无迹卡尔曼滤波法估计锂离子电池的SOC

Estimation of SOC of Li-ion battery based on unscented Kalman filter method

通过电池脉冲放电实验,得到脉冲放电曲线,对曲线回弹段进行二阶指数拟合,结合电压零输入响应,离线辨识锂离子电池二阶RC等效电路模型的参数。为避免非线性函数线性化处理出现的误差,提高算法精度,采用无迹卡尔曼滤波(UKF)估计荷电状态(SOC)。与扩展卡尔曼滤波(EKF)和安时积分法估计相比,UKF的估计误差在1%以内,精度更高。

The pulse discharge curve of the battery was obtained by pulse discharge experiments;the second-order exponential fitting of the rebound curve segment of the pulse discharge curve was carried out.Combined with voltage zero input response,the parameters of the second-order RC equivalent circuit model for Li-ion battery were identified off-line.In order to avoid the error caused by the linearization of the nonlinear function and to improve the accuracy of the algorithm,the unscented Kalman filter(UKF)method was adopted to estimate the state of charge(SOC).Compared with the extended Kalman filter(EKF)and the ampere-hour integration methods,the estimation error of UKF was within 1%,the accuracy was higher.