Research on battery state of charge estimation based on variable window adaptive extended Kalman filter
Zhigang He, Xianggang Zhang, Xurui Fu, Chaofeng Pan, Yingjie Jin
Abstract
Battery state of charge (SOC) estimation is one of the indispensable functions in the battery management system. Among them, the Kalman filter series of algorithms are widely used in the battery SOC estimation process. The adaptive extended Kalman filter uses the difference between the measured voltage and the estimated voltage (error innovation) as the innovative covariance matrix update noise in battery SOC estimation. However, the general adaptive Kalman filter does not consider the error innovation in the estimation process. The change. This will lead to inaccurate battery SOC estimation. In order to solve this problem, we proposes a variable window Kalman filter algorithm that takes into account the changes in the error innovation sequence. The algorithm updates the innovation covariance matrix according to the change of error innovation to improve the accuracy of SOC estimation. The results show that compared with the extended Kalman filter and the fixed window Kalman filter, its calculation efficiency is also improved while its accuracy is improved. Finally, its accuracy is verified under different initial parameters, and the study shows that the algorithm proposed in this article is robust.