Abstract:

The key to fault diagnosis of rolling element bearings is how to find typical characteristic frequencies of rolling element bearings from low SNR mixed signals. A method was presented to combine continuous wavelet transform (CWT) with ICA for diagnosing early faults of rolling element bearings and a method to select wavelet scales with iso-interval frequency was proposed for the first time.Envelope spectrum was analyzed to diagnose the faults of rolling element bearings. Finally, the method has been verified by practical signal analyses of rolling element bearings.

Key words: wavelet transform, independent component analysis(ICA), single-channel signal, iso-interval frequency

摘要：

滚动轴承早期故障诊断的关键在于如何从低信噪比混合信号中检测出显著的轴承故障特征频率。提出以连续小波变换（CWT）和独立分量分析(ICA)相结合的方法来诊断单通道信号的滚动轴承早期故障,提出按频谱等间隔选取伪中心频率的小波分解尺度,并对ICA处理后的信号进行包络频谱分析以确定故障类型。最后,利用实际的滚动轴承实验数据对该方法进行了验证。

关键词: 小波变换, 独立分量分析, 单通道信号, 等频率间隔