Design of Fault Diagnosis System for Balise Cable Based on Machine Learning

  • Xiaoyi CuiEmail author
  • Jingyang Lv
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 550)


Based on the research of the lineside electronic unit in the high-speed rail system, the design of the fault diagnosis system based on machine learning-based of balise cable is completed. For the cable fault characteristics, the impedance method is used to analyze the impedance of the cable and the magnitude and phase of the current is used for actual fault diagnosis. A FPGA-centric platform is designed to calculate the magnitude and phase of the current generated in cable in real time. Analysis by actual current data, the feasibility of the program has been verified. For the diagnosis of fault characteristics, the way of the classifier is adopted: Logistic Regression, and support vector machine (SVM). This paper briefly introduces the implementation principles of the two models. The training set and test set are constructed from the actual collected data. The two models are trained and tested respectively. According to the accuracy of the test results and the complexity of the model, the model suitable for fault diagnosis with FPGA is selected.


SVM Logistic Regression Impedance Fault diagnosis 


  1. 1.
    Kawai, T., Takinami, N., Chino, T.: A new approach to cable fault location using fiber optic technology. IEEE Trans. Power Delivery 10(10), 85–91 (1995)CrossRefGoogle Scholar
  2. 2.
    Wang, L.: Design and implementation of transponder fault cable detection system. Beijing University of Posts and Telecommunications (2017)Google Scholar
  3. 3.
    Shen, Y., Liu, C.: Circuit Analysis. People’s Posts and Telecommunications Press, Beijing (2004)Google Scholar
  4. 4.
    Harrington, P.: Machine Learning in Reality. People’s Posts and Telecommunications Press, Beijing (2017)Google Scholar
  5. 5.
    Introduction to Support Vector Machines [EB/OL], July.
  6. 6.
    Li, H.: Statistical Learning Method. Tsinghua University Press, Beijing (2017)Google Scholar
  7. 7.
    Zhou, Z.: Machine Learning. Tsinghua University Press, Beijing (2017)Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2019

Authors and Affiliations

  1. 1.Beijing University of Posts and TelecommunicationsBeijingChina

Personalised recommendations