Research On Railway Signal Equipment Fault Diagnosis Method Based On Information Fusion

As rail transportation is developing towards high speed, high density and heavy load,people tend to raise higher requirements to the safety performance of the rail signalingequipment. Since the safety of the rail signaling equipment constitutes a quite important factorin ensuring the rail safety, it is very necessary to develop a system diagnosing the railsignaling equipment faults according to the actual situation.At present, the fault diagnosis of the rail signaling equipment usually depends on theexperience of maintainers and the degree of competence in the fault mechanism.Consequently, the problems arise as follows: the inexperience of maintainers may lead to thedelay of maintenance time and train operation accident caused by equipment fault; the singlediagnostic method may reduce the diagnosis accuracy. Therefore, new technology is in urgentneed in order to achieve the intelligent fault diagnosis of rail signaling equipment.The research of the thesis is presented in the following. To begin with, the thesisanalyzes the outdoor rail signaling equipment including its constituents, working principles,fault mechanism, and historical fault data. By analysis, several most typical fault modes areobtained and meanwhile the sample data of outdoor signaling equipment is established. Next,the thesis discusses information fusion technology feature, structure, and its specificprocessing method. Besides, the thesis demonstrates the feasibility and effectiveness ofinformation fusion technology in outdoor signaling equipment fault diagnosis according to theinformation theory. Lastly, based on the D-S evidence theory, the thesis tentatively proposes afault diagnosis model and method of rail signaling equipment by combing fault diagnosis ofthe rail signaling equipment with information fusion. This mode and method works in thefollowing way. Firstly, fault diagnosis are made on the S700K turnout switch machine and theZPW-2000A joint-less track circuit of outdoor signaling equipment by using BP neuralnetwork and fuzzy comprehensive evaluation respectively. Secondly, the probabilityassignment of D-S evidence theory is constructed by using the output of both BP neuralnetwork and fuzzy comprehensive evaluation. Finally, from the angle of D-S evidence theory,the decision fusion is made on the fault diagnosis result got from BP neural network andfuzzy comprehensive evaluation. At last, the feasibility and effectiveness of this method isverified by examples.The diagnosis results show that the information fusion method is of comparatively higheraccuracy in fault diagnosis and the reliability of diagnostic results is significantly improved.