上海大学学报(自然科学版) ›› 2020, Vol. 26 ›› Issue (3): 413-424.doi: 10.12066/j.issn.1007-2861.2048

• 研究论文 • 上一篇    下一篇

一种变参数模型平方根 UKF 锂离子电池SOC 估计方法

高文凯1, 严利民1(), 孙叠2   

  1. 1. 上海大学 微电子研究与开发中心, 上海 200444
    2. 上海大学 机电工程与自动化学院, 上海 200444
  • 收稿日期:2018-03-24 出版日期:2020-06-30 发布日期:2020-07-07
  • 通讯作者: 严利民 E-mail:yanlm@shu.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(61674100)

A square root UKF with variable-parameter model for Li-ion batteries SOC estimation method

GAO Wenkai1, YAN Limin1(), SUN Die2   

  1. 1. Microelectronic Research and Development Center, Shanghai University, Shanghai 200444, China
    2. School of Mechatronic Engineering and Automation, Shanghai University, Shanghai 200444, China
  • Received:2018-03-24 Online:2020-06-30 Published:2020-07-07
  • Contact: YAN Limin E-mail:yanlm@shu.edu.cn

摘要:

荷电状态(state of charge, SOC)估计是现代电池管理系统 的一个重要方面. 扩展卡尔曼滤波(extended Kalman filter, EKF)等基于锂电池的戴维南等效模型的方法已被广泛用于 SOC 估计, 但其在雅可比矩阵的推导和线性化精度等方面存在不足. 提出了基于变参数模型的平方根无迹卡尔曼滤波(square root unscented Kalman filter, SRUKF)方法估算 SOC, 该方法不需要对非线性模型进行线性化, 同时平方根特性改善了状态协方差的数值性质. 变参数模型是在 2 阶戴维南等效模型的基础上令锂电池的各项参数随电量 变化而得到的, 减小了因固定参数模型无法反映不同电量下参数变化造成的误差. 实验验证了该方法的有效性, 与现有的 SOC 估计方法 EKF、常规的 UKF 以及使用固定参数 模型的估计结果进行了比较, 该方法的误差明显小于其他 3 种方法.

关键词: 锂离子电池, 荷电状态, 变参数模型, 平方根无迹卡尔曼滤波

Abstract:

The state of charge (SOC) estimation is an important aspect of modern battery management system. Methods based on Thevenin Equivalent model for Lithium Battery such as extended Kalman filter (EKF) have been widely used in SOC estimation. However, there are some limitations in the derivation of Jacobian matrix and linearization. In this paper, a square root unscented Kalman filter (SRUKF) based on variable-parameter model is proposed to estimate SOC, which does not need to linearize the nonlinear model. At the same time, the square root property improves the numerical properties of the state covariance. The variable-parameter model is based on the second order Thevenin model, which makes the parameters of the lithium battery change with the quantity of electricity and reduces the error caused by the fixed parameter model that cannot reflect the change of parameters under different electric quantity. The experiment results show that the proposed method is effective. Compared with the existing SOC estimation method, the EKF, UKF and the estimation results from the fixed-parameter model, the error of this method is obviously smaller than that of the other three methods.

Key words: Li-ion battery, state of charge, variable-parameter model, square root unscented Kalman filter (UKF)

中图分类号: