基于实车数据的电动汽车电池健康状态估计  

SOH estimation of electric vehicle battery based on real segment charging data

作　　者：龚贤武[1] 丁璐 穆邱倩 李萌 马宇骋 GONG Xianwu;DING Lu;MU Qiuqian;LI Meng;MAYucheng(School of Electronics and Control Engineering,Chang’an University,Xi’an Shaanxi 710064,China)

机构地区：[1]长安大学电子与控制工程学院,陕西西安710064

出　　处：《电源技术》2021年第12期1577-1580,共4页Chinese Journal of Power Sources

基　　金：国家重点研发项目(2019YFB1600800)。

摘　　要：针对实际工况下电动汽车电池充放电电流不稳定,难以获取满充满放数据带来的电池健康状态估计困难的问题,研究以容量作为健康状态评价指标,以安时积分逆过程计算的深度充电片段电池容量及实车运行特征数据为数据样本,建立特征数据和容量之间的机器学习模型,得到剩余充电片段容量的估计,提出了一种结合安时积分法和机器学习模型对全时间段电池健康状态进行估计的方法。模型评价结果表明,该方法合理有效,对实际工况下电动汽车电池健康状态的实时估计有重要意义。In order to solve the problem of state of health(SOH)estimation of electric vehicle batteries that the charging and discharging current data is unstable and the full time current data is difficult to obtain under actual operating conditions,this paper studied the evaluation index of SOH by capacity,and established the machine learning model between feature data of actual operating condition and capacity by a reverse process of ampere-hour integration in deep segment charging data.A method combined with ampere-hour integration and machine learning was completed to estimate the SOH in the whole time period.The model evaluation results show that this method is reasonable and effective and is great significance to estimate the health of electric vehicle battery under actual working conditions.

关 键 词：实际工况 健康状态估计 安时积分 机器学习 

分 类 号：TM912[电气工程—电力电子与电力传动]