基于多退化特征的锂离子电池剩余寿命预测

Remaining useful life prediction of lithium ion battery based on multiple degradation characteristics

针对目前锂离子电池剩余寿命预测模型精度低、泛化性差的问题,在一种基于充放电健康特征提取的锂离子电池剩余寿命估计方法的基础上,增加了健康因子和实际容量之间的相关性分析,具体方法是:从锂离子电池充放电电压、电流、温度曲线变化趋势中提取若干潜在健康因子,并利用主成分分析(PCA)去除数据冗余性,得到代表退化特征的融合健康因子。结合自适应遗传算法(AGA)优化了Elman预测模型。结果表明所建立的PCA-AGA-Elman神经网络预测模型误差控制在1.5%之内,可作为锂离子电池的剩余使用寿命(RUL)预测模型。

Aiming at the low accuracy and poor generalization of the current remaining useful life prediction model for lithium ion battery,a remaining useful life estimation method based on charge-discharge health characteristics was adopted,and the correlation analysis between health factors and actual capacity was added.The specific method was to extract several potential health factors from the changing trend of charging and discharging voltage,current and temperature curve of lithium ion battery.The data redundancy was removed by principal component analysis(PCA)to obtain the fusion health factors representing the degradation characteristics.The Elman prediction model was optimized with adaptive genetic algorithm(AGA).The results show that the error of the PCA-AGA-Elman neural network prediction model can be controlled within 1.5%,which can be used as the prediction model of the remaining useful life of lithium ion batteries.