Agent-based Active Balancing Control Method for Lithium-ion Battery Packs

A battery pack can face energy imbalance among its cells resulting from cell-to-cell variation in capacity, internal resistance, and some manufacture parameters. So the desired operation mode needs the dynamic energy equalizing to adjust each cell’s state-of-charge (SOC) to the certain level. In this paper, the cell equalizing system for a serially connected battery pack has already modeled as a multi-agent system (MAS). In this paper, we present a novel distributed control approach for active balancing of battery cells facilitating a scalable and flexible battery system design to overcome imbalance problems. For the proposed approach, the topology of the underlying battery system is assumed to be not predefined and can change at runtime. In such a system, each battery cell may have its own power electronics and processing unit which are used to manage the control actions for active balancing on cell level. Thus, the distributed balancing approach can be useful in this situation. Each agent would be able to locally computes it’s SOC and optimizes its own control actions based on cooperative control. Each cell is required to find a single partner or multiple partners to form coalitions and exchange energy for active balancing. The working principle and capability of the distributed control approach is evaluated by simulations and compared to a passive balancing approach and other control schemes that has already existed in literature