Bearing fault diagnosis for coupling faults of rotor system using spectral kurtosis and wavelet packet

For a rotor system, when the bearing of induction machine powered the rotor occurs faults with the rotor bearing simultaneously, the vibration signal becomes complex. Aiming at the complexity of multiple faults of a rotor bearing system, spectral kurtosis (SK) is utilized to detect the bearing fault in this paper. An experimental study is performed to explore that it is difficult to identify the multiple faults of rotor system using the spectral kurtosis alone. A method based on wavelet packet and spectral kurtosis is proposed in order to extract different characteristic frequencies. The superiority of the proposed method based on spectral kurtosis and wavelet packet is demonstrated in extracting different fault characteristics of multiple faults from different bearings.