GT-POWER自學(xué)很有用哦

1GT-POWERTRAININGEnginePerformanceAnalysisGammaTechnologies,Inc.Allinformationcontainedinthisdocumentisconfidentialandcannotbereproducedortransmittedinanyformorbyanymeans,electronicormechanical,foranypurpose,withouttheexpresswrittenpermissionofGammaTechnologies,Inc.2GT-POWERContent(Basic)GT-POWERApplicationsSolutionMethodDiscretizingaModelOrificeConnectionsPipesCylinderPortsValvesInjectionIn-CylinderHeatTransferCombustionCylinderPressureAnalysisFlowSplitsConvergenceAdvancedTopics3GT-POWERContent(Advanced)PipeEquations(ref)

TurbochargersHeatExchangersSIFuelPuddlingModelEGRSITurbModelDIJetModelExhaustAftertreatmentAcoustics(Non-linear)Acoustics(Linear)ModelCorrelationTransientSimulation3-DCombustion(KIVA)MeanValue/RealTimeCFDCouplingBasicTopics4GT-POWERApplications5BenefitsOfSimulationSavesTimeandMoneyShorteneddevelopmentcycleReducesnumberofprototypesrequiredOptimizationofdesignwithminimalprototypingandlaboratorytestingExcelsWhereTraditionalMethodsLackProofofconceptValidationandsensitivitystudiesComponentmatchingwhennohardwareyetavailableAnalysisofstubbornperformanceproblemsSimulationofunusualambientconditions:composition,temperature,pressure6ModelFidelityModelDetailReal-TimeModels(controlsmodeling)1-DGasExchange

(GT-POWER)3-DCFD(KIVA)1101001000xRealTimeCPUTime

vs.ModelTypeBlackBoxModels0.11-DMeanValueModel(GT-POWER)OriginalFigureprovidedbyFEV7GT-POWERApplicationsEngineperformanceanalysisBasedon1-DfluiddynamicsVeryflexible,toallowstudiesofadvancedconceptsDetailedthermodynamicsManycombustion&emissionsmodelsThermalanalysisAcousticsEnginecontrolanalysisviaSIMULINKCFDflowanalysisvia STAR-CD,fluentandKIVA8SolutionMethod9EngineCycleSimulationSolves1-DimensionalequationstopredicttheflowratesintheintakeandexhaustsystemsIn-Cylindermodelingofcombustion,pressure,heattransfertocylinders,workDetailedsub-models:turbos,acoustics,catalyst,etc.Enableusertofindbestbalancebetweencomplexityandaccuracy,takingintoaccountavailableinputdataanddesiredoutputs10FlowSolutionWholesystemisdiscretizedintomanysmallsub-volumes,connectedbyboundariesEachenginecycleisdividedintomanysmalltimestepsScalarVariables

densityinternalenergypressuretemperatureetc.VectorVariables

massfluxvelocityetc.StaggeredGrid11FlowSolutionSimultaneoussolutionofthreeequationsateachtimestep:Continuity(conservationofmass)Energy(conservationofenergy)Momentum(conservationofmomentum)Mass&Energysolvedinsub-volumes(scalars)Pressure,Temperature,Speciesconcentrations,etc.Calculatedatthecentroidofeachsub-volumesConsidereduniform(1D)throughoutthesub-volumeMomentumsolvedatboundaries(vectors)Velocity,Massflux,Massfractionfluxes,etc.12FlowSolutionExample13SolutionAtEachTimestepThesolutionisNOTaniterativenumericalprocess(asisthecaseinaCFDtypesimulation)Thesolutionisbasedonthestateofthesystemattimet0andiscalculatedforanewtimet1Thenewtimet1mustbecloseenoughtotimet0toensurethesolutionisvalidThismaximumtimestepisalwayscalculatedateachtimestep14TimeStepsGT-POWERremainsstablebychoosingitstimestepssuchthattheCourantnumberislessthanorequalto.8,wheretheCourantnumberisdefinedas:

Dt = TimeStep Dx = DiscretizedElementLength u = FluidVelocity c = SpeedofSoundTimestepiscalculatedforeachsubvolume:smallestoneisappliedtoentiresystem15PredictionMethodologyStepssimilartophysicalprocessesinengineSolveflowequationsPredictsairandfuelinthecylinderCombustionModelfuelenergyreleaseintopressureandtemperatureCalculateIMEPMEP=Workpercycle/CylinderVolumeDisplacedperCycleSubtractfrictionandauxiliarylossesCalculateBMEP,torque&power16DiscretizingaModel17IntakeManifoldComponents18ModelDiscretizationDiscretizationisthesplittingoflargepartsintosmallersectionsinordertoimprovethemodel’saccuracy.CoarsediscretizationresultsinlargertimestepsandfasterexecutionattheexpenseofmodelaccuracyFinerdiscretizationresultsinbetteraccuracyandfrequencyresolutionattheexpenseofexecutiontimeBeyondacertainlimit,furtherreducingthesub-volumesizedoesnotbringbenefits.19DISCRETIZATIONPRACTICESGeneralperformanceanalysis:Intakesystemdiscretization~=0.4*cylinderboreExhaustsystemdiscretization~=0.55*cylinderboreDetailedacousticanalysisUse?discretizationlengthfromgeneralDonotbuildmodelswith1or2veryshortelements.20ShortConeDiscretizationModelashortconeasasharpcontractionwithanorificeconnectionifeitherofthefollowingcriteriaaretrue:L1<DXHalfAngle>15deg.FlowcoefficientCd=1intheconvergingdirection;Cd=“def”inthedivergingdirection.21ExerciseDiscretization.gtm22OrificeConnections23OrificeConnectionsOrificeconnectionsallowfluidtoflowbetweentwoadjacentcomponentsDischargecoefficientsineitherdirectionmaybeuserdefinedorsetto“def”.“def”dischargecoefficientsarecalculatedassumingthatthepartsareconnectedasshownbelow.CalculateddischargecoefficientscanbeseenintheGT-POSTRLT-ViewerMode24OrificeConnections(Cont.)Adiameterrestriction,d,maybespecifiedorsetto“def”ifthereisnorestriction.OrificesmaybespecifiedwithmultipleholesD1dD2D1dD225DischargeCoefficientsThedischargecoefficientistheratiooftheeffectiveareatothereferencearea.Referenceareaisderivedfromthediameterforthrottles,orificesandvalves. (morelateronvalves)CD=AEFFECTIVE/AREFERENCE26ExpansionandContractionsWhenmodeling,alwayspaycloseattentiontoanypointofareachangeLargesourceofpressuredropintypicalmanifoldsPressurewavesreflectoffoftheseareachangesduetothechangeinvelocityImportanttomanifoldwavedynamics—bothforperformance(breathing)ANDacousticresults27FlowContractionsAtdiscontinuouscontractions,velocitytowardthecenterresultsinavenacontracta.CdinflowconnectionscharacterizesthevenacontractaDiscontinuoustransitionofupstreamareatoboundaryarearesultsinaflowcontractionloss.Cd=defassumesdiscontinuity.28FlowContractionsSmoothtransitionfromupstreamareatoboundaryarearesultsinnocontractionloss.Cd=defCd~1.0.29MeasuringDischargeCoefficientsFlowcoefficientsacrossvalves,throttles,andorificesmaybederivedfrommeasureddatausingtheisentropicvelocityequationforflowthroughanorifice.

=ActualMassFlowRateCD =DischargeCoefficientAR =ReferenceFlowArearo =UpstreamStagnationDensityPR =AbsolutePressureRatio(StaticOutlet/TotalInlet)R =GasConstantTo =UpstreamFluidTemperature=SpecificHeatRatio(1.4forAir@300K)30Pipes31TypesofFlowLossesInapproximateorderofPriority……ExpansionLossesContractionLosses(seeflowconnections)SurfaceFrictionLossesBendsandtaperedpipelosses32ExpansionandContractionsWhenmodeling,alwayspaycloseattentiontoanypointofareachangeLargesourceofpressuredropintypicalmanifoldsPressurewavesreflectoffoftheseareachangesduetothechangeinvelocityImportanttomanifoldwavedynamics—bothforperformance(breathing)ANDacousticresults33FlowExpansionsVelocitybecomesaturbulenteddy.Kineticenergyisturnedintoheatviainternalandsurfacefriction.ExpansionLossVelocityisexpanded,Venturi-like.Kineticenergyisrecoveredasstaticpressureincrease:PressureRecovery34FlowExpansionsExpansionlossdoesnotoccurinthethroat.EffectiveArea=GeometricArea(Cd=def=1.0)MomentumEquationcalculatespressurerecoveryandexpansionlossfromarearatioofboundaryareatodownstreamarea35SurfaceFrictionSurfacefrictionDependsofsurfaceroughnessandcrosssectionNon-Circularcross-sectionsavailable(‘Pipe**Bend’)forincreasedfrictionandheattransferManifoldflowlossesareoftenmistakenlyattributedtosurfacefrictionProbablyduetoourfamiliaritywithflowinliquidswherefrictionisadominantlossfactorItsinfluenceontheheattransfercoefficientistypicallymoreimportantthanpressuredropEspeciallyimportantforaccuratepredictionofturbineandcatalystinlettemperatures36SteadyFlowSimulationGT-POWERissometimesusedtomodelsteadyflow,suchassteadypressuredropthroughamanifoldorairboxFlowControl(RunSetup)hasaGlobalFrictionMultiplierandGlobalHeatTransferMultipliertoaccountfornon-steadyflow(oscillating)non-fullydevelopedflowDefaultsettingsare>1.0toaccountfortypicalenginecharacteristicsForsteadyflowtestsandliquids,setthesevaluesto1.037BentandTaperedPipesPressureLossCoefficient“def”optiontoautomaticallycalculateKdP=(0.5rV2)KMostenginesarecarefullydesignedNOTtohavesharpbends–thereforeminorfactorFlowthroughpipesofincreasingdiametersdonot“perfectly”convertdynamicpressure(i.e.kineticenergy)backintostaticpressure38Accountsfor:Internal/ExternalHeatTransferHeatTransfertoNeighbor(conductance)ThermalCapacitance(transientoption)InitialTemperature(transientoption)RecommendedformostexhaustsystemsBettertoguessexternalheattransfercoefficientthantoguessinternalwalltempEspeciallyimportantforturbinesIfwalltemperatureatalocationisknown,bettertouseittocalibrateexternalconvectioncoefficientthantoimposethewalltemperaturePipe****ThermalWallSolver39ExerciseSteadyFlow.gtm40CylinderPorts41Ports–UsingPipe***LimitationswhenusingPipe***templates:WalltemperatureisimposedasonevalueHeattransferfromthebackofthevalveisnotmodeled(whenusingEngCylTWall–morelater)ValveguideheattransferareaisnotmodeledTurbulencecausedbyvalveguideisnotmodeled 42Ports–UsingPipe***Recommendedpractices:Frictionmultiplier*=0Pressurelosscoefficient*=0Intakewalltemperature~=450KExhaustwalltemperature~=550KIntakeandExhaustheattransfermultiplier=1.5to2

*-Veryimportant!43Valves44Valves-TypesValveliftbasedonrotationalpositionofcrankshaftValveCamConn,ValveCamPRConn,ValvePortConn,ValveCamUserConnValveliftbasedonelapsedtimesincesignaltoopenorcloseValveSolenoidConnCheckvalve-EffectiveareabasedonfloworvalvedynamicsValveCheckConn,ValveCheckSimpleConn

45ReferenceAreaReferenceareamaybeeitherconstant(derivedfromthediameter)orvariable(derivedfromthecurtainarea)forvalves.Forconstantreferencearea,thefirstdischargecoefficient(at0lift)mustbe0.0.Forcurtainarea,thefirstdischargecoefficient(at0.0lift)mustbegreaterthanzero.46CamLiftProfileValveCam***TemplatesLiftprofileinputasfunctionofcamangleorcrankangleProfileshiftedby“CamTimingAngle”47ExerciseSavePorts.gtmasValveTimingAdvanceandchangeintakevalvetiming+2degrees.RepeatasandValveTimingRetardusing-2degreesCompareresultingvolumetricefficiencycurves.48Injection49INJECTION(InjProfileConn)ProfileinjectionTypicaluse:directinjectionInputs:FuelmassperstrokeVaporizedfuelfractionInjectionratevs.crankangleisspecifieddP,P,MassflowrateChecknozzledischargecoefficient (between0.65and0.75)Smoke-limitsmaybeimposed50INJECTION(InjAF-SeqConn)SequentialfuelinjectionTypicallyuse:sequentialportinjectionInputs:InjectordeliveryrateAirFuelRatioNumberofinjectorssharingthesamemassflowsensorStartofInjectionVaporizedfuelfractionInjectiontimingcanbemadeRLTDependentIdealfordevelopingfuelmapsVaporizationeffectsonvolumetricefficiencywillbemodeledmoreaccurately51SIInjectionDeliveryRateEstimateTypically,thelongestpulsedurationincrank-angledegreesisatthehighestRPM,WOTTypically,thelongestpulsedurationincrank-angledegreesisaboutequaltothedurationoftheintakevalveopening:180to210degrees52SIInjectionDeliveryRateEstimatemDelivery =injectordeliveryrate(g/s)hV =volumetricefficiency(fractional)rref =referencedensityforvolef(kg/m3) (typically1.16kg/m3forambient)NRPM =enginespeed(RPM)VD =enginedisplacement(liters)F/A =enginefuel-to-airratio#CYL =numberofcylindersPulseWidth =injectionduration(crankdegrees)Note:The6intheequationcomesfromunitconversions53In-CylinderHeatTransfer54CylinderWallTemperaturesTwallAveragetemperaturespecifiedforhead,piston,andcylinderwallsonly.DoesNOTtransferheatfrombackofvalvestopipeTWallDetailPistonandcylinderwallsspecifiedinzonesHeadandvalvefacetemperaturesspecifiedseparatelyTransfersheatfromheadonlytoPipePortpartsTWallSolnWalltemperaturesolvedusingfiniteelementmodelFEmodelbuiltautomaticallyfromuserspecifieddimensionsTransfersheatfromheadonlytoPipePortparts55CylinderHeatTransferWoschniHeattransfercalculatedfromcorrelativeWoschnimodelFlowMoredetailed,spatiallyresolvedheattransfermodel(Swirldataneeded).UserUsermaywritetheirownheattransfermodelandincorporateitintoGT-POWER.HGProfileHeattransfercoefficientsmaybeimposedasafunctionofcrankangle(typicallyfromCFDanalysis).56Combustion57Combustion-TermsBurnRateRateatwhichfuelandairisconvertedtocombustionproductsInGT-POWERthisistherateatwhichthefuel-airmixtureisaddedtotheequilibriumequationsUsedforALLcombustionmodels(eitherimposedorpredicted)HeatReleaseRateChemicalenergyreleaserateDiffersfromBurnRateduetopartialcombustionHeatreleaseratesareNOTusedinGT-POWER(availableasaresultfromcylinder)58Combustion-Terms59Combustion-TermsNon-predictivecombustionTheinstantaneousburnrateisdirectlyimposedasasimulationinputThetotalamountofenergyreleasedfromthefueldependsonlyonthemassoffuelandairinthecylinder.PredictiveCombustionTheinstantaneousburnrateispredictedbyaphysicalmodelwithinGT-POWERbasedonvarioussimulationinputsandresultsSemi-predictivecombustionTheinstantaneousburnrateisimposedusingnon-predictivecombustionmodelsTheinputparametersthatdefinetheburnraterespondtochangesinmodelinputs/resultsviaalookuporotherrelation(non-physicalmodel)60Combustion-TermsForwardRun–Cylinderpressureispredictedbasedonanimposed(orpredicted)combustionburnrateReverseRun–CombustionburnrateiscalculatedfrommeasuredcylinderpressureFWDREV61CombustionAvailableForwardCombustionModels:62SIWiebeModelNon-predictiveburnrateimposedaccordingtotheSIWiebefunction(50%burned,10-90%duration,exponent)VeryfastexecutionCombustion-SI63Combustion-SISITurbulentFlameModelPredictiveburnratetakingintoaccount:cylindertemperatureandpressurecompositioninthecylinderincludingfuel,freshair,andegr/residualsSparktiming,position,andgapFuelpropertiesFlamearea/WallwettedareaIn-cylinderflowSlowexecutionCombustionchambergeometry(headandpiston)maybereadfroman.STLfileorgeneratedautomaticallyfromdimensionsenteredbytheuser64SINOxandKnockPredictiveNOxModel(extendedZeldovich)TemperaturetrackedinmanyzonesSensitivetoheatreleaserateandcomposition(includingEGR)Sensitivetopressure,temperature,equivalenceratio,anddilutionratioPredictiveKnockModelCorrelatepredictedknockindextomeasurementsofknockinitiationandstrengthKnockindexpredictsknocktrendsCylinderwalltemperaturesshouldbespecifiedbyeitherthedetailedorsolutionreferenceobjects65Combustion-DIDieselDIWiebeModel(Nodefaultentries)Non-predictiveburnrateimposedaccordingtothethree-termDIWiebefunctionVeryfastexecutiontime66Combustion-DIDieselDIWiebemodelwithdefaultentries“def”allowedforallparametersSemi-predictive,includeseffectsofcylinderpressure,fuelinjectionrate,andinjectiontiming.Useonlywhenanaccurateinjectionprofileandtimingareavailable.Notnecessarilyaccurateforidle,highegr,ornon-dieselfuels.Veryfastexecution67Combustion-DIDieselDIJetModelPredictiveburnratemodel:Takesintoaccount:CylinderpressureandtemperatureInjectiontiming,rate,velocity,andplumeshapeCompositioninthecylinderincludingfuel,freshair,andEGR/residualsIn-cylinderflow(SwirlandTumble)SlowexecutionPredictiveNOxModelPredictiveSOOTModel(trendsonly)68Combustion–AnyTypeMulti-WiebeModelNon-predictiveburnrateimposedbasedonsumofmultipleSIWiebefunctions(upto6)DevelopedforDIwithmultiplepulses(butcanbeusedanywhere)Fastexecution69Combustion–AnyTypeCombustionProfileNon-predictiveburnrateimposeddirectlyasfunctionofcrankangleCanbeusedforanytypeoffuelorinjectionVeryfastexecutionGT-POWERcanperformareverseruntocalculatetheburnratedirectlyfrommeasuredcylinderpressure70PredictiveorNon-predictive?The“real”combustionburnrateisdependentonalonglistofvariables(i.e.trappedcylinderconditions,fueling,spark/injectiontiming,in-cylinderflow,etc.)Predictivephysicalmodelisidealbut...MoredifficulttobuildandcalibratemodelSlowercomputationtime (UseMaster/Slavewhenapproprate)UsepredictivecombustionwhenvariablestobestudiedhaveadirectandsignificanteffectonthecombustionburnrateDIInjectionrateandprofileSICombustionchambershape71Usenon-predictivecombustionwhenevertheinfluenceontheburnrateofthevariablestobestudiedcanbe:Neglected(notthedominantfactor)Modeledusinglookups,etc.(semi-predictive)Examplesofnon-predictiveapplicationsOptimizingcamtimingorrunnerlengthIntakeorexhaustbackpressure/acousticsTurbochargermatchingPredictiveorNon-predictive?72Semi-predictiveExampleSIWiebemodelwhere3Wiebeparametersarepredictedusingneuralnetworksasafunctionof:EngineSpeedTrappedMassTrappedBurnedGasFractionTrappedLambdaTrappedTemperature(IVC)TurbulentKineticEnergy(IVC)NeuralnetworkstrainedusingexperimentaldataLargeDOEofoperatingpoints (rpm,load,VVT,lambda)Burnratesdeterminedfrom measuredcylinderpressureWiebeparameters“fit”to apparentburnrate73ExerciseUseburnratefromDI-Injection.gxtodetermineDI-WiebecoefficientsusingWiebeComb.xls74CylinderPressureAnalysis

(ReverseRun)75ReverseCombustionTwomethodsprovidedtocalculatecombustionburnratefrommeasuredcylinderpressurePre-processingmethodSimulationbasedmethod(ThreePressureAnalysis)BothmethodsusesameburnratecalculationsResultofcalculationistheburnraterequiredinGT-POWERtoexactlyreproducethemeasuredcylinderpressureinaforwardrunMostothertoolsperformaclassicalheatreleaseanalysis76ReverseCombustionClassicalHeatReleaseAnalysis(methodusedinmostothercylinderpressureanalysistools):Resultis“apparent”heatrelease(NOTburnrate)HRArequiresassumptionsandsimplificationsDoesnotsolvethesameequationsasforwardanalysisRarelyreproducesexactlytheinputpressure77GT-POWERusessamemethodologyforbothforwardandreverserunsFullChemistryandThermodynamicsIterativemethod–fuelmassburnedisvariedwithineachtimesteptomatchmeasuredcylinderpressure

ReverseCombustion78ReverseCombustion79ResultisGT-POWERburnrateCapableofexactlyreproducingtheinputpressureReverseCombustion80ReverseCombustionDifferencebetweenthetworeverserunmethodsavailableinGT-POWERishowburnratecalculationinputsareacquiredPre-processing:Allcalculationinputsaremodelinputs‘EngBurnRate’templateVeryfastcalculation(noconvergence)TPAMostcalculationinputsaresimulationresults‘EngCylCombPressure’templateNormalmulti-cyclesimulation–requiresconvergence81PRE-PROCESS(EngBurnRate)AllInputsdefinedin‘EngBurnRate’template:Air,Fuel,ResidualsinCylinderHeatLosstoCylinderMeasuredCylinderPressure82Pre-process(EngBurnRate)Pre-processsimulationrunCalculationsdoneif‘EngBurnRate’inprojecttreeOutputsofcalculation:Plotsandtablesmodelname.gxCheckintegrityofinputdataShowresultingburnrateandotherquantitiesCombustionProfilesreadyforuseinforwardrunmodelname_prof.datImportintoGT-ISEContains‘EngCylCombProfile’objectsreadytocopyintoanymodel83ReverseRunPlots/Tables84ReverseRunCombustionProfiles85ThreePressureAnalysis(TPA)SinglecylindermodelofanengineonteststandThreemeasureddynamicpressuresaremodelinputs86Measuredcylinderpressureisenteredin‘EngCylCombPressure’combustionobjectThreePressureAnalysis(TPA)87ThreePressureAnalysis(TPA)Normalsimulationrun(rununtilconvergence)Foreachcycleofthesimulation:ApparentburnrateiscalculatedatIVCusingthecylindertrappedquantitiesandheattransferfromsimulationBurnrateisimposed(exactlylike‘EngCylCombProfile’)Sameoutputsas‘EngBurnRate’plusfolderof“PressureAnalysis”RLT’sinEngCylinderpart

88ThreePressureAnalysis(TPA)89TPAvs.Pre-Process90ExerciseModifyDI-Injection.gtmtosaverequireddataforTPA,thenperformTPAanalysis.91FlowSplits92FlowsplitUsesFlowsplitsareusedtocreatevolumeswhichconnectthreeormorepipesUsedtodiscretizelargevolumesIntakeplenumsAirboxesMufflershellsUsedforotherspecializedvolumesHelmholtzresonatorsPerforatedpipesInlet/outlettocatalystsandheatexchangers93FlowsplitDiscretizationDiscretizationIssuesUnlikePipecomponents,flowpslitsarealways1subvolumeChooseflowsplitsizetobeclosetotargetdiscretizationAlwaysplaceflowconnectionsatlocationsofareadiscontinuity(usecenterline)Beespeciallycarefulaboutmaintainingdiscretizationlengthinsensitiveareasofthemodelandareasathighervelocities:IntakePlenumandExhaustManifoldCollectors94HowToDiscretize

AFlowSystemUsetargetdiscretizationlengthasroughFlowSplitsizePlaceboundariesatveryedgeofpipesFlowmustbenormaltoboundariesConsiderationsFlowSplitis1-Dapproximationof3-Dphenomena:maynotbeperfectMayrequirecouplingwithCFDtoresolvecertain3-Deffects95FundamentalTheoryofFlowsplitsMass,speciesandenergyarecalculatedsameasinpipesMomentumcalculationsareatOrificeConnsAnglesusedtocalculatethetransferofmomentumtootheropenings(directionaleffects)ExpansiondiametersusedtocalculatemomentuminFlowSplitafterflowenters96FlowsplitAttributesVolumeAngles(X,YandZ)usedtocalculatepressurechangesduetodirectionalchangesExpansiondiametersUsedtocalculateflowlossesattheflowsplitopeningsCharacteristicLengthsUsedtocalculatewaveresponseOtherssimilartoPipe*templates97FlowsplitAnglesPortanglesrelativetocoordinateaxes.Anglesmustbespecifiedbetween-180and180degrees.AllvectorseitheralloutoforallintotheFlowSplit.98ExpansionDiameterExpansiondiameterateachportdiametertowhichfluidmayexpanduponenteringtheFlowSplitvolumeusedtocalculatekineticenergylossesduetoexpansionwhenfluidenterstheFlowSplitalsousedtocalculatedischargecoefficientswhenthedischargecoefficientsaresetto“def”ConsiderationsApplyshortcone-discretization(e.g.,catalyticconvertersandintercoolers)Maynotbeapparent(considerlimits&studyrangebetween)99CharacteristicLengthsCharacteristiclengthateachportlengththatfluidenteringtheFlowSplittravelsbeforeithitsawalloraportontheoppositesideoftheFlowSplitusedtocalculatethepropagationand/orreflectionoftravelingwavesMoreinfoSometimesunclear(considerlimits&studyrangebetween)Overallsystemperformanceusuallyinsensitivetoinput100GeneralFlowsplitExamplesExample:Skewedjunctionoftwopipes.ExpansionDiameters:DIAC2=DIAC1,DIAC3seefigureCharacteristicLengths:DX2=DX1,DX3seefigure101GeneralFlowsplitExamplesPulsationsoutofphasefromconnections2&3TypicalcaseExamples:Manifolds,Collectors,etc…ExpansionDiameters:

DIAC2=DIAC3=DIAC1=D1CharacteristicLengths: DX2=DX3=DX1102GeneralFlowsplitExamplesPulsationsinphasefromconnections2&3SpecialflowcaseExample:Bifurcatedflowsuchasmultipleintake/exhaustvalves.ExpansionDiameters: DIAC2=DIAC3~=Diameterforcirclewith?areaofD1 DIAC2=DIAC3<DIAC1CharacteristicLengths: DX2=DX3=DX1Flowfromports2&3expandintosharedvolume,DIAC2,3~=SQRT(0.5*D1^2)103PredefinedFlowsplitsOrificesplacedatentryplanesofanytypeofFlowSplitaredefinedbyOrificeConnconnections.104CommonFlowsplitMistakesFlowsplitAnglesIncorrect:FlowintoandoutoftheflowsplitinthesamedirectionCorrect:DrawvectorspointingeitheralloutoforallintotheFlowSplitthrougheachport(Vectordirectionsdependonlyongeometry,notonflowdirection)ExpansionDiametersIncorrect:ExpansiondiameterissamesizeasinletpipewhenthereisanareadiscontinuitypresentCorrect:Expansionareashouldbedeterminedbycross-sectionINSIDEtheflowsplitboundary,notontheadjacentparts1053-DFlowsplitViewerToolthatdisplaysgraphicallytheinputFlowSplitmustbeapartonmapandconnectedtootherparts.Usebuttonofwireframeboxontoolbar{Demonstration}106Convergence107ConvergenceSimulationmayrunuptothemaximumdefinedsimulationdurationMaybeconfiguredtostopwhenvaluesdonotchangefromcycletocycleChangeinaverageflowratesforeverypartChangeinaveragepressureforeverypartIfdP>1%,dFcriteriaistightenedby4XChangeinaveragefluidtemperatureAnyRLTvalue108PipeEquations(Reference)109Laminarregion,ReD<2000TransitionregionTurbulentregion,ReD>4000Largerof:where: ReD Reynoldsnumberbasedonpipediameter D pipediameter

h roughnessheightPipe****FrictionLosses110ForwardandReversePressureLossCoefficientwhere:

p2 totalpressureatinlet

p1 totalpressureatoutlet

inletdensity

V1 inletvelocityCalculatedbasedonsynthesisofvariouspu