A novel rolling bearing fault detect method based on empirical wavelet transform

Abstract

A new self-adaptive signal decomposition method called empirical wavelet transform (EWT), inherits the advantages of empirical mode decomposition and wavelet transform. In this method, the vibration signal of rolling bearing is decomposed by the EWT, several frequencies modulation components (AM-FM) are obtained. This paper presents a new method for bearing fault detect based on EWT. First applied EWT to rolling bearing vibration signals. Then calculated the normalized correlation coefficients of each order IMF with original signal respectively. The sensitive IMF is selected according to the normalized correlation coefficient and each order IMF kurtosis factor. Finally, the Hilbert transform to the sensitive IMF and using this envelope spectrum to bearing fault diagnosis. The experiment results show that the proposed method provides a good performance in the detection of outer and inner race faults.