Rolling Bearing Fault Diagnosis under Strong Background Noise Based on ACMD and Optimized MOMEDA

Abstract

A method based on adaptive chirp mode decomposition (ACMD) and optimized multipoint optimal minimum entropy deconvolution adjusted (OMOMEDA) is proposed to diagnose the rolling bearing fault in the presence of strong background noise. First, ACMD based on the Gini index is used to separate the low resonance impulse component in the fault signal from the harmonic component and noise. After ACMD, the OMOMEDA process is performed on the low resonance impulse component to enhance the fault impulse. After OMOMEDA, the envelope analysis is conducted on the deconvoluted signal to determine the fault condition. It should be noted that OMOMEDA overcomes the parameters’ artificial dependence of multipoint optimal minimum entropy deconvolution adjusted (MOMEDA) and realizes a superior optimization solution through two aspects of improvement. First, OMOMEDA introduces the augmented gray wolf optimizer and cuckoo search to achieve adaptive determination of the period parameter. Second, OMOMEDA defines balanced permutation entropy (BPE) and uses BPE as a fitness function for finding the best filter length. The simulation and experimental results show that the proposed method can accurately and effectively diagnose the rolling bearing fault in the presence of strong background noise.