基于自适应无迹卡尔曼滤波的锂电池SOC估计  

Estimation of state of charge for lithium battery based on adaptive unscented Kalman filter

作　　者：宫兵 凌六一[1] 何业梁 邢丽坤[1] GONG Bing;LING Liu-yi;HE Ye-liang;XING Li-kun(School of Electrical and Information Engineering,Anhui University of Science and Technology,Huainan Anhui 232001,China)

机构地区：[1]安徽理工大学电气与信息工程学院,安徽淮南232001

出　　处：《电源技术》2020年第11期1594-1599,共6页Chinese Journal of Power Sources

基　　金：安徽省高校自然科学基金项目重点项目(KJ2019A0106)。

摘　　要：荷电状态(SOC)是电动汽车电池管理系统的重要组成部分,SOC的准确估计可以有效降低电池的成本,提高电池的利用效率。为解决扩展卡尔曼滤波算法(EKF)中噪声协方差对SOC估计精度的影响问题,采用噪声协方差实时更新自适应无迹卡尔曼滤波算法(AUKF)对电池SOC进行准确估计。对锂电池开展充放电实验,进行离线参数辨识,得到锂电池二阶RC等效电路模型参数。通过建立电池SOC与开路电压、电池模型参数之间的函数关系,利用Matlab/Simulink仿真验证了等效电路模型的正确性。基于验证的二阶RC电路模型采用AUKF和EKF两种算法对实际工况下的单体锂电池SOC进行估计,最大估计误差分别为0.0114、0.0176,平均误差分别为0.0062、0.0079。结果表明AUKF解决了EKF在复杂工况下估算精度较低的问题。SOC(the state of charge)is an important part of the battery management system(application software)of electric vehicles.Accurate estimation of SOC can effectively reduce the cost of batteries and improve the utilization efficiency of batteries.This paper conducts accurate estimate of the battery SOC through the real time update of AUKF by using noise covariance in order to solve the influence of noise covariance in extended Kalman filter algorithm(EKF)on SOC estimation accuracy.It carries out charge-discharge experiment of the lithium battery and identify the offline parameters so as to obtain the second order RC equivalent-circuit model parameter of the lithium battery.By establishing the functional relation among the battery SOC,open circuit voltage and battery model parameter,the correctness of electrical equivalent circuit model is verified through matlab/simulink simulation.Based on the second-order RC circuit model,the actual working state of a single lithium battery is estimated by using AUKF and EKF.The maximum errors of the two estimates are respectively 0.0114 and 0.0176 and their average errors are 0.0062 and 0.0079.The results show that AUKF solves the problem of low estimation accuracy of EKF under complex conditions.

关 键 词：自适应无迹卡尔曼滤波 荷电状态 二阶RC模型 扩展卡尔曼滤波 

分 类 号：TM912[电气工程—电力电子与电力传动]