欢迎来到天天文库
浏览记录
ID:52970092
大小:281.07 KB
页数:3页
时间:2020-04-05
《基于EMD和AR奇异值的柴油机故障诊断.pdf》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、机械设计与制造第4期230MachineryDesign&Manufacture2011年4月文章编号:1001—3997(2011)04—0230—03基于EMD和AR奇异值的柴油机故障诊断吴虎胜1吕建新战仁军吴庐山(’中国人民武装警察部队工程学院,西安710086)(河南农业大学,郑州450000)FaultdiagnosisbasedonEMDandARsingularvaluesfordieselenginewUHu—shengI,LVJian—xin1,ZHANRen-jun.WULu—shan(EngineeringCollegeofC
2、APF,Xi’an710086,China)(He’nanAgriculturalUniversity,Zhengzhou450000,China),⋯·●⋯·●⋯·●⋯·●⋯·●⋯·●⋯·●⋯·●⋯-●⋯··⋯··⋯··⋯··⋯‘●⋯-●⋯·●⋯·●⋯··⋯··⋯··⋯·⋯‘●⋯-●⋯·⋯·⋯·⋯‘·⋯‘●⋯‘●⋯·●⋯··⋯··⋯‘●⋯‘●⋯‘●⋯’●⋯··⋯·●⋯。●⋯‘●””●⋯,;【摘要】针对柴油机振动信号的非平稳特性和在现实条件下难以获取大量故障样本的实际情;况,提出一种经验模态分解(EmpiricalModeDecomposition
3、,EMD)、自回归(AutoRegression,AR)模型和;支持向量机(Suppo~VectorMachine,SVM)~合的柴油机故障诊断方法。运用经验模态分解方法对柴;?油机失火及气阀机构不同工况下的缸盖振动信号进行分析,计算各个内禀模态函数(IntrinsicMode?iFunctions,IMF)的AR模型参数向量以此组成初始特征向量矩阵,再计算此初始特征向量矩阵的奇异i值,并将其作为支持向量机的输入特征向量以判断柴油机的工作状态和故障类型。试验结果表明:该方i法在小样本情况下也具有较高的精度和较强的泛化能力。i关键词:柴油机;故障诊断
4、;AR模型;支持向量机(SVM);奇异值;经验模态分解(EMD)i【Abstract】Accordingtothenon-stationaritycharacteristicsofthevibrationsignalsfromthedieselienginewithfaultandthesituationthatit’hardtoobtainenoughaltsamples,adieselenginefaultdiag-:hostsmethodbasedonEmpiricalModeDecomposition(EMD),AutoRegression
5、(AR)andSuppo~SectorMachine(SVM)isproposed.~rstly,evibrationsignalsofthreeairvalvecle~r-culATesandthevibrationsnalsunde,.themisfireconditionaredecomposedinto0finitenumberofIntrinsicModeFunctions(1MF),÷thentheAutoRegressive(AR)m。delofeachlMFc。mponen£sareestablished.Theant。一regre
6、ssiveparameters÷;andthevarianATeofremnantareregaredastheinitialfeaturevectormatrixes.Thirdly,byapplyingthesingu一!÷laredecompositi。凡techniq“etheinitialfeat“reecf。rmatrixes,thesingularMesare。btained.一;;natty,thevaluesserveasthefaultcharacteristicvectorstobeinputtoSVMclassiferand
7、theworkingco,卜;●●;ditionandfaultsofthedieselengineareclassifed.Theresultsshowthatthismethodhave劬accuracy;:andgoodgeneralizationeveninthecaseofsmallnumberofsamples.:;Keywords:Dieselengihe;Faultdiagnosis;Autoregressionmodel;Supportvectormachine(SVM);;?Singularvalue;Empiricalmode
8、decomposition(EMD)中图分类号:TH16。TK428,TP206.3文献标识码:A,1引言奇异值熵实现了转
此文档下载收益归作者所有