储能科学与技术 ›› 2023, Vol. 12 ›› Issue (3): 941-950.doi: 10.19799/j.cnki.2095-4239.2022.0697

• 储能测试与评价 • 上一篇    下一篇

基于PNGV模型与自适应卡尔曼滤波的铅炭电池荷电状态评估

陈正1,3(), 王志得1,3, 牟文彪2, 祝培旺1,3, 肖刚1,3()   

  1. 1.浙江省清洁能源与碳中和重点实验室，浙江大学能源工程学院，浙江 杭州 310027
    2.浙江省能源集团有限公司，浙江 杭州 310006
    3.浙江大学嘉兴研究院，浙江 嘉兴 314031
  • 收稿日期:2022-11-25 修回日期:2022-12-06 出版日期:2023-03-05 发布日期:2023-04-14
  • 通讯作者: 肖刚 E-mail:3190102446@zju.edu.cn;xiaogangtianmen@zju.edu.cn
  • 作者简介:陈正（2002—），男，本科，研究方向为铅炭电池建模分析， E-mail：3190102446@zju.edu.cn；
  • 基金资助:
    浙江省“领雁”研发攻关计划(2022C03156);国家自然科学基金项目(52176207);浙江省杰出青年基金项目(LR20E060001)

State-of-charge estimation of lead-carbon batteries based on the PNGV model and an adaptive Kalman filter algorithm

Zheng CHEN1,3(), Zhide WANG1,3, Wenbiao MOU2, Peiwang ZHU1,3, Gang XIAO1,3()   

  1. 1.Key Laboratory of Clean Energy and Carbon Neutrality of Zhejiang, College of Energy Engineering, Zhejiang University, Hangzhou 310027, Zhejiang, China
    2.Zhejiang Provincial Energy Group Company LTD, Hangzhou 310006, Zhejiang, China
    3.Jiaxing Institute of Zhejiang University, Jiaxing 314031, Zhejiang, China
  • Received:2022-11-25 Revised:2022-12-06 Online:2023-03-05 Published:2023-04-14
  • Contact: Gang XIAO E-mail:3190102446@zju.edu.cn;xiaogangtianmen@zju.edu.cn

RichHTML

16

PDF

90

摘要：

储能电池应用广泛，准确估计储能电池的荷电状态(state of charge，SOC)对提高电池健康状态有重要意义。铅炭电池作为一种高性能、低成本、高安全性的新型储能电池，在储能电站等场景受到广泛关注，而目前尚缺少铅炭电池SOC估计相关研究。本工作首先通过静流间歇滴定技术探究铅炭电池的荷电状态与开路电压关系，后通过混合脉冲功率性能试验得到铅炭电池的伏安特征数据，建立一阶Thevenin和一阶PNGV等效电路模型，利用基于代理模型和灵敏度分析的随机算法(surrogate optimization algorithm，SOA)对两种等效电路模型进行参数辨识。在此基础上，利用扩展卡尔曼滤波算法(extended Kalman filter，EKF)估计铅炭电池SOC，估算过程考虑噪声干扰。另外，在铅炭电池SOC初值未知的情况下，EKF算法不能准确估计铅炭电池SOC。因此，本工作提出采用自适应扩展卡尔曼滤波算法(adaptive extended Kalman filter，AEKF)对铅炭电池进行状态估计，来弥补EKF的不足。结果表明，在存在噪声且SOC初值未知的情况下，AEKF算法较EKF算法和安时积分法更能准确估计铅炭电池SOC，在给定SOC初值为0.9时，误差最小，为3.91%，验证了算法的有效性与适用性，提高了铅炭电池荷电状态估计的准确性和可靠性。

关键词: 铅炭电池, 荷电状态, PNGV模型, 自适应卡尔曼滤波

Abstract:

Energy storage batteries are widely used, and accurate estimation of the state of charge (SOC) of these batteries is of great significance in improving the state of battery health. In scenarios like energy storage power stations, lead-carbon batteries have drawn increasing interest as a novel-type energy storage battery with high performance, low cost, and high safety. However, there is still no lead-carbon battery SOC calculation available. Following the galvanostatic intermittent titration technique, this study first explores the relationship between the SOC and open circuit voltage of lead-carbon batteries. Then, voltammetry characteristic data of lead-carbon batteries are obtained through hybrid pulse power characteristic test, the equivalent circuit models of first-order Thevenin and first-order PNGV are established, and the parameters of the two equivalent circuit models are identified using a random algorithm based on the surrogate model and sensitivity analysis. On this basis, the extended Kalman filter (EKF) algorithm is used to estimate the SOC of lead-carbon batteries, and noise interference is considered during the estimation process. When the initial value of the lead-carbon battery SOC is unknown, the EKF algorithm cannot accurately estimate its SOC. Therefore, to address the drawbacks of EKF, this study proposes an adaptive extended Kalman filter to estimate the state of lead-carbon batteries. Results show that in the presence of noise and an unknown initial SOC value, the adaptive extended Kalman filter algorithm can estimate the SOC of lead-carbon batteries more accurately than the EKF algorithm and ampere-hour integration method, and the error is the smallest (3.91%) when the initial SOC value is 0.9, which verifies the effectiveness and applicability of the algorithm and improves the accuracy and reliability of SOC estimation of lead-carbon batteries.

Key words: lead-carbon batteries, state of charge, PNGV model, adaptive Kalman filter

中图分类号: