Fault Diagnosis method for Railway Turnout Control Circuit based on Information Fusion

High speed railway turnout is an important signal equipment that is directly contacted with the high speed train. However, it is still in a simple way to deal with the faults of the control circuit by simple instruments and artificial experience. In order to realize the intelligence of the fault diagnosis method for the turnout control circuit, this paper summarizes 11 typical fault modes and 8 corresponding typical fault features. Then, according to the fuzzy theory and neural network, the fault diagnosis is respectively realized by multi factor fuzzy evaluation and three layer BP neural network model. But both two methods cannot solve this problem well. They still can't satisfactorily deal with the problem of false positives and false negatives, which threatens the safety of railway operations. Therefore, based on the Dempster-Shafer evidence theory, this paper further proposes a comprehensive evaluation method of fault diagnosis on the information fusion decision-making level, and achieves the complementary fusion of the two methods, and also increases the accuracy of fault diagnosis. From the verification of the simulation experiments, the method is more accurate than any of the two simple methods of fault diagnosis, and it is promising to have a good application prospect in this field.