首页|期刊导航|汽车工程学报|基于DBO-BP神经网络增强EKF算法的锂电池荷电状态估计

基于DBO-BP神经网络增强EKF算法的锂电池荷电状态估计OA

State-of-Charge Estimation for Lithium Batteries Based on a DBO-BPNN-Enhanced EKF Algorithm

中文摘要英文摘要

针对锂电池荷电状态(SOC)估计中传统BP神经网络存在的初始权值和阈值敏感性问题,提出一种基于蜣螂优化算法(DBO)的BP神经网络优化方法,并将其与扩展卡尔曼滤波(EKF)算法相结合,以提升SOC估计精度.构建了锂电池二阶RC等效电路模型,并通过HPPC脉冲功率试验进行电池参数辨识.在此基础上,采用DBO算法对BP神经网络进行离线优化,在线补偿EKF算法SOC估计误差.为验证所提方法的有效性,在动态应力测试(DST)工况和联邦城市行驶工况(FUDS)下进行了仿真试验.试验结果表明,与传统BP-EKF算法相比,本文方法在SOC估计精度和收敛性方面均表现出显著优势,能更准确地跟踪SOC变化,均方根误差(RMSE)和平均绝对误差(MAE)分别降低了45%和48%.在较大环境噪声条件下,本文方法的准确性提升更显著.研究结果证实,所提出的DBO-BP-EKF算法能有效克服传统方法的局限性,为锂电池SOC估计提供了一种更可靠的技术方案.

To address the sensitivity to initial weights and thresholds in traditional Back Propagation(BP)neural networks for lithium battery state of charge(SOC)estimation,this paper proposes a BP neural network optimization method based on the Dung Beetle Optimizer(DBO)algorithm,which is integrated with the Extended Kalman Filter(EKF)algorithm to enhance SOC estimation accuracy.First,a second-order RC equivalent circuit model for the lithium battery is constructed,and the battery parameters are identified through hybrid pulse power characterization(HPPC)experiments.On this basis,the DBO algorithm is used to optimize the BP neural network offline,thereby providing online compensation for the SOC estimation error of the EKF algorithm.To validate the effectiveness of the proposed method,simulation experiments were conducted under the dynamic stress test(DST)and federal urban driving schedule(FUDS)conditions.The simulation results demonstrate that,compared to the traditional BP-EKF algorithm,the proposed method exhibits significant advantages in SOC estimation accuracy and convergence,enabling more accurate tracking of SOC changes.Specifically,the root mean square error(RMSE)and mean absolute error(MAE)of the proposed method are reduced by 45%and 48%,respectively.In addition,the improvement in accuracy is more significant under high environmental noise.The results confirm that the proposed DBO-BP-EKF algorithm effectively overcomes the limitations of traditional methods,providing a more reliable technical solution for lithium battery SOC estimation.

余曼;张凯;刘建峰;赵轩

长安大学 机械工程与智能装备学院,西安 710064长安大学 汽车学院,西安 710018长安大学 汽车学院,西安 710018长安大学 汽车学院,西安 710018

交通工程

锂电池荷电状态估计EKF算法蜣螂优化算法

lithium batterystate of charge estimationEKF algorithmdung beetle optimization algorithm

《汽车工程学报》 2026 (5)

805-817,13

国家自然科学基金项目(52472397)陕西省重点研发计划项目(2024GX-YBXM-260)西安市"揭榜挂帅"制技术攻关项目(24JBGS0006)咸阳市重点研发计划项目(L2023-ZDYF-QYCX-032)长安大学中央高校基本科研业务费专项资金资助(300102254101)

10.3969/j.issn.2095‒1469.2026.05.08

评论