基于无迹卡尔曼滤波的动力电池状态估计  

State estimation for power battery based on unscented Kalman filter

作　　者：李锦满 李儒欢 李浩南 李存鑫 邱子桐 郭凯 吴锴[1] 周峻[1] LI Jinman;LI Ruhuan;LI Haonan;LI Cunxin;QIU Zitong;GUO Kai;WU Kai;ZHOU Jun(Department of Electrical Engineering,Xi’an Jiaotong University,Xi’an,Shaanxi 710049,China;Shandong Electric Power Research Institute,Jinan,Shandong 250003,China)

机构地区：[1]西安交通大学电气工程学院,陕西西安710049 [2]国网山东省电力科学研究院,山东济南250003

出　　处：《电池》2024年第3期340-343,共4页Battery Bimonthly

基　　金：国家自然科学基金(52377212);电工材料电气绝缘全国重点实验室项目(EIPE22109)。

摘　　要：准确预测动力电池的荷电状态(SOC)与健康状态(SOH)对电动汽车电池系统的安全运行至关重要。卡尔曼滤波(KF)算法被广泛用于动力电池的状态估计,但非线性误差较大。提出利用无迹卡尔曼滤波(UKF)算法实现对动力电池状态的准确估计。首先,通过分析动力电池实验数据,建立一阶等效电路模型,模型拟合优度达到0.992。随后,加入容量衰退机制模拟锂离子电池老化过程,并对电池进行恒流充电以及随机放电循环,模拟动力电池实际工况。不同初始值下,SOC、SOH估计的均方根误差均小于0.01,且随着循环次数的增加,误差逐渐减小。Accurately predicting the state of charge(SOC)and state of health(SOH)of power battery is vital for the safe operation of electric vehicle battery systems.Kalman filter(KF)algorithm is widely used for power battery state estimation,but with significant nonlinear errors.The unscented Kalman filter(UKF)algorithm is proposed and used to realize accurate power battery state estimation.Firstly,experimental data of power battery are analyzed to establish a first-order equivalent circuit model with a fitting goodness of 0.992.Then,a capacity degradation mechanism is incorporated to simulate Li-ion battery aging.Through galvanostatic charge and random discharge cycles of the battery,actual power battery operating conditions are replicated.The root mean square errors of SOC and SOH estimation are below 0.01 with different initial conditions,and gradually decrease with more cycles.

关 键 词：锂离子电池 状态估计 等效电路模型 荷电状态(SOC) 健康状态(SOH) 无迹卡尔曼滤波(UKF) 

分 类 号：TM912.9[电气工程—电力电子与电力传动]