Fault Diagnosis of Rolling Bearings Based on Wavelet Packet and Spectral Kurtosis

Abstract

In order to effectively identify the weak fault characteristic frequency of rolling bearings under strong background noise, researching on fault signals de-noising processing based on kurtosis coefficients and the segmentation thresholds noise reduction method by autocorrelation analysis of wavelet packet decomposition coefficients, which can improve the signal-to-noise ratio and the composition of high frequency resonance signal components. And combined with the spectral kurtosis theory determines the parameters of band pass filter. Researching on the weak fault diagnosis and detecting the fault characteristic frequency of rolling bearings based on band pass filtering and envelope demodulation method. And engineering application has been done for bearings weak fault diagnosis, which achieved well diagnosis effect.