Turnout Health Assessment Based on Dynamic Time Warping

Abstract

Turnouts are vital components in the rail system. Assessing the healthy state of turnout and ensuring normal operation is essential. However, the faults of rolling element bearings under running occur suddenly. Thus, it is essential to develop a practicable and real-time method to put forward the assess the turnouts’ running state, which could insure the normal operation. This paper presents a novel method based on the dynamic time warping (DTW) to assess the running status. DTW have two superiorities: (1) DTW could calculate the distance between the different state by using calculating the total cost (minimum distance) of two time series, it makes DTW algorithm fast and simply, so the state is assessed timely. (2) The total cost not only be used to diagnose the fault of turnout, but also could be used to transfer to CV (confident values) to assess the state of the turnout. Hence, this method uses DTW to calculate the distance between the different turnout’s states and CV to assess the healthy state by using the distance. Finally, this method is verified by using the power data collected from Guangzhou Metro turnout testbed and Line 4. The result shows that it is accurate for turnout health assessment and suitable for practical scenarios.