Hierarchical optimization of battery state of charge planning and real-time energy management for connected fuel cell electric vehicles

The variability and complexity of driving conditions pose significant challenges to the energy management of fuel cell electric vehicles (FCEVs). The emergence of connected and autonomous vehicle technologies offers new opportunities for predictive energy management strategies (EMS). This paper proposes an advanced hierarchical EMS to enhance adaptability to diverse driving scenarios and minimize energy consumption. In the upper layer, an iterative dynamic programming (IDP) algorithm is developed to plan the reference trajectory of the battery state of charge (SOC), leveraging long-horizon traffic information to guarantee the optimality of the strategy. In the lower layer, the model predictive control (MPC) algorithm is implemented to achieve real-time energy optimization and reference tracking, with a fast-solving algorithm incorporated to reduce computation time to the millisecond level. The simulation results validate the effectiveness of the proposed strategy, demonstrating a reduction in energy consumption by 0.75%–9.12% compared to traditional MPC methods, while the results are close to the theoretical optimal value. • A HMPC for FCEVs is proposed to enhance the adaptability to driving conditions. • A speed prediction algorithm based on GPR is proposed to improve the adaptability. • The IDP is integrated for long-horizon reference trajectory planning . • A fast-solving model predictive control algorithm is developed for real-time EMS.