An effective fault diagnosis method is of great significance to improve the safety of mine production. Despite the considerable success of deep learning the complexity of actual working conditions and the data acquisition correspond to too many parameters to adjust. In this paper, a fault diagnosis method is proposed to solve the problem of parameter explosion. In the method, empirical mode decomposition is used to separate complex vibration signals to obtain Intrinsic Mode Function. Based on Intrinsic Mode Function, the component with large information entropy is selected as the input, and its envelope analysis is carried out to compare the bearing fault frequency to calculate the corresponding characteristics. Taking the feature as the input of the neural network, the number of parameters in the neural network is reduced through the existing prior knowledge, and the phenomenon of parameter explosion is alleviated to a great extent. The performance of the network is better than the end-to-end learning mode, and considerable optimization has been made in accuracy and operation efficiency. It has achieved a comprehensive improvement of more than 7% over the end-to-end network on the laboratory data set.

Rolling bearings;fault diagnosis;empirical mode decomposition(EMD)