基于多尺度残差动态域适应网络的不同工况下转子故障诊断方法

打开文本图片集

关键词：故障诊断；转子；迁移学习；残差网络；动态域适应 中图分类号：TH133；TH165.5 文献标志码：A DOI：10.16385/j.cnki.issn.1004-4523.202312029

Rotor fault diagnosis method based on multi-scale residual dynamic domain adaptive network under different working conditions

XIANG Ling1²，WANG Ning1，BING Hankun1， HU Aijun1²， HAN Zhongquan1 （1.Department ofMechanical Engineering，North China Electric Power University，Baoding O710o3，China; 2.HebeiKeyLaboratoryofElectric Machinery Health Maintenance&Failure Prevention，Baoding O710o3，China）

Abstract：Thedistributionofotordatacoletedunderdiferentworkingconditonsisverydifeent，whichleadstothelowacuracyof traditionalfaultdagosisodelTispperpropossaotorfaultdagosismethdbsedonmulticalesidualdaicdomadaptatio network （MsRDDA）underdiferent workingconditions.Itisusedtosolvethe problem thatthesourcedomaisampleshavelabels but te targetdomainsampleshavenolabels，andrealizeunsupervisedmigrationdiagnosisbetweendierentworkingconditions.TheonedimensionaltiomaisialcoledbeotrtchiseeteyotFurierasfo（Sndoertedtot dimensionalimagewithie-frequecydomainharacteristis.Amulti-alesiduatwokombiingulti-aleonvoutioandpaable convolutionisproposd.Temultisaleonvolutiolerisusedasteinputlyertoetractsalowfeatures，andfourimprovdresidal modulesareusedtoextractdpfeaturestoexpandtenetworksensitivityfield，soatoesurethediversityoffultfeaturesadavodte gradientdisappearancecausedbytheincreaseofnetworkdepthTedynamicdistributiondomainadaptationstrategyisintroducedintothe multi-scaleuaokeporaeofdistrtiododiioaldstrbtioisdamicallmeauedodgtobaae factor，andthefaturedistributionbetwenthesouedomainandtetargetdomainisalignedtoimprovethepformanceofthemigration diagnosisodelTeproosdmethodwasapliedtoonducossconditiontransferdagosisexpermentsonthedatacletedfromthe rotortestbenchndwascomparedwithothertraditioaltrasfermodels.Tfectivenessandsuperoityoftismethdweriied.

Keywords： fault diagnosis;rotor; transfer learning;residual networks;dynamic domain adaptation

转子在旋转机械设备运行过程中起着十分关键的作用，转子发生故障会影响到整个装备的正常运行，造成巨大经济损失的同时还有可能带来人员的伤亡。（剩余13476字）