基于Hilbert谱奇异值的轴承故障诊断

Bearing Fault Diagnosis Based on Hilbert Spectrum and Singular Value Decomposition

针对机械故障振动信号时频特征提取问题,提出一种基于Hilbert谱奇异值的特征提取方法,并将其应用于轴承故障诊断。该方法首先利用经验模式分解方法将振动信号分解为若干个内蕴模式函数之和,接着对每个内蕴模式函数进行Hilbert变换得到振动信号的Hilbert谱,然后对Hilbert谱进行奇异值分解,得到反映机械状态特征的奇异值序列,最后利用奇异值作为特征向量,使用支持向量机进行轴承故障诊断。轴承正常、内圈故障、滚动体故障、外圈故障实测信号实验结果表明,该方法能有效地提取轴承故障振动信号特征。

A new fault diagnosis method based on Hilbert spectrum and singular value decomposi- tion was proposed and applied to bearing falut diagnosis. Firstly, the bearing vibration signals were decomposed into a set of intrinsic mode functions by means of the empirical mode decomposition method. Then, Hilbert transform was applied to each component and get the Hilbert spectrum of the signals. To extract the time--frequency feature of the faulted bearing the singular decomposition val- ue method was used to the Hilbert specturm. Finally,the singular values were used as the feature vec- tore and the support vectore machine method was used to identify the different faults. Experiments were conducted on roller bearing without faults, with inner-race faults, ball and outer-race faults and several levels of fault severity. The experimental results show that the proposed method is effec- tive.