Scholarlink

<p>As a core component of the railway signalling system, the condition assessment and maintenance strategy optimisation of point machines are critical to railway safety. However, current issues include vague condition assessment standards and rigid maintenance strategies. This paper employs image processing techniques to extract current curves, constructs a curve difference assessment model based on PCHIP inter-polation and RMSE, and compares it with commonly used methods such as Fréchet distance and Bray-Curtis dissimilarity to classify the switch machine into five state levels. By introducing maximum likelihood estimation combined with Markov chain calculations to determine state transition probabilities, an ageing model is constructed to predict the state changes of the switch machine throughout its entire lifecycle. Results show that the fault repair strategy can extend the physical lifespan of the switch machine to nearly the design value of 14.73 years, far exceeding the maintenance strategy’s 7.29 years. However, from an economic lifespan perspective, the maintenance effects of the two strategies differ by no more than 4 years. This study provides a technical pathway for intelligent maintenance of railway signal equipment, driving the transformation of critical equipment toward “state prediction” and “precision repair”.</p>

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

Markov chain-based turnout state prediction and lifespan simulation

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月17日

查看全文
  • Ziyi Wu,
  • Wanying He,
  • Ziyue Wu,
  • Shichang Gu,
  • Chao Li,
  • Jingjuan Guo

摘要

As a core component of the railway signalling system, the condition assessment and maintenance strategy optimisation of point machines are critical to railway safety. However, current issues include vague condition assessment standards and rigid maintenance strategies. This paper employs image processing techniques to extract current curves, constructs a curve difference assessment model based on PCHIP inter-polation and RMSE, and compares it with commonly used methods such as Fréchet distance and Bray-Curtis dissimilarity to classify the switch machine into five state levels. By introducing maximum likelihood estimation combined with Markov chain calculations to determine state transition probabilities, an ageing model is constructed to predict the state changes of the switch machine throughout its entire lifecycle. Results show that the fault repair strategy can extend the physical lifespan of the switch machine to nearly the design value of 14.73 years, far exceeding the maintenance strategy’s 7.29 years. However, from an economic lifespan perspective, the maintenance effects of the two strategies differ by no more than 4 years. This study provides a technical pathway for intelligent maintenance of railway signal equipment, driving the transformation of critical equipment toward “state prediction” and “precision repair”.