Battery SOC Estimation Based on IAFFRLS-IGWO-AEKF Method

To ensure the safe and stable operation of energy storage systems (ESS) during peak power supply periods, this study proposes an enhanced state of charge (SOC) estimation framework for lithium-ion batteries. By integrating an Improved Adaptive Forgetting Factor Recursive Least Squares (IAFFRLS) method with an adaptive extended Kalman filter (AEKF) optimized by an Improved Gray Wolf Optimizer (IGWO), the proposed method achieves superior dynamic adaptability. Specifically, the IAFFRLS employs a sliding-window root-mean-square error to dynamically adjust the forgetting factor, effectively mitigating the impact of single-point disturbances. Comparative results indicate that the average voltage estimation error is reduced by 58.62% compared to the conventional AFFRLS method. Four dynamic condition tests demonstrate that the proposed IAFFRLS–IGWO–AEKF method achieves average absolute errors of 0.186–0.209% in SOC estimation and 0.069–0.088% in voltage estimation, significantly outperforming two benchmark algorithms and enabling efficient, accurate, and stable SOC estimation for lithium-ion batteries in energy storage systems.