基于充放电特征和表面温度变化的锂离子电池剩余寿命预测  

Remaining Life Prediction of Lithium-Ion Batteries Based on Charge and Discharge Characteristics and Surface Temperature Variation

作　　者：吴铁洲[1] 李宜安 彭慧刚 WU Tiezhou;LI Yian;PENG Huigang(Hubei Key Laboratory for High-Efficiency Utilization of Solar Energy and Operation Control of Energy Storage,Hubei University of Technology,Wuhan Hubei 430068,China;Wuhan Haichuang Electronic Co.,Ltd.,Wuhan Hubei 430067,China)

机构地区：[1]湖北工业大学太阳能高效利用及储能运行控制湖北省重点实验室,湖北武汉430068 [2]武汉海创电子股份有限公司,湖北武汉430067

出　　处：《电子器件》2025年第1期123-129,共7页Chinese Journal of Electron Devices

基　　金：中国湖北省重大科技创新项目(2018AAA056);国家自然科学基金项目(51677058)。

摘　　要：锂电池剩余使用寿命(RUL)的准确预测对电池的安全稳定运行至关重要。为了较为准确估算电池的RUL,提出了一种采用电池充放电特征和表面温度变化为健康因子,结合改进向量机算法的RUL预测方法。在充放电曲线中,拟合曲线斜率作为表征电池性能退化的健康因子(HI);在电池充放电表面温度变化曲线中,提取温度随时间的变化率作为表征电池性能退化的健康因子(HI)。利用Pearson和Spearman相关性分析法证实所提出的HI与电池容量有关。然后将HI和最小二乘支持向量机(LS-SVM)结合,实现电池RUL预测。最后利用NASA电池数据集对LS-SVM方法进行验证,对比反向传播(BP)和支持向量回归(SVR)方法,实验结果表明,所提出的RUL预测结果的RMSE和MAPE均小于2%,预测结果更加准确。The accurate prediction of the remaining useful life(RUL)of Li-ion batteries is crucial to the safe and stable operation of bat-teries.In order to estimate the RUL of batteries more accurately,an RUL prediction method using battery charge and discharge charac-teristics and surface temperature variation as health factors is proposed,which is combined with an improved vector machine algorithm.In the charge-discharge curve,the slope of the fitted curve is used as the health factor(HI)to characterize the degradation of battery per-formance and the charge-discharge surface temperature variation curve of the battery,the rate of change of temperature with time is ex-tracted as the health factor(HI)to characterize the degradation of battery performance.The proposed HI is confirmed to be related to the battery capacity by using Pearson and Spearman correlation analysis.Then the HI and least squares support vector machine(LS-SVM)are combined to achieve battery RUL prediction.Finally,the LS-SVM method is validated on the NASA battery dataset,and compared with the back propagation(BP)and support vector regression(SVR)methods,the experimental results show that the RMSE and MAPE of the RUL prediction results proposed are less than 2%,and the prediction results are more accurate.

关 键 词：锂离子电池 表面温度 健康因子 剩余寿命 支持向量机 

分 类 号：TM912[电气工程—电力电子与电力传动]