基于局部切空间排列与WSVM的滚动轴承故障诊断  

The Roller Bearing Fault Diagnosis Based on Local Tangent Space Alignment and WSVM

作　　者：陈法法[1,2] 李冕[2] 肖文荣[2] 钟先友[2] 

机构地区：[1]三峡大学新能源微电网湖北省协同创新中心,湖北宜昌443002 [2]三峡大学水电机械设备设计与维护湖北省重点实验室,湖北宜昌443002

出　　处：《组合机床与自动化加工技术》2016年第6期106-110,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金(51405264;51205230);三峡大学人才启动基金(KJ2014B007);湖北省教育厅项目(B2015248)

摘　　要：针对滚动轴承的多类故障特征非线性难以有效辨识的问题,提出基于局部切空间排列和小波支持向量机的滚动轴承故障诊断方法。在由集成经验模式分解(Ensemble Empirical Mode Decomposition,EEMD)处理后的频带能量组成的故障特征集中,首先采用局部切空间排列进行约简降维,提取其中的低维敏感特征,随后将获取的低维敏感特征输入给小波支持向量机进行滚动轴承的多类故障辨识。实验结果表明,基于局部切空间排列(local tangent space arrangement,LTSA)和小波支持向量机(wavelet support vector machine,WSVM)的滚动轴承故障诊断方法能够有效提取多类故障的低维敏感特征,并且相对传统诊断方法而言故障诊断精度更高。Aimed at the problem that the multi-class fault feature of rolling bearings are difficult to be effectively identified,a fault diagnosis method of rolling bearing based on local tangent space arrangement（ LTSA）and wavelet support vector machine（ WSVM） is proposed. The feature sets are constituted of the frequency band energy decomposed by Ensemble Empirical Mode Decomposition（ EEMD）. Firstly, it uses the local tangent space alignment for high dimension reduction. Then, the low-dimensional sensitive fault features are extracted from the original fault feature set. Finally, those low dimensional sensitive fault features are inputted into the wavelet support vector machine for rolling bearing multi-class fault identification. The experimental results show that the rolling bearing fault diagnosis method based on the LTSA and WSVM can effectively extract more low-dimensional sensitive fault features,and the fault diagnosis accuracy is higher than that of the traditional diagnosis method.

关 键 词：局部切空间排列 小波支持向量机 滚动轴承 故障诊断 

分 类 号：TH165[机械工程—机械制造及自动化] TG65[金属学及工艺—金属切削加工及机床]