Open Circuit Fault Diagnosis of Advanced Cophase Traction Power Supply System Based on Neural Network

Lots of power switch are used in advanced cophase traction power supply system, the open fault of power switch are bad for the reliability of the power supply system. Due to the variety of power switch and their non-linear characteristics, it is difficult to establish a mathematical model of the system to diagnose the open faults. This paper presents a fault diagnosis method based on Back propagation (BP) neural network. Firstly, the mechanism that the output level of the cascaded system will change when an open circuit fault occurs is analyzed. Then, According to the modulation strategy, the variation law of the harmonic of output voltage is analyzed, when the open-circuit fault occurs within the module or between modules. The feature quantities of all faults is extracted as training samples, with the trained three-layer neural network structure, the open circuit fault of the system can be diagnosed in real time. The simulation result shows that neural network fault diagnosis method can accurately and reliably diagnose the open fault of the power switch within 0. 02s without adding additional sensor.