A multi time-scale framework for state-of-charge and capacity estimation of lithium-ion battery under optimal operating temperature range

To reduce computation cost and improve state-of-charge (SOC) estimation accuracy over battery's whole lifetime, a multi time-scale framework is proposed to co-estimate SOC and capacity in this paper. With forgetting factor recursive least square, model parameters are online identified firstly, which are later transmitted into adaptive extended Kalman filter to predict SOC in real-time. Subsequently, the difference between two estimated SOC before and after macro time-scale is calculated and innovatively seen as measurement information of extended Kalman filter to further update capacity periodically. Considering battery optimal operating temperatures, Federal Urban Driving Schedule tests under 20°C, 30°C and 40°C are performed to verify the feasibility, co-estimation accuracy and adaptability to different macro time-scales of the presented method. The validation results show that the mean absolute error (MAE) and root mean square error (RMSE) of SOC estimation results with three different macro time-scales under optimal operating temperature range can be roughly limited within 1%, while most MAE and RMSE of capacity prediction results is below 1% and 2%, respectively. Moreover, the comparison with other three typical co-estimation methods is also conducted, whose results indicate that the proposed algorithm has more superior comprehensive performance on co-estimation accuracy and convergence speed.