Battery hysteresis compensation modeling and state‐of‐charge estimation adaptive to time‐varying ambient temperature conditions

Summary Temperature and cell hysteretic effects are two major factors that influence the reliability and safety in long‐term management of battery‐integrated systems. In this paper, a hysteresis‐compensated electrical characteristic model is established to track the terminal voltage of batteries with the uncertain hysteretic effect of the open‐circuit voltage. Then, an autoregressive exogenous model with multi‐feature coupling is employed for the identification of the parameters to make them adaptive to the uncertainties of the temperature and hysteretic effects. After that, a novel method for state‐of‐charge (SOC) estimation based on an adaptive moving window‐square root unscented Kalman filter is constructed to avoid the filtering divergence problem caused by the negative error covariance matrix. Multiple constraints, such as Coulombic efficiency, varying ambient temperatures, and hysteresis voltage, are considered for the SOC estimation. Experimental results show that the root‐mean‐square error for SOC calculation can be limited to 0.0211 when the temperature varied up to 40°C and the root‐mean‐square error of the voltage measurement noise up to 61.9 mV. The proposed method provides an effective way for battery‐integrated management of electric vehicles. A hysteresis‐compensated electrical characteristic model is established for terminal voltage tracking. An autoregressive exogenous model with multi‐feature coupling is employed for parameter identification. An adaptive moving window‐square root unscented Kalman filter is proposed for SOC estimation. The proposed method is adaptive to uncertainties in temperature and cell hysteretic effects.