State of charge estimation of lithium-ion batteries based on chi-square adaptive unscented Kalman filter

Accurate estimation of the State of Charge (SOC) of lithium-ion batteries (LIBs) is a critical issue in battery management systems and is essential for ensuring the reliability and safety of these systems. To mitigate the issues of significant degradation in SOC estimation accuracy caused by unknown or inaccurate statistical characteristics of system noise, this paper proposes a Chi-square Adaptive Unscented Kalman Filter (CAUKF) SOC estimation method. Initially, a chi-square statistic is devised based on measurement residuals to detect whether the measurement values are anomalous; if the values are normal, the system noise remains unchanged; if they are found to be anomalous, the Sage-Husa adaptive algorithm is employed to correct the system noise. Subsequently, this strategy is integrated into the Unscented Kalman Filter (UKF) framework to adjust the system noise during the SOC estimation process. Finally, the effectiveness of the CAUKF algorithm is validated using the Turnigy Graphene 5000mAh 65C LIB dataset. Experimental results demonstrate that the proposed CAUKF maintains a Root Mean Square Error (RMSE) of less than 2.13% under LA92, US06, and mixed operating conditions, achieving an average SOC estimation accuracy improvement of 48.19% compared to the UKF.