电源技术2025，Vol.49Issue(5)：991-1005,15.DOI:10.3969/j.issn.1002-087X.2025.05.015
基于CEEMDAN-IGWO-CNN-BiLSTM模型的锂电池剩余寿命预测
Remaining life prediction of lithium battery based on CEEMDAN-IGWO-CNN-BiLSTM models
摘要
Abstract
To address the challenge of limited or absent large-scale battery aging data,a hybrid pre-diction model is proposed that integrates Complete Ensemble Empirical Mode Decomposition with Adaptive Noise-improved Gray Wolf Optimization-convolutional neural network-bidirectional long short-term memory network(CEEMDAN-IGWO-CNN-BiLSTM).The traditional GWO often suf-fers from local optimality and exhibits slow convergence speeds.To enhance the GWO,we incorpo-rate the Tent chaos mapping,dimension learning-based hunting search strategy,and the Taguchi method,resulting in multi-strategy improvements.Initially,the battery capacity data are decomposed into intrinsic mode functions and residual using the CEEMDAN.Subsequently,a CNN is utilized to extract feature data,which is then fed into a BiLSTM.The optimal parameters for the BiLSTM are determined through the IGWO for accurate prediction.Finally,we validate the model using public datasets and compare its performance with that of other models.The results demonstrate a reduction in the root mean square error and the mean absolute error by 17%and 30%,respectively,with the coefficient of determination increasing by 4%.These findings indicate that the proposed model exhib-its strong robustness and generalization capabilities.关键词
锂离子电池/IGWO/CEEMDAN/BiLSTM/剩余使用寿命预测Key words
lithium ion battery/IGWO/CEEMDAN/BiLSTM/remaining useful life prediction分类
信息技术与安全科学引用本文复制引用
王旭,胡明茂,宫爱红,龚青山,黄正寅,姜宇,李帅雨,姚政豪,陈锐..基于CEEMDAN-IGWO-CNN-BiLSTM模型的锂电池剩余寿命预测[J].电源技术,2025,49(5):991-1005,15.基金项目
国家自然科学基金项目(523755084) （523755084）
湖北省教育厅重点项目(D20211803) （D20211803）
湖北汽车工业学院博士基金(BK202001) （BK202001）
汽车动力传动与电子控制湖北省重点实验室(湖北汽车工业学院)项目(ZDK12023B02) （湖北汽车工业学院）
