重庆理工大学学报2025，Vol.39Issue(9)：185-194,10.DOI:10.3969/j.issn.1674-8425(z).2025.05.023
利用VMD-SWT与ISSA-BiLSTM的充电桩故障诊断方法
Fault diagnosis method for the charging pile using VMD-SWT and ISSA-BiLSTM
摘要
Abstract
To address the low accuracy in fault diagnosis of the DC charging pile for electric vehicles,a charging pile fault diagnosis model combining variational mode decomposition with stationary wavelet transform(VMD-SWT)and improved sparrow search algorithm optimized bidirectional long short-term memory network(ISSA-BiLSTM)is proposed.First,the VMD-SWT method is employed to extract features from the nonlinear fault signals of the charging pile,obtaining time-frequency domain features with better separability.Then,the time-frequency domain characteristics of the charging pile are normalized.Next,Iterative-Tent chaotic mapping,adaptive dynamic inertia weights,and Laplacian operator are introduced to improve SSA.ISSA is employed to optimize the hyperparameters of the BiLSTM model and obtain the optimal model.Finally,the fault diagnosis results of the proposed method is compared with other methods to verify its effectiveness.Results show the proposed method achieves an accuracy of 98.34%in fault diagnosis,outperforming traditional fault diagnosis methods.关键词
充电桩/故障诊断/变分模态分解/双向长短时记忆网络/麻雀搜索算法Key words
charging pile/fault diagnosis/variational mode decomposition/BiLSTM/SSA分类
计算机与自动化引用本文复制引用
冯莉,朱宇环,杨睿,李龙飞..利用VMD-SWT与ISSA-BiLSTM的充电桩故障诊断方法[J].重庆理工大学学报,2025,39(9):185-194,10.基金项目
重庆市自然科学基金面上项目(CSTB2022NSCQ-MSX1456) （CSTB2022NSCQ-MSX1456）
重庆市教育委员会科学技术研究项目(KJZD-K202300705) （KJZD-K202300705）
重庆交通大学研究生课程思政示范项目(KCSZ2023011) （KCSZ2023011）
重庆交通大学课程思政示范课程培育项目(19210096) （19210096）
重庆市高等教育教学改革研究项目(233239) （233239）
重庆交通大学研究生科研创新项目(2023S0054) （2023S0054）
