Scholarlink

<p>With the deepening of research on zinc-air batteries (ZABs) and continuous advancements in battery management systems (BMS), accurate state of charge (SOC) estimation has become increasingly important for the reliable operation and energy management of ZABs. However, due to their complex electrochemical reactions, oxygen diffusion process, and time-varying polarization characteristics, ZABs exhibit significant nonlinear dynamic behavior, making accurate SOC estimation under dynamic operating conditions challenging. To address this issue, this work proposes an adaptive SOC estimation framework based on a dual polarization (DP) model, forgetting factor recursive least squares (FFRLS), and an extended Kalman filter (EKF). The DP model is used to describe the multi-time-scale polarization behavior of ZABs, while the FFRLS algorithm identifies the time-varying model parameters online. These updated parameters are then introduced into the EKF for recursive SOC estimation and correction. Experimental results show that the proposed DP-FFRLS-EKF strategy achieves accurate SOC estimation under different discharge conditions. The maximum absolute SOC estimation errors are kept within 0.004 and 0.012 under intermittent discharge at 20 and 30&#xa0;mA cm<sup>− 2</sup>, respectively, and remain below 0.007 under the dynamic 20-30-20&#xa0;mA cm<sup>− 2</sup> pulse discharge profile. In addition, comparison with the conventional fixed-parameter EKF method shows that the proposed FFRLS-EKF strategy achieves lower root mean squared error and mean absolute error, confirming the advantage of online parameter identification. These results demonstrate that the proposed adaptive estimation framework can effectively improve the accuracy and robustness of SOC estimation for ZABs, providing a reliable algorithmic basis for real-time battery state monitoring and energy management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

State of charge estimation for zinc-air batteries by combining forgetting factor recursive least squares and extended Kalman filter based on the dual polarization model

  • 期刊论文

  • 发表日期：

    2026年5月26日

查看全文
  • Yaxin Liu,
  • Haoran Li,
  • Zhong Qi,
  • Lulu Pang,
  • Yandan Du,
  • Peng Pan,
  • Huayi Li,
  • Jie He,
  • Yahui Cheng,
  • Hong Dong,
  • Rui Zhang,
  • Zhengchun Yang

摘要

With the deepening of research on zinc-air batteries (ZABs) and continuous advancements in battery management systems (BMS), accurate state of charge (SOC) estimation has become increasingly important for the reliable operation and energy management of ZABs. However, due to their complex electrochemical reactions, oxygen diffusion process, and time-varying polarization characteristics, ZABs exhibit significant nonlinear dynamic behavior, making accurate SOC estimation under dynamic operating conditions challenging. To address this issue, this work proposes an adaptive SOC estimation framework based on a dual polarization (DP) model, forgetting factor recursive least squares (FFRLS), and an extended Kalman filter (EKF). The DP model is used to describe the multi-time-scale polarization behavior of ZABs, while the FFRLS algorithm identifies the time-varying model parameters online. These updated parameters are then introduced into the EKF for recursive SOC estimation and correction. Experimental results show that the proposed DP-FFRLS-EKF strategy achieves accurate SOC estimation under different discharge conditions. The maximum absolute SOC estimation errors are kept within 0.004 and 0.012 under intermittent discharge at 20 and 30 mA cm− 2, respectively, and remain below 0.007 under the dynamic 20-30-20 mA cm− 2 pulse discharge profile. In addition, comparison with the conventional fixed-parameter EKF method shows that the proposed FFRLS-EKF strategy achieves lower root mean squared error and mean absolute error, confirming the advantage of online parameter identification. These results demonstrate that the proposed adaptive estimation framework can effectively improve the accuracy and robustness of SOC estimation for ZABs, providing a reliable algorithmic basis for real-time battery state monitoring and energy management.