Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace

Research Article

An Intelligent Fault Diagnosis Model Based on FastDTW for Railway Turnout

Download366 downloads
  • @INPROCEEDINGS{10.4108/eai.27-8-2020.2295306,
        author={WEN-Jiang  JI and Yuan  ZUO and XING-Hong  HEI and rong  Fei},
        title={An Intelligent Fault Diagnosis Model Based on FastDTW for Railway Turnout},
        proceedings={Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace},
        publisher={EAI},
        proceedings_a={MOBIMEDIA},
        year={2020},
        month={11},
        keywords={turnout; switch machine; fault diagnosis; fastdtw},
        doi={10.4108/eai.27-8-2020.2295306}
    }
    
  • WEN-Jiang JI
    Yuan ZUO
    XING-Hong HEI
    rong Fei
    Year: 2020
    An Intelligent Fault Diagnosis Model Based on FastDTW for Railway Turnout
    MOBIMEDIA
    EAI
    DOI: 10.4108/eai.27-8-2020.2295306
WEN-Jiang JI1,*, Yuan ZUO1, XING-Hong HEI1, rong Fei1
  • 1: School of Computer Science and Engineering, Xi’an University of Technology
*Contact email: wjj@xaut.edu.cn

Abstract

The turnout handles the direction of the train which is one of the key equipment in the railway transportation system. In this paper, by using real action current data obtained from switch machine model No.ZD7, a turnout fault diagnosis model based on the FastDTW pattern recognition algorithm was proposed. Firstly, the original current curve was segmented relate to the features of them. Then the warping path distance between the standard sample and the tested current curve was obtained according to FastDTW algorithm. Finally a dynamic optimized threshold was used to confirm whether there is a fault happened in the turnout. According to the experiment results, the proposed diagnose model without the prior knowledge of fault samples can works well both with single and double acting type turnout machines, owning to the following elements: the diagnose accuracy can be more than 96%, the time-cost can be improved more than 5 times compared with traditional DTW based algorithms