基于包络分析的齿轮故障监测及基于人工神经网络的齿轮故障分类.pdf

《基于包络分析的齿轮故障监测及基于人工神经网络的齿轮故障分类.pdf》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、EnvelopeAnalysisforGearFaultsDetectionandClassificationUsingArtificialNeuralNetworkAThesisSubmittedtoChongqingUniversityinPartialFulfillmentoftheRequirementfortheDegreeofMasterofEngineeringBySimonRomliHanSupervisedbyProf.ShaoYiminProfMajor:MachineDesign&TheoryCollegeofMechanic

2、alEngineeringofChongqingUniversity,Chongqing,ChinaJune,2011基于包络分析的齿轮故障监测及基于人工神经网络的齿轮故障分类重庆大学硕士学位论文学生姓名：SimonRomli指导教师：邵毅敏教授专业：机械设计及理论学科门类：工学重庆大学机械工程学院二O一一年六月AThesisfortheDegreeofMScofCQUABSTRACTABSTRACTGearshasbeenwidelyusedasintegralpartsformanyequipmentsandmachinesinvariousi

3、ndustries,Therapidgrowthanddevelopmentofmoderntechnologyhascausedtheequipmentstobecomemoresophisticatedandthereforeimposingaverychallengingtaskinmonitoringtheiroperatingconditions.Beingpartofthesecomplexmechanicalsystemsgearboxshouldbemonitoredcloselyandeffectivelysothatanydef

4、ectcanbedetectedandtimelycorrectedinitsearlystages.Theenvelopeanalysishaslongbeenwidelyusedforbearingsandgearfaultdiagnosisbasicallyinvolvingtheuseofresiduesignalobtainedaftersynchronousaveragingbeforebandpassfilteringaroundthesignificantgearmeshingcomponents.Howeverinsomeappl

5、icationsthesynchronousaveragingofgearvibrationsignalcanhardlybeachievedandthereforethereneedtobeothertechniquetofacilitateenvelopeanalysisforgearfaultdiagnostic.Thisstudyisthereforeaimedatsolvingtheabovementionedproblembyintroducingademodulationtechnique,whichinvolvebandpassfi

6、lteringaroundtheexitedstructuralresonanceawayfromanysignificantgearmeshingharmonics,todetectandclassifyingeachofthetwotypesoflocalgearfaults(geartoothcrackandspall)withdifferentseveritylevels.Itisknownthatalocalfaultongeartoothwillgeneratesalow-levelimpulseeverytimethefaultcon

7、tactsanothergeartoothsurfaceduringmeshingcausingamodulationeffectonthegearvibrationsignalandpossiblyexcitingthestructuralresonances.Theselow-energyimpactsarealwaysoverwhelmedbythehighenergyvibrationsignalfromothersourcesinthegearbox,sothemainideaistodevelopasuitabletechniquewh

8、ichcanextractthesedefectinducedimpulsesfromwhichthepresencean