基于小波包-支持向量机的滚动轴承故障诊断  

Fault Diagnosis of Rolling Bearings Based on Wavelet Packet and Support Vector Machine

作　　者：陈季云[1] 李娟[2] 

机构地区：[1]南通农业职业技术学院机电工程系,江苏南通226007 [2]山东商务职业学院机械工程系,山东烟台264670

出　　处：《化工自动化及仪表》2010年第9期50-52,60,共4页Control and Instruments in Chemical Industry

摘　　要：以轴承在正常、内圈和滚子裂缝、内圈和滚子剥落三种工况下的振动信号为研究对象,采用小波包频带能量特征提取的方法,构成振动信号的特征向量。在此基础上采用支持向量机对特征向量进行故障模式识别,试验结果表明,和神经网络相比,采用支持向量机进行故障诊断可以获得更高的诊断精度。Vibration signals of rolling bearings of petroleum drilling in three statuses such as normal,inner ring and ball crack,and inner ring and ball peeling were studied.The wavelet packet analysis was used to extract energy eigenvector of frequency domain.On the basis of feature extraction,support vector machine（SVM） was proposed to recognize the fault patterns.The result of experiments indicates that this method is effective and has higher diagnosis precision than neural network.

关 键 词：滚动轴承 故障诊断 振动信号 小波包 支持向量机 

分 类 号：TH165[机械工程—机械制造及自动化]