基于时移多尺度散布熵和SVM的滚动轴承故障诊断方法  

Fault Diagnosis Method of Rolling Bearing Based on Time-shifted Multi-scale Dispersion Entropy and SVM

作　　者：王勉 刘勇[2] WANG Mian;LIU Yong(Guizhou Industry Polytechnic College,Guiyang 551400,China;School of Mechanical Engineering,GuizhouUniversity,Guiyang 550025,China)

机构地区：[1]贵州工业职业技术学院机电工程系,贵阳551400 [2]贵州大学机械工程学院,贵阳550025

出　　处：《机械设计与研究》2021年第5期83-87,共5页Machine Design And Research

基　　金：贵州省科技厅资助项目(黔科合LH字2016-7066)。

摘　　要：针对多尺度散布熵(MDE)在对滚动轴承故障信号进行特征提取时会出现信号信息严重损失的问题,提出了时移多尺度散布熵(TMDE)的概念,并由此提出基于TMED和支持向量机(SVM)的滚动轴承故障诊断方法。首先,通过仿真信号对TMDE和MDE进行了对比分析,结果表明,TMDE得到的熵值更稳定且对数据长度依赖小。其次,将所提方法应用到滚动轴承的故障诊断实例中,结果表明,TMDE获得了比MDE更高的滚动轴承不同类型和不同程度故障的诊断精度。Aiming at the problem that the signal information would be seriously lost when the multi-scale dispersion entropy(MDE)was used to extract the feature of rolling bearing fault signals,the concept of time-shifting multi-scale dispersion entropy(TMDE)is proposed,and a rolling bearing fault diagnosis method is proposed based on TMED and support vector machine(SVM).Firstly,TMDE and MDE are compared and analyzed through simulation signals.The results show that the entropy value obtained from TMDE is more stable and less dependent on the data length.Secondly,the proposed method is applied to a rolling bearing fault diagnosis example,and the results show that TMDE can obtain higher diagnosis accuracy for different types and degrees of rolling bearing faults than MDE.

关 键 词：散布熵 时移多尺度散布熵 故障诊断 滚动轴承 支持向量机 

分 类 号：TH113[机械工程—机械设计及理论]