Statistical relationships between numerous retired lithium-ion cells and packs with random sampling for echelon utilization

Retired batteries are widely repurposed in energy storage packs as an economical and eco-friendly method to achieve echelon utilization. However, pack performance is strongly affected by variations in retired cells and pack configuration. Quantifying this effect in various pack configurations considering cell-to-cell variations is crucial for predicting the performance of numerous packs. Therefore, under a random sampling scenario, we developed statistical models of relationships between retired cells and packs in terms of capacity and resistance based on probability and statistics, thereby providing a solid theoretical foundation for designing and optimizing the pack structure. It is proven that parallel configuration improves the utilization efficiency and variation of pack-level capacities. Meanwhile, both parallel and series configurations reduce the pack-level resistance variation. Moreover, the statistical capacity performance of packs with parallel connections in series is superior to that of packs with series connections in parallel, although their statistical resistance characteristics are the same. Furthermore, based on the developed models, a capacity screening criterion is proposed that retired cells with a capacity greater than μC-2σC should be accepted in screening process to randomly compose energy storage packs, thereby reducing the capacity variation of packs while making full use of retired cells.