基于混合卡尔曼滤波和H∞滤波的动力电池SOC估计  

Power Battery SOC Estimation Based on Hybrid Kalman Filter and H_∞ Filter

作　　者：于海波[1] 李贺龙[1] 卢扬 王兴媛 YU Hai-bo;II He-long;LU Yang;WANG Xing-yuan(China Electric Power Research Institute,Beijing 100192,China)

机构地区：[1]中国电力科学研究院,计量研究所,北京100192 [2]国网吉林省电力有限公司白山供电公司,运营监测(控)中心,吉林长春130000

出　　处：《电力电子技术》2018年第12期57-60,共4页Power Electronics

摘　　要：动力电池荷电状态(SOC)估计是动力电池管理系统(BMS)中的关键技术和难点。近年来,基于模型的扩展卡尔曼滤波法(EKF)和H_∞滤波法(HIFF)因具有更好的估计性能备受关注。首先选取了常用的Thevenin模型,改进了常用带遗忘因子的递推最小二乘(RLS)参数辨识方法,提出了偏差补偿递推最小二乘法(BCRLS),对数据的有色噪声有很好的抑制作用。然后设计了一种混合EKF/HIFF算法,该算法综合两种算法的优点,拥有更高的精度和更强的鲁棒性,实现了参数和状态的联合估计,有效提高了实车应用的可能。最后利用电池工况试验验证了算法可靠性。Estimation of the power battery state of charge(SOC) is a key technology and difficulty in the power battery management system(BMS).In recent years, model-based extended Kalman filter(EKF) and H_∞ filtering(HIFF)have drawn attention for their better estimation performance.Firstly,the commonly used Thevenin model is selected to improve the commonly used recursive least squares(RLS) parameter identification method with forgetting factor.The bias compensated recursive least square method(BCRLS) is proposed to suppress the colored noise effectively effect.Then,a hybrid EKF/HIFF algorithm is designed.The algorithm combines the advantages of the two algorithms with higher accuracy and robustness and achieves the joint estimation of parameters and states which effectively increases the possibility of real car applications.At last,the reliability of the algorithm is verified by battery condition test.

关 键 词：电池:荷电状态 动力电池管理系统 

分 类 号：TM911[电气工程—电力电子与电力传动]