建筑电气故障诊断文献综述及外文文献资料.docx

建筑电气故障诊断文献综述及外文文献资料.docx

《建筑电气故障诊断文献综述及外文文献资料.docx》由会员分享，可在线阅读，更多相关《建筑电气故障诊断文献综述及外文文献资料.docx（10页珍藏版）》请在冰豆网上搜索。

建筑电气故障诊断文献综述及外文文献资料

本份文档包含：

关于该选题的外文文献、文献综述

一、外文文献

文献信息

标题:

ReviewstotheResearchonBuildingElectricalIntelligentFaultSelf-diagnosis

作者:

MaileT;VincentiM

期刊名称:

BuildingandEnvironment；第8卷；第3期；页码：

29-42

年份:

2015

ReviewstotheResearchonBuildingElectricalIntelligentFaultSelf-diagnosis

Abstract

Wesummarizesomemethodsofthefaultdiagnosisinthepaper,basedontheFaultSelf-diagnosisresearching.BecauseoftheBuildingElectricalFaultSelf-diagnosisSystemisnotresearchedindepth,ourgroupfoundtheproblemofFaultSelf-diagnosisandproposetakingtheartificialintelligencemethod.Especiallyinthisfewyears,wediscussedtheapplyingofmanykindsofextensionsinBuildingElectricalFaultSelf-diagnosisSystembasedontheneuralnetworkandtheresultofdiscussingcanprovidesomenewideasforfurtherresearchingofFaultSelf-diagnosis.

Keywords：

Faultself-diagnosis Buildingelectrical Detectiontechnology

1 Introduction

Theelectricalandelectronicequipmentinbuildingsbringstheconveniencetopeople’slifeandwork,andhasalsoimprovedthequalityoflife.However,itwillbeahugedisasteriftheelectricalandelectronicequipmentbreakdown.Inordertopreventtheaccidentinbuildingofelectricalequipment,First,theequipment’sproduction,installation,operationconformtothecorrespondingdesignspecificationandprocess,itcanfundamentallyenhanceequipmentreliabilityanddecreasethefailurerate.Second,aneffectivemanagementmeasureincludestheequipment’smaintenanceandcarewhentheabnormalsituationcanbefoundearlier,foravoidingthispreventingaccidentsbeforetheyoccur.

Early,thewayofdetectingthebuildingelectricalequipmentfaultisthroughthemaintenanceworkerstocheckup.Therearealotofsubjectivityanduncertaintyinthisway.Itcannotsatisfytherequirementsthroughthefactitiousempiricalmaintenancealongwiththeelectricalequipmentwerebecomemoreandmorecomplicatedandtheirkindsarebecomingmoreandmore.Self-diagnosissystembecomesrealisticdemand.Atpresent,thecountriesinallovertheworldpayalotofmoneyforequipment’smaintenanceeveryyearbecauseofthemaintenancesystemsthatbeusedontheconstructionofelectricalequipmenthaveseriousdefects,suchasthefrequenttemporaryrepair,inadequatemaintenanceorsuperfluousmaintenanceandblindmaintenance,etc.Itisanimportantproblemthathowtosavemaintenance’scosttoconstructionofelectricalequipmentandatthesametimeensurethatthesystemhashighreliabilityforthesystem’soperationpersonnel.Thefault’smaintenanceresearchbasedontheequipment’sstatemonitoringandadvancedfaultdiagnosistechnologydevelopingbecameanimportantresearchintheconstructionofelectricalsystem,becauseofthesensingtechnology,microelectronics,computerhardwareandsoftware,digitalsignalprocessingtechnology,artificialneuralnetwork,expertsystem,fuzzysettheoryandothercomprehensiveintelligentsystemwareappliedtoconditionmonitoringandfaultdiagnosis.Specifictotheconstructionofelectricalfield,therearefivekindsofdirectbenefitstocarryouttheself-diagnosisoffaultinconstructionofelectricalequipmentsystem:

1.savealotofelectricalequipment’smaintenancefee;2.increasetheavailablecoefficientofelectricenergy;3.extendtheservicelifeofelectricalequipment;4.ensurereliabilityofthepowersupplyforelectricalequipment;5.reducethemaintenancecoastofelectricalequipmentfaultandthemaintenancerisk.Thispapermainlyintroducessystemevolution,thegeneralsituationandstateofdevelopmentandtheproblemofself-diagnosismaintenance.

Thisprojectadoptsmethodsthatarebasedonthemethodofclassification,andconsideredaboutpeopleneedtohaveanaccuratefaultclassificationandfaultlocationintheconstructionofelectricalfailure.Neuralnetwork,akindofclassificationmethod,hasthegoodpatternrecognitionabilityandneuralnetworkisgoodatdealingwiththeproblemthathassomestrongcouplingnonlinearcorrelationaboutmappingofsymptomsetorfaultset.Then,wedecidedtohaveafurtherstudyaboutthetheoryandpracticebasedonthefaultdiagnosisofneuralnetwork.

2 TheMainTypesandProgressoftheApplicationofFaultDiagnosisTheory

Thedevelopmentofthetheoryoffaultdiagnosishascometoanewstagenow.Thecoreisthatthemodernknowledgesystemofvariousdisciplineswassynthesized.Thefaultdiagnosistheoryaboutcontroltheoryandcontrolengineeringinvolvedthemoderncontroltheory,signalandsystem,powerelectronictechnology,modernstatistics,systemidentificationandpatternrecognitionandotherkindsofdiscipline.Inrecentyears,withtheresearchscopeoffaultdiagnosiscontinuestoexpand,itformedthatsomemainstreammethodsoffaultdiagnosis.

2.1 TheApplicationofArtificialIntelligenceintheElectricalSystem’sFaultSelf-diagnosis

TheArtificialIntelligence（AI）systemwasdevelopedandappliedinthefaultself-diagnosisofelectricalequipmentfield.ThereforetheExpertSystemandArtificialNeuralNetworkarethemostconspicuous.Theelectricalequipmentfaultself-diagnosisexpertsystemthathasbeendevelopedhastheknowledgebase,database,reasoningmachine,interpretationsystemandman-machineinterfacefiveparts.Theelectricalequipmentfaultself-diagnosisexpertsystem’sknowledgebaseusesthemodularstructure.AstheFigshows:

Theartificialneuralnetworkalsousesthemodularstructure.ItusestheRBFradialbasisfunctionneuralnetwork,bringinthefuzzylogictheoryandjudgesomefaultinelectricalequipmenttopreventtheoccurrenceoffailure.Theelectricalequipmentsystemcanmakeearlyanticipationforinternallatentfailureofoperationalelectricalequipmentanditcanprovide“consultation”totheoperatingpersonnelwhooperatetheelectricalequipment.Electricalequipmentfaultself-diagnosissystemknowledgebasemodulechart.

Atpresent,thefaultself-diagnosissystemofelectricalequipmentusingtheproducetypemodelsystemthatisthemostcommon,andtheknowledgebaseusethemodularstructure.Theknowledgebasemoduleismutuallyindependentanditisadvantageousinthemodification,extensionandrenewalofknowledgebase.Ittakesagreatconveniencetotheknowledgebase’smaintenanceofinthefaultself-diagnosissystemofelectricalequipment.

Usingthecharacteristicsoftheexpertsystemprogramminglanguagetorealizethereverseinferenceofelectricalequipmentfaultself-diagnosissystem’sobjectandbringinthefuzzylogicreasoningmodelandsuccessfullydealingwiththeproblemofsomefuzziness.

Electricalequipmentfaultself-diagnosissystemdatabaseconsistsofelectricalparameteranalysisanddynamicdatabase.

 2.2 TheMethodBasedontheTheoryModelofElectricalTestingEstimate’sFaultSelf-diagnosis

Aseveryoneknows,peopleusuallyusethestateequationofsinglecontinuousvariationtoaccuratelydescribewhentheelectricalequipmentsystemsendsfailure.Therandomhybridsystemmodelismoreaccuratemathematicalmodelinthefaultself-diagnosissystemofelectricalequipment.Thisrandomhybridsystemmodelisdifferentwithstochasticsystem,becausethestateofrandomhybridsystemcanbediscreteandcontinuous,sowecanknowthattherandomhybridsystemhascharacteristicofrandom.Thepassivefailuremodecontrolsystemthatisakindofelectricaltestsystemisakindoftypicalrandomhybridsystemmodelandtheyaredifferentbetweenthefailuremode’sstructuresoftheinternalelectricaltestsystem,andwecanknowthataredifferentbetweenthefailuremode’sstructureandnormalmode’sstructure.

Intheelectricaltestsystem,themultiplemodelestimationalgorithmincludetherandomhybridsystemmodel’sdesign,filter’sselection,thereinitializationofestimationfusionandfilter.Theprincipleblockdiagramofelectricaltestsystem’srandomhybridsystemmodelestimationalgorithmisshowninFig. 42.2.

MultiplemodelestimationalgorithmisbroughtoutbyMaiglwhenheresearchedtheoptimizationself-adaptionestimationofrandomsampledataprocessing.Butintheearly,multiplemodelalgorithmdidnotconsideraboutthemodelsswitchinganditbelongthemultiplemodelalgorithmwithstaticzeroinput.Inthefewyears,ChinaandothercountriesbegantoresearchtheMultipleModelSelf-adaptiveEstimateandMultipleModelSelf-adaptiveControl,theyareakindofmultiplemodelapproachoftypicalzeroinputinteraction.ThemethodofMultipleModelSelf-adaptiveEstimationmaybegeneratetheelectricalequipmentsystem’sestimationperformancedeteriorationandevenmaketheelectricalequipmentsystem’sestimatorappearthedivergence,becauseofthemethodofMultipleModelSelf-adaptiveEstimationlackthenecessaryinputmutualinformation.

2.3 TheElectricalSystem’sFaultSelf-diagnosisBasedonFuzzyNeuralNetwork

Whentheelectricalequipmentsystemisdoingthefaultself-diagnosis,thefuzzyneuralnetworkcanextractthecurrentelectricalequipmentsystem’sfuzzymodel’sdescriptionwithnonlinearityandcomparewitheachreferencemodel,thuspeoplecangettheaccuratefaultdiagnosis.ThesystemmodeloffuzzyneuralnetworkastheFig. 42.3 shows.Themethodofelectricalequipmentsystem’sfaultself-diagnosisbyusingneuralnetwork,althoughwecangetthedataoffailurecharacteristicsthroughthenetworkself-learningbutitistardinesstocomparewithsomecomplexelectricalequipmentsystemlearning.Andtheneuralnetworkisnotfullyinthenetwork’sstructure,soitisdifficulttoextracttheperformanceofneuralnetworkthroughtoanalysisthedataoftrainingcharacteristics.Inaddition,m