基于充电数据的多阶段锂离子电池健康状态估计  

Multi-Stage State of Health Estimation Based on Charging Phase for Lithium-Ion Battery

作　　者：魏中宝 阮浩凯 何洪文[1] WEI Zhongbao;RUAN Haokai;HE Hongwen(School of Mechanic Engineering,Beijing Institute of Technology,Beijing 100081,China)

机构地区：[1]北京理工大学机械与车辆学院,北京100081

出　　处：《北京理工大学学报》2022年第11期1184-1190,共7页Transactions of Beijing Institute of Technology

基　　金：国家自然科学基金资助项目(52072038)。

摘　　要：锂离子电池的健康状态估计是锂离子电池寿命评估和健康管理的基础.文中针对实际应用场景中充电数据的缺失,提出一种实用的多阶段电池的健康状态估计方法.研究中根据电压大小,将充电过程划分为3个阶段,分别提出了具有针对性的电池的健康状态估计方法.特别是对于恒流电压过渡阶段,在恒流数据和电压数据都严重缺失地情况下,利用卷积神经网络的数据挖掘能力,直接建立了电压电流数据与电池的健康状态的关系,在锂离子电池的长期老化实验数据研究基础上对所提出的方法进行了验证.结果表明,该方法具有估计精度高、应对严重数据缺失的能力强、对电池不一致性鲁棒性强等优点.State of health estimation of lithium-ion battery is the basis of lithium-ion battery life assessment and health management.A practical multi-stage state of health estimation method was proposed to deal with different charging stages,including the scene of serious lack of charging data.According to the voltage,the constant current-constant voltage charging process was divided into three stages and their target state of health estimation methods were proposed respectively.Especially for the constant current-constant voltage transition stage,being a lack of constant current data and constant voltage data heavily,the relationship between raw voltage/current data and battery state of health was directly established taking the strong data mining capability of convolutional neural network.The proposed method was evaluated by long-term aging experiments on lithium-ion battery.The results show that this method possesses the advantages of high estimation accuracy,strong ability to deal with serious data loss,and strong robustness to battery inconsistency.

关 键 词：锂离子电池 健康因子 健康状态估计 机器学习算法 恒流 卷积神经网络 

分 类 号：U469.72[机械工程—车辆工程]