Data-driven design of a derailment detection system for freight wagons

Abstract

Railway transportation is widely known for its high safety level. Nevertheless, derailments still occur, typically causing large service disruptions and economic losses. For these reasons, identifying derailments is of paramount importance for improving the railway transportation safety. This paper presents the design and testing of a derailment detection system for freight trains. A derailment detection algorithm able to process and analyze data in real-time has been designed leveraging the knowledge gained from full-scale derailment tests on freight wagons. Moreover, a LoRa-based wireless communication system has been implemented to relay freight wagon derailment warnings to the locomotive. The effectiveness of the proposed derailment detection algorithm is proved through its deployment to a freight wagon used for common operations. In that, it demonstrated increased robustness against derailment false triggering compared to traditional mechanic-pneumatic detectors. Furthermore, the proposed wagon-locomotive wireless communication system is proven effective in transmitting information across various convoy compositions and freight wagon loading conditions during operational testing on the railway line.