Model-Predictive Energy Management for Optimizing the System Efficiency of a Fuel Cell Truck Considering Fuel Cell Degradation

This work presents a model predictive control (MPC) approach for optimizing the powertrain system efficiency of a fuel cell electric heavy-duty vehicle. The MPC determines the power split between the high-voltage traction battery and the fuel cell system so that the power demand of the drive cycle is met, while a second objective extends the lifetime of the fuel cell stack by including stack degradation in the cost function. Based on driving profile information, the controller adjusts the fuel cell power trajectory within a specified prediction horizon such that a target state of charge of the traction battery is reached at the end of the time-discrete horizon, fuel consumption is minimized, and excessive degradation is avoided. The control variables are computed from discrete-time models of the fuel cell, the truck, and the battery. The resulting non-linear optimization problem is solved with the open-source software package acados, integrated into MATLAB Simulink to achieve real-time capability. The MPC is implemented and tested in a model-in-the-loop environment, and its real-time execution is verified on a dSPACE MicroAutoBox III rapid control prototyping unit. On the Vehicle Energy Consumption Calculation Tool (VECTO) Long Haul cycle, hydrogen consumption is reduced by 6.5%, while membrane thinning and the loss of electrochemically active surface area are reduced by 6.1% and 2.3%, respectively, compared with a rule-based strategy.