摘要
针对锂电池直接预测剩余使用寿命难及预测结果不准确等问题,提出利用锂电池循环充放电监测参数构建间接寿命特征参数的方法。应用一阶偏相关系数分析法验证间接寿命特征参数与直接参数间的相关性,选择等压降放电时间作为锂电池间接寿命特征参数,构建基于ELM的等压降放电时间与实际容量的关系模型和等压降放电时间预测模型,实现锂电池的RUL预测。基于NASA锂电池数据集预测并评估锂电池的RUL,并且与ELM直接预测方法和高斯过程回归间接预测方法相比较,本方法能够有效的预测锂电池的RUL,预测结果的误差范围为5%左右,具备较好的锂电池RUL预测精度。
Since the direct prediction of remaining useful life( RUL) of lithium-ion battery is difficult and inaccurate,monitoring parameters based on lithium-ion battery charge-discharge cycle is adopted to construct the method of indirect characteristic parameters of life,and the first-order partial correlation analysis is used to verify the correlation between indirect characteristic parameters of life and the direct parameters. The time interval to equal discharging voltage is chosen as the indirect characteristic parameters of life. Relation model of ELM based on time interval to equal discharging voltage and remaining capacity is proposed,and the prediction model of ELM based on the time interval to equal discharging voltage is constructed,which predicts RUL of lithium-ion battery. The RUL of lithium-ion battery is predicted and assessed based on lithium-ion battery data sets of NASA. Compared with the direct ELM prediction method and Gaussian process regression( GPR) prediction method,this method can effectively predict the RUL of lithium-ion battery with an error range about 5%,and realize RUL prediction accuracy of lithium-ion battery with satisfactory results.
出处
《电子测量与仪器学报》
CSCD
北大核心
2016年第2期179-185,共7页
Journal of Electronic Measurement and Instrumentation
基金
国家自然科学基金(61401215)
安徽省高校优秀青年人才支持计划重点项目(gxyq ZD2016082)项目资助
关键词
间接预测
剩余寿命特征参数
极限学习机(ELM)
锂电池
indirect prediction
characteristic parameters of remaining life
extreme learning machine(ELM)
lithium battery