面向多步快速充电场景的锂离子电池健康状况估计

Health State Estimation of Lithium-Ion Batteries for Multi-Step Fast Charging Scenarios

针对快速充电对锂离子电池健康状态评估所提出的挑战,构建了适用于多步快速充电场景的锂离子电池健康状态估计模型。先从锂离子电池快速充电电压曲线中提取了12个直接特征,并分析了与这些特征强相关的衰退机制。随后,基于衰退机理并通过相关性分析进行特征筛选,使用径向基函数神经网络建立了健康状态估计模型。验证结果表明,所构建数据特征在电池间衰退路径存在显著差异情况下依然具有良好的泛化能力,相比于传统特征筛选方法估计精度提升超过17%。即使在不同的快速充电协议以及较小训练数据规模下,仍取得了令人满意的估计结果。

In response to the challenges posed by the widespread adoption of fast charging in lithium-ion battery health assessment,this study develops a state-of-health estimation model for dynamic fast-charging scenarios.Twelve direct features are extracted from the partial voltage curve during the fast charging process,followed by a comprehensive analysis of degradation mechanisms strongly correlated with these features.Subsequently,feature selection is conducted based on degradation mechanisms and correlation analysis,and the radial basis function neural network(RBFNN)is deployed to establish the estimation model.The validation results indicate that the constructed data features exhibit excellent generalization across various battery degradation paths,improving accuracy by over 17%compared to traditional feature selection methods.Satisfactory estimation results are obtained even under different fast charging protocols and with a smaller training dataset.