Intelligent Fault Diagnosis of Rolling Bearing Using Adaptive Deep Gated Recurrent Unit

摘要

Rolling bearing plays a significant part in enhancing the reliability and security of locomotive. Therefore, how to accurately and automatically identify the rolling bearing faults is becoming more and more urgent. For this purpose, an adaptive rolling bearing fault diagnosis method is proposed in this paper. Firstly, deep gated recurrent unit is constructed to effectively learn the features of bearing vibration signals. Secondly, artificial fish swarm algorithm is applied to obtain the key parameters of deep gated recurrent unit. Finally, extreme learning machine is used to accurately classify the learned features and provide final diagnosis result. The proposed method is verified by the measured locomotive bearing vibration signals and the results indicate the feature learning ability of deep gated recurrent unit is powerful and the proposed method achieves more accurate and robust performance than other diagnosis methods.