Abstract: When traditional grouping equalization methods used to reduce battery inconsistency, there were inefficient energy consumption problems in the grouping mode. Therefore, a battery grouping equalization method was proposed based on adaptive circuit topology. A novel adaptive topology for grouping equalization was designed with the characteristics of the Buck-Boost circuit and several switches. Based on the analysis of the shortages of initial cluster centers and the membership matrix should be given for the fuzzy C-means(FCM) clustering algorithm, a new density-based fuzzy C-means (DBFCM) clustering algorithm was proposed from a software perspective for implementing batteries grouping. In each equalization period, DBFCM clustering algorithm was used to group the individual battery into several battery clusters. Based on the proposed adaptive circuit, the clustered batteries were balanced among the groups. The experimental results show that the scheme may effectively improve the overall energy efficiency of the battery pack, reducing equalization time and reducing the inconsistency among single batterys.

Key words: grouping equalization, adaptive topology, clustering algorithm, dynamic grouping

摘要： 传统电池分组均衡方法用于减小电池组不一致性时，其分组方式存在低效耗能问题，因此提出了一种基于自适应电路拓扑的电池分组均衡方法。结合开关组与Buck-Boost电路特点，设计了用于电池分组均衡的自适应电路拓扑，在分析模糊C均值（FCM）聚类算法需给定初始聚类中心及隶属度矩阵等不足的基础上，从软件角度提出了用于实现电池分组的基于密度的模糊C均值(DBFCM)聚类算法。每个均衡周期内，采用DBFCM聚类算法实现电池单体聚类分组，依托自适应电路拓扑对聚类完成的电池各组进行组间均衡操作。实验结果表明，该方案可有效提高电池组的整体能量利用率，缩短均衡时间，减小单体电池间的不一致性。

关键词: 分组均衡, 自适应拓扑, 聚类算法, 动态分组