Chinese Journal of Intelligent Science and Technology ›› 2021, Vol. 3 ›› Issue (2): 195-201.doi: 10.11959/j.issn.2096-6652.202120

• Special Topic: Industrial Internet of Minds • Previous Articles     Next Articles

Fault discovery based on text information extraction for on-board equipment of CTCS

Xi CHEN, Runmei LI, Jian WANG, Shuyun DONG   

  1. Beijing Jiaotong University, Beijing 100044, China
  • Revised:2020-11-26 Online:2021-06-15 Published:2021-06-01
  • Supported by:
    The National Key Research and Development Program of China(2018YFB1201500)


The on-board safety computer of Chinese Train Control System generates a large number of Log files during operation every day, which together with the quality analysis record sheet, record the running status of all equipment and provide data support for the fault discovery and processing of the equipment of train control system.The use of these two types of text data is still limited to manual record, query and analysis, which has the problems of low efficiency, strong subjectivity and easy omission and error.The on-board safety computer Log files and the quality analysis record sheet were taken as the research object, and designs an information extraction method and the fault discovery automation scheme to replace the existing manual working mode.The information of Log files and manual recording by the staffs were analyzed, the word segmentation algorithm and information extraction algorithm were selected to find useful information automatically from two kinds of record to avoiding the cumbersome manual search.Then a failure analysis dictionary was built.The automatic fault discovery application was established by using regular expression algorithm.The automatic fault discovery and analysis functions of the application were validated by experiment.

Key words: on-board equipment, on-board safety computer Log file, quality analysis record sheet, fault discovery, in-formation extraction, failure analysis dictionary

CLC Number: 

No Suggested Reading articles found!