Resilience Enhancement Strategies for High Speed Train Traction Motor Power System with Coupled Faults

This paper analyzes the incipient fault detection and isolation issues of high speed train (HST) traction motor power system. The fault features in such a motor are obscure and tight coupling with few fault samples, and the fault separation vary in nonlinear pattern. Up until now, most existing fault diagnosis techniques have been presented under the condition that these faults are exclusive, which means there is only one fault once. However, in HST applications, coupled faults are more common to take place due to the interconnected nature of different components. To enhance the system resilience or the diagnostic accuracy is of great significance to HST safe operation. A novel fault severity model is proposed suitable for the normal and fault conditions. Then electromagnetic torque energy entropy coding is utilized to extract fault feature and construct different feature matrixes. Resilience enhancement strategies with support vector machine model is generated from a novel grey wolf optimizer algorithm. The perform of the proposed work is validated through simulation and experimentation on a fault testing verification platform for HST traction system.