文档名：基于短时傅里叶变换和深度网络的模块化多电平换流器子模块IGBT开路故障诊断
摘要：针对现有模块化多电平换流器(MMC)子模块故障诊断过程中所需传感器较多、测量干扰较大等问题,提出一种基于深度学习的MMC子模块IGBT开路故障诊断方法.在对MMC子模块开路故障特征进行分析的基础上,利用短时傅里叶变换(STFT)提取桥臂电压信号的谐波分量信息作为故障诊断所需的特征参数.将所得到的特征参数进行处理后构建故障诊断样本,在通过深度置信网络实现故障类型快速检测的基础上,依据不同故障类型,构建多个基于卷积神经网络的故障定位网络,进而实现开路故障的检测与定位.通过129电平的MMC系统仿真模型和降功率的MMC实验系统搭建,对该文所提方法进行了验证.仿真和实验结果表明,所提故障诊断方法可以在减少传感器数量的基础上实现子模块开路故障的诊断,提高系统的可靠性.
Abstract：Themodularmultilevelconverter(MMC)isanewtypeofmultilevelconverterthatfeatureslowswitchfrequency,highoutputwaveformquality,andlowconsistencyrequirementsforpowerelectronicdeviceswitching.Therefore,ithasbeenwidelyappliedinvariousfieldssuchasflexibleDCtransmission,powerelectronictransmission,andrailtransittractioncontrol.However,theincreaseinthenumberofsub-modulesalsomeansanincreaseinpotentialfailurepoints.Tomeetthegrowingdemandsforsystemsafetyandreliability,timelyandeffectivediagnosisneedstobecarriedoutaftersystemfailures.ExistingMMCfaultdiagnosismethodsmainlyfallintotwocategories:additionalhardwarecircuit-basedandsoftwarealgorithm-based.Amongthem,thehardware-basedmethodexhibitsexcellentdiagnosticaccuracyandshortdiagnostictimebutmayincuradditionalhardwarecostsandwiringcomplexity,whilethesoftware-basedmethoddoesnotrequireadditionalhardwarecircuitryandisrelativelysimpleandconvenienttoimplement.However,existingsoftware-basedmethodoftenheavilydependsonthecapacitancevoltageinformationofsub-modulesduringthediagnosisprocess.Theexcessivenumberofvoltagesensorsnotonlyincreasesthecomplexityofsystemwiringbutalsoaddstothepotentialfailurepointsofthesystem.Tothisend,adeeplearning-basedIGBTopen-circuitfaultdiagnosismethodforMMCsub-modulesusingaminimumnumberofvoltagesensorsisproposed,whichonlyrequirestwoarmvoltagesensorsforeachphase.Basedontheanalysisofopen-circuitfaultcharacteristicsofMMCsub-moduleIGBTs,thebridgearmvoltageisselectedastherequiredelectricalparameterforfaultdiagnosis.Theshort-timeFouriertransform(STFT)isutilizedtoextractharmoniccomponentinformationfromthebridgearmvoltagesignal.Byarrangingfrequencydomaininformationatdifferentsamplingtimes,thematrixcontainingtime-frequencydomaininformationisconstructedasthediagnosticsample.Bycombiningwiththedeepbeliefnetwork(DBN),anIGBTopen-circuitfaultdetectionmethodbasedonSTFT-DBNisproposed,whichrealizestherapiddiagnosisofthefaulttype.Meanwhile,toaddresstheproblemofrelativelylowlocationaccuracyoftheDBN,aconvolutionalneuralnetwork(CNN)basedfaultysub-modulelocationnetworkisestablishedforeachfaulttypeafterthefaulttypeisidentifiedbythefaultdetectionnetwork.Combiningthesetwomethodsallowsforrapidfaulttypedetectionandhigh-accuracyfaultysub-modulelocation.A129-levelMMCsimulationmodelwasestablishedusingMatlab/Simulink,andarmvoltageinformationwascollectedundernormaloperationandvariousfaultconditions.Byapplyingtheproposedmethodfordatapreprocessing,inputsampleswereconstructed,andthedatasetwasdivided.Thenthefaultdiagnosisnetworksweretrainedandtestedseparately,andtheeffectivenessoftheproposedmethodwasverifiedbasedonsimulationdata.Finally,thehardwareandsoftwareofanMMCexperimentalsystemweredesignedinthelaboratoryenvironment,andanexperimentalsystemwithreducedpowerlevelswasconstructed.ThecontrolunitwasimplementedbyacombinationofDSPTMS320F28335andFPGA,andvariousfaultconditionsweresimulatedbychangingthedrivesignals.Armvoltagesensorswereusedtocollectdatafromtheexperimentalsystemunderdifferentfaultconditions.Theexperimentalresultsshowthattheproposedmethodcanachievehigh-precisionandshort-timefaultdiagnosisbyusingonlyarmvoltageinformation.
作者：朱琴跃 于逸尘 占岩文 李杰 华润恺Author：ZhuQinyue YuYichen ZhanYanwen LiJie HuaRunkai
作者单位：同济大学电子与信息工程学院上海201804
刊名：电工技术学报 ISTICEIPKU
Journal：TransactionsofChinaElectrotechnicalSociety
年，卷(期)：2024, 39(12)
分类号：TM46
关键词：模块化多电平换流器 开路故障诊断 短时傅里叶变换 卷积神经网络
Keywords：Modularmultilevelconverter(MMC) open-circuitfaultdiagnosis short-timeFouriertransform convolutionalneuralnetworks
机标分类号：TM73TP391.41TM46
在线出版日期：2024年7月19日
基金项目：国家自然科学基金基于短时傅里叶变换和深度网络的模块化多电平换流器子模块IGBT开路故障诊断[
期刊论文] 电工技术学报--2024, 39(12)朱琴跃 于逸尘 占岩文 李杰 华润恺针对现有模块化多电平换流器(MMC)子模块故障诊断过程中所需传感器较多、测量干扰较大等问题,提出一种基于深度学习的MMC子模块IGBT开路故障诊断方法.在对MMC子模块开路故障特征进行分析的基础上,利用短时傅里叶变换(STF...参考文献和引证文献
参考文献
引证文献
本文读者也读过
相似文献
相关博文
基于短时傅里叶变换和深度网络的模块化多电平换流器子模块IGBT开路故障诊断 IGBT Open-Circuit Fault Diagnosis of Modular Multilevel Converter Sub-Module Based on Short-Time Fourier Transform and Deep Networks
基于短时傅里叶变换和深度网络的模块化多电平换流器子模块IGBT开路故障诊断.pdf
- 文件大小:
- 9.74 MB
- 下载次数:
- 60
-
高速下载
|
|
