Health-aware Fast-charging Control of Lithium-Ion Battery Based on Reinforcement Learning

Lithium-ion battery stands out from many kinds of energy storage devices due to its promising characteristics and has become the energy supply unit of Electric Vehicles (EVs). However, the charging speed of lithium-ion batteries limits the sustainable range of EVs, resulting in "range anxiety" for drivers and necessitating a faster charging method. To fulfill the safety and economic requirements of battery operations, the temperature and degradation of batteries must be taken into account during fast charging control. Consequently, a reinforcement-learning-based health-aware fast charging control scheme is proposed in this paper. First, a coupling model that considers aging and thermal dynamics simultaneously is established to capture battery behaviors. Then, the health-aware fast-charging problem is formulated as a comprise between charging time and battery degradation. After that, a reinforcement learning control scheme is proposed to find the optimal fast charging solution, where the influence of battery parameter drift on the charging strategy is considered. Finally, simulations are performed to validate the effectiveness and superiority of the proposed method. Compared with the commonly used constant current and constant voltage charging strategy, the proposed charging strategy can improve safety and prolong battery lifetime by sacrificing some charging speed.