现代电子技术2025，Vol.48Issue(16)：31-37,7.DOI:10.16652/j.issn.1004-373x.2025.16.006
基于MVMD-互相关时频域特征的风电变流器开路故障诊断
Wind power converter open circuit fault diagnosis based on MVMDcross-correlation time-frequency domain feature
摘要
Abstract
In allusion to the problems of single-open circuit and double-open circuit faults of power tubes in the grid-side converter of permanent magnet synchronous wind generator,a diagnosis method combining multivariate variational mode decomposition(MVMD)and cross-correlation time-frequency domain features is proposed.The three-phase midpoint voltage of the grid-side converter is collected as the original data.The three-phase midpoint voltage is taken as the three-channel input data of MVMD and is decomposed,to obtain several layers of intrinsic mode functions(IMF).By calculating the mutual correlation between the original signal data and the IMF function of each layer,the IMF that can best reflect the features and trends of the original signal is selected,and then its average envelope and the original data are extracted in time domain and frequency domain.The obtained features are input into the bidirectional long short term memory(BiLSTM)network for the training,realizing the location of the fault power tube.The experimental results show that the proposed method can effectively identify and classify 22 states of the grid-side converter,and the accuracy of testing set can reach 99.55%.关键词
风电变流器/MVMD/时频域特征/开路故障诊断/IMF分量/BiLSTM/IGBTKey words
wind turbine converter/MVMD/time-frequency domain feature/open circuit fault diagnosis/IMF component/BiLSTM/IGBT分类
信息技术与安全科学引用本文复制引用
侯锐,李然..基于MVMD-互相关时频域特征的风电变流器开路故障诊断[J].现代电子技术,2025,48(16):31-37,7.基金项目
华北理工大学研究生创新项目(2024S05) （2024S05）
