储能科学与技术 ›› 2021, Vol. 10 ›› Issue (4): 1454-1462.doi: 10.19799/j.cnki.2095-4239.2021.0124

• 储能测试与评价 • 上一篇    下一篇

基于改进的CDKF锂电池SOC估计方法

张小利1(), 王玥童1, 夏金松1, 张莹莹1,2()   

  1. 1.合肥工业大学电气与自动化工程学院
    2.可再生能源接入电网技术国家地方联合工程实验室，安徽 合肥 230009
  • 收稿日期:2021-03-26 修回日期:2021-04-24 出版日期:2021-07-05 发布日期:2021-06-25
  • 通讯作者: 张莹莹 E-mail:2018110394@mail.hfut.edu.cn;zhangyy@hfut.edu.cn
  • 作者简介:张小利（1989—），男，硕士研究生，研究方向为能源系统设计、能源管理与控制等，E-mail：2018110394@mail.hfut.edu.cn。
  • 基金资助:
    中央高校基本科研业务费专项资金资助项目(PA2020GDKC0019)

Estimation of the SOC of lithium batteries based on an improved CDKF algorithm

Xiaoli ZHANG1(), Yuetong WANG1, Jinsong XIA1, Yingying ZHANG1,2()   

  1. 1.School of Electrical Engineering and Automation, Hefei University of Technology
    2.National and Local Joint Engineering Laboratory of Renewable Energy Access Grid Technology, Hefei 230009, Anhui, China
  • Received:2021-03-26 Revised:2021-04-24 Online:2021-07-05 Published:2021-06-25
  • Contact: Yingying ZHANG E-mail:2018110394@mail.hfut.edu.cn;zhangyy@hfut.edu.cn

RichHTML

39

PDF

244

摘要：

准确估算荷电状态(SOC)可以为电池之间的均衡管理提供依据，延长锂电池组整体的使用寿命。针对中心差分卡尔曼滤波算法(CDKF)存在较大线性误差的问题，提出一种改进的CDKF算法。在原算法中引入迭代滤波思想，多次利用测量信息更新状态量估算值，使得观测信息不断迭代更新，基于LM优化方法不断修正协方差矩阵，有效减小了线性误差。首先基于二阶阻容(RC)电路单元模型，选择最小二乘参数辨识方法，辨识出模型阻容参数；然后进行HPPC实验，验证电池等效模型的准确性；最后分别在恒流放电和动态工况下应用改进后的CDKF算法对电池SOC和电压进行估计，并将估计结果与CDKF算法进行比较。两种工况下验证结果表明改进后的CDKF算法精度更高，SOC估计精度可提升1.16%，最大估计误差小于1.7%，算法收敛时间也比原算法短，改进后的CDKF算法在估计精度和鲁棒性方面均有所提升，更具有应用优势。

关键词: 荷电状态, CDKF, 参数辨识, LM优化方法

Abstract:

Accurate estimation of the SOC can provide a basis for balanced management between batteries and extend the overall service life of a lithium battery pack. In order to address large linear errors in the central differential Kalman filter algorithm (CDKF), an improved CDKF algorithm is proposed. The iterative filtering idea is introduced in the original algorithm, and the measurement information is used to constantly update the estimated value of the state. The observation information is provided iteratively, and the covariance matrix is continuously modified based on the LM optimization method, which effectively reduces the linear error. Based on the second-order resistance-capacitance circuit unit model, the least square parameter identification method is selected to identify the model resistance and capacitance parameters, then HPPC experiments are performed to verify the accuracy of the battery equivalent model. Finally, the improved CDKF algorithm is applied to estimate the SOC and voltage under both constant current conditions and dynamic conditions, and the estimation results are compared to the CDKF algorithm. The results show that the improved CDKF algorithm has higher accuracy, the SOC estimation accuracy can be improved by 1.16%, the maximum error is less than 1.7%, and the algorithm convergence time is shorter than the original algorithm. This improved CDKF algorithm improves the estimation accuracy and robustness and offers numerous application advantages.

Key words: state of charge, central differential Kalman filter, parameter identification, LM optimization method

中图分类号: