Model predictive control for energy saving of hybrid trains in catenary-free scenarios

This paper addresses the energy management control problem for railway systems, focusing on a hybrid train equipped with hydrogen fuel cells and lithium-ion batteries, also enhanced by regenerative braking, capable of operating both with and without a catenary connection. Initially, a detailed electrical model is developed, ensuring an accurate representation of the train. This model is then simplified for control purposes, enhancing computational efficiency while maintaining accuracy. Hence, a model predictive control (MPC) strategy with an economic cost is designed, focusing on optimizing energy management and minimizing power losses, while ensuring that the levels of the batteries and fuel cells remain within their optimal ranges. The work, relying on real data provided by the industrial partner Alstom, concludes with a comparative analysis against a heuristic approach, with satisfactory performance in terms of efficiency and reliability.