基于CARIMA模型的镍钴锰酸锂三元材料动力电池荷电状态估计  

The state of charge estimation of NCM lithium-ion power batteries based on CARIMA model

作　　者：段云鹏 吴刚[1] 石春[1] Duan Yunpeng;Wu Gang;Shi Chun(Institute of Industrial Automation,Department of Automation,University of Science and Technology of China,Hefei 230027,China)

机构地区：[1]中国科学技术大学自动化系工业自动化研究所

出　　处：《电子测量技术》2018年第16期23-27,共5页Electronic Measurement Technology

摘　　要：动力电池的荷电状态估计是电动汽车的关键技术之一。提出基于受控自回归积分滑动平均（CARIMA）模型的锂离子动力电池荷电状态（SOC）估计方法。动力电池的电化学特性复杂,车辆上的电磁干扰严重,车辆运行工况大范围、非线性变化,因此假设电池受到的扰动是增量平稳过程,建立电池的CARIMA模型,给出了电池荷电状态最优估计。电池模型中荷电状态和开路电压（OCV）呈非线性关系,采用非线性多元逐步回归法分别建立SOC-OCV,OCVSOC的多项式模型。实验结果表明,本文提出的方法可以实现SOC的精确估计,SOC估计的绝对误差在±3%以内。The state of charge estimation of power batteries is one of the key technology for electric vehicles. This paper presents a method of the state of charge （SOC） estimation of lithium ion batteries based on CARIMA model. The electrochemical properties of power batteries are complex, the electromagnetic interference on the vehicle is serious, vehicle operating conditions change widely and nonlinearly, so the disturbance of the battery is assumed to be an incremental stationary process and CARIMA model of batteries is built, the optimal estimation of the state of charge is presented. The state of charge and open circuit voltage （OCV） are non-linear correlated. Non-linear multiple stepwise regression is used to build SOC-OCV polynomial model and OCV-SOC polynomial model. In the validation, the estimation results are found to be agreed with the reference SOC within ~ 3 % error bound.

关 键 词：锂离子动力电池 荷电状态 CARIMA 非线性多元逐步回归 

分 类 号：TP2[自动化与计算机技术—检测技术与自动化装置] TN98[自动化与计算机技术—控制科学与工程]