基于自适应多小波与综合距离评估指数的旋转机械故障特征提取-论文.pdf

《基于自适应多小波与综合距离评估指数的旋转机械故障特征提取-论文.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、振动与冲击第33卷第12期JOURNALOFVIBRATIONANDSHOCKV01．33No．122014基于自适应多小波与综合距离评估指数的旋转机械故障特征提取卢娜，肖志怀，张广涛，孙召辉(武汉大学水力机械过渡过程教育部重点实验室武汉430072)摘要：旋转机械设备故障诊断主要包括信号采集、特征提取和故障识别，而特征提取是进行故障诊断的基础和保证诊断结果正确的关键，为_『提高特征参数对故障的敏感性，提出了基于自适应多小波与综合距离评估指数的旋转机械故障特征提取方法。该方法以综合距离评估指数最大值为目标函数，利用遗传算法从

2、CL3自适应多小波库中选择最优多小波，并将该最优多小波用于转子振动信号的特征提取。通过对正常、不对中、不平衡、碰摩四种设备状态下采集的振动信号进行特征提取，并将所提出的方法和传统特征提取方法提取的特征参数输入到K一最邻近分类器进行分析，结果表明，所提出的方法能够大大增强特征参数对故障的敏感性，获得更高的故障诊断准确率。关键词：旋转机械；特征提取；故障诊断；CL3自适应多小波；综合距离评估指数中图分类号：TH165．3文献标志码：ADOI：10．13465／j．cnki．jVS．2014．12．034Rotatingmachi

3、neryfaultfeatureextractionbasedonadaptivemulti-waveletsandsynthesisdistanceevaluationindexLUNa，XIAOZhi—huai，ZHANGGuang—tao，SUNZhao—hui(MOEKeyLaboratoryforHydrodynamicTransientsProcesses，WuhanUniversity，Wuhan430072，China)Abstract：Theprocessofrotatingmachineryfaultdi

4、agnosisiscomposedofsignalacquisition，featureextractionandfaultidentification，amongthemthefeatureextractionisthe~undationoffaultdiagnosisandthekeytoobtaincorrectdiagnosisresults．Toimprovethesensitivityoftheextractedfeaturestofaults，arotatingmachineryfaultfeatureextr

5、actionmethodbasedonadaptivemulti—waveletandsynthesisdistanceevaluationindexwasproposedhere．Inordertoevaluatethesensitivityoffeatureparameters，themaximumvalueofthesynthesisdistanceevaluationindexwastakenastheobjectivefunction，andtheoptimalmulti—waveletswereselectedf

6、romthelibraryofCL3adaptivemulti—waveletwithgeneticalgorithm．Thentheywereusedtoextractfeaturesfromvibrationsignalsofarotor．Toprovetheeffectivenessoftheproposedmethod，K—nearestneighborclassifierwasusedtoanalyzethefeaturesextractedwiththeproposedfeatureextractionmetho

7、d，thesynthesisdistanceevaluationindexfeatureextractionmethodandtheprincipalcomponentanalysisfeatureextractionmethod，respectivelyfromvibrationsignalsofatestedrotatingmachineryundernormal，unbalance，misalignmentandrotor—to-statorrubconditions，respectively．Theresultssh

8、owedthattheproposedmethodcanbeusedtoimprovethesensitivityoffeatureparametersandobtainahigherfaultrecognitionrate．Keywords：rotatingmachinery；featu