储能科学与技术 ›› 2021, Vol. 10 ›› Issue (6): 2334-2341.doi: 10.19799/j.cnki.2095-4239.2021.0205

• 储能测试与评价 • 上一篇    下一篇

基于BCRLS-ACKF的锂离子电池荷电状态估计

苏航1(), 高怀斌1(), 李争光1, 李洪峻2, 刘剑飞2, 佐晓波2, 纪林林2   

  1. 1.西安科技大学机械工程学院,陕西 西安 710054
    2.中国人民解放军32181部队,陕西 西安 710032
  • 收稿日期:2021-05-12 修回日期:2021-05-29 出版日期:2021-11-05 发布日期:2021-11-03
  • 作者简介:苏航(1995—),男,硕士研究生,研究方向为新能源汽车电池管理系统,E-mail:1197657940@qq.com|高怀斌,副教授,研究方向为新能源汽车电池管理、温差发电等,E-mail:gaohuaibin@xust.edu.cn
  • 基金资助:
    国家自然科学基金(5217061383)

State of charge estimation of Li-ion battery based on BCRLS-ACKF

Hang SU1(), Huaibin GAO1(), Zhengguang LI1, Hongjun LI2, Jianfei LIU2, Xiaobo ZUO2, Linlin JI2   

  1. 1.School of Mechanical Engineering, Xi′an University of Science and Technology, Xi'an 710054, Shaanxi, China
    2.32181 Troops of PLA, Xi'an 710032, Shaanxi, China
  • Received:2021-05-12 Revised:2021-05-29 Online:2021-11-05 Published:2021-11-03

摘要:

精确的锂离子电池荷电状态(state of charge,SOC)估计对于电池管理系统至关重要。模型参数辨识是SOC估计的前提,也是影响其估计精度的关键因素。为了有效避免噪声对参数辨识的影响,采用偏差补偿递推最小二乘法(BCRLS)进行在线参数辨识。在此基础上,采用自适应容积卡尔曼滤波(ACKF)算法估计电池SOC,对系统噪声进行实时更新以提高估计精度。此外,对于计算过程中由于协方差矩阵失去正定性而出现平方根无法分解的问题,利用奇异值分解的方法代替Cholesky分解,以提高数值计算的稳定性。最后将BCRLS与ACKF相结合以实现模型参数和SOC的联合估计,并在不同工况和初始值不精确的情况下进行算法验证,结果表明本文所提算法具有较高的精度,平均绝对误差在2%以内。

关键词: 荷电状态, 偏差补偿最小二乘法, 奇异值分解, 自适应容积卡尔曼滤波, 联合估计

Abstract:

Accurate estimation of the state of charge (SOC) of Li-ion batteries is essential for the battery management system. Model parameter identification is the premise of SOC estimation and the key factor affecting its estimation accuracy. Bias compensation recursive least squares (BCRLS) is used for online parameter identification to effectively avoid noise's influence on parameter identification. The adaptive cubature Kalman filter (ACKF) algorithm is used to estimate the battery SOC on this basis, and the system noise is updated in real-time to improve estimation accuracy. In addition, for the problem that the square root cannot be decomposed due to the loss of positive definiteness of the covariance matrix in the calculation process, the singular value decomposition method is used instead of Cholesky decomposition to improve the stability of numerical calculation. Finally, BCRLS and ACKF are combined to realize joint estimation of model parameters and SOC, and the algorithm is validated under various working conditions and with incorrect initial values. The results show that the algorithm proposed in this paper has high accuracy, and the average absolute error is within 2%.

Key words: state of charge, BCRLS, singular value decomposition, ACKF, joint estimation

中图分类号: