PVRC'S POSITION ON ENVIRONMENTALEFFECTS ON FATIGUE LIFE IN LWR

APPLICATIONS

W. Alan Van Der Sluys

STEERING COMMITTEE ON CYCLIC LIFE AND ENVIRONMENTAL EFFECTSOCKETEDSumio Yukawa, Chair USNRC

Robert S. Vecchio, Vice Chair Aunust 12 2nOn Ill .nn!m\

Asada, YasuhideBush, Spencer H.

Chakrabarti, Gauranga S.Chopra, Omesh K.Donavin, Paul R.

Hechmer, John L.Higuchi, Makoto

Hoffmann, Christopher L.lida, KunihiroMehta, Har S.

O'Donnell, William J.VanDerSluys, W. Alan

Vecchio, Robert S.Yukawa, Sumio

OFFICE OF SECRETARYRULEMAKINGS AND

ADJUDICATIONS STAFF

WRC Bulletin 487-DECEMBER 2003

Publication of this report is sponsored byThe Pressure Vessel Research Council of the Welding Research Council, Inc.

and The Board on Nuclear Codes & Standards of The Council on Codes & Standards of theAmerican Society of Mechanical Engineers

WELDING RESEARCH COUNCIL, INC.PO Box 1942

New York, NY 10156www.forengineers.org

U.S. NUCLEAR REGULATORY 1buivnii,5•lU1

In the Matter of £t WJW • AI•u U" " -

Docket No. 5o - 711 Official Exhibit No. -Z-3'-' Y

OFFERED by'i ,ntervenor_....

NRC Staft Other

IDENTIFIED On+I'? 12-' 0% Witness/Panel LJC ? ,

Action Taken: MITTEI REJECTED WITHDRAWN

Reporter/Clerk A1f c--I,

ISBN No. 1-58145-494-5

Library of CongressCatalog Card Number: 85-647116

Copyright © 2003 byWelding Research Council, Inc.

All Rights ReservedPrinted in U.S.A.

ii

FOREWORD

This report describes the activities of the PVRC Steering Committeeon Cyclic Life and Environmental Effects (CLEE) and the PVRCWorking Group S-N Data Analysis. This report presents the PVRCrecommendations to the ASME Board on Nuclear Codes and Stan-dards (BNCS) concerning needed modifications to the ASME fatigueanalysis procedure. The proposed modifications will account for theeffect of the environment on the fatigue properties of the pressureboundary materials. These recommendations are in response to thefollowing request from the BNCS:

"BNCS Looks to PVRC to Obtain, Characterize, and Report inSufficient Detail to ASME Such Data as May be Useful to ASME inits Evaluation of the Fatigue Curves of Sections IIII and XI"

The PVRC Committee has worked closely with, and received com-ments from, investigators in Japan, Europe, and America and has re-viewed essentially all public domain data. We are particularly apprecia-tive of databases and analyses provided by those in Japan working onMITI projects and in America at the Argonne National Laboratory.

We believe we have been successful in guiding the experimentalwork and forging a consensus with regard to the key issues that wereformerly much less than clear. Considering all well characterized,available data, PVRC has drawn the following major conclusions:

1. ASME Section III should adopt a procedure such as proposed inSection 7 of this report to apply an environmental correction factor,Fen; to life fractions calculated using the existing ASME S-N designcurves when anticipated operating conditions are sufficiently severethat it is necessary to account for environmental effects.

2. ASME Section XI should adopt a procedure such as proposed in adraft code case in Section 7 of this report and apply the environ-mental correction factor, Fenw to life fractions calculated using theexisting ASME S-N design curves when it is necessary to accountfor environmental effects.

3. The Fen models are shown to work well in predicting the effect ofthe coolant environments on the low cycle fatigue properties ofstainless steel. The low cycle fatigue information on stainlesssteel in air, collected by the PVRC to perform the evaluation, doesnot appear to support the ASME mean data line for stainlesssteel, and more data are needed to adequately understand behavior.

The above conclusions are based on two principles:

1. The environmental correction factors can be determined using equa-tions developed either by Argonne National Laboratory or by MITI'sinvestigators in Japan. While these equations are somewhat differ-ent; in real situations, they are expected to give similar results,within the bounds of experimental error and operating uncertainties.

111..

2. The factor of 20 on life, originally used in the development of thefatigue design curves to account for uncertainties, is adequate toaccount for reductions in fatigue life due to the environmentunder well controlled operating conditions. Under those condi-tions, provision for further reductions in fatigue life due to theenvironment is not essential.

The PVRC has reviewed the ASME Section III Fatigue Analysisprocedure to determine what modifications are needed to take intoaccount the effects of the coolant environment on the S-N fatigue proper-ties. In performing this review, the PVRC evaluated the following areas:

1. The margins used in the development of the Section III procedure.2. Laboratory data used in the development of the Section III

procedure.3. Laboratory fatigue data on smooth specimens in simulated reac-

tor coolant environments.4. Models to predict the S-N properties in Light Water Reactor

(LWR) coolant environments of the pressure boundary materials.5. Laboratory data on structural tests conducted in water environ-

ments.

This report is divided into 10 sections that describe in detail thedevelopment of the PVRC recommendations and present examples of theCode changes needed to implement the recommendations. The S-Nfatigue data for carbon steel, low alloy steel, and stainless steels, collectedby the PVRC are compared with the available S-N models. Both the modelsdeveloped by Argonne National Laboratory and MITI are shown to ade-quately predict the S-N results in simulated LWR coolant environments.i The available data from laboratory specimens tested in simulated LWRcoolant environments were used to evaluate expected reduction in life inplants. It was determined that the margins applied to laboratory data todevelop the ASME Fatigue Design Curves need not be adjusted whencertain operating thresholds are not exceeded. These thresholds identifiedby PVRC pertain to oxygen level, temperature, stain rate, etc. The PVRCdeveloped thresholds, or more rigorous analysis without thresholds, canbe used to determine the effect of the environment on specific components.

A limited amount of laboratory data exist on the effect of coolantflow rates on carbon and low alloy steels. These data show a reductionin the environmental effect with increasing flow rates. These flow rateeffects need to be incorporated into the Fen models and the thresholdsfor carbon and low alloy steels. At this time no information exists as tothe effect of flow rate on stainless steel.

Available data from the literature on the results of laboratory testsof structural components in water environments were evaluated usingthe proposed procedure. This evaluation supported the concept of amoderate reduction in fatigue life without applying the environmentalcorrection factor, Fen) to the ASME Fatigue Design Curves.

In section 9, of this report a copy of the MITI Guidelines ForEvaluating Fatigue Initiation Life Reduction in LWR Environments isreproduced. These guidelines recommend the use of the Fen factor toaccount for the effect of the environment but do not utilize the conceptof thresholds to deal with moderate environmental effects.

iv

CONTENTS

FOREW ORD ............................................................. iii1. Introduction ........................................................... 1

1.1 BNCS Response and Request to PVRC .............................. 11.2 Proposed Environmental Factor Approach to Account for

Environmental Effects in LWR Applications ........................ 21.3 Summaries of Programs External to PVRC .......................... 21.4 Summary of PVRC Activities ....................................... 3

2. Summary of Technical Basis of Section III Fatigue EvaluationProcedure .............................................................. 3

2.1 ASME Air Curve .................................................... 42.2 M argins ............................................................. 4

3. Early Tests and Results in Simulated Reactor Coolant Environments ... 64. Thresholds ............................................................. 7

4.1 Carbon and Low Alloy Steel .............. .................. 84.1.1 Verification of the Thresholds .............. .................... 9

4.1.1.1 Comparison with Fen Models ................................ 94.1.1.1.1 Strain Rate ............................................... 94.1.1.1.2 Oxygen Content ......................................... 94.1.1.1.3 Temperature ....... ...................................... 94.1.1.1.4 Strain Amplitude ................... , .................... 9

4.1.2 Other Factors .................................................. 124.1.2.1 Sulfur Effect ............................................... 124.1.2.2 Flow Rate Effects ........................................... 12

4.1.3 Comparison of the Threshold Values with Fatigue Crack GrowthInform ation .................................................... 14

4.2. Austenitic Stainless Steels ........................................ 144.2.1 Fen M odels ..................................................... 144.2.2 Strain Rate ..................................................... 154.2.3 Temperature ................................................... 154.2.4 Dissolved Oxygen .............................................. 154.2.5 Strain Amplitude ............................................. 164.2.6 Discussion of Austenitic Thresholds ........................... 17

5. M argins ............................................................... 175.1 Initial Consideration of ASME Code Margins ...................... 175.2 PVRC Review of the Margins ...................................... 175.3 Sub-Factor for Size Effects ......................................... 185.4 Sub-Factor for Surface Finish ..................................... 185.5 Sub-Factor for Data Scatter ........................................ 205.6 Atmosphere Factor ................................................ 205.7 Load History Effects ........................... 205.8 Application of Fen..................................................... 215.9 Conservatism of Fen ............................................. 215.10 Evaluation of the Z Factor ........................................ 22

6. Review of S-N Models....................................... 226.1 ASME Code S-N Models. ............................... . ...... 226.2 Argonne National Laboratory Air Fatigue Data Models .......... 246.3 Japanese Air Fatigue Data Models ................................ 246.4 Comparison of Model Predictions for Air Fatigue Data .......... 246.5 Reactor Water Environmental Effects ....... ................ 276.6 Environmental Model Comparisons .......................... 29

V

7.0 Code Implementation ................................................ 337A Appendix: Recommendation for Code Implementation ............ 34

7J.A1 Potential Section III Implementation .......................... 347.A.2 Potential Section XI Implementation.......................... 357.A.3 Potential Implementation as Section XI Code Case ............ 357.A.4 Nonmandatory Appendix XX ....... ! ........................... 35

8.0 PVRC Data Base and Analysis ....................................... 388.1 Carbon Steel ................................ ...................... 388.2 Low-Alloy Steel .................................................... 428.3 Austenitic Stainless Steel .......................................... 458.4 Structural Test Results ............................................ 50

8.4.1 PVRC Pressure Vessel Tests ...... ...................... 518.4.2 Tests on Butt-Welded Piping ................................... 518.4.3 KWU Tube Tests ................................................ 54

9.0 M ITI Procedure ..................................................... 5710.0 Conclusions and Recommendations ................................ 58

Acknowledgment

This report represents the culmination of ten years of activities within the PVRC. The author would liketo acknowledge the help and contributions of a number of persons:

Throughout most of the ten years the work was preformed under the guidance of Sumio Yukawa. Withouthis guidance and contributions this report could not have been written.

Throughout this effort there has been continuous support from the following committees in Japan:

" Thermal and Nuclear Power Engineering Society (TENPES), EFD Project" Japan Power Engineering and Inspection Corporation (JAPEIC), EFT project" Japan Nuclear Energy Safety Organization (JNES)

These Japanese committees contributed much of the fatigue data collected by the PVRC as well asparticipated in the activities of the PVRC subcommittees. M Higuchi, Ishikawajima-Harima HeavyIndustries Co. Ltd, attended many -of the committee meetings, contributing in the discussions andkeeping the PVRC informed as to activities in Japan.

In addition representatives from Hitachi, Mitsubishi Heavy Industries, Toshiba Corp, Tokyo ElectricPower, Kansai Electric frequently attended meetings and contributed valuable test results.

Another major contributor to the committee activities was 0. Chopra from Argonne National Laboratory.He attended most of the PVRC meetings, supplied significant data to the PVRC, and contributed in thediscussions of the subcommittees.

EPRI contributed several figures from EPRI Report MRP-49.

vi

PVRC's Position on Environmental Effectson Fatigue Life in LWR Applications

W. Alan Van Der SluysI

1.0 Introduction

The rules and requirements provided in SectionIII of the ASME BOILER and PRESSURE VESSELCODE has been widely used in the US and in othercountries for the design, fabrication, and pressureintegrity evaluation of the components for light wa-ter-cooled reactor (LWR) type of commercial nuclearpower systems. Among its many features, Section IIIincludes procedures for analyzing fatigue damageand the possibility of crack formation by fatigue as aresult of pressure and temperature cycling duringoperation.

Beginning in the 1950's, design, fabrication, andconstruction activities related to nuclear power expe-rienced a major increase and the ASME Code in-creased its scope and activities to keep pace with theincrease. Emphasis .on the "Design by Analysis" in-cluded additional effort on fatigue analysis with theformation of a Task Group for the determination ofallowable fatigue stresses chaired by B.F Langer.The Task Group collected and analyzed the availablefatigue test data and developed curves of allowablefatigue stresses as a function of number of imposedcycles.

The methodology utilized by the Langer TaskGroup to formulate Fatigue Design Curves (desig-nated as Figs 1-9.0 in the Code) is described inSection 2 of this report. The work of the Langer TaskGroup was limited by the fact that the technology offatigue testing in elevated temperature water atpressures and chemistries typical of LWR operatingconditions was not well developed, which limited theavailable amount of fatigue test data for LWR cool-ant water environments. This limitation was recog-nized by the ASME in the 1974 and 1992 editions ofthe Code, wherein, Articles NB3120 and NB3121

entitled "SPECIAL CONSIDERATIONS," and "Cor-rosion" stated:

"It should be noted that the tests on which the fatiguedesign curves (Figs 1-9.0) are based did not include testsin the presence of corrosive environments which mightaccelerate fatigue failure."

Within a few years, results for fatigue tests con-ducted in water environments which simulated LWRcoolant water became available in technical. Ex-amples include:

D. Hale, S.A. Wilson, J.W. Kass, and E. Kiss "LowCycle Fatigue of Commercial Piping Steels in aBWR Primary Water Environment," Journal ofEngineering Materials and Technology, Vol. 103,pp. 16-25 (1981)

M. Higuchi and K. Iida, "Fatigue Strength Correc-tion Factors for Carbon and Low-Alloy Steels inOxygen-Containing High-Temperature Water,"Nuclear Engineering and Design, Vol. 129, pp.293-306(1991)

O.K. Chopra, and W.J. Shack, "Environmental Ef-fects on Fatigue Crack Initiation in Piping andPressure Vessel Steels," NUREG 6717, ANL-00277 May 2001, U.S. Nuclear Regulatory Commission

These results all indicated that LWR coolant watercould have a significant detrimental effect on thefatigue life of metals utilized for the pressure bound-ary of LWR Nuclear systems.

1.1 BNCS Response and Request to PVRCThese results produced serious concerns within

the ASME Board on Nuclear Codes and Standards(BNCS) regarding the structural integrity of Nuclearpower plants and BNCS made the following requestto PVRC to assist in resolving the concerns:

"BNCS Looks to PVRC to Obtain, Characterize, and1Consultant, Alliance, OH Report in Sufficient Detail to ASME Such Data as May

WRC Bulletin 487 1

be Useful to ASME in its Evaluation of the FatigueCurves of Sections III and XI"

1.2 Proposed Environmental Factor Approach toAccount for Environmental Effects in LWRApplications

In 1994-95, GE with EPRI support developed theenvironmental factor procedure for ASME Code-typeAnalysis of Environmental Effects in Fatigue UsageEvaluation. In Oct 1999, PVRC forwarded this proce-dure to the BNCS with a recommendation that theprocedures be considered for Code application andimplementation. The Procedure has been utilized forfatigue life evaluation in several License Submittalsto the NRC. This report presents available data,models to predict the environmental factors andsuggested Code Cases for Code implementation.

Starting in 1992, the Pressure Vessel ResearchCouncil (PVRC) has had a continuing activity con-cerned with the effect of the Light Water Reactor(LWR) coolant environment on the fatigue perfor-mance of the pressure boundary materials used inLWR applications. The activity has involved threemain aspects of fatigue performance and applica-tions. These are: (a) cyclic life under repeated stressand strain, the so-called S-N properties, (b) fatiguecrack growth under repeated loading, and (c) evalua-tion of the design procedures and methodology usedto assure performance and life of the structural com-ponents under anticipated cyclic duty. The primaryfocus of this paper concerns the effect of the LWRenvironment on the first aspect of fatigue perfor-mance, namely the S-N properties. The PVRC effortin this area has consisted of compiling and evaluat-ing the available test data and assessing the variouscorrelations of cyclic life and various mechanical andenvironmental parameters. The interim status andfindings of this effort have been reported by Van DerSluys and Yukawa [1-1, 1-2 and 1-3]. It may be notedthat Hechmer [1-4] has presented a summary of thePVRC effort related to the design and evaluationaspects of fatigue performance.

1.3 Summaries of Programs External to PVRCDuring the 1993-1995 period, the US Department

of Energy (DOE) and the US Nuclear RegulatoryCommission (NRC) initiated and supported severalprograms that evaluated and assessed the ASMECode design criterion for fatigue life performance ofoperating LWR nuclear power plants. In addition,the ASME Code adopted an enabling rule for reanaly-sis of usage factor calculations, and the ElectricPower Research Institute (EPRI) supported develop-ment of an approach and procedures that could beimplemented into the ASME Code to perform envi-ronmental effects analysis. The findings and/or ensu-ing actions from these activities included the follow-ing:

0 A DOE supported study [1-5], examined theeffect of applying an early version of a fatiguedesign curve that included an adjustment for

LWR environmental effects on the calculatedfatigue usage factor of representative ASMEClass 1 components. As expected, the lower cy-clic life of the adjusted curve increases the calcu-lated usage factor. However, it was observedthat in a number of instances, conservative andbounding values were utilized in the originalusage factor calculations. Using more realisticvalues in the usage analysis could compensatefor a significant portion of the environmentaleffect on usage factor.

" The NRC program, titled Fatigue Action Plan,included studies of a number of issues associ-ated with the assessment of fatigue perfor-mance of structural components in a LWR envi-ronment. For example, it included a muchbroader and detailed study of situations notedin the DOE study mentioned above; these re-sults are described by Ware et al. [1-6, 1-7]..Based in part on the results of the study, theNRC concluded that no major actions wereneeded by the NRC regarding environmentaleffects for currently operating LWR plants [1-8].

* In the 1996 Addenda to the ASME Code, SectionXI added a new nonmandatory Appendix L titledOperating Plant Fatigue Assessment. In es-sence, the Appendix permits a re-evaluation ofthe original usage factor analysis to determineacceptability for continued service. Addition-ally, the Appendix also contains flaw tolerancebased procedures and acceptance criteria to de-termine acceptability for continued service. AnEPRI supported activity to develop proceduresthat could be used in conjunction with generallyavailable data and information in existing ASMECode stress and fatigue analyses to account forLWR water environmental effects was com-pleted in 1995 by Mehta and Gosselin [1-9, 1-10].The approach and procedures have been re-viewed and evaluated by the PVRC and deter-mined to be a reasonable and workable ap-proach for Code implementation. The detailedevaluation combined with some trial uses of theprocedures has revealed areas where revisionsand modifications are needed, and PVRC effortis being applied to this need. This developmentmade full use of the results of the statisticalmodeling and analysis effort performed by theArgonne National Laboratory [1-11] and by Japa-nese investigators [1-12].

0 In 2000 MITI Guidelines for Evaluating FatigueInitiation Life Reduction in LWR Environmentswere issued by the Nuclear Power Safety Admin-istration, Public Utilities Department, Agencyof Natural Resources, and Energy Ministry ofInternational Trade and Industry. These guide-lines are presented in Section 9 of this report.These guidelines use Fen as a fatigue life correc-tion factor in the same way as recommended bythe PVRC. The equations for the calculation of

2 WRC Bulletin 487

Fe' give very similar results as Fen calculationsdeveloped by Argonne National Laboratory andused in the PVRC approach. Both sets of equa-tions are presented in this report. The MITIapproach differs from the PVRC approach inthat it does not accept a moderate environmen-tal effect which is discussed in sections 4 and 5of this report.

0 In 2001 EPRI published MRP-49 Materials Reli-ability Program (MRP) Evaluation of FatigueDate Including Reactor Water EnvironmentalEffects. This report recommends the use of thePVRC procedure in plant life extension evalua-tions. Many of the figures and some of the text inthis PVRC report are the same as in this EPRIreport.

1.4 Summary of PVRC ActivitiesThis report recommends an approach which en-

tails the use of a life reduction factor, Fen for thecases when the characteristics of the transient beingevaluated exceed a set of threshold conditions for theexistence of an environmental effect on the fatiguelife of the material. It has been shown that thisapproach will account for the environmental effectsobserved in laboratory studies. This approach is ap-plicable because there is no observed effect of theenvironment on the fatigue limit of the material;thus, a factor of fatigue life alone will account for allobserved effects. The laboratory studies have notshown an effect of the environment on the fatiguelimit and the experience in Germany with oxygenwater treatment in fossil boilers, that will be dis-cussed later in this paper, has not observed such aneffect.

The following sections of this report will presentthe technical bases for the life reduction factor ap-proach. These sections will show the development ofthe threshold values for carbon, low alloy and stain-less steel in the environment and present the modelsused to calculate the life reduction factors Fen.

This report will also describe the application ofthis procedure to the results from a number of labora-tory tests programs in which structures were testedunder fatigue loading conditions in water environ-ments until failure. In these cases the recommendedprocedure is shown to work very well to predict thelife of the structures.

References1-1. Van Der Sluys, W.A., "Evaluation of the Available Data on the Effect

of the Environment on the Low Cycle Fatigue Properties in Light WaterReactor Environments," presented at Sixth Int. Sympos. on EnvironmentalDegradation in Nuclear Power Systems-Water Reactors, TMS/NACE,Aug. 1-5, 1993, San Diego.

1-2. Van Der Sluys, W.A. and Yukawa, S., "Studies of PVRC Evaluationof LWR Coolant Environmental Effects on the S-N Fatigue Properties ofPressure Boundary Materials," in: PVP-Vol. 306, pp. 47-58, presented atASME PVP 1995, Honolulu, HI, July 23-27, 1995.

1-3. Van Der Sluys, W.A. and Yukawa, S., "S-N Fatigue Properties ofPressure Boundary Materials in LWR Coolant Environments," in: PVP. Vol.374, pp. 269-276, presented at ASME PVP 1998, San Diego, July 26-30,1998.

1-4. Hechmer, J., "Evaluation Methods For Fatigue-A PVRC Project,"in: PVP-VoI. 374, pp. 191-196, presented at ASME PVP 1998, San Diego,July 26-30, 1998.

1-5. Smith, J.K., Deardorff, A.F., and Nakos, J.T., "An Assessment of the

Conservatisms In Fatigue Evaluation of ASME Class 1 Pressure Vesselsand Piping," PVP-Vol. 286, ASME, 1994, pp. 19-29.

1-6. Ware, A.G., Morton, D.K., and Nitzel, M.E., "Application of Environ-mentally-Corrected Fatigue Curves To Nuclear Power Plant Components,"PVP-Vol. 323, ASME, 1996, pp. 141-150.

1-7. Ware, A.G., Morton, D.K., and Nitzel, M.E., "Application of NUREG/CR-5999 Interim Fatigue Curves to Selected Nuclear Power Plant Compo-nents," NUREG/CR-6260, U.S. Nuclear Regulatory Commission, March1995.

1-8. SECY-95-245, "Completion of the Fatigue Action Plan," James M.Taylor, Executive Director for Operations, U.S. Nuclear Regulatory Commis-sion, Washington, DC, September 25, 1995.

1-9. Mehta, H.S. and Gosselin, S.R., "An Environmental Factor Ap-proach to Account for Reactor Water Effects In Light Water ReactorPressure Vessels and Piping Fatigue Evaluations," PVP-Vol. 323, ASME,1996, pp. 171-185.

1-10. Mehta, H.S. and Gosselin, S.R., "An Environmental Factor Ap-roach to Account for Reactor Water Effects in Light Water Reactor

essure Vessels and Piping Fatigue Evaluations," EPRI Report No. 105759,Dec. 1995.

1-11. Meisler, J. and Chopra, 0., "Statistical Analysis of Fatigue Strain-Life Data for Carbon and Low-Alloy Steels," PVP-Vol. 296, Risk and SafetyAssessments: Where is the Balance? Book No. H00959-1995.

1-12. Nakao, G., Higuchi, M., lida, K., and Asada, Y., "Effects of Tempera-ture and Dissolved Oxygen Contents on Fatigue Lives of Carbon and LowAlloy Steels in LWR Water Environments," Effects of the Environment onthe Initiation of Crack Growth, ASTM STP 1298, ASTM, 1997, pp. 232-245.

2.0 Summary of Technical Basis of Section IIIFatigue Evaluation Procedure

The fatigue evaluation procedure in Section III ofthe ASME Boiler and Pressure Vessel Code wasdeveloped in the early 1960's. It was based on theBureau of Ships Design Bases developed in the late1950's. The S-N fatigue curves and a description ofthe technical basis for the curves for the BuShipsDesign Basis Ref 2-1. The following is taken fromthis reference and is the description of the procedureused to develop the S-N curves.

"This curve was constructed in the following man-ner.

(a) Available strain fatigue data for this generalclass of material were plotted in the form oftotal strain (elastic plus plastic) range versuscycles-to-failure. Machined specimens withoutnotches that were tested at temperatures lessthan 600'F were considered. The mean curvefor each material was drawn.

(b) A lower limit of the mean curves was drawnand then converted to a stress amplitude ver-sus cycles-to-failure curve by multiplying thestrain range by E/2, where E was taken as26 X 106 psi.

(c) The design fatigue curve was then constructedby applying a factor of safety of either 2.0 onstress amplitude of a factor of 20 on cycles,whichever was more conservative at each point.The factor of 20 on life is the product of thefollowing sub-factors:a. Scatter of data (minimum to mean) 2.0b. Size Effect 2.5c. Surface finish, atmosphere, etc. 4.0

(d) The design fatigue curve stress amplitude forless than 100 cycles was taken as the value at100 cycles."

This procedure is essentially the same as used in thedevelopment of the ASME curves as described in Ref.2-2. The data and equations used in this develop-ment are described in the next section.

Environmental Fatigue in LWR Applications 3

2.1 ASME Air CurveIn Reference [2-2] the ASME Sub-Task Group on

Fatigue recommended to the ASME Boiler and Pres-sure Vessel Committee, Special Committee to Re-view Code Stress Basis that the formula given belowbe use in low cycle fatigue. Langer describes theapplication of this equation to 18-8 stainless steels inref [2-3].

E 100Sn 100-RA Se

Where

S = elastic modulus x stain amplitude (psi)E = elastic modulus (psi)N = cycles-to-failure

RA = reduction of area in tensile test (percent)S. = endurance limit or fatigue strength at 107

cycles (psi)

The above formula was used to determine the lowcycle fatigue curves for carbon steel, low alloy steel,and austenitic stainless steel. A best-fit curve ob-tained from the method of least squares, applied tothe logarithms of the measured S and N values,using the above equation as a model. The roomtemperature modulus, E, was known in each case,and the computer code gave the best-fit value for RAand S.. These values are shown on the curves tepro-duced from this report as Figures 2-1, 2-2, and 2-3.

These curves were then corrected for the maxi-mum effect of mean stress using the formula below.This was derived from the Goodman diagram consid-ering the change in the mean stress that is producedby yielding.

S'=S S. ] for S< S

Where

S = value from curveS' = adjusted S valueSu = ultimate tensile strengthSy = yield strength

The results from this correction for the mean stressare shown in the figures as dotted lines. It was feltthat austenitic stainless steels due to their highendurance limit and low yield strength cannot sus-tain a mean stress at a cyclic strain level that wouldproduce failure.

The best-fit lines, developed by Langer, appear tofit the data well. In these cases all of the results arefrom strain controlled experiments and the resultsare all in what is considered the low cycle region.

2.2 MarginsThe last step in the development of the ASME S-N

Fatigue Curve is the introduction of the margins of20 on life and 2 on stress. These are the samemargins as described earlier in this section. In Refer-ence 2-4 W. Cooper describes this process as follows:

"The final step in the process was to shift the curvesin recognition of the fact that laboratory data were tobe applied to actual vessels. Reference [2-2] statesthat the 'design stress values were obtained from thebest-fit curves by applying a factor of two on stress ofa factor of twenty on cycles, whichever was moreconservative at each point.' Unfortunately, these havebeen understood to be factors of safety, and nothingcould be further from the truth. As stated in Reference[2-2] 'it is not to be expected that a vessel will actuallyoperate safely for twenty times its specified life.'

The factor of twenty applied to cycles was devel-oped to account for real effects. Reference [2-1] states

1.E+07 I

CARBON STEELEN-2,A-201

0.

ccJ

I*-II

01.E+06

1.E+05

I~RA = 68.5%o ISE=21,645psi

LagrModel

IMean Stress CorrectionI" 0

1.E+04

1.E+01 1.E+03 1.E+05 1.E+07N

Fig. 2-1-ASME Mean Air Fatigue Curve for Carbon Steel

4 WRC Bulletin 487

Low-Alloy SteelIEn-25,A-225 and A-302 I

1.E+07

CLJ

IIU,1*

1.E+06

1.E+05

M rRA=61.4%SE=38'500psi i

[Langer o. -ti

i I ean Stress Correctiorl

1.E+04

1.E+01 1.E+03 1.E+05 1.E+07N

Fig. 2-2-ASME Mean Air Fatigue Curve for Low-Alloy Steel

18-8 Stainless Steel Curve

1 .E+07

[R)S,4S

1.E+06

.L LaniCI

0

II .E+05

II.E+04 4-1.E+0-1 1.E+02 1.E+03 1.E+i04 1.E+05 1.1E+06

Cycles To FailureFig. 2-3-ASME Mean Air Fatigue Curve for Stainless Steel Comparison of Margins

'The factor of 20 on life is the product of the followingsub factors:

a. Scatter of data (minimum to mean) 2.0b. Size Effect 2.5c. Surface finish, atmosphere, etc. 4.0

Two terms in the last line require definition. 'Atmo-sphere' was intended to reflect the effects of theindustrial atmosphere in comparison with an air-conditioned lab, not the effects of a specific coolant.

"Etc," simply indicates that we thought this factorwas less than four, but rounded it to give the factorof 20.

A factor on the number of cycles has little effect at ahigh number of cycles, so a factor on stress wasrequired at the higher number of cycles. It was foundthat at 10,000 cycles, approximately the border be-tween low- and high-cycle fatigue, a factor of two onstress gave approximately the same result as a factorof twenty on cycles."

Environmental Fatigue in LWR Applications 5

* The subject of the appropriate margins to be appliedto the mean of the fatigue data obtained in thelaboratory on smooth cylindrical specimens tested insimulated reactor coolant environments is one of themost important issues to be resolved by the PVRC inorder to develop an analysis procedure which takesinto account the effect of the coolant environment onthe fatigue life of the material.

2.3 References2-1. "Tentative Structural Design Basis for Reactor Pressure Vessels

and Directly Associated Components (Pressurized, Water Cooled Sys-tems)," dated 1 December 1958, with Addendum dated 27 February 1959.

2-2. "Criteria of the ASME Boiler and Pressure Vessel Code for Designby Analysis in Sections III and VIII, Division 2," ASME International, NewYork, NY, 1969.

2-3. Langer, B.F., "Design of Pressure Vessels for Low-Cycle" ASMEJournal of Basic Engineering, pp. 389-402, 1962.

2-4. Cooper, W.B., 'The Initial Scope and Intent of the Section IIIFatigue Design Procedure," Presented at PVRC Workshop on Cyclic Lifeand Environmental Effects in Nuclear Applications, Jan. 1992.

3.0 Early Tests and Results in SimulatedReactor Coolant Environments

It has been known for some time that under some,test conditions the low cycle fatigue properties ofcarbon and low alloy steels in simulated reactorcoolant environments could be reduced. The GeneralElectric Company conducted two series of experi-ments that showed such effects [3-1, 3-21. The first ofthese was conducted at the Dresden Reactor andinvolved cantilever bending specimens exposed tothe reactor coolant.

In these experiments, a special facility was set upat the Dresden-1 Nuclear Power Station, Morris,Illinois. Primary water from the Dresden-1 test loopBWR system was piped to this special test loop andcirculated at 10 gpm through three test vessels. Atotal of 35,535 loading cycles were applied to thefatigue specimens. Four materials were evaluated,Types 304 and 304L stainless steel, Inconel 600 andA-516 carbon steel. A summary of the results is asfollows:

"The results of this work confirm the adequacy ofthe current ASME Section III fatigue design curves toaccount for the effect of a B WRprimary water environ-ment on the low cycle fatigue behavior of the fourmaterials tested. Specifically:

1. Fatigue performance of non-sensitized stainlesssteel, even with slight chemical or machinednotches, is consistent with the ASME Code MeanData Curve and far exceeds the Design Curve.Performance of 304L stainless steel is compa-rable.

2. There is a slight reduction in fatigue life associ-ated with zero-tension loading in the, Type-304stainless steel in the BWR water environmentand this can be accounted for by use of a meanstress correction.

3. Reduction in cyclic life can be expected for heavilysensitized welded stainless steel. This is due tothe presence of stress corrosion cracking Whensuch welds are subjeit to cycling with long timesand stresses exceeding the yield level.

4. Based on an admittedly few data points, the lowcycle fatigue performance of Inconel far exceedsthe ASME Section III Fatigue Design Curve.However, significant amounts of intergranularcracking were observed in normally welded ma-terial.

5. Carbon steel material, whether welded or non-welded, displayed a reduction in fatigue perfor-mance in the BWR environment. This reductionappears to be related to the surface pitting.However, all data fall above the ASME SectionIII design curve and this material is fully ad-equate for field performance."

These conclusions appear to be inconsistent withthe later results presented in this report. Theresults from this program are, however, consistentwith the results from latter programs. The loadingstrain rates of from 0.03 to 0.06 in/in/sec., used inthe Dresden experiments, are not low enough forthe fatigue lives to be less than the ASME designcurves.

The second series of experiments were conductedon both cylindrical specimens under axial loadingand butt welded pipe samples of carbon steel withinternal pressure and axial loading [3-2]. These ex-periments were conducted in simulated BWR cool-ant with various dissolved oxygen contents. A sub-stantial environment effect was observed in theseexperiments and a Ke environmental correction fac-tor was suggested.

Experiments were conducted on SA333-Gr6 car-bon steel pipe material in room temperature air,550F air and simulated BWR coolant with variousdissolved oxygen contents. In this study, a number ofdifferent specimen geometries was tested includingbutt-welded pipe specimens. The program resultedin a number of recommendations as to changes

* needed in the ASME. Code fatigue analysis proce-dure and a Ke environmental correction factor. Theresults from the butt-welded pipe specimen testsfrom this program will be discussed in more detail inSection 9 of this report.

In 1991 Higuchi and Tida [3-3] proposed a fatiguelife correction factor, Fen' for correcting the low cyclefatigue properties of carbon and low alloy steels forthe effect of the LWR coolant environment. Theseresults stimulated the current concern for the effectof the environment on the low cycle fatigue proper-ties of the pressure boundary materials. A number ofversions of the correction factor has evolved sincethis original proposal but the basic concept has notchanged. This concept is that the effect of the environ-ment on the low cycle fatigue properties of carbonand low alloy steel can be corrected for the effect ofthe environment by applying a correction to thefatigue life as determined from the ASME designcurve. A correction is not needed on the strain ampli-tude.

6 WRC Bulletin 487