Scholarlink

Railway switch systems are of great importance in railway infrastructures, by directing trains along predetermined routes and facilitating the transition between tracks. Given their critical nature, any failure can significantly compromise both the efficiency and safety of train operations, and potentially leading to catastrophic accidents. Thus, the use of predictive maintenance becomes a crucial task to early detect anomalies and guarantee reliable and safe transportation. In this context, the paper aims to develop a diagnosis approach using real data collected during the operation of an electrical railway point machine, specifically the High-Performance Switch System (HPSS), enabling 4 classes of observed faults to be detected and identified. The approach leverages statistical indicators derived from segmented data to train a fault classifier based on Support Vector Machines (SVM). Experimental results obtained from field data validate the effectiveness of the proposed method and demonstrate its ability to accurately detect and classify various faults.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine Learning-Based Approach for Railway Switch and Crossing System Diagnosis

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Zakariya Belghiti,
  • Asma Ladj,
  • Lamiaa Elmenzhi

摘要

Railway switch systems are of great importance in railway infrastructures, by directing trains along predetermined routes and facilitating the transition between tracks. Given their critical nature, any failure can significantly compromise both the efficiency and safety of train operations, and potentially leading to catastrophic accidents. Thus, the use of predictive maintenance becomes a crucial task to early detect anomalies and guarantee reliable and safe transportation. In this context, the paper aims to develop a diagnosis approach using real data collected during the operation of an electrical railway point machine, specifically the High-Performance Switch System (HPSS), enabling 4 classes of observed faults to be detected and identified. The approach leverages statistical indicators derived from segmented data to train a fault classifier based on Support Vector Machines (SVM). Experimental results obtained from field data validate the effectiveness of the proposed method and demonstrate its ability to accurately detect and classify various faults.