基于人工神经网络在电力传输线并联情况下的距离保护设置(IJISA-V4-N12-10)

I.J.IntelligentSystemsandApplications,2012,12,75-85PublishedOnlineNovember2012inMECS()DOI:10.5815/ijisa.2012.12.10Copyright©2012MECSI.J.IntelligentSystemsandApplications,2012,12,75-85DistanceProtectionSettingsBasedArtificialNeuralNetworkinPresenceofTCSRonElectricalTransmissionLineMohamedZellaguiLSP-IEResearchLaboratory,FacultyofTechnology,DepartmentofElectricalEngineering,UniversityofBatnaCampusCUB,StreetMedElHadiBoukhlouf,Batna,05000,[e-mail]-IEResearchLaboratory,FacultyofTechnology,DepartmentofElectricalEngineering,UniversityofBatnaCampusCUB,StreetMedElHadiBoukhlouf,Batna,05000,[e-mail]—Thisresearchpaperstudytheperformanceofdistancerelayssettingbasedanalytic(AM)andartificialneuralnetwork(ANN)methodfora400kVhighvoltagetransmissionlineinEasternAlgeriantransmissionnetworksatSonelgazGroupcompensatedbyseriesFlexibleACTransmissionSystem(FACTS)i.e.ThyristorControlledSeriesReactor(TCSR)connectedatmidpointoftheelectricaltransmissionline.Thefactsareusedforcontrollingtransmissionvoltage,powerflow,reactivepower,anddampingofpowersystemoscillationsinhighpowertransferlevels.ThispaperstudiestheeffectsofTCSRinsertiononthetotalimpedanceofatransmissionlineprotectedbydistancerelayandthemodifiedsettingzoneprotectionincapacitiveandinductiveboostmodeforthreezones.Twodifferenttechniqueshavebeeninvestigatedinordertopreventcircuitbreakernuisancetrippingtoimprovetheperformancesofthedistancerelayprotection.IndexTerms—DistanceProtection,TCSR,InjectedReactance,ArtificialNeuralNetworkI.IntroductionInrecentyears,powerdemandhasincreasedsubstantiallywhiletheexpansionofpowergenerationandtransmissionhasbeenseverelylimitedduetolimitedresourcesandenvironmentalrestrictions.Asaconsequence,sometransmissionlinesareheavilyloadedandthesystemstabilitybecomesapowertransfer-limitingfactor.FlexibleACtransmissionsystems(FACTS)controllershavebeenmainlyusedforsolvingvariouspowersystemsteadystatecontrolproblems[1].TherearetwogenerationsforrealizationofpowerelectronicsbasedFACTScontrollers:thefirstgenerationemploysconventionalthyristor-switchedcapacitorsandreactors,andquadraturetap-changingtransformerswhilethesecondgenerationemploysgateturn-off(GTO)thyristor-switchedconvertersasvoltagesourceconverters(VSCs).ThefirstgenerationhasresultedintheStaticVarCompensator(SVC),theThyristorControlledSeriesCapacitor(TCSC),andtheThyristorControlledPhaseShifter(TCPS)[2-3].ThesecondgenerationhasproducedtheStaticSynchronousCompensator(STATCOM),theStaticSynchronousSeriesCompensator(SSSC),theUnifiedPowerFlowController(UPFC),andtheInterlinePowerFlowController(IPFC)[4].ThetwogroupsofFACTScontrollershavedistinctlydifferentoperatingandperformancecharacteristics.InthepresenceofseriesFACTSdevices,theconventionaldistancerelaycharacteristicssuchasMHOandQuadrilateralaregreatlysubjectedtomal-operationintheformofoverreachingorunder-reachingthefaultpoint[5,6].Therefore,theconventionalrelaycharacteristicsmaynotworkproperlyinthepresenceofFACTSdevice.Artificialneuralnetwork(ANN)basedtechnology,whichisinspiredbybiologicalneuralnetworks,hasdevelopedrapidlyinthepreviousdecadeandhasbeenappliedinpowersystemprotectionapplications.Specificapplicationsincludedirectiondiscriminationforprotectingtransmissionlines[7-8],faultclassificationforfaultsondoublecircuitlines[9],ANNbaseddistancerelays[10],differentialprotectionofthreephasepowertransformers[11]andfaultsongeneratorwindings[12],theANNbaseddesignsofgenericprotectionsystemsproposedsofarworkwellonlyforidealfaultconditionsbutdonotmaintaintheintegrityoftheboundariesoftherelaycharacteristics.SomeofthepublishedresultsrelatedtotheapplicationofANNsindistanceprotectiverelayingimprovements76DistanceProtectionSettingsBasedArtificialNeuralNetworkinPresenceofTCSRonElectricalTransmissionLineCopyright©2012MECSI.J.IntelligentSystemsandApplications,2012,12,75-85arestatedinthesurveygivenby[13-14].Reference[15]demonstratestheuseofneuralnetworksasapatternclassifierforthedistancerelayingalgorithmandshowsthattheproposedschemeimprovesprotectionsystemselectivity.In[16]theuseofamultilayerfeedforwardnetworktoreducetheinfluenceofDCoffsetonfaultdistancecomputationisreported.Saturationofcurrenttransformersduringaheavyfaultcouldcauseincorrectdistancemeasurementbytherelay.PerformanceevaluationofdistanceprotectionschemeinpresenceofFACTScontrollers,whichaffecttheapparentimpedancecalculationsatrelaypoint,hasbeencarriedoutin[17].Theapparentimpedancecalculationsaregenerallycarriedoutusingpowerfrequencycomponentsofvoltageandcurrentmeasuredatrelaypoint.SomeofthepublishedarticlesdealingwiththeimpactonpresenceTCSCondistancerelayandmeasuredimpedancehavebeenreportedinreferences[18-20].Inthispaper,thethreeprotectionzonessettingforaMHOdistancerelayusingtwodifferentalgorithmsi.e.analyticandANNmethodareinvestigated.Theinvestigationconcerna400kVsingletransmissionlineinstalledateasternAlgerianelectricalnetwork.ThelineisequippedwithTCSRseriesFACTSdevices,incapacitiveandinductivemodes.Theperformancesofthe