Battery-thermal-models-for-hybrid-vehicle-simulati

BatterythermalmodelsforhybridvehiclesimulationsAhmadA.PesaranNationalRenewableEnergyLaboratory,1617ColeBlvd,Golden,CO80401,USAAbstractThispapersummarizesbatterythermalmodelingcapabilitiesfor:(1)anadvancedvehiclesimulator(ADVISOR);and(2)batterymoduleandpackthermaldesign.TheNationalRenewableEnergyLaboratory’s(NREL’s)ADVISORisdevelopedintheMatlab/Simulinkenvironment.ThereareseveralbatterymodelsinADVISORforvariouschemistrytypes.Eachoneofthesemodelsrequiresathermalmodeltopredictthetemperaturechangethatcouldaffectbatteryperformanceparameters,suchasresistance,capacityandstateofcharges.AlumpedcapacitancebatterythermalmodelintheMatlab/SimulinkenvironmentwasdevelopedthatincludedtheADVISORbatteryperformancemodels.Forthermalevaluationanddesignofbatterymodulesandpacks,NRELhasbeenusingvariouscomputeraidedengineeringtoolsincludingcommercialfiniteelementanalysissoftware.ThispaperwilldiscussthethermalADVISORbatterymodelanditsresults,alongwiththeresultsoffiniteelementmodelingthatwerepresentedattheworkshopon‘‘DevelopmentofAdvancedBatteryEngineeringModels’’inAugust2001.PublishedbyElsevierScienceB.V.Keywords:Batterythermalmodel;Batterythermalmanagement;Vehiclesimulations1.IntroductionElectricandhybridelectricvehiclesaremoreenergyefficientandcleanerthanconventionalvehicles.Thesevehicles’performancedependsstronglyontheirbatterypackperformance.Ingeneral,temperatureaffectsseveralaspectsofabatteryincluding:
operationoftheelectrochemicalsystem;
roundtripefficiency;
chargeacceptance;
powerandenergycapability;
reliability;
lifeandlifecyclecost.Therefore,itcanbeconcludedthatthebatterytempera-tureimpactsvehicleperformance,reliabilityandlifecyclecost.Abattery’stemperaturechangesfromitsinitialvaluebecauseofinternalheatgenerationduetoelectrochemicalreactionsandresistanceorJouleseffectheatingknownasI2Rheating(I¼currentandR¼internalresistance).Externalheatingorcoolingfromthesurroundings,suchasconduction,convectionand/orradiationwillalsochangethetemperature.Dependingonthegeometry,materialorconstructionandlocationofpositiveandnegativeterminals,therecouldbeatemperaturedistributioninamodule.Asthebatterytemperaturechanges,itsstateofcharge,opencircuitvoltage,internalresistance,powerandavailableenergycanchange.Inabatterypack,thelocationofeachmodule,externalconditionsandtypeofheatingandcoolingcouldcreateanuneventemperaturedistributioninthepack.Uneventem-peraturedistributioninapackcouldleadtoelectricallyunbalancedmodulesandthusleadtolowerperformanceforthepackandvehicle.Eachbatterytypeworksbetterinaparticulartemperaturerange,forexample,lead/acid,nickelmetalhydride(NiMH)andlithiumion(Li-ion)batteriesoperatebestattemperaturesbetween25and408Candatthesetemperaturestheyachieveagoodbalancebetweenperformanceandlife.Itisdesirabletohaveatemperaturedistributionof58Cfrommoduletomodule.Inavehicle,inordertoachievetheoptimumperformancefromabattery,athermalmanagementsystemisnecessaryto:(1)regulatethebatteriestooperateinthedesiredtemperaturerange;and(2)toreduceuneventemperaturedistribution.Inordertoquantifytheimpactoftemperatureontheperformanceofabatteryandthusonthevehicle,tempera-turedependentbatteryperformancemodelsareneeded.TheNationalRenewableEnergyLaboratory’s(NREL’s)advancedvehiclesimulator(ADVISOR)[1]predictsbatteryandvehicleperformanceforconventional(e.g.non-electri-fiedvehiclesontheroadtoday),hybrid,electricandfuelcellvehiclesastheyvarywithdrivecycle.ThereareseveralJournalofPowerSources110(2002)377–382E-mailaddress:[e-mail](A.A.Pesaran).0378-7753/02/$–seefrontmatter.PublishedbyElsevierScienceB.V.PII:S0378-7753(02)00200-8batteryperformancemodelsdevelopedatNREL[2]forADVISOR.Thesebatterymodelsoutlinedinacompanionpaperinthispublicationare:resistance,resistance–capaci-tance,partnershipforanewgenerationofvehicles(PNGV)model,fundamentallead/acidandneuralnetwork.EachoftheseperformancemodelshasanassociatedthermalmodelinADVISOR.Inordertoevaluatethethermalperformanceofbatterymodelsandpacksandtocomeupwithimproveddesigns,NRELhasbeenusingcomputeraidedengineeringtools.WehaveextensivelyusedANSYS,afiniteelementanalysis,toanalyzeanddesignbettermodulesandpacks.Inthispaper,detailsofthethermalmodelsforADVISORwillbediscussedandtypicalresultsareprovided.Interestingresultsfromourfiniteelementthermalanalysiswillalsobeprovided.2.BatterythermalmodelsinADVISORTherearefiveMatlab-basedbatterymodelsavailableinADVISORasdescribedbyJohnson[2]:1.aninternalresistancemodel;2.aresistance–capacitancemodel;3.aPNGVcapacitancemodel;4.aneuralnetworklead/acidmodel;5.afundamentallead/acidbatterymodel.Thefirstthreeareequivalentcircuitmodels.ADVISORhasamainlumpedcapacitybatterythermalmodelandanintegratedthermalmodelinthefundamentallead/acidmodel.Thismainthermalmodelisbasedonasimplelumpedcapacitanceapproach.NothermalmodelwasusedintheneuralnetworkmodelinADVISOR.Batterytestdataatanoperatingtemperatureof258Cwasusedtotraintheneuralnetworkmodel.Duetothelimitedtemperaturerangeoftestdataavailableatthetimeofthetrainingofthemodel,themodeldidnotshowsensitivitytotemperature.Thermalbehaviorofabatteryinthefundam