:II'NUREGICR4731EGG-2469 Volume IJune 1987

Residual Life Assessment of Major LightWater Reactor Components-Overview

Volume i

EditorsV. N. ShahP. E. MacDonald A.

ELIEGn Idaho

Work performed underDOE Contract No. DE-AC07-761D01570

for theU.S. NuclearRegulatory Commission

. . *.1

j

. ; Idaho National AL.,Engineerin~g L~abor~atory

S - -

.5..",.

*.44/*4

'-5-;."

.4

j-.

.�,, *s,

A. ��*4

.,"4- ,.c.

L.*' � '4

..�

r-I

.-c. c'-.. '

Managed by the US D6�partrnent of Energy

-. - - .,,.,... � -. 4,-.'' 44�

.4. -2,-

--.4- 4*44 ...A..P....-.....-.4f.... * �.4A%.j 5� �..

..>. .� ..�

4

4 4/. ** �. � � 4*p��

.4 .5-. .5. ' 4.�. .-.4 .4. 4A4.4�>..*.*45��5.4.

� *4444-�� *5**44 � .. --.s.�

r

INUREG/CR-4731

EGG-2469Distribution Categories: RM, R5

I

RESIDUAL LIFE ASSESSMENT OFMAJOR LIGHT WATER REACTOR

COMPONENTS-OVERVIEWVOLUME 1

V. N. Shah and R E. MacDonald, Editors

I

I

Authors

R. L. CloudJ. F. CookM. A. DayeW. G. HopkinsL. J. HouseV. Malhotra

H. MantleW. R. MikesellG. R. OdetteR. 0. RitchieW. L. ServerV. N. Shah I

Published June 1987

EGtG Idaho, Inc.Idaho Falls, Idaho 83415

Prepared for theDivision of Engineering Technology

Office of Nuclear Regulatory ResearchU.S. Nuclear Regulatory Commission

Washington, D.C. 20555Under DOE Contract No. DE-AC07-761D01570

FIN No. A6389

I

I

I

CONTRIBUTORS

Several specialists, engineers, and scientists have contributed to the dilnames, affiliation, and contributions are:

Name Affiliation

V. N. Shah EG&G IdahoP. E. MacDonald

NV. L. Server Robert L. Cloud Associates

R. L. Cloud

NV. R. Mikesell

G. R. Odette

R. 0. Ritchie

S. A. Traiforos

WV. G. Hopkins

M. A. Daye

P. B. Lindsay

H. Mantle

V. MaIhotra

C. E. Jaske

L. J. House

J. F. Cook

Robert L. Cloud Associates

Robert L. Cloud Associates

University of California, Santa Barbara

University of California, Berkeley

Bechtel National

Bechtel National

Bechtel National

Bechtel National

Bechtel National

Bechtel National

Battelle Columbus

Battelle Columbus

EG&G Idaho

fferent sections of this report. Their

Contribution

Ranking of major PWR and BWRcomponents

PWR RPV, BWR RPV, PWRprimary coolant piping

PWR primary coolant piping

BWR RPV

RPV radiation damage

RPV fatigue

Task Leader (Bechtel)

RPV supports

PWR containment structures

Recirculation piping, steamgenerator

Recirculation piping

Steam generator

Degradation mechanisms, lifeassessment methods, emergingNDE methods

Standard and emerging NDEmethods

ASME Section Xl NDE activities

L

L

[-

ii

I

environment and cyclic loading. Likely degrada-tion mechanisms include hydrogen embrittlementof the anchorages; corrosion of the liner, tendons,;reinforcing steel, and metal shells, including possi-ble microbiologically induced corrosion; andchemical reactions in the concrete. The potentialfailure modes are loss of prestress in the tendons,leakage of radioactive material caused by liner-concrete interactions, and loss of structural integ-rity, mainly because of corrosion: The ISI includesa tendon surveillance program, integrated leak ratetests, and visual inspection of surfaces. Majorunresolved technical issues are a lack of agingrelated data for reinforced'concrete and tendons,the need for improved inspection programs to iden-tify and quantify degradation, and a better under-standing of the potential impact of concrete-linerinteractions in aged containments. A comprehen-sive and standardized 1SI program is needed toidentify and quantify degradation in reinforcedconcrete. Integration and evaluation of the availa-ble research results and information from olderfacilities on the degradation of reinforced concretesubjected to long-term exposure to elevated temper-atures, radiation, and cyclic loading will supportthe development of an appropriate 1SI program.

PWR Reactor Coolant Piping

The key degradation sites for PWR reactor cool-ant piping are the main coolant pipe nozzles, dis-similar metal welds, and cast stainless steelcomponents. The reactor coolant piping is sub-,jected to thermal and pressure loading caused bysystem operating transients. The major degrada-tion mechanisms are low-cycle fatigue and thermalaging. The potential mode of failure is through wallleakage. IS methods include surface and volumet-ric inspection. Two unresolved technical issues are alack of accurate accounting of operating transients,and the development of a better understanding andassessment of the embrittlement of cast stainlesssteels because of thermal aging.

PWR Steam Generators

The key degradation sites in the recirculatingsteam generators are the inside tube surfaces at theU-bends, tube sheet, and tube supports; the out-side surfaces at the tube-to-tube sheet 'creviceregion; and ihe girth welds in the upper shellregion. The corresponding sites for once-throughsteam generators are the outside surface of the

tubes in the upper-tube sheet region, and the weldsattaching the thermal sleeves to the auxiliary feed-water inlets. The major degradation mechanismsinclude intergranular stress corrosion cracking'(IGSCC), intergranular attack (IGA), pitting,wastage, denting, fretting, and thermal fatigue.The likely modes of failure are cracking, localizedor uniform tube thinning, and wear out'of tube.material, which will ultimately lead to leakage ofprimary coolant to the secondary system and possi-bly to the outside environment. 1SI methodsinclude eddy-current testing and leak detectionmethods. Some of the steps taken to mitigate theeffects of the various degradation mechanisms areusing AVT chemistry on the secondary side, shotpeening the U-bend and roll-transition regions ofthe tubes, using thermally treated Inconel 600 tubematerial and a quatrefoil design for tube supportplate, and using 12% chromium ferritic stainlesssteel as tube support material. Additional work isneeded to understand and model the corrosionmechanisms, and to, better monitor the status ofsteam generator tube degradation.

.Reactor Pressure Vessel Supports

The potential degradation sites for the neutronshield tanks and column supports are at the core hori-zontal midplane elevations, and for the cantilever sup-ports they are in the active height of the core. These

*RPV supports are subjected to neutron irradiation,tensile stresses, operating temperatures, and a corrosiveenvironment because of the presence of water. Themajor degradation mechanisms include neutronembrittlement, corrosion, and radiation damage to thelubricant used in the sliding foot assembly: The poten-tial failure mode of the neutron tank, cantilever, orcolumn supports is brittle fracture. Two remainingtypes of supports, i.e, the skirt and bracket types; arenot likely to undergo catastrophic brittle failurebecause they are exposed to very little irradiation.However, the skirt support is subjected to fatiguebecause of the thermal- and pressure-induced expan-sions and contractions of the RPV during startup andshutdown. Currently, there are no requirements for ISIof RPV supports. Additional Fork is'needed todevelop a fracture toughness data base for RPV sup-port steels irradiated at <2320C (4500F), determinethe range of radiation environment conditions for sup-port structures, and investigate the effects of actualradiation levels'on lubricants used in the sliding footassemblies and between RPV nozzles and supports.

I

I

I

I

v

I

I

BWR Reactor Pressure Vessels

The key degradation sites for BWR RPVs arenozzles, safe end welds, closure studs, and beltlineregion. BWR vessels are subjected to mechanicaland thermal loads and neutron irradiation. Themajor degradation mechanisms are low and highcycle fatigue and neutron embrittlement. Thepotential mode of failure is ductile overload leadingto a leakage. A surveillance program similar to thePWR programs is required to assess irradiationdamage. ISI methods include volumetric inspec-tion of weldments, studs, and threads. An unre-solved technical issue is the need for closemonitoring of nozzle fatigue usage.

BWR Recirculation Piping

The key degradation sites for BWR recirculationpiping are the dissimilar metal welds at the safeends, the cast austenitic stainless steel components,and the crevices at the shaft sleeves. The BWR recir-culation piping is subjected to cyclic tensilestresses, an oxygen environment, high tempera-tures, and has sensitized heat-affected regions. Themajor degradation mechanisms are IGSCC, ther-mal fatigue, thermal embrittlement, and crevicecorrosion. The potential mode of failure is a leak-age through a crack in the piping. ISI methodsinclude ultrasonic examination and use ofmoisture-sensitive tape. Two unresolved technicalissues are the lack of an accurate accounting ofoperating 'transients, and the development ofimproved assessments of embrittlement because ofthermal aging. Additional work is also needed todevelop a better understanding of the effects of thehydrogen added to the recirculation piping loop toreduce the oxygen level in the coolant.

Current In-service InspectionMethods

Many of the standard NDE methods employed tosatisfy ISI requirements were developed for the detec-tion and qualitative assessment of fabrication-relateddefects. These methods are not entirely adequate forresidual life assessment. Inspections for life assessmentgenerally require greater detection reliability and amore quantitative determination of defects and accu-mulated damage than traditional ISI. The ISI methodsgenerally practiced by the nuclear industry are visualexamination, 'penetrant and magnetic particle testing,

x-ray radiography, eddy-current testing, ultrasonic test-ing, and occasionally acoustic emission monitoring.

Visual examination is the most widely used NDEmethod. In most cases, it provides an indicationthat damage may have occurred but cannot directlyquantify the amount of material damage. Pene-trant and magnetic particle testing are used toimprove the visibility of surface-connected flaws,and are therefore an extension of the visual exami-nation rhethod. X-ray radiography measures den-sity variations that may be due to cracks,inclusions, porosity, voids,' lack of bonding, anddimensional changes. The application of x-rayradiography is restricted because of the slow rate ofexamination. The use of single- and multifrequencyeddy-current testing is generally limited to inspec-tion of near-surface cracks in simple geometriessuch as PWR steam generator tubes. However, thestandard eddy-current methods are not adequate todetect and characterize circumferential flawscaused by intergranular stress corrosion attack.Ultrasonic testing has been accepted as the mostuseful volumetric examination method, especiallyfor inspection of welds and adjacent base material.The limitations of standard ultrasonic test methodsare due to deficiencies in the available technologyand human factors, such as operator boredom.Acoustic emission techniques are used to detectgrowing flaws in pressure vessels. The main advan-tage of the acoustic emission method is its use as aprecursor to impending failure (rather than amethod for flaw sizing required for residual assess-ment). Another useful application of acousticemission monitoring is leak detection in pressurevessels. Among all the standard NDE methods, theeddy-current and ultrasonic methods are the mostpromising for making quantitative damage relatedmeasurements needed for residual life assessment.

The ISI of major nuclear power plant components iscontrolled by USNRC regulations and Section Xl ofthe ASME code. Improvements in the ASME codeNDE methodology are being made to detect flaws inpiping, RPVs, containments, and steam generators.The current ASME code methodology is not adequateto assess the residual life of the major LNVR compo-nents. More efforts are needed to develop field-usableNDE techniques and equipments. The major unre-solved issues associated with current ISI methods are(a) the need for quantitative sizing of flaws for use infracture-mechanics analyses, (b) the need for methodsto inspect cast stainless steel components, and (c) theneed for techniques to measure the degradation inmechanical properties during long-term service expo-sure.

F,U

L

vi

U7

I

Current Life AssessmentMethods

Life assessment techniques include testing of sur-veillance specimens, monitoring of operationalparameters, evaluating samples removed fromservice-exposed components, and predicting dam-age accumulation processes. The data from the sur-veillance programs used to assess the irradiationembrittlement of the RPV material show a goodcorrelation between the brittle-to-ductile transitiontemperature shifts at 30 ft-lb (41 J) and the tensileyield strength. However, a correlation with a morerelevant material property such as fracture tough-ness of the RPV steel is not well established. Moni-toring of the operational pressures andtemperatures may be used to determine the fatigueusage factors at critical locations in the primaryloop. Metallographic and fractographic examina-tions and fatigue testing of samples removed fromservice-exposed components have been used toassess the structural integrity of LWR piping sys-tems. Analytical prediction methods are required toestimate accumulation of damage during antici-pated future operations. A fracture-mechanicsapproach based on probabilistic life assessmentmethodology can be used to estimate the reliabilityof pressure vessels and piping. Two major unre-solved issues are (a) lack of confidence in the ade-quacy of the current models for material damageaccumulation processes, such as irradiationembrittlement and fatigue crack initiation andgrowth under a spectrum of loadings, and (b) theunavailability of archival mechanical property datafor comparison with the properties of the samematerial after service exposure, so that the degreeof degradation can be assessed.

Emerging Methods for Inspectionand Life Assessment

Recent interest in LWR plant life extension isencouraging the development of new methods forinspection and life assessment. Inspection methodsare needed to accurately determine the size, shape,location, orientation, and type of both surface andinternal flaws, (including microstructural phasechanges), so that fracture mechanics approachesmay be used for life assessment. Current NDEmethods are not capable of detecting time-dependent changes in microstructural features,such as changes in the ferrite phase in austenitic-ferrite stainless steels, and the precipitate phases in

stainless steel; or the radiation embrittlement ofRPV steels. Such changes can be detected by repli-cation, extraction, indentation hardness testing,x-ray diffraction, and electrochemical testing atassessable locations. Another potential method isthe use of eddy-current to measure residual stress.

The synthetic aperture focusing technique(SAFT) for ultrasonic testing (UT) has been devel-oped to provide enhanced visual images of flawsdetected during inspection. While the initial resultsof the SAFT-UT application in the nuclear industryare encouraging, additional work is needed to fullyqualify this technique. Another emerging method,a dc potential drop method, may be used to detectsurface cracks on the inside of a pipe. Computer-aided techniques for eddy-current testing are alsobeing developed to detect flaws in austenitic stain-less steel pipe.

The USNRC is sponsoring programs usingacoustic emission technology for on-line monitor-ing of crack growth in pressure boundaries, andleak surveillance in LWVR systems. On-line monitor-ing appears to be a promising technique to detectfatigue-crack growth in RPVs and stress-corrosioncrack growth in stainless steel piping. Initial evalua-tion results of the acoustic emission leak-detectiontechnique are encouraging, but the method requiresadditional field validation.

The emerging methods for life assessment employminiature test samples that can be removed from acomponent-with negligible damage to the compo-nent, on-line procedures for the calculation of damage,and improved models of material degradation anddamage accumulation. Miniature specimen testinig(MST) can provide a direct measure of the degree ofaging. Miniature specimens can be valuable for surveil-lance testing where only a limited amount of test mate-rial is available and where space available for materialirradiation is restricted. The main constraint to theapplication of MST is that the specimen must be largeenough to be representative of the material from whichit has been removed. MST may be employed to mea-sure stress-strain response, fracture toughness, andfatigue cracking. MST also may be used to measure thethrough-thickness properties of thin steel plate, and tocharacterize the material properties near welds.

On-line damage and remaining life calculationsmay be performed for some component locationsby monitoring key operating parameters on-line.For example, this approach can be applied to calcu-late fatigue usage factors for .RPV nozzles duringstartups, shutdowns, and major operating tran-sients when the nozzle temperatures are monitoredduring operation.

I

I

I

I

I

vii

l

I

Finally, effective engineering models are required forthe successful implementation of life assessment strate-gies. These models must provide a balance betweenengineering sophistication and practical utility so thatgood assessments can be made economically. Using

simplifying assumptions and the results of the emerg-ing improved inspection and monitoring methods,miniature specimen testing and on-line monitoring,these models should be capable of predicting themajor, important features of material degradation.

7

r,

A_.

La.

viii

-ACKNOWLEDGMENTS

The editors and authors of this report acknowledge the important role thatJ. P. Vora of the U.S. Nuclear Regulatory Commission has played in providing pro-grammatic guidance, review, and encouragement for this effort.- The editors and authors also sincerely appreciate the efforts of M. E. Lapides of

the Electric Power Research Institute and D. E. Hostetler, Vice Chairman of theNuclear Utility Plant Life Extension Steering Committee, in arranging an industryreview of this report. In particular, the critical review performed by the technical staffof the Grove Engineering and the Multiple Dynamics Corporation and other consult-ants is acknowledged. The editors and authors also thank A. B. Johnson, Jr., and hisco-workers at the Pacific Northwest Laboratories and D. M. Eissenberg,K. V. Cook, R.WV. McClung, and C. E. Pugh of the Oak Ridge National Labora-tory for providing detailed and useful reviews of this report. All these reviews havehelped make this report more complete and accurate. *

ix

CO NTENTS

CONTRIBUTORS ............................ ............................ ii

ABSTRACT ......................................................... iii

EXECUTIVE SUMMARY ....................................................... iv

ACKNOWLEDGMENTS .. ix

1. INTRODUCTION . ...................................................... I

2. RANKING OF MAJOR PRESSURIZED WATER REACTOR COMPONENTS 3

3. PRESSURIZED WATER REACTOR PRESSURE VESSELS .12

4. PRESSURIZED WATER REACTOR CONTAINMENTS AND BASEMATS .32

5. PRESSURIZED WATER REACTOR COOLANT PIPING ...... ...................... 55

6. PRESSURIZED WATER REACTOR STEAM GENERATORS .66

7. REACTOR PRESSURE VESSEL SUPPORTS FOR PRESSURIZED WATERREACTORS AND BOILING WATER REACTORS .79

8. RANKING OF MAJOR BOILING WATER REACTOR COMPONENTS . .95

9. BOILING WVATER REACTOR PRESSURE VESSELS. 100

10. BOILING WVATER REACTOR RECIRCULATION PIPING ............ .............. 107

11. NONDESTRUCTIVE EXAMINATION METHODS .............. .. ................. 113 r12. ADEQUACY OF ASME CODE IN-SERVICE INSPECTION METHODOLOGY .. 122

13. CURRENT LIFE ASSESSMENT TECHNIQUES ............... .. .................. 129

14. NEW OR EMERGING METHODS FOR INSPECTION AND LIFE ASSESSMENT 143

15. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS ...... .................. 167I.

i-

x

......-

I

* 1 .. II ' I I

RESIDUAL LIFE ASSESSMENT OF MAJOR LIGHTWATER REACTOR COMPONENTS-OVERVIEW

VOLUME 1 I

1. INTRODUCTION

The United States was one of the first nations touse nuclear power to commercially generate elec-tricity and, therefore, has some of the oldest oper-ating commercial reactors. As U.S. light waterreactors have matured, problems associated withtime- or cyclic-dependent degradation (aging)mechanisms such as stress corrosion, radiationembrittlement, fatigue, and other effects have-occurred and have raised questions about the con-tinued safety and viability of nuclear plants and, inparticular, about the integrity of the primary cool-ant pressure boundary. These problems haveincluded cracked piping at boiling water reactors(BWRs), steam generator degradation at pressur-ized water reactors (PWRs), defective valves andrelays and inadequate means -for detecting andcharacterizing flaws.

At the same time, with a continually increasingdemand for electricity and limited new generatingcapacity under construction, the U.S. electric utili-ties are motivated to keep their existing plants oper-ating beyond the original design life at as high acapacity factor as possible. The economics of plantlife extension are clearly favorable. Studies cospon-sored by the U.S. Department of Energy (DOE)and the Electric Power Research Institute (EPRI)show that replacing any single nuclear plant com-ponent can easily be justified, if the life of the plantcan be extended for a number of years.-Extendingthe life of a 1 ,000-MWV plant by 20 years is expectedto realize a net present'worth of about $1 billion.

Therefore, the potential problems of managingaging in older plants and the resolution of technicalsafety issues in consideration of the development ofappropriate life extension criteria have become amajor focus for the research sponsored by the U.S.Nuclear Regulatory Commission (USNRC). This is -reflected in the following Policy and PlanningGuidance (PPG NUREG-0885) provided to-theNRC staff in 1986:

"Requests for an operating license renewalare to be anticipated and will requireadvanced planning and analysis. TheCommission intends to continue develop-

, ment of the policies and criteria to definerequirements for operating license exten-sions to help assure that industry's effortsin this area are focused on the primary reg-ulatory concerns."

An important part of the USNRC research effortis the Nuclear Plant Aging Research (NPAR) Pro-gram that is being conducted at several nationallaboratories, including the Idaho National Engi-neering Laboratory (INEL). One of the NPAR pro-gram tasks at the INEL is to develop theappropriate technical criteria for the USNRC toassess the residual life of the major LWR compo-nents and structures. These assessments will helpthe USNRC resolve certain safety issues anddevelop policies and guidelines for making operat-ing plant license renewal decisions. Most of theeffort for this residual life assessment task isfocused on integrating, evaluating, and updatingthe technical information relevant, to aging andlicense renewal from current or completed NRCand industry research programs. A five-step

-approach is being pursued to accomplish the resid-ual life assessment task: (a) identify and prioritizemajor components, (b) identify degradation sites,mechanisms, stressors, potential failure modes,.and evaluate current in-service inspection (ISI)

"methods, (c) assess current and advanced inspec-tion, surveillance, and monitoring methods andevaluate maintenance programs, (d) develop resid-ual life assessment models, and (e) develop criteriafor-license renewal. This report addresses the pro-gress made toward gaining a qualitative, under-standing of the degradation mechanisms active inseveral light water reactor components and it repre-sents a completion of the first step and partial com-pletion of the second step of. the approachdescribed above.- Virtually all the major components and structureswithin a nuclear plant complex must be evaluated in alife-extension program. However, from the USNRC'sperspective of ensuring the health and safety of the

-public, the first step in the residual life assessment taskis to identify those major components that are critical

I

I

IiI

i :I I

I

I

I

to nuclear power plant safety during normal operation,off-normal conditions, design-basis accidents such as ahypothetical large-break, loss-of-coolant accident orthe design-basis earthquake, or a severe accident. Thisreport presents appropriate criteria for selecting thecomponents and structures of interest and then ranksthese components in order of importance. The nextstep is to integrate the available technical informationso as to identify the degradation sites, stressors, degra-dation mechanisms, failure modes, and ISI methodsassociated with each of these major components. Thisreport identifies the degradation sites and correspond-ing processes for seven major components mentionedlater in this chapter. The remaining components will beaddressed in a later edition of this work. Future workwill be concerned with developing recommendationsfor quantitative residual life models'

A major factor affecting the life of the compo-nent is the amount of aging, i.e., degradation ofperformance over time, in the materials from whichthe component is fabricated. This is critical forthose components that have a relatively small origi-nal safety margin. The degree of aging is oftenreferred to as the damage state of the material.Material that has not been degraded by servicewould have zero damage, whereas a failed materialwould have- I00%o damage. Knowledge of the cur-rent damage state of the material is essential toassess the residual life because it establishes the ini-tial conditions for the remaining life. ISIs are per-formed to measure the current state' of damage.This report discusses the current nondestructiveexamination (NDE) methods used to detect andcharacterize defects in materials. This report alsodiscusses certain new or emerging examinationmethods that are being successfully used in otherindustries and that could be used to quantify dam-age states and measure material properties innuclear power plant components. With an ade-quate characterization of the damage state and aquantitative model, an accurate assessment of theresidual life of selected major structural compo-nents will be possible.

In addressing this task on residual life assess-ment, major emphasis has been placed on integrat-ing the available technical information on therelevant degradation mechanisms and applicableNDE methods. The sources for the technical infor-mation are technical reports from the USNRC andEPRI, technical workshops and conferences,American Society of Mechanical; Engineers(ASME) Section XI meetings, technical seminars,technical journals, technical publications inEurope and Japan, technical discussions with

experts, and contributions from subcontractor spe-cialists with expert knowledge of specific majorequipment or NDE activities. Several additionalsources of information will be added to this list.One source that is being currently pursued is theplans and results of certain ongoing research activi-ties at the national laboratories and other researchorganizations. Also, operating experience data,i.e., licensing event reports and NPRDS, includinginformation on operating transients and NDEactivities taking place at nuclear power plants willbe used.

This report is organized into three major parts. Thefirst part, Chapters 2 through 7, discusses major PWRcomponents. The second part, Chapters 8 through 10,discusses major BWR components. The third part,Chapters 11 through 14, discusses NDE methods.

Chapter 2 presents the criteria used to select themajor PWR components for this residual life assess-ment task. These criteria are also used to select themajor BWR components. Chapter 2 ranks the selectedmajor components according to their relevance toplant safety. Chapters 3 through 6 discuss the aging ofthe four highest ranked PWR components: the reactorpressure vessel, the containment, the reactor coolantpiping, and the steam generators, respectively. The dis-cussion for each component includes a description ofthe component, the stressors acting on the component,the degradation sites and their rankings, degradationmechanisms active in the component, potential failuremodes of the component during normal operation andtransient and accident conditions, and current ISImethods used for the component. A prioritized list ofdegradation sites and associated degradation mecha-nisms, potential failure modes, and ISI methods arepresented in a table at the end of each section. Thediscussion for the other components follows a similarsequence. Chapter 7 discusses the aging of the reactorpressure vessel supports in both PWR and BWRplants. The major BWR components are identifiedand prioritized in Chapter 8. Chapters 9 and 10 dis-cuss the aging of the BWR reactor pressure vessels andrecirculation piping, respectively. Chapter 11 discussesthe standard ISI methods. Chapter 12 discusses thecurrent NDE activities in the ASME Section Xl com-mittees. Chapter 13 discusses the current life assess-ment methods. Chapter 14 evaluates the applicabilityof advanced NDE and life assessment methodsemployed in the fossil power plant and petrochemicaland aerospace industries as well as those being devel-oped for LWR applications. Finally, Chapter 15presents the conclusions and recommendations for fur-ther work to assess the residual life of selected LWRcomponents.

T

U

L

U

2

12

I

2. RANKING OF MAJOR PRESSURIZEDWATER REACTOR COMPONENTS

V. N. Shah

IResidual life assessment of the major components isof primary interest for continuing safe operation andlife extension of a nuclear power plant. This chapteridentifies the major components of interest in commer-cial pressurized water reactors (PWRs). In identifyingthese components, it is assumed that the lifetime of thesensors, controls, batteries, certain types of pumps,valves, motors, etc. is much' shorter than 40 years andthese smaller, less expensive components will be main-.tained, refurbished, and/or replaced frequently. There-fore, the residual life of these small components' willnot significantly affect the life extension of the plantand is not addressed in this report.

The major components reported here are identi-fied and prioritized according to their relevance toplant safety'only. The industry-sponsored Surry-lpilot plant project has also identified componentsfor plant life extension studies.I However, the iden-tification and prioritization criteria used by indus-try pilot project were based on plant safety,reliability, and cost.

2.1 Key PWR Components

The selection of the key components is based ontwo main safety criteria: to contain the release offission products that may take place during an acci-dent, and to maintain an acceptable'low'level ofradioactivity 'in the-containment'during normaloperation. There are several barriers to the releaseof fission products. Two'of iheii barriers are asso-ciated with the fuel rods. Because these rods arereplaced several times''du'ring the lifetime of a reac5

tor, they are not-included in'the selected compo-nents. The 'remaining two'barriers are the pressureboundary components and the' containmenit.Therefore, the pressure boundary components,'-containment, and their supporting 'structures areincluded in-the list of selected major components.The control 'rod driv&e mechanism (CRDM) isincluded in the list because its failu'e'mrnay lead to ananticipated transient 'without scrarm "(ATWS) orreactivity insertion accident." The s~afety-relatedcables and connectors are included beeause theyprovide sigials related 6t plant co'nditini't6 ihe'sen'sors that control the' operation 'of the safety sys-tems. The failure of the cables' and connectorsbecause of aging may lead to common-cause failure

of a number of systems. The emergency diesel gen-erator is included in the list because it is requiredfor plant safety during a station blackout event..-The reactor pressure, vessel (RPV) internals areincluded in this list because their failure may pre-vent control rod insertion or cause fuel failure. The.RPV supports and biological shields maintain thelevel of radiation in the containment at an accept-able level and provide support to the' RPV and pri-mary coolant piping, and therefore they areincluded in the list.

Major PWR Components

1. Reactor pressure vessel2. Containment and basemat3. Reactor coolant piping and safe ends4. Steam generators5. Reactor coolant'pump body6. Pressurizer7. Control rod drive mechanisms8. Cables and connectors9. Emergency diesel generators

10. RPV inte'rnals11. RPV supports and biological shields.

The RPV is ranked Number I because it is themost' critical major component 'as far as plan''safety is concerned. It is one of the major compo-nents of a nuclear power plant for which there is noredundant member. The catastrophic failure of apressure vessel could lead to a rapid core'meltdowvn,generating high pressure and temperature loadingsfor'which light water re'actor containments ar 'not'designed. Therefore, the RPV is ranked higher thanIthe c'ontainrnient. The' vessel is"subjected to high'

system pressures,' temperatures, neutron irradia-tion, and cyclic fatigue. The most likely degrada-tion' sites 'are' the circumferential rand axial'weldments in the beltline region of the vessel.' Otherpossible degradation siies'of concern are the hot-'leg' and cold-leg nozzles, the weldment at the lowershell to bottom head junction, the vessel flanges''and studs, and'the instrument penetrations in thebottom head. The'most important potential degra-dation mechanism' is neutron embrittlement'. Thefactors affecting neutron embrittlement of RPVsteel are fast' neutron fluence, irradiation

I

I

I

I

,1

3

I

I

temperature, and the chemical composition of thesteel and weld metal (mainly Cu and Ni con-tent).2 '3 The fast neutron fluences and the chemi-cal composition of the weld metal varycircumferentially and through the thickness of theRPV wall. The irradiation effects on the RPV steelsare well documented, required to be considered inoperation and test limits, and monitored by surveil-lance programs during service. A less critical degra-dation mechanism is fatigue. However, the damagecaused by fatigue is irreversible, while irradiationdamage of pressure vessel steels may be reversible.

The containment and basemat are ranked higherthan the remaining components because they arethe major barriers protecting the public during anaccident from released fission products. Under asevere accident scenario, it is possible that the con-tainment internal pressure can rise to a value thatmay cause leakage through the containment wall.LWR containment structures may be divided intothree types: prestressed concrete containments,reinforced concrete containments, and steelcylinder/steel sphere containments. The mostlikely degradation site for a prestressed concretecontainment is the posttensioning system, consist-,ing of tendons and anchors. The potential degrada-tion mechanisms for the posttensioning system aretime-dependent relaxation of tendon material, andhydrogen embrittlement and stress corrosion crack-ing of the heat treated anchor heads.4 Corrosionr ofthe tendon material is also possible. The liner con-stitutes the weakest element in the prestressed andreinforced containment structures during a designbasis accident (DBA) and severe accidents.5 Aliner-concrete interaction may take place at loca-tions of major concrete cracks and distortions andintroduce highly localized strains in the liner. Thisstrain concentration may cause localized rupture ofthe liner that will lead to leakage from the contain-ment at a relatively early time in a pressurizationtransient. This leakage to the environment repre-sents the most likely failure mode for the dry andsubatmospheric types of containment but not forthe dual containment, and it is more likely in thereinforced structures than in the prestressed struc-tures. Prestressed structures see relatively minorcracking until nearly twice the design pressure, butreinforced structures develop widespread crackingat about 75%. of the design pressure. The relevantdegradation modes affecting the concrete materialproperties are deterioration because of aggressiveenvironments and internal chemical reactions. Theaggressive environments include nuclear heating,extensive cycling of wetting/drying and freezing/

thawing of the exterior walls, acid rains and sulfate-bearing ground water, and carbonation. 6 ' 7

Carbonation is the process by which the alkalinecalcium hydroxide present in the concrete is neu-tralized by carbon dioxide and atmospheric mois-ture.8 Cracked concrete can provide a direct path ofwater through or near to the rebars that may lead tocorrosion of reinforcing steel and subsequently toadditional cracking and spalling of the concrete.The internal chemical reactions include alkali-aggregate, cement-aggregate, and carbonate-aggregate reactions. The most likely degradationsites for steel cylinder/steel sphere containmentsare the base metal and the weldments in the con-tainment wall. The major degradation mechanismis the corrosion of the steel walls.

The four components representing the pressureretaining boundary are ranked right below the con-tainment and vessel because they constitute the firstbarrier to the release of the fission products follow-ing the failure of fuelrod cladding during an acci-dent. These components are the reactor coolantpiping and safe ends, the steam generator, the reac-tor coolant pump body, and the pressurizer. Thefailure of any of these four components may lead toa large-break or small-break loss-of-coolant acci-dent (LOCA) that may challenge the integrity ofthe containment by imposing high thermal andpressure loadings on it. The reactor coolant piping.and safe ends are ranked third, i.e. the highestamong these four components, because the impactof a failure is severe. A break of the reactor coolantpiping will impose significant mechanical loads onthe component supports and other structural mem-bers and may damage them. The weldments at thesafe ends have dissimilar metal welds, and are themost likely locations for degradation. The branchnozzles and the cast stainless steel components, i.e.elbows and T-connections, etc., in the reactor cool-ant piping are also possible locations for degrada-tion. The potential degradation mechanisms arethermal fatigue and thermal embrittlement.

The steam generator is ranked next to the reactorcoolant piping because the rupture of its tubes will pro-vide a passage for primary coolant to outside the con-tainment by way of the relief valves. In addition, amajor steam generator tube rupture event is of seriousnature because of the rapid depressurization of the pri-mary coolant system and the complications that resultin terminating the pressure transients that ensue. Infact, relatively small tube rupture accidents have takenplace at several nuclear power plants, 9 and the steamgenerators in the operating PWRs have been plaguedby failures because of mechanical and chemical

I,

Fr1

4

U'

I

problems. The majority of problems have been associ-ated with the recirculating type units manufactured byWestinghouse and Combustion Engineering. However,the once-through units manufactured by Babcock &Wilcox have also'experienced some operational diffi-culties. 1I The locations of the degradation sites within -a recirculating type steam generator include: the out-side surfaces of the tubes near the support plate in theregion of the cold leg, I the inside surface of the Inconeltubing in the region of the hot leg, 1 1 the divider plate,the tube support plates, the feedater nozzle, theJ-tubes attached to the feedring,12 and the upper shellto transition cone girth welds. The J-tubes are nor-mally fabricated from carbon steel and'have beenobserved to degrade through wall thinning and perfo-ration by a corrosion-erosion mechanism. The J-tubesin the recirculating type units manufactured byWestinghouse have been replaced,'and therefore they'probably will not have a significant impact on the lifeextension of those steam generators. A potential degra-dation mechanism for the girth welds is -corrosionfatigue. 13 The normal operating water level for thesteam generators is in the vicinity of those welds andthe normal oscillations of the water level may imposecyclic thermal loads on the welds. Cracks have beendiscovered in the girth welds of the Surry-2 and IndianPoint-3 steam generators.

The locations of the degradation'sites within aonce-through steam generator (OTSG) include thesecondary face of'the upper tube sheet and theuppermost tube support plate in tube rows adjacentto the inspection lane.lU Another likely degrada-tion site is the weld at the auxiliary feedwater nozzlethermal sleeve in the OTSG. The cracking of thisweld will increase the potential for thermal shock tothe OTSG shell. The potential degradation mecha-nisms for steam generator tubes are wastage, dent-ing, intergranular stress corrosion cracking,intergranular attack,' erosion corrosion, 'pitting,and fretting. The use of all-volatile treatment of thesecondary coolant has significantly reduced thedamage due to wastage and denting in the recircu-lating type steam generators. 14 The potential deg-radation mechanism for the divider plate andfeedwater nozzle is thermal fatigue. '

The massive size'of the reactor coolant pumpcasing provides primary protection to the pressureretaining components within the reactor coolant-system and containment building from missilesgenerated by internal failure of'pump elements(e.g. impeller parts, shafts, fasteners). Therefore, itis ranked higher than the pressurizer. The reactor'coolant pump casings in the PWR systems are thick(8 in. or greater) and constructed of cast austenitic

steel. Because of the'casing size, the pumps manu-factured by Westinghouse are welded by the elec-troslag process that introduces high residualstresses in the heat-affected zones near the weld-ments.' 5 The pump is powered by a large electric'motor (4000 hp and larger) to maintain the coolantflow through the reactor vessel, and the pump cas-ing is subjected to the high vibratory forces of the

.rotating pump impeller assembly.' The vibratoryforces, in 'combination with operating 'stresses,fluid dynamic cyclic loadings, and the primarycoolant environment, may. contribute to'flawgrowth. The potential degradation mechanisms forthe pump casing are corrosion fatigue and thermal*embrittlement.L6 The other possible locations fordegradation'sites are flange and seal housing bolt-ings. Corrosion fatigue is the potential degradationmechanism for the boltings.

The pressurizer is the last key component thatrepresents the'primary coolant pressure retainingboundary. A' surge line connects the pressurizer tothe hot leg, while a spray line connects it to the coldleg. The surge and spray nozzle connections-at thepressurizer vessel are'subjected to cyclic tempera-ture changes resulting from the transient conditionsin the reactor. The spray nozzles'are subjected tolarger temperature changes than the surge nozzle.Thermal fatigue is the potential degradation m'ech-anism for these two nozzles. The'spray head, uppershell barrel, and seismic lugs are the other possiblelocations for degradation. The potential degrada-tion mechanism for the spray head is erosion, andfor the shell barrel and seismic lugs it is fatigue.Another potential degradation mechanism for the

,spray head is thermal embrittlement, if it is made ofcast stainless steel.

The next three components play critical roles in miti-gating LWR operational transients and accidents.' Thecontrol rod drive mechanism is ranked highest amongthese three components because its failure may lead toan ATWS, or'a reactivity initiated accident (RIA), or asmall-break LOCA. The key subcomponents that maydegrade are the drive r6d assembly and the control rodpressure vessel. Electrical failure caused by closed con-tacts in a switch may result in a control rod withdrawalevent leading to a RIA. Mechanical binding of thelatch mechanism may result in the failure of the scramsystem and lead to an ATWS accident. The pressurevessel houses the latch assembly and drive rod assem-bly, and provides a pressure barrier for'the CRDMbetween the primary coolant and containment. Failureof the control rod drive pressure vessel may result in arod ejection accident along with a small-break LOCA.Eleven failures of pressure vessels and two failures of

I

I

I

5

I

drive rod assemblies are reported in the data base thatcovers the information from 31 Westinghouse-designed, operating plants up to and including 1983.17The potential degradation mechanism for those pres-sure vessels having cast stainless steel latch housings, isthermal embrittlement. I Further research is required todetermine the impact of the thermal embrittlement onthe residual life of the CRDM. Wlear is the most likelydegradation mechanism for the drive rod assembly.

Safety-related cables and connectors are active dur-ing normal operation of the reactor, while the emer-gency diesel generators are needed if offsite power islost. Therefore, the safety-related cables and connec-tors are ranked higher. The key degradation sites are thecable insulation, especially in the higher temperatureregions of the containments, and the connectors.Cable insulation becomes brittle because of thermalaging, radiation, and oxidation, and may break duringa seismic event causing electrical shorts. Cables mayalso fail by creep shortout, which occurs when amechanically-stressed cable is bent over a corner andthe conductors are pulled toward each other or towarda grounded electrode through insulation softened byaging.18 Certain types of connectors have inserts thatare made of carbon, which can become embrittled andfracture because of aging. The Westinghouse-designedreactor, San Onofre-l, which has been in operationsince January 1968, has experienced the problem ofembrittled connectors. 1I Moisture intrusion hascaused corrosion of some connector material.

The emergency diesel generators must providebackup power to operate critical reactor safetyequipment in the event of a loss-of-offsite power tothe emergency bus bars. The failure of the dieselgenerators to provide backup power will lead to astation blackout transient. A station blackout tran-sient followed, by the failure of the steam-drivenemergency feedwater pump to provide coolingwater to the steam generators and the absence ofoperator intervention, is identified as a dominantcore melt sequence by the recent USNRC reactorrisk study. 19 The diesel generators have instrumen-tation and a control system that should start themautomatically on loss-of-offsite power, low reactorwater level, high contamination level, or otheremergency condition signals, and automaticallyapply the required load. A review of the variousfailure modes of the diesel generators indicates thatthe governor in the instrumentation and controlsystem is the subcomponent most susceptible toaging degradation. 20 The other subcomponentssusceptible to aging degradation are the-coolingsystem pumps and piping, the fuel system injectorpumps and engine piping, and the turbocharger.

The potential degradation mechanisms are metalfatigue and loosening of the fasteners.

The main function of the reactor internals is toprovide orientation and support for the reactorcore, and guide and protect the reactor control rodassemblies. It also provides a passageway, support,and protection for any in-vessel instrumentation.One of the potential failures of the RPV internals,i.e. baffle jetting, has led to fuel rod cladding deg-radation and disbursement of fuel into the coolantin certain reactors. Failure of the RPV internalsalso may relocate fuel away from the control rodsor prevent the control rods from inserting properlyand lead to an operational transient without scram.The key RPV internal components susceptible toaging degradation are the lower core plate, thebaffle-former assembly, the upper support columnbolts, the control rod guide tube sheaths and sup-port pins, the thermal shield bolts, the core barrelbolts, the in-core instrument nozzles, and the fluxthimble tubes. I Thermal shield and core barrel boltfailures have occurred in a number of olderWestinghouse plants. Twenty-one out of 54 in-coreinstrument nozzles in the Oconee-l RPV had beencracked and broken off in the region of the weld.The loose broken pieces of the in-core instrumentnozzles subsequently caused extensive damage tothe tube ends and to the tube-sheet welds in theupper head of one of the two steam generators. 21 Itshould be noted that some of the past RPV internalproblems were caused or aggravated by originaldesign errors. The potential degradation mecha-nisms are neutron embrittlement, stress-corrosioncracking, mechanical I wear, low- and high-cyclefatigue, and stress relaxation. The low-cycle fatigueis caused by the loads due to changes in power lev-els, vessel inlet and outlet temperature differences,and coolant pressures and flow rates. The high-cycle fatigue is a result of the flow-induced vibra-tions. The control rod guide tube sheath andsupport pins, and flux thimble tubes are locationsthat may experience significant mechanical wear.

The last major components are the RPV sup-ports and biological shields. There are several dif-ferent designs of RPV supports and biologicalshields employed in the commercial nuclear powerplants. One designed by Stone & Webster includes asliding foot assembly and a neutron tank. The slid-ing foot assembly supports both the vertical andtangential loads of the RPV and its permanentlylubricated foot slides radially to allow free thermalexpansion of the RPV. The neutron tank supportsthe reactor and, in addition, stores water in theannular space between the RPV and the tank wall

I

I_

F-1

6

I

and provides shielding against neutrons. The slid-ing foot assembly is made of a maraging steel andcoated with Heresite, an air-drying phenolic coat-ing. The potential degradation mechanism for thesliding foot assembly is the stress-corrosion crack-ing of its threaded parts. 1,9 The potential degrada-tion mechanisms for the neutron tank are neutronembrittlement and corrosion at the inside surfaceof the tank walls.

2.2 Summary, Conclusions, andRecommendations

The major PWR components are selected andranked such that the release of the fission productsthat may take place during an accident is con-tained. The RPV has been identified as the mostimportant component in PWR plants. This meansthat the failure of a reactor pressure vessel wouldhave a very significant impact on the safety of thepower plant. The other key components, accordingto their ranking, are containment, reactor coolantpiping, steam generators, reactor coolant pumpbody, pressurizer, control rod drive mechanisms,cables and connectors, emergency diesel generator,

RPV interna,''ahd RPV supports. The insightsand results presented in this chapter are based onpast experience as reported in the available litera-ture. These results will be updated as more data aremade available.

Table 2.1 summarizes the aging related informa-tion on the key PWR components discussed in thissection. It lists the components and gives reasonsfor their ranking. Table 2A identifies the mostlikely degradation site and other degradation sitesfor each component. Table 2.1 also lists the mostlikely degradation mechanism and other potentialdegradation mechanisms.

The major components reported in this sectionwere selected and prioritized according to their rele-vance to plant safety only. These criteria are differ-ent than the ones used in the industry pilot study.Therefore, the ranking of the major componentspresented in this section is somewhat different thanthe one in the pilot study. However, all major com-ponents identified in this section were also rankedhigh in the industry pilot study. Detailed discus-sions of the possible degradation processes ofPWR reactor pressure vessels, PWR containmentsand basemats, reactor coolant piping, steam gener-ators, and RPV supports are presented inChapters 3 to 7, respectively.

0

r:

7

r

I

Table 2.1. Key PWR components for residual life assessment

Rank Component Reasons for RankingDegradation Sites

(most likely, others)Degradation Mechanisms

- (most likely, others)

II. Reactor pressurevessel

2. Containmentand basemat

3. Reactor coolantpiping and safeends

4. Steam generator

Severe safety impactof catastrophic failureof an embrittledvessel

Public protectionduring an accident

Severe safety impactof a large break

Thbe rupture willprovide a passage fromthe primary systemdirectly to theenvironment for theprimary coolant,relatively pooroperating experience

Weldments in the beltlineregion, hot-leg andcold-leg nozzles, weldmentat the lower shell to bottomhead junction, vessel flangesand studs, instrumentpenetrations in lower head

Posttensioning system(tendons and anchors),concrete, metal liner,reinforcing steel

Weldments at the safeends, branch nozzles,cast stainless steelcomponents (elbows andt-connections), completeprimary loop piping in newWestinghouse plants

Inside tube surfacesat U-bends and tubesheet, outside surfacesat tube-to-tube sheetcrevices, divider plate,tube support plate,feedwater nozzle, girthweld

Neutron embrittlement,corrosion fatigue, thermalfatigue

Hydrogen embrittlementand SCC of anchor heads,corrosion and relaxationof tendon material,environmental degradation ofconcrete, corrosion ofreinforcing steel and liner

I

Fatigue, thermalembrittlement

I.

Wastage, denting, IGSCC,pitting, fretting, IGA,thermal fatigue, corrosionfatigue

5. Reactor coolantpump casing

6. Pressurizer

7. Control roddrivemechanism

Primary protectionfrom any internalfailure of pumpelements (impellerparts, shafts), primarycoolant pressureboundary

Primary coolantpressure retainingboundary

Failure may lead to areactivity initiatedaccident, an operationaltransient without scram.or a loss-of-coolantaccident

Casing wall, flanges,seal housing bolts

Surge and spray nozzles,spray head, upper shellbarrel, seismic lugs

Drive rod assembly, controlrod pressure vessel

Corrosion fatigue, thermalembrittlement

Thermal fatigue, erosion,thermal embrittlement

Wear and thermal embrittlement

I

8

IU

I

Table 2.1. (continued)

Rank Component

8. Safety-relatedcables andconnectors

9. Emergencydieselgenerator

10. Reactorinternals

11. RPV supports

Degradation SitesReasons for Ranking (most likely, others)

Active during normal Cable insulation, insertsoperation in mitigating in connectorsoperational transientsand accidents

Needed to operate Governor in thecritical safety instrumentation andequipment in the event control system, coolingof a loss of offsite system pumps and piping,power fuel injector pumps,

turbocharger

Failure may lead to Lower core plate,dispersement of fuel baffle-former assembly,into the coolant, or a upper support columnreactivity-initiated bolts, control rod bolts,accident control rod guide tube

sheaths and support pins,thermal shield bolts, in-coreinstrument nozzles, fluxthinible tubes bolts, in-coreinstrument nozzles, fluxthimble tubes

Failure will challenge Inside surface of thethe integrity of the support structure (neutronprimary coolant piping tank or column support)and the small lines at the core horizontalconnected to the RPV. midplane level; lubricantThe neutron shield in sliding foot assemblytanks maintain of neutron tank supportacceptable levels ofradioactivity in thecontainment

Degradation Mechanisms(most likely, others)

Thermal aging and creep ofinsulation, thermalembrittlement andcorrosion of connectors

Fatigue and vibrations

Neutron embrittlement, wear,high cycle fatigue, SCC,relaxation

r-

Neutron embrittlement,corrosion, degradation oflubricant

I

I,

L.

9

172

2.3 References

1. Seminar on Nuclear Power Plant Life Extension, Seminar Proceedings, cosponsored by EPRI, North-ern States Power, U.S. DOE and Virginia Power, Alexandria, Virginia, August 25-27, 1986.

2. J. Ahlf, "Irradiation Experiments Related to Reactor Safety in Particular Pressure Vessel Steels,"Nuclear Engineering and Design, 81, 1984, pp. 231-245.

3. U.S. Nuclear Regulatory Commission, "Effects of Residual Elements on Predicted Radiation Dam-age on Reactor Vessel Materials," Regulatory Guide 1.99, Revision 2, February 1986.

4. "Farley Tendon Problem Blamed on Water as NRC Mulls Generic Impact," Nucleonics J'eek, 26, 10,March 7, 1985, pp. 1-2.

5. R. S. Dunham et al., Methodsfor UltimateLoadAnalysisof ConcreteContainment, EPRI NP-4046,June 1985.

6. C. A. Negin et al., Extended Life Operation of Light Water Reactors: Economic and TechnologicalReview, Volume 2, EPRI NP-2418, June 1982.

7. D. J. Naus, "Concrete Material Systems in Nuclear Safety Related Structures-A Review of FactorsRelating to Their Durability, Degradation Detection and Evaluation, and Remedial Measures forAreas of Distress," Appendix B in The Longevity of Nuclear Power Systems, EPRI NP-4208,August 1985.

8. G. H. Lewis, Degradation of Building Materials Over a Lifespan of 30-100 Years, Volume 1, Commis-sion of the European Communities-Nuclear Science and Technology, EUR 10020 EN, 1985.

9. V. L. Sailor et al., Summary of Barrier Degradation Events and Small Accidents in U.S. CommercialNuclear Power Plants, NUREG/CR-4067, BNL-NUREG-5 1842, March 1985.

10. H. R. Carter, M. T. Childerson, and T. E. Moksal, Model Tests ofa Once-Through Steam Generatorfor Lane Blocker Assessment and THEDA Code Verification, EPRI-NP-3042, June 1983.

11. G. Taylor, "Maintenance Practices for Longer Plant Life," Nuclear News, 29, 3, March 1986, pp. 50-59.

12. J. D. Roarty, "Corrosion-Erosion of Steam Generator J-Tubes," International Meeting on NuclearPower Plant Maintenance, Salt Lake City, Utah, March 23-27, 1986.

13. C. J. Czajkowski, "Investigation of the Shell Cracking on the Steam Generators of Indian Point 3,"Proceedings of the 9th International Congress on Metallic Corrosion, Toronto, 1984, pp. 229-234.

14. 0. S. Tatone and R. S. Pathania, "Steam Generator Tube Performance: Experience with Water-Cooled Nuclear Power Reactors During 1982," Nuclear Safety, 26, 5, September-October 1985,pp. 623-640.

15. D. T. Umino and A. K. Rao, Long-Term Inspection Requirements for PWR Pump Casings, EPRINP-3491, May 1984.

16. J. F. CopelandandA. J. Giannuzzi, Long-Term Integrity ofNuclear Power Plant Components, EPRINP-3673-LD, October 1984.

10

I

17. G. K. Roberts and G.'R. Ahdre,' A 'Component Wiear-Out Analysis'on Control Rod Drive Mecha-nism, Westinghouse Electric Corporation, July 1985.; -.

18. 0. M. Stuetzer, Correlation of Electrical Cable Failure with Material Degradation, Sandia NationalLaboratories, NUREGICR-4548, March 1986.

19. U.S. Nuclear Regulatory Commission, ReactorRiskl"ReferenceDocument, Volume1, NUREG-1150,February 1987.'

20. J. V.' Vause, D. A. Dingee, and J. F. Nesbitt, Aging of Nuclear Station Diesel Generators: Evalua-tion of Operating and Expert Experience, to be published by the Pacific Northwest Laboratories forthe U.S. Nuclear Regulatory Commission.

21. H. W. Bertini, Descriptions of Selected Accidents that Have Occurred at Nuclear Reactor Facilities,ORNL/NSIC-176, Ap'ril 1980.

r

r

,_

.,

… , .- .: i i .

. .. .;, , ,.

,'','.'t. ' .. - .. - . ..

W .

II

I

3. PRESSURIZED WATER REACTOR PRESSURE VESSELSW L. Server, G. R. Odette, and R. 0. Ritchie

3.1 Description

The design of pressure vessels for commercialpressurized water reactors (PWRs) is very similar forthe different domestic nuclear steam supply systemvendors: Westinghouse (W), Combustion Engi-neering, Inc. (CE), and Babcock & Wilcox Co.(B&W). Both CE and B&W manufactured their ownvessels, while W procured vessels from CE, B&NV,Chicago Bridge and Iron (CB&I), or RotterdamDockyard Co. The vessels for Prairie Island, Units 1and 2, were manufactured by Societe des Forges atAteliers du Cresot. Also, at least one B&W vesselwas completed by Rotterdam Dockyard Co. for W,and this vessel is part of the Surry-l plant that isbeing studied in detail for life extension under Elec-tric Power Research Institute (EPRI)/Departmentof Energy (DOE) funding. Vessels were designedand built in accordance with the applicableAmerican Society of Mechanical Engineers (ASME)Boiler and Pressure Vessel Code, Section IIII at thetime of fabrication, except for the earliest vesselsthat were constructed before Section III existed.These earlier vessels were built to ASME Boiler andPressure Vessel Code, Sections I or VIII. Differentversions or updates to the Code apply to differentvessels depending upon the vintage. The structuralintegrity of the Sections I and VIII vessels has beenjudged to be comparable to Section III vessels ofapproximately the same era.2 Section III requiresmore limiting nondestructive examination of weldsso that the probability of manufacturing defects issmall.

The CE, early B&W, and CB&I production ves-sels used steel plates formed and welded to producethe vessel structure. The earliest commercial reac-tor was Yankee-Rowe (operational in 1961), whichhad a W vessel made by B&W of SA302B steel.SA302B was used in several early vessels untilSA533B-I steel plate became the industry standard.(Nickel-modified SA302B is essentially the samecomposition as SA533B-1). Automatic submergedarc welding was typically employed for both longi-tudinal and circumferential welds. The materialsconsumed in this welding process are a manganese-molybdenum-nickel filler wire and a granulatedflux that minimizes atmospheric contaminationand provides ingredients to form a slag to removeoxides during the welding process. The type of flux

material is important because the mechanical prop-erties of the weldment can differ depending uponwhat flux is used. B&W welds employed a Linde 80flux that typically gives lower values of upper-shelfCharpy V-notch properties than other fluxes. CEand CB&I used Linde 0091, 1092, and 124 fluxes;those. three fluxes produce similar mechanicalproperties. The Linde flux designations are propri-etary codes used by the Linde Division of UnionCarbide Corporation. Some shielded metal arc(SMA) welding may be used for complex geometrywelds (nozzles), fit-up, backwelding, and repairs.The SMA electrode is a wire coated with a bondedflux (which provides a protective arc environment);the welding current is automatically controlled, butthe arc voltage and welding speed are manuallycontrolled. An E8018 electrode is typically used forSMA welds by all fabricators. Minimum preheatand interpass temperatures (%3000F or 1491C) arecontrolled, and intermediate postweld heat treat-ment may be employed. All vessel welds arepostweld heat treated to reduce residual stresses.

For later B&W vessels, an SA508-2 forging steelwas used for fabrication. The main advantage ofthe forging is that the entire ring section is made ofone piece, eliminating the need for longitudinalweld seams. The W vessels are only slightly modi-fied versions of the CE and B&W vessel designs.

W has three slightly different versions of vessels,depending on whether a two-, three-, or four-loopplant was built; the number of loops dictate theoverall size (diameter) of the vessel as well as thenumber of nozzles. B&W has three slightly differ-ent versions with the main difference being thechange in fabrication from plate to forging mate-rial. The CE vessels change slightly in terms ofdiameter, but a more significant change is that thenewer System 80 design is larger to reduce neutronexposure and has fewer welds.3 Additionally, thenonsystem 80 CE-design vessels have instrumenta-tion nozzles on the top head; all others (includingthe System 80) have bottom-head instrumentation.

A typical reactor vessel section is shown inFigure 3.1 to illustrate the important areas of con-cern. These regions will be discussed in more detaillater. All vessels are lined with a stainless steel clad-ding (usually Type 308 or 309 stainless steel) that isapplied in one or two layers by either multiple wireor strip cladding SMA processes. 4 The cladding

12

I

, :' l ., ., SE->;

studs

flange

r-Vessel wallthickness transition

I-

I In-core- Instrumentation

pentrations7.3042

Figure 3.1. Typical PWR vessel showing important degradation sites. U.-

13

I

thickness can vary, but is typically around 0.2 in.(5 mm).

Many of the issues described in the next subsec-tions are also directly applicable to the discussionsof Chapters 5 and 9 on primary system piping andBWR RPVs, respectively.

3.2 Stressors

The radiation environment is a major stressorbecause it causes a degradation in mechanicalproperties. The other. primary stressors aremechanical pressure loads during operation, peri-odic thermal transients that induce through-wallbending moments, dead weight loads, residualstresses, and other abnormal loadings that cancause damage (such as pressurized thermal shock).Other stressors that can be important are waterchemistry and mechanical contact that result inwear.

The effects of neutrons from the nuclear reactionprocess bombarding the inside wall of the reactorvessel are a serious issue for maintaining the integ-rity of the vessel. As a result of irradiation expo-sure, the ferritic steel strength increases with anattendant decrease in ductility. In terms of theCharpy V-notch (CVN) energy properties, theductile-brittle transition temperature increases tohigher temperatures as shown in Figure 3.2 (belowthe transition temperature, the material behaves ina brittle manner and its fracture requires only littleenergy, whereas above the transition temperature,the material behaves in a ductile manner), and the

upper shelf level decreases in magnitude as the irra-diation exposure increases (upper shelf level in theCVN energy versus temperature curve representsthe fracture energy that corresponds to approxi-mately 100°1o ductile shear and remains approxi-mately a maximum for the upper range of testtemperatures). The fracture energy of the pressurevessel steel thus changes continuously with radia-tion exposure at a given temperature. Therefore,10 Code of Federal Regulations (CFR) 50 hasestablished a criterion that defines the transitiontemperature as the temperature at the fractureenergy level of 30 ft-lb (41 J). Monitoring of thecondition of the reactor beltline region is con-ducted through the use of surveillance specimentesting, which is covered later in Section 3.4.1.

Operating transients that occur with regularitywere used in the original ASME Section III fatiguedesign analysis. Typical normal, test, and upsettransients are given in Table 3.1. The actual num-bers of occurrences of these transients are usuallymuch less since conservative numbers were esti-mated for the design. Other transients that areabnormal also are important since the fatiguedesign did not consider their occurrence. Theseevents, such as a main-steam-line break or a small-l4reak loss-of-coolant accident, can create a poten-tial challenge to the vessel integrity (but notbecause of frequent occurrences).

Not only are many of the number of design tran-sients unlikely to occur, but the magnitude of actualevents are not generally as severe as those designedagainst. For example, plant operating proceduresgenerally limit the heatup and cooldown rates to a

I

I

1.

III

Temperature (OF)0 200-400 -200 400

C)

CD

C.)0

Cd

0.

D.L

160 >.0)C)a)

80 0

c.

30 coC ,

ITemperature ("C) 7.3197

Figure 3.2. Effect of irradiation on the Charpy impact energy for a nuclear pressure vessel steel.

14

Table 3.1. Typical plant transients and assumed design occurrences5 6

Transient

Plant heatup at 100l F (38QC)/hPlant cooldown at 1000F (380C)/h

Plant loading at 5 .full power/miPlant unloading at 5%yo full power/min

Step load increase of 10010 full powerStep load decrease of 10% full power

Reactor trip from full powerLoss of flow and abnormal loss of load

Loss-of-secondary pressure ' - :Hydrotest to 3125 psig (21.55 MPa), 400'F (204'C)

Operating basis earthquakeNormal plant variation 1100 psi and 100F (-120C)]

Number of Events

500'500

15,00015,000

2,0002,000

40080,

510

200>106

maximum of 50'F/h (281C/h), and actual practicereduces the rate even lower. Therefore, the originalfatigue design is generally very conservative.

The primary transients leading to fatigue dam-age (in terms of the cumulative fatigue usage factorused in Section III of the code) are plant heatup/!cooldown, plant loading/unloading, reactor trips,loss of flow, and abnormal loss of load. Other tran-sients, such as inadvertent safety injections, canalso contribute to fatigue damage even though theyare not explicit design transients. If these transientscan be controlled and properly monitored to indi-cate the actual level'of loadings, estimates of theactual fatigue usage can be obtained: This issue isdiscussed further in Section 3.4.2.

The other issues of water chemistry and mechan-ical wear are less important. The water chemistryissue pertains 'mainly to' the possibility ofcorrosion-assisted fatigue-crack 'propagation if acrack were to exist tha' penetrates the innercladding layer' and extended into the ferritic steelvessel material. Fatigue-crack growth rates for fer-ritic steels in water environments at operating tem-peratures may be an order of magnitude higherthan those measured in air environments at thesame temperatures. 7 The mechanical wear problemconcerns surfaces that contact and periodicallymove relative to each other. The action of the forcesbetween the surfaces creates potential wear of thesurfaces that can lead to eventual inoperability.

3.3 Degradation SitesThe sites of highest degradation depend on the

degradation mechanisms. Obviously, the sites ofdegradation relative to radiation embrittlement arein or near the reactor beltline region. Therefore, thewelds within that region become possibly the weak-est link because the welds are more likely to containdefects that can become cracks. Additionally, theproblem exists of higher copper (and nickel) con-tent in many of the older vessel welds leading tohigher radiation damage sensitivity. The steel fillermaterial used in the fabrication of older submergedarc welds was copper plated to enhance electricalcontact during welding and resistance to corrosionduring storage. The copper content in the weldsresults directly from the, copper plating and theresidual copper in the consumable steel filler mate-rial. It is unfortunate that the effects of copper(and nickel) were not recognized to be detrimentalto neutron exposure resistance earlier. Appendix Gto 10 CFR 50 defines the beltline region as includ-ing all shell material directly surrounding the effec-tive height of the fuel element assemblies plus an

-additional volume of shell material both below andabove the active core, with a predicted referencetemperature shift (ARTNDT) at the end of servicelife of 50°F (>280C). In some earlier CE vessels,(especially considering license renewal) the10 CFR 50-defined beltline region could eveninclude the nozzles, the nozzle welds, and the

I

r

15

r

I

nozzle shell. The base metals should not be ignoredbecause the copper content in older plates and forg-ings was not controlled to a minimum level; how-ever, there appears to be less radiationembrittlement in base materials as compared towelds with the same copper/nickel concentrations.Also it is possible that flaws can exist in the vesselcladding which could potentially lead to flaws inthe ferritic vessel wall.

In the case of fatigue, most of the problemsinvolve geometric discontinuities or changes.Informal review of CE-designedvesselsa reveals thefollowing order for the areas'of relatively higherusage factors: closure head studs (inside bottomsurface), outlet nozzles (two per vessel), inlet noz-zles (four per vessel),' and instrumentation nozzles(penetrations).

The instrumentation nozzles for the older CEdesign (at the top) did not have high fatigue usagefactors, but the System 80 design (at the bottom)does. Other areas considered include the closurehead itself and the vessel flange, control rod drivemechanism (CRDM) housings, vessel supports (atthe nozzles), vessel wall transitions, veisel'wall tobottom head juncture, surveillance holders, corestabilizing lugs, and a flow baffle. The CE designfor integrally welded surveillance holders is uniquebecause none'of the other designs allow attachmentto the vessel inner wall (cladding surface). The CEsurveillance holders are made of' a nickel-basedalloy welded to the stainless steel cladding. Corestabilizing lugs are located on the bottom: shell aswell as core stop lugs. A flow baffle is located at thebottom head 'shell course near the core stop lugs. Avent tube through the top head also is present' butdoes not present a fatigue problem. The support ofthe vessels for all W and CE vessels is at the nozzles,but the usage factors at these support areas are rela-tively low.

The W_ design is somewhat similar to CE designand yields similar results.b The CRDM housing inthe W design also appears to have relatively highfatigue usage. The CRDM nozzles in some casesare made'of cast stainless steel that has the poten-tial to degrade because of thermal aging. The sameareas of potential degradation turn up regardless ofwhether a two-, three-, or four-loop plant (i.e.,

number of nozzles) is considered. Also, no differ-ence between vessel fabricators is seen.

For the B&W design, a list of high usage factorlocations for a typical B&W system were obtained.cThose regions for the reactor pressure vessel werethe closure studs, instrumentation nozzles, and thecore flood nozzle venturi sleeve. Interestingly, theinlet and outlet nozzles were not included, nor werethe CRDM housings.

Therefore, in general, the regions of greatest con-cern for fatigue damage are:

1. Closure studs2. Outlet nozzles3. Inlet nozzles;'4. Instrumentation nozzles (penetrations)5. CRDM housings/penetrations (nozzles).

Comparing fatigue usage factor values for differentregions of a component or calculations betweendifferent vendors can be misleading because differ-ent simplifying assumptions are made in their com-putations. The ASME Code, Section 111, I specifiesthat the cumulative fatigue usage factor cannotexceed a value of one. Therefore, if a value less thanunity is obtained by making simplifying assump-tions to allow an easy calculation, there is no needto continue. Refined calculations can significantlyreduce the usage' factor values reported. Theserefinements include extensions from simple elasticcalculations to simplified' elastic-plastic calcula-tions and to highly detailed fully plastic calcula-tions. Other refinements in the way that transientsare grouped and counted can also make a signifi-cant change.

Corrosion fatigue can occur anywhere fatigue ingeneral is a consideration. Therefore, the sites indi-cated above are appropriate.

For wear problems, EW has identified thimbletubes as an area of concern.8 Evidently, this con-cern stems from the fact that some French reactorshave experienced wear problems induced by vibra-tion. Replacement appears to be an economicallyviable option if and when problems arise, although_W indicates that there is no significant effect on lifeextension. The other vendors did not indicate anyproblems relative to their equivalent of thimbletubes. Thus, wear is not considered to be a signifi-cant problem.

a. Private communication with E. Siegel, Combustion Engi-neering, 1986.

b. Private communication with S. L. Abbott, WestinghouseElectric Corporation. 1986.

c. Private communication with A. D. McKim, Babcock &Wilcox Company, 1986.

16

I

3.4 Degradation Mechanisms

The primary degradation mechanisms covered inthis section are irradiation damage (embrittlement)and fatigue (both classical design fatigue and crackgrowth using fracture mechanics). The effects ofenvironmentally assisted fatigue growth also arediscussed. The list of potential degradation mecha-nisms can be very large, because many mechanismsare always present but specific conditions suppresstheir direct influence. General corrosion and stress-corrosion cracking are not a problem in PWR ves-sels because the water is kept very clean (andessentially free of oxygen). Erosion and cavitationare not a problem because flow velocities are withinacceptable limits. High-temperature creep is not aproblem because the temperatures in the vessel are,well below one-half the melting point (on an abso-lute temperatures scale) of the materials. Thermalaging of cast stainless steels is not considered aproblem (CRDM housings), because thetemperatures are not sufficiently high- to promotelarge changes in properties. The following discus-sions reflect only those mechanisms that theauthors feel to be most significant.

3.4.1 Irradiation Embrittlement. Irradiation-induced increases in yield strength and decreases inthe fracture toughness of low alloy steels may limitthe operating life of PWR pressure vessels. Themajor variables controlling such "irradiationembrittlement" are the copper and nickel contentof the steels and the total neutron exposure orfluence. Other variables believed to be importantinclude the irradiation temperature, neutron spec-trum and flux, phosphorous content, thermome-chanical history, concentrations of some otherimpurity, and minor alloying elements. Indeed,embrittlement may be controlled by the particular,combination of these first- and second-order varia-bles; this .combination greatly complicates theproblem of developing reliable embrittlement pre--dictions based on a limited experimenital data base.'.

Embrittlement is normally monitored by testingsamples of base metal, heat-affected zone '(HAZ),and weld in the formn of CVN and tensile specimensthat have been irradiated in surveillance capsuleslocated near or adjacent to a vessel wall. This repre-sents an effort to come as close as possible t6 siniu'-lating the actual conditions experienced by the,vessel itself. Federal regulations 9 require that thesamples be'selected from the mosI limiting'mate-rial. There are usually several surveillance capsules

which are tested at specified intervals over the lifeof a vessel;'the neutron flux in the surveillance cap-

'sules is greater than that in the actual vessel by alead factor that is intended to provide an 'earlywarning of unacceptable levels of property degra-dation. The CVN specimens provide two measuresof embrittlement: changes in the' upper-shelfenergy (USE) and shifts in the transition tempera-

-ture at the 30 ft-lb (41 J) energy level,' AT. Theformer is used as a signal of significant degradationin the upper-shelf ductile toughness level based on

'a criterion of the USE falling below 50 ft-lb (68 J).The latter is used to shift RTNDT, the temperatureused to reference a lower bound toughness curve,i.e., AT = ARTNDT.

Excellent descriptions of most of the vendors'surveillance programs are presented inReference' 10. Certain nuclear power plants do nothave any active program because of the structuralfailure of capsule holders. One approach to dealingwith this situation is to use data from other plants,including irradiation of extra capsules in selectedhost reactors. The use of host reactors is anapproach used by B&W after the damage of all sur-'veillance capsule holder tubes in their reactors inoperation in 1976.11 There are two categories of

-plants having surveillance programs of limited-applicability because the materials in the programdo not adequately represent the limiting physicalproperties and chemical compositions of the RPVmaterials. In the first category of plants, the mate-rial in one representative utility surveillance pro-gram is actually part of the material from anotherutility's RPV. In the second category of plants,'there may be an undesirable weld material in anumber of reactor vessels that is reflected in thesurveillance program material of only one of thevessels. The plants in the above two categories maybenefit by the industry sponsored programs thatprovide the unified data from the pre- and postirra-diation testing of the specimens.

In general, however, estimates of USE changesand shifts in such circumstances are based on corre-lations (of the surveillance data base) that are con-tained in the USNRC Regulatory' Guide 1.99.12 Asecond revision of this regulatory guide has beenproposed and iscurrently out 'for' public com-'ment. 13 The differences between the two versionsof the Reguiatory Guide are- copper and 'nickelconcentrations are now taken into account for both'weld and base metal, a distinct margin is used toaccount for uncertainties in the data base, and thedamage change through the wall 'uses a displace-ments per atom (dpa) approach rather than taking

I

I

II

r.LI

17

I

fluence for energies > I MeV. RegulatoryGuide 1.99, Revision 2, provides methods for cal-culating numerical estimates of changes in USE asa function of copper and fluence, and shifts in tran-sition temperature as a function of copper andnickel content and fluence. Extrapolations of radi-ation damage into the vessel wall are based on aprototypical fluence and spectral gradient and theassumption that calculated dpa provides an appro-priate estimate of spectrum weighted radiationdamage exposure. Limitations of the application ofthe numerical estimates are specified for classesand-grades of steels, temperatures, and composi-tion ranges.

The consequences of the new regulatory guideaffect at least two areas: the pressurized'thermalshock screening criteria and heatup/cooldowncurves. With regard to the pressurized thermalshock issue, the higher magnitudes of the newshifts could potentially lead to a reduced residuallife in some utility vessels depending upon theirexact chemistry and fluence conditions. The lowerattenuation through the vessel wall calculated usingRegulatory Guide 1.99, Revision 2, than that cal-culated using Revision I means that the heatupcurves can now become more limiting.

A major source of uncertainty in current correla-tions is the limited accuracy and variable range ofthe surveillance data base. For example; there arelittle or no data for high-fluence and low-flux com-binations that may characterize the end-of-life orlife extension conditions of some vessels.

One method'of recovering CVN toughness andtensile properties is to perform postirradiation heattreating (annealing). This technique has been suc-cessfully applied to two vessels at temperatures lessthan or equal to 650 0F (3430C). 13 For higher tem-peratures [up to 850'F (455 0C)J, the feasibility ofthis technique has been shown analytically 14 but itrequires a major engineering effort to actually per-form the annealing and sophisticated calculationsof the postannealing reembrittlement rate. Thecomplexities in estimating irradiation' damage areincreased when the effects of an annealing areincluded. 15 Adjustments in surveillance testingalso are required. Even though annealing may bephysically possible, the economics have only dic-tated two attempts: the Army SM-IA vessel in1967 and the Belgian BR-3 vessel in 1984. 14 Both ofthese successful annealings 'were conducted at alower temperature than is' needed for commercialUnited States reactor pressure vessels. Therefore,more work leading to a demonstration of a higher-

temperature annealing for a full-size vessel isneeded.

There have been a very large number of acceler-ated materials test reactor (MTR) embrittlementexperiments over the last several decades. It is nowgenerally accepted that because of the high flux lev-els in these studies, the MTR data should be used asa supplement rather than as a direct substitute forsurveillance data. Most of this research has beenaimed at generating engineering data (e.g.,CVN shifts) for establishing selected effects such asthe influence of copper content on shifts or the'degree of recovery because of postirradiationannealing treatments.

Sustained efforts to establish'the underlyingmechanisms of embrittlement have been initiatedonly recently, primarily by groups in theUnited States, Great Britain, and Germany. How-ever, substantial improvements in our understand-ing of embrittlement have been achieved throughthese efforts. 16 -27 Indeed, this research hasresulted in significant improvements in the reliabil-ity of data correlations. For example, based on theguidance of early theoretical models, correlationsof weld surveillance data were developed withstandard errors of only .1607o of the mean shift; 15

more recent analysis has improved the physicalbasis for the models while maintaining a similarcorrelation accuracy for an expanded surveillancedata base for weldments. The basic insights fromthe recent studiesl7-19 can be summarized as fol-lows:

I. Embrittlement is primarily caused byirradiation-iniduced increases in the yieldstrength, which can be quantitativelyrelated to CVN USE changes and tempera-ture shifts.

2. Yield strength increases are because ofirradiation-induced, fine-scale microstruc-tures (.l nm) that act as obstacles to dislo-cation motion.

3. The most likely candidate microstructuresare precipitates, vacancy clusters (micro-voids), and interstitial clusters (dislocationloops). Copper-rich precipitates, possiblyalloyed with either elements such as man-ganese and nickel or vacancies, arebelieved to be a dominant hardening fea-tu're in sensitive steels containing signifi-cant 'concentrations of this impurityelement. Other, possible types ofirradiation-induced or -enhanced precipi-tates are phosphides and small carbides.

l

18

I'

I

4. The major effects of, irradiation' areenhanced diffusion rates' and defect clus-tering.

5. These microstructural evolutions arekinetic phenomena and are functions offlux and temperature, as well as composi-tion and microstructure. In general, thechanges in the microstructure can beunderstood from basic principles of alloythermodynamics and precipitation kinet-ics, coupled with rate theories'of radiationdamage. 17-20-

A number of detailed issues remain to beresolved, including: the composition of copper-rich phases and the balance between defect and pre-cipitate microstructures, the mechanisms of nickelenhancement of embrittlement, quantitative treat-ment of neutron flux effects, the role of factorssuch as other elements and thermomechanicaltreatment history, and the kinetics and mechanismsof postirradiation anneal recovery and reembrittle-ment. Also, the effects of such variables as flux,composition, 'and annealing' conditions onrecovery/reembrittlement need further resolution.

In principle, some other mechanisms may con-tribute to embrittlement in a synergistic,' additive,or competitive fashion. These include: thermalaging of ferritic materials caused by precipitation(which is not irradiation enhanced), segregation(temper embrittlement and strain aging), andhydrogen embrittlement/hydrogen attack. Forexample, there is some evidence that fine-scale cop-per precipitation can take place at vessel operatingtemperatures over periods on the order of 25,000 to250,000 h.17 Hydrogen effects could be promotedby circumstances that may or may not be present inPWR environments, 'related to 'nonequilibriumpartial pressures or local chemical reactions thatyield high local hydrogen fugacities.

There are'a number of'other unresolved ques-tions. For example,' the accuracy of using CVN-based RTNDT shifts to reference toughness curves(ASME Code' Section Ill, Appendix G) has notbeen demonstrated: Theoretical consideratioiis'suggest that in'some sensitive'steels this practicemay be nonconservative, and that relatively low.cleavage fracture toughness levels may persist up toelevated temperaiures in steels subject to very largeirradiation-induced yieldI strength inicreases.16 In'other words, the actual RT1DT shifts in some pres-sure vessel steels may be significantly larger thanthe CVN-based RTNDT shifts. Indeed, the serious-'ness of this'issue may be mitigated only by the'

uncertain conseri'atisrri of using RTNDT-referencedlower bound toughness curves. (It should be notedthat some work is being pursued within the HeavySection Steel Technology (HSST) Program spon-''sored by NRC to actually measure shift and shapechange in the Kic and Kia curves from large-specimen fracture toughness tests).28 Further, noreliable basis exists for characterizing irradiationeffects on ductile upper-shelf static fracture tough-ness and crack arrest toughness from surveillancedata. All these properties may be required in a com-prehensive integrity analysis.

Finally, this section has not covered the topics ofdefect monitoring and evaluation or application ofthe material properties in fracture mechanics analy-ses, including the treatment of elastic-plasticeffects an'd multidimensional part-through flawssubject to combinations of primary and secondarystresses (which may be rapidly time varying). Infor-mation on the effects of irradiation on additionalproperties of ferritic steel, such as fatigue crackgrowth rates (including the effect of the environ-ment), stress corrosion cracking, strain rate effects,strain hardening exponents, and various measuresof ductility are generally required to carry out suchanalysis. Indeed the effect of radiation damage onthe overall constitutive behavior and macroscopic/microscopic strain distribution may be important.It is noted that if large increases in yield strengthare correlated with large decreases in cleavage frac-ture toughness, a greatly expanded regime of brittlevessel operation might be anticipated. 17

3.4.1.71 Uncertainties. The prior discussionsets the context for uncertainties in characterizingthe material state as a function of service life and

,history for use in analysis of pressure vessel safetymargins. Summarizing, uncertainties are relatedto:

1. The accuracy of either plant specific orcorrelation based estimates of CVN transi-'tion temperature shifts for specified valuesof thedominant variables

2. The existing surveillance programs do notinclude the effects of high-fluence and'low-flux conditions, 'and treat the effectsof spectrum and irradiation temperature,approximately.

3. The actual versus nominal value of signifi-cant variables in the most critical locationin the vessel

4. The questionable validity of using refer-ence toughness curves and shifting these

1.

II

I

19

I

curves by transition temperature increasesmeasured in CVN tests

5. The essential lack of information on anumber of properties, such as static upper-shelf toughness (which may be needed inan integrity analysis in general);' and aneven greater lack of information on theeffects of in-service irradiation on theseproperties. In addition, there are uncer-tainties in the initial properties such as theunirradiated RTNDT.

6. Gaps in basic understanding and detailedempirical characterization of many aspectsof embrittlement, postirradiation anneal-ing, and reembrittlement phenomena.

While this admittedly incomplete list of technicaluncertainties is imposing, it may not be all that sig-nificant. Currently, it is simply not possible toassess the combined effect of the issues listed aboveon the general reliability of the integrity assess-ments. However, a change in attitude about how totreat the aggregated uncertainties,'or discovery ofnew issues (e.g., an unanticipated accidentsequence) could change the assessment of residuallife far more drastically than'would be expectedfrom further resolution of some technical issueslisted above (e.g., the accumulation of additionalCVN surveillance data or the identification-of anew second-order variable or embrittlement mech-anism). Clearly, the magnitudes of the technicaluncertainties increase with the length'of the pro-jected time frame; and may be paramount to theissues of relicensing and life extension.

3.4.1.2 Practical Issues. There are a numberof practical issues that constrain the range ofapproaches for resolving the uncertainties indi-cated above. Some of the technical issues are beingaddressed by the USNRC research29 and others,but it is not totally clear that all issues'are integratedrelative to life extension considerations. Some ofthe pertinent practical technical issues include:

1. The limited range of potentially importantvariables and properties contained in theexisting data base and the reliability ofsubsets of that data base.' ' ''

2. The limited amounts of critical materialand limited numbers of surveillance cap-suies. (For example, if plant life extensionis anticipated, how should this influencethe schedule and approach to testing sur-

veillance capsules? Or, what can be doneafter all the capsules are used up?)

3. The limited amount of property informa-tion that can be obtained from currentCVN-based surveillance programs relativeto the data needed for comprehensiveintegrity assessments.

4. The flexibility (or inflexibility) of existingregulation and ASME code requirements.

5: Questions regarding the suitability of theuse of MTR irradiations to simulate radia-tion damage under service conditions; andthe validity of using (prototypical) surro-gate alloys to represent the behavior of ves-sels materials (e.g., if archival samples arenot available).

6. Lack of a pertinent data base to assess thepotential for a long-time, irradiation-induced aging effects in ferritic steel.

3.4.1.3 Issues to be Resolved. Clearly, themost pressing need is the development and timelyexecution of a comprehensive, integrated, and bal-anced plan for dealing with the array of issues asso-ciated with residual life assessment and lifeextension of reactor pressure vessels. The NRCMaterials Branch has such a plan underway for thereactor pressure vessel, i.e., the HSST Program.30

Timely execution of this plan will allow resolutionof several issues associated with aging of the reactorpressure vessels. It is not' appropriate to attempt todesign such a plan here; however, it is useful to elab-orate on the approach and to identify a few keyelements that can be expected to emerge from adetailed study;

1. Comprehensive-After identifying the keyvariables, develop a means of obtainingand validating the necessary array of prop-erties for alloys and conditions (i.e., theirradiation, thermal and chemical environ-ment) reasonably representative of therealities of long-term in-service behavior.

2. Integrated-The experimental effort toobtain appropriate properties in suitablyconditioned alloys should be closely tied tothe overall basis for integrity analysis. The'integration with the mechanics componentof the integrity analysis should probablyanticipate significant advances in compu-tational capabilities. On one hand, suchdevelopment of computational capabili-ties may permit the further exploitation of

Iii

I

I

20

I

I

miniature specimen testing and reconsti-tuted specimen testing techniques '(seeChapter 14 for additional discussion),greatly extending the potential sources ofinformation about embrittlement; and, onthe other hand, these advances will permitmore direct and reliable application offundamental property data to practicalstructural analysis. Significant improve-ments in the basic theoretical understand-ing of the mechanisms of embrittlementand techniques to characterize theembrittling microstructures (as well asproperty changes) should also be antici-pated.

3. Balanced-The efforts should be carefullybalanced, because the impact of a fewweak links in the sequence of analysis willfar outweigh high precision in one oranother step in the process. For example, itwould do little good to develop methods toaccount for each and every neutronimpinging on a surveillance capsule or ves-sel if the irradiation temperature or alloymicrostructure and composition were stillpoorly characterized.

Some of the important specific elements in this pro-gram are listed below (the HSST program has someactivities planned in many of these elements):

1. Rapidly resolve critical outstanding issues,e.g., the potential for high-temperaturecleavage fracture in sensitive steels withhigh levels of irradiation strengthening -

2. Comprehensively define the property arrayneeded to carry out reliable integrity analy-sis, including ranking of the overall sensitiv-ity to uncertainties in the properties

3. Develop improved ways of extracting anarray of mechanical property informationfrom material in' surveillance-programs,e.g. miniature specimen techniques

4. Validate available RT:NDT shift data fromaccelerated MTR irradiations'as a' basis forestablishing long-term in-service 'behaviorfor RPV steels. These data will supplementthe surveillance data from the power reactors

5. Hold additional surveillance specimens inpower reactor internals above'or below thecore as well as on the sides to get the appro-priate flux and fluence levels, and tempera-tures

6. Initiate`sorme long-time studies of otherenvironmental effects (e.g. thermal aging),and their potential synergistic interactionswith irradiation embrittlement

7. Resolve postirradiation annealingrecovery/reembrittlement kinetics andmechanisms

8. Optimize surveillance programs with long-term objectives clearly defined

9. Continue efforts to develop fundamentalunderstanding of embrittlement and pos-tirradiation annealing recovery/reembrittlement phenomena in terms of

- physically based models and techniques tocharacterize the state of the 'alloy (such assmall-angle neutron scattering).

I

3.4.2 Fatigue. The fatigue analysis of Class Icomponents in nuclear plants is 'defined by theASME Boiler and Pressure Vessel 'Code, specifi-cally in Section Ill for the fatigue design evaluationprocedurel and in Section XI for the flaw accept-ance standards and evaluation procedures. 31 TheSection Ill procedures are based on the classicalstress-strain/life (S/N) approach, using an experi-mentally determined relationship between the elas-tically calculated "stress range and fatigue life.Under variable amplitude cyclic loading, a Miner'srule approach of accounting for damage accumula-tion in the component at different stress ranges isused by linearly summing the fraction'of life con-sumed at each stress range with the cumulativeusage factor set at less than, or equal to, unity. Sec-tion XI procedures, conversely, rely on fracture:mechanics-based, damage-tolerant analyses,' wherecontinued operation of components containingdefects can be permitted in-service, provided pre-dictions of fatigue-crack growth show that thedefect will remain subcritical until the next inspec-tion. .

-From the perspective of improved life-prediction; rocedures, the Section III nd Xl approaches maybe open to question. First, they represent funda-mentally different'fatigue approache's (one basedprimarily on crack initiation and the other based oncrack growth)' which'may cause problenii of con-sistency wheri applied to' the same'.component. 2

'Second, with regard to the S/N analysis, no allow-ance is made for the sequence of loading cycles,

'with'the Miner's 'linear damage' law, and furtherseparate analyses are required for high- and low-cycle fatigue. Third, the fatigue-crack-growth,damage-tolerant analysis does not fully account for

'such factors as near-threshold behavior, variable

1.

I

I

I

21

[

I

amplitude loading, and the anomalous behavior ofsmall flaws (typically of a size < I mm).33

Described 'next are various corisiderations thatmay be used to improve the life-prediction capabil-ity of the Section 111 and Xl fatigue design analy-ses. (Note that the NRC Pressure Vessel ResearchProgram is conducting work on both environmen-tal fatigue-crack-growth behavior and cumulativefatigue-damage modeling.) Such advanced lifetimecomputation procedures are designed for life exten-sion and rely on:

I. A modified S/N procedure for crack initi-ation based on the local-strainapproach, 3 4 which incorporates theeffects of both stress concentrations atnotches and the sequence of variableamplitude cycles

2. An updated version of the damage-tolerant procedures that incorporate animproved appreciation of environmentalfactors, low growth rate behavior, and therole of small flaws.

3.4.2. 1 Modified Life Prediction Approachfor Crack Initiation. The local strain-approachrepresents an advanced version of the classical S/Napproach, and provides a more consistentapproach to fatigue life prediction, based primarilyon crack initiation. Specifically, it incorporatesimproved analyses of the roles of mean stress andstress concentrations and in the summation ofcumulative'damage. 3 4 The analysis is essentiallybased on total strain, and so as such, does notrequire any distinction between high- and low-cyclefatigue. Typical S/N analyses generally define low-cycle fatigue (LCF) as failure in the'ran'ge of'102-105 cycles' (depending upon the maiterial) atstresses near or above yield, and high-cycle fatigue(HCF) as failure at a high number of cycles undernominally elastic stresses. The Coffin-Manson lawcan be used for life prediction in LCF, and the Bas-quin relationship can be used in HCF.3 5

The first commonly used equation for the local-strain approach is the experimentally-determninedcyclic stress-strain curve, which relates cyclic stressrange (Ao) to 'strain range (Ae) in terms of Young'smodulus (E) and the cyclic strain hardening expo-nent (n'), and is written

Ae AO /aa \In'_2 = _2E + ~Vk') (3.1)

where k' is the cyclic-strength coefficient repre-senting the slope of the log stress/log plastic straincurve, and n' is typically 0. 15. Use of this flow ruleallows a far better estimate of actual materialbehavior (e.g., cyclic hardening or softening). Thesecond commonly used equation is the basic stress-strain/life equation, which relates total strain range(Ae) to life (Nr) by incorporating modified versionsof the Coffin-Manson law for LCF, the Basquinrelationship for HCF, and the Goodman relation-ship for mean stress (or) into a single equation thatgives3 4 ' 3 5

1.

Ae = ar_ a 2__ +Er_ 2

2 Eo + f)(3.2)

where af and ef are the fatigue-strength and fatigue-ductility. coefficients, respectively (approximatelyequal to the true fracture strength and fracturestrain in a tensile test), and b and c are the fatigue-strength and fatigue ductility coefficients, respec-tively (with approximate values of -0.1 and -0.6). Itshould be noted here that no fatigue limit isassumed.

Equation (3.2) is also used for the prediction offatigue life in the presence of stress concentrations,where the values of Aa and AE now define the localstresses and strains at the notch root. Instead ofemploying the classical approach of computingthese local stresses merely in terms of the theoreti-cal stress concentration factor (K,) or the fatiguenotch strength reduction factor (Kr), Neuber'srule,36 or equivalent numerical calculations, can beused to compute the local strain history. For exam-ple, Neuber's rule gives

I

I

Kr = (KOKE )I/ (Aa Ae E)AS

(3.3)

where K0is the actual stress concentration (ratio oflocal to nominal stress, Aav/AS) and K, is the actualstrain concentration (ratio of local to nominal totalstrain, Ae/Ae). Such an approach can result inmuch improved life prediction, particularly inlower strength steels where local yielding at thenotch results in significant strain concentration aswell as stress concentration.

For irregular cyclic loading conditions, a modi-fied Palmgren-Miner linear-cumulative-damagerule can be employed to allow for consideration ofthe sequence of variable amplitude loading cycles.(Sequence effects, overstressing, and understress-ing are not accounted for in classical S/N

22

analyses.) The basic linear-cumulative-damage'la'w' 3.4.2.2 Modified Life Prediction Approachcan be stated as ' ' ''for Crack Propaigation. The flaw-evaluation

-procedures in Appendix' A of Section XI of the'.' ASME Code3 ' use linear elastic fracture mechan-

= usage factor (3.4) ics to predict the growth of existing flaws underi "'e-cyclic loading. For these procedures to be effective,

crack extension rates (da/dN) must be predicted onwhere n1 is the number of cycles of a given stress the basis of a realistic crack growth 'relationship,range (Aat), Nr, is the number of cycles to failure 'which generally has the form39had all the cycles been of magnitude (au1), and the'usage factor is set not to exceed unity. In the modi-' da F ' ' '- (35fied approach, which has been shown to yield supe- dN = f (ak, Kmax* (rior fatigue-life'estimates with'complex loading'histories,37 the primary input is to account for' - where Kmax and AK'are the maximurn value andmean stress and sequence effects by using a moret' range of he stress intensity factor ( - 'aappropriate means for' cycle counting. There are F is the frequency, and C and m are'experimentallyseveral procedures that have been used with '' -"determined constants appropriate to thermaterialsuccess, but the so-'called rainflow method is gener- - the frequency, and C and m are experimentallyally the most widely used.37 This method involves: 'determined constants appropriate to the' material

3 ' '- ' ' ' '' ' ' '''' intermediate cyclic crack growth rates and can' be1. The determination'of the 'number of corm- used with altered (realistic) constants' for some por-

plete hysteresis loops in the loading spec '. 'tions outside the intermediate'crack growth range.trum .' - ' ' ' ' .The approach of James and Jones40

2. Measurement of'a mean-and alternating stainless steels represents an ideal example. Thesestress (or strain) for each loop authors employed an extensive data base for stain-

3. Computation of the value of Nf for each less steels in order to apply least squares regressionloop from Equation (3.2) '' techniques to derive a power-law, crack-growth

4. Calculation of the usage factor, using relationship [Equation (3.5)] for use in' life calcula-Equation (3.4) in the usual manner. tions. However, because in this particular case no

'attempt was' made to allow for environmentalBecause these procedures require a great deal of effects, the"equation can only be used for cracktedious bookkeeping to implement, several'- 'growth behavior in air." 'algorithms have beerf published to reduce computa- The approach of integrating an appropriate.,tion time.38 , crack growth'relationship 'from initial to final

Finally, as with all fatigue analyses, it is vital to defect size41 is' generally considered to be in'her-consider the effect of service environment,-as it is ently conservative because it assumes negligible ini-'well known that fatigue life can be drastically tiation life. However, based on recent studies ofaltered by the presence of an active environment variable-armiplitiide, near-th eshold, 'and small-(such as PWR water). This generally entails defin- crack behavior (e'j., Refereii6es 33,' 42, 43), 'thereing the life equation [Equation (3.2)] in simulated are several factors that should be considered addi-service environment with loading applied at a real- tionally to improve the accuracy of life prediction. [istic frequency, 'if possible.'' ' ''; ' '' 'These factors are discussed later.'

Compared to classical S/N analyses,' the local-"' *strain approach has 'two principal 'advantages. ' '; -

First, life predictions f6r a wide ivariety'of situa- '' "3.4.2.3 'ita Base' 'Establishmnent. Cracktions can be made'from a limited amount'of small- ''pr6pagation' data should be generated experimen-specimen test data. Secorid,' it provides a rationial' 'tally in as realiitienvironments and frequencies asbasis for considering various material'belhaviotr 'possible, over 'i' ikide spectrum of growth rate's;.related interaction 'effects that 'are -importanti in ' Sufficient data must be obtained to g'uaranitee con- 'i

metal fatigue."For' these ieasons,-the approach' 1 " fidenceintheres~Iltingeqiations. Dataintheultra-' ' '

yields superior life pOedictions compared with clas- low 'growth-rate regime below 10.6 mm/cycle,sical analyses;, 'and'could be successfully used to 'approaching the so-called fatigue threshold, 'arebetter define usage'factors for life-extension con-'' pairticularly important as'fatigue' lifetimes; espe-siderations. ''. i!: ' " - ' cially those of practical importance to' life

23

I

I

extension, can be dominated by crack extension inthis region. For example, using a linear extrapola-tion of higher growth rate data (above 10-5 mm/cycle 4 l may lead to heavy penalties on predictedlife and indicate component retirement long beforeits safe useful end-life is reached). The need forultralow growth rate data is further amplified byestimates of reactor component service loads. Forexample, service loads at the Zion-I pressurizedwater reactor plant have been estimated to includeover 2 x 1010 cycles of vibrational stress cycles [ofmagnitude typically 1000 psi (6.89 MPa)1, 3 2

which over the 40-year life of a particular compo-nent could lead to extensive crack growth at theseemingly insignificant 'near-threshold growthrates. Because near-threshold growth' rates can beseverely retarded by the phenomenon of crack clo-sure42 in laboratory tests (e.g., using comjpact-typespecimens) at low load ratios, it is essential thatsuch data: be determined at high' load'. ratiosabove 0.5 to generate a conservative'crack-growthrelationship (see also small crack discussed later)..

3.4.2.4 Variable Amplitude Behavior.Unlike crack initiation under irregular'cyclic load-'ing, as described by Miner's law [Equation (3.4)],crack growth under variable amplitude loads can besubjected to severe load interaction effects, particu-larly for long cracks under broad band spectrumloading.4 4 Such interaction effects generally takethe form of transient' crack-growth retardations,following, say, single tensile overload'cycles orhigh-low block loading sequences, which contrib-ute significantly to enhanced fatigue lifetime.45

Although physical explanations for suclibeliaviorare still mechanistically uncertain, the aerospaceindustry (in particular) has been successful in devis-ing several modified damage-tolerant analyses thatcan account, (albeit empirically) for the increasedlife caused by such transient behavior. The' mostwidely used models are those of Wheeler 46 andWillenborg et al.,4 7 although-nore sophisticatedapproaches based on numerical modelin'g'ofcirackclosure have recently been developed.4 8 Theapproach in all these models is essentially to. useconstant amplitude data to predict variable ampli-.tude behavior through modification of the integra-

- tion of the crack growth relationship, e.g., throughdetermining a retardation factor on crack growthwhile the crack is propagating within the: plasticzone of the overload cycle. Incorporation of suchprocedures in the life prediction analyses may, wellcontribute to increased, life (which correspondsessentially to decreased usage factors).

3.4.2.5 Small-Crack Effects. One further.factor, which should be considered with advanceddamage-tolerant life prediction procedures, is anappreciation of the possible anomalous behavior of.small flaws. Cracks are considered small when theyare of a size comparable with the scale of the micro-structure, or the scale of the local plastic zone size,or when they are simply physically small (i.e.,<I mm).33 In general, their behavior is to displaygrowth rates in excess of those of conventional longcracks (e.g., over 10 mm in length), when exposedto the same stress intensity range (although mildreverse effects have been reported). In addition,small crack growth can occur below the fatiguethreshold, below which long cracks are presumeddormant. Moreover, under severe environmentalconditions, the accelerated growth behavior ofsmall cracks may be further enhanced.49 Short ofgenerating experimental data on such small cracks,damage-tolerant procedures to account for suchbehavior have yet to be developed. However, sincemechanistically the behavior of small flaws is pri-marily influenced by a diminished role of crack clo-sure, by generating long crack data at high loadratios, some of the discrepancy between long- andshort-crack growth rates can be diminished.

In summary, it is felt that by incorporatingadvanced S/N analyses to replace the typicalapproaches outlined in Section III of the code, andby incorporating revised damage-tolerant, crack-growth procedures to supplement the Section XIprocedures, significantly improved life-predictioncapabilities can be obtained. Such approachescould further be embellished by consideration ofthe small-crack issue, which represents the inter-face between crack initiation and crack propaga-tion analyses.

3.5 Potential Failure Modes

The eventual failure modes resulting from degrada-tion during service life will generally emerge when asevere abnormal transient or event occurs. For exam-ple, in the reactor beltline region, if a pressurized ther-mal shock event were to develop and the materialproperties were already severely degraded by irradia-tion embrittlement (i.e., the RTNDT was raised to hightemperatures), the potential for failure of the vessel in abrittle manner is possible. However, there must also bea defect or crack of critical size present to produce thisfailure scenario. The possibility of an, irradiation-induced low upper-shelf material in the beltline regioncould lead to a low-energy ductile tearing overload

I

I

I

I

rL

24

:I

I

situation, if again a crack is present. Therefore in thebeltline region, the necessity of having a defect is para-mount for producing failure. However, the regions ofinterest in the beltline have extremely low fatigue usagefactors indicating the srmall probability of having agrowing defect or crack. Defects are possible'as evi-denced by the concern over underclad cracks in the1970s;50,51 these manufacturing-induced cracks canresult from high heat input multipass strip claddingtechniques (at the overlap points), especially forSA508-2 forging steel. The United States processestypically use a single- or double-layer submerged arcpass;- and there is no embrittlement of the HAZ, lead-ing to cracks. The United States vessels have not shownin-service evidence of underclad cracking. These smallcracks, even if present, would not pose an integrityproblem because a series of them would have to linktogether to form a continuous defect of any significantsize. 6

In the areas where fatigue is a potential problem,the type of final failure will almost always resultfrom a plastic overload situation caused by thestress concentration and reduced ligament load car-rying ability (i.e., the resultant material left fromthe fatigue crack tip to the surface.) These ductiletearing conditions are not necessarily failires, butare potential through-wall leakers. However, theinlet nozzles could be exposed to cold water duringsome accident conditions that could lead to a brit-tle fracture there (especially if some irradiationdamage also was present). The CRDM housingsand instrumentation penetrations would probablybe exposed to only ductile overload situations, andthe closure studs also would be exposed to onlyoverload conditions resulting in ductile failure forconditions caused by internal pressure, boltuploads, and some thermal transients.

In planning for extended life, the closure studsshould be replaced near the end of the design life ofthe vessel because of the high fatigue usage. TheCRDM housings, instrumenitation penetrations,and outlet/inlet nozzles should be, monitoredclosely and will most likely limit the vessel extendedservice life before repair or replacement is required.The beltline region may be most limiting, but eachindividual vessel will require independent assess-ment depending on material chemistries, weldtypes, and fluence conditions.

3.6 In-service Inspection and:Surveillance Methods

In-service inspection (ISI) is. required by theASME Code, Section XI.31 The general philoso-phy lof Section XI ISI and the status of theimprovements in the code requirements for theRPV are presented in Chapter 12. There are basi-cally four inspection intervals, and certain welds'must be examined during those intervals. All shell,head, shell-to-flange, head-to-flange, and repairwelds must be subjected to a lOO1o volumetricexamination during the first inspection interval(over 3 to 10 years). Successive inspection intervalsrequire fewer beltline region, head, and repair weldexaminations. The nozzle-to-vessel welds must allbe subjected to a volumetric examination during allfour inspection intervals. Twenty-five percent ofthe partial-penetration nozzle welds (CRDM andinstrumentation) are required to have a visual,external surface examination during each inspec-.tion period (leading to total coverage of all noz-zles). All of the nozzle-to-safe end butt welds withdissimilar metals (i.e. ferritic steel nozzle to stain-less steel or Inconel) must be subjected to volumet-ric and surface examinations at each interval. Thesewelds are covered in Chapter 5 (primary systempiping) of this report. All studs and threaded studholes in the closure head studs need surface andvolumetric examinations at each inspection inter-val. Any integrally welded attachments are requiredto have surface (or volumetric) inspections of weldsat each inspection interval. A brief description ofthe currently used volumetric examination methodsis given in Chapter 11.

Thus, the inspection plan is very complete andresults in close monitoring of potentialfatigue-crack formation and growth. The additional moni-toring and recording of transients are usually donein accordance with the plant technical specifica-tions; however, this area could be greatly enhancedby more accurate monitoring and assessment of thetransients (identified earlier) leading to fatigueusage.

The surveillance program for monitoringchanges in the vessel wall and weld propertiescaused by neutron embrittlement has been dis-cussed in Section 3.4. Coupling these surveillance

r

I

I.

I

25

I

I

results (and existing data) with the accurate charac-terization of flaws or defects: through ISI wouldallow prudent calculation of vessel integrity. If bet-ter logging of the critical fatigue transients also isundertaken, even further confidence in the reliabil-ity of the pressure vessel is possible. More confi-dence now in the safety of PWR vessels will extenddirectly into later life extension.

3.7 Summary, Conclusions, andRecommendations

A summary of the important degradation sites,mechanisms, and consequences for PWVR vessels isshown in Table 3.2. Note that the ranking of sites isbased on the importance of that area, not on the possi-bility of that site having severe degradation. For exam-ple, the closure studs are probably the most likely site tohave significant damage (except for possibly the reac-tor beltline region), but the studs are listed at the endbecause they are easily replaced. Other conclusions andrecommendations are:

I. Radiation embrittlement is the primarysafety concern that may limit vessel lifeextension; flux reduction programs used toreduce fluence levels need to becoordinated with overall life extensionplans; Charpy referencing of toughnessneeds to be resolved, fundamental studiesof irradiation damage and further work onphenomenological correlation modelsshould be developed for both transitiontemperature shift and upper shelf energydrop (rate effects need to be better under-stood and correction factors developed);surveillance program schedules may needadjustment and revision to meet life-extension goals.

2. Thermal annealing needs to be pursued as acontingency plan (a long-term option);annealing mechanisms and kinetics need tobe better understood and models developed.

3. Application of miniature specimen testingand reconstituted specimen testing need to

be evaluated. Use of these testing tech-niques is valuable because the availablespace to irradiate the test specimens andthe available test material are limited. Thisis discussed further in Chapter 14.

4. Other synergisms associated with irradia-tion may be present and need to be studiedalong with some simple long-time thermalaging studies, the total package of infor-mation needed for structural integrityanalysis should be reviewed and uncertain-ties quantified.

5. Nozzles' and penetrations represent.,themajor fatigue problems associated withlife extension; better monitoring of fatiguetransients is needed to properly assessactual usage factors; more sophisticatedfatigue usage factors can be developedusing the local-strain approach presentedearlier; as actual usage factors becomelarge (i.e,'approach unity), fatigue-crackpropagation calculations using the state ofthe art should be used to predict usablelife.

6. Flange closure studs should be replacedwhen considering life extension; also, thethimble tubes in W plants should beinspected and assessed for replacement.

7. Inspection and transient logging records aswell as' all material properties (unirra-diated and surveillance results), designdata, and other key information, shouldbe maintained for future reference; earlycollection of these records will allow lifeextension plans to be developed.

8. It is probable that the reactor pressure ves-sel, in most cases, can be used for longerthan the 40-year license period dictated byfederal regulations. 9 Actual ,decisions areplant specific, however, due to slightly dif-ferent designs, different limiting materials,and operating conditions (fluence levelsand transient history). Continued closemonitoring of the material condition in

the vessels is required for life extension.

I

I

r

26

I

Table 3.2. Summary of degradation processes for PWR reactor pressure vessels

Rank ofDegradation

SiteDegradation

Site Stressors Degradation Mechanisms Potential Failure Modes, ,I ;

I Beltline region Neutronirradiation,mechanicaland thermalstresses

* Irradiation embrittlement(degree is dependent onindividual vesselmaterials and fluxspectrum history)

I , fI ,

Ductile high-energytearing leading toleakage (net sectionover-load)

ISI Surveillance Methods

100% volumetric duringfirst inspection; oneweld for subsequentinspection

Surveillance program forassessing irradiationdamage is required by law

ifBrittle fracture (i.e..

- ' pressurized thermalshock)

, Ductile low-energytearing (low upper-shelf toughness)

Environmentalfatigue

Ductile overload leadingto a leak; possible brittlefracture if PTS occurs

rL

2 Outlet/inlet Mechanicalnozzlesa and thermal

stresses

Fatigue crack initiationand propagation

, Fatigue crack initiationand propagation

Ductile overload leadingto a leak; possiblebrittle fracture ifpressurized thermalshock occurs'withsome irradiationembrittlement

Ductile overload leadingto a leak

3 Instrumentationnozzles(penetrations)and control roddrive mechanismhousing nozzles

Mechanicaland thermalstresses

All nozzle welds inspectedvolumetrically at eachinterval

Visual, inspection ofexternal surface; 25%o ofnozzles inspected at firstinterval; remaining 755ospread out over next threeintervals

Volumetric and surfaceinspection of all studsand threads in flange studholes at each interval

I

4 Flange closure Mechanicalstuds ' Iand thermal

stresses

Fatigue crack initiation Ductile overload failureand propagation (possibly (can be replaced)corrosion assisted)

L

a. Welds at the vessel to safe end are covered in Chapter Son primary system piping; the comments here are also applicable exceptthat volumetric and surface'examinations ire required. '

I

E

27

3.8 References

1. American Society of Mechanical Engineers, ASMDE Boiler and Pressure Vessel Code, Section 111,Nuclear Power Plant Coinponents, Division 1, New York, 1983.

2. K. Hoge, Evaluation of the Integrity of Reactor Vessels Designed to ASME Code, Sections I and/orVIII, NUREG-0031, NRC-I, U.S. Nuclear Regulatory Commission, 1976.

3. J. J. Koziol, "Radiation Surveillance Program for Combustion Engineering, Inc. Reactor VesselMaterials," in Status of USA Nuclear Reactor Pressure Vessel Surveillance for Radiation Effects,L. E. Steele (ed.), ASTM Special Technical Publication 784, Philadelphia: American Society forTesting and Materials, 1983, p. 216.

4. NV. R. Corwin, R. G. Berggren, and R. K. Nanstad, "Charpy Toughness and Tensile Properties of a |Neutron-Irradiated Stainless Steel Submerged-Arc Weld Cladding Overlay," Effects of Radiation onlMaterials: Twelfth International Symposium, WVilliamsburg, Virginia, June 18-20, 1984, ASTM Spe-

cial Technical Publication 870, American Society for Testing and Materials, 1985, pp. 951-971.

5. G. T. Yahr, R. C. Gwaltney, A. K. Richardson, and NV. L. Server, "Case Study of the Propagationof a Small Flaw Under PWR Loading Conditions and Comparison with the ASNIE Code DesignLife: Comparison of ASME Code Sections I11 and Xl," ASMEPressure Vessel and Piping Confer-ence, Chicago, July 20, 1986. C

6. T. J. Griesbach, "Section XI Future Trends in Equipment Requalification of Nuclear Components,"ASMfE Pressure Vessels and Piping Conference, San Antonio, Texas, June 17-21, 1984, PVPVolume 90.

7. NV. H. Bamford, "Technical Basics for Revised Reference Crack Growth Rate Curves for PressureBoundary Steels in LWR Environment," Journal of Pressure Vessel Technology, 102, 1980, p. 433.

8. Seminar on Nuclear Power Plant Life Extension, Seminar Proceedings, cosponsored by EPRI,Northern States Power, U.S. DOE, and Virginia Power, Alexandria, Virginia, August 25-27, 1986.

9. Code of Federal Regulations, Title 10-Energy, Part 50-"Domestic Licensing of Production andUtilization Facilities."

10. L. E. Steele (ed.), Status of USA Nuclear Reactor Pressure Vessel Surveillancefor Radiation Effects,ASTM Special Technical Publication 784, Philadelphia: American Society for Testing and Materi-als, 1983.

r11. A. L. Lowe, K. E. Moore, and 1. D. Aadland, "Integrated Reactor Vessel Material Surveillance

Program for Babcock and Wilcox 177-FA Plants," Effects of Radiation on Materials: Twelfth Inter-national Symposium, Williamsburg, Virginia, June 18-20, 1984, ASTM Special TechnicalPublication 870, American Society for Testing and Materials, 1985, p. 931.

12. U.S. Nuclear Regulatory Commission, "Effect of Residual Elements on Predicted Damage to Reac-tor Vessel Materials," Regulatory Guide 1.99, Revision 1, 1977.

13. U.S. Nuclear Regulatory Commission, "Effects of Residual Elements on Predicted Radiation Dam-age on Reactor Vessel Materials," Regulatory Guide 1.99, Revision.2, February 1986.

14. NV. L. Server, In-Place Thermal Annealing of Nuclear Reactor Pressure Vessels, NUREG/CR4212,EGG-MS-6708, 1985.

28

15. W. L. Server and A. 7Iboada, "An Approach for Estimating Post-Anneal Reirradiation Embrittlement ofReactor Vessel Steels," Proceedingsof the Second International Symposium on Environmental Degradation ofMaterials in Nuclear Power Systems- Water Reactors, Monterey California, September 9, 1985, AmericanNuclear Society, 1986, p. 496. -

16. G. R. Odette and P. M. Lombrozo, Physically Based Regression Correlations of Embrittlement DataFrom Reactor Pressure Vessel Surveillance Programs, EPRI NP-3319, 1984.

17. G. R. Odette, "Recent Advances in Predicting Embrittlement of Reactor Pressure Vessel Steels,"Proceedings of the Second International Symposium on Environmental Degradation of Materials inNuclearPowerSystems- WaterReactors, Monterey, California, September9,1985, American NuclearSociety, 1986, p. 121.

18. G. R. Odette, "On the Dominant Mechanism of Irradiation Embrittlement of Pressure VesselSteels," ScriptaMetallurgica, 17, 1983, p. 1183.

19. G. R. Odette and G. E. Lucas, "Irradiation Embrittlement of Reactor Pressure VesselSteels: Mechanisms, Models and Data Correlations," in Radiation Embrittlement of Reactor Pres-

,, sure Vessel Steels: An International Review, ASTM Special Technical Publication 909, L. E. Steele(ed.), American Society for Testing and Materials, 1986, p. 206.

20. G. E. Lucas and G. R. Odette, "Recent Advances in Understanding Radiation Hardening andEmbrittlement in Pressure Vessel Steels," Proceedings of the Second International Symposium onEnvironmental Degradation of Materials in Nuclear Power Systems-Water Reactors, Monterey,California, September 9, 1985, American Nuclear Society, 1986, p. 345.

21. G. R. Odette, P. M. Lombrozo, and R. A. Wullaert, "Relationship Between Irradiation Hardeningand Embrittlement of Pressure Vessel Steels," Effects of Radiation on Materials: Twelfth Interna-tional Symposium, Williamsburg, Virginia, June,. 18-20, 1984, ASTM Special TechnicalPublication 870, American Society for Testing and Materials, 1985, p. 840.

22. G. E. Lucas, G. R. Odette, P. M. Lombrozo, and J. W. Sheckherd, "Effects of Composition,Microstructure and Temperature on Irradiation Hardening of Pressure Vessel Steels," Effects ofRadi-ation on Materials: Twelfth International Symposium, Williamsburg, Virginia, June 18-20, 1984,ASTM Special Technical Publication 870, American Society for Testing and Materials, 1985, p. 900.

23. G. E. Lucas, G. R. Odette, R. Maiti, and J. W. Sheckherd, "Tensile Properties of Irradiated Pres-sure Vessel Steels," Proceedings of the Thirteenth International ASTM Symposium on the Effects ofRadiation on Materials, Seattle, Washington, 1986. - .

e-b~,'ThEe of *rr a resr

24. G. R. Odette, G. E. Lucas, and T. Griesbach,"TheEffectof Irradiation and Aging on PressureVessel Steel Embrittlement," Proceedings of the International Conference on Nuclear Plant Aging,Availability Factor, and ReliabilityAnalysis, San Diego, California, July 7,1985, American Society forMetals, 1986, p. 375.

25. G. E. Lucas and G. R. Odette, "Application of Small Specinene Test Techniques to Power PlantReliability and Life Extenision 'Pr6blemis," Proceedings of the International Conference on NuclearPlant Aging, Availability Factor, and Reliability Analysis, San Diego, California, July 7, 1985,American Society for Metals, 1986, p. 385. .

29

I

26. J. T. Buswell, E. A. Little, R. B. Jones, and R. N. Sinclair, "Analysis of Microstructural Changesin Irradiated Pressure Vessel Steels Using Small Angle Neutron Scattering," Proceedings of the Sec-ond International Symposium on Environmental Degradation of Materials in Nuclear Power Systenms-WHater Reactors, Monferey, California, September 9-12, 1985, pp. 139-152.

27. P. A. Beaven, F. Frisius, R. Kampmann, and R. Wagner, "SANS/TEM Studies of the Defect Micro-structures of Test Reactor Irradiated Fe-Cu Alloys and Cu Containing RPV Steels," Proceedings ofthe Second International Symposium on Environmental Degradation of Materials in Nuclear PowerSystems- Water Reactors, Monterey, California, September9-12, 1985, pp. 400-408.

28. Vessel Integrity Review Group Meeting, Review of NRC Pressure Vessel Research Program, Oak RidgeNational Laboratory, December 1986.

29. M. Vagins and A. Taboada, Research Program Plan: Reactor Vessels, NUREG-I 155, Volume 1, U.S.Nuclear Regulatory Commission, 1985.

30. Engineering Technology Division, Heavy-Section Steel Technology Programn-Five-Year Plan,FY 1985-1989, NUREG/CR-4673, ORNL/TM-10108, August 1986.

31. American Society of Mechanical Engineers, ASME Boiler and Pressure Vessel Code, Section XI,Rulesfor In-Service Inspection of Nuclear Power Plant Components, New York, 1983.

32. G. T. Yahr, R. C. Gwaltney, A. K. Richardson, and W. L. Server, Case Study ofthe Propagation ofa rSmall Flaw under PWYR Loading Conditions and Comparison with the ASME Code Design Life,NUREG/CR-3982, ORNL-6099, 1984.

33. S. Suresh and R. 0. Ritchie, "Propagation of Short Fatigue Cracks," International Metals Review,29, 1984, pp. 445476.

34. N. E. Dowling, NV. R. Brose, and W.- K. Wilson, in Fatigue under Complex Loading-Analysis andExperiments, Warrendale, Pennsylvania: Society of Automotive Engineers, 1977, p. 55. E

35. M. R. Mitchell, "Fundamentals of Modern Fatigue Analysis for Design," in Fatigue and Microstruc-ture, Metals Park, Ohio: American Society for Metals, 1979, pp. 385437.

36. H. Neuber, "Theory of Stress Concentration for Shear-Strained Prismatical Bodies with ArbitraryNonlinear Stress-Strain Law," Journal ofApplied Mechanics, Transactions of theAmerican SocietyofMechanical Engineers, 28, 4, 1961, pp. 544-550.

37. N. E. Dowling, "Fatigue Life and Inelastic Strain Response under Complex Histories for an AlloySteel," Journal of Testing and Evaluation, 1, 4, 1973, p. 271. L

38. R. M. WMetzel (ed.), Fatigue under Complex Loading: Analyses and Experiments, Warrendale,Pennsylvania: Society of Automotive Engineers, 1977.

39. P. C. Paris and F. Erdogan, "A Critical Analysis of Crack Propagation Laws,"Journal of BasicEngineering, Transactions of the American Society of Mechanical Engineers, 85, 4, 1963, pp. 528-534.

40. L. A. James and D. P. Jones, "Fatigue Crack Growth Correlations for Austenitic Stainless Steels in MAir," in Predictive Capabilities in Environmentally Assisted Cracking, Proceedings of the Winter UAnnual Meeting of the American Society of Mechanical Engineers, Miami Beach, Florida, November17-22, 1985, PVP Volume 99, p. 363.

30

41. AU. A. Van Der Sluys, Corrosion Fatigue Characteristics of Reactor Pressire Vessel Steels, EPRI NP-2775,1982.

42. S. Suresh and R. 0. Ritchie, "Near-Threshold Fatigue Crack Propagation: a Perspective on theRole of Crack Closure," in Fatigue Crack Growth Threshold Concepts, D.L. Davidson and S. Suresh F(eds.), Warrendale, Pennsylvania: The Metallurgical Society/American Institute of Mechanical LEngineers, 1984, pp. 227-261.

43. J. Schijve, in Fatigue Crack Closure, ASTM Special Technical Publication, J. C. Newman andNv. Elber (eds.), American Society for Testing'and Materials, in press.

44. W. Elber, "Significance of Fatigue Crack Closure," in Damage Tolerance in Aircraft Structures,ASTM Special Technical Publication 486, American Society for Testing and Materials, 1971,pp. 230-242.

45. C. M. Ward-Close and R. O. Ritchie, "On the Role of Crack Closure Mechanisms in Influencing FatigueCrack Growth Following Tensile Overloads in a Titanium Alloy: Near-Threshold vs. Higher AK Behavior,"ASTM International Symposium on Fatigue Crack Closure, Charleston, South Carolina, May 1986.

46. 0. E. Wheeler, Spectrum Loading and Crack Growth, New York: American Society of MechanicalEngineers, 1971.

47. J. Willenborg, R. M. Engle, and H. A. Wood, A Crack Growth Retardation Model Using an Effec- Ltive Stress Concept, AFWPL Report No. AFFDlTM-71-1-FBR, Wright-Patterson Air Force Base,Dayton, Ohio, 1971.

48. J. C. Newman, "A Crack-Closure Model for Predicting Fatigue Crack Growth under Aircraft Spec-'trum Loading," in Methods and ModelsforPredictingFatigue Crack Growth under Random Loading,ASTM Special Technical Publication 748, Philadelphia: American Society for Testing and Materi-als, 1981, p. 53.

49. R. P. Gangloff and R. O. Ritchie, "Environmental Effects Novel to the Propagation of Short [Fatigue Cracks," in Fundamentals of Deformation and Fracture, Eshelby Memorial Symposium,.B. A' Bilby, K. J. Miller, and J.'R. Willis (eds.), Cambridge, England: Cambridge University'Press, 1985, pp. 529-558.

50. A. G. Vinckier and A. W. Pense, A Review of Underclad Cracking in Pressure Vessel Components,WRC Bulletin 197, Welding Research Council, New York, 1974.

51. A. Disooge et al., "A Review of Work Related to Reheat Cracking in Nuclear Reactor Pressure Vessel 'r

Steels," International Journal of Pressure Vessels and Piping, 6,5, 1978, p. 329.

., ., ,, : .~~~~~~

. . . . . . . .~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~L

31

I

4. PRESSURIZED WATER REACTOR CONTAINMENTS AND BASEMATSM. A. Daye

4.1 Description of ContainmentStructures

The majority of the pressurized water reactor(PWR) containments in the United States are pre-stressed concrete containment vessels (PCCV) withsteel liners covering the inside surfaces. Reinforcedconcrete containment vessels (RCCV) with steel lin-ers are also commonly used. Other types of con-tainments constructed as steel cylinders or steelspheres were used for some PWRs.

The function of the containment structure is tohouse the nuclear steam supply system (NSSS) andlimit the release of radioactive material to the environ-ment during normal operation and in the event of acci-dents. In particular, the containment structure providesthe much needed shielding against radiation in theevent of an accident so that necessary emergency mea-sures can be provided to minimize the consequencesand deal with the causes in a protected environment.The structure also provides protection to the reactorsystem from the effects of severe and extreme environ-mental loads such as tornadoes, external accidents,and penetration by moving objects.

Containment structures are designed to with-stand design earthquakes, loss-of-coolant accidentpressures and temperatures, and operating loadsand temperatures for a continued, predesignatedlife span of normally 40 years.

4.1.1 Prestressed Concrete Containment Ves-sels.The principal components of a PCCV arereinforced concrete, steel liner plate, and postten-sioning wire or strand tendons..

In general, the containment structure consists of aflat reinforced concrete slab, an upright posttensionedconcrete cylinder and posttensioned concrete dome.All internal surfaces are lined with steel plates anchoredto the concrete. Figures 4.1 (a) and 4.1 (b) represent twoconfigurations for the PCCV containments.

4.1.2 Reinforced Concrete Containment Ves-sels. The RCCVs are basically similar to the pre-stressed concrete containment vessels except thatthe cylinder and dome are not posttensioned andrely completely on the reinforcing steel and con-crete for load resistance (Figure 4.2).

4.1.3 Steel CylinderlSteel Sphere ContainmentVessels. Steel cylinder containments consistmainly of a freestanding steel cylinder with a steeldome on top and an ellipsoidal steel bottom that isanchored to a concrete foundation. The inside ofthe steel bottom is filled with concrete to approxi-mately the transition line between the ellipsoidalbottom and the base of the cylinder. Within thebottom concrete inside the containment, the reac-tor cavity is formed to house and support the reac-tor vessel. The freestanding cylinder and dome areenclosed inside a reinforced concrete reactor build-ing that forms the biological shield around the con-tainment (Figure 4.3). The PWR containmentswith ice condensers, i.e., Sequoyah-I and -2, andWatts Bar-I and -2, have this type of containment.Some of the other plants having this type of con-tainment are Indian Point-I, St. Lucie-l and -2,Prairie Island-I and -2, and Davis-Besse-l.

The steel spherical containment structure is simi-lar in concept to the freestanding steel cylinderexcept that the cylindrical portion is omitted and acomplete sphere is used. The bottom of the sphereis anchored to a concrete foundation. The inside isfilled with concrete to a height sufficient to providestructural stability and form the reactor cavitywithin. There is one commercial PWR plant thathas this type of containment structure in the UnitedStates, San Onofre-I. Yankee-Rowe, a 175-MIVedemonstration plant, also has a steel spherical con-tainment and it is supported by concrete/steelcolumns through the lower hemisphere.

4.1.4 General Discussion. Exterior and internalwalls and floors of the containment structure areinvariably of thick concrete primarily for shielding pur-poses. Occasionally, high-density aggregate such asilmenite, is used to provide additional shielding with-out increasing the thickness of the concrete members.

The base slabs for both PCCV and RCCV con-tainment structures are essentially the same andconstructed of reinforced concrete. Some varia-tions in the geometry of the base slabs may existdepending on the NSSS system used. The slab iseither completely flat or has a discontinuity wherethe reactor vessel is located in a cylindrical reactorcavity compartment. formed into the slab.

Posttensioning systems consist of high-strengthsteel wire or strand tendons, end anchorages, andcorrosion protection filler material. The tendons

L

L_

7,L.

32

I

200 ft.

I

I

L

: r,

Figure 4.1(a). PWR prestressed concrete containm'ent Type l and Typel11.

:~~~~~~ I

i - .

i,

L

33

I

210 ft.

U

rU

Secondary shieldingand pressure boundary

S 11 BPimay shieldin

' (~~~~~~ - D ~~~~~~It.~~a~: -10 ft.

7-3054

Figure 4. 1 (b). PWR prestressed concrete containment Type I I l-three buttresses and hemispherical dome.Li

34

E'

I

..

-Secondary shieldingand pressure boundary

V'O ft.

-10 ft.

7-3076

I

Li

- . I . l | } , l j .; .~~~~~.

-. ~~Figure 4 2. PWR reinforced concrete containrnent.

.. , . ^ , .A. , ,,i.

35

I

Primary shielding,

Figure 4.3. PWR steel cylinder containn

are placed inside ducts within the containment walland dome, which were formed during construc-tion. The tendons provide prestressing forces in twoorthogonal directions within the curved shell struc-ture. For types I and II prestressed containments,[both types are shown in Figure 4.1 (a)], the tendonanchorages are provided at the top and bottom ofthe cylinder for the vertical tendons, at six (Type I)or three (Type 11) equally spaced buttresses on thecircumference of the cylinder for the hoop tendonsand around a ring girder at the top of the cylinder

4 ft annular space

ndary Secondary shielding

v85 f t.

'v 0 ft.

7.3041

with reinforced concrete reactor building.

for the dome tendons. In Type Ill containments,[Figure 4.1 (b)], the tendon anchorages are providedat the bottom of the cylinder and the vertical ten-dons extend from one side of the cylinder over thedome and down the opposite side. Tendon anchor-ages also are provided at three equally spaced but-tresses on the circumference of the cylinder andextending vertically into the dome for the hoop ten-don in the cylinder and the dome. All tendons andanchors are encased in a corrosion protective mate-rial and tightly capped at the ends.

7.

lent

Li

36

I

The PWR containment structures are designed in,accordance with the American Society of Mechani-cal Engineers (ASME) Section III, Division 2 codeor equivalent requirements approved by theUSNRC.I Material used in the construction of the'containment structure is controlled by' specificrequirements specified by the American Society forTesting and Materials (ASTM) and other industrystandards to guarantee a quality product. Qualitycontrol and quality assurance programs are set to'control all steps of design, production, and con-struction of the containment.

The typical values of design/capability pressuresand free volumes for the three PWR containmentsare as follows:

Design/Capability Free VolumeContainment Pressure (10 E+03

(plant) (Psig) cubic feet)

PCCV (Zion) 47/134 2,600RCCV (Surry) 45/119 1,800Steel cylinder (Sequoyah) 12/50 1,200

4.2 Containment Environmentand Degrading Factors

Containment structures are subjected to a num-ber of stressors that influence the characteristicsand physical properties of the structural and non-structural components of the containment. In' gen-eral, exposure can be categorized as external,caused by the outside environment, and internal,caused by the environment created by the operatingsystems inside the containment building. An addi-tional deteriorating factor unique to concrete is theadverse chemical reactions 'in concrete that are'related to certain characteristics of the concrete.mixture and its constituents.

4.2.1 External Environment. Exposure to 'theexternal environment affects mainly the concreteand in advanced cases of cracking and spailling it'can possibly affect the embedded reinforcement.The adverse effects are due to wetting and drying,and freezing aind thawing cycles.'Acid rain, carbon-ation, salt spray on containment structures locatednear the sea, and harmful sulfates in ground wateraffecting the concrete foundation are the majordegradation factors for both prestressed and rein-

forced concrefe containments, and for concretereactor buildings and steel shells in the case of steelcontainments.

The degree 'of deterioration caused by each ofthese factors varies, depending on the intensity ofeach and on the quality of the concrete and the cor-rosion 'protection system for the steel.

The' external environment also affects the post-tensioning system anchorage components and ten-dons mainly in' the form of corrosion or hydrogen

',embrittlement caused by-either rain or groundwater seeping in, through the' concrete' or endanchorage caps, to the tendon ducts. The differentdegradation mechariisms are in most cases relatedto moisture conditions and accumulation of freewater near the anchors.

These effects and degradation mechanisms willbe discussed in detail in Section 4.4 of this report.

4.2.2 Internal Environment. 'The effects of theinternal environment can be categorized asmechanical deterioration because of vibration

'loads, high temperatures, and fatigue stress anddeterioration resulting from exposure to moistureand nuclear heating'. The internal environmentstressors affect both steel and concrete componentsin different degrees, depending on the severity ofthe stressor and the proximity of the components tothe source of deteriorating factor. The posttension-ing system' is only slightly affected by the internal'environment because of its location, embedded inthe concrete wall. However, the steel liner may suf-fer from corrosion caused by the hot, moist atmo-sphere' inside the containment, coupled with anacidic environment.

4.2.3 Chemical Reactions in 'Concrete. Asidefrom the 'exterril and internal environmentaleffects, concrete can be subject to adverse internalchemical reactions in the form of alkali-aggregatereactions, carbonate-aggregate reactions -orcement-aggregate reactions resulting from adversecharacteristics of the concrete mix ingredients. 2

Outside factors such as moisture and temperatureconditions can enhance such reactivities. Reactivityin concrete may not occur or be noticeable for along time'after concrete' placement and 'culd 'varyin the degree of seriousness from insignificant scal-ing to complete spalling and deterioration.

4.2.4 Degradation Environment During Con-tainment Service. Degradation factors'also canbe categorized as continuously acting stressors dur-ing normal operation of the plant, stressors

L

L

I

E.

37

r

I

occurring for short durations during accidents ormalfunction of certain components, and as stres-sors naturally acting during the life of the plant.

Stressors during normal. operation consist, ofradiation and nuclear heating, causing strengthdegradation and changes in physical and chemicalproperties of steel, concrete, and other materialssuch as gaskets and wires. Thermal stressors causedby elevated temperature, vibrating loads causingfatigue stresses, moisture and condensation caus-ing corrosion, and acidic environments causingconcrete and metal deterioration also exist duringnormal operation. In addition, normal operatingprocedures and scheduled pressure testing and refu-eling shutdowns can result in load and temperaturecycling leading to fatigue and ratcheting effects.Naturally occurring stressors such as freeze andthaw, wetting and drying, acid rain, etc., may beconsidered as stressors occurring during normaloperation. The degree of deterioration that mayoccur depends on how well concrete and othermaterials are constructed. The stressors that occurduring accidents or malfunction of certain equip-ment can cause similar effects to those present dur-ing normal operation but at an accelerated rate andhigh magnitude that can possibly result in over-stress conditions.

Structures are designed to a specified service life.Under normal conditions, most structures meetsuch a design criteria. However, beyond the designlife span, degradation may alter the function, char-acteristics, and physical properties of componentsof the structure. Thus, reevaluation and possiblyreanalysis of the load-bearing capability of thestructural members becomes necessary. using theproperties of the members at the time of the reanal-ysis. Main stressors affecting the structures beyondthe design life are those causing concrete deteriora-tion, metal fatigue, and corrosion.

4.3 Degradation Sites

The degradation sites are best identified by exam-ining each PWR containment type separately.However, some of these sites may be common tomore than one containment type. The degradationsites for each containment type are identified inorder of importance to the functionality and safetyof each structure.

4.3.1 Prestressed Concrete Containment Ves-sels. The major structural components for thePCCV are the posttensioning system for the wall

and dome, and the reinforced concrete wall, dome,and base slab as discussed above. The steel linerplate covering the internal surfaces of the contain-ment, although it is not a structural load-bearingelement, is important in providing a leak-tight bar-rier that will prevent radioactive releases.

Because the wall and dome are prestressed andprestressing induces a state of compression in boththe steel liner and concrete, stress cracks in the steelliner are less likely to occur in this type of contain-ment, and the possibility for release of radioactivematerial through the shell under normal operationis highly unlikely. Corrosion may develop in theliner or liner welds resulting in through-wall cracks.However, as long as the concrete is still in compres-sion, the danger of radioactive material release issmall. Therefore, degradation in the shell steel linercannot be considered to be as important as loss ofstress in the prestressing tendons. On the otherhand, corrosion or cracks in the steel liner over thebase slab can be important because radioactivewater or other liquids may seep through the baseslab and contaminate the ground water in the vicin-ity of the containment structure. Degradation inthe base slab liner could become more significant ifcracks were to have developed in the slab as a resultof a seismic event or any other reason.

Loss of the prestressing stress (below the limitsaccounted for in the design) can be caused by brokenwires or strands, cracked anchor heads resulting in par-tial or complete release of the tendons, excessive wireor strand relaxation or creep, and/or shrinkage in theconcrete. Although the possibility of anchorage failureis remote, it has happened at various plants, after vary-ing lengths of service times and different environments.This type of degradation can have important conse-quences because accident pressures inside a contain-ment may cause the concrete shell and steel liner toenter a state of tension, at which time any cracks in thesteel liner may possibly permit leakage of radioactivegases to the outside environment.3 Therefore, the fol-lowing degradation sites for prestressed containmentare identified and ranked in the following order ofimportance:

I. Loss of prestressing force caused by:a. Loss of tendon anchorageb. Loss of tendon wires or strandsc. Excessive tendon wire or strand relax-

ationd. Excessive concrete creep and shrinkage

2. Corrosion of the reinforcing steel embed-ded in the concrete

r

38

I

3. Cracks or corrosion in the steel liner ovejthe base slab

4. Deterioration of the tendon corrosion pro.tection material

5. Cracks or corrosion in the steel liner oveithe walls and dome

6. '.Deterioration in the concrete caused by theoutside environment or internal reactions.

4.3.2 Reinforced Concrete Containment Vessel. RCCVs rely on reinforcing steel in the con.crete to resist tensile stresses. With the absence of aprestressing system, the state of stress in the stee'liner and concrete is not always in compression.Tension zones exist both in the wall, dome, and irthe base slab. Therefore, cracks in the steel lineican be significant. Although seepage of radioactivematerial through cracks in the base slab is of firsiimportance; leakage of radioactive gases throughcracks in the steel liner on the wall and dome arealso important. Because the tensile loads are car.ried by the reinforcing steel, corrosion of the rein.forcement steel is of concern. Cracks and 'highporosity in the concrete cover over the reinforcingsteel may permit moisture and oxygen to permeateand enhance the corrosion process. Therefore,good quality concrete is essential for both protec.tion and strength. The possible degradation sitesfor the reinforced concrete containment can, there.fore, be ranked as follows:.

l. Corrosion of the reinforcing steel in thewall and dome

2. Corrosion of the reinforcing steel in theslab

3. -Cracking and deterioration in the wall aniddome concrete caused by'the outsidelenvi.ronment and internal chemical reactions

4. Cracks'or corrosion in the steel liner'oveithe base slab

5. Cracks or corrosion in the steel liner oveithe walls and dome

6.' Deterioration inthe base slab cbncrete' caused by groundw'ater sulfate and inter-nal chemical reactions.

4.3.3 Steel CylinderlSteel Sphere Containmen,Vessel. For this type of containment, the prime protection against leakage of radioactive gases'is achievecby the same steel shell that also provides the structurabarrier' against pressure buildup' inside the contain

ment. Therefore, the most important degradation sitesare associated with the shell itself. Degradation can be

- in the form of cracks or corrosion, either in the basemetal or weld seams of the steel shell. Because the bot-tom part 'of the containment is embedded in concreteand the rest of the shell is freestanding, there exists aregion of discontinuity atfthe top of the concrete where

' stress concentrations and vibrating loads in the'steelshell may cause fatigue cracking. In addition, the sameregion is also susceptible to either stress-enhanced orgalvanic corrosion. Other stress concentration zonesexist in this type of containment in all other areas ofdiscontinuities such as hatches and penetrations. Cor-rosion of the bottom portion of the containmenf that is'embedded in the concrete is likely to occur and might

* eventually lead' to escape 'of radioactive fluids to theground outside the boundaries of the containment.The corrosion of this embedded portion could possibly

e be the result 'of attack by sulfates in the ground water,which reach the steel by penetrating through the out-side concrete.

Elevated temperatures affecting the steel shell,'either in areas near high-temperature equipment orat penetrations for high-temperature mechanicalpiping, can cause deterioration in the strength andother physical properties of the shell material.

The degradation sites for the steel cylinder/steelsphere containments can be ranked as follows:

I . Cracks'or corrosion in the welds or basemetal of 'the 'shell because of vibratingequipment and aggressive environments'

2. Stress cracks or corrosion in the shell at theconcrete' embedment interface at the baseof the containment

3. Stress cracks or corrosion in the shell at dis-continuities near'h'atches and penetrations''

4. Corrosion of the bottom steel of the shellembedded in concrete

5.'- Deterioration and cra'cks in the base slabconcrete inside the containment

6. Regions of elevated temperature 'on thesteel shell.'

4.4 Degradation Mechanisms

The discussion of the 'potential degradationmechanisms is separated into 'three parts.Section 4.4.1 deals exclusively with the 'effects of

i ; radiation and nuclear heating on the physical prop-I' ' erties of concrete. This 'is -the only phenomena

unique to structures that house nuclear reactors

L

Il

FII

39

I

and associated systems. Several other phenomenathat affect all concrete structures and are notunique to nuclear power plants are described inSection 4.4.2. For this reason, there is an abun-dance of research and literature on degradationcaused by temperature, freezing-thawing cycles,sulfate attack, corrosion of the reinforcing mate-rial, etc. The potential degradation mechanismsassociated with the various steel components arediscussed in Section 4.4.3.

4.4.1 Effects of Radiation on Concrete.Because of the dual role of nuclear power plantconcrete as a structural support and radiationshielding material, an evaluation of extended serv-ice life must address the continued effectiveness ofthe concrete in both its functions. From the stand-point of shielding, water is the most desirable ingre-dient in concrete, because water's high hydrogencontent makes it an effective neutron moderatorand attenuator. Because concrete generally servesas a structural support as well as a shield, the initialwater content is maintained at a minimum becausethe structural strength of the concrete decreaseswith increasing initial water content: Thus, in theinitial selection of the concrete constituents, shield-ing design considerations must be balanced withstructural considerations. Similarly, the effects ofaging and the various concrete degradation mecha-nisms must be evaluated in light of both shieldingand structural requirements..

The most significant radiation degradationaspect of concrete results from nuclear heating.Nuclear heating produces an increase in concretetemperature causing evaporation of the free waterin the concrete, 5 that is, water that is not chemi-cally bound to the hydrated cement. Both the struc-tural and shielding properties of concrete aresomewhat. affected by increased temperatures andloss of the free water.

As concrete temperatures increase, decreases in com-pressive strength, tensile strength, bonding strength,and the modulus of elasticity have been observed. 6 '7

For gamma-ray shielding, the water loss as a result ofnuclear heating has little impact because of the rela-tively small change in density associated with the loss.Because of the role hydrogen plays as a neutron moder-ator, the water lost at high temperatures can alter theneutron attenuation properties of concrete.. However,the amount of the water loss during a 40- year operating,life is not expected to be substantial enough to impairstructural or shielding properties during extended oper-ation.7

In contrast to the potential degradation of con-crete resulting from nuclear heating and water loss,direct radiation damage is a less serious concern.This is because the fluxes that can cause direct radi-ation damage to concrete are higher by orders ofmagnitude than those that cause damage fromnuclear heating. Hungerford7 suggests that even afast neutron exposure of 1021 nvt (n/cm2) has littleapparent effect on concrete aside from reducing itsrupture strength by a factor of 2. Figure 4.48shows the axial distribution of the fast neutron fluxat the boundary of the reactor pressure vessel for atypical PWR. Because of attenuation, only theconcrete closest to the surface directly exposed tothe source is expected to approach a fast neutronexposure of 1018 nvt, which is well below expecteddamage levels.

Evidence of radiation-induced dissociation of thewater (in small concrete samples) into hydrogen andoxygen, accompanied by gas evolution has also beenreported. 7 This phenomenon is not expected to depletethe concrete's water content significantly, and, is mostlikely to occur near (i.e., within A.l-1/2 to 2 feet of) thesurface adjacent to the source.

Concrete samples obtained from the NuclearWaste Tank Farm built in 1953 (Rockwell HanfordOperations, Richland, Washington) have beentested for the effects of long-term exposure to tem-perature and radiation.9"0 ' 1 1 The results indi-cated limited reduction in strength and modulus ofelasticity but did not indicate any reduction in theintegrity of the concrete. Thus, even in extendedplant operation, degradation of concrete becauseof radiation is not expected to be life limiting.

4.4.2 Effects of Other Concrete DegradationMechanisms. The degradation of reinforcedconcrete can be described in terms of elementsaffecting the deterioration of the concrete and thesubsequent corrosion of the reinforcing steel afterits protective cover is degraded. The potential deg-radation of concrete can be divided into two cate-gories. The first category of potential degradationis related to the initial material selection, propor-tioning and mixing procedures, and placement andcuring of the concrete. The second category ofpotential degradation is related to the environmentto which the concrete will be subjected during thelife of the plant.

4.4.2.1 Initial PotentialDegradation of Con-crete. This category of potential degradation isinfluenced by such factors as physical properties of

I

I

3

40

U'

I

C.a)

0C0,

U)0

I-L

r~

L

0 50 100 150 200 250 300

Distance (cm) 7.3053

Figure 4.4. Isoflux contours for neutrons with energy > I MeV in the PWR (R,Z) model (neutrons/cm2 s).9

41

I

the aggregate, water-cement ratio, type and quanti-ties of cement, shrinkage and settlement cracking,temperature effects related to heat of hydration,and adverse chemical reactions. These potentialdegradations can be controlled by the proper choiceof materials, mix ratios, and curing procedures.However, if any of the key parameters affecting theconcrete integrity are not handled properly duringmixing and placing, the quality of the concrete canbe reduced which will increase the potential forlater degradation.

These degradation modes usually result first insurface cracking of the concrete. Cracking of con-crete can occur while the concrete is in either a plas-tic or hardened state. The presence of cracksnormally does not affect the ability of the structureto carry load, but can affect structural durability byexposing the embedded reinforcements to a corro-sive environment. Excessive shrinkage resulting incracks of unacceptable sizes usually results fromuse of materials and techniques that are not inaccordance with good engineering practice and thatwould not pass standard construction inspections.

4.4.22 Environmental Degradation of Con-crete.

1. Chemical Reactions. A number of delete-rious chemical reactions can produce con-crete cracking. These generally are relatedto the concrete aggregate andinclude: alkali-aggregate reactions,cement-aggregate reactions, and the reac-tion of carbonate aggregates. Sulfate-bearing waters also present problems, butby attack of the cement paste rather thaninvolvement of the aggregate.a. Alkali-Aggregate Reactions., Cracking

can occur as a result of expansive reac-tions between aggregates containingactive silica and alkalies usually derivedfrom cement. Problems with structuresbuilt since about 1950 have been signifi-cantly reduced through proper selectionof aggregate materials (petrographicexamination to identify potentially activematerials), use of low alkali cements(<0.6qo equivalent NaO), and additionof pozzuolanic materials that restrictexpansion.

b. Cement-Aggregate Reactions. Con-crete deterioration attributable tocement-aggregate reactions has beenobserved with some sand-gravel aggre-

gates. These materials are highly sili-ceous and produce map cracking inconcrete. This type of distress is notprevented by the use of pozzuolana orlow-alkali cements. Tests have beendevised to indicate potential damagefrom this phenomenon.2 A recom-mended remedy, if these materialsmust be used, is to dilute the aggregatewith crushed limestone.

c. Carbonate-Aggregate Reactions. Cer-tain dolomitic limestone aggregatescontaining some clay react with alkalisto produce an expansive reaction. Thealkali-carbonate reaction can be con-trolled by keeping the alkali content ofthe cement low (<0.4%) or by dilut-ing the reactive aggregate with a less-susceptible material.

d. Sulfate-Bearing Waters. The sulfatesof sodium, potassium, and magne-sium have caused deterioration ofmany concrete structures thoughchemical reaction of the sulfates withthe hydrated lime and hydrated cal-cium aluminate in the cement paste.Calcium sulfate and calcium sul-foaluminate are formed along withconsiderable expansion that disruptsthe concrete. Resistance of concrete tosulfate attack can be improved by theuse of admixtures such as pozzuolanaand blast furnace slag or the use ofsulfate-resisting cements (Types V,VP, VS, IP, or IS).12

2. Elevated-Temperature Effects. Elevatedtemperature has an important effect onconcrete structures because it affects con-crete's compressive strength, stiffness, anddurability.

A review of research and testing per-formed on concrete at elevated tempera-tures indicates that temperature effects area function of a number of factors, some ofwhich are related to the type of aggregateand mix proportions and others of whichare related to the limits and duration of the'temperature exposure. Other importantfactors influencing the results of testedconcrete are the temperature of the con-crete at the time of test and the conditionof the concrete during heating (whethersealed or open). Thermal cycling is alsoimportant because it can cause cracking

I

G.

r

I'U

42

I

and separation between' the concretematrix and aggregate and between the con-crete and the reinforcing steel.

The effects of the type of aggregate oncompressive strength do not become sig-nificant until the temperature reaches%600'F (3160C). Neville2 reported thatthe compressive strength exhibits a slightincrease between 400 and 600'F (204 and3160 C) (Figure 4.5). The modulus of elas-ticity is reported by Neville to drop withincreased temperature, Figure 4.6.

Bertero and Polivkal3 concluded thatmoisture in sealed specimens has a detri-mental effect on concrete exposed to hightemperature. They reported a 71 % and5SIo reduction in compressive strengthand elastic modulus, respectively in sealedspecimens exposed to 3000F (1490C).However, other specimens from the sameconcrete experienced insignificant reduc-tions in mechanical properties when mois-ture was allowed to escape during thethermal exposure.

3. Wetting and Drying. In structures-withcracks, improperly treated constructionjoints, or areas of segregated or porousconcrete, water may enter and exit. Aswater passes through the concrete, it dis-solves (leaches) some of the calciumhydroxide and other solids so that in timeserious disintegration of the concrete mayoccur. As the concrete dries, salts (sulfatesand carbonates of sodium, potassiun, orcalcium) are deposited through evapora-tion of the water or interaction with car-bon dioxide in the atmosphere. Extensiveleaching' causes an increase in concreteporosity, which leads to reduced strengthand increased vulnerability to other corro-sive environments.

4. Freezing and Thawing. Repeated cyclingof wet concrete between freezing and thaw-ing can lead to severe damage or destruc-tion of concrete structures. The principalforce responsible for concrete damageresulting'from this phenomenon is internalhydraulic pressure created-by the expand--ing ice-water system during freezing. Con-struction of highly freeze-thaw resistantstructures requires attention to these fac-tors: (a) use of air entrainment, (b) selec-tion of aggregate with durability adequate

for the' proposed exposure, and (c) use ofinitial low water-to-cement ratio concrete.

5. Cracking. Cracks in concrete surfaces canbe indicative of normal processes such asdrying shrinkage or of unusual (andunwanted) processes such as foundationsettlement, structural overload, chemicalreactions, and other degrading processes.Because the opening of cracks allowsaccess by hostile environments to reinforc-ing steel and other embedded items, it isimportant to investigate cracks of signifi-cant size.

6. Aggressive Chemical Environrnent.Aggressive environments of high tempera-ture, moisture, and acid-saturated fluidsor gases are common in power plants.Such environments are detrimental to bothconcrete and steel containment types.When concrete is exposed to an acidic envi-ronment, a reaction occurs between theacids and the calcium hydroxide of thehydrated portland cement. Some of theacid reactions result in water-soluble cal-cium compounds that are then leachedaway by aqueous solutions. In the case ofsulfuric acid, the deterioration is morepronounced because the reaction productoccupies a larger volume than the acid andthe calcium hydroxide, resulting in spall-ing and cracking in the concrete.

7. Corrosion of Embedments. Cracking andspalling of concrete can result from the corro-sion of embedments, primarily mild steelreinforcing bars. Concrete cracking and spall-ing results from the buildup of corrosionproducts that can act as a Wedge to delami-nate and spall the adjacent concrete.

Continuous exposure of aluminum tomoist concrete produces severe corrosionbecause of the destruction of its passive filmat high alkalinities. Lead and zinc also 'cor-rode in moist concrete, with zinc being moreresistant than aluminum or lead. In all thesecases, the corrosion process is aggravated inthe presence of chloride ions. -

8. Vibration and Structural Loads. Concreteand steel can both fail in fatigue. After a fewvhundred thousand reversals at a significant

- stress level, the concrete strength can bereduced to half its original value. The effectsof fatigue on concrete are more pronounced ifthe stress reversals alternately place the con-crete in tension and compression.2 ,4 The

I

;.

L

L.

43

I

Temperature (OF)1000 1200 1400 1600 1800

'aL-

a)

C

0a)

Cq

a)Ua)Cu

-cCn

I

L

20 100 200 300 400 500 600 700 800 900 1000Temperature (@C) 7-3052

Figure 4.5. Compressive strength of concrete after heating to high temperatures. 3

Temperature (OF)-300 .200 -100 0 100 200 300 400 500 600 700

151 1 1 | I I I I I I |

C.)

c)Cu

'5 1a)-a

0E 05at

:0co

a:

-200 -100 0 100 200 300Temperature (*C)

Figure 4.6. Influence of temperature on modulus of elasticity of concrete.3

400

7-3051

Fn

Li

44

U.

I

durability of the concrete' can be improved lated at the lower end of the vertical tendons in thesignificantly if the stress reversals are such area of the anchorage. This problem has beenthat the concrete always remains in compres- encountered in Types I and 11 postiensioned PWRsion. containments, Figure 4.1(a), where the configura-

Excessive 'vibration can lead to cracking tion of the vertical tendon anchorage possiblyof the concrete. The presence of such allows some rain water ponding around the domecracks may expose the reinforcing steel to to seep through the top and travel to the bottomcorrosive elements. anchorage. Another reason for pitting corrosion of

*the wires and anchor heads may have been a poorconstruction practice where the tendons were either

4.4.3 Steel Degradation Mechanisms and Their stored at the site without proper protection for longEffects. This section will discuss the potential periods or were not properly protected after instal-Effects. Ths s n wlation and before application of the permanent cor-degradation mechanisms associated with reinforc- rosion protection material.ing steel, posttensioning system steel, steel liner

A few anchor heads on different containmentsplate, and steel material used for the steel cylinder/ a eweancho hea onadint ontainmesteel sphere containment shells. Unlikec have experienced either cracking or partial failure

themajr.dgraatin.mchac r et e - and in one case a complete failure because ofthe major degradation mechanism 'for steel is cor- hydrogen embrittlement. The time at which theserosion. Elevated temperatures and radiation alsohave the potential to degrade steel. ' failures occurred ranged from a few days to a fewv

years after stressing. The chemical composition ofthe anchorhead material and heat treatment proce-

4.4.3.1 Corrosion of Steel. The potential for dures used are thought to have contributed to thecorrosion of the reinforcing steel is highly depen- failures. It has not been positively confirmeddent on the quality of the concrete in'which it is whether such failures are time dependent andembedded and on the presence of cracks and poros- whether they are related to a specific environmentity in the concrete that may allow moisture and near the anchorage.other harmful substances to reach the steel. Corro-sion, when it occurs, will result in iron oxides that

4.4.3.2 Effects -of Elevated -Temperatures onexpand In volume causing spalling of the concreteexpand in volume causing saigofthecoSteel. In general, elevated temperature decreases theand further exposure,~ allowing more corrosion to'andcfurtWhera eosureallcowing'mren corxosin to an yield strength and modulus of elasticity of the steel andoccur. 'Whenvrne an s otee ntainmens meosured th ans increases its ductility. However, these effects are onlyacidic environment and protective measures such ascoatings or a dry environment are not provided (or pronounced when the steel is exposed to temperatureshave worn out), the acid attacks the steel. This of-3000F (1491C) and above.14,15 One characteristic

resuls in cof special importance, even at temperatures of 100results in chemical corrosion. Scaling of the corro- to 150'F (38 to 660C), is the relaxation of the postten-sion product will result in a reduced structural sec- sionin(g tendon steel that increases with increased tem-tion leading to increased stress 'intensities -and

reduc.d leak-tightness cap.bility. Similarly, 'co- perature resulting in loss of the prestressing force.reduced leak-tightness capability. Similarly, corro- ---sion of the steel linr en cnt Fortunately, the tendons are located in the containment

sion of the steel liner m i concrete containments..could' j e ishell within the concrete which acts as an insulator andcould'Jeopardize its 'leak-tightness capability,whih iprotects the tendons from exposure to harmful temper-which is the primary function of the liner.

Corrosion in posttensioning system components atures.may be either localized or uniform. Mosi prestress-ing, corrosion-related failures have been the result 4.4.3.3 Effectof Radiation on Steel. This is notof localized corrosion in the form of pitting, stress a potentially 'damaging mechanism for reinforcingcorrosion, or hydrogen embrittlement. A review of' - bars, 'liner, structural steel, anid posttensioning ten-the prestressing tendon surveillance reports indi-t dons. Reinforcing steel and posttensioning tendons incates that corrosion incidents are extremely'rare' the containment shell are normally exposed to limitedHowever, corrosion has been detected at a number ' amounts of radiation compared to the structural steelof plants during'the tendon system inspection. The' inside the 'containment. Exposing test 'specimens tomost common incidents of corrosion are pitting on '- radiation flux levels higher than that which postten-the tendon wires or strands. The pitting in most sioning tendons are normally exposed to has resulted incases has been limited and attributed to the press harmless effects.7 Therefore, this mechanism is notence of small amounts of water that have accumu-' " important.

PI

Li

I

I

45

I

I

4.5 Potential Failure Modes

Possible containment failure modes during nor-mal operation, accidents, and postdesign life arediscussed in this section.

4.5.1 Prestressed Concrete Containment Vessels

4.5.1.1 Failure Modes During Operation.Structural failure under load is not anticipated dur-ing normal operation because load-carrying mem-bers are designed with considerable margin to carrysuch applied loads. Therefore, any failure modesare expected to be the result of exposure to adverseenvironmental conditions. The failure modes areranked in order of importance as follows:

I. Loss of prestressing force. There are anumber of factors that can lead to differ-ent levels of losses in the prestressing force.The most severe occurrence is the failure ofan anchor, resulting in either partial orcomplete loss of the tendon stress. In addi-tion, dealing with the situation of a parti-ally failed anchor presents a safety hazardto workers replacing the damaged anchor.In most cases, failure of anchors has beencaused by hydrogen embrittlement result-ing from either improper chemistry orimproper heat treatment of the anchormaterials. Failure of individual wires orstrands in the tendons is also of concern.This type of failure can be caused by eitheroverstress in individual wires or strands,localized inclusions of weak material,localized brittle fracture because of inclu-sion of an embrittling type of material, orreduction in cross section because of pit-ting and corrosion. Corrosion in postten-sioning components has occasionallyoccurred and may be the result of break-down in the effectiveness of the corrosionprotection material (grease) around thetendons and anchor components or thepresence of moisture and water in areasaround the tendons and anchors. Microbi-ologically influenced corrosion of the car-bon steel tendon wires has been reported atsome PWR sites. 16 .Prestressing lossescaused by creep and shrinkage of the con-crete and relaxation of the tendon material

-* might also result in failure of the postten-

sioning system, if the losses exceed thoseaccounted for in the design.

2. Corrosion of the reinforcing steel. Thismode normally occurs as a result of poorconcrete quality or concrete deterioration.The consequences can be serious andrepairs are complicated.

3. Corrosion of the steel liner over the baseslab. This mode of failure can lead to leak-age of radioactive fluids through the linerinto possible cracks or porosity in the con-crete slab and out to the outside environ-ment.

4. Deterioration of the concrete. During thelife of the plant, the concrete may be sub-jected to the degradation mechanismsdescribed in Sections 4.4.1 and 4.4.2. Theresistance of the concrete to degradation isrelated to the quality of the concrete andthe construction procedures used. Deterio-ration of concrete can have a number ofside effects, ranging from serious struc-tural consequences to superficial scaling.The primary shield wall is among the con-crete structures that should be inspected.

I

I

[.

4.5.1.2 Failure Modes During Accidents.Accidents which could adversely affect contain-ment structures are those which result in largeincreases in pressure or a hydrogen burn explosioninside the containment. In the case of prestressedconcrete containments, the hoop stresses in the ves-sel control the failure modes. Resistance of the con-tainment to this type of loading is dependent on theadequacy of the prestressing system. When theinternal pressure generated by the accident exceedstwice the design pressure, it could cause sizable,cracks in the concrete while structural stability, isstill maintained. A liner-concrete interaction maytake place at locations of major concrete cracks anddistortions and introduce high strain concentra-tions that may lead to localized rupture of the steelliner plate provided the liner is firmly bonded to theconcrete. The ruptured liner plate will then allowsome leakage of radioactive gases to the outsideenvironment. This failure mechanism has been pro-posed by Rashid and is sometimes referred to as theRashid mode of failure.17 Other failure modeshave been identified for containment failure atpressure levels higher than design pressure. 18

4.5.1.3 Failure Modes Beyond Design Lifeof Containment. Most of the possible failure

I

rL

LI

46

I

I

s ~ ~~~~ fal 'cno esmodes for this stage'are similar to the failuremodespossible during normal operation: However, prob-lems associated with the loss of the prestressingforce because of creep and shrinkage of the con-crete and relaxation of the tendon' material maybecome more important because these' changes,with time, will eventually consume all or most ofthe design margin provided. Load cycling resultingfrom scheduled and forced'shutdowns, and inte-'grated leak rate tests, may also develop fatiguecracks that jeopardize the integrity of the contain-ment components.'

4.5.2 Reinforced Concrete Containment Vessels. 'The potential failure modes for RCCVs aresimilar to those for prestressed conicrete contain-ments except for the failure modes related, to theposttensioning system (that does not exist in thereinforced concrete containment). However, rein-forced concrete containment failure in the hoopdirection under accident pressures can be more pro-nounced than in the prestressed concrete contain-ment. RCCVs develop widespread cracking 'atabout 75% of the design pressure because of theabsence of the prestressing forces. Therefore, local-ized strains in the steel liner plate 'due to liner-concrete interaction can' be large and may lead toleakage of radioactive gases.

,capability. The SIT and ILRT consist of pressuriz-ing the containment to 15% above its design pres-sure. Associated with the SIT are measurements ofcontainment growth under pressure for evaluationof actual containment behavior'in comparison withdesign predictions. The ILRT includes instrui-mented and visual examitiation' of surface-crackinitiation and growth at critical locations such ashatches, the ring girder, and the posttensioning sys-tern anchorage regions. After the ILRT, the surfaceof the steel liner plate also is visually inspected forany cracks, separation, hollow sounding or inwardbuckling regions, and corrosion that might indicate.any degradation in the system. A brief descriptionof the visual'examination is given in Chapter 11.The ILRT is repeated every five years to ensurestructural integrity and continued safety of the con-tainment. 'Additi6nal'discussion on the visualexaminations and leak testing of containments isgiven in Chapter 12.'

In addition to the ILRT tests, ISI of the postten-sioning system in the prestressed concrete'contain'-ment vessels is performed one, three, and five yearsafter the initial SIT-ILRT in-spection and every fiveyears thereafter.-This inspection is required by the'operating license and ASME Section XI and is con-ducted to monitor the level of prestressing forces inthe postternsioning tendons and 'inspect the condi-tion of theltendons, anchorages, 'shims, bearing

EL~

4.5.3 Steel Cylinder/Steel Sphere Containment 'plates, and corrosion protection materials. In addi-Vessels. The loads in the steel containment tion, wire or strand samples' are obtained duringstructures are entirely resisted by the structural steel each ISI from one or more tendons and tested forof the shell. Cracks and corrosion in welds and base tensile strength The concrete surface conditionmetal of the shell dominate the possible failure ' * near the anchorages, where it is stressed the most, ismodes in this type of containment structure during visually examined for the presence of any cracks ornormal 'operation. Containment cracks present .) ,deterioration. The ISI surveillance is also expectedboth structural and safety hazards because of stress to identify any'adverse effects related to tendonconcentrations and leakage of radioactive gases, relaxation or creep of the concrete.respectively. Corrosion and fatigue are the poten -' mProblems identified during these'tests' are' nor-tial failure modes for operation beyond the design y corrected immediaely. Gaskets, seals, welds, 'life. Of special concern are the shell areas near dis- etc. are provided when measured leak rates exceedcontinuities such as the base of the shell, hatches, the standard design allowables. Prestressing levelsand 'enetraiions. ' are restored when tendon losses exceed those pre-

penetrations. - ' -''''''''dicted. Tendons, anchors, and corrosion protection'material are replaced when' corrosion or cracking is'

4.6 In-service Inspection ' ' observed on' the tendons or anchors" and when' ' ' ' ''''chemical and physical test results indicate any deg-

In-service inspection (ISI) of the critical compo, radation in the corrosion 'pro materials.nents of the containment structures area manda-'' In addition'to the 'SIT, ILRT, and ISI, atory part of the operating licenses of nuclear power walkdown and inspection of all 'components,' sup-plants.1Eachl'contairinnent is subjected to a struc-' ports, structural elements, and system irains is per-tural integrity test (SIT) and an-integrated leak rate formed during each refueling outage. A punch listtest (ILRT) before operation of the plant to ensure is generated and executed for correction of defi-complete leak tightness and pressure resistance ciencies discovered during the walkdown.

r,~L

I

I

t[.

47

[

I

A great deal of information is collected on thecondition of each containment as a result of all theabove described inspection and testing programs.However, in light of the degradation mechanismsdescribed in Section 4.4 and the failure mode's pos-tulated in Section 4.5, additional monitoring maybe required to maintain the' containment in a reli-able condition beyond the design life.

General monitoring and ISI procedures for extendedlife operation should be established to identify thedamage due to any likely and significant degradationmechanisms in a timely manner. Such procedures mayinclude both nondestructive and destructive test meth-ods. Nondestructive methods may consist of visualinspection, ultrasonic inspection and testing, radiogra-phy, infrared thermography, rebound method for' sur-face hardness, etc.19 20 Destructive 'methods shouldbe used only when the nondestructive methods are notsufficient to quantify the extent of the: degradationdamage. Destructive methods may consist of obtainingcore samples from the in situ concrete for compressive-strength tests and petrographic 5caminationsA or theuse of penetration techniques sich as'the WindsorProbe and pullout testing for determination of in situconcrete strength. Core samples from the degradedregions would also provide a direct means to evaluatethe width and depth of cracking or the extent of anyvoids. Steel samples from both reinforcing steel andstructural steel components also may be required forlaboratory testing.

Establishment of inspection procedures to cover crit-ical areas where adverse environmental conditions suchas high temperature, humidity, and/or radiation, andlocations subjected to an acidic environment, will be anecessary measure to determine the exient of degrada-tion. Representative samples. of steel 'and concreteshould be obtained from critical areas'and tested todetermine the existing chemical and physical propertiesof the materials. Specifically, inspection of concreteagainst deterioration from'environmental factors'andunderground water attack'and monitoring of concretecompressive strength by testing some core samples maybe necessary. Specific examination of steel membersfor corrosion, cracking, and distress because of cyclicloading conditions may also be necessary.

In summary, by implementing a well planned,meaningful, and comprehensive inspection pro-gram followed by a firm program of repairs andmaintenance, the service life of containment struc-tures can be extended, probably to at least'doublethe initial licensed term and possibly much longer.

4.7 Summary, Conclusions, andRecommendations

The nuclear power plant containment degrada-tion sites include areas and components where deg-radation may cause either structural damage or asafety hazard or both. Degradation mechanismsunique to nuclear power plants such as those result-ing from radiation and nuclear heating and ele-vated temperature have been 'discussed in detail.Other degradation mechanisms that affect generalstructural components have also been discussed.

The available research and test results indicatethat radiation and nuclear heating have little effecton the shielding and structural capability of con-crete. However, activation of primary shield con-crete surfaces in the reactor vessel cavity byneutrons, even to a few inches in depth, can be areal radiation protection' problem after 40 years ofoperation.2 2 The effects of radiation on the steelused in PWR containmen'is appear to be insignifi-cant within the dose range experienced during nor-mal operation of the plant. Although it appearsthat elevated temperature does affect concrete, itsdegradation effects are unique to each plant basedon the concrete material and quality used.I Degradation caused by freezing and thawing, wet-ting and drying, and chemical reactions is a conditionthat is related to the quality and durability of the con-crete selected for the"plant. Careful testing, evaluation,and selection of the concrete material with proper mixproportioning and construction procedures could min-imize or eliminate degradation related to these factors.However, some remedial mneasures may be taken tolimit or slow the degradation process if it occurs. Suchmeasures can be in the form'of actions to prevent mois-ture and free water from entering and leaving the con-crete.

Corrosion of the structural or reinforcing steelcan be a major degradation mode; however, itshould be noted that'corrosion is initiated as theresult of adverse environmental conditions. Dis-tress caused by corrosion whether in structural steelor reinforcing steel can have a great structuralimpact. Cracks and fatigue caused by vibratingloads and dynamic loads can be serious in terms ofthe structural integrity of the affected members.

Failure modes'are closely related to the describeddegradation mechanisms. The failure modes postu-lated are general because they are presented on a

L

rim

48

E1

I.

generic basis. More unique failure modes such asliner-concrete interaction are expected to be identi-fied when evaluations are performed on a plant-by-plant basis.

The importance of ISI has been recognized andreflected in the presently imposed inspection pro-grams for the containment structures and compo-nents, in particular, the posttensioning system forthe prestressed concrete containment vessels. Itshould be realized that great safety and economicbenefits can be derived if an expanded ISI is imple-mented to cover identified degradation sites thatmay not be frequently inspected.2 3 Immediaterepair or remedial action is of great importance inarresting any degradation before it progresses to apoint where other components begin to degrade orother degradation mechanisms start that otherwisewould have not started. A simple example of this isdeterioration of the concrete cover over the rein-forcement. If deterioration of the cover material ispromptly detected and repaired, degradation of thereinforcing steel will not occur.

Tables 4.1,- 4.2, and 4.3 provide a summary ofthe degradation sites, stressors, mechanisms, fail-ure modes, and necessary inspection methods forthe three types of the PWR containments discussedin this report.

There are several issues associated with aging andlife extension of the PWR containments that needto be resolved. A comprehensive and standardizedISI program is needed to identify and quantify deg-.radation in reinforced concrete. The state-of-the-art methods used in Europe, Japan, and the U.S.A.for evaluating the condition of reinforced concreteneed to be assessed and included in the program. Astandardized program implies the same ISI pro-'gram for all LWRs. The ASME Section XI com-mittees on metal and concrete' containments areevaluating the current ISI methods for reinforced

concrete and developing specific requirements forthorough visual inspections.

Additional information about the long-termdegradation of reinforced and prestressed concretecontainments is needed. Data should be collectedfrom the older LWR containments. Also, facilitiesthat have been shut down after an extended period'of service (such as Shippingport, Dresden, orHumbolt Bay) could be used to obtain additionalaging-related data on reinforced concrete materi-als. Accelerated aging techniques could also be

'.investigated and, if found to be appropriate, usedto obtain posttensioning system and concrete mate-rials aging data. The results from these investiga-tions will contribute to the development of anaging-related material property data base and sup-port the development of quantitative models of thedegradation:of reinforced concrete subjected tolong-term exposure to elevated temperatures, radia-:tion, and cyclic loadings. Two items in the postten-:sioning system that need particular attention aretendon relaxation and degradation of the greaseused as the corrosion inhibitor. Also,' a monitoringmethod' is needed to detect degradation of theanchorages because of hydrogen embrittlementand aging of the grease caused by impurities. Otherless important recommendations include evalua-tion of the potential interactions between chemicalreactions in concrete and radiation, and determin-ing the extent of the deterioration of the shieldingproperties of concrete caused by radiation heating.

The structural integrity of aged metal containmentand analysis'of concrete-liner interactions in 'agedRCCVs and PCCVs during severe accidents are twoother issues that need to be resolved. The results fromthe testing of containment models subjected to pres-sure and thermal loadings and the results from theEPRI-sponsored research'programs to determine thesignificance of concrete-liner interactions in aged con-tainments will be helpful.

I

I

i

I

49

I

a

Table 4.1. Summary of degradation processes for prestressed concrete containmentvessel

DegradationRank Site

I Posttensioningsystemanchorage

2 Posttensioning. tendon wire orstrand

Stressor

Materialproperties andtrapped water

Moisture, trappedwater, orbreakdown ofgrease material

Moisture, acidicenvironment, andstress

DegradationMechanisms

Hydrogenembrittlement

Pitting,microbiological-inducedcorrosion

PotentialFailureModes

Loss of stress

Loss of stress

1SIMethod

Tendonsurveillanceprogram

Tendonsurveillanceprogram

.

3 Steel liner. dome and wall

Corrosion Liner-concreteinteraction,leakage ofradioactivegases

Leakage testing(10 CFR 50,Appx. J)

4 Steel linerover base slab

Moisture, acidicenvironment, andstress -

Corrosion Leakage ofradioactivematerial

Leakage testing(10 CFR 50,Appx. J)

5 Dome, wall, andbase slabreinforcingsteel

Aggressiveenvironment

Corrosion Loss ofstructuralintegrity

Visual

6 Concrete Aggressiveenvironment andinternal chemicalreactions

Cracking andspalling

Loss ofintegrity,corrosion ofreinforcingsteel

Visual, reboundmethods, coresamples ifrequired

Li

50

a

Table 4.2. Summary of degradation processes for reinforced concrete containment vessel

DegradationSite

- DegradationMechanisms

PotentialFailureModes

ISIMethodRank Stressor

I Dome and wallreinforcingsteel

2 Base slabreinforcingsteel

Aggressiveenvironment

Aggressiveenvironment

Corrosion - Lossofstructural

! integrity

Visual

VisualCorrosion Loss ofstructuralintegrity .

3 Steel linerover dome andwall

Moisture, acidicenvironment, andstress

Corrosion Liner-concreteinteraction,leakage ofradioactivegases -

Leakage ofradioactivematerial

Leaking testing(10 CFR 50,Appx. J) [;

4 Steel linerover base slab

*Moisture, acidicenvironment, andstress

Corrosion Leakage testing(1OCFR 50,Appx. J)

5 Dome and wallconcrete

Aggressiveenvironment,internal chemicalreaction

Cracks andspalling

Loss of .integrity,corrosion ofreinforcingsteel

Loss ofintegrity,corrosion ofreinforcingsteel

Visual, reboundmethods, coresamples ifrequired

I

6 Base slabconcrete

Aggressiveenvironment,internal chemicalreactions

. Cracks andspalling

Visual, reboundmethods, coresamples ifrequired

[

51

a

Table 4.3. Summary of degradation processes for steel cylinderlsteel sphere containmentvessel

DegradationRank Site Stressor

DegradationMechanisms

PotentialFailureModes

5l1Method

I

I Shell welds andbase metal

2 Interfacebetween shelland concreteslab at baseof shell

3 Discontinuitiesin the shellsuch as hatchesand penetrations

Stresses,vibration, cyclicloading,aggressiveenvironment

Stresses,vibration,cyclic loading,aggressiveenvironment

Stresses,vibration,cyclic loading,aggressiveenvironment

Corrosion

Corrosion

Corrosion

Loss ofstructuralintegrity,leakage ofradioactivegases

Loss ofstructuralintegrity,leakage ofradioactivegases

Leakage ofradioactivegases

Visual, leakagetesting

Visual, leakagetesting

FLeakage testing(10 CFR 50,Appx. J)

4 Steel bottomof shellembedded inconcrete

Aggressiveenvironment

Corrosion Leakage ofradioactivematerial

Leakage testing(10 CFR 50,Appx. J)

5 Base slabconcrete

Aggressiveenvironment,internal chemicalreactions

Cracks andspalling

Corrosion ofreinforcingsteel,corrosion ofsteel bottomofcontainmentshell

Visual, reboundmethods, coresamples ifrequired

I

52

I

4.8 References

I. American Society of Mechanical Engineers, ASME Code for Concrete Reactor Vessels and Contain-ments, Section ill, Div. 2, New York.

2. A. M. Neville, Properties of Concrete, third edition, London: Pitman Publishing Limited, 1981.

3. N. Hanson, D. Schultz, and J. Roller, "Results of Larger Scale Test of a Discontinuity Region in aPrestressed Concrete Containment," Construcdion Technology Laboratories, Third Workshop onContainment Integrity, May 21-23 1986, Washington, D.C.

4. American Concrete Institute, Considerationsfor Design of Concrete Structures Subjected to FatigueLoading, ACI 215R-74, 1981.

5. G. E. Troxel and H. E. Davis, Composition and Properties of Concrete, New York: -McGraw-HillBook Company, Inc., 1968.

6. American Nuclear Society, Guidelines on the Nuclear Analysis and Design of Concrete RadiationShieldingforNuclearPowerPlants, Americani National Standard, ANSI/ANS-6, 4-1985, La GrangePark, Illinois, 1985.

7. H. E. Hungerford et al., "Concretes, Cements, Mortars and Grouts," in Engineering Compendiumon Radiation Shielding, New York: Springer-Verlag New York, Inc., 1975. X

8. M. L. Gritzner, G. L. Simmons, T. E. Albert, and E. A. Straker, PWR and BW'R Radiation Envi-ronmentsfor Radiation Damage Studies, EPRI NP-152, September 1977, pp. 30, 49.

9. Construction Technology Laboratories of the Portland Cement Association, Strength and ElasticProperties of Concretes from Waste Tank Farms, report prepared for Rockwell Hanford Operations-Energy Systems Group, Richland, Washington, RHO-C-22, December 1978.

10. Construction Technology Laboratories of Portland Cement Association, Effects of Long-Term Expo-sure to Elevated Temperature on Mechanical Properties of Hanford Concrete, report prepared forRockwell Hanford Operations-Energy Systems Group, Richland, Washington, RHO-C-54,October 1981.

11. "Strength and Elastic Properties of Concrete Exposed to Long-Term Moderate Temperatures and HighRadiation Fields," American Concrete Institute, Fall Convention 1984, New York, RHO-RE-SA-55P.

12. American Society for Testing and Materials, Standard Specification for Portland Cement, ASTM C 150. r

13. V. V. Bertero and M. Polivka, "Influence of Thermal Exposures on Mechanical Characteristics ofConcrete," in Concrete for Nuclear Reactors, Volume 1, ACI Special Publication SP-34,Detroit: American Concrete Institute, 1971.

14. T. T. Lie, "Fire Resistance of Structural Steel," Engineering Journal/AISC, fourth quarter, 1978.

15. M. Fujimoto, F. Formura, and T. Ave, "Primary Creep of Structural Steel (SM50A) at High Temper-atures," Transactions of the Architectural Institute of Japan, 306, August 1981. E

16. D. H. Pope, A Study ofMicrobiologically lnjluenced Corrosion in Nuclear Power Plants andaPracti-cal Guidefor Countermeasures, EPRI NP-4582, May 1986, pp. 1-2, 1-3.

53

17. R. S. Dunham et al., Methodsfor Ultimate LoadAnalysis of Concrete Containment, EPRI NP-4046,June 1985.

18. 'M. A. Daye and B. L. Meyers, "Progression of Stress Development in Pressurized ContainmentStructures," Transactions of the 8th International Conference on Structural Mechanics in ReactorTechnology, Volume J, Brussels, Belgium, August 19-23, 1985, North-Holland for the Commission ofthe European Communities, pp. 7-12.

19. American Society for Testing and Materials, Aarnual Book of ASTM Standards, Section 00,Volume 00.01.

20. D. J. Naus, Concrete Component Aging and its Significance Relative to Life Evtension of NuclearPower Plants, NUREG/CR-4652, ORNL/TM-10059, September 1986.

21. American Society for Testing and Materials, Standard Recommnended PracticeforPetrographicExarm-ination of Hardened Concrete, ASTM C 856.

22. R. G. Jaeger et al. (eds.), Engineering Compendium on Radiation Shielding, Volume 11, ShieldingMaterial, New York: Springer-Verlag New York, Inc., 1975.

23. J. Cook, "Review of Containment Examination and Testing Requirements," 3rd Workshop on Con-tainment Integrity, Washington, D.C., May 21-23, 1986.

54

I

5. PRESSURIZED WATER REACTOR COOLANT PIPINGR. L. Cloud and W L. Server

5.1 Description coolant piping for the CE and B&W plants iswrought ferritic steel clad with stainless steel. Thematerials for all piping are shown in Table 5. 1.- 7hr^ ferriti- nini;npn iicpr! hy nRurWnnri (7F in rindj

- ~ ~ ~ ~~~~~~~~~~~~~~~~~~~ . .31 tsssJz 2rps ua*u u ~ *svwLa o sla t.

5.1.1 Geometry, Design, Vendors. Reactor with austeritic stainless steel, as noted in Table 5.1.coolant piping systems for the three principal pres- B&W uses weld-deposited cladding, and CE 'usessurized water reactor (PWR) types are shown in ' roll-bonded cladding. B&W uses explosivelyFigures 5.1, 5.2, and 5.3. The PWR designers are bonded cladding in the piping elbows. The elbowsBabcock & Wilcox (B&W), Combustion Engineers - are fabricated with two half elbows (clam shells)ing (CE), and Westinghouse (W). and the clad is applied before the two halves are

As can be seen, the B&W and CE designs are welded together.similar in that there are two heat 'exchange loops. There is no need for cladding in the W pipe becauseEach loop has a' hot leg from the reactor to the it is entirely stainless steel. As noted in Table S.1, how-steam generator, and two cold legs that return the ever, there are special circumstances with the weldcoolant from the steam generator to the reactor. between the austenitic pipe and the ferritic reactor ves-There are four pumps, one in each cold leg. sel and steam generator. Because any weld to the fer-

The only major conceptual difference between ritic steel would require stress relief in these wallthe CE and B&W design is that the B&W steam thicknesses, a layer of Inconel 600 is deposited on thegenerator (SG) is a once-through straight tube type, nozzle ends before final vessel stress relief. Then, theand the CE design is a recirculating U-tube type. stainless steel pipe can be welded to the Inconel and noCirculation of primary coolant in the U-tube SG is additional stress relief is required. The Inconel was notfrom the bottom hot-leg side up to the top of the thought to become sensitized in the vessel stress relief.tube bundle, then down to the bottom of the cold- CE piping is constructed of roll-bonded claddedleg side. The B&W straight-tube SG path for the plates. The plates that are cladded during the roll-primary coolant flow is from the top headv'ertically ing process are formed into rounds and seamdownward to the bottom head. ' welded. Careful weld preparations permit austen-

Consistent with the SG designs, the primary sys- itic weld material to be used on the cladding andtem hot-leg piping for the B&W design goes from ferritic weld material on the ferritic plate formingthe reactor vessel outlet nozzle upward to the top of the main pipe wall.the SG. The hot-leg piping for the CE design goesdirectly across the compartment from the reactorvessel outlet nozzle to the bottom' head of the SG. 5.2 StressorsThe overall arrangement of the cold-leg piping for - -

these two designs is quite sirnilar. " - ' t ~~~Nuclear plants are most economical when theyThe third major PWR design, .xV. is conceptually -* - .. operate at steady state,' full power from one refuel-

different from the CE and B&W designs 'in that - v~-ing to the next. Certainly, all plant owners strive foreach heat exchange loop has a SG and a 'reactor this type of operation. From a technical point ofcoolant pump. A given reactor plant may have twiO, view; this mode of operation is also optimum. Thethree, or four heat exchange loops. The number of only significant source of stress is the internal pres-loops in any given plant is dependent on a number sure. In'the steady-state condition, thermal stressesof factors, but the most important is the capacity of do exist because of minor degrees of restrainedthe plant. Each loop increases the amount of power expansion, but they are minimal.that may be extracted from the reactor. However, for one reason or another, nuclear

plants rarely operate for a complete fuel cycle at5.1.2 Materials and Fabrication. The later 3Y full power in an undisturbed condition. The con-plants have main coolant piping made from centrif- - cern about safety is so great that the plants are fre-ugally cast austenitic stainless steel, and the fittings quently shut down and examined for off-normalare statically cast stainless steel. However, the main situations. Equipment failures often require shut-coolant piping in the early plants was wrought aus- down to cold conditions. For various reasons,'tenitic stainless steel (such as in Surry-l). The main reductions in power are common. Testing of

1.

I

I

55

Reactorcoolantpump/motor---,

Steamgenerator

Reactorcoolantpump/motor --

Reactorcoolantpump/motor

._Steam)- generator

- Reactorcoolantpump/motor

Steamgenerator

Reactorcoolantpump/motor

7-3068

Figure 5.1. Babcock & Wilcox system.

56

Pump 1B Pump2A

SteamgeneratorNo.2 I

I

I7-3069

Figure 5.2. Combustion Engineering system.

fL

57

I

Two loops

Three loops

Four loops

Reactor

7-3067

Figure 5.3. Westinghouse system.

58

Table 5.1. 'PWR main coolant pipe material. I .; ,. i,. . *I

Vendor'

Bib

CE

Piping ':

CF8ACCF8M ;- i

SA516GR70

Fittings

CF8A

*- - SA516GR70! Type 309L SS

Claddinga

N.A.

Type 308L SS

B&W ;. . SA 106GRC . l: SA 516 GR 70Type 308L SSType 309L SS

Type 304L SS

a. Welds between austenitic stainless steel (SS) and ferritic material normally have an Inconel 600 layer to overcome the need forsensitizing stress relief heat treatment.

b. Early W PWR plants had wrought Type 304 SS main coolant piping.

c. CF8A and CF8M are cast grades of l'x 304 and Type 316 SS respectively.

systems and equipment is a major cause of tran,-,,sients. These transient conditions contribute addi-_-tional thermal stresses to the main piping.

In addition, the plants are designed to be able toaccept unusual conditions, such as severe earth-quakes and dynamic effects resulting from cata-;strophic pipe failure in connected systems; Thesevarious conditions are discussed in the followingparagraphs. : -

5.2.1 Steady-State Stress. All PWR primarysystems are designed to freely expand as their tem-perature is increased from ambient to operatingtemperature, a rise of "500'F (280'C). Inall cases,the reactor vessel is rigidly attached to the reactorbuilding and serves as a fixed point when the sys-..tem heats up and expands. The hot-leg pipingattached to the reactor vessel nozzles expands alongits axis to the steam generator. All three PWR--designs have SGs that move freely to accommodate.the pipe expansion. L : -

The W SG is supported on columns with pinnedconnections that rotate freely and permit the SG toassume the natural or unconstrained hot position,-.of the piping. The CE and B&W designs have SGs la-supported on skirts with sliding bases that accom-plish the same result. The SGs for all three vendorsmove '.2 in. (50 mm) during the plant heatup fromambient to operating temperature. The reactor

coolant pumps also are supported in such a mannerthat cold-leg expansion is essentially unrestrained.

In the steady-state condition then, the state ofstress in the main coolant piping is dominated bythe internal pressure contribution. There are smallamounts of thermal stress that result from second-ary effects because perfect free expansion cannotbe achieved. In addition, there are small amountsof gravity-induced stress that are inconsequentialbecause of the short, stiff nature of the primarypiping. Because the thermal expansion is uncon-strained, there is little or no steady-state bendingstress. The piping itself is of relatively large diame-ter with very short lengths..

For the reasons given above, the main coolantpiping may be more accurately thought of as aseries of pressure vessels than piping. This conceptis important from the viewpoint of life extension,because the measures to be taken with main coolantpiping will be'the san'xe measures to be taken withthe connected pressure vessels, in contrast withthose measures needed for long runs of relativelyflexible piping.

As discussed above, the steady-state stress is essen-tially given by the 'ressure6 contribution. The stress isuniform tension in the pipe wall and is written as

I

I

I

I

Sb=prSh = Pr (5.1)

59

1'

I

S2 = Pr (5.2)2t

S. = 2 (5.3)

where

p = pressure

r = pipe mean radius

t = pipe thickness

Sh = circumferential or hoop stress

SI= longitudinal or axial stress

Sr = radial stress.

The American Society of Mechanical Engineers(ASME) code stress intensity (Pm) is equal to twicethe maximum shear stress, shown as

p pr + p (5.4)t 2

and is limited to the ASME code design stress Sm.Sm in turn is limited to the smaller of 2/3 S. or

1/3 Su

where

Su = specified minimum, ultimate tensilestrength

Sy = specified minimum yield strength

and these quantities are the minimum specificationvalues for the steel in question. Because of the highstrain hardening nature of some austenitic stainlesssteels, Sm may be >2/3 Sy but never >90% of S.nor 1/3 Sul, where S., and Su, are the yield and ulti-mate strengths, respectively, at temperature.

The foregoing paragraphs describe the steady-state stress on the pipe wall. As mentioned, there

may be small amounts of additional stress caused.by secondary constraints, but these are not signifi-cant. The stress picture is quite different, however,at discontinuities or changes in section. Stresseswill be much higher at openings in the pipe wallwhere auxiliary pipe nozzles are attached and alsoat terminal ends where the pipe is welded into theprimary vessels. In general, the elevated stresses atthese locations are the result of purely geometricaleffects that tend to produce incompatible deforma-tions under the pressure load. The continuity of thestructure requires compatibility, however, andresults in locally higher stresses at geometrical dis-continuities. As will be discussed later, thesestresses are secondary in nature and are importantmainly from the viewpoint of fatigue.

5.2.2 Transient Stresses. Each PWR primarysystem of the current generation is designed for aspecific set of transient conditions. That is, the*stresses resulting from a specific set of transientssatisfy the limits of the ASME code. Earlier plantsthat did not fall under the requirements ofSection III of theASME code of 1971 or later or ofAmerican National Standards Institute B31.7 werenot specifically required to be analyzed for tran-sient stresses.

In this segment, nuclear plant transient condi-tions and the stresses resulting therefrom are dis-cussed. As mentioned earlier, different generationsof plants had to meet increasingly stringent analyti-cal requirements, but as a practical matter the phys-ical configurations of the plants changed but little,and the actual transient conditions were also verysimilar. Therefore, little or no generality is lost bydiscussing transient loads and stresses in current-generation plants.

5.2.2.1 Transient Conditions. The manufac-turer of the PWR system is required by Section IIIof the ASME code to compile a design specifica-tion that lists all' transient'design conditions. Thedesign specification contains a description of thetransient conditions, usually in the form of atemperature-time relation, and lists the number of'occurrences for which each transient conditionmust be considered. Design conditions and tran-sients have historically been grouped into the fol-lowing categories:

0

0

0

Design'TestNormal operation

60

I'

0

Upset conditions ! 7

Emergency conditions' ' -Faulted conditions.

Faulted conditions normally include the safeshutdown earthquake and loss-of-coolant 'acci-dents. If a' nuclear plant is afflicted by either ofthese events, there is no doubt that it will receiveconsiderable special attention, probably far exceed-ing any life extension program procedures. In viewof this consideration, and the extremely low proba-bility of occurrence of faulted plant conditions,these conditions will be omitted from this discus-sion on life extension.' .

All the system transients in the remaining catego-'ries are considered in the plant analysis. Most ofthe transients arc in the category of normal andupset conditions: To understand the effect of thetransient condition, a specific transient, the systemheatup and cooldown is discussed in some'detail.'This transient is especially appropriate for severalreasons: it certainly must occur for the plant tooperate, it is one of the simplest transients, and it'contributes significant fatigue damage.-

Ideally, in a plant designed for 40 years, the primarysystem would need to be shut down only during refuel-ing, say conservatively once a year or on the order of40 times during life. For design purposes, usually 200'(or even '500) heatup, cooldown cycles are specified toaccount for nonideal operation.

Normally the heatup would begin with the cool-'ant at ambient temperature and pressure. In PWRplants, the 'heatup 'is controlled by a pressure-temperature relation that limits the ' pressureaccording to the temperature elevation. Thispressure-temperature limit ensures that the reactorvessel will not be pressurized in a cold condition'when the fracture toughness'of the 'feiritic steel is'depressed, particularly later in life after some radi-!ation embrittlement has occurred. Heat is added to'the'system" gradually, usually the heat-loss fromoperation of the 'reactorr'coolant pumps and thetemperature and pressure' increase gradually, usu-ally at a rate'on the'order of 50'F/h'(280C/h)although specification limits' may be higher, sayl00F/h'(380C/h). Once the temperature is ele-vated sufficiently,' 'the control rods can be !pulledand the heatup completed with reactor heat. In a'cooldown, the p'rocess is reversed.'} ''"

Temperature gradients 'develop throughout theprimary system in the c6urse'of the heatups.'From'a stress viewpoint, the most significant temperaturegradients are those'from the inside diameter to theoutside diameter of the'piping and nozzles. The

warmer inside' diameter "seeks 'to :expand and isrestrained by the cooler outside diameter. The situ-ation is exacerbated at nozzles and other locationsof changing' geometry and wall thickness. Aftersteady state is reached,' the gradients even out andthe transient stresses vanish. Each change in system'operation produces 'these effects in a greater orlesser degree. For example, once the system is at fullpower, a change in power level lowers the hot-legtemperature, which develops gradients in the pipeand associated transient stresses.'

The significance of the transient stresses is thatthey, being of a cyclic nature, can cause fatigue.This is further discussed in Section 5.3.1. As can beseen from the foregoing discussion, the damage canbe limited by reducing the number of transients andthe rate of change of temperature. ''

I

5.3 Degradation Mechanismsand Sites

The primary-coolant piping consists of the hot-legpiping, cold-leg piping, pressurizer surge-line piping,and the several nozzles in the pipe connected to, forexample, the charging line, safety-injection line,residual-heat-removal line, etc. In this section, variousmechanisms are discussed that tend to reduce the life ofpiping and the nozzles in the general case. For the sake'of completeness, an attempt has been made to discussthe applicability of most such mechanisms to the pri-mary coolant piping of the PWR. No particular orderis followed.

I

5.3.1 Fatigue. There is an entire list of mecha-nisms that can cause failure of pressure vessels andpipings and have done so in the past. As has beendiscussed in previous paragraphs, very few of thesemechanisms are applicable to the main coolarit'pip-ing of PWRs. However, one of the most widespreadand best known, namely fatigue, is applicable. Sofar as is known, there have been no failures, noteven 'discovered cracks, in existing PWR 'reactorcoolant piping. That does not mean, however, that':there will not be fatigue problems in the future.'

The significant fatiguie mechanism is low-cyclefatigue caused by 'a combination of, pressure andtransient 'thermal stresses. The points' of higheststress throughout the systems are the most vulnera-ble. These sites include the nozzles welded into theprimary-coolant piping and the'terminal ends ofthe piping. Each transient event (that is,' eachheatup :cycle,'each test of the safety injection

I

F

r

61

I

system, each operation of the charging pump, etc.)produces a cycle of stress throughout the system.At the geometrical irregularity of a nozzle or metal-lurgical irregularity of a. dissimilar, metal weld,these stress cycles may be significant. It is perhapsworthwhile to digress briefly to describe fatiguedamage and the means of calculating it.

At a given point subject to a cyclic stress ofsufficient magnitude, a crack will form. Then, dur-ing succeeding cycles the crack will grow until itprogresses through the wall. It was shown byCoffinl that the fatigue life of a small, smooth-polished specimen was proportional to the range ofstrain imposed on the specimen in a cyclic constantstrain range test. Since most of the fatigue life of asmall smooth specimen is in the precrack phase, itmay be inferred that the range of strain determinesthe onset of crack initiation.

Later it was shown by Paris2 that the amount ofcrack growth of a given crack under a given cyclicstress correlated almost exactly with the appliedfracture mechanics stress intensity range, AK,which is a function of the crack size, shape, stresslevel, and geometry. Detailed discussions of crackinitiation and growth during low-cycle fatigue areprovided in Section 3.4.2 of Chapter 3 and inChapter 13 of this report.

The primary system piping is designed and ana-lyzed on the basis of Section III of the ASME code.The ASME code provides a fatigue design curvederived directly from smooth, small specimen fail-ure data by imposing safety factors of 2 on stressor 20 on cycles, whichever is lower. These factorsare intended to account for size, environment, andother differences between laboratory specimensand actual equipment.3 Large-scale vessel fatiguetests have been performed for the express purposeof checking the ASME fatigue design curves.4 Itwas shown by these tests that cracks, may beexpected to initiate below the fatigue curve, butthat, wall penetration is not expected until fatigueusage exceeds the design curve.

The ASMtE code also employs Miner's lineardamage rule to combine fatigue damage that accu-mulates at different stress levels. In this way, fatigueis accumulated according to a cumulative usagefactor that is equal to unity when the accumulatedfatigue usage is just equal to the ASME code designfatigue curve. See Section 3.4 for a detailed discus-sion on this topic.

Referring then to the tests of Reference 4, it maybe inferred that for those locations in the primarysystem that: have design fatigue usage factors nearunity and in-service actually experience the design

usage, then it would not be unreasonable to findfatigue cracks initiating near the end of design life.Of course, there are mitigating factors such as vari-able sequences of load cycles, etc. that couldimprove the situation. Nevertheless, it is clear thatlow-cycle fatigue cracking is definitely one of themost important damage mechanisms that will haveto be dealt with in any life extension program forprimary piping.

5.3.2 Thermal Aging. Cast austenitic-ferritic(duplex) stainless steels with significant amounts ofdelta ferrite will experience a reduction of tough-ness when aged at elevated temperatures. The maxi-mum effect occurs at 8850F (4750 C) and thegeneral phenomenon is often identified as 8850F or4750C embrittlement. This temperature is wellabove the maximum piping temperatures of PWRmain coolant piping; nevertheless, the phenome-non obeys a time-temperature relationship suchthat at lower temperatures the toughness propertiesare still affected, but a longer time is required. Thegeneral phenomenon and the effect on PWR main-coolant piping is discussed at length inReferences 5, 6, and 7.

In the references cited, substantial investigationshave been carried out and it has been suggested thatin the design lifetime of existing PWR plants, thesafety and integrity of the materials affected hasnot been compromised. The reason for this state-ment is that the cast stainless steels are very toughinitially, and the amount of aging that occurs at thereduced temperatures of PWR operating systemsdoes not lower the properties enough to render thematerials susceptible to brittle fracture, nor toaffect the crack-growth rates significantly to com-promise the design capability.

It turns out that the degree of aging is related tothe volume percentage of delta ferrite in the duplexmaterial. It would definitely seem that piping orcomponents such as pump or valve casings thathave high delta ferrite content that are in the hot-legside of the plant would become susceptible to brit-tle fracture at some period during an extended life-time. These components will in the future requirecareful monitoring. Just as the reactor vessels aremonitored today, both as to property changes andthe appearance of cracks.

The NRC Office of Research is sponsoring a pro-gram at the Argonne National Laboratory to study:the long-term aging effects on cast duplex stainlesssteels. 8 Service-aged, as well as laboratory-aged,specimens are being examined in this program.Close monitoring of this program is recommended.

62

I'

Additional information on the thermal aging ofcast stainless steel components may be found inChapter 10, Boiling Water Reactor RecirculationPiping, of this report.,

5.3.3 Stress-Corrosion Cracking. As is wellknown, intergranular stress-corrosion 'cracking(IGSCC) occurs in metals when tensile'stress'ispresent, the environment is conducive to stress cor-rosion, and the material has been sensitized. Thereare different conducive environments but the halo-gen salts promote stress corrosion cracking in aus-tenitic stainless steels, and caustic 'environmentspromote stress-corrosion cracking in ferritic steels.

In the PWR primary piping, to date, there havebeen no known cases of stress corrosion cracking.Cracking is suppressed by good control of the envi-ronment including scavenging of oxygen to verylow levels. Also, the boric acid used for reactivitycontrol is believed to contribute to suppression ofstress-corrosion cracking. In any event, unlike theboiling water reactor (BWR) plants, those. fewPWRs with wrought stainless steel piping haveshown no evidence, so far, of susceptibility tostress-corrosion cracking.

Later W plants with main coolant piping of caststainless have also been impervious to stress-corrosion cracking; however, this resistance to*cracking was definitely expected. The cast stainlesssteel is a duplex alloy with significant percentagesof delta ferrite that renders these alloys greatlyresistant to stress-corrosion cracking.* The B&W and CE plants have ferritic steel pipingand therefore are free from the threat of stress-corrosion cracking because there is little or no pos-sibility of caustic or related compounds enteringthe primary system. Conceptually, the stainless-steel cladding would be vulnerable to cracking insome circumstances, but there has not been a prob-lem; nor is it expected in the wrought stainless pip-ing of the early 3M plants.

In summary, stress-corrosion cracking does notseem to be a problem in the primary piping systemof any of the PWRs. A caveat on this conclusion,however, is that the piping should nevertheless con-tinue to be inspected closely, since it is to be remem-bered that no stress corrosion was expected in theInconel steam generator tubes, but after a few yearsit appeared. Similarly, for BWR recirculation pip-ing, IGSCC was not expected but occurred.

5.3.4 Corrosion. Ordinary metal wastagebecause of any of the well-known electrochemicalmechanisms of corrosion has never been a problem

in the primary system piping of any of the PWRs;It is not expected to be in the future either. The'main reasons for this good experience are that oxy-gen is completely removed from the system, andother impurities that could lead to electrochemicalaction are likewise removed. Primary PWR coolantis extremely pure water.

5.3.5 Brittle Fracture. Brittle fracture is the mostfeared!of all mechanical-failure mechanismsbecause these types of failure are invariably sud-

- den, complete, and catastrophic. This mode of fail-ure is not considered possible in PWR primarypiping because all the piping materials that haveever been used in this application are more thansufficiently tough. Indeed, recent calculations forall plants have proved that leaks will occur wellbefore fracture is possible.

5.3.6 Erosion and Cavitation. Flow velocities inprimary systems are well within acceptable limitsfor austenitic stainless steel, whether used for pip-ing or cladding. As a result, erosion of main cool-ant piping has never been a problem, nor is itexpected to be one in the future.

Cavitation erosion is not an uncommon problemin the process industries. In the PWR primary sys-tems, however, there are operating restrictions toensure net pump suction head limits are neverexceeded. Similarly, the static overpressure is morethan enough to control cavitation at other flow' dis-turbances in the system.

5.3.7 Creep. In light-water nuclear. plants, allsystem temperatures are below the creep range, i.e.,l less than 0.3 Tm, where Tm is the melting point ofthe material. Creep and other time-dependentdeformations are insignificant.

5.3.8 Summary. In summary, the most signifi-cant degradation mechanism in the primary PWRpiping is low-cycle fatigue. The locations at whichthis occurs first are the primary piping nozzles. Ingeneral, the larger-diameter nozzles tend to be mostsusceptible, as well as those nozzles with the mostsevere transients (such as the safety-injection noz-zles). However, the true ranking of the nozzles isdifficult to determine because the results are depen-dent on the type and exactness of the analyses(which is variable) and the actual transient usage,which is somewhat unknown. In addition, eachmanufacturer analyzes the system somewhat

1.

I

I

I

63

differently. The general results for all significantdegradation sites are shown in Table 5.2.

5.4 Recommendations for FutureResearch

The fundamental and most significant damageparameter is fatigue because of system transient opera-tion. However, nuclear plant operators rarely keep acareful accounting of the number, type, and severity oftransients. The truth is that very few nuclear plantsknowv with any certainty how much of their design lifehas been consumed at any given time.

Therefore, it is strongly recommended here thatone or two plants, at least on a pilot basis, institutean in-depth study of past operations in order togain some insight into the rate that the design life isbeing consumed. The issue sounds simple, but it isnot. The plant was qualified to idealized descrip-tions of the transients. Actual plant transients donot conform to the idealization and are rarely assevere as the design descriptions.

The second recommendation is that programs beput in place now to monitor the actual degree ofthermal-aging embrittlement. This issue is not seen as acurrent problem, but it could turn out to be a problemif very long-term operation were to be achieved.

Table 5.2. Summary of degradation processes for PWR primary system piping

Rank ofDegradation Site

DegradationSite

DegradationMechanisms

PotentialFailureModes

IS'MethodsStressors

I Main coolantpipe nozzlesa

Terminal enddissimilarmetal weldb

Systemoperatingtransients

Systemoperatingtransients

Fatigue crackinitiation andpropagation

Fatigue crackinitiation andpropagation

Throughwallleakage

Throughwallleakage

Volumetricinspectionfor diameter:-4-in.

Surfaceinspectionfor diameter<4-in.

2

3 Caststainlesssteel

Temperature Thermalembrittlement

Throughwallleakage

Volumetric

a. The nozzles in the primary coolant piping are the highest ranking degradation sites. Additional specificity is avoided because themost severely loaded nozzles are difficult to determine. Reported results are heavily dependent on type of analysis performed. Actualdamage will completely depend on the actual transient usage that occurs in the plant. This is the basis for the recommendation.

b. In W plants, the dissimilar metal welds are at the reactor vessel and steam generator nozzles.

64

I a. ! ~5.5 References' r

1. L. F. Coffin, "A Study of the Effects of Cyclic Thermal Stresses on a Ductile Material," Transactionsof the American Society of Mechanical Engineers, 76, 1954, p. 931.

2. P. C. Paris, "The Fracture Mechanics Approach to Fatigue," in Fatigue-An InterdisciplinaryApproach, J. J. Burke, N. L. Reed, and V. Weiss (eds.), Syracuse, N.Y.: Syracuse University Press,1964, pp. 107-132.

3. "Criteria of the ASME Boiler and Pressure Vessel Code for Design by Analysis in Sections IIIand VIII, Division 2,". in Pressure Vessels ,and Piping:, Design' and Analysis, Volume 1, NewYork: American Society of Mechanical Engineers, 1972, pp. 61-83.

4. L. F. Kooistra, E. A. Lange, and A. G. Pickett, "Full Size Pressure Vessel Testing and Its Approachto Fatigue," Journal of Engineering for Power, Transactions of the Amnerican Society, of MechanicalEngineers, 86, 4, 1964, pp. 419428.

5. WV. H. Bamford, E. 1. Landerman, and E. Drar, "Thermal Aging of Cast Stainless Steel and ItsImpact on Piping Integrity," Journal of Engineering Materials and Technology, Transactions of theAmerican Society of Mechanical Engineers, 107, 1985, p. 53.

6. E. 1. Landerman and W. H.; Bamford, "Fatigue Characteristics of Centrifugally Cast Type 316Stainless Steel Pipe after Simulated Thermal Service Conditions," in Ductility and Toughness Consid-erations at Elevated Temperature Service, American Society of Mechanical Engineers, WinterAnnualMeeting, 1978, MPC-8.

7. G. Slama et al., "Effect of Aging on Mechanical Properties of Austenitic Stainless Steel Castingsand Welds," SMIRTPostconference Seminar 6, Monterey, California, 1983.

8. M. Vagins and J. Strosnider, Research Program Plan: Piping, NUREG- 1155, Volume 3, U.S. NuclearRegulatory Commission, July 1985. , .

65.

[

6. PRESSURIZED WATER REACTOR STEAM GENERATORSV Malhotra

6.1 Description

The steam generator is an important part of the pres-surized water reactor (PWR) primary system pressureboundary. Failure of the steam generator tubes willprovide a passage for the primary system coolant to theoutside of the'containment by'way of the relief valves.The recirculating and the once-through steam genera-tors shown in Figures 6.1 and 6.2, respectively, are thetwo types of steam generiators commonly used. In therecirculating type steam generator, the secondary sys-tem water is fed into the downcomer (Figure 6.1)where it is mixed with recirculating water that is drain-ing from the moisture separators. The downcomerwater flows to the bottom of the steam generator,across the tube sheet, and then upward through thetube bundle where steam is generated. The primarysystem coolant flows through U-tubes. It enters thesteam generator at 599 to 621OF (N-315 to 3270C) onone'side and leaves at 5501F (\2881C)'on the otherside.

A variation of the above recirculating design isthe use of a section of the secondary side as a feed-water preheater to achieve'greater thermal effi-ciency. In this, a section of tubes on the outlet orcold-leg side of the U-bends is partitioned off fromthe rest of the secondary side to form a preheatsection. Water is then fed to the preheat sectionrather than into the downcomer.

In the once-through steam generator, thesecondary-system water enters a feed annulus (Fig-ure 6.2) above the ninth tube support plate level.Here it is mixed with steam aspirated from the tubebundle area and preheated to saturation. The satu-rated water flows down the annulus, across thelower tube sheet, and upwards into the tube bundlewhere it becomes steam. This superheated steamflows radially outward and then down the annulusto the steam outlet connection.

Operating experience has demonstrated thatsteam generators have had and will probably con-tinue to have aging-degradation problems. Thesedegradation problems are mainly associated withthe recirculating type of unit that is manufacturedby either the Westinghouse Electric Company orthe Combustion Engineering Company. Some ofthese steam generators have developed thousandsof degraded tubes after just seven or eight years ofoperation. I The five- to ten-year span of operation

appears to be the critical timeframe in which manynuclear plant steam generators begin to exhibitsevere tube degradation.

In the existing steam generators, the tubes aremade of Inconel 600 with carbon steel as the sup-port material. Over the years, various defects havecaused tube failure as shown in Figure 6.3. A defec-tive steam generator tube is defined as any tubewhich has been plugged for whatever reason. Alarge number of tubes have been plugged, however,the tubes that actually leaked are a small propor-tion of the tubes plugged. Phosphate wastage wasthe major cause of steam generator defects in theearly 1970s. 2 From 1976 to 1979, denting was themajor cause of tube failures in PWR steam genera-tors. In 1980, stress corrosion cracking/intergranular attack (SCC/IGA) was the majorcause of tube defects (31qo), followed by denting(29To), and wastage (9016). In 1982, the majorcauses of tube'defects were primary side SCC(54.5%), pitting corrosion (11.7%), and secondaryside SCC/IGA (10.3%). The major causes of tubedefects in 1983 and 1984, were primary side SCCand secondary side SCC/IGA.3

6.2 Stressors

Stressors in steam generators are any load or.environment that tends to affect the functionalcapability of a part or component. Stressorsinclude chlorides, copper compounds, air leakageinto the system, type of water treatment, appliedstresses during fabrication, tube support design,pH, temperature, velocities, and crevices. Thesefactors, either in combination or separately, canproduce intergranular stress corrosion cracking(IGSCC), SSC/IGA wastage, erosion-corrosion,pitting, fretting, or denting.

6.3 Degradation Sites

The type of degradation acting on steam generatortubes varies; however, it can be divided into the twobroad categories of chemical and mechanical degrada-tion. The following is a short description of the variouschemical and mechanical degradation mechanismsand the sites at which they have been observed:

66

I'

viw) L11 F- iFE ~ aa~~~ts L]~~~~~~~~~F ~~DeformationQ ~~of tube U-bend

rged) ~~~~~~~~Denting[

LocalAe ~~~~~~~~~~~~~~~concentration

of Impurities

_,. _To blowdown(removal ofImpurities)

rng T~ube

ng lntergranularattack

Primary Inlet Primary outlet7-3079

Figure 6.1. Sketch of recirculating steam generator with indicated problem areas. 4

i

67

I

Upper spanlane regioncircumferentialcracks

- Uppertubesheet

- Auxiliaryfeedwaterinlet

-Steamoutlet

-Feedwaterinlet

Aspiratingsteam

Firstsupportplate

Lower tubesheet

7-3066

Figure 6.2. Sketch of once-through steam generator with indicated problem areas.4

68

C

IU

0)CL

76 ' 78 ' '80 ' 82 '84Year ' ' '

7-3050F u 63 H t y r t g r t f l m a

Figure 6.3. History of steam generator tube failure me'chanisms.3

* Wastage (thinning): General corrosion ofthe outside surface of tubing in the regionabove the tube sheet when phosphatechemistry is used. Sludge provides a mech-anism for the concentration of chemicalsthat attack the Inconel 600 tubing.

* Denting: The plastic deformation oftubes resulting from buildup of carbonsteel support-plate corrosion products(magnetite) in the tube-to-tube supportplate annuli.

* Pitting: A localized attack on the outsidesurface of tubing resulting from nonuni-form corrosion rates caused by the forma-tion of local cells.' Pitting can be severe inthe area of the sludge pile or where scalecontaining copper'deposits is'found. Pit-ting of Inconel '600 tubes takes place in thetemperature range'' 212 -to 392IF(100 to 200 0C), hence, pitting is 'foundmainly on the cold-leg side of the steamgenerator. However, Pacific NorthwestLaboiatories (PNL) has reported the pres-ence of pitting'on the hot-leg side in theoriginal Surry steam generator.

Intergranular Attack (IGA): Grainboundary attack on the outside surface ofInconel tubing with no stress orientation.This attack is noted in the hot-leg tube-to-tube sheet crevice region and tube-to-tubesupport plate-crevice region.

* Intergranular Stress Corrosion Cracking(IGSCC): The primary side IGSCCsometimes initiates on the inside surface ofthe tubing at the U-bends of tubes in theinner rows with small bend radiu' and theroll-transition region where the tube is

'expanded 'into 'the tube sheet. The'secondary side IGSCC initiates on the out-side'suiface'of the iubing'in the crevice"formed by the tube and the tube sheet, andtube and tube support plate. '

* Fretting: The loss of tube material causedby excessive rubbing of tubes'against theirsupport structure. This can be caused byeither primary side or secondary side flow-'induced vibrations 'of the tubes.' Severefretting has occurred at baffles in the inte-

''gral preheater nearjthe feedwater inlet, andat antivibration bars in the U-bend region.

69

.

* Erosion-Corrosion: The mechanicaldamage caused by a corrosive environmentand impingement of liquid or solid parti-cles in a system.. The carbon steel J-tubesin the early Westinghouse-designed steamgenerators have experienced wall thinningand perforation caused by an erosion-corrosion mechanism. The carbon steel J-tubes have been replaced by the J-tubesmade of improved corrosion resistantmaterials to avoid this type of degrada-tion, and therefore they will probably nothave a significant impact on the life exten-sion of these steam generators.

* Environmental Fatigue: This degradation isa result of synergistic effects of corrosion andfatigue. The locations of the degradation siteswithin the once-through steam generatorsinclude the secondary face of the upper tubesheet and the uppermost tube support platenear the open lane. The upper shell to transi-tion cone girth welds in a recirculating steamgenerator have also been degraded by corro-sion fatigue.

* Thermal Fatigue: The welds at the auxil-iary feedwater nozzle thermal sleeve inonce-through steam generators are sub-jected to thermal fatigue. The cracking ofthis weld will increase the potential of ther-mal shock to the steam generator shell.

6.4 Degradation Mechanismsand Failure Modes

Corrosion in recirculating steam generatorstends to occur in stagnant regions, where impuritiescan concentrate and attack the tubes. These regionsinclude the tube-to-tube sheet crevices, the tubesupport crevices, and the sludge zone above thetube sheet. The following sections describe themajor types of degradation mechanisms and theirfailure modes.

6.4.1 Wastage. Wastage, or tube thinning, orcorrosion of the tube material primarily takes placeon the outside of the tubes in the area of a sludgepile. The sludge collects in the steam generator andcauses chemicals to concentrate.4 ' 5 The result isthat the tube wall thinning occurs in the regionabove the tube sheet when phosphate chemistry isused. It is believed that the rate of thinning, in part,

is a function of the chloride and oxygen concentra-tions, as well as the sulfate concentration. Resinleakage from the condensate polisher bed is postu-lated to provide an acidic sulfate environment con-taminated with chlorides. The progression of tubewall thinning appears to be a serious problem onlyin those plants that are still on phosphate waterchemistry control. About 200 tubes were pluggedin. 1982 because of the wastage problem. Closercontrol of the phosphate water treatments wasattempted by a number of utilities in order to avoidsteam generator tube wastage and cracking, butthese efforts were unsuccessful. Thus in order tohalt these problems, many utilities changed the sec-ondary water treatment from phosphate to ammo-

-nia. Hydrazine chemistry [all-volatile treatment(AVT)h, in essence, eliminates phosphate problems.

6.4.2 Denting. Oxygen, chloride, and copperoxide and their compounds can form a voluminous(nonprotective magnetite) corrosion product on thecarbon steel support plate and subsequentlydeform some of the steam generator heat transfertubes at the tube-to-tube support plate intersec-tions. This reduction in tube diameter was subse-quently termed denting. The forces that cause theindentation result from the accelerated corrosion ofthe carbon steel support material because of chlo-ride contamination of the steam generator water.Chloride ingress from condenser leaks concentrateat the tube/support structure intersection and leadto fast linear growth of nonprotective magnetite.

Several nuclear power plants experienced steamgenerator tube-denting problems in the early 1970s.The incidents of denting were observed both inplants that were originally on phosphate chemistrycontrol and then switched to an AVT and plantsthat were on AVT from start-up. However, theplants experiencing tube denting generally used seawater or brackish water for condenser cooling andwere therefore susceptible to higher and moreacidic levels of chloride contamination from con-denser inleakage than were plants that used freshwater. Some steam generator tube denting has beenobserved in recent years in plants that are on AVTcontrol and fresh cooling water; however, onlyplants with sea water cooling have shown extensivetube denting to date.

Research has been undertaken to understand therole of certain parameters such as chemistry, tem-perature, heat flux, crevice condition, and materi-als in the tube-denting mechanism, and todetermine the threshold concentration levels atwhich denting occurs. Devices such as the

70

isothermal capsule, 'single-tHbe-mode[ boileiheated-crevice apparatus, and inleakage concentra-tor were used in these studies. Additionally,'existingdata from various operating plants were analyzedto understand 'the tube-denting mechanism. 5 Asummary of some of the conclusions derived 'fr6m'the above work follows. - '' '

An acidic chloride environment is required at thecrevice surface of the carbon steel support 'for theoccurrence'of denting'or nonprotective magnetite -formation (NPM). The threshold concentration of''strongly acid-forming chloride solutions (copperchloride or hydrogen chloride) for the formation ofNPM is between 175 and 700 ppm 'chloride.Sufficient NPM was produced in solutions withconcentrations 'of 700 to' 1500 ppm chloride 'tocause tube denting. The threshold concentration ofa weak acid-forming chloride solution for the for-"mation of NPM is also between 175 and 700 ppmchloride. The threshold chloride concentration insea water for NPM formation is between 2000 and10,000 ppm and for tube denting is 20,000 and90,000 ppm chloride. Oxygen in the makeup watermade a sea water environment more aggressive withrespect to the production of NPM. ' ' ': -'-'

In heated-crevice tests with prepacked crevices ina I1% sea water (200 ppm chloride) and copperoxide environment, a crevice superheat of only0.3O F (0.17'C), which yields'a concentration of2000 ppm chloride (based on boiling point eleva-tion) produced 'significant 'denting.'The 'denting'rate increased with initial crevice superheat in therange of 0 to I IF (0.56 0C). Above this, the dentingrate decreased with increasing superheat up 'to avalue of 23IF (12.81C). As the size of an initialopen crevice was reduced in these tests, a width of4 mils' was reached -for isothermal capsules 'ahd2 mils for heated crevice tests at which tube dentingis reduced or stopped.'

The rate of tube denting increased in isothermalcapsules containing 5800 ppm chloride 'as thetemperature increased from 450 to 600°F (2320C to316°C). Also, when the pH of the chloride solutionwas above 4, NPM was not formed even with chlo-ride levels of 35,000 ppm chloride.

Cooling tower water containing 0.34 ppm chlorideand silicon dioxide, which' is mildly alkaline-forningwhen heated to ste'am generator temperature, pro-'"duced denting-when copper oxide was added as an oxi-dant. Silicon'dioxide by itself produced slight dentingin a nondenting environment."'" -' ' '

Increasing the chromium content of "steeldecreased 'the tendency for NPM formation.Type 405 stainless steel in place of carbon steel did

not produce'NPM. Therefore the support plates inthe new' steam 'generators are made' of Type 405stainless steel.'Quatrefoil designs for support plates'are currently used to reduce further the damage due,to denting. ' ' '

Large quantities of hydrazine (40 ppm) in theblowdown were not immediately effective in stop-ping denting in either a copper chloride or an oxy-'gen environment.

Corrosion rates are a function of available super-heat, but they'also depended on the concentration ofchloride 'present. The existing plant sludge "datarevealed that'plants with a copper-to-iron ratio'greaterthan' I developed accelerated denting, while those with'a copper-to-iron ratio less than I had little or no indi-cation of denting. Current trends 'are for isea-water-'and brackish-water-co6led plants to use AVT uith con-.densate demineralization and titanium instead of cop-per as the condenser tube material.3

6.4.3 Pitting. Pitting'is an extremely localizedattack that results in -holes in the metal (pittingoccurs because of a localized breakdown in passiv:'ity).' In' most cases these holes are relatively smallwith a surface diameter about the same or less thanthe depth. It is difficult to detect pits because oftheir small size and because they are often coveredwith corrosion products.'Pitting usually requiresseveral months to years to show up. Once initiated,however, it progresses 'at an accelerating rate, andfailures often occur with extreme suddenness'.- Forthis'reason, it is difficult to detect pitting''and thebest fix for life extension is to eliminate the coridi-tions that lead to pitting:

With'the exceptions of Millstone-2 and-lndianPoint-3,' only shallow pitting has been observed inPWR steam'generator tubes removed from service.The maximum pit'depthsnireasured in'tubesremoved were "s0.005-in. Pitting 'was sufficientlyminor that these pits were not detected by'eddy-'current testing nor'did the existence of these pitspromote 'concern about'possible primary-to-secondary leakage. .

As a result of the extensive pitting; in theMillstone-2 and Indian'Point-3 steam generators(40%b throughwall pits), the Electric Power'-Research'Institute (EPRI) conducted a workshop 6

on steam generator tube' pitting. It 'was reportedthat'both affected 'plants had pits'of similar mor-phology (chromiuiri'oxide and copper metal-filledundercut pits). These pits were found under scale'and in similar locations (e.g., the cold-leg side near'the sludge ' pile). 'PNL has also 'reported the

71 ;"

________________________ I

presence of pitting in the hot-leg side in the originalsteam generator removed from the Surry plant.

It has been shown in the laboratory that it is possibleto pit Inconel 600 under a variety of conditions butonly a fewv of the laboratory-grown pits have the samemorphology that is found in operating plants.Electricite de France has produced pits in low-oxygen,high-chloride (740 ppb), high-sulfate (760 ppb) water.These pits had chromium-rich, copper-included scabssimilar to those found in the Millstone and IndianPoint steam generators. There is no universal agree-ment as to whether steam generator pit initiation andgrowth is primarily an operating temperature. (hot)phenomenon or a layup (cold) phenomenon. However,chromium oxide-filled pits are seen in high-temperature tests, but not in low-temperature tests.Also, laboratory tests indicated that the chloride con-tent of water and pitting are closely related, and thatpits grow under sludge or scale and are associated withcopper deposition. PNL reports that oxygen inleakagewill also cause pitting of steam generator tubes. Thereis no straightforward explanation about why pits gen-erally appear on the cold-leg side and what role, if any,resin ingress may play in the initiation and growth ofpits. Also, no mention is made of the effect of pittingon corrosion fatigue in the literature.

6.4.4 Intergranular Attack. IGA has beenobserved in Inconel 600 tubing in the hot leg andtube-to-tube sheet region of some of the operatingsteam generators. From analytical determinations,it appears that for IGA to occur, a highly alkalinecondition must exist in the crevice caused by theconcentration of alkaline species present in the sec-ondary side water. It was noted that a combinationof IGA and SCC is often present close to each otherin failed tube samples, with the: IGA being moreextensive than the SCC. Tubes removed from exist-ing plants7 indicate that both the outside diametersurface and the intergranular fracture face had inaddition to. the three major elements,(nickel,chromium, and iron), the presence of, sodium,potassium, calcium, phosphorus, sulfur, alumi-num, and chloride. . .

Tests. were conducted; using high-temperatureelectrochemical measurementsto identify condi-tions leading to this kind of IGA.8. The results ofthese tests indicated that in 10° Oocaustic media at6080F,,(3200 C), IGA is commonly observed inInconel 600. in the mill-annealed: condition.Thermally-treated material at 1292°F, (7000C)showed. definite improvement, over mill-annealedmaterial in both resistance to IGA and SCC.

The presence of various anions strongly influ-ences the corrosion attack. Carbonates and sul-fates, and to a lesser extent phosphates, are verydeleterious and can develop deep IGA and SCC,depending on the value of the electrochemicalpotential. The electrochemical potential of theInconel 600 tubes is governed by the compositionof the secondary water during operation. WhenAVT control is used, reductive conditions areencountered, and in the case of caustic pollution,IGA should preferentially occur in crevice regions.On the other hand, when oxygen enters the steamgenerator (either in the form of oxygen or metallicoxides), the potential is raised and favors an SCCmechanism. For this reason, the composition of thesludge and, in particular, the oxidizing potentialcould be the deciding factor in whether IGA or SSCwill occur.

Occurrence of SCC/IGA at or within the tubesheet can be treated by sleeving affected tubes. Thelarge-scale sleeving at Point Beach-2 andSan Onofre-l included many tubes with SCC/IGA. The possibility of SCC/IGA at the tube sheetmay also be reduced by expanding the tubes for thefull depth of the tube sheet and, thus, eliminatingthe crevices.

Secondary side SCC/IGA has also been found,recently, in the crevice formed by the tubes and tubesupport plates.3 This degradation mechanism atthe support plates caused several failures in 1983.The SCC/IGA at the support-plate intersectionscannot be eliminated by closing the crevices and it isof greater concern because it can occur at morethan one level in the steam generator.

6.4.5 Intergranular. Stress Corrosion Cracking.Increasingly, tubes have been removed from service as aresult of primary-to-secondary leaks at points of highstress concentration: the Row I U-bends, the rolledtransition regions in the tube sheet, and dented inter-sections. These cracks appear to have initiated on theinside surfaces of the tubing. It is not clear whetherinside initiated cracking of nondented units is a. spo-radic occurrence related to some flaw in the manufac-turing process, which imparts excessive stresses to thatsurface, or if it is a generic problem4 characterized by along initiation time and slow rate of progression orboth. Inside surface initiation of. tube cracking, isbelieved to be related to pure water or the Corion phe-..nomenon. In the first Corion tests, pure water wasshown to crack Inconel 600 at temperatures well abovethe primary side operating conditions. Laboratoriesare trying to define quantitative relationships between

72

failure time and variables such as 'strain rate, 'stress,environment, and temperature.

The outside surface initiation of SCC in the second-ary side of the tube-to-tube sheet crevice region w asdiscussed in Section 6.4.4.

6.4.6 Fretting. Tube fretting and 'wear is caused byvibration. Tube vibration can be induced by fluidcrossflow perpendicular to the tubes or by parallel flowalong the tubes.4 In the PWR steam generators, aclearance exists between the tube and the tube supportplate through which the tube passes. This clearance isrequired for manufacturing and design considerations.Vibration of the tubes can cause'impact and tangentialsliding of the tube against the tube support plate andadjacent tubes, resulting in local wear.

Fretting regions are sensitive to fatigue cracks.Under fretting conditions,' 'fatigue cracks can'initiateclose to the material 'surface at very low stresses, wellbelow the fatigue limit of nonfretted tube material. Ifthe vibration cyclic stresses in the tube' are sufficient tocause these fretting-induced cracks to grow and propa-gate, early failure of the tube can occur. ' ' ' '

Fretting has caused extensive tube damage at theantivibration bars in several steam generators. Thisdamage has been blamed on the poor design of theantivibration bars that were in point contact withtubes. The design has now been corrected, and thenew design provides a broader region of contactwith the tubes. ' -

6.4.7 Erosion-Corrosion. Impingement of liquidor solid particles is known to'cause mechanicaldamage, especially to protective surface films.This can happen to a variety of materials for certainimpinging velocities, sizes, shapes, and hardness ofparticles' and impinging angles.' When the' erosionfactor is' combined with a' corrosive environment,the mechanical damage can be accelerated.'

The impingement of solid particles on steam gen-erator tubes under operating conditions can'ncausemechanical damage. The probable sequence ofevents is that solid particles entrained in the bulkflow impact against the tube; promoting'erosion.In a noncorrosive environment, the erosion processremoves metal from the tube wall. In a corrosiveenvironment, the erosion process may first removea protective' film from the -tube, thus making thetube susceptible to corrosion.- In both cases, wallmetal loss occurs, either directly or by acceleratedcorrosion of the tube surface. Sources of solid par-ticles can include any particles entering the 'steam"generator and those particles entrained from withinthe steam generators.

Erosion-corrosion has occurred in 'once-throughsteam generators'around the fourteenth supportplate. This process has caused more than 40% wallreduction in some steam generators.

6.4.8 Environmental Fatigue. Some once-through steam generators, particularly early units,have an open lane, i.e., an untubed lane in the steamgenerators. 9 The steam flow carries entrained waterdroplets up the open lane to the upper, tube sheetregion. Evaporation of water concentrates any contam-inants in these droplets and also carries the chemicalsup the open lane to the upper tube sheet region. Herethe droplets impact the heat transfer tubes and dryout,thus further concentrating the chemicals on the steamgeneratortubes. Even chemicals that are generallyinnocuous in the feedwater can cause metal loss in highconcentrations.;

Tube samples removed from alongside the untubedlane revealed a serpentine band of shallow metal loss-sometimes containing microcracks on the tube outsidesurface just below the secondary face of the upper tubesheet or near the highest tube support plate. Severaltube failures, attributed to fatigue, are associated withthis band of metal loss. Circumferential transgranularcracks have also led to primary to secondary leaks. Theserpentine metal loss acts as sites for fatigue crack initi-ation, since the steam generator, tubes do undergocyclic vibration.

Laboratory tests have indicated that an acid solu-tion containing sulfates, silicates, and chlorides canproduce metal loss and microscopic indications(microcracks), closely resembling those observedfrom field-examined tubes.

6.5- In-service rinsp'ectio'n'.Methods''

A 'host of in-service equipment has beendesigned for steam generator inspection, some ofwhich have been' used 'on existing power plants.This equipment provides information about tubedenting,- intergranular. 'stress corrosion' cracking,pitting,'tube-to-tube support gap," etc.'"A briefdescription of some of ihe'techniques available tothe industry follows: ' '

* Hydrogen Evolution Monitoring as a Mea-:sure of Steam' Generator Corrosion.10High-temperature aqueous corrosion ofcarbon'steel in steam generators produceshydrogen and magnetite. Measuring the

73

hydrogen content of the exiting steam is, atpresent, the only available on-line meansof monitoring steam generator corrosion.It also can provide information about theeffectiveness of the corrosion inhibitorsand the effect of changes in water chemis-try. There is no available information as towhether or not this technique has beenused at any operating plant.

* Eddy-Current Nondestructive Examina-tion for IGA. Routine inspection of steamgenerator tubing is conventionally per-formed using a bobbin-type 'eddy-current'coil. I I However,- IGA defects have notbeen detected successfully in'all the'cases.Extensive copper deposits on the second-ary side of the U-bends and the'hot-legside of the tubes interferes with the eddycurrent inspection. Successful detectionhas been achieved by the use of pancakecoils that are arrayed with their axes paral-lel to the tube radius and multiplexed to'permit use of available electronic pack-ages. Field eddy-current data show thatvolumetric intergranular attack can bedetected when'the 'flaws are from 20 to30% through the wall, whereas an orientedIGA crack can only be detected when the

'flaws are about 60% through the wall. Ori-ented IGA flaws tend to be orienied'alongthe axial direction' within the tube sheetcrevice region, and along the'circumferen-tial direction at dented tube sheet intersec-tions and in the roll transition regions.(Additional discussions on the eddy-current examination technology are givenin Section 11.2.4 in Chapter 11 and Sec-tion 14.1.1 in Chapter 14 in this report.)

* EMAT System. This method addresses ultra-sonic inspection of steam generator tubingusing an electromagnetic acoustic transducer(EMAT).12 It was developed to detect. flawsin certain areas of tubes where conventionalsingle-frequency differential-coil eddy-current has had difficulty, including circum-ferential cracks, defects at dents or supportplates, and defects in U-bends.

The EMAT system has a good defect-detection capability for circumferentialcracks and other defects that provide afairly wide and sharp circumferential-oriented cross section (e.g., dented areas);however, the system has difficulty detect-ing flaws with small cross sections, i.e.,

cracks that are tight. For example, the sys-tem can inspect U-bends with its advancesignal but does not effectively detect theaxial cracks that are found in U-bends. Inaddition, the system has limited defectdepth-sizing capability.

* Optical Probe for Steam Generator TubeDent Measurement. 1 3 The optical profilome-ter has been tested successfully under labora-.tory conditions. It can measure inside tubingprofiles in the range of radii from 0.32 to0.40 in., with an average calibration error of0.005 to 0.008 in. on nominal inner diameterand + 0.006 in. on dents. Preliminary investi-gations suggest that the calibration error ismost probably caused by variation in interiorsurface finish.

* Pulse-Echo Ultrasound for Steam Genera-tor Tube-to-Support Plate Gap Measure-ment. 14 This equipment uses an ultrasonictechnique for determining the condition ofthe gap between steam generator tubes andsupport plates, which-allows monitoringcorrosion product buildup that might leadto denting. It also can determine the effi-ciency of chemical cleaning for removingthis buildup..

Other nondestructive methods include sonic leakdetectors and helium leak testing to search for airleakage into the system:

* Sonic leak detectors are used to check allof the potential leak locations during theperiod after a vacuum has been drawn onthe condenser but before the unit gets toload. This technique is used regularly, infinding steam leaks in boilers and has beenused on industrial vacuum systems. Its useon power plant vacuum systems is new.

* Helium leak testing is used to search for airleaks after the unit reaches partial load. Aload of at least 20% is necessary for effi-cient leak testing.

6.6 Summary, Conclusions, andRecommendations

IThe major causes of PWR steam generator tub-

ing failures were identified in this chapter alongwith the probable causes of the identified failure

74

I

modes. Conclusions were drawn,v here applicable,at the end of the discussion of each failure mode.

The degradation mechanisms for the steam gen-erator tubes are prioritized based on operatingexperience as summarized in Table 6.1. For exam-ple, in 1982, IGSCC was the number one cause oftube defects, followed by IGA, and so on. To con-trol or mitigate each failure mode, the followingrecommendations are made:

I. Wastage. Wall thinning of tubes has beenreported in plants that use phosphate waterchemistry. The number of tubes affected byphosphate water chemistry is small; this prob-lem can' be controlled by periodic lancing ofsludge to minimize tube degradation, affectedtubes may be sleeved, or the plant changed toAVT control.

2. Denting. Denting is no longer an impor-tant cause of tube failures. ln'1982, only1.4%o of the defective tubes were dented,mainly because of more stringent controlof the secondary side chemistry by reactoroperators. Some of the rmeasures that havehelped 'to control denting in older' steamgenerators include sludge lancing, mini-mizing condenser and air leaks, eliminat-ing copper alloys from the feed train, usingboric acid, installing deep bed polishers,and reducing the oxygen in the condensatestorage tank. The steam generators shouldbe subjected to the soak procedure toremove as much chloride as possible. Inthis method, the steam generators aresoaked at about 300'F (1490C) for at least8 h, with maximum bleed and feed, andthen the steam generators are cooled tobelow 200'F (930C), drained under anitrogen blanket and refilled with cleanwater. Newer steam generators that are bet-ter designed and equipped with morecorrosion-resistant tube supports made offerritic stainless steel (SS405) should notsuffer denting if good water chemistry ismaintained.

3. Pitting. Two reactors (Indian Point-3'andMillstone-2) reported severe secondaryside pitting of their Inconel 600 tubing.The pitting occurred in the cold legbetween the tube sheet'and the' first sup-port plate. Ingress of brackish water, air,and copper is believed to'have assisted thepitting process. To mitigate pitting, removeall copper material, minimize chloride and

air 'ingress,' change' to"AVT control, andmaintain strict chemistry control.

4. Intergranular Stress Corrosion Cracking.The only demonstrated method to preventinside-surface-initiated IGSCC in an over-'rolled tube is to peen the inside surface.This places all of the inside surface in com-pression where parts of it'after over-rollingare in tension and therefore subject toIGSCC. A Westinghouse method calledRotopeening has not been satisfactory insteam generators that have been operated;however, it has been successful in newsteam generators. Shot-peening has beenhighly successful in both'new and previ-ously operated steam generators.

The cracks resulting from inside-tubing-surface-initiated stress corrosion, caused byover-rolling, are difficult and in some casesimpossible to detect by nondestructive'exami-nation of the installed tubes because the tubesheet provides interfering signals. If crackingdoes occur, the tubes should be sleeved ortheir lower ends replaced.

1GSCC has also been reported in'Row Iof the U-bends in'several steam generators.Development of an in situ stress-relief pro-cedure is recommended to mitigate suchcracking. Testing has shown that thermallytreated Inconel'600 provides better resist-ance to IGSCC than, ,.annealedInconel 600. If leaks do develop in the U-bends, they should be plugged. The use ofthermally treated Inconel 690 for tubingfor new steam generators should be con-sidered because the available metallurgicaland corrosion data show that it is the mostcorrosion-resistant Inconel tube alloy.15

The French are already using this alloy fortheir steam generator" tubings.Kraftwerkunion AG uses Inconel 800 forsteam generator tubing and shot-peens itafter bending to introduce compressivestresses in the U-bend that eliminate stress'corrosion cracking.

'On -the secondary side, -SCC/IGA arewidespread problems that affected 14 reac-tors in 1982.; This problem led to replacementof the steam generators at Point Beach-l andlarge-scale sleeving at Point Beach-2. Tubedegradation occurred within the deep tube-to-tube sheet crevice in 11 reactors. The short-term solution to SCC/IGA is to install sleevesof either thermally treated Inconel 600

75 -

Table 6.1. Summary of degradation processes for steam generator tubes

DegradationRanka Site

DegradationMechanisms

FailureModesStressors

Is'Method

Inside surface ofU-bends and roll-transition regions

2 Outside surfaceof hot-leg tubesin the tube-to-tube sheet creviceregion

3 Cold-leg side insludge pile orwhere scalecontaining copperdeposits is found

Tube rolling stresses,Corion phenomenon,denting

IGSCC Cracking Eddy-current testing

Alkaline environment,presence of SO 4 andCO3 anions

Brackish water, air,and copper

IGA, IGSCC

Pitting

May eventuallyresult in cracking

Local attack andtube thinning mayeventually lead toa hole

Eddy-current testing

Eddy-current testing,optical scanner system,sonic leak detectorsystem

4 Outside surfaceof tubing abovetube sheet

Phosphate chemistry,chloride concentration,resin leakage fromcondensate polisher bed

Wastage (thinning) Uniform attack, tubethinning mayeventually wear outthe material

Eddy-current testing

5 Tubes in the tube-support regions

6 Contact pointsbetween tube and

- antivibration bar

Oxygen, copper oxide,chloride, temperature,pH, crevice conditions

Flow-induced vibrations

Denting

Fretting

Flow blockage intubes caused byplastic deformation

Wear out of materialcaused by rubbingand/or fatigue

Helium leak and sonicleak testing, opticalprobes, hydrogenevaluation monitoring,pulse echo ultrasoundmethod

Eddy-current testing

Once-through steamgenerator tubes

Velocities, sizes.shapes, impact angle,and hardness ofparticles

Chemicals, (localizedcorrosion) vibrations

Erosion-corrosionfrom impingementof particles

Wear out of material Eddy-current testing

8 b Once-through steamgenerator tubes inthe upper tubesheet region

Fatigue Primary to secondaryleaks

Eddy-current testing

ita. Based on operating experience for steam generator defects.

b. Denotes once-through steam generator (Items 7 and 8 do not reflect rank order). First six items are for recirculating steamgenerators.

or 690. A change to AVT control chemistryalso is recommended.

5. Fretting. Seven reactors in 1982 experi-enced fretting damage to their steam gen-erator tubes, caused by flow-inducedvibration. In two- cases, severe frettingoccurred at baffles in the integral pre-

heater near the feedwater inlet. Remedialactions included installation of modifiedimpingement baffles and plugging ofaffected tubes. Fretting caused by antivi-bration bar movement in the U-bendregion was observed in other units, whichis considered a design problem in the older

I

76

I'

units. In newer units, the problem is elimi-nated by reducing the gap size between theantivibration bars and the tubes..,

6. Erosion-Corrosion. Affected tubes should'be plugged. Stringent control on waterchemistry should also be applied. -

7. Environmental Fatigue. Most of the'previ-ously afflicted plants were partially by-passing their polisher during and shortlybefore most failures occurred. Action wastaken to reduce the vibrational loads on thetubes, and to provide strict control of waterchemistry (including restoration of lO0ocondensate demineralization) to preventingress or buildup of aggressive chemicals.

Blockage of the open lane with a lane-flow blocker is used now. This forces theopen lane's steam-water mixture out intothe tube bundle at a lower elevationthus,dispersing the damaging chemicals beforethey can concentrate at a weak point in theupper tube sheet location.

To mitigate the various degradation mechanisms asso-ciated with steam generators, the manufacturers havetaken or recommended the following steps:

Change from drilled tube support plates toother designs, such as, the quatrefoil

design in the Westinghouse Model F unitand the trifoil design for B&W steam gen-erators. Additionally, shifting from car-bon steel tube-support material to .l2ochromium ferritic stainless steel.

- Strict control on water chemistry on thesecondary side and a change from copperto titanium condenser tube material.,Wesiinghoiuse now recommends 'mostlythe AVT chemistry over phosphate treat-ment.,

* To control IGSCC, heat-treatedInconel 600 or 690 now is used instead ofannealed Inconel 600.

; *The use of shot peening has been highlysuccessful in the U-bend and rolled areasof the new and operating steam genera-

- tors. The fretting problem is eliminated inthe new steam generators by redesigningthe antivibration bars. The new design hasreduced the gap size between the tubes andthe antivibration bars.

There are'several issues associated with aging andlife extension of steam generators that require follo-wup. Life-assessment models are needed to determinethe degradation of steam generator tubes by the two

' potentially most tstatus of corrosion within nuclearsteam generator tubes. The significance' damagingmechanisms: IGSCC and IGA. Monitoring methods

' are needed to determine of the changes made in thedesigns and material selectio&'of the tubes and tubesupport plates needs to be evaluated. Also, usefulresults are expected from the EPRI program that isdeveloping improved nondestructive-evaluation meth-ods for steam generators, the USNRC steam generatorprogram, and similar programs in France, Germany,and Japan.

I I

77

6.7 References

1. C. A. Creary and E. L. Houchens, "Economics of Sleeving Steam Generators," Proceedings of theAmerican Power Conference, 45, 1983, pp. 811-817.

2. 0. S. Tatone and R. S. Pathania, "Steam Generator Tube Performance: Experience with Water-Cooled Nuclear Power Reactors During 1982," Nuclear Safety, 26, 5, September-October, 1985,pp. 623-639.

3. 0. S. Tatone, "Worldwide Tube Performance: Analysis of the 1983-84 Statistics," Nuclear Engineer-ing International, June 1986, pp. 81-83.

4. Steam Generator Owner Group, Steam Generator Technical Status Report, EPRI, January 27, 1982.

5. Causes of Denting, Volumes I through 5, EPRI NP-3275.

6. M. J. Angwin, Workshop Proceedings: Pitting in Steam Generator Tubing, EPRI NP-3574-SR.

7. G. P. Airey and F. W. Pement, "A Comparison of Intergranular Attack in Alloy 600 Observed in theLaboratory and in the Operating Steam Generators," National Association of Corrosion Engineers,Houston, Texas, March 22-26, 1982.

8. G. Pinard-Legry and G. Plante, Intergranular Attack of Alloy 600: High Temperature Electrochemi-cal Tests, EPRI NP-3062, May 1983.

9. M. 0. Speidel (ed.), Corrosion in Power Generating Equipment, Proceedings of the Eighth Interna-tional Brown Boveri Symposium, Baden, Switzerland, September 19-20, 1983, New York: PlenumPress, 1984.

10. C. R. Wilson, Hydrogen Evaluation Monitoring as a Measure of Steam Generator Corrosion, EPRINP-2650, November 1982.

11. S. D. Brown, Eddy-Current NDEfor IntergranularAttack, EPRI NP-2862, February 1983.

12. R. B. Thompson and R. K. Elsley, A Prototype EMAT System for Inspection of Steam GeneratorTubing, EPRI NP-2836, January 1983.

13. D. L. Oberg, Optical Probefor Steam Generator Tube Dent Measurement, EPRI NP-2863, February 1983.

14. E. S. Furgason and V. L. Newhouse, Evaluation of Pulse-Echo Ultrasound for Steam GeneratorTube-to-Support Plate Gap Measurement, EPRI NP-2285, March 1982.

15. C. E. Shoemaker et al. (eds.), Proceedings: Workshop on Thermally Treated Alloy 690 Tubes forNuclear Steam Generators, EPRI NP-4665M-SR, July 1986.

78

7. REACTOR PRESSURE VESSEL SUPPORTS FOR PRESSURIZEDWATER REACTORS AND BOILING WATER REACTORS

W. G. Hopkins

7.1 Review of Types of Supports,the Materials andEnvironment

There are five major types of reactor pressurevessel supports used in light water reactor (LWR)plants: neutron shield tanks, columns, cantileversupports, brackets, and skirt type supports asshown in the Figures 7.1 to 7.5, respectively. Thefirst four types of supports are used to supportpressurized water reactor (PWR) vessels, while thefifth type is used to support boiling waterreactor (BWR) vessels. The design responsibilityfor the four PWR types of support varies from ven--dor to architect/engineer, while the design respon-sibility of the skirt support lies solely with thereactor vendor because it is an integral part of theRPV. The major degradation mechanism for theRPV supports is irradiation embrittlement causedby neutron irradiation. Therefore,- the designs ofthe RPV supports'are discussed with respect to thesusceptibility of their materials, e.g., ferritic steeland weld metal, to irradiation embrittlement. Adetailed discussion of the damage due to irradia-tion embrittlement was presented in the Chapter 3of this report and therefore, is not repeated here.

A major factor in producing nil-ductility-transition-temperature (NDTT) shifts in supportmaterials is direct exposure to the reactor corebeltline neutron flux. The neutron shield tank is askirt-mounted, annular, water-filled tank formedby two concentric shells extending above and belowthe reactor core (as shown in Figure 7.1), and isfully exposed to the reactor core beltline neutronflux. WVater is circulated through the tank for cool-ing and shielding. Similarly, the column type sup-port is also fully exposed to the reactor core beltlineneutron flux. The cantilever type supports may ormay not experience significant neutron exposure;depending on a'design's specific location relative tothe core's location. The-bracket type supports aresufficiently elevated above the region of significantflux and are, therefore, not likely to experiencegreat shifts in their NDTT temperatures. The skirtsupports used in BWRs do not experience signifi-cant shifts in their NDTT because the BWR reactorvessels are larger in size than the PWR vessels andhave much more water between the reactor core and

pressure vessel that absorbs neutrons and shieldsthe pressure vessel, skirt and other hardware out-side the vessel. However, the BWR skirt supportsare subjected to fatigue damage when the RPVschange temperature and pressure during startupand shutdown.

In order for crack propagation to occur, the sup-port member must' be in tension at some point.With the neutron shield tank, column land skirttype supports, tension is experienced only underdynamic loadings, such as 'safe-shutdown-earthquake and loss-of-coolant-accideni (LOCA)loadings. The bracket and cantilever supports expe-rience tensile forces under their deadweight loads.

The Ndrth Anna units 3 and 4 have neutron shieldtanks. In 1979,'Virginia Electric and Power Companynotified the USNRC (in accordance with 10 CFR 21)that predicted end-of-life NDTT shifts would exceed

* the operating temperature of the supports. TheUSNRC commissioned the Naval Research Labora-tory (NRL) to perform a confirmatory study. In thatstudy, the NRL used displacements per atom (dpa) topredict NDTT shifts. This parameter was chosen sothat embrittlement or hardening data taken from onerieutron spectrum could be applied to the same mate-rial in a different spectrum. The end-of-life level (reac-tor life of 30 full-power years) of damage in dpa.wascalculated for the shield tank. The NDTT shift wascalculated based on the assumption that the elevationin transition temperature is proportional to the squareroot of dpa. From this result, it was found that theestimated end-of-life NDTT exceeded the service tem-perature of the shield tank. In addition, it was pre-dicted that the Charpy V-notch upper shelf energy levelwould be reduced to 30 ft-lb (41 J).

From this study, NRL recommended to theUSNRC that additional investigation's were war-ranted in the following areas1:

"1. Determination of the range of radiation envi-ronment conditions (neutron spectra and flux

l levels) for other shield tank support structuresand for support structures of the opencolumn design (column support). Compari-son of the findings with those for NorthAnna Units 3 and 4, relative to potential fornotch ductility reduction.

79

Incotr Inst

Neteon Reactor vessel supporte assembly unneltransfer1pipe ElI11.9 151161

I 2 tEI.31511816'11

I1~~~~ ~Neutron1detector Reactor neutron shield lank Plan

sahroud 0 2 4 6 81 _ l ~~~Ventilation Primary shield cooler Scale (It)

manlfotd 2.999 5pmn Reanif.... EL14.10'Nom 2.998' O

0.0005 I A(5 ) 43/8 _t24J\/Realort~015. ~ 14.Reactor vessel

shbiding 4. 4t 12-* 8' 3v _ J8 \-'~p ~ t471±0 w 1 UN 3A Del G

Access sil

1tank

X-X \ /support assembly D0 2 4 6 8 0 1 2

Scat.'4 (It)Req'd Seite tt}

0123456

8 -10 34 nom-Btil o Matertials

r .4 1.2eto - Reactor neutron shield tank

t2 112- No. Name Material

Dust cover 2 * 1mC () Plate A ASTMAA51t GA 60 FBX (normalizedM

ttont ' Plate B ASTM-A516 GR 60 FBX laustemlzd)2@ I _Et8 3 5/16 ( . PtateC -ASTMA240Type316L(usefor7 IDtubes)

k V (D~~~~~~~~ Pipe A ASTM.A312type3l8LSST®( PlpeC ASTM.AlO GR A or B carbon steel

Iry L.,, Reactor vessel support assembly.16 1t4 ItemL_-3 .1nom No. Name Material

0D0 D Bl Maraged steel 330.00 PSI min Y.R

0 1 2 3 O Ptate Maraged steel 270.000 PSI min Y.P.

Scale (ItI 4 Forging Maraged steel 270.000 PSI min YR sun

Figure 7.1. Neutron shield tank RPV support.

I

80

"6. If possible, conduct experimental irradia-tions, which will directly compare radia-tion effects of spectra in far-core versusnear-core irradiation locations to verifythe usefulness of the dpa approach formaking radiation damage projections."

If the above recommendations by NRL are fol-lowed, the estimate for the shift in NDTT will beimproved. An improved estimate for the NDTTshift will most probably result in increased marginsfor the plant life extensions.

Table 7.12,3 lists the steels most commonly usedin RPV supports. The NDTT shifts are a'functioniof irradiation temperature and neutron'fluence(E > 1.0 MeV). At lower irradiation temperatures[<4500 F (2320C)] the NDTT shifts are more pro-nounced because of the inability of the material toself-anneal. In addition, those steels with relativelyhigh copper contents will experience greater transi-tion temperature shifts. Table 7.24 shows the com-puted relationship between copper content andirradiation temperature on:the' expected NDTTshift: The effect of neutron fluence levels at a con-stant irradiation temperature on the expectedNDTT shift can be seen from Table 7.3.

The typical range of service temperatures for theRPV supports is 120 to 250IF (49 to 1210IC). Thetemperature can fall to values as low as 501F (280C)when the reactor is shut down. This low temperature,combined with the neutron environment'and coppercontent of the support material, can lead to significantshifts in NDTT. In addition, it has been found thatwhen using dpa as a neutron exposure parameter ratherthan fluence, (in both cases E > 1.0 MeV), greaterNDTT shifts are calculated (see Reference 5, Table 1).Comparison with test results suggests that dpa mightbe a better neutron exposure parameter to use.5

Table 7.4 is taken from examples presented in Refer-ence 5 and illustrates the effect of copper content andfluence versus dpa calculations of NDTT.

7.2 Review of the Design Basisfor RPV Supports

In considering plant-life extension (PLE) for theRPV supports a brief look must first be taken attheir design basis. The General Design Criteria of10 CFR 50 address the minimum design require-ments with regard to licensing that must be met forstructures, systems, and components important tosafety. In particular, Criterion 2 states that thedesign of a component should be able to withstand

Figure 7.2. Column RPV support.

"2. Development of more detailed informa-tion on the upper shelf notch toughness ofsupport structure materials in the trans-verse (weak) 'test orientation in the as-fabricated condition.

"3. Development of fracture toughness andstrength assessments for low upper shelfsteels irradiated at <4501F and a correla-tion of fracture toughness with Charpy-Vproperties in this condition.

"4. Assessments of the significance of radia-tion reductions in notch ductility and ele-'vation in yield strength to supportstructure fracture safe reliability.

"5. Develop, for the steels of. application,more detailed information on the potentialfor and sources of variable steel sensitivityto <450'F irradiation.

81 ':

.0 ~~91 09o k ~~~~~~~El. 566' L-~~~~~~.

Ad sold' >, 5'2'

@~~~~~~~~~~~~~~~- 151!X 57181rll4' grout

<,~~~'6 Section (E

rn support 2

El. 571' AO

tie 1 1/2' web

1 112 El. 568'-7/8'

82~~~~52

L~~~~~~~~~~~~

h-~~~~~~~~16- - 2- 6X

& ~~~~~2'0' 2'0' .

<3 ~~~~~~~~~~Section 2) 7-3071

Figure 7.3. Cantilever RPV support.

82

I'

. I ' I . :. I .i . . I

! � 1 ; , -

I : I ,

I ; , iI ;

I I

/

, Ventopening

I . ~Reactor ..1;:

\ ~ ' Anchior bolts openingsShear lugs

.. - -. l . ,,'.1 ' ... * -, , : 7.~~~~~73049

Figure 7.4. Bracket RPV support.

83

Control rod drivehydraulic lines

7.3070

Figure 7.5. Skirt RPV support.

84

I

Table 7.1 Common RPV support steel characteristics 2,3,a.b

Type of Steel

A572(normalized)

A302-B

SA533-BI

Classification

High-strengthlow alloy

Low alloy (notquenched andtempered)

Quenched andtempered

Initial NDTT(°F)

-'50

0

0

IrradiationcNDTT Shift

(OF)

85

85

I .8.5

MinimumYield Strength

(ksi)

40

50

so

SA516-70 Carbon-manganese 10

A572 High-strength 25(as-hot rolled) low alloy

A36 Carbon-manganese , 60-100

a. Private communication with P. B. Lindsay, Bechtel National, November 4, 1974.

b. Private communication with G. A. McCoy, Bechtel Power Corporation, January 28, 1976.'

c. Fast fluence (E > I MeV) = 1.23 x 101 n/cm2.

100

85

38

40

100 36

Table 7.2. NDTT shifts at a neutronfluence of 1018 njcm2

(E> 1 MeV) for various copper'levels, wvith 0.012% maximum 'phosphorus4

550°F Irradiation'' 450°F IrradiationCopper ND1r Shift! NDTT Shift(wt°0 o) (°F) (°F)

0.35 !105 .158

0.25 75 113

0.20 58 ; 87 ..,

0.10 : 25 38

Table 7.3. 'Dependency of NDTT change's: i- ' of A302B steel upon various

neutron fluence'levelsa

.

- ' Neutron'Fluence*NDTT.Increase .. at 250°Fr

(O°F) (n/cm2 > 1 MeV)

- 80 1 x 1018

85. 1.25 x 1018 .

100 . 2 x 1018 i

150, I; 4.1 x IV8 ,

' ';: '170 '' ''-'x5 XA1018

a. Private communication with P. B. Lindsay, BechtelNational, November 4, 1974.

85

Table 7.4. Computed NDTT at PWR supports based on neutron fluence (E > 1 MeV) anddpa (E > 1 MeV) 5

Fluence atCore Midplanea(E > I MeV)

(neutrons/cm2 )(A)

1.08 x 1018

4.86 x 1017

3.08 x 1017

dpa at CoreMidplane

(B)

9.03 x 10-4

3.96 x 104

2.92 x 10-4

CopperContent

(No)

0.4

0.1

0.05

NDTT `F(A)

110

23

11

NDTT 0F(B)

127

26

13

a. Assumed 40-year plant life and 80% capacity factor.

the most-severe-case effects of natural phenomenarecorded for the plant site and not lose the capabil-ity to perform its safety function. In addition, theeffects of natural phenomena must be appropri-ately combined with the effects of normal and acci-dent conditions. Criterion 4 deals with theenvironmental- and missile-design bases. It statesthat the design of a component should "accommo-date and be compatible with the effects of environ-mental conditions associated with normaloperation, maintenance, testing, and postulatedaccidents, including loss-of-coolant accidents."Also, "the components should be appropriatelyprotected against dynamic effects that may resultfrom equipment failures and from events and con-ditions outside the nuclear power unit."

In any combined event, the design-basis limita-tion and primary-failure mode is catastrophic brit-tle failure of the RPV support. Catastrophic brittlefailure is a result of three conditions being metsimultaneously. These conditions are (a) that aflaw of critical size is present, (b) that there issufficient load on the support member to develop acritical stress level at the crack tip of the flaw and,(c) that the service temperature is low enough topromote a cleavage fracture at the crack tip of theflaw.6 The NRL has developed a method that cor-relates the above needed conditions for cata-strophic brittle failure. The method is known as thefracture analysis diagram (FAD) 7 and is based on ageneralized stress-temperature diagram for fractureinitiation and crack arrest.

Use of the FAD to determine fracture-safe operationrequires determination of the NDTT, which can be

accomplished by either Charpy V-notch tests or bydrop-weight tests. The NDTT is defined as the highesttemperature at which the stress needed for fracture ini-tiation becomes the same as the stress taken from thecurve representing the' yield strength of the steel. Thenominal stress needed for fracture initiation decreasesin the presence of a flaw. As the flaw increases in size,the nominal stress required decreases. The FAD pre-dicts that, for a given level of stress experienced abovethe NDTT, fracture initiation is possible only with flawsizes equal to or greater than a critical flaw size. Thecritical flaw size is proportional to the square of thestress intensity KIc. The crack-arrest temperature(CAT) curve relates the temperature to the appliednominal stress at which propagation of brittle fractureis arrested. For stress levels at one half the yield strengthof a material and at temperatures 30 to 350F above theestablished NDTT, it was found that there was a defi-nite cutoff in fracture propagation. The CAT was,therefore, found to be directly relatable to the NDTT.Thus, the FAD method can be used to determinefracture-safe operation based on the NDTT of thematerial and expected stress levels. If the service tem-perature of the material in use never falls below theNDTT plus some safety margin temperature, then cat-astrophic brittle failure cannot occur. Should the serv-ice temperature fall below the NDTT plus safetymargin under certain loading conditions, the determi-nation of the critical crack size and location is requiredfor the analysis, using linear elastic fracture mechan-ics,8 to ensure continued safe operation.

Regarding the stress levels for RPV supports, thevarious loading combinations used in the design ofRPV supports should be reviewed. These loading

1.

rI.

9.1

Li

86

combinations are an explicit interpretation of the pre-viously discussed NRC General Design Criteria 2 and4 of 10 CFR 50 as implemented in detail bySection III of the American Society of MechanicalEngineers (ASME) code.9 Table 7.510 shows the var-ious loading combinations that must be accountedfor under the four designated design conditions.Table 7.610 lists the appropriate parts of Section IIIof the ASME code for obtaining the appropriatestress limits allowed for each design condition.

With regard to brittle failure, special attentionmust be directed to those design conditions whenthe service temperature is at its lowest, i.e., duringshutdown. If the applied loads, such as deadweightand safe shutdown earthquake, are high enoughwhen combined with the low temperature, then twoconditions for brittle failure are present.

Also of particular interest are the loadings asso-ciated with a postulated LOCA. It was discoveredin May of 1975 that asymmetric loadings on thereactor vessel supports because of a postulateddouble-ended rupture of the reactor-coolant piping'had not been taken into consideration in the origi-

Table 7.5. Design loading combinationsfor supports for ASME CodeClass 1, 2, and 3 components1 0

.. ...*Sa ......i .nal design of North Anna units I and 2 RPV sup-ports. These asymmetric loadings result fromforces induced on the reactor internals by transientdifferential pressures across the core barrel and byforces on the vessel caused by transient differentialpressures in the reactor cavity. I I Because of thesenewly identified loads, the USNRC on January 20,1978,12 required that all PWR plants evaluate theadequacy of their reactor system components andsupports under the effects of asymmetric LOCAloadings. Analysis of postulated guillotine breakshas shown that the additional loads experienced bythe supports are approximately equal to the com-bined loads of deadweight and SSE.

In addition to the above environmental parame-ters that influence the fracture toughness of RPVsupports and their design lifetime, one otherimportant environmental parameter must also beconsidered. This parameter is the degradationeffects of neutron radiation embrittlement' on theRPV support material discussed above and in'Chapter 3. Only the RPV and the RPV supportstructures in a nuclear power plant experience thisunique type of degradation mechanism. Conse-quently, Section III of the ASME Boiler and Pres-sure Vessel Code, Part NF-2121, requires that thematerials used in reactor pressure vessel supports"... shall be made of materials that are not injuri-ously affected by ... irradiation conditions to whichthe item wkill be subjected." A recent report by theElectric Power Research Institute13 has shown thatthere are significant fast neutron fluences (i.e.,> 1018 neutrons/cm2 at energies' greater -than> I MeV) in the annulus between the outer surfaceof the reactor pressure vessel and the inner surface'of the primary shield, in the region opposite thereactor core. As discussed previously, supports arelocated within'this region in some pressure-vessel-support system designs and are therefore subject toshifts in'their NDTT and subsequent decrease infracture toughness. In addition, supports'in thisregion are usually maintained'at or below '150 0F(660C);so thai the concrete in the priniary shieldmeets the concrete-codel 4 requirements for maxi-mum temperatures. Therefore,' any inherentannealing in the supports is nil compared'to theannealing in ihe reactor pressure vessel that is main-tained at temperatures of -5501F (2880C).'

Studies1 5 at the NRL have demonstrated the impor-tance 'of reucing the copper content in steels to sup-press both the magnitude of the NDTT shift and thedecrease in upper-level fracture toughness. The NRLalso has developed the drop-:weight test to ascertain thetrue NDTT. The USNRC has incorporated the results

DesignConditions

Normal

Upset

Emergency

Faulted

Design Loading'Combinationsa

DW + OBE'DW+ RVCDW + FV-DW + OBE + RVO

'DW + DU

DW + DE : -

DWDWDW

+ SSE + RVO+SSE ,+ DF

a. Legend:DW Piping dead weight . . ::OBE Operating basis earthquake (inertia portion)SSE Safe shutdown earthquake (inertia portion)FV Fast valve closureRVO Relief valve - open system {sustained)DU Other transient dynamic events associated with the

upset plant conditionDF Dynamic events defined as a faulted condition

87

Table 7.6. Allowable stress limits for Class 1 component supports10

Conditions

Support Type

Plate and shelldesign by analysis

Linear type supportsby analysis

Componentstandardsupports design byanalysis

Componentsupportsdesign by loadrating

Design

NF-3221

NF-3231

NF-3240

NF-3260

Normal

NF-3222

NF-3231

NF-3240

NF-3260

Upset

NF-3223

NF-3231

NF-3240

NF-3260

Emergency

NF-3224

NF-3231

NF-3240

NF-3260

Faulted

- NF-3225

NF-3231

NF-3240

NF-3260

of this research into Regulatory Guide 1.99,Revision 216 for prediction of radiation damage inreactor pressure vessels.

With the emergence of the North Anna syndrome17

and the' additional loads it imposes on reactor vesselsupport systenms, it is important that any decrease infracture toughness and increase in NDT1 in reactorpressure vessel supports be assessed in PLE consider-ations.

Regulatory Guide 1.99, Revision 1,4- was used tocalculate the increase in NDTT and 'assesses thepotential lifetime of a RPV support in the followingexample.

1. Deterrnine the actual current NDTT of theRPV support material by way of samplingand monitoring techniques as describedbelow in Section 3 orby,',ay of analyticalestimates as described in the following steps.

2. Assess the fast neutron fluence to which theRPV support material will be exposed over its'iormhal design' basis ' lifetime" (e.g.,1018 neutrons/cm2 've a 40-year lifetimewith an 80% plant capacity factor. This isequivalent to 'a fast neutron' flux of109 neutrori/cM2-s). This can be done usingreactor physics 'calculations' as outlined inAmerican Society' for Testing andMaterials (ASTM) Standard E 482.!

3. Deternine the lowest temperature the RPVsupport material will experience [e.g., 50F

(28 0C)]. This temperature is usually set in atechnical specification in the license of thefacility.

4. Determine the initial NDTr of the RPV sup-port material from the material mill certifica-tion records [e.g. -230F (-21.5 0C)J.

5. Determine the maximum loading conditionsat the lowest service temperature. This estab-lishes what safety margin must be used basedon Pellini's FAD procedures, e.g., at half theyield strength, the FAD indicates that thematerial must be at NDTT + 300F(NDTT + 170C). For this example, it isassumed that the location of the maximumtensile stress (half of yield strength) in thesupport coincides with the location of maxi-mum fluence. Then, for this example, themaximum NDTT shift must be given by

I

I

500F - 30 0F - (-23 0F)

(initialNDTT)

(lowestoperatingtemperature)

(FADsafetymargin)

Ii= 43 0F

(maximumNDTT shift)

88

6. Using Regulatory Guide 1.99 (see Table 7.4),assess the NDTT shift of 251F (14'C) for thenormal operating lifetime for -the supportmaterial's weight percent copper. However,because the. supports are irradiated below450'F (2320C), the 251F (141C) shift must beincreased by 50%, (i.e., the supports in theexample with 0.10°7 copper and 0.012%phosphorus, and a fast neutron fluence of1018 n/cm2 have experienced an NDTT shiftof 380F (20'C) over their 40-year lifetime).

7. The available PLE margin for temperaturesless than 4500F, can now be determined in thefollowing manner:

NDTT margin for PLE NDTrmamum

':--- ND ;TT4O years

NDTT margin for PLE = 430F - 380F

=50 F

- 8. Referring to Figure 7.6, a 30F (20C) increase15SF (30C), decreased by 50°7 for irradiationat temperatures greater than 450'F] in NDTTwould result in an increase of the fast neutronfluence from the 40 year value of %l.0 x 1018' ieutrons/cm2, to' a fast'' fluence' of'1.2 x 1018 neutron/cm 2. Recalling fromStep 2 that the neutron flux was 109

- neutrons/cm 2-s for our example, the PLEmargin is given by i .

1.2 x JOB I - 10X1o8PLE margin = 1

.^109.

.O- ; .._A~kt) ,; , , ,

recent exemption granted by the USNRC when con-sidering loads from post-LOCA guillotine breaks.Recent advances in fracture mechanics has shownthat there is an extremely low probability of adouble-ended :break occurrinig in PWR primarycoolant loop piping. Instead, it has been found thatthe PWR primary coolant pipe is much more likelyto leak before it breaks, so that detection of smallflawvs by either in-service inspection or leakagemonitoring systems is assured. This has led to amodification of General Design Criterion 4Requirements by' the'USNRC. 'A statement hasbeen added as follows:

"... the dynamic effects 'associated withpostulated pipe ruptures of primary cool-ant loop piping in pressurized water reac-tors may be excluded from the design basiswhen analyses demonstrate the probabilityof rupturing such piping is extremely lowunder'design basis conditions." 18

The analysis referred to 'above is outlined inNUREG-1061, Volume 3, dated November 1984.19Once an extremely low probability (104 per reactoryear) of rupturing is ascertained, redesign of the RPVsupports is possible. In particular, support loadingsbecause of the dynamic effects of postulated pipe'breaks can be eliminated. This'resuit, in effect, extendsthe lifetime of the RPV supports because the remain-ing loads (dead weight + earthquake) are about onehalf of the total design basis loads caused by the asym-metric loads from a'LOCA. Thus, if a support steelhad been chosen with a yield strength of 100 ksi with50 ksi ensured for NDTT + 300F (NDTT + 170C),dropping the loading from 50 ksi to 25 ksi implies thata larger fast neutron fluence may be tolerated.' In theprevious example, where the final NDTT must be430 F (60C) because' of the, NDTT + 30 0F(NDTT + 171C),' a new -design basis can be used of!NDTT'+ 150F (NDTT +'91C) (for one-quarter'yield strength from the FAD) or the final NDTT is 50'-15°F- .c.(-231;) ='581F (14.50 C) . Compared to the

old value of 430F (61C), an additional 150F (90C) PLEmarginris thereby realized:.Again using the aboveexample, with a 1018 fast neutron' fluehce at 40 years,this PLE margin of 151F (91C) can be translated fromRegulatory Guide :1.99 as before into an extended life-,time.The 150F(90C)isonlyworth 10F(5.50C) recallthe S0%'perialty for <4500F (<2320C) irradiation];therefore, 'a250F (141C) NDTT 1380F (201C), forT <4500 F (2320C)] -corresponding' 'to a startingfluence of 1018 would result in a final NDTT of 35F(1.70 C). This corresponds'to fast'fluenice from!Figure 7.6 of 2 x' 1018 neutron/cm 2. Proceeding asbefore -

= 2.0 x 10' seconds

where .'' ,

- - ! i, ; -d' neutronfast flux

Ot ' = neutron fast fluence

PLE margin '= 6.3 years. ' ' '

A reexaminiation of design'-basis loads can also' -result in a similar PLE gain. This is because of a

89

C,0Ua00)a

QC30

(C

0U

0

400 11fir" 1000 (% Cu 140'-� 0.08) + 5000(0/0 P -0.008)JIf/10191.1 M 11W Will Till M11111111111i 11 [M.P111-MM U '4- -1;114114111111111 JiMMU1111i ii !W

300 lit 1.111111 111111 I IM110 11 ININ 4� 14" lull III 11MI HIM I IM 6,11 R.URM1111 11 11MIM-1141 iMW1111111iii 1$11 t 11111MIt II

200 M1110111111 ME N $1,1111 1111_6010 M1111 oil

S IR Hills 1110-1100 JWrI 11111111t Iffillp! 111111111IM1141111

i Aforl I1. I I I III 11111110 0 Wli!' 4- J-44 I II IMIN1104 1111till :17 : 1:111 1111111Di"llit1h Hilllit 111IM-1- UK 11111") D�Hnl I 1-1:11 .1 11 I 111p, 11k I-1 t.air, 11111IM111-111 101 ft Lf IWIT NM 41444N 11 I

so 1111 111MIM HT111 illi "IN t'11111M Hill.35 0.30 11-0�% 10.20%cul 0.15% CU 0.10% Cu

419141;11111114= 44441-1-11� a/.p n niq TTM MIR life Lower limit1� MHH-1:L-r jtdf ExtrapolI I dI I IT.,

2 x 1017 4 6 8 1018 ,. 2 4 6 8 1019 2Fluence, n/cm2 (E > 1 MeV)

4 6

7.304B

Figure 7.6. Predicted adjustment of reference temperature, "A", as a function of fluence and copper content. Forcopper and phosphorus contents other than those plotted, use the expression for "A" given on the figure.

PLE margin = = 2 x 1018- 1014) 1019

= I x 109 seconds

Additional PLE margin = 32 years.

The above two examples show that certain RPVsupports may have a PLE margin of several years.However, to develop a valid and more acceptableapproach for calculating RPV' support margins,more research is needed to determine the NDTTshifts :due to both low energy neutrons(E < I MeV) and lower irradiation temperatures[<4500 F (232 0C)]. The purpose of the aboveexamples was to show the importance of the load-ing assumptions. However, it should be noted thatthe correlations for the temperature shifts in Regu-latory Guide 1.99, Revision- l, are based on highenergy neutrons (E > I MeV), while the neutronspectra in the RPV supports have a very large com-ponent of low energy neutrons (E <I MeV).Therefore, the NDTT shifts as developed from dpagradients (similar to Equation 3, RegulatoryGuide 1.99, Revision 215) may be more appropri-ate then the use of Regulatory Guide 1.99,Revision 1.

7.3 Recommended Strategies forRPV Supports PLE

There are two basic options that can be simulta-neously pursued for RPV supports PLE. They areadministrative relief and monitoring and sampling.

The first option, administrative relief by way ofthe new GDC4 will result in the elimination ofdesign basis loadings because of a double-endedbreak of the reactor coolant piping. As previouslymentioned, this one tactic alone provides a goodmeans of extending the use of RPV supports.

The second option of monitoring and samplingallows an actual NDTT evaluation of the supportsthat have experienced operational neutron radia-tion exposures. Once the neutron environment ofthe RPV support has been assessed per the follow-ing ASTM Standard E 1035,20 the correspondingshift in NDTT can be calculated, 5 based on thetype of material employed in the support.

In ASTM Standard E 1035, it is recommendedthat all pressurized water reactors whose supportssee a lifetime neutron. fluence (E > I MeV) thatexceeds I x 1017 neutrons/cm 2or3 x 104 dpa con-sider monitoring as a first step in evaluating anypossible NDTT in the supports.

I

I

90

U

I

The standard outlines state-of-the-art methodsin dosimetry that are now possible to properly char-acterize the neutron field near the RPV supports.In particular, it describes how scrapings from theRPV or the supports themselves can be used as pas-sive dosimeters through the use of a relatively newand very sensitive technique21 known as heliumaccumulation 'fluence monitors (HAFM). In thistechnique, the helium atoms released from the reac-tion Cu(n,a) are counted and are a measure of the,neutron field.

Furthermore, the standard suggests the refine-ments in the reactor physics calculations that areneeded to properly assess the analysis of the inte-gral dosimetry data and the final prediction of-NDTT. For example, if the vessel supports do notlie within the core's'active height, then an asymmet-ric quadrature set'rmust be chosen' for discrete-ordinates calculations' that' will accuratelyreproduce the neutron transport in the direction of-the supports. Care also has to be exercised in con-structing the quadrature set to ensure that raystreaming effects in the cavity air gap do not distortthe calculation of the neutron transport. If the sup-port system is so large or geometrically complexthat it actually perturbs the general neutron field inthe cavity, then the analytic method of choice is acoupled discrete ordinates/Monte Carlo calcula-tion. The normal coupling for this type of problemis to perform the two-dimensional discrete ordi-nates analysis only within the vessel. The neutroncurrents generated by this analysis are then used tocreate the appropriate cumulative distributionfunctions in the final Monte Carlo analysis.

After the monitoring data have been unfolded, esti-mates of the exposure in dpa can be made using ASIMpractices E 944 and E 1018. The dpa then can be con-'verted to NDTT using the newest relations outlined byMcElroy.22 If these monitoring results indicate apotential for the PLE margin, then it also may be pru-dent to take actual samples of the'RPV support toconfirm the'current NDTT by way of NDTT tests(cf. ASTM E 208). The structural considerations ofremoving suitable samples without 'any subsequentrepair are discussed below.

Four basic RPV support types have potential forsampling if the reactor cavity is large enough toallow access. These include the column, bracket,cantilever, and neutron shield tank type (seeFigures 7.1 through 7.4). Samples are assumed tobe removed from an area adjacent to the mid-planeof the core usually considered to be the location of

highest neutron fluence. However, it should bepointed out that this should be confirmed by waykofreactor physics calculations and by way of thedosimetry monitoriig outlined above or both. Thisis because new low-leakage fuel patterns may causethe maximum neutron fluence to occur off the reac-tor core's horizontal midplane. Based on the above'assumptions, a preliminary review indicates thatpermanent removal of a specimen from the neutronshield tank type is not feasible because it is a fluidretaining vessel. -,

However, for the bracket, cantilever, or columntype supports removal of a sample may be practi-cal. The design basis for existing RPV supports are,in general, extremely conservative. This is, in part,because of the unrealistic loads or load combina-tions that may have been assumed in their designbases. Large design margins such as those previ-'ously explained regarding pipe break exist, as wellas state-of-the-art refinements in analytical tech-niques used in the original design.:

7.4 Other Factors Affecting PLEfor RPV Supports

The above sections have discussed the primaryarea of concern with RPV supports for PLE, i.e.,the potential catastrophic failure of embrittled sup-ports. This section will address those other. factorsthat, although they are not as significant as brittlefailure, are still worth considering primarilybecause their synergistic effects in a nuclear powerplant environment have not been the subject ofdetailed research programs.-

I

I

* Corrosion. One of the factors present inall LWR containments is a greater thanaverage humidity'a an elevated tempera-ture coupled with the presence of a radia-tion environment. -The effect of thegamma radiation can be important incausing increased corrosion rates synergis-tically with the high humidity and temper-atures. The long-term effects of such anenvironment on the integrity of the RPVsupports should be examined because thetotal gamma dose alone on the supportswill be over 5000 Megarads in the first40 years of life. RPV support corrosionduring the original license period is notexpected to be a problem; only minor localpitting has been observed to date.

[

r

91

I

I

* Radiation damage to nonferritic parts ofthe RPV support system. Some' supportsystems depend on a dry lubricant, e.g.Lubrite, which is located between the sup-port and the RPV nozzle. This, materialhas a radiation threshold dose of2.2 x 103 Megarads. Therefore, any PLEplans should evaluate the actual radiationlevels the lubricant has received and willreceive over a PLE period. This evaluationwould determine if supplement or replace-ment of the lubricant is needed.

* Stress corrosion cracking (SCC) of threadedparts in the sliding foot assembly. The threadsare coated with Heresite. The threshold stressfor initiation of SCC is 150 ksi while theapplied tensile stresses are calculated to be'about 20 ksi, therefore SCC of the slidingfoot assemblies is not expected.

* Fatigue. The skirt type supports for BWRreactor pressure vessels are subjected tofatigue because the expansion and contrac-tion cycles associated with thetemperature- and pressure-induced expan-sion and contraction of the vessels duringthe plant startup/shutdown cycles. How-ever, the fatigue usage factor during thefirst 40 years of operation is expected to bewell below 1. (This 'degradation mecha-nism for the skirt-type supports is also dis-cussed in Chapter 9 of this report.)

7.5 Summary, Conclusions, andRecommendations

This chapter of the report has discussed the primarycontrolling phenomenon for PLE for reactor pressurevessel supports. Five basic types of supports used inLWRs were examined: neutron shield tank, column,cantilever, bracket and skirt type supports.

The major conclusion reached was that the onlyimportant potential failure mode for the neutron shield

tank, column, and cantilever type supports is cata-strophic brittle failure due to irradiation embrittle-ment. Table 7.7 summarizes these findings and alsopoints out other possible degradation mechanisms(corrosion), stressors, and in-service inspection meth-ods. The bracket and skirt type supports will experi-ence no significant NDTT shifts due to their location.Therefore, BWRs as a class of LWRs will have toaddress only fatigue for PLE because they are all skirtsupported. Because some Babcock & Wilcox PWRsare also skirt supported, they, too, should be consid-ered to be in this class.

A method to predict PLE margins for RPV sup-ports based on a combination of the USNRC Regu-latory Guide 1.99, various ASTM standards ondosimetry for RPVs, and Pellini's FAD was dis-cussed in Section 7.2. In addition, it may be possi-ble to get actual support material for NDTT tests.,It was demonstrated that by using only the adminis-trative relief for leak-before-break (GDC-4 of10 CFR 50) that margins of several years couldprobably be realized.

As a result of the. review conducted for thisreport, the following are recommendations for fur-ther study:

1. Develop fracture toughness and strengthassessment data for the RPV support steelsirradiated at temperatures less than 450'F(2320C). Develop a correlation of fracturetoughness versus Charpy V-notch proper-ties at temperatures less than 450'F(2321C) and as a function of dpa and neu-tron fluence (E >I MeV). '

2. Determine the range of radiation condi-tions (neutron spectra and flux levels) inand around shield tanks, and cantileverand column-type support structures.

3. Investigate the'effects of the expected radi-ation levels due to extended operation onthe lubricants between the RPV nozzlesand supports.

I

I

I

I

I

92

i.

I

Table 7.7. Summary of degradation processes for PWR and BWR RPV supports

Rank Degradation Sitea StressorDegradation ' PotentialMechanism Failure Modes ISI Methodb

I Neutron shield tank atthe core horizontalmidplane elevation

2 Column support at thecore horizontalmidplane elevation

Neutron irradiation, ' Neutron embrittlement, Catastrophic brittle IMonitorinigCoperating temperature, corrosion failurewater chemistry L 1 -, :

Neutron irradiation, _Neutr n embrittlement Catastrophic brittle Monioring and samplingoperating temperature ' ' ' I ' failure

I

3 Cantilever support inthe active height ofthe core

Neutron irradiation,tensile stresses, operatingtemperature

Neutron embrittlement Catastrophic brittlefailure

Monitoring and sampling

4 Threaded parts insliding foot assembly

5 Dry lubricant insliding foot assembly

6 Skirt support

Tensile stresses, operating Stress corrosion Binding that causestemperature cracking possibly excessive

stresses in the primarycoolant system duringheatup and cooldown.

Neutron irradiation, Degradation caused by Binding that causesoperating temperatures neutron irradiation possibly excessive

stresses in the primarycoolant system duringheatup and cooldown

' Mechanical and thermal Fatigue Ductile overloadstresses ' ' - ' ' - ' : -' failure

rL

I

a. The bracket and skirt RPV supports will experience no neutron embrittlement.a.. Th ...c .sup

b. There are no national standard in-service inspection methods per se to determine state of degradation.

c. Monitoring of neutron field near RPV support.I_

1; ;'.: "I'

1`1

93

I7.6 References

1. R. E. Johnson and R. P. Snaider, "Radiation Effects to Reactor Vessel Support Structures," U.S.Nuclear Regulatory Commission memorandum to D. G. Eisenhut and S. H. Hanauer,November 23, 1979:

2. L. E. Steele, Neutron Irradiation Embrittlement of Reactor Pressure Vessel Steels, Technical ReportsSeries No. 163, International Atomic Energy Agency, Vienna, 1975.

3. G. A. Knorovski, R. D. Krieg, and G. C. Allen, Jr.; Fracture Toughness of PWR Components Sup-ports, NUREG/CR-3009, SAND78-2347, February 1983.

4. U.S. Nuclear Regulatory Commission, "Effect of Residual Elements on Predicted Damage to Reac-tor Vessel Materials," Regulatory Guide 1.99, Revision 1, 1977.

5. W. C. Hopkins and NV. L. Grove, "A Study of the Embrittlement of Reactor Vessel Steel Supports,"Reactor Dosimetry, Volume 2, J. P. Genthon and H. Rottger (eds.), Dordrecht,Netherlands: D. Reidel Publishing Company, 1985, pp. 621-628.

6. W. S. Pellini, Principles of Fracture-Safe Design, Pressure Vessels & Piping: Design and Analysis, 1,American Society of Mechanical Engineers, 1972, pp. 275-293.

7. W. S. Pellini and P. P. Puzak, FractureAnalysisDiagram ProcedurefortheFracture-SafeEngineeringof Steel Structures, NRL 5920, March 15, 1963.

8. G. R. Irwin, Fracture of Metals, Cleveland: American Society for Metals, 1948.

9. American Society of Mechanical Engineers, ASME Boiler and Pressure Vessel Code, Section III,Subsection NF, July 1, 1974, Summer 1975 Addenda.

10. Bechtel National, Incorporated, Bechtel Standard SafetyAnalysis Report, Chapter 3.9, Tables 3.9-11and 3.9-12, October 2, 1975.

11. U.S. Nuclear Regulatory Commission, Resolution of Generic Task Action Plan A-2, AsymmetricBlowdown Loads on PWR Primary Systems, NUREG-0609, January 1981.

12. Victor Stello, Jr., letter to all PWR licensees from U.S. Nuclear Regulatory Commission Director ofthe Division of Operating Reactors, Office of Nuclear Reactor Regulation, January 20, 1978.

13. M. L. Griztner, G. L. Simons, T. E. Albert, and E. A. Straker, PWR and BWR Radiation Environ-mentsfor Radiation Damage Studies, EPRI-NP-152, September 1977. r

14. American Concrete Institute, Code Requirements for Nuclear Safety Related Concrete Structures,ACI 349-76, Appendix A, June 1977.

15. J. R. Hawthorne, Further Observations on A533-B Steel Plate Tailored for Improved RadiationEmbrittlement Resistance, NRL 7917, September 22, 1975.

16. U.S. Nuclear Regulatory Commission, "Effects of Residual Elements on Predicted Radiation Dam-age on Reactor Vessel Materials," Regulatory Guide 1.99, Revision 2, February 1986. P

17. ."Docket 50338-207 North Anna Power Station, Units I and 2," Summary of Meeting Held onSeptember 19, 1975 on Dynamic Effects of LOCAs, September 22, 1975.

94

I

18. U.S. Nuclear Regulatory Commission, "Modification of General Design Criterion 4 Requirementsfor Protection Against Dynamic Effects of Postulated Pipe Ruptures," Federal Register, 51, 70,April 11, 1986.

19. U.S. Nuclear Regulatory Commission, Report of the U.S. Nuclear Regulatory'Commission'Piping'Review 'Committee,' Evaluation of Potential for Pipe Breaks, NUREG-1061, Volume 3,November 1984. . - ;

20. American Society for Testing and Materials, Standard Practice for Determining Radiation Exposuresfor Nuclear Reactor Vessel Support Structures, ASTM E 1035-85, Volume 12.02, 1985.

21. American Society for Testing and Materials, Method ofA pplication andAnalysis of Helium Accumu-lation Fluence Monitors for Reactor Vessel Surveillance,' ASTM Standard E 910-82, Volume' 12.02,1982.. - - 1 - :

22. W. N. McElroy et al.;"Trend Curve Exposure Parameter Data Development and Testing," ReactorDosimetry, Volume I, J. P. Gentihon and H. Rottger (eds.), Dordrecht, Netherlands: D. ReidelPublishing Company, 1985, pp. 113-135.-.

Ii

r

.I I ,

I..1

. . I - : .

T I .

I . I I I .

- . . . : . . . -- : :; !:; ' - . .-

;:, , ;, ' ' -. ''' ', . :- '; * ' :!. ' '

: :' ' .- . .'. . .'. :t ' ' ' : ' : ' ' - .'.'

_ ', 1 | i .'' . I ' . . ;

* t * - *, ! . . ); . .q . i t . Pa I !- t : i e .: . ......... g . f ) i i i . ., ......... . s -; # .: ,' P , . 1

-- , ' , ' i ' *. ; 1 " f .' J. '- t "'' '' $

. > t ; ; * , z -; - s , t S

I . ' ' ' '-'.1..'. . ' ..-,'. . '' ' ,''. i'' .;."'''- .'.''.

.: -' t ...... : x . t . -;: i: *t. A-: S f ?i. J ' '!;,' ', _ '' i' ......... _-; . > s; 's'.; ' . '". '. . .

, ': ' ' ,'',', ' * S ,' . ; , - , ....... ,., , . i

.; 1- (- . - -:, - -,. ! -. , -- - . S - ; . :;

.. ; .J

, ,, : :: ! . '

95

r

TI

I -~~~~~~~~~~~~~~~~

8. RANKING OF MAJOR BOILING WATER REACTOR COMPONENTSV N. Shah

This chapter identifies the major components ofinterest in commercial boiling water reactors (BWRs)for continuing safe operation and life extension. [Pres-surized water reactors (PWRs) were discussed in Chap-ter 2.1 In identifying these components, it is assumedagain that the lifetime of the small, less-expensive com-ponents, i.e. sensors, controls, batteries, certain typesof pumps, valves, motors etc., is much shorter than40 years and these components will be maintained,refurbished, and replaced frequently. Therefore, theresidual life of these small components is not signifi-cant to life extension of the plants and is not addressedin this section.

The major components reported here are identi-fied and prioritized according to their relevance toplant safety only. The industry-sponsoredMonticello pilot plant project has also identifiedcomponents for plant life extension study. 1,2 How-ever, the identification and prioritization criteriaused by the industry pilot project were based onplant safety, reliability, and cost.

8.1 Key BWR Components

The two criteria employed to select the key BWRcomponents are the same as those employed inChapter 2 to select the key PWR components: tocontain any release of fission products that maytake place during any accident, and to maintain anacceptable level of radioactivity in the containmentduring normal operation. The pressure boundarycomponents, containment, control rod drive mech-anism (CRDM), reactor pedestal, and biologicalshield are included in the selected key components.

Major BWR Components

Many of the reasons for the above ranking of themajor BWR components are the same as those dis-cussed in Chapter 2 for the PWR components. How-ever, the degradation of BNVR pressure vessels causedby neutron embrittlement is less severe than that ofPWR vessels, and therefore it is not ranked Number 1.Also, some of the degradation sites and the potentialdegradation mechanisms of the major BNVR compo-nents are different than those identified for the PWRcomponents and are discussed below.

The containment and basemat are'ranked firstbecause they are the most critical major components,as far as public safety is concerned. N-ost of the earlyBWRs have a Mark I pressure suppression contain-ment system. This 'contairini-nt concept features a reia-tively small drywell connected by a system of vent pipesto another containment compartment commonlyreferred to as a wetwell. The other BWR containmentconcepts (Mark 11 and Mark III) also have a wetwelland drywell but are of larger size. The foundation ofthe drywell of the Mark I containment is made of con-crete. Two demonstration BWR plants, i.e., Big RockPoint (71-MWe capacity) and Dresden-i (207-MWecapacity) have spherical metal containments withoutwater suppression.

The potential degradation mechanisms for theMark I drywell metal shells are corrosion of the out-side surface and fatigue. The potentially life-limitingmechanisms for the suppression chamber are corrosionof inside and outside surfaces and fatigue of the torusmetal shell and vent pipes.

The BWR reactor pressure vessel differs from thePWR vessel in the way it provides support to the reac-tor core. In the PWR vessel, the lower core supportstructure carries the weight of the reactor core andtransmits it to the reactor vessel head flange. In theBWR vessel, the control rod drive (CRD) stub tubesprovide vertical support for the CRD housings, guidetubes and mechanisms, and the fuel assemblies. Inter-granular stress corrosion cracking (IGSCC) has beenfound at the weldments joining the stub tubes to thelower head. Also, the degradation of BWR reactorpressure vessels by neutron embrittlement is less severethan that of PWR vessels because there is a lot morewater between the BWR fuel and vessel because of thelocation of the jet pumps. Most BWR vessels have anexpected end-of-life fluence (E >I MeV) of about5 x 1O17 n/cm2 , while typical PWR vessels have anexpected end of life fluence of I x 1019 n/cm2 .Therefore, vessel fluence and pressurized thermal

I

1.2.3.

4.5.6.7.8.9.

10.

Containment and basematReactor pressure vesselRecirculation piping, safe ends, and safetysystem pipingRecirculation pump bodyControl rod drive mechanismsCables and connectorsEmergency diesel generatorsReactor pressure vessel (RPV) internalsReactor pedestalBiological shield.

96

I

shock are not significant issues in BWRs.3 However,irradiation embrittlement cannot be totally discountedwhen assessing the residual life of BWR vessels. Thenew Regulatory Guide 1.99, Revision 2, takes intoaccount Cu and Ni content in the vessel materials, anddictates higher transition temperature shifts than thosepreviously determined by Regulatory Guide 1.99,Revision 1, for BWR vessels. Fatigue is the most likelydegradation mechanisms for BWR reactor vessels.

The feedwater nozzles of the BWR reactor vesselconstitute an important part of the primary coolantpressure boundary. Cracks have been found in thefeedwater nozzles at Monticello and 17 other BWRplants.4 The feedwater nozzles are subjected to rela-tively severe cyclic thermal loading because of turbu-lent mixing of incoming cold water 1340 to 4350F (171to 2240C)] and the hot water [5450F (2850C)] returningfrom the steam separators and dryers. The thermalloading is more severe during startup and shutdownbecause the incoming cold water is at a temperature ofabout 1000F (381C). The main degradation mecha-nism is high-cycle thermal fatigue. The stainless steelcladding on the nozzle surface contributes' to thefatigue cracking because of the different thermalexpansion'coefficients of the stainless steel claddingand the carbon steel base metal.

The key BWR piping components that are sus-ceptible to aging are the recirculation piping, thesafe ends, the feedwater piping, and the core sprayspargers. After more than ten years of operation,several piping components in the Monticello planthave been repaired or replaced. 1 The core spraysafe end was replaced in 1981, and some recircula-tion piping welds were weld overlayed in 1982. Therecirculation piping and safe end were replaced in1984. Similarly, the recirculation piping in thePilgrim power station has been replaced.5 Severalother BWRs have experienced similar failures at therecirculation piping safe ends.6 The potential deg-radation mechanisms for the recirculation pipingand its safe ends are IGSCC, crevice corrosion, andthermal fatigue.3 The feedwater spargers distributethe feedwater such that the flow is evenly distrib-uted to the jet pumps. Sparger heads have experi-enced cracking at the flow holes because of fatigueand IGSCC.4 BWR spray headers also have experi-enced similar fatigue and IGSCC problems. Forexample, the Peach Bottom spray header experi-enced significant cracking and has been repaired. 7

The key locations of the potential degradation sitesand potential degradation mechanisms for the recir-culation pumps, CRDMs, cables and connectors, andemergency diesel generators are likely to be similar tothose discussed for the corresponding PWR compo-

nents. If there are any variations in the degradationprocesses, they will be reported later.

The key RPV internal components susceptible toaging degradation are the core shroud, jet pumps,fuel support pieces, steam separators, core plate,and steam dryers. The CRDM,- feedwater pipingand spargers, and core spray spargers that were dis-cussed above also are key RPV internal compo-nents. The reactor internals provide positioningand support of control rods, fuel rods, and othercomponents as in PWRs. However, the design ofthe BWR internals is quite different than that of thePWR internals. Their failure may cause binding ofthe control blades, thus prohibiting their insertioninto the core. This binding -may lead to an opera-tional transient without scram. Failure of the BWRinternals also may cause fuel rod cladding degrada-tion, fuel relocation, wet steam because of failureof dryers, and loss of forced recirculation flowbecause of failure of the jet pumps. Type 304 stain-less steel and Inconiel 600 are 'the principal materi-als of the reactor internals that see high fluences.Neutron irradiation of Type 304 stainless steel willincrease its yield and tensile strengths, decrease itsuniform elongation, cause relaxation, lead toIGSCC, and reduce the Charpy impact upper shelftoughness. Neutron irradiation also reduces thelow-cycle fatigue life of Type 304 stainless steelcomponents. Among all the reactor internal com-ponents, the core shroud is most susceptible tothermal shock following a design-basis accident.3

The main degradation locations in the jet pumpsare the holddown beams, restrainer brackets, andsensing lines.8 The holddown beam is fabricated.from heat-treated Inconel 600 and has failed in sev-eral BWR plants because of lGSCC.9 The restrain-ing brackets have failed because of improperassembly, and sensing lines have failed because ofvibrations. The jet pumps have been successfullyrepaired under field conditions, and therefore, thedegradation of their subcomponents may not havea significant impact on the life extension of BWRreactor internals. The failure of the steam separa-tors and steam dryers will result in poor qualitysteam passing through the steamlines to the tur-bines but probably will have no safety significance.The wet steam may cause erosion and corrosion inthe steam piping and the first stage of the turbineblades.10 Other potential degradation mechanismsfor the BWR internal components are low-cyclefatigue, stress-corrosion cracking, and high-cyclefatigue.

The reactor pedestal and biological shield are madeof concrete. The key locations and potential

TI

.S

97

I

degradation mechanisms are the same as those for theconcrete in the PWR containment structures. How-ever, the reactor pedestal is subjected to high-cycle,low-amplitude vibration and thermal cycling. Anadditional loading on the biological shield is gamma-ray heating resulted from the proximity of this com-ponent to the core.

8.2 Summary, Conclusions, andRecommendations

The major BWR components were selected andranked such that any release of fission productsduring an accident will be properly contained. Thecontainment and basemat have been identified asthe most important safety components in BWRplants. This means that a failure of the contain-ment and basemat will have a major impact on thesafety of the power plant. The other key compo-nents, according to their ranking, are reactor pres-sure vessel, recirculation piping and safe ends,recirculation pump body, CRDMs, cables and con-nectors, emergency diesel generator, RPV inter-nals, and RPV supports and biological shield. Theresults presented in this chapter are based on past

experience, as reported in the available literature.These results will be updated as more data are madeavailable.

Table 8.1 summarizes the aging related informa-tion on the key BWR components discussed in thischapter. It lists the components and gives reasonsfor their ranking. Table 8.1 identifies the mostlikely degradation site and other degradation sitesfor each component and also lists the most likelydegradation mechanism and other potential degra-dation mechanisms.

The major components reported in this chapterare selected and prioritized according to their rele-vance to plant safety only. These criteria are differ-ent than the ones used in the industry pilot study.Therefore, the ranking of the major componentspresented in this chapter is somewhat different thanthe one in the pilot study. However, all the majorcomponents identified in these chapters are alsoranked high in the pilot study. Detailed discussionsof the degradation processes for BWR pressure ves-sels and recirculation piping are presented in Chap-ters 9 and 10, respectively. Discussion of thedegradation processes of both PWR and BWRRPV supports was presented in Chapter 7.

I

rL.

I

i

98

I

Table 8.1.. Key BWR components for residual life assessment

Degradation Sites Degradation MechanismsRank Component Reasons for Ranking (most likely, others) (most likely, others)

I Containment andbasemat

Public protection duringan accident

,I

Concrete, metal liner, and Environmental degradation ofreinforcing steel in suppression concrete, fatigue and corrosionpool and primary containm'ent of metal containmentswall, base metal, welds and ventpipes in metal containment

2 Reactor pressurevessel

Primary pressure boundary, Nozzles, closure studs, beltline Thermal fatigue, neutronpoor operating experience region, welds at the stub tubes embrittlement, IGSCCwith RPV nozzles

3 Recirculationpiping, safeends, and safetysystem piping

Primary pressure boundary, Safetends, austenitic stainlessrelatively poor operating steel fittingsexperience

IGSCC, thermal aging

4 Recirculationpump body

Primary protection fromany internal failure ofpump elements, primarycoolant pressure boundary

Heat-affected zones nearweldments in the wall

Thermal aging, corrosionfatigue, crevice corrosion

5 CRDM

6 Safety-relatedcables andconnectors

Failure may lead to areactivity-initiatedaccident, an anticipatedtransient without scram,or a small-break loss-of-coolant accident

Active during normaloperation in mitigatingoperational transients andaccidents

Drive rod assembly, control rod Wear and thermal embrittlementpressure vessel

[

Cable insulation, inserts inconnectors

Thermal aging of insulation,thermal embrittlement andcorrosion ofconnectors

E7 Emergency diesel

generator

8 Reactorinternals

Needed to operate criticalsafety equipment in theevent of a loss-of-offsitepower

Failure may cause fuelfailure or problems inscraming the reactor

Governor in the instrumentation Fatigue and vibrationsand control system, cooling andlubrication system pumps andpiping, fuel injector pumps,turbocharger, generatorwindings

Core shroud, top guide plate, Irradiation-assisted stresscore plate, holddown beam in a corrosion cracking,jet pump, feedwvater and core intergranular stress corrosionspargers cracking, fatigue

r

L

9 Reactor pedestal Failure will challengeintegrity of RPV

Concrete and reinforcing steel Thermal cycling of concrete,corrosion and fatigue ofreinforcing steel

10 Biologicalshield

Maintains acceptable levelof radioactivity incontainment

Inside surface of the biological Neutron irradiation, gamma-rayshield at the core horizontal heating, environmentalmidplane level degradation of concrete,

corrosion of reinforcing steelsI.

99

8.3 References

1. Nuclear Plant Life Extension, EPRI/DOE Workshop, Dearborn, Michigan, August 1985.

2. Seminar on Nuclear Power Plant Life Extension, Seminar Proceedings, cosponsored by EPRI, NorthernStates Power, U.S. DOE, and Virginia Power, Alexandria, Virginia, August 25-27, 1986.

3. Northern States Power Company, Updated Safety Analysis Reportfor the Monticello Plant, April 30, 1982.

4. R. Snaider, BIVR Feedwater Nozzle and Control Rod Drive Return Line Nozzle Cracking, NUREG-0619(issued for comments), April 1980.

5. L. C. Chan and G. H. Hart, "Experience of Lowering A High Drywell Air Temperature at thePilgrim Nuclear Power Station," International Meeting on Nuclear Power Plant Maintenance,Salt Lake City, Utah, March 23-27, 1986.

6. R. A. Carnahan and R. MN. Horn, Evaluation of Safe-End Weld Materials and Safe-End ReplacementExperience, EPRI NP-4443, February 1986.

7. N. P. Alexakos, "Peach Bottom Atomic Power Station U/3 Core Spray Header TEE Box BracketInstallation," International Meeting on Nuclear Power Plant Maintenance, Salt Lake City, Utah,March 23-27, 1986.

8. J. E. Charnley, "Repair of BWR Jet Pumps," InternationalMeeting on NuclearPolverPlant Mainte-nance, Salt Lake City, Utah, March 23-27, 1986.

9. R. S. Mills et al., Closeout of LE Bulletin 80-07: BWR Jet Pump Assembly Failure, Parameter, Inc.,Elm Grove, Wisconsin, for the U.S. Nuclear Regulatory Commission, NUREG/CR-3052,November 1984.

10. G. A. Delp, J. D. Robison, and M. T. Sedlack, Erosion/Corrosion in Nuclear Plant Steam Piping:Causes and Inspection Program Guidelines, EPRI NP-3944, April 1985.

100

r

I

9. BOILING WATER REACTOR PRESSURE VESSELSW. R. Mikesell and W. L. Server

9.1 Description

This discussion for boiling water reactor (BWR)pressure vessels is not as detailed as that for thepressurized water reactor (PWR) pressure vesselpresented in Chapter 3. However, the same impor-tant degradation mechanisms (namely fatigue andirradiation damage) are considered for the BWRvessels, as will be discussed. Primarily, the recollec-tions and experience of one of the authors are pre-sented; therefore, some of the details and specificsshould be considered as close approximations.

BWR vessels have been manufactured throughoutthe years with the same basic geometry shownrinFigure 9.1. They consist of a cylindrical shell and ahemispherical bottom head containing both the con-trol rod drive (CRD) penetrations and the in-coreinstrumentation penetrations. The top head also ishemispherical and includes nozzles with boltingflanges attached. One of the nozzles is for venting(pressure relief) and top head cooling spray, and theothers are spares. The top head is bolted to the cylindri-cal shell by means of ^.-in. (150-mm) diameter studs.The flanges are butt welded to the shell and to the headand are furnished in one piece. The earliest BWRs wereof several sizes/geometries and represented demonstra-tion and small power reactors (i.e.,'Dresden-1,Big Rock Point, Humboldt Bay, and La Crosse). Theearly commercial-generation BWRs consisted of threesizes (not including Duane Arnold which was smaller).These sizes were \.201-in. (5.11-nm), 218-in. (5.54-r),and 251-in. (6.38-m) diameters. For example, theVermont Yankee and Monticello vessels were 201-in.(5.11-m) diameter, Zimmer and Brunswick-I and -2were 218-in. (5.54-m) diameter, and Quad Cities-I and-2, Susquehanna-l and -2, and Limerick-] and -2 wereall 251-in. (6.38-m) diameter. These are just examplesand do not include all of the plants of any given size.

The latest BWR model, which was designated as;.BWR/6, was built in two sizes. These were 218-(5.54-m) and 238-in. (6.05-m) diameters. Examples ofplants of these sizes are River Bend-l and Clinton-l forthe 218-in. (5.54-m) diameter and Grand Gulf- IandPerry-l for the 238-in. (6.05-m) diameter. Quite a fewof the larger diameter vessels were constructed for utili-ties in this country (many of which have been can-celled), while the other 218-in (5.54-m) diameter vesselswere built for overseas plants (i.e., Laguna Verde andKuo Sheng).

The predominant vendors of BWR vessels in thiscountry have been Babcock & Wilcox (B&W), Com-bustion Engineering (CE), and Chicago Bridge andIron (CB&I). 'One foreign-produced vessel fromBabcock/Hitachi was fabricated for Hope Creek. Oneof the vessels finished by CB&I was partially fabricatedby Rotterdam Dockyard. Rotterdam Dockyard didhave a contract to construct one of the vessels for theBlack Fox units that were cancelled. In total, CB&Ibuilt most of the BWR vessels used in theUnited States.

All of the major nozzles of the vessel where placed inthe cylindrical shell, but they were 'not located at theelevation of the reactor core (see Figure 9.1). The noz-zles in the lower portion of the shell were, in general,limited to the primary coolant recirculation inlet andoutlet nozzles which were located in a circumferentialband a short distance above the attachment of thecylindrical shell to the bottom head. Nozzles above thecore region were designed for core spray, feedwater, andthe steam outlet. The largest nozzles were the two forthe recirculation outlet lines (primary coolant outlet) atthe bottom part of the shell; they were typically 26 in.(660 mm) in diameter. Few, if any, of the other nozzlesexceeded 12- to 15-in. (305- to 380-mm) diameters.

All the nozzles in the vessels.built by CB&I wereforged and provided a lip at the shell end of the forgingto permit butt welding of the nozzle to the shell. Reac-tor vessels furnished by at least one other fabricatorincluded nozzles which were forged but did not containa lip so that the attachment weld of the nozzles to theshell was a corner weld.

The earliest BWR reactor vessel was Dresden-I. Theplates used in its construction were procured to therequirements of American Society for Testing and-Materials (ASTM) A300 for fracture toughness(Charpy V-notch) testing and American Society ofMechanical Engineers (ASME) Code Case 1280 whichprovided requirements for fine grain melting. Before

__the introduction of the SA533 Grade B, Class I(SA533B-I) plate specification, SA302 Grade B(SA302B) plate was used. All vessels since about 1965have been constructed of SA533B-I material (thisincludes the plates used in the shell and in both heads).It appears that all BWR vessels have been constructedusing plates for the shell and the heads. That is, therehave been no forged rings, nor forged head segments,used in this country.

The forging materials used for the nozzles inDresden-l were A105 11 (equivalent to present A105).

I

I

I

FL

U.

101

I

Vent andtop headspray _ ,Vessel head

Closure studs

V

, Main steamoutlet

r

Jet pump/recirculatingwater Inlet ,

Skirtsupports -

Control roddrive -

Recirculatingwater outlet [

-Shield wall

E.1.

7.3045

Figure 9.1. lypical BWR pressure vessel.

102

I

Probably, this material was used for most of theSA302B vessels, before the introduction of the SA508specifications. All of the more recent vessels have usedSA508, predominantly Class 2, (SA508-2) althoughsome Class 3 material (SA508-3) has been used.

Before the introduction of the SA540 bolting specifi-cation, the studs were furnished to an A193-B7 specifi-cation. Subsequently, the studs have been SA540, B22,or B23 for the 130 ksi (900 MPa) yield strength level. Itis possible that some studs were furnished at the140 ksi (965 MPa) yield strength level.

Until about 1974, the CRDs were attached to a stubtube that was welded to the inside of the bottom headof the reactor vessel. Until about 1966 or 1967, the stubtubes were made of stainless steel. Thereafter, they weremade of Inconel (Alloy 600). From about 1974 to the'present, the CRDs were welded directly to the bottomhead without an intervening stub tube. The in-coreinstrumentation penetrations were always weldeddirectly to the bottom head. The vessel is supported bya skirt welded to the bottom head just below theattachment of that head to the cylindrical shell, as isshown in Figure 9.1. The design of that skirt in termsof load-carrying capability is governed by the weight ofthe vessel and its contents and includes considerationof the operating basis earthquake loadings.

As mentioned earlier,'all fabricators made the headsand cylindrical shells of BWRs with formed andwelded plate. CB&I purchased the heat-treated platefrom the mill, and then cold formed and welded ittogether. The welding used proprietary welding fillermetals that generally matched the chemistry and prop-erties of the base material. The other primary fabrica-tors of BWR vessels in this country, B&W and CE,purchased unheat-treated materials from the platemills, hot formed the material, electroslag welded thejoints, and heat treated the complete assembly after therings were complete. This heat treatment consisted ofaustenitizing followed by tempering. All vessel maiiu-facturers welded the circumferential seams by the sub-merged arc process and completely stress relieved thevessel in accordance with ASME code requirementsafter the welds were made. In some cases, an interme-diate postweld heat treatment may have been 'made forindividual ring assemblies before joining' them in acomplete vessel.'

The earliest vessels were designed to the require-ments of the ASME Code, Section 1, and the 1270series of code cases. All vessels manufactured after1963 were furnished to the 'requirements 'ofSection III of the ASME Code. The mechanics ofmaking the design and satisfying the code analyti-cal requirements are pretty much uniform and con-ventional regardless of the vendor. Of course, the

analytical methods did change as available com-puter programs became more developed. An exam-ple of this evolution would be the evaluation of thediscontinuity analysis'from the old beam on anelastic foundation analysis to computer programsfor closed form numerical integration of the shellequations and finite clement methods.

Problems that have occurred throughout the yearshave resulted in changes to both design and materials.For instance, the safe ends on the nozzles of BWRswere originally forged of austenitic stainless steel andstress relieved after welding to the vessel. The occur-rence of intergranular stress corrosion cracking(IGSCC) in the primary coolant lines of BWR plantsled to the removal of these postweld heat-treated aus-tenitic safe ends' and replacement with safe ends madeof austenitic materials that had not been stress relieved.Additionally, some safe ends were fabricated ofInconel. Similarly, all of the nozzles of BWRs wereoriginally stainless steel overlayed. However, thermalfatigue cracking as 'experienced at Millstone-I as aresult of bypass flow and a loose thermal sleeve'in thefeedwater line. Subsequently, the BWR vessel nozzleshave not been overlayed.

There also has been an evolution in the design of theattachment to' the inside of the nozzles for thermalsleeves and the recirculation inlet lines. As mentionedabove, there was cracking at Millstone-l because ofbypass of feedwater flow around the safe end of thatnozzle. The thermal sleeve was not integrally attachedto the nozzle. All of the lines in these vessels have sincebeen repaired to provide for a welded attachment topreclude'this problem. Similarly, the attachment to theinside of the recirculation inlet nozzle provides flowfrom the inlet line to the jet pump risers. Initially, thesejet pump riser lines were close-fitted to the inside of thenozzle and welded at the end (a detail that was a combi-nation of a groove and a fillet weld). :However, stresscorrosion cracking of the safe end (that is actuallywhere these lines are welded) at the Duane Arnoldplant created a concern about IGSCC of Inconel anrdcrevice conditions. Therefore, most of the safe endshave been replaced, eliminating this crevice condition.

9.2 Stressors

The basic design requirements for the BWR ves-sel were'1250 psi (8.62 MPa)'pressure and v5501F(2880C) temperature. The'actual operating condi-tions at steady 'state are around 1,000 psi(6.89 MPa) and 5401F (2820C). The design specifi-cations included the transients necessaiy to heatupand cooldown the vessel from normal operating

I

I

I

r[

[1

103

r

I

conditions and a variety of normal and abnormalconditions that might be expected throughout thelife of the vessel. The magnitude of the changeswere provided as were the rate at which theyoccurred. Pressure loading in excess of 1250 psi(8.62 MPa) design level occurs during safety valvedischarge when the pressure is permitted to rise to1375 psi (9.48 MPa).

The skirt support for BWR vessels is shown sche-matically in Figure 9.1. One of the key design consider-ations here is that of earthquake loadings.Additionally, the attachment of this skirt to the bottomhead requires a degree of refinement beyond that nor-mally encountered in pressure vessel constructionbecause of the necessity for meeting the code stresslimits for the primary plus secondary stresses andfatigue stresses that occur as a result of the pressure/temperature cycles at that intersection.

All of the nozzles were designed to carry specifiedpipe loads and the bimetallic or trimetallic safe endattachment welds to the nozzle forging area were ana-lyzed for differential thermal stresses, as well.

The core support structure in a BWR consists of anannular plate welded at its outer periphery to the bot-tom of the cylindrical shell immediately above the shellto the bottom head attachment weld. The inner periph-ery of this annular plate is supported by a series ofheavy Inconel posts or columns. The attachment of theoutside periphery of the annular plate and the bottomof the Inconel columns (at the attachment to the bot-tom head) is analyzed in detail for all mechanical, ther-mal, and pressure conditions. Likewvise, theattachments at the upper portion of the cylindricalshell for the core spray spargers and the feedwaterspargers are thoroughly analyzed for specifiedmechanical and thermal loads.

9.3 Degradation Sites

As mentioned earlier, three of the highest rank-ing potential degradation sites have been elimi-nated in the newer plants by design changes andrepaired in most of the older plants' These include:austenitic stainless steel safe ends that have beensensitized by postweld heat treatment, crevice areaswhere thermal sleeves are attached to Inconel orstainless steel safe ends, and austenitic cladding onthe inside corners and surrounding regions of thenozzles (that experience fairly wide temperatureswings); Another possible area of concern that hasbeen eliminated in later design modifications is theCRD penetration stub tube to vessel weld, heat-affected zone; earlier plants that used stainless steel

stub tubes were susceptible to stress corrosioncracking at this weld (leading to through-wall leak-age). Failures did occur at two plants. The mostrecent BWR-6 reactor pressure vessels do not havestub tubes and the guide tube is welded directly tothe bottom head. This decreases the number ofpotential degradation sites.

Regardless of whether the above design changeshave been made, the ranking of potential degrada-tion sites would still vary from plant to plant,depending on how they are operated. For example,one of the more critical items would be the closurehead studs at the main closure flange. Because thegreatest range of stress amplitude in the studs isproduced during installation and removal of thehead (and the use of stud tensioners associated withthose operations), the number of times these occurin a specific plant would have a great effect onwhere the studs fall in the ranking of degradationsites. Thus, real operating history rather thandesign numbers and magnitudes of transients isnecessary in assessing critical degradation sites.

The threaded stud holes in the flanges withoutbushings are subjected to wear because the flangematerial is considerably softer than the studs. Inflanges with bushings, the threads for the closurestuds and flange bushings could be subjected tofretting. However, the studs are removed from theflange infrequently. In addition, the use of studtensioners for tightening makes fretting during thisoperation unlikely.

The design of the flange assembly for most, ifnot all, BWRs includes provision for sufficient pre-load to prevent slippage between the flange surfacesduring operation. Therefore, the flange surfacesare unlikely to suffer physical damage except frommishandling during removal and placement of thetop head. The O-rings are placed in the O-ringgrove between the flange surfaces to provide seal-ing. The O-rings have a soft metal layer on the out-side to enhance sealing and reduce wear, and areintended to be replaced when the head is removed.

Another important potential degradation site isthe outer end of the nozzle forgings, and theattached safe end, for those nozzles that have waterwith large temperature variations flowing throughthem. The most severe nozzle of this type is thefeedwater nozzle. The specified flow temperaturesfor this nozzle go from essentially the freezingpoint of water to tfie operating temperature of thevessel. The best operating example of problemscaused by flow through this nozzle is Millstone-Iwhere nozzle corner cracking was found. Thebypass of the feedwater around the thermal sleeve

I

E

104

I

at this nozzle resulted in a temperature swing of,\.20 0'F (I 10C) at approximately 1-Hz frequencydeveloping high-cycle fatigue cracking. Of course,the upper end of that temperature swing was whenthere was no bypass flow; therefore, this tempera-ture range was larger than the variation in the linetemperature itself.

In a general sense, all locations of dissimilarmetal welds will be most sensitive to thermalcycling. The effect of this thermal cycling will becompounded by any mechanical loads occurring at,the same location. Therefore, the most importantareas in this regard would be the nozzle safe endattachments and the attachment welds to the pres-sure boundary for the core support structure.

The welds and the base metal in the reactorbeltline region are the main sites subjected to radia-tion embrittlement. It should be noted that repairwelds formed during construction and fabricationare also present sometimes in the beltline region ofBWR vessels.

External attachments such as the support skirt andthe refueling bellows skirt are subjected to cyclic load-ing caused by differential radial thermal-inducedexpansion during startup/shutdown. However, neitherof these locations are fatigue limited, i.e. fatigue usagefactors for these locations are far smaller than 1.0.Rather, the limiting condition is primary plus second-ary stresses being controlled by the 3-S. limit. Thereare some corner weld joints in the refueling bellowsskirt assembly that may be fatigue limited for somevendor's designs.

9.4 Degradation Mechanisms

As indicated in the previous discussion, the pri-mary degradation mechanism will be fatigue(induced through a combination of thermal andmechanical cycling). This cycling also has an effecton the main closure studs and is compounded bythe additional cycles of tensioning and untension-ing during top head removal. The threads for clo-sure studs, flange bushing, and stud holes inflanges without bushing are subjected to wear, fret-ting, and corrosion, as discussed earlier. Externalattachments such as the support skirt and the refu-eling bellows are subjected to cyclic loading, as dis-cussed in the preceding section.

There should not be any locations in a BWR ves-sel except the earlier ones having stainless steel stubtubes, where IGSCC should be a problem, becausemost (if not all) of these problem areas have beenpreviously identified and modified. However, con-

tinued monitoring 6f these same areas should bemaintained to ensure'against :any reoccurrence.Water chemistry control used to mitigate the poten-tial for IGSCC in the recirculation piping alsowould reduce any likelihood for problems in thereactor pressure vessel.

' ' All of the welds in the ferritic steel componentshave been postweld heat treated. Therefore, resid-ual stress effects should not contribute to degrada-tion. However, the welds joining Inconel, stainlesssteel,' and the two materials to each other have notbeen stress relieved. An exception to this last state-ment would be the build up of these materials that'is made for the attachment of Inconel (or stainlesssteel) to the ferritic material. These locations wouldbe at the outboard end of the nozzle forgings andwhere the core support structure is attached. How-

, ever, even in the locations where some residualstresses may exist, there is no need to be concernedwith excessive fatigue damage.

Radiation embrittlement is not as severe a prob-lem with BWR vessels as it is with PWR vessels (seeChapter 3). The fairly large diameters of the BWRvessels, and the amount of water shielding the shellfrom the reactor core, help alleviate this problemfor BWRs. However, radiation damage can eventu-ally become a limiting factor, depending on mate-rial chemistries and actual accumulated fluence.Most BWR vessels have an end-of-life fluence(E>1 MeV) of about 5 x 1017 n/cm2 , althoughsome of the older vessels can be a bit higher; typicalPWR vessels have end-of-life fluences aboveI x 1019 n/cm2 . However, irradiation embrittle-ment should not be totally discounted when assess:ing life extension of BWR vessels. From anoperational standpoint, the effects of radiationdamage can be more of a nuisance. For example,with the new Regulatory Guide 1.99, Revision 2(discussed in detail in Chapter 3), hydrotesting ofBWR vessels after experiencing moderate radiationexposure becomes more difficult because of thehigher temperatures required for the hydrotests.However, the data used to develop the RegulatoryGuide 1.99, Revision 2 strongly dictate highertransition temperature shifts than those previouslyused for BWR vessels at these low fluence levels.

9.5 Potential Failure Modes

Because pressurized thermal shock and low-energyductile tearing are extremely unlikely in a BWR vesselbecause of the low neutron irradiation conditions inthe beltline region, the only final failure mode is plastic

I

r

[

r

105 '

I

overload leading to a leak in a region where a crackexists (or failure of a stud). Thus, the conditions lead-ing to this final failure mode are fatigue or corrosioninduced. All parts of the reactor vessel for which cyclicloading has been specified have taken fatigue intoaccount and have fairly low usage factors (almostalways below 0.6 over the intended 40-year life). How-ever, as indicated in Chapters 3 and 5, the calculationof fatigue usage factors can vary and exact comparisonof numbers can be misleading. Consideration ofabnormal loading conditions that currently would beconsidered as Levels C or D (emergency or faulted)did not control any of the design features of a BWRvessel. In the earlier BWR vessels with stainless steelstub tubes, IGSCC-induced failures have resulted inleakage through the CRD penetrations.

9.6 In-service Inspection andSurveillance Methods

The PWR vessel in-service inspection (ISI) meth-ods and requirements of Section XI of the ASMECode were presented in Chapter 3 and are the samefor BWR vessels. However, many older BWRs havevery limited accessibility for external ISI of the ves-sel. Typically, 75 to 90%o of the vessel weld lengthsare exempted because of inaccessibility. The onlyalternative is ISI methods of examination from theinside surface. This is of particular importance atthe beltline welds. There were minor and majorrepairs to shell plates during construction but some

of these cannot be inspected because of limited.accessibility. This is also true for some of the pipe-to-nozzle welds that were not configured to facili-tate ISI. This was changed in later reactors after therequirements of ASME Section XI were published.

Surveillance for irradiation embrittlement is dic-tated by federal law for BWRs as well as PWRs.Therefore, monitoring of actual changes in CharpyV-notch and tensile properties with regard to accu-mulated fluence for the most critical vessel materi-als is underway.

9.7 Summary and Conclusions

A summary of the degradation sites and mecha-nisms, stressors, potential failure modes, and ISI meth-ods for BNVR vessels is presented in Table 9.1. Most ofthe extensive conclusions listed in Chapter 3 also areapplicable here. The main difference is the much lessimportant effect of irradiation embrittlement for BNVRvessels. However, this damage mechanism is presentand should never be ignored.

Other conclusions worth restating here are: noz-zle fatigue usage should be monitored more closelywith regard to refined cycle counting and transientseverity so that better estimates of actual fatigueusage factors can be determined, the closure studsshould be examined closely for replacement nearthe normal 40-year end of life, and the large-nozzle, CRD, and other small-nozzle penetrationwelds should be carefully inspected.

I

I

I

I.

I

106

I

I

Table 9.1. Summary of degradation processes for BWR reactor pressure vessel.

Rank ofDegradation Site

DegradationSite

- . ., : . .I .~Degradation Potential Failure

Stressors Mechanisms Modes* ISI/Surveillance

Methods [Nozzles (includinginstrument and CRDpenetrations) plussafe end welds

2 Closure studs,flange bushings,stud holes

3 Beltline region

4 External attachments

Mechanicaland thermalstresses

Fatigue crackinitiation andpropagation,IGSCC

Ductile overload I All large nozzle weldsl Ieading to a leak . inspected volumetrically

at each interval; visual,external surfaceinspections of small

- nozzles/penetrations

Mechanical Fatigue crackand thermal . initiation andstresses ' propagation,

fretting,corrosion

Irradiation Neutronembrittlement irradiation

(extentdepends onvesselmaterials)

Ductile overloadfailure (can bereplaced)

Ductile high-energytearing leading to aleak (not a seriousproblem):

Ductile overloadfailure

Volumetric and surfaceinspections of all studs,threads in flange studholes and bushings

100 volumetricinspection; surveillanceas required by federallaw

Volumetric and surfaceinspections

E

IMechanical Iand thermalstresses : .

Fatigue

F

: - , I I I r.~~~~7

107

I

10. BOILING WATER REACTOR RECIRCULATION PIPINGH. Mantle

There has been an increasing incidence of inter-granular stress corrosion cracking (IGSCC) failuresin boiling water reactor (BWR) recirculation pipingsystems in the past 10 years: first in highly stressedsmall-diameter piping and later in large-diameterpiping (12 in. and greater). While an actual pipeseverance has never occurred, the problem hasimpacted plant availability and resulted in a sub-stantial increase in radiation exposure to mainte-nance and inspection staff. Although IGSCC hasbeen the Achilles' heel of BWR plants, great strideshave been made in understanding the conditionsunder which it occurs and in developing both short-and long-term solutions.' Other concerns that mayaffect life extension have been raised (such as low-temperature aging of duplex stainless steel castings,crevice corrosion, and thermal fatigue), but thesehave not yet been shown to be significant problemsin BWR recirculation piping systems.

10.1 Description

The recirculation system draws coolant from thebottom of the reactor vessel through an outlet noz-zle and downcomer spool into the recirculationpump. The pump discharges this coolant into aring manifold that feeds the coolant by way of mul-tiple risers to the associated inlet nozzles and jetpumps, thus completing the loop. The reactorrecirculation system consists of two independentloops. Primary piping in each recirculation loop ina large BWR reactor (21100 NIWe) consists of a28-in. pump suction line, a 28-in. pump dischargeline, a 22-in. pipe riser manifold, and five 12-in.taps (ten total taps, five for each loop) to the ten jetpumps inside the reactor. A simple schematic of thebasic arrangement is shown in Figure 10.1. Keycomponents include safe end nozzle welds, suctionand discharge valves, the recirculation pump, andthe ring manifold.

Historically, BWR recirculation piping systemshave been fabricated from wrought, AISI types 304and 316 (piping) and cast types, CF8 and CF8M,(pumps and valves) stainless steels. Replacementpiping has been made from Type 316 NG (nucleargrade) that has been shown to be much more resist-ant to IGSCC than the conventional grades.

10.2 Stressors

Stressors include welding heat (sufficient to pro-duce a sensitized microstructure), tensile stresses(both residual and applied), oxygenated coolingwater, temperature and cyclic thermal stresses.These factors, either in combination or separately,can produce IGSCC, thermal fatigue, thermalembrittlement, and crevice corrosion.

10.3 Degradation Sites

Degradation sites include weld heat-affected zonesand furnace sensitized components (safe ends), weldmetal, duplex stainless steel castings (elbows, valves,T-joints), areas in the system with high thermallyinduced stresses predicted by stress rule index analysisI(nozzles, safe ends), and crevices at safe ends shieldedfrom the bulk fluid environment. Additionally, basedon the interpretation of the North American ElectricReliability Council's 1984 Component Cause CodeReport, 2 the recirculation pump also is a primary deg-radation site (see Table 10.1). The report data indicatethat the recirculation pump and driver have a greateroutage frequency than the reactor coolant systemvalves and piping. The cause for greater outage fre-quency is not reported in Reference 2 and is probablynot related to a loss in pressure boundary integrity.

10.4 Degradation Mechanisms

As discussed in previous surveys of aged nuclearpower plant facilities,3 degradation mechanismsmost responsible for current aging-related failurescan broadly be categorized as erosion, corrosion,vibration (fatigue), and foreign material related.Within the recirculation piping system of BWRs,failures (leaks) have primarily been related toIGSCC. Of secondary concern are crevice corro-sion, low-temperature aging (embrittlement) ofduplex stainless steel, and thermal fatigue.

Degradation resulting from IGSCC has essen-tially been controlled by way of a multidirectionaleffort, because IGSCC of austenitic stainless steelsrequires the combined existence of the followingthree conditions:

I

I

r

I

108

I

a

,,,,,, .1

'- Pumpdischarge valve

74048

Figure 10.1. Schematic of a BWR recirculation piping system.

Table 10.1. Outage frequency: valves and piping versus pumps and drivers(years 1975-1984, 29 units, 230 unit years)

Reactor CoolantSystem Valves

and Piping

Forced 0.44

* Reactor CoolantRecirculation PumpI and Drivers

. .I 0 5

0.56

1 .89

2.79 .

Unplanned derating

Forced maintenance,planned andequipment derating

0.26

* . 0.93.i

109

.U.

I

1. A sensitized microstructure2. A local tensile stress that exceeds the bulk

alloy's yield stress3. A corrosive (oxygenated) environment.

Even though these conditions theoretically existin the heat-affected zones of most BWR recircula-tion piping welds, only a very small percentage ofthese welds have cracked. The primary factor thatdetermines whether a particular site will developIGSCC can therefore be inferred to be the existinglocal tensile stress since it is the most variable fac-tor. It is also the most difficult to quantify, becauseit is made up of both applied (design loads) andresidual (weld shrinkage, forming, machining, andgrinding) stresses. Although IGSCC can be elimi-nated by controlling any one of the above three fac-tors,4 the current practice is to attack two or (insome cases) all three of the factors.

The first factor, sensitization of the base metaland weld heat-affected zone, is affected by thechemical composition of the base material (primar-ily carbon content) and the total thermal history(high weld heat inputs increase sensitization). Thedevelopment of nuclear grade stainless steels withmaximum 0.02 wtqo carbon, nitrogen additions,and tightly controlled levels of impurities has pro-duced a material much less susceptible to sensitiza-tion and thus more resistant to IGSCC. Installationcontractors and nuclear steam system suppliersalso have attempted to control sensitization by rap-idly cooling the pipe interior after deposition of theroot pass. This technique is called heat sink weld-ing. This treatment reduces the degree of weld-zonesensitization and produces a more favorable pat-tern of residual stress on the pipe inside surface.

The second factor, tensile stress, can be mitigatedby way of induction heat stress improvement(IHSI) and last-pass, heat sink welding. The first ofthese two techniques must be closely controlled toachieve the desired effect (residual compressivestress at the inside weld surface), particularly withwelds made by the gas tungsten arc process. 5 It alsohas been suggested that IHSI may sensitize pumpand valve body castings adjacent to the weld zone.

The third factor, oxygenated cooling water, canbe attacked by careful control of water chemistry.Under steady-state conditions, BWR coolant con-tains 200 to 400 ppb dissolved oxygen. This levelcan rise to -.8000 ppb under start-up conditions,thus further increasing the susceptibility to IGSCC.Because oxygen is constantly being created fromthe water coolant by way of radiolysis in the reactorcore, the only practical method to control the oxy-

gen to levels below the threshold for IGSCC withinthe recirculation loop is to inject hydrogen into thecooling water. This produces an environment favor-ing the recombination of hydrogen and oxygen.Ljungberg6 has shown that levels of 3 to 5 ppboxygen are achievable with this method. IGSCC iscompletely inhibited at levels below 5 ppb oxygen.

Also of concern for life extension of BWR systems isthe effect that low-temperature aging has on duplex(austenitic-ferritic) stainless steel castings (and possiblyweld metal) at operating temperatures and the potentialfor thermal embrittlement in localized areas. Exposureto temperatures in the 560 to 700'F (293 to 371IC)range promotes the formation of secondary phases(such as the G-phase, Type X, and alpha-prime phase)and carbide growth in these materials. The embrittlingeffect can be observed on materials with > 10°70 ferrite.Materials with >20% ferrite show drastic reductionsin impact energy when aged at higher temperatures[i.e., 750 to 850IF (399 to 4550C)]. Low-carbongrades in general show greater resistance to thisembrittling effect, probably because of a reduction inthe formation of chromium carbides and other second-ary carbide phases. There also is some evidence to sug-gest that careful initial heat treatment of cast duplexstainless steel may produce a casting with superiorunaged impact energies 7 such that even after aging sig-nificant fracture toughness (measured by room-temperature, Charpy V-notch energies) remains.Typically, these duplex stainless steel components areextensively weld repaired with attendant ferrite dilutionin the weld zone. Detailed records of welded repairsand detailed ferrite surveys are not available. Ferritemeasurements and monitoring should be done on theinside surfaces, which requires disassembly and exten-sive isolation provisions or draining the vessel afterunloading the core. Further study of the thermalembrittling of the duplex stainless steel castings andparticularly the weld material is recommended. 8 Addi-tional information on the thermal embrittling of theduplex stainless steel components is discussed in Sec-tion 5.3.2, Chapter 5.

Crevice corrosion has been mentioned as apotential degradation mechanism that could occurover time at locations such as shaft sleeves and theunderside of socket welds if a corrosive environ-ment and a difference in oxygen concentrationexisted. To date, crevice corrosion has not beenreported as a problem in BWR recirculation piping.This is probably a result of the tight controls main-tained on water conductivity and oxygen levels inthis system.

Thermal fatigue and corrosion-assisted fatiguealso have been mentioned as potential degradation

E

I.tL

a

110

a

mechanisms. Frequent reactor scrams andunscheduled outages produce transient thermalstresses on both the cooldown and heatup cycles.Highly stressed hot spots need to be defined and adata base established to confirm whether this effecthas significance for BWR recirculation piping. Theweld overlays probably deserve specific attention.-Fatigue considerations are currently being studiedfor carbon steel systems, but stainless steel systems(such as the BWR recirculation system) have not yetbeen addressed.9

10.5 Failure Modes

Failure modes produced by the failure mecha-nisms above would be primarily through walldefects resulting from cracks (IGSCC, embrittle-ment, or fatigue) or material wastage (crevice cor-rosion). So far, cracks in recirculation pipingmaterials have satisfied a leak-before-break criteriaand no guillotine breaks (complete separation)have occurred. This latter type of failure is unlikelyto occur unless it can be shown that the duplexstainless steel castings in actual components haveembrittled to such an extent that a brittle fracturecould occur at or near the operating temperature.

10.6 In-service InspectionMethods (1S1)

The currently required in-service inspection(ISIs) have not always been able to detect cracksthat have developed during service. Detection ofIGSCC using the currently available ISI techniquesis particularly difficult. Examination of cast stain-less steel by ultrasonic testing is considered unrelia-ble. The USNRC is currently sponsoring researchinto improved methods for continuously monitor-ing pressure boundary components. 10 On-lineacoustic emission is being considered, but a ques-tion exists as to its ability to judge the severity of aparticular cracking event. Proposals have also beenmade to install moisture-sensitive tape on the out-side surface of critical piping to detect leaks. 1I

10.7 Summary, Conclusions, andRecommendations

The degradation sites and mechanisms, stres-sors, potential failure modes, and ISI methods forthe BWR recirculation piping are summarized byTable 10.2. IGSCC has been a major problem and

a variety of fixes and controls (discussed above) arein place. Other degradation mechanisms are possi-ble and should be evaluated, but to date they havenot been associated with actual outages or failuresin BWR recirculation piping systems.

As might be expected, the nuclear industry isextremely optimistic about the possibility ofextending the life of nuclear power plants, at leastfrom the technical point of view.12' 1 3 The pre-dominant view is that no insurmountable technicalobstacles currently exist that would preclude theextended use of existing BWR plants, provided thegroundwork is laid now to support continued use.The careful monitoring and recording of compo-nent performance from all operating plants is a keyfactor in this effort. The actual degree'of thermalembrittlement in the duplex stainless steel castingsshould be monitored. The power industry, throughindustry groups like the Electric Power Research'Institute, and the nuclear steam supply systemmanufacturers through programs like GeneralElectric's COMPASS computerized data base andWestinghouse's Reliability Data Base, are attempt-ing to develop useful data bases that will pinpointweak links and identify trends useful in the life-extension effort. These data bases will be essentialin the effort to ensure and demonstrate the safetyand reliability of each plant being considered forlife extension.

Information on failure mechanisms and sitesshould continue to be gathered from individualutilities, nuclear steam supply system manufactur-ers (such as General Electric), and reviews ofUSNRC reportable incident records to update thedata base of known weak links and material fail-ures. Additional work needs to be performed toanalyze the existing data bases for trends in compo-nent failure mechanisms and locations. Forinstance, failure locations may be one of the keys inverifying the existence of so-called fatigue hot spotsin BWR recirculation systems. The existence ofsuch hot spots and their effect on overall system lifebecomes of greater concern as the number andseverity of startup/shutdown cycles increase.

Additional work also should be done on duplexstainless steel castings to determine whether (a) a BWRenvironment will produce significant thermal agingand (b) typical castings contain sufficient delta ferriteto exhibit a significant embrittling effect.

The possible detrimental effects of hydrogenadditions to the BWR recirculation piping loopshould be studied to determine what reactionsoccur between hydrogen and other system compo-nents such as feedwater piping.

I!

F7

t U

111

II

a

Table 10.2. Summary of degradation processes for BWR recirculation piping

Degradation Degradation Failure ISIRank Site Stressors Mechanisms Modes Method

WVeld heat-affected Tensile stress, oxygenzone furnace environment, sensitizedsensitized safe heat-affected zoneends

IGSCC Cracks, leaks Ultrasonic examination,moisture-sensitive tape

Cracks, leaks Ultrasonic examination,moisture-sensitive tape

2 High thermal stress Cyclic tensile stress, Thermal fatigue,regions predicted corrosive environment corrosion fatigueby stress ruleindex analysis

3 Austenitic-ferritic High temperature, Thermal embrittlestainless steel tensile stress, shockcastings with high loadsdelta ferritelevels

IT!ment Cracks, leaks Ultrasonic examination,

moisture-sensitivetape, impact testspecimens

4 Shaft sleeves,other crevices

Corrosive environment, Corrosionoxygen-starved areas

Materialwastage,leaks

Moisture-sensitive tape

L

U

112

10.8 References

1. P. W. Marriott, "Nuc!ear Plant Life Extension-A Comprehensive Approach to Plant Improve-ments," Nuclear Europe, 3, 1984, pp. 29-32.

2. North American Electric Reliability Council, 1984 Annual Report, Generating Ava'lability Data Sys-tem (GADS), Component Cause Code Report, Princeton, New Jersey.

3. J. A. Rose, R. Steele, Jr., B. C. Cornwallj and K. D. DeWall, Survey ofAged PowerPlantFacilities,NUREG/CR-3819, EGG-2317, June 1985.

4. J. C. Danko, "IGSCC-Recent Experience Reviewed," Nuclear Engineering International, 29, 356,June 1984, pp. 27-30.

5. G. T. Krause, "Stress Improvement of GTAW-P-ME Welds in Austenitic Stainless Steel Pipe," Weld-ing Journal, 65, 5, May 1986, pp. 21-29..-

6. L. G. Ljungberg, D. Cubicciotti, and M. Trolle, "Materials Behavior in Alternate (Hydrogen)Water Chemistry in the Ringhals-l Boiling Water Reactor," Corrosion, 42, May 1986, pp. 263-271.

7. 0. K. Chopra and H. M. Chung, Long-term Embrittlement of Cast Duplex Stainless Steels in LWERSystems: Annual Report October 1984-September 1985, NUREG/CR-4503, Volume 1, ANL-86-3, rJanuary 1986. -

8. 1. Spiewak, R. S. Livingston, and C. A. Negin, The Longevity of NuclearPobwer Systems, EPRI NP4208,August 1985.

9. M. Vagins and J. Strosnider, Research Program Plan-Piping, NUREG-1 155, Volume 3, July 1985,p. 4.

10. J. Muscara, Research Program Plan-Nondestructive Examination, NUREG-1155, Volume 4,July 1985, p. 1.

11. P. Riddle, "Localized Leak Detection Using Moisture Sensitive Tape," NuclearEngineering Interna-tional, June 1984, pp. 32-34.

12. D. R. Eggett and A. E. Meligi, "Extending the Life of Nuclear' Power Plants: Technical and Eco-nomic Issues," American Nuclear Society Transactions, 46, June 1984, pp. 583-584.

13. P. W. Marriott and J. P. Higgins, "The Impact of Power Plant Aging on PlantAvailability: I (Managing Nuclear Plant Aging)," American Nuclear Society Transactions, 46,June 1984, pp. 579.

113

a

11. NONDESTRUCTIVE EXAMINATION METHODSL. J. House

Safety and reliable performance typically areachieved during the design life of nuclear reactorsby using an engineering approach based on overde-sign and incorporation of generous safety factors.Therefore, the demands on methods for detectingand measuring flaws using nondestructive exami-nation (NDE) are modest. Historically, many of theAmerican Society of Mechanical Engineers(ASME) code requirements and standard NDEmethods were developed for detection and qualita-tive assessment of fabrication related defects gener-ally associated with the quality of workmanshipand then adapted to in-service inspection require-ments within the normal design life of the compo-nents. On the other hand, reactor plant-lifeextension beyond the normal design life involvesengineering calculations for cases where safety fac-tors may be reduced because of in-service degrada-tion of critical reactor components. Thus,life-extension inspection generally requires greaterdetection reliability and a more quantitative assess-ment of defects and accumulated damage than tra-ditional fabrication and in-service inspection.Successful application of residual-life predictionprograms, therefore, depends on the availability ofadequate methodologies to detect and measure theeffect of accumulated damage on performance-related properties of the key components.

The objective of this review is to address the efficacyof standard NDE methods for detecting and evaluatingdamage produced by degradation mechanisms occur-ring in reactor environments. It is well known that thereliability and capability of NDE methodologies aredirectly related to operator performance factors1 andspecific characteristics of the instrumentation, dataacquisition, and signal processing employed.24 How-ever, this review is intended to be of a general naturecovering the standard methodologies rather thanattempting to assess the many specific embodiments ofthe techniques derived from those methods. Also, theadequacy of NDE for any given application dependson many other factors besides the performance of thetechnique. These factors include cost-benefit analysis(e.g., does inspection cost more than periodic compo-nent replacements?), availability of trained manpower,accessibility, and inspection rates. Only the technical-performance-related factors (i e., can the methodologyreliably provide the necessary measurement?) areaddressed in this review.

The methodologies generally practiced by thenuclear and other industries and included in this revieware visual examination, penetrant and magnetic parti-cle testing, x-ray radiography, eddy-current testing,ultrasonic testing, and acoustic emission monitoring.These standard NDE methods are discussed withrespect to their adequacy for providing the informationabout accumulated damage in reactor componentsnecessary for residual-life estimation. The format forthis review consists of a background discussion of whatresidual-life estimation requires of NDE, the type ofinformation standard NDE methods can provide, anda summary of the general deficiencies in the standardmethods for use in residual-life estimation.

11.1 Background

In the transition from the task of providing asimple quality control function to supplying quan-titative assessment of accumulated damage forfracture mechanics analysis and residual-life esti-mates, the reliability and adequacy of standardNDE methodologies have rapidly diminished.Many of the standard NDE methods were intendedto provide only an indication of the presence of aflaw or developed specifically to detect a particulartype of flaw in a given component during fabrica-tion and assembly. For residual-life estimates, thosesame methods must be employed not only to detectflaws but to determine the location, size, shape,orientation, and type of flaw or damage in compo-nents with various service histories, complex geom-etries, and limited access. The inspection is oftenperformed in harsh environments with very severeconstraints on the time available to do the inspec-tion. Consequently, a number of inadequacies instandard practices of NDE have, and are continu-ing, to appear.

To estimate residual life using either a deterministicor probabilistic approach, information about the effectof various degradation mechanisms on performance-related properties is needed. As illustrated inFigure 11.1, the approach for each critical componenthas been to identify the principal degradation sites,establish the associated damage mechanism, determineperformance-related physical and material propertiesaffected by the damage mechanism, select the appro-priate measurement technique to detect the relevantproperty changes, and finally to assess the extent and

114

I

I I . . I . I

Il I I Accessibility. . environment

Degradation site I . - configurationI I material

I Degradation mechanismI'

IPhysical and materialproperties of damage

produced by degradationmechanism - . W

Damage measurementI . methods H

. .

-{.

Standard NDE methods

l

Miniature specimen testing I

l

Emerging measurement andmonitoring techniques F

7-3044

Figure 11.1. Typical approach to selecting a method for detection and measurement of accumulated damage incritical reactor components. I

severity of the damage.5- 12 The selection of the mea-surement methodology may include standard NDE,miniature specimen testing, or one of the many emerg-ing measurement and monitoring technologies. Theselection of the appropriate method depends on acces-sibility, environment, component configuration, andthe material at the potential degradation site.

The properties affected by degradation, particu-larly those detected and measured by standardNDE methods, are generally only indirectly relatedto the actual damage. For example, in the earlystages of plastic deformation, there is a markedincrease in dislocation density and associatedstacking faults. This causes a decrease in the electri-cal conductivity of the material. Eddy-current.methods can measure changes in conductivity and,thus, in principle can be used to measuredeformation-related properties, such as yield,:strength. However, other factors, including varia-tions in chemical composition, temperature, sur-face condition, and unrelated degradationmechanisms, such as thermal aging, also can affectconductivity. Also, changes in the concenitration ofinterstitial solid solution hydrogen may distort thecrystal lattice of the metal alloy, causing a change in

its electrical conductivity. Therefore, the othersources of conductivity changes preclude thisapproach for most practical applications. Thus, ifthe NDE method can not discriminate between thetypes of damage or if a knowledge of the degrada-tion mechanism causing the damage is unavailable,then the nondestructive measurement-based onempirical correlation of a physical property likeelectrical conductivity with a particular type ofmaterial degradation such as hydrogen damage orplastic'deformation-can give unreliable results.This issue becomes very complicated because of thelarge number'of potential damage mechanisms thatexist in the'reactor service environment.'

The principal degradation mechanisms in keyreactor copronents include: embrittlement (neu-tron, hydrogen, and thermal), fatigue (corrosion,thermal, low cycle, high cycle), corrosion (erosion,wastage, pitting, denting,' and stress corrosioncracking), wear and fretting, and aging (thermaland environmenital). The issue for residual-lifeassessment is whether' stauidard NDE methods candetect the accumulated damage produced by thesevarious mechanisms and provide a measure'of theseverity and extent of the damage in a sufficiently

Tr

115

a

quantitative manner. For some of the degradationmechanisms, like neutron embrittlement, it isunclear what to measure using NDE methods inorder to detect and characterize the resulting dam-age. Ideally, however, NDE should be able to dis-criminate between or identify the type of damage,measure the flaw size or extent of damage, measurethe flaw shape and orientation, and determine thelocation of the flaw or damage. Although standardNDE has many deficiencies in being able to provideall of these necessary measurements for residual-life estimates, the information NDE can provide,despite the attendant uncertainties and reliabilityissues, is nevertheless essential.

11.2 Standard NDE Methods

Summarized in this section is the type of informa-tion provided by each of the standard NDE methods.More specifically, the flaws and damage that can bedetected, the properties that can be measured, and thedeficiencies of each method with respect to residual-lifeassessment are discussed.

11.2.1 Visual Examination.13 Inspectionbyvis-ual examination is the most widely used nondes-tructive method. Visual examination is importantfor reasons other than its simplicity and low cost.First, visual examination provides a qualitativeindication of surface conditions. In welds, visualexamination can reveal the presence of cracks, sur-face porosity, and other potential sources ofmechanical weakness, such as the extent of pene-tration, undercutting, sharp notches, or misalign-ment. Furthermore, visual examination may assistin the determination of other nondestructive teststhat should be applied to a potential site of degra-dation in a reactor component and how they shouldbe applied. As with all nondestructive methods, theusefulness of visual examination methods dependon proper application and correct interpretation.In visual examinations, more directly so than in theother nondestructive methods, the results are sub-jective and dependent on the operator. The resultsof visual examination are generally qualitative andlimited to assessing damage that manifests itself atan accessible, surface of a component. Visual aidssuch as optical microscopy to examine the micro-structure of a prepared surface and a borescope toexamine the interior walls of tubing increase theutility of this method. In tube bundles, borescopicinspection provides information about the restric-tion of flow because of scale and deposit build-up

and wall thinning caused by wastage, pitting, oruniform corrosion. The development of computer-enhanced vision systems to automate visual exami-nation and to permit remote visual examinations inhostile or otherwise inaccessible environments havebeen successful for viewing structures, but theinformation provided is often qualitative in natureand of limited value to residual-life estimation. Vis-ual examination in most cases provides an indica-tion that damage may have occurred but cannotdirectly quantify the amount of material damage.

11.2.2 Penetrant and Magnetic ParticleTesting. Penetrant and magnetic. particle testingare essentially an extension of visual examination inthat they are used as an enhancement technique toimprove the visibility of surface-connected flaws.Magnetic particle testing also can indicate the pres-ence of near-surface defects, such as cracks andinclusions. Its principal utility as a NDE method isto increase the inspection rate and probability ofdetection of surface flaws, compared with unaidedvisual examination. Liquid penetrants delineatesurface discontinuities making the inspection lesssubject to the visual acuity of the operator. Theyprovide no direct information about the depth ofthe crack although some very experienced inspec-tors have been able, in some cases, to make fairlyreliable estimates of the depth and size of defects.The accuracy of these estimates depend on themagnitude of the defect opening at the surfacebeing related to the depth of the defect. Penetrantmethods can be used to locate grinding cracks,welding cracks, casting cracks, fatigue cracks,shrinkage, blowholes, seams, laps, pores, coldshuts, porosity, lack of bonding, pinholes in welds,through cracks, forging laps, bursts, gouges, toolmarks, and die marks, provided the defects open tothe surface. In some work with radioactive pene-trants, 14 residual activity of the penetrant has beensuccessfully correlated with crack size but completecapillary diffusion of the penetrant into the flaw isessential. In general practice, the correlation is nota reliable indication of crack size. The deficienciesof penetrant and magnetic particle testing are thesame as with visual examination, except that thedetection probability for surface-connected flaws isimproved over unaided visual examination.

11.2.3 Radiographic Examination. The standardpractice in radiographic examination is generally x-rayfilm radiography. The method essentially provides arecording of density variations that may occur because

r

L

FL

116

I

of a variety of causes: cracks, inclusions, porosity,voids, lack of bonding, and dimensional changes. Theradiograph can'provide quantitative information aboutflaws that affect the mass density or x-ray absorptioncoefficients of the component; for example, it can indi-cate the length of cracks. A film'densitometer also cangive an estimate of the actual extent of damage givensuitable calibration and precise control of exposuretime and source intensity. However, two-sided access,'-except in 'those few cases where autoradiography (atechnique using the intrinsic radioactivity of the com-ponent) is possible, the slow rate of examination, back-ground radiation, and the time between theexamination and when the results are available afterfilm development and evaluation severely restrict theapplication of this method. This is particularly true forperiodic in-service inspections of reactor componentswhere, to minimize costly down-time, the NDE activitymust not delay decisions to replace or return a compo-'nent to service.

11.2.4 Eddy-Current Testing. The use of single-and multifrequency eddy-current testing is gener-'ally limited to inspection of simple geometries fornear-surface defects. The only extensive app'lica-'tion of this method in the'nuclear industry has beento inspect pressurized water reactor steam genera-tor tubing. Although the eddy-current signal isaffected by'the metallurgical condition of the mate-rial being tested and can provide an indication ofdegradation or damage, the effect of changes inphase and chemical composition, as well as othermetallurgical properties, on the eddy-current signal'is neither quantitative nor predictable. A correla-tion between the measured impedance change ofthe eddy-current coil and the desired structural orserviceable characteristics must be established forall of the various degradation mechanisms. Normalmetallurgical or chemical variations can mask theeffects produced by service-related damage in thecomponent structure and any condition affectingthe permeability or conductivity of the'materialbeing tested affects the measured eddy-current sig-nal. Discriminating between changes in the mea-'surement caused by normal material variations oractual damage is often not possible. Single fre-quency eddy-current examination of steam genera--tor tubing 'in 'accordance with the ASME -codeprocedures has not always been reliable because, inaddition to the reasons cited above, other inspec-tion variables, such as tube supports, tube sheets,denting, and eddy-current'probe wobble, affect the'results. Multifrequency methods remove the influ-ence of many of these variables, but improvements

are still needed in the sensitivity and reliability offlaw detection and characterization. One particulardeficiency is that standard eddy-current methodsand probes for tubes and methods are inadequatefor detecting and characterizing circumferentialtype flaws, like intergranular cracks and generalintergranular attack, in tubes in the neighborhoodof and within the tube-sheet crevice. The design ofpancake coils for tube inspections enables the

-detection of circumferential flaws, but the use ofthis type of probe is not presently a standard prac-tice for tube inspections.

11.2.5 Ultrasonic Testing. After visual examina-tion, the next most widely used volumetric nondestruc-tive method for in-service inspection of components isultrasonic testing. The ASME code requires ultrasonicinspection of welds and adjacent base material.Although presently accepted as the most useful volu-metric examination method, the reliability of standardultrasonic procedures and methods of flaw detectionhas been increasingly questioned during the last severalyears.3' 15 17 This is because'the ultrasonic examina-tion methods, particularly procedures and practicesspecified by the various ASME 'code requirements,were developed primarily to assess overall quality bydetecting fabrication-related defects that correlated

- 'with workmanship. 16 However, the informationrequired for residual-life assessments and fracture-mechanics-based integrity assessments requires detect-ing, locating, and sizing of more subtle flaws. Sizing offlaws has always been a source of error because ASMEcode sizing techniques are typically based on empiricalcorrelations of ultrasonic-echo-amplitude 'measure-ments with flaw size and not on basic physical princi-ples.

The standard ultrasonic examination methodsgenerally use recording of pulse-echo amplitude

'and transducer position as the basis for flaw detec-tion and sizing. The functional aspects of ultra-sonic testing involve two operations. The firstoperation is generally a rapid initial search mode todetect the presence of a flaw in a component and asecond operation in which a more detailed inspec-tion is performed to determine whether the flaw isbenign or represents a performance-limiting condi-tion. Standard procedures, although satisfactoryfor their original purpose, are not adequate and areoften unreliable for, providing the detection andcharacterization of flaws needed to assess the resid-ual life of reactor components. Some issues con-tributing to the inadequacy of standard ultrasonictesting practices include'theoretical limits of detec-tability and resolution; external factors such as

r

I I

L.

117

17

a

operator training, equipment, and calibration defi-ciencies; acoustic coupling variabilities; and inter-nal factors such as material variabilities, partgeometry, and flaw characteristics. An example ofinternal factors that affect flaw-detection reliabilityis the variability of ultrasonic echo strength withmechanical load. A mechanical load of sufficientmagnitude can compress the two faces of a cracktightly enough so that the ultrasonic reflection isreduced or, in the extreme case, the crack becomestransparent. Rough-clad weldments and compo-nents fabricated from cast stainless steels are virtu-ally'uninspectable by standard ultrasonic methods.Other deficiencies include unreliability of detectionof cracks under cladding, detection and sizing ofintergranular stress corrosion cracking (IGSCC),and detection of thermal fatigue cracks and IGSCCin wrought stainless steel welds.

It is clear that the limitations of standard ultra-sonic test methods are in part caused by deficien-cies and physical limitations of the availabletechnology. On the other hand, a previous evalua-tion of the reliability of Air Force field inspectionsindicated that inspection results were poorer thancould be explained on the basis of the inherentphysical limitations of the technique.15,18 Poorperformance was attributed to human factors suchas operator boredom and the awkward positions inwhich the operator was required to perform theinspection, because of limited access.

A good summary of the capabilities of standardultrasonic NDE methods is given in Figure 11.2,3 inwhich diagrams of the probability of defect detec-tion (DD), correct rejection (CR), and correctacceptance (CA) as a function of defect categoriesare shown. The defect categories included-smallacceptable defects (a), small rejectable defects (b),large continuous defects (c), and composite(multiple-type) defects (d). The ultrasonic test pro-cedures fell into three categories. Category A com-plied with the requirements of the Program for theInspection of Steel Components (PISC)-formerlyPlate Inspection Steering Committee-whichclosely follow the requirements-of ASME XI. Cate-gory B involved minor modifications to the PISCprocedure, and Category C was markedly differentultrasonic techniques providing'truly alternativeprocedures. Only the large continuous defects areproperly detected and rejected by the Category Band C ultrasonic test procedures. The Category Cprocedures do result in better, but not perfect,small-defect detection.

11.2.6 Acoustic Emission Monitoring. Acous-tic emission measurement techniques have beenused extensively as an alternative to ultrasonicexamination for detecting flaws in pressure vesselsin the nuclear and petrochemical industry. Acousticemission is a passive technique that detects flawsonly during growth. One very useful application ofacoustic emission monitoring is leak detection inpressure vessels. A major deficiency in acousticemission testing is the possibility of missing a largeflaw if it happens to not be growing during moni-toring. Advocates of acoustic emission testingcounter with the argument that only flaws that aregrowing can induce failure and continuous acousticemission should preclude missing the large flawsduring growth. A number of major evaluation pro-grams have been performed on pressure vesselswith mixed results. Difficulties are associated withdiscriminating between damage- or flaw-relatedacoustic emission events and ambient backgroundnoises, low-level emissions produced by flawgrowth in certain alloys, and accurate location ofthe acoustic emission event in complex structures.An intrinsic deficiency of standard acoustic emis-sion monitoring is the absence of a reliable correla-tion between measurement of acoustic emissionevents and the severity of the flaw. A presentadvantage of most acoustic emission monitoring isits use as a precursor to impending failure ratherthan a method for quantitative measurements ofcomponent damage for use in fracture-mechanicsestimates of flaw severity or residual-life assess-ment . Another advantage of acoustic emissionmonitoring, unlike the previous methods dis-cussed, is that a more global examination or large-volume interrogation is provided. A majordisadvantage of acoustic emission testing forresidual-life estimation at present is the absence ofa correlation between acoustic emission propertiesand fracture-mechanics-related parameters. Forexample, acoustic emission parameters can be usedfor crack-growth detection or location, but themeasured parameters have not been reliably correl-ated with crack size.

11.3 Summary of GeneralDeficiencies of StandardNDE Methods

NDE methods, as embodied in standard practice,offer capabilities that are primarily limited to flawdetection. Standard NDE methods are presently

_.

.,I I

t

118

a

DD1

0.5

. 0

: ; -. I 11 .. ,_ ' .

1..CA, ,CR

B

PISC procedure

DO1

0.5 :

'- A -'

I..,

I I

''a b c d a b c dNational standard or procedure in thespirit of ASMECode (for preserviceor manufacturing inspection)

DD - CA CR

c- 1 1 r

0.5 cdc

a bc d a bc d

E-

LAlternative techniques 7-3065

Figure 11.2. Average cumulative diagrams of defect detection (DD) and correct rejection or acceptance [CR(CA)] asa function of the defect categories a, b. c, and d (from Reference 3).

measurements of physical properties that, based onempirical correlation, are related to defects andcomponent damage. In other words, most NDEmeasurements are only indirectly related to the flawsize or damage state. These measurements are oftenaffected by geometry and material properties unre-lated to damage and flaws. Changes in test condi-tions may produce entirely false indications andsome comparisons of damage state using differentinspection methods or simply different inspectiontimes may be invalid. Having to rely on flaw cali-bration standards and empirical correlations alsomay make it prohibitively expensive to developNDE methods for all types of damage, materialsand heat treatments, component geometries, andthe various service histories of each component.

The high cost of down time and slow inspectionrates (area or volume of material inspected per unittime) associated with the high-resolution tech-niques necessary for detecting small critical flawsand providing flaw characterization representsanother deficiency of standard NDE methods.Global inspection methods, even with low resolu-tion, are needed to initially determine whether amore detailed inspection is warranted and to mini-mize NDE bottlenecks in returning a system on-line. Acoustic emission is one of the few NDEmethods offering considerable potential for pro-viding a global inspection approach.

Residual-life estimation and fracture-mechanicsanalysis require more than flaw detection. Ensuringstructural reliability when extending components

119

I

beyond their original design life necessitates a quantita-tive determination of the effect of accumulated damn-age on the performance related properties of variouscomponents. When a NDE operator is asked to makethese types of measurements, an entirely different setof problems arise than those encountered in simpleflaw detection. Standard NDE methods are presentlyincapable of handling this new set of problems.

Eddy-current and ultrasonic methods are among themost promising for making the type of damage-relatedmeasurements needed for residual-life estimation, butmodels for quantifying the effect of material propertiesand associated degradation mechanisms are still beingdeveloped. Presently, there is poor agreement betweenmodels and experiments in actual structural compo-nents because models have been well developed onlyfor idealized defects in infinite media. There are defi-ciencies in the present understanding of the interactionmechanisms between the various manifestations ofperformance-related structural damage and nondes-tructive measurements. This requires considerableempirical correlation to establish a reliable relationshipbetween the presence or extent of damage and a non-destructive measurement. The constraints of in-serviceinspection limit most standard NDE to providing onlyqualitative assessments. -

One approach for handling the uncertainties in thereliability of flaw detection and sizing is a probabilistic

treatment of the NDE data. The probabilisticapproach estimates residual life with a specified level ofconfidence given the uncertainty associated with thedetection and measurement of flaws. This does notremedy the inadequacies of standard NDE methodsbut provides an-approacli for assessing the risk associ-ated with relying on the information that NDE canpresently provide.

The ultimate goal is to develop an NDE capabilityadequate for a more quantitative evaluation of residuallife. Realization of this goal requires significantimprovements in flaw detection reliability and the abil-ity to make quantitative measurements of the effectaccumulative, damage has on performance relatedproperties of a component. Discrimination betweentypes of flaws, accurate quantitative flaw sizing, andquantitative measurement of performance-relatedmaterial properties is presently not adequately pro-vided by standard NDE methods. Furthermore, amore fundamental deficiency affecting the develop-ment of the necessary NDE methods for reliable esti-mation of residual life of components is related to thelimited understanding of the relation between the phys-ical properties of the material, the degradation mecha-nisms, and their effect on component performance. Abetter understanding of the effect of damage on thesephysical properties is needed in order to identify thephysical measurements that need to be made.

T

ri

120

.IT

11.4 References

1. H. R. Chin Quan afd'l. G. Scott, "Operator Performance and 'Reliability in NDI," in ResearchTechniques in Nondestructive Testing, Volume'3, New York: Academic Press, 1977.

L2. S. H. Bush, Reliability of Nondestructive Examination, NUREG/CR-3 110, PNL-4584, Volumes 1-3,

October 1983.

3. 0. Buck and S. Wolf (eds.), "Nondestructive Evaluation: Microstructural Characterization and Reli-ability Strategies," Conference Proceedings of The Metallurgical Society of the American Institute ofMechanical Engineers, October 5-9, 1980, p. 405.

4. Y. Segal et al., "A Systematic Evaluation of Nondestructive Testing Methods," in Research Tech-niques in Nondestructive Testing, Volume 3, New York: Academic Press, 1977.

5. R. B. Clough (ed.), Quantitative NDE in the Nuclear Industry, Metals Park, Ohio:- American Soci-ety for Metals, 1982.

6. J. T. Fong and 0. F. Heddon (eds.), Nondestructive Evaluation (NDE): Reliability and Human Fac-tors, National Bureau of Standards Special Publication, Gaithersburg, Maryland.

7. M. J. Davidson et al., "Monitoring for Life Extension," Journal of Pressure Vessel Technology, 107, rAugust 1985.

8. 3. Muscara, Research Program Plan-Nondestructive Examination, NUREG-1 155, Volume 4, July 1985.

9. F. A. Simonen, "The Impact of Nondestructive Examination Unreliability on Pressure Vessel Frac-ture Predictions," Journal of Pressure Vessel Technology, 107, February 1985.

10. G. D. Gupta et al., Remaining Life Evaluation and Extension of Boilers, FWC/FWDC/TR-84/14,May 1984. rL

11. G. H. Harth and T. P. Sherlock, "Monitoring the Service-Induced Damage in Utility Boiler PressureVessels and Piping Systems," Journal of Pressure Vessel Technology, 107, August 1985.

12. G. D. Gupta et al., "A Practical Strategy for Life Evaluation and Extension of Fossil-Fuel-FiredBoilers," Journal of Pressure Vessel Technology, 107, August 1985.

13. W. J. McGonnagle, Nondestructive Testing, second edition, New York: Gordon and Breach SciencePublishers, 1971. L

14. G. P. Hayward, Introduction to Nondestructive Testing, Milwaukee, Wisconsin: American Society ofQuality Control, 1978.

15. R. B. Thompson and D. 0. Thompson, "Ultrasonics in Nondestructive Evaluation," Proceedingsof the Institute of Electrical and Electronics Engineers, 73, 12, December 1985.

16. NondestructiveEvaluation in theNuclearlndustry-1980, Metals Park, Ohio: American Society forMetals, 1981. fl

121

17. D. 0. Thompson and D. E. Chimenti (eds.), Review ofProgress in Quantitative Nondestructive Eval-uation, Volumee2, New York: Plenum Press, 1983.

18. W. H. Lewis et al., Reliability and Nondestructive Inspections, Report SAALC/MME 76-6-38-1, SanAntonio Air Logistics Center, Kelly Air Force Base, Texas, 1978.

122

.U

I

12. ADEQUACY OF ASME CODE IN-SERVICEINSPECTION METHODOLOGY

J. F Cook

Federal regulations1 require that the require-ments of Section XI of the American Society ofMechanical Engineers (ASME) Codea for in-service inspection (ISI) of nuclear power plantClass 1, 2, and 3 components and systems be used..The edition and addenda of Section XI that mustbe followed also are specified in the regulations.,The code edition/addenda to be used for the ISI isdetermined by the date of the'start of plant opera-tion and is normally updated whenever' a.newaddenda is issued. However, the updating of ISIplans to later editions/addenda can be delayed byextended shutdowns. Therefore, the code edition/addenda followed varies among the operatingplants. This is further discussed in Reference 2.Because there is no one set of current ISI require-ments, this chapter discusses the general philoso-,phy of the ISI specified"in' Section'XI, and thestatus of improvements' in the ASME code require-ments relative to selected components important toplant safety and plant life extension. AdvancedNDE methodologies are also discussed.

12.1 ASME Section XI Code Rules

The acceptance criteria in ASME Section XI forflaws found by nondestructive examination (NDE)are based on fracture-mechanics analysis. This is incontrast with the fabrication code (Section 111)NDE acceptance criteria that are based more 'onworkmanship standards. However, the NDE.requirements initially put in Section XI were basedon the same requirements used in the past for thefabrication code NDE. This has resulted -in anincompatibility between the Section XI acceptancestandards and the NDE capabilities. In many cases;the NDE is not capable of detecting the Section XIacceptance limit indications. For, example, inci--dents' of through-wall intergranular stress corro--sion cracking (IGSCC) have been found in piping,

systems by evidence of leakage that were notdetected by standard ASME code ultrasonic inspec-.tion. Studies performed under the Program for theInspection of Steel Components3 (PISC) (formerlythe Plate Inspection Steering Committee) onheavy-section specimens showed inadequacies instandards, ASME code minimum flaw sizes, andvessel ultrasonic examination procedures. Improve-ments have been made and work is in progress tocorrect some of these limitations. Those improve-ments and work in progress are discussed below asrelated to certain components identified as impor-tant to life extension.

The Section XI nondestructive examinations serveas a complement to pressure tests specified inSection XI that must be applied to all safety-class pres-surized components. The pressure tests act as tempo-rary stressors providing for: (a) some verification ofcontinuing structural integrity, and (b) identificationof flaws that cause leakage. Unlike NDE, which is gen-erally applied only to welds and adjacent base metal,the pressure tests check the complete pressure bound-ary. Also, the NDE is done on a sampling plan while allcomponents receive the pressure test. However, thesepressure tests only evaluate the integrity of the pressureretaining boundaries; they do not check other compo-nents important to safety, such as supports, and theydo not'detect small defects that may later grow to acritical size.

The NDE specified in Section XI is done on a sam-ple of the welds. For example, 25°7 of the reactor cool-ant piping system welds must be examined during eachISI interval. If flaws are found, then additional exami-nations are required.'Also components containing flawindications are required to be 'examined more fre-quently in the future (e'g., reexamined during the nextthree successive inspection periods). The more frequentexaminations are designed to determine whether theflaws are growing. Examples of these requirements arecontained -in Sections IWB-2420 and IWB-2430'ofSection XI.'

II

A code requirement that needs to be resolved is theinspection schedule for times beyond the present 40-year plant life. Two different' inspection programs are

a. Wvhen used in this Chapter of~this report, unless otherwisestated, references to the Code or Section Xl means Rulesfor In- presently allowed by the code. One program (Inspec-service -Inspection of Nuclear Power Plant Components, tion Program B) uses four 10-year intervals. InspectionSection Xl, Division I ("Rules for Inspection and Testing of program A uses, from the start of operation, 3 yearsComponents of Light-Water Cooled Plants") of the ASMEBoiler and Pressure Vessel Code. ' - , for the first interval, 7 years for the second interval,

123

I

13 years for the third interval, and 17 years for thefourth interval. Most plants now use Program B. Thisis fortunate because Program A reduces the ISI fre-quency as the plant operates longer and, thus, wouldgive less ISI information for assessing plant conditionrelative to life extension.

The code has historically specified NDE require-ments that included details on how to perform theexaminations (a cookbook approach). The philoso-phy of Section XI of the code is now changing tospecify NDE performance standards, leaving theprocedures up to the performing organizations.The performance demonstration requirements areintended to verify that the NDE system (equip-ment, procedure, and personnel) is capable ofdetecting the minimum size flaws identified in theSection XI acceptance standards. Additional per-formance demonstrations are being required forflaw sizing. Examples of these requirements forpiping welds are contained in a draft of CodeCase N-409, Revision 1, now under considerationby the Section XI ASME Code Committee. Imple-mentation of the performance demonstrationsshould result in more uniformity in the perform-ance of NDE systems (procedures, personnel, andequipment) for flaw detection and flaw length anddepth sizing.

12.2 NDE Improvements

Reference 4 contains extensive information onthe degradation of bolting in nuclear plants. Partlyas a result of the problems that are described in thisreference, the first code document that requiresNDE performance demonstrations (Section XI,Appendix VI, "Ultrasonic Examination of Boltsand Studs") was included in the Winter 1983Addenda. The licensee is required to demonstratethat the procedures and personnel can detect aqualification, notch representing the maximumacceptable size flaw. Data on field experience usingthese requirements should be available followingUSNRC reference of this code addenda in the regu-lations and subsequent update of plant ISI plans.

Piping-ultrasonic examination practices havereceived great attention in the code because theproblems. with IGSCC in boiling water reactorplants have increased. As a result of the IGSCC ofcertain larger pipes,. first discovered at the NineMile Point Plant, two USNRC Inspection andEnforcement (I&E) bulletins have been issued. 5 ' 6

These. bulletins require qualification, of pipingexamination procedures and, personnel on blind

specimens containing IGSCC. These practices weresubsequently covered by Code Case N-409. 7 Addi-tional ASME code changes that are being devel-oped may add performance demonstrationrequirements for all piping weld ultrasonic exami-nations. This activity will significantly improve theeffectiveness of ultrasonic examination practicesfor finding flaws in piping welds. However, thereare still problems in developing effective ultrasonicprocedures that apply to all piping. The geometryof welds and adjacent fittings often prevent theaccess required for conducting a complete ultra-sonic examination. 8 In many cases, even specialultrasonic techniques will not penetrate cast stain-less steel piping and fittings. 9

Changes were published in the Winter 1983Addenda to Section XI that simplified selection of,and reduced the number of, Class 2 welds to beincluded in examination samples. This change alsoimproved examination methodology in that allsurface-only examinations were eliminated. Datacollected from piping failures show that most prob-lems (e.g., IGSCC and thermal fatigue cracks)originate from the inside surface. Similar changesshould be made to Class I examination require-ments. At present, Class I piping with less than a4-in. nominal pipe size receives only a surfaceexamination. Confirmatory data on the effective-ness of ultrasonic examinations on small pipingwould help in obtaining code committee approvalof changed Class I requirements.

Improvements in vessel ultrasonic examinationrequirements are presently being acted on by theASME Boiler and Pressure Vessel Committee. It iscommon practice now for most foreign and domes-tic plants to apply additional examinations beyondthe code minimum. As a result of internationalcooperative studies on ultrasonic vessel examina-tion practices known collectively as PISC and rec-ommendations from the Pressure Vessel ResearchCommittee, requirements for an additional exami-nation (70-degree examination angle) for near-surface flaws (e.g. underclad cracks) are beingadded to the code. In this same action, caution isbeing added to the code language regarding usingamplitude/search unit movement techniques forflaw sizing. The Section XI code action in progressis scheduled to be published as Appendix I to Sec-tion XI. Another significant improvement in thiscode action is a requirement for more sensitiverecording that will reduce the chance of missingflaw indications.

Containment structure leakage rate testing is con-trolled by Appendix J of the USNRC regulations. As

rL;

3

124

I

well as the leak testing, a general visual examination ofaccessible internal and external surfaces is required.However, minimum accessibility requirements are notgiven. Thus, the extent of the examinations probablyvaries among the plants. Subsection IWE ofSection XI of the code contains rules for the examina-tion of metal containment structures. USNRC actionto require use of these rules is in progress. When imple-'mented, these subsection IWE requirements, which areconcentrated on welds and other important parts of thecontainment structure (e.g., bolting, seals, and gas-.,kets), should result in more consistent metal contain-ment examination practices. Most of the examinationsare visual and conducted on painted surfaces. Refer-ence 10 indicates that visual examinations can appro-priately detect degradation in painted metal structures.

Rules for concrete containment examinations arenow approved for Code' publication asSubsection IWL of Section XI in the Winter 1986Addenda. Additional tests that could enhance thenew subsection IWL rules relative to life assessmentfor concrete containments are given inReference 10. These tests, controlled by Britishstandards, include checking carbonation depth,ultrasonic pulse velocity, covermeter, and reboundhardness measurements. These tests will 'bereviewed and compared with the American Societyfor Testing and Materials practices. Following this'review, recommended changes will be presented tothe ASME for improving concrete containmentexamination practices relative to plant-life exten-sion. Previous work at the Idaho National Engi-neering Laboratory involved correlating ultrasonicpulse velocity measurements with changes in con-crete properties following exposure to radiation. 1

Reference 10 emphasizes the importance ofdestructive tests for determining life forecasts ofconcrete structures.

The Section XI steam generator tubing eddy-current examination requirements are based on thesingle-frequency methodology published in RegulatoryGuide 1.183.12 However, most examinations are nowperformed using multiple-frequency methods, whichcan enhance detection of flaws in areas such as thoseadjacent to tube-support structures. The code has notyet required that the eddy-current procedures be quali-fied for detection of cracks in tubing. Calibration usingflat bottom holes (simulating wall thinning) is speci-fied. The eddy-current response to a flat-bottom hole isusually significantly different than the eddy-currentresponse to a crack. Other steam generator examina-tion practices that are now being used by industry suchas ultrasonic testing of tubing, profilometry measure-ments and radiography are scheduled to be covered by

the ASME; work has been started at the task grouplevel of the ASME code committee. Further discussionof steam generator'examination practices is included inChapter 6 of this report.

The Section Xl NDE requirements are focused onfinding discrete flaws in components; detection of gen-eral degradation (e.g. radiation damage, thermalaging) of material properties is not covered per se.However, control specimens for destructive testing arespecified for inferring reactor vessel material propertiesafter periods of operation. As part of checking materi-als for aging relative to life extension, it may be desir-able to have material property data on'othercomponents such as piping sections, supports, pumps,etc. Samples of these components are not now availa-ble for destructive tests. However, miniature specimentechniques (see Chapter 14 of this report) may proveuseful in this regard. Also, some of the required prop-erties such as yield strength, tensile strength,: andductile-to-brittle transition temperature may beinferred from NDE data using advanced techniques, asreported in References 13 through 17. These referencesindicate that measurements of ultrasonic attenuationland velocity can be used to infer certain mechanicalproperties because the mechanical and ultrasonic prop-erties are all dependent, to a degree, on material micro-structure. Measurement of changes in the mechanicalproperties because of environmental factors, (such asthermal aging and radiation damage) should be possi-ble for a specific material once calibration data for thematerial have been obtained. The results to date havebeen obtained on laboratory test specimens; applica-tion of the techniques to actual components in the fieldmust await perfection of ultrasonic instrumentationand data processing methods now being developed thatwill allow measurements where only one material sur-face is accessible and will minimize surface and ultra-sonic coupling variations. Further confirmatory testsand field demonstrations will be required before any ofthese techniques can be included in the ASME code.Addition of material property NDE requirements toSection XI would significantly improve the knowledgeof the actual in-service properties, and also mightreduce the need for testing surveillance specimens.

12.3 ASME Code ISI and LifeExtension

The current Section XI ISI requirements canprovide only part of the information needed formaking decisions regarding life extension. Severalweaknesses in NDE practices need to be resolved.For example, in most cases, the code does not

I.-

6F

fL

125

I

require any backup examinations to confirm resultsfrom the primary examination. However, selectedutilities are applying backup examinations. If anultrasonic examination finds indications of cracksin a pipe, it may be desirable to use an inspectionport to go inside the pipe with a remote visual orsurface examination to backup the ultrasonicexamination. Geometric conditions in componentssometimes produce indications that are confusedwith indications from flaws. It is common practiceto always specify a backup examination for theUnited States Air Force nondestructive inspectionprograms. Backup examinations will increase theconfidence in NDE and provide better informationon the condition of components relative to plant-life extension.

Improvements are being made in the NDE meth-odology to enhance the reliability of flaw detec-tion. After a flaw is found in a component, the flawmust be characterized (sized) to determine theeffect of that flaw on the life of that component, orelse the component must be repaired or replaced.Experience with the IGSCC problems in piping hasshown that qualifying equipment and personnel forsizing is more difficult than qualifying for detec-tion. Applying standard ASME code rules for siz-ing of flaws in vessels can in reality end up giving

the size of the ultrasonic soundbeam. Correctiveaction is underway as shown in Supplement 12 ofAppendix I of the code. A significant problem indeveloping flaw sizing requirements will be defin-ing the qualification specimens. The economics offabricating an adequate number of vessel speci-mens will slow the code action on this problem.However, life-extension considerations mayincrease interest in solving this problem.

12.4 Summary, Conclusions, andRecommendations

ISIs of nuclear power plants are controlled by theUSNRC regulations and Section XI of the ASMEcode. Improvements in the ASME code NDE meth-odology relative to flaw detection are evolving andare summarized in Table 12-1. These improvedcode rules should be applied as part of evaluatingplant conditions relative to license renewal. Currentcode methodology is not wholly adequate to assessthe residual life of the major light water reactorcomponents. More efforts are needed to developfield-usable nondestructive examination tech-niques and equipment that can provide valid infor-mation on the sizing of defects and the materialproperties of the major components.

126

I

I

Table 12.1. Summary of code NDE improvements

Component NDE Methoda NDE Improvements Status

Bolting UT

Piping UT

Demonstration of procedure andpersonnel qualification

Procedture and personnelqualification using statisticallybased specimen sets, containingflaws

Near surface examinationrequirement, correction of flawsizing techniques

Addition of Code examinationrules to supplement theAppendix J leak testing

Published, W-83 Addenda

In committee, may be in;W-87 Addenda

In committee, may be in-W-86 Addenda .

NRC review in progress

Vessels UT rL

Containments VT, LT

Steam generators ET, RT, UTProfilometry

Code rules to cover all types ofNDE used by the industry

In committee

Ia. . I .

a. UT, VT. LT, ET, and Kr refer respectively to ultrasonic, visual, leak testing, eddy-current, and radiograph nondestructive examination methods.

L-

r

127

12.6 References

1. Code of Federal Regulations, Title 10-Energy, Part 50-"Domestic Licensing of Production andUtilization Facilities."

2. J. F. Cook and B. W. Brown, Evaluation of Nondestructive Examination Requirement Changes withUpdates of Section XI, Division I of the ASME Code, EGG-SD-6893, November 1985.

3. Commission of the European Communities, A Summary of the PISC-II Project, PISC-II Report No. I,June 1985, General Directorate XII, Joint Research Centre, Ispra Establishment, S.A./l.07/Cl.85.13.

4. Proceedings of the Bolting Degradation and Failure in Nuclear Power Plants Seminar, Revision 1,EPRI, December 1983.

5. U. S. Nuclear Regulatory Commission, "Stress Corrosion Cracking in Thick-Wall, Large Diameter,Stainless Steel Recirculation System Piping at BWR Plants," USNRC Inspection and EnforcementBulletin 82-03, October 1982.

6. U. S. Nuclear Regulatory Commission, "Stress Corrosion Cracking in Large Diameter Stainless SteelRecirculation System Piping at BWR Plants," NRC Inspection and Enforcement Bulletin 83-02,March 1983.

7. American Society of Mechanical Engineers, "Procedure and Personnel Qualification for Ultrasonic Detec- Ltion and Sizing of Intergranular Stress Corrosion in Austenitic Piping Welds," Code Case N-409,December 1984.

8. B. L. Baker, J. F. Cook, and H. A. Wilkin, Evaluation and Recommendationsfor Ultrasonic Exami-nations of ASME Class I and 2 Piping Melds, EGG-FM-5034, October 1979.

9. J. Muscara, Research Program Plan-Nondestructive Examination, NUREG-1 155, Volume 4, 1985.

10. G. H. Lewis, Degradation of Building Materials Over a Lifespan of30-100 Years, Commission of the [European Communities, Final Report, Contract No. DE-A-001-UK, 1985.

11. A. D. Perkins, Analysis of Datafrom Irradiated Concrete Testing, EG&G Idaho, Inc., RE-M-80-002,March 1980.

12. U. S. Nuclear Regulatory Commission, "Inservice Inspection of Pressurized Water Reactor SteamGenerator Tubes," USNRC Regulatory Guide 1.83, Revision 1, July 1975.

13. R. Klinman, G. R. Webster, F. J. Marsh, and E. T. Stephenson, "Ultrasonic Prediction of Grain LSize, Strength, and Toughness in Plain Carbon Steel," Materials Evaluation, 38, 1980, pp. 26-32.

14. R. Klinman and E. T. Stephenson, "Ultrasonic Prediction of Grain Size and Mechanical Propertiesin Plain Carbon Steel," Materials Evaluation, 39, 1981, pp. 1116-1120.

15. B. E. Droney, "Use of Ultrasonic Techniques to Assess the Mechanical Properties of Steels," inNondestructive Methodsfor Material Property Determination, C. 0. Ruud and R. E. Green (eds.),New York: Plenum Press, 1984.

128

16. A. Vary, "Correlations Among Ultrasonic Propagation Factors and Fracture Toughness- Properties ofMetallic Materials," Materials Evaluation, 36, 1978, pp. 55-64.

17. A. Vary, "Correlation Between Ultrasonic and Fracture Toughness Factors in Metallic Materials," inFracture Mechanics, ASTM Special Technical Publication 677, 1978.

L

U

129

I

13. CURRENT LIFE ASSESSMENT TECHNIQUESC. E. Jaske

Both experimental and analytical techniques areused to assess the remaining life of components thatare in-service. These include testing of surveillancespecimens, monitoring of operational parameters,evaluation of samples removed from service-exposedcomponents, and prediction of damage accumulationprocesses. In actual practice, one or some combinationof these techniques typically are used in makingresidual-life assessments.

Surveillance programs are used to assess theinfluence of neutron irradiation on the mechanicalproperties of materials used in the beltline region ofreactor pressure vessels (RPVs). Owners of lightwater reactors (LNVRs) are required by federal regu-lations to maintain such programs, and AmericanSociety for Testing and Materials (ASTM) Desig-nation: E 185 defines the "Standard Practice forConducting Surveillance Tests for Light-WaterCooled Nuclear Power Reactor Vessels," whichprovides guidelines for a minimum surveillanceprogram. Samples of base metal, weld metal, andheat-affected-zone metal are enclosed in surveil-lance capsules. The capsules are placed within theRPV at locations such that the amount of irradia-tion leads is greater than that at the inside surfaceof the beltline region by a factor of I to 3. A sched-ule for capsule removal and sample testing is estab-lished so that the long-term effects of irradiationcan be assessed. Both tension and Charpy V-notchtests are required. Other tests, such as hardness andfracture toughness, are optional.

The status of surveillance programs in the UnitedStates has been described in detail elsewhere. I The inte-grated surveillance program used for Babcock &Wilcox plants has been described by Lowe.2 Data fromthe surveillance programs provide indications of theeffects of radiation on tensile strength and Charpy-impact properties. Figure 13.1 shows the correlation ofbrittle-to-ductile transition temperature shifts at30 ft-lb (41 J) with tensile yield strength. A similar cor-relation of fractional decrease in upper-shelf energy isshown in Figure 13.2. These two figures are from thework of Odette,3 and they show that consistent trendsexist. The major shortcoming of such data is that theirrelationship to more relevant properties for fracture-mechanics analyses, such as fracture toughness andcrack-arrest resistance, is not well established.

Operational parameters, such as temperatures,pressures, and amount of cycling, can be moni-tored.to provide an assessment of remaining life.

These parameters are not direct indicators of thematerial's condition or state, so some type of dam-age accumulation algorithm is needed to estimatetheir impact on material degradation. For example,if thermal fluctuations were monitored at a criticalnozzle region, one would have to compute the cor-responding thermal strain history in that regionand then use an appropriate fatigue-damage accu-mulation model to calculate estimates of expendedlife and remaining life. A practical example of suchon-line monitoring of the condition of an oil-fired1000-MW boiler is presented by Davidson.4 Asrecently pointed out by Toman,5 equipment moni-toring techniques should be (a) nondisruptive, (b)noninvasive, (c) repeatable, (d) accurate, and (e)cost-effective, in addition to yielding data that arecorrelatable with the deterioration process. As anexample, he discusses the monitoring of the condi-tion of an emergency diesel-engine generator sys-tem for a nuclear power plant.

Another approach to life assessment is to directlymeasure the properties of samples that have beenremoved from components. Typically, a boat, core,or ring sample is removed from the component,and the component is then repaired using proce-dures that satisfy the applicable regulations andsafety codes. As will be discussed later inChapter 14 of this report, there has been somerecent interest in removing miniature samples toeither eliminate or greatly reduce the amount ofrequired repairs. The main advantage of thisapproach is that it provides a direct measure of themechanical properties of the material after serviceexposure. However, archival data for that samematerial are not usually available for comparisonand assessment of the degree of degradation. Usingtypical properties for that comparison and assess-ment adds to the uncertainty of the analysis.Another problem with this approach is that thetests must be accelerated to obtain information in areasonable time period during an outage. If a time-dependent degradation mechanism, such as corro-sion or corrosion-fatigue behavior, is beingevaluated, the results of the accelerated testingmust be extrapolated to make remaining service-lifepredictions, and the extrapolation introduces addi-tional uncertainty.

As an example, the direct properties measure-ment approach was used by Battelle to provide datafor assessing the structural integrity of Type 316

11

L

L

130

U:

I

Li

0

i>

a Avy (MPa) Aay (MPa)7-3077

Figure 13. 1. Plots of transition temperature' shifts indexed at 30 ft-lb'(Al J) versus static' yield stress changes for (a)weld and (b) plate and forgings.

0.7

ILi

IC

a)V)co

0

0

ILL

USEO =

USEO =

Upper shelf energy forunirradiated material:

Upper energy forIrradiated material

[

EI.

00 :' hAUy (MPa) 400

7-3037

Figure 13.2. Fractional decreases in C, upper-shelf energy versus yield stress changes... .. .. .. .:?/ - .1 ' , .. *.,

* L

131

I

stainless steel weld metal from the Maine Yankeereactor coolant system.6 Fatigue test bars wereremoved from boat samples of weld metal that con-tained microfissures. Low-cycle fatigue specimenswere machined from those bars and tested usingstandard procedures comparable to the test proce-dures that are used to develop fatigue data for theestablishment of American Society of MechanicalEngineers (ASME) code design fatigue curves.-Thefatigue data then were used to evaluate structuralintegrity in terms of the expected fatigue usage fac-tors and the safety factors incorporated in theASME code. As another example, tensile, fatigue,fracture toughness, and Charpy specimens havebeen taken from steam-turbine shafts to measureproperties for use in remaining life analysis.

Metallographic and fractographic examinationof samples from in-service components also is usedto help'assess remaining life. For example, if volu-metric nondestructive evaluation (NDE) has indi-cated the presence and growth of a defect or both, asample of material containing the indicated defectmay be removed and sent to a laboratory fordetailed examination using conventional metallur-gical techniques to ascertain the nature and extentof the defect. Suspected regions of intergranularstress corrosion cracking (IGSCC) in boiling waterreactor (BWR) piping systems have been examinedin this manner. In fossil plants, ring samples ofsuperheater and reheater tubes are often examinedmetallographically to measure the degree of fire-side and steam-side corrosion and the amount ofcreep damage.8 Similar samples of furnace tubesfrom petrochemical plants have been examinedmetallographically to measure the amount of creepvoids and fissures, where an empirical correlationhas been established between those factors and theservice-failure rate of the furnace tubes.8

The analytical prediction of damage accumulation isan important technique used in remaining life assess-ment. In fact, some type of prediction of the accumu-lation of damage during anticipated future operationsis required to make any estimate of remaining life. Themodels of material degradation processes discussedpreviously are used to make the required calculationsof damage propagation starting from the current dam-age state of the material.

If no measure of the current damage, state isavailable, one can estimate it by calculating thedamage accumulation during the past operatinghistory. However, knowledge of the past operatinghistory and the condition of the material that wasoriginally put into service is often incomplete, add-ing a great deal of uncertainty to this calculation-

of-current-condition approach. Thus, somedestructive or nondestructive measurement of thecurrent condition of the material is preferred forremaining life analysis.

Analytical techniques for the prediction ofremaining service life have been pioneered by theaircraft industry, where NDE and a fracture-mechanics-based, damage-tolerant approach isemploy~ed. 9

. Complex computer programs havebeen developed to compute fatigue-crack growthand assess the potential for fracture in structuralcomponents throughout the course of variousexpected loading spectra. Because of the historydependence inherent in niaterial degradation proc-esses, the damage accumulation'calculations areperformed on event-by-event (cycle-by-cycle forfatigue) basis throughout the course-of the simu-lated loading. For this reason, the computationsare usually time-consuming and demanding ofcomputer resources.

The type of analytical procedure described byJaske8 is often used to assess the remaining life ofsteam-turbine rotors. A flow diagram showing thetechnologies used in performing the life-assessmentanalyses is presented in Figure 13.3. Examples ofthe results are presented in Figure 13.4. Figure 13.4(a) shows the ultrasonic indications that areobtained from boresonic inspection of a rotor.Cluster analysis is used to group the indications inregions that are considered to be single flaws in theflaw-growth analysis. The remaining parts of thatfigure show the calculated stress distributions attwo flaw locations, the expected loading spectrumfor the rotor, and the predicted flaw-growth behav-ior during future service.

In performing plant-wide remaining-life assess-ments, it is useful- to have simplified, rule-basedapproaches that can be applied rapidly and. easily tocompute conservative life estimates. Then the detailedand extensive analyses are performed only for the caseswhere the simplified rules show that a potential prob-lem may exist. Simplified methods are usually applica-tion specific, so such procedures should be developedfor key reactor components, such as RPV welds, noz-zle welds, piping Xwelds, coolant pumps, diesel genera-tors, and internals, where the possibility of fatiguedamage is of concern.

Engle, 10 for example, has developed a simplifiedapproach for estimating fatigue-crack growth intypical aircraft structural components under spec-trum loading. Figure 13.5 shows a comparison ofpredicted versus experimentally measured fatigue-crack growth for several design limit stress levelsthat were made using his simplified approach.

r

F

I,

132

I

Analysis ofInspection data

* G.E. inspection 1979

* Data plots/analysis

Stress and thermalanalysis

(NASTRAN/AXISOL)

* Finite element models/analysis

* Isostress contours

* Flaw stresses and peakstresses

. , i

.. , 1 r :

Identify NDI flaws

* Cluster analysis

.* Potential flawI link-ups~~ ~~~~~~~~~~~~~~~~ I

Rotor steel materialfracture response data

* Literature/in-housedata daldN, Kic, KC

Residual lifeI analysis

* NDI and assumedflaws

* Service stressspectrum-dutycycle

Operational -

history data

* Cold/hot starts

* Cold/hot overspeeds

'- Structural Integrity- evaluation

* Run/retire decision

* Reinspection Intervals

* Operation guides

L

7-3043

Figure 13.3. Life assessment analysis approach for steam-turbine components.

I

133

I

I

150 -140 -130 _

's 120 -a 110 P

0,100CDa 90 _

80 -70 -60 _50 2

42

I._ i

Area of holding indic,

Sl 3

Flaw 3

* : 4.X Individual indications0

a

43 44 45 46 47 48 49 50 51 52Axial position (in.)

54 55

I(a) Boresonic data for rotor area showing flaw Indications

._

ci,0

80 l l l

60

40-Growth

20 _

0 10 1 2 3 4 5

r (In.)Flaw 1 section

601I-I

0t

. 80

401

I I I I

-W Growth

I I I

1 2 3 4r (In.)

Flaw 3 section

I20

5

(b) Circumferential stress distributions

I

=1

0.

&V

FractureFlaw 1 IFlaw 3>Zz

Mao = 0.125 in./12co = 2.00 in. /

.- As~~~~~~~~~.1

I I II. . . . . .

0 2Cycles 4 6 .8Years of operation

10 12

(c) Yearly service spectrum for turbine rotor (d) Spectrum loading fatigue crack growthbehavior of bore flaws 1 and 3

7.3080

Figure 13.4. Example of results from a life assessment of a steam-turbine rotor.

134

I

I

I

7000.-Flights. 30

7-3060

Figure 13.5. Correlation of predicted with measured fatigue-crack growth rates under spectrum loading. 1.Saxena et al.1I applied fracture-mechanics tech-.nology to predict the residual life of a ships serviceturbine generator (SSTG) casing using the generalapproach illustrated in Figure 13.6. The region ofconcern had a complex configuration as is shown inFigure 13.7 and was subjected to cyclic thermalstresses. They concluded that casings with cracks<6.3 mm deep could be operated safely'-foranother 800 cycles, and they recommended inter-vals for future periodic inspections. This same typeof residual-life analysis could be applied to nuclearreactor components with appropriate modifica-tions to the materials properties, configurationalmodels, and failure criteria.

Broek9 has summarized the six basic types ofinformation that are needed to apply a fracture-mechanics approach:

* The minimum detectable flaw size and.type

* A prediction of the residual strength of thestructure with a flaw present and of thecritical flaw size associated with the fail-safe load

* A description of the expected future load-ing history

* A determination of the crack growth responsefrom the minimum detectable size to the criti-cal size associated with failure

A knowledge of the critical locations in thestructure where cracks are likely todevelop, such as geometric discontinuitiesassociated with fatigue-loaded weldsA reliable inspection of the various impor-tant regions of the structure or compo-nent, including consideration of theaccessibility to those regions.

With current NDE methods, defects can be detectedwith relatively high confidence. However, as discussedin chapters 11 and 12, !the sizing and classifying of'those defects has much lower confidence and is an areawhere major improvements are needed and wheremuch NDE research is concentrating.

Predictions of residual strength and critical flawsize cani be made with reasonable confidence forbrittle fracture where linear elastic fracturemechanics can be applied. For nuclear pressureboundary comnponents, however, the possibility ofductile fracture is often of concern. Inelastic frac-ture mechanics methods for dealing with ductilefracture problems in piping and pressure vessels arestill evolving. Those methods need to model leak-before-break and crack-arrest, so that fail-safe con-ditions occur at the critical crack size. Forapplications to actual components, these methodsneed to be developed and verified for dynamic aswell as static loading conditions.

I

I

135

I

Elastic finite element Cyclic stress-strainanalysis behavior

* Constrained p lasticity analysisusing Neuber approach Analytical

verification of* Obtain elastic-plastic stress Neuber approach

distribution

* Calculation A J

* Convert to equivalent AK

= I

* Obtain A K versus crack length

Selection of failure criteria

th Crack growth analysis Crack sizer da f (AK) input fromie dN nondestructive

N= lai f (^ K) ,inspection, ai

P red ictedremaining life

as a function ofinitial crack

depth /7-3074

Figure 13.6. Methodology for remaining life prediction of SSTG casings.

136

1

I

Plane C-

Plane B

IPlane A,

2114% Cr. 1% MOLY

r

2-ft 3 114-In. dia.1.0 in. = 2.54 cm

I

7.3059

Figure 13.7. Illustration or configuration of region where remaining life of SSTG casing was assessed. 'r

Descriptions of expected loading histories are The cyclic crack growth behavior of pressure ves-obtained from past experience and from postulated ' sel steels in LWR environments has been well char-worst-case scenarios that may' be encountered! acterized and that wo'rk is being xtended to pipingBecause damage accumulation is strdngly history' materials. One area of concern is at very low cyclicdependent, it is important that such loadinghisto- growth rates where there is so'me debate over theries be Well defined.- Unfortunately,`as pointed out existence of 'a threshold for 'cyclic crack growth.by Yahr et al.,'12 this is noi the case in' the nuclear Additional work is needed to resolve this question.industry, so they had to develop a unit histogram Asdiscussed in the earlier chapters of this report,for the Zion-I plait for usein iheir fatiguedamage ',',the areas where cracks are likely to 'develop inanalyses. This type of work needs to be pursued'sot nuclear-plant 'components have 'been 'identifiedthat sets of reference histograms can be developed from operating experience and original design cal-for routine use in remaining life calculations. ' culations. These include transition regions in the

L;

137

_. _1

I

RPV, closure studs, safe end welds, nozzle connec-tions, elbows, control rod drive mechanism pene-trations, and other welded regions in the pressurevessel and piping system.'.'

Inspections of important regions of the nuclear-plant components are. performed on a scheduledbasis. Improved quantitative descriptions ofdetected flaws are required for use of a fracture-mechanics approach.

All six of the items needed to apply a fracture-mechanics approach are actually probabilistic innature. Ideally, they should be quantified andemployed in a probabilistic remaining life assess-ment methodology.

Recently, Sundararajan1 3 has critically reviewedthe state of the art in probabilistic assessment of thereliability of pressure vessels and piping. Most ofthe applications of that methodology have been inthe nuclear power industry. Figure 13.8 shows thepredicted failure rate of PWR RPVs for variousreactor shutdown pressure-temperature paths at theend-of-life (EOL) fluence. For life extension, eitherthe core loading patterns would have to be modi-fied to prolong the attainment of the EOL fluenceor the allowable pressure-temperature path wouldhave to be modified to provide an acceptable risk offailure. In residual-life assessment, this type ofanalysis should be made on a plant-specific basis toallow for different materials properties and NDEresults.

As is illustrated in Figure 13.9, Sundararajan13

also has predicted the probability of protective bar-rier penetration from random fragment strike. Ifthe effective thickness or the penetration resistanceof the barrier is reduced during service exposure,then the probability of penetration is increased. Forlife extension, the barrier must be maintained andrefurbished as required to maintain an acceptablylow level of perforation probability.

Predictive methods for environmentally assisteddegradation and cracking are not as well established asthose for fatigue. A recent ASME symposium14 wasdevoted to documenting the current state of the art inthis technical field. The typical approach to solvingcorrosion'and stress-corrosion cracking problems hasbeen to adopt fixes that will eliminate or minimize theoccurrence of those problems in the future. 15,16 Thatapproach has been largely successful, so the predictivemethods may not be required. However, in order toextend plant life, the components and regions of com-ponents that are'potentially susceptible to environmen-tally assisted degradation must be identified byinspection so that the appropriate fixes can be imple-mented.

In the area of environmentally assisted fatiguecracking, a great deal of work has been devoted todeveloping predictive models14, 17 ,18 andSection XI of the ASME Code requires assessmentof environmental effects on fatigue-crack growth.Bamford1 9 has reviewed the basis for referencefatigue-crack-growth rate curves that are presentlyincluded in Section XI. Based on detailed reviewsoftthe EPRI Database on Environmentally AssistedCracking (EDEAC), revisions of those referencecurves are being considered. 20 The EDEAC alsocontains data from slow-strain-rate and stress-corrosion-cracking tests, as well as the data fromfatigue-crack-growth tests. Thus, it provides animportant resource of very relevant materials prop-erties data that can be used in developing methodsfor residual-life assessment.

Gilman21 ' 22 used EDEAC to develop a usefulengineering model for predicting corrosion-fatiguecrack growth rates in LWR environments. The basisfor that model is illustrated in Figures 13.10 and13.11. The lower bound of the cyclic crack growthrate in BWR water, which coincides with the upperbound of the cyclic crack growth rate in air, isdefined as the base rate. The actual time rate ofcrack growth in LWR water then is correlated withthe base rate. Figure 13.10 shows that there is agood correlation between the growth rates in BWRwater and the base rate times frequency, andFigure 13.11 shows that there is a good correlationbetween the upper bound of the growth rates inpressurized water reactor (PWR) water and thebase rate times frequency. These correlations pro-vide a useful tool to help estimate the remainingservice life in areas of the RPV where corrosion-fatigue crack growth is a potential problem.

13.1 Summary

The major life-assessment techniques availablefor application to LWR components are summa-rized in Table 13.1. For each technique, its charac-teristic features and its applications and/orpotential applications are briefly highlighted.

Testing of specimens from surveillance capsulesis used to assess the amount of radiation embrit-tlement that occurs in RPV steels during serviceand is an important technique for ensuring thestructural integrity of the RPV at its EOL fluence.Monitoring of operating parameters to establishaging trends is not widely practiced, but with thedevelopment of appropriate baseline data it canprovide a viable, economic tool to assist in

I

I

I

I

I:

138

I

I

^ff^n

2000 g

n 1500a-

0 0DE 1000

I.

104 failure perevent Act . // -

10-5 failure per event , //10-6 failure per'event //10 7 failure perevent' j , < / /

.0-~

, " t'0 Code limit curve0.23% Cu and 23.0 F IRTNDT

EOL fluence

[

5001

l 1I I.----

100 200 300Temperature (OF)

400 500

743058

Figure 13.8. Constant failure rate and code-allowable pressure temperature paths for reactor shutdown at EOLfluence.

I

Z

conr0Cl-

0.0

a)-0-

I

- -Barrier thickness (T).7-3038

F i Iur e 3 P o r f o r a t i o n . b hFigure 13.9. Probability of perforation versus barrier thickness.

139

I

Crack growth rates, A533B-1, 0.025% S in 288°C BWR water, R = 0.7, 0.8

10-3.

105-4

10-5

2 10-6

", aK %S Keyp 1 o 6 / O / ~~~~~~~~~~~~~111 .2

010-7 - 16.5 025 vS / 4 ~~~~~~~~~~~22 0.2

33 0

108 10-7 10-6 10-5 10-4 10-3

Base rate, 6.2 x 10-8 AK 2 -5 x frequency (mm/s)7-3081

Figure 13.10. Time rate of fatigue-crack growth in BWR water as a function of base rate.

Crack growth rates, A533B-1,in 2880C PWR water, R 0.5-0.8

10~-3

3.1 ~~~~~AK %S Key20-6 6 0 1 1 0.025 A

S ~~~~~16.5 0.025 vA ~~~ ~~22 0.025

O 10-7 33 0.025 0A ~~~ ~~44 0.025 0

44 0.01310-8 * ''it'1 II1,11 I I I Iufill

1 0.8 10-7 10.6 i0-5 10-4 1-Base rate, 6.2 x i0-8 A K2-5 x frequency (minis)

7.3057

Figure 13.1 1. Time rate of fatigue-crack growth in PWR water as a function of base rate.

140

I

I

I,I

I

I'

I

Table 13.1. Summary of current life assessment techniques

CharacteristicsTechnique Applications

Surveillance programs

Monitoring ofoperations

Testing of specimensremoved fromcomponents

Metallographic andfractographicexamination of samplesremoved fromcomponents

Analytical predictionof damage accumulation

Uses specimens exposed to ' ' Measuring the radiation embrittlementneutron irradiation in capsules of RPV steels. Limited amount ofwithin the RPV to measure test material is usually available.tensile and impact properties. The relation of tensile and impact

properties to fracture toughness is notwell defined.

Measure and record tempera- Used to assess condition oftures, pressures, and cycling mechanical equipment, suchfor trending of performance. as pumps, bearings, engines, and

generators.'Need baseline data fornormal operating conditions.

Measure mechanical properties Used on piping welds in regionsof boat, core, or ring samples where repair is possible. Not widelyremoved from the component. used because material removal is

destructive. Need to develop methodsfor testing miniature specimens.

Measure metallurgical aging, Used on piping welds in regions wherevoid formation, and amount of repair is possible. Not favored becausemicrocracking. Identify removal of material is destructive.mechanisms'of cracking. Need to develop methods that use

miniature samples or surface replicas.

Compute the amount of Widely used for areas where highdamage accumulation expected amounts of fatigue damage mayduring future operations.. occur. Need to develop realistic

engineering approaches.

I

I

Iremaining life assessment. The destructive testingand examination of samples removed from com-ponents provides useful data, but is not useful forroutine applications because repairs must bemade.

Analytical predictions of damage accumulationsare widely used in both design and remaining-lifeassessments. Current approaches typically use very

simplified damage accumulation algorithms andlarge safety factors to ensure conservative predic-tions. More realistic approaches that allow theincorporation of history effects, fracture'mechan-ics, and probabilistic factors need to be developed,especially to resolve differenice between the ASMECode Section III and Section XI approaches tofatigue damage analysis.

r

EL.

141

I

I

13.2 References

1. American Society for Testing and Materials, Status of USA Nuclear Reactor Pressure Vessel Surveil-lancefor Radiation Effects, ASTM Special Technical Publication 784, Philadelphia, 1983.

2. A. L. Lowe et al., "Integrated Reactor Vessel Material Surveillance Program for Babcock & Wilcox177-FA Plants," in Effects of Radiation on Materials, Twelfth International Symposium, AmericanSocietyfor Testing and Materials, ASTM Special Technical Publication 870, 1985, pp. 931-950.

3. G. R. Odette, P. M. Lombrozo, and R. A. Wullaert, "Relationship Between Irradiation Hardeningand Embrittlement of Pressure Vessel Steels," in Effects of Radiation on Materials, Twelfth Interna-tional Symposium, American Society for Testing and Materials, ASTM Special Technical Publication870, 1985, pp. 840-860.

4. M. J. Davidson et al., "Monitoring for Life Extension," in Residual-Life Assessment, Nondestructive Exam-ination, and Nuclear Heat Exchanger Materials, Proceedings of The American Society of Mechanical Engi-neers 1985 Pressue Vessels and Piping Conference, New Orleans, Louisiana, June 23-26, 1985, PVP-Volume98-1, pp. 65-71.

5. G. J. Toman, "The Measurement of Equipment Degradation," Proceedings of the Thirteenth WaterReactor Safety Research Information Meeting, NUREG/CP-0072, Volume 3, 1985, pp. 319-329.

6. R. D. Buchheit et al., Laboratory Studies of Type 316 Stainless Steel Weld Metal From the MaineYankee Reactor Coolant System, Final Report to Maine Yankee Atomic Power Company, 1971.

7. B. N. Leis et al., "Fatigue and Fracture Properties of Steam Turbine Shaft Materials," FractureMechanics: Fourteenth Symposium- Volume II: Testing and Applications, ASTM Special TechnicalPublication 791, 1983, pp. 11-101-II-I 19.

8. C. E. Jaske et al., "Life-Assessment of High-Temperature, Fossil-Plant Components," EPRI Con-ference on Life Extension and Assessment of Fossil Plants, Washington, D.C., June 1986.

9. D. Broek, Elementary Engineering Fracture Mechanics, Boston: Martinus Nijhoff Publishers, 1978.

10. R. M. Engle, Jr., "A Simplified Approach for Assessing Spectrum Loading Effects in PreliminaryDesign," Advances in Life Prediction Methods, New York: The American Society of MechanicalEngineers, 1983, pp. 223-227.

11. A. Saxena et al., Residual Life Prediction and Retirement for Cause Criteria for SSTG Casings, PartII: FractureMechanicsAnalysis, Westinghouse Research Report 84-8D7-LARGE-P2, 1984.

12. G. T. Yahr et al., "Case Study of the Propagation of a Small Flaw Under PWR Loading Conditionsand Comparison With the ASME Code Design Life: Comparison of ASME Code Sections III andXI," American Society of Mechanical Engineers 1986 Pressure Vessels and Piping Conference andExhibition, Chicago, Illinois, July 20-24, 1986.

13. C. Sundararajan, "Probabilistic Assessment of Pressure Vessel and Piping Reliability," Journal ofPressure Vessel Technology, 108, 1986, pp. 1-13.

14. Predictive Capabilities in Environmentally Assisted Cracking, Proceedings of the WinterAnnual Meet-ing of the American Society of Mechanical Engineers, Miami Beach, Florida, November 17-22, 1985,PVP-Volume 99.

142

I'

15. NV. E. Berry, "Stress-Corrosion Cracking in Light-Water Reactors," Thermal and EnvironmentalEffects in Fatigue: Research-Design Interface, PVP-Volume 71, ASME, New York, 1983, pp. 217-230 .

16. B. M. Gordon and G. M. Gordon, "Materials Aspects of BWR Plant Life Extension," Proceedingsof the Thirteenth W'ater Reactor Safely Research Information Meeting, NUREG/CP-0072, Volume 3,1985, pp. 293- 317.

17. Thermal and Environmental Effects in Fatigue: Research-Design Interface, PVP-Volume 71, ASME,New York, 1983.

18. Proceedings of the Second International Atomic Energy Agency Specialists' Meeting on SubcriticalCrack Growth, Volumes I and2, NUREG/CP-0067, April 1986.

19. NV. H. Bamford, "Technical Basis for Revised Reference Crack Growth Rate Curves for Pressure BoundarySteels in LWR Environment," Journal of Pressure Vessel Technology, 102, 1980, pp. 433-442. ..

20. R. Rungta, H. Mindlin, and J. D. Gilman, "Applications of EPRI Database on Environmentally -Assisted Cracking in the Nuclear Industry," Paper No. 42, Corrosion '86, National Association ofCorrosion Engineers, Houston, March 17, 1986.

21. J. D. Gilman, "Application of a Model for Predicting Corrosion Fatigue Crack Growth in ReactorPressure Vessei Steels in LWR Environimen ts," Predictive Capabilities in Environmentally AssistedCracking, Proceedings of the Winter Annual Meeting of the American Society of Mechanical Engi-neers, MiamiBeach, Florida, November17-22, 1985, PVP-Volume99, pp. 1-16.

22. J. D. Gilman, "Application of a Model for Predicting Corrosion-Assisted Fatigue Crack Growth inLWR Environments," in Proceedings of the Second internationalAtomic Energy Agency Specialists'Meeting on Subcritical Crack Growth, Volume?2, NUREG/CP-0067, April 1986, pp. 365-384.

.43

.. . ..

' ' ''!, ' , .':;i 143

____ __ __J

I14. NEW OR EMERGING METHODS FOR

INSPECTION AND LIFE ASSESSMENTC. E. Jaske and L. J. House

As a direct result of the recent increased interestin extending the operating life of all types of majorindustrial plant equipment, including nuclearpower-plant equipment, new methods for inspec-tion and life assessment of that equipment arebeing developed. The inspection methods arerequired to evaluate the current damage state of thematerials in such equipment without introducing asignificant amount of additional damage from themeasurement technique that is used. Currentremaining life-assessment methods, such as thoseof Section XI of the American Society of Mechani-cal Engineers (ASME) Code, are based on inspec-tion to find cracks or crack-like defecis' and using afracture-mechanics approach.

14.1 Nondestructive ExaminationTechniques

The principal thrust of current research and develop-ment in nondestructive examination (NDE) is toextend the state of the art from qualitative flaw detec-tion and estimation of the size of large continuousflaws to quantitative measurements of performance-related material properties and more accurate quantita-tive flaw characterization. Emerging NDEtechnologies consist essentially of variations andimprovements in existing measurement or NDE meth-ods rather than new sensor technologies.

The fundamental basis of all NDE is predominantlythe measurement and analysis of either electromag-netic or mechanical radiation. Of the emerging tech-nologies, ultrasonics continues to be the dominantsource of new inspection techniques and advances inquantitative inspection technology.

The development of automated and expert sys-tems for ultrasonic, eddy-current, and radio-graphic inspection also represents a majorcontribution to advanced NDE. Improvements indigital data acquisition of eddy-current and ultra-sonic inspection data are now enabling real-timeanalysis and archiving of inspection results andpermitting the application of adequately sophisti-cated models to provide more meaningful andquantitative interpretation of the inspection data.Advances in imaging systems also offer promise forgreater realization of the potential of radiographicexamination. Improvements in digital radiographic

image processing include qualitative imageenhancement and quantitative image analysis. Thedevelopment of video systems with improved reso-lution presents the possibility of practical real-timeradiography offering the advantages of immediatepresentation of a flaw image and potentiallyenhanced flaw detection probability because manyangles can be examined. If combined with digitalprocessing, real-time radiography can provide apowerful quantitative inspection tool.

A shift from the development of periodic inspectionmethods to in situ monitoring techniques offers someadvantages. Because monitoring techniques rely ondetecting changes in material parameters rather thanmaking absolute measurements, a number of calibra-tion problems that limit the usefulness and reliability ofperiodic NDE techniques based on making absolutemeasurements are avoided. With in situ monitoring,the possibility arises for detecting microstructural ormetallurgical changes in a material, measuring changesin component stresses, and measuring internal temper-atutre. Of the techniques that provide large area moni-toring capabilities, acoustic emission is one of the morepromising emerging technologies. Advances in analyti-cal models for more quantitative analysis of acousticemission events, the development of small aperturearrays, and greater sophistication in the hardware fordata acquisition offer considerable promise for acous-tic emission to become a useful tool to monitor struc-tural integrity.

Table 14.1 shows Broek's I summary of the sevenbasic nondestructive inspection techniques thatwere discussed in Chapter II of this report and areused in conjunction with the fracture-mechanicsapproach. The basic physical principles and thetypical application are summarized for each tech-nique. The major challenge for quantitative use ofthese techniques in a fracture-mechanics context isto accurately determine the size, shape, location,orientation, and type of crack or flaw. Further-more, it would be useful to extend these techniquesto other types of material aging or damage.

As is illustrated in Figure 14.1, Masuyama 2 summa-rizes methods that could be used for the NDE of mate-rials that have been in service for some period of time,pointing out that both surface and internal cracks areof concern. Most'of the traditional NDE methods canbe used to detect surface cracks. However, only

144

I

Table 14.1. Inspection methods (from Broek) I J.,

Method Principles

Visual

Penetrant

Naked eye, assisted by magnifying glass, lowpower microscope lamps, mirrors.

Colored liquid (penetrant) is brushed onmaterial and allowed to penetrate into cracks.Penetrate is washed off and quickly drying -suspension of chalk is applied (developer)..Remnants of penetrant in crack are extracted bydeveloper and give colored line.

Part to be inspected is covered with a layer of afluorescent liquid containing iron powder. Partis placed in strong magnetic field and observedunder ultraviolet light. At cracks, the magneticfield lines are disturbed.

Application

Only at places easily accessible.Detection of small cracksrequires much experience.

Only at places accessible.Sensitivity of same order asof visual inspection.

Magnetic particles Only applicable to magneticmaterials. Parts have to bedismounted and inspected in'special cabin. Also notches andother irregularities giveindications.-Sensitive method.

X-ray X-rays emitted by (portable) x-ray tube passthrough structure and are caught on film.Cracks, absorbing less x-rays than surroundingmaterials, are delineated by black'line on film.

Ultrasonic

Eddy current

Acoustic emission

Probe (piezo-electric crystal) transmits high-frequency wave into material. The wave isreflected at the ends and also at a crack. Theinput-pulse and the reflections are displayedon an oscilloscope. Distance between firstpulse and reflection indicates position of crack.Interpretation: Reflections of cracks disappearon change of direction of wave.

Coil induces eddy current in the metal. In turn,this induces a current in the coil. Under thepresence of a crack the induction changes; thecurrent in the coil is a measure for thesurface condition.

Method with great versatilityand sensitivity. Interpretationproblems if cracks occur infillets or at the edge'of -reinforcements. Small surfaceflaws in thick'plates difficult todetect.

Universal method because avariety of probes and inputpulses can be selected.Information about the size andthe nature of the defect (whichneed not be a crack) aredifficult to obtain.'

Cheap method (no expensiveequipment) and easy to apply.Coils can be made smallenough to fit into holes..Sensitive method when appliedby skilled personnel. Little orno information about natureand size of defect.

Inspection while structure isunder load. Continuoussurveillance is possible."Expensive equipment required.Interpretation of signals isdifficult.

I" I' I I .

Measurement of the intensity of stress waves are: emitted inside the material as a result of plastic' deformation at crack-tip-, and as a result of

* crack growth.

I

145

1

I

Damage mode Nondestructive detecting method

- For surface cracks_ _ Cracking * Penetrant test

( Surface cracks 0 Magnetic particle testInternal cracks&* Ultrasonic test® Internal cracks * Electric resistance method

- For internal cracks* Radiographic test

_ Jo * * Ultrasonic test

* Replication method\ } Micro cracks * Physical test (not established)I UltrasonicElectricMagneticRadiographic,

* Replication methodMicro voids * Physical test (not established)

UltrasonicElectricMagneticRadiographic

0 Replication methodStructural * Extraction method

\* ....... ^- .8;-. degradation * Physical test

( Hardness* "* X-raydiffraction

< P * : .. - nearlite d* Ele ctro-chemical tost

Virgin' ExposedFerritic steel

\/ 5 i *tN 'Examples of precipitates analysis

a7 phase Fe Fe Cr Exposed. I S (precipitationJ MVg

Virgin Exposed Carbide a phase IntensityAustentic steel depends on

exposed time

Ii

I

I

I

I7-3078

Figure 14.1. Methods for nondestructive examination of materials after service exposure.

146

U

I

volumetric methods, such as radiographic andultrasonic testing, can be used to'detect internalcracks. Metallographic replication can be used todetect microstructural damage, such as-crackingalong grain boundaries or voids, at accessible sur-faces, but physically based NDE methods for-thedetection of such microstructural daimiage have notbeen developed for general use. "Time-dependentchanges in microstructural features, such as the fer-rite phase in austenitic-ferritic stainless steels andprecipitate phases in stainless steels, can be detectedby replication, extraction, hardness testing, x-raydiffraction, and electrochemical testing at accessi- --

ble locations.Several general review articles on the current sta- !,

tus and future prognosis of NDE technology havebeen published recently.3 -5 Posakony3 pointed outthat the future thrusts in NDE development include(a) defect characterization by means of advancedsystems, (b) continuous monitoring of equipmentby means of acoustic emission and vibration analy-ses, and (c) assessment of materials properties andmonitoring of material condition by various NDEtechniques for use in life extension efforts.McClung4 indicated that improvements in NDEare coming from (a) the increased use of computersfor acquisition, processing, and interpretation of -

large quantities of data in the field, while theinspection is taking place, (b) the development ofvisual display systems that provide enhancedimages for examination and interpretation, (c) abetter understanding of the physical phenomenaemployed in NDE, and (d) the development of newNDE techniques. Bernard5 described the basicprinciples of time-of-flight-ultrasonics, and dis-cussed its use in inspecting a pressure vessel withsteel walls that were almost one foot thick.

Muscara 6 has reviewed the current status and futureplans for the USNRC research program on NDE. Thatreport describes projects that are developing arid e&lu-''ating methods both for in-service irispection (ISI) ofkey reactor components usinig ultrasonics'and eddy-"currents and for continuous monitoring of key reactorcomponents using acoustic emissions. Each of the fool-lowing subsections'discus'ses in more detail the emerg'-ing NDE methods that could be applied to key reactorcomponents.' Relevant methods being develojed bothby the nuclear power iridustry and by other industries,'such as fossil po&er, petrochemnical,4 and aerospace, areincluded.

14.1.1 Advanced Processing of Data. Ultra-sonic inspection is used for the reactor pressure ves-sel (RPV), piping, and nozzles. The synthetic

aperture focuiing'technique (SAFT) for ultrasonictesting (UT)' has been developed to provideenhanced visual images of flaws detected in com-ponents during inspection.6 The authors of Refer-ence 7 reviewed their field experience with theapplication of SAFT-UT to the inspection of weldsin the recirculation piping system at the Dresden-3and the Vermont Yankee boiling water

*reactors (BWRs). The authors' initial results areencouraging, but they indicated that destructiveexamination of samples from the sites of the flawindications are required to fully qualify the resultsof this new technique..

Figure 14.2 (from Reference 7) shows a concep-tual diagram of a SAFT field system. It is evidentthat the real-time processing involved requires amajor commitment of computer resources. How-ever, with the rapid advances being made in digitalcomputer technology and parallel processing, suchresources will become readily available at reason-able costs within the next few years.

Bemard5 described a case history where SAFT-UTwas successfully used to inspect a refinery's internallystsinless-steel-clad, ferritic-steel pressure vessel. Thevessel was inspected through the thickness of its'-12-in. wall. SAFT was used to distinguish betweenindications from reflectors and flaws.

The authors in Reference 8 reviewed emergingcomputer-aided UT inspection systems for theautomated -examination of nuclear piping. Thatreview provides a'detailed summary of the hard-ware and software'that is used in six of those pipe-inspection systems. The systems can be controlledautomatically and remotely, and the hardware isassembled in reasonably small packages. An adapt-able track-mounted, automated pipe scanner isused with'four of ihe systems. All of the systemsprovide fairly extensive data recording. The com-puterized iultrasibihic' data acquisition arid process-ing system and the ultrasohic'data' recording' andprocessing system also provide data'analysis capa-bilities. As these systems'evolve,'it is expected thatexpert software systems will be incorporated intothe data analysis to help the inspector in real-timeinterpretation of the UT information'.

The authors in'Reference 9'developed a proto-type syktern for 'detecting surface cracks' inp'ipingusing ' the' dc-poienitial-drop m'ethod and haveapplied it to detecting sirface cracks oni the insideof a 12-in.'diameter austehitic' stainless' steel pipe.Their equipment includes scannerslfor both theinside and outside surfaces and a personal com-puter. They performed a large number of solutionsfor the electric potential distributions around

147

I

Pre-amplifierand mechanical

Operatorconsole

Figure 14.2. Conceptual diagram of SAFT field system.

various sizes and shapes of cracks using the finite-element method. During an inspection, they mea-sured a distribution and compared it with theirlibrary of solutions to estimate the crack configura-tion using the personal computer..

Pressurized water reactor (PWR) steam genera-tor tubes are normally inspected by. means ofeddy-current testing. 6 The state-of-the-art eddy-current techniques are being evaluated and statis-tically based reliability data are being developed inan extensive evaluation of a steam generator thathas been retired from service. 6 To overcome prob-lems associated with detecting circumferentialcracking and intergranular attack, a 16-elementpancake-coil eddy-current probe has been con-

structed and is being evaluated in conjunctionwith that work.

Whaley and Flora1 0 discussed the application ofmicrocomputers to eddy-current testing. They havedeveloped a computerized system for detectingflaws in steam generator tubes near the supportplates. They also have performed a preliminaryevaluation of the use of a deep-penetration, eddy-current method for detecting flaws in the welds ofaustenitic stainless steel pipe, with encouraging ini-tial results. However, the computer-aided tech-niques for eddy-current testing are not as welldeveloped as those for UT.

Thomel I discussed the application of acoustic(or ultrasonic) holography to defect imaging. To

rI

148

I

I

date, this technique has been'liiniate to laboratoryevaluation, where it provides excellent images ofdefects. However, 'significant development ofangle-beam methods is required before this methodcan be employed to inspect components.

14.1.2 On-Line Monitoring. The 'USNRCresearch plan6 includes two major programs thatdeal with on-line monitoring of reactor'compo-nents using acoustic emission (AE) technology.One program is aimed at using AE to monitor crackgrowth in pressure boundaries, 12 and the otherprogram is aimed at using AE for leak surveillancein light water reactor (LWR) systems. 13

Methods for AE monitoring of crack growthhave been developed in laboratory studies, evalu-ated during a long-term fatigue test of a pressurevessel, and demonstrated by field surveillance onthe Watts Bar-I reactor.12 That work has shownthat AE sensors and 'instrumentation can' bedeployed successfully in a reactor environment.Fatigue-crack growth in pressure'vessels and, insome cases, stress-corrosion crack growth in stain-:less steel piping can be detected by AE monitoring,even with interference from background reactorflow noise. Thus, AE monitoring appears to be apromising technique for on-line monitoring of crit-ical areas in key reactor components.

Kupperman' 3 reviewed the current leak detec-tion practices for the'cooling systems at 74 nuclearplants. He concludes that the current methods maynot be sensitive and reliable enough for detectingsome types of possible leaks, especially with regardto the Duane Arnold safe end cracking incident.Work is underway to improve the existing leak-detection systems and to develop two new types ofsystems-moisture-sensitive tape and AE moni-tors. Moisture-sensitive tape has been installed in afew reactors for evaluation; it has a tendency to givefalse alarms and, does not provide quantitativeinformation about the leak rate. AE leak detectionis being evaluated both through laboratory testingand field demonstration. The initial results' areencouraging, but more software development and,field validation of the technique are needed beforeAE leak detection can be used with confidence.,

Methods for on-line monitoring of pressureboundary components also are being developed forfossil plants. Davidson' 4 describes a prototype'sys-tem for on-line monitoring of boilers to record andaccumulate data during typical operating condi-tions, during startup, shutdown, and duringoperating transients. That system is being evaluatedat Consolidated Edison's Ravenswood-30 Unit.

Similar systerfs'coiild be installed in a nuclear plantto obtain important and valuable information onkey parameters during typical operating histories.

Fryl51 6 discusses the use of neutron-noise anal-ysis for monitoring the conditions in a reactor core.Ex-core neutron noise 'measurements have beenmade at 13 LWRs, and a library of data for futurereference has been 'established. This 'approachshows promise for monitoring vibrations in PWRinternals and possibly inadequate core cooling inPWRs.' :Neutron noise could -be used to detectabnormal vibration of instrument tubes and fuelboxes in BWRs, but other vibrations in BWRs weredifficult to detect because of the large backgroundnoise from boiling. This technique is not yet appro-priate for on-line use, but work is continuing todevelop it further.

.GalpinI7 has outlined'a'practical, economicapproach to either manual or automatic monitor-ing of component performance to develop infor-mation on aging trends. He emphasizes thatstandard procedures should be employed to moni-tor parameters, such as temperature, flow, pres-sure, vibration, and acoustic emission.' Repeatable,accurate, and reliable data are needed to documentaging trends and use those trends to establish safeoperating limits.

14.1.3 Nondestructive Measurement of Materi-als Properties.; Most of the current and emergingNDE techniques are limited to the detection andcharacterization of flaws and cracks that are in theform of mechanical or elastic discontinuities incomponents. Recently there has been an increasedinterest in the nondestructive measurement ofmaterials properties (NDMMP), such as tensilestrength, fracture toughness,' impact toughness,'creep strength, grain size, and precipitate morphol-ogy, 'that 'heretofore have traditionally been mea-sured by destructive 'testing and examination ofspecimens. 'As illustrated by the papers in a recentvolume,1 8 the techniques being evaluated forNDMMP include ultrasonic testing, eddy-currenttesting, and x-ray diffraction analysis. Most of theNDMMP work performed to date has been'restricted to laboratory studies. Field applications'will depend upon future development of apparatusand techniques. ' i ' - --

Generaziol 9 investigated the use"of ultrasonicattenuation for measuring the mean grain size ofcopper and nickel. Based on the results of his study,he has suggested a nondestructive method forchecking the heat treatment of metals. In caseswhere aging causes microstructural changes to

I

I

I

149 '

I

I

occur during service at LWR operating tempera-tures, this type of technique might be used to mea-sure those changes. For example, in studies ofgas-turbine alloys, it was. found that ultrasonicattenuation could be used to detect and quantifymicrovoid creep damage.a It may be possible todevelop similar correlations for radiation-induceddamage in key reactor component materials.

Swanson 20 and Vary21 studied the use of ultra-sonic attenuation measurements for ranking mate-rial fracture toughness. They have foundcorrelations between ultrasonic measurements andthe plane strain fracture toughness of variousgrades of steel and certain titanium alloys. How-ever, the measurements made by the ultrasonicmethods used in these studies required sampleswith smooth plane parallel surfaces and a veryaccurate measurement. of the sample thickness.Other methods, such as those employing backscat-tering,2 2 ' 23 are being developed to eliminate thisrestriction, but they require spatial- or directional-averaging techniques to eliminate the effects on thedata produced by the coherent nature of grain scat-tering. The principal drawback of this approach isthe complex scanning requirements necessary forspatial and directional averaging. Digital process-ing methods based on frequency averaging mayeventually provide an alternative to the complexscanning requirements of spatial or directionalaveraging necessary for eliminating the coherent-grain scattering effects.

Researchers 4 have successfully made ultrasonicmeasurements of stress intensity factors caused bysurface cracks, but the study was limited to brittlenonmetallic materials. The method was based onthe measurement of the reflection coefficient of aRayleigh wave incident to the crack. Clark2 5

extended the use of ultrasonic phase velocity mea-surements to the measurement of stress and stressintensity factors around crack tips in aluminum. Inthat work, the ultrasonic method used the birefrin-gence of a horizontally polarized shear, wave thatoccurs in a slightly orthotropic plate in a state ofplane stress.

Bussiere26 reviewed the techniques used forNDMMP, with emphasis on electromagnetic andultrasonic methods for on-line measurement of theproperties of steel. He indicated that eight tech-niques are useful for NDMMP:

a. Unpublished results of studies by C. M. Jackson at BattelleColumbus, Columbus, Ohio, 1982.

* Magnetic, including Barkhausen noise-for hardness,' grain size, precipitate size,morphology, and tensile strength

* Ultrasonic-for formability, grain size,tensile strength, and fracture toughness

* X-rays-for residual stress and stressrelaxation

* Resistivity/conductivity-for chemicalcomposition, phase transformations, andstate of cold work

* Thermal and thermal imaging-for grainstructure and extent of hardening

* Small-angle neutron scattering-for particle,precipitate, and microvoid characterization

* Mossbauer spectroscopy-for microstruc-ture and phase characterization

* Position annihilation-for damage causedby plastic deformation, fatigue, and creep.

None of these techniques have been widely appliedto characterize material-aging processes for LWRcomponents, but with further development and theestablishment of reference standards some of themmay prove useful for that purpose in the future. Forexample, it may be possible to use magnetic tech-niques to measure the volume fraction of ferriteand mean ferrite spacing (Chopra and Chung27

have shown that both of these parameters areimportant) in stainless steel castings to help assesstheir potential for embrittlement by thermal agingduring lons-term service.

Yavelak2 8 used Barkhausen noise analysis(BNA) to characterize the service-induced damagein a secondary superheater outlet header that wasremoved from a 240-MW boiler after 24 years ofservice. The results indicate that this method haspotential for application as a tool to nondestruc-tively measure damage. BNA has been used. toinspect and locate damaged regions in headers andsteam lines in the field during outages.29

Eddy-current methods for measuring materialproperties are being developed by Clark andMetala.3 0 They have used eddy-current techniquesto measure applied and residual stress, tensilestrength, the amount of temper embrittlement, andthe degree of creep damage. Figure 14.3 shows theirmeasured correlations between the yield strengthand eddy-current response for a low-alloy steelheat-treated to different strength levels. They pointout that a sufficient catalog of standard back-ground data needs to be developed before the eddy-current method can be employed as tool forresidual-life assessment.

e

I

150

I

I

8

6

5

4

8

7

Test frequency2.0-i n. (5.08 cm a. kHzprobe 1. ~Ground surfaceksi = 6.9 MPa

30 100 120 140 160 180 201

5.0 kHz

30 100 120 140 160 180 20

0

IU,C0

CLU,

C.ID

~0w1

6

5

4

8

0

I *I l l I1.0 kHz7 _- 1.

6

5

4 L.80

I I I I I

100 120 140 160 180 08

7

6

5

484

-

I I i I I

I0.5 kHz

_ I - .

. . I;

I

5 100 * 120. 140 160Yield strenath (ksi)

. 180 - 200

- . .. ''' 7-3056

Figure 14.3. Eddy-current response versus yield strength (2-in.-diameter probe).1.. . . p ., .

Electrochemical methods can be used to nofides-'tructiely measure material properties: Because ofthe 'problem of intergranular stress-corrosioncracking (IGSCC) in BWR piping,- an extensiveeffort was directed-toward developing a nondesI-tructive'electrochemical testing procedure for mnea-suring the susceptibility of austenitic stainless steelsto IGSCC. This work resulted in the developmentof the electrochemical potentiokinetic reactivation(EPR) test, which has been described and evaluatedby Majidi and Streicher.31 In most cases, the EPR

test gives good, quantitative measures of the degreeof sensitization with a variety of operators andinstruments, and portable equipment is availablefor performing it in the field. A double-loop EPRtest procedure has been developed to simplify themaking of field measurements.3 2

Shoji and Takahashi33 used an electrochemicalmethod similar to the EPR test to evaluate the shiftin Charpy transition temperature and the creepdamage in steam-turbine rotor and casing steelsthat result from material degradation during

I

151

r

I

long-term service. They found a correlationbetween the measured electrochemical parametersand the shift in Charpy transition temperaturecaused by temper embrittlement and suggested thattheir method provides a userul tool for nondestruc-tively measuring the degree of temper embrittle-ment. It may be possible that a similar approachcould be employed to nondestructively evaluate thedegree of radiation embrittlement in RPV steels.

In heavy-section components, indentation hard-ness measurements may provide a nondestructiveevaluation of the degree of material aging. It maybe possible to correlate hardness readings with thedegree of radiation embrittlement in RPV steels34

and then make hardness measurements to assessradiation embrittlement during extended service.As an example, Cane35 outlined a procedure forusing hardness measurements to assess the remain-ing life of thick-section steel components in fossilpower plants.

14.1.4 Surface Replication. Making celluloseacetate replicas of sample surfaces for detailedexamination at magnifications up to about1O,OOOX using electron microscopy has been a typi-cal laboratory procedure for many years. Recently,there has been a great deal of interest in using repli-cation as a tool for field NDE2 ,36-43 of compo-nents as part of remaining life-assessmentactivities. That work has concentrated on headersand steam lines in fossil plants rather than onnuclear plant components, but the same techniquescould be adapted for use in LWR applications.

Figure 14.4 shows an illustration of the replica-tion and extraction methods from the paper ofMasuyama. 2 The surface to be examined must beaccessible. The area to be replicated is first cleaned,metallurgically polished, etched, and moistenedwith an organic solvent. The cellulose-acetate tape(plastic film) is then applied and allowed to sit inplace for a time while it softens and adapts to thesurface. The replica is then removed to provide anegative image of the surface. For extraction, theetching is deeper so the particles in the microstruc-ture can be removed for examination. To producehigh-resolution images, the replicas are coated withcarbon or gold in the laboratory and examinedusing optical or electron microscopy.

GPU Nuclear Corporation has developed a sim-plified replication technique where mechanical pol-ishing time is minimized by the use of a finalelectropolishing step and the replica is shadowedwith a felt-tip marking pen while it is 'drying.4 2

This approach permits replicas to be made morerapidly (in \15 min) at the sacrifice of high resolu-

tion. These replicas are examined optically at lowmagnifications using a portable microscope toobtain preliminary information in the field and atmagnifications up to I,OOOX in the laboratory. Apotential danger of such a high-speed technique(electropolishing) is that it may produce artifactsthat might be incorrectly interpreted as damage.

For many nuclear-plant applications, a remotereplication apparatus would be required. A proto-type remote replication unit has been developed forthe metallographic examination of steam-turbinerotor bores,43 and a more advanced rotor-boreremote replication system is planned for develop-ment as part of a future Electric Power ResearchInstitute project.a Also, a replication manipulatorwas developed for the remote examination of steamgenerator tube bores.44 That equipment.was suc-cessfully used to inspect regions down to 6 feet inthe 0.57-in.-inner-diameter steam generatortubes.45 In another study, replication was used todetect and document creep-fatigue cracking nearwelds in a test pressure vessel.4 The feasibility ofusing remote replication apparatus has been dem-onstrated, but specific equipment needs to be devel-oped for use in LWR applications.

Replication may be a valuable inspection tool forareas near welds in LWR vessels, piping systems,and internals. It also may prove useful for inspec-tion of steam generator tubes. It can be used todetect very small cracks or to assess changes inmicrostructure because of aging. Replicas of caststainless steels could document the amount, distri-bution, and spacing of ferrite and might indicatechanges in the microstructure related to long-termembrittlement of the ferrite phase. For these andother possible applications, development and dem-onstration work will be required before replicationcan be used as a routine ISI tool.

14.2 Life-Assessment Methods

As reviewed previously, the current life-assessment techniques are both experimental andanalytical. The newer or emerging methods for lifeassessment employ, miniature test samples that canbe removed from a component with negligible dam-age to the component, on-line procedures for thecalculation of damage, and improved models ofmaterial degradation and damage accumulation,

a. The project. Steam Turbine Rotor Life Assessment andExtension, EPRI RP2481-5, is scheduled to begin in 1987.

I

I

I

I

152

I

I

" 1.

Carbide Crack

\ / Surface of partafter polishIng,77~k7~Iz and etching

Plastic film

| Softened filmadapts

-to surface

Carbide Crack

< Replica withnegative of

Rep n microstroctur

Replication method

Electron microscopeand analyzer

Precipitates

Surface of partalter deep

Aetgtchlng

Plastic film~~~ ~~ Softened filmadaptsto surface

Precipitates

- I I. I~ Replicawith

precipitates

Extraction method7Tao

I

Figure 14.4. Illustration of replication and extraction methods.

especially with the inclusion of probabilisticparameters that are needed to assess 'safety on arealistic basis. Each of these topics is'addressed in'the remainder of this section of the report. -

14.2.1 Miniature Specimen Testing. Testing ofspecimens removed from an actual component dur-ing service, as mentioned earlier, can provide adirect measure of the degree of aging. If miniature'samples are used, then the amount of material thatmust be removed from the component will be smallso that little or no repairs will be required. Minia-'ture specimens also can be valuable for surveillancetesting where only a limited amount of test materialis available and where space available for materialirradiation is restricted. For example, it may be pbs--sible to make miniature specimens from the- ;'untested grip portions of standard size specimensto obtain additional data. Studies are in progress toevaluate the applicability of- miniature specimen'technology (MST) to residual-life evaluation.

As discussed by Jaske,4 7 there are two main con-:straints to the application of MS1T

The specimen to be tested must be largeenough that it represents the material fromwhich it has been removed

* 'The specimen to be tested must be smallenough that its removal from the compo-nent being evaluated does not impair thestructural integrity of that component.

The first requirement (minimum size) expresses theneed for the miniature specimen to exhibit proper-ties that are the same as those of the bulk material.Experience has shown that the requirement can besatisfied when the minimum specimen dimension isfive to ten times as large as the largest strength-controlling microstructural feature. For steels usedin LWR components, that feature usually, is thegrain size, which rarely is greater than 0.1 mm. Theminimum-size criterion is applicable, however, onlyto yield strength, and maybe to ultimate strength.

A shear-punch test has shown promising results formeasuring stress-strain response. 48 That test uses smalldisks of the same size as those used for transmissionelectron microscopy (-3.0 by 0.25 mm). During test-ing, the load versus deflection response of ihe disk ismeasured. Finite element stress analysis then is used todevelop stress-strain curves. For Type 316 stainlesssteel, -this approach gave results comparable to thoseobtained from conventional tensile tests. Similarlypromising results have been obtained using hardnesstests.4 9

l

!

I

153

I

I

Fatigue cracking and fracture toughness, veryimportant properties in remaining life evaluationand reliability analyses, are strongly influenced bygrosser microstructural features, such as inclusionsand banding.47 Both of these properties also arenotch sensitive, so predictions based on data fromunnotched specimens can be extremely misleading.It also is well known that fracture toughness is sizedependent; the problem is that small specimensprovide nonconservative data compared with datafrom large specimens. For these reasons, the car-bon segregation in specimens from an AmericanSociety for Testing and Materials (ASTM) A 508pressure vessel steel limited the minimum specimensize to 5 mm for obtaining successful measure-ments of plane-strain fracture toughness, JIc,where the criterion for success was comparison ofthe results with those obtained from larger speci-mens that satisfied the criteria of ASTM E 813.44

Recently, Misawa and Hamaguchi50 evaluatedtwo types of miniature specimens for measuring thefracture resistance of ferritic steels. They used a24-mm-diameter, disk-shaped, compact-type speci-men with thicknesses from 1.8 to 5.0 mm to mea-sure the tearing modulus. Also, they usedsmall-punch specimens (0.25xlOx1O mm) takenfrom the ends of broken Charpy specimens to mea-sure the ductile-to-brittle transition temperatureand obtained a linear correlation with transitiontemperature measured using Charpy specimens.

The fracture-toughness data obtained using MSTshow promise. However, one must be careful inselecting a specimen size for the measurement offracture toughness and fatigue cracking. Severaltask groups of the ASTM currently are working onguidelines for the application of MST to fracture-toughness testing.

Battelle's Columbus Division is currently under-taking a major in-house study of MST. 51 Theobjective of that laboratory study is to evaluate theranges of parameters over which miniature speci-mens can be successfully employed. Applicationsbeing evaluated include characterization of defor-mation' behavior, fracture toughness, and crack-growth resistance.

Miniature specimens also are being evaluated for usein assessing the remaining life of fossil-plant compo-nents. Askins 52 described the development and valida-tion of a miniature-specimen testing technique forevaluating the tensile creep and stress-rupture proper-ties of steels. As is illustrated in Figure 14.5, their speci-men design employs a 3-mm-diameter by 13-mm-longpiece taken from a component in service. Extensionpieces are electron-beam welded to each end of that

piece to provide material for gripping in the test fix-tures. The final specimen has a configuration with a2-mm-diameter by 10-mm-long gauge section. Foraccelerated testing at temperatures near 10000F(600'C), an argon gas environment is used to avoiderroneous results from excessive oxidation of the testspecimen. For a lower temperature of 10000F (5380C),Saxena 53 tested similarly sized creep specimens from aretired l-l/4Cr-I/2Mo steel header in air and obtainedgood agreement with comparable data from tests ofstandard-size specimens, as is shown in Figures 14.6and 14.7. Thus, miniature tensile creep testing has beendemonstrated to be a viable tool for residual-life assess-ment.

Although validation work would be required, itis expected that the same type of miniature tensilespecimen configuration used for creep testingwould provide valid results for the measurement oftensile strength and deformation response of sam-ples removed from reactor components. For exam-ple, a study by VanEcho and Williamsa showedthat miniature tensile specimens are useful for mea-suring the through-thickness properties of thin('.12 mm or less) steel plate. Miniature specimens(including the disk configuration discussed earlier)would be particularly useful for characterizing theproperties near welds because samples could betaken from the weld metal and heat-affected zonein cases where standard-size specimens would betoo large to sample the properties in the desiredregion.

142.2 Reconstituted Specimen Testing. Anotherprocedure developed to fully use the limited amount ofirradiated material that usually is available for evalua-tion is to reconstitute a Charpy V-notch impact speci-men from one half of a previously testedspecimen.54 ' 55 Figure 14.8 illustrates the sequence ofsteps involved in reconstitution. 54 The fracture face iscut off one half of the tested specimen, and two endtabs are stud welded to it. Final machining and notch-ing of the specimen then is completed. Properly recon-stituted specimens yield results that agree with thoseobtained from the original specimens, as is shown for atypical set of data in Figure 14.9.54 Thus, the use ofreconstituted specimens can triple the amount ofCharpy data obtained from the same volume of mate-rial when only original specimens are used.

a. Unpublished research results, Battelle Columbus, Colum-bus, Ohio, July 1981.

154

I

I

;

. Ex-servicesample: '

EIBweld

II

* J~-13mm -71 N Extension piecefor grip

;Collarto providerun-in material for thisweld to prevent undercut

Figure 14.5(a). Schematic diagram of an electron beam welded blank and test piece for tensile testing.

I

I

2mAm 1 ,/ ~~~~~~~~~~6mm

Parallel gauge 7.3062

Figure 14.5(b). Schematic diagram of a test specimen machined from the assembly, shown in Figure 14.5(a)._~~~~~ ! - -

I

I

155

l

I

MPa40 60 80 10010 2010-2

10 i- 4

0.

S!0)C.,0)

0)

1 2 4 6 8 10 20Applied stress, a (Ksi) 7.3198

IFigure 14.6. Minimum creep rate as a function of stress for hot- and cold-end materials.

156

I

I

N.t

CO

104

7.3199F r , i b. .

Figure 14.7. "Comparison between creep strain 'versu's te'st time data from standard and miniature specimens.I ,; ,, i ; ,- I , -. , ,,,, ; , . I - I --: . -. . 1 . . . , I , '' I :1 I

157

I

I

Original specimen

As-tested specimen

* Half of brokenspecimen

Aluminum ball

:��' C': )End tabs, machinedhalf of broken specimen,and arc shields(exploded view)

I.

Fli

As-welded specimen

iished specimen

7-3061

Figure 14.8. Sequence of stages in preparing a reconstituted Charpy specimen. I

14.2.3 On-Line Damage Calculation. On-linedamage and remaining life calculations are beingevaluated as a possible method for residual lifeassessment.14 This approach uses a digital com-puter and thermal, stress, creep, and damage anal-ysis software to compute the damage accumulationduring service in conjunction with input data froman on-line temperature monitoring system. All ofthe equipment is installed at the plant and must beable to survive and perform satisfactorily in anindustrial environment. If the calculations are per-formed in full detail, this approach is quitedemanding on computer resources. Thus, in prac-tice, appropriate simplifications must be made tothe analyses. If these analyses are performed prop-erly, then the approach appears to be promisingand can provide useful information about equip-ment operation. The major shortcoming, at

present, is the lack of validated simple analyticalmodels of the damage accumulation processes.

It would be possible to apply a similar on-line dam-age accumulation approach to nuclear systems wherekey operating parameters are being monitored on-line.For example, if local temperatures near a nozzle werebeing monitored during operation, then on-line dam-age analysis could be employed to compute the thermalstresses and strains and the resulting fatigue damageaccumulation during startups, shutdowns, and majoroperating transients. Such a system would requiredevelopment and field demonstration and validationbefore it could be implemented with confidence.

14.2.4 Effective Engineering Models. Success-ful implementation of life extension and life assess-ment strategies requires engineering models thatincorporate reasonable simplifying assumptions,

158

I

V._ _

I

Temperature (CC)

a,ZCw

a)Cw

Original specimenReconstituted specimen

-250 -150 -50 50 - ; 150 ! 250Temperature (°F)

350 450

7-3055

Figure 14.9. Correlation between Charpy data for original and reconstituted specimens.'

but still predict the major, important features ofmaterial damage evolution (aging). In other words,the models must provide a balance between sophis-tication and practical utility so that good assess-ments can be made economically. Wheresimplifying assumptions a-re necessary; they shouldbe made on the conservative side to ensure a highprobability of safe future operation.

Balkey and Furchi 56 outlined an engineeringapproach for the life extension of PWR vessels, asis shown in Figure 14.10. First, onemakes an initialassessment of the RPV using the current guidelines(Regulatory Guide 1.99). If the required remaininglife can be achieved with the. current operatingparameters, then operations can be continued inthe present manner. If not, then fuel managementmay be evaluated using traditionral or-newapproaches. Other modifications to reduce the>neutron flux, such as shielding,-may beconsidered.The appropriate actions are then taken to obtainthe desired remaining safe life. Thistype of assess-'ment is carried out in a probabilistic fashion to help.ensure that safe and reliable decisions are made.

Sidey57 used a model for combined corrosionand stress-rupture damage of fossil-plant super-

heater and reheater tubes to develop a simplifiedengineering procedure for estimating the remaininglife of reheater tubes. When a reheater tube failedafter 78,000 h of service, he used that result to cali-brate the model and develop a simple plot of wallthickness versus remaining life, as is shown in Fig-ure 14.11. Now, when ultrasonic testing is used tomeasure tube-wall thicknesses during an outage,that relationship is used to estimate remaining lifeof other reheater tubes. Using detailed models topredict remaining life trends based on past operat-ing experience provides a valuable,' straight-forward engineering approach.

Probabilistic factors need to be incorporated inresidual-life estimations. Clark58 u'sed probabilis-tic relationships to illustrate the role of NDE in fail-ure risk assessments, as is shown in Figure 14.12.The probability of failure increases as the inspec-tion interval and as the probability of crack initia-tion, Pi, increase. Increased probability of defectdetection,'PF, reduces the probability of failure.This type of approach provides a rational and real-istic method for evaluating and minimizing the riskof failure.

I

159

I

I

Implementationphase

7.3073

Figure 14.10. A cost-effective approach to address reactor vessel life extension.

T4 -

0)

CxY

0)

0.

21

Material - 1.25 Cr 0.5 MoOutside diameter - 54 mmUncorroded wall thickness - 4.18 mm

I I I I

III

III

I

A\

0 20 40 60 80 100 IUseful remaining life (h) x

7.3040

Figure 14.11. Remnaining life graph for TI I reheater tubing after 78,000 hours in service.

160

I

I

1 I I I I I

o1 o-

1 -2 -

a.-

.0

0

2 -

k-

10-5 I

10-6

I10o- 71L.

0I I I

5 10 15 20Inspection interval, years

25

7-3039

Figure 14.12. Probability of failure versus inspection interval for example cases... ..

14.3 Summary

The major emerging methods for nondestructiveinspection and life assessment of key primary sys-tem pressure boundary components are summa-rized in Table :14.2.. For each method, both itscharacteristics and its applications are briefly high-lighted. The applications indicate areas where themethod is used and areas where the method haspotential use if further advancements'and develop-ments are achieved.

Advances -in traditional flaw-detectionapproaches are being achieved through automationand computerized processing of data. With state-* of-the-art systemsand well-trained, knowledgeableoperators, flaws can be located with high probabil-ity and sized with moderately good probability.Clearly, there is a need to develop better methods ofquantifying the size, nature, and distribution of

flaws for use in fracture-mechanics analyses and'development 'work is underway in this area.

Acoustic emission monitoring of crack growthand pressure boundary leaks needs to be furtherdeveloped for routine field application. On-lineneutron-noise and material damage analysis offerlong-range potential, but a great deal of develop-ment work is required before they can be routifielyused with a high degree of confidence.'

The nondestructive measurement of materials prop-erties shows a great deal of promise for use in residual-life'assessmerits, but a large amount of research anddevelopment is needed before the laboratory tech-niques can be made commercially viable. Surface repli-cation and miniature specimen testing also offer somepotential uses. For real-world use, effective engineeringmodels of material-aging processes are essential. Thesemodels must provide a balance between ease of use andcomplex representations of physical phenomena.

I

161 '

I

I

Table 14.2. Summary of emerging methods for inspection and life assessment

Method Characteristics Applications

SAFT-UT Computerized processing ofultrasonic data to provide enhancedsignificant visual images of flaws inreal time.

Welds in the RPV, in piping,and at nozzles. Requirescomputer resources resourcesresources and has not been fullydeveloped and verified for fieldapplications.

V

Computer-aided UT pipeinspection systems

DC potential drop pipeinspection systems

Computerized eddy-current inspectionsystems

Acoustic emissionmonitoring

On-line damage analysis

Automated ultrasonic inspection ofpiping by means of state-of-the-arthardware and software.

Automated inspection andcomputerized analysis of DCpotential distributions around crackson both inner and outer surfaces ofpiping.

Advanced eddy-current coil designscoupled with real-time computeranalysis of data.

Uses computerized processingof data to monitor noise fromcrack growth or water leakage.

Monitors temperature and pressureduring operation and computesdamage accumulation on-line inreal time.

Piping and IGSCC cracks nearwelds in piping. Need morereal-time analysis andinterpretation of data to aidoperators in decision making.

Promising new method forinspection of pipes. Prototypesystem needs to be developed,demonstrated, and verified forroutine commercial applications.

More reliable inspection ofPWR steam generator tubes. Theadvanced methods need to befully evaluated anddemonstrated. Deep penetrationmethods show some potential forpipe inspection, but they needconsiderable development.

Monitoring of fatigue-crackgrowth in RPVs, possiblystress-corrosion crack growthin piping, and water leaks in theprimary pressure boundary.Requires advanced techniques toremove background noise.Results of initial field studies areencouraging, but furtherdemonstration and validation areneeded.

Creep-fatigue damage in fossilboiler headers and fatigueusage factors in nuclearplant nozzles. Only prototypesystems are currently available.Much uncertainty with damageaccumulation models that areused.

fI

162

I1

I

Table 14.2. (continued)

Method Characteristics Applications

Neutron-noise analysis

Nondestructivemeasurement ofmaterials properties

Surface replication

Measures the ex-core neutron noiseand analyzes the recorded data todiagnose potential problems withinthe RPV.

Uses ultrasonic, eddy-current,x-ray, and electrochemical testingmethods to measure properties, suchas tensile strength, fracturedemonstrated, toughness, and impactstrength.

Produces plastic film replica ofmaterial surface that can be . cexamined at high magnificationsusing both optical and electronmicroscopy to detect small cracks,voids, and microstructural changes.

Shows promise for monitoringvibration of PWR internalsand BWR instrument tubesand fuel boxes. Need referencedata for normal operations and'hardware and software foron-line use.

Most applications have beenlimited to laboratory studies.Offers great potential whenmethods are developed,and verified by means of futurebasic and applied research. Needto develop a library ofwell-characterized correlationsbetween nondestructivemeasurements and properties.

Widely used to inspect for creepdamage in fossil-plant headersand steam piping. Has potentialuse for the detection of smallcracks and microstructuralchanges caused by aging in keynuclear components. Could beused to measure ferrite amountand spacing in cast stainlesssteels. Needs to be developed fornuclear-plant applications.

Laboratory studies are beingperformed to define the range ofuse for this technique. Showspromise for the measurement ofdeformation, such as creep andstress-strain response. May haveapplications for measurement offracture toughness and crackgrowth. Further verification ofthe technique is required for fieldapplication.

Trending the aging of keycomponents from a knowledge

, of the operating history. Mostcurrent approaches seem eithertoo detailed and cumbersome touse or too simplistic to providerealistic predictions.

1.

IMiniature specimentesting

Uses a small specimen removed from'a component to measure materialsproperties.

. I

Effective engineeringmodels

Use of simplified engineering modelsthat provide reasonable descriptionsof material aging as a -,

function of operating history.

I

163

i

14.4 References

1. D. Broek, Elementary Engineering Fracture Mechanics, Boston: Martinus Nijhoff Publishers, 1978.

2. F. Masuyama et al.,- "Findings on Creep-Fatigue Damage in Pressure Parts of Long-Term Service-Exposed Thermal Power Plants," Journal of Pressure Vessel Technology, 107, 1985, pp. 260-270.

3. G. J. Posakony, "Future Thrusts in NDE and Measurements," ASTM Standardization News, 14, 7,1986, pp. 30-35.

4. R. WV. McClung, "Improved Applications of NDT," ASTM Standardization News, 14, 7, 1986, pp.52-55.

5. L. Bernard, "Time of Flight Ultrasonics," ASTM Standardization News, 14, 7, 1986, pp. 56-58.

6. J. Muscara, Research Program Plan-Nondestructive Examination, NUREG-I 155, Volume 4, 1985.

7. S. R. Doctor et al., "SAFT-UT Field Experience," Residual-Life Assessment, Nondestructive Exami-nation, and Nuclear Heat Exchanger Materials, Proceedings of The American Society of MechanicalEngineers 1985 Pressure Vessels and Piping Conference, New Orleans, Louisiana, June 23-26, 1985,PVP-Volume 98-1, pp. 203-210.

8. G. J. Dau et al., "Status of Computer-Aided Pipe Inspection," Residual-Life Assessment, Nondes-tructiveExamination, and NuclearHeatExchangerMaterials, Proceedings of TheAmerican Society ofMechanical Engineers 1985 Pressure Vessels and Piping Conference, New Orleans, Louisiana, June 23-26, 1985, PVP- Volume 98-1, pp. 195-201.

9. M. Hayashi et al., "Detection of Configuration of Fatigue Surface Crack in Inside Surface of Stain-less Steel Piping with Potential Drop Method," EPRINDE CenterAdvanced Pipe Inspection SystemWorkshop, Charlotte, North Carolina, May 6-8, 1986.

10. H. L. Whaley and J. H. Flora, "Microcomputer Based Electromagnetic Inspection," Residual-LifeAssessment, Nondestructive Examination, and Nuclear Heat Exchanger Materials, Proceedings of TheAmerican Society of Mechanical Engineers 1985 Pressure Vessels and Piping Conference, New Orleans,Louisiana, June 23-26, 1985, PVP-Volume 98-1, pp. 223-229.

11. D. K. Thome, "Random Scan Ultrasonic Holography-Ultrasonic Imaging Without a PrecisionScanner," Residual-Life Assessment, Nondestructive Examination, and Nuclear Heat ExchangerMaterials, Proceedings of The American Society of Mechanical Engineers 1985 Pressure Vessels andPiping Conference, New Orleans, Louisiana, June 23-26, 1985, PVP-Volume 98-1, pp. 211-215.

12. P. H. Hutton and R. J. Kurtz, "Current Capabilities for Continuous AE Monitoring to Detect Flawsin Reactor Pressure, Boundaries," Residual-Life Assessment, Nondestructive Examination, andNuclear Heat Exchanger Materials, Proceedings of The American Society of Mechanical Engineers1985 Pressure Vessels and Piping Conference, New Orleans, Louisiana, June 23-26, 1985, PVP-Volume98-1, pp. 217-222.

13. D. Kupperman, "Assessment of Leak Detection for Nuclear Reactors," Meeting of the Metal Compo-nent Subcommittee of the Advisory Committee on Reactor Safeguards, Columbus, Ohio, July 1-2,1986.

14. M. J. Davidson et al., "Monitoring for Life Extension," Residual-Life Assessment, NondestructiveExamination, and Nuclear Heat Exchanger Materials, Proceedings of The American Society of

164

. I'~-

I

Mechanical Engineers 1985 Pressure Vessels and Piping Conference, New Orleans, Louisiana, June 23-26, 1985, PVP-Volume 98-1, pp. 65-71.

15. D. N. Fry et al., "Analysis of Neutron-Density Oscillations Resulting From Core Barrel Motion in aPWR Nuclear Power Plant," Annals of Nuclear Energy, 2, 1975, pp. 341-351.

16. D. N. Fry et al., Use of Neutron NoiseforDiagnosis of In- VesselAnomalies in Light-WMater Reactors,NUREG/CR-3303, 1984.

17. D. S. Galpin, "Performance Monitoring for Assessment of Equipment Aging Trends," EPRI Con-ference on Life Extension and Assessment of Fossil Plants, Washington, D. C., June 1986.

18. C. 0. Ruud and R. E. Green, Jr., (eds.), Nondestructive Methods for Material Property Determina-tion, New York: Plenum Press, 1984.

19. E. R. Generazio, "Scaling Attenuation Data Characterizes Changes in Material Microstructure,"Materials Evaluation, 44, 1986, pp. 198-202.

20. G. D. Swanson, Development of Fracture Toughness Measurement Capability by Ultrasound, BDX-613-2625, June 1981.

21. A. Vary, "The Feasibility of Ranking Material Fracture Toughness by Ultrasonic Attenuation Mea-surements," Journal of Testing and Evaluation, 4,14, 1976, pp. 251-256.

22. K. Coebbels and P. Holler, "Quantitative Determination of Grain Size and Detection of Inhomoge-neities in Steel by Ultrasonic Backscattering Measurements,"' Proceedings of the First InternationalSymposium on Ultrasonic Materials Characterization; Gaithersburg, Maryland, June 7-9, 1978, NBSSpecial Publication 596.

23. A. Hecht et al., "Nondestructive Determination of Grain Size in Austenitic Sheet by UltrasonicBackscattering," Materials Evaluation, 39, 1981, pp. 934-938.

24. M. T. Research et al., "The Acoustic Measurement of Stress Intensity Factors," Applied PhysicsLetters, 34, 3, 1979, pp. 182-184.

25. A; V. Clark et al., "Acoustoelastic Measurement of Stress and Stress Intensity Factors Around CrackTips," Ultrasonics, 21, 1983, pp. 57-64.

26. J. F. Bussiere, "On-Line Measurement of the Microstructure and Mechanical Properties of Steel,"Materials Evaluation, 44, 1986, pp. 560-567.

27. 0. K. Chopra and H. M. Chung, "Investigations of the Mechanisms of Thermal Aging of CastStainless Steels," Meeting of the Metal Component Subcommittee of the Advisory Committee on

. Reactor Safeguards, Columbus, Ohio, July 1-2, 1986. -

28. J. J. Yavelak, "Correlation of Diametral Measurements and Cracking with Barkhausen Noise Pat-terns in a Thick-Wall Croloy 1-1/4 (1-1/4Cr-1/2Mo) Header," Residual-Life Assessment, Nondes-tructiveExamination, and NuclearHeatExchangerMaterials, Proceedings of TheAmerican Society ofMechanical Engineers 1985 Pressure Vessels and Piping Conference, New Orleans, Louisiana, June 23-26, 1985, PVP-Volume 98-1, pp. 73-78.

29. T. P. Sherlock et al., "Advanced Techniques for Assessing the Remaining Life of Boiler Compo-nents," EPRI Conference on LifeExtension and Assessment of Fossil Plants, Washington, D.C., June1986.

165

1

30. W. G. Clark and M. J. Metala, Jr., "Eddy Current Determination of Material Properties," EPRIConference on Life Extension and Assessment of Fossil Plants, Washington, D.C., June 1986.

31. A. P. Majidi and M. A. Streicher, "Potentiodynamic Reactivation Method for Detecting Sensitiza-tion in AISI 304 and 304L Stainless Steels," Corrosion, 40, 8, 1984, pp. 393-408.

32. A. P. Majidi and M. A. Streicher, "The Double Loop Reactivation Method for Detecting Sensitiza-tion in AISI 304 Stainless Steels," Corrosion, 40, 11, 1984, pp. 584-593.

33. T. Shoji and H. Takahashi, "Non-Destructive Evaluation of Materials Degradation During ServiceOperation by Means of Electro-Chemical Method," EPRI Conference on Life Extension and Assess-ment of Fossil Plants, Washington, D.C., June 1986.

34. G. E. Lucas et al., "Effects of Composition, Microstructure, and Temperature on Irradiation Hard-ening of Pressure Vessel Steels," Effects of Irradiation on Materials: Twelfth International Sympo-sium, ASTM Special Technical Publication 870, Philadelphia, 1985, pp. 900-930.

35. B. J. Cane et al., "A Mechanistic Approach to Remanant Creep Life Assessment of Low AlloyFerritic Components Based on Hardness Measurements," Journal of Pressure Vessel Technology, 107,1985, pp. 295-300.

36. B. Neubauer and U. Wedel, "Restlife Estimation of Creeping Components by Means of Replicas,"Advances in Life Prediction Methods, New York: The American Society of Mechanical Engineers,1983, pp. 307-313.

37. P. Auerkari, "Remanant Creep Life Estimation of Old Power Plant Steam Piping Systems,"Advances in Life Prediction Methods, New York: The American Society of Mechanical Engineers,1983, pp. 353-356.

38. R. Viswanathan, "Dissimilar Metal Weld and Boiler Creep Damage Evaluation for Plant Life Exten-sion," Residual-Life Assessment, Nondestructive Examination, and Nuclear Heat Exchanger Materi-als, Proceedings of The American Society of Mechanical Engineers 1985 Pressure Vessels and PipingConference, New Orleans, Louisiana, June 23-26, 1985, PVP-Volume 98-1, pp. 9-18.

39. G. H. Harth and T. P. Sherlock, "Monitoring the Service Induced Damage in Utility Boiler PressureVessels and Piping Systems," Residual-Life Assessment, Nondestructive Examination, and NuclearHeat Exchanger Materials, Proceedings of The American Society of Mechanical Engineers 1985 Pres-sure Vessels and Piping Conference, New Orleans, Louisiana, June 23-26, 1985, PVP-Volume 98- 1, pp.19-23.

40. J. F. Henry and F. V. Ellis, Plastic Replication ProceduresforDamageAssessment, progress report onEPRI Project RP2253-1, September 1983.

41. S. 0. Hilton and M. F. Steakley, "Steam Line Inspection Program for Tennessee Valley Authority'sHigh Temperature Steam Piping Systems," EPRI Conference on Life Extension and Assessment ofFossil Plants, Washington, D.C., June 1986.

42. J. A. Janiszewski et al., "A Production Approach to On-Site Metallographic Analysis," EPRI Con-ference on Life Extension and Assessment of Fossil Plants, Washington, D.C., June 1986.

43. T. Takigawa et al., "Life Assessment and Extension of Steam Turbine Rotor-Overview of MHI'sActivities, " EPRI Conference on Life Extension and Assessment of Fossil Plants, Washington, D. C.,June 1986.

166

I

44. T. R. Padden and C. A. Daniels, "Replication Manipulator and Procedure for Tube Bore Surfaces,"Annual Meeting of Scanning Electron Microscopy Inc., Dallas, Texas, April 1981.

45. T. R: Padden and C. A. Daniels, In Situ Examination'of the Small Scale Steam Generator Model,WARD-1i10134-1, Westinghouse Electric Corporation, October 1976.

46. T. R. Padden, Metallurgical Characterization of the Creep Ratcheting Test Article Failure, WAESD-HT-94000-28, Westinghouse Electric Corporation, June 1983.

47. C. E. Jaske et al., "Life-Assessment of High-Temperature, Fossil-Plant Components,' EPRI Con-ference on Life Extension and Assessment of Fossil Plants, Washington, D. C., June 1986.

48. M. P. Manahan et al., "The Development of a -Miniaturized Disc Bend Test for Determination ofPost Irradiation Mechanical Properties," Journal of NuclearMaterials, 103-104, 1981, pp. 1545-1550.

49. F. M. Haggag et al., "The Use of Miniaturized Tests to Predict Flow Properties and Estimate Frac-ture Toughness in Deformed Steel Plates," International Conference and Exhibition on Fatigue, Cor-rosion Cracking, Fracture Mechanics, and Failure Analysis, Salt Lake City, Utah, December'1985.

50. T. Misawa and Y. Hamaguchi, "Miniature Specimen Tests for Evaluating Fracture Resistance WithRadiation and Hydrogen Embrittlements in Ferritic Steel," in TheMechanism of Fracture, V. S. Goel(ed.), Metals Park, Ohio: American Society for Metals, 1986, pp. 521-527.

51. Battelle Columbus Division, Nuclear Tech Notes, 3, Spring 1984. -

52. M. C. Askins et al., "Estimating the Remanant Life of Boiler Pressure Parts," Meeting of the EPRIAdvisory Group on Remaining Life of Boiler Pressure Parts, Palo Alto, California, June 4, 1985.

53. A. Saxena et al., "Evaluation of Remaining Life of High Temperature Headers: A Case History,"EPRI Conference on Life Extension and Assessment of Fossil Plants, Washingion, D.C., June 1986.

54. -J. S.-Perrin et al., "Preparation of Reconstituted Charpy V-Notch Impact Specimens for GeneratinglPressure Vessel Steel Fracture Toughness Data," Effects of Radiation on Materialsi. Eleventh Confer-'ence, American Societyfor Testing andMaterials, ASTM Special Technical Publication 782, 1985, pp.582-593.

55. R. P. Shogan, S. E. Yanichko, and W. S. Galloway, "The Use of Reconstited Charpy Specimens to.Extend R. E. Ginna Reactor Pressure Surveillance Data," Nuclear Technology 72, 3, March 1986.

56. - K. R. Balkey and E. L. Furchi, "A Cost Effective Approach for Extending the Life of Pressurizedwater Reactor Vessels," Residual-Life Assessment, Nondestructive Examination, and Nucleai HeatExchanger Materials, Proceedings of :The American Society of Mechanical Engineers 1985 PressureV Vessels and Piping Conference, New Orleans, Louisiana, June 23-26, 1985, PVP-Volume 98-1, pp.175-176. .-

57. D. Sidey, "Utility Plant Life Assessment as a Continuous Procedure," EPRI Conference on LifeExtension and Assessment of Fossil Plants, Washington, D.C., June 1986.

58. -W. G. Clark, Jr.-et al., "The Role of NDE in Structurail Risk Assessment," in Advances in Life- :Prediction Methods, New York: The American Society of Mechanical Engineers, 1983, pp. 293-299.

*, ,6* ' ,- ...

167

I

15. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS

The first step in the residual life assessment task wasto identify the major light water reactor (LWR) compo-nents that are critical to nuclear plant safety duringnormal operation, off-normal situations, design-basisaccidents, or severe accidents. This report has identi-fied and prioritized these components according totheir role in preventing the release of fission products tothe public. Table 15.1 presents the list of selected pres-surized water reactor (PWR) components for the resid-ual life assessment task and reasons for their ranking.This table also identifies the most likely and otherpotential degradation sites and aging mechanisms foreach component. Similarly, Tible 15.2 presents the listof selected boiling water reactor (BWR) components.

The next step in the residual life assessment task wasto perform a qualitative analysis of the degradationprocesses active in each of these components and care-fully identify corresponding degradation sites, stres-sors, degradation mechanisms, and potential failuremodes during normal operation and accident condi-tions. This report presented the results of the analysesfor seven major components: four PWR components(vessel, containment, primarycoolant piping, andsteam generator), two BWR Components (vessel andrecirculation piping), and supports for both PWR andBWR vessels. The currently employed in-serviceinspection (ISI), surveillance, monitoring, and life-assessment methods were also evaluated. Similar stud-ies will be completed in, 1987. of the, degradationprocesses associated with the remaining major compo-nents: four PWR components (reactor coolant pumpbodies, pressurizers, control rod drive mechanisms,and reactor internals), four BWR components (con-tainments, recirculation pump bodies, control roddrive mechanisms, and reactor internals), cables andconnectors, and emergency diesel generators in PWRand BWR plants.

This review of potential or, in many cases, actualdegradation sites, mechanisms, and potential fail-ure modes associated with the major LWR compo-nents and structures has raised a number ofquestions or issues deserving further attention. Asummary of the results, presented in this report,conclusions based on these results, and the recom-mendations for further work are grouped by com-ponent and listed below. (these recommendationsare indicative of general needs and not necessarilytasks that might be accomplished at the Idaho-National Engineering Laboratory as part of theNuclear Plant Aging Research program). Finally, asummary of the results, corresponding conclu-

sions, and recommendations related to the currentand emerging nondestructive examination (NDE)and life-assessment methods is presented.

15.1 PWR Reactor PressureVessels

Table 15.3 summarizes the important degradationsites, stressors, degradation mechanisms, potential fail-ure modes, and current ISI and surveillance methodsassociated with PWR pressure vessels. Note that theranking of sites is based on the safety impact of thatsite and not on the possibility of that site having severedegradation. The major conclusions and recommen-dations are as follows:

1. Radiation embrittlement is the primarysafety concern that: may limit vessel lifeextension. Therefore, continued closemonitoring of the pressure vessel materialcondition is required. The current correla-tions used to estimate the shifts in RTNDTand upper-shelf-energy changes, areuncertain because of the limited accuracyand range of the surveillance data base.The data base does not include the effectsof high-fluence and low-flux conditions,and treats the effects of spectrum and irra-diation temperature in an approximatemanner. A comprehensive plan to obtainreactor pressure vessel (RPV) materialsdata from environments (i.e. irradiation,thermal, and chemical) that reasonablyrepresent the realities of long-term in-service behavior shouid be developed.

2. Use of the temperature shift at 30 ft-lb(41 J) level, ARTNDT, from the Charpy V-notch curve for certain RPV materials mayunderpredict the actual temperature shiftat the measured transition-temperaturefracture toughness of 100 ksi %/W There-fore, the validity of using Charpy testresults for estimating temperature shifts todetermine transition-temperature fracturetoughness needs further assessment. How-ever, the conservatisms in the overall refer-ence toughness curves and in the shiftingof these curves may outweigh this differ-ence between Charpy V-notch and fracturetoughness tests.

.

168

I

Table 15.1. Key PWR components for residual life assessment

Rank Component

I Reactorpressurevessel

Reasons for Ranking

Severe safety impactof catastrophic failureof an embrittledvessel

Degradation Sites(most likely, others)

.Weldments in the beltline region, hot-leg andcold-leg nozzles, weldmentat the lower shell tobottom head junction,vessel flanges and studs,instrument penetrations inlower head

Degradation Mechanisms(most likely, others)

Neutron embrittlement,corrosion fatigue, thermal-fatigue

2 Containment Public protectionand basemat during an accident

Posttensioning system(tendons and anchors),concrete, metal liner,reinforcing steel

Hydrogen embrittlement andstress corrosion cracking ofanchor heads, corrosion andrelaxation of tendon material,environmental degradation ofconcrete, corrosion ofreinforcing steel and liner

3 Reactorcoolantpiping andsafe ends

4 Steamgenerator

Severe safety impactof a large break

Tube rupture willprovide a passage fromthe primary systemdirectly to theenvironment for theprimary coolant, -relatively pooroperating experience

Weldments at the safeends, branch nozzles, caststainless steel components(elbows andt-connections), complete

* primary loop piping innew Westinghouse plants

Inside tube surfacesat U-bends and tubesheet, outside surfacesat tube-to-tube sheetcrevices, divider plate,tube support plate,feedwater nozzle,girth weld

Fatigue, thermal embrittlement

I.

Wastage, denting,intergranular stresscorrosion cracking, pitting,fretting, intergranular attack,thermal fatigue, corrosionfatigue

I

5 Reactorcoolantpump casing

6 Pressurizer

7 Control roddrivemechanism

Primary protectionfrom any internalfailure of pumpelements (impellerparts, shafts), primarycoolant pressureboundary

Primary coolantpressure retainingboundary

Failure may lead to areactivity initiatedaccident, ananticipated transientwithout scram, or aloss-of-coolantaccident

Casing wall, flanges,seal housing bolts

Surge and spray nozzles,spray head, upper shellbarrel, seismic lugs

Drive rod assembly,control rod pressurevessel

Corrosion fatigue, thermalembrittlement

Thermal fatigue, erosion,thermal embrittlement

Wear and thermalembrittlement

!

I

169

-I

I

Table 15.1. (continued)

Rank Component

8 Safety-relatedcables andconnectors

9 Emergencydieselgenerator

10 Reactorinternals

II RPVsupports

Reasons for Ranking

Active during normaloperation in mitigatingoperational transientsand accidents

Needed to operatecritical safetyequipment in the eventof a loss-of-offsitepower

Failure may lead todisbursement of fuelinto the coolant, or areactivity-initiatedaccident

Failure will challengethe integrity of theprimary coolantpiping and the smalllines connected to theRPV; the neutronshield tanks maintainacceptable levels ofradioactivity in thecontainment

Degradation Sites(most likely, others)

Cable insulation, insertsin connectors

Governor in theinstrumentation andcontrol system, coolingsystem pumps and piping,fuel injector pumps,turbocharger

Lower core plate,baffle-former assembly,upper support columnbolts, control rod bolts,control rod guide tubesheath and support pins,thermal shield bolts,in-core instrumentnozzles, flux thimbletubes bolts, in-coreinstrument nozzles, fluxthimble tubes

Inside surface of thesupport structure (neutrontank or column support)at the core horizontalmidplane level, lubricantin sliding foot assemblyof neutron tank support

Degradation Mechanisms(most likely, others)

Thermal aging and creep ofinsulation, thermalembrittlement andcorrosion of connectors

Fatigue and vibrations

Neutron embrittlement, wear,high-cycle fatigue, SCC,relaxation

Neutron embrittlement,corrosion, degradation oflubricant

I

i

I

170

1

Table 15.2. Key BWR components for residual life assessment

Degradation Sites(most likely, others)

Degradation Mechanisms(most likely, others)Rank Component Reasons for Ranking

I Containment Public protection during anand basemat accident

Concrete, metal liner, andreinforcing steel insuppression pool andprimary containment wall,'base metal, welds and ventpipes in metal containment

Nozzles, closure studs,beltline region, welds atthe stub tubes

Safe ends, austeniticstainless steel fittings

Environmental degradationof concrete, fatigue andcorrosion of metalcontainments

2 Reactorpressurevessel

3 Recirculationpiping, safeends, andsafety systempiping

Primary pressure boundary, pooroperating experience with reactorpressure vessel nozzles

Primary pressure boundary,relatively poor operatingexperience

Thermal fatigue, neutronembrittlement, intergranularstress corrosion cracking

-Intergranular stresscorrosion cracking,thermal aging

4 Recirculation Primary protection from anypump body internal failure of pump elements,

primary coolant pressure boundary

Heat affected zones nearweldments in the wall

Thermal aging, corrosionfatigue, crevice corrosion

5 Control roddrivemechanism

6 Safety-relatedcables andconnectors

7 Emergencydieselgenerator

8 Reactorinternals

9 - Reactorpedestal

10 Biologicalshield

Failure may lead to a reactivityinitiated accident, an anticipatedtransient without scram (ATWS),or a small-break, loss-of-coolantaccident

Active during normal operationin mitigating operational transients-and accidents

Needed to operate criticalsafety equipment in the eventof a loss-of-offsite power

Failure may cause fuelfailure or problems inscraming the reactor

Failure will challenge integrityof reactor pressure vessel

Maintains acceptable level ofradioactivity in containment

Drive rod assembly,control rod pressurevessel

Cable insulation, insertsin connectors

Governor in theinstrumentation andcontrol system, cooling andlubrication system pumpsand piping, fuel injectorpumps, turbocharger,generator windings

Wear and thermalembrittlement

Thermal aging of insulation,thermal embrittlement andcorrosion of connectors

Fatigue and vibrations

Irradiation-assistedstress corrosioncracking, intergranularstress corrosioncracking, fatigue

Core shroud, top guideplate, core plate,holddown beam in ajet pump, feedwaterand core spargers .

Concrsteel

Insidebiologcore hilevel

ete and reinforcing ---Thermal cycling of concrete,corrosion and fatigue ofreinforcing steel

surface of the Neutron irradiation,ical shield at the gamma-ray heating;.orizontal midplane 'environmental

; degradation of concrete,corrosion of reinforcing steels

i, -! ; 'i�-

: . : I

171 . ;

Table 15.3. Summary of degradation processes for PWR reactor pressure vessels

Rank ofDegradation

SiteDegradation

Site Stressors Degradation Mechanisms Potential Failure Modes ISI Surveillance Methods

I Beitline region Neutronirradiation,mechanicaland thermalstresses

Irradiation embrittlement(degree is dependent onindividual vesselmaterials and fluxspectrum history)

Ductile high-energytearing leading toleakage (net sectionover-load)

Brittle fracture (i.e.,pressurized thermalshock)

100% volumetric duringfirst inspection; oneweld for subsequentinspection

Surveillance program forassessing irradiationdamage is required by law V

Ductile low-energytearing (low upper-shelf toughness)

Environmentalfatigue

Ductile overload leadingto a leak, possible brittlefracture if PTS occurs

2 Outlet/inlet Mechanicalnozzlesa and thermal

stresses

3 Instrumentationnozzles(penetrations)and control roddrive mechanismhousing nozzles

Mechanicaland thermalstresses

Fatigue crack initiationand propagation

Fatigue crack initiationand propagation

Fatigue crack initiationand propagation (possiblycorrosion-assisted)

Ductile overload leadingto a leak; possiblebrittle fracture ifpressurized thermalshock occurs withsome irradiationembrittlement

Ductile overload leadingto a leak

Ductile overload failure(can be replaced)

All nozzle welds inspectedvolumetrically at eachinterval

Visual, inspection ofexternal surface; 253% ofnozzles inspected at firstinterval; remaining 750espread out over next threeintervals

Volumetric and surfaceinspection of all studsand threads in flange studholes at each interval

4 Flange closure Mechanicalstuds and thermal

stresses

a. Welds at the vessel to safe end are covered in Chapter 5 on primary system piping; the comments here are also applicable exceptthat volumetric and surface examinations are required.

II3. Current RPV materials surveillance pro-

grams have limited space and limited num-bers of specimens of critical material.Therefore, miniature and reconstitutedspecimen testing should be performed, sothat additional mechanical property datacan be extracted from the given limitedamount of critical material. Also, the cur-rent surveillance programs should be mod-ified as necessary to ensure that specimensare available for surveillance testing duringthe renewed license period.

4. Thermal fatigue of the RPV nozzles and pen-etrations is an important safety concern asso-ciated with license renewal. Techniques forbetter monitoring of fatigue transients fornozzles and penetrations should be developedso that the actual usage factors may be deter-mined. To improve the life-prediction capabil-ity of the American Society of MechanicalEngineers (ASME) Section III and XIfatigue design analyses, a modified procedurefor crack initiation should be developed,based on the local-strain approach that

I

172

I

I

accounts for the effect of both stress andstrain concentrations at notches and of thesequence of variable amplitude cycles. Inaddition, an updated version of the damage-tolerant procedures that incorporate animproved appreciation of environmentalfactors, low-growth-rate behavior, and therole of small flaws should be developed.

5. Inspection and transient logging records aswell as all material properties (unirra-diated and surveillance results), and designdata need to be maintained. Collection ofthese records is needed to develop life-extension plans.

6. Thermal annealing of reactor pressure ves-_sels should be pursued as a long-termoption for license renewal. Annealingmechanisms and kinetics need to be betterunderstood and corresponding models,need to be developed.

15.2 PWR Containments andBasemats

Tables 15.4, 15.5, and 15.6 (for prestressed con-crete, reinforced concrete, and steel PWR contain-ments and basemats, respectively) summarize theimportant degradation sites, stressors, degradationmechanisms, potential failure modes, and the currentISI, surveillance, and testing methods. The major con-clusions and recommendations are as follows:

1. Hydrogen embrittlement, pitting, andmicrobiological-inrduced corrosion are themajor degradation mechanisms associatedwith the posttensioning system anchors,tendon wires, and grease in the prestressedconcrete containments. Monitoring meth-ods that will detect the degradation of theanchors and decomposition of the tendongrease are needed.

2. Aggressive environments and internalchemical reactions are the major stressorsdamaging the concrete in concrete contain-ments. Corrosion is the major degradationmechanism associated with the reinforcingbars and steel liners in concrete contain-ments. Additional information about thelong-term degradation of reinforced andprestressed concrete containments isneeded. Data should be collected from theolder LWR containments and facilities

that have been shut down after extendedservice. Accelerated aging techniquescould also be investigated and, if foundappropriate, used to obtain additionaldata. The results from these studies shouldsupport the development of quantitativeresidual life models.

3. A comprehensive and standardized ISIprogram can and should be developed toidentify and quantify degradation in rein-forced and 'prestressed concrete contain-ments. State-of-the-art ISI methods used

- - in Europe,-Japan, and the United'Statesneed to be evaluated and included in theprogram where appropriate.

4. The impact of the concrete-liner interactionon the integrity, of the aged reinforced-concrete and prestressed-concrete contain-ments during severe accidents should beevaluated. The relevant research programssponsored by the Electric Power ResearchInstitute are expected to provide useful infor-mation.

15.3 PWR Coolant Piping andSafe Ends

Table 15.7 summarizes important degradation sites,stressors, degradation mechanisms, potential failuremodes, and current ISI methods associated with thePWR coolant piping and safe ends. The major conclu-sions and recommendations are as follows:

1. Low-cycle fatigue is the main degradationmechanism for the main coolant pipe nozzlesand the dissimilar metal welds at terminalends. However, the number, type, and severityof transients are generally not recorded.Therefore, an in-depth study of past opera-tions in selected nuclear power plants wouldprovide some insight into the rate at which thedesign life is being consumed because offatigue.

2. Cast austenitic-ferritic (duplex) stainlesssteel components, i.e., main coolant pipe,elbows, T-connections, and valve bodies,are subjected to thermal embrittlement.The actual degree of thermal-agingembrittlement of these cast stainless com-ponents should be monitored.

173 -

I'

Table 15.4. Summary of degradation processes for prestressed concrete containmentvessel

DegradationRank Site

I Posttensioningsystem anchorage

2 Posttensioningtendon wire orstrand

DegradationMechanisms

PotentialFailure ModesStressor

Material propertiesand trapped water

Moisture, trappedwater, or breakdownof grease material

ISI Method

Hydrogenembrittlement

Loss of stress Tendon surveillanceprogram

Pitting.microbiological-induced corrosion

Loss of stress Tendon surveillanceprogram

3 Steel liner domeand wall

Moisture, acidicenvironment, andstress

Corrosion andcracking

Liner-concreteinteraction,leakage ofradioactive gases

Leakage ofradioactivematerial

Leakage testing(10 CFR 50,Appendix J)

4 Steel liner overbase slab

Moisture, acidicenvironment, andstress

Corrosion Leakage testing(10 CFR 50,Appendix J)

5 Dome, wall, andbase slabreinforcingsteel

Aggressiveenvironment

Corrosion Loss of structuralintegrity

Visual

6 Concrete Aggressiveenvironment, andinternal chemicalreactions

Cracking andspalling

Loss of integrity,corrosion ofreinforcing steel

Visual, reboundmethods, coresamples if required

174

I'

I

Table 15.5. Summary of degradation processes for reinforced concrete containmentvessel

DegradationRank Site

DegradationMechanisms

PotentialFailure ModesStressor ISI Method

I Dome and wallreinforcingsteel

2 Base slab' reinforcing

steel

Aggressiveenvironment

Aggressiveenvironment

Corrosion

Corrosion

*Loss ofstructuralintegrity

Loss ofstructuralintegrity

Visual

Visual

3 Steel linerover dome andwall

Moisture, acidic-environment, andstress

Corrosion Liner-concreteinteraction,leakage ofradioactivegases

Leakage testing(10 CFR 50,Appendix J)

4 Steel linerover base slab

5 Dome and wallconcrete

6 Base slabconcrete

Moisture, acidicenvironment, andstress

Aggressiveenvironment,internal chemicalreaction

Aggressiveenvironment,internal chemicalreactions

Corrosion

Cracks andspalling

Cracks andspalling

Leakage ofradioactivematerial

Loss ofintegrity,corrosion ofreinforcingsteel

Loss ofintegrity,corrosion ofreinforcingsteel

Leakage testing(10 CFR 50,Appendix J)

Visual, reboundmethods, coresamples ifrequired

Visual, reboundmethods, coresamples ifrequired

175

I

Table 15.6. Summary of degradation processes for steel cylinderlsteel spherecontainment vessel

RankDegradation

Site

I Shell welds andbase metal

2 Interfacebetween shelland concreteslab at baseof shell

Stressor

Stresses, vibration,cyclic loading,aggressiveenvironment

Stresses, vibration,cyclic loading,aggressiveenvironment

DegradationMechanisms

Corrosion

Corrosion

PotentialFailure Modes

Loss of structuralintegrity, leakageof radioactivegases

Loss of structuralintegrity, leakageof radioactivegases

ISI Method

Visual, leakagetesting

Visual, leakagetesting

3 Discontinuitiesin the shellsuch as hatchesand penetrations

4 Steel bottomof shellembedded inconcrete

5 Base slab concrete

Stresses, vibration,cyclic loading,aggressiveenvironment

Corrosion Leakage ofradioactivegases

Leakage testing(IO CFR 50,Appendix J)

Aggressiveenvironment

Corrosion Leakage ofradioactivematerial

Leakage testing(10 CFR 50,Appendix J)

Aggressiveenvironment,internal chemicalreactions

Cracks andspalling

Corrosion ofreinforcingsteel, corrosionof steel bottomof containmentshell

Visual, reboundmethods, coresamples ifrequired

Table 15.7. Summary of degradation processes for PWR primary system piping

Rank ofDezradation Site

DegradationSite

I Main coolantpipe nozzlesa

2 Terminal enddissimilarmetal weldb

Stressors

Systemoperatingtransients

Systemoperatingtransients

Systemoperatingtransients

DegradationMechanisms

Fatigue crackinitiation andpropagation

Fatigue crackinitiation andpropagation

Thermalembrittlementleakage

PotentialFailure Modes

Through wallleakage

Through wallleakage

Through wallleakage

ISI Methods

Volumetricinspectionfor diametera 4-in.

Surfaceinspectionfor diameter<4-in.

Volumetric3 Cast stainlesssteel

a. The nozzles in the primary coolant piping are the highest ranking degradation sites. Additional specificity is avoided because themost severely loaded nozzles are difficult to determine. Reported results are heavily dependent on type of analysis performed. Actualdamage will completely depend on the actual transient usage that occurs in the plant. This is the basis for the recommendation.

b. In Westinghouse plants, the dissimilar metal welds are at the reactor vessel and steam generator nozzles.

176

I

15.4 Steam Generators:

Table 15.8 summarizes the important degrada-tion sites, stressors, mechanisms, potential failuremodes, and current ISI methods associated withthe recirculating and once-through steam genera-,

. . tors. The major conclusions and recommendations' are as follows:

.. . .

1. Intergranular stress corrosion cracking* (IGSCC) and intergranular attack (IGA) are

- the major degradation mechanisms for the

Table 15.8. Summary of degradation processes for steam generator tubes

Degradation Degradation u

Ranka Site Stressors i Mechanisms Failure Modes ISI Method

I Inside surface ofU-bends and roll-transition regions

Tube rolling stresses,Corion phenomenon,denting

IGSCC Cracking Eddy-current testing

2 Outside surface of*hot-leg tubes inthe tube-to-tube

*sheet crevice region

3 Cold-leg side insludge pile or wherescale containing

.copper deposits isfound

Alkaline environment, IGA, IGSCCpresence of S0 4 and

-CO3 anions

Brackish water, air, Pittingand copper

May eventually result incracking ,

Local attack and tubethinning may eventuallylead to a hole

Eddy-current testing

Eddy-current testing,optical scannersystem, sonic leakdetector system

4 Outside surface oftubing above tubesheet l

Phosphate chemistry,chloride concentration,resin leakage fromcondensate polisher bed

Wastage (thinning) Uniformn attack, tubethinning may eventuallywear out the material

Eddy-current testing

5 Tubes in the tube-support regions

Oxygen, copper oxide, Dentingchloride, temperature,pH, crevice conditions

Flow blockage in tubescaused by plasticdeformation

Helium leak andsonic leak testing,optical probes,hydrogen evaluationmonitoring, pulse -echo ultrasoundmethod

6 Contact between tubeand anti-vibrationbar

Flow-inducedvibrations

Fretting Wear out of materialcaused by rubbing and/or

; fatigue -

'Eddy-current testing

7b Once-through steamgenerator tubes -

Velocities, sizes, __ ' . Erosion-corrosion'shapes, impact angle, :',frorn impingementand hardness of . - of particlesparticles "

Wear out of material Eddy-current testing

8b Once-through steam - Chemicals, (localized -Fatiguegenerator tubes in ,,corrosion) vibrationsthe upper iube sheet .region

a. Based on operating experience for steam generator defects.

Primary to secondaryleaks

Eddy-current testing

_ . .' . . .b. Denotes once-through steam generator (Items 7 and 8 do not reflect rank order). The first six items are for recirculating steamgenerators.

177

I'

recirculating steam generator tubes. Pitting isanother important degradation mechanismwhich progresses at an accelerated rate onceinitiated, and failures often occur suddenly.Therefore, methods to monitor the degrada-tion of steam generator tubes caused byIGSCC, IGA, and pitting are needed. Also, abetter understanding and better models ofthese corrosion mechanisms are needed.

2. Several changes in the nuclear power plantwater chemistry, and the design and mate-rials of construction of the recirculatingsteam generators (SG) and/or other plantcomponents have been effective in mitigat-ing the degradation damage. The waterchemistry changes include use of an all-volatile treatment for the secondary-sidewater chemistry rather than a phosphatetreatment, addition of boric acid to thesecondary coolant, use of full-flow con-densate polishers, and use of tube sheetcrevice flushing. The changes related to theplant design and materials include achange from copper to titanium or stain-less steel as condenser tube material, use ofstainless steel for the feedwater lines, useof a quatrefoil design and Type 405 stain-less steel for the SG tube support plates,SG tube rolling to the full depth of the tubesheet, shot peening of the tubing insidesurfaces in the U-bend and roll-transitionregions of the SG tubes, and use of ther-mally treated Inconel 690 in place ofInconel 600 for the SG tubes. The gap sizebetween the antivibration bars and the SGtubes has also been reduced in somedesigns to avoid fretting. The effects ofthese mitigating techniques on the failurerates of the SG tubes and tube supportplates should be monitored.

3. Erosion-corrosion and environmental fatigueare the major degradation mechanisms asso-ciated with the once-through steam genera-tors. The damage caused by environmentalfatigue has been reduced by blockage of theuntubed lane with a lane blocker.

4. Bobbin-type eddy-current coils have not beensuccessful in all cases in detecting IGAdefects. However, successful detection hasbeen achieved by the use of pancake coils thatare arrayed with their axes parallel to the tuberadius and multiplexed to permit use of avail-able electronic packages.

15.5 Reactor Pressure VesselSupports

Table 15.9 summarizes the important degrada-tion sites, stressors, degradation mechanisms,potential failure modes, and current ISI methodsassociated with the five different types of RPV sup-ports: neutron shield tank, column, cantilever,bracket, and skirt-type supports. The first fourtypes of supports are used with PWR RPVs, whilethe fifth type of support is used on BWR RPVs.The major conclusions and recommendations areas follows:

1. The neutron shield tanks, columns, andcantilever-type supports may be suscepti-ble to catastrophic brittle failure caused byirradiation embrittlement. The neutronspectra have a very large component oflow-energy neutrons (E < 1 MeV), and thesupport temperatures are less than 4500F(2320C). Both the fracture toughness andstrength of the RPV support steels at highfluence and low irradiation temperaturesneed to be determined. Also, a correlationof fracture toughness with Charpy V-notchproperties at temperatures less than 4500F(2320C), and as a function of displace-ments per atom is needed.

2. A better determination of the radiation envi-ronment (neutron spectra and flux levels) isrequired for the neutron shield tank, canti-lever, and column-type support structures.

3. The dry lubricant in the sliding foot assem-bly of the neutron tank support degradesbecause of neutron irradiation. The effectsof the radiation levels on the lubricantsthat are used in the RPV nozzle supportsshould be investigated.

15.6 BWR Reactor PressureVessels

Table 15.10 summarizes the important degradationsites, stressors, degradation mechanisms, potential fail-ure modes, and current ISI and surveillance methodsfor the BWR pressure vessel. The major conclusionsand recommendations are as follows:

1. Low-cycle fatigue of the RPV nozzles isthe major degradation mechanism of con-cern. Refined cycle counting and transient

178I,

I

Table 15.9. Summary of degradation piocesses for PWR a'nd BWR RPV supports

Rank Degradation SiteaDegradation Potential

Stressor - Mechanism Failure Modes ISI Methodb

I Neutron shield tank atthe core horizontalmidplane elevation

2 Column support atthe core horizontalmidplane elevation

3 Cantilever support inthe active height ofthe core

4 Threaded parts insliding foot assembly

Neutron irradiation,tensile stresses, operatitemperature, waterchemistry

Neutron irradiation,tensile stresses, operatitemperature

Neutron irradiation,tensile stresses, operatitemperature

Tensile stresses,operating temperature

ngNeutron embrittlement, Catastrophiccorrosion ' : brittle failure

I ' Monitoringc

Monitoring and sampling

Monitoring and sampling

Neutron embrittlementng

Neutron embrittlementng

Stress corrosioncracking

Degradation caused by; neutron irradiation

Catastrophicbrittle failure

Catastrophicbrittle failure

5 Dry lubricant in Neutron irradiation,sliding foot assembly operating temperatures

Binding that causespossibly excessivestresses in the primarysystem during heatupand cooldown

Binding that causespossibly excessivestresses in the primarysystem during heatupand cooldown

Ductile overloadfailure

6 Skirt support Mechanical and thermal Fatiguestresses

a. The bracket and skirt RPV supports will experience no neutron embrittlement. x ,

b. There are no national standard ISI methods per se to determinie state of degradation.

c. Monitoring of neutron field near RPV support.

severity estimates should be used to moni-tor nozzle fatigue usage so that actualfatigue usage factors can be determined.

2. Closure studs, 'flange bushings,' and studholes in flanges without bushing are sub-jected to wear, fretting, and corrosion.These components should be examinedand replaced 'or repaired as appropriatenear the 40-year end of life.

3. Brittle failire of a BWR vessel caused byirradiation embrittlement is extremelyunlikely because of the relatively low neutronflux and fluence in the beltline region. How-

- ever, the effects of irradiation embrittlementof BWR vessels should be monitored. Cur-rent RPV materials surveillance programsshould be modified for license renewal con-sideration.

4. Many of the older BWRs have limitedaccessibility for external ISI of the welds inthe beltline region. ISI methods for inspec-tion of these welds from the vessel insidesurfacerneed to be developed.'

15.7 BWR Recirculation Piping

Table 15.11 summarizes the important degrada-tion sites, stressors, mechanisms, potential failuremodes, and current ISI methods for the BWR'recir-culation piping. The'major conclusions and recom-mendations are as follows:

1. Aging degradation of the recirculationpiping caused by IGSCC has been a signif-icant problem at a number of BNVRs. The

179

I'

Table 15.10. Summary of degradation processes for BWR reactor pressure vessel

Rank of Degradation Degradation Potential ISI/SurveillanceDegradation Site Site Stressors Mechanisms Failure Modes Methods

I Nozzles (including Mechanical Fatigue crack initiation Ductile overload All large nozzle weldsinstrument and CRD and thermal and propagation, IGSCC leading to a leak inspected volumetricallypenetrations) plus stresses at each interval;safe end welds visual, external surface

inspections of smallnozzles/penetrations

2 Closure studs,flange bushings,stud holes

3 Beltline region

Mechanical Fatigue crack initiation Ductile overload Volumetric and surfaceand thermal and propagation, failure (can be inspections of allstresses fretting, corrosion replaced) studs, threads in flange

stud holes and bushings

Irradiation Neutron irradiation Ductile high- 100%6 volumetricembrittlement (extent depends on vessel energy tearing inspection; surveillance

materials) leading to a leak as required by federal(not a serious lawproblem)

4 External attachments Mechanical Fatigueand thermalstresses

Ductile overload Volumetric and surfacefailure inspections

Table 15.11. Summary of degradation processes for- BWR recirculation piping

Degradation DegradationRank Site Stressors Mechanisms Failure Modes SI Method

I Weld heat-affected Tensile stress, oxygenzones furnace environment, sensitizedsensitized safe ends heat-affected zones

IGSCC Cracks, leaks Ultrasonic examination,moisture-sensitive tape

2 L High thermal stress . Cyclic tensile stress,regions predicted by corrosive environmentstress rule indexanalysis

Thermal fatigue,corrosion fatigue

Cracks, leaks Ultrasonic examination,moisture-sensitive tape

3 Austenitic-ferriticstainless steelcastings with highdelta ferrite levels

High-temperature,tensile stress, shockloads

Thermal embrittlement Cracks, leaks Ultrasonic examination,moisture-sensitive tape,impact test specimens

4 Shaft sleeves, other Corrosive environment, Corrosioncrevices oxygen-starved areas

Material wastage, Moisture-sensitive tapeleaks

180

heat-affected zones of the recirculation-piping welds have been particularly sus-ceptible. Three conditions are required forIGSCC of austenitic stainless steel to exist:a sensitized microstructure, high tensilestresses, and an oxygenated environment.Several techniques have effectively miti-gated IGSCC damage. These techniquesinclude solution heat treatment of thewelds, heat sink welding, induction heat-ing stress improvements of the welds, last-pass heat-sink welding, mechanical stressimprovement processes, weld overlays,and use of hydrogen water chemistry.Nuclear-grade stainless steels are much lesssusceptible to sensitization and moreresistant to IGSCC, and therefore are usedas alternate materials for the recirculationpiping material.

2. The possible detrimental effects, if any, ofhydrogen additions to the BWR recircula-tion piping loop need to be evaluated.Addition of hydrogen will reduce the oxy-gen level in the BWR coolant and may leadto degradation of the feedwater linesbecause of a erosion-corrosion mechanismsimilar to that which caused the recentSurry pipe break.

3. A program should be developed to moni-tor the actual degree of thermal-agingembrittlement in the cast stainless steelcomponents in the recirculation pipingloop.

4. The long-term fatigue behavior of the weldoverlays needs to be better understood.

15.8 Nondestructive ExaminationMethods

The NDE methods used in standard practice are vis-ual examinations, penetrant and magnetic particle test-ing, radiographic examinations, eddy-current testing,ultrasonic testing, and acoustic-emission monitoring.These standard NDE methods, which are employed tosatisfy the present ASME code ISI requirements, arenot entirely adequate for residual life assessments.Inspection methods are needed to accurately determinethe size, shape, location, orientation, and type of bothsurface and internal flaws, so that fracture mechanicsapproaches may be used for life assessment. Severalemerging methods being developed to satisfy theseneeds are summarized in Table 15.12. Several emerging

inspection methods that can be used to detect time-dependent changes in microstructural features (such asthe ferrite phase in duplex stainless steels) are also sum-marized in Table 15.12. Recommendations for furtherwork on emerging inspection methods are as follows:

1. Development of the synthetic aperture focus-ing technique (SAFI) should be continued.SAFT ultrasonic testing (UT) can provideenhanced visual images of flaws detected dur-ing inspection. The initial results of the appli-cation of SAFT-UT in the nuclear industryare encouraging.

2. The dc potential drop method and computer-aided eddy-current testing techniques need tobe evaluated and their ability to detect flaws inpipe needs to be determined.

3. The use of acoustic emission technologyfor on-line monitoring of crack growth inpressure boundaries, and leak surveillancein LWR systems should be evaluated.

4. NDE methods for measuring the materialproperties changes caused by aging of themajor LWR components are needed. Someof the promising methods are eddy-currentmethods to measure residual stresses, elec-trochemical methods to measure the sus-ceptibility of austenitic stainless steel toIGSCC, and indentation hardness mea-surements to assess radiation embrittle-ment of RPV steel.

15.9 Emerging Life-AssessmentMethods

Emerging life-assessment methods include min-iature specimen testing (MST), reconstituted speci-men testing, on-line damage monitoring, 'andeffective engineering models applications.Table 15.12 summarizes two of these life-assessment methods. These methods may providevaluable information concerning several importantdegradation mechanisms. For example, MST canbe used for surveillance testing where only a limitedamount of test material is available. On-line dam-age monitoring can be applied to calculate fatigueusage factors for RPV nozzles during startups,shutdowns, and major operating transients, wherenozzle temperatures are monitored during opera-tion. The engineering models can then be used-along with the information from the improved ISImethods, data from MST activities, and on-line

181

monitoring data-to properly estimate the residuallife of various components. These estimates willform the basis for LWR license renewals. Recom-mendations for further work on the emerging life-assessment methods are as follows:

1. The usefulness for RPV surveillance test-ing of the miniature specimen and recon-stituted specimen testing methods needs tobe evaluated. The validity of the tensileproperty, fracture toughness, and fatigue-

crack initiation and growth data obtainedwith these two methods also needs evalua-tion.

2. Development, field demonstration, andvalidation of the on-line damage monitor-ing methods are required before they canbe implemented with confidence.

3. Simplified engineering models should bedeveloped to provide reasonable estimatesof material aging as a function of operat-ing history.

;

182

Table 15.12. Summary of emerging methods for inspection and life assessment

Method Characteristics Applications

SAFT-UT Computerized processing of ;ultrasonic data to provide enhancedsignificant visual images of flaws inreal time.

.Welds in the RPV, in piping,and at nozzles. Requirescomputer resources resourcesresources and has not been fullydeveloped and verified for fieldapplications.

Computeinspectio

r-aided UT pipe Automated ultrasonic inspection ofn systems piping by means of state-of-the-art

hardware and software.

itial drop pipe Automated inspection andn systems computerized analysis of DC

- potential distributions around crackson both inner and outer surfaces ofpiping.

DC poteiinspectio

Piping and IGSCC cracks nearwelds in piping. Need more

-- real-time analysis andinterpretation of data to aidoperators in decision making.

Promising new method forinspection of pipes. Prototypesystem needs to be developed,demonstrated, and verified forroutine commercial applications.

More reliable inspection of: PWR steam generator tubes. The

advanced methods need to be- fully evaluated and

demonstrated. Deep penetrationmethods show some potential forpipe inspection, but they needconsiderable development.

IComputerized eddy-current inspectionsystems

Advanced eddy-current coil designscoupled with real-time computeranalysis of data. -

.. I ....

Acoustic emissionmonitoring

On-line damage analysis~~~~

Uses computerized processingof data to monitor noise from

-crack growth or-water leakage. :

Monitoring of fatigue-crackgrowth in RPVs, possiblystress-corrosion crack growth -in piping, and water leaks in theprimary pressure boundary.Requires advanced techniques toremove background noise.Results of initial field studies areencouraging, but furtherdemonstration and validation areneeded.

, Monitors temperature aniduring operation and condamage accumulation onreal time. ;

d pressure Creep-fatigue damage in fossil2putes boiler headers and fatigue-line in ; usage factors in nuclear

i- plant nozzles. Only prototype- ; systems are currently available.

Much uncertainty with damageaccumulation models that areused.

183

Table 15.12. (continued)

Method Characteristics Applications

Neutron-noise analysis

Nondestructivemeasurement ofmaterials properties

Surface replication

Miniature specimentesting

Effective engineeringmodels

Measures the ex-core neutron noiseand analyzes the recorded data todiagnose potential problems withinthe RPV.

Uses ultrasonic, eddy-current,x-ray, and electrochemical testingmethods to measure properties, suchas tensile strength, fracturedemonstrated, toughness, and impactstrength.

Produces plastic film replica ofmaterial surface that can beexamined at high magnificationsusing both optical and electronmicroscopy to detect small cracks,voids, and microstructural changes.

Uses a small specimen removed froma component to measure materialsproperties.

Use of simplified engineering modelsthat provide reasonable descriptionsof material aging as afunction of operating history.

Shows promise for monitoringvibration of PWR internalsand BWR instrument tubesand fuel boxes. Need referencedata for normal operations andhardware and software foron-line use.

Most applications have beenlimited to laboratory studies.Offers great potential whenmethods are developed,and verified by means of futurebasic and applied research. Needto develop a library ofwell-characterized correlationsbetween nondestructivemeasurements and properties.

Widely used to inspect for creepdamage in fossil-plant headersand steam piping. Has potentialuse for the detection of smallcracks and microstructuralchanges caused by aging in keynuclear components. Could beused to measure ferrite amountand spacing in cast stainlesssteels. Needs to be developed fornuclear-plant applications.

Laboratory studies are beingperformed to define the range ofuse for this technique. Showspromise for the measurement ofdeformation, such as creep andstress-strain response. May haveapplications for measurement offracture toughness and crackgrowth. Further verification ofthe technique is required for fieldapplication.

Trending the aging of keycomponents from a knowledgeof the operating history. Mostcurrent approaches seem eithertoo detailed and cumbersome touse or too simplistic to providerealistic predictions.

III

I

184

I

NRC FORM 335 U.S. NUCLEAR REOULATORY COMMISSION I REPORT NUMBER IAslgs,.ob, TIDC add V2 ho. Ac."l

C BIBLIOGRAPHICDAAS NUREG/CR-4731,,0,, 3702 BIBLIOGRAPHIC DATA SHEET EGG-2469 Volume 1SEE NSTIRUCTIONSON THE REVERSE

J TITLE AND SUBTITLE J LEAVE BLANK

Residual Life Assessment of Major Light Water ReactorComponents--Overview Volume 1

A DATE REPORT COMPLETED

MONTH ',EAR

6 I 87S AUTHORISI

V. N Shah and P. E. MacDonald e DATE REPORT ISSUED

MONTH YEAR

6 1. 87AGE ,4R oGANf12ION NAME AND MAILING ADDRESS I'Iciud. ZMI Coad) I B PROJECTITASK.WORK UN?7 NUMBER

calm 9 ano, lnc.P. 0. Box 1625 9 FIN OR GRANT NUMBER

Idaho Falls, Idaho 83415 A6389

10 SPONSORING ORGANIZATION NAME AND MAILING ADDRESS Ilcluoe Za Cod.l II TYPE OF REPOF°Final, TechnicalU.S.Nuclear Regulatory CommissionWashington, D. C. 20555

b PERIOD COVERED liu~ doal/

17 SUPPLEMENTARY NOTES

13 ABSTRACT (ZO.1d. e,-.t.

This report presents an assessment of the aging (time-dependent degradation) ofselected major light water reactor components and structures. The stressors, possibledegradation sites and mechanisms, potential failure modes and currently used non-destructive examinations, in-service inspection (ISI) and life assessment methods arediscussed for seven major light water reactor components: pressurized water reactor(PWR) and boiling water reactor (BWR) pressure vessels, PWR containment struc-tures, PWR reactor coolant piping, PWR steam generators, BWR recirculation pip-ing, and reactor pressure vessel supports. Unresolved technical issues related to lifeextension of these components, including requirements for advanced ISI and lifeassessment methods, are also discussed.

%A DOCUMENT ANALNSIS -. KENOORDSDESCRIPTODS 15 AVAILABILITISTATEMENT

Unlimited

16 SECURITY CLASSIsICATIO%

IT-., D.,.

II IODNTI IERS.OPEN ENDED TERMaS

rhs t*oo,)

UnclassifiedI7 N.MBEF O 'KGES

194IS PRICE

I

I