重庆理工大学学报2025，Vol.39Issue(7)：9-16,8.DOI:10.3969/j.issn.1674-8425(z).2025.04.002
GWO-CNN框架下的锂电池健康状态估计方法
Lithium battery health status estimation method under GWO-CNN framework
摘要
Abstract
The state of health(SOH)of lithium-ion batteries is a key indicator for evaluating the electric vehicles' performance and safety.To accurately and efficiently estimate State of Health(SOH)in real-time,this paper proposes a SOH estimation method based on constant voltage charging segments and Grey Wolf Optimizer(GWO)optimized Convolutional Neural Network(CNN).The method simplifies the feature engineering steps in traditional methods by extracting partial current and time series data during the constant voltage charging process and automatically extracting features using CNN.It overcomes the dependence on complete charging data and significantly reduces the cost of data acquisition and storage.The introduction of GWO to optimize the hyperparameters of CNN improves the estimation accuracy of the model.Results on NASA and CALCE battery datasets show the method achieves high accuracy in SOH estimation.关键词
锂离子电池/健康状态估计/恒压充电/卷积神经网络/灰狼算法Key words
lithium-ion batteries/health status estimation/constant voltage charging/CNN/GWO分类
动力与电气工程引用本文复制引用
郭鹏旭,赵理,马占潮,王震,李玉琦..GWO-CNN框架下的锂电池健康状态估计方法[J].重庆理工大学学报,2025,39(7):9-16,8.基金项目
国家自然科学基金项目"基于频繁模式挖掘和概念漂移的新型动力电池检测机制研究"(52077007) （52077007）
