摘要
针对锂离子电池使用过程中,随着循环充放电后健康状态不断变化的特点,以及当前研制阶段和使用阶段数据未得到充分融合的现状,分析反映锂离子电池健康状态的特征参数,参照IEEE 1188—2005《推荐用于站用阀控铅酸电池的维护、测试和更换方法》等标准划定锂离子电池健康等级,结合堆叠自编码深度学习和迁移学习算法进行建模,提出了利用试验数据进行模型训练和运行数据的参数进行模型微调的理念,并利用试验数据进行了模型验证。经计算表明,该模型对锂离子电池状态的评估精度达到97.1%,可为其健康状态确认提供参考和借鉴。
The capacity of a lithium battery gradually declines during use,as the recharging cycle changes the health status of the battery.And the data generated in producing the batteries and in consuming the batteries have not been fully integrated.Our research aims to address these problems.In this paper,the character parameters of Lithium batteries are analyzed,the health grade is classified according to IEEE 1188—2005 Recommended practice for maintenance,testing,and replacement of valve-regulated lead-acid(VRLA)batteries for stationary application,and a model combining stack auto-encoders and transfer learning is proposed.Furthermore,the model verification is accomplished using testing data,which reveals that the accuracy of the proposed model reaches 97.1%,serving as an effective indicator for monitoring Lithium batteries'health status.
作者
龙东腾
王庆杰
刘家赫
李昆儒
Long Dongteng;Wang Qingjie;Liu Jiahe;Li Kunru(China Astronautics Standards Institute,Beijing 100071,China;Guizhou Meiling Power supply Co.Ltd,Zunyi Guizhou 563000,China)
出处
《质量与可靠性》
2025年第1期60-65,共6页
Quality and Reliability
关键词
锂离子电池
健康评估
迁移学习
lithium battery
health evaluation
transfer learning
