基于V2G与XGBoost技术的锂电池SOH估计与RUL预测分析  

Analysis of SOH Estimation and RUL Prediction for Lithium Batteries Based on V2G and XGBoost Technology

作　　者：张文龙 高昕[1] ZHANG Wenlong;GAO Xin(School of Electrical and Information Engineering,Anhui University of Technology,Anhui 232001,China)

机构地区：[1]安徽理工大学电气与信息工程学院,安徽232001

出　　处：《集成电路应用》2024年第3期18-21,共4页Application of IC

摘　　要：阐述针对锂离子电池特征提取困难导致SOH预测精确度较低的问题,提出一种基于V2G技术的充放电过程与XGBoost的锂电池SOH预测及RUL预测方法。通过从充电曲线中提取的滤波后曲线峰值与峰值对应点电压作为充电过程中的健康因子,以及放电过程中的周期放电电压达到最低点的时间和平均电压衰减,作为放电过程中的健康因子XGBoost模型的输入,进行电池SOH预测,结合SOH估计值和估算模型实现RUL的长期预测。实验结果表明,改进后的模型具有更高的估算精度,SOH估计和RUI预测精度较高。This paper describes the problem of low accuracy in SOH prediction caused by the difficulty in feature extraction of lithium-ion batteries.A charging and discharging process based on V2G technology and XGBoost based method for SOH prediction and RUL prediction of lithium-ion batteries are proposed.By extracting the filtered peak value and corresponding point voltage from the charging curve as the health factor during the charging process,as well as the time and average voltage attenuation of the periodic discharge voltage reaching its lowest point during the discharge process,as input for the XGBoost model of the health factor during the discharge process,battery SOH prediction is carried out.Combined with SOH estimation values and estimation models,long-term prediction of RUL is achieved.The experimental results show that the improved model has higher estimation accuracy,as well as higher accuracy in SOH estimation and RUI prediction.

关 键 词：V2G 锂电池SOH RUL预测 

分 类 号：TM912[电气工程—电力电子与电力传动]