故障程度鲁棒的滚动轴承智能诊断方法

Intelligent Diagnosis Method of Rolling Bearings Robust to Fault Severity

本文中提出了一种基于共振滤波和包络小波包分解的鲁棒故障特征提取方法,首先对信号进行共振频带滤波,再对滤波信号的包络进行小波包分解并提取重构子带信号的标准偏差值作为特征向量。该方法考虑了包络信号能充分描述轴承局部故障引起循环冲击的特点,并利用支持向量机进行轴承故障类型自动识别。实验结果表明,该方法能有效地实现故障程度鲁棒的滚动轴承智能诊断,具有较高的诊断速率,效果优于传统滚动轴承诊断方法。

A novel feature extraction is presented to obtain features which are insensitive to fault severity levels. Signals are first filtered by a band-pass filter around their resonance frequency, and then envelops are obtained by the Hilbert transform. Wavelet packet transform are employed to decompose the envelops to a third level followed by calculating the standard deviation of each node signals as fault severity robust features. These features are then input to support vector machines （ SVMs ） to identify fault type. The results verify the effectiveness and quickness of the proposed approach and this method is better than the results of traditional methods for the practical problems.