An Intelligent Fault Diagnosis Method For JTC Based On CSM System

As an important part of the train control system,the Jointless Track Circuit(JTC)plays an important role in train occupancy and ground-to-vehicle information transmission.Its working status is directly related to the normal operation of the train control system,which would affect the safety of train operation and the efficiency of railway transportation.At present,the Centralized Signal Monitoring System(CSM system)is widely used on the railway to realize the condition monitoring and fault diagnosis for the JTC.Therefore,studying the CSM system’s fault diagnosis algorithm for JTC to improve its diagnostic performance has high practical significance and research value.The main research work of this paper is as follows:1.Based on the transmission line theory,establish the monitoring data model of the CSM system,verify the simulation model through the JTC semi-physical simulation platform,and analyze the influence law of the monitoring data of CSM system.Based on the basic structure and working principle of JTC and CSM,establish the existing monitoring points of the CSM system under the JTC adjustment state,namely equivalent models of voltages such as transmitting cable equipment and cable side,main rail receiving cable cable and equipment side,and main rail and small rail receiver.And use the corresponding JTC semi-physical simulation platform to verify the correctness of the built models.Then through the normalization the effective value of the monitoring voltage,analyze the influence law of five main failure modes on CSM monitoring data,and the five modes are main equipment disconnection,small cable simulation network distributed capacitance or inductance value,small matching transformer capacitance or inductance value,small rail line compensation capacitor value and ballast resistance fluctuations.Finally,quantitatively analyze the influence law of different fault forms on the monitoring data based on the analysis methods of variance,chi-square and mutual information.2.Using CSM system monitoring data,the JTC fault diagnosis algorithm based on random forest is proposed.This method uses the random characteristics and output characteristics of the random forest to enhance the generalization of the algorithm.First,design the corresponding fault diagnosis algorithm based on the random forest;then,build the fault feature set based on the models of CSM system monitoring data to generate the initial model of the random forest,determine the basic parameters of the random forest and the threshold of the retraining process;finally,based on the JTC semi-physical simulation platform,the algorithm is verified for function and performance.The verification results show that the accuracy of the algorithm for known faults is above 92.18%,and for unknown faults the algorithm can trigger the retraining process.It can be seen that this method can accurately locate common JTC faults,and has certain processing capabilities for unknown faults,which improves the generalization and intelligence of the diagnosis algorithm.This method can accurately locate common JTC faults,and has certain processing capabilities for unknown faults,which improves the generalization and intelligence of the diagnosis algorithm.3.Design and implement JTC fault intelligent diagnosis software based on CSM system.By analyzing the requirements of JTC intelligent diagnosis,the overall structure and required functions of the software are designed and realized.Experiments show that the software can collect six monitoring voltages of the CSM system in real time and display them graphically;it analyzes the monitoring data through the diagnosis algorithm model,diagnoses the fault status of JTC in real time,and reflects the fault results,occurrence time and faults treatment measures in the software interface timely.The software has good operability and practicality,which can meet the requirements of the railway site.The amount of figures is 48,the amount of tables is 5,the amount of reference articles is 53.