State-of-charge estimation for lithium-ion battery pack based on Unscented Kalman Filter considering cell inconsistency

Abstract

The accuracy of battery state of charge (SOC) estimation has tremendous value in the safe use of battery. However, inconsistencies between batteries will affect the power and usable capacity. When vehicle battery pack contains hundreds of single cell, more efforts should be devoted to develop a high-precision and low-complexity method of SOC estimation. In the article, a mean-difference model (MDM) has been established to precisely characterize the battery characteristics, in which the thevenin model is chosen to be mean model (MM) to characterize the whole battery pack features, and the rint model is used to evaluate the inconsistency between batteries. The mean difference model parameters are identified by means of forgetting factor least square (FFRLS), and then the SOC were estimated by the unscented kalman filter (UKF). The validity of MDM and the UKF is verified on a small series battery pack at last.