ML003770849

Draft Regulatory Guide DG-1096, Transient and Accident Analysis Methods

ML003770849
Person / Time
Issue date:	12/31/2000
From:	Office of Nuclear Regulatory Research
To:
Lauben N (301)415-6762
References
-nr, DG-1096 SRP Section 15.0.2
Download: ML003770849 (39)

Text

Thisregulatoryguideisbeingissuedindraftformtoinvolvethepublicintheearlystagesofthedevelopmentofaregulatorypositioninthisarea.It hasnotreceivedcompletestaffrevieworapprovalanddoesnotrepresentanofficialNRCstaffposition.Publiccommentsarebeingsolicitedonthisdraftguide(includinganyimplementationschedule)anditsassociatedregulatoryanalysisorvalue/impactstatement.Commentsshouldbeaccompaniedbyappropriatesupportingdata.WrittencommentsmaybesubmittedtotheRulesandDirectives Branch,OfficeofAdministration,U.S.NuclearRegulatoryCommission,Washington,DC20555-0001.Commentsmaybesubmittedelectronicallyor downloadedthroughtheNRC'sinteractivewebsiteat<WWW.NRC.GOV>throughRulemaking.CopiesofcommentsreceivedmaybeexaminedattheNRCPublicDocumentRoom,11555RockvillePike,Rockville,MD.CommentswillbemosthelpfulifreceivedbyFebruary15,2001.Requestsforsinglecopiesofdraftoractiveregulatoryguides(whichmaybereproduced)orforplacementonanautomaticdistributionlistforsinglecopiesoffuturedraftguidesinspecificdivisionsshouldbemadetotheU.S.NuclearRegulatoryCommission,Washington,DC20555, Attention:ReproductionandDistributionServicesSection,orbyfaxto(301)415-2289;orbyemailtoDISTRIBUTION@NRC.GOV.ElectroniccopiesofthisdraftguideareavailablethroughNRC'sinteractivewebsite(seeabove),ontheNRC'swebsite<www.nrc.gov>intheReferenceLibraryunderRegulatoryGuides,andinNRC'sPublicElectronicReadingRoomatthesamewebsite,underAccessionNumberML003770849.U.S.NUCLEARREGULATORYCOMMISSIONDecember2000OFFICEOFNUCLEARREGULATORYRESEARCHDivision1DraftDG-1096DRAFTREGULATORYGUIDEContact:N.Lauben(301)415-6762DRAFTREGULATORYGUIDEDG-1096TRANSIENTANDACCIDENTANALYSISMETHODSA.INTRODUCTIONIn10CFRPart50,"DomesticLicensingofProductionandUtilizationFacilities,"Section50.34,"ContentsofApplications;TechnicalInformation,"requiresthat:1.SafetyAnalysisReportsbesubmittedthatanalyzethedesignandperformanceofstructures,systems,andcomponentsprovidedforthepreventionofaccidentsandthe mitigationoftheconsequencesofaccidents,and2.Analysisandevaluationofemergencycorecoolingsystem(ECCS)coolingperformancefollowingpostulatedloss-of-coolantaccidents(LOCAs)beperformedinaccordancewith therequirementsof10CFR50.46.Thetechnicalspecificationsforthefacility(10CFRPart50.36)aretobebasedonthesafetyanalysis.ThisregulatoryguideisbeingdevelopedtodescribeaprocessthatisacceptabletotheNRCstaffforthedevelopmentandassessmentofevaluationmodelsthatmaybeusedtoanalyze transientandaccidentbehavior.Chapter15oftheStandardReviewPlan(SRP)(NUREG-0800, Ref.1)andtheStandardFormatandContentGuide(RegulatoryGuide1.70,Ref.2)describe theseevents(transientsandaccidents),whichareasub-setofthoserequiredby10CFR50.34to beaddressed.TheseeventsarepresentedinSections15.1through15.6oftheSRP,exceptfor thefuelassemblymisloadingeventandallradiologicalconsequenceanalyses.Anappendixto thisregulatoryguideisprovidedforECCSanalysis.Asappropriate,otherappendiceswillbe developedforotherspecificclassesofeventsthataredescribedinSRPSections15.1through 15.6toaddressphenomena,assessment,uncertaintyanalyses,andotherfactorsimportantor uniquetoaparticularclassofevents.Thisregulatoryguideisintendedtoprovideguidanceonrealisticaccidentanalyses,whichwillprovideamorereliableframeworkforrisk-informedregulationandabasisforestimatingthe uncertaintyinunderstandingtransientandaccidentbehavior.

2Section15.0.2oftheSRP(Ref.1)providesguidancetoNRCreviewersoftransientandaccidentanalysismethods.ThisregulatoryguideandSRPSection15.0.2 coverthesamesubjectmaterialandaremeanttobecomplementarydocuments,with Section15.0.2providingguidancetoreviewersandthisguideprovidingpracticesand principlesforthebenefitofmethodsdevelopers.Chapter15oftheSRPrecommends thatapprovedevaluationmodelsorcodesbeusedfortheanalysisofmostidentifiedevents.TheSRPsuggeststhatevaluationmodelreviewsbeinitiatedwheneveranapprovedmodelforaspecifiedplanteventdoesnotexist.Iftheapplicantorlicenseeproposestouseanewmodel,anevaluationmodelreviewshouldbeinitiated.Thisguideisintendedtobeusedbyconstructionpermitapplicantsthatmustmeetthedesignbasesdescriptionrequirementsof10CFR50.34andtherelationofthedesign basestotheprincipaldesigncriteriadescribedinAppendixAto10CFRPart50.Chapter 15oftheSRP(Ref.1)describesthetransientsandaccidentsthattheNRCstaffreviews aspartoftheapplication,andthecriteriaofAppendixAthatspecificallyapplytoeach classoftransientandaccident.Chapter15alsostatesthatacceptableevaluationmodels shouldbeusedtoanalyzethesetransientsandaccidents.Thisguideisalsointendedtobeusedbyoperatinglicenseapplicantsthatmustmeetthedesignbasesdescriptionrequirementsof10CFR50.34andtherelationofthedesignbasestotheprincipaldesigncriteriadescribedinAppendixAto10CFRPart50.ThisguidewouldbeapplicabletonewevaluationmodelsorchangestoexistingevaluationmodelsproposedbyoperatingreactorlicenseesthattheNRCstaffundertakes toreview.RegulatoryguidesareissuedtodescribetothepublicmethodsacceptabletotheNRCstaffforimplementingspecificpartsoftheNRC'sregulations,toexplaintechniques usedbythestaffinevaluatingspecificproblemsorpostulatedaccidents,andtoprovide guidancetoapplicants.Regulatoryguidesarenotsubstitutesforregulations,and compliancewithregulatoryguidesisnotrequired.Regulatoryguidesareissuedindraft formforpubliccommenttoinvolvethepublicindevelopingtheregulatorypositions.Draft regulatoryguideshavenotreceivedcompletestaffreview;theythereforedonotrepresent officialNRCstaffpositions.Theinformationcollectionscontainedinthisdraftregulatoryguidearecoveredbytherequirementsof10CFRPart50,whichwereapprovedbytheOfficeofManagementandBudget, approvalnumber3150-0011.Ifameansusedtoimposeaninformationcollectiondoesnot displayacurrentlyvalidOMBcontrolnumber,theNRCmaynotconductorsponsor,anda personisnotrequiredtorespondto,theinformationcollection.B.DISCUSSIONThetwofundamentalfeaturesoftransientandaccidentanalysismethodsare(1)theevaluationmodelconceptand(2)thebasicprinciplesimportantforthedevelopment, assessment,andreviewofthosemethods.EVALUATIONMODELCONCEPT 3Thebasisforanalysismethodsusedtoanalyzeaparticulareventorclassofeventsiscontainedintheevaluationmodelconcept.Thisconceptisdescribedin10CFR 50.46forLOCAanalysisbutcanbegeneralizedtoallanalyzedeventsdescribedin Chapter15.Anevaluationmodel(EM)isthecalculationalframeworkforevaluatingthe behaviorofthereactorsystemduringapostulatedtransientordesignbasisaccident.It mayincludeoneormorecomputerprograms,specialmodels,andallotherinformation necessaryforapplicationofthecalculationalframeworktoaspecificevent,suchas:1.Proceduresfortreatingtheinputandoutputinformation,particularlythecodeinputarisingfromtheplantgeometry,theassumedplantstateat transientinitiation,2.Specificationofthoseportionsoftheanalysisnotincludedinthecomputerprogramsforwhichalternativeapproachesareused,and3.Allotherinformationnecessarytospecifythecalculationalprocedure.Itistheentiretyofanevaluationmodelthatultimatelydeterminesthattheresultsareincompliancewithapplicableregulations.Therefore,theentireevaluationmodel mustbeconsideredduringthedevelopment,assessment,andreviewprocess.Inthisregulatoryguide,thetermmodelisalsousedandshouldbedistinguishedfromtheevaluationmodelorEM.IncontrasttoEMasdefinedhere,modelwithoutthe evaluationmodifierisusedinthemoretraditionalsensetodescribetherepresentationof aparticularphysicalphenomenonwithinacomputercodeorprocedure.MostevaluationmodelsusedtoanalyzetheeventsinChapter15oftheSRP(Ref.1)relyonasystemscodethatdescribesthetransportoffluidmass,momentum,and energythroughoutthereactorcoolantsystems.Theextentandcomplexityofthephysical modelsneededinthesystemscodearestronglydependentonthereactordesignandthe transientbeinganalyzed.Foraparticulartransient,asubsidiarydevicelikeasub-channelanalysiscodemayactuallybemorecomplexthanthesystemscode.Regardless ofitscomplexity,thesystemscodeplaysakeyroleinorganizingandcontrollingother aspectsofthetransientanalysis.Eachcomputercode,analyticaltool,orcalculational procedurethatcomposetheevaluationmodelisreferredtoasa"calculationaldevice"in thisguide.Insomecases,asmanyas7or8calculationaldevicesmaybeusedtodefineanevaluationmodelforaparticularevent,althoughthetrendtodayistointegratemanyof thesecomponentsintoasmallersetofcomputercodes,usuallywithintheframeworkof thesystemscode.Sometimes,ageneralpurposesystemscodemaybedevelopedtoaddresssimilarphenomenologicalaspectsofseveraldiverseclassesoftransients.Thispresentsunique challengesinthedefinition,development,assessment,andreviewofthosecodesasthey applytoaparticulartransientevaluationmodel.AseparatesectionoftheRegulatory Positionisdevotedtotheissuesinvolvedwithgeneralpurposecomputercodes.BASICPRINCIPLESOFEVALUATIONMODELDEVELOPMENTANDASSESSMENTRecentreviewshaveshowntheneedtoprovideguidancetoapplicantsandlicenseesregardingtransientandaccidentanalysismethods.Byprovidingsuch guidance,thereviewprocessshouldbestreamlinedbyreducingthefrequencyandextent 4ofiterationsbetweenthemethodsdevelopersandNRCstaffreviewers.Toproduceaviableproduct,certainprinciplesshouldbeaddressedduringthemodeldevelopmentand assessmentprocess.Therearesixbasicprinciplesthathavebeenidentifiedasimportanttofollowintheprocessofevaluationmodeldevelopmentandassessment.Theyare:1.Determinerequirementsfortheevaluationmodel.Thepurposeofthisprincipleistoprovideafocusthroughouttheevaluationmodeldevelopmentandassessment process(EMDAP).Animportantoutcomeshouldbetheidentificationof mathematicalmodelingmethods,components,phenomena,physicalprocesses, andparametersneededtoevaluatetheeventbehaviorrelativetothefiguresof meritdescribedinChapter15oftheSRPandderivedfromtheGeneralDesign Criteria(GDC)inAppendixAto10CFRPart50.Thephenomenaassessment processiscentraltoensuringthattheevaluationmodelcananalyzetheparticular eventappropriatelyandthatthevalidationprocessaddresseskeyphenomenafor thatevent.2.Developanassessmentbaseconsistentwiththedeterminedrequirements.

Sinceanevaluationmodelcanonlyapproximatephysicalbehaviorforpostulatedevents, itisimportanttovalidatethecalculationaldevices,individuallyandcollectively, usinganappropriateassessmentbase.Thedatabasemayconsistofalready existingexperimentsoritmayrequiretheperformanceofnewexperiments, dependingontheresultsoftherequirementsdetermination.3.Developtheevaluationmodel.Thecalculationaldevicesneededtoanalyzetheeventsinaccordancewiththerequirementsdeterminedinthefirstprincipleshould beselectedordeveloped.Todefineanevaluationmodelforaparticularplantand event,itisalsonecessarytoselectpropercodeoptions,boundaryconditions,and thetemporalandspatialrelationshipamongthecomponentdevices.4.Assesstheadequacyoftheevaluationmodel.Basedontheapplicationofthefirstprinciple,especiallythephenomenaimportancedetermination,anassessment shouldbemaderegardingtheinherentcapabilityoftheevaluationmodelto achievethedesiredresultsrelativetothefiguresofmeritderivedfromtheGDC.

Someofthisassessmentisbestmadeduringtheearlyphaseofcode developmenttominimizetheneedforcorrectiveactionslater.Akeyfeatureofthe adequacyassessmentistheabilityoftheevaluationmodeloritscomponent devicestopredictappropriateexperimentalbehavior.Onceagain,thefocus shouldbeontheabilitytopredictkeyphenomenaasdescribedinthefirst principle.Toalargedegree,thecalculationaldevicesarecollectionsofmodels andcorrelationsthatareempiricalinnature.Therefore,itisimportanttoassure thattheyareusedwithintherangeoftheirassessment.5.FollowanappropriatequalityassuranceprotocolduringtheEMDAP.Qualityassurancestandards,asrequiredinAppendixBto10CFRPart50,areakey featureofthedevelopmentandassessmentprocess.Whencomplexcomputer codesareinvolved,peerreviewbyindependentexpertsshouldbeanintegralpart ofthequalityassuranceprocess.

56.Providecomprehensive,accurate,up-to-datedocumentation.ThisisanobviousrequirementforacredibleNRCreview.Itisalsoclearlyneededforthepeer reviewdescribedinthefifthprinciple.Sincethedevelopmentandassessment processmayleadtochangesintheimportancedetermination,itismostimportant thatdocumentationofthisactivitybedevelopedearlyandkeptcurrent.TheprinciplesofanEMDAPweredevelopedandappliedinastudyonquantifyingreactorsafetymargins(Ref.3).Inthatreport,thecodescaling,applicability,and uncertainty(CSAU)evaluationmethodologywasappliedtoalarge-breakLOCA.The purposeofthatstudywastodemonstrateamethodthatcouldbeusedtoquantify uncertaintiesasrequiredbythebest-estimateoptiondescribedinthe1988revisiontothe ECCSRule(10CFR50.46).Whilethegoalwasrelatedtocodeuncertaintyevaluation, theprinciplesderivedtoachievethatgoalinvolvedtheentireprocessofevaluationmodel developmentandassessment.Thusmanyofthesameprincipleswouldapplyevenifa formaluncertaintyevaluationwasnotthespecificgoal.Sincethepublicationof Reference3,therehavebeenseveralapplicationsoftheCSAUprocesswith modificationstofiteachparticularcircumstance(SeeReferences4-12).InReferences4and5,aprocesswasdevelopedusinganintegratedstructureandscalingmethodologyforsevereaccidenttechnicalissueresolution(ISTIR).ISTIRdefined separatecomponentsforexperimentationandcodedevelopment.Althoughacode developmentcomponentisincludedinISTIR,theISTIRdemonstrationdidnotinclude codedevelopment.AnimportantfeatureofReference4istheuseofhierarchicalsystem decompositionmethodstoanalyzecomplexsystems.IntheISTIRdemonstration,the methodswereusedtoinvestigateexperimentalscaling,buttheyarealsowellsuitedto providestructureintheidentificationofevaluationmodelfundamentals.Reference6wasanadequacyevaluationofRELAP5forsimulatingAP600small-breakLOCAs(SBLOCAs).Mostofthateffortfocusedondemonstratingtheapplicability andassessmentofadevelopedcodeforanewapplication.ThesubjectsaddressedinReferences3-6arecomplex,andthestructuresusedtoaddressthesesubjectsareverydetailed.TheEMDAPdescribedinthisguideisalso detailed,sothatitcanbeappliedtothecomplexeventsdescribedinSRPChapter15.

Thisisparticularlytrueiftheapplicationisneworthemethodsproposedarenew.The risk-importanceoftheeventorthecomplexityoftheproblemshoulddeterminethelevel ofdetailneededtodevelopandassessanevaluationmodel.Forsimplerevents,manyof thestepsintheprocessmayonlyneedtobeaddressedbriefly.Also,ifanewevaluation modelonlyinvolvesanincrementalchangetoanexistingevaluationmodel,theprocess maybeshortenedaslongastheeffectofthechangeisthoroughlyaddressed.Anoverall diagramoftheEMDAPprocessandtherelationshipofitselementsisshowninFigure1.

6Gui danceonmeth odsfor calculati ngtrans ient and acci dent behaviorisprovidedinthefollowingRegulatoryPosition.AppendixAprovidesadditional informationimportanttoECCSanalysis.TheRegulatoryPositionaddressesfourrelated aspectsofevaluationmodeldevelopmentandassessment.Theyare:

71.DescriptionofthefourelementsandincludedstepsintheEMDAPbasedonthefirstfourprinciplesdescribedaboveandshowninFigure1.2.Therelationshipofacceptedqualityassurancepracticestothisprocessandtheincorporationofpeerreviewasdescribedinthefifthprinciple.3.Adescriptionofwhatshouldbeincludedinevaluationmodeldocumentationtobeconsistentwiththesixthprinciple.4.Theuniqueaspectsofgeneralpurposecomputerprograms.C.REGULATORYPOSITION1.EVALUATIONMODELDEVELOPMENTANDASSESSMENTPROCESS(EMDAP)ThebasicelementsdevelopedtodescribeanEMDAPdirectlyaddressthefirstfourprinciplesdescribedintheDiscussionsectionandareshowninFigure1.ThisRegulatoryPositionaddressesthefourelementsandtheadequacydecisionshowninFigure1.AdherencetoanEMDAPfornewapplicationsoracompletelynewevaluationmodels couldinvolvesignificantiterationswithintheprocess.However,thesameprocessapplies evenifthenewevaluationmodelistheresultofrelativelysimplemodificationstoan existingevaluationmodel."Feedback"loopsarenotshown;rather,theyareaddressed intheadequacydecisiondescribedinRegulatoryPosition1.5.1.1Element1-EstablishRequirementsforEvaluationModelCapabilityItisveryimportanttodetermine,atthebeginning,theexactapplicationenvelopefortheevaluationmodelandtoidentifyandagreeupontheimportanceofconstituent phenomena,processes,andkeyparameterswithinthatenvelope.Figure2illustratesthe stepswithinthiselement.1.1.1Step1.SpecifyAnalysisPurpose,TransientClass,andPowerPlantClassThefirststepinestablishingevaluationmodelrequirementsandcapabilitiesisspecificationoftheanalysispurposeandidentificationoftheclassofplantsandclassof transientstobeanalyzed.Specificationofthepurposeisimportantbecauseanyspecific transientmaybeanalyzedfordifferentreasons.Forinstance,aSBLOCAmaybe analyzedtoassessthepotentialforpressurizedthermalshock(PTS)ortoassess compliancewith10CFR50.46.Thestatementofpurposeinfluencestheentireprocessof development,assessment,andanalysis.Evaluationmodelapplicabilityisscenario-dependentbecausethedominantprocesses,safetyparameters,andacceptancecriteria changefromonescenariotoanother.Thetransientscenario,therefore,dictatesthe processesthatmustbeaddressed.Acompletescenariodefinitionisplant-specific becausethedominantphenomenaandtheirinteractionsdifferinvaryingdegreeswiththe reactordesign.

8 ForeventsdescribedinChapter15oftheSRP,thesestepsshouldbestraight-forward.Thepurpose iscompliancewiththeGDC;theeventsandeventclassesaredescribedinChapter15.

Thelicenseeorapplicantandevaluationmodeldevelopershouldthenspecifytheir applicabilitytoplantsandplanttypes.Asexamples,fueldesign,coreloading,number anddesignofsteamgenerators,numberanddesignofcoolantloops,safetyinjection systemdesign,andcontrolsystemscandiffersignificantlyfromplanttoplantandwill significantlyinfluencescenariobehavior.1.1.2Step2.SpecifyFiguresofMeritFiguresofmeritarethosequantitativestandardsofacceptancethatareusedtodefineacceptableanswersforasafetyanalysis.TheGDCinAppendixAto10CFRPart 50describegeneralrequirementsformaintainingthereactorinasafeconditionduring normaloperationandduringtransientsandaccidents.Chapter15oftheSRPfurther definesthesecriteriaintermsofquantitativefuelandreactorsystemdesignlimits (departurefromnucleateboilingratio(DNBR)limits,fueltemperatures,etc.)fortheevents ofinterest.ForECCSdesign,fivespecificcriteriadescribedin10CFR50.46mustbe metforLOCAanalysis.Thus,forChapter15events,figuresofmeritaregenerally synonymouswithcriteriadirectlyassociatedwiththeregulations,andtheirselectionis usuallyasimplematter.Duringevaluationmodeldevelopmentandassessment,a temporary"surrogate"figureofmeritmaybeofvalueinevaluatingtheimportanceof phenomenaandprocesses.Section2.5ofReference7describesahierarchyofcriteria thatwasusedinSBLOCAassessment,inwhichvesselinventorywasjudgedtobemore 9valuableindefiningandassessingcodecapability.Justificationforusingasurrogatefigureofmeritshouldbeprovided.1.1.3Step3-IdentifySystems,Components,Phases,Geometries,Fields,andProcessesThatMustBeModeledThepurposeofthisstepistoestablishtheevaluationmodelcharacteristics.InReferences4and5,hierarchicalsystemdecompositionmethodsareusedtoinvestigate scalingincomplexsystems.Thesemethodscanalsobevaluableintheidentificationof evaluationmodelcharacteristics.Theingredientsateachhierarchicalleveldescribedin References4and5are,inorderfromtoptobottom:

1.System--Theentiresystemthatmustbeanalyzedfortheproposedapplication.

2.Sub-systems--Majorcomponentsthatmustbeconsideredintheanalysis.Forsomeapplications,thesemayincludetheprimarysystem,secondarysystem,and containment.Forotherapplicationsonlytheprimarysystemwouldneedtobe considered.

3.Modules--Physicalcomponentswithinthesub-system,i.e.,reactorvessel,steamgenerator,pressurizer,pipingrun,etc.

4.Constituents--Chemicalformofsubstance,e.g.,water,nitrogen,air,boron,etc.

5.Phases--Solid,liquid,orvapor.

6.GeometricalConfigurations--Thegeometricalshapethatisdefinedforatransferprocess,e.g.,pool,drop,bubble,film,etc.

7.Fields--Thepropertiesthatarebeingtransported(mass,momentum,energy).8.Processes--Mechanismsthatmovepropertiesthroughthesystem.Ingredientsateachhierarchicallevelcanbedecomposedintotheingredientsatthenextleveldown.InReferences4and5,thisprocessisdescribedinthefollowing way:1.Eachsystemcanbedividedintointeractingsubsystems.2.Eachsubsystemcanbedividedintointeractingmodules.

3.Eachmodulecanbedividedintointeractingconstituents.

4.Eachconstituentcanbedividedintointeractingphases.

5.Eachphasecanbecharacterizedbyoneormoregeometricalconfigurations.6.Eachgeometricalconfigurationcanbedescribedbythreefieldequations,thatis,byconservationequationsformass,energy,andmomentum.7.Eachfieldcanbecharacterizedbyseveralprocesses.Bycarefullydefiningthenumberandtypeofeachingredientateachlevel,theevaluationmodeldevelopershouldbeabletoestablishthebasiccharacteristicsofthe evaluationmodel.Animportantprincipletonoteisthatifadeficiencyexistsatahigher level,itisusuallynotpossibletoresolveitbyfixingingredientsatlowerlevels.For relativelysimpletransients,thedecompositionprocessshouldalsobesimple.

101.1.4Step4.IdentifyandRankKeyPhenomenaandProcessesProcessidentificationisthelaststepinthedecompositiondescribedaboveandprovidesthelogicalbeginningtothisstep.Plantbehaviorisnotequallyinfluencedbyall processesandphenomenathatoccurduringatransient.Anoptimumanalysisreducescandidatephenomenatoamanageablesetbyidentifyingandrankingthephenomenawithrespecttotheirinfluenceonfiguresofmerit.Eachphaseofthetransientscenarioandsystemcomponentsareseparatelyinvestigated.Theprocessesandphenomena associatedwitheachcomponentareexamined.Causeandeffectaredifferentiated.After theprocessesandphenomenahavebeenidentified,theirimportanceshouldbe determinedwithrespecttotheireffectontherelevantfiguresofmerit.Theimportancedeterminationshouldalsobeappliedtohigh-levelsystemprocesses,whichmaybemissedifthefocusissolelyoncomponents.High-levelsystem processes,suchasdepressurizationandinventoryreduction,areoftenveryclosely relatedtofiguresofmerit.Focusonsuchprocessescanalsohelptoidentifythe importanceofindividualcomponentbehavior.AsnotedinStep2,itmaybepossibletoshowthatafigureofmeritotherthantheapplicableChapter15acceptancecriterionismoreappropriateasastandardfor identifyingandrankingphenomena.Thisisacceptableaslongasitcanbeshownthat, forallthescenariosbeingconsideredforthespecificrankingandidentificationactivity,the alternativefigureofmeritisconsistentwithplantsafety.Theprincipalproductoftheprocessoutlinedaboveisaphenomenaidentificationandrankingtable(PIRT)(seeReferences3,6,9,and12).Evaluationmodel developmentandassessmentshouldbebasedonacredibleandscrutablePIRT.The PIRTshouldbeusedtodeterminetherequirementsforphysicalmodeldevelopment, scalability,validation,andsensitivitiesstudies.Ultimately,thePIRTisusedtoguideany uncertaintyanalysisorintheassessmentofoverallevaluationmodeladequacy.The PIRTisnotanendinitself,butisratheratooltoprovideguidanceforthesubsequent steps.Theprocessesandphenomenathatevaluationmodelsshouldsimulatearefoundbyexaminingexperimentaldata,experienceandcodesimulationsrelatedtothespecific scenario.Independenttechniquestoaccomplishtherankingincludeexpertopinion, selectedcalculations,anddecisionmakingmethods(suchastheAnalyticalHierarchical Process(AHP)).ExamplesofthefirsttwoarefoundinReference12,andanexampleof thelastisfoundinReference13.Comparisonoftheresultsofthesetechniquesprovides assuranceoftheaccuracyandsufficiencyoftheprocess.TheinitialphasesofthePIRTprocessdescribedinthisstepcanrelyheavilyonexpertopinion,whichcanbesubjective.Therefore,iterationofthePIRTbasedon experimentationandanalysisisimportant.Althoughtheexperienceislimited, developmentofotherlesssubjectiveimportancedeterminationmethodsisencouraged.SensitivitystudiescanhelpdeterminetherelativeinfluenceofphenomenaidentifiedearlyinthePIRTdevelopmentandforfinalvalidationofthePIRTastheEMDAP isiterated.Examplesofsensitivitystudiesusedforthispurposeareprovidedin References3,6,9,11,and12.Theidentificationofprocessesandphenomenaproceedsasfollows:

111.Thescenarioisdividedintooperationallycharacteristictimeperiodsinwhichthedominantprocessesandphenomenaremainessentiallyconstant.2.Foreachtimeperiod,processesandphenomenaareidentifiedforeachcomponentfollowingaclosedcircuitthroughoutthesystem.Thisisdonetodifferentiatecause fromeffect.3.Startingwiththefirsttimeperiod,theactivitiescontinue,componentbycomponent,untilallpotentiallysignificantprocesseshavebeenidentified.4.Theprocedureisrepeatedsequentially,fromtimeperiodtotimeperiod,untiltheendofthescenario.Whentheidentificationhasbeencompleted,therankingprocessbegins.Thereasontonumericallyranktheprocessesandphenomenaisbasedontheneedtoprovide asystematicandconsistentapproachtoallthesubsequentEMDAPactivities.SufficientdocumentationshouldaccompanythePIRTtoadequatelyguidetheentireEMDAP.Developmentandassessmentactivitiesmayberevisitedduringthe process,includingtheidentificationandranking.Intheend,however,theevaluation model,thePIRT,andalldocumentationshouldbe"frozen"toprovidethebasisfora properreview.Withwelldefinedrankingofimportantprocesses,evaluationmodel capabilities,andcalculatedresults,theprioritizationoffurthermodelingimprovementscan bemademoreeasily.Animportantprincipleistherecognitionthatthemorehighly rankedphenomenaandprocessesrequiremodelingwithgreaterfidelity.References6 and7describetheroleofthePIRTprocessinexperiments,codedevelopment,andcode applicationsassociatedwithreactorsafetyanalysis.1.2Element2-DevelopAssessmentBaseThesecondcomponentofISTIR(Refs.4and5)isascalingmethodologythatincludesacquiringappropriateexperimentaldatarelevanttothescenariobeing consideredandassuringthattheexperimentalscalingissuitable.InReferences4and5, therelationshipofthesevereaccidentscalingmethodology(SASM)componenttocode developmentisshownbutnotemphasizedintheSASMdemonstration.FortheEMDAP, thepurposeistoprovidethebasisfordevelopmentandassessmentasshownpreviously inFigure1.Figure3showsthestepsinthiselementandtheirrelationship.Itshouldbe notedthatforsimpletransientsortransientswherethescalingissuesandassessmentare wellcharacterized,theimplementationofthiselementshouldalsobesimple.The numberingofstepsinthisandsubsequentelementscontinuesfromeachprevious element.

121.2.1Step5.SpecifyObjectivesforAssessmentBaseForanalysisofChapter15events,theprincipleneedforadatabaseistoassesstheevaluationmodeland,ifneeded,todevelopcorrelations.Theselectionofthedata baseisadirectresultoftherequirementsestablishedinElement1.Thedatabaseshould

include:1.Separateeffectsexperimentsneededtodevelopandassessempiricalcorrelationsandotherclosuremodels,2.Integralsystemsteststoassesssysteminteractionsandglobalcodecapability,3.Benchmarkswithothercodes(optional),

4.Planttransientdata(ifavailable),and 5.Simpletestproblemstoillustratefundamentalcalculationaldevicecapability.

13Itshouldbenotedthatitems3and5intheabovelistarenotmeanttobesubstitutionsforobtainingappropriateexperimentaland/orplanttransientdatafor evaluationmodelassessment.1.2.2Step6.PerformScalingAnalysisandIdentifySimilarityCriteriaAllexperimentsarecompromiseswithfull-scaleplantsystems.Evennominallyfull-scaleexperimentsdonotincludecompletesimilitude.Scalinganalysesshouldbe conductedtoensurethatthedata,andthemodelsbasedonthedata,willbeapplicableto thefull-scaleanalysisoftheplanttransient.Scalingcompromisesthatareidentifiedhere shouldultimatelybeaddressedinthebiasanduncertaintyevaluationinElement4.

Scalinganalysesareemployedtodemonstratetherelevancyandsufficiencyofthe collectiveexperimentaldatabaseforrepresentingthebehaviorexpectedduringthe postulatedtransientandtoinvestigatethescalabilityoftheevaluationmodelandits componentcodesforrepresentingtheimportantphenomena.Thescopeofthese analysesismuchbroaderthanforthescalabilityevaluationsdescribedinElement4 relatingindividualmodelsandcorrelationsorscaling-relatedfindingsfromthecode assessments.Here,theneedistodemonstratethattheexperimentaldatabaseis sufficientlydiversethattheexpectedplant-specificresponseisboundedandthatthe evaluationmodelcalculationsarecomparabletothecorrespondingtestsinnon-dimensionalspace.Thisdemonstrationallowsextendingtheconclusionsrelatedtocode capabilities,drawnfromassessmentscomparingcalculatedandmeasuredtestdata (Element4),tothepredictionofplant-specifictransientbehavior.Thescalinganalysesemploybothtop-downandbottom-upapproaches.Thetop-downscalingapproachevaluatestheglobalsystembehaviorandsystemsinteractions fromintegraltestfacilitiesthatcanbeshowntorepresenttheplant-specificdesignunder consideration.Atop-downscalingmethodologyisdevelopedandappliedinwhich:1.Thenon-dimensionalgroupsgoverningsimilitudebetweenfacilitiesarederived, 2.Thesegroupsareshowntoscaletheresultsamongtheexperimentalfacilities,and 3.Itisdeterminedwhethertherangesofthegroupvaluesprovidedbytheexperimentsetencompassthecorrespondingplant-andtransient-specificvalues.Thebottom-upscalinganalysesaddressissuesraisedintheplant-andtransient-specificPIRTrelatedtolocalizedbehavior.Theseanalysesareusedtoexplain differencesamongtestsindifferentexperimentalfacilitiesandtousetheseexplanations toinfertheexpectedplantbehavioranddeterminewhethertheexperimentsprovideadequateplant-specificrepresentation.ApplicationofthisscalingprocessisdescribedinSection5.3ofReference6.Inmostapplications,especiallythosewithalargenumberofprocessesandparameters,itisdifficult,ifnotimpossible,todesigntestfacilitiesthatpreservetotal similitudebetweentheexperimentandtheNPP.Therefore,basedontheimportant phenomenaandprocessesidentifiedinStep4andthescalinganalysisdescribedabove, theoptimumsimilaritycriteriashouldbeidentified,andtheassociatedscalingrationales developedforselectingexistingdataordesigningandoperatingexperimentalfacilities.1.2.3Step7.IdentifyExistingDataorPerformIETsandSETsToCompleteData Base 14Basedontheresultsofthepreviousstepsinthiselement,itshouldbepossibletocompletethedatabasebyselectionandexperimentation.Tocompletetheassessment matrix,thePIRTdevelopedinStep4isusedtoselectexperimentsanddatathatbest addresstheimportantphenomenaandcomponents.Inselectingexperiments,arangeof testsshouldbeemployedtodemonstratethatthecalculationaldeviceor phenomenologicalmodelhasnotbeentunedtoasingletest.Acorrelationderivedfroma particulardatasetmaybeidentifiedforinclusionintheevaluationmodel.Insuchcases, aneffortshouldbemadetoobtainadditionaldatasetsthatmaybeusedtoassessthe correlation.Forintegralbehaviorassessment,counterparttests(similarscenariosand transientconditions)indifferentexperimentalfacilitiesatdifferentscalesshouldbe selected.Assessmentsusingsuchtestsleadtoinformationconcerningscaleeffectson themodelsusedforaparticularcalculationaldevice.1.2.4Step8.EvaluateEffectsofIETDistortionandSETScaleupCapability8A-IETDistortions.Distortionsintheintegralexperimentaldatabasemayarisefromscalingcompromises(missingoratypicalphenomena)insub-scalefacilitiesor atypicalinitialandboundaryconditionsinallfacilities.Theeffectsofthedistortionsshould beevaluatedinthecontextoftheexperimentalobjectivesdeterminedinStep5.Ifthe effectsareimportant,areturntoStep7isprobablyneeded.8B-SETScaleup.AsnotedinStep7,correlationsshouldbebasedonSETsatvariousscales.Inthecaseofpoorscaleupcapability,itmaybenecessarytoreturnto Step6.AppendixCofReference3describesrationaleandtechniquesassociatedwith evaluationofscaleupcapabilitiesofcomputercodesandtheirsupportingexperimental databases.1.2.7Step9.DetermineExperimentalUncertaintiesasAppropriateItisimportanttoknowtheuncertaintiesinthedatabase.Theseuncertaintiesarisefromsuchitemsasmeasurementerrorsandexperimentaldistortions.Ifthequantified experimentaluncertaintiesaretoolargecomparedtotherequirementsforevaluation modelassessment,theparticulardatasetorcorrelationshouldberejected.1.3Element3-DevelopEvaluationModelAsdiscussedearlier,anevaluationmodelisacollectionofcalculationaldevices(codesandprocedures)developedandorganizedtomeettherequirementsestablishedin Element1.ThestepsfordevelopingthedesiredevaluationmodelareshowninFigure4.1.3.1Step10.EstablishanEvaluationModelDevelopmentPlanBasedontherequirementsestablishedinElement1,adevelopmentplanshouldbedevisedthatincludesdevelopmentstandardsandproceduresthatwillapplyduringthe developmentactivity.Specificareasoffocusshouldinclude:1.Calculationaldevicedesignspecifications,2.Documentationrequirements(seeRegulatoryPosition3ofthisguide),

3.Programmingstandardsandprocedures, 154.Transpo rtability requirem ents,5.Quality a ssuranc eprocedures(see Regulat ory Position 2ofthis guide),

and6.Configurationcontrolprocedures1.3.2Step11.EstablishEvaluationModelStructureTheevaluationmodelstructureincludesthestructureoftheindividualcomponentcalculationaldevicesandthestructurethatcombinesthedevicesintothetotalevaluation model.ThisstructureisbasedontheprinciplesofElement1,especiallyStep3.Thestructureforanindividualdeviceorcodeconsistsof:

1.Systemsandcomponents--Astructureshouldbepresentthatcananalyzethebehaviorofallthesystemsandcomponentsthatplayaroleinthetargeted application.

2.Constituentsandphases--Thecodestructureshouldbeabletoanalyzethebehaviorofallconstituentsandphasesrelevanttothetargetedapplication.

3.Fieldequations--Fieldequationsareequationsthataresolvedtodeterminethetransportofthequantityofinterest(usuallymass,energy,andmomentum).

4.Closurerelations--Closurerelationsarecorrelationsandequationsthatprovidecodecapabilitytomodelandscaleparticularprocesses;theyareneededtomodel thetermsinthefieldequations.

5.Numerics--Numericsprovidecodecapabilitytoperformefficientandreliablecalculations.

6.Additionalfeatures--Theseaddresscodecapabilitytomodelboundaryconditionsandcontrolsystems.

16Ofcourse,thecodestructureshouldbebasedontherequirementsestablishedinElement1andStep10.Becauseoftheimportanceofselectingproperclosure relationshipsforthegoverningequations,thesemodelsaretreatedseparatelyinStep12.

Thesixingredientsdescribedaboveshouldbesuccessfullyintegratedandoptimizedifa completedcodeistomeetitsobjectivesdeterminedinStep10.Therearespecialconcernsrelatedtotheintegrationofthecomponentcalculationaldevicesintoacompleteevaluationmodel.Thisisfrequentlyreferredtoastheevaluation modelmethodology.Thewayinwhichthedevicesareconnectedspatiallyandtemporally shouldbedescribed.Howclosethecouplingneedstobewould,inpart,bedetermined bytheresultsoftheanalysisdoneinStep3,butitisdeterminedbythemagnitudeand directionoftransferprocessesbetweendevices.Thehierarchicaldecomposition describedinReferences4and5wouldapplytohowtransferprocessesareanalyzed betweendevices.Sincemostdevicesincludeuseroptions,allselectionsmadeshouldbe justifiedasappropriatefortheevaluationmodel.1.3.3Step12.DeveloporIncorporateClosureModelsModelsorclosurerelationsthatdescribeaspecificprocessaredevelopedusingSETdata.Thisincludesmodelsthatcanbeusedinastandalonemodeorcorrelations thatcanbeincorporatedinacalculationaldevice(usuallyacomputercode).Onrare occasions,sufficientexperimentaldetailmaybeavailabletodevelopcorrelationsfromIET experiments.Thescalabilityandrangeofapplicabilityofacorrelationmaynotbeknown apriorithefirsttimeitisdevelopedorselectedforuseinthisstep.Aniterationofscaleup evaluation(Step8)andadequacyassessment(Element4)maybeneededtoensure correlationapplicability.ItshouldbenotedthatapathisshownfromElement2tothis step,sincecorrelationsmaybeselectedfromtheexistingdatabaseliterature.Modelsdevelopedherearekeytosuccessfulevaluationmodeldevelopment.Thebasis,rangeofapplicability,andaccuracyofincorporatedphenomenologicalmodels shouldbeknownandtraceable.Justificationshouldbeprovidedforextensionofany modelsbeyondtheiroriginalbasis.1.4Element4-AssessEvaluationModelAdequacyEvaluationmodeladequacycanbeassessedafterthepreviouselementshavebeenestablishedandtheevaluationmodelcapabilityhasbeendocumented.Figure5isa diagramofElement4.TheevaluationmodelassessmentisdividedintotwopartsasshowninFigure5.Thefirstpart(Steps13through15)pertainstothebottom-upevaluationoftheclosure relationsforeachcode.Thesecondpart(Steps16through19)pertainstothetop-down evaluationsofcode-governingequations,numerics,theintegratedperformanceofeach code,andtheintegratedperformanceofthetotalevaluationmodel.Inthefirstpart,importantclosuremodelsandcorrelationsareexaminedbyconsideringtheirpedigree,applicability,fidelitytoappropriatefundamentalorseparate effectstestdata,andscalability.Thetermbottom-upisusedbecausethereviewfocuses onthefundamentalbuildingblocksofthecode.

17Itisimport antto notethatanychangestoanevaluationmodelshouldincludeatleastapartialassessment toassurethatthesechangesdonotproduceunintendedresultsinthecodepredictive capability.1.4.1Step13.DetermineModelPedigreeandApplicabilityToSimulatePhysical ProcessesThepedigreeevaluationisrelatedtothephysicalbasisofaclosuremodel,assumptionsandlimitationsattributedtothemodel,anddetailsoftheadequacy characterizationatthetimethemodelwasdeveloped.Theapplicabilityevaluationis relatedtowhetherthemodel,asimplementedinthecode,isconsistentwithitspedigree orwhetheruseoverabroaderrangeofconditionsisjustified.1.4.2Step14.PrepareInputandPerformCalculationsToAssessModelFidelityorAccuracy 18Thefidelityevaluationisrelatedtotheexistenceandcompletenessofvalidationefforts(throughcomparisontodata)orbenchmarkingefforts(throughcomparisontoother standards,forexample,aclosedformsolutionorresultsobtainedwithanothercode)or somecombinationofthesecomparisons.SETinputforcomponentdevicesusedinmodelassessment(usuallycomputercodes)shouldbepreparedtorepresentthephenomenaandtestfacilitybeingmodeled andthecharacteristicsofthenuclearpowerplantdesign.Inparticular,nodalizationand optionselectionshouldbeconsistentbetweentheexperimentalfacilityandsimilar componentsinthenuclearpowerplant.WhenthecalculationsoftheSETsare completed,thedifferencesbetweencalculatedresultsandexperimentaldataforimportant phenomenashouldbequantifiedforbiasanddeviation.1.4.3Step15.AssessScalabilityofModelsThescalabilityevaluationhereislimitedtowhetherthespecificmodelorcorrelationisappropriateforapplyingtotheconfigurationandconditionsoftheplantandtransient underevaluation.References5and14-17documentrecentapproachestoscaling, rangingfromtheoreticalmethodstospecificapplicationsthatareofparticularinterest here.Inthesecondpartoftheassessment,theevaluationmodelisevaluatedbyexaminingthefieldequations,numerics,applicability,fidelitytocomponentorintegral effectsdataandscalability.Thispartoftheassessmenteffortiscalledthetop-down reviewbecauseitfocusesoncapabilitiesandperformanceoftheevaluationmodel.1.4.4Step16.DetermineCapabilityofFieldEquationsToRepresentProcessesandPhenomenaandtheAbilityofNumericSolutionsToApproximate EquationSetThefieldequationevaluationconsiderstheacceptabilityoftheequations.Anassessmentofthegoverningequationsineachofthecomponentcodesshouldconsider theirpedigreeandthekeyconceptsandprocessesculminatingintheequationsetsolved bythecode.Theobjectiveofthisassessmentistocharacterizetherelevanceofthe governingequationsforthechosenapplication.Thenumericsolutionevaluationconsidersconvergence,propertyconservation,andstabilityofcodecalculationstoasolutionoftheoriginalequationswhenappliedtothe targetapplication.Theobjectiveofthisreviewistosummarizeinformationregardingthe domainofapplicabilityofthenumericaltechniquesanduseroptionsthatmayimpactthe accuracy,stability,andconvergencefeaturesofeachcomponentcode.Acompleteassessmentwithinthisstepcanonlybeperformedafterasufficientfoundationofassessmentanalysesiscomplete.Section3andAppendixAofReference 6provideanexampleforapplicationofthisstep.1.4.5Step17.DetermineApplicabilityofEvaluationModeltoSimulateSystemsandComponentsThisapplicabilityevaluationconsiderswhethertheintegratedcodeiscapableofmodelingtheplantsystemsandcomponents.Beforeintegratedanalysesareperformed, itshouldbedeterminedthatthevariousevaluationmodeloptions,specialmodels,and inputhavetheinherentcapabilitytomodelthemajorsystemsandsubsystemsrequired fortheparticularapplication.

191.4.6Step18.PrepareInputandPerformCalculationsToAssessSystemInteractionsandGlobalCapabilityThefidelityevaluationconsidersthecomparisonofevaluationmodel-calculatedandmeasuredtestdatafromcomponentandintegraltestdataand,wherepossible,plant transientdata.Forthesecalculations,theentireevaluationmodeloritsmajor componentsareusedtocompareagainsttheintegraldatabaseselectedinElement2.AswasdoneinStep14fortheSETassessments,theevaluationmodelinputforIETsshouldbestrepresentthefacilitiesandshouldrepresentthecharacteristicsofthe nuclearpowerplantdesign.Asbefore,nodalizationandoptionselectionshouldbe consistentbetweenexperimentandnuclearpowerplant.WhentheIETsimulationsare complete,thedifferencesbetweencalculatedresultsandexperimentaldataforimportant processesandphenomenashouldbequantifiedforbiasanddeviation.Theabilityofthe evaluationmodeltomodelsysteminteractionsshouldalsobeevaluatedinthisstep.

Section5ofReference6providesanexampleapplicationofthisstep.Inthisstep,plantinputdecksshouldalsobepreparedforthetargetapplications.Sufficientanalysesshouldbeperformedtodetermineparameterrangesexpectedinthe nuclearpowerplant.Theseinputdecksalsoprovidethegroundworkfortheanalyses performedinStep20.1.4.7Step19.AssessScalabilityofIntegralCalculationsandDataforDistortionsThescalabilityevaluationhereislimitedtowhethertheassessmentcalculationsandexperimentsexhibitotherwiseunexplainabledifferencesamongfacilities,orbetween thecalculatedandmeasureddataforthesamefacility,thatindicateexperimentalorcode scalingdistortions.1.4.8Step20.DetermineEvaluationModelBiasesandUncertaintiesTheanalysispurposeestablishedinStep1andthetransientcomplexitywilldeterminethesubstanceofthisstep.Forbest-estimateLOCAanalysis,uncertainty determinationdescriptionandguidanceareinReferences3and18andAppendixAof thisguide.Intheseexamples,theuncertaintyanalysesdiscussedhavetheultimate objectiveofprovidingasingularstatementofuncertaintywithrespecttothe10CFR50.46 acceptancecriteriawhenusingthebest-estimateoptioninthatrule.Thissingular uncertaintystatementisaccomplishedwhentheindividualuncertaintycontributionsare determined(seeRegulatoryGuide1.157,Ref.18).ForotherChapter15events,acompleteuncertaintyanalysisisnotrequired.However,inmostcasestheSRPguidanceistouse"suitablyconservative"input parameters.Thissuitabilitydeterminationmayinvolvealimitedassessmentofbiasesand uncertaintiesandiscloselyrelatedtotheanalysesperformedinStep16.Basedonthe resultsofStep4,individualdevicemodelscanbechosenfromthoseobtainedinStep9.

Theindividualuncertainty(intermsofrangeanddistribution)ofeachkeycontributoris determinedfromtheexperimentaldata(Step11),inputtothenuclearpowerplantmodel, andtheeffectonappropriatefiguresofmeritevaluatedbyperformingseparatenuclear powerplantcalculations.Thefiguresofmeritanddeviceschosenshouldbeconsistent.

Inmostcasestheanalysiswouldinvolvetheentireevaluationmodel.Thelastpartofthis stepistodeterminewhetherthedegreeofoverallconservatismoranalyticaluncertaintyis appropriatefortheentireevaluationmodel.Thisisdoneinthecontextoftheanalysis purpose(Step1)andtheregulatoryrequirements.

201.5AdequacyDecisionThedecisionontheadequacyoftheevaluationmodelistheculminationoftheEMDAPdescribedinRegulatoryPositions1.1through1.4.ThroughouttheEMDAP, questionsconcerningtheadequacyoftheevaluationmodelshouldbeasked.Attheend oftheprocess,theadequacyshouldbequestionedagaintoassurethatalltheanswers aresatisfactoryandthatinterveningactivitieshavenotinvalidatedpreviousacceptable responses.Ifunacceptableresponsesindicatesignificantevaluationmodelinadequacies, thecodedeficiencyiscorrectedandtheappropriatestepsintheEMDAParerepeatedto evaluatethedeficiencycorrection.Theprocesscontinuesuntiltheultimatequestion regardingadequacycanbeansweredpositively.Ofcourse,thedocumentationas describedinRegulatoryPosition3shouldbeupdatedascodeimprovementsandassessmentareaccomplishedduringtheprocess.Analysis,assessmentandanysensitivitystudiescanalsoleadtoare-assessmentofthephenomenaidentificationandranking.Therefore,thatdocumentationshouldalsoberevisedasappropriate.Itishelpfultodevelopalistofquestionstobeaskedduringtheprocessandagainattheend.Toanswerthesequestions,standardsshouldbeestablishedbywhichthe capabilitiesoftheevaluationmodelanditscompositecodesandmodelscanbejudged.

Section2.2.2ofReference6providesanexampleofthedevelopmentofsuchstandards.2.QUALITYASSURANCEMuchofwhatisdescribedthroughoutthisregulatoryguiderelatestogoodqualityassurancepractices.Forthatreasonitisimportanttoestablish,earlyinthedevelopment andassessmentprocess,appropriatequalityassuranceprotocol.Thedevelopment, assessment,andapplicationofanevaluationmodelareallactivitiesthatarerelatedtothe requirementsofAppendixBto10CFRPart50.SectionIIIofAppendixBisakey requirementforthisactivityandrequiresthatdesigncontrolmeasuresbeappliedto reactorphysics,thermal,hydraulic,andaccidentanalyses.SectionIIIstatesthat:Thedesigncontrolmeasuresshallprovideforverifyingorcheckingtheadequacyofdesign,suchasbytheperformanceofdesignreviews,bytheuseofalternateor simplifiedcalculationalmethods,orbytheperformanceofasuitabletesting program.SectionIIIalsostatesthatdesignchangesshouldbesubjecttoappropriatedesigncontrolmeasures.ItisimportanttonotethatotherpartsofAppendixBarealsorelevant,suchasSectionV(whichrequiresdocumentedinstructions,e.g.,userguidance);SectionXVI (correctiveactions,e.g.,errorcontrol,identification,andcorrection);andSectionVIand XVII,whichaddressdocumentcontrolandrecordsretention.TocapturethespiritandintentofAppendixB,independentpeerreviewshouldbeperformedatkeystepsintheprocess,suchasattheendofamajorpassthroughan element.Intheearlystagesofevaluationmodeldevelopment,itisrecommendedthatareviewteambeconvenedtoreviewevaluationmodelrequirementsasdevelopedin 21Element1.Peerreviewshouldalsobeemployedatthelaterstagesduringmajorinquiriesassociatedwiththeadequacydecision.Inadditiontoprogrammers,developers,andendusers,itisrecommendedthatthepeerreviewteamhaveindependentmemberswithrecognizedexpertiseinrelevant engineeringandsciencedisciplines,codenumerics,andcomputerprogramming.Expert peerreviewteammembers,whowerenotdirectlyinvolvedintheevaluationmodel developmentandassessment,canenhancetherobustnessoftheevaluationmodels.

Further,theycanbeofvalueinidentifyingdeficienciesthatarecommontolargesystem analysiscodes.Throughoutthedevelopmentprocess,configurationcontrolpracticesshouldbeadoptedthatprotectprogramintegrityandallowtraceabilityofthedevelopmentofboth thecodeversionandtheplantinputdeckusedtoinstructthecodeinhowtorepresentthe facilityornuclearpowerplant.Configurationcontrolofthecodeversionandtheplant inputdeckareseparatebutrelatedelementsoftheevaluationmodeldevelopmentand requirethesamedegreeofqualityassurance.Responsibilityforthesefunctionsshould beclearlyestablished.Attheendoftheprocess,onlytheapproved,identifiedcode versionandplantinputdeckshouldbeusedforlicensingcalculations.3.DOCUMENTATIONProperdocumentationallowsappraisaloftheevaluationmodelapplicationtothepostulatedscenario.Thedocumentationfortheevaluationmodelshouldcoverallthe elementsoftheEMDAPprocessandshouldincludethe:1.EvaluationModelrequirementsdocument2.EvaluationModelmethodologydocument 3.Codedescriptionmanuals 4.Usermanualsanduserguidelines 5.Scalingreports 6.Assessmentreports 7.Uncertaintyanalysisreports3.1RequirementsDocumentTherequirementsdeterminedinElement1shouldbedocumentedsotheevaluationmodelcanbeassessedagainstknownguidelines.Inparticular,a documented,currentPIRTisimportantindecidingwhetheraparticularevaluationmodel featureshouldbemodifiedbeforetheevaluationmodelcanbeappliedwithconfidence.3.2MethodologyDocumentMethodologydocumentationshouldincludetheinter-relationshipofallthecomputationaldevicesusedfortheplanttransientbeinganalyzed,includingthe descriptionofinputandoutput.Thisshouldalsoincludeacompletedescriptionand specificationofthoseportionsoftheevaluationmodelnotincludedinthecomputer programs.Adescriptionofallotherinformationnecessarytospecifythecalculational procedureshouldalsobeincluded.Averyusefulpartofthisdescriptionwouldbea diagramtoillustratehowthevariousprogramsandproceduresarerelated,bothintime 22andinfunction.Thismethodologydescriptionisneededtoknowexactlyhowthetransientwillbeanalyzedinitsentirety.3.3ComputationalDeviceDescriptionManualsAdescriptionmanualisneededforeachcomputationaldevicethatiscontainedintheevaluationmodel.Thereareseveralimportantcomponentstothemanual.Thefirstis adescriptionofthemodelingtheoryandassociatednumericalschemesandsolution models,includingadescriptionofthearchitecture,hydrodynamics,heatstructure,heat transfermodels,tripsystemsandcontrolsystems,reactorkineticsmodels,andfuel behaviormodels.Akeyingredientofthedocumentationisamodelsandcorrelationsqualityevaluation(MC/QE)report.TheMC/QEreportprovidesabasisforthetraceabilityofthe modelsanddetailedinformationontheclosurerelations.Informationoncorrelationand modelsources,databases,accuracy,scale-upcapability,andapplicabilitytospecific plantandtransientconditionsshouldalsobedocumentedintheMC/QEreport.TheMC/QEreportrepresentsaqualityevaluationdocumentthatprovidesablueprintastowhatisinthecomputationaldevice,howitgotthere,andwhereitcamefrom.TheMC/QEdocumenthasthreeobjectives:1.Toprovideinformationonthesourcesandqualityofclosureequations,thatis,oncorrelationsandmodelsorothercriteriaused.2.Todescribehowtheseclosurerelationsarecodedinthedeviceandtoassurethatthedescriptionsinthemanualconformtothecoding,andthecodingconformsto thesourcefromwhichtheclosurerelationswerederived.3.Toprovideatechnicalrationaleandjustificationforusingtheseclosurerelations;thatis,toconfirmthatthedominantparameters(pressure,temperature,etc.)

representedbythemodelsandcorrelationsreflecttherangesexpectedintheplant andtransientofinterest.Consequently,forcorrelations,models,andcriteriaused,theMC/QEshould:1.Provideinformationontheoriginalsource,thesupportingdatabase,theaccuracyandapplicabilitytotheplant-specifictransientconditions.2.Provideanassessmentofeffectsifusedoutsidethesupportingdatabase.Adescriptionofandjustificationfortheextrapolationmethodshouldbeprovided.For certainapplications,recommendationsmaybegiventouseoptionsotherthanthe defaultoptions.Insuchcases,instructionsshouldbeprovidedtoensurethat appropriatevalidationisperformedforthenonstandardoption.3.Describetheimplementationinthedevice(i.e.,actualcodingstructure).

4.Describeanymodificationsrequiredtoovercomecomputationaldifficulties.

5.Provideanassessmentofeffectscausedbyimplementation(item3)ormodifications(item4)ontheoverallcodeapplicabilityandaccuracy.

23References19and20areexamplesoftheMC/QEdocumentsgeneratedtomeettherequirementslistedabove.3.4UsersManualandUserGuidelinesTheusersmanualshouldbeacompletedescriptionofhowtoprepareallrequiredandoptionalinput.Theuserguidelinesshoulddescriberecommendedpracticesfor preparationofallrelevantinput.Tominimizetheriskofinappropriateprogramuse,the guidelinesshouldinclude:1.Theproperuseoftheprogramfortheparticularplant-specifictransientoraccidentbeingconsidered,2.Therangeofapplicabilityforthetransientoraccidentbeinganalyzed,3.Thecodelimitationsforsuchtransientsandaccidents,4.Recommendedmodelingoptionsforthetransientbeingconsidered,theequipmentrequired,andthechoiceofnodalizationschemes.Plantnodalizationshouldbe consistentwithnodalizationusedinassessmentcases.3.5ScalingReportsReportsshouldbeprovidedforallscalinganalysesusedtosupporttheviabilityoftheexperimentaldatabase,thescalabilityofmodelsandcorrelations,andthescalability ofthecompleteevaluationmodel.Section5.3ofReference6providesanexampleand referencestoscalinganalysesdonetosupportadequacyevaluations.3.6AssessmentReportsAssessmentReportsaregenerallyofthreetypes:1.Developmentalassessment 2.Componentassessment 3.IntegraleffectstestssessmentMostdevelopmentalassessment(DA)reportsshouldbeasetofcodeanalysesthatfocusonalimitedsetofrankedphenomena.Thatis,thecodeorotherdeviceshould analyzeexperimentsorplantdatathatdemonstrateinaseparateeffectsmannerthe capabilitytocalculateindividualphenomenaandprocessesdeterminedtobeimportantby thePIRTforthespecificscenarioandplanttype.Acodeorotherdevicemaymodelcertainequipmentinaspecialway;assessmentcalculationsshouldbeperformedforthesecomponents.Integraleffectstests(IET)shouldshowtheevaluationmodel'sintegralcapabilitybycomparisontorelevantintegraleffectsexperimentsorplantdata.SomeIETassessments maybegeneralinnature,butforevaluationmodelconsideration,theIETassessments shouldincludeavarietyofscaledfacilitiesapplicabletotheplantdesignandtransient.

24Forsomeplantsandtransients,code-to-codecomparisonscanbeveryhelpful.Inparticular,ifanewcodeordeviceisintendedtohavealimitedapplication,theresultsmay becomparedtocalculationsusingapreviouscode.However,thepreviouscodeshould bewellassessedtointegralorplantdatafortheplanttypeandtransientbeingconsidered forthenewdevice.Differencesinkeyinputsuchassystemnodalizationshouldbe explainedsothatfavorablecomparisonsareprovidingtherightanswersfortheright reasons.Suchbenchmarkcalculationswouldnotbeareplacementforassessmentofthe newcode.Asignificantamountofevaluationmodelassessmentmaybeperformedbeforeselectionoftheplant-specifictransienttobeanalyzed.Inothercases,theassessment maybedoneoutsidethecontextoftheplant-andtransient-specificevaluationmodel.In stillothercases,theassessmentmaybedonebyorganizationsotherthanthose responsiblefortheplant-specificanalysis.Ifitisdesiredtocredittheseassessmentsto theplantandtransientunderconsideration,greatcareshouldbetakeninevaluatingthe applicabilityofthoseassessments.Theapplicabilitytothepresentcaseshouldbe thoroughlyevaluatedanddocumented.Togainconfidenceinevaluationmodelpredictivecapabilitywhenappliedtoaplant-specificevent,itisimportantforassessmentreportsto:1.Assesscalculationaldevicecapabilityandquantifyaccuracytocalculatevariousparametersofinterest,inparticularthosedescribedinthePIRT.2.Determinewhetherornotthecalculatedresultsareduetocompensatingerrorsbyperforminganappropriatescalinganalysisandsensitivityanalysis.3.Assesswhetherornotthecalculatedresultsareself-consistentandpresentacohesivesetofinformationthatistechnicallyrationalandacceptable.4.Assesswhetherthetimingofeventscalculatedbytheevaluationmodelareinagreementwiththeexperimentaldata.5.Assesstheevaluationmodelcapabilitytoscaletotheprototypicalnuclearplant.Almostwithoutexception,suchassessmentalsoaddressestheexperimentaldata baseusedindevelopmentorvalidationoftheevaluationmodel.6.Explainanyunexpectedor,atfirstglance,strangeresultscalculatedbytheevaluationmodelorcomponentdevices.Thisisparticularlyimportantwhen experimentalmeasurementsarenotavailabletogivecredencetothecalculated results.Insuchcases,rationaltechnicalexplanationswillgreatlysupport generationofcredibilityandconfidenceintheevaluationmodel.Wheneverthereisadisagreementbetweencalculatedresultsandexperimentaldata,assessmentreportsmust:7.Identifyandexplainthecauseforthediscrepancy,thatis,identifyanddiscussthedeficiencyinthedevice(or,ifnecessary,discusstheinaccuracyofexperimental measurements).8.Addressthequestionofhowimportantthedeficiencyistotheoverallresults,thatis,toparametersandissuesofinterest.

259.Explainwhyadeficiencymaynothaveanimportanteffectonaparticularscenario.Withrespecttoacalculationaldeviceinputmodelandsensitivitystudies,itisnecessaryforassessmentreportsto:10.Provideanodalizationdiagramalongwithadiscussionofthenodalizationrationale.11.Specifyanddiscusstheboundaryandinitialconditions,aswellastheoperationalconditionsforthecalculations.12.Presentanddiscussresultsofsensitivitystudies(ifperformed)onclosurerelationsorotherparameters.13.Discussmodificationstotheinputmodel(nodalization,boundary,initialoroperationalconditions)resultingfromsensitivitystudies(ifperformed).14.Provideguidelinesforperformingsimilaranalyses.3.7UncertaintyAnalysisReportsDocumentationshouldbeprovidedforanyuncertaintyanalysesperformedaspartofStep20oftheEMDAP.4.GENERALPURPOSECOMPUTERPROGRAMSVeryoftenageneralpurposetransientanalysiscomputerprogram,suchasRELAP5,TRAC,orRETRAN,isdevelopedtoanalyzeanumberofdifferenteventsfora widevarietyofplants.Thesecodescanconstitutethemajorportionofanevaluation modelforaparticularplantandevent.Genericreviewsareoftenperformedforthese codestominimizetheamountofworkrequiredforplant-andevent-specificreviews.A certainamountofgenericassessmentmaybeperformedforsuchacodeaspartofthe genericcodedevelopment.TheEMDAP,ontheotherhand,startswithidentificationof plant,event,anddirectlyrelatedphenomena.Thisprocess,aspreviouslydescribed,may indicatethatagenericassessmentdoesnotincludealltheappropriategeometry, phenomena,orthenecessaryrangeofvariablestodemonstratecodeadequacyforsome oftheproposedplant-specificeventanalyses.Evidenceofthisisthefactthatsafety evaluationsforgenericcodereviewsoftencontainalargenumberofqualificationsonthe useofthecode.Toavoidsuchproblems,itisimportanttoqualifytheapplicabilityofthe genericcode,includingitsmodelsandcorrelations,andtheapplicabilityofany"generic" assessmentthataccompaniesthecode.D.IMPLEMENTATIONThepurposeofthissectionistoprovideinformationtoapplicantsandlicenseesregardingtheNRCstaff'splansforusingthisdraftregulatoryguide.

26Thisdraftguidehasbeenreleasedtoencouragepublicparticipationinitsdevelopment.Exceptinthosecasesinwhichanapplicantorlicenseeproposesan acceptablealternativemethodforcomplyingwiththespecifiedportionsoftheNRC's regulations,themethodstobedescribedintheeffectiveguidereflectingpubliccomments willbeusedintheevaluationofsubmittalsinconnectionwithevaluationmodelsusedto analyzetransientsandaccidents.

27DEFINITIONSThesedefinitionsareinthecontextofthisregulatoryguideandmaynotapplytootheruses.AHPAnalyticalHierarchicalProcess--Ananalyticalandsoftwarebasedmethodologyusedtocombineexperimentaldatawithexpert judgmenttoefficientlyranktherelativeimportanceofphenomenaand processestotheresponseofanNPPtoanaccidentorothertransient inaconsistentandtraceablemanner.AP600AdvancedPassive600MwePWRdesignedbyWestinghouseElectricCo.Bottom-upTheapproachtoasafety-relatedanalysissimilartotop-down(seebelow),butinwhichthekeyfeatureistotreatallphenomenaand processes,includingallthoseassociatedwiththeanalysistoolsfor modeling,asequallyimportanttothefacility'sresponsetoan accidentortransient.Therefore,thephenomenaandprocessesare quantifiedindepth.CalculationalComputercodesorothercalculationalproceduresthatcomposeandevicesevaluationmodel.Chapter15Inthisregulatoryguide,Chapter15eventsrefertothetransientsandeventsaccidentsthataredefinedinChapter15oftheSRP,NUREG-0800(Ref.1)tobeanalyzedtomeettherequirementsoftheGeneral DesignCriteria(GDC)ofAppendixAto10CFRPart50,exceptfor thefuelassemblymisloadingeventandallradiologicalconsequence analyses.CFRCodeofFederalRegulations ClosurerelationsEquationsandcorrelationsrequiredtosupplementthefieldequationsthataresolvedtoobtaintherequiredresults.Thisincludesphysical propertydefinitionsandcorrelationsoftransportphenomena.ConstituentsChemicalformofanymaterialbeingtransported,e.g.,water,air,boron.CSAUCodescaling,applicability,anduncertainty--Aprocesstodeterminetheapplicability,scalability,anduncertaintyofacomputercodein simulatinganaccidentorothertransient.APIRTprocessisnormally embeddedwithinaCSAUprocess.SeeReference3.DADevelopmentalAssessment--Calculationsperformedusingtheentireevaluationmodeloritsindividualcalculationaldevicesto validateitscapabilityforthetargetapplication.DNBRDeparturefromnucleateboilingratioEMDAPEvaluationmodeldevelopmentandassessmentprocess 28ECCSEmergencycorecoolingsystemEvaluationmodelCalculationalframeworkforevaluatingthebehaviorofthereactor(EM)systemduringapostulatedChapter15event,whichincludesoneormorecomputerprogramsandallotherinformationneededforusein thetargetapplication.FieldsThepropertiesthatarebeingtransported(mass,momentum,energy).FieldequationsEquationsthataresolvedtodeterminethetransportofmass,energy,andmomentumthroughoutthesystem.FrozenTheconditionwherebytheanalyticaltoolsandassociatedfacilityinputdecksremainunchanged(andunderconfigurationcontrol) throughoutasafetyanalysis,therebyensuringtraceabilityofand consistencyinthefinalresults.GDCGeneralDesignCriteria--DesigncriteriadescribedinAppendixAto10CFRPart50.GeometricalThegeometricalshapethatisdefinedforatransferprocess,e.g.,configurationspool,drop,bubble,film.H2TSHierarchicaltwo-tieredscaling--Methodologythatuseshierarchicalsystemsanalysismethodstoevaluateexperimentalscaling.

DescribedinReferences4and5.IETIntegralEffectsTest--Anexperimentinwhichtheprimaryfocusisontheglobalsystembehaviorandtheinteractionsbetweenparameters andprocesses.ISTIRIntegratedStructureforTechnicalIssueResolution--Methodologyderivedforsevereaccidentissueresolution.Describedin References4and5.LBLOCALarge-breakloss-of-coolantaccident LOCALoss-of-coolantaccident LWRLightwaterreactorMC/QEModelsandcorrelationsqualityevaluation--Areportdocumentingwhatisinacomputercode,thesourcesusedtodevelopthecode, andtheconditionsunderwhichtheoriginalsourceofinformationwas developed.Model(Without"evaluation"modifier)--Equationorsetofequationsthatrepresentsaparticularphysicalphenomenonwithinacalculational device.

29ModulesPhysicalcomponentswithinthesub-system,e.g.,reactorvessel,steamgenerator,pressurizer,pipingrun.

.

MYISAMaineYankeeIndependentSafetyAssessmentNPPNuclearpowerplant PCTPeakcladdingtemperature PhaseStateofmatterinvolvedinthetransportprocess,usuallyliquidorgas.Anotableexceptionisheatconductionthroughsolids.PIRTPhenomenaIdentificationandRankingTable--Mayrefertoatableortoaprocess,dependingonthecontextofuse.Theprocess relatestodeterminingtherelativeimportanceofphenomena(or physicalprocesses)tothebehaviorofanNPPfollowingtheinitiation ofanaccidentorothertransient.APIRTtableisalistingofthe resultsofapplicationoftheprocess.ProcessesMechanismsthatmovepropertiesthroughthesystem.

QAQualityAssurance SASMSevereaccidentscalingmethodology SBLOCASmall-breakloss-of-coolantaccident ScalabilityTheprocessinwhichtheresultsfromasubscalefacility(relativetoa(scaling)NPP)orthemodelingfeaturesofacalculationaldeviceareevaluatedtodeterminethedegreetowhichtheyrepresentaNPP.ScenarioDescriptionandtimesequenceofevents SensitivitystudiesThetermisgenerictoseveraltypesofanalyses;however,thedefinitionofmostinteresthererelatestothosestudiesassociated withthePIRTprocessandusedtodeterminetherelativeimportance ofphenomenaorprocesses.Thismayalsoinvolveanalysisof experimentaldatathatareasourceofinformationusedinthePIRT process.SETSeparateEffectsTest--Anexperimentinwhichtheprimaryfocusisonasinglephysicalphenomenaorprocess.SRPStandardReviewPlan--AcceptableplanforNRCreviewers,NUREG-0800,"StandardReviewPlanfortheReviewofSafety AnalysisReportsforNuclearPowerPlants."SystemTheentiresystemthatmustbeanalyzedfortheproposedapplication.

SystemscodeTheprincipalcomputercodeofanevaluationmodelthatdescribesthetransportofmass,momentum,andenergythroughoutthereactor coolantsystems.

30Sub-systemsThemajorcomponentsthatmustbeconsideredintheanalysis.Forsomeapplicationsthiswouldincludetheprimarysystem,secondary system,andcontainment.Forotherapplicationsonlytheprimary systemwouldneedtobeconsidered.TargetapplicationThesafetyanalysisforwhichaspecificpurpose,transienttype,andNPPtypehasbeenspecified.Top-downTheapproachtoasafety-relatedanalysisinwhichonesequentiallydeterminesorperforms(1)theexactobjectiveoftheanalysis (regulatoryaction,licensingaction,desiredproduct,etc.),(2)the analysisenvelope(facilityorNPP,transients,analysiscodes,facility-imposedgeometricandoperationalboundaryconditions,etc.),(3)all plausiblephenomenaorprocessesthathavesomeinfluenceonthe facilityorplantbehavior,(4)aPIRTprocess,(5)applicabilityand scalabilityoftheanalysistools,and(6)theinfluenceofvarious uncertaintiesembeddedintheanalysisontheendproduct.Akey featureofthetop-downapproachistoaddressthosepartsofthe safetyanalysisassociatedwithitems5and6inagraduatedmanner basedontherelativeimportancedeterminedinitem4.Theapproach items1through5areindependentofanalysistools.Items5and6 requiretheapproachtobecomedependentonanalysistools.UncertaintyTherearetwoseparatebutrelateddefinitionsofprimaryinterest:(1)Theinaccuracyinexperimentallyderiveddatatypicallygenerated bytheinaccuracyofmeasurementsystemsand(2)Theinaccuracyofcalculatingprimarysafetycriteriaorrelatedfiguresofmerittypicallyoriginatingintheexperimentaldataorassumptions usedtodeveloptheanalyticaltools.Theanalyticalinaccuraciesare relatedtoapproximationsanduncertaintiesinvolvedwithsolving theequationsandconstitutiverelations.

1ElectroniccopiesarepostedonNRC'swebsite,<

WWW.NRC.GOV>,throughRulemaking,andfromNRC'sDistributionSection,PublicDocumentRoom,andPublicElectronicReadingRoom.Seefootnotes below.2Singlecopiesofregulatoryguides,bothactiveanddraft,anddraftNUREGdocumentsmaybeobtainedfreeofchargebywritingtheReproductionandDistributionServicesSection,OCIO,USNRC,Washington,DC20555-0001, orbyfaxto(301)415-2289,orbyemailto<DISTRIBUTION@NRC.GOV>.Activeguidesmayalsobepurchased fromtheNationalTechnicalInformationServiceonastandingorderbasis.Detailsonthisservicemaybeobtained bywritingNTIS,5285PortRoyalRoad,Springfield,VA22161;telephone(703)487-4650;online

<http://www.ntis.gov/ordernow>.CopiesofcertainguidesandmanyotherNRCdocumentsareavailableelectronicallyontheinternetatNRC'shomepageat<

WWW.NRC.GOV>intheReferenceLibrary.DocumentsarealsoavailablethroughthePublicElectronicReadingRoom(NRC'sADAMSdocumentsystem,orPERR)atthe samewebsite.

3CopiesareavailableatcurrentratesfromtheU.S.GovernmentPrintingOffice,P.O.Box37082,Washington,DC20402-9328(telephone(202)512-1800);orfromtheNationalTechnicalInformationServicebywritingNTISat5285 PortRoyalRoad,Springfield,VA22161;(telephone(703)487-4650.Copiesareavailableforinspectionorcopying forafeefromtheNRCPublicDocumentRoomat11555RockvillePike,Rockville,MD;thePDR'smailingaddressis USNRCPDR,Washington,DC20555;telephone(301)415-4737or(800)397-4209;fax(301)415-3548;emailis

PDR@NRC.GOV.

4ElectroniccopiesareavailableinNRC'sPublicElectronicReadingRoom,whichcanbeaccessedthroughtheNRC'swebsite,<

WWW.NRC.GOV>.31 REFERENCES1.DraftSection15.0.2,"ReviewofAnalyticalComputerCodes,"December2000,ofNUREG-0800,"StandardReviewPlanfortheReviewofSafetyAnalysisReports forNuclearPowerPlants,"USNRC,updatedbysection.

12.RegulatoryGuide1.70,"StandardFormatandContentofSafetyAnalysisReportsforNuclearPowerPlants(LWREdition),Revision3,USNRC,November1978.

23.B.Boyacketal.,"QuantifyingReactorSafetyMargins,ApplicationofCode Scaling,Applicability,andUncertaintyEvaluationMethodologytoa LargeBreak,Loss-of-CoolantAccident,"NUREG/CR-5249,USNRC, December1989.

34.B.Boyacketal.,"AnIntegratedStructureandScalingMethodologyforSevereAccidentTechnicalIssueResolution,"DraftNUREG/CR-5809,USNRC,November 1991.25.N.Zuberetal.,"AnIntegratedStructureandScalingMethodologyforSevereAccidentTechnicalIssueResolution:DevelopmentofMethodology,"

NuclearEngineeringandDesign,186(1-21),1998.6.C.D.Fletcheretal.,"AdequacyEvaluationofRELAP5/MOD3,Version3.2.1.2forSimulatingAP600SmallBreakLoss-of-CoolantAccidents,"INEL-96/0400 (nonproprietaryversion),April1997.

4(AvailableinPERRbyAccessionNumberML003769921)7.G.E.WilsonandB.E.Boyack,"TheRoleofthePIRTProcessinExperiments,CodeDevelopmentandCodeApplicationsAssociatedwithReactorSafety Analysis,"NuclearEngineeringandDesign,186(23-37),1998.

328.H.Holmstrometal.,"StatusofCodeUncertaintyEvaluationMethodologies,"inProceedingsoftheInternationalConferenceonNewTrendsinNuclearSystemThermohydraulics,DipartimentodiCostruzioniMeccanicheNucleari,Pisa,Italy, 1994.4(AvailableinPERRunderAccessionNumberML003769914)9.M.G.OrtizandL.S.Ghan,"UncertaintyAnalysisofMinimumVesselLiquidInventoryDuringaSmallBreakLOCAinaBabcockandWilcoxPlant,"

NUREG/CR-5818,USNRC,December1992.

310.W.Wulffetal.,"UncertaintyAnalysisofSuppressionPoolHeatingDuringanATWSinaBWR-5Plant,"NUREG/CR-6200,USNRC,March1994.

311.G.E.Wilsonetal.,"Phenomena-BasedThermalHydraulicModelingRequirementsforSystemsAnalysisofaModularHighTemperatureGas-CooledReactor,"

NuclearEngineeringandDesign,136(pp.319-333),1992.12.R.A.Shawetal.,"DevelopmentofaPhenomenaIdentificationandRankingTable(PIRT)forThermal-HydraulicPhenomenaDuringaPWRLarge-BreakLOCA,"

NUREG/CR-5074,USNRC,August1988.

313.J.C.WatkinsandL.S.Ghan,"AHPVersion5.1,UsersManual,"EGG-ERTP-10585,IdahoNationalEngineeringLaboratory,October1992.

4(AvailableinPERRbyAccessionNumberML003769902)14.J.ReyesandL.Hochreiter,"ScalingAnalysisfortheOSUAP600TestFacility(APEX),"NuclearEngineeringandDesign,pp.53-109,November1,1998.15.S.Banerjeeetal.,"ScalingintheSafetyofNextGenerationReactors,"

NuclearEngineeringandDesign,pp.111-133,November1,1998.16.V.Ransom,W.Wang,M.Ishii,"UseofanIdealScaledModelforScalingEvaluation,"NuclearEngineeringandDesign,pp.135-148,November1,1998.17.M.Ishiietal.,"TheThree-LevelScalingApproachwithApplicationtothePurdueUniversityMulti-DimensionalIntegralTestAssembly(PUMA),"NuclearEngineeringandDesign,pp.177-211,November1,1998.18.RegulatoryGuide1.157,"Best-EstimateCalculationsofEmergencyCoreCoolingSystemPerformance,"USNRC,May1989.

219."RELAP5/MOD3CodeManual,ModelsandCorrelations,"NUREG/CR-5535,Volume4,USNRC,August1995.

320.J.Sporeetal.,"TRAC-PF1/MOD2,TheoryManual,"NUREG/CR-5673(Draft),July1993.(Availableelectronicallyat<www.nrc.gov/RES/TRAC-P>

.)

33AppendixAADDITIONALCONSIDERATIONSINTHEUSEOFTHISREGULATORYGUIDEFORECCSANALYSISA.1BACKGROUNDSection50.46of10CFRPart50,asitexistedpriortoSeptember1988,providedtherequirementsfordomesticlicensingofproductionandutilizationfacilitiesusing conservativeanalysismethods.Theacceptancecriteriaforpeakcladtemperature, claddingoxidation,hydrogengeneration,andlong-termdecayheatremovalwerelistedin 10CFR50.46(b).AppendixKto10CFRPart50providedspecificrequirementsrelatedto ECCSevaluationmodels.Therequirementsof10CFR50.46wereinadditiontothe requirementsofCriterion35of(GDC35)inAppendixAto10CFRPart50.GDC35 statesrequirementsforelectricpowerandequipmentredundancyforECCSsystems.

Chapter15.6.5.ofNUREG-0800,theStandardReviewPlan,describesforreviewersthe scopeofreview,acceptancecriteria,reviewprocedures,andfindingsrelevanttoECCS analysessubmittedbylicensees.Chapter15.0.2ofNUREG-0800isthecompanionSRP sectiontothisregulatoryguide.InSeptember1988,theNRCamendedtherequirementsof10CFR50.46andAppendixKsothattheregulationsreflectedtheimprovedunderstandingofECCSperformance duringreactortransientsthatwasobtainedthroughextensiveresearchperformed betweenthepromulgationoftheoriginalrequirementsinJanuary1974andSeptember 1988.ExamplesofthatbodyofresearchcanbefoundinReferenceA-1.Further guidancetolicenseesorapplicantswasprovidedinMay1989byRegulatoryGuide1.157, "Best-EstimateCalculationsofEmergencyCoreCoolingSystemPerformance."The amendmentto10CFRPart50andRegulatoryGuide1.157nowpermitlicenseesor applicantstouseeithertheAppendixKconservativeanalysismethodsorarealistic evaluationmodel(commonlyreferredtoasbest-estimateplusuncertaintyanalysis methods).Thatis,theuncertaintyinthebest-estimateanalysismustbequantifiedand consideredwhencomparingtheresultsofthecalculationswiththeapplicablelimitsin10 CFR50.46(b)sothatthereisahighprobabilitythatthecriteriawillnotbeexceeded.It maybenotedtheacceptancecriteriaforpeakcladdingtemperature,claddingoxidation,hydrogengeneration,andlong-termdecayheatremovaldidnotchangewiththeSeptember1988amendment.A.2NEEDFORREGULATORYGUIDANCEUPDATEFORECCSANALYSISTheregulatorystructuredescribedabovewasstronglyfoundedonthesupportingworkdocumentedinReferenceA-2.Therefore,itisimportanttoupdatetheregulatorystructure toreflectthelastelevenyearsofadvancementinbest-estimateplusuncertaintyanalysis methods.Examplesoftheextensionofevolvingbest-estimateplusuncertaintyanalysis methodstoboththeoldandnewadvancedreactordesignscanbefoundinReferences A-3throughA-9ofthisappendix.A.3UNCERTAINTYMETHODOLOGY 34Thebest-estimateoptionin10CFR50.46(a)(1)(i),allowedsince1988,requiresthat:Uncertaintiesintheanalysismethodandinputsmustbeidentifiedandassessedsothattheuncertaintyinthecalculatedresultscanbeestimated.Thisuncertainty mustbeaccountedfor,sothat,whenthecalculatedECCScoolingperformanceis comparedtothecriteriasetforthinparagraph(b)ofthissection,thereisahigh levelofprobabilitythatthecriteriawouldnotbeexceeded.Tosupporttherevised1988ECCSrule,theNRCanditscontractorsandconsultantsdevelopedanddemonstratedanuncertaintyevaluationmethodologycalledcodescaling, applicability,anduncertainty(CSAU)(Ref.A-2).Whilethisregulatoryguideisoriented towardtheCSAUapproach,includingitsembeddedPIRTprocess,itisrecognizedthat otherapproachesexist.SincetheCSAUdemonstrationwasnotaplant-specific application,evaluationofinputuncertaintiesrelatedtoplantoperationwasnot emphasized.Proprietarymethodologieshavebeensubmittedtoandapprovedbythe NRCthatfullyaddressuncertaintiesinanalysismethodsandinput.Thus,other approachestodeterminethecombineduncertaintyinthesafetyanalysisarerecognized ashavingpotentialadvantages,aslongastheevaluationmodeldocumentationprovides thenecessaryvalidationofitsapproach.Thesafetycriteria(PCT,H 2generation,etc.)specifiedin10CFR50.46remainunchangedregardlessoftheuncertaintymethodologyusedinalicensingorregulatory submittal.Similarly,thegeneralguidelinesinRegulatoryGuide1.157withregardtothe phenomena,components,andcomputermodelsalsoremainunchanged.Thus,thefocus oftheremainderofthissectionisthoseconsiderationsprimarilyrelatedtodetermining the:Relativeimportanceofthephenomenaorprocessesandcomponents,andthosethatshouldbeincludedintheuncertaintyanalysis,Methodofestablishingtheindividualphenomenonorprocesscontributiontothetotaluncertaintyinthesafetycriteria,andMethodtocombinetheindividualcontributionstouncertaintyintothetotaluncertaintyinthesafetycriteria.CSAUandothermethodsaddresstherelativeimportanceofphenomenaorprocesses,thedifferencebeingintheapproach.CSAUusesthePIRTprocessinwhichrelative importanceisestablishedbyanappropriategroupofexpertsbasedonexperience, experimentalevidence,orcomputer-basedsensitivitystudies.Whenfinalized,the resultingPIRTsguidethedegreeofefforttodeterminetheindividualphenomenonor processuncertaintyinthesafetycriteria.ThePIRTprocessresultsalsoguidethemethod usedtocombinetheindividualcontributionsintoanestimateofthetotaluncertaintyinthe safetyanalysis.Commonly,butitisnotrequired,aresponsesurfaceisdevelopedtoact asasurrogateforthecomputercodesusedinestimatingthetotaluncertainty.The responsesurfacecanthenbeextensivelyMonteCarlosampledtodeterminethetotal uncertainty.Theuseoflimitedcomputercalculationstodevelopanaccurateresponse surfaceisfollowedbysufficientMonteCarlosamplingoftheresponsesurfaceinaneffort tobeasthoroughasnecessaryyetaseconomicalaspossible.Therefore,themajorcost oftheCSAUmethodologyisrelatedtotheextensiveexpertstaff-hoursnormallyrequired bytheexpertpaneltoperformthePIRTprocess.AdditionaladvantagesoftheCSAUare thatithasbeenusedbytheUSNRC,andthedetailsofthemethodologyhavebeenwell documented(Ref.A-2).

35ApotentialdisadvantageisrelatedtothedependencyofthenumberofcomputersimulationsonthenumberofphenomenaorprocessesdeterminedinthePIRTthatmay beneededtoestimatethetotaluncertainty.Thatis,atleasttwo"singleparameter change"runsmustbemadeforeachrequiredphenomenonorprocess.Inaddition, cross-productrunsmustbemadewhenseveralofthephenomenaorprocesseshave significantcovariance.Thecross-productrunsmayinvolvechangerunsoftwo parameters,threeparameters,orfourparameterstoadequatelydeterminetheeffectof nonindependentphenomenaorprocesses.Incontrast,othermethods(Ref.A-7)mayonlyuseapanelorindividualexperiencetodeterminewhatphenomenaorprocessesmaycontributetothetotaluncertaintyinthe safetycriteriaandadequateestimatesofthevariabilityofthosephenomenaorprocesses.

SimilartoCSAU,theestimatesoftheindividualparametervariationsarebasedonexpert experience,experimentaldata,andavailablesensitivestudies.Alargenumberof computersimulationsmayberequiredbecausethenumberofcomputercalculations neededtodeterminethetotaluncertaintyisindependentofthenumberofcontributors.

Thatis,thenumberofcomputersimulationsisdependentonlyontheprobabilityand confidencelimitsdesirableinthefinalresults.Forexample,95%/95%limitsrequire approximately90simulationsregardlessofthenumberofphenomenaorprocesses selectedascontributors.Thisfeatureisachievedthroughtheuseofuniquestatistical assumptionswithrespecttohowtheindividualcontributoruncertaintydomainissampled.

Thereisnotastrongnon-proprietaryprecedencethatcouldbeusedaprioribythe USNRCinapprovingsuchalicensingorregulatorysubmittaltoevaluateoverall uncertainty.Accordingly,suchsubmittalswouldinitiallyrequiresignificantvalidationofthe methodology.Thesameisconsideredtobetrueofuncertaintymethodologiesdescribed inReferenceA-7thatmightbeused.Anuncertaintymethodologyisnotrequiredfortheoriginalconservativeoptionin10CFR50.46.Rather,therequiredfeaturesofAppendixKprovidesufficientconservatism withouttheneedforanuncertaintyanalysis.ItshouldbenotedthatSectionII.4of AppendixKrequiresthat"Totheextentpracticable,predictionsoftheevaluationmodel,or portionsthereof,shallbecomparedwithapplicableexperimentalinformation."Thus,AppendixKrequirescomparisonstodatasimilartothoserequiredforthebest-estimateoption,butwithouttheneedforanuncertaintyanalysis.However,poor comparisonswithapplicabledatamaypreventNRCacceptanceoftheAppendixKmodel.

1CopiesareavailableatcurrentratesfromtheU.S.GovernmentPrintingOffice,P.O.Box37082,Washington,DC20402-9328(telephone(202)512-1800);orfromtheNationalTechnicalInformationServicebywritingNTISat5285 PortRoyalRoad,Springfield,VA22161;(telephone(703)487-4650).Copiesareavailableforinspectionorcopying forafeefromtheNRCPublicDocumentRoomat11555RockvillePike,Rockville,MD;thePDR'smailingaddressis USNRCPDR,Washington,DC20555;telephone(301)415-4737or(800)397-4209;fax(301)415-3548;emailis

PDR@NRC.GOV.

2ElectroniccopiesareavailableinNRC'sPublicElectronicReadingRoom,whichcanbeaccessedthroughtheNRC'swebsite,<

WWW.NRC.GOV>.36APPENDIXAREFERENCESA-1"CompendiumofECCSResearchforRealisticLOCAAnalysis,"NUREG-1230,USNRC,December1988.

1A-2B.Boyacketal.,"QuantifyingReactorSafetyMargins,ApplicationofCodeScaling,Applicability,andUncertaintyEvaluationMethodologytoaLarge-Break Loss-of-CoolantAccident,"NUREG/CR-5249,USNRC,December1989.

1A-3G.E.Wilsonetal.,"PhenomenaIdentificationandRankingTablesforWestinghouseAP600SmallBreakLossofCoolantAccident,MainSteamLine Break,andSteamGeneratorTubeRuptureScenarios,"NUREG/CR-6541, USNRC,June1997.

1A-4A-4.M.G.OrtizandL.S.Ghan,"UncertaintyAnalysisofMinimumVesselLiquidInventoryDuringaSmallBreakLOCAinaBabcockandWilcoxPlant,"

NUREG/CR-5818,USNRC,December1992.

1A-5U.S.Rohatgietal.,"BiasinPeakCladTemperaturePredictionsDuetoUncertaintiesinModelingofECCBypassandDissolvedNon-CondensableGas Phenomena,"NUREG/CR-5254,USNRC,September1990.

1A-6C.D.Fletcheretal.,"AdequacyEvaluationofRELAP5/MOD3,Version3.2.1.2forSimulatingAP600SmallBreakLoss-of-CoolantAccidents,"INEL-96/0400 (Nonproprietaryversion),April1997.

2(AvailableinPERRbyAccessionNumberML003769921)A-7H.Holmstrometal.,"StatusofCodeUncertaintyEvaluationMethodologies,"ProceedingsoftheInternationalConferenceonNewTrendsinNuclearSystemThermohydraulics,DipartimentodiCostruzioniMeccanicheNucleari,Pisa,Italy, 1994.2(AvailableinPERRunderAccessionNumberML003769914)A-8G.E.WilsonandB.E.Boyack,"TheRoleofthePIRTProcessinExperiments, CodeDevelopmentandCodeApplicationsAssociatedwithReactorSafetyAnalysis,"NuclearEngineeringandDesign,186(pp.23-37),1998.A-9"RELAP5/MOD3CodeManual,ModelsandCorrelations,"NUREG/CR5535,Volume4,USNRC,August1995.

1 1Inthisregulatoryanalysis,an"applicant"meansanapplicant,alicensee,avendor,amethodsdeveloperorotherentitythatpetitionsforapprovalofananalyticalmodeltobeusedonbehalfofalicenseeorapplicant.

37REGULATORYANALYSIS1.PROBLEMSection50.34,"ContentsofApplications;technicalinformation"of10CFRPart50,"DomesticLicensingofProductionandUtilizationFacilities,"requiresthat:1.SafetyAnalysisReportsbesubmittedthatanalyzethedesignandperformanceofstructures,systemsandcomponentsprovidedforthe preventionofaccidentsandthemitigationoftheconsequencesofaccidents,2.AnalysisandevaluationofECCScoolingperformancefollowingpostulatedloss-of-coolantaccidents(LOCAs)shallbeperformedinaccordancewith therequirementsofSection50.46,and3.Thetechnicalspecificationsforthefacility(Section50.36)willbebasedonthesafetyanalysis.VarioussectionsofChapter15oftheStandardReviewPlan(SRP)(NUREG-0800)instructtransientandaccidentreviewerstoinitiategenericreviewsofmodelsusedby applicantsorlicenseesiftheanalyticalmodelshavenotbeenpreviouslyreviewedand foundacceptablebythestaff.WhiletheSRPdiscussesreviewoftheinputtothesemodels,noguidanceisprovidedregardingreviewoftheanalyticalmodels.Exceptfor RegulatoryGuide1.157onbestestimateECCSanalysis,noguidanceexistsfor applicants 3onthedevelopmentandassessmentoftransientandaccidentanalysismethods.Recentreviewshaveshownthatsuchguidancecouldhaveapositiveeffectin termsofclarifyingexpectationsandstreamliningthereviewprocess.Toproduceaviable product,certainprinciplesshouldbeaddressedduringthemodeldevelopmentand assessmentprocess;thesewouldbedescribedinDraftRegulatoryGuideDG-1096.The accompanyingnewsectionofChapter15oftheSRPaddressesthesameprinciplesas DG-1096butfocusesonresponsibilitiesoftheanalyticalmodelreviewer.Thisregulatory analysisappliestoboththeproposedDG-1096andtheproposedsection15.0.2ofthe

SRP.2.ALTERNATIVEAPPROACHESTwoalternativeapproacheswereconsidered:1.Takenoaction 2.Provideguidanceonthedevelopment,assessment,application,andreviewofmethodsusedtoanalyzetransientsandaccidentsasdescribedin10CFR50.34.Thefirstalternative,takenoaction,wouldrequirenoadditionaldirectcostfortheNRCstafforapplicantsovercurrentconditions,sincenochangetotheprocesswouldoccur.

Thisprocesswouldinvolvesignificanteffortonthepartofapplicantstoanticipatethetype ofinformationthatwouldbeacceptabletotheNRCstafftodemonstratethecapabilityof theanalysismethods.Inaddition,theNRCstaffreviewwouldinvolveconsiderableeffort 38anditerationwiththeapplicanttodeterminewhethertheproposedmethodsareacceptable.Thesecondalternative,providetheanalyticalguidancetoindustryandreviewer,wasconsidered.ProvidingguidanceshouldreducetheNRCstaffeffort.Definingacceptable analyticalmodelingprinciplesshouldstreamlinethereviewprocess.Itmayappearto involvemoreeffortonthepartofapplicants,butifoneconsidersthetimeandeffortspent oniterationswiththestaff,realsavingscouldberealizedinthelongrun.The development,assessment,application,andreviewprocessdescribedintheproposed regulatoryguideandSRPchapterarebasedonaninitialidentificationoftheplantand transientfollowedbyassessmentofphenomenaandprocessesandapplicationofthese phenomenaandprocessestocodedevelopmentandassessment.Akeyprinciplehereis thatguidancewouldfocusthedevelopment,assessment,andreviewasclearlyas

possible.3.VALUESANDIMPACTSInthisanalysis,theprobabilityofguidancehavingapositiveeffectandtheprobabilityofthateffectontheachievementofoverallsafetygoalsarenotknownquantitatively.Inthe summarybelow,animpactisacostinschedule,budget,staffing,oranundesiredattribute thatwouldaccruefromtakingtheproposedapproach.3.1Alternative1-TakeNoActionThisalternativehasaperceivedcostbenefitsincethereareno"start-up"activities.Italsoprovidesflexibility,sinceeachapplicantwoulddeviseitsownprocessforanalytical methodsdevelopment.However,theNRCwouldcontinuetoreceiverequeststoreview andapproveanalyticalmethodsthatarepreparedwithnoclearguidanceonwhatthe NRCstaffconsiderstobeanacceptablemodel.Thelackofanidentifiedsetofguidelines andpracticeswouldhaveadverseeffectsonthelevelofstaffeffortrequiredtoconduct modelreviewsandtoassureconsistencyofprincipleamongreviews.Itwouldtakea longertimefortheapplicanttounderstandstaffreviewerexpectationsandforthestaffto haveaclearunderstandingofwhattheapplicantwasprovidingtosatisfythose expectations.Thus,althoughtheinitialcostwouldapparentlybelow,takingnoaction couldresultingreatertotalcosts,toboththeNRCstaffandtheapplicant,duringthe reviewprocess.Value-Novaluebeyondthestatusquo Impact-Schedule,budget,andstaffingcost,tothestaffandapplicantassociatedwithregulatoryuncertainty.3.2Alternative2-ProvideGuidanceonAnalyticalMethodsDevelopmentforChapter15EventsAbenefittotheNRCstaffofthisalternativewouldbeamorecomprehensiveunderstandingofafocusedbasisforthedevelopmentoftheanalyticalmethodsunder review.Fromtheapplicant'sperspective,therewouldbeareductionintheneedtoiterate withthestaffasthereviewprogressed.Theresultsofusingsuchaprocesswouldalso resultinabetterproductmoreclearlysuitedtothespecificanalyticaltaskathand.The costsinvolvedwouldbesomeadditionalworkanddocumentationinitially.Thisalternative wouldhavethevalueofpromotingapredeterminedcommonunderstandingthathas 39gainedacceptanceoverthelastdozenyearswiththeaffiliatedtechnicalcommunity.Theresultwouldbehigherconfidenceinresultsoftransientandaccidentanalysis.Value-Commonunderstandingofgoodpracticesfordevelopment,assessment,andapplicationofanalysismethods.1.Lessburdensomedevelopmentalandreviewiterationsbetweenstaffandapplicant.2.Minimizationofregulatoryuncertainty.Impact-Additionalinitialworkanddocumentationbytheapplicant.4.CONCLUSIONSExperiencewithrecentmodelreviewshasdemonstratedtheneedforguidanceinthisarea.Onbalanceitisbelievedthatguidanceintheformofgoodprinciplesoftransient andaccidentcodedevelopmentandassessmentoutweighstherelativelysmallcostof initialworkanddocumentation.Therefore,DraftRegulatoryGuideDG-1096andthe proposednewStandardReviewPlanSection15.0.2shouldbeissuedforpubliccomment.ADAMSAccessionNumberforDG-1096:ML003770849