Portal de Periódicos da CAPES

Crossing‐Gate Rod Breakage Detection in a Railway Telemeter System Using Machine Learning

2022; Wiley; Volume: 18; Issue: 1 Linguagem: Inglês

10.1002/tee.23710

1931-4981

Tomoaki Kashiwao, Hiroya Tanoue, Noriyuki Shiraishi, Yuki Misaki, Takashi Ando, Kenji Ikeda, Daisuke Tanaka,

Electrical Contact Performance and Analysis

Abstract Shikoku Railway Company (JR Shikoku) has installed a telemeter system whose network spreads across all the railway lines in the Shikoku area. The telemeter system monitors and collects real‐time data, e.g. current, voltage, and relay signals of railway equipment. We propose a method of detecting the breakage of a crossing‐gate rod based on big data using representative machine‐learning techniques, namely, random forest and support vector machine. Moreover, we evaluate the proposed method for rod breakage cases at seven locations and demonstrate its effectiveness and versatility. © 2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.