An Improved Equivalent Circuit Model and a Variable Scale Unscented Kalman Filter for State-of-Charge Estimation of Lithium-Ion Batteries at Variable Ambient Temperatures

Abstract

State of charge (SOC) is a key index of lithium-ion batteries, and the equivalent circuit model is widely used in engineering to estimate the SOC of batteries, but its estimation accuracy is poor under complex dynamic conditions. To solve this problem, an improved equivalent circuit model (IECM) is first proposed, which considers the long-term polarization effect due to solid-phase diffusion of lithium-ion, and corrects the voltage by the lithium-ion concentration difference. Then, the parameters of IECM at different ambient temperatures are identified for temperature compensation. Finally, a variable scale unscented Kalman filter (VSUKF) is proposed to improve the dynamic performance of SOC estimation by changing the sampling interval distribution through predicted error. The accuracies of terminal voltage and SOC estimation are compared and validated under different working conditions and temperatures. The results show that the maximum root mean square error (RMSE) and the mean absolute error (MAE) of the predicted terminal voltage of the proposed IECM are 38.3 mV and 15.2 mV, and the maximum RMSE and the MAE of the SOC estimation results of VSUKF are 1.70% and 1.46%. This results show that the proposed model and method have high accuracy and can provide some technical support for SOC online estimation.