Addressing the start-stop decision in Multi-Stack Fuel Cell vehicle energy management with Model Predictive Control

Abstract

The total cost of ownership of fuel cell powered heavy-duty vehicles depends largely on hydrogen consumption and stack lifetime. Particularly in the context of Multi-Stack Fuel Cell Systems, the number of start-stop events is a key factor influencing the degradation of the system. Since stack temperature affects system efficiency, and the actual system power impacts waste heat generation, power distribution, and temperature trajectories should be planned simultaneously. In the present paper, a Model Predictive Controller is proposed to optimize the trade-off between hydrogen consumption and start-stop induced degradation by determining optimal inter-stack power splits and stack temperatures. The start-stop decision introduces binary variables into the optimal control problem, rendering it a computationally demanding Mixed-Integer Program. A simulation study shows that the results of the Model Predictive Controller form a Pareto front in the multi-objective trade-off. Compared to simulation results of rule-based approaches, the solutions of the Model Predictive Controller demonstrate a potential for hydrogen savings of up to $\text{14}\text{\%}$ and a reduction in start-stop events of up to $\text{86}\text{\%}$. On an exemplary low-demand cycle, this translates to a projected lifetime fuel cost saving of $\text{\$}4320$. The predictive capabilities of the Model Predictive Controller are best demonstrated in a Hot-Land and a Heat-Up scenario, where it achieves a seamless supply of power through predictive temperature planning. This paper highlights the benefits of using road prediction data in operating strategies of Multi-Stack Fuel Cell Systems.