基于动态优选遗忘因子最小二乘在线辨识的磷酸铁锂电池SOC估算  

Lithium Iron Phosphate Battery SOC Estimation Based on the Least Square Online Identification of Dynamic Optimal Forgetting Factor

作　　者：王浩[1] 郑燕萍[1] 虞杨 Wang Hao;Zheng Yanping;Yu Yang(Nanjing Forestry University,Nanjing 210037)

机构地区：[1]南京林业大学,南京210037

出　　处：《汽车技术》2021年第10期23-29,共7页Automobile Technology

基　　金：江苏省重点研发计划项目(BE2017008)。

摘　　要：为提高行驶中电动汽车的动力电池荷电状态(SOC)估算精度,以磷酸铁锂电池为例,提出了一种基于粒子群优选遗忘因子最小二乘(PSO-FFRLS)在线辨识模型的电池SOC估算方法。首先建立电池二阶等效电路模型,然后运用粒子群优化(PSO)算法实时为遗忘因子递推最小二乘法(FFRLS)优选最佳遗忘因子,最后,在动态工况下,对比了常用定遗忘因子最小二乘和PSO-FFRLS的在线辨识端电压误差,并分别与扩展卡尔曼滤波(EKF)算法联合,对比2种方法的估算效果。结果表明,PSO-FFRLS的端电压在线辨识结果能更好地跟随实测电压且误差极小,其与EKF的联合算法对SOC的估算精度也更高。In order to improve the accuracy of estimating the State of Charge(SOC)of the power battery of electric vehicles in driving,this paper takes lithium iron phosphate battery as an example,and proposes an online identification model battery SOC estimation method based on Particle Swarm Optimization Forgetting Factor Recursive Least Squares(PSO-FFRLS).This method first establishes a second-order equivalent circuit model of the battery,then uses the Particle Swarm Optimization(PSO)algorithm to optimize the forgetting factor in real time for the Forgetting Factor Recursive Least Square(FFRLS)method.Finally,under dynamic conditions,the commonly used fixed forgetting factor least square and PSO-FFRLS on-line identification terminal voltage errors are compared,and they are respectively combined with the Extended Kalman Filter(EKF)algorithm to compare the battery SOC estimation effects of the two methods.The results show that the online identification terminal voltage of PSO-FFRLS can better follow the measured voltage with minimal error,and the joint algorithm of PSO-FFRLS and EKF can estimate the SOC with higher accuracy.

关 键 词：磷酸铁锂电池 粒子群优化 动态优选遗忘因子 在线辨识 扩展卡尔曼滤波 

分 类 号：TM912.2[电气工程—电力电子与电力传动]