An Intelligent Identification Approach Using VMD-CMDE and PSO-DBN for Bearing Faults

Abstract

In order to improve the fault diagnosis accuracy of bearings, an intelligent fault diagnosis method based on Variational Mode Decomposition (VMD), Composite Multi-scale Dispersion Entropy (CMDE), and Deep Belief Network (DBN) with Particle Swarm Optimization (PSO) algorithm—namely VMD-CMDE-PSO-DBN—is proposed in this paper. The number of modal components decomposed by VMD is determined by the observation center frequency, reconstructed according to the kurtosis, and the composite multi-scale dispersion entropy of the reconstructed signal is calculated to form the training samples and test samples of pattern recognition. Considering that the artificial setting of DBN node parameters cannot achieve the best recognition rate, PSO is used to optimize the parameters of DBN model, and the optimized DBN model is used to identify faults. Through experimental comparison and analysis, we propose that the VMD-CMDE-PSO-DBN method has certain application value in intelligent fault diagnosis.