Abstract:Roller element bearing is an important part of mine belt conveyor, which directly determining working condition of the belt. Therefore, there are important theoretical and practical significance in studying fault diagnosis of the roller element bearing. The learning algorithm, structure and principle of artificial neural network were studied, and the network was applied in fault diagnosis of roller element bearing of the belt conveyor. Firstly, fault signals of different types of roller bearings were collected, and the signals were pre processed. Then, the neural network was trained, when the training errors met settings, the training was completed. Finally, the roller element bearing was fault diagnosed by the well trained neural network. Experimental results show that the neural network technique can diagnose rapidly and precisely the fault types of the bearing of belt conveyor.