Fault Diagnosis Of ZPW-2000A Track Circuit Based On Set Pair Analysis

ZPW-2000A track circuit is one of the basic equipment of railway signal.Whether the equipment can work normally directly affects the railway transportation efficiency and traffic safety.Therefore,the detection and diagnosis of track circuit is very important.However,at present,the state assessment and fault identification of track circuit use manual analysis of monitoring curve and data,which has poor real-time detection,ignores the early weak change of fault symptoms,and fails to consider the uncertainty problems caused by incomplete,random and fuzzy information of track circuit state information and the dynamic relationship between uncertainty problems.In view of this,it is of great significance to study the operation status and fault diagnosis of track circuit.Based on the actual track circuit monitoring data of a certain railway Bureaus’ electricity section,the combination of set pair analysis,membership function and clustering principle is proposed in this thesis to realize the research of track circuit fault diagnosis,and gives the corresponding maintenance suggestions.The main contents of this paper are as follows.Firstly,the basic theory of ZPW-2000A track circuit is introduced,analyzes the monitoring system and extract required data,summarizes the fault types and the impact on the running state,extracts the track circuit characteristic parameters according to the technical specifications and the actual situation,and selects the typical fault types and corresponding fault symptoms that can represent the running state,so as to provide reference for the research of track circuit.The theoretical foundation of circuit fault diagnosis is established.Then,aiming at the lack of description between voltage symptoms and operation state of track circuit at present,the relationship between fault symptoms and fault types is analyzed,fault symptom set is established,set pair analysis is introduced,and fault type connection number is established.According to the classification of track circuit operation state level,the connection number element is determined,and the connection number expression of operation state level is constructed: Aiming at the boundary fuzziness of track circuit state level classification,the same,different and opposite evaluation matrix is determined by combining with membership function,and the connection membership degree is established,and the average method is used to take the value of difference coefficient to form the difference coefficient matrix.For the calculation of the weight coefficient,the entropy weight method is used to establish the index weight interval.Combined with the connection number,the weight is calculated from the three aspects of the same,the different and the opposite,and the final accurate weight value is obtained,which improves the accuracy of fault diagnosis and solves the influence of experts’ subjective opinions on the determination of the weight value.According to the contact value of the overall operation state and various fault types of the track circuit,the current operation state of the track circuit is determined by comparing the division of the state level.Finally,each reference system and sample connection number are established.According to the results of state analysis,the principle of identity,difference and opposition is combined with the idea of clustering.Fault state is diagnosed based on distance measurement,and fault type is determined quickly.Make full use of the monitoring data,so that the equipment affecting the site can be transferred from fault repair to state repair,and improve the operation efficiency.The results show that the proposed method has advantages and higher accuracy in track circuit.