Further development of an AI driven Rolling Stock Fleet Maintenance scheduling optimisation tool in partnership with end users

Rolling Stock maintenance (RSM) is crucial for safe and reliable railway systems. Preventative RSM is scheduled at a fleet level based on fixed intervals (time or train usage), considering maintenance resources and operational requirements. RSM scheduling is largely a manual process making it difficult for planners to develop the most optimal schedules considering all variables. The rail industry needs moving towards prognostic maintenance where component condition is continually monitored, predictions are developed, and components are replaced at the last possible date, to maximise component life. There has been a significant technological advance in condition monitoring; however, there is no widespread implementation of prognostic maintenance, as this significantly increases the complexity of RSM scheduling. To implement prognostic maintenance, advanced decision tools are required to develop optimised schedules considering asset life, safety and operational risk, and resource availability. An initial tool for optimising RSM scheduling has been developed as part of the SRSMRF project. To develop this further, deeper collaborative work with an end-user is required.