ML18040A226

Technical Evaluation Rept,Technical Basis for Reduction of Torus Shell Condensation Oscillation Loads for Niagra Mohawk Power Corp Nine Mile Point Unit 1.

ML18040A226
Person / Time
Site:	Nine Mile Point
Issue date:	05/31/1994
From:	ECONOMOS C, LEHNER J, LIN C C BROOKHAVEN NATIONAL LABORATORY
To:	Office of Nuclear Reactor Regulation
Shared Package
ML17059A382	List: ML18040A226
References
CON-FIN-L-1331 NUDOCS 9408100168
Download: ML18040A226 (106)

Text

TECHNICAL EVALUATION REPORTTECHNICAL BASISFORREDUCTION OFTORUSSHELLCONDENSATION OSCILLATION LOADSFORTHENIAGARAMOHAWKPOWERCORPORATION NINEMILEPOINTUNIT1byC.Economos, J.Lehner,andC.C.LinAccidentAnalysisGroupSafetyandRiskEvaluation DivisionDepartment ofAdvancedTechnology Brookhaven NationalLaboratory Upton,NewYork11973January1994RevisedMay1994PreparedforOfficeofNuclearReactorRegulation NuclearRegulatory Commission Washington, D.C.20555UnderContractNo.DE-AC02-76CH00016 NRCFINL-13119408i00168 9408051PDRADQCN05000220PDR,

'P4 LISTOFFIGURESFigure1.Figure2.Figure3.MOIpredicted Variation ofPressureinComputation CellPlaneofSymmetry-P(x,0,ZO/2).................................

19Comparison ofBNLandNMPCEstimates forPressureAmplitude Reduction Factors...............................

~.....20.ResultsofSensitivity StudiesforNMPviatheBNLMethodfImages.............................................

210 P

ABSTRACTBNL'sevaluation ofthetechnical basissubmitted byNMPCtojustifyareduction intheNMPtorusCOloadsisdocumented inthisreport.Thereduction wasrequested becausethinningoftheNMPtorusshellduetocorrosion impliesthatstresslevelsinducedbytheseDBAloadswouldexceedallowables.

Thetechnical basisutilizedinBNL'sreviewincludesaseriesoftopicalreportsprovidedbytheapplicant aswellasresponses toRAIsgenerated duringthecourseoftheevaluation.

Inaddition, theevaluation involvedreviewof.documents inwhichdevelopment oftheoriginalCOloadspecification anditsbasisaredescribed.

Alsofactoredintotheevaluation aretheresultsofindependent calculations performed byBNLtoconfirmtheadequacyoftheapplicant's analytical results.Finally,thereview'sscopewasexpandedtoincludetheimpactofshellthinningonallDBA-related hydrodynamic loads.BNL'sfindingsbasedontheaboveisthattherequested reduction inCOloadsisappropriate andhasasoundtechnical basis.

~I'~~'

TABLEOFCONTENTSSTRACT~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ABLISTOFFIGURES.................,...............

0Vl~~LISTOFTABLES.....................................,..........

vllLISTOFACRONYMSANDABBREVIATIONS

...,.....

~~~~~~~~~~~~VillACKNOWLEDGEMENT...........................................

lx

1.0INTRODUCTION

ANDBACKGROUND

.......................

12.0DESCRIPTION OFTHEPROPOSEDMETHODOLOGY

3.0 EVALUATION

OFTHEPROPOSEDMETHODOLOGY

......~~~~~~~~53.13.2Evaluation BasedontheInitialSublnittal...

Evaluation BasedonOtherConsiderations

.,3.2.1Acceptability ofTotalStructural ResponseMethod.........,.....

3.2.2BNLConfirmation ofGeometryEffectonCOBoundaryPressures

........3.221Description oftheMethodology

.3.Z2.2Geometric Considerations.....

3.2.Z3Presentation ofResults.......

3.22.4Discussion ofResults..........5~.5.6..6..7..7..8..94.0IMPACTOFSHELLTHINNINGONOTHERDBA-RELATED HYDRODYNAMIC LOADS.................................,

.115.0CONCLUDING REMARKS...............................,....

1

36.0REFERENCES

.............................................

14APPENDIXATHEBNLTECHNICAL LETTERREPORT...............

A-1APPENDIXBTHEBNLINTERNALMEMORANDUM

.................

B-1APPENDIXCTHEDESCRIBING EQUATIONS FORTHEBNLMETHODOFIMAGES..e~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~oC-1APPENDIXDTHENRCREQUESTFORADDITIONAL INFORMATION ANDTHENMPCRESPONSE~..................,...

D-1

LISTOFTABLESTable1.ValuesofParameters UsedtoDefineCalculation CellGeometry....17Table2.ResultsofBNLMOICalculations

...........................

18 y~v LISTOFACRONYMSANDABBREVIATIONS ACABSSBNLBWRBWROGCDICODBADNEFSTFIBALDRMOINEPNMPNMPCNRCPCPUARRAIRFSERSRSSSRVTERTESTLRTSAcceptance CriteriaAbsoluteSumBrookhaven NationalLaboratory BoilingWaterReactorBWROwnersGroupContinuum

Dynamics, Inc.Condensation Oscillation DesignBasisAccidentDepartment ofNuclearEnergyFullScaleTestFacilityIntermediate BreakAccidentLoadDefinition ReportMethodofImagesNon-Exceedance Probability NineMilePointNuclearStationUnit1NiagaraMohawkPowerCorporation NuclearRegulatory Commission PersonalComputerPlantUniqueAnalysisReportRequestforAdditional Information Reduction Factor(forpressureamplitude)

SafetyEvaluation ReportSquareRootoftheSumofSquaresSafety/Relief ValveTechnical Evaluation ReportTeledyneEngineering ServicesTechnical LetterReportTechnical Specification g~I ACKNOWLEDGEMENT Theauthorswouldliketoexpresstheirappreciation toA.D'Angelo, theNRCLeadEngineerforthisproject.Theguidanceandsuggestions heprovidedwereinsightful andconstructive.

Hiscontributions wereessential tothesuccessful completion oftheevaluation documented inthisreport.

ILJtCp

1.0INTRODUCTION

ANDBACKGROUND ThegenericCOloaddefinition anditsgenesisaredescribed intheMarkILDR.'twassynthesized frompressures recordedduringtheworstcaseblowdown(TestNumberMS)fromthefirstFSTFtestseries.'his testsimulated alargeliquidbreakbutwasconducted atapooltemperature belowthecurrentTechnical Specification (TS)forcontinuous operation (70'Fvs90-95'F).

TheseloadswereapprovedbytheNRC,subjecttotheresultsofadditional confirmatory tests.'ncreased pressures wereobservedintheselatertests'hich wereconducted athigherpooltemperatures (95'FforTestM12).However,theoriginalloadspecification wasdeemedacceptable'ased onafavorable comparison betweenpredictions andthestresslevelsobservedduringthehightemperature test,Insomecases,theprediction exceededmeasurements byasmuchas150%.Tounderstand whytheLDRloadsexhibitthisconservatism itisnecessary todescribehowtheloadspecification isderivedandhowitistobeapplied.Reference 6providesadetaileddescription ofthedesignload'sdevelopment.

Herewenoteonlythefollowing:

asinglepressuresignature wasselectedforprocessing (Figure2-6ofReference 6).AFourierseriesrepresentation ofthissignalwasthendeveloped.

Thiswasfollowedbyasomewhatcomplexprocedure thatconverted theFouriercoefficients tocorresponding "rigidwall"values.Fromthese,atableofrigidwallFouriercoefficients/pressure amplitudes wasgenerated asafunctionofdiscretefrequency bands(Table4.4.1-2ofReference 1).TheLDRthendirectsthattheseharmonicexcitations beapplied,withineachfrequency band,tostructural modelsthatrepresent eachplant-specific torus,toestablish thestructure's response.

Becausethedesignforcingfunctionhasbeendecomposed intoaseriesofdiscreteharmonicexcitations, awayofcombining thecorresponding stresseshastobeprescribed.

Wenotethatiftheexcitation wasgivenasasingle,continuous pressurewaveformas,forexample,isspecified fortheMarkISRVload(Section5.2.2oftheLDR),thisrequirement doesnotarise.Thus,tocompletetheCOloadspecifi...tion, theLDRrequiresthatthestressesbecombinedbytheABSSmethodwhichisthesimplestandmostconservative approach.

Thelargemarginsbetweenmeasuredandpredicted stressesnotedaboveareadirectresultofthisapproach.

Theexcessive conservatism ofthisapproachwasrecognized bytheBWROGevenbeforetheLDRloadswereappliedtospecificplants.Accordingly, aseriesofstudieswerecommissioned todevelopimprovedprocedures thatreducedtheexcessive marginsbutretainedanappropriate levelofconservatism.7@'ased onareviewofthesestudies,theNRCstaffagreedthatstrictapplication oftheABSSmethodwasnotrequiredandrelaxeditsoriginalAC,ForNMP,inparticular, amodifiedCOloadwas,approved duringreviewoftheirPUAR."Thismodification involvedapplication oftheABSSmethodtoonlythefourhighestharmonicresponses andaddition, byaSRSSmethod,oftheremaining ones.Notethatthisprocedure doesnotmodifytheforcingfunctionitselfwhichconsistsofthepressureloadsthatareappliedtothewettedtorusboundaries.

Therevisedmethoddoes, lrfP however,reducethetotalstressexperienced byanyparticular structural elementrelativetothatresulting fromfullABSSapplication.

FormostBWRplants,useoftheLDRspecified ABSSmethodwasacceptable despiteitsinherentconservatism.

InthecaseofNMP,however,theneedtoreducetheloadsaroseduetoNMP'snon-prototypically thintorusshell.Withthepassageoftime,therehasbeenafurtherreduction intheshellthickness duetocorrosion.

Thiscorrosion isacontinuing processwhichNMPCanditsconsultant estimateoccursatarateof0.00126inchesperyear."IftheCOloadsarenotreducedevenfurther,controlling stresslevelsareexpectedtoexceedallowables during1994.Todelaytheneedtostructurally reinforce thetorus,NMPChasproposedafurtherreduction intheloadspecification.

Thetechnical justification forthisreduction isdescribed andevaluated intheensuingsections.

~lJt14"

2.0 DESCRIPTION

OFTHEPROPOSEDMETHODOLOGY Theinformation suppliedbytheapplicant tojustifyaloadreduction wastransmitted inavarietyofformsandatvarioustimes.Thegiveandtakebetweenthesesubmittals andthestaffsresponses extendedoveraconsiderable chronological period.Inthissectionwewilldescribetheapplicant's methodanditsbasisinawaythatparallels thishistorical development.

ThemethodsproposedbyNMPCtodemonstrate thatareduction inCOloadsisjustified werefirstdescribed intwodocuments preparedbyaconsulting firm.'~'his

material, aswellasthatprovidedinReference 11,constituted theinitialsubmittal totheNRCstaff.Thekeyelementsoftheinformation suppliedtherewereasfollows:FSTFdataareusedtodemonstrate thatsignificant correlation oftheCOprocessattheexitoftheeightdowncomers occursonlyinthe5-6Hzfrequency rangeandthat,atotherfrequencies, theprocessanditscontribution toboundarypressures israndom.2.ItisnotedthattheFSTFgeometry, whichconsistsofasingle,torus-like baywitheightdowncomers (cf:Figure3.2-5ofReference 2),doesnotcorrectly simulatetheNMPtorussince,inthelatter,fourdowncomer baysalternate witheightdowncomer bays(cf:Figure3ofReference 13).Theconsequence ofthisgeometric featureisthattheFSTFpressures areexcessive forboththefourandeightdowncomer NMPbays.Thisistrueovertheentirefrequency rangeofthepressuresignature including thesynchronous 5-6Hzvalue.3.ItisalsonotedthattheFSTFdoesnotcorrectly simulateanactualMarkItorusbecauseoftherelatively rigidendcapswhichactasplanesofsymmetrybetweenadjacentbays.Inadditiontoimplyingthatadjacentbayshavethesamenumberofdowncomers astheFSTFasnotedabove,anotherconsequence ofthisgeometric featureisthatasynchronous contributions tothemeasuredpressures areamplified.

4.Anacousticmodelappliedtoanidealized versionoftheNMPtorus(horizontal cylinderhalffilledwithwater)isdeveloped andutilizedtoquantifytheeffectsenumerated above.Theresultsofthisanalysisarepresented asreduction factors(cf:Table1ofReference 13)thataretobeappliedtotheLDRpressureamplitudes (cf:Table4.4.1-2ofReference 1).Thesefactorsdependprimarily onbaygeometryandthenatureoftheCOprocess:i.e.:whetheritiscoherentorrandom.Thereduction factorsalsoexhibitaslightdependence onfrequency.

Foruncorrelated COtheirvaluesareabout60%and80%forthefourandeightdowncomer arrangements, respectively.

Thecorresponding valuesforcoherentCOare70%and95%%uo.Theseallrepresent bayaveragedvalues.

J~~I1Jz' 5.Thecorrelated reduction factorsaretobeappliedonlyforthe5to6Hzpressureamplitude (about3psia).Forthebalanceofthefrequency spectrumtheuncorrelated valuesaretobeutilized.

6.Theprocedure todevelopthestructural response(stresses) totherevisedhydrodynamic loadsisalsoaddressed.

Reference 13states(p.14)that"thestructural analysisshouldbeundertaken aspertheLoadDefinition Report".Theanalogous citationfromReference 11appearsonpage11.Itstates"total...stress wasdonebyaddingtheabsolutevalueofthefourhighestharmoniccontributors totheSRSScombination oftheothers...".

SincetheLDRdictatestheuseoftheABSSmethodforcombining

stresses, twocontradictory procedures forcombining stressesarespecified inthisoriginalsubmittal according towhatiscitedinItem6.Thus,inBNL'soriginalevaluation'4 thedistinction betweentheLDR'sABSSmethodandthealternative ofcombining onlythefourpeakresponses byABSSandtheremaining responses bySRSS'ashighlighted andtheacceptability oftheproposedmethodmadecontingent ontheassumption thattheABSSmethodwastobeused.ThispositioncarriedoverintotheSERissuedbytheNRC.'ollowing theissuanceoftheSER,theNMPCtookexception totherequirement thatABSSbeusedtocomputetotalstructural response."

Itclarified theambiguity impliedinItem6bystatingthattheintentwastoutilizethe4ABSS+SRSS methodaswasdoneintheiroriginalPUAR.'dditional information insupportofthisapproachwasalsoincludedinthissubmittal.

Adescription andevaluation ofthislaterinformation isincludedinSection3.2.1below.Insummary,therevisedmethodology consistsofasetofmuttipliers (Table1ofReference 13)thatareusedtoreducetheLDRpressureamplitudes (Table4.4.1-2ofReference 1).Allotheraspectsofthemethodareidentical tothoseusedintheoriginalNMPPUAR."Forconvcnicncc inthcensuingdiscussion, thismethodofcombining thcindividual harmonicresponses willbedenotedbythcacronym4ABSS+SRSS.

V~Cgl

3.0 EVALVATION

OFTHEPROPOSEDMETHODOLOGY 3.1Evaluation BasedontheInitialSubmittal Asindicated above,anevaluation basedontheinitialsubmittal wascompleted anddocumented viaaBNLTLRearlyin1992.AcopyofthisTLRisincludedinthisreportasAppendixA.Itwasfoundthattheproposedreduction was"reasonable, conservative andtechnically defensible".

Thebasisforthisconclusion restedprimarily onBNL'sconcurrence thattheLDRpressureloadswereoverlyconservative forthereasonscited(therandomness oftheexcitation formostoftheobservedfrequency spectrumandthegeometric differences betweentheFSTFandtheNMPtorus)andtheacceptability oftheanalytical procedure utilizedtodevelopanappropriate reduction.

However,theincorrect assumption thatitwasNMPC/YES's intenttodeveloptotalstressesviaanABSSmethodalsoplayedapartindeveloping theoverallfindinginthatitimpliedasourceofadditional conservatism.

Thispositionwasevenmoreemphatically statedintheNRCstaff'sSER.Finally,theTLRhighlighted thefactthatBNLdidnotcritically reviewimplementation oftheanalytical methodnoraccuracyofthenumerical resultsthatweregenerated.

Theconfirmatory analysispresented inSection3.2.2represents anindirectwayofevaluating thecorrectness oftheNMPCmethodandresults.Summarizing thissection,thefindingsfromtheinitialevaluation regarding thepressureamplitude reduction factorsremainqualitatively validbutrequireadditional confirmation oftheirquantitative acceptability.

Thisadditional requirement aswellasotherconsiderations thathaveevolvedsincetheissuanceofthestaff'sSER'saddressed inSection3.2.3.2Evaluation BasedonOtherConsiderations AsaresultofNMPC'sresponsetotheTER,furtherevaluation wasundertaken basedontheadditional information thatwassuppliedthereandinReferences 8,18,19,and20.Themainfocusofthisnewinitiative wastoestablish thesuitability ofusingthe4ABSS+SRSS methodincombination withthereducedpressureloads.However,becauseitcouldbeanticipated thatacceptance ofthiscombination inevitably wouldreduceexistingmargins,thestafffeltthatamorethoroughexamination ofthenewlydeveloped excitation wasappropriate.

Specifically, theNRCformallyrequested BNLtoexpandthescopeofitsefforttoincludeanindependent, confirmatory setofcalculations todemonstrate theloadreducingeffectofthegeometrydifferences citedearlier(ie:Items2and3listedinSection2.0).Forcompleteness, thedecisionwasalsomadetoincludeanexamination oftheimpactofreducedshellthickness ontheabilityoftheNMPtorustowithstand allotherhydrodynamic (ie:besidesCO)loads.Inthenextsub-section, theacceptability ofthemethodproposedtodeveloptorusstructural response(the4ABSS+SRSS method)isaddressed.

Then,theBNLmethodforestimating 1~lytt, theeffectofgeometryonpressureisdescribed andnumerical resultspresented.

Finally,BNL'sfindingsresulting fromexamination oftheNMPtorusstructural capability vis-a-vis allDBAhydrodynamic loadsarediscussed.

3.2.1Acceptability ofTotalStructural ResponseMethodInSection1.0itwasnotedthattheNRCstaff'soriginalACwererelaxedregarding useoftheLDRABSSmethodforcombining stresses.

Thebasisforaccepting alessconservative versionwasdocumented inanAugust1983BNLInternalMemorandum."

AcopyofthismemohasbeenincludedhereasAppendixB.Themethodapprovedtherewasintendedtobegenerically applicable toallMarkIplantsbuthasbeenutilizedbyrelatively fewutilities otherthanNMPC.Theevaluation wascarriedoutbythelateG.Bienkowski, ofPrinceton University actingasconsultant totheContainment SystemsGroupofBNL'sDNE.Itreviewedessentially thesamedocumentation NMPCsuppliedmorerecently.

Usingconventional, industryacceptedstatistical considerations, methodsweredeveloped theretoobtainimprovedagreement betweenmeasuredFSTFstructural responses (stresses, displacements, forces)andthosepredicted usingtheLDRharmonicpressure".

Commontoallthesemethodswasthenotionthatsomewhere betweenpureABSSandpureSRSSexistsawayofcombining theresponses inamorerealistic way.The"NavalSum"P~whichcombinesthetwohighestpeaksbyABSSandtheremainder bySRSS(2ABSS+SRSS) isoneexample.InReference 7therecommended procedure was3ABSS+SRSS implyinganon-exceedance probability (NEP)of84%%uo.Althoughimprovedagreement wasdemonstrated, someexceedances werefound,primarily intheareaofmembranestresses.

Toprovidesufficient conservatism toboundallthemeasuredresponses, itwasrecommended inReference 21thattheproposedmethodbemodifiedtothe4ABSS+SRSS methodthatwasacceptedbythestaffandapprovedforusebyNMPCintheNMPPUAR.Insummary,the4ABSS+SRSS methodthatNMPChasusedtodeveloptotalstructural responsetotheCOexcitation wasapprovedbythestaffearlier.Nothingthathastranspired sincethatapprovalwarrantswithdrawal ofthisapprovaland/ormodification oftheprocedure.

3.2.2BNLConfirmation ofGeometryEffectonCOBoundaryPressures Themethodology usedbyBNLtocomputeboundaryloadsonsimulated versionsoftheNMPtorusandtheFSTFduetoCOatdowncomer exitsisdescribed inthissection.Numerical resultsarealsopresented here.Theyincludecomparisons withcorresponding NMPCresultsandsensitivity studiesthatexhibitthedependence oftheloadsonkeygeometric parameters thatcharacterize theNMPsuppression pool.

)~lyi15,AqV<<P' 3.2.2.1Description oftheMethodology Themethodusedderivesfromanapplication oftheclassical MethodOfImages(MOI)technique.

Thetechnique isparticularly suitablefordescribing thehydrodynamic phenomenon occurring duringtheCOphaseofaDBAblowdown.

BNL'smethodisvirtually identical tothatemployedbytheGeneralElectricCo.toestimateramsheadrelated,SRVhydrodynamic loads(Section3.3.1,ofReference 23),Thesoledifference isthatarectangular arrayofimagesisusedbyBNLratherthanGE'sdiamondpattern.Thisisbecausecomputerstoragecapacityandexecution timeshaveimprovedconsiderably sincethen(1978).Thus,thegreaterefficiency providedbythediamondshapedarrayisunnecessary.

Wewereabletocarryoutthesecalculations onaPC(Gateway2000).Abriefdescription oftherelevantdescribing equations usedherearepresented inAppendixC.Itshouldbenotedthatthesegivethealgorithm fordeveloping thepressureatanypoint(x,y,z)duetotheexcitation inducedbya~siniedowncomer/source.

Tocomputethepressureduetomultiplesources,thecomputercodeloopsoverallsourcesandcombinesthepressurecontribution fromeacheitherbyABSSfor"correlated" pressureorbySRSSfor"uncorrelated" results,3.2.2.2Geometric Considerations ThegeometryoftheFSTFwasmodelledasasingle,rectangular parallelopiped withplatformXObyZOanddepthYO(seeFigure1).Thespecificvaluesusedfortheseparameters aregiveninTable1andweredeveloped usingtheinformation giveninReference 20asfollows:YOwastakenequaltotheFSTFtorusradius(aofReference 20);XO,corresponding tothecircumferential lengthoftheFSTFbay,wastakentobefourtimesthedowncomer pairspacing(lofReference 20);ZO,thelateralwidthofthecomputation cell,wasselectedsothatthecross-sectional areaofthecellsequaledthatoftheFSTF;ie:wetookZOsuchthat(YO)(ZO)=m(YO)~/2.

Fourpairsofsourceswithlateral/radial spacingDS,weresymmetrically locatedwithinthecelladistanceHOabovethetorusbottom.HOandDSderivefromthevaluesgivenforrand0inReference 20todefinethelocation/submergence ofthedowncomer exitplanes.Thissinglecomputational cellwasutilizedtodevelopestimates ofboththecorrelated anduncorrelated pressureloads.ThisisvalidfortheFSTFsince,asnotedearlier,therigidendcapsrepresent planesofsymmetrysothatasynchrony oftheCOpulsescanonlyoccuramongtheeightdowncomers contained withinthesinglecell.Modelling oftheNMPgeometrydifferedfromthatfortheFSTFbecauseoftheneedtocorrectly represent conditions whentheCOprocessisasynchronous.

Incontrasttothesituation fortheFSTF,whenthiscondition prevailsintheNMPtorusitimpliesthattheCOpulsesatall120downcomers (10bayswith4pairs;10bayswith2pairs)areoutofphaseratherthanjustatthefouroreightlocatedinasinglebay.Theloadreduction thatwouldresultfromsuchalimitednumberofuncorrelated sourceswouldbeunrealistic.

I~

Inviewoftheabove,twotypesofgeometrywereemployedfortheNMPsimulations.

Forthecorrelated case,asinglecomputational cellwasemployedanalogous tothatusedfortheFSTFexceptforthenumberandlocationofthesources/downcomers.

Thesewerearrangedsothatthecalculation cellextendedfromthecenter(hence,planeofsymmetry) ofanon-vent,eightdowncomer baytothecenterofafourdowncomer bay.Thus,onlythreepairsofsourceswereusedforthissimulation.

Allotherpertinent dimensions forthesingle,NMPcomputational cellaregiveninTable1.Thesevaluesalsoderivefromtheinformation giveninReference 20.Referring toTable1,itisinteresting tonotethattheFSTFandNMPgeometries arecomparable exceptfordowncomer pairclearance (HO)andspacing(DS).Ascanbeseen,theNMPdowncomer exitsaresignificantly closertoeachotherandtothebottomofthetorus.Thesegeometric differences haveasignificant impactontheboundarypressures aswillbediscussed inSection3.2.2.4.Forthecaseofuncorrelated sourcesintheNMP,thegeometrymustreflectthefactthatincoherence betweendowncomers isnotlimitedtothoseresidentinasingleoreveninapairofbays.Solongasrigidwallsarespecified attheendsoftheselectedcalculation cell,coherence betweenthesourcesinthatcellandthearrayofimagesthatareemployedbytheMOIisimposed.Unlessthiseffectisproperlyaccounted for,misleading resultscanbeobtained.

Thisisaccomplished herebymodelingtheNMPtorusasrealistically aspossiblewithrespecttothetotalnumberofdowncomers, Asnotedabove,forNMPthisnumberis120.Ourmodelling hasutilizedhalfthisnumberwhichwouldyieldconservative results;i.e.:thepressureatanyparticular spatiallocationdecreases astherigidboundaries withinwhichincreasing numbersofuncorrelated sourcesareembeddedrecedefromthatlocation.

Insummary,twotypesofgeometries areemployedinBNL'scalculations.

Forcorrelated pressures, asinglecellinwhicheitherfourorthreepairsofsources/downcomers arelocatedisused.Foruncorrelated pressures corresponding toNMP,thecellextendsinthecircumferential (X)direction approximately 180feetcorresponding to10bays.'ach ofthesecellshasthreepairsofsourcesclustered insuchawaythatthealternating 8-4-8-4patternintheNMPisreproduced (cf:Figure3ofReference 13)~TheoriginoftheXcoordinate isatthecenterofoneortheotherofthesebays(bothcalculations weremadewithnosignificant difference found)corresponding toaplaneofsymmetry.

ForFSTFpressures, thecalculation cellforbothcorrelated anduncorrelated resultsisidentical.

Thisisconsistent withtheactualgeometric configuration ofthefacilityandcorrectly modelsthepresenceoftherigidwalls.3.2.2.3Presentation ofResultsItwasnotedinSection2.0thattheNMPCmethodultimately involvesreduction oftheLDRCOpressureamplitudes bywhatarereferredtoinReference 13as"Harmonic Amplitude LoadReduction Factors".

Thecalculations performed hereprovideanalogous reduction factorsbygenerating boundarypressures forthemodeledFSTFduetoaunitexcitation ateachsourceandformingtheratiowiththecorresponding valuesobtainedwhenidentical strengthsourcesarelocatedinasimulated NMPtorusgeometry.

4' ResultsofBNL'scalculations aregiveninthreedistinctways.First,themaximumpressurecomputedwithinagivencomputational cell(PMAX)istabulated asinTable2.Thetableincludestheresultsofasensitivity studywherekeygeometric parameters havebeenvariedfromthebasecasevaluesgiveninTable1.Byfocusingonthesepeakvalues,improvedinsightregarding thetrendsassociated withchangesingeometryisprovided.

Inadditiontotabulating PMAX,thespatialvariation ofpressureatthebottomofthecellattheverticalplaneofsymmetrybetweenpairsofdowncomers hasbeengenerated andisplottedinFigure1,Notethatthisfigureincludesanindication ofthecalculation cellgeometryasithasbeensimulated here.Figure2comparestheNMPCreduction factors(RF)withthosedeveloped byBNL.ThelatterderivefromtheresultsshowninFigure1byformingtheappropriate ratios.Finally,graphical representation ofthefindingsfromthesensitivity studiesisshowninFigure3.3.2.2.4Discussion ofResultsFromtheperspective ofjustifying aloadreduction forNMPrelativetotheloadsderivedfromFSTFtests,thekeyfindingisthecomparison betweenthevaluesofPMAXobtainedforCasesN1andF1forcorrelated resultsandN1(U)andF1(U)foruncorrelated COsources,Thereduction factor(RF)impliedbythefirstoftheseisessentially unity;thatis,PMAX=1.45forbothCaseF1andCaseN1asindicated inTable2.Foruncorrelated sources,RF=0.76sincePMAX=0.42forCaseN1(U)andPMAX=0.55forCaseF1(U).Thecorresponding valuesproposedbyNMPC(fromTable1ofReference 13)are0.98and0.83.Weconsiderthistobereasonable agreement particularly whenthecomparison showninFigure2isalsofactoredin,Referring nowtoFigure1,themostsignificant oftheresultsshownthereisthelargereduction inpressurethatisobtainedwhenthesourcesarenolongercorrelated.

EvenfortheFSTF,thepressures arereducedtoonlyaboutonethirdofthecorrelated values.Interestingly enough,thisismoreorlesstheArderofmagnitude ofreduction inpressureamplitude betweenthefundamental frequency (about3psifor5Hz)andtheothernon-synchronous values(1psimaximum)asindicated inTable4.4.1-2ofReference 1.Comparison oftheBNLandNMPCvaluesoftheRFsshowninFigure2indicateclearlythattheyareinverygoodagreement.

Noteespecially thattheNMPC'suncorrelated RFsaremoreconservative thanBNL's.Thatis,theLDRpressureamplitudes arereducedlesswhentheNMPCRF'sareused.Somenonconservatism isexhibited forcorrelated RF'sbutthisdifference is,atmost,6%.Thesedifferences areconsidered minorand,inourjudgement, donotinvalidate theacceptability oftheproposedmodifications.

Withrespecttothesensitivity studiesthatwereperformed, wenotefirstthattheyweremotivated bytheresultobtainedforCaseN2ofTable2corresponding toacomputational cellwiththeNMPgeometrybutwithfourratherthanthreepairsofdowncomers.

Although Jtr~

thiscasedoesnothavedirectapplicability here,itwasperformed outofacademicinterestandforthesakeofcompleteness.

Ascanbeseenbyreferring toTable2,thevalueofPMAXcorresponding toCaseN2(1.66)notonlyexceedsthatforCaseN1(1.45),whichistobeexpected, butalsoexceedsthevaluepredicted forthebasicFSTFcase(CaseF1).Thisincreasecanonlybeattributed togeometric differences sincesourcenumberandstrengthareidentical forthosetwocases.Thesamesortofdifference isexhibited betweentheCaseF2(PMAX=1.24)andtheCaseN1(PMAX=1.45)results.Notethattheselattertwocasescorrespond to8-4-8-4typeconfigurations.

Thesefindingswerethemotivation forthesensitivity studiesthatwereconducted; ie:todetermine whatfeatureoftheNMPtorusgeometrygivesrisetopressureloadshigherthanthoseexpectedintheFSTFforthesamenumberofdowncomers.

AscanbeseenfromtheresultsshowninFigure3,thenotedincreases areprimarily theresultofthesignificantly smallerclearance (7vs7.8ft)thatprevailsintheNMPtorus.Oneimplication ofthisfindingisthattheFSTFgeometrywasnotstrictlyapplicable forNMPbothwithrespecttothelatter's8-4-8downcomer arrangement (aconservatism),

butalsowithregardtodowncomer clearance (anon-conservatism)

~Thetrade-off betweenthesetwoopposingeffectssuggeststhattheoriginalNMPdesignloadsweresuitable, notwithstanding that,totheauthors'nowledge, noadjustment fortheeffectofreducedclearance wasmadeorconsidered.

Amorepositiveinterpretation ofthisfindingwouldbethatCOloadscouldbereducedbyanincreaseinclearance atNMP;ie:byshortening thelengthofthedowncomers.

According toTable2(CaseN8vsCaseN1),adecreaseinPMAXofalmost20%couldbeachievedbyremoving12inchesfromthedowncomer ends.Ofcourse,suchamodification wouldinvolvesignificant expensebutmightbeacost-effective alternative tothestructural modifications currently underconsideration byNMPCintheeventtheybecomenecessary.

Notethatareduction ofHOimpliesacorresponding reduction indowncomer submergence whichtendstomitigatepoolswellloads(Section3ofReference 3)~However,italsopotentially reducesthesteamcondensing performance ofthesuppression pool.Thusatotalsystemanalysiswouldberequiredtodetermine themeritofthisconcept.10 I

4.4IMPACTOFSHELLTHINNINMi&RODYNAMIC LOADS6ONOTHERDBA-RELATED Theapproachherewastore-review theNMPCPUARfortheNMPToruswiththefocusonhowthinningofthetorusshellcouldpotentially effecttheearlierevaluation.~

Asaresultofthisreview,anRAIwasdeveloped andtransmitted toNMPC.~NMPC'sresponsetothisRAIwasprovidedtotheNRCstaffvialetterdatedSeptember 28,1993.BNLreceivedandreviewedthisinformation inOctoberof1993.Itsfindings" werethat"itwasresponsive andcomplete.

Noopenissuesorconcernsrelatedtothissubmittal were'dentified."

ThecontentoftheRAIandNMPC'sresponseareincludedinthisreportasAppendixD.Thegeneralthrustofthequestions posedintheRAIwastoconfirmthatsufficient marginwasavailable toaccommodate theincreaseinstresslevelsassociated withreducedshellthickness foralltheloadcasesmandatedbytheNRCstaff'sAC.Atotalof27suchloadcasesarespecified asenumerated inFigures4.3-1,4.3-2and4.3-3ofReference 3.TheRAIalsorequested thatthecontribution tostresslevelduetoeachevent(eg:theDBACOload)beitemizedtodetermine theirrelativeimportance.

Examplesoftheinformation obtainedinthiswayissummarized below.InresponsetothefirstRAIquestion, theapplicant indicated thatforEventCombination 20(thecontrolling loadcase)onlyabout30%ofthetotalstressisduetotheCOload.Mostofthestress(almost60%)stemsfromtheDBAinternalpressureload.Thus,anyreduction intotalstresscanonlybeasmallfractionofthecorresponding reduction inCOload.Thisclarifies theseemingly anomalous resultthatupto60%reductions inthelatterleadto,atmost,a10%reduction intotalmembranestressasreportedinNMPC'soriginalsubmittal."

Anotherquestionaskedthattheloadcombination inducingthenexthigheststressbeidentified.

Theresponseindicated thistobeEventCombination 14,Thiseventcombines, amongotherloads,thosestemmingfromSRVactuation andtheIBACOload."Considerable reduction intotalstressrelativetoEventCombination 20isreported(from16ksitoabout13ksi)withtheinternalpressurebeingevenmoredominant(almost75%ofthetotal).TheIBACOloadcontribution isonly6%,anamountequaltothatfromSRVactuation, asurprisingly smallvalue.Insummary,theresponses indicatethatthemarginbetweenexpectedandallowable stresslevelsforallotherEventCombinations aremuchgreaterthanforthatcitedasthelimitingcase.Basedonthisinformation, itcanbeconcluded thatthecapability oftheNMPtorusItisimportant tonotethattheoriginalIBACOloadisusedherc.Themodification rcriucstcd byNMPCappliesonlytotheDBACOload.11 I4 tomaintainitsintegrity duringpostulated DBAeventsisassuredprovidedthatthisisdemonstrated forEventCombination 20.Finally,duringreviewoftheNMPtorusPUAR,aconcernrelatingtothewayinwhichtheCOloadsareappliedtodeveloptheringgirderstructural response(Section5.0ofReference 24)wasidentified.

Thisconcernarisesduetotheasymmetry introduced bythe8-4-8downcomer arrangement.

Specifically, thePUARstatesthatthehalfbaystructural modeloftheNMPtorus(Figure3-3ofReference 24)isusedfortheringgirderresponsefor"allshelldynamicloads."SincetheaverageCOpressureamplitudes thatareappliedinalternate baysdifferbymorethan20%,thequestionofhowthegradientacrosstheringgirderwasaccommodated arose.Theissuewasraisedandresolvedviatelephone conference withNMPCpersonnel andconsultants.

Firstitwasestablished thattheissuehadnotbeenaddressed.

Toresolvetheissue,theapplicant committed toutilizing aboundingapproach~

whereinthehigher,non-ventbayloadswouldbeappliedonbothsidesoftheringgirdertodevelopthestresseswithintheregionimmediately adjacenttothetorusmiterjointandmiteroffset.Sincethisloadingcreatesthehighestbendingmomentacrosswhatis,ineffectarigidconnection, thestructural responsewillbemaximized.

Weconsiderthisaconservative and,therefore, acceptable approach.

12 I4

5.0 CONCLUDING

REMARKSBasedontheevaluation documented here,BNLconcludes thatthereduction intheDBACOloadsthathasbeenrequested byNMPCisappropriate andtechnically justified.

Thegeometric restraints imposedbytheFSTFfromwhichtheLDRloadsderivedidintroduce conservatisms thatcansafelybereduced.Theabsenceofcoherence formostofthefrequency spectrumisalsoclear.TheCDIanalysisbasedonacoustics represents arationalprocedure forestimating andquantifying theseeffects.Ourindependent calculations confirmthatthisanalysiswascorrectly implemented.

Thesoleconcernthatwewouldhighlight hererelatestoBWRplantoperating procedures/technical specifications/emergency operating procedures.

Specifically, itwasnotedintheintroductory remarksthattheFSTFtestresultsexhibited anincreasing trendoftheCOloadswithincreasing pooltemperature.

AlthoughtheLDRloadsandtheNMPmodifiedversioncanaccommodate theobservedincrease, anyfurtherincreaseintheTSfortheinitiation ofsuppression poolcooling~canpotentially invalidate theiracceptability.

Inthisconnection wenotethatarequesttopermitasubstantial increaseinthisTSisnowbeingconsidered bytheNRCstaff."Wewanttoemphasize heretheneedtokeeptheconnection betweenDBAloadsandplantoperating conditions intheforefront whenconsidering anyfurthermodifications tocurrently acceptable designhydrodynamic loads.Additional evaluation and/oraugmentation oftheexistingsuppression poolhydrodynamic databasetogetherwithadditional analysiscouldverywellbeneededtoprovidesoundjustification forsuchmodifications.

13 I4

6.0REFERENCES

GeneralElectricCompany,"MarkIContainment ProgramLoadDefinition Report,"GE TopicalReportNEDO-21888, Revision2,November1981.2.Fitzsimmons, G.W.,etal.,"MarkIContainment ProgramFull-Scale TestProgramFinalReport,TaskNo.5.11,"GEProprietary ReportNEDE-24539-P, April1979.3.U.S.NuclearRegulatory Commission, "SafetyEvaluation Report,MarkILongTermProgram,Resolution ofGenericTechnical ActivityA-7,"NUREG-0661, July1980.GeneralElectricCompany,"MarkIContainment ProgramLetterReport:Supplemental Full-Scale Condensation TestResultsandLoadConfirmation,"

MI-LR-81-01-P, April1981.U.S.NuclearRegulatory Commission, "SafetyEvaluation Report,MarkILongTermProgram,Resolution ofGenericTechnical ActivityA-7,"NUREG-0661, Supplement 1,August1982.6.GeneralElectricCompany,"MarkIContainment ProgramAnalysisofFullScaleTestFacilityforCondensation Oscillation Loading,"

GEReportNEDE-24645, July1979.7.GeneralElectricCompany,"MarkIContainment ProgramEvaluation ofHarmonicPhasingforMarkITorusShellCondensation Oscillation Loads,"GEReportNEDE-24840(prepared byStructural Mechanics Associates),

October1980.8.Kennedy,R.P.,"Response FactorsAppropriate forUsewithCOHarmonicResponseCombination DesignRules,"SMAReport12101.04-R002D, preparedbyStructural Mechanics Associates forGeneralElectric, March1982.9.Kennedy,R.P.,"AStatistical BasisforLoadFactorsforUsewithCOHarmonicResponseCombination DesignRules,"SMAReport12101.04-R003D, preparedbyStructural Mechanics Associates forGeneralElectric, March1982.10.Bienkowski, G.,Lehner,J.R.,andEconomos, C.,"Technical Evaluation oftheNineMilePointUnit1NuclearGenerating StationPlantUniqueAnalysisReport,"BNL-04243,September 1984.TeledyneEngineering Service,"NineMilePointUnit1Reduction inMarkITorusProgramCondensation Oscillation LoadDefinition andResulting EffectonMinimumShellThickness Requirements,"

TESTR-7353-1, Revision1,April1991.14 lI~,

12.Bliss,D.B.andTeske,M.E.,"FSTFShellCondensation Oscillation Correction Factors-Uncorrelated Vents,"CDIReport79-1,Revision2,preparedbyContinumDynamics, Inc.forGeneralElectric, August1980.13.Continuum

Dynamics, Inc.,"Reduction ofTorusShellCondensation Oscillation Hydrodynamic LoadsforNineMilePointUnit1,"CDITechnical Note90-11,preparedforTeledyneEngineering

Services, November1990.14.Economos, C.,Lehner,J.andLin,C.C.,"Evaluation ofNMCTechnical BasisforReduction ofNMPTorusCOLoads,"BNLTechnical LetterReport,February1992.15.U.S.NuclearRegulatory Commission, "SafetyEvaluation bytheOfficeofNuclearRegulation RelatedtoProposedDeferment ofTorusModifications NiagaraMohawkPowerCorporation NineMilepointNuclearStationUnitNo.1,DocketNo.50-220,"August1992.16.TeledyneEngineering

Services, "Technical ResponsetoSafetyEvaluation bytheOfficeofNuclearRegulation RelatedtoProposedDeferment ofTorusModifications NiagaraMohawkPowerCorporation NineMilepointNuclearStationUnitNo.1,DocketNo.50-220,"November1992.17.TeledyneEngineering

Services, "MarkIContainment Program,Plant-Unique AnalysisReportoftheTorusSuppression ChamberforNineMilePointUnit1NuclearGenerating Station",

TESTR-5320-1, Revision1,OctoberSeptember 1984.18.NiagaraMohawkPowerCorporation, "Reduction inMarkITorusProgramCondensation Oscillation LoadDefinition andResulting EffectonMinimumShellThickness Requirements,"

Presentation toNRC,March23,1993.19.Kennedy,R.P.,etal,"StudytoDemonstrate theSRSSCombinedResponsehasgreaterthan84PercentNonexceedance Probability WhentheNewmark-Kennedy Acceptance CriteriaareSatisfied,"

GEReportNEDO-24010-03, Supplement 3,August1979.20.Bilanin,A.J.,"NRCRequestforDocumentation oftheNineMileCondensation Oscillation AcousticTorusLoadReduction AnalysisLimits,"Continuum DynamicsLetterReportNo.TELEDYNE/0073, April1993.21.Bienkowski, G.,"ReviewoftheValidityofRandomPhasingRulesasAppliedtoCOTorusLoads",InternalBNLMemorandum, August1983.22.Scanlon,R.,Professor

Emeritus, Princeton University, PersonalCommunication.

15 ItkAll'I 23.GeneralElectricCompany,"MarkIIContainment DynamicForcingFunctions Information Report,"GEReportNEDO-21061, Revision3,June1978.24.TeledyneEngineering

Services, "PlantUniqueAnalysisReportoftheTorusSuppression ChamberforNineMilePointUnit1NuclearGenerating Station,"

TESTechnical ReportTR-5320-1, October1983.25.Triolo,S.andLe,A.,"AuditforMarkIContainment LongTermProgram-Structural AnalysisforOperating

Reactors, NiagaraMohawkPowerCorporation NineMilePointNuclearStationUnit1",FranklinResearchCenterTechnical Evaluation ReportTER-C5506-331, September 1984.26.U.S.NuclearRegulatory Commission, "RequestforAdditional Information Regarding Re-Review ofNineMilePointNuclearStationUnit1TorusLoadReduction Submittal ofMay14,1991.",August1993.27.TeledyneEngineering

Services, "Response totheNRCRequestforAdditional Information Regarding Re-Review ofMay14,1991TorusLoadReduction Submittal NiagaraMohawkPowerCorporation NineMilePointNuclearStationUnit1,DocketNo.50-220",TESLetterReport7519-28,September 1993.28.Brookhaven NationalLaboratory, "MonthlyBusinessReportforOctober1993-Nine MilePoint,Unit1:Suppression PoolDynamicLoadsRevisedMethodology."

29.Attachment 1toNMPCLetterDatedNovember30,1993fromC.D.Terry,NMPCtoUSNRCDocumentControlDeskUnderDocketNo.50-220.

30.Mintz,S.,"BWRSuppression PoolTemperature Technical Specification Limits.",

GeneralElectricReportNEDO-31695.

31.U.S.NuclearRegulatory Commission, "StandardTechnical Specifications GeneralElectricPlant,BWR/4,"NUREG-1433, January1991.16 f/

Table1ValuesofParameters UsedtoDefineCalculation CellGeometryITEMXOYOZODSHOFSTFVALUE19.513.821.78.07.84.956.37.2NMPVALUE19.6(176.4)13.521.26.07.04.965.27.2NOMENCLATURE DSHOlNDPMAXp/XOYOZOeUNITSSPACINGBETWEENDOWNCOMER PAIRSINTHEZDIRECTION VERTICALDISTANCEBETWEENTORUSBOTTOMANDDOWNCOMER EXITSPACINGBETWEENDOWNCOMER PAIRSINTHEXDIRECTION NUMBEROFDOWNCOMER PAIRS(SEETABLE2)MAXIMUMPRESSUREIN8DOWNCOMER BAYINFLUENCE COEFFICIENTS

-dPMAX/dDS, dPMAX/dYO, etc.RADIALDISTANCETORUSCENTERTODOWNCOMER EXITCENTERSIMULATED LENGTHOFTORUSBAYDEPTHOFSUPPRESSION POOL=TORUSRADIUSSIMULATED TORUSDIAMETERPOLARCOORDINATE ANGLEMEASUREDFROMHORIZONTAL TODOWNCOMER EXITCENTERPMAX-SOURCEUNITS(SU)P-SU/FTDS,HO,I,r,XO,YO,ZO-FEET8-DEGREES17 I

Table2ResultsofBNLMOICalculations CASEF1F1(U)F2YO13.8ZO21.7ND8.0HO7.8PMAX1.450.551.24N1(U)N2N3N4N5N6N7N8N9N1013.512.514.513.521.2176.43021.224.027.06.07.08.06.07.06.08.07.01.450.421.661.251,631.421.391.701.211.341.27Notes:SeeTable1fornomenclature andunits.Allresultsareforcorrelated sourcesunlessotherwise indicated bythenotation(U)following theCaseIdentifier Number.

I III1~NonventBayA2XOVentBay(NMP)DS:FSTFIPjO+O+O++,'ypical Computation Cell+a+++,s+,'I,--,IIIOnly0+Y=Z=Ox=z=o2.0P'.an~HO~YOSectionA-A1.61.2~CaseN2~CaseF1~CaseN1P0.80.4~CaseF1(U)~CaseN1(U)0.000.51.0X/XO2.0Figure1-BNLMethodofImagesPredicted Variation ofPressureintheComputation CellPlaneofSymmetry-P (X,O,ZO/2) 19!II l4

'.0BNLRF080.6Correlated NMPC////r/X/XO1.0RF0.80.6NMPC~\~srBNLUncorrelated rr//I/////0.40.00.51.0X/XO1.52.0Figure2-Comparison ofBNLandNMPCEstimates forPressureAmplitude Reduction Factors20 IJC 2.0PMAX1.00.0125~ppWP'=0.1914YO2.01.00.0P'=-0.037DS2.02.0PMAX1.01.0~~ppP'=-0.25P'=-0.030.07HO0.02024ZO28Figure3-ResultsofSensitivity StudiesforNMPviatheBNLMethodofImagesSolidSymbolsDenoteCaseN1Parameters SeeTable1forNomenclature andUnits21

I4A-1 4

a~~Evaluation ofNMCTechnical Basisfor~Reduction ofNMPTorusCOLoadsC.Economos, J.Lehner,andC.C.Linjanuary1992RevisedFebruary1992~r~SummaBNL'sevaluation ofthetechnical basissubmitted byNMCtojustifyareduction intheNMPTorusCOloadsisdocumented viathisletter.report.Theevaluation includesareviewofthehistorical developments thatprecededthecurrentsubmittal.

Thesearepertinent becausetheyrepresent thepointofdeparture fortheproposedmodiGcations.

BNL'sGndingisthatthemethodology usedtodemonstrate thatareduction intheseloadsisappropriate is,technically soundandjustiGestherequested modiGcations.

BackroundThegenericCOloaddeGnition anditsgenesisaredescribed intheMarkILDR'.Itwassynthesized fromthepressures recordedduringtheworstcaseblowdown(TestNumberM8)fromtheGrstFSTFtestseries'.Thistestsimulated alargeliquidbreakbutwasconducted attherelatively lowpooltemperature of70'F,avalueless'hanthecurrentTechnical SpeciQcation (TS)forcontinuous operation (theLCO).TheseloadswereapprovedbytheNRC,subjecttotheresultsofaddi~anal conGrmatory tests3.Thepressures observedintheselatertests'ere higherforliquidblowdowns conducted atsomewhathigherpooltemperatures.

SpeciGcally, TestNumberM12,conducted ataninitialpooltemperature of95F,gaverisetopressures thatwereabout15%higherthanpeakM8values.Notethatthistemperature levelisroughl~eual tothecurrentTSontheLCO(90to95'F)andissomewhatlessIt)anthemohiGedValueof100FthattheBWROGhasrequested theNRCtoapprove.Notwithstanding theincreased loadsobsertj'ed duringTestM12,theoriginalloads'peciQcation wasWundacceptable6 basedonafavorable comparison betweenthemeasuredand'.predicted stressNvelsfortheFSTF.Insomecases,theprediction exceededmeasurements byasmuch.as150%%uo'.

I'heconservatism oftheLDRloadspeciGcation stemsprimarily fromtherequirement thatalloftheharmoniccomponent responses beaddedbyabsolutesum.Thisisequivalent toassumingthattheexcitation createdbyoscillation ofthesteam-water interface attheendofeachoftheeightdowncomers issynchronized overtheentirefrequency rangethatwasobserved(upto50Hz).Thestaffrecognized thatthisapproachisconservative andrelaxedtheACbasedonseverallaterstudiessubmitted byGEanditsconsultants "9.ForNMP,inparticular, amodiQedCOloadwasapprovedduringreviewoftheirPUAR'.Thismodification accounted fortheabsenceofcompletecorrelation betweenventsbytakingtheabsolutesumofonlythefourhighestharmonicresponses andaddingtheSRSSof'the~~'SeeListofAbbreviations fordetinition o(neronyms.

A-2 II

~remaining ones.Notethatthisprocedure reducescriticalstressesbutdoesnotexplicitly changetheforcingfunctionitselfwhichconsistsofthepressureloadingonthesubmerged boundaries.

Thebasisforapproving thisapproachwasthatitstillboundedthemeasuredresponsewhenappliedtotheFSTF..WhenappliedtoNMP,thecriticalstressesintheshellremainedbelowallowables.

l/TheneedtoreducetheCOloadsbelowthegenericLDRvaluesarose'because ofNMP'sthintorusshell.Withthepassageoftime,therehasbeenafurtherreduction intheshellthickness duetocorrosion.

Thisreduction isacontinuing processwhichNMCanditsconsultant estimateoccursatarateof.00126inchesperyear".IftheCOloadsarenotchanged,criticalstresslevelsareexpectedtoexceedallowables during1994.Todelaytheneedtostructurally reinforce thetorus,NMChasproposedareduction intheloadspeciGcation.

Thetechnical justiGcation forthisreduction isdescribed andevaluated intheensuingsections.

DescritionoftheProosedMethodolo ThemethodsproposedbyNMCtodemonstrate thatareduction inCOloadsisjustiGedaredescribed intwodocuments preparedbyaconsulting Grm'~".Keyelementsofthepresentation areasfollows:1.FSTFtestdataareusedtodemonstrate thatsigniGcant correlation oftheCOprocessattheexitoftheeightdowncomers occursonlyinthe5-6Hzfrequency rangeandthat,atotherfrequencies, theprocessanditscontribution tothepressuresignature israndom.2.3.ItisnotedthattheFSTFtestfacilityisnotprototypical ofanactualMarkItorusbecauseoftheendcapswhich'act asplanesofsymmetrybetweenadjacentbays.Itisclaimedthattheconsequence ofthisgeometric featureistiiattheincoherent contributions totheobservedpressures areampliGed.

Itisfurther-notedthattheFSTFfacilityisalsonotprototypical oftheNMPtorussince,inthelatter,fourdowncomer baysalternate witheightdowncomer bays".Inthiscaseitisclaimedthattliisgeometric discrepancy impliesthatthe'FSTFpressures areexcessive forboththefourandeightdowncomer NMPbays,andthatthisistrueovertheentirefrequency rangeincluding thesynchronous 5-6Hzvalue.4.Anacousticmodelappliedtoanidealized versionoftheNMPtorus(horizontal cylinderhalffilledwithwater)isdeveloped andutilizedtoquantifytheeffectsenumerated above.Theresultsfrom'this analysisarepresented asreduction factors'hat aretobeappliedtotheLDRpressureamplitudes".

Thesefactorsdependprimarily onbaygeometryandthenatureoftheCOprocess,ie.,coherentorrandom.Theyalsoexhibitaslightdependence onfrequency.

Thereduction factors're about60%forthefourdowncomer geometiyand80%fortheeightdowncomer bayconGguration foruncorrelated CO,Thecorresponding factorsforthecorrelated caseareapproximately 70%and95%,-respectively.

Theserepresent bayaveragedvalues.'Theterm"reduction factor"isusedhereandinReference 12toindicateamultiplier oftheoriginalvalue.A-3 lIt'Ay'4II5'

~5.Correlated reduction factorsaretobeappliedonlytothe5-6Hzpressureamplitude.

Fortheremaining frequency

spectrum, uncorrelated valuesaretobeutilized.

AftertheLDRpressures arereducedbythesefactors,thestructural analysisisto"be.undertaken aspertheLDR.""/~Withrespecttotheoriginalanalysis",

theseprocedures yielda4%%uoreduction ofthecontrolling stress(membrane) foraneightdowncomer bayanda10%%uoreduction forafourdowncomer bay".Intermsofshellthickness, thesecorrespond toreductions of16and44mils,respectively.

Thecorresponding valuesgiveninamorerecentsubmittal're 18and37mils.Itisstatedthere,thatthesecorrespond toa17%and30%reduction intheLDRCOloads,respectively.

Evaluation oftheProosedMethodolo InBNL'sjudgement, thereduction intheCOloadsthatNMChasrequested are,reasonable, conservative, andtechnically defensible.

Thebasisforthisconclusion areasfollows:1.TheFSTFdatasupportthenotionthattheCOprocessisrandomovermostofthefrequency spectrumconsidered intheloadmethods,2.Becauseofthegeometric differences, particularly the4-8-4downcomer arrangement, thepressureloadsduringaCOblowdownwilltendtobegreaterintheFSTFrelativetotheNMPtorusforthesamethermodynamic Qowconditions.

3.Theprocedure usedtoquantifytheeffectofItems1and2represents astraightforward application ofaconventional hydrodynamic method.Theresultsarereasonable andprobablyconservative becauseofthehighsoundspeedusedinthenumerics.

Wealsoconsidertheassumption thatacorrelation existsbetweenbaystobeasigniQcant conservatism.

4.Theoverallreduction oftheloadsfromLDRvaluesissigniQcantly lessthanthatapprovedearlierbythestafP.Thisreduction wasfoundacceptable becauseitwasabletoaccommodate allofthestressesobservedduringtheFSTFtests.Concludin RemarksTherearethreepointswewanttoemphasize here.Thefirstisthattheprocedure wehaveevaluated represents amorerigorous, almostQrstprinciples way,toaccomplish whatwasdonebeforeinanapproximate way.Aswealreadynotedinourbackground discussion, themodification thatwasutilizedbyNMPearlierdidnotinvolveanychangeintheLDRpressures.

ReliefwasobtainedbynotsummingthestressesinducedbyeachandeveryoneoftheQftyharmonicexcitations byabsolutesumasrequiredbytheLDRmethodology.

Thatthiswasanacceptable procedure couldonlybedemonstrated bycomparing predicted FSTFPe~=Modelingofthetorusashalffilledwithwaterisaminornonconservatism (NWLinMarkIplantsiswellbelowthetoruscenterline),

bui.isareasonable simplification ofananalysiswhichisalreadyquitecomplex.A-4

'lld>>;,'

~stresseswithmeasuredFSTFstresses.

Indistinctcontrast, thepresentmethodprovidesreliefbyreducingtheexcitation (pressures) itself.ThesecondpointisthatthebasisforItem4restsonourassumption thatwhentheapplicant refersto"LDRvalues"whatismeantarethestressesthatresultbyapplyingthe,LDRpressureamplitudes andthencombining alloftheindividual peakstressesbyabsolutesum.Thedocuments thatwehaveinhandaresomewhatambiguous onthispointanditwould.beprudenttoobtaindocumented conGrmation thatourinterpretation iscorrect.Finally,wenotethatourreviewoftheanalysisdoesnotincludedirectconfirmation ofanyofthenumerical resultsthatwerepresented, e.g.,the:reduction factors.Itisassumedthatthesederivefromacorrectapplication ofthemethodology.

A-5

'IIIS.r References 1.GeneralElectricCompany,"MarkIContainment ProgramLoadDefinition Report,"GeneralElectricTopicalReportNEDO-21888, Revision2,November1981.~r~2.Fitzsimmons, G.W.,etal.,"MarkIContainment ProgramFull-Scale TestProgramFinalReport,TaskNumber5.11,"GeneralElectricProprietary ReportNEDE-24539-P, April1979.3.U.S.NuclearRegulatory Commission, "SafetyEvaluation Report,MarkILongTermProgram,Resolutioh ofGenericTechnical ActivityA-7,"NUREG-0661, July1980.4GeneralElectricCompany,"MarkIContainment ProgramLetterReport:Supplemental Full-Scale Condensation TestResultsandLoadConfirmation,"

MI-LR-81-01-P, April1981.5.Mintz,S.,"BWRSuppression PoolTemperature Technical Specification Limits,"GeneralElectricReportNEDO-31695, May1989.6.U.S,NuclearRegulatory Commission, "SafetyEvaluation Report,MarkILongTermProgram,Resolution ofGenericTechnical ActivityA-7,"NUREG-0661, Supplement 1,August1982.7."MarkIContainment ProgramEvaluation ofHarmonicPhasingforMarkITorusShellCondensation Oscillation Loads,"NEDE-24840, preparedbyStructural Mechanics Associates forGeneralElectricCompany,October1980.8.9.Kennedy,R.P.,"Response FactorsAppropriate forUsewithCOHarmonicResponseCombination DesignRules,"SMA12101.04-R002D, preparedbyStructural Mechanics Associates forGeneralElectricCompany,March1982.IKennedy,R.P.,"AStatistical BasisforLoadFactorsAppropriate forUsewithCOHarmonicResponseCombination'Design Rules,"SMA12101.04-R003D, preparedbyStructural Mechanics Associates forGeneralElectricCompany,March1982.10.11.Bienkowski, G.,Lehner,J.R.andEconomos, C.,"Technical Evaluation oftheNineMilePointUnit1NuclearGenerating StationPlantUniqueAnalysisReport,"BNL-04243, September 1984.7"NineMilePointUnit1Reduction inMarkITorusProgramCondensation Oscillation LoadDefinition andResulting EffectonMinimumShellThickness Requirements,"

Technical ReportTR-7353-1, Revision1,preparedbyTeledyneEngineering ServicesforNiagaraMohawkPowerCorporation, April1991.12."Reduction ofTorusShellCondensation Oscillation Hydrodynamic LoadsforNineMilePointUnit1,"C.D.I.Technical NoteNo.90-11,preparedbyContinuum

Dynamics, Inc.forTeledyneEngineering

Services, November1990.A-6 I

13."FSTFShellCondensation Oscillation LoadingCorrection Factors-Uncorrelated Vents,"C.D.I.ReportNo.79-1,Revision2,preparedbyD.B.BlissandM.E.TeskeofContinuum

Dynamics, Inc.forGeneralElectricCompany,August1980.14.Figure3ofReference 12.15.Table1ofReference 12.16.Table4.4.1-2ofReference 1.17.p.14ofReference 12.18.TESReportTR-5230-1, Rev.1,"MarkIContainment Program,Plant-Unique AnalysisReportoftheTorusSuppression ChamberforNineMilePointUnit1NuclearGenerating Station,"

datedSeptember 21,1984.19.NMCletterNMP1L-0628 fromC.D.Terry(VPNuclearEngineering) toU.S.NRC,datedDecember13,1991.20.Bienkowski, G.,"ReviewoftheValidityofRandomPhasingRulesasAppliedtoCOTorusLoads,"InternalBNLMemo,August1983.A-7 I

LISTOFACRONYMSACBNLAcceptance CriteriaBrookhaven NationalLaboratory BWROGBoilingWaterReactorOwnersGroupCOFSTFGELCOLDRNMCNMPNRCPUARSRSSTSCondensation Oscillation FullScaleTestFacilityGeneralElectricLimitingCondition forOperation LoadDefinition ReportNiagaraMohawkPowerCorporation NineMilePoint-Unit1NuclearRegulatory Commission NormalWaterLevelPlantUniqueAnalysisReportSquareRootoftheSumofSqaresTechnical Specification A-8 Ilf4

'B-1 I

ReviewoftheValidityofRandomPhasingRulesasAppliedtoCOTorusLoadsGeorgeBienkowski August25,1983Containment SystemsGroupDepartment ofNuclearEnergyBrookhaven NationalLaboratory Upton,NewYork11973B-2 l4lS INTRODUCTION TheLDR<'>specification forCOTorusloadsisbasedonFSTFdata(primarily testM-8).Inordertoresolvepotential uncertainties intheconservatism ofthedata,supplementary testsM-11BandM-12wereconducted intheFSTFfacility.

WhileM-12wasnottotallyboundedbytheLDRspecification, thestafffeltthattheLDRprocedure ofsummingtheabsolutevaluesoftheharmoniccomponents wassufficiently conservative toboundanyuncertainties inthedata(Supplement toMarkISER-NUREG-0661).<'>

Manyoftheindividual MarkIplantshavechosentodeviatefromtheLDRprocedure andhavereducedtheconservatism inherentintheabsolutesumloadapplication throughsomeuseofrandomphasingbetweenharmonics oftheLDRCOrigid-wall pressurespecification.

Thebasisforallofthesealternate loadapplication procedures comesfromGEreportNEDE-24840<'~

andsomesubsequent reportsbyStructural MechanicAssociates (SMA12101.04-RODID, SMA12101.04-ROOZD, SMA12101,04-R003D).<4

">Whileindividual plantsobtainareduction inloadduetotheeifectofrandomphasingindifferent matter,agenericevaluation ofthebasefortheseprocedures isnecessary inordertoestablish theadequacyofeachplant'sexception totheAcceptance Criteria.

A.ReviewofGENEDE24840,"Evaluation ofHarmonicPhasingforMarkITorus"Theprimaryobjective ofthisreportistoreducetheexcessive conservatism inthetorusshellresponseduetotheuseoftheabsolutesumofharmonicamplitudes.

Thereportdemonstrates, byexamining throughMonteCarlocalculation boththeFSTFdataandanactualfacility(OysterCreek),thatrandomphasingleadstoamorerealistic response.

Thereportfurtherproposesadesignrulethatisrelatively easytoapplyandprovides90%confidence of50%non-exceedance probability.

Thereportfurtherjustifies thischoiceasbeingappropriate topreserve, attheresponselevel,thenon-exceedance probability orthedegreeofconservatism contained withintheloaddata.Sevenresponses (BDCaxialandhoopstress,BDCradialdisplacement, andfourcolumnforces)attheFSTFfacilityareanalyzedonthreedifferent bases:(a)Fouriercomponents ofthemeasuredspatially-averaged pressuretimehistories over5(second)intervals ofRunNumberM-8areusedasloadinput;(b)MonteCarlotrialsbasedonrandomphasingbetweenthe50harmoniccomponents representing thehistories in(a)areapplied;(c)MonteCarlotrialsusingrandomphasingamongthe50harmonics oftheLDRloadspecification areused.Thepeakresponses resulting fromthesesanalysesarethencomparedtothemeasuredpeaksintheFSTFtests.Acomparison oftheresultsof(a)tothemeasuredresponses suggeststhatthemodelling ofthefacilityandarepresentation ofthedataisadequatetomatchthecolumnforcesandradialdisplacement butyieldspeakmembranestressesthatarefrom13%to30%(hoop)toolow.Thereportgoesintoanumberofexplanations forthereasonsforthisdiscrepancy.

Whilemostofthesuggested causeswouldnotbeapplicable inarealfacility, thesuggestion B-3

\I~

thatshellmembranestresseswillrespondtolocalpressures whiletheinputloadhasbeenaveraged, canbeassumedtobetransferable toaplantcalculation.

Thispotential non-conservatism iseventually recognized inthefinaldesignrule.Thepeakresponses atthe50%NEPlevelresulting from200MonteCarlotrialswithrandomphasingbetweenharmonics (optionb)generally eitherboundtheresultsusingactualphasingorareveryclosetothem.Theratiooftheresponses, basedon(a)dividedbythe50%NEPresultof(b)rangesfrom0.88forthecolumnforcesto1.03fortheradialdisplacement, withthemembranestressesat0.94.The50%NEPresultof(b)comesclosertoboundingtheexperimental databutthemembranestressesarestilllow(9%axialand15%hoop).The200MonteCarlotrialsarealsoperformed fortheLDRspecification.

Becauseofsomeadditional conservatisms intheloadharmonicamplitudes, the50%NEPnowboundsthecolumnforcesandradialdisplacement substantially, essentially matchestheaxialmembranestress,andunderpredicts theFSTFpeakhoopstressbyonlyabout6%.Thereportthenproceedstoperform200MonteCarloresponsecalculations forthemodelofarealfacility(OysterCreek).ClearlyinthisexerciseonlytheLDRharmonics canreasonably beappliedandnodirectcomparison toexperiments canbeperformed.

Theresults,however,suggestthatthecumulative probability distributions (CDP's)fortherealfacilityareverysimilartothosefortheFSTFfacility.

Theratioofthe50%NEPleveltotheabsolutesumisaboutthesameasinFSTFandliesinthevicinityof50%forthemonitored responses.

Thereport'ssubsequent discussion oftheproperwaytocombinestressesisoutsidethescopeofthisreviewandnotdirectlyrelevanttotheloadspecification issue.Onthebasisoftheinformation summarized above,thereportrecommends asimpledesignrulethatwillyield90%confidence of50%NEP,Theresultsof(b)and(c)forFSTF,andthecalculations forOysterCreekdemonstrate thattakinganabsolutesumofthethreehighestharmonics (atresponselevel)coupledtoasquarerootofthesumofthesquares(SRSS)oftheremaining harmonics alwaysboundsandcloselyapproximates the50%NEPlevel.Thereport,therefore, suggeststheuseofthissimplealgorithm fortheadditionoftheharmoniccomponents inthefrequency domain.Inordertoprovideadditional conservatism inarealfacility, theharmonicloadcomponents thatspanastructural naturalfrequency aretunedtothenaturalfrequency ratherthanappliedattheaveragefrequency intheinterval.

Acomparison oftheapplication ofthisdesignruletotheFSTFfacility(wherefrequency tuningisnotused)tothemeasureddatashowsthatallpeakresponses arebound,except'he hoopstresswhichisabout5%low.Thereportsuggestsanumberofconservatisms intheloadingthatwouldcompensate forthissmalldiscrepancy.

Theprimaryeffectsuggested isrelatedtothedampingof2%usedinanalysis.

Inarealfacility, whereloadsarecombinedandarethushigher,the2%dampingisaconservative representation ofthestructure andwouldthusleadtoconservative responses.

B-4

/

B.ReviewofSMAreport12101.04-R001D, R002DandR003DReportSMA12101.04-R001D, "Evaluation ofFSTFtestsM12andM11BCondensation LoadsandResponses,"

wasnotavailable andthusnotdirectlyreviewed.

Themajorresultsandconclusions ofthatreportare,however,summarized inSMA12101.04R002D, andwerefoundtobeconsistent withboththeoriginalreportNEDE-24840 andtheFSTFSupplemental TestLetterReportM1-LR-81-01P.

ReportSMA12101.04-R002D, "Response FactorsAppropriate forUsewithCOHarmonicResponseCombination DesignRules,"summarizes alloftheconclusion ofNEDE-24820 andupdatesthecomparison toincludeFSTFtestsM12andM11B.WhentestM12isincludedinthecomparison, thedesignruleapplication oftheLDRharmonics totheFSTFtorusunderestimates peakmeasuredmembraneaxialstressby11%andunderestimates thehoopstressby14%,whileboundingtheotherresponses.

Onthebasisofthiscomparison, thereportsuggestsmodifying thedesignrulebyusinga"response factor:R,=1.0forotherresponses.

Inaddition, thereportaddsanadditional designruleforthosecircumstances wherethecombination ofabsolutesumandSRSSisnotconvenient, suchasinthetimedomain.Inthiscasethereportstatesthata90%confidence of50%NEPlevelcanbeachievedbymultiplying thepeakresponseresulting fromasinglerandomphasedtrialby1.15.Notethatformembranestressesandstrainsthereisanadditional 1.15"response factor"described above,Theconclusions providecriteriafordesignanalysesalongthelinesjustdiscussed, butanadditional simplification ofneglecting harmoniccomponents above30HZissuggested forstructures withsimilarnaturalfrequency contenttotheFSTForOysterCreek.ReportSMA12101.04-R003D, "Statistical BasisforLoadFactorsAppropriate forUsewithCOHarmonicResponseCombination Rules,"reiterates thedesignrulesdescribed above.Inaddition, recognizing potential uncertainties inthedata,thereportattemptstoprovidesomejustification forneglecting anyadditional factortoprovideadequateconservatism.

Thereportshowsthat,considering thespecification isaresultofthreedatapoints(M8,M12,M11B),theincreaseinresponsetoachieve75%confidence of84%NEPrangesfrom2%forinsidecolumnforceto33%forthehoopstress.Thereportfurtherquotesanunreferenced communication fromDr.AlanBilaninasstatingafactorof1.33fortheratiooftheFSTFdatatothatexpectedinarealfulltorus.Thiseffectispurported tobetheresultoftherigidendeffects,butnofurtherexplanation isprovided.

InAppendixA,thiseffectisexamined.

Weconcludethatforthesefrequencies thatarenotcorrelated betweenbays,theFSTFshouldproduce32%to35%higherloadsthanwouldexistinarealfacility.

Anexamination oftheFSTFdata(inSupplemented LetterReportM1-LR-81-01-P) showsthatonlythefundamental frequency near6Hzshowsanycorrelation betweendowncomers.

Ifoneassumescorrelation betweenbaysatthatfrequency andrandomphasingbetweenbaysatallotherfrequencies, theoverallconservatism fortheaveragepressuremaybeaslowas17%,whileattheresponseleveltheFSTFconservatism willrangeform18%forthehoopstressto38%%uofortheaxialstress.Ifwenowbalancethisversusthemaximumexpecteduncertainty factorforhoopstress(1.33)asinreportR003D,wecouldexpectaB-5 I~~)~t%~

maximumdegreeofnonconservatism ofabout13%.Thisisnotseriousfortwodistinctreasons.Theadditional conservatisms associated withtherealstructure duetothetuningofharmoniccomponents tothenaturalfrequencies andtheclosermatchtothe2%dampingfactorcaneasilycompensate fortheslightnonconservatism.

Secondly, theuncertainty

estimate, usingonlythreepeakresponses formthetestsM8,M11BandM12isprobablyexcessively conservative.

Ifoneused1secondaveragedRMSpressures from8-secondhighmassflowintervals, aswasdoneintheSERSupplement, theratioofmeantopeakR=0.72andthestandarddeviation iso,=0.172.Theresultant loadorresponseat2o,fromthemean(nowproviding averyhighconfidence levelofnon-exceedence) isonly7%abovethedesignruleandcanbeeasilycompensated bythe1.18conservatism factorfortheFSTF.C.SummaryandConclusions Onthebasisofthereviewofthesereportsthestaffconcludes thatadirectapplication ofdesignrulesasgiveninreportSMA12101.04R002D onpage41orinreportSMA12101.04-R003Donpages1-2isacceptable.

Ifharmonics above30Hzareneglected, assuggested forstructures similartoFSTForOysterCreek,aspecificjustification intheformtorusresponsefrequency characteristic mustbepresented.

Anyvariation thatproducesatleastashigharatioofresponsetothatproducedbyabsolutesumasthehighestobservedintheFSTFandOysterCreekanalyses(63%)isalsoacceptable.

Usingthedesignruleasinitially statedonNEDE-24840 (withoutthe1.15factorforshellstresses) isnotacceptable, butamodification using4harmonics summedabsolutely addedtotheremaining summedSRSSismarginally acceptable, providedthereportedshellstressesarenotwithinafewpercentofallowables.

Theadditionof1harmonic, tobesummedabsolutely, providesonlyabouta10%increaseintheresponses ratherthanthe15%neededtoboundFSTFmeasurements.

Theeffectissufficiently small,however,thatfurtherevaluation wouldbenecessary onlyintheeventtheresultant stressesapproached allowable valuesveryclosely.Insummary,thestafffindstheanalysispresented intheseriesofreportsreasonable.

Anyconservative application ofthoseresultsisthusacceptable.

Thedirectapplication ofthedesignrules,asstatedinthefinalreportSMA12101-04-R003D isconsidered adequately conservative.

Anyalternate isacceptable, provideditsapplication toFSTFdatawouldboundallthemeasuredstresses.

B-6 1'I I~IC-1 IR METHODOFIMAGESBythemethodofimages,theimagelocations aredefinedbyx;=2L~x,y;=2H;~y, z),=2'zgi,j,k=0,~1,+2...Thetankdimensions areL,H,Dinthex,yandzdirection respectively.

yistheverticaldirection, y=0isthetankbottomandy=Histhefreesurface.Theoriginofthecoordinate systemisatthelowerleftcorner.xyandz,definethelocationofthesourcewithstrengthP,.Thepotential atanypoint(x,y,z)canbeexpressed asDefine$g=x0-xpg,=-x0-x,g,=-x,+x,g4=xa+xgi=-yo+ygz--2H-y,-y,7t,=y,+y,g,=2H+y0-y1=zoz>(20z03="0-z04=0+ZandliiBnan(&,.4,)'+(4Hj+g

)'2D1yklmn~(2L,Q,)+(4H,.+2H+g

)+(2D~Q)C-2 4~~l sothatNNKLNlP=P.ZZZZAWheii+lilhi4EZ(4i-I;;up(1)-"jsOmal+aihhlPJ,!a siMilPJWwhereNI=4NI=2NK=4NK=2ifififi~1ifi=1kw1k=1andL,MandNdefinethenumberofimagesusedintheimagearray.C-3

~e~f~'~E APPENDIXDTHENRC.REQUESTFORADDITIONAL INFORMATION-

-'::-',;

'.A'ND'THENMPCRESPONSE I'aji gVN~IAAII$HIAOAIIAICWhWkMWR5COlfPOllAYIONl30I PLAINFIELD ROAD.SYRAE.NY.I3212'ELEPHONE I3'I5I47*15IISeptember 29,lNMPrr.0VU.B.NuclearReguhgozy Commission Attn:Dociunent ContxolDesk%aswzlgton, DC35'B:NineMilePointUzIh1DoclretNo.$0-220DPR43S~ed:MCRcqucdfarAAQheati+onzzcofozz BeguSaBe~&wqfAVqrMltPorcNuclearSafesVn01ToneLoadctfazsSubmittal q/"MtyX4gNNByletterdatedAuguat26,l993,theNRCnqIIestod additional

'nnecessary tocompletethere-mvievofam'ayl4,1991re@estfoeducethecondensation oscillation loadsintheNineMihPointUnitITorus.Attachmezlt ltothislprovidesourresponsetotherequested infarntatha.

pyoiI+~anyquestimIs

~QgthezIMponsct pleasecontact~DavidELh5rat(315)428-7029.

VerytrulyyorC.D;-TerryVicepzeideat-NuclearipzlecrizIg xc:RcgiolQ1Adlni51tslzator Rcg4ÃiIMr.B.Norris,SeniorReaMentInspector Mr.R.A.Capra,Director, projectNzeetorate X-i,NRRMr.D.3.BzinRznan, 8caiorprojectManatpr,NRRRecozdsMIzlagemezIt 0-2

)~I

..g~~RESPONSETOTHENRCREQADBXHONAL PiFORIVtATKON RERE-M',VIEW OFMAT14,1TORUSLOADREBUCHONAJSNIAGARAMOHAWKPOWERCOMjNEMIlUEPOP%MJCLEARSTATIO90CKEYNG.S{}-2207FORORATIONUNl'O.1R8dD-39$:68266'l-ZL'-O'C

~<~-!tP LetterReport7519-28,Rev.1September 17,1993~Attachment Page1SERVtCESpgyle'Tt~tltOWNtN&iC4NHo InSectian3.3,1aftheNinebaflePointPlant-Unique TorusSuppression Chamber(Teledyne Engineering SerRev.1;September 21,19S4),itisstatedthatcantraforthetorusshellfsthatwhichcombinesOBACOwfpressure(P),deadweight (M)andtheOBE(Case20).thefol'lowing additional fnformation and/arclarifithisstatement:

nalysisReportofthefees{TES)TR-5320-1, linglaadcambfnatfan htheOBAhydrostatic heHRCstaffrequestsationwithrespecttoQhatfractionofthetota1shellstress(membrane, local,etc.)derivesfromtheCOloadingtfromthePlaadf97Etc.FarEventCombfnatfan 20,ElementNo.19(theostlimitingelement),

thestressesfromeachof'hecantrfbuting loasareasfollows:OriginalAnalysis, Unreduced CmembranepercentStressoi'l}~xLMembranePercent+Bond'9of&51:M&CaLOeadweight OBESeismicInternalPressure(QBA):Total1)7562059,219&2K.16,150(16,025)10.951.3%57.0%'100.051;8122079,72216,751(16,618)10.P41.2X58.0XMJ5100.(4CodeAllowable Stress16,500'4,750herepart(TR-7353-1) ulatedafterallthesasaresuaraed,Theseaddingtheprincipal 5QhksThestressesinparentheses areframtandaretheprincipal stressescalcomponent stressesframthefaurloadaarelowerthanthetotalsobtainedstressesfromeachloadcase.D-49L'-602667-ZE'-QC

~lrar$,

LatterReport7519-28,Rev.1September 17,1993Attachment Page2IJN'%ERvCES W~oYNI~IKWCKCWNRfterCase20,whatloadcombination involving nexthfgheststressfnthetorusshell2WhatpstressisduetaCQ2hmmne:Thenextcontrolling eventcombination isEvelieitfngflelentHo.19,whichincludesdeainternalpressure, SRUandIHACO.Thestrecontributfng loadsareasfollows:COloadsinducesthercentagaofthetotaltCaibfnatfon 14forwefght,08Esefsmfc,sasfromeachoftheMembranepercentStressofMB1~aLNeabrano+Bend'g~UPercentofZetaLOeadwe)ght 08KSoismfcLnternalPressure(lSA)SRV1,7562059,928821'3.534(13,232) 13.0%1,5L73.4%6.0X100.OSCodeAllowable Stress16,500$@Q!Thestressesinparentheses arethcalculated afterallthecol'ponant strescasesaresummed.Thoseara,'ower thanaddingtheprfncfpal stressesframeach1,81220710,4702,385~i'll16470(15,148)24,75011.0%1.3%83,6%14,5'00.l7Ii pr'incfpalstressesesfromthafiveloadhototalsobtainedbyloadcase,vfngbothCOandSNitforthiscaso2kfaftingelement19,ation14,Seeanswer'Whatistheworstcasaloadcombfnatfan involoads'hat arethestresslave>sandtheirs18mme:Theworstcaseloadcombfnatfon, fortheinvolving bothCOandSRVloadsisEventCondftoQuestion2.D-5

~>~it,,~,~'4 LetterReport7619-28,Rev.1September 17,1993Attachment Page3SERVICESNVSCHQf%CUP~QKWNOcwlQWo4.Qhatfstheworstcaseloadcombination forthenotfnvolveC07Howarethestressesforthatthinnfngofthetorusshall2Howdotheycornstresses'mm:

Theworstcaseeventcombination thatdoesnoCambfnatian 18farElementHo.l9,whichfncsefsmfc1andpoolswell,Thestressesfromealoadsareasfollows;torusshellthatdoescaseaffectedbythearewiththeCase20involveCQisEventudesdeadweight, OBEhofthecontrfbut1ng HlabranaStressNeabrane+8endfngOeadwefght OBESefsmfcTotalCodeAllowable Stress1,756205MKR71929(7,812)16,50011812207k2JQ.S,222(8,103)24,760fhgysThestressesfnparentheses arethcalcu'lated aftera)1thecomponent stloadcasesaresummed.Thesearelobtafnedbyaddingtheprfncfpal stressThesestrlsseswouldincreaseslightlyduettorusshell.However,ascanbeseen,approxfmatoly halftheEventCambfnatfan 20swouldcontrolbyawidemarg1n.\Referring nowtoSection5.3ofTESTR-7353-1, Rev.provfdethefallowfng fnformatfon/ciarfffcatfon:

5.IsthIstatement that"...Event Camb1natfo validforboth8and4downcomer bays'owfsBQSSE:Yes.Itfsestablfshed bycomparing thesCoebfnatfons presented herein.(Seeresponse6.Hhatshellthfckness fsusedtacalculate eachlevelstabulated fnthissectfanThnaMfJ::Theor1gfnalthickness 1sused,t0.46inchprincipal stressesessesfromthethreewerthanthetotalssfromeachloadcase.thethfnnfngofthethesestressesareresses,sothelattar2(January14,1992),20...fscontrolling" thisestablfshed7 ressesfromtheEventoquestion8).afthe"actual"stressD-699'd j4 LetterRaport7519-28,Rev.1September 17,1993Attachment Pago4ITthmzae:Yes.Itauldbethesame.Isthesplitinstressesduetothevarioidentical forthethinnershel1case?Iffractional distribution, scantHbuting loadsitisnot,state'thl8.Isthesplitinstressesduetothevariascontributing loadsidentical fartha4derncomar beycases?Iitisnot,statethefractional distribution?

km)nil'.:

FaradventCombination 20,BeatentNo.19,withreducedCO,thestressesframeachofthecontributing loadsaasfollows:ReducedC0,SOowncamer QaysOeadweight OBESeismicInternalPressure(OBA)TotalCadeAIIovable StressHeehraneStress~D1,756'205g,z)gMBZ)5,522{15,452)16,500PercentofU8aL)1.3%1.3'A59-4%MMK100.0XHeibrane+Band'g~Q1,8122079,722~0.16,122()6,044)24,750Percentotll..P).3%60.3%~5100.%ReducadCO,4Oelecaeer BaysHoebranaPercentStressat&%ELZSalHeabranlPercent+Send'gof'Ml~aLGeaheight OBKSeisaicInternalPressure(QSA)Tota1,7562059,2)9M2614,529()4,460}12.)X,).4%63.5'5~5100.N1,8122079,722M2B/,15,116()5~040))2.Ã).4%64.3%~El100.(4CadeAllowable Stress16,50024,750therepgrt(TR-7353-I) culatadafteralltheasesaresumaed.Theseyaddingtheprincipal 5gia!Thestressesinparentheses arefroandaretheprincipal stressesccomponent stressesframthefavrloadcarelaierthanthetotalsobtainedstressesfromeachloadcase.I60IdD-7LT:68266'-0C-6<

0~<,l%p~g LetterReport75}9-28.RevSeptember l7,1993Attachment Pago59.Pravfdetheequfvalent responsetaquestfondawncaaar bays.hmsz:EventCombfnatfon 20fstheonlycasefnvoIvfnextcontrallfng eventcombfnatfan fsEventfnc)udesISACO,Thisfsalsotheworstcfnvalv$ngbathCQandSRYloads.Sfncothosoughtfsanlyfar'BACO,therearenochvaluesforthelBACOresultsforEventCombfsay,thora'fsnadfffsrantfatfon batmenthe.'oreventcambfnatians otherthanthe:rsvfsed Theresultsarepresented fn'guestfan..2~.

I~yygygOITILaOYNtIhCSNL?N+E~

and3farthe4g08ACOlaads.TheCombfnatfon l4whichsIQvontCONhfnatfon loadreductfan befngngestotheorfgfnalatfanl4.Thatistaand8.4amceaer bays.":,"2EventComhfnatfan RQ.D-8 e~I