Real-Time Evaluation of Ground Insulation Degradation and Fault Warning Method Under Multiple Operating Conditions for Train Traction Drive System.

Aiming at the problem that the main circuit grounding fault in the traction drive system of locomotives and high-speed trains can only be diagnosed under a single operating condition and cannot be warned about early, a mechanism and data-driven real-time evaluation and full operating condition fault warning method for ground insulation degradation is proposed. Firstly, based on the mechanism of grounding faults, the circuit characteristics of the main circuit of the traction transmission system under different grounding fault conditions are analyzed, and mathematical models are established for the detection of various grounding faults and sensor signals under different operating conditions, as well as for evaluating the degree of degradation of grounding faults. Secondly, based on engineering application experience, a feature index set that can accurately classify different types of grounding faults is extracted. Combined with on-site fault case data, a decision tree method is used to establish a classification model between the feature index set and typical grounding fault sources under different operating conditions, which is then converted into a fault diagnosis rule library. Finally, real-time collection of relevant sensor signals, based on the fault diagnosis rule library and the degradation degree evaluation model of grounding faults, enables real-time detection and warning of grounding faults under all operating conditions to ensure train safety and provide key information support for optimal degraded operation in the future. The test result based on controller hardware in the loop shows that the method proposed in this paper can achieve accurate detection and localization of grounding faults under different operating conditions and can provide real-time warning of the severity of grounding faults, which has good engineering application value.