Rolling Bearing Fault Diagnosis of Train Running Gear Based on Optimized Deep Residual Network

Abstract

With the development of the railway in China, fault diagnosis of rolling bearings is important in high-speed and heavy-load trains. Aiming at the shortcomings of traditional fault diagnosis methods, such as low recognition rate, poor robustness, and relying on manual extraction of features, this paper proposes a method based on the optimized deep residual network SE-RESNET-26 for the rolling bearing fault diagnosis method of train running parts. The deep residual network is composed of stacking residual blocks, which can effectively improve the gradient disappearance and network performance degradation of the traditional deep learning models. In this paper, the channel attention mechanism, Squeeze-and-Excitation block (SE), is embedded into the residual block, and the model structure is adjusted according to the characteristics of fault data to further improve the recognition accuracy. Experimental results show that the proposed method can adaptively extract fault features of bearings under complex working conditions and realize high-precision intelligent fault diagnosis, which is superior to the comparison method.