变速器齿轮啸叫CAE分析与样件试验的相关性评估

400CommonwealthDrive,Warrendale,PA15096-0001U.S.A.Tel:[phone]Fax:[phone]Web:(3)reviewsbyqualifiedreviewers.Allrightsreserved.Nopartofthispublicationmaybereproduced,storedinaretrievalsystem,ortransmitted,inanyformorbyanymeans,electronic,mechanical,photocopying,recording,orotherwise,withoutthepriorwrittenpermissionofSAE.Forpermissionandlicensingrequestscontact:SAEPermissions400CommonwealthDriveWarrendale,PA15096-0001-USAEmail:[e-mail]:724-772-4891Tel:[phone]Formultipleprintcopiescontact:SAECustomerServiceTel:877-606-7323(insideUSAandCanada)Tel:[phone](outsideUSA)Fax:[phone]Email:[e-mail]0148-7191Copyright©2005SAEInternationalPositionsandopinionsadvancedinthispaperarethoseoftheauthor(s)andnotnecessarilythoseofSAE.Theauthorissolelyresponsibleforthecontentofthepaper.AprocessisavailablebywhichdiscussionswillbeprintedwiththepaperifitispublishedinSAETransactions.PersonswishingtosubmitpaperstobeconsideredforpresentationorpublicationbySAEshouldsendthemanuscriptora300wordabstractofaproposedmanuscriptto:Secretary,EngineeringMeetingsBoard,SAE.PrintedinUSA2005-01-2459CorrelationofCAEModelstoPrototypeTestsForTransmissionGearWhineAssessmentYangDai,DavidJ.SchipperandDasM.RamnathFordMotorCompanyCopyright©2005SAEInternationalABSTRACTRecentadvancesinAutomaticTransmissionEngineeringOrganization(ATEO)atFordMotorCompanyhaveintegratedCAEgearwhineanalysisintothetransmissiondesignprocess.Detailedmodelscanbedevelopedtopredictradiatednoiseundermanytypesofloadingconditions,includingdynamicmeshforceexcitation.OftencorrelatingtheanalyticalFRFswiththeresultsobtainedfromtestingaprototypevalidatestheanalyticalmodelanditsassumptions.TheevaluationanddeterminationofcorrectandappropriateCAEpropertiescanbeextremelydifficultforeventheexperienceddesignengineer.Thispaperdiscussestheapplicationofstructuraloptimizationtechniquestotheassumedmaterialpropertiesandmodelparameterstominimizethedifferencebetweentheanalyticalresultsandtestdataoveralargefrequencyrange.Intheprocedurethematerialpropertiesandthemodelparametersaremanuallyandautomaticallyupdateduntiltheanalyticalresultshaveabest-fitmatchtothetestdata.INTRODUCTIONRadiatednoiseinducedbythepowertrainsystemisbeingincreasinglystressedatFordMotorCompany.Inthepowertrain-inducednoise,gearwhineisoneoftheconcerns.GearwhineisthenoiseproducedattheharmonicsofthetoothpassingfrequencyandisgeneratedbyTransmissionError(TE)attheloadedgearmeshes.TheradiatednoisecanbecomeunacceptableifthemeshforcegeneratedbytheTEexcitesgearboxresonanceswhichpassintothepassengercompartmentthoughair-borneorstructure-bornepaths.Twotypesofgearsystemsarefoundinfrontwheeldriveautomatictransmissions:planetarysets,whichareusedforshiftingschemes;andtransfersets,thatmovethepowerfromtheplanetaryaxistothedifferential/halfshaftaxis.Thelaterisaddressedinthispaper.RecentadvancesinComputerAidedEngineering(CAE)makeitpossibletomodeltheinternalgearsystemandpredictthemagnitudeofthemeshforcedevelopedatthegearmeshinthefrequencydomain.Unlikeprototypetesting,CAEanalysiscanbeperformedinthebeginningstagesofthedesigntoidentifythepartsofthesystemthataffecttheforcemagnitudeandprovideearlyperformanceassessmentanddesignguidance.However,numericalsimulationofdynamicresponsesbasedonCAEmodelisofteninaccurateduetosimplifyingassumptions,uncertainty,andcomplexity.Thismakesitnecessarytocorrelatethemodelwithprototypetestswhenpartsareavailabletoascertainthevalidityofthemodelingassumptions.Thecorrelationprocedureiscomposedoftwoportions,correlationandmodelupdating.ThecorrelationisessentiallythecomparisonbetweentestprototypeandCAEmodelusingsomecertaincriterion,whiletheupdatingischangestothemodelneededtogetbettercorrelationorchangestoimprovethedynamics.ThemostcommonlyusedcorrelationcriteriaareCross-OrthogonalityCheck(COC),ModalAssuranceCriteria(MAC),andFrequencyResponseFunction(FRF)-BasedCriterion(FBC)usedinthepaper.TheCOCusestheorthogonalitypropertyofmodalanalysistheory,wherethemassmatrixpre-andpost-multipliedbythemodalmatricesyieldsadiagonalmatrix.TheCOCisdefinedas]][[][aTtMFF=COC(1)where][tFand][aFaretestandanalyticalmodeshapematrix,respectively.][Misthemassmatrixofadynamicallyreducedsystem,calledTest-Analysis-Model(TAM).Thecrossorthogonalitybetweentestandanalysismodesisthetraditionalcriteriathatiscommonlyusedintest-analysiscorrelationformodaltesting[1].TheCOCmatrixshouldbegreaterthan0.9onthediagonalandlessthan0.1foroffdiagonalterms.TheadvantageoftheCOCisthatitisabletocorrelateboththenaturalfrequenciesandthemodeshapes.OneofdisadvantagesoftheCOCisthatitneedstocreateanappropriateTAM.Manyr