基于小波分解和SVM的滚动轴承故障程度研究

Research on Rolling Bearing Fault Degree Based on Wavelet Decomposition and Support Vector Machine

提出一种采用小波分解和支持向量机(Support Vector Machine,简称SVM)提取故障特征的滚动轴承故障诊断方法。首先对原始振动信号进行小波降噪处理,以减小误差,然后进行小波分解,并利用分解得到的小波重构系数计算其能量特征,归一化后作为特征向量,输入SVM中进行故障诊断。实验结果表明,所提出的方法能有效地提取出故障特征,并且具有较高的故障诊断准确率,能准确地区分出滚动轴承不同故障的严重程度。

A rolling bearing fault diagnosis method is proposed in this paper that based on wavelet decomposition and SVM to extract fault feature. First,the original vibration signals were processed by wavelet de-noising in order to reduce error,and then wavelet was decomposed and the decomposed wavelet was used to calculate the energy characteristics using the wavelet reconstruction coefficient. After normalization,they were input into SVM as feature vector to realize fault diagnosis. The experiment results show that the proposed method is effective in fault feature extraction,and has high accuracy of fault diagnosis. The method can accurately distinguish the severity of rolling bearing faults.