基于神經(jīng)網(wǎng)絡(luò)方法的齒輪箱故障診斷研究.pdf

1、河北科技大學(xué)碩士學(xué)位論文基于神經(jīng)網(wǎng)絡(luò)方法的齒輪箱故障診斷研究姓名：時志素申請學(xué)位級別：碩士專業(yè)：測試計(jì)量技術(shù)及儀器指導(dǎo)教師：龐志鋒20090501河北科技大學(xué)碩士學(xué)位論文AbstractWiththedevelopmentofscience，Machineryhasdevelopedtothedirectionofhi曲performance，highefficiencyhighautomationandhighreliabilityT

2、hegearboxisusedinchangingtherotationalspeedandthetransmissionpowerforfixationrate，hightorquetightandstructurevirtueetcThegearboxisanimportantcomponentofmechanicaldeviceandbreakdowneasily；itsrunningStatushastheverytremend

3、ousinfluencetothecompletemachineoperatingperformance。Sotheimportanceoftheconditiontestingandfaultdiagnosisofitstransmissionsystemconsistsnotonlyincuttingtimeandcostofrepairing，butalsoinraisingdiagnosingaccurate，whichimpr

4、ovesrepairingqualityandacquireseconomicbenefitsdramaticallyVibrationsignalofgearboxisverycomplexItisdifficultifweonlyusetime—domainorfrequency—domainmeansanalyzingvibrationsignaltofindfaultaccuratelyIftimedomainanalysisa

5、ndfrequencydomainanalysisspecialtyareprovidedatthesametime，diagnosticveracityandreliabilitywillbegreatlyimprovedSoweputforwardandstudyanewfaultdiagnosistechniqueuneuralnetworkdiagnosisbasedontimedomainanalysisandfrequenc

6、ydomainanalysisVibrationmechanismofthefailuregearboxisanalyzedThebasictheoryofmulti—resolutionanalysisareintroducedThispaperproposedafeatureextractionmethodthroughtimedomainanalysisandfrequencydomainanalysisThispaperbuil

7、dsagearboxtestingsysteminlaboratoryandsimulates5modesofgearboxrepresentativefaults，andcollectsthesignalfromthegearboxAnalyzingthevibrationsignalbytimedomainanalysisandfrequencydomainanalysis，thenextractscharacteristicsas

8、theinputofneuralnetwork，andestablishesnetworkmodelThroughprocessingandanalysisofmanydataItshowsthatthenewmethodpresentedinthispapercansupplyaconvincinganalysismeansforgearboxfaultdiagnosisKeywordsgearbox；faultdiagnosis；t