Denoising method of vibration signal of rolling bearing based on multi—criteria fusio

Abstract

In view of problem that early weak fault features of rolling bearings were difficult to extract, a denoising method of vibration signal of rolling bearing based on multi—criteria fusion was proposed. EEMD method was used to decompose original vibration signal to obtain a set of IMF components, then correlation coefficient of each IMF component and original vibration signal, J divergence of envelope spectrum of each order and original vibration signal, and kurtosis value of each IMF component are calculated. Effective IMF components are selected according to correlation coefficient criterion, J divergence criterion and kurtosis criterion, and the simultaneously retained IMF components are used as the effective component for signal reconstruction. The experimental results show that the can effectively suppress modal aliasing problem in EMD, and at the same time weaken low frequency noise and highlight high frequency resonance component, and has good adaptability.

Keywords

• rolling bearing
• fault feature extraction
• denoising
• eemd
• correlation coefficient
• j divergence
• kurtosis