Scholarlink

<p>Lithium-ion batteries (LIBs) exhibit decoupled time-varying dynamic characteristics between slowly-varying equivalent circuit model (ECM) parameters and rapidly-varying state-of-charge (SOC). Conventional frameworks enforce uniform update periods, inducing redundant ECM recalculations that consume excessive computational resources without commensurate accuracy benefits. This paper develops an adaptive extended Kalman filter (AEKF) algorithm with residual covariance matching for multi-time-scale SOC estimation. A theoretical maximum update period for ECM parameters is established using the Nyquist-Shannon sampling theorem and the battery’s time constant. The impact of different parameter update periods on SOC accuracy is systematically analyzed. An adaptive strategy dynamically adjusts the update frequency based on discharge rate and SOC change, triggering updates only when deviations exceed a significant threshold. Results demonstrate that the method maintains high SOC estimation accuracy while optimizing computational efficiency and provides an optimal balance for real-time applications. Under the Dynamic Stress Test (DST) conditions, the highest SOC estimation accuracy is achieved with the ECM parameters calculated at 4&#xa0;s period. The Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) decrease by 8.91% and 20.87%, respectively, and the algorithm’s running time decreases by more than 8%.</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

Adaptive extended Kalman filter with residual covariance matching for multi-time-scale state of charge estimation

  • 期刊论文

  • 发表日期：

    2026年7月30日

查看全文
  • Dan Liang,
  • Bo Yang,
  • Chang Cao

摘要

Lithium-ion batteries (LIBs) exhibit decoupled time-varying dynamic characteristics between slowly-varying equivalent circuit model (ECM) parameters and rapidly-varying state-of-charge (SOC). Conventional frameworks enforce uniform update periods, inducing redundant ECM recalculations that consume excessive computational resources without commensurate accuracy benefits. This paper develops an adaptive extended Kalman filter (AEKF) algorithm with residual covariance matching for multi-time-scale SOC estimation. A theoretical maximum update period for ECM parameters is established using the Nyquist-Shannon sampling theorem and the battery’s time constant. The impact of different parameter update periods on SOC accuracy is systematically analyzed. An adaptive strategy dynamically adjusts the update frequency based on discharge rate and SOC change, triggering updates only when deviations exceed a significant threshold. Results demonstrate that the method maintains high SOC estimation accuracy while optimizing computational efficiency and provides an optimal balance for real-time applications. Under the Dynamic Stress Test (DST) conditions, the highest SOC estimation accuracy is achieved with the ECM parameters calculated at 4 s period. The Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) decrease by 8.91% and 20.87%, respectively, and the algorithm’s running time decreases by more than 8%.