滚动轴承故障程度识别

针对实际运行滚动轴承的故障程度问题,提出一种诊断滚动轴承故障程度的方法.深入研究滚动轴承的故障机理、振动信号的时域特征以及不同程度故障对滚动轴承运行的影响进行了,广泛分析振动特征提取方法和支持向量机的算法,采用了小波包能量法提取状态特征,使用新型二叉树支持向量机的多类分类算法.实验结果表明采用小波包提取状态特征和支持向量机可以滚动轴承故障程度识别,模型的学习、泛化能力强.

Rolling Bearing Fault Degree Identification

A special structure for diagnosis of rolling element bearing fault degree is proposed. The failure mechanism of the rolling element bearing vibration signal, time domain characteristics , the influence of different degree of fault on the rolling element bearing operation , vibration feature extraction method and support vector machine algorithm are discussed. Based on this, the wavelet packet decomposition method is used to extract fault feature of rolling element bearing and a new structure of binary tree support vector machine （SVM） classification algorithm is adopted. The experimental results show that the wavelet packets extraction state feature and support vector machine can identify rolling bearing fault degree with strong ability of learning, generalization