Fault diagnosis of rail turnout system based on case-based reasoning with compound distance methods

With the rapid development of the high-speed railway, the safety and reliability problem of the railway bring more rigid requirement to the high-speed railway industry. This paper focuses on the diagnosis problem of railway turnout system, which is the most important infrastructure elements of the rail system. According to the characteristics of the turnout mechanism and in order to improve the diagnosis performance, CBR method with compound distance metrics is proposed. First, different distance metrics (Euclidean metric, dynamic time warping metric) are calculated in the CBR to get the single result; then a group decision making strategy is adopted to combine the two metrics to get the final result. The experimental results indicate that the proposed method achieves a better performance, which prove that the proposed could improve the diagnosis performance, and guarantee the operation safety of high-speed rail way.