimizing the cost-effectiveness of the steam generator repair and inspection program A-1
I A.2 STEAM GENERATOR MANAGEMENT STRATEGY Managing steam generator tube degradation requires attention in three areas: >
- 1. Mechanism management (preventive maintenance)
- 2. Degradation management (near-term corrective maintenance)
- 3. Life extension / replacement (long-term corrective maintenance)
A.2.1 Mechanism Management Mechanism management determines the cause of and remedies for the steam generator degradation forms applicable to a particular plant. The remedies are mechanism-specific, may be applied at various times in the plant's operational period, and include such items as enhanced water chemistry, in situ thermal stress relief, application of stress modification processes (peening), etc.
A.2.2 D_egradation Management The goal of degradation management is to remove only those tubes from service that are required to be removed for reliability and/or safety considerations. Support for this area is generated from in-service inspection, tube repair limits and the capability to monitor primary-to-secondary leakage. - (This document focuses on the elements of degradation management as they apply to EZ PWSCC.)
A.2.3 Life Extension /Ren1peement Life extension / replacement is targeted toward long-range planning involving the economic analysis surrounding the repair / replacement decision. The ability to make an appropriate repair / replacement dccision depends on information generated as part of the mechanism and degradation management elements of the program.
A.3 DEFENSEIN DEPTH The defense-in-depth concept is used to assure that components which form the reactor coolant pressure boundary have a high degree of integrity in order to minimize the transport of radioactivity from the fuel and maintain core coolant inventory under a variety of assumed failure or accident modes.
Defense-in_-depth is based on using multiple elements to minimize the potential for:
catastrophic failures , and is broadly addressed by, ag ' ag proper prirdples; selecting appropriate materials; specifying design loads, limits and safety factors; and identifying nondestructivc evaluation and maintenance requirements.
A-2 v
9
h For steam generator tubing, defense-in dep:n has the following elements:
- 1. -Initial design and material selection
- 2. Applicction of remedial measures for known degradation mechanisms
- 3. Limits on the maximum allowable primary-tcesecondary leakage ,
- 4. Periodic in service inspection (eddy current testing) to monitor the steam ,
generator tubing for degradation
- 5. Repair of those tubes exceeding a specified degradation limit by:
-Removal from service (plugging).
--Repair of the defective tube segment by sleeving or similar techniques.
Historically, specific interrelated criteria and requirements have been developed for each element of the defense in-depth approach to steam generator tube degradation.
~.In principle,it should be acceptable to change the relationship among the five elements without impacting defense-in-depth or changing the overall outcome of steam generator tube integrity assurance; i.e., with fixed design and material condi- -
tions, a change in the tube repair limit is acceptable assuming suitable analyses are completed to establish new. allowable leak rate and to identify new steam generator -
tube inspection requirements.
Figure A-1 is a decision tree / logic chart that could be applied to any steam generator tube degradation mechanism to assure that defense-in-depth is adequately addressed. Table A-1 is a definition of the terms used in figure A-1, and table A-2 describes the management options. The following is a general discussion of the application of figure A-1 to the management of any specified degradation form.
For existing steam generators the design features and material selection were-established as part of the original design. Tube degradation is monitored by periodic -
in service inspection (ISI). Flaw-types are characterized using specialized inspection -
techniques and are further quantified through tube removals and laboratory tests.
A cause can be postulated for the mode of tube degradation and degradation progress can be assessed. ,
'The result of this combined field inspection and laboratory assessment is the development of a degradation-specific progression correlation, and the prediction of future steam generator tube degradation. The progression correlation serves as the input to the decision-analysis process regarding the impact of the degradation on.
tube rupture potential, the effectiveness of remedial measures, primary-to-secondary leak potential and the adequacy of inspection requirements.
A-3
The results of the inspection program are then used to improve the quantitative flaw assessment model which then can be used to improve the overall tube
-degradation progression correlation.
As shown in figure A 1, there are 12 possible outcomes as a result of application of the decision tree and each outcome could be distinct with appropriate methods for managing a specific steam generator tube degradation mode.
A.4 EZ PWSCC DEGRADATION MANAGEMENT Application of the decision tree of figure A-1 to EZ PWSCC, using existing laboratory and field data, is shown in figure A-2 and leads to the following: (1) multiple axial cracks can form in the roll transition region of alloy 600 tubes hard rolled into the tube sheet; (2) operating experience has shown these cracks are short, often less than long, and (3) the cracks generally do not propagate upward beyond the stressed region of the roll transition. With appropriate analysis and approved alternate tube repair criteria, tubes containing these short, axial cracks can be safely left in service.
Leaving such tubes in service is predicated on the following:
- Application of EZ PWSCC remedies to limit the number of cracks expected in service
- Evaluation of the primary-to-secondary leak rate limit with respect to leak-before-break (LBB) capability
- Commitment to perform rotating pancake coil (RPC) eddy-current inspections of- of tube expansion zones in regions of the steam generator (i.e., the hot leg) where the tube expansion zones are susceptible to PWSCC
- Repairing of tubes with axial defects longer than a conservatively established repair limit, which includes the following elements:
- Use of a tube rupture curve which bounds published tube burst test correlations.
-- Reg. Guide 1.121 safety factors applied to tube loads.
-- Use of lower bound tube material properties.
I - Correction for tubesheet constraint.
I A-4 i-
-- Allowance for crack growth between inspections.
-- Allowance for eddy-current crack length measurement uncertainty.
- Calculation of the leak rate expected during postulated accident loads from the cracked tubes that remain in service.
- While this document and the logic described above is focused on permitting the existence of through wall EZ PWSCC cacks up to and including the final expansion roll transition, earlier industry work has allowed revised limits for EZ PWSCC cracks that exist more than a specified depth below the tubesheet secondary face.
These criteria are referred to as P+ and F* criteria and are discussed in paragraphs A.4.1 and A.4.2 below. .
.A.4.1 Tubesheet Region Repair (P') Criterion Existing tube repair criteria apply throughout the tube length and do not take into account the reinforcing effect of the tubesheet on the external surface of the tube.
The presence of the tubesheet constrains the tube and improves its integrity in that region by essentially precluding tube deformation beyond its initial outside diameter. The resistance to both tube rupture and tube collapse behavior is strengthened by the tubesheet, and the use of an alternate plugging criterion is justified.
The tubesheet repair criterion applies to steam generators that have been full-depth mechanically rolled in the tubesheet and obviates the need to repair a tube (by sleeving) or to remove a tube from service (by plugging) due to eddy-current indications for most of the length of tubing within the tubesheet. The basic premise is that if a severed tube is constrained such that it cannot come clear of the tubesheet, the integrity of the tube bundle is conserved, and if leakage is limited, the safety consequences of leakage have been appropriately addressed. In the develop-ment of the tubesheet repair criterion, three aspects of bundle integrity are addressed: (1) maintenance of a fixed tube end condition in the tubesheet with' the limiting case of a circumferential indication near the top of the tubesheet, (2) limitation of primary-to-secondary leakage consistent with accident analysis assumptions, and (3) maintenance of tube integrity under postulated conditions of primary-to-secondary and secondary-to-primary differential _ pressure.
The tubesheet repair criterion for repairing and removing tubes from service is considered to be conservative. The basis for the analysis is the assumption that tube -
' degradation is oriented circumferentially. If an indication on a tube within the' tubesheet has been determined to be axially oriented, the structural integrity of the tube would be maintained, and _ tube pullout from the tubesheet is not an issue.
' Additional margin is provided by the resistance to vertical displacement of a separated tube that develops due to interference between the tube and the tubesheet, A-5
the tube support plates and antivibration bars. However, no credit is taken for this in the development of the tubesheet plugging criterion.
The tube sheet repair criterion identifies a distance (1-0, designated P'
, below the top of the tubesheet for which tube degradation of any extent does not necessitate remedial action since even a severed tube is constrained by other tubes such that it cannot come clear of the tubesheet. The uncertainty in position of the eddy-current indication must be included in the final calculation of P' criterion. For example, for a standard bobbin coil probe, a conservative eddy-current uncertainty would be ; therefore, P' would have a value of between A.4.2 E' Criterion An alternate steam generator tube repair criterion, designated the F' criterion (1-0, has been developed for use in determining whether or not repairing or plugging of hardroll expanded tubes is necessary for degradation which has been detected in that portion of the tube located within the tubesheet. The F' criterion represents a length, F', of a mechanically rolled tube in the tubesheet which provides for sufficient engagement of the tube in the tubesheet hardroll such that tube pullout during both normal operating and postulated accident condition loadings would be successfully resisted by the clastic preload between the rolled tube and the tubesheet.
Existing steam generator tube repair criteria apply throughout the tube length but do not take into account the reinforcing effect of the tubesheet on the external surface of the tube. The presence of the tubesheet acts to constrain the tube and complement its integrity in that region by essentially precluding tube deformation beyond its expanded outside diameter. In addition, the proximity of the tubesheet significantly affects the leak behavior of through-wall tube cracks in this region; the joint is expected to be leak-tight. The conservativeness of the F' criterion has been demonstrated to result in tube integrity consideration commensurate with R.G.1.121 criteria, both empirically and analytically.
In developing the F* criterion, it is postulated that the circumferential severance of the tube can occur. Implicit in assuming a circumferential severance to occur in the ,
development of the F' criterion is the consideration that degradation of any extent or orientation within the tubesheet is demonstrated to be acceptable below the F' regions.
The required engagement length, F*, below the bottom of the hardroll transition or the top of the tubesheet, whichever is greater relative to the top of the tubesheet,is calculated from a derived preload force, as well as an indirect measure of the static coefficient of friction. The F* criterion varies with tube size and has been verified as conservative based on tube pullout and hydraulic proof testing. Also,in assessing the F' criterion, it is postulated that the radial preload resulting from the hardroll would be sufficient to significantly restrict leakage during normal operating and A-6 l
postulated accident-condition loadings. No leakage from any of the hydraulic a..of tests for pressures in excess of faulted condition loadings has been detected.
A.5 REFERENCES ,
- 1. "Tubesheet Region Plugging Criterion for Full Depth Hardroll Expanded Tube " WCAP-10950. September 1985.
- 2. "Tubesheet Region Plugging Criterion." WCAP-11225, July 1986.
- 3. "Tubesheet Region Plugging Criterion." WCAP-11225, Rev.1, October 1986.
- 4. Docket Nos. 50-369 and 50-370. Issuance of Amendment No. 59 to Facility Operating License NPF-9 and Amendment No. 40 to Facility Operating License -
NPF-17 for the McGuire Nuclear Station, Units 1 and 2, August 1986, s
A-7 l
e .
I in I.
l 1
l Figure A-L Degradation-SpeciGc Management Decision Tree;
, . . . . . . . . . . . . . . . . - . ~ . .
t 6
Figure A-2. EZ PWSCC Degradation Management Decision Tree 6
-Table A-1 DEFINITIONS Terms Degradation Management: Remove only those tubes (or tube segments) from service which are required to Sc removed for safety and/or reliability and no others.
Elements / Decision Points Progression Correlation: Relates material condition, fabrication / techniques and environmental factors to determine degradation initiation point, growth rate and orientation as a function of time. (The degree of sophistication /
complexity of this element is dictated by the need to separate rupturable from nonrupturable tubes, which,in turn, is dictated by the need to pursue defect-specific degradation management.)
Rupture Potentiah Flaw-type can grow over time and reach a length that would allow tube to " rupture" under a specified load condition (e.g., typical steam line break or normal operating condition with safety factors).
Effective Remedies: Remedial measures such as improved water chemistry, sludge removal, stress relief, stress redistribution (peening), plating, sleeving, etc., exist that prevent flaw-type (s) from reaching a rupture-critical size.
.LBB: Flaw-types that can grow.to a rupture-criticallength will reach a__ length that will produce a reliably detectable leak under normal operating loads -
(with a suitable response interval) prior to reaching the rupture-critical. length such that the plant can be shut down. ,
i Effective Inspection: The capability exists to accurately detect, characterize and -
size the flaw-type at a specified inspection interval such that flawed tube sec-
. tions can be removed from service prior to reaching a rupture-critical length.
Regulatorv Issues: The existence of the flaw-type regardless of rupture potential or capability for preventing rupture by LBB considerations or effective inspection procedures, violates a regulatory and/or technical specification position such as maximum through-wall penetration.
F Leak Potential: Flaw-types that cannot rupture can leak at measurable / detectable rates during normal operation.
A-10
Table A-2 ,,
MANAGEMENT OPTIONS .
Management 1 Conditions:
(a) Flaw type (s) can rupture (b) Effective remedial measures (exclusive of plugging) prevent flaw types from growing to a rupture-critical condition.
Actions:
(a) Apply appropriate remedial measures __
(i) environmental control (ii) stress alteration (iii) sleeving (b) Manage as a "nonrupture" flaw type Management 2 Conditions:
(a) Flaw-type (s) can rupture (b) Remedial measures not effective or are not used (c) LBB argument is technically sound (i) acceptable analysis method (ii) supporting leak model data (iii) supporting rupture / burst data (d) Acceptable leak rate less than that for existing Tech Specs (e.g., ) _
Actions:
(a) Accept lower leak limit (i) Tech Spec change (ii) administrative limit (b) Install suitable leak-detection equipment (c) Develop (plant-specific) leak-monitoring / response procedures (d) Solve any regulatory issue (s) per Management 4 A-11
Table A 2 (Cont'd)
Management 3 Conditions:
(a) Flaw type (s) can rupture (b) Remedial measures not effective or are not used (c) LBB argument is technically sound (i) acceptable analysis method-(ii) supporting leak model data (iii) supporting rupt_ure/ burst data (d) Acceptable leak rate less than that for existing Tech Specs =
(e.g., . )
Actions:
(a) Manage as a "nonrupture" flaw-type Management 4 Conditions:
(a) Flaw-type (s) can rupture (b) Remedial measures not effective or are not used (c) LBB argument is technically sound
. (i) acceptable analysis method (ii) supporting leak model data (iii) supporting rupture / burst data (d) Acceptable leak rate less than that for existing Tech Specs (e.g., )
(e) Operation with identifiable /locatable through-wall flaws not accepted by USNRC Actions:
(a) Obtain NRC acceptance of the position nr proceed to Management 5 or 6 A-12
Table A-2 (Cont'd) -
Management 5' Conditions:
(a) Flaw-type (s) can rupture -
(b) Remedial measures not effective or are not used (c) LBB argument is a supporting basis (defense-in-depth)
(d) Effective inspection capability exists (e) Operation with identifiable /locatable through-wall flaws acceptable to USNRC (i.e., length-based plugging limit @ ' BV is accepted)
Actions:
, (a) Commit to conduct required inspection (i)- plug / sleeve as required (ii) leak reduction measures (such as nickel plating) may be acceptabla to reduce total primary-to-secondary leakage Management 6 Conditions:
(a) Flaw-type (s) can rupture (b)- . Remedial measures not effective or are not used (c) LBB argument is a supporting basis (defense-in-depth)
(d) Effective inspection capability exists (e) Operation with identifiable /locatable through-wall flaws not acceptable to USNRC (i.e., length-based plugging limit - -BV is not accepted) '
Actions:
(a)- Obtain NRC acceptance of the position 91 proceed to Management 7 or 8 oi A-13L J
= . - - .ir. - _. , . __ t,_, , ,
Table A-2 (Cont'd)
. Management 7 Conditions:
-(a) Flaw-type (s) can rupture b) Remedial measures not effective or are not used (c) LBB argument is a supporting basis (defense-in-depth) d) Effective inspection capability not used (i) doesn't exist (ii) choose not to use (c) Consequence analysis shows ihat the probability of a tube rupture and-resultant consequences are acceptable 4
Actions:
(a) Obtain NRC acceptance of the position (b) No further action Management 8 Conditions:
(a) Flaw-type (s) can rupture (b)- Remedial measures not effective or are not used (c) LBB argument is a supporting basis (defense-in-depth)
(d) Effective inspection capability not used (1) doesn't exist (ii) choose not to use (e) - Consequence analysis shows that the probability of a tube rupture and resultant consequences are not acceptable Management 9 -
Conditions:
(a) Flaw-type (s) can't rupture
- (i) -- during life of plant (ii) during next operating cycle (b) _ Flaw-type (s) can leak at rate that could result in a forced shutdown by approaching Tech Spec leakage limit Actions:--
(a) Apply appropriate _ remedial measures OR (b) Accept periodic shutdowns (c) Resolve any regulatory issues
_ _ _ _ =_
A-14 l,
i Table A-2 (Cont'd)
Management 10 Conditions:
(a) Flaw-type (s) can't rupture (i) during life of plant (ii) during next operating cycle (b) Flaw-type (s) can leak but not at a rate likely to result in forced shutdown (c) Existence of flaw-type (s) violates a technical specification position such as maximum TW penetration 3 Actions:
(a) Obtain NRC acceptance of the position, OR ~
(b) Apply appropriate remedial measures Management 11 Conditions:
(a) Flaw-type (s) can't rupture (i) during life of plant (ii) during next operating cycle (b) F.aw-type (s) can't leak (c) Existence of flaw (s) violates a technical specification position such as maximum TW penetration Actions:
(a) Obtain NRC acceptance of the position, OR (b) Apply appropriate remedial measures Management Option 12 Conditions:
(a) Flaw-type can't rupture (i) during plant life (ii) during next operating cycle (b) Flaw-type can't leak (c) Existence of flaw-type is acceptable to plant technical specifications Actions:
(a) Monitor flaws for growth A-15
1
.I l
Appendix B BACKGROUND OF PWSCC ,
B.1 CAUSES OF PRIMARY WATER STRESS CORROSION CRACKING (PWSCC)
Primary water stress corrosion cracking (PWSCC) of alloy 600, and the related phenomenon of pure water stress corrosion cracking, are both forms of inter-granular stress corrosion cracking (IGSCC). This degradation has been studied exten-sively since originally observed by Coriou (1), and both laboratory research (2,3) and
- plant experience (1) have shown that PWSCC occurs when a high tensile stress, a susceptible microstructure and an aggressive environment occur simultaneously. .
Each of the three necessary elements are discussed in more detail below:
B.1.1 Tensile Stress
- PWSCC occurs most readily when the tensile stress at the exposed surface is above the yield strength of noncold-worked tube material, i.e., above about 55 ksi. Typi-
- cally, such levels occur in areas where there are high residual stresses as a result of -
tube fabrication or installation processes. Regions experiencing these residual stresses include the expanded tubesheet region (particularly where rolling anoma-lies have occurred), the expansion transition zone into the unexpanded region of the tube, tube dents such as can occur at TSP intersections, and the small radius U-bends. The total stresses are due to pressure, and thermal and residual stress. The other stresses are usually not as high as the residual stresses which can be in some locations (5,6).
B.I.2, Susceptible Material Susceptibility of alloy 600 to PWSCC appears to be related to the carbide microstruc-ture which is a function of the carbon content, heat treatment and cold-work history of the tubing. Highest susceptibility is characterized by a microstructure with few -
carbides at grain boundaries and many carbides within the grains (i.e., many intra-granular carbides). High susceptibility also tends to correlate with small grain size and high yield strength. The susceptible microstructure is the result of low tempera-tures (within the acceptable range) used for the final mill anneal and for the mill armealing performed before the last stage of cold-work. The least susceptible condition is that associated with a microstructure characterized by continuous carbides at grain boundaries and few carbides within the grains (Z,8).
B-1
B.1.3 Aggressive Environment With a susceptible microstructure and high tensile stresses, elevated temperature pure water is sufficiently aggressive to cause IGSCC of alloy 600, i.e., PWSCC. The main environmental factor that increases aggressiveness is high temperature (about increase in the progression of degradation for each ) (9,10). A secondary factor is the concentration of dissolved hydrogen. Current research indicates that at high temperatures (e.g., ), increasing hydrogen concentration increases the rate of cracking (H). However, at operating temperatures characteristic of plant hot leg temperatures (e.g., ), the occurrence of PWSCC does not appear to be strongly dependent on the concentration of dissolved hydrogen (H).
The effects of other environmental variables are not well known. The impact of lithium concentration on PWSCC--tests have yielded mixed results, with some data indicating that higher lithi. a accelerates cracking, while other data indicate that-lithium has little effect (H). The most recent data indicate that higher lithium concentrations can accelerate PWSCC of alloy 600 in certain conditions (H). Recent .
studies have shown that the addition of zine to primary coolant may increase the time to PWSCC initiation significantly, by altering the protective oxide film. Conse-quently, zinc additions may be useful for reducing the incidence of PWSCC as well as corrosion product buildup on the primary side. However, substantial additional study is required before such steps could be taken (H). There is also some infor-mation indicating that impurities such as sulfate can affect PWSCC; however, at the levels currently maintained in PWR primary system, the impact of such impurities is considered negligible when compared to the other factors discussed above (stress, material and temperature).
B.2 PLANT EXPERIENCE This section provides a brief chronology of the more notable occurrences of PWSCC in PWR steam generators.
i
- The earliest occurrence of what is now considered PWSCC in a PWR steam generator was cracking observed at roll transitions and in row I U-bends at Obrigheim in 1971 (M) . (In that case, lithium hydroxide was not added to the primary coolant, as is normally done to control the pH. Also, hydrogen was not added, and the tubing was pickled). Later, PWSCC appeared in the expansion zone and,in 1983, the Obrigheim steam generators were the first to be replaced (in 1983) as a result of PWSCC at roll transitions within the tubesheet.
- About 1976, PWSCC was identified at U-bends and distorted TSP intersections at several plants affected by severe denting. This PWSCC with dents was associated with high dynamic straining 1
B-2
, . . . _ __ _ _ _ - _ . . _ - - _ _ _ _ _ _ . _ _ _ _ _ - = _ _ _ . _ _ _
(e.g., up to ), and occurred at plants with relatively low hot leg temperatures ( and below).
- In 1978, Doel 2 experienced PWSCC at a roll transition (1Z) that was detected as a result of primary to-secondary leakage. Doel 2 has part.
depth rolled tubes, and has low temperature mill-annealed tubing which appears to be especially susceptible to PWSCC. Since that time, a large percentage of the tubes in Doel 2 have been diagnosed by ECT as having PWSCC at the roll transition (1Z). However, tubes in Doel 2's sister unit, Doel1, from the same manufacturer, have not experienced PWSCC.
Other plants around the world have also reported PWSCC at roll __
transitions (18). In many cases, PWSCC reported by eddy-current test (ECT) examination methods has been confirmed by destructive examination of pulled tubes (several hundred tubes pulled from U.S., Spanish, French, Belgian and Swedish plants). PWSCC has occurred in the standard roll transition found in both U.S. and European plants, as well as in " kiss roll" transitions found in many French and Belgian units and one Swedish unit (see figure B-1).
- In 1977 at Takahama 1, and in 1978 at Trojan, PWSCC was identified in row 1 U-bends (19,20). The PWSCC was at the discontinuity that can occur at the location where the ball mandrel used to fabricate row 1 and 2 U-bends was located at the end of the bending opera-tion. Measurements of the residual surface stresses at this location indicate that they are in the neighborhood of (6).
- Since the late 1970s, reporting of PWSCC has become more common. A number of plants with small radius row-1 U-bends, susceptible tubing material, and hot leg temperatures over have either experienced leaks due to PWSCC or have preventively plugged the row-1 tubes (18), and many plants with hot leg tempera-tures over have also reported PWSCC at hot leg roll transi-tions. This has been especially true in European plants with kiss-roll transitions, but has also been true in plants with standard rolls.
As plants age, older plants with lower temperatures are starting to report roll transition cracking (18) (see table B-1).
- PWSCC has also been observed at explosive expansion transitions and at minor dents in a few plants with high hot leg temperatures (18).
B-3 s
A
x i
~
I' Figure B-1. Standard and " Kiss" Roll Transitions, e -With Typical PWSCC - _
1; l-L 1;
I L.
- - B .
, 4 ,
Table B-1 REPORTED EXTENT OF EZ PWSCC AT PLANTS WITH TUBES AFFECTED (September 1991)
Tubes with PWSCC Steam Percent Year Plant Generator (note 1) Reported Notes Notes:
- 1. Average percent PWSCC in each steam generator, based on RPC inspection unless otherwise noted.
- 2. Value based on RPC inspection of more than tubes.
- 3. Based on RPC inspection of a tube bundle sample.
B-5
- Recently, PWSCC has been reported in the expansion zone in the cold leg tubesheet region. Based on inspection information, the cold leg cracking is similar to the hot leg cracking, but the time to appearance of cold leg roll transition PWSCC is delayed as a result of the lower cold leg temperatures (21).
- While the large majority of EZ PWSCC degradations observed to date by ECT and/or destructive examination have been axially oriented, small but increasing numbers of circumferentially oriented PWSCC are also being observed in the EZ region. In France, isolated circumferential PWSCC has been identified in kiss roll transitions in tubes in the sludge pile region, and in tubes with dents at the top of the tubesheet. Isolated circumferen-tial PWSCC has also been observed in standard roll transitions in some Spanish and Swedish units. These isolated circumferential cracks can typically be identified using RPC-type inspections. In addition to the occurrence of isolated circumferential cracking, some plants have expe--
rienced mixed mode, or combined axial and circumferential PWSCC in close proximity, in the EZ. Mixed mode EZ PWSCC has been identified in French, Swedish, and Belgian plants with kiss roll transitions, and in French, U.S., and Swedish plants with standard roll transitions. Mixed mode cracking can generally be identified using RPC-type inspections,if the circumferential component is of significant length (2.2).
B.3 IMPACT OF ROLL YRANSITION PWSCC B.3.1 Number of Plants Affected To date, reported roll transition PWSCC has been more extensive in Europe than in the United States. One reason for this is believed to be the fact that plants with the combination of susceptible mill-anneal material, mechanical rolls, and high hot leg temperatures came into operation in large numbers earlier in Europe than in the United States. Since the fraction of tubes affected by PWSCC increases with time, the longer operating times of the European plants results in more tubes with reported . degradations. Another major factor is that many European units have kiss rolls. The cold work associated with kiss rolling appears to accelerate the development of detectable cracks.
B-6
+Ew., ew , g . * - - . -
- 9h'm t-- e=m-
I i
Table B 2 STATUS OF PWSCC AT PART-DEPTH-ROLLED PLANTS IN USA (September 1991)
PWSCC Experience Indications Tubes Comm T Hot Tube not repaired- Repaired
- Ejant Oper DegE Supplier & a E- a Remarks Notes:
- 1. Data taken from reports to NRC Public Document Room and information provided to EPRI by utilities.
- 2. H = Huntington, W = Westinghouse.
B-7'
.. . - - ~ a , - ~ , . , , . . . . ,
Table B-3 STATUS OF PWSCC AT FULL-DEPTH ROLLED PLANTS IN USA (September 1991)
PW9CC Experience Indications Tubes Comnt T Hot Tuk not repaired Repaired Peening _
Plant Oper. I_ Sunplier liL G HL G HL G Remarks Notes:
- 1. List limited to plants with LTM A tubing.
- 2. Data taken from reports to NRC Public Document Room and inforrnation provided to EPRI by utilities.
- 3. Peened before operation.
- 4. Indications not repaired data for does not include inspection results, B-8 l
1
Plants that have reported roll transition PWSCC in the United States are listed in tables B-2 and B-3, respectively. The magnitude of reported PWSCC may increase with time as discussed below:
- Part-Depth-Rolled Plants. Older plants with low hot leg temperatures , e.g., Conn Yankee, San Onofre, Ginna, Zion, and Cook 1, have reported the roll transition PWSCC in the past few years. Accordingly, other plants with part-depth rolls may eventually develop roll transition PWSCC.
- Full-Depth Rolled Plants. The plants in table B-3 have material in the mill-annealed condition and many also have high hot leg temperatures. This combination is prone to PWSCC (18,21).
Expected behavior at these plants is as follows:
-Plants that operated prior to peening report EZ PWSCC.
Depending on the length of operation before peening, this PWSCC may be severe.
--For plants that were shot peened before operation, PWSCC in the tubesheet am is expected to be negligible (21).
- Based on experience in Belgian plants (25), the occurrence of reported PWSCC will be limited to relatively low numbers of tubes, e.g., a few percent, at plants that rotopeened or shot peened prior to operation. This SCC is expected to occur mainly at relatively severe rolling abnormalities, and will probably become detectable early in life, within three or four cycles of operation, since rolling irregularities have high stresses and initiate defects earlier than other locations.
B.4 REMEDIAL MEASURES FOR PWSCC IN EXPANDED AREAS This section briefly describes remedial measures that have been developed or explored for PWSCC in expansion zones. Preventive measures are first described, and then the repair methods.
B.4.1 Peening Peening of metal surfaces has been used for many years to change the stress state of the surface to increase resistance to stress corrosion cracking. This effect is achieved by developing a thin layer of compressively stressed material on the surface as a -
result of impacting by small balls or beads. Laboratory testing and experience ihow -
that if this compressive stress is uniform, and is sufficiently high to overcome all in-service tensile stresses, SCC will not occur even in aggressive environments.
B-9
.- -- - . .. - . - - - - - - - - . - . . - - . ~ . -
l l
l Peening is directed at preventing initiation of new cracks. If applied to a surface that has already developed cracks, it will be less effective. It is probable that shallow. .
cracks, with a depth less than the thick compressive layer developed by- l peening, will be prevented from growing by peening. However, defects with depths of over probably can continue to grow since the crack tip is in a tensile area. If a plant has operated for a few years and peening is then performed, it is - !
possible that there will be a population of cracks present with depths that exceed the thickness of the compressive layer. .
Experience at many plants which peened after operation and which have now operated for cycles subsequent to peening show that the population of tubes with cracks present at the time of peening can be substantial. Some units have exhibited up to an additional of their tubes cracked following peening (2J).
B,4.2 Siress Relief Tests have shown that stress relief is effective at preventing roll transition PWSCC (2Z1 However, a practical stress relief process has not been developed and qualified to the point where it is commercially available.
If performed, the stress relief would involve heating the roll transition area to about-for up to . Such heat treatments have been found to be effective
- for U bends and have been qualified for use at dented TSP intersections. The heat treatments reduce tensile residual stresses from over-yield levels to values of -
yield or less. This amount of stress reduction is expected to prevent PWSCC from occurring during the steam generator operation.
B.43 Plugging Tube plugging has been the standard steam generator repair approach u' sed in the United States for PWSCC degradations reported in the roll transitions.- However, -
many plants have limited plugging margins, in the range of (the plugging-margin is the fraction of tubes that can be plugged beare plant power must be reduced because of reduced primary flow or reduced steam pressure). Since the -
number of tubes that may be affected by PWSCC at some plants'will approach or exceed applicable plugging margins, and since other tube defect mechanisms are also likely to cause some level of plugging, plugging is not the desirable long-term =
n solution.
, B.4.4 Sleeving Sleeving has been widely used for problems such as pitting and IGA /IGSCC in tube -
sheet sludge pile and crevice areas. Recently, several _ vendors have qualified sleeving for use in' PWSCC susceptible tubing (28), and several plants are now using .
sleeving on a regular basis to repair tubes with roll transition PWSCC. - These -
, B ,
, < - -s , .,,,--x ., .ev. >---,' 1.-' m- -
ev ww-- - - - - y e
1 l
sleeves involve stress relief of the top joint to prevent residual stresses associated with joint fabrication from causing PWSCC. The lower joint is not stress-relieved because the configurations result in a low potential for problems even without
- stress relief, as shown by mockup tests. Alloy 690 is used for sleeving because tests have shown it to provide additional resistance to PWSCC.
Sleeving, while useful, has some limitations:
- Installing sleeves is relatively expensive and slow (relative to plugs).
- Sleeves increase resistance to flow (though not as much as plugs) and reduce heat transfer area.
- Sleeves interfere with repairs that may become necessary in the future at higher elevations, for example, at TSP intersections.
- Some portions of the sleeved joint are difficult to inspect, especially the tube area immediately adjacent to the top joint.
B.4.3 Nickel Plating Nickel plating of roll transitions with axial cracks has been developed by Belgatom-and Framatome, and has been used on a trial basis in Doel 2 (part-depth-rolled) and Doel 3 (full-depth rolls)(22), and Ringhals 3. The nickel plating is typically applied for a length of about , including the roll transition, and has a thickness of
. It is applied remotely using electroplating techniques.
. Tubes with nickel plating have been in operation since 1985, and examination of pulled tubes shows that nickel plating can effectively seal PWSCC Belgatom and Framatome indicate that it will provide long-term repair, but the possibility remains that crack growt? might occur from the secondary side (22). Ultrasonic examination after three years of operation at Doel 3 has shown no evidence of such growth.
Because of its thinness, nickel plating has the advantage of not interfering with other repairs or with inspections of other areas.' On the other hand, roll transitions with nickel plating are difficult to inspect using ECT techniques, and Laborelec together with Framatome, has developed an ultrasonic technique to monitor cracks in nickel-plated roll transitions (10). 0 In addition, unlike sleeving, nickel plating is
- not considered a_ structural repair, and can only be used for tubes with relatively short axial cracks.
Nickel plating has been developed as an alternative repair method to plugging or sleeving. It can be used to seal PWSCC that has grown to a length that could lead to excessive leakage or, if further growth occurred, might impin'ge on allowed crack B-11
i length limits. Use of nickel plating is practical only where approved plugging limits allow short through-wall cracks to remain in service. j An alternate approach for nickel plating would be to use it in a preventativ_e mode. l In this mode, nickel plating would be applied to roll transitions in the absence of detected cracking in an effort to prevent cracking from occurring.
An additional repair approach using nickel plating has been used at Doel 2, which has part-depth rolls. At Doel 2, a new expansion was made above the original roll transition, and tlw ID surface was their nickel plated to prevent PWSCC.
B.4.6 Tube Section Replacement KWU has developed, and demonstrated in nonnuclear heat exchangers, a technique for cutting out a defective length of tube and replacing it wh a new tube section.-
The technique is applicable to tube sections extending to a short distance above the top of the tube sheet. The new tube section would be made of a more corrosion-resistant material, such as alloy 690, and would be installed in a manner to mini-mize residual stresses. The technique has not been used in operating PWR steam generators, and there is no experience on which to judge whether installing new tube sections would be practical and economical in nuclear applications.
B.4.7 Temperature Reduction As described in paragraph B.2, PWSCC is strongly affected by temperature, with an activation energy of about This means that a reduction in temperature would reduce the number of tubes initiating cracking to of the-value experienced at the higher temperature. If plants have a sufficient design margin to accommodate such temperature reductions, the temperature reductions-may be cost-effective. Some utilities (e.g., Commonwealth Edison at Byron) have
, implemented such temperature reductions. Temperature reduction as a PWSCC remedial measure is dependent on many plant-specific factors such as available design margins, secondary system design features, etc., and some utilities have =
concluded that it is not their preferred remedial approach.
B.5 REFERENCES
- 1. H. Codou et al. " Corrosion Fissurante Sous Contrainte de l'Inconel Dans l' Eau a Haute Temperature." Third Colloque de Metallurgie Corrosion, .
Centre d' Etudes Nucleaires de Saclay, France. Amsterdam: North Holland Publishing Co.,1959, p.161.
- 2. D. van Rooyen. _ " Review of the Stress Corrosion Cracking of Inconel 600."
Corrosion,1975, vol. 31, p. 327.
B-12 i
- 3. R. Bandy and D. van Rooyen. " Stress Corrosion Cracking of Inconel Alloy 600 in High Temperature Water-An Update " Corrosion,1984, vol. 40, p. 425.
4.- A. R. McIlree et al. 'Trimary Side Stress Corrosion Cracking." Chapter 7, Steam Generator Reference Book. Palo Alto, Calif.: Electric Power Research Institute, May 1,1985.
- 5. h Boudot, P. Vidal and G. Zacharie. " Evaluation of Residual Stresses Involved in PWSCC at Roll Transitions." In Proceedings: 1987 EPRI Woiksbor) on Mechanisms of Primary Water Intergranular Stress Corrosion Cracking. Palo Alto, Calif.: Electric Power Research Institute, September 1988.
NP-5987SP.
- 6. E. C. Ruud. " Residual and Applied Stress Analysis of Inconel Alloy 600 7/8-Inch U-Bend Tubing." Proceedings: 1987 EPRI Workshop on Mechanigms of Primary Water Intergranular Stress Corrosion Cracking. Palo Alto, Calif.:
Electric Power Research Institute, September 1988. NP-5987SP.
- 7. S. M. Bruemmer et al. " Microstructure and M!crodeformation Effects on IGSCC of Alloy 600 Steam Generator Tubing." Paper 88 presented at Corrosion 87, March 1987, San Francisco, California, NACE.
- 8. T. Yonezawa-and K. Onimura. "Effect of Chemical Compositions and Microstructure on the Stress Corrosion Cracking Resistance of Nickel Base -
Alloys in High Purity Water." Proceedings: 1987 EPRI Workshop _on Mechanisms of Primary Water Intergranular Stress Corrosion Cracking.
Palo Alto, Calif.: Electric Power Research Institute, September 1988.
NP-5987SP.
- 9. A. S. Stein. "Use of Activation Energy for SCC to Predict Life of Alloy 600 in Primary Water." Proceedings: 1985 Workshop on Primarv-Side Stress Corrosion Crackine of PWR Steam Generator Tubing. Palo Alto, Calif.:
Electric Power Research Institute, June 1987. NP-5158.
- 10. D. van Rooyen. " Update on SCC of Alloy 600 in High Temperature Water."
Proceedings: 1985 Workshop on Primary-Side Stress Corrosion Cracking of PWR Steam Generalor Tubing. Palo Alto, Calif.: Electric Power Research Institute, June 1987. NP-5158.
ill. Preliminary resuits of Studsvik tests at 10 and 25 cc/Kg hydrogen (on file at Electric Power Research Institute, Steam Generator Project Office.
- 12. F. W. Pement et al. "Effect of Operational Dissolved Hydrogen Concentration on Alloy 600 PWSCC in PWR Reactor Coolant at 315 and 330*C (599 and 626*F)." Presented at Fifth Intl'l. Symp. on Envir. Degr. of Mat'Is. in Nuclear
- l. Power Systems, August 1991, Monterey, California.
B-13
i
- 13. Report on May 9 and May 18,1989, visits by Electric Power Research Institute !
representatives to Studsvik and Electricite de France,"Effect of Lithium on PWSCC of Steam Generator Tubing". (On file at Electric Power Research !
Institute, Steam Generator Project Office.)
i
- 14. R. B. Rebak and Z. Szklarska Smialowska. "Strets Corrosion Cracking of j Alloy 600 in Pressurized Water Reactors - Primary Water. The influence of ;
pH and Stress Intensity." Presented at Fifth Intl'l. Symp. on Envir. Degr. of .
hiat'Is. In Nuclear Power Systems, August 1991, Monterey, California.
- 15. J. N. Esposito et al. "The Addition of Zine to Primary Reactor Coolant for .
Enaanced PWSCC Resistance." Presented at Fifth Intl'l. Symp. on Envir.
Degr. of Mat'Is. in Nuclear Power Systems, August 1991, Monterey, California.
s
- 16. H. J. Sc' enk. " Investigation of Tube Failures in Inconel 600 Steam Generator
, Tubing a 1 WO Obrigheim." Materials Performance, March 1970, p. 25,
- 17. R. Houben and P. Hernalsteen. "Part Depth Rolled Plants, Status of the Doel2 Unit." Proceediner 1989 Electric Power Research Institute PWSCC Remedial Measures Workshop. Palo Alto, Calif.: Electric Power Research Institute, .
April 1990. NP-6719 SD.
- 18. A. R. McIlree. " Status of PWSCC-USA." Proceedings: 1989 Electric Power Researebi 'tute PWSCC Remedial Measures Workshop. Palo Alto, Calif.:
Electric Poi Tesearch Institute, April 1990. NP-6719 SD.
- 19. .iteam Genera.or Progress Report. Palo Alto, Calif.: Electric Power Research Institute, November 1989. EPRI Research Project S405-3, Rev. 5. ,
- 20. J. W. Carter. " Operating Hist < ,, and Field Inspection Results of Trojan Steam i Generators." Workshop Proceedings: U-Bend Tube Cracking in Steam Generators. Palo Alto, Calif.: Electric Power Research Institute, June 1981.
WS-80-136.
- 21. D. B. Mayes. "When to Peen the Cold Leg Tubesheet." Proceedinns: 1989 EPRI PWSCC Remedial Measures Workshop. Palo Alto, Calif.: Electric Power Research Institute, April 1990. NP-6719-SD.
- 22. Preceedings: 1990 EPRI Workshop on Circumferential Cracking of Steam Generator Tubes. Palo Calif.: Electric Power Research Institute, March 1991.
- NP-7198-S.
l B 14
- 23. E. S. Hunt and J. A. Gorman. St.alus and Sugcested Course of Action for Nondenting Related Primary Side IGSCC of Westinghouse-Type Steam Generators. Palo Alto, Calif.: Electric Power Research 'nstitute, May 1986. i NP 4594-LD. l
Measures Workshop, Clearwater Beach, Florida. ;
i
- 25. P. Hernalsteen and C. Leblois. " General Status of PWSCC in Belgian Plants."
Proceedings: 1989 EPRI PWSCC Remedial Measures Workshop. Palo Alto, Calif.: Electric Power Research Institute, April 1990. NP-6719 SD.
- 26. P. Hernalsteen. Belgian Utility Experience Reports, October 1990 and June 1991, On File at EPRI.
i
. 27. J. Woodward and D. van Rooyen. Stress Relief to Prevent Stress Corro-lon in the Transition Region of Exoanded Alloy 600 Steam Generator Tubinc. 1 Palo Alto, Cahf.: Electric Power Research fnstitute, May 1983. NP 3055.
~28. Proceedings: 1989 EPR! PWSCC Remedial Measures Workshop. Palo Alto, "
Calif.: Electric Power Research Institute, April 1990. EPRI-6719-SD.
- 29. J. Stubbe and C. Leblois. " Belgian Utilities Experience with Nickel Electroplating and Sleeving." Proceedings: 1989 EPRI PWSCC Remedial Measures Work 3 hop. Palo Alto, Calif.: Electric Power Research Institute, i April 1990. NP-6719-SD.
1
- 30. D. Degreve and D. Dobbeni, "UT Developments for the Inspection of Electroplated and Sleeved Tubes," Proceedings: 1989 EPRI PWSCC Remedial Measures Workshop. Palo Alto, Calif.: Electric Power Research institute, April 1990. NP-6719-SD.
M d
I
}
t
- B-15 ;
., E -e- , e,m -e v -w+- r-, .,,wr.- .e-, - . , - ,1cr,g---%-1.-,-r,--ye -,=v-, --- v - --w e , e , w -*-r-*,-- -,y- c-gy>
)
4 Aroendix C NDE OF ROLL TRAN . W.NS-INDUSTRY EXPERIENCE This appendix provides (1) a listing of tubes removed from operating steam generators that has been used to evaluate ECT reliability for detecting and sizing EZ PWSCC, and (2) a discussion of some specific and unique aspects of the information developed as a result of the tube pulls.
C.1 TUDE PULL DATA Table C-1 is an incomplete but representative listing of tubes removed from operating steam generators for investigation of PWSCC, principally from steam i generatort with standard roll expansions
- For each tube, all known references which discuss metallographic rt ults and/or the correlation with inplant eddy-current results are also given. tubes are identified, which represent the bulk of the tubes removed for possible PWSCC at a standard roll expansion. For tubes which are actually degraded, axial PWSCC is the predominant
' degradation. Several exceptions in which circumferential PWSCC has been observed include tubes from Ringhals 4, Oh! 1, and V. C. Summer.
C.2 SPECIFIC DETAILS C.2.1 Correlation Between Pulled Tube Metallography and in Plant Eddy Current Data ,
C.2.1.1 Detection of Axial Cracking.
Ringhals 4 (SG2 Tube R22-C97). This tube was removed because of a long axial eddy-current indication at the roll transition (see figure C-1). The destructive exam confirmed a axial crack from the roll transition zone downwards. The eddy-current crack length was approximately . No other cracking was identified during metallography. A comparison of test results using various examination methods is shown in figure C-2. The long .
length of the crack was probably the result of incomplete expansion of the tube >
in the region just below the top of the tubesheet, which lead to an abnormal tapered profile (see figure C-2). The whole length of crack was below the top.of -
the tubesheet.
'Tuttes from Ringhals 3 and Doel 3 have kiss-roll transitions.
C-1
Figure C-1, Ringhals 4 - Tube R22-C97.- RPC Isometric - High and Low Sensitivity. Indication Just Below Roll Expansion,-
C-2
i r
l E
? !
i i
f I
1-t t
9 .p ,
w :
- i 1-s
[
i t
i l
r P
r i I l
t t
i i
. . t Figure C-2. Ringhals 4 -Tube R22-C97. Comparison of Composite Test Results. !
e i r M i t-
, * ,~ ,;-.'
, . - . r + , . ., , .- ~ . - - - - . . + - -' - - - - - - - ,,---- -- __ _ _ _ - - - --
Qampierre 1 (SG2 Tube L12 C43). This tube (kiss rolled) was removed because of multiple (seven) arial eddy current indications. Eddy current data was analyzed using the French *ESTELLE' analysis system. Figure C 3 shows a comparison of the RPC graphic analysis of tube condition with that determined destructively. A direct comparison of crack lengths is given in the table shown in the figure. ESTELLE indicates the lengths of longitudinal cracks within accuracy, provided the cracks are through wall in depth. System perfor-mance is such that it enables detection of isolated, longitudinal through-wall cracks in length or greater.
Doel 2 (SG Tube R13 C30). Figure C-4 shows a comparison of tube condition as determined using eddy-current RPC with destructive metallography results.
The numbers at the upper part of the figure identify crack location whereas the individual cracks are identified by the numbers one through ten shown in mid-figure. The lower set of numbers are the eddy-current and metallography crack lengths. Correspondence between the two data sets is generally good both in terms of identifying individual cracks and predicting crack length. One crack
(#10 on the far left of the figure)is not recognized by the eddy current because of masking effects by the larger crack adjacent to it (#1).
Doel 3 (SG Tubes R23 C32 & R27 C52). Figure C-5 shows the actual shape of two axial cracks from two tubes removed from the Doel 3 steam generators (a kiss-rolled unit). Both cracks were through-wall with different lengths on the tube inner and outer wall. Both cracks were detected using RPC eddy current with the length estimate shown below each crack cross-section for comparison with the "true length" which is different on the two sides of the tube. As expected, the eddy current provides some sort of estimate of crack length.
C.2.1.2 Detection cf Citcumferential Cracking.
Rinehals 4 (SG1 Tube R47-C50). This tube with a standard roll transition was sele'cted for removal because of an eddy-current indication of an abnormallong circumferential crack, _with a small axial component at one end. An isometric -
plot (end-on and oblique views) of the inplant eddy-current RPC data is shown in figure C-6. Flaw orientation is deduced by observing the orientation of the major axis of the eddy-current indications. In this case, two orthogonalindi-cations are present, suggesting both axial and circumferential cracking Metal-lography results did confirm the presence of both cracks; the length of the circumferential crack was -the axial crack length was not given.
Figure C-7(a) shows the circumferential crack while figure C-7(b) shows the axial extension emanating from one end of the circumferential crack.
C-4
I l
, .J I
b i
i.
1 i
?
9 7
3 Figure C-3. Dampierre 1 - Tube L12-C43. Comparison of RPC and Destructive .
. Test Results.
C-5 I
I
Figure C-4. Doel 2 Tube R13-C30. Comparison of RPC and Destructive Test Results.
C-6
- - . . . . . - . . . - _ . . . - . . - . - - . _ - - . . . ~ . . . . . . - - . - . - - - -
i
-f 1
E i
1 t
t t
t I
i i
i i
- i. .
t
)
l 9
5
?
I 4
i
.f f
1 ,
Figure C-5. Doel 3 Tubes R27-C52 and R23-C53. Actual -
Shape of Through-Wall Cracks and Comparison with RPC Length Measurement.- ~
')
C-7
+maww + w> w s - e -- -r-c.,.+w- -d ,r-re-e,,,.v-w-.wa-> n, rm + iv,-n on s,,,- -oww.,,-esm-y'~~ t , e h s w, w.v rw---, , w.v ,r s , w- r,--vr +- w~~~-
piilllIn 4 i. . -
F gure IUnghals 4 - Tube R47-C50. Axial and Circumferential C .
.)
_.. . ,.i..,.
i h
i 4
jL t I
d 7
i I 6
- t i >
t a
t n.. t
'o i 6
1 i 4
t 1 '
1 i,
4
' i L
k
.h .'
il
, L i ;
4z '
Figure C-7. Ringhals 4 -Tube R47-C50. Axial and Circumferential Cracks. l t
b y h
, t
., ,4 . .,-- .- , m . . , , ,_ , , , . . , . , , - .. . _ . , , . - .,__ . _.
Dampierre 1 (SG Tube L8 C48). Another example of a tube kiss rolled with an isolated circumferential crack is shown in figure C-8 removed from Dampierre 1. Shown is a comparison of the destructive analysis results and the initial inplant eddy-current prediction. For circumferential cracks, the detection threshold is approximately arc length, with a depth of at least of the tube wall thickness.
C.2.1.3 Nondetection of Short Axial Crackine.
Ringhals 3 (SG2 Tube R18-C79). This tube had been kiss-rolled. No eddy-current indications were reported by the field analyst. Metallography confirmed the presence of numerous ID-initiated cracks; they were located in the area between the top of the main roll and the kiss roll. Maximum crack length was approximately . Figure C-9 shows a view of the ID surface of the tube; the multiple short cracks are apparent. A pictorial summary of various test results for this tube are shown in figure C 10.
C.2.1.4 Mitsking Effccts.
Rinehals 4 (SG1 Tube R17 C81). This tube was removed because it contained the ' greatest number of axial indications around the tube circumference. No circumferential indications were identified. Metallography revealed axial cracks, the longest of which was , with another five approximately
. An unusual circumferential crack was also identified with a length of . Figure C-11 shows the RPC isometric; the presence of multiple cracking in close proximity is apparent with most indications appearing to have their major axes aligned in the axial direction. No clear evidence of circumferential cracking is indicated. In this case, the eddy-current coil is averaging over several coil diameters, causing a masking of individual indications. Figure C-12 shows composite test results for this_ tube, using _ _
various methods. The circumferential crack was only observed after the tube was flattened. It should be noted that short circumferential cracks of the type not detected in this tube do not significantly affect burst or leakage behavior.
C-10
__j
t b
2 1
i
?
1 i
b
.I Y
l 4
l Figure C-8, Dampierre 1 - Tube L8-CE Circumferential Cracks'in Kiss-Roll Transition Companson of _RPC and' s l
Destructive Test Results. ,
9 t
7 C-11 e i e .c..*,', e . >w,c u.v., - g ~ -: - E ., cr,n r--9,:sp-,-- r{<*
i r
i t
1
.i i
i a
i k
't
't
)
i i I 3
k 4
r J
- Figure C 9. Ringhals 3 - Tube R18-C79. ' Short ID Axial Cracking..
C-12 ~
wt 3 w& wre- u.+== , e-- 4-w..,,r.,.-,-- --,ef.-, - - - . - , ..;m-e-,-+- . ,.--.ms4.--.--,,,.,. .. . - - . - - ,,,c w s- i.-*-,... ,wi4.vww4 . a,,4.
- c
> ;L ,7 It ,;t,ttlIl.;Fr(t[;(ib! r!t Ei;p,b\f>!L.>( it if [: y *- ;r i[ 1 i ,5 ' u 1PI n
,~
t s .
l u
s -
e _
R .
t _
s e .
Te v t
i
. s o
p s m
o >
C ~
9 c 7 -
C -
8 1
R e .
b ~
u T
3 .a l
s a
h g .
.~
i n ,
R ,
e 0 -
1 C- ~
e r
u g
i F
e n
n,pw '-
v.
4 s
J bE
. ; o .
ii i Li i '
i1!i,
- , 4 j: .
$.f ! !i:!.:r4 1111ii1i;1j. :,l4-
". y
pIiEBie6i Figure C-11. Ringhals 4 - Tube R17-C81. Multiple RPC Indications.
C-14 iruri ii
.- 4.4-em4 A422wa2._--ma_.-.m4.._ .
-, +4us a s 4- n :-4-.~ .--a----w s -"AA ma me . a A_ mL w -um - - mam..._AAm a ux 4a . .w 2,.. Jeu.* _eJ==+- A4 s e m. ..
P 1
i e
h
'k 6
s
?
h
- I 0
M >
H v
.5 L
E o
V
[
I s.
e-
$a
. b.
T-2 ra
.C to c
s- 'y e1 0
V o>
w U
to
.C e n
5
?
1 r
h C-15
.e l'
. _ . . . . . . _ ~,. , .~-.
C.2.2 Improved Detectlan of Ciromifnential Cracks in the Presence of Multiole ~
Axlal Cracks Observations derived from laboratory studies'and tubes removed from operating plants have provided estimates on isolated crack detectability, resolution problems associated with the occurrence of multiple cracking, as well as highlighting special difficulties associated with the detection of circumferential crack initiation in the presence of multiple axial cracking.
The resolution properties of an eddy current pancake coil are determined by the mean coil radius. The field of view of the coil-as determined by the average coil radius-is finite which means that multiple cracks present simultaneously within the coil field of view will be averaged as a single composite response. This effect can cause the output response to appear as a single crack even though several cracks are present. The occurrence of small circumferential cracks 'altiating and radiating from larger preexisting axial cracks has been a concern because of masking effects by the larger axial cracks. Similar to detection of single cracks, this situation can be improved somewhat by using coils of smaller diameter.
A series of laboratory experiments were conducted to define the ability of the RPC eddy-current probe to detect circumferentially oriented degradations in the presence of through-wall axial degradations. Those experiments are described in appendix H.
They show that the conventional pancake coil clearly resolves a circumferential notch between axial notches separated by , with a degraded capability beginning at spacing between axial notches of or slightly greater. The (shielded, i.e. focused) pancake coil offers significantly improved resolution, with the circumferential notch visible down to a axial notch separation.
i
\
l ,
C-16 ,
I I
i Table C-1 ,
LISTING OF TUBES REMOVED FOR PWSCC f J
SG/ Row Plant Tube # Column Pulled Mechanism Location i i
I
' Full-Depth Roll Documents ,
Tubes (1-3) i
'(,
C-17
n i
1 Table C-1 (Cont'd) -
)
l Plant Tube #_ Coluntn Pulled Mechanism Location ;
i
' Full-Depth Roll ;
Documents f w
+
I l
I C-18 1
, ere wm, w-.w., ,',cw- m -
l l
i f
Table C-1 (Cont'd) i T
SG/ Row l Plant Tube # Column Pulled Mechanbm Location j i
t i
t t
i
- Full Depth Roll
- Documents i
b i
T t
L C-19
< _a _ -_ - _ . - - _ , . _ . - - - _ . - _ _ _ -
_ _ _ . _ _ - - . ._u.,- _. .__ _ _ -
i Table C 1 (Cont'd) l SG/ Row Plant Tube # Column Pulled Mechanism Location ,
r.
I I
' Partial Depth Roll- SKIP ROLL ;
Documents t
w 5
F 1
b 5
t I
+
3 J
C-20
+ t<'-, - o=~a w s,v v , w r -r,-~vw- w, vew,r.,w,-rw-reer,,,r-a,-ww+r--w, ,--~wv, r--w-,w -
r,w.,--wr-w,wn -v,+ ~-e------w, --
.. - . . ~ . . . . . . - . . . . - - . - . . _ - . - -.- - - - . . . . . . .- -. .... . .
l l
l 4
l i
l Table C-1 (Cont'd) i SG/ Row i Plant T u b e #. Column Pulled Mechanism Location F
Partial Depth Roll ,
1 Documents ,
E s
. I p
i I
+
5 l
l i.
I l
l-l l'
1.
i +
C-21 W"-rv -T =Yes= y'Merf e-e-ev 5 W--h+-r9n. ema.er-.-.r,m-emy--e-----tmW-:vtn++m+em*r4-+rr------me-w eW e--e'-0-=r-T--'w'T- ' tert'CTt-t we--T'3+T Y WW=+w$d+" * - -'I ~ -W'" TF '" Y V'M TN ' W
t Table C-1 (Cont'd)
SG/ Row Plant Tube # Column Pulled Mechanism 1.ocation -
F i
I Partial Depth Roll-l t 2 Sleeved ivith B&W mististeeve 3Re-expanded .
C-22
Table C 1 (Cont'd) 4Electroplated SUsed for NDE qualification (No destructive examination)
Documents C-23
. . _ _ _ _ . _ _ . . _ _ _ . . _ . ~ . . . --_-- - _ . ..-- -- -
4 A
i i
i h
i Table C-1 (Corit'd) ,
l 1 ?
I e
_b 1 I i
i l
f i
.I
)~ t
.t ;
a i
i b
'I t
i 4
M t -' !
1' t
i
.I t
.i '
I
}4 I,
'1 C-24 M g c e y 1--pu-sy yer p -p g g-eN -WyhpWp gepqm ww yw wmi g gagy_
r Table C-1 (Cont'd)
SG/ Row Plant Tube # Column Pulled Mechanism Location >
9
' Full-Depth Roll -.
t Documents I
4 s
P P
.C-25-~ ;
I n%
- 4 Table C-1 (Cont'd)
SG/ Row Plant - Tube' # ' Column Pulled Mechanism' Location h
3
' Partial-Depth Roll
- Documents t
C-26
_., . , .,_....m, . . _ . . . ~ , - . . . . . . _ . . . . . , . . . .
..... - . . . . . - . . . .. . .. . -- . .. ..- - .- . . ... _ .-.-. -. ,~. -. . .. -
3 E
p P
Table C-1.(Cont'd).
l N
r k
j, s
i h
..f.
I e
9 e
Y h
- t %
1
.r
-j iC-27.
.s i ? T *' -gy =*v w
- y g yg. g . , =.-qage yg a+m-y-w Iqw.- O y 7
t Table C-1 (Cont'd)
SG/ Row-Plant .T u b e # Column Pulled hiechanism Location ,
1 o
b
' Full-Depth Roll with' of tubes kiss-rolled.
- Kiss-rolled. -
Documents -i C-28 1
. .. .. u -.
1 1
l Table C 1 (Cont'd) -
SG/ Row . .
Plant Tubs # Coltimn Fulled hiechanism Loca tion '
7 i
Ft.ll Depth Roll Documents .
C-29
Table C-1 (Cont'd)-
SG/ Row Plant Tube # Column Pulled Mechanism -Location
- Full Depth Roll Documents
-1 C-30 4 I
l
-Table C-1 (Cont'd)
SG/ Row-
_ Plant __ Tube # Column Pulled Mechanisrn Location -
t oPartial-Dept isoll - Double Roll Documents .
2 C --
- W-
Appendix D BURST DATA FOR PART-THROUGHWALL AND THROUGHWALL FLAWS IN STEAM GENERATOR TUBES
- 94 J-
'These data are in addition to data from major tube burst studies used to develop the tube burst equation as described and referenced in section 2.0.
D-1
Table D-1 NRC PHASE 11 BURST TESTS, TUBES WITH AA1 AL SCC a (Nominal Tube Size: -
Thickness)
Maximum (Sy + Su )b Axial Crack Range of Normalized Pressure (ksi) Length (inches) Crack Depth, a/t (psi)
Ref: a. R. J. Kurtz et al., " Steam Generator Tube Integrity Program Phase II Final Report."
NUREG/CR-2336, U.S. Nuclear Regulatory Commission, Washington D.C., August 1988.
- b. J. M. Alzheimer et al., " Steam Generator Tube Integrity Program Phase I Report",
NUREG/CR-0718, U.S. Nuclear Regulatory Commission, Washington, D.C.,
September 1979.
D-2
Table D 2 CE S/G TUBE BURST DATA,565F AX1AL IG A CRACKS SIMULATING THOSE IN ST. LUCIE 1 (Nominal Tube Size : Thickness)*
Specimen Maximum Crack Range of Pressure Length Normalized (psi) (inch) Crack Depth, a/t Comments p
s Ref: W. J. Heilker et al., " Tube Burst and Leakage Testing of IG A and IGSCC Defects Representative of Those Found in the St. Lucie, Unit One Steam Generator." CE Report CEN C-1740, Combustion Engineering, August 1986.
(CE Specification)
(Mill Tests)
D-3 1
Table D-3 l
WESTINGHOUSE S/G TUBE BURST DATA, l EDM NOTCHES IN ROLL TRANSITION REGION AXIAL THROUGH-WALL CRACKS 1 Axial Through-Wall' No. of Slots Maximum Slot Length Around Pressure (inches) Circumference (psi)
L Westinghouse reports (EPRI contract no, RPS 301-9) these burst data for EDM axial throughwall flaws in mill-annealed - . tubes with wall thickness. . The material yield and ultimate strengths were
, respectively.
l' D-4
Appendix E LEAK RATE DATA FOR STEAM GENERATOR TUBES WITH SCC IN THE EXPANSION ZONE ROLL TRANSITION REGION E-1
-j Table E c CE S/G TUBE LEAKAGE TEST RESULTS AXIAL ROLL TRANSITION CRACKS,-
Specimen Nominal Tube Axial Crack Length Leak Rate Number - Diameter * -(inches) @
(inches) OD ID 1
I
-Ref: D. E. Powell et al., " Steam Generator 3 Tube Leakage Experiments," EPRI Report -
NP-6897-L, Electric Power Research Institute, Palo Alto, CA, July 1990.
i E-2
I Table E-2 WESTINGHOUSE LEAK RATE TESTS 1
AXIAL ROLL TRANSITION CRACKS AT Nominal Specimen Crack Length Leak Rate Number (inches) @
IWestinghouse reports (EPRI contract no. RPS 301-9)these leak rate data for SCC axial cracks in mill-annealed diameter tubes with wall thickness. The material yield and ultimate strengths were , respectively.
E-3
l Table E-3 IN-SERVICE LEAK RATES - AXIAL ROLL TRANSITION CRACKS 1
Nominal or Average Axial Crack Length Leak Rate Plant (inches) a Laborelec reports this leak rate for a tube leaking in service and crack length.
measured _by RPC.
b EdF data provided to EPRI.
E-4
. - . - . - -. .- ~.
l
?
Appendix F CRACK GROWTH RATE DATA F.1 INTRODUCTION The crack growth rate of EZ PWSCC can be estimated from sequential measurement
. of EZ PWSCC cracks in service. Data to estimate crack growth are available for two plants with kiss roll transitions and a program has been initiated to generate data for a plant with standard roll transitions (D. In this appendix the results of analyses of the kiss roll data are presented along with estimates for standard roll plants and plans for generation of standard roll data. In the future,it is anticipated that utili-ties will generate plant specific crack growth rate data for estimating an allowance for crack growth based on repeated measurement of EZ PWSCC cracks left in service.
F.2 KISS ROLL DATA -
The Belgians have performed RPC of kiss roll transitions in Doel 3 and Tihange 2, using a highly accurate automated system, for three fuel cycles. These inspections provide EZ PWSCC crack length measurement data for three points in ,
time. By taking the difference between crack length measurements and dividing by the time interval between measurements, crack growth rates have been calculated.
. Statistical analysis of the resulting crack growth data has been carried out by
' LABORELEC. Their results are shown in table F-1. Table F-1 gives the average, standard deviation and maximum growth rates as a function of nominal crack ilength for individual plants, steam generators and time periods.
To develop a better picture of the growth rate of EZ PWSCC cracks as a function of .
crack size , the individualplant, steam generator and time period data points were combined in table F-2. ;In addition, these data were fitted to a quadratic equation
- using a least squares method. The values were weighted by the number of observa-
-tions at each crack length. Growth values were converted to EFPY by dividing by ,
a typical capacity factor for Belgian plants. After reaching minimum growth rates at
' lengths of about ' , the growth rate was modeled as being constant out to a length of ). The results of this " fitting" or " smoothing" are shown in figure F-1.
l l
!- F-1 l
Table F-1 LABO.RELEC ANALYSIS OF EZ PWSCC CRACK GROWTH DATA FROM DOEL-3 AND TIHANGE-2 X=
F-2
Table F-2 CRACK GROWTH DATA FROM BELGIAN PLANTS WITH KISS ROLLS Average Growth Standard Initial Lennth Per Annual Cycle No. of Deviation mm in mm in Tubes mm in Note: Length shown as actually ranged from .
F-3
-i t
! -'I
.1 i
Figure F-1. Average Crack Growth Rates Based on LABORELEC Repeated Measurements of Kiss and Standard Roll Transitions With EZ-PWSCC In'dications..
1 4
4
'F-4
~!
It should be pointed out that there has been no attempt to separate from the apparent variability of crack growth rate data, the variability due to the uncertainty in crack length measurements. Referring to table F-2, the standard deviation in
~
crack growth rate ranges form - _
. As the reported maximum uncertainty for crack length measurement for the Belgian RPC inspection process is on the order of , it is likely that apparent crack growth rate variability is significantly influenced by crack length measurement uncertainty.
F.3 STANDARD ROLL DATA As there was not a similar database for EZ PWSCC growth rates in standard roll transitions, EPRI initiated a program to generate the needed data. Specifically, EPRI contracted Belgatom to_ have three successive RPC inspections analyzed by LABORELEC on a large sample of standard roll transitions in a Doel 2 steam generator. The first of the three inspections was performed in 1989 (1). The second was performed in 1990 (2).
Doel 2 steam generators have partial-depth roll tubes with the standard roll transition occurring well within the tubesheet. Roll transitions have not been shot peened and significant EZ PWSCC has been identified. As the PWSCC is confined by the tubesheet, crack indications have been allowed to remain in service. A total of steam generator A tubes with track indications in 1989 were again inspected' in 1990.
Average crack growth rates as a function of initial crack length are shown plotted in figure F-1. The reported growth rates in terms of extension per year have been converted to extension per effective full-power year (EFPY) by dividing by Referring to figure F-1, note that growth rates for standard roll cracks are significantly lower at small crack lengths and at large crack lengths than that of cracks in kill roll transitions. These differences are likely due to different residual stress patterns in kiss and standard roll transitions (3). The average growth rate for initial crack lengths in the range of is similar to that observed in kiss rolls and is approximately .
With limited data, the conservative growth rate curve for kiss roll cracks shown in figure F-1 and is used in section 4 to calculate allowable crack lengths. A change can be considered when additional standard roll crack growth rate data are available.
F-5
l F.4 REFERENCES
-1. D. Dobbeni, G. P. Benoit, and W. Ferket. Roll Transition Inspection of Doel-2 Steam Generator Tubes. Volume 1: Tune 1989 Inspection. Palo Alto, Calif.:
Electric Power Research Institute, March 1990. NP-6716-L, vol.1.
- 2. D. Dobbeni, G. P. Benoit, and W. Ferket. Roll Transition Inspection of Doel-2 Steam Generator Tubes, Volume 2: October 1990 Inspection. Palo Alto, Calif.:
Electric Power Research Institute, June 1991. NP-6716-L, vol. 2.
- 3. R. Houben and P. Hernalsteen. "Part Depth Rolled Plants, Status of the Doel 2 Unit." Proceedines: 1989 Electric Power Research Institute PWSCC Remedial Measures Worksiloo. Palo Alto, Calif.: Electric Power Research Institute, -
April 1990. NP-6719-SD.'
F-6 l
- - - - _ ________2___._______ ____ ___ ____ _ _ ___ __,_ ___ _ ______
Appendix G ANALYSIS METHOD FOR LEAK RATE CALCULATION G.1 INTRODUCITON This appendix describes a methodology for evaluation of the primary to secondary leakage in a steam generator with a large number of through-wall axial cracks. The methodology comprises a probabilistic approach (Monte Carlo type) combined with a simple deterministic model for leakage from an axial crack.
G.2 BASE DETERMINISTIC MODEL The deterministic model for leakage from an axial crack in a tube with internal pressure is given by the following formula:
Q=
where Q is leak rate (volumetric)
K is the discharge coefficient (dimensionless)
S is leakage area P is effective differential pressure P is liquid density The key parameters AP, A and K are established as follows:
G.2.1 P (Effective Differential Pressure)
For " single-phase" flow of water, the formula holds true if P is taken as the actual
~ differential pressure acting across the tube. When the primary saturation pressure is larger than the actual secondary pressure, as is the case for accident conditions, the -
leak flow will flash and the leak rate will be limited to a critical value by a choking l process. A simple way to approximate this behavior is to use an " equivalent" or l
" effective" differential pressure by assuming a down stream back pressure equal to l _the saturation pressure; this approximation is valid for subcooling down to a range of G-1
. G.2.2 _SJLeakage Ar. cal The leakage area is obtained by summation of an elastic and plastic component S=
- Elastic Component (Sei). The elastic crack opening area (Set) of a through wall axial crack is Set =
M=
X=
where a = crack length a = value of X for the crack length under consideration o = circumferential membrane stress E = modulus of elasticity m = " bulging" factor as used in the " collapse load" theory of tube failure il = Poisson's ratio R = average tube radius t u tube wall thickness
. Plastic Component (Spi). The plastic crack opening area of the crack (Sp t)is a f unction of the tube material properties. This component gets larger when the differential pressure approaches the burst pressure. Sp1 can be expressed as a nonlinear function of crack length and the ratio of either actual pressure to burst pressure or crack length to critical crack length. This is illustrated by figure G-1 on the basis of laboratory tests.
G-2
7 -
,1
. Figure G-1. Plastic Component of the Crack Lea kage Area i
l G-3
.i
G.2.3 K (Flow Discharge Coefficient)
Under normal operating conditions, the crack opening area is small and there is a-major effect of the tortuosity of the leakage path on flow rate. This is caused by the intergranular nature of the stress corrosion process and is dependent on the material-grain size, resulting in multiple flow restrictions and deviations. The consequence is a low apparent value of K which is difficult to estimate other than by curve fitting to the available field data (such as the curve of figure 5-1).
For larger crack opening areas, as should be expected under accident conditions, the value of K will gradually increase up to a maximum value of about Thus, the K value to be used is defined as a function of the calculated average crack width (leak area 8 (6 = divided by crack length a). It should be realized that such a function is more reliable for larger crack widths and K values; significant uncertainties are associated with tight cracks under normal operating conditions, which are prone to clogging up to the point of being almost leak tight. Figure G-2 shows three functions (Curves 1,2 and 3) used to model K over the range of interest in this work.
Note: There are not yet reliable criteria for selecting the "best curve." Curve 3 is based on theoretical considerations and is believed to be conservative. Curves 1 and 2 result from approximations of actual S.G. leakage behavior. The examples (figures G-4 through G-10) are based on Curve 2.
Figure G-2. Flow Discharge Coefficient Assumptions ]
1 G-4 !
i l
I i
G.3 MODEL VALIDATION ,
An integrated validation of the model was obtained through performance by Belgatom (Laborelec) of large scale testing of ~ tube specimens with artificial flaws; the test rig uses the feedwater pump of a Belgian fossil fueled plant and allows bursting of test specimens with through wall degradations, without any scaling bladder (pressure up to ; flow rate in excess of ).
Figure G-3 illustrates the results from two identical test specimens, provided with a through-wall EDM flaw. One specimen did burst at .
, while the other did not, so that it was possible to register the leak rate during the pressure decrease; from comparison of the flaw deformation with prior laboratory experiments, it was concluded that of the burst pressure had been reached.
Figure G-3 also shows, in solid line, the leak rate predictions for the full-cycle (pressure up and down); the agreement with test data is quite good.
The calculation process is detailed hereafter for a pressure equal to of the burst pressure l
l.
Figure G-3. Leakage Model Validation G-5
The leak area is made up of three components:
- Initial Flow Area. For this case of an EDM slit with a width, S in =
- Elastic Eaw Opening. For the considered pressure level (P = ) and crack length ( , leading to ). =
and M = ), the clastic component is given by Sei =
valid for both the pressurizing and depressurizing phases.
- Plastic Flaw Opening. The plastic component is taken from figure G-1 at two pressure levels
--P/Per = for the pressurizing phase, leading to Spl =
--P / Per = for the depressurizing phase, leading to Spi =
The total leak area is thus
. S+ =
S=
In both cases, the crack width is large enough for using the asymptotic value
~
(K = ) of the flow discharge coefficient.
By using the formula Q=
the calculated leak rates are thus:
Q+=
Q=
This validation applies to the range of large plastic deformation of " rectangular" shaped flaws (ID length = OD length); more testing is intended with variable flaw shapes (shape factor = ratio of OD length to ID length).
- However, validation about the reduced value of flow discharge coefficient, associated with small crack widths, can only be produced by a different kind of tests, -
measuring small leak rates through actual corrosion cracks; such a program is also underway as mentioned in paragraph 5.2.2.
G-6
. G.4 PROBABILISTICMETHODOLOGY The probabilistic methodology consists of a stochastic combination of (1) the statistical
. distributions assumed for each of the parameters involved in the base deterministic
-model, and (2) the population of crack lengths established by in-service inspection of the steam generator. It results in a Monte Carlo type evaluation, with repeated calculation of the base formula for every possible combination of crack length and associated set of relevant parameters. Appropriate statistical distributions are assumed for:
- Mechanical properties (Sy + Su) at operating temperature.
- Correlation coefficient between " flow stress" (c) and mechanical properties.
Cs =
- Tube wall thickness.
- - Sizing uncertainty on crack length (based on RPC inspection technique).
- Expected crack lengths at the end of next (or current) operating cycle.
The first four distributions can be approximated by Gaussian curves while the last distribution needs inspection results and some knowledge of the crack growth of the population of interest.
Deterministic algorithms and input values are further used to take into account:
- Reinforcing effect of the tubesheet in accordance with figure 4-4 (Eqs. 4-4 and 4-5).
- Shape Factor. Empirical formula derived from-LABORELEC laboratory experiments (plastic deformation of specimens with axial artificial
- flaws of trapezoidal shape).
- Repair Limit. Arbitrary value;in the examples a value of 11 mm has been used as stated in table 5-1.
l G-7
.=_ _
G.5 COMPUTER PROGRAM A version of the above methodology has been developed into a software program, ,
written in APL language and implemented on a personal computer (L1)._ The-program is user-friendly and does not require any special computer training.
A simple " input data screen" can be edited to select the particular set of parameters representative of the steam generator under consideration. These include:
- Mechanical Properties (S y + Su). Average and standard deviation values.
- Flow Stress Coefficient (Cs). Average and standard deviation values.
- Eddy-Current (RPC) Crack Length Sizing Error. Average and standard deviation values.
- Tube Wall Thickness. Average and standard deviation values.
- Crack Length Distribution. 3 options are available:
- Uniform distribution. Used only to get information about the relative " weight" of each crack length in contributing to the total leak rate.
- Given distribution of crack lengths (histogram). This allows the input of any kind of distribution expected at the end of the operating cycle, as derived from the inspection data at the start of cycle.
- Probabilistic model of steam generator denradation. as validated for PWSCC in kiss-rolled and peened roll transitions (figure G-3). This includes input of 3 characteristic parameters of the initial (start of cycle) crack length and crack growth distributions and the required repair limit.
- Flow Shane Factor. Coefficients of the linear relationship between ID and OD crack lengths.-
- Flow Discharge Coefficient. Selection of either constant K value or three optional functions of crack width.
- Primary Pressure and Temperature: Secondary Pressure.
- Number of Cracks.
1 l
1 G-8
The output data are produced as a variety of useful plots; typical plots are illustrated below.
Ege:1 lup31LData This is a summary table of all input data.
En. G-5 12is,ribution of Crack Lengths This is a graphic representation of the specified input (inspection results at the start of cycle).
Fig. G-6 Distribution of Actual and Critical Crack tengths This illustrates the range of:
- Critical crack lengths for the specified canditions; variation is caused by variations in (S y+ Su), Cs, and t (iube wall thickness).
- Predicted crack lengths at the end of cycle.
- Comparison of these two distributions gives a picture of the overall safety margin with respect to tube failure.
- The calculated " probability of steam generator tube rupture"is also indicated.
Eg. G 7 Average Crack Openinn Width as a Function of Crack Length This illustrates the nonlinear relationship between width and length of cracks under the specified conditions; the elastic component of the crack width is also indicated.
Fig. G-8 Distrihntion of Crack Opening Widths This illustrates, for any particular crack length, the variation of crack widths from the average value given by figure G-7.
Fig. G-9 Leak Rate as a Function of Crack Length This illustrates the contribution of each crack length category to the total leak rate. No contribution is observed for the shorter cracks while the peak contribution is influenced by the crack length distribution (figure G-6). The "best-estimate" value of the total leak rate is also indicated.
Fie. G-10 Distribution of Tube Leak Rate This illustrates the relative participation of individual leakers.
Bg;_ Cell Distribution of Steam Generator Leak Rate This illustrates the expected variation of the total leak rate for the given number of cracxs. It allows selection of leak rate values at any predefined confidence level.
G-9
. . _ . . .. _ _ _ _ . _ _ _ __. _______.._m._. _ . _ . _ _ _ . . _ _ . _
I i
I J
i i
INPUT DATA ,
t Nominal Tube Diameter = ,
Gauss Distributions Mean Sigma :
Tube Wall Thickness (mm) hkchanical Properties YS + UTS (MPa)
Flow Stress Coefficient Eddy-Current Sizing Error 'mm) ,
Number of Cracks =
Crack Length Distribution Model: ; Plugging Limit =
Crack Shape Factor: OD length = l Flow Discharge Coefficient According to Curve 2_ _ _ _
Primary Pressure (Temp) = Secondary Pressure =
Figure G-4. Input Data t
G-10
l
\
t i
i l
l l
i F
i i
i Figure G-5. Distribution of Crack Lengths t
i Figure G-6. Distribution of Actual (End of Cycle)
Critical Crack Lengths -
G-11
_._._...__-.,.-_____-.._..-..:_...-___c-....__..
4 1
P i
i J
t i
i 6
i t
i i
Figure G-7. Average Crack Opening Width as a l Function of Crack Length -
r i
Figure G-8. - Distribution of Crack' Opening -
Widths (Illustrated for a crack)-
. G-12 ;
- _ .._,_.:. _ . _ _ _ _ , , . . , _ _ . . . . . . _ . _ - . . . , . . . _ _ _ . , . _ . , _ . _ _ , . _ _ , , _ _ _ _ _.._1.._..
_ . .. a; -
i l
1 l !
t I
h t
i i
L i
I e
f Figure G-9. Contribution of Each Crack Length '
Category to the Total Leak Rate
-t t
i t
l~
\
-i f
3 Figure G-10.' Distribution of Tube Leak-Rates G-13 ,
f r T'- y e-- N'T -- 1r- -g y yyg b 9- g '% $ y- etyv ge wny y g auer - gp wdwv*-' W w wgwgreiseV'g- --r--Pprr+gs amat v gwq d 'y
i e
L:
i s
i i
i i
i i
i i
.4 Figure G-11. Distribution of Steam Generator Leak Rate .l t
- G-14 a
heMm e . medie
__hw tme e- wilt-r-wsermev m 4-+ hwtede ww es-welveh --- -- - - - - - - -
t i
G.6 REFERENCES
- 1. P. Hernalsteen. "LABOLEAK: Laborelec Leaktate Probabilistic Assessment -
Program for PWSCC Cracks in Steam Generator Tube Roll Transitions - User's '
Manual." LABORELEC Report M00 90-023/R PH, May 1990.
- 2. P. Hernalsteen. "LABOLEAK: Laborelec Leakrate Probabilistic Assessment Program for PWSCC Cracks in Steam Generator Tube Roll Transitions -
Description and Validation." LABORELEC Report M00-90-024/R.PH, May 1990.
J l
G-15 i
)
- . ., ,,L._._..--,-.,-.___.=__,,n, - -
1
i Appendix H COMBINATION AXIAL AND CIRCUMFERENTIAL CRACK DETECTION UTILIZING RI'C EDDY-CURRENT PROBES
11.1 INTRODUCTION
A series of laboratory experiments were conducted to define the ability of the rotating pancake coil (RPC) eddy-current probe to detect circumferentially oriented cracks in the presence of through-wall axial cracks simulated by using EDM notches.
The results of these experiments are summarized in this appendix.
In principle, the RPC is equally sensitive to both axial and circumferential cracking.
Crack orientation is inferred by noting the direction of the major axis of the indication; an axial crack-of sufficient length--will have its major axis aligned or directed along the tube longitudinal axis. A circumferential crack--again of sufficient length--will have its major axis directed along the tube circumference.
H.2 NDE EXPERIMENTS U-shaped EDM notch test samples in expanded tubes with tubesheet collars were fabricated in order to investigate the detection and resolution capability of the RPC.
U-shaped notch geometry and dimensions were specified based on structural considerations. A drawing of the notch geometry showing typical shape and location relative to the upper roll expansion is shown in figure H-1. A summary of U-shaped notch dimensions for the test sample matrix is given in table H-1.
Test data was taken using three pancake coil configurations. These included and unshielded coils and a shielded coil.
Front and rear view oblique RPC graphics for the unshielded and shielded pancake coils--for five U-shaped EDM notch geometries summarized in table H-1-are shown in figures H-2 through H-11. A summary of the test results, e.g., whether the circumferential notch which links the two adjacent axial notches is discernible, as inferred from a review of the graphic area, is given in table H 2. The unshielded pancake coil data is not included since the results are not significantly improved over the coil data.
H-1
l' m
1 Figure H 1. Typical Test Sample Geometry 7 (Expanded Tube With U Shaped Notch at Roll)
Table H 1 TEST SAMPLE MATRIX Length, (Depth )
Iest Sample # A Il C _
Z-9280 Z-9300 Z-9299 Z-9289 Z-9281 H-2
U-Snaped Notch Closed-End View
^
U-Shaped Notch Open-End View Figure H-2. Pancake Coil- Diameter, Sample # Z-9300 H3
U-Shaped Notch Closed End View 9
5 e
U-Shaped Notch Open End View L
Figure H-3. Pancake Coil- Diameter (Shielded), Sample # Z-9300
/ H-4
U-Shaped Notch Closed-End View U Shaped Notch Open-End View Figure H-4. Pancake Coil - Diameter, Sample # Z-9299 H-5
U-Shaped Notch Closed-End View U Shaped Notch Open-End View Figure 11-5. Pancake Cc4 - Diameter (Shielded), Sample # Z-9299 H-6
U Shaped Notch Closed-End View U-Shaped Notch Open-End View Figure H 6. Pancake Coil - Diameter, Sample # Z-9289 H-7
U-Shaped Notch Closed-End View U-Shaped Notch Open-End View Figure H 7. Pancake Coil - Diameter (Shielded), Sample # Z-9289 H-8
i U Shaped Notch Closed End View U Shaped Notch Open End View Figure 11-8. Pancake Coil - Diameter, Sample # Z-9281 li-9
U-Shaped Notch Closed-End View U-Shaped Notch Open-End View Figure H-9. Pancake Coil - Diameter (Shielded), Sample # Z-9281 H-10
U-Shaped Notch Closed-End View U-Shaped Notch Open-End View Figure H-10. Pancake Coil - Diameter, Sample # Z-9280 H-11
U-Shaped Notch Closed-End View U-Shaped Notch Open-End View Figure H-11. Pancake Coil - Diameter (Shielded), Sample # Z-9280 H-12
i 11.3 RESULTS The conventional diameter pancake coil clearly resolves the circumferential notch at spacings, with a degraded capability be;; inning at or slightly greater. The (shielded) pancake coil offers significantly improved resolution, with the circumferential notch visible down to a axial notch separation. Even the circumferential notch linking the two axial notches is resolvable.
Table H-2
SUMMARY
OF PANCAKE COIL DATA Note: 1 - Clearly discernible.
2 - Marginally discernible.
3 Not reliably discernible.
H-13
Appendix I l'RESSURE TESTS OF STEAM GENERATOR TUBES WITH COMBINATION AXIAL AND CIRCUMFERENTIAL CRACKS
1.1 INTRODUCTION
A number of pressure tests have been conducted with steam generator tubes containing combinations of axial and circumferential degradations. These tests included machined through walllinear degradations inclined at an angle with respect to the tube axis, machined "L" shaped cracks with an axial and circumferential through wall component, machined "U" shaped cracks with axial through wall components, and part-through-wall and through-wall circumferential components. A limited number of tests with tubes containing laboratory-produced stress corrosion combination cracks have also been conducted.
Pressure tests involve pressurizing a section of steam generator tube containing either machined degradations or stress corrosion cracks. For through wall degradations, a crack sealing system was used to maintain tube pressure. Tubes were pressurized until the maximum pressure that the tube can support was achieved.
Generally, maximum pressure occurs when part-through-wall ligaments tore or the crack scaling system was lost. This would result in a tube leak, and with a limited makeup capacity pressurization system, higher pressures could not be applied.
1.2 LABORELEC DATA LABORELEC pressure-tested the degradation geometries shown in figure 11.
Inclined axial crack specimens had through-wall electrodischarge-machined (EDM) slits with overall lengths from to inclined with the tube axis from . "L" shaped crack specimens had EDM through-wall slits with axial components and circumferential legs with gaps between the axial and circumferential components for to The LABORELEC test results are provided in table I-1 and are shown plotted in figures 1-2 and 1-3. These figures show the dimensionless length (O of the axial component of the degradation plotted with dimensionless maximum pressure F. These dimensionless parameters are defined in Eqs. 2-8 and 2-9.*
%e section 2.
1-1
Also shown in figures I-2 and I-3 is the "EPRI" burst curve for a simple through-wall axial crack. This curve is defined by Eq.2-11.* For reference, the dimensionless pressure corresponding to times normal differential pressure .
and MSL13 differential pressure , along with th'e dimensionless axial crack length corresponding to the repair limit and the repair limit defined in section 4.0, are shown in each of the these figures as well.
Referring to figure I 2, note that the inclined axial cracks all had maximum test pressures above the burst curve. This was true even for a degradation inclined to the tube axial by with a projected axiallength and a projected circumferential length. Figure I 3 shows that the "L" shaped degradation tests also produced maximum pressures that fall on or above the burst curve. These tests show that for these geometries one could neglect the circumferential component of the "L" shaped degradations and estimate the tube maximum pressure using the burst curve for simple axial cracks.
1.3 TECIINATOM DATA TECNATOM tested the EDM degradation geometries shown schematically in figure I-4 (D. The test degradation configurations generally included two axial through-wall degradations separated by either or (approximatelty
). The circumferential degradations were part through-wall and intersected the axial defects. In a few of the TECNATOM tests, the circumferential degradation was at or below the top of a simulated tubesheet. TECNATOM tests results are summarized in table 1-2 and are shown plotted in figure 1-5. Examination of figure I-5 shows that unless the depth of the circumferential crack was deeper than of the wall thickness, the maximum pressure was above the burst curve.
I.4 WESTINGIIOUSE DATA Westinghouse pressure test results for "L" and "U" shaped cracks are summarized in table I-3 and are shown plotted in figures I-6 and I-7.
Through-wall "L" shaped cracks had an axial length of with circumferential lengths of .
Figure I-6 shows that with circumferential length of the "L" shaped crack equal to
, the maximum test pressure falls above the burst curve, while with the circumferential lengths, they fall below the curve.
Westinghouse conducted tests with machined "U" shaped degradations where the circumferential component was not completely through-wall and where the circumferential component was through-wall with shims inserted in the circum-ferential slots to create slot face interference. The results of these tests are shown in figure I-7. Maximum pressures for machined "U" shaped specimens through-wall and through-wall are below the burst curve. Specimens with machined I-2
l l
"U" shaped degradations through wall achieved maximum pressures equal to or above the burst curve. Note that in spite of being below the burst curve, "U" shaped specimens with axial through wall components longer than the repair limit defined in section 4 and through-wall circumferential components supported pressures greater than MSLB differential pressures.
Westinghouse tested steam generator tubes with combination stress corrosion crack geometries produced in model boiler tests Q). Sketches of the resulting stress cor-rosion crack geometries are shown in figure 1-8. Crack geometries which can be characterized as "L" and "U" shaped are included. The depth of the circumferential portion of the stress corrosion cracks was reported as through-wall. A collar was used to simulate the tubesheet. The specimens were fabricated with the roll transition flush, to slightly above (about ), the simulated tube-sheet. The test results are summarized in table I-3, and maximum test pressures are plotted in figure I 9. It is important to note that in all cases, maximum test pressure was above the burst curve and significantly above the maximum pressure supported by EDM degradations with similar dimensions.
1.5 REFERENCES
- 1. Jiyaliation of the impact of the Presence of Circumferential Orientation Eissures in the Transition Zone of the Expanded Portion of Steam Generator Tubes. TECNATOM Report, December 21,1990.
- 2. J. A. Begley. Fagluation of the Leak and Burst Characteristics of Roll Transitions Containine Primary Water Stress Corrosion Cracks. Palo Alto, Calif.: Electric Power Research Institute (publication pending). NP 7414.
13
-- ---'- '-- ._-___,-___._.m.__ ___
Table 1-1 LABORELEC COMBINED AXIAL AND CIRCUMFERENTIAL CRACK PRESSURE TEST DATA t=
R=
No. 1 (mm) 0 (*) c (mm) I"/P i' A "L" Shaped Cracks (see figure I-1)
S (mm)
I-4
Table 1-2 TECNATOM COMBINED AXIAL AND CIRCUMFERENTIAL CRACK PRESSURE TEST DATA t =
R=
Sy + Su =
No. 1 (mm) e d/t P* (Bar) f) A (1)No. circ. crack.
(2)Cire. crack at the top or within tubesheet.
c)No axial crack.
(4)Only one axial crack, 1-5
Table I-3 WESTINGHOUSE COMBINED AXTAL AND CIRCUMFERENTIAL CRACK PRESSURE TEST DATA t =
R =
Sy + Su =
No. 1 (in.) c (in.) P' (psi) I' ).
'These maximum pressure numbers are preliminary and approximate.
" Wedge inserted in circumferential slot to simulate interference. -
1-6 m
Figure I-1. LABORELEC Combination Notch Geometries I-7
Figure 1-2. LABORELEC Through-Wall Angle Degradation Pressure Tests 1-8
__ _ _-_-__________________.____._____.__-_____m_ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ . _ _ _ . . _ _ _ _ . - _ _ _ _ _ _ _ _ _
a t
i.
Figure I-3. LABORELEC Through-Wall "L" Shaped Degradation Pressure Tests I-9
iI "1
s C
l -.
E f
4 Figure I-4. .Spanish Combination Degradation
- Pressure Test - Specimen Configuration-1-10 o a 4
, ,1.g y
w y
l -L i:-
' Figure I-5. TECHNATOM Part-Through-Wall Degradation Pressure Tests I-11;
, p , , y , p-, .g. PM '?Y
. - .. .. ._- _ ~. . . . .__ . . . . ,,
1 i.
i.
Y b
i e
- Figure I-6; Westinghouse Through-Wall "L" Shaped Degradation Pressure Tests ; -
I-12
- I
^ - - - - - . . . . _ _ .
l t
I -
- Figure 1-7. Westinghouse Modified Through-Wall and Part-Through-Wall-
"I!' Shaped Degradation Pressure Tests 1-13
ru, t;;; -
,~f F
=t
^
i
+
5.
Figure I-8; Sketches of Axial and Circumferential Cracks in Westinghouse
- RP301-9. Specimens
-I.14; y q .. v--- n
_. ... . , . ~ _ . _- _.. . . _ . - . . . . _ _ _ _ , . ____ ._ . _._ _ . . _ _ _ _ _ . .__ .._ . ,
- -'t
-f F
4 Y
5 1
( -
I l
Figure I-9. Westinghouse Stress Corrosion Crack Pressure Tests.
1-15
~. . _ , . -
x- _. . , ,, .