Early Fault Diagnosis of Rolling Bearing based Empirical Wavelet Transform and Spectral Kurtosis

An adaptive mode decomposition method based on empirical wavelet transform (EWT) and spectral kurtosis (SK) is proposed for rolling bearing fault diagnosis. First, by applying EWT the raw vibration signal of rolling bearing is decomposed into a set of intrinsic mode functions (IMFs) are obtained. Then the IMFs which mostly characterizes fault information is selected according to energy entropy. Finally, the selected IMF signal is further filtered by an optimal band-pass filter based on SK to extract impulse signal. Fault identification can be deduced by the appearance of fault characteristic frequencies in the squared envelope spectrum of the filtered signal. The experiment results show that the proposed method certain advantages and provides better effect in the detection of outer race, inner race and rolling element faults.