摘要
提前对锂离子电池的剩余使用寿命(RUL)进行预测可以保证电池及其应用设备安全稳定运行。针对目前预测方法的结果滞后且缺乏实际意义等问题,为了利用少量循环数据实现RUL的提前预测,本文基于原始极限学习机和自动编码器构建了分层极限学习机(H-ELM)预测模型。然后选取丰田研究所(TRI)的实验数据集对H-ELM完成了仿真实验验证。实验结果表明,本文提出的H-ELM预测模型可以在电池使用初期预测出RUL,同时预测结果的平均绝对百分比误差(MAPE)仅有10.14%。
Predicting the remaining useful life(RUL)of lithium-ion batteries in advance can ensure the safe and stable operation of the battery and its application equipment.Aiming at the problem that the results of current prediction methods are lagging behind and lack of practical significance,in order to use a small amount of cyclic data to realize the advance prediction of RUL,a hierarchical extreme learning machine(H-ELM)prediction model based on the original extreme learning machine and the autoencoder is constructed.Then,the experimental data set of Toyota Research Institute(TRI)is selected to complete the simulation experiment verification of H-ELM.The experimental results show that the H-ELM prediction model can predict RUL in the early stage of battery use,and the mean absolute percentage error(MAPE)of the prediction results is only 10.14%.
作者
史永胜
洪元涛
丁恩松
施梦琢
欧阳
SHI Yongsheng;HONG Yuantao;DING Ensong;SHI Mengzhuo;OU Yang(School of Electrical and Control Engineering,Shaanxi University of Science and Technology,Xi’an Shaanxi 710021,China;Jiangsu Runyin Graphene Technology Co.,Ltd.,Yangzhou Jiangsu 225600,China)
出处
《电子器件》
CAS
北大核心
2021年第3期652-658,共7页
Chinese Journal of Electron Devices
基金
国家自然科学基金项目(61871259)。
