首页|期刊导航|东北电力技术|一种改进的AEKF算法估算电池的SOC

一种改进的AEKF算法估算电池的SOCOA

An Improved AEKF Algorithm for Estimating SOC of Batteries

中文摘要英文摘要

为提高电池荷电状态(state of charge,SOC)计算的准确性,提升电池参数辨识准确性与 SOC 估算精度是核心关键.为解决传统方法中电池参数固定不变的局限性,首先,采用带遗忘因子的递推最小二乘法(forget factor recursive least squares,FFRLS)结合改进粒子群优化算法(improved particle swarm optimization,PSO),对电池二阶等效电路参数进行辨识与寻优,以获取最优参数值.其次,优化扩展卡尔曼滤波(extended kalman filter,EKF)方法,利用自适应扩展卡尔曼滤波(adaptive extended kalman filter,AEKF)对电池 SOC 进行准确估算,通过混合脉冲功率特性(hybrid pulse power characteristic,HPPC)测试采集 18650 电池数据.最后,将该方法与已有 FFRLS-EKF 方法进行试验数据对比.结果表明,在参数辨识阶段,所提算法对端电压的估计值更接近真实值,优于已有的方法;在 SOC 估算阶段,所提算法可显著提升 SOC 估算准确性.综上,所提 FFRLS-PSO-AEKF 算法有效提升了电池参数辨识与 SOC 估算的精度,为高性能电池管理系统的研发提供了可靠的技术支撑.

In order to improve the accuracy of battery state of charge(SOC)calculation,enhancing the accuracy of battery parameter identification and the precision of SOC estimation is the core key.To address the limitation of fixed battery parameters in traditional methods,first of all,it adopts the forget factor recursive least squares(FFRLS)method combined with the improved particle swarm optimization(PSO)algorithm.It identifies and optimizes the second-order equivalent circuit parameters of the battery to obtain the optimal parameter values.Secondly,it optimizes the extended kalman filter(EKF)method and uses the adaptive extended kalman filter(AEKF)to accurately estimate the battery SOC.It collects data of 18650 batteries through the hybrid pulse power characteristic test.Finally,it compares the experimental data of this method with those of the existing FFRLS-EKF methods.The results show that in the parameter identification stage,the adopted optimized algorithm makes the estimation of terminal voltage closer to the actual terminal voltage compared with the proposed method.In the SOC estimation stage,the proposed algorithm can improve the accuracy of SOC estimation.In conclusion,the FFRLS-PSO-AEKF method improves the accuracy of battery parameter identification and SOC estimation,providing reliable technical support for the development of high-performance battery management systems.

徐俊江;袁学庆;刘军

沈阳理工大学自动化与电气工程学院,辽宁 沈阳 110159||中国科学院沈阳自动化研究所,辽宁 沈阳 110167中国科学院沈阳自动化研究所,辽宁 沈阳 110167沈阳理工大学自动化与电气工程学院,辽宁 沈阳 110159

信息技术与安全科学

参数辨识电池荷电状态粒子群优化算法自适应扩展卡尔曼滤波

parameter identificationSOCPSOAEKF

《东北电力技术》 2026 (8)

28-34,7

评论