改进VMD去噪与多特征融合的声发射信号识别方法-豆柴文库

您所在位置：网站首页 / 改进VMD去噪与多特征融合的声发射信号识别方法.docx / 文档详情

在线预览结束，喜欢就下载吧，查找使用更方便

5 金币

下载文档

/ 2

下载提示

如果您无法下载资料，请参考说明：

1、部分资料下载需要金币，请确保您的账户上有足够的金币

2、已购买过的文档，再次下载不重复扣费

3、资料包下载后请先用软件解压，在使用对应软件打开

文本预览

改进VMD去噪与多特征融合的声发射信号识别方法
Abstract
Vibration-basedMachineDiagnostics(VMD)isapromisingtechniquefornoisereductionandfeaturefusioninsoundemissionsignalrecognition.ThispaperproposesanimprovedVMDalgorithmfortheclassificationofsoundemissionsignals.TheimprovedVMDalgorithmcombinestheprincipalcomponentanalysis(PCA)andGrayLevelCo-occurrenceMatrix(GLCM)featurestoidentifysoundemissionsignalseffectively.Theresultsshowthattheproposedalgorithmcaneffectivelyidentifythesoundemissionsignals.
Introduction
Soundemissionsignalrecognitionisacrucialtaskinthefieldofmachinerydiagnostics.Vibration-basedMachineDiagnostics(VMD)hasrecentlybeenusedasaneffectivetechniquefornoisereductionandfeaturefusionforsoundemissionsignalclassification.Thetraditionalsignalprocessingtechniquessufferfromseveralproblems,suchaspooridentificationaccuracy,lowsignal-to-noiseratio,andpoorfeatureextractionefficiency.VMDalgorithmisappliedtodecomposeasignalintomultiplespectralcomponentsandcanretainthefrequencyandmodalinformationofthesignal,thusachievingexcellentperformanceinsoundemissionrecognition.However,VMDalgorithmsstillhaveproblemsinidentifyingweaksignalinformationandextractingeffectivefeaturesfromthesignal.
Therefore,thispaperproposesanimprovedVMDalgorithmcombinedwithPCAandGLCMfeaturestoachieveeffectivenoisereduction,featureextractionandsoundemissionrecognition.
Methodology
Inthispaper,weproposeanimprovedVMDalgorithmwithPCAandGLCMfeatures.TheimprovedVMDalgorithmconsistsofthefollowingsteps:
-Preprocessing:Therawdataispreprocessedusingthewavelettransform.Thehigh-frequencynoiseisremovedfromthedata.
-VMDdecomposition:ThesignalisdecomposedintomultiplemodesusingtheVMDalgorithm.
-PCAfeatureextraction:Principalcomponentanalysis(PCA)isperformedoneachVMDmodetoextractimportantfeatures.PCAaimstotransformthemodeintoasetoflinearlyindependentprincipalcomponentsthatdescribethemode'svariation.
-GLCMfeatureextraction:GrayLevelCo-occurrenceMatrix(GLCM)featuresareextractedfromthePCAcoefficientstorepresentthetexturepropertiesofthemodes.