A Novel Dual-Model MPC-Based Energy Management Strategy for Fuel Cell Electric Vehicle

Carbon neutrality policies stimulate hydrogen technologies, making fuel cell electric vehicles (FCEVs) a potential solution for efficient travel. The integration of advanced energy management strategies (EMSs) into FCEVs can tap into the energy-saving potential of the nonlinear powertrain across diverse driving environments. In this article, a novel data-driven model-predictive-control (MPC)-based EMS with a dual-model framework (DDMPC) is proposed for FCEVs to substantially enhance economic performance while maintaining control robustness. First, to improve the control robustness, a dual-model framework for MPC, constituted by a nominal and a practical system, is established. Next, the nominal system is constructed using a mathematical equivalent circuit model with precise internal parameters. Subsequently, a Gaussian process (GP) capable of real-time application is employed to establish the practical system, which aids in accurately predicting state transitions within the predictive horizons, thus bolstering tolerance against external disturbances. In addition, a novel control input fusion function is designed to be integrated into DDMPC. It is purposed to cooperatively process the control vectors of both subsystems, thereby effectively merging the control inputs to enhance the economic performance of the powertrain. Finally, the simulations and hardware-in-the-loop (HIL) tests evaluate the energy-saving potential and the feasibility of the proposed DDMPC strategy.