ML17332B013
| ML17332B013 | |
| Person / Time | |
|---|---|
| Site: | Cook  |
| Issue date: | 11/10/1995 |
| From: | INDIANA MICHIGAN POWER CO. (FORMERLY INDIANA & MICHIG |
| To: | |
| Shared Package | |
| ML17332B012 | List: |
| References | |
| NUDOCS 9511160295 | |
| Download: ML17332B013 (55) | |
Text
ELECTRICALPOWERSYSTEMSSURVEILLANCEREUIREHENTSContinued4.8.1.1.2EachdieselgeneratorshallbedemonstratedOPERABLE-InaccordancewiththefrequencyspecifiedinTable4.8"1onaSTAGGEREOTESTBASISby:1.2.3.5.fromstandtrycondidonsandacldevesInlessthanorequalto10seconds.vottage~at60na20V,andfrequencyi60nl.2Hz.Verifyingthefuellevelinthedaytank,Verifyingthefuellevelinthefuelstoragetank,Verifyingthatthefueltransferpumpcanbestartedandthatittransfersfuelfromthestoragesystemtothedaytank,Verifyingthatthedieselstartsb.C.nd6.'"Verifyingthatthedieselgeneratoris.alignedtoprovidestandbypowertotheassociatedemergencybusses.Bgremovingaccumulatedwater"+1)Fromthedaytankatleastonceper31.:daysandaftereachoccasionwhenthedieselisoperatedforgreatei'han1hour,and2)Fromthestoragetanksatleastonceper31days.Bysamplingnewfueloil"<inaccordancewiththeapplicableguidelinesofASTM04057-81priortoaddingnewfueltothestoragetanksandVerlfylngthergesetlssynchroAltedandloadedandcneratesforgreaterthanor~qualto60minutesataloadol3500Inv,1)Byverifying,inaccordancewiththetestsspecifiedinASTH0975-81andpriortoaddingthenewfueltothestoragetanks,thatthesamplehas:Thedieselgeneratorstart10secondsfromconditionsshabeperformedatleastonceper184daysinthesesurveillancetests.Allotherenginestartsforthepurposeofthissurveillancetestingandcompensatoryactionmaybeatreducedaccelerationratesasrecommendedbythemanufacturersothatmechanicalstressandwearonthedieselengineareminimized.kkMomcntnyloadtransientsdonotinvalidatethistest.+""Theactionstobetakenshouldany'ofthepropertiesbefoundoutsideofspecifiedlimitsare'definedintheBases.O.C.COOK"UNIT13/48-3AMENOMENTNO.'si5111'60295951110,'PDRADOCK,05000515P-.,"PDR..~484NUI~Jl"lIIe'1%ihr'I~46l~C4,r,HM:~alllallslate~
'~IRlWt~tl'+IP phag~T/5NILKuPstopowsrfsolofoftossthonor~rtosltoO.BBforsttssstBfactors.+SCoeuedc)Verifyingthatallaueomaeicdiesel.generaeortrips,exceptengineoverspeedandgeneratordifferential,areautomaeicallybypasseduponlossofvoleageontheemergencybusand/orSafetyInjecei.onaccuaeionsignal.7.VerifyingthatthedieselgeneraeoroperatesDuringthistestthedieselgeneracorshallbeloadedto3500kw.Within5minutesaftercompleeingthis~~hourtest,performSurveillanceRequiremene4.8.1.1.2.a.4(atexistingconditions).*+8.Determinethattheaueo-connectedloads.toeachdieselgeneratordonotexceed3500kw.9.Verifyingthedieselgeneraeor'scapabilityeo:a)Synchronizewiththeoffsieepowersourcewhileehegeneraeorisloadedwithitsemergencyloadsuponasimulatedrestorationofoffsieepower.Ib)Transferitsloadseotheoffsitepowersource,andc)Bereseoredeoitsseandbyseaeus.10.Verifyingthaewiththedieselgeneraeoroperaeinginatesemodeawhileconnectedtoitstestload,asimulatedSafetyInjectionsignaloverridesthetestmodeby:a)ReturningthedieseLgeneraeoreoseandbyoperaeion,andb)Verifyingtheemergencyloadsareservicedbyoffsitepower.11.VerifyingthaetheautomaticsequencetimingrelaysareOPERABLEwitheachloadsequencetimewiehinplusorminus5%ofiesrequiredvalueand'thaeeachloadissequencedonwithinthedesignallowabletimeLimit.Aeleaseonceper10yearsby:1)Employingoneofthefollowingcleaningmethodseocleanthefueloilstoragetanks:a)DraineachfueLoilseorageeank,remove.theaccumulaeedsedimene,andcleantheeank,or*+IfSurveillanceRequirement4.8.L.1.2.a.4isaoesaeisfaceorilycompleeed,ieisnoenecessaryeorepeatthepreceditqp'Zdo-hourtest.Instead,thedieselgeneraeormaybeoperaeadae3500kwfor2hoursorunciloperatingtemperaturehasseabilized.Momsntaytrsnslsntsotrtstdotftofoodsnc'owerfstsorrsntfodonotfrwsMstotnfsttst.COOKNUCLEARPLANT-UNIT13/48-6AMENDMEHTNO.424 P.ly1IC ELECTRICALPOWERSYSTEMSSURVEILLAHCEREUIREMEHTSContinuedpftesqfrf1/srnftfticvp4.8.1.1.2EachdieselgeneratorshallbedemonstratedOPERABLE:InaccordancewiththefrequencyspecifiedinTable4.8-1onaSTAGGEREOTESTBASISby:l.2.3.Verifyingthefuellevelinthedaytank,VerifyingthefuellevelinthefuelstorageVerifyingthatthefueltransferpumpcanbeittransfersfue1fromthestoragesystemtoVerifyingthatthedieselstartstank,startedandthatthedaytank,Fromstandbyconrstfonsand~VivosInlessthanorcorralto10seconds,volteqeidt60aa20V.sndfrequency~60*1.2Ht.b.and6.:"-Verifyingthatthedieselgeneratorisalignedtoprovidestandbypowertotheassociatedemergencybusses.Bfremovingaccumulatedwater"W~~1)Fromthedaytankatleastonceper33,-daysandaftereachoccasionwhenthedieselisoperatedforgreatirthan1hour,andVerffyfntfthedfeseffssynclvonrtedandloadedandoperatesfortfreaterthenor~qrrafto60mfnrrtesataloadof3500fnv,C.2)Fromthestoragetanksatleastonceper31days.Bysamplingnewfueloil~+inaccordancewiththeapplicableguidelinesofASTM04057-81priortoaddingnewfueltothestoragetanksand1)Byverifying,inaccordancewiththetests.specifiedinASTH0975-81andpriortoaddingthenewfueltothestoragetanks,thatthesamplehas:"hedieselgeneratorstart10secondsromconditionsshaeperformedatleastonceper184daysinthesesurveillancetests.Allotherenginestartsforthepurposeofthissurveillancetestingandcompensatoryactionmaybeatreducedaccelerationratesasrecommendedbythemanufacturersothatmechanicalstressandwearonthedieselengineareminimized.Morncntatyloadtransientsdonotinvalidatedtistest""TheactionstobetakenshouldanyofthepropertiesbefoundoutsideofspecifiedlimitsareldefinedintheBases.0.C.COOK-UNIT23/48-3AHEHOHEHTHO.
Ef1,~,l~tv<A otued7.c)Verifyingthatallautomaticdieselgeneratortrips,exceptengineoverspeedandgeneratordifferential,areautomaticallybypasseduponlossofvoltageontheemergencybusand/orSafetyIn]ectionactuationsignal.et~powerfooter~qoetto0.88lorVerifyingthatthedieselgeneratoroperates~etreset8horrre.$Duringthistestthedieselgeneratorshallbeloadedto3500kv.Within5minutesaftercompletingthis~-hourtest,performSurveillanceRequirement4.8.1.1.2.a.4(atexistingconditions).*f-8.Determinethattheauto-connectedloads'toeachdieselgeneratordonotexceed3500kv.9.Verifyingthedieselgenerator'scapabilityto:a)Synchronizewiththeoffsitepowersourcevhilethegeneratorisloadedvithitsemergencyloadsuponasimulatedrestorationofoffsitepower.bfTransferitsloadstotheoffsitepowersource,andc)Berestoredtoitsstandbystatus.10.Verifyingthatviththedieselgeneratoroperatinginatestmodewhileconnectedtoitstestload,asimulatedSafetyInjectionsignaloverridesthetestmodeby:a)Retuz~igthedieselgeneratortostandbyoperation,andb)Verifyingtheemergencyloads'araservicedbyoffsitepover.11.VerifyingthattheautomaticsequencetimingrelaysareOPERABLEwitheachloadsequencetimewithinplusorminus5%ofitsrequiredvalueandthateachloadissequencedonwithinthedesignallovabletimelimit.rAtleastonceper10yearsby:1)Employingoneofthefollowingcleaningmethodstocleanthefueloilstoragetanks:a)Draineachfueloilstoragetank,removetheaccumulatedsediment,andcLeanthetank,or*+IfSurveillanceRequirement4.8.1.1.2.a.4isnotsatisfactorilycompleted,itisnotnecessarytorepeatthepreceding+-hourtest.Instead,thedieselgeneratormaybeoperatedat3500kvfor2hoursoruntiloperatingtemperaturehasstabilized.powerrectorrangeoonotlrrregdetethleteel.COOKNUCLZMPLANT-UNIT23/48-6AMENDMEHTNO.4XQ+66-ATTACHMENT3TOAEP:NRC:0896KPROPOSEDREVISEDTECHNICALSPECIFICATIONPAGES
,03/4LIMITINGCONDITIONSFOROPERATIONANDSURVEILLANCEREQUIREiiTS3/4.8ELECTRICALPOWERSYSTEMSSURVEILLANCEREUIREMENTSContinued4.8.1.1.2EachdieselgeneratorshallbedemonstratedOPERABLE:a0InaccordancewiththefrequencyspecifiedinTable4.8-1onaSTAGGEREDTESTBASISby:1.Verifyingthefuellevelinthedaytank,2.Verifyingthefuellevelinthefuelstoragetank,3.Verifyingthefueltransferpumpcanbestartedandthatittransfersfuelfromthestoragesystemtothedaytank,Verifyingthedieselstartsfromstandbyconditionsandachievesinlessthanorequalto10seconds,voltage=4160~420V,andfrequency=60+1.2Hz,Verifyingthedieselissynchronizedandloadedandoperatesforgreaterthanorequalto60minutesataloadof3500kw",and6.Verifyingthedieselgeneratorisalignedtoprovidestandbypowertotheassociatedemergencybusses.Byremovingaccumulatedwater1)Fromthedaytankatleastonceper31daysandaftereachoccasionwhenthedieselisoperatedforgreaterthan1hour,and2)Fromthestoragetanksatleastonceper31days.Bysamplingnewfueloil"inaccordancewiththeapplicableguidelinesofASTMD4057-81priortoaddingnewfueltothestoragetanksand1)Byverifying,inaccordancewiththetestsspecifiedinASTMD975-81andpriortoaddingthenewfueltothestoragetanks,thatthesamplehas:Thedieselgeneratorstart(10seconds)fromstandbyconditionsshallbeperformedatleastonceper184daysinthesesurveillancetests.Allotherenginestartsforthepurposeofthissurveillancetestingandcompensatoryactionmaybeatreducedaccelerationratesasrecommendedbythemanufacturersothatmechanicalstressandwearonthedieselengineareminimized.Momentaryloadtransientsdonotinvalidatethistest.TheactionstoUetakenshouldanyofthepropertiesbefoundoutsideofspecifiedlimitsaredefinedintheBases.COOKNUCLEARPLANT-UNIT1Page3/48-3AMEXDiWIENT~
il~I'il 3/4LIMITINGCONDITIONSFOROPERATIONANDSURVEILLANCEREQUIREMENTS3/4.8ELECTRICALPOWERSYSTEMSSURVEILLANCEREUIREMENTSContinuedc)Verifyingthatallautomaticdieselgeneratortrips,exceptengineoverspeedandgeneratordifferential,areautomaticallybypasseduponlossofvoltageontheemergencybusand/orSafetyInjectionactuationsignal.Verifyingthatthedieselgeneratoroperatesatapowerfactoroflessthanorequalto0.86foratleast8hours.Duringthistestthedieselgeneratorshallbeloadedto3500kw.Within5minutesaftercompletingthis8-hourtest,performSurveillanceRequirement4.8.1.1.2.a.4(atexistingconditions)."Determinethattheauto-connectedloadstoeachdieselgeneratordonotexceed3500kw.9.Verifyingthedieselgenerator'scapabilityto:a)Synchronizewiththeoffsitepowersourcewhilethegeneratorisloadedwithitsemergencyloadsuponasimulatedrestorationofoffsitepower.b)Transferitsloadstotheoffsitepowersource,andc)Berestoredtoitsstandbystatus.10.Verifyingthatwiththedieselgeneratoroperatinginatestmodewhileconnected,toitstestload,asimulatedSafetyInjectionsignaloverridesthetestmodeby:a)Returningthedieselgeneratortostandbyoperation,andb)Verifyingtheemergencyloadsareservicedbyoffsitepower.11.VerifyingthattheautomaticsequencetimingrelaysareOPERABLEwitheachloadsequencetimewithinplusorminus5%ofitsrequiredvalueandthateachloadissequencedonwithinthedesignallowabletimelimit.Atleastonceper10yearsby:Employingoneofthefollowingcleaningmethodstocleanthefueloilstoragetanks:a)Draineachfueloilstoragetank,removetheaccumulatedsediment,andcleanthetank,orMomentarytransientsoutsidetheloadandpowerfactorrangedonotinvalidatethistest.IfSurveillanceRequirement4.8.1.1.2.a.4isnotsatisfactorilycompleted,itisnotnecessarytorepeatthepreceding8-hourtest.Instead,thedieselgeneratormaybeoperatedat3500kwfor2hoursoruntiloperatingtemperaturehasstabilized.COOKiNUCLEARPLAiVI'-UNIT1Page3/484Ail'IENDiWENT42$,49, piEgtj 03/4LIMITINGCONDITIONSFOROPERATIONANDSURVEILLANCEREQUIREMENTS3/4.8ELECTRICALPOWERSYSTEMSSURVEILLANCEREUIREMENTSContinued4.8.1.1.2EachdieselgeneratorshallbedemonstratedOPERABLE:InaccordancewiththefrequencyspecifiedinTable4.8-1onaSTAGGEREDTESTBASISby:1.Verifyingthefuellevelinthedaytank,2.Verifyingthefuellevelinthefuelstoragetank,3.Verifyingthatthefueltransferpumpcanbestartedandthatittransfersfuelfromthestoragesystemtothedaytank,4.Verifyingthatthedieselstartsfromstandbyconditionsandachievesinlessthanorequalto10seconds,voltage=4160~420V,andfrequency=60+1.2Hz,Verifyingthedieselissynchronizedandloadedandoperatesforgreaterthanorequalto60minutesataloadof3500kw,andVerifyingthatthedieselgeneratorisalignedtoprovidestandbypowertotheassociatedemergencybusses.Byremovingaccumulatedwater":1)Fromthedaytankatleastonceper31daysandaftereachoccasionwhenthedieselisoperatedforgreaterthan1hour,and2)Fromthestoragetanksatleastonceper31days.Bysamplingnewfueloil"inaccordancewiththeapplicableguidelinesofASTMD4057-81priortoaddingnewfueltothestoragetanksand1)Byverifying,inaccordancewiththetestsspecifiedinASTMD975-81andpriortoaddingthenewfueltothestoragetanks,thatthesamplehas:Thedieselgeneratorstart(10seconds)fromstandbyconditionsshallbeperformedatleastonceper184daysinthesesurveillancetests.Allotherenginestartsforthepurposeofthissurveillancetestingandcompensatoryactionmaybeatreducedaccelerationratesasrecommendedbythemanufacturersothatmechanicalstressandwearonthedieselengineareminimized.Momentaryloadtransientsdonotinvalidatethistest.TheactionstobetakenshouldanyofthepropertiesbefoundoutsideofspecifiedlimitsaredefinedintheBases.COOKNUCLEARPLAiVT-UNIT2Page3/48-3AMENDMENT444, L)
~03/4LIMITINGCONDITIONSFOROPERATIONANDSURVEILLANCEREQUIREMENTS3/4.8'ELECTRICALPOPOVERSYSTEMSSURVEILLANCEREUIREMENTSContinuedc)Verifyingthatallautomaticdieselgeneratortrips,exceptengineoverspeedandgeneratordifferential,areautomaticallybypasseduponlossofvoltageontheemergencybusand/orSafetyInjectionactuationsignal.Verifyingthatthedieselgeneratoroperatesatapowerfactoroflessthanorequalto0.86foratleast8hours.Duringthistestthedieselgeneratorshallbeloadedto3500kw.Within5minutesaftercompletingthis8-hourtest,performSurveillanceRequirement4.8.1.1.2.e.4(atexistingconditions)."Determinethattheauto-connectedloadstoeachdieselgeneratordonotexceed3500kw.Verifyingthedieselgenerator'scapabilityto:a)Synchronizewiththeoffsitepowersourcewhilethegeneratorisloadedwithitsemergencyloadsuponasimulatedrestorationofoffsitepower.b)Transferitsloadstotheoffsitepowersource,andc)Berestoredtoitsstandbystatus.10.Verifyingthatwiththedieselgeneratoroperatinginatestmodewhileconnectedtoitstestload,asimulatedSafetyInjectionsignaloverridesthetestmodeby:=a)Returningthedieselgeneratortostandbyoperation,andb)Verifyingtheemergencyloadsareservicedbyoffsitepower.11.VerifyingthattheautomaticsequencetimingrelaysareOPERABLEwitheachloadsequencetimewithinplusorminus5%ofitsrequiredvalueandthateachloadissequencedonwithinthedesignallowabletimelimit.Atleastonceper10yearsby:I)Employingoneofthefollowingcleaningmethodstocleanthefueloilstoragetanks:a)Draineachfueloilstoragetank,removetheaccumulatedsediment,andcleanthetank,orMomentarytransientsoutsidetheloadandpowerfactorrangedonotinvalidatethistest.IfSurveillanceRequirement4.8.1.1.2.a4isnotsatisfactorilycompleted,itisnotnecessarytorepeatthepreceding8-hourtest.Instead,thedieselgeneratormaybeoperatedat3500kwfor2hoursoruntiloperatingtemperaturehasstabilized.COOKNUCLEARPLANT-UNIT2Page3/48-6AMENDMENT~,468, ATTACHMENT4TOAEP:NRC:0896XCOOKNUCLEARPLANTEMERGENCYDIESELGENERATORLOAD-RUNPERFORMANCEANDRELIABILITYDURINGSHORTANDLONGDURATIONTESTPERIODS r
Attachment4pageIABSTRACTThefollowingstudycomparesCookNuclearPlantemergencydieselgeneratorperformanceduringrelativelyshorttestrunstoperformanceduringsignificantlylongertestdurations.A,statisticaltreatmentofdataoveratwentyyearperiodisusedtodevelopanestimateofcumulativefailureprobabilitywithrespecttotestruntime.Inaddition,areviewofpreviousemergencydieselgeneratorperformancestudiesisusedtoprovideanindustrywidecomparisonofload-runfailureprobabilities.ThestudyconcludesthatanendurancetestdurationofeighthoursinlieuofthecurrenttwentyfourhourswouldbeacceptablefortheemergencydieselgeneratorsatCookNuclearPlant.AmericanElectricPower Attachment4page2TABLEOFCONTENTSSection1.0Introduction2.0CookNuclearPlantDieselGenerators2.1SpecificationsEcMajorComponents2.2OperationsOverview3.0ReviewofCookNuclearPlantTestData3.1ValidLoad-RunCriteria3.2InvalidLoad-RunCriteria3.3LongDurationLoad-RunData3.4ShortDurationLoad-RunData3.5FailureCharacteristics3.6DataAnalysis4.0ReviewofSelectedStudies4.1BNLTechnicalReport¹A-31341-854.2NSAC-1084.3NUREG/CR-5994155.0IndustrySurveys6.0Conclusions7.0References171718AmericanElectricFower
'lIt>kr Attachment4.1.0Introductionpage3Thisreportfocusesonemergencydieselgenerator(EDG)load-runperformanceatCookNuclearPlant.Thereportconsidersbothlongdurationload-runs(greater60minutes)andshortdurationload-runs(60minutesorless).ThereportdoesnotconsiderEDGstartingreliability.Thisreporthasfourprimaryobjectives:1)Reviewtheload-runperformanceofemergencydieselgenerators(EDGs)atCookNuclearPlantwithemphasisonthelongdurationtestruns.2)Reviewexistingstudies,evaluations,anddatatodetermineameasureofEDGload-runperformanceonanindustrywidebasis.3)SampleEDG24hourendurancetestrunperformanceatothernuclearpowerplantstoprovidecomparisondata.4)Constructameaningfultranslationofthisdatainconjunctionwithexistingstudies,evaluations,andotherdatatodetermineaqualitativebenchmarkforevaluatingtheacceptabilityofreducingthe18monthEDGendurancetestdurationlromthecurrent24hourrequirementforCookNuclearPlant.AmericanElectricPopover pt Attachment4page42.0CookNuclearPlantDieselGeneratorsThedatainsections2.1and2.2summarizeimportantEDGspecificationsandoperatingrequirements.BothunitsatCookNuclearPlantareWestinghouse,4-loop,pressurizedwaterreactors.Unit1israted3250MWtandUnit2israted3411MWt.FourEDGs,twoperunit.EachdieselengineisaWorthingtonTypeSWB-12,12cylinder,heavydutyturbochargeddieselengine,withacontinuousratedoutputof4900BHPat514RPM.EachgeneratorisaGeneralElectric,4375KVA,3500KWat0.8P.F.,514RPM,3-phase,60-cycle,4160V,25%voltageregulation,directengine-drivensynchronoustypegenerator.EachgeneratorisequippedwithaGEbrushlessexcitertypeSARwithafieldrated5.8Aand100VDCand,astaticvoltageregulatorwithitsassociatedpotentialandsensingtransformers.TheEDGsystemforeachunitconsistsoftworedundant,Class1E,identicaldieselgeneratorswhichareindividuallycapableofsupplyingsufficientpowertooperateonecompleteredundanttrainofengineeredsafetyfeatures(ESF)andprotectionsystemsrequiredforsafeshutdownoftheunit.TheEDGsaredesignedtostartautomaticallyuponreceiptofasafetyinjectionsignalancVoralossofoffsitepowersignalandbereadytoacceptloadswithin10secondsofreceivingastartsignal.Duringnormalplantoperation,theEDGsareonstandbyandareautomaticallyavailableifofBitepowerislost.AmericanElectricPopover IJCUNp'hEp)
Attachment4page53.0ReviewofCookNuclearPlantTestDataThissectionconsidersload-runtestdatafrom1974to1994.Theobjectivesofthissectionincludeexaminationofload-rundataspecificallyassociatedwith24hoursurveillancetests.Accordingly,thestudycouldhavefocusedstrictlyon24hoursurveillancetestruns.However,itwasfeltthatdoingsocouldeliminate,ormask,insightfulobservationsaboutthe"quality"ofEDGperformanceduringotherextendedfullload-runsandthecharacteristicbehaviorofload-runfailures.Inaddition,thelimiteddataavailablefiomthe24hoursurveillancetestswasnotstatisticallysignificant.Evenbyincludingalllongduration(greaterthan60minutes)fullload-runs,thedatabasewasstilllimitedto128totaltests.Therefore,dataRomshortduration(60minutesorless)load-runsisprovidedbelowforcomparisonandtofacilitateanalysisofthelongdurationfullload-runs.InevaluatingEDGtestdata,theboundaryandsupportsystemsoftheEDGsystemwereconsideredtoincludethedieselengine,thegenerator,theexciterandvoltageregulatorsystem,thecontrolandprotectionsystem,theEDGlubricatingoilsystem,EDGfueloilsystem,EDGjacketwatersystem,EDGstartingairsystem,EDGintakeairsystem,EDGexhaustgassystem,andthecontrolcircuitryuptotheimmediatecontrolpowersource.ThecriteriainSections3.1and3.2wereusedtoevaluatethevalidityofEDGtestruns.Thefollowingcriteriawerebasedonreferences5,10,and11.Aload-runwascountedasavaliddemand,success,orfailureifitsatisfiedeitheroneofthefollowingcriteria:a.Termination'oftheload-runpriortomissioncompletion,duetoabnormalconditionsthatwouldultimatelyhaveresultedinthefailureoftheEDG,wascountedasavaliddemandandfailure.b.Terminationaftercompletionoftheintendedmissionwascountedasavaliddemandandsuccess.AmericaElectricPointer Attachment4page6Thefollowingcriteriawerebasedonreferences5,10,and11.Aload-runwasnotcountedasavaliddemand,success,orfailurewhentheload-runwasprematurelyterminatedforanyofthefollowingreasons:a.Aspuriousoperationofatripthatisbypassedintheemergencyoperatingmode.b.Amalfunctionofequipmentthatisnotoperativeintheemergencyoperatingmode.c.AnoperatingerrorthatwouldnothavepreventedtheEDGGombeingrestartedandloadedwithinafewminutesandwithoutcorrectiverepairs.d.ObservableabnormalconditionsthatwouldnothavepreventedtheEDGfromcompletingitsmissionduringanactualemergencydemand.33rain-RnaThissectionexaminedEDGload-rundatawhichmetthefollowingcriterion:TheEDGwassuccessfullystartedwithanintentiontooperateatfullloadforadurationgreaterthan60minutes.Theabovecriterionwasused.toeliminatedatafromthe1hourmonthlyruns,whichareperformedathalfload(1750KW),andtocapturefull(3500KW)load-runmissionsofsignificantduration(greaterthan60minutes).Also,thecriteriondistinguishesbetweenEDGload-rundataandEDGstartingdatabydefiningtheload-runphasetobeginonlyafterasuccessfulstart.Forcomparisonandanalysispurposes,thedataf'romshortduration(60minutesorless)load-runsareprovidedseparatelyinSection3.4below.Usingtheabovecriterion,apopulationof128load-runswasexamined.Fromthispopulation,2load-runsweredeletedbecausetheywereactualdemandswhichoperatedtheEDGatlessthanfullload.Also,3additionalload-runsweredeletedbecausetheEDGwasnotoperatedatfullloadcontinuouslyduringthesurveillanceduration.AmericanElectricPopover Sic Attachment4page7Examinationoftheremainingdatashowedthat16fullload-runmissionswereprematurelyaborted.Thisdatarequiredengineeringevaluationbasedonthecriteriapresentedinsections3.1and3.2.Ingeneral,prematureterminationwasassociatedwithminorfueloilleaks,personnelerror,orspuriousoperationormalfunctionofequipmentthatwouldnotbeoperativeduringanemergencycondition.Outofthe16prematureterminations,zeroterminationsweredeterminedtobevalidload-runfailures.Theremaining107load-runsprovidedapproximately2,010hoursofload-rundataforanaverageruntimedurationofapproximately19hours.3.4ShortDurationLoad-RunDataThissectionexaminedEDGload-rundatawhichmetthefollowingcriterion:TheEDGwassuccessfullystartedwithanintentiontooperateforadurationof'notmorethan60minutes.AspartoftheIndividualPlantExamination(IPE),datacollectionandanalysiswerecompletedforEDGsovertheperiodJanuary1983throughDecember1992.Thetestdataprimarilyexaminedmonthlysurveillancerunsandothershortdurationtestswhichinvolveloadingthedieseltoatleast1750KW.Thestudyreviewedatotalof929demandsandfound12startfailuresand3load-runfailures.Pertheabovecriterion,the12startfailuresweresubtractedRomthe929demandstoyield917validload-rundemands.Also,basedonreviewofthedataandthecumulativeruntimehours,itwasdeterminedthatsomeofthe917demandsincludedrunsover1hour(someofthedemandswereactually24hourfullload-runs).InSection3.3above,thelongdurationfullload-runmissionswereexamined.Therefore,topreventdoublecountingandtoconditionthesedemandstoaonehouraverageduration,44successfuldemandsweredeletedGomthe917remainingdemands.Thus,873demandsprovidedatotalof873hoursofoperation.'he3load-runfailuresoccurredwithinthefirst30minutesofoperationduringroutinetesting.AmericanElectricPower
~IN,pP Attachment4page835'IraeriiPriortoanalyzingtheabovedata,itisimportanttoreviewtheunderlyingcharacteristicsofEDGfailures.Ingeneral,EDGfailuresmayoccurasaresultofthefollowing:a)Endofnor7rrallifeofaconrponentorpieceofeqrripment.Ingeneral,preventativemaintenanceprogramswillreducethenumberofage-relatedcomponentfailures.However,thereissomestatisticaluncertaintyassociatedwithEDGcomponentlifetimes.Thisuncertaintycreatesthepossibilityofage-relatedfailures.Thus,EDGcomponentsaresubjecttofailureaftera,requisitenumberofcumulativeservicehours,regardlessofthelengthofthesurveillancerunduringwhichthefailureoccurs.Therefore,toproperlyinterprettheobservationofacomponentfailureduringanygivensurveillance,itisimportanttocarefullystudythenatureofthefailuretodetermineifthecomponentfailurewasprecipitatedbytheactualsurveillancedurationor,rather,bythenormalexpectedlifetimeofacomponent(AppendixBofreference12,providesusefulinformationonage-relatedfailuremechanisms).Ifthefailuredistributionofaparticularcomponentisknown,thenthetimeoffailurecanbepredicted.Therefore,itcanbeconcludedthatexcessivetestingincreasestheprobabilityofsuchfailuresandthatthesurveillancetestdurationshouldbeminimizedtoreducetheriskoffailureduringanactualdemand.b)Rcmdonrnralfirrrctr'onofcomponents,desigrrarromalies,orincor7ectnraintenarrce.Theseproblemsmaysurfaceoveranytimeinterval.Inparticular,itisintuitivelyobviousthattheeffectsofsmalldeviations6'ommaintenanceproceduresmaybemoredifficulttoidentifyandmayrequirealongertimetoappear.Thissuggeststhat,toidentifyallpossibleproblemsofthisnature,thesurveillancetestdurationshouldbemaximized.Notethatitems(a)and(b)presentcompetingrequirements.Accordingly,theobjectiveofEDGsurveillancetestingshouldbetoexposeasmanyofthetype"b"failuresaspossiblewithoutacceleratingthetype"a"failures.If,however,thetesttimeexceedsanoptimalduration,T,thenthenumberoftype"a"failuresbeingintroducedcouldtheoreticallyincreasebeyondtheexposurerateoftype"b"failures.Ideally,thesurveillancetestdurationwouldbeselectedtominimizetype"a"AmericanElectricPopover E~~Agt Attachment4page9failures.Inthiscase,itcanbeshownthatagoodapproximationoftheoptimaltestduration,T,canbedeterminedbyusinganexponentialprobabilitydensityfunctionfortheEDGfailureprob'ability.TheformofthisfunctionisshownbelowandisfurtherdevelopedinSection3.6.wheret=timeand,A.anduareconstants.Itisalsoworthnotingthatitisreasonabletoexpectthat,foramachinethatisdesignedforextendedoperationatfullyloadedconditions,amajorityoftheprimaryfailuremodeswouldoccurbeforetheenginereachesthermalequilibrium(usuallywithin2hours).Engineequilibriumtemperatureisdefinedasthetimeatwhichthejacketwaterandlubeoiltemperaturesarebothwithin+10'Foftheirnormaloperatingtemperaturesestablishedbytheenginemanufacturer(reference11).Theexpectancyofahigher.'ailureprobabilityduringthefirsttwohoursofoperationisconsistentwithCookNuclearPlantactualperformancedata,asindicatedinSections3.3and3.4.3~6DAdInconsiderationofthefailurecharacteristicsdiscussedinSection3.5,thefollowingbasicquestionswereaddressedinthisstudy:a)HotvdoesfhefarlureprobabilityChrlingshortdurafionrunsconrparewilhlongdurationruns?b)OvertvhatfestdurationtvouldthemnjorityofFDGload-sanfailuresbeexpected?Toanswertheabove'questions,itisnecessarytodevelopatechnicalapproachforevaluatingEDGunreliabilityasafunctionofsurveillancetesttime.'herefore,thepurposeofthissectionistouseactualperformancedatatoestablisha..statisticalmodelthatconservativelycharacterizesEDGfailureprobabilitiesovertime.AmericanElectricPoimr s44 Attachment4page10Table1:DataSan>maryShortDurationLoad-RunData1983-1993TotalDemands873TotalFailuresAverageDurationHoursTotalHours873L'ongDurationLoad-RunData1974-1994TotalDemands107*cppQxUM1cvllllcTotalFailuresAverageDurationHours19'otalHours2,010VTheabovedatacanbereasonablycharacterizedasasetofBernoullitrials.Thatis,eachvalidtest(ortrial)caneitherbeasuccessorfailure,eachsuccesshasaconstantprobabilityp,andeachtestisindependentoftheprevioustest.Itisrecognizedthatsomeofthetestsmaynotbeindependentandidentical,however,inmanycasestheassumptionsofaseriesofBernoullitrialswillprovideagoodapproximation..Therefore,thebinomialdistributionisareasonableselectionforevaluatingthistypeofdata.BinomialDistributionZ=(n!/[(n-x)!*x!])*p"*(I-p)'whereZequalstheprobabilitythatoutofntrialsxfailureswillbeobservedgiventhefailureprobability,p.Also,itcanbeproventhatthemostprobablevalueofpisp=x/nForexample,fortheshortdurationruns,p~=3/873=0.0034andZ~=0.2244AmericanElectricPoiver Attachment4page11However,forthelongdurationrunswithxt,W,adifferentapproachmustbetakentocalculatep(i.e.,theresultpt,Wisnotstatisticallymeaningfultoanswertheabovequestions).Byutilizingthebinomialdistributionfunctionandtheabovedata,arangefortheparameter,p,canbedeterminedforthelongdurationload-runs.Foranygivenvalueofn,itcanbeshownthattheinequality,Z(x=0)>Z(x=1)isaboundingcaseandresultsinpp~and,therefore,0.0034<pt<0.0093Forconservatism,assumept&.0093,thentheratioptgp~=274Toanswerquestion"a"above,thisresultmeansthatthefailureprobabilityassociatedwiththelongdurationload-runsis2.74timesgreaterthanthefailureprobabilityforshortdurationload-runs.TableZ:$0mmaryofFnilareProbabilitiesShortOurationLoad-RunOataProbabilities1983-1993¹Trialsn873FailuresXFailureProbabilityshort0.0034FailureRate/Hour0.0034Zshort0.2244LongOurationLoad-RunProbabilities1974-1994¹TrialsnFailuresXFailureProbabilitypLongFailureRate(/Hour)ZLong107a'calculatedvalue.0.00930.0005'.3680AmericanElectricPopover 8lLi4 Attachment4page12AsdiscussedinSection3.5,theoptimaltestdurationwouldproducefailurespredominatelyduetotype"b"occurrences.(random),asopposedtotype"a"(componentwear).Assuch,selectinganoptimaltestdurationwouldmeanthatnon-failureinprevioustestperiodswouldnotsignificantlychangethefailureprobabilityinthefutureperiodsandalsothatthefailureratewouldbeapproximatelyconstantovertime.Todeterminetheoptimaltestduration,theintegraloftheprobabilitydensityfunction,introducedinSection3.5,canbeusedtomodeltheprobabilityofEDGfailuresthatcouldbeobservedoveratimeperiod;(O,t}.EDGFailureProbabilityFunction,P(t)P(t)=(afA,)[l-e']wherecandA,areconstants.TheconstantA,isdependentonthefailureprobabilityoftheshortandlongdurationload-runswhichweredefinedasshortdurationruns560minutes<longdurationnmsFromTables1and2above,att=1hour,P=0.0034andatt=19hours,P=0.0093.ThesevaluescanbesubstitutedintotheaboveexpressiontofirstcalculateXandthenct.Theresultingvaluesareshownbelow.0,=0.004294X,=0.4637Figure1,below,showsaplotofP(t)usingtheabovevaluesofaandA,overatimeperiod,t,fromzerototwenty-fourhours.AmericanElectricPopover
\1k$
Attachment4-page13EDGFailureProbability0.00950.00900.00850.00800.00750.00700.00650.00600.0055P(t)0.00450.00400.00350.00300.00250.00200.00150.00100.00050.0000024681012141618202224TIME(hours)Figurel:PlotofI'(t)Figure1showsthatafterapproximatelytenhourstheEDGfailureprobabilitywould,increaseataveryslowrate(i.e.,dP(t)/dt=0).Tobetterunderstandthisresultandtoanswerquestion"b"above,aplotofP(t)normalizedbyP(M4hrs)isshownbelow.AtnericanElectricPower 4h~J~lIl4 Attachment4page14k1.00000.95000.90000.85000.80000.75000.70000.65000.60000.5500P(t)IP(24)05000OA5000.40000.35000.30000.25000.20000.15000.10000.05000.00000CumulativeFailureProbability24681012141618202224TIME(hovrs)Fignre2:EDGCasnllativeFailureI'robabilityFromfigure2,itcanbeseenthatapproximately98%ofthefailuresoccurwithinthefirst8hourswhileonly60%occurwithinthefirst2hoursofoperation.Thismaybeanoverlyconservativeestimationfiomthestandpointthat,atCookNuclearPlant,alloftheactualobservedfailuresoccurredwithinthefirst2hoursofoperation.AmericanElectricPopover 4~1r~1'Wp'IIP4y'l~rtW Attachment4page154.0ReviewofSelectedStudiesThepurposeofthissectionistoprovideameasureofEDGload-runperformanceonanindustrywidebasis.ThestudiesreviewedbelowwereselectedbecausetheyprovideddiscernibleEDGperformancedataduringtheload-runphaseofoperation.(reference7)In1985,BrookhavenNationalLaboratorycompletedastudytoexamineEDGperformanceatNuclearPowerPlantsintheUnitedStatesduringafouryearperiod,from1980through1983.Thereportdidnotdistinguishbetweenstartandload-runfailuresbut,rather,calculatedoverallfailureratesnormalizedbytotalEDGyearsofoperation.DatawascompiledfromLERfiles,10CFR50.55(e)reports,NuclearPlantReliabilityDataSystem,andElectricPowerResearchInstitutedocumentfiles.Thestudyincluded158EDGs,producedby12differentmanufacturers(includingWorthington),foratotalof588EDGyears.Thereportedtotalcompositenumberoffailureswas396whichcorrespondstoafailurerateofapproximately0.7failuresperEDGyear.ThereportalsostudiedthefailuremodesofdifferentEDGmanufacturers,withtheexceptionofTrans-AmericaDelaval.Also,thestudyprovidedaqualitativepresentationofpredominantfailuremodesbasedonthenumberofreportedfailures.Reviewofthereportindicatesthatthefivemostdominantfailuremodesincludedinstrumentationandcontrolssystems,lubricatingoilsystems,speedandloadcontrol,coolingwatersystems,andstartingsystems.In1986,EPRIcompletedacomprehensivestudyofEDGsuccess/failureexperienceatNuclearPowerPlantsintheUnitedStatesduringathreeyearperiodfrom1983through1985.Thestudyreviewedstartandload-runreliability.Thecriteriausedtoevaluateload-rundatawassimilartothatusedinSection3.0above.Theload-rundatabaseincludedbothplannelandunplanned(actual)demandsthatinvolvedEDGoperationatgreaterthan50%ofplantessentialsafetyfunctionloadratingforonehourorlonger.Fortheperiodunderstudy,adatabaseof13,808load-runswasassembled.Thestudyreported138load-runfailureswhichcorrespondedtoanEDGunreliabilityof0.0100failuresperdemand.AmericanElectricPoiver C
Attachment4page16/R-54(reference4)In1994,NUREG/CR-5994presentedrecentoperatingexperiencetoassessEDGunavailabilityduetotesting,maintenance,andfailuresduringreactorpoweroperationandduringplantshutdown.Section3ofthereportprovidedEDGindustry-wideperformancedataoverfouryears,1988through1991,andusedtheempiricalHayes.methodtoanalyzeEDGMuredata.Thedatacovers195EDGsat63plantsitesandincludesbothactualandtestload-rundata.Forplantsites,themeanload-runfailureprobabilitywas9.5E-03failuresperdemandwithavarianceof2.9E-05.Tnble3:EDGIndustry->videFnilareProbabilityHistory1980-1983(reference7)numberofFailures396'umberofEDGs158numberofEDGears588failuresperEDGear0.71983-1985(reference5)numberofFailures138numberofEDGs154numberofload-runs13,808failuresperdemand0.011988-1991(reference4)numberofFailures182numberofEDGs195numberofload-runs19,520failuresperdemand0.0095'Includesbothstart4loadiefailures.AmericanElectricPotver Attachment4page175.0IndustSurvesInanefforttoprovidemorerecentEDGperformancedata,alimitedinformalindustrysurveywasconductedtospecificallydeterminetheapproximatenumberoffailuresoccurringbefore8hoursandthenumberoffailuresoccurringafter8hoursduringthe18month24hoursurveillancetest.Reviewofthedataindicatedthattheindustiy-wideEDGfailureratesduringthe24hoursurveillancetendtosupportthepreviousstudiesdiscussedinSection4.0(i.e,.failures/load-rundemand<0.01).6.0ConclusionsExaminationofthefullloadrundataontheEDGsatCookNuclearPlantrevealedthat3validload-runMuresoccurredduringshortdurationruns(t~60minutes)andzerofailuresoccurredduringlongdurationruns(t>60minutes).Aconservative,statisticallybased,mathematicalmodelwasdevelopedtodeterminethereliabilityoftheEDGduringtheload-runphaseoftestingasafunctionoftime.Themodelpredicatedthat60%ofthevalidload-runfailureswouldoccurwithinthefirsttwohoursof.operation,95%inthefirst6.5hours,and98%inthefirsteighthours.Also,themorerecentindustrysurveys,includingtheinformalrandomsurveysdiscussedinSection5.0,areveryconsistentwiththeresultsfoundatCookNuclearPlant(i.e.,failures/load-rundemand<0.01).AsexplainedinSection3.5,theobjectiveoftheendurancetestshouldbetoexposeasmanyoftherandomtypefailuresaspossibleandtominimizeEDGcomponentwear.Thus,theoptimalsurveillancetestduration,T,shouldbeselectedtoprovidereasonableassurancethatthemajorityofincipientrandomfailuresareexposedwithoutadverselyimpactingEDGavailability,foranactualemergencydemand.Also,consistentwithactualperformancedataandengineeringjudgment,theperiodofliigheststresswouldoccurduringstartupandbeforeequilibriumconditionsareestablished(05T<2hours).Therefore,basedonactualperformancedataandrelatedindustry-widesurveys,anendurancetestdurationofT=8hoursisrecommendedtoprovidethenecessaryinsightaboutEDGreliability.Testdurationsgreaterthaneighthourswouldnotprovideasignificantoperatingsafetybenefitandwouldonlyservetoincreasecumulativeruntimeandthelikelihoodofage-relatedcomponentfailures.AmericanElectricPosver
~~iQt Attachment47.0Referencespage181)"AReliabilityProgramforEmergencyDieselGeneratorsatNuclearPowerPlants-ProgramStructure,"NUREG/CR-5078,SAND87-7176,Vol.1,April1988.2)"AReliabilityProgramforEmergencyDieselGeneratorsatNuclearPowerPlants-Maintenance,Surveillance,andConditionMonitoring,"NUREG/CR-5078,SAND87-7176,Vol.2,December1988.3)"ImprovementstoTechnicalSpecificationSurveillanceRequirements,"NUIREG-1366,December1992.4)"EmergencyDieselGenerator:MaintenanceandFailureUnavailability,andTheirRiskImpacts,"NUREG/CR-5994,BNL-NUREG-52363,November1994.5)"TheReliabilityofEmergencyDieselGeneratorsatU.S.NuclearPowerPlants,"NSAC-108,September1986.6)"ASimulationStudy:EmergencyDieselGeneratorAvailability,"EPRINP-1759,Interim.Report,March1981.7)"AReviewofEmergencyDieselGeneratorPerformanceatNuclearPowerPlants,"BrookhavenNationalLaboratory,TechnicalReportA-31341-85,January1985.8)"GuidelinesandTechnicalBasesforNUMARCInitiativesAddressingStationBlackoutatLightWaterReactors,"NU1VIARC87-00,Rev.1,August1991.9)"PeriodicTestingofDieselGeneratorUnitsUsedasOnsiteElectricPowerSystemsat.NuclearPowerPlants,"RegulatoryGuide1.108,Revision1,August1977.10)"Selection,Design,Qualification,andTestingofEmergencyDieselGeneratorUnitsUsedasClass1EOnsiteElectricPowerSystemsatNuclearPowerPlants,"RegulatoryGuide1.9,Revision3,July1993.11)"IEEEStandardPeriodicTestingofDiesel-GeneratorUnitsAppliedasStandbyPowerSuppliesinNuclearPowerGeneratingStations,"IEEEStd749-1983.12)"IEEEStandardCriteriaforDiesel-GeneratorUnitsAppliedasStandbyPowerSuppliesinNuclearPowerGeneratingStations,"IEEEStd387-1984.AmericanElectricPower 1~lll'lai Attachment4page1913)."IEEEGuidetotheCollectionandPresentationofElectrical,Electronic,SensingComponent,andMechanicalEquipmentReliabilityDataforNuclear-PowerGeneratingStations,".IEEEStd500-198414)"ResolutionofGenericSafetyIssueB-56,DieselGeneratorReliability,"SECY-93-044,~02/22f93.AmericanElectricPointer Pph ATTACHMENT5TOAEP:NRC:0896XDONALDC.COOKNUCLEARPLANTCOSTBENEFICIALLICENSINGACTIONEDGSURVEILLANCETESTINGTECHNICALSPECIFICATIONCHANGES I4Erf,'%jg Attachment5toAEP:NRC:0896KPage1RegulatoryRequirement:TechnicalSpecificationT/S4.8.1.1.2.e.7requiresschedulingofEDG18monthsurveillanceduringshutdown(refuelingoutages)andperforminga24hourendurancetestrun.EffectofRequirement:Thepresent18month24hourEDGendurancetestdoesnotprovideasafetybenefitcommensuratewithitscost.Inaddition,performingtherequiredtestingcausescriticalpathcomplicationsanddelaysduringtheoutage.RationaleforRegulatoryChange:TherationalefortheproposedchangesisthattheintentofregulatorysurveillancerequirementtodemonstratetheabilityoftheEDGtooperateforanextendedperiodoftimeunderfullyloadedconditionswillbepreservedbymaintainingan8hourendurancetestrun.Atthesametime,thesechangeswillbetterutilizeplantresourcesandpreventcriticalpathcomplicationsduringtheoutages.Unnecessarydieselgeneratorstressandwearwillalsobereduced.Additionally,thereductionfrom24to8hourswillincreasedieselavailabilityand,thereby,reduceshutdownrisk.ApproximateCostofRequirement:Thecostsavingsassociatedwithreducingthe18month24hourEDGsurveillancetestdurationisprovidedbelow.Thecostsavingswascalculatedusinganestimateoflaborandmaterialsrequiredforperformingthe24hourendurancerun(ontwodieselsperunit)andmultiplyingthisestimatebytheproposedreductioninsurveillanceduration.Unit1:remainingcycles13Unit2:remainingcycles152*13*16,000*16/24$277,3332*15*16,000*16/24$320,000Totalplantlifetimesavingsisapproximately$600,000.P
~e4