ML17333A691

Rev 0 to, Review of Donald C. Cook Nuclear Power Plant Methodology for Analysis of Fire Barrier Ampacity Derating Factors, Ltr Rept

ML17333A691
Person / Time
Site:	Cook
Issue date:	06/28/1996
From:	NOWLEN S SANDIA NATIONAL LABORATORIES
To:	NRC (Affiliation Not Assigned)
Shared Package
ML17333A671	List: ML17333A691
References
NUDOCS 9612060351
Download: ML17333A691 (32)

Text

AReviewoftheDonaldC.Cook4i'uclearPlantMethodologyfortheAnalysisofFireBarrierAmpacityDeratingFactorsALetterReporttotheUSNRCRevision03une28,1996Preparedby:SteveNowlenSandiaNationalLaboratoriesAlbuquerque,NewMexico87185-0737(505)845-9850Preparedfor:RonaldoJenkinsElectricalEngineeringBranchOfficeofNuclearReactorRegulationU.S.NuclearRegulatoryCommissionWashington,DC20555AttachmenttotheEnclosureqbl206035i9bl202P.----,-PDRPDRADGCK05000315 TABLEOFCONTENTS:SectionPa~eFORWARD)n1OVERVIEW...1.1Objective1.2ReportOrganizationUTILITYAMPACITYDERATINGAPPROACH2.1OverviewofAnalysisApproach...2.2TheUtilityPart1AnalysisMethod2.2.1BasisandUnderlyingAssumptions2.2.2TheoreticalDevelopment2.2.3Validation.........~2.2.4SummaryofTechnicalConcerns....2.3TheUtilityPart2AnalysisMethod2.3.1BasisandUnderlyingAssumptions2.3.2TheoreticalDevelopment2.3.3Validation...2.3.4SummaryofTechnicalConcerns.2.4ConduitApplicationsoftheModel.....2.5CoordinationoftheUtilityAnalysisPackageElements..3UTILITYEXAMPLECALCULATIONS....3.1Overview..3.2CableTray1AZ-P8.3.3TheCNPAppendixRConduits......3.4SummaryofInsightsandFindings...4SUMMARYOFFINDINGSANDRECOMMENDATIONS....3.34...,5.6.8.8.&13...14...15...171919.1920.22,...,245REFERENCES.28 TheUnitedStatesNuclearRegulatoryCommission(USNRC)hassolicitedthesupportofSandiaNationalLaboratories(SNL)inthereviewofutilitysubmittalsassociatedwithfireprotectionandelectricalengineering.ThisletterreportdocumentstheresultsofaSNLreviewofasubmittalfromtheDonaldC.CookNuclearPlant(CNP).Thissubmittaldealswiththeassessmentofampacityloadsforfirebarrierprotectedcables.ThisdocumentwassubmittedbytheutilityinresponsetoUSNRCGener'tt92-iceeraninresponsetoasubsequentUSNRCRequestforAdditionalInformation(RAI).ThisworkwasperformedasTaskOrder9,Subtask4ofUSNRCJCNJ20l7.

1,0OX'ERVIE1.1ObjectiveInresponsetoUSNRCGenericLetter92-08,andasubsequentUSNRCRequestforAdditionalInformation(RAI)theDonaldC.CookNuclearPlant(CNP)provideddocumentationofamethodologyfortheassessmentofitscableampacityloadingfactorsincludingtheeffectsoffirebarrierampacityderatingimpacts.Th'als.issumittcabletraewosetsofspecificcaseexamples,oneforaseriesofconduitsandfoneoraeray,toillustratethemethodology.Theutilityhasalsoincludedtstoincueatestreportwiciscitedassupportingthevalidityoftheutilitycalculationmethod.Thesubmittalreviewedwasdocumentedinautilityletter:Letter,May12,1995,(docketnos.50-315,50-316itemAEP:NRC:0692DF),E.E.Fitzpatrick,IndianaMichiganPowerCo.,totheUSNRCDocumentControlDeskincludingsixattachmentsasfollows:-Attachment1:SummaryofAmpacityDeratingAnalyses-Attachment2:AIEETransactionsPapers57-660&50-52-Attachment3:CableTrayAllowableFillDesignStandard-Attachment4:AnalysesandMathematicalModels-Attachment5:RepresentativeAmpacityDeratingCalculationResults-Attachment6:ResultsfromTestReportPCL-542SNLwNLwasrequestedtoreviewthissubmittalunderthetermsofthegeneraltechnicalsupportcontractJCNJ-2017,TaskOrder9,Subtask4.Thisletterreportdocumentstheinitialresultsofthisreview.1.2ReportOrganizationSection2providesareviewofthetechnicalaspectsoftheutilityanalyticalapproachtoampacityassessments.Inparticular,twoseparateanalyticalmodelsaredevelopedyeutility,andeachisreviewedindependently.Inaddition,considerationisgiventothevalidation,orlackthereof,ofeachofthetwomodels.Section3providesareviewoftheexamplecalculationsprovidedbyCNP.Thisincludescalculationsforonecabletrayandforanundeterminednumberofconduits.SNL'sreviewhasincludedacomparisonoftheutilityresultstomoreconventionalapproachesinvolvingecratingoftabulatedampacitylimits.Section4providesasummaryofthemajorissuesandconcernsidentifiedinthisreview.Section5identifiedthereferenceddocuments.

2.0UTILITYA!vG'ACITYDERATIiGAPPROACH2.1OverviewofAnalysisApproachTheapproachtakenbyCNPisbasedonalargelyanalyticalassessmentofactualcableampacitylimitswithlimitedexperimentalvalidationoftheoverallanalysismethod.TheandysisapproachtakenbyCNPisquiteuniqueincertainrespects.Inparticular,theutilityhascitedthatitscableinstallationproceduresforpowercablesincabletraysrequiredtheuseof,ineffect,amaintainedspacingapproach.Thatis,powercablesincabletraysatCNPwereallinstalledsoastoconformtothefollowingfeatures:nomorethanasinglelayerofpowercablesisinstalledinanytray,eachcableinthetraytobeseparatedfromitsneighborswitha'gapofnolessthanI/3ofthe.diameterofthelargerofthetwocables,thesumoftheinstalledcableouterdiametersshallnotexceed75/oofthefulltraywidth.Thisisanimportantobservation.Thebulkoftheutilitymodeldevelopmentisaimedataddressingsparselyloadedcabletrays,althoughthefinalresultsareapparentlyappliedtoconduitsaswell.Theutilityanalysesareactuallyperformedintwoparts.Theexactintentofeachstepremainssomewhatunclear,andmayhavebeenmisinterpretedinthisreview(apointrequiringclarification).Theutilitydiscussionimpliesthefollowinginterpretation:PartIAnalysis:Givenanoverallheatrejectioncapacityforthecabletray,calculatetheallowableampacitylimitforindividualcables.ThemethodusedinthispartoftheanalysisisdocumentedinAppendixAoftheutilityAttachment4.Part2Analysis:Calculatetheoverallheatrejectioncapacityforagivencabletray(orconduit)basedonheattransfercorrelationsandcalculations.ThemethodusedinthispartoftheanalysisisdocumentedinAppendixBoftheutilityAttachment4.However,theutilitydescriptionofthePart2analysisstepappearstoincludebothaspectsoftheproblemtosomeextent.Inparticular,the*Part2theoreticaldevelopmentpresentsequationsforthecalculationofampacitylimitsnotfrthcducculationoftotalheatrejectioncapacity.Infact,themethodappearstobeintendedtoassesstheampacitylimitsofindividualcables.Hence,itremainsunclearastotheintentoftheAppendixAandAppendixBmethodsandhoweachworkswithththtoertoprovideacompletesolution.Itisalsounclearhowthemethodshavebeenappliedtoconduits.Thatis,theutilityprovidesexamplesinvolvingconduits,butthePartIpartitioningmethodwouldclearlynotapplytoconduitapplications.Hence,howtheseconduitswereanalyzedremainsunclear.

Eachofthetwoparisoftheoverallutilityanalysisapproachwillbereviewedindetailinthefollowingsubsections.Thesediscussionswillincludeareasofpotentialconcernidentifiedduringthereviewandtheacceptabilityoftheutilityvalidationarguments.However,overall,theutilityshouldprovideforsomediscussionofhowthetwopansoftheanalysisworktogetherinpractice,andshouldprovidemoredetailedexampleswhichil!ustratehowthetwopartsoftheanalysiswereusedindividuallytoobtainactualcableampacitylimitestimates.NotethatSection3ofthisreportwilldiscussthisfinalpoint,theexampleanalyses,inmoredetail.2.2TheUtilityPart1AnalysisMethod2.2.1BasisandUnderlyingAssumptionsAsnotedabove,itisourinterpretationthattheintentofmethodologydescribedinSectionAofutilityAttachment4(whichwillbereferredtohereasthePart1analysisforconvenience)istodeterminetheampacityofindividualpowercablesgivenanestimateoftheoverallheatrejectioncapacityofagivencabletray.(NotethattheheatrejectioncapacityiscalculatedseparatelyasdiscussedinSection2.3below.)Theutilitymethodology,ineffect,partitionstheoverallheatrejectioncapacitytoindividualcables.Thispartitioningisultimatelybasedonaratiooftheavailablethehetsurfaceareaofagivencabletothetotalsurfaceareaforallcablesinthtr0sineay.nceeeatrejectioncapacityforthegivencableisestablished,thecorrespondingampacitylimitiscalculatedusingasimple"I'R"typecalculation.Thecriticalunderlyingassumptioninthisanalysisderivesdirectlyfromtheutilitycableinstallationpracticewhich,asdiscussedinSection1above,requiredthatonlyasinglelayerofcablebeinstalled,andthatthosecablesbeinstalledwith,ineffect,amaintainedspacingbetweencables.Basedonthispractice,theutilityconcludesthattheheattransferbehaviorofitscabletraysisdominatedbyconvectionandradiation,ratherthanbyconduction.Thisisanimportantdistinctionbecauseconvectionandradiationarebothdrivenprimarilybytheavailableheattransfersurfacearea,inthiscase,theactualsurfaceareaoftheinstalledcables.Incontrast,forgeneralcabletrays,suchasthosetestedbyStolpefl]inhispioneeringcabletrayampacitywork,heattransferwithinthecablemassisdominatedbyconduction.This,CNPnotes,isreflectedinthefactthatopentopcabletrayampacitybtablesinthattheampacitylimitsarecitedasafunctionofthecabledepthofflldasedonthelimitingheatgenerationrateperunitvolumeofthecablemassforagivencablesize.Insomesenses,theutilityappearstotakethisanalogyalittlefatherthanitshouldbetaken,althoughthisisnotconsideredtobedirectlyrelevanttotheultimatevalidityoftheutilitymethodology.'hatis,CNPcitesthecomparisonmadebyStolpebetweenthevolumeoccupiedbyseven412AWGcablesasessentiallyequaltothatofasingle4l0cable.Theutilitythenstatesthat"thustheheatgeneratedbythetwoconfigurationsshouldbeequivalentunderuniformheatdistributionconditions."However,thisidealismdidnotholdtrueforStolpe'sfinalresults.AreviewoftheICEAP-54<40ampacitytableswillreveal

",Z.ZThoreticalDevelopmentForthesparselyloadedtraysatCNPconvectionandradiationareconsideredthegoverningmodesofheattransfer.ThisassumptionisconsideredtoaccuratelyreflectthesparselyloadedcabletraysatCNP(withonepossibleexceptionaswillbediscussedinSection2.2.4below).Basedonthisassumption,theutilitypostulatesthatindividualcableampacitylimitscanbeestablishedbasedonconsiderationoftheheatgenerationrateperunitofcablesurfacearea.Itisthisassumptionwhichdrivestheutilitymethod.TheutilitydevelopmentofitsthermalmodelispresentedinSectionsA.3andA.4ofAttachment4toitssubmittal.Thedevelopmentfirstbeginsbyconsideringacaseinwhichallthecablesareassumedtobeofthesamesize.Thetotalavailablesurfaceareaofthecablesinthetrayisthengivenby:Where(A,)isthecabletotalsurfacearea,(n)isthenumberofcables,and(d)istheouterdiameterofthecable.Theutilitythendefinesthe"percentagefillofthetray"(F)as:F=nd(2)Theutilityacknowledgesthatthisdefinitionofpercentagefillisnotconsistentwithgeneralindustrypractice.SNLfurthernotesthattodescribethisvalueasapercentagefillisnotaccurate.Thequantity(n~d)isalineardimensionwhichquantifiesthetotalwidthoftheinstalledcables.Hence,(F)willhaveunitswhicharethesameasthoseusedtoexpress(d),inchesforexample.Thisisbynomeansa"percentagefill"termA"becausea"percentage"termisaratioofonequantitytoanotherofequaldimens'o.percentage"termshouldbedimensionless.Itwouldbemoreaccuratetosimplycallthisvaluethe"totalfill"ratherthanthe"percentagefill."Thisisaminorpointofnomenclature,andsolongasthevalueisusedconsistently,shouldintroducenoerrors.Theutilitythengoesontoexpressthetotalheatgenerationrate(Q)foratrayasawholebasedonsimple"I'R"heating:Q=3nZ~R~~(3)where(I)isthecablecurrentand(R)istheACresistanceofthecableofinterest.Notethattheconstant(3)appearsbecausetheutilityassumesthateachofthecablesisathree-conductorcable.Combiningequations(2)and(4),theutilityeliminatesthevariable(n):thatthelimitingheatgeneratingcapacityfora4/0cableisontheorderof12-18timesthatofa12AWGcable,not7timesasthisidealizationwouldsuggest.Thisisonlyaveryminorpointandhasnorealimpactontheutilitymethodology.

CNPthenrearrangesthisexpressiontosolveforthecurrent(I)asafunctionofheatrejectioncapacity(Q).Whilethissameequationiswritteninatleastthreeformatsthroughthedevelopment,thefinalexpressionispresumablyutilityequation9asfollows:gd3FRacorbyrearrangingthetermsinthisexpression:(6)This,then,istheultimatemethodofpartitioningthetotalheatrejectioncapacityofacabletrayasawholedowntoindividualcables.Notethatasshownin(6),thegrouping([d/R]'isafunctionoftheindividualcable;thediameterdividedbytheACresistance.Forlargercables,'thediameterincreasesandtheACresistancedecreases,sothevalueofthisgroupingincreaseswithcablesize.Thegrouping([Q/3F]'")isafunctionofthetray;thelimitingtrayheatloaddividedbythetotalcablefill.2.2.3UalidationTheutilityvalidationofthisthermalmodelispresentedinFigureA-1oftheutilityAttachment4.Thisfigureplotscurrent(I)versusthe([d/R]'cablecharacteristicgrouping.Consistentwithequation(6),theutilityanticipatesalinearrelationshippassingthroughtheorigin.Thisis,infact,demonstratedforthreespecificcabletrayconfigurations,eachinvolvinga67%fillofthetray.Theconfigurationsinvolveaventilatedtraywithaventilatedcover,asolidtraywithasolidcover,andaventilatedtraywitha1hrbarriersystem.Ineachcasethelinearrelationshipholdstrue.Whilethisplotdoesprovidesomeencouragingresults,itisnotsufficienttofullyvalidatetheutilityPart1analysismethod.Themostimportantfactorwhichhasnotbeenestablishedbytheutilityisrelatedtotheheatloads(Q)whichweremeasuredintheseexperiments.Thatis,theutilitymethodassumesthatforagivencabletray,thetotallimitingheatrejectioncapacityisafunctionofthetotalavailablecablesurfacearea,whichisinturnafunctionoftheloadingfactor(F}.Hence,inordertovalidatethisassumption,theutilitymustdemonstratethatinitsexperiments,foranyonecombinationoftrayconfiguration(e.g.foranopentray)loadingfactor(F)andtemperaturedifference(T>>-T,),thesameoverallheatloadwasmeasuredindependentofthesizeofcablestested.Forexample,alltrayswereapparentlyloadedto67%oftheavailablewidth.Wemustassumethatcableswerecurrent-loadedtoestablishahot-spot temperatureof90'C(anassumptionwhichneedstobe'fdbCbl'veriieyi'.Tovalidateitsmodel,ChPshouldshowthatthetotalresistanceheatingloadmeasuredforthevariouscablesizesremainedconstantforeachofthetrayconfigurationstestedregardlessofthecablesizeinstalled.WithoutadditionalinformationregardingthenatureofthetestscitedbyCNP,thisassessmentcannotbemade.Inordertocompletethisassessmenttheutilitymustprovideadditionalinformationonitstestprotocolandresults.Forexample,providingacopyofthemissingAppendixCmayprovesufficient.Attheleast,theutilityshouldensurethatthefollowinpointsareaddressed:aeoowingThephysicaldescription'oftheexperimentalsetupandtestarticlesshouldbeprovided.Thisshouldincludeaphysicaldescriptionofthetestarticlesthemselves(traywidths,heightandlength),aphysicaldescriptionofecalestested(sizeandconductorcount),adescriptionoftestinstrumentation,andadescriptionoftheinstallationofcablesinthetestarticle(types,spacingandlocations).Adiscussionofthetestingprotocolisneeded.Thisshouldi1dthadiscus'ionofthespecificobjectiveofthetest.Forexample,werecableissouincueecurrentsadjustedtoachieveadesiredtemperaturestate?Whatwasthedesiredfinal.endstateoftheexperimentsin'termsoftemperatureandcurrent?Wereifferentsizedcablesinstalledinthesametestarticle,andifsohowwerethecurrentsforthesedifferentsizedcablesdetermined?Howdidthetestsverifythatasteadystateconditionwasachieved?Asummaryofthetestresultsinneeded.Thisshouldincludethefinalmeasuredvaluesofcurrentandtemperatureforeachofthecablesinanygiventestarticle.TheutilityshoulddescribetheunitsassociatedwithitsFigureA-1anddescribehowtestedcableparameterscorrespondtotheplottedvalues.Thecurrent(I)showninFigureA-1isassumedtobegiveninunitsofAmps,butthiswouldimplythatmostlyverylargecableswereusedintesting.Themeasuredcurrentvaluesrangeashighas700Awhichshouldgenerallyimplyacableofontheorderof750kcmil,whichappearstobemuchlargerthanthecablesconsideredelsewhere(thelargestcableidentifiedelsewhereinthesubmittalisaP4i0cable).Thephysicalcablesizestestedshouldbedescribed,andtheapplicabilityofthetestedcablestothecablesconsideredelsewhereintheanalysisshouldbe'established.2.2.4SummaryofTechnicalConcernsTheprimarytechnicalconcernregardingtheutilityPart1analysismethodwhichmustbeaddressedisoneofvalidation.Theutilityanalysisisbasedontwofundamentalassumptions:

0Jl~,~

Thetotallimitingheatloadforagivencabletravisafunctionofthetotalsurfaceareaofthecables,and-Thistotalheatloadcanbepartitionedtoindividualcablesbasedoneachcablescontributiontotheoverallavailablecablesurfacearea.CNPhasnotprovidedsufficientinformationtodeterminethatithas,infact,validatedtesetwocriticalassumptions.Additionaldetailregardingtheexperimentsitperformedtovalidatethisaspectofthethermalanalysismodelsisrequired.Section2.2.3aboveprovidesadescriptionoftheadditionalinformationrequired.'Thereisoneparticularfactorwhichmaybeinadequatelytreatedintheutilityanalysis.Thatis,thecabletemperatureratingof90'Cisbasedontheconductortemperature,notthesurfacetemperatureofthecable.Iwouldagreethatoncetheheatreachesthecablesurface,convectionandradiationbecomepredominant.However,withinthecableitself,conductionofheatthroughtheinsulationandjacketmaterialsremainsacriticalfactor.WhiletheutilityPart2analysisincludesathermalresistanceelementassociatedwiththeinsulation,theconcernhereiswiththePartlpartitionialp'oninganysisinwhichincreasingdiameteriscreditedforincreasingampacityinproportiontothecablediameter.Thiswouldbecorrectifthesurfacetemperatureofthecableremainedfixed(forexampleat90'C)butwillnotholdtrueifthecablesurfacetemperaturechangeswithdiameter.Infact,becausetheelectricalinsulationisalsoathermalinsulation,thethickerthatinsulationbecomes,thehigherthetemperaturedropthroughtheinsulationwillbeforagivenoverallrateofheatflow,andhence,thelowerwillbethesurfacetemperatureofthecable.Thiswouldcausesomeerrorintheutilitymethod,althoughbasedontheinformationprovided,itisimpossibletoassessthemagnitudeoftheerrorintroduced.Thepotentialpitfallcanbeillustratedasfollows.Fortwocablesofthesamewiregage,thethicknessoftheinsulationwoulddependprimarilyonthevoltageratingofthecable.Thus,a600Vcablewouldhaveasmalleroveralldiameterthanwoulda5000Vcable.Intheutilitymethod,thiswouldimplythatifthesetwocableswereinthesametray,theallowableampacityofthe5000Vcablewouldbepredictedashigherthanthatofthe600Vcableinaccordancewiththeratiooftheirdiameters.Inreality,theincreasedthicknessoftheinsulationforthe5000Vcablewouldgenerallyresultinareducedampacityforthemore"open"(maintainedspacing)applicationsuchasthoseatCNP.(Fortightlypackedcabletrays,asperStolpe,thevolumeistheoverridingfactorandampacitydoesincreasewithcablevoltagerating.However,moreopenconditionsallowmoregenerousampacitylimits,andcablevolumeisnottheoverridingfactorintheseratings.)Thisisparticularlytrueforthelargercablesizes.Note,forexample,thattheopenairampacitytablesinIPCEAP-46-426willgenerallydecreasetheallowableampacitylimitsforsingleconductorcableswithincreasingvoltageratingforcableslargerthanabout1/OAWG(thetransitionpointdependsonthetableconsideredandthetemperatureratingofthecable).Theutilitymethodwouldallowjusttheoppositeeffect.Theutilityshouldprovideforsomeassessmentofthisquestion.Oneapproachwhichmightprovefruitfulwouldbefortheutilitytoexaminetheopenairampacitytables forthetypesofcablesinuseatC4P,andforCbiPioattmpttovalidatetheirassumptionsregardingampacityasafunctionofcablediameter.Forexample,withthethreeconductorcables(whichseemtobequitecommonatCNP)thetransitionpointinthetablesisatmuchlargercablesizes.Alternatively,theutilitytestsmayprovidesomefurtherinformationonthisquestionforthespecificcablesinuseatCNP.Theinformationprovidedinthesubmittalisinsufficienttodeterminewhetherornotsuchacomparisoncanbemade.Inanycase,theutilitymustdemonstratethatitsdiameter-basedpartitioningapproachappropriatelyreflectstheseampacitylimitbehaviorswithincreasingvoltageratingandincreasingcablediameter.Thatis,theutilityshouldshowthat,infact,foritscablesampacitylimitsincreasewithcableoveralldiameter(voltagerating),andthatthisincreaseisconsistentwiththeirdiameterscalingassumptions.Theutilitymightalsoconsultmanufacturerrecommendedampacitylimitsofcablesincomparisontocablediameters.2.3TheUtilityPart2Analysis2.3.1BasisandUnderlyingAssumptionsSectionBoftheCNPAttachment4documentsasecond,separatebutrelated,thermalanalysismethod(thismethodwillbereferredtohereastheutility"Part2"analyses).ThismethodisbasedlooselyontheNeher-McGrathapproachtocablethermalanalysis,althoughconsiderablelibertiesaretakenwiththatmethodology.Theintentofthemethodis,apparently,toprovideestimatesofthetotallimitingheatrejectioncapacityofagivencabletraybasedontheconfigurationofthetray,thenumberofcablesinstalledinthetray,andthenatureoftheprotectivefeatures(barrier)installedonthetray.Inexecutingthiscalculationtheutilityisinherentlyassumingthatitcanaccuratelymodelthephysicalphenomenaofinterestandtherebyestimatetheactualcableheatrejectioncapacityofagivencabletrayorconduitsystem.ThisoverallsystemheatrejectioncapacityisthenpartitionedtoindividualcablesinthetraythroughthePart1analysismethodasdiscussedin2.2above.2.3.2TheoreticalDevelopmentThebasicstructureofthePart2thermalmodelderivesdirectlyfromNeher-McGrath.Thegeneralexpressionisquitesimpleandstraightforward.Theapproachisbasedonforcingasteadystatebalancebetweentherateofheatgenerationinthecablesandtherateofheatdissipationfromthecablestotheambient.Therateofheatgenerationcanbeexpressedas:where(R)isthecableACresistance,(R)isthecableDCresistance,and(Y,)istheAC/DCresistanceincrement.Initsmostsimpleform,therateofheatdissipationcanbeexpressedas:

Thisassumptionisacknowledgedbytheutilitytobeofquestionablevalidity.Ingeneral,thetreatmentofarectangularsystemusingcirculargeometrycanintroduceanomalousbehavior,oftenassociatedwiththeparadoxofsurfaceareachanes.Thisassystemofnestedannularbodiesexperiencesanincreaseintheavailablesurfaceareaasonemovesoutwardthroughthesystem.Thisisnotacharacteristicofatypicalone-dimensionalrectangularsystem,andhence,greatcaremustbetakentoensurethatthesurfaceareasassumeddonotdistorttheheattransferbehavior.Becausetheutilityhasprovidednodetailedcalculationexamplesitisimpossibletodeterminehowsignificantthiseffectwouldbefortheutilityanalyses.WhileCNPstatesthattheyachieved"excellentcorrelationbetweencomputerdataandtestdata"noevidenceofthisvalidationisprovided.Thisaspectoftheutilitymodelisstillconsideredsuspect.Thisquestionwillberevisitedin2.3.4below.Theutilitycalculationofthecable-to-ambientthermalresistanceapparentlyincludesconsiderationofthevariouslayersinthesystem(thecableinsulation,thetrayorconduit,thebarrieritself)andtheairgapsbetweenlayers.However,theexactmannerinwhichthesevariousfactorsareimplementedinthemodelisunclearbecausetheutilitysimplycitescertaincorrelations,butdoesnotprovideanydescriptionoftheoverallfinalstructureofitsthermalmodel.WhiletheoverallmethodistakenfromNeher-McGrath,thecorrelationsusedtocharacterizethevariouslayersareactuallytakenfromapaperbyBullerandNeher[3].Whilethispaperwaspublishedin1950,itcitesa1933textbyMcAdams[4]aseoriginalsourceofthecorrelationspresented.Hence,theseheattransfercorrelationsarequitedated;andinfact,predateanybutthemostelementaryofheattransferinvestigations.Thisisapointofsignificantpotentialconcern.Since1933thefieldofheattransferhasadvancedsignificantly,andmuchimprovedcorrelationsforh'eattransferbehaviorarenowavailable.Inparticular,itisonlyintheperiodsinceattimethatextensivescientificinvestigationsofcriticalheattransferphenomenahavebeenundertakenwhichtookadvantageofsophisticatedinstrumentationandelectronicdatagatheringmethods.Infact,SNLwasunabletoobtainacopyoftheMcAdamstext,nowbeinglongoutofprint.Theuseofsuchdatedcorrelationswouldappearinappropriate.Ofparticularconcerninthisregardwouldbetheconvection,conductionandradiationtermsassociatedwiththeairgapsasgivenbyutilityequations17-19.Thesearequitecrudeestimatesofairgapehavior,andwouldnotbeconsideredgoodpracticebytoday'sstandards.Arelated,butsomewhatdifferentfindingisaparticularconcernrelatedtothemodelingofheattransferfromtheoutersurfaceofthebarriertotheambientenvironment(utilityequation21).Inthisparticularcasetherearetwopointsofconcern.First,thisexpressionactuallyderivesfroma1929work[5],andhence,mustalsobeconsideredquitedated.Second,thecorrelationappliesonlytopipes,andderivesfromstudiesofblackpipesthatrangedinsizefrom1.3to10.8inchesindiameter.(Giventhevintageofthecorrelation,thisisnotverysurprisingasmostofthethencurrentstudieswerefocussedonthefairlysimplecylindricalandsphericalgeometries.)Innocaseshouldcorrelationsbasedonheattransferfromapipebeappliedtoflatplatesurfacessuchasthosewhichwouldbeexperiencedaroundacable10 tray.ThephysicalconfigurationofapipeisquitedifferentfromthatofaflatplateIandhence,thebuoyancydrivenconvectiveairflowssurroundingapipearequitedifferentfromthosesurroundingarectangularobject.Ingeneral,theconvectivecurrentsaroundapipewouldbefarmoreefficient,andhence,convectiveheattransferisalsotypicallyfarmoreeffectiveforapipethanforaflatplateonaperunitofsurfaceareabasis(theaverageheattransfercoefficientforapipewouldbesignificantlylargerthanthatofarectangularbox).Thisshouldbeparticularlytruefordownwardfacingheatedsurfacessuchasthoseonthebottomofatray.Thisisoneareawherethetreatmentofthecabletraysusingcylindricalassumptionswouldbebothinappropriateandoptimistic(tendtooverestimatetheefficiencyofheattransfer).OverallIwasunabletoassessthepotentialimpactthatuseofmoremoderncorrelationswouldhaveontheresultsbecausenolistingoftheactualprogramhasbeengiven,becausenospecificvaluesforthevariousinputshavebeenprovidedtosupporttheexamplecalculationscitedbyCNP,andbecauseofotheruncertaintiesintheutilityanalysismethod.Asageneralpoint,theutilitymightbeabletoovercomethisshortcomingthroughathoroughvalidationofthemodelresults.Thatis,solongasthemodelyieldsconservativeresultsforafullrangeofapplications,thentheissueofusingdatedorinappropriatecorrelationsmightbeconsideredofsecondaryimportance.AswillbediscussedinSection2.2.3below,theutilityhasnotmetsuchaburdenofvalidation.Oneotherpointregardingthevariousheattransfercorrelationsishowtheutilityaccountedfortheeffectsoftheactualavailablesurfaceareaineachlayer.Thatis,theutilitytreatseachlayerasaprogressivelylargercylinder.Thisisnotanaccuraterepresentationofacabletraysystem.Inparticular,acabletrayis,inreality,alargelyone-dimensionalheattransferproblemwithheatflowingupwardsanddownwardsawayfromthecentrallylocatedcables.Hence,thesurfacearearemainsfixedasonepassesfromlayertolayer.Intheutilitymodel,thesurfaceareacontinuouslyincreasesasonemovesfromlayertolayer.Howthiswouldimpacttheutilityresultsinagainunclear.Ingeneral,theutilitytreatmentmightyieldresultswhichareinsignificanterror(eitherconservativeornonconservative).Theactualrateofheattransferfrom(convectionandradiation)orthrough(conduction)asurfaceisdirectlyproportionaltotheavailablesurfacearea.Hence,ifthesurfaceareasarenotmaintainedatthesamevalueinthemodelasthoseoftheactualphysicalsystem,thensignificanterrorscouldresult.Thiswoulddirectlytranslateintoerrorsintheampacitylimitspredictedbecausetheampacitylimitsaredirectlyrelatedtothelimitingrateofheattransfer.Theutilitytreatmentprovidesnoassurancethatthesurfaceareashavebeenappropriatelytreated.Thisisespeciallyimportantinadirectcalculationoflimitingheatloadssuchasthatimplementedbytheutility.Anotherpointofconcernrelatedtoaparticularfactor(variable)introducedintoequations16-21.Thisconcernraisesadditionalquestionsregardingtheintentandgeneralvalidityoftheutilityanalysis.Thatis,thestatedintentofthePart2analysisistocalculatethelimitingheatloadforacabletrayasasystem.However,thetreatmentappearstobebasedoncalculationofindividualcableampacitiesdirectly,andquitepossiblyusesathermalpartitioningassumptionwhichisindirectconflict viththestatedassumptionsofthePart1analysis.Forexample,considerutilityequation16:'2=0.0104pn'-)D-C(CNP16)Thisequationiscitedascharacterizing"thethermalresistancethroughrelativelythincylinders(i.e.,cablejacket,tray,firebarrier)".Thevariable(n')isstatedtobe"thenumberofconductorswithinthesection."Thiswouldappeartobearatherstrangerelationshipforexpressingthethermalresistanceofthefirebarrier,forexample.Whatpossibleimpactwouldthenumberofconductorswithinthefirebarrierhaveonthethermalresistanceofthefirebarrieritself?Thereisnodirectcontactbetweenthetwo,andhence,thefirebarriersystemshouldnotbeimpactedbythenumberofcablesinsideofit.InreviewingtheBuller-Neherpaperasomewhatd'ffaierenteinitionisgiven,buttheexactmeaningremainsunclear.Itwouldappearthatthisis,ineffect,a"thermalpartitioningfactor."Thatis,itwouldappearthatBuller-Neherusethisfactortopartitiontheoverallsystemheatloadtoindividualconductorsbasedsimplyontheconductorcount.Solongasthecablesareallidentical,suchanapproachmightbeconsideredappropriate.However,theutilityanalysisinvolvescablesofdifferenttypesandsizesinthesamecabletray.Hence,partitioningbasedonthesimpleconductorcountwouldbeinappropriate.ThiswouldalsoappeartobeindirectconflictwiththeassumptionsmadeinthePart1analysis.Thatis,inthePart1analysis,CNPhasassumedthatthethermalloadforthecabletrayasasystemcanbepartitionedtoindividualcablesbasedoneachcablescontributiontotheavailablesurfacearea.Clearly,theuseofthefactor(n')isnotconsistentwiththisstatedassumption.Howtheutilityimplementedthisfactorisentirelyunclear.Asarelatedfinding,therewasonquitepuzzlingstatementmadeduringthediscussionofutilityequation20aswell.Intheparagraphimmediatelyprecedingthatequationtheutilitystates:"Thethermalresistanceperconductorwillbethetotalnumberofconductorsdividedbythetotalthermalconductance."Theintent,basis,andimpactofthisstatementiscompletelyunclear.Hereagain,theutiityintroducestheparameter(n')asthenumberofelectricalconductors.Ineffect,theutilityseemstobepartitioningthetotalthermalresistanceassociatedwiththislayer(theairgap)bythe'numberofconductorspresenttoestimatetheresistanceperconductor.(Ineffect,thisisatreatmentinwhichtheoverallsystemthermalresistanceistreatedasasetofindividualresistanceelements,oneforeachconductor,arrangedinaparallelresistorconfiguration.Hence,thethermalresistancefortheoverallsystemisactuallylessthanthatassociatedwithasinglecable.)th'&isapparentpartitioningofthethermalresistanceisnecessaryisunclear.Howthisassumptionaccountsforcableswithdifferentphysicaldiametersisalsounclear.Finally,thisapparentpartitioningappearstobeindirectconflictwiththePart1analysisassumptioninwhichthepartitioningisassumedtobebasedonsurfacearearatios.Again,thelevelofdocumentationisinsufficienttodetermineexactlyhowthis12 particularfactorwasimplementedintheFinalutilitymodel(nolistingisprovided)andwhetherornottheproblemisself-correcting(forexample,thecomputermodelmaysimplymultiplyby(n')inonespotandthendivideby(n')somewhereelseandtherebyneutralizetheassumptionentirely).InanycasesomeadditionalexplanationofhowtheseequationsrelatetoeachotherandtothePart1analysis,andthebasis,intent,andimpactofthefactor(n')isneeded.Onefinalfactorwhichappearstobelackingintheutilitymodelisatreatmentoftheeffectsofspacingontheradiativeheattransferbehaviorofthecables.Thatis,theutilityassumesthatthefullsurfaceofeachcableisequallyeffectiveasaradiatingbodyforthedissipationofheat.Thiswouldnotbecorrect,andwouldbecomemoreincorrectasthespacingbetweencablesbecamesmaller.Thatis,ineffect,thesidesofthecable"see"theneighboringcablesandexchangeheatwiththoseneighboringcablesaswellaswiththenextlayerofthethermalsystem(theinsideofthefirebarrierforexample).Thisisnormallytreatedthroughtheapplicationofaradiationviewfactor.Thiswouldbeavaluebetween0.0and1.0whichrelatestherelativefractionoftheradiatingbodywhicheffectively"sees"theotherparticipatingsurface.Inthiscase,thisvaluewouldlikelybeontheorderof0.5-0.7dependingonthecable-to-cablespacing,andontheuniformityofthecablesizes.ThisfactorwouldInesignificantlyreduceradiantheattransferincomparisontothatassumedbthtil'ty.neffect,theutilityhasassumedanidealviewfactorof1.0,andthisiscertainlynotcorrect.2.3.3ValidationAsnotedabove,theutilitystatesthatitobtained"excellentcorrelationbetweencomputerdataandtestdata".SNI.wasunabletofindevidenceofthisintheutilitysubmittal.Thesubmittaldoesincludeadescriptionof6specificexperimentsperformedtomeasurecableoperatingconditionsunderveryspecificampacityloads.Theutilityhasalsoprovidedthefinalresultsofitsthermalanalysisofonespecificcabletrayandanundeterminednumberofconduits(possiblyasmanyas12conduitsorasfewasthreedependingonhowtheresultsareinterpreted).However,thereisnodirectcorrelationbetweentheexperimentsperformedandthecabletraysorconduitsanalyzed.Suchacomparisonmaybedifficulttodrawconsideringthenatureoftheexperimentsdocumented.Theutilitytestreportcitesthattheobjectiveoftheexperimentswas"tosimulateascloselyaspossibletheactualconditionsoftrayandconduitrunsproposedforCookPlantanddeterminethefinalconductortemperatureforthespecifiedamperageandtrayfill."Thisobjectiveisnotcompatiblewiththegoalofvalidatingananalyticalmodelwhichpurportstoestimateultimatecableampacitylimits.Validationoftheampacitylimitmodelwouldrequirethat"theutilityshowthatitsestimatedampacitylimitsconservativelyboundactualmeasuredampacitylimitsinthecorrespondingphysicalsystem.Thiswouldrequirethatatestbeperformedonaparticularphysicalsystem,andthattheampacitylimitofeachofthecablesinthatsystemwhenoperatedsimultaneouslybedeterminedexperimentally.Theseexperimentalvaluesshouldthenbecomparedtotheestimatedampa'citylimitsasa13 finalvalidation,Incontrast,theutilityhasperformedtestsinwhichtheampaciryloadswerepredetermined,andthetestsimplymeasuredtheoperatingtemperaturesofthecables.Thiswouldbeausefultestforassessingtheactualoperatingconditionsofspecificin-plantinstallations,butwouldnotbeusefulinthevalidationoftheutilitymodelbecausethereisnodeterminationoftheactualampacitylimitsforthetestedcase.ItispossiblethatthetestsoriginallydescribedinAppendixCoftheutilityanalysisevelopmentdocumentwouldbemorehelpfulinresolvingthisuncertainty.However,becauseAppendixCwasnotprovided,itisimpossibletomakethisassessment.Giventhedocumentationprovided,SNLmustconcludethattheutilityhasprovidednomeaningfulvalidationofitsanalysismethodforcalculatingcableampacitylimitsandcabletraylimitingheatrejectioncapacities.23.4SummaryofTechnicalConcernsAnumberofareasofspecificconcernwereidentifiedinthisreview.Overall,thethermalmodelfortheestimationoftotallimitingheatrejectioncapacitieswasfoundtobepoorlyfounded,andpoorlyvalidated.Itisnotrecommendedthatthismodelbeaccepteduntilthetechnicalconcernsidentifiedareresolved,andthemodelreceivesanadequatevalidationtreatment.Inparticular,thefollowingitemswereidentified:Theutilitytreatsallcabletraysystemsusingamodelbasedoncirculargeometries.Thispracticeissaidtohavebeenvalidatedbasedonthe"excellentcorrelationbetweencomputerdataandtestdata".However,noevidenceofsuchvalidationhasbeenprovided.Ingeneral,thetreatmentofaninherentlyrectangulargeometrybasedonfcylindricalcorrelationscanleadtosignificanterrors.Thisisparticularltruorconvectioncorrelations.Theconvectiveheattransferfromacylinderismoreefficientthanthatfromaflatrectangularbox.Theutilityuseofconvectionandradiationcorrelationsbasedonheattransferfrompipes(utilityequation21)isinappropriateandnonconservative.Thisutilitytreatmentofequivalentannularregionsalsoappearstoignoretheimportanceoftheavailablesurfaceareainheattransfercorrelations.Thatis,heattransferratesaredirectlyproportionaltothesurfacearea.Theutilityhasprovidednoassurancethatingeneratingitsequivalentannularregions,appropriateheattransferareasrepresentativeoftheactualphysicalsystemhavebeenmaintained.Thiscouldeasilydistortthemodelingresultsineitheraconservativeornonconservativemanner.Thisisespeciallytruesincetheutilityisattemptingtodirectlycalculateabsoluteheattransferrates,whichinturndeterminestheampacitylimitsoftheinstalledcables.Inamoregeneralcontext,giventhedocumentationprovided,SNLmustconcludethattheutilityhasprovidednomeaningfulvalidationofitsanalysismethodforcalculatingcableampacitylimitsandcabletraylimitingheat14 rejectioncapacities.Suchvalidationshouldberequiredbeforetheanalysismethodologyisaccepted.Theutilitymodelisbasedonheattransfercorrelations'whichwereorigniallypublishedinthe1929-1933timeframe.Thesecorrelationsarebadlydated,andtheimpactofusingmoremoderncorrelationsinthemodelshouldbeassessed.Theutilityhasmadeaveryconfusingstatementregardingtherelationshipbetweenthethermalresistanceofindividualcablesandthatofthesystemasawhole(seethefirstsentenceintheparagraphimmediatelyprecedingequation20oftheutilityanalysis:"Thethermalresistanceperconductorwillbethetotalnumberofconductorsdividedbythetotalthermalconductance").Thebasis,intent,andimpactofthisstatementneedstobefurtherexplainedandclarifiedparticularlyinthecontextofthestatedobjectiveofestimatingthelimitingheatloadforthecabletrayasasystemratherthanindividualcableheatloads(theapparentpurviewofthePart1analysismethod).Theutilitycorrelationfor"thethermalresistancethroughrelativelythincylinders(i.e.,cablejacket,tray,firebarrier)"(utilityequation16)includesafactor(n')describedas"thenumberofconductorswithinthesection."Asimilartreatmentisalsonotedinthecaseofutilityequation21,the"thermalresistancefromthelastsurfacetoambient."Itisunclearhowthisfactorhasbeenappliedintheutilityanalysis.Itwouldappearthatthisis,ineffect,apartitioningfactorforheattransferfromthesystemdowntoindividualconductors.Ifthisisacorrectinterpretation,thenthispartitioningisindirectconflictwiththePart1analysisinwhichheatispartitionedonthebasisofavailablesurfacearea.Theuseofthisfactorinthecontextoftheutilityanalysisisunclear,andmaybeinappropriate.Inparticular,theutilityisdealingwithsituationsinvolvingcablesofvarioussizes,types,insulationthicknessesandconductornumber.Howthisfactorisimplementedbytheutilityandtheneteffectofthisfactoronitsanalysisneedstobeclarified.Thebasis,intent,andimplementationoftheseequations,andthefactor(n')inparticular,shouldbeclarified.Theutilitymodelhasprovidednotreatmentoftheeffectsofspacingontheradiativeheattransferbehaviorofthecables.Theutilityshouldincludeconsiderationofradiationviewfactorsinthedevelopmentofitsradiativeheattransfercorrelationswhichmightsignificantlyreducethepredictedratesofradiantheattransfer.ConduitApplicationsoftheModelThedevelopmentofboththeutilityPart1andPart2thermalmodelsispresentedprimarilyinthecontextofcabletrays.However,itappearsthatthesamethermalmodelisalsobeingappliedtoconduits.Thisraisescertainuniquequestionswhichshouldbeaddressedbytheutility.15 4~,Itmustbenotedattheoutsetthatthelevelofinformationprovidedwithregardstoitsconduitapplicationswasevenmoresparsethanthatprovidedforthecabletrays.Inparticular,theutilityhasprovidednodiscussionofhowthemodelwasappliedtoconduits,andyet,exampleresultsforcablesinconduitsarepresented(seetheutilityAttachment5).Withregardtotheseexampleresults,whiletheutilityhasprovidedseveralexampleanalysesofcablesinconduits,itisunclearwhethereachcableishousedinanindividualconduit,orwhethermorethanonecablemightbehousedinacommonconduit.Giventhesizeofthecables(thelargestis1.14"indiameter),andthesizeoftheconduits(upto4"),itisquitepossiblethatmorethanonecableisinstalledinagivenconduit.First,recallthattheuseofamaintainedspacinginstallationprocedureforitscabletrayswascitedinthisreviewasacriticalfactorinthethermalmodelingofthecabletrays.Thissamefactorwouldcertainlynotapplyingeneraltoconduits.Ofparticularconcernwouldbeanyconduitwhichhousesmorethanonecable.Foraconduitwithasingleinstalledcable,theutilitythermalmodelmightbeconsideredappropriate.Thatis,inadirectcalculationofampacitylimits,themostconservativeapproachwouldbetoassumethatthecableislocatedinthecenteroftheconduitandthatthereisnocontactbetweenthecableandtheconduititself.'nthecaseofasinglecable,thefullsurfaceofthecablewouldbeavailableforheattransfer.However,inthecaseofanyconduitwithmorethanonecableinstalled,thecableswouldbearrangedinabundleofsometype,andhence,onlyafractionofthetotalcablesurfaceareawouldactuallybeactiveinthesurfaceheattransferprocesses.Theutilitymodelwould,apparently,assumethatthefullsurfaceareaofthecableswasavailablefordirectthermalexchangewiththenextlayersinthesystem(convectiontotheairgapandradiationtotheinnersurfaceoftheconduit).Thiswouldbeanonconservativeassumptionformultiplecableinstallations.ThelackofmaintainedspacingforcablesinconduitswouldalsoeffectthevalidityoftheutilityPart1partitioninganalysis.Thatis,forconduitswithmorethanonecableinstalled,thepartitioningofthetotalallowableheatlimitforthesystemtoindividualcablesbasedonthesurfaceareaofthecablescouldnotbejustifiedonthesamebasisasthatappliedtothecabletrays(maintainedspacing).This,again,isbecausetheactivesurfaceareawouldnotbeequaltothetotalsurfacearea.Infact,itisquitepossibletohaveacablefullysurroundedbyothercables,andhence,tohaveineffectnoactiveheattransfersurfaceareafordirectexchangewiththenextlayerinthesystem(convectiontotheairgapandradiativeexchangewiththeinnersurfaceoftheconduit).2Recallthatthisisincontrasttothecaseinwhichamodelisattemptingtoassessarelativederatingimpact.Inthecaseofarelativecalculation,assumingthatthecableisincontactwiththeconduitisbothmorerealisticandmoreconservativebecausethismaximizestherelativechangeduetoadditionofthefirebarrier.CNPhasnotperformedarelativecalculation,butrather,anabsolutecalculationofampacitylimits.Hence,thecentrallocationassumptionwouldbeconservative.16 Asecondfactorwhichis,again,notclearrelatestothegeneraltreatmentofsurfaceareaintheutilitymodel.Asnotedaboveinthecontextofthecabletrays,heattransferratesaredirectlyproportionaltotheavailablesurfacearea.Hence,itiscriticalthatthesurfaceareavaluesassumedinthemodelaccuratelyrepresentthephysicalsystem.Thiswouldbeespeciallytruefortheconduitsbecausetheavailablesurfaceareaisgenerallysmall,andminorincreasesintheassumeddiameterofaconduitwouldseverelyeffecttheheattransferrates.Theinputvariablesidentifiedbiieytheutilityimplythatsuchanartificialincreasemightinadvertentlyresultfromthestructureoftheutilitymodel.ThatisbecausethemodelinputsaregivenintermsoftheinitialdiameteroftheinnerlayerandthenthethicknessoftheairgapsbetweenIayers.Hence,dependingonhowthemodelwasimplemented,theselayerthicknessesmightsimplybeaccumulatedtodeterminetheequivalentdiameterofeachlayer.Becausethecomputercodehasnotbeensupplied,itisimpossibletotellhowthiswasimplementedbytheutility.Intheutilitymodelitisadirectcalculationofheattransferrateswhichdeterminesampacitylimits,andhence,errorsintheestimationofheattransferratestranslatedirectlyintoampacityerrors.2.5CoordinationoftheUtilityAnalysisPackageElementsAsdiscussedinSections2.2and2.3above,theutilityanalysismethodologyispresentedintwoparts.Thetextaccompanyingthemodeldescriptionsimpliesthatthefirstpartoftheanalysis(asdocumentedinutilityAttachment4,SectionA)providesabasisforassessingindividualcableampacitiesbasedontheoverallheatrejectioncapacityofthethermalsystemasawhole.Similarly,itisimpliedthatthesecondpartoftheanalysis(asdocumentedinutilityAttachment4,SectionB)isintendedtoprovidefortheestimationoftheoverallheatrejectioncapacityofthethermalsystemasawhole.Thatis,thePart2analysisestimatesoverallheatrejectioncapacity,andthePart1analysispartitionsthatoverallcapacitytotheindividualcablesinthesystem.However,uponreviewofthePart2analysisdescription,itisnotentirelyclearhowthetwopartsoftheanalysis,infact,worktogether.Thedevelopmentintheparttwoanalysisprovidesanexpressionfortheampacityofthecablesdirectlyasafunctionoftheenvironmentalandelectricalconditions.ThiswouldappeartomakethePart1analysismethodentirelyobsolete.ThePart2analysisalsoincorporatesaconductorcountfactor,(n'),whichappearstoactasathermalpartitioningfactor.Ifthisinterpretationiscorrect,thenthispartitioningisindirectconflictwiththeassumptionsmadeinthePart1analysis.Thatis,inthePart1analysisthermalpartitioningisassumedtobeinproportiontothecablesurfacearea.InthePart2analysis,thermalpartitioningappearstobebasedonasimpleconductorcount(eachconductorispartitionedequallyregardlessofsize).Thisisaninconsistencywhichmustberesolved.Theutilityshouldbeaskedtoclarifytheintentofeachpartofitsoverallanalysismethod,andtodescribehowthetwopartsoftheanalysisworktogether.Theexamplesprovidedshouldillustratebothaspectoftheanalysis(thePart1andPart2analyses)andshouldprovidesufficientinformationtoverifythecalculations.17 Asafinalpointofgeneralconcern,Imustremainskepticalofanypurelyanalyticalpredictionsofactualampacitylimits.Inourownwork',itwasfoundthatwhilepredictingarelativechangeinampacitylimitsduetoadditionofafirebarriersystemwasrelativelysimple,predictingactualampacitylimitsbasedondirectthermalmodelingwasmuchmoredifficult,andledtomuchgreateruncertainties.Theproblemsgenerallyarisefromtheratherlargeuncertaintiesinherentingeneralheattransfercorrelationsforsuchfactorsasconvectiveheattransfer.Thisisaparticularconcerngiventhattheutilityisbasingitsanalysisoncorrelationsof1929-1933vintage.Thismeansthatthereisasignificantinherentuncertaintyintheutilitycalculations,andthatsignificantvalidationagainstknownconditionsshouldbeprovided.Onepotentialapproachwhichmightprovefruitfulwouldbefortheutilitytocompareitsmodelingpredictionstoother,morecommon,ampacityderatingapproachesandtestresults.ThisisdiscussedfurtherinSection3inconjunctionwithareviewoftheutilityexamplesprovidedinCNPAttachment5.'SeetheresultsofUSNRCJCNJ2018,TaskOrder2.18 3.L"TILITYEXAMPLECALCL'LATIONS3.1OverviewTheutilityhasprovidedtwosetsofexampleampacitycalculations,oneforaparticularcabletray,andasecondsetassociatedwithcertainconduits.Ingenealthlevi.ngener,eForevelofdetailprovidedbytheutilityisinsufficienttofullyreviewthesecalculaticua'ons.orexample,innocasehastheutilityidentifiedthenatureofthefirebarrierinstalled(nominal1hror3hr,norinstallationcharacteristicssuchasthickness,materials,upgrades,etc.).Forthecabletrayexample,noinformationonthephysicalcharacteristicsofthecabletrayareprovided(width,height,covers,orconfigurationsuchassolidbottom,ventilatedbottom,orladdertype).Fortheconduits,itisnotpossibletodeterminewhethereachcableishousedinaseparateconduitorwhethermultiplecablesmightbelocatedinacommonconduit(groupingfactorsforconduitsareaparticularlyimportantconsideration).Asapartofthisreviewoftheutilityexampleresults,SNLhasattemptedtocomparetheutilityresultstothoseonemightobtainusingmore"conventional"approachestoampacityderating.Inparticular,amoretypicalapproachtothederatingwouldinvolveaninitialassessmentofthebaselineampacityofthecablesfrompublishedampacitytables,andthederatingofthosevaluesbasedonfactorssuchasambienttemperature,groupingandthefirebameritself.Inthisreview,SNLhasattemptedtomakesuchcomparisonsasappropriatetotheparticularexample.3.2CableTray1AZ-P8InthecaseofsparselyloadedcabletrayssuchasthoseatCNP,themoreconventionalapproachtoampacityderatingwouldbetobeginwiththebaseampacityvaluesfromthetablesusingtheapproachof"maintainedspacing"asperIPCEAP-46-426.Thenagenericestimateoftheampacitycorrectionfactor(ACF)forthefirebarriercouldbeapplied,andanominalampacitylimitfortheprotectedcablesfound.Thisvaluecouldthenbecomparedtothepredictionsoftheutilitymodelforaroughassessmentofhowwellthemodel~ouldreflectthecurrentampacitytables.Oneobvioussourceofuncertaintyinthisapproachis'thattheremaynotbegoodtestsuponwhichtoestimatetheACFofafirebarriersysteminstalledonasparselyloadedtray.Ingeneral,onemightexpectamarginallyhigherACFforasparsetraythanonewouldforaheavilyloadedtrayduetothemoreprofoundeffectonconvectiveaircurrentsforthesparselyloadedtray.However,asafirstorderapproximationthisapproachwouldcertainlyhelplendconfidenceto,orhighlightdeficienciesof,theutilityanalysismodels.SNLhasperformedsuchacomparisonfortheonecabletrayidentifiedinCNPAttachment5(TraylAZ-P8).UnfortunatelyforthistrayCNPhasgivennoinformationregardingthenatureoftheinstalledfirebarriersystem,andfromtheinformationprovideditisimpossibletodeducewhetherthisisanominallhror3hrsystem.Hence,thefollowinganalysiswillremainspeculative.Itisprovidedforillustrativepurposesonly.Theutilityhascited12,3ICcablesinthistray.Elevenoftheseare12AWGcopperconductorcables(oneofwhichisunpowered"cutintrayandtaped"),andoneisa42AWGaluminumconductorcable.Allaretype"TC"19 cables(indicatingtheserviceconditionsallowableforthiscable).Table3.1summarizestheampacityservicefactorsforthiscabletray.Table3.1:AcomparisonofCNPpredictionsandanominalanalysisbasedontabulatedampacityvalues,maintainedspacingandanominalfirebarrierACFforCNPcabletray1AZ-P8.812AWG,3/C,Cu,TC@2AWG,3/C,Al,TCOpenairampacity'aintainedspacingACF'eratedopenampacityNominalfirebarrierACF'ominalderatedampacityUtilityestimatedampacityUtilityhighestactualloadcited32A.8226.2A.68417.9A21.58A20A108A.8288.6A.68460.6A90.67A60A1.fromNECTableB310-3,19962.MaximumvaluefromIPCEAP-46-426,TableVIIforasinglelayerofcables3.BasedonnominallhrcabletrayfirebarriersystemtestedbyTexasUtilities.Thisisacrudeestimateforillustrativepurposesonly.TheactualfirebarrierconfigurationatCNPisunknown.'Notethattheresults,evenassuminganominalampacitycorrectionfactor(ACF)basedonalhrbarriersystemtestedbyTexasUtilities,showsthattheCNPestimatedampacitylimitsmaybeoverlyoptimistic,andthatcertainofthecablesmaybeoperatingatorabovetheiractualampacitylimits.Thisexampleillustratestheimportanceofpropermodelvalidation,andtheimportanceofcomparisonsofthemodelpredictionstopublishedampacitylimits.3.3TheCNPAppendixRConduitsAswasnotedabove,theutilityhasnotprovidedanydiscussionofhowitstwothermalmodelswereappliedtoconduits,andyet,resultsforcertaincablesinconduitsarepresentedasapartoftheutilitypackage.Hence,onemustconcludethatthesamethermalmodelwasusedforconduitsaswell.TheconduitresultsarealsopresentedinCNPAttachment5.Hereagaintheutilityhasprovidedonlyaminimalamountofinformationuponwhichtobasethiscomparison.Forexample,theutilityhasfailedtoidentifywhetherornoteachcableislocatedinitsownconduit,orwhethermorethanonecablemightbeinstalledinacommonconduit.Insomecases,theanswerisobvious(itisdifficulttogetmorethanonecableof0.32"diameterintoa1/2"conduit,forexample).However,inothercasesseveralcablesmightwellbelocatedinacommonconduit(a4"cablemightwellholdmorethanone1.14"diametercable).Further,thenatureofthefirebarriersinstalledontheseconduitsisalsounknown.20 Hence,asabove,thisdiscussionisforillustrativepurposesonly.Inallcases,ithasbeenassumedthatonlyasinglecableislocatedinanygivenconduit.Hence,noACFforgroupingofcablesinacon'duithasbeenapplied.Ifthisassumptionisincorrect,thentheestimatedampacitylimitsgivenherewouldbetoogenerousandwouldrequirereductionforgroupingofcables.Table3.2summarizestheresultsofthiscomparison.Table3.2:AcomparisonofCNPpredictionsandanominalanalysisbasedontabulatedampacityvaluesandanominalfirebarrierACFforCNPconduits.812AWG,3/C,Cu,TC42AWG,3/C,Al,TCNominalconduitampacity'ominalfirebarrierACF'ominalderatedampacityUtilityestimatedampacityUtilityhighestactualloadcited24.6A22.1A25.85A2.7A84.6A76.2A99.04A71.9A1.fromNECTableB310-1,1996.Includescorrectionofampacitytoambienttemperatureof40'C.2.Basedonnominal1hrconduitfirebarriersystemstestedbyTexasUtilities.Thisisacrudeestimateforillustrativepurposesonly.TheactualfirebarrierconfigurationatCNPisunknown.NotethattheampacitylimitspredictedbyCNParelargerthantheestimatedderatedampacitylimitsforthesecables,evenusingnominalvaluesforthefirebarrierACF.Infact,thederatedampacitylimitspredictedbyCNParelargerthanthenominalampacitylimitsspecifiedintheNECtableswithoutconsiderationofadditionalfirebarrierderating.ThisdiscrepancyindicatespotentialproblemsintheCNPthermalmodel,andclearlyindicatethatthemodelmaybegeneratingnonconservativeestimatesofcableampacitylimits.Thesediscrepanciesmustberesolved.Inthisparticularcase,thein-plantserviceloadsremainboundedbytheestimatedderatedampacityvalues.However,thisisbasedononlyanominalanalysis.Asnotedabove,theinclusionofcablegroupingfactors,ifsuchfactorswouldbeapplicableatCNP,orthepresenceofa3hrratedbarriersystemmightsignificantlyalterthefinalampacityestimates.ItisalsounclearwhetherornotthespecificcablescitedbyCNPareeitherall-inclusiveofconduitfirebarriersorarerepresentativeofboundingapplications.AsnotedinSection2.4above,thereisconsiderableuncertaintyregardinghowtheconduitmodelingapplicationswereimplemented.Itisinterestingtonoteherethattheutilityresultsprovideauniformvalueofampacityforagivencablesize.Providedthatallofthecablesareinstalledintheexactsameconfiguration,thiswouldbeanappropriateresult.However,itthenumberofcablesinaconduitvariesfrom21 applicationtoapplication,oriftheconduitsizevariedfromcasetocaseforthesamecable,thenoneshouldseesomedifferencesintheallowableampacitylimits.Oneshouldexpectthatanychangeinthephysicalsystemwouldbereflectedinachangeintheampacitylimitscalculated.Theutilityresultsforagivencablesizeareallidentical,andhence,onemustassumethattheinstallationsareallidentical(foragivencablesize).Thisisnotthecaseforatleastoneofthetwocablesizesconsidered.Considerthe3/C412ANGresultsfora1/2"conduit(cable8026R)incomparisontothoseforthesamesizecableina1"conduit(cable8505Rforexample).Inthiscasetheutilityhascitedtheexactsameampacitylimitdowntofoursignificantfigures(25.&5A).Thisclearlyindicatessomesortoferrorintheutilitymodel,orintheimplementationofthemodel.Giventhesamecablesintwodifferentconduitsofsignificantlydifferentsizeonewouldcertainlyexpectsignificantdifferencesinthenumericalmodelingresults.Whiletheampacitytableswouldnotdistinguishbetweenthesetwocases,thethermalmodelcertainlyshould.Thefactthatthetworesultsarelistedasidenticalindicatesthatthethermalmodelisnotproperlyaccountingforthephysicalcharacteristicsofthesystem.3.4SummaryofInsightsandFindingsAnominalcomparisonwasmadebetweentheampacityresultsprovidedbytheutilityandthosewhichmightbeobtainedusingmoreconventionalapproachestotheampacityassessment.Inboththecabletrayandconduitcasescited,itwasfoundthattheCNPestimatedampacitylimitswerenonconservativeincomparisontonominalampacitylimitsderivedfromderatingofthepublishedcableampacitytables.Inthecaseoftheconduits,theutilityestimatedampacitylimitsincludingderatingforthe'firebarriersystemwereinexcessofthetabulatedampacitylimitsforcablesinconduitswithoutafirebarrierassetforthintheNECtables.TheseresultsindicatepotentialproblemsintheCNPthermalmodel.Thepredictionofactualcableampacitylimitsbasedondirectthermalmodelingisquitedifficult,andwouldbeexpectedtoholdconsiderableuncertainty.TheresultsofthecomparisonsmadehereindicatethattheCNPthermalmodelmaywellbegeneratingunrealisticandnonconservativeestimatesofactualcableampacitylimits.Additionalvalidationoftheutilitythermalmodelisneeded.Asapartofthevalidationprocess,theutilityshouldprovideadirectcomparisonofitsownmodelingresultstotheresultsobtainedusingmoreconventionalampacityderatingapproaches,and/ortoactualtestresultsinwhichampacitylimitsweremeasureddirectly.Thisshouldincludebothcabletrayandconduitapplicationsifthemodelistobeappliedtobothtypesofinstallations.TheSNLcomparisonsmustbeviewedinthecontextofillustrativeexamplesonly,dueprimarilytothefactthatinsufficientinformationhasbeenprovidedbytheutilityuponwhichtobasemoredefinitiveanalyses.Initssubmittal,theutilityshouldalsoprovideasufficientbaseofinformationonitsparticularapplicationstoallowforacompletereviewandassessmentoftheresults.Thismustincludemoredetaileddescriptionsofthephysicalcharacteristicsofeachsystem,andthecharacteristicsoftheinstalledfirebarriersystem.22 Theexaminationoftheconduitresultsalsoraisedaparticularpointofconcernwhichindicatesthatthereareerrorseitherinthethermalmodelorintheutilityimplementationofthatmodel.Inthecaseofthe3/C12AWGwires,theutilitypredictedthesameampacitylimitsforacableina1/2"conduitandforthesamecableina1"conduit.Whiletheampacitytableswouldnotdistinguishbetweenthesetwocases,thethermalmodelcertainlyshould.Thisisaclearindicationofanerrorofsometype.Basedontheinformationprovided,itisimpossibletoidentifythesourceofthiserror.23 4.SL'iiMARYOFFINDINGSANDRECOMMEM3ATIQNSWithrespecttotheadequacyoftheoverallutilitydocumentation,SNLfindsthat:Thelevelofdocumentationprovidedisnotadequatetocompletefullevaluationoftheutilityampacityassessments.Specificareasinwhichfurtherdocumentationisrequiredaredocumentedbelow.Ingeneral,thereisnodiscussionofhowthetwopartsoftheutilityanalysismethodologyaremadetoworktogether,theexamplecalculationsdonotprovideenoughinformationtoverifythe'calculations,andtheexperimentspurportedtosupportvalidationofthethermalmodelsareeithernotprovided,ornodirectone-to-onecomparisonoftheexperimentstomodelingresultsisprovided.Whiletheutilityhasdocumentedtheresultsofitsanalysisforonecabletrayandanindeterminatenumberofconduits,nosummaryofthebalanceoftheplantresultshasbeenprovided.Theutilityshouldprovideasummaryoftheampacityassessmentresultsforitsinstalledfirebarriersystems.Nodiscussionhasbeenprovidedastohowin-plantcableserviceloadsweredetermined,whichcableshavebeenconsideredintheanalysis,andthebasisfortheeliminationofothercablesfromconsideration.Thisinformationisneededtoassesstheadequacyoftheutilitytreatment.Withrespecttotheutilityampacity"partitioning"analysismethodology(referredtointhisreviewasthePartIanalysis)SNLfindsthat:Theutilityhasprovidedaninadequatebasisforvalidationofitsassumptionthattheoverallheatrejectioncapacityofasparselyloadedcabletraycanbepartitionedtoindividualcablesinproportiontothecablediameter.Thelimitedinformationprovidedbytheutility(oneplotwithnosupportingdataandnoindicatedunits)isunconvincing.Theutilityhascitedasetofexperimentsasthebasisforthisplot,andhence,forthevalidationofthismethodology(seereferencetoAppendixCintheutilityAttachment4).However,nodocumentationoftheseexperimentshasbeenprovided.Documentationofthevalidationexperimentscitedinthisportionoftheutilityanalysisisneeded.Thisshouldincludeadiscussionoftheutilityanalysisandapplicationofthetestdata.Withrespecttotheutilitythermalheatrejectioncapacitycalculationmethod(referredtointhisreviewasthePart2analysis):Theutilitytreatsallcabletraysystemsusingamodelbasedoncirculargeometries.Thispracticeissaidtohavebeenvalidatedbaseonthe"excellentcorrelationbetweencomputerdataandtestdata".However,noevidenceofsuchvalidationhasbeenprovided.Theutilitymustvalidateitsassumptionsinthisregard.24 Theutilityassumptionofequivalentannularregionsappearstogiveinadequatetreatmenttotheimportanceofsurfaceareainheattransfercalculations.Theactualratesofheattransferaredirectlyproportionaltosurfacearea,hence,itisimportantthatthethermalmodeluseactualavailablesurfaceareasinitsformulation.Theutilityshouldprovideexamplestoillustratetheeffectiveheattransferareasassumedforeachofthelayersinitsmodelingandcomparethoseassumedareastotheactualheattransferareasavailableinthephysicalsystem.Theseexamplesshouldcoverbothconduitsandcabletrays.(Forcabletraysitisrecommendedthat,consistentwithothermodelingefforts,theutilityshouldassumethatonlytheupperandlowersurfacesofthetrayandfirebarrierareactiveintheheattransferprocess.Bothexperimentsanddetailedmodeling(Stolpe,e.g.)haveshownthatthesidesofacabletrayarerelativelyunimportantintheoverallheattransferprocess.)SNLfindsthattheutilityhasprovidednomeaningfulvalidationofitsanalysismethodforcalculatingcableampacitylimitsandcabletraylimitingheatrejectioncapacities.Theutilityshouldprovideforthedirectcomparisonofpredictedcableampacitylimitstothosemeasuredinexperimentsonthecorrespondingsysteminordertovalidateitscalculations.Theutilitythermalmodelisbasedonheattransfercorrelationswhichwereoriginallypublishedinthe1929-1933timeframe.Thesecorrelationsarebadlydated,andtheimpactofusingmoremoderncorrelationsinthemodelshouldbeassessed.(Thisissuemightbeconsideredofsecondaryimportanceprovidedthatasufficientbaseofvalidationwereprovided.Suchavalidationbasehasnotbeenprovidedaspertheprecedingfinding.)Thesubmittalstates,inthedevelopmentofcorrelationforheattransferfromthecablestothesurroundingairgap,that"thethermalresistanceperconductorwillbethetotalnumberofconductorsdividedbythetotalthermalconductance"(seethefirstsentenceintheparagraphimmediatelyprecedingequation20oftheutilityanalysis).Thebasis,intent,andimpactofthisstatementneedstobefurtherexplainedandclarified.Directlyrelatedtotheprecedingcomment,theutilitycorrelationsfor"thethermalresistancethroughrelativelythincylinders(i.e.,cablejacket,tray,firebarrier)"(utilityequation16)andforthe"thermalresistancefromthelastsurfacetoambient"(utilityequation21)eachincludeafactor(n')describedas"thenumberofconductorswithinthesection."Thebasis,intend,andimplementationofthisfactor(n')shouldbeclarifiedasitisappliedtoeachofthesetwoequations.Thiswouldappeartobe,ineffect,athermalpartitioningfactorwhichisbasedonasimpleconductorcount,andassuch,maybeindirectconflictwiththestatedassumptionsofthemodeldescribedin"AppendixA"ofutilityAttachment4(whereitisassumedthatsurfaceareawillbethebasisforthermalpartitioning).Thebasis,intent,andimpactofthisfactormustbeclarifiedandjustified.Also,theutilitymustexplainhowthisfactorapplieswhencablesofdifferentphysicaldimensionsarepresentinacommoncabletray.25 Thecorrelationcitedforthethermalresistancebetweentheoutersurfaceofthebarrierandtheambientenvironment(utilityequation21)appliesonlytopipes.Thesecorrelationsshouldnotbeappliedtoflatplatesurfacessuchasthosewhichwouldbeexperiencedaroundacabletray.Theutilitymodelprovidesnotreatmentoftheeffectsofspacingontheradiativeheattransferbehaviorofthecables.Reducedviewfactorsduetocableproximitymightsignificantlyreducethepredictedratesofradiantheattransfer.Thisshouldbeaccountedforintheradiativeheattransfercorrelations.Itisnotpossibletoverifytheutilitycalculationsbecausenoinformationonthespecificinputsusedinanyexamplecasehavebeenprovided,andthelistingoftheutilitycomputercodewasnotprovided.Theutilityhasapparentlyappliedoneorbothofitsthermalmodelstotheanalysisofcablesinconduitsaswellasthoseintrays.Theapplicabilityofbothpartsoftheutilityanalysismodeltoconduitsneedstobeaddressed.Inparticular,theutilitymodelforcabletraysisbasedontheuniqueconfigurationof"maintainedspacing"foritspowercablesincabletrays.Thissamefactorcannotgenerallybeassumedtoexistforcablesinconduits.Theutilityshouldprovideanexplicitdiscussionoftheapplicabilityofitstwothermalmodels(the"AppendixA"surfaceareabasedheatloadpartitioningmodelandthe"Appendix8"directthermalanalysismodel)toconduits,andshoulddiscusshowthemodelswereimplementedforconduitanalyses.WithregardtothespecificexamplecalculationsprovidedinthesubmittalSNLfindsthat:AnominalcomparisonbetweentheutilityampacitypredictionsandthoseobtainedusingmoreconventionalapproachestotheampacityassessmentshowedthattheCNPestimatedampacitylimitswerenonconservative.Inthecaseoftheconduits,theutilityestimatedampacitylimitsincludingderatingforthefirebarriersystemwereinexcessofthetabulatedampacitylimitsforcablesinconduitswithoutafirebarrierassetforthintheNECtables,evengiventhemostgenerousinterpretationoftheconduitloadings(onlyasinglecableperconduitwithnomorethanthreeconductors).TheseresultsindicatethattheCNPthermalmodelmaybegeneratingunrealisticandnonconservativeestimatesofactualcableampacitylimits,ThesediscrepanciesmustberesolvedbyCNP.Asapartofthevalidationprocess,theutilityshouldprovideadirectcomparisonofitsownmodelingresultstotheresultsobtainedusingmoreconventionalampacityderatingapproaches.Initssubmittal,theutilityshouldalsoprovideasufficientbaseofinformationonitsparticularapplicationstoallowforacompletereviewandassessmentoftheresults.Thismustincludemoredetaileddescriptionsofthe26 physicalcharacteristicsofeachsystem,andthecharacteristicsoftheinstalledfirebarriersystem.Theexampleresultsforconduitspredictedtheexactsameampacitylimitsdowntofoursignificantfigures(2S.8SA)fora3/C412AWGcableinbotha1/2"conduit(cable8026R)anda1"conduit(cable8505Rforexample).Whiletheampacitytableswouldnotdistinguishbetweenthesetwocases,giventhedifferencesinthephysicalconfigurations,thethermalmodelcertainlyshould.Thisisaclearindicationofanerroreitherinthemodelorintheimplementationofthemodel.CNPshouldidentifythesourceof,andresolve,thisdiscrepancy.ItisrecommendedthatanRAItotheutilitybepreparedtoclarifythesepoints.Ingeneral,itislikelythatsignificantadditionalconsiderationwillberequiredontheartPoftheutilitytoresolvetheconcernsidentifiedinthisreview.Significantlymorecompletedocumentationoftheutilitymodels,theexamplecalculations,andasummaryoftheoverallanalysisresultswillalsobeneededbeforeafinalassessmentoftheutilityampacityloadfactorscanbemade.27

5.0REFERENCES

1.Stolpe,J.,"AmpacitiesforCablesinRandomlyFilledTrays,"IEEETransactionsonPowerApparatusandSystems,Uol.PAS-90,Pt.I,PP962-974,1971.2.J.H.NeherandM.H.McGrath,"TheCalculationsoftheTemperatureRiseandLoadCapacityofCableSystems",AIEETransactionspp752-772Oct.19573.F.H.Buller,J.H.Neher,"TheThermalResistancBetweenCablesandaSurroundingPipeorDuctWall,"AIEETransactions,V69,pp342-349,1950.4.W.H.McAdams,HeatTransmission,McGraw-Hill,1933.S.R.H.Heilman,"SurfaceHeatTransfer,"ASMETransactions,U51,pp287-302,1929.28