摘要
高效的电池筛选技术是保障储能电池系统安全性、经济性的关键技术之一。为了满足液态金属电池储能系统构建对单体一致性的要求,提出一种组合聚类方法,通过电池恒流放电期间获得的分选指标进行有效聚类,快速实现液态金属电池筛选。首先,提出一种特征提取框架来准确表征电池放电曲线,主要包括数据采集、数据预处理、特征参数生成、筛选指标生成4个阶段,生成的3个筛选指标,分别是电池曲线第2个拐点电压、对应的时间以及电池在第1个拐点电压前的放电能量。然后,提出基于密度的噪声应用空间聚类和均值漂移的组合聚类电池筛选方法,并提出一种一致性差异指标量化评估聚类效果。最后,通过分析212个200 Ah级液态金属电池的聚类优化效果可知,所提方法可显著减弱差异性较大电池对电池筛选的影响,能够同时实现离群值检测及电池自适应快速精准筛选,并通过动态工况测试证实所提方法可有效提升电池组响应一致性。
Efficient battery screening technology is crucial for ensuring safety and economy in battery system integration applications.To achieve consistent results for integrated liquid metal batteries(LMBs),this paper proposes a combined clustering method to effectively cluster the feature indicators obtained during the constant current discharge of the battery to quickly achieve liquid metal battery screening.First,a feature extraction framework is proposed to accurately characterize the battery discharge curve.It consists of four main stages:data acquisition,data pre-processing,feature indicator generation,and screening indicator generation.Three screening indicators are generated,i.e.the second change point voltage of the battery curve,the corresponding time,and the discharge energy of the battery before the first change point voltage.Then,a combined clustering method based on density-based noise applied spatial clustering and mean shift is proposed,and a consistency difference indicator is proposed to quantitatively evaluate the clustering effect.Finally,through the analysis of the clustering optimization effect of 212200 Ah-level liquid metal batteries,it is shown that the proposed method significantly reduces the influence of cells with large variability on cell screening,and it can simultaneously achieve outlier detection and fast and accurate cell adaptive screening.The dynamic profile test verifies that the battery screening method in this paper can effectively improve the response consistency of the battery pack.
作者
张娥
樊磊
徐成
王晟
李浩秒
蒋凯
李波
王康丽
ZHANG E;FAN Lei;XU Cheng;WANG Sheng;LI Haomiao;JIANG Kai;LI Bo;WANG Kangli(State Key Laboratory of Advanced Electromagnetic Engineering and Technology(School of Electrical and Electronic Engineering,Huazhong University of Science and Technology),Wuhan 430074,Hubei Province,China;Institute of Electric Power Science of Guizhou Power Grid Co.,Ltd.,Guiyang 550000,Guizhou Province,China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2024年第20期8073-8085,I0013,共14页
PROCEEDINGS OF THE CHINESE SOCIETY FOR ELECTRICAL ENGINEERING
基金
国家自然科学基金项目(52277217,52177215)
中国南方电网有限公司项目(GZKJXM20222324)。
关键词
电池筛选
液态金属电池
特征提取
组合聚类方法
离群值处理
battery screening
liquid metal batteries(LMBs)
feature extraction
combined clustering method
outlier point processing
