Robust real-time energy management of flexible traction substation with energy storage and PV for heavy-haul railways

Flexible traction substation (F-TSS), comprising a power flow controller (PFC), energy storage system (ES), and photovoltaic (PV) system, serves as a critical energy nexus for heavy-haul railways. However, the stochastic nature of heavy-duty electric locomotives and PVs poses significant challenges to ensuring the safe and efficient operation of the railway system. Correspondingly, a real-time energy management system framework (R-EMS) is constructed in this paper. At the R-EMS optimization stage, a day-ahead robust optimization approach is introduced to handle worst-case scenarios, ensuring the resilient operation of the F-TSS under uncertainties in freight train power and PV output while maximizing energy savings. At the R-EMS control stage, a direct-type Model-Free Adaptive Predictive Control method with error differential characteristics (ED-dMFAPC) is proposed to enable rapid and precise tracking of day-ahead scheduling plans for the PFC and ES. This method features a simple incremental structure, equivalent to an ideal predictive controller, and incorporates a compensation strategy to mitigate voltage imbalance and support reactive power in real-time. These two stages operate sequentially on day-ahead and intra-day timescales, effectively balancing robustness and economic performance. Simulation-based comparative studies, built on realistic FTPS configurations, are conducted to evaluate the proposed strategy, achieving a 13.39% reduction in costs and a significant improvement in the Integral of Squared Error (ISE) and Integral of Absolute Error (IAE), with reductions of at least 89%.