Rolling element bearing fault diagnosis using adaptive Morlet wavelet filter

摘要

Rolling clement bearings are critical components responsible for most of (he failures of rotating machinery. The impulses generated by the rolling element hearing fault with relatively low energy is spread across a wide frequency band. These periodic impulses may be modulated and masked by noise and low frequent-} effects in the measured vibration signal. Wavelet transform can he used for the multi-resolution time-scale analysis of vibration signals for revealing the hidden impulses. This paper presents a method using Morlet wavelet filter (MWF) for defect diagnosis of rolling element bearings. The parameters of the Morlet wavelet are optimized using Shannon entropy and kurtosis. The method is demonstrated using both the simulated and the actual vibration signals for bearings with localized defects in the inner race and the rolling element. The same signals are then decomposed using discrete wavelet transform (DWT). The results obtained using MWFs are more promising compared to those of DW'T decompositions.