BP neural network based bearing fault diagnosis with differential evolution & EEMD denoise

In mechanical equipment, rolling bearing is frequently used. Its running state directly affects the performance of the whole machine and it is also the main cause of mechanical equipment failure. This paper focuses on the fault diagnosis of the rolling bearings. The method of fault diagnosis based on the self-adaptive denoise of ensemble empirical mode decomposition (EEMD) and improved back propagation (BP) neural network with differential evolution (DE) algorithm is studied. Firstly, the signal is decomposed by EEMD, and then the reconstructed signal of adaptive noise reduction is acquired based on the threshold of distance measurement. Secondly, utilizing the error of BP neural network as the objective function, weights and thresholds of the network are optimized by DE algorithm. Finally, the optimized BP network is used for fault diagnosis. Experimental results show that the proposed method is more effective and accurate than the traditional BP neural network.