Transient Fault Diagnosis of Track Circuit Based on MFCC-DTW

With the rapid development of high-speed railway at home and abroad, the sudden failure of track circuit will seriously affect the safety and transportation efficiency. In this paper, a fault diagnosis method of track circuit is proposed based on Meier frequency coefficient and dynamic time regulation model. The fault state of track circuit equipment is analyzed by using transient theory, and the state of track circuit equipment is classified into multiple states. Mayer frequency coefficient and principal component analysis are used to extract features. K-means clustering is used to construct template libraries for different faults. The matching distance between test data and template libraries is compared by DTW model to diagnose faults. Using the measured transient voltage data of track circuit, the performance of the model is tested, and the realization and validation of fault diagnosis are completed. The results show that compared with other machine learning methods, the MFCC-DTW algorithm improves the diagnosis time greatly, and the correct rate is more than 90%. The method classifies the transient state of track circuit and provides an economical and feasible solution for the real-time diagnosis of multiple faults in centralized monitoring system.