Multi-Cell-to-Multi-Cell active equalization method based on k-means clustering and battery pack SOC estimation

Abstract

In this paper, a MC2MC (Multi-Cell-to-Multi-Cell) battery pack active equalization scheme based on reconfigurable circuits and Buck-Boost converters is proposed and validated for the problem of inconsistent single cell power in battery packs. This scheme can simultaneously equalize multiple non-adjacent cells in a single equalization process, improving the equalization time efficiency. Then, based on the relationship between duty cycle and voltage difference, the MC2MC battery pack active equalization control strategy based on adaptive duty cycle is designed. In addition, a k-means based clustering algorithm is used to cluster and reorganize batteries with different charges, and the SOC (State of charge) estimation algorithm of the battery pack is improved and introduced into the equalization control among multiple batteries. Finally, AC2AC (Any-Cell-to-any -Cell), MC2MC, and MC2MC simulations considering the SOC estimation of the battery pack are validated and analyzed for the unbalanced battery packs under MATLAB/Simulink environment, respectively. On the basis of verifying the accuracy of the simulation process, the effectiveness of the MC2MC equalization control method considering the SOC estimation of the battery pack is verified through the comparative analysis of four aspects, namely, equalization time, computational workload, resultant variance, and energy efficiency.