Research On Intelligent Fault Diagnosis For ZPW-2000A Track Circuits

As crucial ground work equipment in the modern high-speed railway signal system, the Track circuit is the core component of the security critical information for on-board safety computers. The failure of track circuit function will affect the transport efficiency, it may even lead to catastrophic consequences. For this reason,it is necessary to ensure the normal operation of the track circuit. The ZPW-2000A track circuit is widely used in Chinese passenger line and high-speed railway, and it is the fault-prone equipment due to the perennial working in the outdoor complex environment. At present, the track circuit maintenance use the way of manual check at regular intervals. Therefore, the theoretical level and practical experience of maintenance staff member will directly affect the efficiency and accuracy of fault handling. How to diagnose the orbital circuit fault quickly and accurately is still a difficult problem to be solved. So, there is great practical significance to study the intelligent fault diagnosis method of ZPW-2000A track circuit.In this thesis the main intensive study and contributions are as follows:Firstly, ZPW-2000A track circuit is used as the research object, distinguish the type of track circuit failure by analyzing the principle of the composition of the track circuit and combining with field failure; Simulate the type of track circuit red band failure basing on the automation laboratory of Southwest Jiaotong University; establish single wheel occupancy model which based on uniform transmission theory and four port network.Secondly, the study takes red tape fault of ZPW-2000A Track Circuit as the research subject, and aims at the fault of the diversity and complexity to come up with a method for track circuit intelligent fault diagnosis, which based on Fault Tree Analysis (FTA) with improved BP neural network. According to the track circuit and fault relations establish fault tree, to perform the qualitative analysis of FTA and analyze the fault causes. So that the diagnosis model could extract the fault diagnosis rules and input-output could be confirmed. Constructing of two parallel BP neural subnets to compose diagnostic model and adjusting the model parameters by LM and Genetic Algorithm. The simulation analysis has proved the that is an effective and feasible method. The model provides a new idea for intelligent red tape fault diagnosis of the track circuit.After that, the study aims at shunt malfunction of ZPW-2000A Track Circuit combine with the specific data provided by TCR equipment.Since the faults in different locations have huge effects on the induced voltage.In this paper,the Empirical Mode Decomposition and Fuzzy Entropy were used to extract the parameters which diagnosis of shunt malfunction needed, at the same time,achieve fault diagnosis of shunt malfunction of track circuit through the hybrid algorithm which improved PSO optimizes support vector machine. Comparing with SVM,PSO-SVM, GA-SVM algorithm, this paper provides a fresh diagnose method which is much more rapid and accurate. What's more this paper provides new ideas for shunt malfunction.Finally, this paper summarizes the content of this study, gives an outlook for the future research.