Real-Time Energy Management Strategy of a Fuel Cell Electric Vehicle With Global Optimal Learning

Abstract

This paper proposes a novel energy management strategy for a fuel cell electric vehicle. The strategy combines the offline optimization and online algorithms to guarantee optimal control, real-time performance and better robustness in an unknown route. In particular, dynamic programming is applied in a database with multiple driving cycles to extract the theoretically optimal power split between the battery and fuel cell with a priori knowledge of the driving conditions. The analysis of the obtained results is then used to extract the rules to embed them in a real-time capable fuzzy controller. In this sense, at the expense of certain calibration effort in the offline phase with the dynamic programming results, the proposed strategy allows on-board applicability with sub-optimal results. The proposed strategy has been tested in several actual driving cycles, and the results show energy savings between 8.48-10.71% in comparison to rule-based strategy and energy penalties between 1.04-3.37% when compared with the theoretical optimum obtained by dynamic programming. In addition, a sensitivity analysis shows that the proposed strategy can be adapted to different vehicle configurations. As the battery capacity increases, the performance can be further improved by 0.15% and 1.66% in conservative and aggressive driving styles, respectively.