基于RFMRA和改进PNGV模型的锂离子电池SOC估算  

Estimation of Li-ion battery SOC based on RFMRA and improved PNGV model

作　　者：刘冬雷 范永存 王顺利[1] 夏黎黎 LIU Dong-lei;FAN Yong-cun;WANG Shun-li;XIA Li-li(School of Information Engineering,Southwest University of Science and Technology,Mianyang,Sichuan 621010,China)

机构地区：[1]西南科技大学信息工程学院,四川绵阳621010

出　　处：《电池》2021年第5期470-473,共4页Battery Bimonthly

基　　金：国家自然科学基金项目(61801407);西南科技大学自然科学基金(17ZX7110)。

摘　　要：提出基于渐消记忆的递推算法(RFMRA)和改进的新一代汽车合作伙伴(PNGV)等效电路模型的锂离子电池荷电状态(SOC)在线估算方法。RFMRA在传统递推最小二乘算法的基础上加入遗忘因子,使在线参数辨识能更实时地反映模型参数;在PNGV模型上串联RC回路,可更精确地体现电池特性。与离线SOC估算相比,算法的最大误差降低35.106%,平均绝对误差(MAE)降低7.21%,平均绝对值百分比误差(MAPE)降低22.42%,均方根误差(RMSE)降低25.38%。An online estimation method for state of charge(SOC)of Li-ion battery based on recession-fading memory recursive algorithm(RFMRA)and the improved partnership for a new generation of vehicles(PNGV)equivalent circuit model was proposed.In RFMRA,the forgetting factor was added to the traditional recursive least squares algorithm to make online parameter identification be able to reflect model parameters in real time.In PNGV model,RC loop was connected in series to reflect the characteristics of battery more accurately.Compared with off-line SOC estimation,the maximum error,mean absolute error(MAE),mean absolute percentage error(MAPE)and root mean square error(RMSE)were reduced by 35.106%,7.21%,22.42%and 25.38%,respectively.

关 键 词：锂离子电池 新一代汽车合作伙伴(PNGV)模型 荷电状态(SOC) 渐消记忆的递推算法(RFMRA) 

分 类 号：TM912.9[电气工程—电力电子与电力传动]