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'uclear PlantMethodology fortheAnalysisofFireBarrierAmpacityDeratingFactorsALetterReporttotheUSNRCRevision03une28,1996Preparedby:SteveNowlenSandiaNationalLaboratories Albuquerque, NewMexico87185-0737 (505)845-9850Preparedfor:RonaldoJenkinsElectrical Engineering BranchOfficeofNuclearReactorRegulation U.S.NuclearRegulatory Commission Washington, DC20555Attachment totheEnclosure qbl206035i 9bl202P.----,-PDRPDRADGCK05000315 TABLEOFCONTENTS:

SectionPa~eFORWARD)n1OVERVIEW...

1.1Objective 1.2ReportOrganization UTILITYAMPACITYDERATINGAPPROACH2.1OverviewofAnalysisApproach...

2.2TheUtilityPart1AnalysisMethod2.2.1BasisandUnderlying Assumptions 2.2.2Theoretical Development 2.2.3Validation.........

~2.2.4SummaryofTechnical Concerns....

2.3TheUtilityPart2AnalysisMethod2.3.1BasisandUnderlying Assumptions 2.3.2Theoretical Development 2.3.3Validation...

2.3.4SummaryofTechnical Concerns.2.4ConduitApplications oftheModel.....2.5Coordination oftheUtilityAnalysisPackageElements..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 TheUnitedStatesNuclearRegulatory Commission (USNRC)hassolicited thesupportofSandiaNationalLaboratories (SNL)inthereviewofutilitysubmittals associated withfireprotection andelectrical engineering.

Thisletterreportdocuments theresultsofaSNLreviewofasubmittal fromtheDonaldC.CookNuclearPlant(CNP).Thissubmittal dealswiththeassessment ofampacityloadsforfirebarrierprotected cables.Thisdocumentwassubmitted bytheutilityinresponsetoUSNRCGener'tt92-iceeraninresponsetoasubsequent USNRCRequestforAdditional Information (RAI).Thisworkwasperformed asTaskOrder9,Subtask4ofUSNRCJCNJ20l7.

1,0OX'ERVIE1.1Objective InresponsetoUSNRCGenericLetter92-08,andasubsequent USNRCRequestforAdditional Information (RAI)theDonaldC.CookNuclearPlant(CNP)provideddocumentation ofamethodology fortheassessment ofitscableampacityloadingfactorsincluding theeffectsoffirebarrierampacityderatingimpacts.Th'als.issumittcabletraewosetsofspecificcaseexamples, oneforaseriesofconduitsandfoneoraeray,toillustrate themethodology.

Theutilityhasalsoincludedtstoincueatestreportwiciscitedassupporting thevalidityoftheutilitycalculation method.Thesubmittal reviewedwasdocumented inautilityletter:Letter,May12,1995,(docketnos.50-315,50-316itemAEP:NRC:0692DF),

E.E.Fitzpatrick, IndianaMichiganPowerCo.,totheUSNRCDocumentControlDeskincluding sixattachments asfollows:-Attachment 1:SummaryofAmpacityDeratingAnalyses-Attachment 2:AIEETransactions Papers57-660&50-52-Attachment 3:CableTrayAllowable FillDesignStandard-Attachment 4:AnalysesandMathematical Models-Attachment 5:Representative AmpacityDeratingCalculation Results-Attachment 6:ResultsfromTestReportPCL-542SNLwNLwasrequested toreviewthissubmittal underthetermsofthegeneraltechnical supportcontractJCNJ-2017,TaskOrder9,Subtask4.Thisletterreportdocuments theinitialresultsofthisreview.1.2ReportOrganization Section2providesareviewofthetechnical aspectsoftheutilityanalytical approachtoampacityassessments.

Inparticular, twoseparateanalytical modelsaredeveloped yeutility,andeachisreviewedindependently.

Inaddition, consideration isgiventothevalidation, orlackthereof,ofeachofthetwomodels.Section3providesareviewoftheexamplecalculations providedbyCNP.Thisincludescalculations foronecabletrayandforanundetermined numberofconduits.

SNL'sreviewhasincludedacomparison oftheutilityresultstomoreconventional approaches involving ecratingoftabulated ampacitylimits.Section4providesasummaryofthemajorissuesandconcernsidentified inthisreview.Section5identified thereferenced documents.

2.0UTILITYA!vG'ACITY DERATIiGAPPROACH2.1OverviewofAnalysisApproachTheapproachtakenbyCNPisbasedonalargelyanalytical assessment ofactualcableampacitylimitswithlimitedexperimental validation oftheoverallanalysismethod.TheandysisapproachtakenbyCNPisquiteuniqueincertainrespects.

Inparticular, theutilityhascitedthatitscableinstallation procedures forpowercablesincabletraysrequiredtheuseof,ineffect,amaintained spacingapproach.

Thatis,powercablesincabletraysatCNPwereallinstalled soastoconformtothefollowing features:

nomorethanasinglelayerofpowercablesisinstalled inanytray,eachcableinthetraytobeseparated fromitsneighbors witha'gapofnolessthanI/3ofthe.diameter ofthelargerofthetwocables,thesumoftheinstalled cableouterdiameters shallnotexceed75/oofthefulltraywidth.Thisisanimportant observation.

Thebulkoftheutilitymodeldevelopment isaimedataddressing sparselyloadedcabletrays,althoughthefinalresultsareapparently appliedtoconduitsaswell.Theutilityanalysesareactuallyperformed intwoparts.Theexactintentofeachstepremainssomewhatunclear,andmayhavebeenmisinterpreted inthisreview(apointrequiring clarification).

Theutilitydiscussion impliesthefollowing interpretation:

PartIAnalysis:

Givenanoverallheatrejection capacityforthecabletray,calculate theallowable ampacitylimitforindividual cables.Themethodusedinthispartoftheanalysisisdocumented inAppendixAoftheutilityAttachment 4.Part2Analysis:

Calculate theoverallheatrejection capacityforagivencabletray(orconduit)basedonheattransfercorrelations andcalculations.

Themethodusedinthispartoftheanalysisisdocumented inAppendixBoftheutilityAttachment 4.However,theutilitydescription ofthePart2analysisstepappearstoincludebothaspectsoftheproblemtosomeextent.Inparticular, the*Part2theoretical development presentsequations forthecalculation ofampacitylimitsnotfrthcducculationoftotalheatrejection capacity.

Infact,themethodappearstobeintendedtoassesstheampacitylimitsofindividual cables.Hence,itremainsunclearastotheintentoftheAppendixAandAppendixBmethodsandhoweachworkswithththtoertoprovideacompletesolution.

Itisalsounclearhowthemethodshavebeenappliedtoconduits.

Thatis,theutilityprovidesexamplesinvolving

conduits, butthePartIpartitioning methodwouldclearlynotapplytoconduitapplications.

Hence,howtheseconduitswereanalyzedremainsunclear.

Eachofthetwoparisoftheoverallutilityanalysisapproachwillbereviewedindetailinthefollowing subsections.

Thesediscussions willincludeareasofpotential concernidentified duringthereviewandtheacceptability oftheutilityvalidation arguments.

However,overall,theutilityshouldprovideforsomediscussion ofhowthetwopansoftheanalysisworktogetherinpractice, andshouldprovidemoredetailedexampleswhichil!ustratehowthetwopartsoftheanalysiswereusedindividually toobtainactualcableampacitylimitestimates.

NotethatSection3ofthisreportwilldiscussthisfinalpoint,theexampleanalyses, inmoredetail.2.2TheUtilityPart1AnalysisMethod2.2.1BasisandUnderlying Assumptions Asnotedabove,itisourinterpretation thattheintentofmethodology described inSectionAofutilityAttachment 4(whichwillbereferredtohereasthePart1analysisforconvenience) istodetermine theampacityofindividual powercablesgivenanestimateoftheoverallheatrejection capacityofagivencabletray.(Notethattheheatrejection capacityiscalculated separately asdiscussed inSection2.3below.)Theutilitymethodology, ineffect,partitions theoverallheatrejection capacitytoindividual cables.Thispartitioning isultimately basedonaratiooftheavailable thehetsurfaceareaofagivencabletothetotalsurfaceareaforallcablesinthtr0sineay.nceeeatrejection capacityforthegivencableisestablished, thecorresponding ampacitylimitiscalculated usingasimple"I'R"typecalculation.

Thecriticalunderlying assumption inthisanalysisderivesdirectlyfromtheutilitycableinstallation practicewhich,asdiscussed inSection1above,requiredthatonlyasinglelayerofcablebeinstalled, andthatthosecablesbeinstalled with,ineffect,amaintained spacingbetweencables.Basedonthispractice, theutilityconcludes thattheheattransferbehaviorofitscabletraysisdominated byconvection andradiation, ratherthanbyconduction.

Thisisanimportant distinction becauseconvection andradiation arebothdrivenprimarily bytheavailable heattransfersurfacearea,inthiscase,theactualsurfaceareaoftheinstalled cables.Incontrast, forgeneralcabletrays,suchasthosetestedbyStolpefl]inhispioneering cabletrayampacitywork,heattransferwithinthecablemassisdominated byconduction.

This,CNPnotes,isreflected inthefactthatopentopcabletrayampacitybtablesinthattheampacitylimitsarecitedasafunctionofthecabledepthofflldasedonthelimitingheatgeneration rateperunitvolumeofthecablemassforagivencablesize.Insomesenses,theutilityappearstotakethisanalogyalittlefatherthanitshouldbetaken,althoughthisisnotconsidered tobedirectlyrelevanttotheultimatevalidityoftheutilitymethodology.'hat is,CNPcitesthecomparison madebyStolpebetweenthevolumeoccupiedbyseven412AWGcablesasessentially equaltothatofasingle4l0cable.Theutilitythenstatesthat"thustheheatgenerated bythetwoconfigurations shouldbeequivalent underuniformheatdistribution conditions."

However,thisidealismdidnotholdtrueforStolpe'sfinalresults.AreviewoftheICEAP-54<40ampacitytableswillreveal

",Z.ZThoreticalDevelopment ForthesparselyloadedtraysatCNPconvection andradiation areconsidered thegoverning modesofheattransfer.

Thisassumption isconsidered toaccurately reflectthesparselyloadedcabletraysatCNP(withonepossibleexception aswillbediscussed inSection2.2.4below).Basedonthisassumption, theutilitypostulates thatindividual cableampacitylimitscanbeestablished basedonconsideration oftheheatgeneration rateperunitofcablesurfacearea.Itisthisassumption whichdrivestheutilitymethod.Theutilitydevelopment ofitsthermal modelispresented inSectionsA.3andA.4ofAttachment 4toitssubmittal.

Thedevelopment firstbeginsbyconsidering acaseinwhichallthecablesareassumedtobeofthesamesize.Thetotalavailable surfaceareaofthecablesinthetrayisthengivenby:Where(A,)isthecabletotalsurfacearea,(n)isthenumberofcables,and(d)istheouterdiameterofthecable.Theutilitythendefinesthe"percentage fillofthetray"(F)as:F=nd(2)Theutilityacknowledges thatthisdefinition ofpercentage fillisnotconsistent withgeneralindustrypractice.

SNLfurthernotesthattodescribethisvalueasapercentage fillisnotaccurate.

Thequantity(n~d)isalineardimension whichquantifies thetotalwidthoftheinstalled cables.Hence,(F)willhaveunitswhicharethesameasthoseusedtoexpress(d),inchesforexample.Thisisbynomeansa"percentage fill"termA"becausea"percentage" termisaratioofonequantitytoanotherofequaldimens'o.percentage" termshouldbedimensionless.

Itwouldbemoreaccuratetosimplycallthisvaluethe"totalfill"ratherthanthe"percentage fill."Thisisaminorpointofnomenclature, andsolongasthevalueisusedconsistently, shouldintroduce noerrors.Theutilitythengoesontoexpressthetotalheatgeneration rate(Q)foratrayasawholebasedonsimple"I'R"heating:Q=3nZ~R~~(3)where(I)isthecablecurrentand(R)istheACresistance ofthecableofinterest.

Notethattheconstant(3)appearsbecausetheutilityassumesthateachofthecablesisathree-conductor cable.Combining equations (2)and(4),theutilityeliminates thevariable(n):thatthelimitingheatgenerating capacityfora4/0cableisontheorderof12-18timesthatofa12AWGcable,not7timesasthisidealization wouldsuggest.Thisisonlyaveryminorpointandhasnorealimpactontheutilitymethodology.

CNPthenrearranges thisexpression tosolveforthecurrent(I)asafunctionofheatrejection capacity(Q).Whilethissameequationiswritteninatleastthreeformatsthroughthedevelopment, thefinalexpression ispresumably utilityequation9asfollows:gd3FRacorbyrearranging thetermsinthisexpression:

(6)This,then,istheultimatemethodofpartitioning thetotalheatrejection capacityofacabletrayasawholedowntoindividual cables.Notethatasshownin(6),thegrouping([d/R]'isafunctionoftheindividual cable;thediameterdividedbytheACresistance.

Forlargercables,'the diameterincreases andtheACresistance decreases, sothevalueofthisgroupingincreases withcablesize.Thegrouping([Q/3F]'")

isafunctionofthetray;thelimitingtrayheatloaddividedbythetotalcablefill.2.2.3Ualidation Theutilityvalidation ofthisthermalmodelispresented inFigureA-1oftheutilityAttachment 4.Thisfigureplotscurrent(I)versusthe([d/R]'cablecharacteristic grouping.

Consistent withequation(6),theutilityanticipates alinearrelationship passingthroughtheorigin.Thisis,infact,demonstrated forthreespecificcabletrayconfigurations, eachinvolving a67%fillofthetray.Theconfigurations involveaventilated traywithaventilated cover,asolidtraywithasolidcover,andaventilated traywitha1hrbarriersystem.Ineachcasethelinearrelationship holdstrue.Whilethisplotdoesprovidesomeencouraging results,itisnotsufficient tofullyvalidatetheutilityPart1analysismethod.Themostimportant factorwhichhasnotbeenestablished bytheutilityisrelatedtotheheatloads(Q)whichweremeasuredintheseexperiments.

Thatis,theutilitymethodassumesthatforagivencabletray,thetotallimitingheatrejection capacityisafunctionofthetotalavailable cablesurfacearea,whichisinturnafunctionoftheloadingfactor(F}.Hence,inordertovalidatethisassumption, theutilitymustdemonstrate thatinitsexperiments, foranyonecombination oftrayconfiguration (e.g.foranopentray)loadingfactor(F)andtemperature difference (T>>-T,),thesameoverallheatloadwasmeasuredindependent ofthesizeofcablestested.Forexample,alltrayswereapparently loadedto67%oftheavailable width.Wemustassumethatcableswerecurrent-loaded toestablish ahot-spot temperature of90'C(anassumption whichneedstobe'fdbCbl'veriieyi'.Tovalidateitsmodel,ChPshouldshowthatthetotalresistance heatingloadmeasuredforthevariouscablesizesremainedconstantforeachofthetrayconfigurations testedregardless ofthecablesizeinstalled.

Withoutadditional information regarding thenatureofthetestscitedbyCNP,thisassessment cannotbemade.Inordertocompletethisassessment theutilitymustprovideadditional information onitstestprotocolandresults.Forexample,providing acopyofthemissingAppendixCmayprovesufficient.

Attheleast,theutilityshouldensurethatthefollowinpointsareaddressed:

aeoowingThephysicaldescription

'oftheexperimental setupandtestarticlesshouldbeprovided.

Thisshouldincludeaphysicaldescription ofthetestarticlesthemselves (traywidths,heightandlength),aphysicaldescription ofecalestested(sizeandconductor count),adescription oftestinstrumentation, andadescription oftheinstallation ofcablesinthetestarticle(types,spacingandlocations).

Adiscussion ofthetestingprotocolisneeded.Thisshouldi1dthadiscus'ionofthespecificobjective ofthetest.Forexample,werecableissouincueecurrentsadjustedtoachieveadesiredtemperature state?Whatwasthedesiredfinal.end stateoftheexperiments in'termsoftemperature andcurrent?Wereifferentsizedcablesinstalled inthesametestarticle,andifsohowwerethecurrentsforthesedifferent sizedcablesdetermined?

Howdidthetestsverifythatasteadystatecondition wasachieved?

Asummaryofthetestresultsinneeded.Thisshouldincludethefinalmeasuredvaluesofcurrentandtemperature foreachofthecablesinanygiventestarticle.Theutilityshoulddescribetheunitsassociated withitsFigureA-1anddescribehowtestedcableparameters correspond totheplottedvalues.Thecurrent(I)showninFigureA-1isassumedtobegiveninunitsofAmps,butthiswouldimplythatmostlyverylargecableswereusedintesting.Themeasuredcurrentvaluesrangeashighas700Awhichshouldgenerally implyacableofontheorderof750kcmil,whichappearstobemuchlargerthanthecablesconsidered elsewhere (thelargestcableidentified elsewhere inthesubmittal isaP4i0cable).Thephysicalcablesizestestedshouldbedescribed, andtheapplicability ofthetestedcablestothecablesconsidered elsewhere intheanalysisshouldbe'established.

2.2.4SummaryofTechnical ConcernsTheprimarytechnical concernregarding theutilityPart1analysismethodwhichmustbeaddressed isoneofvalidation.

Theutilityanalysisisbasedontwofundamental assumptions:

0Jl~,~

Thetotallimitingheatloadforagivencabletravisafunctionofthetotalsurfaceareaofthecables,and-Thistotalheatloadcanbepartitioned toindividual cablesbasedoneachcablescontribution totheoverallavailable cablesurfacearea.CNPhasnotprovidedsufficient information todetermine thatithas,infact,validated tesetwocriticalassumptions.

Additional detailregarding theexperiments itperformed tovalidatethisaspectofthethermalanalysismodelsisrequired.

Section2.2.3aboveprovidesadescription oftheadditional information required.

'Thereisoneparticular factorwhichmaybeinadequately treatedintheutilityanalysis.

Thatis,thecabletemperature ratingof90'Cisbasedontheconductor temperature, notthesurfacetemperature ofthecable.Iwouldagreethatoncetheheatreachesthecablesurface,convection andradiation becomepredominant.

However,withinthecableitself,conduction ofheatthroughtheinsulation andjacketmaterials remainsacriticalfactor.WhiletheutilityPart2analysisincludesathermalresistance elementassociated withtheinsulation, theconcernhereiswiththePartlpartitioni alp'oninganysisinwhichincreasing diameteriscreditedforincreasing ampacityinproportion tothecablediameter.

Thiswouldbecorrectifthesurfacetemperature ofthecableremainedfixed(forexampleat90'C)butwillnotholdtrueifthecablesurfacetemperature changeswithdiameter.

Infact,becausetheelectrical insulation isalsoathermalinsulation, thethickerthatinsulation becomes,thehigherthetemperature dropthroughtheinsulation willbeforagivenoverallrateofheatflow,andhence,thelowerwillbethesurfacetemperature ofthecable.Thiswouldcausesomeerrorintheutilitymethod,althoughbasedontheinformation

provided, itisimpossible toassessthemagnitude oftheerrorintroduced.

Thepotential pitfallcanbeillustrated asfollows.Fortwocablesofthesamewiregage,thethickness oftheinsulation woulddependprimarily onthevoltageratingofthecable.Thus,a600Vcablewouldhaveasmalleroveralldiameterthanwoulda5000Vcable.Intheutilitymethod,thiswouldimplythatifthesetwocableswereinthesametray,theallowable ampacityofthe5000Vcablewouldbepredicted ashigherthanthatofthe600Vcableinaccordance withtheratiooftheirdiameters.

Inreality,theincreased thickness oftheinsulation forthe5000Vcablewouldgenerally resultinareducedampacityforthemore"open"(maintained spacing)application suchasthoseatCNP.(Fortightlypackedcabletrays,asperStolpe,thevolumeistheoverriding factorandampacitydoesincreasewithcablevoltagerating.However,moreopenconditions allowmoregenerousampacitylimits,andcablevolumeisnottheoverriding factorintheseratings.)

Thisisparticularly trueforthelargercablesizes.Note,forexample,thattheopenairampacitytablesinIPCEAP-46-426willgenerally decreasetheallowable ampacitylimitsforsingleconductor cableswithincreasing voltageratingforcableslargerthanabout1/OAWG(thetransition pointdependsonthetableconsidered andthetemperature ratingofthecable).Theutilitymethodwouldallowjusttheoppositeeffect.Theutilityshouldprovideforsomeassessment ofthisquestion.

Oneapproachwhichmightprovefruitfulwouldbefortheutilitytoexaminetheopenairampacitytables forthetypesofcablesinuseatC4P,andforCbiPioattmpttovalidatetheirassumptions regarding ampacityasafunctionofcablediameter.

Forexample,withthethreeconductor cables(whichseemtobequitecommonatCNP)thetransition pointinthetablesisatmuchlargercablesizes.Alternatively, theutilitytestsmayprovidesomefurtherinformation onthisquestionforthespecificcablesinuseatCNP.Theinformation providedinthesubmittal isinsufficient todetermine whetherornotsuchacomparison canbemade.Inanycase,theutilitymustdemonstrate thatitsdiameter-based partitioning approachappropriately reflectstheseampacitylimitbehaviors withincreasing voltageratingandincreasing cablediameter.

Thatis,theutilityshouldshowthat,infact,foritscablesampacitylimitsincreasewithcableoveralldiameter(voltagerating),andthatthisincreaseisconsistent withtheirdiameterscalingassumptions.

Theutilitymightalsoconsultmanufacturer recommended ampacitylimitsofcablesincomparison tocablediameters.

2.3TheUtilityPart2Analysis2.3.1BasisandUnderlying Assumptions SectionBoftheCNPAttachment 4documents asecond,separatebutrelated,thermalanalysismethod(thismethodwillbereferredtohereastheutility"Part2"analyses).

ThismethodisbasedlooselyontheNeher-McGrath approachtocablethermalanalysis, althoughconsiderable liberties aretakenwiththatmethodology.

Theintentofthemethodis,apparently, toprovideestimates ofthetotallimitingheatrejection capacityofagivencabletraybasedontheconfiguration ofthetray,thenumberofcablesinstalled inthetray,andthenatureoftheprotective features(barrier) installed onthetray.Inexecuting thiscalculation theutilityisinherently assumingthatitcanaccurately modelthephysicalphenomena ofinterestandtherebyestimatetheactualcableheatrejection capacityofagivencabletrayorconduitsystem.Thisoverallsystemheatrejection capacityisthenpartitioned toindividual cablesinthetraythroughthePart1analysismethodasdiscussed in2.2above.2.3.2Theoretical Development Thebasicstructure ofthePart2thermalmodelderivesdirectlyfromNeher-McGrath.

Thegeneralexpression isquitesimpleandstraightforward.Theapproachisbasedonforcingasteadystatebalancebetweentherateofheatgeneration inthecablesandtherateofheatdissipation fromthecablestotheambient.Therateofheatgeneration canbeexpressed as:where(R)isthecableACresistance, (R)isthecableDCresistance, and(Y,)istheAC/DCresistance increment.

Initsmostsimpleform,therateofheatdissipation canbeexpressed as:

Thisassumption isacknowledged bytheutilitytobeofquestionable validity.

Ingeneral,thetreatment ofarectangular systemusingcirculargeometrycanintroduce anomalous

behavior, oftenassociated withtheparadoxofsurfaceareachanes.Thisassystemofnestedannularbodiesexperiences anincreaseintheavailable surfaceareaasonemovesoutwardthroughthesystem.Thisisnotacharacteristic ofatypicalone-dimensional rectangular system,andhence,greatcaremustbetakentoensurethatthesurfaceareasassumeddonotdistorttheheattransferbehavior.

Becausetheutilityhasprovidednodetailedcalculation examplesitisimpossible todetermine howsignificant thiseffectwouldbefortheutilityanalyses.

WhileCNPstatesthattheyachieved"excellent correlation betweencomputerdataandtestdata"noevidenceofthisvalidation isprovided.

Thisaspectoftheutilitymodelisstillconsidered suspect.Thisquestionwillberevisited in2.3.4below.Theutilitycalculation ofthecable-to-ambient thermalresistance apparently includesconsideration ofthevariouslayersinthesystem(thecableinsulation, thetrayorconduit,thebarrieritself)andtheairgapsbetweenlayers.However,theexactmannerinwhichthesevariousfactorsareimplemented inthemodelisunclearbecausetheutilitysimplycitescertaincorrelations, butdoesnotprovideanydescription oftheoverallfinalstructure ofitsthermalmodel.WhiletheoverallmethodistakenfromNeher-McGrath, thecorrelations usedtocharacterize thevariouslayersareactuallytakenfromapaperbyBullerandNeher[3].Whilethispaperwaspublished in1950,itcitesa1933textbyMcAdams[4]aseoriginalsourceofthecorrelations presented.

Hence,theseheattransfercorrelations arequitedated;andinfact,predateanybutthemostelementary ofheattransferinvestigations.

Thisisapointofsignificant potential concern.Since1933thefieldofheattransferhasadvancedsignificantly, andmuchimprovedcorrelations forh'eattransferbehaviorarenowavailable.

Inparticular, itisonlyintheperiodsinceattimethatextensive scientific investigations ofcriticalheattransferphenomena havebeenundertaken whichtookadvantage ofsophisticated instrumentation andelectronic datagathering methods.Infact,SNLwasunabletoobtainacopyoftheMcAdamstext,nowbeinglongoutofprint.Theuseofsuchdatedcorrelations wouldappearinappropriate.

Ofparticular concerninthisregardwouldbetheconvection, conduction andradiation termsassociated withtheairgapsasgivenbyutilityequations 17-19.Thesearequitecrudeestimates ofairgapehavior,andwouldnotbeconsidered goodpracticebytoday'sstandards.

Arelated,butsomewhatdifferent findingisaparticular concernrelatedtothemodelingofheattransferfromtheoutersurfaceofthebarriertotheambientenvironment (utilityequation21).Inthisparticular casetherearetwopointsofconcern.First,thisexpression actuallyderivesfroma1929work[5],andhence,mustalsobeconsidered quitedated.Second,thecorrelation appliesonlytopipes,andderivesfromstudiesofblackpipesthatrangedinsizefrom1.3to10.8inchesindiameter.

(Giventhevintageofthecorrelation, thisisnotverysurprising asmostofthethencurrentstudieswerefocussedonthefairlysimplecylindrical andspherical geometries.)

Innocaseshouldcorrelations basedonheattransferfromapipebeappliedtoflatplatesurfacessuchasthosewhichwouldbeexperienced aroundacable10 tray.Thephysicalconfiguration ofapipeisquitedifferent fromthatofaflatplateIandhence,thebuoyancydrivenconvective airflowssurrounding apipearequitedifferent fromthosesurrounding arectangular object.Ingeneral,theconvective currentsaroundapipewouldbefarmoreefficient, andhence,convective heattransferisalsotypically farmoreeffective forapipethanforaflatplateonaperunitofsurfaceareabasis(theaverageheattransfercoefficient forapipewouldbesignificantly largerthanthatofarectangular box).Thisshouldbeparticularly truefordownwardfacingheatedsurfacessuchasthoseonthebottomofatray.Thisisoneareawherethetreatment ofthecabletraysusingcylindrical assumptions wouldbebothinappropriate andoptimistic (tendtooverestimate theefficiency ofheattransfer).

OverallIwasunabletoassessthepotential impactthatuseofmoremoderncorrelations wouldhaveontheresultsbecausenolistingoftheactualprogramhasbeengiven,becausenospecificvaluesforthevariousinputshavebeenprovidedtosupporttheexamplecalculations citedbyCNP,andbecauseofotheruncertainties intheutilityanalysismethod.Asageneralpoint,theutilitymightbeabletoovercomethisshortcoming throughathoroughvalidation ofthemodelresults.Thatis,solongasthemodelyieldsconservative resultsforafullrangeofapplications, thentheissueofusingdatedorinappropriate correlations mightbeconsidered ofsecondary importance.

Aswillbediscussed inSection2.2.3below,theutilityhasnotmetsuchaburdenofvalidation.

Oneotherpointregarding thevariousheattransfercorrelations ishowtheutilityaccounted fortheeffectsoftheactualavailable surfaceareaineachlayer.Thatis,theutilitytreatseachlayerasaprogressively largercylinder.

Thisisnotanaccuraterepresentation ofacabletraysystem.Inparticular, acabletrayis,inreality,alargelyone-dimensional heattransferproblemwithheatflowingupwardsanddownwards awayfromthecentrally locatedcables.Hence,thesurfacearearemainsfixedasonepassesfromlayertolayer.Intheutilitymodel,thesurfaceareacontinuously increases asonemovesfromlayertolayer.Howthiswouldimpacttheutilityresultsinagainunclear.Ingeneral,theutilitytreatment mightyieldresultswhichareinsignificant error(eitherconservative ornonconservative).

Theactualrateofheattransferfrom(convection andradiation) orthrough(conduction) asurfaceisdirectlyproportional totheavailable surfacearea.Hence,ifthesurfaceareasarenotmaintained atthesamevalueinthemodelasthoseoftheactualphysicalsystem,thensignificant errorscouldresult.Thiswoulddirectlytranslate intoerrorsintheampacitylimitspredicted becausetheampacitylimitsaredirectlyrelatedtothelimitingrateofheattransfer.

Theutilitytreatment providesnoassurance thatthesurfaceareashavebeenappropriately treated.Thisisespecially important inadirectcalculation oflimitingheatloadssuchasthatimplemented bytheutility.Anotherpointofconcernrelatedtoaparticular factor(variable) introduced intoequations 16-21.Thisconcernraisesadditional questions regarding theintentandgeneralvalidityoftheutilityanalysis.

Thatis,thestatedintentofthePart2analysisistocalculate thelimitingheatloadforacabletrayasasystem.However,thetreatment appearstobebasedoncalculation ofindividual cableampacities

directly, andquitepossiblyusesathermalpartitioning assumption whichisindirectconflict viththestatedassumptions ofthePart1analysis.

Forexample,considerutilityequation16:'2=0.0104pn'-)D-C(CNP16)Thisequationiscitedascharacterizing "thethermalresistance throughrelatively thincylinders (i.e.,cablejacket,tray,firebarrier)".

Thevariable(n')isstatedtobe"thenumberofconductors withinthesection."

Thiswouldappeartobearatherstrangerelationship forexpressing thethermalresistance ofthefirebarrier,forexample.Whatpossibleimpactwouldthenumberofconductors withinthefirebarrierhaveonthethermalresistance ofthefirebarrieritself?Thereisnodirectcontactbetweenthetwo,andhence,thefirebarriersystemshouldnotbeimpactedbythenumberofcablesinsideofit.Inreviewing theBuller-Neher paperasomewhatd'ffaierenteinitionisgiven,buttheexactmeaningremainsunclear.Itwouldappearthatthisis,ineffect,a"thermalpartitioning factor."Thatis,itwouldappearthatBuller-Neher usethisfactortopartition theoverallsystemheatloadtoindividual conductors basedsimplyontheconductor count.Solongasthecablesareallidentical, suchanapproachmightbeconsidered appropriate.

However,theutilityanalysisinvolvescablesofdifferent typesandsizesinthesamecabletray.Hence,partitioning basedonthesimpleconductor countwouldbeinappropriate.

Thiswouldalsoappeartobeindirectconflictwiththeassumptions madeinthePart1analysis.

Thatis,inthePart1analysis, CNPhasassumedthatthethermalloadforthecabletrayasasystemcanbepartitioned toindividual cablesbasedoneachcablescontribution totheavailable surfacearea.Clearly,theuseofthefactor(n')isnotconsistent withthisstatedassumption.

Howtheutilityimplemented thisfactorisentirelyunclear.Asarelatedfinding,therewasonquitepuzzlingstatement madeduringthediscussion ofutilityequation20aswell.Intheparagraph immediately preceding thatequationtheutilitystates:"Thethermalresistance perconductor willbethetotalnumberofconductors dividedbythetotalthermalconductance."

Theintent,basis,andimpactofthisstatement iscompletely unclear.Hereagain,theutiityintroduces theparameter (n')asthenumberofelectrical conductors.

Ineffect,theutilityseemstobepartitioning thetotalthermalresistance associated withthislayer(theairgap)bythe'number ofconductors presenttoestimatetheresistance perconductor.

(Ineffect,thisisatreatment inwhichtheoverallsystemthermalresistance istreatedasasetofindividual resistance

elements, oneforeachconductor, arrangedinaparallelresistorconfiguration.

Hence,thethermalresistance fortheoverallsystemisactuallylessthanthatassociated withasinglecable.)th'&isapparentpartitioning ofthethermalresistance isnecessary isunclear.Howthisassumption accountsforcableswithdifferent physicaldiameters isalsounclear.Finally,thisapparentpartitioning appearstobeindirectconflictwiththePart1analysisassumption inwhichthepartitioning isassumedtobebasedonsurfacearearatios.Again,thelevelofdocumentation isinsufficient todetermine exactlyhowthis12 particular factorwasimplemented intheFinalutilitymodel(nolistingisprovided) andwhetherornottheproblemisself-correcting (forexample,thecomputermodelmaysimplymultiplyby(n')inonespotandthendivideby(n')somewhere elseandtherebyneutralize theassumption entirely).

Inanycasesomeadditional explanation ofhowtheseequations relatetoeachotherandtothePart1analysis, andthebasis,intent,andimpactofthefactor(n')isneeded.Onefinalfactorwhichappearstobelackingintheutilitymodelisatreatment oftheeffectsofspacingontheradiative heattransferbehaviorofthecables.Thatis,theutilityassumesthatthefullsurfaceofeachcableisequallyeffective asaradiating bodyforthedissipation ofheat.Thiswouldnotbecorrect,andwouldbecomemoreincorrect asthespacingbetweencablesbecamesmaller.Thatis,ineffect,thesidesofthecable"see"theneighboring cablesandexchangeheatwiththoseneighboring cablesaswellaswiththenextlayerofthethermalsystem(theinsideofthefirebarrierforexample).

Thisisnormallytreatedthroughtheapplication ofaradiation viewfactor.Thiswouldbeavaluebetween0.0and1.0whichrelatestherelativefractionoftheradiating bodywhicheffectively "sees"theotherparticipating surface.Inthiscase,thisvaluewouldlikelybeontheorderof0.5-0.7depending onthecable-to-cablespacing,andontheuniformity ofthecablesizes.ThisfactorwouldInesignificantly reduceradiantheattransferincomparison tothatassumedbthtil'ty.neffect,theutilityhasassumedanidealviewfactorof1.0,andthisiscertainly notcorrect.2.3.3Validation Asnotedabove,theutilitystatesthatitobtained"excellent correlation betweencomputerdataandtestdata".SNI.wasunabletofindevidenceofthisintheutilitysubmittal.

Thesubmittal doesincludeadescription of6specificexperiments performed tomeasurecableoperating conditions underveryspecificampacityloads.Theutilityhasalsoprovidedthefinalresultsofitsthermalanalysisofonespecificcabletrayandanundetermined numberofconduits(possibly asmanyas12conduitsorasfewasthreedepending onhowtheresultsareinterpreted).

However,thereisnodirectcorrelation betweentheexperiments performed andthecabletraysorconduitsanalyzed.

Suchacomparison maybedifficult todrawconsidering thenatureoftheexperiments documented.

Theutilitytestreportcitesthattheobjective oftheexperiments was"tosimulateascloselyaspossibletheactualconditions oftrayandconduitrunsproposedforCookPlantanddetermine thefinalconductor temperature forthespecified amperageandtrayfill."Thisobjective isnotcompatible withthegoalofvalidating ananalytical modelwhichpurportstoestimateultimatecableampacitylimits.Validation oftheampacitylimitmodelwouldrequirethat"theutilityshowthatitsestimated ampacitylimitsconservatively boundactualmeasuredampacitylimitsinthecorresponding physicalsystem.Thiswouldrequirethatatestbeperformed onaparticular physicalsystem,andthattheampacitylimitofeachofthecablesinthatsystemwhenoperatedsimultaneously bedetermined experimentally.

Theseexperimental valuesshouldthenbecomparedtotheestimated ampa'city limitsasa13 finalvalidation, Incontrast, theutilityhasperformed testsinwhichtheampaciryloadswerepredetermined, andthetestsimplymeasuredtheoperating temperatures ofthecables.Thiswouldbeausefultestforassessing theactualoperating conditions ofspecificin-plantinstallations, butwouldnotbeusefulinthevalidation oftheutilitymodelbecausethereisnodetermination oftheactualampacitylimitsforthetestedcase.Itispossiblethatthetestsoriginally described inAppendixCoftheutilityanalysisevelopment documentwouldbemorehelpfulinresolving thisuncertainty.

However,becauseAppendixCwasnotprovided, itisimpossible tomakethisassessment.

Giventhedocumentation

provided, SNLmustconcludethattheutilityhasprovidednomeaningful validation ofitsanalysismethodforcalculating cableampacitylimitsandcabletraylimitingheatrejection capacities.

23.4SummaryofTechnical ConcernsAnumberofareasofspecificconcernwereidentified inthisreview.Overall,thethermalmodelfortheestimation oftotallimitingheatrejection capacities wasfoundtobepoorlyfounded,andpoorlyvalidated.

Itisnotrecommended thatthismodelbeaccepteduntilthetechnical concernsidentified areresolved, andthemodelreceivesanadequatevalidation treatment.

Inparticular, thefollowing itemswereidentified:

Theutilitytreatsallcabletraysystemsusingamodelbasedoncirculargeometries.

Thispracticeissaidtohavebeenvalidated basedonthe"excellent correlation betweencomputerdataandtestdata".However,noevidenceofsuchvalidation hasbeenprovided.

Ingeneral,thetreatment ofaninherently rectangular geometrybasedonfcylindrical correlations canleadtosignificant errors.Thisisparticularl truorconvection correlations.

Theconvective heattransferfromacylinderismoreefficient thanthatfromaflatrectangular box.Theutilityuseofconvection andradiation correlations basedonheattransferfrompipes(utilityequation21)isinappropriate andnonconservative.

Thisutilitytreatment ofequivalent annularregionsalsoappearstoignoretheimportance oftheavailable surfaceareainheattransfercorrelations.

Thatis,heattransferratesaredirectlyproportional tothesurfacearea.Theutilityhasprovidednoassurance thatingenerating itsequivalent annularregions,appropriate heattransferareasrepresentative oftheactualphysicalsystemhavebeenmaintained.

Thiscouldeasilydistortthemodelingresultsineitheraconservative ornonconservative manner.Thisisespecially truesincetheutilityisattempting todirectlycalculate absoluteheattransferrates,whichinturndetermines theampacitylimitsoftheinstalled cables.Inamoregeneralcontext,giventhedocumentation

provided, SNLmustconcludethattheutilityhasprovidednomeaningful validation ofitsanalysismethodforcalculating cableampacitylimitsandcabletraylimitingheat14 rejection capacities.

Suchvalidation shouldberequiredbeforetheanalysismethodology isaccepted.

Theutilitymodelisbasedonheattransfercorrelations

'whichwereorignially published inthe1929-1933 timeframe.Thesecorrelations arebadlydated,andtheimpactofusingmoremoderncorrelations inthemodelshouldbeassessed.

Theutilityhasmadeaveryconfusing statement regarding therelationship betweenthethermalresistance ofindividual cablesandthatofthesystemasawhole(seethefirstsentenceintheparagraph immediately preceding equation20oftheutilityanalysis:

"Thethermalresistance perconductor willbethetotalnumberofconductors dividedbythetotalthermalconductance").

Thebasis,intent,andimpactofthisstatement needstobefurtherexplained andclarified particularly inthecontextofthestatedobjective ofestimating thelimitingheatloadforthecabletrayasasystemratherthanindividual cableheatloads(theapparentpurviewofthePart1analysismethod).Theutilitycorrelation for"thethermalresistance throughrelatively thincylinders (i.e.,cablejacket,tray,firebarrier)"

(utilityequation16)includesafactor(n')described as"thenumberofconductors withinthesection."

Asimilartreatment isalsonotedinthecaseofutilityequation21,the"thermalresistance fromthelastsurfacetoambient."

Itisunclearhowthisfactorhasbeenappliedintheutilityanalysis.

Itwouldappearthatthisis,ineffect,apartitioning factorforheattransferfromthesystemdowntoindividual conductors.

Ifthisisacorrectinterpretation, thenthispartitioning isindirectconflictwiththePart1analysisinwhichheatispartitioned onthebasisofavailable surfacearea.Theuseofthisfactorinthecontextoftheutilityanalysisisunclear,andmaybeinappropriate.

Inparticular, theutilityisdealingwithsituations involving cablesofvarioussizes,types,insulation thicknesses andconductor number.Howthisfactorisimplemented bytheutilityandtheneteffectofthisfactoronitsanalysisneedstobeclarified.

Thebasis,intent,andimplementation oftheseequations, andthefactor(n')inparticular, shouldbeclarified.

Theutilitymodelhasprovidednotreatment oftheeffectsofspacingontheradiative heattransferbehaviorofthecables.Theutilityshouldincludeconsideration ofradiation viewfactorsinthedevelopment ofitsradiative heattransfercorrelations whichmightsignificantly reducethepredicted ratesofradiantheattransfer.

ConduitApplications oftheModelThedevelopment ofboththeutilityPart1andPart2thermalmodelsispresented primarily inthecontextofcabletrays.However,itappearsthatthesamethermalmodelisalsobeingappliedtoconduits.

Thisraisescertainuniquequestions whichshouldbeaddressed bytheutility.15 4~,Itmustbenotedattheoutsetthatthelevelofinformation providedwithregardstoitsconduitapplications wasevenmoresparsethanthatprovidedforthecabletrays.Inparticular, theutilityhasprovidednodiscussion ofhowthemodelwasappliedtoconduits, andyet,exampleresultsforcablesinconduitsarepresented (seetheutilityAttachment 5).Withregardtotheseexampleresults,whiletheutilityhasprovidedseveralexampleanalysesofcablesinconduits, itisunclearwhethereachcableishousedinanindividual conduit,orwhethermorethanonecablemightbehousedinacommonconduit.Giventhesizeofthecables(thelargestis1.14"indiameter),

andthesizeoftheconduits(upto4"),itisquitepossiblethatmorethanonecableisinstalled inagivenconduit.First,recallthattheuseofamaintained spacinginstallation procedure foritscabletrayswascitedinthisreviewasacriticalfactorinthethermalmodelingofthecabletrays.Thissamefactorwouldcertainly notapplyingeneraltoconduits.

Ofparticular concernwouldbeanyconduitwhichhousesmorethanonecable.Foraconduitwithasingleinstalled cable,theutilitythermalmodelmightbeconsidered appropriate.

Thatis,inadirectcalculation ofampacitylimits,themostconservative approachwouldbetoassumethatthecableislocatedinthecenteroftheconduitandthatthereisnocontactbetweenthecableandtheconduititself.'n thecaseofasinglecable,thefullsurfaceofthecablewouldbeavailable forheattransfer.

However,inthecaseofanyconduitwithmorethanonecableinstalled, thecableswouldbearrangedinabundleofsometype,andhence,onlyafractionofthetotalcablesurfaceareawouldactuallybeactiveinthesurfaceheattransferprocesses.

Theutilitymodelwould,apparently, assumethatthefullsurfaceareaofthecableswasavailable fordirectthermalexchangewiththenextlayersinthesystem(convection totheairgapandradiation totheinnersurfaceoftheconduit).

Thiswouldbeanonconservative assumption formultiplecableinstallations.

Thelackofmaintained spacingforcablesinconduitswouldalsoeffectthevalidityoftheutilityPart1partitioning analysis.

Thatis,forconduitswithmorethanonecableinstalled, thepartitioning ofthetotalallowable heatlimitforthesystemtoindividual cablesbasedonthesurfaceareaofthecablescouldnotbejustified onthesamebasisasthatappliedtothecabletrays(maintained spacing).

This,again,isbecausetheactivesurfaceareawouldnotbeequaltothetotalsurfacearea.Infact,itisquitepossibletohaveacablefullysurrounded byothercables,andhence,tohaveineffectnoactiveheattransfersurfaceareafordirectexchangewiththenextlayerinthesystem(convection totheairgapandradiative exchangewiththeinnersurfaceoftheconduit).

2Recallthatthisisincontrasttothecaseinwhichamodelisattempting toassessarelativederatingimpact.Inthecaseofarelativecalculation, assumingthatthecableisincontactwiththeconduitisbothmorerealistic andmoreconservative becausethismaximizes therelativechangeduetoadditionofthefirebarrier.CNPhasnotperformed arelativecalculation, butrather,anabsolutecalculation ofampacitylimits.Hence,thecentrallocationassumption wouldbeconservative.

16 Asecondfactorwhichis,again,notclearrelatestothegeneraltreatment ofsurfaceareaintheutilitymodel.Asnotedaboveinthecontextofthecabletrays,heattransferratesaredirectlyproportional totheavailable surfacearea.Hence,itiscriticalthatthesurfaceareavaluesassumedinthemodelaccurately represent thephysicalsystem.Thiswouldbeespecially truefortheconduitsbecausetheavailable surfaceareaisgenerally small,andminorincreases intheassumeddiameterofaconduitwouldseverelyeffecttheheattransferrates.Theinputvariables identified biieytheutilityimplythatsuchanartificial increasemightinadvertently resultfromthestructure oftheutilitymodel.Thatisbecausethemodelinputsaregivenintermsoftheinitialdiameteroftheinnerlayerandthenthethickness oftheairgapsbetweenIayers.Hence,depending onhowthemodelwasimplemented, theselayerthicknesses mightsimplybeaccumulated todetermine theequivalent diameterofeachlayer.Becausethecomputercodehasnotbeensupplied, itisimpossible totellhowthiswasimplemented bytheutility.Intheutilitymodelitisadirectcalculation ofheattransferrateswhichdetermines ampacitylimits,andhence,errorsintheestimation ofheattransferratestranslate directlyintoampacityerrors.2.5Coordination oftheUtilityAnalysisPackageElementsAsdiscussed inSections2.2and2.3above,theutilityanalysismethodology ispresented intwoparts.Thetextaccompanying themodeldescriptions impliesthatthefirstpartoftheanalysis(asdocumented inutilityAttachment 4,SectionA)providesabasisforassessing individual cableampacities basedontheoverallheatrejection capacityofthethermalsystemasawhole.Similarly, itisimpliedthatthesecondpartoftheanalysis(asdocumented inutilityAttachment 4,SectionB)isintendedtoprovidefortheestimation oftheoverallheatrejection capacityofthethermalsystemasawhole.Thatis,thePart2analysisestimates overallheatrejection

capacity, andthePart1analysispartitions thatoverallcapacitytotheindividual cablesinthesystem.However,uponreviewofthePart2analysisdescription, itisnotentirelyclearhowthetwopartsoftheanalysis, infact,worktogether.

Thedevelopment intheparttwoanalysisprovidesanexpression fortheampacityofthecablesdirectlyasafunctionoftheenvironmental andelectrical conditions.

ThiswouldappeartomakethePart1analysismethodentirelyobsolete.

ThePart2analysisalsoincorporates aconductor countfactor,(n'),whichappearstoactasathermalpartitioning factor.Ifthisinterpretation iscorrect,thenthispartitioning isindirectconflictwiththeassumptions madeinthePart1analysis.

Thatis,inthePart1analysisthermalpartitioning isassumedtobeinproportion tothecablesurfacearea.InthePart2analysis, thermalpartitioning appearstobebasedonasimpleconductor count(eachconductor ispartitioned equallyregardless ofsize).Thisisaninconsistency whichmustberesolved.

Theutilityshouldbeaskedtoclarifytheintentofeachpartofitsoverallanalysismethod,andtodescribehowthetwopartsoftheanalysisworktogether.

Theexamplesprovidedshouldillustrate bothaspectoftheanalysis(thePart1andPart2analyses) andshouldprovidesufficient information toverifythecalculations.

17 Asafinalpointofgeneralconcern,Imustremainskeptical ofanypurelyanalytical predictions ofactualampacitylimits.Inourownwork',itwasfoundthatwhilepredicting arelativechangeinampacitylimitsduetoadditionofafirebarriersystemwasrelatively simple,predicting actualampacitylimitsbasedondirectthermalmodelingwasmuchmoredifficult, andledtomuchgreateruncertainties.

Theproblemsgenerally arisefromtheratherlargeuncertainties inherentingeneralheattransfercorrelations forsuchfactorsasconvective heattransfer.

Thisisaparticular concerngiventhattheutilityisbasingitsanalysisoncorrelations of1929-1933 vintage.Thismeansthatthereisasignificant inherentuncertainty intheutilitycalculations, andthatsignificant validation againstknownconditions shouldbeprovided.

Onepotential approachwhichmightprovefruitfulwouldbefortheutilitytocompareitsmodelingpredictions toother,morecommon,ampacityderatingapproaches andtestresults.Thisisdiscussed furtherinSection3inconjunction withareviewoftheutilityexamplesprovidedinCNPAttachment 5.'SeetheresultsofUSNRCJCNJ2018,TaskOrder2.18 3.L"TILITYEXAMPLECALCL'LATIONS 3.1OverviewTheutilityhasprovidedtwosetsofexampleampacitycalculations, oneforaparticular cabletray,andasecondsetassociated withcertainconduits.

Ingenealthlevi.ngener,eForevelofdetailprovidedbytheutilityisinsufficient tofullyreviewthesecalculati cua'ons.orexample,innocasehastheutilityidentified thenatureofthefirebarrierinstalled (nominal1hror3hr,norinstallation characteristics suchasthickness, materials,

upgrades, etc.).Forthecabletrayexample,noinformation onthephysicalcharacteristics ofthecabletrayareprovided(width,height,covers,orconfiguration suchassolidbottom,ventilated bottom,orladdertype).Fortheconduits, itisnotpossibletodetermine whethereachcableishousedinaseparateconduitorwhethermultiplecablesmightbelocatedinacommonconduit(grouping factorsforconduitsareaparticularly important consideration).

Asapartofthisreviewoftheutilityexampleresults,SNLhasattempted tocomparetheutilityresultstothoseonemightobtainusingmore"conventional" approaches toampacityderating.

Inparticular, amoretypicalapproachtothederatingwouldinvolveaninitialassessment ofthebaselineampacityofthecablesfrompublished ampacitytables,andthederatingofthosevaluesbasedonfactorssuchasambienttemperature, groupingandthefirebameritself.Inthisreview,SNLhasattempted tomakesuchcomparisons asappropriate totheparticular example.3.2CableTray1AZ-P8InthecaseofsparselyloadedcabletrayssuchasthoseatCNP,themoreconventional approachtoampacityderatingwouldbetobeginwiththebaseampacityvaluesfromthetablesusingtheapproachof"maintained spacing"asperIPCEAP-46-426.

Thenagenericestimateoftheampacitycorrection factor(ACF)forthefirebarriercouldbeapplied,andanominalampacitylimitfortheprotected cablesfound.Thisvaluecouldthenbecomparedtothepredictions oftheutilitymodelforaroughassessment ofhowwellthemodel~ouldreflectthecurrentampacitytables.Oneobvioussourceofuncertainty inthisapproachis'thattheremaynotbegoodtestsuponwhichtoestimatetheACFofafirebarriersysteminstalled onasparselyloadedtray.Ingeneral,onemightexpectamarginally higherACFforasparsetraythanonewouldforaheavilyloadedtrayduetothemoreprofoundeffectonconvective aircurrentsforthesparselyloadedtray.However,asafirstorderapproximation thisapproachwouldcertainly helplendconfidence to,orhighlight deficiencies of,theutilityanalysismodels.SNLhasperformed suchacomparison fortheonecabletrayidentified inCNPAttachment 5(TraylAZ-P8).Unfortunately forthistrayCNPhasgivennoinformation regarding thenatureoftheinstalled firebarriersystem,andfromtheinformation provideditisimpossible todeducewhetherthisisanominallhror3hrsystem.Hence,thefollowing analysiswillremainspeculative.

Itisprovidedforillustrative purposesonly.Theutilityhascited12,3ICcablesinthistray.Elevenoftheseare12AWGcopperconductor cables(oneofwhichisunpowered "cutintrayandtaped"),andoneisa42AWGaluminumconductor cable.Allaretype"TC"19 cables(indicating theserviceconditions allowable forthiscable).Table3.1summarizes theampacityservicefactorsforthiscabletray.Table3.1:Acomparison ofCNPpredictions andanominalanalysisbasedontabulated ampacityvalues,maintained spacingandanominalfirebarrierACFforCNPcabletray1AZ-P8.812AWG,3/C,Cu,TC@2AWG,3/C,Al,TCOpenairampacity'aintained spacingACF'erated openampacityNominalfirebarrierACF'ominal deratedampacityUtilityestimated ampacityUtilityhighestactualloadcited32A.8226.2A.68417.9A21.58A20A108A.8288.6A.68460.6A90.67A60A1.fromNECTableB310-3,19962.MaximumvaluefromIPCEAP-46-426, TableVIIforasinglelayerofcables3.BasedonnominallhrcabletrayfirebarriersystemtestedbyTexasUtilities.

Thisisacrudeestimateforillustrative purposesonly.Theactualfirebarrierconfiguration atCNPisunknown.'Notethattheresults,evenassuminganominalampacitycorrection factor(ACF)basedonalhrbarriersystemtestedbyTexasUtilities, showsthattheCNPestimated ampacitylimitsmaybeoverlyoptimistic, andthatcertainofthecablesmaybeoperating atorabovetheiractualampacitylimits.Thisexampleillustrates theimportance ofpropermodelvalidation, andtheimportance ofcomparisons ofthemodelpredictions topublished ampacitylimits.3.3TheCNPAppendixRConduitsAswasnotedabove,theutilityhasnotprovidedanydiscussion ofhowitstwothermalmodelswereappliedtoconduits, andyet,resultsforcertaincablesinconduitsarepresented asapartoftheutilitypackage.Hence,onemustconcludethatthesamethermalmodelwasusedforconduitsaswell.Theconduitresultsarealsopresented inCNPAttachment 5.Hereagaintheutilityhasprovidedonlyaminimalamountofinformation uponwhichtobasethiscomparison.

Forexample,theutilityhasfailedtoidentifywhetherornoteachcableislocatedinitsownconduit,orwhethermorethanonecablemightbeinstalled inacommonconduit.Insomecases,theanswerisobvious(itisdifficult togetmorethanonecableof0.32"diameterintoa1/2"conduit,forexample).

However,inothercasesseveralcablesmightwellbelocatedinacommonconduit(a4"cablemightwellholdmorethanone1.14"diametercable).Further,thenatureofthefirebarriersinstalled ontheseconduitsisalsounknown.20 Hence,asabove,thisdiscussion isforillustrative purposesonly.Inallcases,ithasbeenassumedthatonlyasinglecableislocatedinanygivenconduit.Hence,noACFforgroupingofcablesinacon'duithasbeenapplied.Ifthisassumption isincorrect, thentheestimated ampacitylimitsgivenherewouldbetoogenerousandwouldrequirereduction forgroupingofcables.Table3.2summarizes theresultsofthiscomparison.

Table3.2:Acomparison ofCNPpredictions andanominalanalysisbasedontabulated ampacityvaluesandanominalfirebarrierACFforCNPconduits.

812AWG,3/C,Cu,TC42AWG,3/C,Al,TCNominalconduitampacity'ominal firebarrierACF'ominal deratedampacityUtilityestimated ampacityUtilityhighestactualloadcited24.6A22.1A25.85A2.7A84.6A76.2A99.04A71.9A1.fromNECTableB310-1,1996.Includescorrection ofampacitytoambienttemperature of40'C.2.Basedonnominal1hrconduitfirebarriersystemstestedbyTexasUtilities.

Thisisacrudeestimateforillustrative purposesonly.Theactualfirebarrierconfiguration atCNPisunknown.Notethattheampacitylimitspredicted byCNParelargerthantheestimated deratedampacitylimitsforthesecables,evenusingnominalvaluesforthefirebarrierACF.Infact,thederatedampacitylimitspredicted byCNParelargerthanthenominalampacitylimitsspecified intheNECtableswithoutconsideration ofadditional firebarrierderating.

Thisdiscrepancy indicates potential problemsintheCNPthermalmodel,andclearlyindicatethatthemodelmaybegenerating nonconservative estimates ofcableampacitylimits.Thesediscrepancies mustberesolved.

Inthisparticular case,thein-plantserviceloadsremainboundedbytheestimated deratedampacityvalues.However,thisisbasedononlyanominalanalysis.

Asnotedabove,theinclusion ofcablegroupingfactors,ifsuchfactorswouldbeapplicable atCNP,orthepresenceofa3hrratedbarriersystemmightsignificantly alterthefinalampacityestimates.

ItisalsounclearwhetherornotthespecificcablescitedbyCNPareeitherall-inclusive ofconduitfirebarriersorarerepresentative ofboundingapplications.

AsnotedinSection2.4above,thereisconsiderable uncertainty regarding howtheconduitmodelingapplications wereimplemented.

Itisinteresting tonoteherethattheutilityresultsprovideauniformvalueofampacityforagivencablesize.Providedthatallofthecablesareinstalled intheexactsameconfiguration, thiswouldbeanappropriate result.However,itthenumberofcablesinaconduitvariesfrom21 application toapplication, oriftheconduitsizevariedfromcasetocaseforthesamecable,thenoneshouldseesomedifferences intheallowable ampacitylimits.Oneshouldexpectthatanychangeinthephysicalsystemwouldbereflected inachangeintheampacitylimitscalculated.

Theutilityresultsforagivencablesizeareallidentical, andhence,onemustassumethattheinstallations areallidentical (foragivencablesize).Thisisnotthecaseforatleastoneofthetwocablesizesconsidered.

Considerthe3/C412ANGresultsfora1/2"conduit(cable8026R)incomparison tothoseforthesamesizecableina1"conduit(cable8505Rforexample).

Inthiscasetheutilityhascitedtheexactsameampacitylimitdowntofoursignificant figures(25.&5A).

Thisclearlyindicates somesortoferrorintheutilitymodel,orintheimplementation ofthemodel.Giventhesamecablesintwodifferent conduitsofsignificantly different sizeonewouldcertainly expectsignificant differences inthenumerical modelingresults.Whiletheampacitytableswouldnotdistinguish betweenthesetwocases,thethermalmodelcertainly should.Thefactthatthetworesultsarelistedasidentical indicates thatthethermalmodelisnotproperlyaccounting forthephysicalcharacteristics ofthesystem.3.4SummaryofInsightsandFindingsAnominalcomparison wasmadebetweentheampacityresultsprovidedbytheutilityandthosewhichmightbeobtainedusingmoreconventional approaches totheampacityassessment.

Inboththecabletrayandconduitcasescited,itwasfoundthattheCNPestimated ampacitylimitswerenonconservative incomparison tonominalampacitylimitsderivedfromderatingofthepublished cableampacitytables.Inthecaseoftheconduits, theutilityestimated ampacitylimitsincluding deratingforthe'firebarriersystemwereinexcessofthetabulated ampacitylimitsforcablesinconduitswithoutafirebarrierassetforthintheNECtables.Theseresultsindicatepotential problemsintheCNPthermalmodel.Theprediction ofactualcableampacitylimitsbasedondirectthermalmodelingisquitedifficult, andwouldbeexpectedtoholdconsiderable uncertainty.

Theresultsofthecomparisons madehereindicatethattheCNPthermalmodelmaywellbegenerating unrealistic andnonconservative estimates ofactualcableampacitylimits.Additional validation oftheutilitythermalmodelisneeded.Asapartofthevalidation process,theutilityshouldprovideadirectcomparison ofitsownmodelingresultstotheresultsobtainedusingmoreconventional ampacityderatingapproaches, and/ortoactualtestresultsinwhichampacitylimitsweremeasureddirectly.

Thisshouldincludebothcabletrayandconduitapplications ifthemodelistobeappliedtobothtypesofinstallations.

TheSNLcomparisons mustbeviewedinthecontextofillustrative examplesonly,dueprimarily tothefactthatinsufficient information hasbeenprovidedbytheutilityuponwhichtobasemoredefinitive analyses.

Initssubmittal, theutilityshouldalsoprovideasufficient baseofinformation onitsparticular applications toallowforacompletereviewandassessment oftheresults.Thismustincludemoredetaileddescriptions ofthephysicalcharacteristics ofeachsystem,andthecharacteristics oftheinstalled firebarriersystem.22 Theexamination oftheconduitresultsalsoraisedaparticular pointofconcernwhichindicates thatthereareerrorseitherinthethermalmodelorintheutilityimplementation ofthatmodel.Inthecaseofthe3/C12AWGwires,theutilitypredicted thesameampacitylimitsforacableina1/2"conduitandforthesamecableina1"conduit.Whiletheampacitytableswouldnotdistinguish betweenthesetwocases,thethermalmodelcertainly should.Thisisaclearindication ofanerrorofsometype.Basedontheinformation

provided, itisimpossible toidentifythesourceofthiserror.23 4.SL'iiMARY OFFINDINGSANDRECOMMEM3ATIQNS Withrespecttotheadequacyoftheoverallutilitydocumentation, SNLfindsthat:Thelevelofdocumentation providedisnotadequatetocompletefullevaluation oftheutilityampacityassessments.

Specificareasinwhichfurtherdocumentation isrequiredaredocumented below.Ingeneral,thereisnodiscussion ofhowthetwopartsoftheutilityanalysismethodology aremadetoworktogether, theexamplecalculations donotprovideenoughinformation toverifythe'calculations, andtheexperiments purported tosupportvalidation ofthethermalmodelsareeithernotprovided, ornodirectone-to-one comparison oftheexperiments tomodelingresultsisprovided.

Whiletheutilityhasdocumented theresultsofitsanalysisforonecabletrayandanindeterminate numberofconduits, nosummaryofthebalanceoftheplantresultshasbeenprovided.

Theutilityshouldprovideasummaryoftheampacityassessment resultsforitsinstalled firebarriersystems.Nodiscussion hasbeenprovidedastohowin-plantcableserviceloadsweredetermined, whichcableshavebeenconsidered intheanalysis, andthebasisfortheelimination ofothercablesfromconsideration.

Thisinformation isneededtoassesstheadequacyoftheutilitytreatment.

Withrespecttotheutilityampacity"partitioning" analysismethodology (referred tointhisreviewasthePartIanalysis)

SNLfindsthat:Theutilityhasprovidedaninadequate basisforvalidation ofitsassumption thattheoverallheatrejection capacityofasparselyloadedcabletraycanbepartitioned toindividual cablesinproportion tothecablediameter.

Thelimitedinformation providedbytheutility(oneplotwithnosupporting dataandnoindicated units)isunconvincing.

Theutilityhascitedasetofexperiments asthebasisforthisplot,andhence,forthevalidation ofthismethodology (seereference toAppendixCintheutilityAttachment 4).However,nodocumentation oftheseexperiments hasbeenprovided.

Documentation ofthevalidation experiments citedinthisportionoftheutilityanalysisisneeded.Thisshouldincludeadiscussion oftheutilityanalysisandapplication ofthetestdata.Withrespecttotheutilitythermalheatrejection capacitycalculation method(referred tointhisreviewasthePart2analysis):

Theutilitytreatsallcabletraysystemsusingamodelbasedoncirculargeometries.

Thispracticeissaidtohavebeenvalidated baseonthe"excellent correlation betweencomputerdataandtestdata".However,noevidenceofsuchvalidation hasbeenprovided.

Theutilitymustvalidateitsassumptions inthisregard.24 Theutilityassumption ofequivalent annularregionsappearstogiveinadequate treatment totheimportance ofsurfaceareainheattransfercalculations.

Theactualratesofheattransferaredirectlyproportional tosurfacearea,hence,itisimportant thatthethermalmodeluseactualavailable surfaceareasinitsformulation.

Theutilityshouldprovideexamplestoillustrate theeffective heattransferareasassumedforeachofthelayersinitsmodelingandcomparethoseassumedareastotheactualheattransferareasavailable inthephysicalsystem.Theseexamplesshouldcoverbothconduitsandcabletrays.(Forcabletraysitisrecommended that,consistent withothermodelingefforts,theutilityshouldassumethatonlytheupperandlowersurfacesofthetrayandfirebarrierareactiveintheheattransferprocess.Bothexperiments anddetailedmodeling(Stolpe,e.g.)haveshownthatthesidesofacabletrayarerelatively unimportant intheoverallheattransferprocess.)

SNLfindsthattheutilityhasprovidednomeaningful validation ofitsanalysismethodforcalculating cableampacitylimitsandcabletraylimitingheatrejection capacities.

Theutilityshouldprovideforthedirectcomparison ofpredicted cableampacitylimitstothosemeasuredinexperiments onthecorresponding systeminordertovalidateitscalculations.

Theutilitythermalmodelisbasedonheattransfercorrelations whichwereoriginally published inthe1929-1933 timeframe.Thesecorrelations arebadlydated,andtheimpactofusingmoremoderncorrelations inthemodelshouldbeassessed.

(Thisissuemightbeconsidered ofsecondary importance providedthatasufficient baseofvalidation wereprovided.

Suchavalidation basehasnotbeenprovidedasperthepreceding finding.)

Thesubmittal states,inthedevelopment ofcorrelation forheattransferfromthecablestothesurrounding airgap,that"thethermalresistance perconductor willbethetotalnumberofconductors dividedbythetotalthermalconductance" (seethefirstsentenceintheparagraph immediately preceding equation20oftheutilityanalysis).

Thebasis,intent,andimpactofthisstatement needstobefurtherexplained andclarified.

Directlyrelatedtothepreceding comment,theutilitycorrelations for"thethermalresistance throughrelatively thincylinders (i.e.,cablejacket,tray,firebarrier)"

(utilityequation16)andforthe"thermalresistance fromthelastsurfacetoambient"(utilityequation21)eachincludeafactor(n')described as"thenumberofconductors withinthesection."

Thebasis,intend,andimplementation ofthisfactor(n')shouldbeclarified asitisappliedtoeachofthesetwoequations.

Thiswouldappeartobe,ineffect,athermalpartitioning factorwhichisbasedonasimpleconductor count,andassuch,maybeindirectconflictwiththestatedassumptions ofthemodeldescribed in"Appendix A"ofutilityAttachment 4(whereitisassumedthatsurfaceareawillbethebasisforthermalpartitioning).

Thebasis,intent,andimpactofthisfactormustbeclarified andjustified.

Also,theutilitymustexplainhowthisfactorapplieswhencablesofdifferent physicaldimensions arepresentinacommoncabletray.25 Thecorrelation citedforthethermalresistance betweentheoutersurfaceofthebarrierandtheambientenvironment (utilityequation21)appliesonlytopipes.Thesecorrelations shouldnotbeappliedtoflatplatesurfacessuchasthosewhichwouldbeexperienced aroundacabletray.Theutilitymodelprovidesnotreatment oftheeffectsofspacingontheradiative heattransferbehaviorofthecables.Reducedviewfactorsduetocableproximity mightsignificantly reducethepredicted ratesofradiantheattransfer.

Thisshouldbeaccounted forintheradiative heattransfercorrelations.

Itisnotpossibletoverifytheutilitycalculations becausenoinformation onthespecificinputsusedinanyexamplecasehavebeenprovided, andthelistingoftheutilitycomputercodewasnotprovided.

Theutilityhasapparently appliedoneorbothofitsthermalmodelstotheanalysisofcablesinconduitsaswellasthoseintrays.Theapplicability ofbothpartsoftheutilityanalysismodeltoconduitsneedstobeaddressed.

Inparticular, theutilitymodelforcabletraysisbasedontheuniqueconfiguration of"maintained spacing"foritspowercablesincabletrays.Thissamefactorcannotgenerally beassumedtoexistforcablesinconduits.

Theutilityshouldprovideanexplicitdiscussion oftheapplicability ofitstwothermalmodels(the"Appendix A"surfaceareabasedheatloadpartitioning modelandthe"Appendix 8"directthermalanalysismodel)toconduits, andshoulddiscusshowthemodelswereimplemented forconduitanalyses.

Withregardtothespecificexamplecalculations providedinthesubmittal SNLfindsthat:Anominalcomparison betweentheutilityampacitypredictions andthoseobtainedusingmoreconventional approaches totheampacityassessment showedthattheCNPestimated ampacitylimitswerenonconservative.

Inthecaseoftheconduits, theutilityestimated ampacitylimitsincluding deratingforthefirebarriersystemwereinexcessofthetabulated ampacitylimitsforcablesinconduitswithoutafirebarrierassetforthintheNECtables,evengiventhemostgenerousinterpretation oftheconduitloadings(onlyasinglecableperconduitwithnomorethanthreeconductors).

TheseresultsindicatethattheCNPthermalmodelmaybegenerating unrealistic andnonconservative estimates ofactualcableampacitylimits,Thesediscrepancies mustberesolvedbyCNP.Asapartofthevalidation process,theutilityshouldprovideadirectcomparison ofitsownmodelingresultstotheresultsobtainedusingmoreconventional ampacityderatingapproaches.

Initssubmittal, theutilityshouldalsoprovideasufficient baseofinformation onitsparticular applications toallowforacompletereviewandassessment oftheresults.Thismustincludemoredetaileddescriptions ofthe26 physicalcharacteristics ofeachsystem,andthecharacteristics oftheinstalled firebarriersystem.Theexampleresultsforconduitspredicted theexactsameampacitylimitsdowntofoursignificant figures(2S.8SA)fora3/C412AWGcableinbotha1/2"conduit(cable8026R)anda1"conduit(cable8505Rforexample).

Whiletheampacitytableswouldnotdistinguish betweenthesetwocases,giventhedifferences inthephysicalconfigurations, thethermalmodelcertainly should.Thisisaclearindication ofanerroreitherinthemodelorintheimplementation ofthemodel.CNPshouldidentifythesourceof,andresolve,thisdiscrepancy.

Itisrecommended thatanRAItotheutilitybepreparedtoclarifythesepoints.Ingeneral,itislikelythatsignificant additional consideration willberequiredontheartPoftheutilitytoresolvetheconcernsidentified inthisreview.Significantly morecompletedocumentation oftheutilitymodels,theexamplecalculations, andasummaryoftheoverallanalysisresultswillalsobeneededbeforeafinalassessment oftheutilityampacityloadfactorscanbemade.27

5.0REFERENCES

1.Stolpe,J.,"Ampacities forCablesinRandomlyFilledTrays,"IEEETransactions onPowerApparatusandSystems,Uol.PAS-90,Pt.I,PP962-974,1971.2.J.H.NeherandM.H.McGrath,"TheCalculations oftheTemperature RiseandLoadCapacityofCableSystems",

AIEETransactions pp752-772 Oct.19573.F.H.Buller,J.H.Neher,"TheThermalResistanc BetweenCablesandaSurrounding PipeorDuctWall,"AIEETransactions, V69,pp342-349, 1950.4.W.H.McAdams,HeatTransmission, McGraw-Hill, 1933.S.R.H.Heilman,"SurfaceHeatTransfer,"

ASMETransactions, U51,pp287-302, 1929.28