An energy management strategy for fuel-cell hybrid electric vehicles based on model predictive control with a variable time domain

Fuel-cell hybrid electric vehicles (FCHEVs) are emerging as a pivotal model for future new energy vehicles due to their high efficiency and zero pollution. Crafting an effective energy management strategy is crucial for enhancing the robustness and economic viability of these vehicles. Therefore, an energy management strategy (EMS) for FCHEVs was proposed based on model predictive control (MPC) with a variable time domain. This approach integrated driving pattern recognition and model predictive control to overcome the robustness challenges faced by conventional EMS. Variational modal decomposition (VMD) was integrated with a radial basis function neural network for velocity predictions across various frequencies, which mitigated the effect of prediction accuracy on vehicles’ energy efficiency. Additionally, driving pattern recognition (DPR) was used to determine the driving condition online, which enhanced the influence of prediction duration on energy efficiency. This system adapted the prediction duration to the current driving condition to optimize hydrogen consumption. The velocity prediction model achieved 75.61 % of error reduction, which saved 4.40 % of hydrogen based on simulations. Besides, the model demonstrated enhanced adaptability to unpredictable driving environments.