Fault Diagnosis of High-Speed Train Traction Converter System Based on Random Forest Algorithm

Traction converter system is a key part of the energy conversion of high-speed trains, and the fault diagnosis of the traction converter system has become crucial with the increasing requirements on its reliability. However, the system operates in a complex environment and is susceptible to the coupling of multiple physical fields, which makes the mechanism of its operational state behaviour difficult to describe. In addition, the insufficiency of actual fault data also poses a challenge to the fault diagnosis of this system. Therefore, this paper proposes a fault diagnosis method for the traction converter system of high-speed trains using the random forest (RF) algorithm as an analysis tool, which is simple and robust with sample classification capability. Firstly, time domain, frequency domain and time frequency domain features that can reflect the operating characteristics of the system are extracted from the original data. Secondly, RF is used to obtain the importance size of the extracted features and select the feature subset that is most beneficial to the classification task. Finally, the RF fault classification model is built on the dataset corresponding to the feature subset, and the constructed model is used to achieve fault diagnosis. In order to verify the effectiveness of the proposed method, the open-circuit faults of the traction inverter IGBTs in the traction converter system are considered, and a traction inverter simulation model is established using Simulink, and the three-phase current signals of the inverter IGBTs in various states are collected as the original data for experimental validation. The results show that the method proposed in this paper is an effective fault diagnosis method for traction inverter system.