ML17309A610

Provides Response to RAI & Resubmits Level 1 PSA of Ginna Station

ML17309A610
Person / Time
Site:	Ginna
Issue date:	01/15/1997
From:	MECREDY R C ROCHESTER GAS & ELECTRIC CORP.
To:	VISSING G S NRC (Affiliation Not Assigned), NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
Shared Package
ML17264A794	List: ML17264A793 ML17309A610
References
GL-88-20, NUDOCS 9701170095
Download: ML17309A610 (80)

Text

CATEGORYj.REGULAT.INFORMATIONDISTRIBUTIONISTEM(RIDS)ACCESSIONNBR:9701170095DOC.DATE:97/01/15NOTARIZED:NODOCKETIFACIL:50-244RobertEmmetGinn'aNuclearPlant,Unit1;RochesterG.05000244AUTH.NAMEAUTHORAFFILIATIONMECREDY,R.C.RochesterGas6ElectricCorp.RECIP.NAME'ECIPIENTAFFILIATIONVISSING,G.S.

SUBJECT:

ProvidesresponsetoRAI6resubmitslevel1PSAofGinnaStation.DISTRIBUTIONCODE:A001DCOPIESRECEIVED:LTRENCL3SIZE:TITLE:ORSubmittal:GeneralDistributionNOTES:LicenseExpdateinaccordancewith10CFR2,2.109(9/19/72).gg)05000244RECIPIENTIDCODE/NAMEPD1-1LAVISSING,G.INTERN:FILENRR/DE/EMCBNRR/DSSA/SPLBNUDOCS-ABSTRACTEXTERNAL:NOACCOPIESLTTRENCL11111111111111RECIPIENTIDCODE/NAMEPD1-1PDNRR/DE/ECGB/ANRR/DRCH/HICBNRR/DSSA/SRXBOGC/HDS3NRCPDRCOPIESLTTRENCL111111111011EGRY1DENOTETOALL"RIDS"RECIPIENTS:PLEASEHELPUSTOREDUCEWASTE!CONTACTTHEDOCUMENTCONTROLDESK,ROOMOWFNSD-5(EXT.415-2083)TOELIMINATEYOURNAMEFROMDISTRIBUTIONLISTSFORDOCUMENTSYOUDON'TNEED!TOTALNUMBEROFCOPIESREQUIRED:LTTR13ENCL121 ANDROCHESTERGASANDElECTRICCORPORATION~89EASTAVENUE,ROCHESTER,N.Y.Id6d9-000IAREACODE7/65d6-27ÃROBERTC.MECREDYVicePresidentNvefeorOperotionsJanuary15,1996U.S.NuclearRegulatoryCommissionDocumentControlDeskAttn:GuyS.VissingProjectDirectorateI-1Washington,D.C.20555

Subject:

GenericLetter88-20,Level1ProbabilisticSafetyAssessment(PSA)RochesterGas&ElectricCorporationR.E.GinnaNuclearPowerPlantDocketNo.50-244

DearMr.Vissing,

ByReference1,RG&EsubmittedaLevel2PSAinresponsetoGenericLetter(GL)88-20,IndividualPlantlaminationforSevereAccidentViIlnerabilifies.Subsequenttothatsubmittal,RG&EidentifiedseveralinconsistenciesandunwarrantedconservatismswithinthePSAmodelsanddeterminedthatare-assessmentoftheLevel2PSAwasappropriate(Ref.2).Inthemidstofperformingthisnewanalysis,theNRCprovidedRG&Ewitharequestforadditionalinformation(RAI)relatedtotheMarch15,1994submittal(Ref.3).RG&ErespondedtothisRAIwithseveralletterswhichultimatelyextendedRG&EsresponseuntilJanuary15,1997(Refs.4,5,and6).Therefore,thepurposeofthisletteristoprovidearesponsetotheRAIandtoresubmitaLevel1PSAofGinnaStation.Inaddition,anupdatetoalloutstandingresponsestotheremainingGL88-20supplementsisbeingprovided.WithrespecttotheRAIdocumentedinReference3,AttachmentAtothislettercontainstheresponsetoallNRCquestionsandcomments.TheresponsetothesequestionsisprimarilybasedonRevision1totheGinnaStationPSAascontainedinAttachmentB.ThisdocumentreplacestheLevel1PSAprovidedinReference1initsentirety.TheLevel2andflooding~portionsofthePSAwillbeupdatedandre-submittedtotheNRCinaccordancewiththeschedulelistedbelow.GL88-20iscomprisedofmultiplesupplementsrelatedtosevereaccidentvulnerabilities.AdescriptionofeachsupplementandRG&E'sunderstandingofitsstatusisprovidedbelow:-970i-i70095970ii5".PDR"'"ADOCK05000244PDR

@I~I.Ik( GenericLetter88-20-Thisrequestedanindividualplantexamination(IPE)ofeachplantthatwasbasicallycomprisedofanevaluationofcoredamagefrequency(Level1PSA),containmentperformance(Level2PSA),andafloodingevaluation.RGEcEoriginallyrespondedtothisrequestinReference1;however,AttachmentBtothislettersupersedesthepreviousLevel1PSA.AnewLevel2PSAwillbesubmittedtotheNRCbyMay1,1997(Ref.6).ThenewfloodinganalysiswillbepresentedwiththerestoftheexternaleventsbySeptember30,1997(seeSupplement4below).b.SupplementJ-ThisannouncedtheavailabilityofNURI:G-1335whichprovidedreportingguidelineswithrespecttotheIPEasdiscussedina.above.ThisNUREGwasusedinthepreparationofthePSAsuchthatnofurtheractionisrequired.C.Supplement2-Thisrequestedconsiderationofsevereaccidentmanagement(SAM)strategiesintheIPEprocess.Subsequenttothisdocument,thenuclearindustrydevelopedgenericSAMguidanceandcommittedtoimplementSAMcapabilitiesbyDecember31,1998.RG8cErespondedtothesupplementinReference7andstatedthatwewouldmeetthisdeadline;however,wealsoidentifiedthatRGBwas"activelypursuingcompletionofthesetasksin1997."Duetothedelays.incompletingtheLevel2PSAwhichprovidesaninputintotheplant-specificSAMguidelines,RGBdoesnotexpecttocompletetheseactivitiesin1997.However,theindustrydeadlineofDecember31,1998willstillbemet.Supplement3-ThisannouncedthecompletionoftheNRC'scontainmentperformanceimprovementprogram.RelevantinformationwillbeincorporatedintotheLevel2PSAandSAMguidelinessuchthatnofurtheractionisrequiredonthisspecificsupplement.Supplement4-ThisrequestedanIPEforexternalevents(IPEEE)andprovidedNUREG-1407forreportinginformation.Thefollowingprovidesastatusofeachrequiredexternaleventevaluation:Seismicevents-ThisisbeingaddressedaspartofthecloseoutofGL87-02,VerificationofSet'smicAdequacyofMechanicalandElectricalEquipmentinOperatingReactors,UnresolvedSafetyIssue(USI)A-46.AfinalreportonthesetopicswillbesubmittedtotheNRCbyFebruary1,1997(Reference8).Internalfires-RGEcEwillsubmitafireanalysisbySeptember30,1997(Ref.9).PerReference10,RGB'lanstousetheguidanceofNUI&G-1407,Section14.1andtheEPRIFIVEpropagationanddamageassessmentmodelsforscreeningpurposes,andtheFIVEwalkdownprocedurestoaddressFireRiskScopingStudy(FRSS)issues. Highwindsandtornadoes-InReference11,RG&EstatedGinnaStationwasdesignedtowithstanda1E-05/yrtornadowithstructurescapableofwithstandinga1E-06/yrtornado.ThiswasconsideredtomeetSection5.2.4ofNUREG-1407.TheNRCrespondedinReference12thatsinceRG&Ewasusing"thescreeningapproachdescribedinNUREG-1407forhighwinds...",thiswasacceptable.Consequently,RG&Econsidersthisexternalevaluationcomplete.Externalfloods-InReference11,RG&EstatedthataProbableMaximumFloodwitharecurrenceintervalof5E-04/yrwasusedasthebasisfortheSystematicEvaluationProgram(SEP)ofGinnaStation.NecessaryplantmodificationswereimplementedtomeetthisfloodingeventsuchthatSection5.3ofNUREG-1407wasconsideredtohavebeenmet.TheNRCrespondedinReference12thatsinceRG&Ewasusing"thescreeningapproachdescribedinNUREG-1407forhighwinds,floods...",thiswasacceptable.Consequently,RG&Econsidersthisexternaleventevaluationcomplete.Transportationandnearbyfacilityaccidents-InReference11,RG&EstatedthattheNRCsafetyevaluationforSEPTopicII-1.CconcludedthattheStandardReviewPlanwasmetfortheseissues.Therefore,RG&EconsideredthatSection5.2.3ofNUREG-1407wasmet.TheNRCrespondedinReference12thatsinceRG&Ewasusing"thescreeningapproachdescribedinNUREG-1407forhighwinds,floods,andtransportationandnearbyfacilityaccidents,"thiswasacceptable.Consequently,RG&Econsidersthisexternaleventevaluationcomplete.Supplement5-ThisannouncedthatseismicreviewrequirementsasoutlinedinNUREG-1407couldbedowngraded'basedonNRCre-assessmentofpotentialseismicevents.RG&ErespondedinReference13thatRG&EhadalreadydowngradedourseismicreviewsfromthoseinNUI&G-1407.TheseismicevaluationwillbeprovidedbyFebruary1,1997asdescribedine.1above.InadditiontoGL88-20anditssupplements,theNRCrequestedthatRG&EspecificallyevaluatetwootherissuesinthePSAasfollows:DuringtheSafetySystemFunctionalInspectionoftheResidualHeatRemovalsystemin1989,theNRCidentifiedapotentialscenariowherethedischargelineforServiceWatercoolingtothedieselgeneratorscouldcrimpduringaduringadesignbasisseismicevent(Ref.14).RG&ErespondedtothisbeyonddesignbasisconcernbyagreeingtoevaluatethescenariointhePSA(Ref.15).ThisevaluationisprovidedinSection9.4.1ofAttachmentBtothisletter.

b.InReference16,theNRCidentifiedapotentialconcernrelatedtohydrogenstorageadjacenttothedieselgeneratorbuildingsandrequestedthatRG&EevaluatethisissueaspartofSupplement4toGL88-20.ThisevaluationwillbeprovidedbySeptember30,1997consistentwiththediscussionabove.PleasecontactGeorgeWrobel,ManagerofNuclearSafetyandLicensingat(716)724-8070ifyouhavefurtherquestions.Verlyyours,RobertC.Mecredy

References:

LetterfromR.C.Mecredy,RG&E,toA.R.Johnson,NRC,GenericLetter88-20,datedMarch15,1994.2.LetterfromR.C.Mecredy,RG&E,toA.R.Johnson,NRC,'Level2ProbabilisticRiskAssessment,datedMarch10,1995.LetterfromA.R.Johnson,NRC,toR.C.Mecredy,RG&E,GinnaProbabilisticRiskAssessmentProjectReportofMarch15,1994,totheNRCinResponsetoGenericLetter88-20;RequestforAdditionalInformation(TACNo.M74414),datedFebruary23,1995.4LetterfromR.C.Mecredy,RG&E,toA.R.Johnson,NRC,GenericLetter88-20,ResponsetoRequestforAdditionalInformation(TACNo.M74414),datedApril19,1996.5.LetterfromR.C.Mecredy,RG&E,toA.R.Johnson,NRC,GenericLetter88-20,UpdatedResponsetoRequestforAdditionalInformation(TACNo.M74414),datedJuly31,1996.6.LetterfromR.C.Mecredy,RG&E,toA.R.Johnson,NRC,GenericLetter88-20,UpdatedResponsetoRequestforAdditionalInformation(TACNo.M74414),datedNovember15,1996.7.LetterfromR.C.Mecredy,RG&E,toA.R.Johnson,NRC,SevereAccidentManagement,datedMarch24,1995.8.LetterfromR.C.Mecredy,RG&E,toG.S.Vissing,NRC,GenericLetter87-02,Supplement1(TACNo.M69449)andGenericLetter88-20,Supplement4(TACNo.M83624),datedNovember15,1996. LetterfromR.C.Mecredy,RG&E,toG.S.Vissing,NRC,GenericLetter87-02,Supplement1(TACNo.M69449)andGenericLetter88-20,Supplement4(TACNo.M83624),datedAugust26,1996.10.LetterfromR.C.Mecredy,RG&E,toA.R.Johnson,NRC,'IndividualPlantExaminationofI<<<eternalEvents(IPI<I<I<),180DayResponsetoGenericLetter88-20,Supplement4,datedDecember26,1991.LetterfromR.C.Mecredy,RGB',toA.R.Johnson,NRC,IndividualPlantExaminationsforE<xternalEvents(IP~<<<),datedNovember3,1992.12.LetterfromA.R.Johnson,NRC,toR.C.Mecredy,RGEcE,RE.Ginna-ReviewofResponsetoGenericLetter88-20,Supplement4-IndividualPlantExaminationsforExternalEvents(TACNo.M83624),datedAugust11,1993.13.LetterfromRC.Mecredy,RG8cE,toG.S.Vissing,NRC,ResponsetoGenericLetter88-20,Supplement5,datedNovember7,1995.14.LetterfromM.W.Hodges,NRC,toR.C.Mecredy,RG&E,NRCSafetySystemFunctionalInspectionTea)nReportNo.50-244/89-81,datedMay9,1990.15.LetterfromR.C.Mecredy,RG&E,toT.T.Martin,NRC,ResponsetoInspectionReport50-244/89-81,SafetySystemFunctionalInspection-RHRSystem,datedJune8,1990.16.LetterfromA.R.Johnson,NRC,toR.C.Mecredy,RiskAssociated5'ithHydrogenStorageFacilityatGinna,June24,1993.AttachmentsMDF7897xc:U.S.NuclearRegulatoryCommissionMr.GuyS.Vissing(MailStop14C7)PWRProjectDirectorateI-1Washington,D.C.20555U.S.NuclearRegulatoryCommissionRegionI475AllendaleRoadKingofPrussia,PA19406GinnaSeniorResidentInspector

~~

AttachmentAResponsetoFebruary23,1996RAI'9701170095 RcsponsctoFebruary23,1996RAIThefollowingpagesprovidearesponsetothequestionsdocumentedintheFebruary23,1996requestforadditionalinformation(RAI)letterfroniA.R.Johnson,NRC,toR.C.Mecredy,RGAE.ThisRAIwasprovidedwithrespecttotheGinnaStationLevel2PSAsubmittedbyRGREonMarch15,1994.However,theLevel1PSAhasbeenrevisedandisbeingresubmittedtotheNRCaspartofthisletter(seeAttachmentB).Inaddition,theLevel2PSAwillberevisedandresubmittedbyMay1,1997.Therefore,theNRCquestionsandcommentsdonotmatchupwiththenewpagenumbersandsections,andinsomecases,maynolongerbeapplicable.Assuch,RGBsresponsetotheRAIconsistsofthefollowing:.TheNRCquestionisrepeated(initalics)followedbyaRGREresponse.FortheLevel1PSAandhumanerrorquestions(RAIEnclosure1and2,respectively),theRGkEresponseconsistsofareferencetoaspecificsectionwithintherevisedLevel1PSAwhichaddressesthesubject.TherevisedLevel1PSAwasspecificallywrittentoincluderelevantdiscussionofthosetopicswhichtheNRCraised'issuewithintheFebruary23,1996RAI.Theonlyexceptiontothisiswithrespecttoflooding(seeEnclosure1Question10andEnclosure2Question14)whichwillbeprovidedbySeptember30,1997whentheexternaleventPSAissubmitted.FortheLevel2PSAquestions(RAIEnclosure3),theRGEcEresponseiswithrespecttotheMarch15,1994submittalonly.ThisresponseindicateswhetherRGBplanstomakeachangetotheapproachusedinthenewLevel2PSAorcontinuetousetheoriginaltechnique.TheresultsofusinganynewapproachwillbeprovidedwhentherevisedLevel2PSAissubmittedtotheNRC(seeresponsetoEnclosure3Questions5,6,7,9,10,and17,andEnclosure2Question15).

RcsponsctoFebruary23,1996RAIENCLOSURE1LEVELIQUESTIONSThefollowingquestionconcernstheBayesianupdatingschemeusedintheindividualplantexainination(IPI'):Inthemodelingoflossofoffsitepower(LOOP)events,yourBayesianupdatingschemeleadstoanorderofmagnitudereductionintheLOOPfrequencyoverthegenericnuclearpowerplantLOOPfrequency.ThetotalLOOPfrequencycalculatedisabout3.5I3lyr.Consideringtherelativelyjequentweatherphenomena(e.g.,icestorms)intheGinnaareathatcouldconceivablyleadtoalossofthewholegrid,thisnumberseeinsverylow.ItisalsosurprisingthatBayesianupdatingcouldleadtosuchalargereductionintheinitiatingeventfrequency.Thediscussioninthesubmittalindicatesthatyouareattemptingtoupdateapriordistributionwithevidenceofzerofailuresin9years.Pleasenotethatsuchevidenceisnotverystrongsincethepriormeanofoccurrencerat'eis0.0271yr,oronefailureinabout50years.1~orcasesofweakevidence,thepriordistributionshoulddominatethebehavioroftheposteriordistribution.However,thisisnotthecasefortheapplicationresultspresentedinthesubmittal.ItwouldappearthatyouhavereplacedtheLognormalpriorwithanequivalentGammaprior,whichisconjugatedwithaPoissonlikelihoodunderthe13ayesalgorithm.ThemethodusedfortranslatingLognormaltoGammaisthemethodofmatchingthefirsttwomoments(meanandvariance).However,theGammadistributionisnotagoodapproximationforaLognormaldistributionwhentheresultingparametersoftheGammadistribution(namelyaandp)fallincertainregions,specifically,iftheparameterubecomeslesst'hanorequaltoone.Insuchacase,1heGanmiadistributionwouldnothaveany>naximuni,andifusedasapriorwouldheavilyweightthelowvaluesoftheoccurrencera1e,contrarytotheLognormaldistribution.Thisappearstobethecaseforyourapplication.Itisalsoiinportanttonotethattheparameteruisunitlessandwouldnotchangeifanannualoccurrencerateora100-yearoccurrencerateisused(incontrast,thepparameterhasdimensionsoftime).UseoftheGanimadistributionisnotreconmiendedforcasesinwhichtheparameterufallsbelowone.Inpractice,theGainmadistributionshouldonlybeusedasanapproximationtotheLognormaldistributionwhenuisgreaterthantwo.CalculationsshowthatifaLognormaldistributioninsteadofaGanmiadistributionisused,yourpriormeanof0.027lyranderrorfactorof23.7resultinaposteriormeanof0.015lyrandanerrorfactorof8.5,amuchmoremodestreductionthanthatindicatedinthe 0

Responsetolcbruary23,1996RAIsubmittal.Forthesereasons,andthefactthattheupdatingwasdonewiththerelativelyweakevidenceofnofailureinriyears,theposteriordistributionobtainedappearstobeinerror.PleaserecalculateyourBayesianupdateusingaLognormalpriordistributionandcomparetheresiiltsivithyoursubmittal.Pleaseprovidetheimpactthisadjustmentinmodelinghasonthecoredamagepequency(CD&)andontheimportantseqiiences.SeeSection7.3.1.2.Thenewlossofoffsitepowerfrequencyis6.32E-02/ryr.(b)PleaseexplainyourBayesianupdatingtreatmentofreliabilitydataforotherconiponentswhereconditionssimilartotheonesfoundinLOOPtreatmentexisted(i.e.,u<landsparseorzeroplant-specificevidence).Forexample,thefollowingcomponentfailureratesmayfallintothiscategory:120-voltacbusfailures,auxiliaryfeedwater(AF0)andresidualheatremovalpuinpfailurerates,andsafetyinjectionandservicewater(SN)demandfaiiurerates.PleasediscusshowtheBayesianupdatingofthesecoinponentswasdone(e.g.,hownianyfailureswereexperienced,detailsofBayesianupdatingcalculations,andthefinalposteriornumbersusedintheIPE).Ifanadjustmentinfailuredataisnecessary,pleaseprovideanestimateoftheimpactontheCDFandimportantsequences.AppendixCcontainsadetaileddiscussionoftheplant-specificdatacollectioneffort.TableC-3providesalistingofthefailuredatausedinthePSAmodelsforwhichplant-specificdatawascollectedagainst.Thetableprovidesplant-specificfailuredata("Plant"),genericdata("Aggregated"),andfinalvalueusedinthePSA("I'inal").Essentially,ifnocomponentfailureswereobservedinagivenpopulation,genericdatawasused.Ifoneortwocomponentfailureswereobserved,Bayesianupdatingwastypicallyperformed.Ifasignificantnumberoffailureswereobserved(i.e.,morethantwo),onlyplant-specificdatawasused.2.ThefollowingquestionconcernsthemodelingofLOOPevents:(a)Itisnotclearhowthepossibilityofapost-tripLOOPwasmodeled.Thesubmittalstatesthati'econservativeassumptionismadeini'hatareactortripwouldleadtoaLOOPasaresultofagridtransientcausedbya-lossoftheGimiageneratingcapacity.Doesthatincanthatanyinitiator(e.g.,aloss-of-coolantaccident(LOCA),atransient)willalsoleadtoaLOOPandademandfor

/

ResponsetoFebruary23,1996RAIdieselgenerators?Inthatcase,alossofS8'wouldleadfirsttoaLOOP,andthentoastationblackout(SBO),becauseSN'isusedtocoolthediesels.Therefore,alossofSP'initiatorshouldhavearelativelyhighconditionalcoredamageprobability,whichisnotsupportedbytheresults.Pleaseclarifythetreatmentofthepost-initiatorLOOPandprovidedata(andthebases)fortheconditionalprobabilityofaLOOPfollowingareactortrip.Ifanadjustmentinmodelingisnecessary,pleaseprovidetheimpactontheCDFresultsandtheimportantcoredamagesequences.Section3.4.2.3discussesthepost-tripLOOPeventwhileSection7.3.1.2discussesthedatausedfortheseevents.Insummary,apost-tripLOOPwasconsideredforanyinitiator(similartoUFSARChapter15assumptions)sincethelossofGinnaStation~uldleadtosufficientgridinstabilitysoastoresultinasubsequentLOOP.ThereisnodiscussionofLOOPandSBOsequences,andnoseparateeventtreeisprovided,eventhoughastatementismadethattheseeventsaretreatedseparatelyPomothertransientsbecauseoftheirspecialnature.PleasediscussexactlyhowyoutreatedSBOandprovidetheSBOeventtree,ifavailable.ASBOeventtreehasbeengeneratedasshowninSection5.2.AllSBOscenariosaretransferredtotheSBOeventtreeforevaluation.TheonlyexceptionsareformediumandlargeLOCAsandATWSeventswhichareassumedtodirectlyresultincoredamage.Itisnotclearhowtheturbine-drivenauxiliaryfeedwater(TDAFS)pumpisusedunderSBOconditionsandhowitismodeled.ThedependenciesofthispmnpincludeSS';heating,ventilation,andairconditioning(HVAC);anddcvoltage.ItisstatedthatatestrunofIhourand45nunuteswasmadetoshowthatthepinupcansurviveatotallossofacpower(i.e.,lossofS8'coolingtothebearingoilcoolers).However,notestwasrunbeyondthattime.Anothertestshowedthatthetemperatureswillreachl45'FintheTDA1<5'roomaftera4-hourlossofHVAC.Thereisnoindicationastowhathappensbeyondthe4hours,exceptforastatementthattherewouldbenodamageforalleast24hours,butthatforconservational0-hourpowerrecoveryismodeled.ThereisnoindicationofthedepletiontimeforthebatterysupportingTDAb"5'peration(otherthanthe1,200amp-hrcapacitygiven),ofhowthisconclusionwas<<iodeled,orwhatbatterysupportsTDAF8'operation(e.g.,isitbatteryIB?).TheTDAF0'pumpsaretestedtoshowoperationovera2-hourperiod.4

ResponsetoFebruary23,1996RAIPleasedescribethesequenceofeventsforSBOinaniannersiinilartothatforotherinitiators,whichspecificallyaddi'essestheoperationandmodeling(includingthetimeassumedforoperation)oftheTDAP"5'pump.ThesuccesscriteriausedfortheSBOeventtree(andTDAFWpump)isfoundinSection4.2.2.4,~r~lin,andAppendixB.Certaindetailsarenotclearaboutthedesignandoperationofthe125V-dcsysteni.Canthetechnicalsupportcenter(TSC)batterysupporttheoperationofthesafeguardsequipmentinanSBOifbattery1AorIBfailed,andhowisthiseventualitymodeled?CantheIAor1Bbatterysupporttheoperationoftheotherdivisionandhowisthisfimctionmodeled?5'hyistheTSCbatterytheonlyonetested(apparently)andtestedonlyfor2hours?Isthisperiodtheassumedrunningtimeofthe.TDA1$'pumpinanSBO?8%atarethedepletiontimesforthe1A,IB,andTSCbatteries?TheTSCbatteryhasthecapabilitytotieintoDCsafeguardstrainsasshowninFigure6-8.However,duetothesignificantamountofeffortrequiredtoutilizethiscapability,creditwasnotassumedinthePSA(orintheSBOmitigationplansper10CFR50.63).Withrespecttocross-supportamongBatteryAandB,thereisthecapabilityforcertainloadsonaDCtraintobesuppliedbytheoppositetrainduringemergencyconditions(e.g.,dieselgeneratorsupport).ThisisanautomaticfeatureincorporatedintothePSA.ItisstatedthattheTSCbatteryandlorthestandbyelectricpowerTSCdiesel(s)canbeusedasalimitedbackupshouldthemainemergencydieselgenerators(I"'DGs)fail.8%ichloadscanthesedieselssupport?Pleaseexplainifthesedieselsarecreditedinthemodel,andifso,explainhowtheyweremodeled(includingoperatoractions,unavailability,andfailriredata).CreditfortheTSCdieselgeneratorwasnotassumedinthePSAmodel.Itseemsthatthehardwareexistsforcross-connectingtheeniergencybuses.IsthisactionproceduralizedIfso,iscredittakenforitintheIPL<",andhowwasitmodeled?Thecapabilityforcross-connectingthe480Vsafeguardsbusesexistswithintheemergencyplans.However,thisactionwasnotdeemedcriticalforthePSA,andassuch,wasnotcredited.PleaseexplainwhytheeventsofJanuary21,1985(LER85-002),airdJuly16,1988-(LLR88-006),werenotcountedasplant-specificLOOPeventsinthe

RcsponsctoPcbruory23,1/96RAIBayesianupdatingcalculationofLOOPfiequency.TheformereventwasanincipientLOOP,inwhichbothEDGswerestartedandtiedtotheirsafeguardsbusesoflowgridfrequencycausedbyextremelycoldweather.Thelattereventwasvialossofnormalpower,includingpowertoallfoursafeguardsbuses,suchthatbothEDGswerestartedandloadedontothesafeguardsbuses.SeeSections3.3.4and3.3.9fordiscussionoftheseevents.TheEDGfideloil(ransfersystemisapparentlymodeledseparatelyfromtheL<DGs.However,theCommonCauseTables(3.3.4-Iand3.3.4-2)donotseemtoshowthesepumps,eventhoughtherehasbeenatleastoneeventinvolvingacommoncausefailureofthesepumps.TheeventofE~'ebruary20,1987(LER87-001),involvedpluggingofthestrainersinbothpuInpsasaresultoftheuseoftheinappropriatematerialstocleanthefueloiltankPleaseexplainhowthiseventwasincorporatedinthemodelingofthecommon-causefailureoftheEDGfueltransfersystem.Iftheeventwasnotaccountedfor,pleasejustifytheomission.CCFsofthedieselfueloilpumpshasbeenaddedtothemodel(seeTable7-3).3.Thisquestionconcernsthetreatmentofthemainfeedwater(MF8)sys(em:(a)PleaseprovideadescriptionoftheA8<8'systemandhowitwasmodeled(e.g.,whichcomponentsweretakenintoaccount?).SeeSection6.14.(b)AstatementismadethattheMl0'pumpscannotbeusedbelow4percentpower.Thepowerlevelwillfallbelowthisvalueveryquicklyaftershutdown,yetM/~5'perationseemstobecreditedinmanysequences,andthereisnodiscussionofanytimingconcernswithrespecttotheMFKPleaseclarifyhowandwhenMFV'pera(ionwascreditedintheanalysis.IfanadjustmentintheInodelingofMF8'isnecessary,pleaseprovideanestimateoftheimpactontheresultsandontheimportantsequences.SeeSection6.14fordiscussionsofhowMFWwascredited.NotethattheMFWsystem~cnbeusedbelow4%power(andistypicallyusedinitiallypost-trip).However,duetothedifficultiesinmaintainingSGlevelsundertheselowpowerconditions,themotor-drivenAFWpumpsarenormallyusedduringstartupandplannedshutdownactivities.ThesedifficultiesdonotpreventuseoftheMFWsystemduringaccidentconditionsasinstructedbytheEOPs."4PleaseprovidethebasesforusingthecutsettruncationlimitofS.I"8lyr,whichis

ResponsetoFebruary23,1996RAIrelativelyhighcomparedtoindustrypractice,andprovideanestimateoftheresidual.Iftheresidualissigvificant(e.g.,greaterthan5percentoftheCDF),pleaseprovideanestimateoftheimpactontheimportantsequencesandontheresults.Anewtruncationlimitof1.0E-10wasusedasdiscussedinSection9.3.1.9.Oneofthesmall-smallLOCAsuccesspathsutilizesarapidcooldownofthereactorcoolantsystem(RCS)tolowpressureinjection(LPI)conditions.Thisprocessissupportedbymodularaccidentanalysisprogramruns,accordingtothesubmittal.ArethereproceduresinplacetoutilizethisoptionsIfnot,providethebasesforcreditingthisaction.Ifavailable,pleaseprovideanestimateoftheimpactontheresultsandimportantsequencesifthisoptionwasnotcredited.PP.,,~ll-IICAPPCI*PfChIffoperatoraction.6.Thisquestionconcernsthetreatmentofinitiatingevents:(a)Thesinall-smallLOCAinitiatingeventjequencyintheGinnaIPI7issignificantlysmallerthanthatusedinNU1KGICR-4550.TheIPL~"S3frequencyis7.3L-4versustheNURK<GICR-4550frequency-of1.3E-2.Itdoesnotseemtoincludesucheventsasspuriousreactorcoolantpump(RCP)sealLOCAS,whichareamajorcontributortotheS3frequencyinNU1K~GICR-4550andotherstudies.PleaseprovidethebasesforexcludingRCPsealLOCAsPomtheS3frequencyand,ifavailable,provideanestimateoftheimpactontheresultsandtheimportantsequencesifamoretraditionalsmall-smallLOCAfrequencyestimatelsused.SpuriousRCPsealLOCAshavebeenincludedwithinthesmall-smallLOCAfrequencyasdiscussedinSection7.3.1.6.ThetotalfrequencyofsmallLOCAsduetopipebreaksandspuriousRCPsealLOCAsisnow6.6E-03(notethatGinnaStationonlyhastwoRCPs).(b)Pleaseexplainwhythelossofa4kV-busisnotconsideredasaninitiatingevent.Forexample,lossofbus12Aor128mightcauseaninitiatingeventwhilefailingsafeguardsequipmentonthatbus.SeeSections3.3.4and3.3.9.(c)AtleasttwoeventsatGinnahaveinvolvedthepossibilityofPeezing,leadingtoalossofSKThissituationcouldbeduetofrazilicebuildupontheintakescreens.

RcsponsctoFebruary23,1996RAIInadditiontotheeventdiscussedintheIPL~(LER83-006)causedbyloweringofthevoltagetotheintakeheaters,aInorerecenteventoccurredaftertheIPL<"submittal.Thiseventinvolvedazebramusselbuildupontheheaters,againcausingfrazilicebuildup.IncalculatingthefrequencyofthelossofSS'initiatorintheIPL,wasthepossibilityoficebuildupaccountedfor?Ifso,pleasesunrmarizehowitwastakenintoconsideration.Ifitwasnot,pleaseprovideanestimateoftheimpactontheCDFandonimportantsequences.SeeSection7.3.1.8foradiscussionofthefrequencyforlossofSW.(d)ExtremelycoldweathercouldconceivablycauseaLOOPbecauseofahighloadonthegrid(aprecursoreventoccurredinJanuary1985;seequestion2(gJabove),inconjunctionwithalossofSS'becauseoficebuildupasdiscussedinpart(c).LossofSS'would,inturn,leadtofailureofthedieselgenerators,1husleading1oanSBO.Pleasediscusshowyouconsideredsuchaninitiatingevent.Ifthiseventhasnotbeenadequatelyconsidered,pleaseprovideanestimateoftheimpactofthiseventontheCDFandonimportantsequences.ThefrequencyofLOOPwasbasedonindustrydatabetween1980and1995suchthatextremelycoldweatherisaconsideration.AcoincidentfailureofSWisconsideredveryunlikelyduetothefactthatfraziliceoccursfollowingrapidchangesinwinddirection(i.e.,fromthenorth).Therefore,theextremetemperatureswouldhavetobefollowedbyarapidwindchange,withasubsequentfailureofallRG8.'Epreventativemeasures.(e)Pleaseprovidethefrequencyestimateusedforinitiationofananticipatedtransientwithoutscram(AT0'S)intheIPE~analysis.SeeSection7.3.1.11.7.Thisquestionconcernsthedatausedinthemodel:(a)Indiscussingcomponentswithrelativelyhighfailureratescomparedtogenericdata(e.g.,containmentspraypumps),astatementismadethatthehighrateoffailureisdue1oalimitedtestexposuretime,andnotnecessarilyt'oaplantvulnerability.However,certainfailuremodes(e.g.,sedimentbuildup)mayappearincomponentsthatareidleforlongperiods.Pleaseverifythattheplantexperienceforthesecomponentswitharelativelyhighfailureratewasretainedindevelopingthedatabase.--"----'-

RcsponsctoFebruary23,1996RAISeeSection7.2.1.3.1andAppendixC.(b)PleaseprovidethefinalreliabilitydatalistingsofAppendixL.IfnotincludedalreadyinAppendixI,pleasealsoprovidethegenericdataused,aswellastheplant-specificexperience(e.g.,numberofdemands,ornumberofhours,andnumberoffailures)foreachfailuremodeandeachcoinponent.SeeTables7-1andAppendixC,TableC-3.(c)Thecoinmon-causeanalysisseemstohaveoinittedsomepotentiallyimportantcomponents,whichmightinfluenceyoursearchforvulnerabilities.Thefollowingcomponentswereapparentlynotconsidered:'ICircuitbreakersRelays(engineeredsafetyfeaturesactuationsystem)ElectricalswitchgearTransmittersVentilationfansAircompressorsInvertersInaddition,conmion-causefailureswithintheAF0'andstandbyauxiliaryfeedwater(SAI<"5)systems,alsoinvolvingtheturbinedrivenpump(i.e.,driverindependentfailures),coiildbepostulatedPleaseprovidethebasesforoinitlingthesepotentialconrmon-causefailuresanddiscusshowyouensuredthatnovulnerabilitiesweremusedasaresultoftheseomissions.SeeSection7.2.2andTable7-3fordetaileddiscussionsrelatedtocommoncausefailures.(d)Pleaseshowhowthepossibilityoffreezing,alongwithanyplantspecificdatainvolvingfreezing,hasbeenaccountedforinfailuredataandcommon-causefailuredataforcomponentsthatareviilnerabletothisphenomenon.Apparently,thisphenomenonincludestheSAFE'systeni,theEDGs,andtheSWsystem.SeeSections3.3.5and6.11.8.ItisnotclearfromthesubmittalifplantchangesasaresiiltoftheSBOwerecreditedinitheanalysis.Pleaseprovidethefollowinginformation:

RcsponsctoFebruary23,1996RAI(a)Discusswhetherplantchanges(e.g.,proceduresforloadshedding,acpower)madeinresponsetotheSBOwerecreditedintheIPEandwhichplant-specificplantwascredited.TheGinnaStationdesignwithrespecttomeeting10CFR50.63wasusedinthe-GinnaStationPSA.(b);Ifavailable,providethetotalimpactoftheseplantchangestothetotalplantCDFandtotheCDFcontributionfromSBO(i.e.,thereductionintotalplantCDFandSBOCDF).SincetherevisedPSAbeganwiththecurrentGinnaStationdesign(i.e.,postimplementationofanySBOrelatedplantchanges)no~canbeprovided.(c)Ifavailable,providetheimpactofeachindividualplantchangeonthetotalplantCDFandtheSBOCDF(i.e.,thereductionintotalplantCDFandSBOCDF).SincetherevisedPSAbeganwiththecurrentGinnaStationdesign(i.e.;postimplementationofanySBOrelatedplantchanges)no~canbeprovidedwithoutsignificantmodelingchanges.(d)'iscussanyotherchangestotheplantthatareseparatefromthosemadestrictlyinresponsetotheSBOrulethatnonethelessmayreducetheSBOCDF.Inaddition:(i)Describewhetherthesechangesareimplementedorplanned.(ii)IndicatewhethercreditwastakenforthesechangesintheIPE.(iii)Ifavailable,discusstheimpactofthesechangesontheSBOCDF.AllplantchangesmadeasaresultoftheGinnaStationPSAarediscussedinSection11.1.2.ThisquestionconcernsthetreatmentofHVACfailures,eitherasaninitiatingeventorsubsequenttoaninitiator.AdescriptionoftheHVACsystemisprovidedinSection3.2.l.8,alongwithsuccesscriteriaandadescriptionofoperationundernormalandaccidentconditions.-Itisstatedinthesubmittalthatlossofcontrolbuildingventilationwillnotleadtoaninitiatorbecauseoperatorinspectionsareperforniedonaregularbasis.ThereisnodiscussionaboutotherHVACareascausinganinitiatingevent.PleaseprovideamorecompletedescriptionofyourinvestigationintotheimpactoflossofHVACtotheroomscontainingsafety-relatedequipment.Discusstheequipment e

1<esponsctoFebruary23,1996RAIsensitivetotemperaturechange,wherethatequipmentislocated,methodsofassessment(e.g.,calculationsorteststodeterminethetemperaturesandtiming),andcreditsforoperatoractionsandtiming.Givethisinformationfortemporaryequipment,asweII.PleaseprovidetherationaleforeliminationoflossofHVACasaninitiatingeventorassupporttospecificequipment.Consider,thefactthatequipmentmaybetrippedonhightemperaturebeforethedamagethresholdisreached.SeeSections3.3.5and11.1.1.10.Thefollowingquestionconcernsthetreatmentofflooding:(a)Pleasediscussyourconsiderationofdrains(includingbackfloodingtootherareasandtheprobabilityoffailure,i.e.,duetoblockage)andofdoorsallowingfloodingofotherareas.Astheflrezonesareusedfordelineationoffloodzones,discusswhetherallflredoorsarewaterproofatGinnaandwhetherfailureofthesedoorstobeinaclosedpositionisaccountedforinthemodel.AresponsewillbeprovidedbySeptember30,1997followingre-analysisofthefloodingPSA..(b)Pleasediscusswhetherinadvertentactuationoftheflresuppressionequipment(i.e.,notjustpipefailuresinthissystem)isaccountedforintheanalysisandestimateitsimpactonthefloodingscenarioresultsifit'isnot.AresponsewillbeprovidedbySeptember30,1997followingre-analysisofthefloodingPSA.(c)Pleasediscusstheoperatoractionsneededforisolationandmitigationofthemostimportantfloodscenariosandprovidethebasisforflood-affectedhumanerrorprobabilities(HEPs)used.(ItseemsthesameHEPsasintheinternaleventsanalysiswereusedforsomeactions,disregardingtheadditionalstressthatwouldbeplacedontheoperator.)Discussionofanyalarmsoranyothermeanstheoperatorswouldusetodetectandstoptheflood.AresponsewillbeprovidedbySeptember30,1997followingre-analysisofthefloodingPSA.(d)Discusshowmaintenanceerrorsweretreatedinthefloodinganalysis.Includeerrorscommittedwhileincoldshutdownthatwereleftundiagnoseduntilthefloodeventoccurredwhiletheunitwasatpower.AresponsewillbeprovidedbySeptember30,1997followingre-analysisof

RcsponscloFebruary23,1996RAIthefloodingPSA.11.Fromthedescriptionofthesystem,"PrimaryPressureControlSystem,"itisnotclearhowthepressurizerpower-operatedreliefvalves(PORVS)andtheblockvalvesaremodeled.Pleaseprovidethefollowinginformation:(a)Whatfractionoftimearetheblockvalvesclosed?SeeAppendixC,TableC-4(eventsRCMVD00515andRCMVD00516).(b)Howareclosedblockvalvesaccountedforinthemodel(forexample,inmodelingATWS,feedandbleed,andinmodelingRCSintegrityaftertransient)?WhatistheestimatedimpactonCDFandimportantsequencesifblockvalveoperationisnotconsidered?Thepotentialfortheblockvalvesbeingclosed'sspecificallyaddressedwithinthemodelsforanyeventinwhichautomaticormanualoperationofthePORVsisrequired.OnlyforcaseswheremanualopeningofthePORVsiscreditedaretheoperatorsallowedtoopenaclosedblockvalve(i.e.,aclosedblockvalvefailsautomaticactuationofthePORVs).(c)DiscusstheoperatoractionsrequiredtoopentheblockvalvesandthePORVswhenneeded.OperatoractiontoopenthePORVsisaddressedwithinvariousoperatoractionsdescribedinAppendixF,TableF-4(e.g.,RCHFD01BAF,RCHFDCDDPR).12.Thestatusofsomeofthepotentialplantimprovementstoreducethelikelihoodofcoredamageandlorimprovecontainmentperformancediscussedinthesubmittalisnotclear.Pleaseclarifthesubmittalinformationbyprovidingthefollowing:(a)Thespecificimprovementsthathavebeenimplemented,arebeingplanned,orareunderevaluation.SeeSection11.1.2.(b)Thestatusofeachimprovement,thatis,whethertheimprovementhasactuallybeenimplemented,isplanned(withscheduledimplementationdate),orisbeingevaluated.

II Rcsj)011scioFcbn)ary23,1996RAISeeSection11.1.2.(c)Theimprovementsthatwerecredited(ifany)inthereportedCDF.SeeSection11.1.2.(d)Ifavailable,thereductiontotheCDFortheconditionaLcontainmentfailure.probabilitythatwouldberealizedfromeachplantimprovementiftheimprovementwastobecreditedinthereportedCDF(orcontainmentfaiLureprobability),ortheincreaseintheCDF(orconditionalcontainmentfailureprobability)ifthecreditedimprovementwastoberemovedfronithereportedCDF(orcontainmentfailureprobability).SeeSection11.1.2.(e)Thebasisforeachimprovement,thatis,whetheritaddressedavulnerability,wasotherwiseidentifiedfromtheIPEreview,wasdevelopedaspartofotherNRCrulemaking,suchastheSBOrule,andsoon.SeeSections11.1.2and11.1.3.13.NUREG-1335,Section2.1.6,Part4,requests"athoroughdiscussionoftheevaluationofthedecayheatremovalfunction."Section3.4.5oftheIPE,DecayHeatRemoval(DHR)Evaluation,doesnotprovidespecificsandinsightsonvulnerabilitiesofDHRsystems.PleasediscussinsightsderivedforDHRanditsconstituentsystemsandprovidethecontributionofDHRanditsconstituentsystems(includingfeedandbleed)toCDFandtherelativeimpactoflossofsupportsystenisonthefrontlinesystemsthatperformthatfunction.SeeSection9.2.14.Inmanyprobabilisticriskassessments,RCPsealLOCAisasignificantcontributortotheCDFeitherasaninitiatingeventorasasystemfailureconsequentialtoanotherinitiator.AlthoughthesubmittaldiscussesRCPsealLOCA,pleaseprovidethefollowingadditionalinformation:(a)AdiscussionoftheRCPsealLOCAniodelused.Includetheprobabilityversusleakagerateversustimedataandanyspecifictestresults.SeeSection4.2.2.3.2.(b)Adiscussionofoperatoractionsthatareproceduralizedandtheirtiminginthe I,10 ResponsetoFebruary23,1996RAteventofalossofoneortheothermethod(orboth)ofsealcooling.SeeAppendixF,TableF-4(eventsCCHFDCCWAB,CCHFDSTART,CVHFDPMPST,andRCHFDOORCP).15.NUREG-l335requeststhatthefollowinginformationbeincludedforimportantaccidentsequences:"aListofmajorcontributorstothoseaccidentsequencesselectedusingthescreeningcriteria.Majorcontributionssuchasthosefromfront-linesystemsorfuiictionsandsupportstates,aswellascontributionsfromunusuallypoorcontainmentperformance,areimportantforinclusion."TheIPEsubmittalprovidesatableofimportantsequences,aswellastheirdescription.Pleasediscussofimportantcontributors(e.g.,'failureofoperatortoswitchovertorecirculation",or"common-causefailureoftheresiduaLheatremoval(RHR)pumps")tothefailureoffunctionsindominantsequences.SeeSections9.1.1and9.3.2.gfqcyp(-2persteamgeneratorwouldbeexpected),eventhoughtheeventontheB.generatorwasincludedaspartofBayesianupdating.PleaseprovidethebasesfortheSGTRinitiatingeventfrequencyused.l6.Thisquestionconcernsthemodelingofsteamgeneratortuberupture(SGTR)events:(a)TheSGTRinitiatineventreuenissomewhatsmallerthanexected1.ESeeSections3.3.6and7.3.1.7.(b)TheresultsindicatearelativelyhighcontributionfromSGTRevents,thusimplyingthatrelativelyhighoperatorfailurerateswereusedforthisevent.YetitseemsthatinlightoftheSGTReventthatdidoccurattheplant,operatortrainingandprocedureswouldemphasizethiskindofevent.PleasediscusshowtheHEPsforthiseventwerederived.SeeAppendixF,TableF-4(eventsMSHFDISOLR,RCHFDCDDPR,RCHFDCDOVR,RCHFDTR2,andRCHFDCOOLD).(c)Ifanyadjustmentsintheinitiatingeventfrequencyorpost-initiatormodelingarenecessaryinordertoreflectthe"as-built,as-operatedplant,'leaseprovideanestimateoftheimpactontheCDFandthedominantsequences.Section4.2.2.3.3~1/.discussesthesuccesscriteriaforaSGTR,Section5.7describestheSGTReventtree,andSection7.3.1;7discussestheSGTR RcsponsctoFebruary23,1996RAIfrequencydetermination.Nochangestothesesectionsarerequired.17.Thepressurizersafetyvalvesmightbechallengedwhencertaintransientinitiators(J.OOP,lossofinstrumentair)occurbecausethePORVsdependoninstrumentair.(a)PleaseprovidetheconditionalprobabilitiesofPORVchallengesforvariousclassesoftransients,particularlytheonesleadingtoalossofinstrumentair.Pleaseprovidethebasesforthenumbersused.SeeSection4.2.2.3.1foradescriptionoftheeventswhichchallengethePORYsandSection6.15foradescriptionofthePORVsasused'nthemodels.(b)Pleaseprovidetheconditionalprobabilityusedforthesafetyvalvesstickingopenoncechallengedinscenariosunder(a).above,alongwiththebasesforthenumbersused.SeeTable7-1,eventsRYQ,RYT,RZQ,andRZT.

RcspollsctoFebruary23,1996RAIENCLOSURE2HUMANRELIABILITYANALYSISQUESTIONSPRE-INITIATORHUMANERRORSThesubmittalisnotcompletelyclearontheorganizationsthatparticipatedinthehumanreliabilityanalysis(HRA)portionoftheanalysis.PleaseclarifytheextenttowhichtheHRAwasperformedbythelicensee'sstagversuscontractorsandwhichcontractorswereinvolved.Also,pleasedescribeanyindependentpeerreviewperformedfortheHRAandindicatetheextenttowhichHRAexpertswereinvolvedinthe,review.SeeSections11.3and11.4.Thesubmittaldoesnotclearlydiscusstheprocessthatwasusedtoidentifyandselectpre-initiatorhumanfailureevents(HFEs)involvingmiscalibrationofinstrumentation.Theprocessusedtoidentifyandselectthesetypesofhumaneventsmayincludethe.reviewofprocedures,anddiscussionswithappropriateplantpersonneLoninterpretationandimplementationoftheplant'scalibrationprocedures.'leasedescribetheprocessusedtoidentifyhumaneventsinvolvingmiscalibrationofinstrumentation.Pleaseprovideexamplesillustratingthisprocess.SeeSection7.4.1andAppendixF,TableF-3.Itshouldbenotedthatmiscalibrationissuesaretypicallyaddressedwithinthefailuredatafortransmitters,indicators,etc-(see"failslow,""failshigh,"and"failstorespond"inTable7-1).Thesubmittaldoesnotclearlydiscusstheprocessusedtoidentifyandselectpre-initiatorHFEsinvolvingthefailuretoproperlyrestoretoserviceaftertestormaintenance.Thisprocessusedtoidentifyandselectthesetypesofhumaneventsmayincludethereviewofmaintenanceandtestprocedures,anddiscussionswithappropriateplantpersonnelontheinterpretationandimplementationoftheplant'stestandmaintenanceprocedures.Pleasedescribetheprocessthatwasusedtoidentifyhumaneventsinvolvingfailuretorestoretoserviceaftertestormaintenance,andexamplesillustratingthisprocess.SeeSection7.4.1andAppendixF,TableF-3.ThesubmittalisunclearondetailsofthequantitativescreeningapproachusedforHFEsinvolvingrestorationofequipmentandinstrumentmiscalibration.InSection3.3.3,onpage3.3.3-1,thesubmittalnotesthatalLHFEswereinitiallyquantifiedwithscreeningvalues.AreviewofTable3.3.3-4indicatesthatallpre-initiatorshadahwnanerrorprobability(HEP)of0.003.However,adiscussionofthebasisforthisvalueisnotprovided.'leaseprovidetherationaleforthechoiceofthescreeningvalue

Rest)onsctoFebruary23,1996RAIanddiscusswhetheranyadditionalanalysesofpre-initiatorswereconducted.Inaddition,providetherationaleforhowtheselectedscreeningvaluedidnoteliminate(ortruncate)importanthumanevents.Finally,ifTable3.3.3-4doesnotpresentallthepre-initiatorsmodeled,pleaseprovideaListofeventsthatwerescreened.SeeSection7.4.1andAppendixF,TableF-3.IfTable3.3.3-4presentsallthepre-initiatoreventsmodeled,itisnotclearwhyaretherenoeventsrepresentingmiscalibrationofleveltransmitters.Sucheventsareusuallymodeledinprobabilisticriskassessmentsofnuclearpowerplantsandinsomecasesarefoundtobeimportant.PleaseprovideadiscussionofwhytheseeventswerenotmodeledintheR.E.GinnaIndividualPlantExamination(IPE).SeeSection7.4.1andAppendixF,TableF-3.Asdiscussedearlier,miscalibrationissuesaretypicallyaddressedwithinthefailuredatafortransmitters,indicators,etc(see"failslow,""failshigh,"and"failstorespond"inTable7-1).Thesubmittalisunclearonhowdependenciesassociatedwithpre-initiatorhumanerrors(restorationfaultsandinstrumentmiscalibrations)wereaddressedandtreated.Thereareseveralwaysdependenciescanbetreated.In.thefirstexample,theprobabilityofthesubsequenthumaneventsisinfluencedbytheprobabilityofthefirstevent.Forexample,intherestorationofseveralvalves,aboltmustbe"tightened."Itisjudgedthatiftheoperatorfailsto"tighten"theboltonthefirstvalve,hewillsubsequentlyfailontheremainingvalves.Inthisexample,subsequentHEPSinthemodel(i.e.,representingthesecondvalve)willbeadjustedtoreflectthisdependence.Inthesecondexan>pie,poorlightingcanresultinincreasingthelikelihoodofunrelatedhumanevents;thatis,thepoorlightingconditioncanaffecttheabilitiesofdifferentoperators'oproperlycalibrateortoproperlyrestoreacomponenttoservice,althoughtheseeventsaregovernedbydifferentproceduresandperfor'medbydifferentpersonnel.ThistypeofdependencyistypicallyincorporatedintheHRAmodelby"grouping"thecomponentssothattheyfailsimultaneously.Inthethirdexample,pressuresensorsxandymaybecalibratedusingdifferentprocedures.However,iftheproceduresarepoorlywrittensuchthatmiscalibrationislikelyonbothsensorsxandy,theneachindividualHEPinthemodelrepresentingcalibrationofthepressuresensorscanbeadjustedindividuallytoreflectthequalityoftheprocedures.Pleaseprovidethefollowinginformationconcerningthetreatntentofpre-initiatordependencies:(a)Aconcisediscussionofhowdependencies(andhumanactioncommoncausefactors,whereappropriate)wereaddressedandtreatedinthepre-initiatorHRA.Dependencieswerenotspecificallyaddressedwithinpre-initiatorhumanerrors.AsdescribedinAppendixF,TableF-3,theavailableindicationto 0

RcsponsctoFcbntary23,1996RAIoperatorsofthestatusofinstrumentationandmajorequipmenteliminatesthisconcerntoalargedegree.Inaddition,thecommoncausefailuresincludedwithinthemodel(seeTable7-3)specificallyaddressthepotentialconsequencesofanycommonfailuremechanisms.(b)Specificexamplesillustratinghowdependencieswereconsideredforpre-initiatoreventsmodeledinthe1PE.Seeresponseto(a)above.(c)Efdependenciesandhumanactioncommon-causeissueswerenotaddressedforbothmiscalibrationsandrestorationevents,pleasejustify.Seeresponseto(a)above.

RcsponscioFebruary23,1996RAIPOST-INITIATORHUMANERRORSIThesubmittaldistinguisheshumanfailurepost-initiatoreventsfromrecoveryactionsbutdoesnotclearlydescribethemethodusedtoidentifyandselectpost-initiatorhumanfailureeventsforanalysis(onlyareferencetothe"HRATaskProcedure"isprovided).Themethodutilizedshouldconfirmthattheplantemergencyprocedures,design,operations,andmaintenanceandsurveillanceprocedureswereexaminedandunderstoodtoidentifypotentialsevereaccidentsequences.Pleasedescribetheprocessthatwasusedforidentifyingandselectingthepost-initiatorhumanfailureeventsincludedintheeventandfaulttreemodels.Section4addressesthesuccesscriteriausedinthePSAwithrespecttofourcoreprotectionfunctionsandthespecificneedforoperatoractionforeachofthesefunctions(seeTable4-10).Section6describesallhumanactionsrelatedtoeachsystem.Finally,Table7-15andAppendixF,TableF-4describethepost-initiatorhumanerrorsindetail.Thesubmittalisunclearonthebasisforthequantitativescreeningapproachthatwasusedforpost-initiatorhumanfailureevents.Table3.1.1-12(orTable3.1.2-12)indicatesthatscreeningvaluesof0.1,1.0,0.21,andIE-4wereused.Pleaseprovidethefollowing:(a)Thebasisforthescreeningvalue(s)usedandtherationalethatledtoassigningagivenhumanactionaparticularscreeningvalue.Pleaseuseseveralexamples(atleasttwoexamplesforeachofthefourscreeningvaluesused)toillustratehowitwasdeterminedthataparticularactionwould,beassignedoneofthefourvalues.SeeSection7.4.2.IntherevisedPSA,theonlyscreeningvalueusedwas0.1.(b)Inadditiontotheexamplesusedabove,pleaseprovidetherationaleforthescreeningvaluesassignedtothefollowingevents:(1)AFHFDSAFWX-operatorsfailtostartstandbyauxiliaryfeedwater(SAFW)PumpICandID;(2)AFHFDO4297-operatorsfailtocloseair-operatedvalve4297toisolatesteamgenerator(SIG)A,and(3)RCHFDOLBAF-operatorsfailtoinitiatefeedandbleed.EventsAFHFDSAFWX,MSHFDISOLR(versusAFHFD04297)andRCHFD01BAF(versusRCHFDOLBAF)allusedascreeningvalueof0.1.TheirfinalvaluesarediscussedinAppendixF,TableF-4.(c)Providearationaleforhowtheselectedscreeningvalue(s)ensuredthat IL ResponsetoFebruary23,1996RA[importantpost-initiatorhumaneventswerenoteliminatedandlorimportantsequencestruncated.SeeSection7.4.2.Pleaseprovideadetaileddiscussionofthebasisforassumingthatanactionwastimeindependent.Apparently,anactionwithmorethananhouravailablewasassumedtobetimeindependent.Fortheseactions,whatdid"timeavailable"referto?WerethetimerequiredtoperformtheactionandthetemporaLoccurrenceofcuesrelevanttoacorrectdiagnosisconsidered?Thisassumptionisnolongerused(seeSection7.4.2).AppendixF,TableF-4providesadescriptionofthetimesusedforhumanactions.ThevaluesfromTechniquesforHumanErrorRatePredictionPERP)thatwereusedtoquantifythetime-independenteventsappeartoignorepotentialdiagnosiserrorsandtheassociatedperformanceshapingfactors(PSFS)thatmightinfluencesuchdiagnoses.Inaddition,Chapter15oftheTHERPmethodologydiscussesthetablefromwhichthetime-independentHEPSwereapparentlytakeriandnotesthatthevalues.maynotbeappropriatewhensymptom-basedproceduresareused.PleasediscussindetailwhyitwasunnecessarytoconsiderpotentialdiagnosiserrorsandtheassociatedPSFsthatmightinfluencesuchdiagnosesinquantifyingtime-independentevents.Also,provide.,thebasisfortheuseofvaluesfromTable15-3oftheTHERPwhensymptom-basedproceduresarebeingused.TheTHERPapproachisnolongerused;insteadtheASEPmethodperNUREG/CR-4772wasused(seeSection7.4.2).Indiscussingthetime-independentquantijicationtechniqueinSection3.3.3.1.6onpage3.3.3-3ofthesubmittal,itisstatedinthethirdparagraphthat"typically,thebasicvaluesgivenabovewerereducedbyafactorofthreeinordertoaccountfordependenciesbetweenevents."Pleaseexplainwhatismeantbythisstatementandillustratehowthereductionisusedtoaccountfordependencies.Pleaseprovideseveralexamplesthatillustratetheprocess.Thisapproachisnolongerusedfordependicies(seeSection7.4.2.).Instead,allcutsetswithmultiplepost-initiatoroperatoractionswereidentifiedandfurtherevaluatedtodetermineifthehumanfailureprobabilitiesshouldbeadjusted(seeTables7-14and7-,15).ThesubmittaLisunclearonhowthe"time-dependent"quantificationtechniquewasappliedtothosepost-initiatorhunraneventssurvivinginitiaLsequencequantification.

RcsponsctoFcbniory23,1996RAIThesubmittalpresentstwo"time-dependent"quantificationtables(Tables3.3.3-land3.3.3-2),whichwereusedtogenerateHEPSforhumaneventsdependingonwhetheragivenactioncouldbeconsideredrule-basedwithhesitationorrulebasedwithouthesitation.Pleasedescribethemeaningoftheparameterslistedatthetopofthesetables(m=2min.,EF.EFJandthedigits0through9atthetopofthecolumnsofthetables.So,providethefollowing:(a)..Usingthreeormoreexamples,pleaseillustratehowthevariousparametersofthetableswereconsideredindeterminingHEPS.Pleaseprovideexamplesthatillustratewhetherthevaluesinthecolumnheadingswererelevantanddiscuss'owitwasdeterminedwhetherornotaparticularactionwas"withorwithouthesitancy."Thisapproachis'nolongerused(seeSection7.4.2).Onpage3.3.3-7,itisnotedthatfortheoperatoractiontocooldowntoresidualheatremoval(RHR)aftersafetyinjectionfails,"explicitguidanceonproceduretransitionsisnotprovided"andthatsuccessfulperformancewouldrequirea"circumvention."Yet,theHEPof1.8E-3wouldseemtobeoptimisticforanactionwithoutclearproceduralguidance.PleaseprovideadetaileddescriptionofthederivationofthisHEPandajustijicationforwhatappearstobeanoptimisticHEP.Thisapproachisnolongerused(seeSection7.4.2).Also,theabovestatementisincorrectinthatthereisproceduralguidanceforthisaction(seeAppendixF,TableF-4,eventsRCHFDCDOSSandRCHFDCDTR2).Thesubmittalisunclearonwhatplant-specificPSFswereconsider@iindeterminingHEPSfortime-independentandtime-dependenthumanactions.Thisplant-specificinformationcouldincludethesizeofthecrew,theavailabilityofprocedures,andthetraining,stress,andhumanfactorsaspectsofthecontrolroom,andsoon.OnthebasisofthediscussioninSection3.3.3.2,itwouldappearthatmayofthesetypesofplantspecificPSFswerenotexplicitlyconsdered.Ifanyofthesetypesoffactorswereconsidered,pleaseprovideaListandshow(byexample)howtheirinfluencewasfactoredintodeterminingtheHEPSforthevariousevents.(Includeexamplesofbothtime-independentandtime-dependentevents.)Ifnoneofthesefactorswereconsidered,pieazprovideaconcisediscussionofthe,rationaleandjustificationfornotconsideringsuchplant-specificinformationduringthequantificationofpost-initiatorhumanactions.ThistypeofinformationwasincorporatedusingtheASHPmethodology(seeSection7.4.2).

ResponsetoFebruary23,1996RAI8.HRAmethods,ingeneral,attempttoronsiderboththediagnosisportionorphaseofpost-initiatoroperatoractionsandtheexecutiondemandsoftheaction.Pleasediscusshowthesetwodifferentaspectsofhumanfailureeventswereconsideredindeterminingpost-initiatorhumanfailureprobabilitieswiththetime-dependenttechnique.IftheresponseexecutionphaseoftheactionandtheassociatedPSFsarenotexplicitlyconsideredpleaseprovideajustificationforhowthevaluesobtainedwiththetime-dependenttechniqueaccuratelyreflecthumanfailureprobability.Boththediagnosisandexecutionrequirementsofeachhumanactionwereconsidered(seeSection7.4.2).9.Onpage33.3-2,thesubmittalstatesthat"estimatesfortherequiredtimingofoperatoractionsweredeterminedwiththeassistanceoftheAccidentSequenceAnalysisTaskLeader."Pleaseprovideadetaileddiscussionoftheprocessfordeterminingthetimerequiredtocompleteoperatoractionsandindicatehowitwasensuredthattheresultingestimateswerenotoverlyoptimistic.ThetimingsusedasdiscussedinhumanactionsaredescribedinAppendixF,TableF-410.Itisnotclearfromthesubmittalhowdependencieswereaddressedandtreatedinthepost-initiatorHRA.Theperformanceoftheoperatorisbothdependentontheaccidentunderprogressionandthepastperformanceof.theoperatorduringtheaccidentofconcern.Impropertreatmentofthesedependenciescanresultintheeliminationofpotentiallydominantaccidentsequencesand,therefore,theidentificationofsignificantevents.Pleaseprovideaconcisediscussionandexamplesillustratinghowdependencieswereaddressedandtreatedinthepost-initiatorHRAforalltypesofactionstoensurethatimportantaccidentsequenceswerenoteliminated.Thediscussionshouldaddressthetwofollowingpoints:Humaneventsaremodeledinthefaulttreesasbasiceventssuchasfailuretomanuallyactuate.Theprobabilityoftheoperatorisperformingthisfunctionisdependentontheaccidentinprogression-whatsymptomsareoccurring,whatotheractivitiesarebeingperformed(successfidlyandunsuccessfully),andsoon.Whenthesequencesarequantified,thisbasiceventcanappearnotonlyindifferentsequencesbutindiferentcombinationswithdifferentsystemsfailures.Inaddition,thebasiceventcanpotentiallybemultipliedbyotherhumaneventswhenthesequencesthatshouldbeevaluatedfordependenteffectsarequantified.Humaneventsaremodeledintheeventtreesastopevents.Theprobabilityoftheoperator'speformingthisfunctionisstilldependentontheaccidentprogression.Thequantificationofthehumaneventsneedstoconsiderthedifferentsequences'22-RcsponsctoFebruary23,1996RAIandtheotherhumanevents.SeeSection7.4.2.11.ThesubmittaLstatesthatthespecificHFEsthatrequireddetailedanalysiswereall,associatedwithin-controlroomactions.Yet,atleastonesuchaction(CTHFDISOLA-operatoractiontoisolaterupturedSGEMSOlA)apparentlyalsorequiredactionstobeperformedoutsidethecontrolroom.Inthediscussionofthiseventonpage3.3.3-5,itseemsthatthepotentialfailureofactionsoutsidethecontrolroomwereignored.Pleasediscusshowex-controlroomactionswerequantifiedforthiseventorprovideajustificationforwhyitMasunnecessarytodoso.ProvidethesamediscussionforanyotherHFEsthatincludedex-controlrooniactions.Bothin-controlroomandex-controlroomactionsweremodeled.Inaddition,allactivitiesrequiredbyoperatorsineitherscenariowereconsideredingeneratingthefinalhumanerrorfailureprobability(seeSection7.4.2andAppendixF,TableF-4).Thesubmittalisunclearonhowrecoveryactionsmrequantified.Althoughthediscussiononpage3.3.7-12regardingthedeterminationoftheindicesforperformanceinfluencingfactorsisclear,thebasisforthe"multi-factoredapproach"isnotprovided.Thatis,howwasitensuredthatthesummationoftheindicesandtheinsertionoftheoverallindexintotheformulaonpage3.3.7.12producedvalidestimateofhumanfailureprobability?ThereisnoevidenceprovidedthatthemethodhasbeenusedoutsideoftheGinnaIPEandnoindicationthatithasbeenpeerreviewedor"benchmarked"inanyway.Pleaseprovideadiscussionaddressingthevalidityofthequantificationapproach.Inaddition,pleaseprovidethefollowing:(a)AtleastfourexamplesthatillustrateallaspectsoftheapplicationofthequantificationtechniquecorrespondingtotherecoveryeventsnrodeledintheIPE.InparticuIar,ilimtratehowHEPSarederivedusingtheformulaonpage3.3.7.12.Fortwooftheexamples,addressevents(1)PRFISOALTCD-failuretocooldownaftersteamgeneratortuberuptureusingsteamdumporrupturedSIGand(2)NRHLETDOWN-failuretolocallyisolateletdownvalveAOV-371using204A.Thisapproachisnolongerbeingused.AsdescribedinSection7.4.2,recoveryeventsweretreatedthesameaspost-initiatorhumanerrors.Inmostcases,recoveryeventswereaddedtothefaulttreemodelsinordertoallowtheeventstobeaddedonlytothecorrectscenarios.Withrespecttothetwoeventsabove,thesearedescribedinAppendixF,TableF-4,eventsRCHFDCDTR2andCVHFD00371,respectively.(b)Abriefdescriptionofeachoftherecoveryeventsmodeled(apparentlysevenof 1<csponsctoFcbniory23,1996RAIthem)andtheHEPSassignedtothesehumanactions.Asdescribedabove,recoveryactionsweretreatedthesameaspost-initiatorhumanerrors.SeeAppendixF,TableF-4.13.GuidancefromNUREG-1335requeststheidentificationofcoredamagesequencesthatdropbelowthecoredamagefrequency(CDF)screeningcriteriabecausethefrequencywasreducedbymorethananorderofmagnitudebytakingcreditforoperatoractions.Inaddition,informationwasalsorequestedonthetimingandcomplexityoftheassociatedhumanactions.Pleaseidentifytherelevantsequencesandprovideadiscussionoftherelatedoperatoractions.SeeSections9.1.2,9.3.1.1,and9.3.2.2.14.Onpage3.3.8-17,thesubmittaLnotesthatthesametechniquesusedtorecoverinternallyinitiatingsequenceswereusedtorecoverflood-relatedsequencesaefthatpreviouslyrefinedHFEprobabilitieswereincorporated.WasitnecessarytomodifyanyoftheexistingHFEstoreflectfloodingconditions?WereanyHFEsaddedtoaddresspotentiaLhumanisolation.offloodsources?Iftheanswertoeitherquestionisyes,usingexamplespleaseaddresshowtheoperatoractionswerequantified.IftheexistingHFEHEPSwerenotmodifiedtoreflectfloodingconditionsorflood-specifihumanactionswerenotincluded,pleasediscusswhyitwasunnecessarytodoso.AresponsewillbeprovidedbySeptember30,1997followingre-analysisofthefloodingPSA.15.ThesubmittalisunclearonwhathumanreliabilityanalysiswasperfumedduringtheLevel2analysis.PleaseprovidethefollowingregardingtheHRAfortheLevel2-analysis:(a)Onpage4-11ofthesubmittal(lastparagraph),itisimpliedthattherecoverymeasuresconsideredintheLevel1analysisaregenerallyappliedintheLevel2analysis.PleasediscusshowthiswasdoneandprovidealistoftherelevantrecoveryactionsandtheirassocritedHEPS.IftheHEPSdigeredfromthoseusedintheLevel1analysis,pleasedescribehowtheHEPSwerecalculated.AresponsewillbeprovidedbyMay1,1997followingthenewanalysisoftheLevel2PSA.(b)PleaselistanyadditionaloperatorirecoveryactionsconsideredintheLevel2analysis(e.g.,"certaincontainmentisolationrecoveries"Pnddescribethetechniqueusedtoquantifytheevent(s)throughexamples.

ResponsetoFebruary23,1996RAIAresponsewillbeprovidedbyMay,1,1997followingthenewanalysisoftheLevel2PSA.

RcsponscloFebruary23,1996RAIENCLOSURE3LEVEL2QUESTIONSContainmentWallLinerIsolation.AccordingtotheGinnaIndependentPlantExamination(IPE)submittal,thereisa1.25-inchthicklinerinsulationonthesidewallstoapoint15feetabovethespringline.Thelinerinsulationisaclosed-cellpolyvinylchloridefoanrinsulationwithlowconductivity,lowwaterabsorption,andhighstrengthandiscoveredwithmetalsheeting.Pleasediscusshowthisinsulationismodeledinthemodularaccidentanalysisprogram(MAAP)codemodelandtheeffectthisinsulationhasontheabilityofthecontainmentstructuretoabsorbheat.TheinsulationwasmodelledintheMAAPcodebydevelopingacontainmentliner/wall"gapresistance"thatofferedequivalentresistancetoheatflow.ThenominalvalueforthethicknessandthermalconductivityoftheinsulationwereobtainedfromtheUFSAR,andthegapresistancewascalculatedtobetheirratio.Forconservatism,thethermalpropertiesoftheinsulationwereassumednottodegradeundersevereaccidentconditions.SinceMAAP3.0Bonlyallowsoneouterwalltoexistineachcontainmentcontrolvolume,theuninsulatedportionoftheouterwallofthecontainmentwasrepresentedintheMAAP"uppercompartment"andtheinsulatedportionoftheouterwallwasrepresentedinthe~MAAP"annularcompartment".Sincetheupperandannularcompartmentsexperienceverysimilarthermal-hydraulicconditions,thisslightsimplificationdoesnotimpacttheresults.Thepresenceoftheinsulationlargelydecouplestheinsulatedpartsoftheouterwallfromthecontainmentatmosphereoverthetimeframesofinterestinsevereaccidentcalculations.Thisispartlyresponsiblefortherelativelyshortcontainmentfailuretimescalculatedforsequenceswithquenchedcoredebrisandnocontainmentheatremoval,ReactorCavityandDepthofDebrisintheSumpforaFloodedCavity.(a)ItisstatedintheGinnaIPEsubniittalthat"theconcretethicknessinthecavityawayfromthesumpis2.0feetabovetheimbeddedlinerand2.0feetbelowtheliner.Belowthecavitysumpthetotalthicknesscfthebasematconcreteis1.5feet."However,Figure4.1-1indicatesthatthetotalbasematthicknessinthecavityis2feet.Pleaseclarifythisapparentdiscrpancy.Ifthetotalthicknessis2feetinsteadof4feet,pleasediscusshowthiswouldacctyouranalyisofcontainmentbasematmeIt-throughresultingfromcore-concreteinteraction.Thethicknessoftheconcretebasematisasfollows:Inthereactorcavitysump-1.5feet ResponsetoFebn>ary23,1996RAIb.Inthereactorcavity(awayfromthesump)-2feetc.Thecontainmentflooroutsideofthecavity-2feetabovethelinerand2feetbelowtheliner.Sincethebasematwaspoureddirectlyontobedrock,basematmelt-throughisasomewhatarbitrarydefinitionofcontainmentfailure,andtherateatwhichfissionproductswouldbelostafterthedebrispenetratesthebasematisnotclear.Thebasematthicknessaffectstheanalysisonlyinthedefinitionofthetimingoflatecontainmentfailuresduetobasematmelt-through.Suchsequencesarerepresentedbysourcetermcategory(STC)case12,forwhichcontainmentfailurewasassumedtooccuratthetimewhen1.5feetofconcreteattackhadoccurred,aswouldbeappropriateforcasesinwhichdebrisentersthesump.Incaseswheredebrisdoesnotenterthesump;thetimingofcontainmentfailurewouldthusbeslightlyunderpredictedincaseSTC12(i.e.,isconservative);otherwise,theerrorinthesubmittaldoesnotaffecttheresults.TheGinnacavityconsistsofacylindricalportionwithanattachedrectangularvolume,whichcontainsthesump.AmodeldevelopedbyF.Moodywasusedinthe.GinnaIPEtoestiniatedebrisspreadingandtheprobabilityofthedepthofthedebrisinthesumpforafloodedcavitywithnosteamexplosion.Table4.5-7oftheIPEshowsthatif50percentofthetotalcoredebrisisreleaseduponvesselfailure,thedebrisspreadradiusis4.Imeters(m).Sincethedistancefromthecenterlineofthereactorvesseltotheclaestedgeofthecavitysumpis7.8m,itisstatedintheIPEthat"spreadingtocovertheeitirecavityforhighdebrismasses(40-60%)maybepossible,butnotlikely.Hence,aprobabilityof0.1isassignedtotheSUMPFULLbranchandaprobabilityof0.2isassignedtothePARTFULLbranch."However,the4.I-mdebrisspreadradiusissigficantlygreaterthanboththeradiusofthecylindricaLportionandthehalfwidthoftherectangularportionsofthecavityregion,whichisabout2m.Thetotalareacoveredbyacirclewitha4.1-mradiusisabout53m',whichisalsosignijicantPgreaterthanthetotaLcavityareaof29it'iPleasediscusshowtheeffectoftherestrictedspreadareainthereactorcavityisconsideredintheassignmentoftheprobabilityvaluesfordebrisdepthinthesumpandwhattheeffectofhigherprobabilityvaluesfordebrisinthesumpwouldbeoncontainmentfailureprobabilities.AfterinvestigatingtheassumptionsusedinthedevelopmentoftheMoodymodel,RGEcEagreeswiththeinferencebythereviewerthatitwouldhavebeenmoreappropriateinanirregulargeometrysuchastheGinnaStationreactorcavitytoreformulatethemodeltocalculatethespreadarearatherthanspreadr'adius.Ifthisisdone,themodelcalculatesthatdebrisshouldspreadtofillanareanearlytwicethatrepresentedbythecavityfloor.Thiswouldimplythatdebriscouldbe

ResponsetoFcbrunry23,1996RAIexpectedtoenterthecavitysump.However,itshouldalsobenotedthattheheatfluxbetweenthedebrisandthewateriscalculatedintheMoodymodelbyassumingthatenhancedfilmboilingoccursatthesurfaceofthedebris.TheparametersusedinthePSAanalysisweredrawnfromMoody'spaperandresultinanassumedheatfluxofabout800kW/m'.SmallscaleexperimentsperformedbyHenryetal.toaddresstheMarkIlinermelt-throughissuemeasuredheatfluxesduringtheperiodofsimulateddebrisdischargethatwereanorderofmagnitudelargerthanthisvaluef1].Ifsuchanaugmentationinheatfluxalsooccurredinthereactorcaseovertheperiodofdebrisdischarge,thiswouldoverwhelmtheproblemnotedbythereviewerandpreventthedebrisfromflowingtothesump.Theofanousetal.[2]hascriticizedtheexperimentsofHenryetal.,sincenoanalysisofdebrisflowregimewasconducted.Giventhepressurized(150psi)dischargeofsimulantcoredebrisintheexperiments,Theofanousetal..notedthatintensemixingcreatedbymeltsplashingoffthefloorandrunningupthesidewallscouldexplainthehighmeasuredheatfluxes.However,fortheMarkItransient(inwhichthewaterlevelisonly-0.9mdeep,muchlessthanthelevel.intheGinnaStationcavityiftheRWSTisdischarged)Theofanousetal.agreedthatrapidfragmentationandquenching-ofthecoredebriswouldbeexpectedasthedebristraversesthewaterpool,andthatalocalpileupofdebrisimmediatelyundertheRPVfailurelocationcouldoccur.TheprobabilityofthecavitysumpbeingeitherpartiallyorcompletelyfullofdebriswasestimatedinthePSAforthreeseparaterangesofdebrismassdischargedatvesselfailure:Fractionofdebrisdischarged1ProbabilityofProbabilityof<20percent>20and<40percent>'40and<60percent0.010.10.2Impossible0.050.1Consideringboththeirregularcavitygeometryandthelikelihoodofdebrisjetbreak-upandlargerdebris/waterheatfluxesduringthedebrisejectionphase,theseprobabilityvaluesarestillconsideredreasonable.Recognizingthatthiswasanuncertainissue,asensitivitycasewasrunaspartofthePSAinwhichitwasassumedthattheprobabilityofhavingafullsumpwasunity(exceptforthe-<20percentdischargeddebriscaseinwhichthesump RcsponsctoFcbniary23,1996RAIcannotfillandtheprobabilityofapartlyfullsumpwasinsteadconsideredunity).AsdiscussedinSection4.8.1.4andTable4.8-4oftheMarch15,1994submittal,theeffectofthischangewastoincreasetheprobabilityofsourcetermcategory(STC)12(whichrepresentslatecontainmentfailureduetobasematmelt-through)from13percentto30percentofthetotalCDF;thenocontainmentfailureSTCwasreducedbythesameamounttoabout16percentofCDF.Notethatevenifheattransferwasassessedusingthefilmboilingassumption,theprobabilityofafullcavityinthemediumdebrismasscaseswouldbelessthanunity,andthissensitivitycalculationthereforesomewhatoverstatesthemagnitudeoftheeffectnotedbythereviewer.Itisalsoworthnotingthatlatebasematmelt-throughsequencessuchasSTC12havelowsourcetermsandthatthefissionproductreleasesintheGinnaStationPSAweredominatedbycontainmentbypasssequenceswhichareunaffectedbythespreadingissue(seeQuestion9).AlthoughthecalculationsfordebrisspreadintheMarch15,1994submittalweredeterminedtobenon~nservative,considerationoftheseotherfactorsleadsRGkEtobelievethatourresultsareacceptable.Cutset(CSET)StructureandPowerRecovery.AccordingtotheIPEsubmittal,theCSETtreestructureofthebranchforrecoveryofacpowerpriortovsIfailure.(PRV)isidenticaltothatshowninthetreestructureleadingtoendpoints1-46inFigure4.3-1.Ontheotherhand,thestructureofthebranchfortherecoveryofpowerpriortocontainmentfailure(PRC)isassumedtobeidenticaltothatIeadingtoendpoints1-8.Figure4.3-1showsthatallcoreinjectionandrecirculationsystems(i.e.,low-pressureinjectioendrecirculationandhigh-pressureinjectionandrecirculation)areavailableforendpoints1-8;whileoneormoreofthesesystenisarenotavailableforendpoints9to36:PleasediscusswhyallthesesystenisareassumedtobeavailableforthecasesofpowerrecoverypriortocontainmentfailurebutsomeofthesystemsmaynotbeavailableforthecasesofpowerrecoverypriortovesseLfailure.InFigure4.3-1oftheMarch15,1994submittal,branch49showsablackoutsequencewherepowerisnotrecoveredatall,branch48showsablackoutsequencewherepowerisrecoveredpriortocontainmentfailureandbranch47showsablackoutsequencewherepowerisrecoveredpriortovesselfailure.Sincebranches1-46arenon-blackoutsequences,powerrecoveryisnotanissueinanyofthesebranchesandthequestionsofwhatinjectionsystemsareavailableareinsteadaddressed.LxternalCoolingoftheReactorPressureVesseLItisstatedinthesubmittal(page4-4)that"wheneverthecontentsoftheRWST(refuelingwaterstoragetankJareinjectedintothecontainmentthecavitywillbecompletelyfilledandwillremainfilled."Basedonthisstatement,itseeinspossiblethattheex-vesselwater]nayprovidesuffVcientcoolingtothecoredebrisinsidethevesselsothatvesselfailurecouldbeavoidedorsignijicantlydelayed.Asaresult,fissionproductproductionandreleasepathscouldbeaffected(e.g.,in-vessel Rcsi)OI1SC10FC4fUllry23,1996RAIreleasefromadrydebrisbedversusex-vesselreleasefromadebrisbedcoveredbywater).ThereleaseoffissionproductstotheenvironmentmayactuallyincreaseifthecontainmentfailsandexternaLcoolingwasaccountedforinthesourcetermcalculation.PleasediscussthepotentiaLofex-vesselcoolingforGinnaanditseffectonsourcetermdefinition.Becauseexternalcoolingmaymaintainthereactorcoolantsystem(RCS)athightemperatureforalongertime,pleasealsodiscussthe@ctofexternaLvesselcoolingontheprobabilityofcreepruptureoftheRCSboundariesandthesteamgeneratortubesand,consequently,theeffectoncontainmentperformanceandsourcetermsforGinna.WhenthePSAwasinitiallyperformed,thesubjectofex-vesselcoolingofcoredebrisinthelowerheadwasbeingactivelystudied.Anotherareaofactiveresearchatthattimewastheexpectedmodeofvessel.failure,primarilythroughtheTMIVesselInspectionProgram(VIP).ThePSAanalystsjudgeditlikelythatvesselfailurewouldoccuratapenetrationsoonaftercorerelocation,whichwastheassumedmodeofvesselfailureintheIDCORprogram.Thismadeitunlikelythatthedebriscouldberetainedinthelowerhead,andforthisreason,nocreditwastakenfordebrisretentionintheRPV.ThisassumptionwasconsideredapotentialconservatismofthePSA.Morerecently,aconsensusappearstobedevelopingthatretentionofdebrisinthelower.headisfairlylikely.ForsequenceswithinjectionoftheRWST,andcontainmentheatremovalavailable,thiswouldservetopreventvesselfailureandthusthepotentialforcontainmentfailure.RG8cEisunawareofanycalculationsoffissionproductreleasefromdebrisbeingcooledinthelowerheadofthereactorvessel.However,suchreleaseswouldbeexpectedtobeverysmallinthelongterm,sincedebristemperatureswouldberelativelyconstant,the.surfaceareawouldbesmall,andmostimportantly,therewouldbenospargingofthedebrisbygassesasisthecaseforthein-coreandex-vesselfissionproductreleasemechanisms.Thus,coolingofthedebrisinthelowerheadisnotexpectedtorepresentanewsafetyissuefromthestandpointoffissionproductreleasefromdebris.WithrespecttoinducingthefailureofRCScomponents,inducedruptureofthehotlegsisexpectedtooccurlongbeforedebrisrelocation'othelowerheadinhighpressureaccidentsequences.HotlegrupturewoulddepressurizetheRCSandremoveanythreattothesteamgeneratortubes.RecentcalculationsperformedbybothNRCcontractorsaswellastheindustryindicatethathotlegorsurgelineruptureisalsolikelyinmediumpressuresequences.ForsequencesinvolvingRCSpressureslowenoughnottoresultinaninducedruptureofthehotlegorsurgelinepriortocorerelocation,subsequentthermally-inducedruptureofthesteamgeneratortubesisonlycredibleandofconcernif:a.Thesequenceisnotabypasseventandthesteamgeneratorsareinitiallyintact(i.e.sequenceisnotinitiatedbyasteamgeneratortuberupture);

RcsponsctoFebruary23,1996RAIb.OneorbothsteamgeneratorsaredepressurizedtopressureslessthantheRCSpressure(otherwisethesteamgeneratortubeswouldbeincompressionratherthantension);c.TheRCSpressureissignificantlygreaterthanthesecondarysidepressuresoastothreatenthetubesbutsufficientlysmallasnottohavealreadycausedaninducedruptureofthehotlegorsurgeline;d.'heaffectedsteamgenerator(s)secondarysideisdrysothatthetubesarenotcooled(otherwisethermally-inducedrupturecannotoccur);ande.Thereisnosupplyofwatertothedebris(acontinuoussupplyofwater,i.e.fromaRHRpump,wouldcoverthedebrisandcooltheRCS).PastexperienceindicatesthatthemostlikelyscenariosfittingthesecriteriaaretotallossoffeedwatersequencesinvolvingLOCA(s)(e.g,largeRCPsealLOCAsoropenpressurizersafetyorreliefvalves)whichalsoresultinoneormoredepressurizedsteamgenerators.GiventhepresenceofLOCA(s),onewouldnotexpectRCSpressuretobemaintainedsignificantlyabovethesteamgeneratorpressureafterthewaterinitiallyinthelowerheadorthatsuppliedbytheaccumulatorsisboiledawayandtheRCSdepressurizes.Forthisreason,itisbelievedthatinducedruptureofthesteamgeneratortubescausedbydebrisretentioninthelowerheadisveryunlikely.Therefore,weconcludethatfailingtocreditdebrisretentioninthelowerheadoftheRPVrepresentsaconservatismofthePSAratherthanaconcern.TheAvailabilityofContainmentFanCoolersandContainmentSprays.IntheIPEmodel,ifthecontainmentfancoolersareavailable,containmentspraysinboththeinjectionandrecirculationmodesaremodeledasfailed(page4-13).Althoughbothsystemshavethesameeffectoncontainmentheatremoval,theireffectonsourcetermdefinitionmaybediferent.Spraysareusuallycreditedwithbeingmoreejlicmtthanfancoolersinremovingfissionproductsfromthecontainmentatmosphere.PleasestatewhethercontainmentspraysareincludedintheMAAPmodelforsourcetermcalculation(forcasesinwhichcontainmentheatremovalisavailable).Iftheyare,pleasejustifytheuseofcontainmentspraysforsourcetermcalculationsinallsequenceswithontainmentheatremovaLIftheyarenot,pleaseestimatethepotentialsourcetermmitigationeffectachievedbytheoperationofcontainmentsprays.NoneoftheMAAPcalculationsofrepresentativesourceterms(MAAPcalculationsSTCl-STC20)creditedthesprays,consistentwiththeassumptionsmadeinthesequencequantification.Forsequencesinwhichthedebrisiswater-coveredinthecontainmentandfancoolerswerecredited,theonlysubstantialbenefitfromassumingthatthespraysalso

ResponsetoFebruary23,1996RAIoperatewouldbeexpectedinnon-containmentbypasssequencesinvolvingearlycontainmentfailure(especiallythosewithlargecontainmentbreechareas).Thehighrateofremovalofaerosolsaffordedbythesprayscouldbeexpectedtoreducethesourcetermsinthesecasessincethefissionproductreleaseisdeterminedbytherelativemagnitudeoftheratesofaerosolleakageanddeposition.Bypasssequencesareunaffectedbycontainmentsprays(whichusuallydonotoperateanyway),andfancoolerswouldhavesufficienttimetobeeffectiveinsequenceswithlatecontainmentfailure.ThesourcetermcategoriesinvolvingearlycontainmentfailurewithcontainmentheatremovalareSTCs2,3,.5,13,and14.Ofthese,STC2and3involvesmallcontainmentleakareasandonlySTC13hasaconditionalprobabilitygreaterthan0.1percent.STC13represents3percentoftheCDFandiscausedbyafailureofcontainmentisolation.Thecalculatedsourcetermsforthissequence,showninTable4.7-2oftheMarch15,1994submittal,aremoderate,butaremuchlessthanthose.resultingfromcontainmentbypassscenarios(STC16,18,and20)whicharealsomuchmorelikely(over40percentofCDF).Thus,whileneglectofthespraysrepresentsaconservatism,primarilyinthatSTC13mayhavetoolargeasourceterm,inclusionofthesprayswouldhaveanegligibleeffectontheresultsasawhole.Iftheseresultschangesignificantlyduetothere-quantificationoftheLevel1sequences,operationofthespraysmayneedtobeevaluatedintheLevel2update.SteamGeneratorTubeRupture(SGTR)Releases.TheIPEresultsshowthatthesteamgeneratorpower-operatedreliefvalve(PORV)cyclesduringandaftercoredamagein'abouthalfoftheSGTRevents,andthereisastuck-openPORVintheremaininghalfoftheSGTRevents.Pleasediscusshowtheprobabilityofsteamgenerator(SG)valvefailureisdeterminedintheanalysisandhowtheeffectoftheharshconditions(e.g.,theflowofextremelyhightemperaturegaseswithentraineddebris)ontheoperationoftheSGvalvesisconsideredintheanalysis.ThesteamgeneratorPORVoratmosphericreliefvalve(ARV)failureprobabilityisdeterminedintheLevel1dataanalysisportionofthePSA.NoadjustmenttotheARVfailureprobabilitywasmadebasedonharshconditions;however,thiswillbeconsideredintheupdatetotheLevel2PSA.TheProbabilityofPowerRecovery.AccordingtoTable4.3-2oftheIPEsubmittaL,theprobabilityofpowerrecoveryprioItovesseLfailureis0.622,andtheprobabilityofpowerrecoveryaftervesselfailurebutpriortocontainmentfailureis0.127.However,theplantdamagestate(PDS)resultsshowthatamongallstationblackoutsequences(includedinPDS2through8),S4percenthavepowerrecoveryaftervesselfailurebutpriortocontainmentfailure(PDS7).TherecoveryprobabilityofO.S4ismuchgreaterthanthevalueobtainedinthepowerre'coveryanalysis(0.127).Pleaseexplainthisapparentdiscrepancy.

RcsponsctoI'cbrnnry23,1996RAITheprobabilityofpowerrecoveryfortheLevel2analysiswasbasedontheLevel1powerrecoverycurve.Thishassincebeenre-performed(seeAppendixC).ThiscurvewillbeusedwhentheLevel2portionofthePSAisre-quantified.CapacityoftheContainmentVessel.Section4.4oftheIPEsubmittaldiscussestheevaluationofcontainmentultimatestrength.AccordingtotheIPEsubmittal,theultimatestrengthandfailuremodesoftheGinnacontainmentweredeterminedwhyafiniteelementanalysisperformedbyEbascoServices.However,thecriteriausedtodeterminetheultimatefailurepressurearenotdiscussedinthesubmittal.IntheIPE,thefailurepressuresobtainedwereassignedanuncertaintyof5percentforcontainmentfailureevaluation.The5percentuncertaintyusedintheGinnaIPEseenistobelessthanthatusedinotherIPEs.Thedifferencebetweenthe5thpercentilefailurepressureandthemedianfailurepressureforGinnaismuchlessthanthatfoundinotherIPEsorintheNUREG-II50analyses.Pleasediscussthecriteriausedtodeterminethecontainmentfailurepressuresandthebasisforthe5pecentuncertaintyassociatedwiththesepressures.TheEbascostudyseparatelyconsideredgrossfailureofthecontainmentandlocalfailurescausedbylinertearsatcontainmentpenetrations.ThestudyusedtheABAQUSandHEATING6computercodes,andwasbasedonminimumsteelandconcretepropertiesasdefinedintheUFSARandrelevantmaterialspecifications.Grosscontainmentfailurewasassociatedwitharapidincreaseinplasticstrainat155psia;specifically,thisfailurepressurecorrespondstotheachievementof3percentstraininthecircumferentialreinforcementatmid-containmentheight.Linertearingwasevaluatedbyestimatingthefailurestrainnearpenetrations.Suchlocationsareempiricallyassociatedwithathree-dimensionalstateofstress,forwhichthevalueofthe"triaxialityfactor"canbeboundedby5.Thetriaxialityfactorallowstheeffectivefailurestraintoberelatedtotheknownuniaxialelongationfailurestrainmeasuredforthelinersteel.Inthiscase,theuniaxialfailurestrainof9.8percentresultsinaneffectivefailurestrainof0.6percent.Suchastrainwasobservedinthefiniteelementresultsatapressureof145psia;thecalculatedstrainincreasesrapidlyabovethispressure,soanyerrorsintheeffectivefailurestrainwillresultinonlyasmallchangeinfailurepressureassociatedwithlinertears.Ebascojudgedthattheuncertaintyintheresultsofthefiniteelementanalysisitselfwas5percent.AsdiscussedinSection4.4.3oftheMarch15,1994submittal,thisvaluewastakentobethestandarddeviationofanormaldistributioncenteredaroundthecalculatedfailurepressurewhichwastakento.bethemedian.Asstatedabove,theEbascoanalysiswasbasedonminimummaterialproperties,andnouncertaintyinmaterialpropertieswasconsideredintheanalysis.Forthisreason,thefragilitycurvedevelopedinthePSAshouldberegardedasarepresentationoftheuncertaintyintheminimumfailurepressurethatisconsistentwiththevariousmaterialsspecifications.

RcsponsotoFebruary23,1996RAIQualitatively,considerationofmaterialpropertyuncertaintieswouldshiftthefragilitycurvetowardhigherpressuresandwouldbroadenthedistribution;theexpectedmagnitudeofthebroadeningisnotknown.TheGinnaStationcompositefragilitycurve,Figure4.4-3,spansapressuredifferenceofabout12psidingoingfromthefifthpercentiletothemedian.ComparisonoftheGinnaStationresultstothoseshownforZioninFigure1-1ofReferencef3]indicatesthatthismeasureofuncertaintyintheGinnaStationcurveisactuallylargerthanthatestimatedbyoneanalyst(expertC),comparabletothatestimatedbyasecond(expertA),andmuchlessthanthatestimatedbyathird(expertB).Basedonthiscomparison,itisbelievedthattheGinnaStationfragilitycurveisconservative,butnottoasubstantialdegree.AsnotedinSection4.8.1.3,sensitivitycalculationsindicatethatthePSAresultsareinsensitivetoreasonablechangesinthefragilitycurve.ContainmentIsolationFailure.-AccordingtotheIPEsubmittaL,lossofcontainmentisolationsequencesrepresents3.0percentofthetotalcuedamagefrequency(CDF),or5.2percentofthefrequencyfornoncontainment-bypassCDF.Section3.2.1.3providesadescriptionofthecontainmentisolationsystemandtheoperatingexperienceofthecontainmentisolationsystemattheGinnaplant.Section4.3.1.2.4mentionsthatafaulttreewasusedforcontainmentisolationquantification.However,detailsoftheanalysisandtheresultsarenotprovidedinthesubmittal.Withrespecttotheanalysisofcontainmentisolationfailureprobability,NUREG-1335(Section2.2.2.5,page2-11)statesthat"theanalysesshouldaddressthefivea@asidentifiedintheGenericLetter,i.e.,(1)thepathwaysthatcouldsignificantlycontributetocontainmentisolationfailure,(2)thesignalsrequiredtoautomaticallyisolatethepenetrations,(3)thepotentialforgeneratingthesignalsforallinitiatingevents,(4)theexaminationofthetestingandniaintenanceprocedures,and(5)thequantificationofeachcontainmentisolationfailuremode(includingcommon-modefailure)."The5percentprobabilityofcontainmentisolationfailureissignificantlygreaterthanthatofmostIPES.Pleasediscussthesignificantcontainmentisolationmodes(e.g.,thepenetrationsthatfailtoisolateandthecausesforisolationfailure)obtainedjomtheIPEanalysisandforthesemajorcontainmentisolationmodesdiscusshowthefiveareaslistedabovewereaddressed.ThefiveareasrequiredtobeevaluatedbyNUREG-1335were,infact,consideredintheLevel2PSA.TheseareaswereevaluatedindetailintheContainmentIsolationSystemWorkPackageandsummarizedinSection3.2.1.3oftheMarch15,1994submittal.The5percentprobabilityofcontainmentisolationfailurewasduealmostentirelytoISLOCAsequences.Thesesequenceshavesincebeenre-evaluated(seeSection8.2ofthenewsubmittal)withsignificantlydifferentresults.ThesenewresultswillbeincludedintheforthcomingLevel2re-analysis.Additionaldetailsofthecontainmentisolationanalysiswillalsobeprovidedatthattime.

ResponsetoPebniory23,1996RAI10.PenetrationSealFailure.FailureofcontainmentpenetrationsisdismissedintheGinnaIPEasapotentialcontainmentfailuremodebecausetheanalysisinNUREG-1150indicatedthatthisfailuremodewassignificantlylessimportantthantheoverpressurefailureofthecontainmentcylinderwall(page4-28ofthesubmittal).PleaseprovideadescriptionofthesealmaterialsusedforthepenetrationsinGinna,theirproperties,andthepotentiaLharshcontainmentconditionstowhichtheycouldbeexposed.Onthebasisofthisplant-specificinformation,pleaseexplainhowyouconcludedthatthefindingsintheNIJREG-1150analysiscitedinthesubmittalwereapplicabletotheGinnaplant.11.SystematicEvaluationProgram(SEP)TopicVIII-4,ElectricalPenetrationsofReactorContainmentprovidesdetailedinformationaboutthepenetrationsealmaterialsandtheirproperties.Reference[4]providestheNRC'ssafetyevaluationofthistopic.UFSARSection3.11andtheMAAPrunsmadefortheLevel2analysisdetailthecontainmentconditionstowhichthecontainmentpenetrationsareexpectedtobeexposed.ThecontainmentpenetrationswillbeevaluatedusingthisinformationandtheinformationinNUREG-1037,"Containment'PerformanceWorkingGroupReport".TheresultsofthisevaluationwillbeincludedintheforthcomingLevel2re-analysis.ContainmentPerformanceImprovementandHydrogenIssues.Thegenericlettercontainmentperformanceimprovementrecommendationforpressurized-waterreactordrycontainmentsistheevaluationofcontainmentandequipmentvulnerabilitiestolocalizedhydrogencombustionandtheneedforimprovements(includingaccidentmanagementprocedures).Pleasediscusswhetherplantwalkdowninspectionshavebeenperformedtodeterminetheprobablelocationsofhydrogenreleasesintothecontainment.Includingtheuseofwakdowninspections,discusstheprocessused5assurethat(1)localdeflagrationswouldnottranslatetodetonationsgivenanunfavorablenearbygeometryand(2)thecontainmentboundary,includingpenetrations,wouldnotbechallengedbyhydrogenburns.Pleaseidentitypotentialreactorhydrogenreleasepointsandventpaths.Estimatesofcompartmeritfreevolumesandventpathflowareasshouldalsobeprovided.Specificallyaddresshowthisinformationisusedinyourassessmentofhydrogenpocketinganddetonation.Yourdiscussion(includingimportantassumptions)shouldcoverthelikelihoodoflocaldetonationandthepotentiaLformissilegenerationasaresultoflocaldetonation.AcontainmentwalkdownwasperformedonMay4,1992forthepurposeofassessingcontainmentfeaturesimportantforunderstandingtheresponseoftheGinnaStationcontainmentduringsevereaccidents.Onefocusofthewalkdownwastoestablishthemostimportantinter-compartmentgasflowpaths.Theresultsofthewalkdownindicatethat:a.Substantialglobalnaturalconvectioncurrentswouldbeexpectedinasevere

,0 Rcs~)ollsctoFcbnlary23,1996RAIaccident,drivenprimarilybyheatloadsfromtheRCSandwhereapplicable,fromdebrisinthecontainment,andbyheatremovalbythewallsandactivesafeguardssystems.GasflowisexpectedtooccurupwardinthelowercompartmentregionsurroundingtheRCScomponentsandthroughgratingoverthereactorcoolantpumpsandtheannularregionsaroundthesteamgeneratorsandpressurizer.Thegaswillflowintotheupperdomewhereitwillcoolbycontactwithuninsulatedheatsinks(andspraydropletswhereavailable),andflowdowntheannulusoutsidetheshieldwall.Fromtheoutercontainmentannulus,thegaswillflowbacktothelowercompartmentregionthroughtwostairwells,grating,andopenareasaroundHVACductinglocatedatthe253'levation.b.Thelimitingflowareaonthiscirculationloopisquitelarge,about450squarefeetf5].DrivensolelybythenominalheatlossesfromtheRCS,the.overallcontainmentmixingtimeiscalculatedtobeonly-6minutes[5].Inactuality,thiscalculationunderpredictscontainmentmixingsincetheeffectofcontainmentheatremovalsystemshasbeenneglected;sequenceswithoutcontainmentheatremovaloperationalwillresultintoohighasteamconcentrationtoresultindeflagrations,muchlessdeflagrationtodetonationtransition(DDT).TheprincipalhydrogenreleasepointstothecontainmentpriortoRPVfailureareRCSbreaksintheregionssurroundingthecoolantloopsorthepressurizerrelieftankrupturediskforsequenceswithnocyclingoropenpressurizersafetyorreliefvalves.Alloftheselocationsarewell-ventilated,havesubstantialflowareasavailabledirectlyov'erhead,andthusdonotposeathreatoflocalizedhydrogenbuild-up.Hydrogenreleasedtotheselocationswouldbeexpectedtobeentrainedintotheoverallcontainmentcirculationandmixedwiththeothergassesastheyflowintotheupperdomeregion.d.Diffusionflamesoccurringatanyofthesehydrogenreleasepointsinsidetheshieldwallwouldnotendangercontainmentpenetrations.Globalhydrogenburnswouldbeofshortdurationandwouldnotbeexpectedtothreatenthepenetrations.Ifoneassumesthathydrogenreleasedtocontainmentiswell-mixed,thepotentialforDDTcanbequalitativelyassessedbycalculatingthemaximumaveragehydrogenconcentrationthatcanbeachieved.Evenifallthezirconiuminthecoreisoxidized(amorerealisticupperboundisabout75percent),thehydrogenconcentrationinthecontainmentwouldbeonly11percentinthecompleteabsenceofsteam.Inactuality,somesteamwouldalwaysbeincontainmentduringasevereaccident.Also,itisveryunlikelythatsuchalargeamountofhydrogencouldbereleasedtocontainmentsoquicklyastoavoiddeflagrationsatlowerhydrogenconcentrations.Alloftheseconsiderationswouldlimitthepeakachievablehydrogenconcentration.Inanyevent,11percentistoosmalltosupportDDT,evenforgeometriesthatpromoteflameacceleration.

ResponsetoFebruary23,1996RAIAdditionalhydrogenandcarbonmonoxidecouldbereleasedduringcore-concreteinteractions(CCI)assteelinthecoredebrisorbasematreinforcementisoxidized.However,thereleaseofcombustiblegasfromsteeloxidationinCCIwouldoccurrelativelyslowly,wouldbeaccompaniedbydiluentssuchascarbondioxideandsteam,andwouldpromotemixinginthecontainment.ThusCCIwouldalsobeveryunlikelytoresultincombustiblegasconcentrationssufficienttocauseDDT.High.combustiblegasconcentrationscouldpotentiallybecreatedifcontainmentheatremovalwasrestoredafteralongperiodofCCI.InthePSA,STC9representsasequenceinwhichalatedeflagrationisinitiatedthatcausescontainmentfailure.Infact,experimentssuggestthatburnswillbeinitiatedassoonassufficientsteamiscondensedthat.thecontainmentgasisnolongerinerted[6].Theseburnsareobservedtobesoincompletethatnosignificantpressurizationofthecontainmentresults(exceptwhennear-stoichiometricmixturesarecreated).Again,detonationsareunlikely.Basedontheseconsiderations,thethreatofhydrogendetonationsisjudgedtobeverysmallintheGinnaStationcontainment.ThisisconsistentwithpaststudiesonotherPWRlarge,drycontainments[7].EquipmentSurvivability.Theavailabilityofcontainmentfancoolersandcontainmentspraysisconsideredintheplantdamagestate(PDS)definitionoftheIPE.Theeffectofharshenvironmentalconditionsontheoperationofthisequipmentarenotdiscussedinthecontainmenteventtree(CETIquantificationofthe'submittal.Pleaseprovideadescriptionofhowthesurvivabilityofthisequipmentundersevereaccidentconditionswasevaluated.Pleaseincludeinthediscussiontheenvironmentalconditions(e.g.,temperature,pressure,radiation,anddebris)derivedandusedintheevaluation.TheLevel2PSAdidnotconsidertheeffectofenvironmentalconditionsontheoperabilityofthesesystems.Toourknowledge,previousanalyseshavenotconcludedthatcontainmentsprayoperationwouldbethreatenedbycontainmentconditions,exceptperhapsthoseresultingincatastrophiccontainmentfailure.Forexample,intheNUREG1150studyonZion[8],onlya10percentlikelihoodwasassignedtofailureofcontainmentspraysfollowingcatastrophiccontainmentrupture;sprayswouldberelativelyineffectiveinreducingsourcetermaftergrosscontainmentfailureinanyevent.TheZionNUREG1150studydidnotexplicitlyconsiderthedeleteriouseffectofcontainmentconditionsonthefancoolersatall(seeAPETquestions24,43,45,46,53,65,and66)[8].ToassesswhethertheGinnaStationfancoolerswouldcontinuetooperateinasevereaccident,theMAAPcalculationsusedtocharacterizethevarioussource-termcategorieswerereviewed.Thissetofcalculationsisusefulforrepresentingthespectrumofaccidentcalculationsthatcouldbeencounteredina-severeaccident.Asnotedpreviously,

RcsponsctoFcbnmry23,1996RAIcontainmentsprayswerenotcreditedinthesecalculations;forthisreason,gastemperatureswouldbesomewhathigherthanexpectedinsequenceswherethesprayswereactuallyavailable.STCs2,3,4,and12involveoperablefancoolerspriortovesselfailure.TheMAAP-calculatedconditionsinthesesequenceswerecomparedtothedesignbasisenvelopeforthefancoolersdescribedinChapter6.1.2oftheUFSAR.ThepressureandradiationconditionsspecifiedfortheDBAaregenerallymoreseverethanthesevereaccidentconditionsthatwouldbeencounteredforallsequencesnotinvolvingalaterecoveryofcontainmentheatremoval.STC9involvesalaterecoveryoffancoolerswhichgivesrisetoalatefailureofcontainmentfromahydrogenburn;failureofthefancoolerstooperateinthissequencewouldnotthereforeleadtohigherrisk,ascalculatedbythePSA.Sincethefancoolers.arelocatedintheannularregionoutsidetheprimarycomponentshieldwallsandtakesuctionthere,theMAAP-calculatedga'stemperaturesinthisregionwereusedtoassessthethreatfromhightemperatures.ForallbutSTC9,gastemperaturesarelessthantheDBAenvelope(286Ffor2hours)exceptforbriefspikesobservedduringhydrogenburns.Sincethespikesareofshortduration,itisjudgedthatthefancoolerswouldsurvivetheseevents.Inthisregard,itshouldbenotedthattheGinnaStationfancoolermotorsareactivelycooledbyservicewaterandalsothattheTMIfancoolerscontinuedtooperateaftertheglobalhydrogenburnthatoccurredduringtheaccident[9].Itisnotedthatapossibilityexiststhatthefissionproductfilterscouldbecomecloggedduetohighaerosolloadingsduringasevereaccident.Thisshouldnotthreatentheeffectivenessofthefancoolerssinceonly2oftheunitshavefiltersandsincethefilterscanbemanuallybypassed.Therefore,weconcludethatoperationofthefancoolersandspraysisunlikelytobethreatenedbytheenvironmentalconditionsinthecontainmentandnochangesarerequired.EVNTREEvents.IntheGinnaIPE,asmallCETisdevelopedforaccidentprogressionanalysis.ThetopeventsoftheCETaredeterminedintheIPEbytheuseofdecompositioneventtreesPETs).TheeventprogressionanalysiscodeEVNTRE,whichwasdevelopedandusedintheNUREG-1150analyses,isusedforeventtreequantificationintheIPE.Tousethecode,thetopeventsdevelopedintheGinnaeventtreesarenumberedandincorporatedintotheEVNTREmodel.ThehighestnumberfortheEVNTREeventsthatcanbeidentifiedfromthetreesintheGinnasubmittalis47.Thisseemstoindicatethatthereare47events(orquestions)intheEVNTREmodeLHowever,examinationoftheeventtreesintheGinnaIPEshowsthatthetotalnumberofeventsinthetreesislessthan47.Asaresult,someEVNTREquestions(e.g.,14,17,18,etc..)cannotbeidentifiedfrom

RcsponsctoFebruary23,1996RAItheGinnaeventtreespresentedinthesubmittaLPleaseprovidealistofallENTREquestionsusedintheGinnaanalysisanddiscussthequestionsthatareintheEVNTREmodelbutnotintheGinnaeventtrees.AcompletelistoftheEVNTREeventsintheCETisprovidedinTable1attheendofthissection.Thefirst12eventsinthemodelcorrespondtothetopeventsshowninFig4.3-2oftheMarch15,1994submittal.Theremainingeventsareeitherdescribedinthetextofthesubmittalorare"dummy"eventsincludedfortheconvenienceoftheanalystthatdonotaffecttheresults.Thedummyeventsarenumbered14,17,18,27,29,31,35,37,38,40,41,and45.Allnon-dummyeventsweredescribedinthesubmittal.ModularAccidentAnalysisProgram(MAAP)CalculationResults.AnumberofMAAPcalculationswereperformedintheGinnaIPEtoprovidedatafortheaccidentprogressionanalyses.AbriefdescriptionofeachoftheseaccidentprogressioncasesisgiveninTable4.6-1ofthesubmittaLMAAPcalculationswerealsoperformedtoderivereleasefractionsforthevarioussourcetermcategories.BriefdescriptionofthesesourcetermcasesarealsoprovidedinTable4.6-1.ItcanbeseenthattheconditionsofCaseMLOCA03(page4-150)andSourceTennCaseSTC12(page4-161)aresimilar.Forbothcases,theauxiliaryfeedwater(AFW)isavailableandthesafetyinjectionsystemisunavailable.However,thecoreuncoverytimeandthevesselfailuretimeshowninTable4.6-1aresignificantlydifferentforthesetwocases(theyare4.27and5.96hours,respectively,forCaseMLOCA03,and0.7and1.55hours,respectively,forCaseSTCI2).Pleasediscussthereasonsforthetimedifferencebetweenthesetwocases,vhereandhoweachcasewasusedintheIPEanalysis,andtheimpactofthedatafromeachcaseonIPEquantification.AsusedinTable4.6-1,"NoSI"referstotheunavailabilityofhighpressureinjection.Caseswithnosafetyinjectionatall(i.e.,highandlow)arelabelled"Noinjection"(seeforexamplecaseSLOCA00).TheapparentdiscrepancybetweenthetwocasescitediscausedbytheavailabilityoflowpressureinjectioninMLOCA03whichdelayscoreuncoveringandvesselfailurerelativetocaseSTC12whichinvolvesacompletefailureofallactiveinjectionsystems.TheresultsofMAAPcalculationswereusedintheGinnaStationPSAtoassesssequencetiming(see,forexample,thediscussioninSection4.3.1.1.2),toobtaininsightsusefulforthequantificationoftheCET(e.g.,eventssuchasinducedruptureoftheRCS,directcontainmentheating,deinertingthecontainmentatmosphere,etc.),and.toassignexplicitsourcetermstosourcetermcategories.ThisisdiscussedinSections4.5,4.6,and4.7oftheMarch15,1994submittal.TheresultsfromMAAPcalculationMLOCA03werenotusedexplicitlyinthePSAandwereperformedsimplytoprovidetheanalystsaqualitativeunderstandingoftheeffectofbreaksizeandsafetyinjectionsystemavailabilityonsequencetiming.Theresultsof

RcsponsctoFebruary23,1996RAIcalculationSTC12weredirectlyusedtoassignsourcetermstotheassociatedsourcetermcategory.SequenceSelectionforSourceTermDetermination.ItisstatedintheIPEsubmittal(Section4.7.3)that"specificaccidentprogressionsequenceswerechosentobestapproximatetherepresentativesourcetennresultsforeachrelevantSourceTermCategory(STC)endstate.Basedonconsiderationofthedominantsequenceforeachendstateandbasedonotherfactorswhichinfluencet@sourcetermresults,representativesequencedescriptionsweredevelopedtoperformMAAPcalculationstoquantifythesourceterms."However,inthesubmittal,thePDSsthatcontributetotheSTCsarediscussedonlyforafewSTCS,andforsomeofthesecasesthesequencesselectedforMAAPcalculationsarenotthedominantsequencesinthePDSS.Forexample,accordingtothesubmittal;PDS15andPDS17representthemajorityofSTC2.TherepresentativesequencechosentorepresentthisSTCisamedium-breakloss-of-coolantaccident(LOCA)sequence.However,accordingtoTable4.3-5ofthesubmittal,thesequencesthatcontributetothesePDSsaresmallLOCAorreactorcoolantpump(RCP)sealLOCAsequences,andmediumLOCAsequencesarenotinvolvedineitherofthesetwoPDSS.ThereisnodiscussioninthesubmittalwhyamediumLOCAinsteadofasmallLOCAsequencewaschosentorepresentthisSTC.PleaseprovideacompletelistofthecontributingPDSsforalltheSTCsandprovidetherationalefortheselectionofeachofthesequenceschosento-representtheSTCS.ContributingPDSsarenotdiscussedforeachoftheSTCsbecauseofthemanySTCsthataPDSmightcontributetoandbecauseofthelowprobabilitiesassociatedwithanumberoftheSTCs.Forexample,PDS2comprisesonly0.35%ofthePDSsperTable4.3-4.Almost69%ofPDS2contributestoSTC15,butitisstillnotasignificantcontributionsincethePDSisofsuchalowfrequency(about23%ofPDS2contributestoSTC1andtheremaining8%contributesto6otherSTCs).Thefollowingisalistingofthetop11PDSsperTable4.3-4andtheSTCstheyprimarilycontributetoinapproximatepercentages:PDS12contributestoSTC1(72%)andSTC12(28%)PDS22contributestoSTC18(100%)PDS24contributestoSTC20(100%)PDS20contributestoSTC16(100%)PDS15contributestoSTC1(59%)andSTC12(41%)PDS17contributestoSTC1(97%)PDS11contributestoSTC15(95%)PDS1contributestoSTC13(100%)PDS9contributestoSTC15(65%),STC1(21%)andSTC12(13%)PDS7contributestoSTC9(48%)andSTC14(48%)PDS14contributestoSTC12(100%)-"

ResponsetoPcbnmry23,1996RAIThesourcetermcharacterizationwasconsideredlessimportantthanthecharacterizationofthecoredamagesequenceandtheachievementofanoverallunderstandingofcontainmentperformance.SelectionofsequencestorepresentSTCswasprimarilyguidedbyadesireto:RepresentthekeyphenomenologicaleffectsthatwereusedtocharacterizetheSTCs.Theseeffects,e.g.theoccurrenceofCCI,availabilityofwateronthedebris,availabilityofcontainmentheatremoval,sizeandtimingofcontainmentfailure,etc.areshowninFigure4.7-1oftheMarch15,1994submittal.Breaksizewasnotconsideredsufficientlyimportanttobeusedintheclassificationofsourcetermcategories,althoughitdoesaffectthedegreetowhichfissionproductsareretainedintheRCS.b.MinimizeanalyticaleffortforsequenceswhichdonotcontributesignificantlytotheriskAsdescribedinSection4.7.3.1,STC2representssequenceswithvesselfailure,aleak-typeearlycontainmentfailure,containmentheatremoval,andnosustainedCCI.Assuch,thesourcetermisdominatedbythereleaseoffissionproductsfromthevesselintheperiodjustleadinguptoandfollowingvesselfailure.AmediumLOCAsequencewaschosentoreducein-vesselretentionoffissionproductssoastoprovideaconservative,yetreasonableassessmentforthevariousPDSscomprisingthisreleasecategory.NotealsoinTable4.7-1thatSTC2representslessthan0.1percentoftheCDF.Bycontrast,STCs16,18,and20representnearlyhalftheCDFandallinvolvebypassofthecontainment.Allthreeofthelattersequencescanaprioribeexpectedtohavethelargestsourceterms,atleastasmodeledbyMAAP.Thesesourcetermsareprobablyover-predictedbytheMAAPcalculationsincethecodeneglectsturbulentdepositioninpiping(STC16)andonthesecondarysidesofthesteamgenerators(STCs18and20)andsinceretentionoffissionproductsintheauxiliarybuildingwasnotmodeled(STC16).Inanyevent,giventheMAAP-calculatedsourceterms,negligibleerrorsareincurredintheresultsasawholebyfailingtomorepreciselymodelsequence-specificaspectsofthemuchlower-probabilityendstatessuchasSTC2thathaverelativelysmallfissionproductreleases.16.InducedSGTR-TheLikelihoodofinducedhotlegorsteamgeneratortuberupturesinhighpressurescenariosisanalyzedintheGinnaIPEusingMAAPanalysisdataandanempiricaLformula,developedbyLarsonandMiller,relatingtheexpectedtimetorupturewithtemperatureandstress.SincetheconclusionsfromthisanalysisareconsistentwiththeresultsfromtheNUREG-1150dataforSurry,theprobabilityvaluesusedintheNUREG-1150SurryanalysisareusedintheGinnaIPE.ItshouldbenotedthatinsomeIPEStheprobabilityofinducedSGTRduetoforcedcirculationcausedbytherestartofthe 0

ResponsetoFebn>ary23,1996RAIRCPsisaddressedbecausetheinsujlicientcorecooling(ICC)guidelinescallfortheRCPstoberestarted.PleasediscusswhetherthereareproceduresatGinnathatcallfortherestartoftheRCPsand,ifthereare,pleasediscusstheirgectontheprobabilityofinducedSGTR.GinnaStationProcedureFR-C.1,ResponsetoInadequateCoreCooling,potentiallycallsforrestartoftheRCPsafterthecoreisuncovered.Duringthetimeframe"whenthePSAwasbeingperformed,concernwasexpressedthatthismightcauseanincreaseinheattransferfromthecoretothesteamgeneratortubes,potentiallycausingtheirfailure.Thisconcernwouldonlyberelevantintherelativelysmallnumberofsequencesinwhichthesecondarysideofthesteamgeneratorswasdry,RCPswereavailable,andRCSpressurewaselevatedcomparedtothesteamgeneratorpressure.Inanyevent,thisissuewasnotaddressedintheGinnaStationPSA.Thecurrentversionoftheprocedure,revision10dated11/29/95,callsforanRCPtoberestartedinagivenlooponlyifthenarrowrangelevelexceeds5percentintheassociatedsteamgenerator(seestep23).Thisensuresthattheaffectedtubeswillbecooledandeliminatesconcernsforthermally-inducedtuberupture.17.TypographicalErrors(a)ACpowerrecoveryconsidersonlytherecoveryofoff-sitepower.TherecoveryofDGsisnotcredited.ItseemsthatthereiatypographicalerrorinTable4.3-2.,Theprobabilityofpowernon-recoveryat21hoursshouldbe0.00$66insteadof0.0$66showninthetable.ThetypographicalerrorwillbecorrectedduringthenewLevel2analysis.(b)ThetitleofReference4.9-I9ofthesubmittalisgivenas"EvaluationofSevereAccidentRisks:GinnaUnitI".Itshouldbe"SurryUnit1"insteadof"GinnaUnitgItThetypographicalerrorwillbecorrectedduringthenewLevel2analysis.(c)Event9oftheCETtopevent"Ex-VesselCCI"isdescribedinthesubmittal(p4-74)asCAVWATF.However,itisdescribedasLRWSTintheCCIdecompositioneventtree(Figure4.$-6,Paragraph4-l96).ThetypographicalerrorwillbecorrectedduringthenewLevel2analysis.

RcsponsctoFebruary23,1996RAI1.E.Fuller,ed.,ContainmentPerformanceandFissionProductReleaseDeterminationforIndividualPlantExaminations(IPES),ProceedingsofaWorkshopinChicago,IL,July1989,ElectricPowerResearchInstitute.TheseresultsarealsoreportedinR.E.Henryetal.,"ExperimentsRelatingtoDrywellShell-CoreDebrisInteractions",Proc.16thWaterReactorSafetyInformationMeeting,Gaithersburg,October24-27,1988,NUREG/CR-0097.2.T.G.TheofanousandH.Yan,"TheProbabilityofLinerFailureinaMark-IContainment,PartII:MeltReleaseandSpreadingPhenomena",Nucl.Tech.,101,March1993.3.R.J.Breedinget.al,EvaluationofSevereAccidentRisks:QuantificationofMajorInputParameters,NUREG/CR-4551,Vol.2,Rev.1,Part3.4.LetterfromD.M.Crutchfield,NRCtoJ.E.Maier,RG&E,

Subject:

SEPTopicVIII-4,ElectricalPenetrationsofReactorContainmentSafetyEvaluationReportforR.E.GinnaNuclearPower'lant",datedOctober8,1981.5.MemorandumfromM.Kenton(D&M)toDaphneMays(RG&E)datedJuly16,1996.~~~6.T.BlanchatandD.Stamps,"DeliberateIgnitionofHydrogen-Air-SteamMixturesUnderConditionsofRapidlyCondensingSteam",SAND94-3101C,January1995.7.J.W.Yang,Z.Musicki,andS.Nimnual,HydrogenCombustion,Control,andValue-ImpactAnalysisforPWRDryContainments,NUREG/CR-5662,BNL-NUREG-52271,June1991.8.C.Parket.al,EvaluationofSevereAccidentRisks,Zion,Unit1:AppendixA.,NUREG/CR-4551Vol.7Rev1Part2A,March1993.9.J.O.HenrieandA.K.Postma,AnalysisofTheThreeMileIsland(TMI-2)HydrogenBurn,RHO-RE-SA-8P,RockwellInternational,October1982.

RcsponsctoFebruary23,1996RAITable112345678910111213141516171819202122232425262728293031323334353637383940ContainmentBypassSequenceContainmentIsolationStatusTransientorLOCATypeReactorShutdownStationBlackoutPowerRecoveryRCSPressureatCoreDamageStatusofIn-VesselInjectionContainmentFanCoolersContainmentSprayStatusSteamGeneratorIsolatedSteamGeneratorBreakCoveredModeofInducedRCSFailureDummyRCSPressureatTimeofRPVFailureDebrisCooledIn-VesselDummyDummyIn-VesselSteamExplosionFailsContainmentRWSTInjectedEarlyContainmentAtmosphereInertMassDebrisExpelledEarlyFractionDebrisInvolvedinDCHFractionDebrisDispersedOutsideLowerCompartmentFractionDebrisMetalReactedHydrogenBurnDummyEarlyContainmentFailureDummyRWSTInjectedLateDummyDebrisDepthAgainstLinerContainmentLinerMeltthroughSteamExplosionDispersesDebrisDummyDepthofDebrisinSumpDummyDummyTypeofEx-VesselCCIDummy ResponsetoFebruary23,1996RAITable1Continued41424344454647DummyPowerAvailablePriortoRVFailurePowerRecoveryLateContainmentHeatRemovalDummyLateH2BurnFailsContainmentModeofLateContainmentFailure-45-