通信与信息技术Issue(2)：98-103,6.
基于GAF与Swin Transformer的逆变器故障诊断
Inverter fault diagnosis based on GAF and Swin Transforme
摘要
Abstract
Inverter failure will reduce the power generation efficiency and stability of the power station.For inverter fault diagnosis,this paper takes the grid-connected inverter with LCL-type filter as the research object,introduces the structure and mathematical model of the inverter in detail,and summarizes and organizes the common types of faults of PV inverters and their characteristics;builds a model of the PV grid-connected inverter using the simulation platform,and carries out an investigation on the inverter switching tubes(IGBT)open-circuit faults,which is the most likely to occur among the inverter faults.IGBT)open-circuit faults in the inverter are simulated to collect one-dimensional fault current signals;the one-dimensional time series fault data are transformed using the Gram angle field(GAF)to obtain two-dimensional images.Based on the advantages of convolutional neural network in image processing,a fault diagnosis model based on Gram angle field and Swin Transformer is proposed.This fault diagnosis model is tested on samples with different noise ratios after iterative training to optimize the model parameters.The results showed that the accuracy of fault diagnosis reached 96.81%,which proves the effectiveness and noise resistance of the fault diagnosis model.关键词
并网逆变器/格拉姆角场/Swim TransformerKey words
Grid-connected inverter/GAF/Swim Transformer分类
信息技术与安全科学引用本文复制引用
杨龙欢,董效杰,王一帆,刘承毅,周运磊..基于GAF与Swin Transformer的逆变器故障诊断[J].通信与信息技术,2025,(2):98-103,6.基金项目
湖北省教育厅科学技术研究计划项目(项目编号:B2016092) （项目编号:B2016092）
