基于改进BP-EKF算法的SOC估算  

Research on SOC estimation based on improved BP-EKF algorithm

作　　者：田冬冬 李立伟 杨玉新 TIAN Dong-dong;LILi-weil;YANG Yu-xin(School of Elctrical Egincering,.Qingdao University,Qingdao Shandong 266071,China;Library of Qingdao University,Qingdao Shandong 266071,China)

机构地区：[1]青岛大学电气工程学院,山东青岛266071 [2]青岛大学图书馆,山东青岛266071

出　　处：《电源技术》2020年第9期1274-1278,共5页Chinese Journal of Power Sources

基　　金：山东省自然科学基金(Y2008F23);山东省科技发展计划项目(2011GGB01123);山东省重点研发计划项目资助(2017GGX-50114)。

摘　　要：准确估测电池当前荷电状态(SOC)是电池储能系统是否安全可靠的重要指标。根据锂电池内部实际动态特性,提出一种改进BP神经网络和扩展卡尔曼滤波(EKF)相结合的锂离子动力电池SOC估计方法。优化BP神经网络前馈分析计算解决传统BP信噪比低的问题,将训练成功的改进BP神经网络用于补偿EKF算法的估计误差,最后在Matlab/Simulink上搭建仿真模型进行实验。结果表明,与单纯的EKF算法相比,所提出的改进SOC估计方法的估算误差在2%以内,具有良好的矫正性和鲁棒性,能有效提高SOC的估计精度。Accurate estimation of the battery's current state of charge(SOC)is an important indicator to judge whether the battery energy storage system is safe and reliable.Based on the actual dynamic characteristics of lithium batteries,an SOC estimation method for lithium ion power batteries based on an improved BP neural network and extended Kalman filter(EKF)was proposed.The BP neural network feedforward analysis calculation was optimized to solve the problem of low traditional BP signal-to-noise ratio.The successfully improved BP neural network was used to compensate the estimation error of the EKF algorithm.Finally,a simulation model was built on Matlab/Simulink for experiment.The results show that compared with the simple EKF algorithm,the estimated error of the proposed improved SOC estimation method is within 2%,which has good correction and robustness,and can effectively improve the accuracy of SOC estimation.

关 键 词：荷电状态 改进BP神经网络 扩展卡尔曼滤波 

分 类 号：TM912.9[电气工程—电力电子与电力传动]