Scholarlink

<p>In this article, a data-driven model based on the Markov decision process approach is applied to wheelsets from freight wagons, providing a method to support condition-based maintenance for these components. This study analyses observed degradation data of freight wheelsets and develops a Markov decision process model. A comparison between key operating variables, namely, mileage since the last maintenance and gross ton mileage since the last maintenance, is also analysed to determine which key operating variable is more appropriate for developing a bi-dimensional degradation model with wheel tread diameter. An estimation of Markov transition matrices for various actions is performed. An optimal strategy is computed, along with a decision map that determines the best actions based on the current state of the wheelset. Decision maps are generated for linear and quadratic terms of the key operating variables. We then compare the optimal cost-minimizing policies with a simplified policy that fixes the gross ton mileage threshold (the key variable). This comparison highlights the trade-off between practical applicability and cost efficiency. The findings provide railway companies with decision rules that can reduce renewal costs and prevent in-service failures, supporting more efficient maintenance planning.</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

Optimization of condition-based maintenance for wheelsets in freight wagons: a Markov decision process approach

  • 期刊论文

  • 发表日期：

    2026年1月21日

查看全文
  • Abhishek S. Bhadouria,
  • Joaquim A. P. Braga,
  • Rajesh P. Mishra,
  • António R. Andrade

摘要

In this article, a data-driven model based on the Markov decision process approach is applied to wheelsets from freight wagons, providing a method to support condition-based maintenance for these components. This study analyses observed degradation data of freight wheelsets and develops a Markov decision process model. A comparison between key operating variables, namely, mileage since the last maintenance and gross ton mileage since the last maintenance, is also analysed to determine which key operating variable is more appropriate for developing a bi-dimensional degradation model with wheel tread diameter. An estimation of Markov transition matrices for various actions is performed. An optimal strategy is computed, along with a decision map that determines the best actions based on the current state of the wheelset. Decision maps are generated for linear and quadratic terms of the key operating variables. We then compare the optimal cost-minimizing policies with a simplified policy that fixes the gross ton mileage threshold (the key variable). This comparison highlights the trade-off between practical applicability and cost efficiency. The findings provide railway companies with decision rules that can reduce renewal costs and prevent in-service failures, supporting more efficient maintenance planning.