Fault feature extraction method using the lifting wavelet transform and its applications

In order to diagnose and prognose incipient fault of rotating machinery, a new feature extraction method using sliding window is proposed to extract transient impacts from original signals. Firstly, on the basis of the lifting wavelet transform, a wavelet with impact characteristics, constructed by using the lifting scheme, is employed to reveal the impulsive fault feature from the original signals. Then a sliding window, designed according to the revolution cycle of rotating machinery, is applied to process detail signals of decomposition. Modulus maxima obtained from these windows are extracted as expected features in the time domain. An incipient impact malfunction and a bush broken fault of a plain bearing have been successfully detected by using the proposed method.