A Lithium Battery State of Charge Estimation Method Based on Informer Model and Adaptive Kalman Filter

Power batteries are widely used in electric vehicles, energy storage power stations and other fields. The State of Charge (SOC) of a battery indicates the remaining power of the battery. Accurate estimation of the remaining battery power can optimize the performance of the battery and improve the service life of the battery. Due to the influence of internal chemical reactions, temperature and other factors, the state of charge of the battery is difficult to accurately estimate. Moreover, the state of charge estimation of battery has the need of real-time in practical application. In this paper, a combination model of Informer-AKF is formed by coupling between them. The single Informer model has a large fluctuation in SOC prediction results, and AKF filters the output results of Informer to make the results more accurate. The experimental results show that on the INR 18650-20R battery public dataset from the University of Maryland, the model can predict the SOC of batteries under different temperatures. Additionally, on the public dataset from Panasonic18650, the Informer model achieved an average absolute error of 1.78%, while the Informer-AKF model achieved an average absolute error of 1.00%, representing a reduction of 0.78% compared to the Informer model. Comparative analysis was also conducted with the LSTM model. Furthermore, the practicality of the proposed method was validated through data collected from the lead-carbon battery CNF-500H using a BMS system.