基于FA-DSAEKF算法的车用动力电池荷电状态估计

车用发动机 ›› 2026, Vol. 0 ›› Issue (1) : 71-80.

康恒心¹，王计广²，许建忠³，谭泽飞¹，李加强¹，易乾坤¹

作者信息 +

SOC Estimation for Automotive Power Battery Based on FA-DSAEKF Algorithm

KANG Hengxin¹,WANG Jiguang²,XU Jianzhong³,TAN Zefei¹,LI Jiaqiang¹,YI Qiankun¹

Author information +

文章历史 +

摘要

针对扩展卡尔曼滤波（EKF）在车用动力电池荷电状态（SOC）估计中存在的收敛速度慢、精度不高和鲁棒性较差的问题，提出了一种基于萤火虫算法优化的双对称自适应扩展卡尔曼滤波方法（FA-DSAEKF）。在EKF算法的基础上，通过智能优化初始参数、增强算法对称性与稳定性，并实现噪声协方差矩阵的双参数自适应调整，显著提升了SOC估计性能。试验结果表明，在不同工况、温度与初始状态下，该算法均能快速稳定收敛，最大绝对误差、均方根误差和平均绝对误差均低于0.28%，收敛时间在200 s以内。相较于传统EKF算法，估计误差降低约80%，相较于DSAEKF算法，收敛速度提高83%以上，体现出优异的准确性、适应性和鲁棒性。

Abstract

To address the issues of slow convergence, low accuracy and poor robustness in the estimation of state of charge (SOC) for automotive power batteries using the extended Kalman filter (EKF), a dual-symmetric adaptive extended Kalman filter method optimized by firefly algorithm (FA-DSAEKF) was proposed. Based on the EKF algorithm, the initial parameters were intelligently optimized, the symmetry and stability of the algorithm were enhanced, and the noise covariance matrix was adaptively adjusted using dual parameters, significantly improving the SOC estimation performance. The experimental results show that under different operating conditions, temperatures and initial states, the algorithm can converge quickly and stably with maximum absolute error, root mean square error, and mean absolute error all below 0.28%, and convergence time within 200 seconds. Compared to the traditional EKF algorithm, the estimation error is reduced by about 80%, and compared to the DSAEKF algorithm, the convergence speed is increased by over 83%, demonstrating excellent accuracy, adaptability, and robustness.

导出引用

康恒心, 王计广, 许建忠, 谭泽飞, 李加强, 易乾坤. 基于FA-DSAEKF算法的车用动力电池荷电状态估计[J]. 车用发动机. 2026, 0(1): 71-80

KANG Hengxin, WANG Jiguang, XU Jianzhong, TAN Zefei, LI Jiaqiang, YI Qiankun. SOC Estimation for Automotive Power Battery Based on FA-DSAEKF Algorithm[J]. Vehicle Engine. 2026, 0(1): 71-80