Two-Step Sorting and Regrouping Method of Retired Lithium-Ion Battery Cells Based on Branches’ Topological Structures

Abstract

The current consistency of electrochemical performance and the life of retired lithium-ion battery cells for echelon utilization is poor. The existing sorting methods are unable to meet the increasingly stringent performance and safety requirements. In order to improve the consistency and eliminate the small variability of retired cells, this paper proposes a two-step sorting and regrouping method of retired lithium-ion battery cells based on branches' topological structures. Analyzing the main factors affecting the performance and life of the products for echelon utilization, a two-step method is proposed to solve the problem. The iterative self-organizing data analysis technique algorithm is used for preliminary sorting of cells by reservice life and energy efficiency cell-level indicators. With the branch-level indicators of the consistency of branch performance in different topological structures, the results of preliminary sorting are further regrouped by a differential genetic algorithm to ensure consistency. The method proposed in this paper is validated in the public data set. The average value of the silhouette coefficient is 0.6732, the Davies−Bouldin index is 0.4737, and the Calinski−Harabasz index is 976.1366. The discard rate of retired cells for echelon utilization is less than 5%. Compared to the traditional method, the range of applications is wider. The branch-level indicators of topological structures' performance consistency are proposed innovatively, the reservice life and energy efficiency cell-level indicators and branch-level indicators are integrated. The sorting of cells is completed by combining clustering and intelligent optimization algorithms. The different branches' topological structures are given for cells with different performance. It can accommodate more cells with different capacities and improve the utilization efficiency of cells.