Improved chaotic particle butterfly optimization-cubature Kalman filtering for accurate state of charge estimation of lithium-ion batteries adaptive to different temperature conditions

Abstract

Accurate state of charge (SOC) estimation of lithium-ion batteries can effectively help battery management system better manage the charging and discharging process of batteries, providing important reference basis for the use planning of power vehicles. In this paper, an improved chaotic particle butterfly optimization-cubature Kalman filtering (CPBO-CKF) algorithm is proposed for accurate SOC estimation of lithium-ion batteries. Considering the hysteresis characteristics and polarization effects, an improved hysteresis characteristics-dual polarization (HC-DP) equivalent circuit model is established, which can more accurately characterize the internal characteristics of battery. To achieve high-precision SOC estimation, an improved chaotic particle butterfly optimization algorithm is introduced for dynamic optimization of noise in the cubature Kalman filtering algorithm, and the proposed CPBO-CKF algorithm can more accurately describe the actual noise characteristics, thereby reducing estimation errors. The proposed algorithm is validated under complex working conditions at different temperatures, and the results show that it has good accuracy. Under BBDST condition at 15 °C, 25 °C, and 35 °C, the mean absolute errors (MAEs) are 0.80%, 0.56%, and 0.71%, while the root mean square errors (RMSEs) are 1.09%, 0.70%, and 0.88%. Under DST condition, the MAEs are 0.73%, 0.49%, and 0.52%, and the RMSEs are 0.86%, 0.67%, and 0.63%.