Low complexity state-of-charge estimation for lithium-ion battery pack considering cell inconsistency

The estimation accuracy of State-of-Charge (SOC) has great significance for cell charge/discharge management, balance control, and safety management. However, some inconsistencies among cells may lead to model mismatch, which will have a significant impact on the estimation results. Furthermore, it is difficult to achieve an accurate SOC estimation with less computational burden when the battery pack contains hundreds (or thousands) of cells. This paper constructs a battery pack model that consists of a mean-model and several difference-models. The mean-model represents the overall performance of the battery pack. The difference-model describes the inconsistencies of SOC, internal resistance, and coulombic efficiency among cells. In addition, a joint algorithm is proposed to estimate SOC with low complexity. It includes a dual time-scale adaptive extended Kalman filter (AEKF) and a SOC correction method. Comparing with the conventional filter method of SOC estimation, it can effectively simplify a large number of matrix operations, and greatly reduce the overall computational complexity. The experimental results show that the established model can well describe the dynamic characteristics of the battery pack, and the proposed method can track the changing of SOC at a low computational cost.