A Rolling-Horizon Approach for Predictive Maintenance Planning to Reduce the Risk of Rail Service Disruptions

• Published in: IEEE transactions on reliability Vol. 70; no. 3; pp. 875 - 886
• Main Authors: Consilvio, Alice, Febbraro, Angela Di, Sacco, Nicola
• Format: Journal Article
• Language: English
• Published: New York IEEE 01-09-2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Decision support systems (DSS); Degradation; dynamic planning; Horizon; Integer programming; Linear programming; Maintenance engineering; Maintenance management; Mixed integer; Optimization; Planning; Predictive maintenance; Predictive models; Probability theory; Rail transportation; Rails; railway predictive maintenance; Real time; Risk; Scheduling; Stochastic processes; Strain; Tamping; track degradation stochastic models
• ISSN: 0018-9529; 1558-1721
• DOI: 10.1109/TR.2020.3007504

This article proposes a model for the risk-based scheduling of predictive maintenance activities on a railway line to intervene when a track segment has reached a certain state of degradation, thus preventing faults and possible failures. With the aim of taking into account the stochastic nature of real environments, the rail-track degradation process is represented as a stochastic process, and the failure probability is evaluated as the probability of reaching a degradation threshold. Moreover, a rolling-horizon framework is introduced to manage newly available real-time information and unpredicted faults or maintenance activity delays. Whereas the traditional scheduling models are offline models that cover the long-term horizon but neglect operational disturbances, the presented model allows for dynamic day-to-day planning and adaptation of the maintenance plan to real-time information, thereby responding to the increasing understanding of real-world processes. The optimization problem on maintenance scheduling is formulated as a mixed-integer linear programming problem based on risk minimization, in adherence to ISO 55 000 guidelines. Finally, the application of the approach to a real rail network is reported and discussed, with a focus on the planning of tamping activities at the operational level.