Accurate SOC Estimation for Zinc–Bromine Flow Batteries Under Sampling Mismatch Using an Improved Adaptive Extended Kalman Filter

ance the accuracy of state of charge (SOC) To enhance the flow batteries, this study addresses estimation for zinc-brommetry between their charging and the pronounced asymmery discharging behaviors. An improved adaptive extended Kalman filter (AEKF) is proposed specifically for the charging phase. The method introduces a sliding-window cross-entropy criterion along with two diagnostic mechanisms to effectively identify SOC initialization errors and sampling mismatches. Based on the identified anomaly types, the algorithm adaptively adjusts the process and measurement noise covariance matrices, enabling dynamic confidence assignment to observation signals. This targeted adjustment significantly improves estimation robustness and accuracy under practical operating conditions. Simulation results demonstrate that the proposed method outperforms the fixed covariance EKF in scenarios involving the initial SOC deviation and sampling mismatch.