Smart Reconfigurable Battery Packs with Scalable AI Based Cell Balancing

Imbalance between state of charge (SoC) of cells in battery packs can cause numerous issues, including reduction of usable capacity level, degradation of performance, and shortening of lifetime. Successful approaches to mitigate such issues employ cell balancing techniques. Building on our preliminary results, in this paper, we expand our novel cell balancing technique to larger battery packs to investigate scalability aspects and to explore additional machine learning (ML) models that we employ in the proposed cell balancing algorithm. More specifically, we adopt a divide-and-conquer approach, in which the battery pack is divided into smaller partitions to which we apply the proposed ML based cell balancing. Extensive simulation experiments conducted on a 24 cells battery pack demonstrate good scalability and improved battery runtime achieved with the proposed balancing approach.