Optimal Scheduling of Railway Power Supply Systems Integrated with Microgrids: A CCAH-RL Approach

Globally, the pace of railway system electrification and decarbonization is accelerating to support the worldwide sustainability goals for energy and transportation. Increased electricity demand from transport decarbonization, notably within the railway sector, places additional stress on already constrained power grids. Utilizing local renewable generation and energy storage systems to meet this increased demand has emerged as a crucial strategy to enhance the flexibility and resilience of railway power supply systems. This paper proposes to deploy railway microgrids at weak nodes of the existing co-phase traction power supply lines, maximizing the use of current infrastructure (including traction networks and local power grids) to address energy gaps while mitigating exacerbated three-phase voltage imbalance issues caused by increasing electricity demands. A computational cost-aware heuristic reinforcement learning method is developed first for optimal energy scheduling. Then, two agents operating with the proposed algorithm are employed to coordinate energy management between railway microgrids and co-phased feeder stations. Case studies, including a test scenario based on real operating conditions, show that the proposed solution not only accommodates rising electricity demand but also reduces daily operating costs by 15.1% in representative test scenarios, and 19.8% on average under real-world operation conditions. The proposed reinforcement learning algorithm is shown to save up to 81% computation time compared to other conventional approaches. It further substantially strengthens the ability of co-phase traction power supply systems to manage three-phase voltage imbalance, offering a cost-effective mitigation strategy.