State of charge prediction of power battery based on dual polarization equivalent circuit model and improved joint algorithm

Abstract

To improve the accuracy of state of charge (SOC) prediction for electric vehicle batteries under dynamic conditions, this paper proposes a novel joint algorithm combining forgetting factor recursive least squares with adaptive extended Kalman filtering (GS-IFFRLS-AEKF). The GS-IFFRLS method is applied for real-time parameter identification of the battery dual-polarization equivalent circuit model, ensuring accurate representation of the battery’s dynamic behavior. Furthermore, the AEKF algorithm is integrated to handle uncertainties and noise caused by varying battery conditions. Simulation and experimental results show that the GS-IFFRLS-AEKF algorithm achieves high accuracy and robustness under HPPC and DST conditions, with a maximum voltage error of 0.08 V and SOC errors reduced to 0.5 %. The method demonstrates excellent performance in dynamic and complex load scenarios, providing an efficient and accurate solution for SOC prediction in electric vehicle.