Fault Diagnosis of High-Speed Train Bogie Based on Capsule Network

Intelligent fault diagnosis of bogie is fundamental for the reliability, stability, and security of a high-speed train (HST). However, thorny issues, including complicated structure of bogie and complex motion states of HST, aggravate the difficulty of research, where traditional approaches for fault diagnosis might fail. In this article, CapsNet, a contemporary novel neural network architecture, in which a single unit called capsule known for its uniqueness of multidimension and abundant spatial information, is adopted to accomplish the recognition and classification of seven working conditions of a HST bogie, comprising both single and compound faults. The experimental accuracy achieved is 96.65%, which proves the efficiency and potency of CapsNet in this regard. The ability of CapsNet to diagnose faults of a bogie in this article exceeds that of a convolutional neural network (CNN), especially for compound ones. Moreover, the method employed is capable of extracting features from raw data automatically and is independent of expert experience or knowledge about signal processing.