电力系统及其自动化学报2025，Vol.37Issue(5)：131-140,10.DOI:10.19635/j.cnki.csu-epsa.001599
基于特征综合评价和模型优化的锂离子电池健康状态估计方法
State-of-health Estimation Method for Lithium-ion Battery Based on Comprehensive Feature Evaluation and Model Optimization
摘要
Abstract
Aimed at problems such as the single performance of a feature evaluation index,insufficient feature captur-ing capability of prediction models and difficulty in determining the hyperparameters,a state-of-health(SOH)estima-tion method for lithium-ion battery based on comprehensive feature evaluation and model optimization is proposed in this paper.First,the comprehensive evaluation indexes for features are constructed from the perspectives of principle and statistics,and the features with higher index scores are selected as the model input.Second,by combining the con-volutional neural networks(CNN),efficient local attention(ELA)and bi-directional gated recurrent unit(BiGRU),a CNN-ELA-BiGRU prediction model is established,which enhances the model's feature capturing capability.Finally,the golden jackal optimization algorithm is used for hyperparameter optimization of the model,thus improving its predic-tion accuracy.Results of comparative experiments show that the proposed SOH estimation method has good stability and robustness.关键词
锂离子电池/特征综合评价指标/高效局部注意力/金豺优化算法/健康状态估计Key words
lithium-ion battery/comprehensive feature evaluation index/efficient local attention(ELA)/golden jack-al optimization algorithm/state-of-health(SOH)estimation分类
信息技术与安全科学引用本文复制引用
黄凯,郝润凯,郭永芳..基于特征综合评价和模型优化的锂离子电池健康状态估计方法[J].电力系统及其自动化学报,2025,37(5):131-140,10.基金项目
中央引导地方科技发展资金项目(236Z4408G) （236Z4408G）
天津市自然科学基金资助项目(23JCYBJC00810). （23JCYBJC00810）
