Gear Fault Pattern Recognition based on the Optimum Wavelet Packet Decomposition and HMM

The key point of gear fault pattern recognition is the fault feature extraction of vibration signal. Aiming at feature extraction of gear fault pattern recognition, a method based on the Optimum Wavelet Packet Decomposition (OWPD) and Hidden Markov Model (HMM) is proposed in this paper. Processing of the vibration signals in the time domain is considered, using the wavelet packet. The characteristic energy automatically selected by OWPD is then employed as the input of HMM model for training and test. Finally the effect and accurate of the new method is validated by experiments.