Dynamic fleet maintenance management model applied to rolling stock

•Manage dynamically the RUL, existing maintenance limitations, and operation demand.•Considers not a generic predictive maintenance RUL prognosis but also the failure mode.•Programming problem is applied to a case study on a fleet of high-speed trains in spain.•The model can be adapted for other sectors with similar operational constraints. This paper presents a model for optimising fleet maintenance management with a particular application to train rolling stock fleets. The proposed model produces a joint schedule for train operations and opportunistic predictive maintenance activities with an aim to maximise operational useful life. The model opportunistically allocates predictive maintenance interventions to existing preventive maintenance schedule considering the estimated remaining useful life (RUL) of critical components whilst ensuring fleet availability to meet operational demands as well as resource and time constraints at the maintenance depots. The proposed methodology is described in three phases: (i) definition of the operating context and maintenance resources; (ii) evaluation of feasible opportunistic maintenance timeslots; (iii) optimal maintenance and operations scheduling. The optimisation model, developed as a Mixed Integer Linear Programming problem, is applied to a real industrial case study on a fleet of high-speed trains in Spain. The results show significant improvement in the utilisation of operational life of components compared to the current policies used by the company. Although the model was developed with particular consideration to the train fleets, it can be adapted for other sectors such as bus fleets and airlines with similar operational constraints.