Fault feature extraction based on optimal energy using lifting wavelet packet

Abstract

Fault prediction is the key technology to guarantee the safe operation of large mechanical equipment,and fault feature extraction is a key issue in fault prediction. To extract fault feature from the non-stationary fault signals, this paper proposed a fault feature extraction method using lifting wavelet packet, and constructed the fault feature vector of optimal energy. The fault feature extraction analysis shows that the proposed method can highlight the energy change within the optimal decomposition frequency band, and effectively reflect the fault status.