Research On Fault Diagnosis In Track Circuits Using Case-based Reasoning

Railway signaling equipment plays a big role in ensuring traffic safety, improving transport efficiency, improving working conditions and enhancing the operational management level. As the important infrastructure of railway signaling equipment, track circuit has a pivotal role in the insurance of the railway’s traffic safety and transport efficiency. Track circuit has an extremely wide application in China and with the increase of the railway running mileage, the number of applications continues to expand. According to the fault statistics of the railway signaling equipment, the fault of track circuit occurs most frequently. And once it happens, it not only affects the transportation badly, and sometimes even causes huge damage to human life and property. However, currently track circuit fault diagnosis technology is of low efficiency and develops slowly. Therefore, it is necessary to carry out research on the fault diagnosis of track circuit.Case-based reasoning is a problem solving method based on empirical knowledge. It works especially well in the kind of knowledge domain which has abundant experience and large historical records. The characteristics of track circuit fault areas are accordance with the characters described before. In addition, red light is one of the most frequent track circuit faults. It occurs when the track circuit is under unoccupied condition. Therefore, red light is chosen as the main research object and case-based reasoning is adopted to carry out research on the fault diagnosis of the track circuit in this paper. The main contents and results of this research are as follows:(1) Through the study and analysis of the abundant expert knowledge about the red light, we can get a summary of the main fault modes leading to red light. Then combining the common causes of red light to analyze the relationship between the fault modes and causes, we can choose the characteristic attributes required for case reasoning of red light.(2) According to the principle and process of case-based reasoning, we can conduct requirement analysis to the fault diagnosis system and define the functional requirements and performance requirements of the system completely. After the study and analysis of the system from the level of structure and function, the system architecture and functional structure are obtained, then giving explanations to the specific responsibilities of each functional module. Combining the cycle of CBR, we can design the workflow of the system. (3) Case representation, case retrieval and case retain are the key technologies of CBR. In this paper, there is a lot of study and discussion of those key procedures and after that, the detailed and specific algorithm of each key step is obtained. Rough set is applied to calculate the weights of the attributes. In order to get more objective and reasonable results, different similarity function is designed to different types of attributes. The nearest neighbor method is used to calculate the similarity between cases. The operation of case deletion and case restoration is based on the long-term case reuse condition.(4) Finally, by means of Visual Studio2000, MATLAB and myCBR, the design of the track circuit fault diagnosis system is completed. Through applying the system to diagnose actual track circuit fault cases, the feasibility and effectiveness of the system is verified.