Motor bearing fault diagnosis based on multi-feature fusion and PSO-BP

This paper proposes fault a diagnosis method based on the combination of multi-feature fusion, particle swarm optimization and BP neural network(PSO-BP). In this method, the vibration signal of motor bearing is decomposed by wavelet packet, then the fuzzy entropy of each frequency band is solved, and finally the mixed feature vector is formed with the kurtosis of the original signal, which is input into PSO-BP for fault diagnosis. The simulation experiment is carried out using bearing data from CWRU, and the results prove that this method can more effectively distinguish the types of motor bearing faults.