基于自适应无迹卡尔曼滤波算法的锂电池荷电状态预测  

Prediction of Li-ion Battery Charge State Based on Adaptive Untraced Kalman Filtering Algorithm

作　　者：蒙永龙 艾学忠[1] 郑巍 王明达 汪冬冬 MENG Yong-long;AI Xue-zhong;ZHENG Wei;WANG Ming-da;WANG Dong-dong(College of Information and Control Engineering,Jilin Institute of Chemical Technology)

机构地区：[1]吉林化工学院信息与控制工程学院

出　　处：《化工自动化及仪表》2024年第2期294-300,共7页Control and Instruments in Chemical Industry

基　　金：吉林省科技发展计划(批准号:20190302063GX)资助的课题。

摘　　要：针对无迹卡尔曼滤波在噪声不稳定和工况复杂的情况下锂电池荷电状态预测准确度低的问题,提出基于二阶等效RC电路模型,采用遗忘因子递推最小二乘法对模型参数进行辨识,使用自适应无迹卡尔曼滤波算法(AUKF)对锂电池荷电状态进行预测,最后在DST数据工况下,验证预测模型的准确性。对无迹卡尔曼滤波(UKF)算法和提出的AUKF算法进行仿真对比,结果表明:所提算法的最大误差在±0.02之内,预测精度更高、适用性更强。Considering the fact that unscented Kalman filter(UKF) has low accuracy in predicting Li-ion battery charge state in the case of unstable noise and complex working conditions,a model based on second-order equivalent RC circuit was proposed,including having the forgetting factor recursive least square(FFRLS) method employed to identify model parameters,and the UKF algorithm adopted to predict Li-ion charge battery state,as well as verify the accuracy of the prediction model in DST data condition.The simulation results show that,the maximum error of the proposed UKF algorithm stays within ±0.02,and its accuracy is higher and the applicability is stronger.

关 键 词：锂电池 荷电状态 自适应无迹卡尔曼滤波 遗忘因子递推最小二乘 

分 类 号：TM912[电气工程—电力电子与电力传动]