基于FER融合算法的锂电池SOC估计及仿真验证  

Battery SOC estimation and simulation verification based on FER fusion algorithm

作　　者：崔本清 李少林[1] 刘明亮 张晨阳 魏红燕 CUI Benqing;LI Shaolin;LIU Mingliang;ZHANG Chenyang;WEI Hongyan(School of Mechanical and Electrical Engineering,Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]桂林电子科技大学机电工程学院,广西桂林541004

出　　处：《现代电子技术》2021年第22期116-120,共5页Modern Electronics Technique

基　　金：国家自然科学基金(61745126)。

摘　　要：为了提高电池荷电状态估计的准确性,文中提出一种基于带遗忘因子的递推最小二乘法、扩展卡尔曼滤波法与相关向量机算法相结合的FER融合算法。在确定电池复合经验模型的基础上,利用带遗忘因子的递推最小二乘法对其进行参数辨识,利用相关向量机算法建立误差修正模型,并借此修正扩展卡尔曼滤波测量噪声协方差,以实现当模型误差较小时只进行测量更新,而当模型误差较大时只进行过程更新,克服了由于模型误差和系统噪声统计特性的不确定引起滤波发散的问题。仿真结果表明,该算法能有效消除由于模型误差和测量噪声统计特性不确定而引起的荷电状态估计误差,并且具有较好的收敛性和鲁棒性,适用于电动汽车的各种复杂工况,应用价值较高。In order to improve the estimation accuracy of the battery SOE(state of charge),a FER fusion algorithm based on the recursive least⁃squares method with forgetting factor,extended Kalman filtering method and relevance vector machine algorithm is proposed in this paper.On the basis of determination of the battery composite experience model,the recursive least squares method with forgetting factor is used to identify the parameters of the battery model.A relevance vector machine algorithm is used to build the error correction model,by which the measurement noise covariance of extended Kalman filtering is revised to perform only the measurement update when the model error is small,and perform only the process update when the model error is big,which can overcome the problem of filtering divergence caused by the uncertainty of the model error and the statistical characteristics of the system noise.The simulation results show that the proposed algorithm can effectively eliminate the SOC estimation error caused by the uncertainty of the model error and the noise statistical properties,has good convergence and robustness,and is applicable to various complicated driving cycles for electric vehicles.

关 键 词：锂电池 荷电状态估计 仿真验证 FER融合算法 参数辨识 误差修正模型 SOC精度预测 

分 类 号：TN36-34[电子电信—物理电子学] TK018[动力工程及工程热物理]