改进的UKF算法估算锂离子电池SOC

Li-ion battery SOC estimation based on improved UKF algorithm

采用锂离子电池戴维南模型,通过恒流脉冲放电实验结合递推最小二乘法(RLS)辨识模型参数,针对无迹卡尔曼滤波算法(UKF)存在的缺陷,提出改进算法。改进的UKF算法估算锂离子电池荷电状态(SOC)具有更高的精度,估算误差降到1%以内,可加快算法的收敛速度,收敛时间减少200s,并提高算法的稳定性以及自适应能力。

Thevenin model of the Li-ion battery was adopted.The parameters of the model were identified by the galvanostatic pulse discharge experiment and the recursive least square method(RLS).An improved algorithm was proposed for the flaws of the unscented Kalman filter(UKF).The state of charge(SOC)of Li-ion battery was estimated by the improved UKF algorithm with higher accuracy,the estimation error was reduced to less than 1%.The convergence speed of the algorithm was accelerated,the convergence time was reduced by 200 seconds.The stability of the algorithm and adaptive ability also were improved.