| It is inevitable for electric locomotive to malfunction in its routine use, and serious result might be caused by its faults, so on-line fault diagnosis is a very important proof of its running safety. Complexity of the electric locomotive and the range and the difficulty of the faults identifying are very high, it is difficult to diagnose all the faults only use one fault diagnosis technique.Based on the comprehension of the theory of expert system (ES), Park's vector method and Radial-Basis Function (RBF) neural network, the paper research a new faults diagnostic method based on the expert system and the artificial neural network (ANN). The paper presented the faults diagnosis method of expert system and neural network based on the communication network, and the design method based on distributed data gathering and focused faults diagnosis. Objected electric system of 8K electric locomotive, the paper design Rule Pick-Up method and Regulation Strategy of the expert system to diagnose the electric control equipment, and after analyze the PWM harmonics, design diagnosis method based on Park's Vector Method and RBF neural network to diagnose the faults of all kinds of auxiliary induction motors in 8K electric locomotive.The paper analyzed the electric locomotive fault diagnosis method by field experiment. The content of experiment include testifying the reasonability of the designed topology structure, the network protocol and the data frame format of the communication network of the fault diagnosis system and the efficiency of the expert system Knowledge Base design method and Regulation Strategy, and testifying the efficiency of the fault diagnosis method of induction motors based on Park's Vector and RBF neural network.The experiment results indicate that the electric locomotive faults diagnostic method based on the expert system and the artificial neural network is reasonable and efficient. |
