Abstract
In order to guarantee the safety and efficiency of the train, a dynamic time warping (DTW) algorithm was introduced to train axle temperature monitoring in this paper with respect to abnormal data of the train axle temperature. After the corresponding data's preprocessing, different axle temperature sequences were recorded into pattern bases as reference sample in the training phase. In the process of recognition, we compared the similarity of real-time train axle temperature and reference sample in the pattern bases. And dynamic time warping distance according to the ideology of dynamic programming is calculated, then we can get two quantitative index of sequences similarity, so as to obtain the condition monitoring results. The simulation results show that the response time is less than 0.4 second and the false alarm rate is less than 3%., meanwhile, it also eliminates the influence of noise and jump.
