State of charge estimation for lithium-ion batteries under varying temperature conditions based on adaptive dual extended Kalman filter

Abstract

Accurate estimation of the state-of-charge (SOC) is of extreme importance for the reliability and safety of lithium-ion battery operation, for prevention of overcharge, deep discharge, and irreversible damage to batteries. Traditional SOC estimation methods do not consider the effect of temperature on estimation, which may lead to significant errors in the SOC estimation. Considering the effect of temperature change on SOC estimation for lithium-ion batteries, this paper presents a SOC estimation method based on adaptive dual extended Kalman filter (ADEKF). First, the radial basis function neural network (RBFNN) and the forgetting factor recursive least square (FFRLS) methods are adopted to characterize the relationship between battery temperature and polarization resistance and capacitance (RC) of the dual-polarization model based on experimental lithium-ion battery data, then the relationship between identified RC parameters and temperature is described by the temperature characteristic function. Secondly, an estimation method based on the ADEKF is proposed to update the RC parameters online. This method could reduce the SOC estimation error caused by the mismatch between the set RC parameter value and the actual RC parameter value due to ambient temperature changes. Finally, the experimental data of federal urban driving schedule (FUDS) at -10°C, 25°C, and 50°C are selected to simulate and verify the SOC estimation method proposed in this paper. The results indicate that, compared with the method not considering the effect of ambient temperature, the method developed in this paper is able to achieve accurate SOC estimation with a smaller root mean square error and mean absolute error in a wide range of temperatures.