基于EEMD-SC的机械故障诊断方法研究  

Research on Machinery Fault Diagnosis Based on an EEMD-SC Method

作　　者：谭航[1] 梁雪松[1] 万丽萍 吴兆耀[1] TAN Hang;LIANG Xuesong;WAN Liping;WU Zhaoyao(College of Physics and Engineering,Chengdu Normal University,Chengdu Sichuan 611130;Educational Administration Department,Chengdu Normal University,Chengdu Sichuan 611130)

机构地区：[1]成都师范学院物理与工程技术学院,四川成都611130 [2]成都师范学院教务处,四川成都611130

出　　处：《河南科技》2019年第17期50-53,共4页Henan Science and Technology

基　　金：四川省教育厅自然科学一般项目(18ZB0096);成都师范学院人才引进项目(YJRC2014-5)

摘　　要：针对实际工程中,装备长期处于正常运行状态,故障样本稀少,数据标记困难,导致智能诊断往往无训练样本可用的问题,提出一种EEMD-SC的机械故障诊断方法。首先利用EEMD将已知故障类型的参考样本与待诊断样本数据进行分解,得到若干个IMfs分量。接着,计算出每个IMFs的概率密度。然后利用相关计算得到待诊断样本IMF与不同故障的参考样本对应IMF的相关性,并求出所有IMFs相关性之和,即为所要求的SC值。最后,求出SC最大时的参考故障样本,待检测样本的故障即为此参考样本所含故障。利用包含不同故障程度的内圈、外圈、正常、滚动体故障的轴承振动监测试验对提出方法进行验证。试验结果表明,在每种故障的参考样本均只有一个时,最后诊断结果仍可达到令人满意效果,从而证实了本方法的有效性。Aiming at the problem that equipment was in normal operation for a long time,fault samples were scarce and data labeling was difficult,which led to intelligent diagnosis without training samples available,a mechanical fault diagnosis method based on EEMD-SC was proposed.To solve this problem,this paper proposed a new machin?ery fault diagnosis based on EEMD-SC.In this method,first,the EEMD was used to decompose the data of samples with labels of fault types and tested samples into several IMfs,respectively.Then,the probability density distribution of each IMF could be calculated.Afterwards,the cross correlation between the IMFs of tested samples and referenced samples of different fault types was calculated and the SC value could be obtained by summing up all the cross corre?lation values.Finally,the fault type of one tested sample could be determined according to the SC value and the type was the same with the type of the referenced sample,when the maxima SC value was obtained.One experiment about the condition monitoring of bearing were used to verify the effectiveness of the proposed method.In this experiment,bearing health conditions including inner race fault,outer race fault,ball fault and normal with different fault severity were considered.The results show that the proposed method can still achieve a high fault diagnosis accuracy even though there is only one referenced sample of each fault type,which demonstrates the effectiveness of the proposed method.

关 键 词：故障诊断 EEMD 概率密度分布 互相关 

分 类 号：TG156[金属学及工艺—热处理]