基于并行卡尔曼滤波器的锂离子电池荷电状态估计  

State of charge estimation of lithium ion battery based on parallel Kalman filter

作　　者：朱奕楠 吕桃林[2] 赵芝芸 杨文[1] ZHU Yi'nan;LÜTaolin;ZHAO Zhiyun;YANG Wen(School of Information Engineering,East China University of Science and Technology,Shanghai 200237,China;Shanghai Space Power Research Institute,Shanghai 200245,China)

机构地区：[1]华东理工大学信息学院,上海200237 [2]上海空间电源研究所,上海200245

出　　处：《储能科学与技术》2021年第6期2352-2362,共11页Energy Storage Science and Technology

基　　金：国家重点研发计划项目(2018YFB0905300,2018YFB0905301)。

摘　　要：针对新能源电动汽车的电量显示与安全管理问题,对其锂离子电池的荷电状态展开研究,提出了基于并行卡尔曼滤波器的全寿命下的电池荷电状态(state of charge,SOC)估计算法。建立了电池Thevenin一阶RC等效电路模型,通过开路实验的数据处理获取静态OCV-SOC关系表达式,并利用具有动态遗忘因子的最小二乘法对模型参数进行了辨识。以安时积分法为状态传递方程,在扩展卡尔曼滤波的基础上利用最大似然估计准则使模型噪声协方差具有自学习能力。考虑模型参数随电池寿命衰减而改变的问题设计并行结构的滤波器来分别进行电池状态估计和参数修正,保证了数据传递中的纯洁性和独立性,从而实现了全寿命下的SOC估计。经过仿真实验验证算法的快速收敛性与实时性,估计精度在2%以内。The state of charge(SOC)of lithium-ion batteries is studied,and a parallel Kalman filter-based SOC estimation algorithm is proposed,with the goal of solving the problem of power display and life prediction in new energy electric vehicles.The Thevenin battery's firstorder RC equivalent circuit model is defined.Data processing of open circuit experiments results in the static OCV-SOC relationship expression.The least-square method with a dynamic forgetting factor is used to identify the model's parameters.The maximum likelihood estimation criterion is used to make the model noise covariance self-learning,using the Ampere-hour integral method as the state transfer equation and the extended Kalman filter as the state transfer equation.Given that the model parameters change as the battery life declines,a parallel structure filter is designed to estimate the battery state and modify the parameters accordingly,ensuring the purity and independence of the data transmission and allowing SOC estimation throughout the life.The simulation results show that the algorithm has fast convergence and real-time performance,and the estimation accuracy is less than 2%.

关 键 词：电动汽车 锂离子电池 扩展卡尔曼滤波 SOC估计 动态遗忘因子 

分 类 号：TP185[自动化与计算机技术—控制理论与控制工程]