,396 北京科技大学学报 第32卷 率均在96.259%以上;对叠加不同程度的高斯白噪 tionaries EEE Trans Signal P rocess 1993 41(12):3397 声数据进行了实验,当信噪比为一5db时,信号的总 [6]Fan H.Meng Q F.Zhang YY.et al Matching pursuit based on 体识别率平均值为89.3%,表明本文的方法具有 nonpanmetric wavefom estination DigitalSignal P rocess 2009, 19.583 较强的鲁棒性, [7]Engan K.Aase S O.Husoy JH.Method of optialdirections for frame design EEE Int Conf Acoust Speech SignalProcess 1999 参考文献 5.2443 [1]W an S T Lii L Y.He Y L Fault diagnosis method of molling [8]Engan K.Aase $O.Husoy JH.Multi-frme campression theory bearing based on undecmated wavelt transfomation of lifting and design Signal Pmocess 2000.80.2121 schane JVib Shocke 2009.28(1):170 [9]Zhang J F Huang Z C Kemel fisher discrin nant analysis for (万书亭,吕路勇,何玉灵.基于提升模式非抽样小波变换的滚 bearing fault diagnosis EEE Proceedings of the Fourth Intema- 动轴承故障诊断方法研究.振动与冲击,2009.28(1):170) tional Confennce on Machine Leaming and Cybemetics 2005. [2]Zhang L J Xu JW,Yang JH.et al Adaptive multiscale mor 3216 phology analysis and its application in fiult diagnosis of bearngs [10]Yang H Y.Mathew J Ma L Fault diagnosis of molling element JUniv Sei Technol Beijing 2008 30(4):441 (章立军,徐金梧,阳建宏,等.自适应多尺度形态学分析及其 bearings using basis pursuit Mech Syst SignalP mcess 2005.19 在轴承故障诊断中的应用.北京科技大学学报,200830(4): (2):341 441) [11]Yu X T.Chu F L Hao R J Fault Diagnosis appmach for rolling [3]Pan Z W,Xu JW.A supervised fuzzy ART neural network for bearing based on support vector machine and soft momphological pattem classification JUniv Sci Technol Beijing 2000 22(3): filters Chin J Mech Eng 2009.45(7):75 262 (于湘涛,褚福磊,郴如江·基于柔性形态滤波和支持矢量机 (潘紫薇,徐金梧.一种用于模式分类有监督的模糊ART神经 的滚动轴承故障诊断方法.机械工程学报,200945(7): 网络,北京科技大学学报,200022(3):262) 75) [4]Sananta B.AlBahshiK R.AlAmainiS A.Artificial neural ner [12]Yang B S HwangW W,K in D J et al Condition classification works and support vectormachines with genetic akorithm for bear of mall reeipmeatng campressor for refrigemtors using artificial ing fault detection Eng Appl Artif Intell 2003.16:657 neural neworks and support vector machines Mech Syst Signal [5]Mallat SG.Zhang Z F.Matching pursuit w ith tme-frequency dic- Pmcess200519:371北 京 科 技 大 学 学 报 第 32卷 率均在 96∙25%以上;对叠加不同程度的高斯白噪 声数据进行了实验当信噪比为 —5db时信号的总 体识别率平均值为 89∙35%表明本文的方法具有 较强的鲁棒性. 参 考 文 献 [1] WanSTLüLYHeYL.Faultdiagnosismethodofrolling bearingbasedonundecimatedwavelettransformationoflifting scheme.JVibShock200928(1):170 (万书亭吕路勇何玉灵.基于提升模式非抽样小波变换的滚 动轴承故障诊断方法研究.振动与冲击200928(1):170) [2] ZhangLJXuJWYangJHetal.Adaptivemultiscalemor- phologyanalysisanditsapplicationinfaultdiagnosisofbearings. JUnivSciTechnolBeijing200830(4):441 (章立军徐金梧阳建宏等.自适应多尺度形态学分析及其 在轴承故障诊断中的应用.北京科技大学学报200830(4): 441) [3] PanZWXuJW.AsupervisedfuzzyARTneuralnetworkfor patternclassification.JUnivSciTechnolBeijing200022(3): 262 (潘紫薇徐金梧.一种用于模式分类有监督的模糊 ART神经 网络.北京科技大学学报200022(3):262) [4] SamantaBAl-BalushiKRAl-AraimiSA.Artificialneuralnet- worksandsupportvectormachineswithgeneticalgorithmforbear- ingfaultdetection.EngApplArtifIntell200316:657 [5] MallatSGZhangZF.Matchingpursuitwithtime-frequencydic- tionaries.IEEETransSignalProcess199341(12):3397 [6] FanHMengQFZhangYYetal.Matchingpursuitbasedon nonparametricwaveformestimation.DigitalSignalProcess2009 19:583 [7] EnganKAaseSOHusoyJH.Methodofoptimaldirectionsfor framedesign.IEEEIntConfAcoustSpeechSignalProcess1999 5:2443 [8] EnganKAaseSOHusoyJH.Multi-framecompression:theory anddesign.SignalProcess200080:2121 [9] ZhangJFHuangZC.Kernelfisherdiscriminantanalysisfor bearingfaultdiagnosis.IEEEProceedingsoftheFourthInterna- tionalConferenceonMachineLearningandCybernetics2005: 3216 [10] YangHYMathewJMaL.Faultdiagnosisofrollingelement bearingsusingbasispursuit.MechSystSignalProcess200519 (2):341 [11] YuXTChuFLHaoRJ.FaultDiagnosisapproachforrolling bearingbasedonsupportvectormachineandsoftmorphological filters.ChinJMechEng200945(7):75 (于湘涛褚福磊郝如江.基于柔性形态滤波和支持矢量机 的滚动轴承故障诊断方法.机械工程学报200945(7): 75) [12] YangBSHwangW WKimDJetal.Conditionclassification ofsmallreciprocatingcompressorforrefrigeratorsusingartificial neuralnetworksandsupportvectormachines.MechSystSignal Process200519:371 ·396·