Research On ZPW-2000A Non-insulated Track Circuit Fault Diagnosis Method Based On Rough Set And Graph Theory

ZPW-2000 A non insulated track circuit is an important part of train operation safety assurance and railway transportation equipment.If ZPW-2000 A uninsulated track circuit fails during train operation,it is likely to cause train accidents,which will not only cause economic losses,but also damage life and health.In order to improve the efficiency of railway transportation and make the train run safely,it is very important to maintain ZPW-2000 A non insulated track circuit.Up to now,manual test and analysis is the main discrimination monitoring mode of ZPW-2000 A non insulated track circuit.This monitoring method has many defects,such as difficult discrimination,long time and low efficiency.After studying the monitoring data of ZPW-2000 A uninsulated track circuit,and based on the combination of rough set and graph theory,this thesis makes fault diagnosis of ZPW-2000 A uninsulated track circuit under the condition of making up for the above defects as much as possible.The research in this thesis is as follows:First of all,the structure and working principle of ZPW-2000 A non insulated track circuit are analyzed,and the monitoring type and detection principle of railway signal centralized monitoring system are analyzed.Based on the general fault diagnosis method,this thesis presents the fault diagnosis method of ZPW-2000 A track.Secondly,based on the microcomputer monitoring and recording data of ZPW-2000 A uninsulated track circuit,by analyzing the historical fault data and field application of the track circuit,the voltage conditions of transmission power,track out 1,track out 2,main track input voltage,small track input voltage,transmission power,xgj voltage,XG voltage,GJ voltage and transmission power are extracted as the characteristic attributes.According to the expert experience and historical fault data,the fault diagnosis information system is constructed,and the decision table of fault diagnosis is constructed by combining rough set and graph theory.The weighted multipartite decision table graph and its adjacency matrix are constructed,and then the core attributes are extracted and reduced,the reduced attributes are obtained,and the most simplified attribute reduction table is obtained.Then,the classification of decision rules,the calculation of fault information coverage and the rule network structure of Rete algorithm are defined.According to the coverage and confidence of fault information,the fault rules are extracted from the core attributes.A rule matching algorithm based on Rete algorithm is proposed to complete the fault diagnosis.Finally,Labview and Matlab are used to complete the design and implementation of the ZPW-2000 A non-insulated track diagnostic system.Using the field test data to test the fault diagnosis system in this thesis,it can be seen from the test results that the number of misdiagnosis of the fault diagnosis system in this thesis is improved compared with the faultdiagnosis method of the track circuit based on fuzzy cognitive map,and the misdiagnosis rate is significantly lower the misjudgment rate based on fuzzy cognitive map.