小样本条件下动力电池容量和内阻协同估计  

Data Driven Estimation of Capacity and Internal Resistance of Lithium-Ion Batteries

作　　者：朱建功[1] 张仁杰 姜波 戴海峰[1] 魏学哲[1] ZHU Jiangong;ZHANG Renjie;JIANG Bo;DAI Haifeng;WEI Xuezhe(School of Automotive Studies,Tongji University,Shanghai 201800,China)

机构地区：[1]同济大学汽车学院,上海201800

出　　处：《西南大学学报（自然科学版）》2024年第12期14-23,共10页Journal of Southwest University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(52107230,52377211)。

摘　　要：锂离子电池的容量和内阻分别是表征电池系统能量和功率特性的重要参数,准确估计容量和内阻对于电动汽车电池管理系统有至关重要的作用.提出了一种数据驱动的电池容量和内阻估计方法.该方法首先从电池的恒流充电片段提取容量增量(incremental capacity,IC)特征,经过重采样方法构建特征向量.然后将IC特征输入弹性网络模型进行训练,从而实现容量和内阻估计.应用KIT数据集对所提方法进行验证.结果表明:该模型可以在相同特征输入下实现高精度的容量和内阻协同估计,且在小样本训练(仅使用2个电池作为训练集)时保持容量估计误差在2%以内,内阻的估计误差在3%以内.最后提出了基于容量和内阻双参量的“三段式”电池评价方法.Capacity and internal resistance of lithium-ion batteries are important parameters that characterize the energy and power characteristics of battery system.Accurate estimation of capacity and internal resistance is crucial for electric vehicle battery management systems.Therefore,this article proposes a data-driven method for estimating battery capacity and internal resistance.This method first extracts incremental capacity(IC)features from the constant current charging segment of the battery,constructs feature vectors through resampling,and then inputs IC features into Elastic Net model for training,thereby achieving capacity and internal resistance estimation.The proposed method is validated using the KIT dataset.The results show that the model proposed in this paper can achieve high-precision capacity and internal resistance estimation under the same feature input,and maintain capacity estimation error within 2%and internal resistance estimation error within 3%during small sample training(using only 2 batteries as the training set).Finally,a“three-stage”battery evaluation method based on dual parameters of capacity and internal resistance was proposed.

关 键 词：锂离子电池 数据驱动 容量和内阻估计 小样本训练 电池评价方法 

分 类 号：TM912[电气工程—电力电子与电力传动]