SOC estimation of EV batteries based on Temperature-Compensation model and Robust Extended Kalman Filtering

State of Charge (SOC) is a key part of Electric drive vehicle (EV) battery management system. Aiming at the dynamic nonlinear characteristics of SOC, an improved method based on Temperature-Compensation Model (TCM) and Robust Extended Kalman Filtering (R-EKF) was proposed to accurately estimate the changes of battery SOC at different temperatures. Firstly, the corresponding parameters are obtained through the battery test experiments at different temperatures, and then a TCM representing the equivalent circuit of the battery at different temperatures is established in MATLAB. Different from the traditional EKF algorithm ignoring the high-order terms, this paper increases the highorder terms when the EKF performs Taylor series expansion to improve the estimation accuracy. Experimental results show that TCM and R-EKF algorithm (R-EKF-T) can improve the accuracy, convergence speed and robustness of SOC estimation. The estimation error can be controlled effectively.