外文翻译--多传感器信息融合技术在液压系统的故障诊断上的应用 英文版.pdf
Applicationofmulti-sensorinformationfusiontechnologyonfaultdiagnosisofhydraulicsystemLQZhang1,2,GLYang1,2,LGZhang3andSYZhang41SchoolofEnergyandPowerEngineering,LanzhouUniversityofTechnology,QiliheDistrict,Lanzhou,730050,China2WenzhouAcademyofPumpandValveEngineering,LanzhouUniversityofTechnology,MadaoWestRoad,Oubei,Yongjia,Wenzhou,325105,China3HandanSpecialSinkingLimitedCompanyofChinaCoal,ChinaCoalFifthConstructionCompany,FuxingDistrict,Handan,056003,China4ChineseAcademyofAgriculturalMechanizationSciences,ChaoyangDistrict,Beijing,100083,ChinaE-mail:izlq163.comAbstract.Thestructurallayersandmethodsofmulti-sensorinformationfusiontechnologyareanalysed,anditsapplicationinfaultdiagnosisofhydraulicsystemisdiscussed.Aimingathydraulicsystem,amodelofhydraulicfaultdiagnosissystembasedonmulti-sensorinformationfusiontechnologyispresented.Choosingandimplementingthemethodofinformationfusionreasonably,themodelcanfuseandcalculatevariousfaultcharacteristicparametersinhydraulicsystemeffectivelyandprovidemorevaluableresultforfaultdiagnosisofhydraulicsystem.1.IntroductionHydraulicsystemplaysanimportantroleinengineeringindustry.Toensurethathydraulicsystemisworkingsafely,reliablyandwithoutanypotentialaccident,itsfaultdiagnosisisveryimportant.Butengineeringpracticeshowsthatfaultdiagnosisbasedononeparametercannotmakesurewhetherthesystemisoutoforderornotallthetime.Andbymulti-sensorinformationfusiontechnology,differentparametersabouttheoperatingconditionsofthehydraulicsystemfromdifferentanglescanbeobtained.Integratingandfusingalloftheparameterseffectively,thefaultdiagnosisofhydraulicsystemissuccessfullycarriedoutandthefaultofhydraulicsystemcanbeidentifiedandlocatedmoreaccurately1,2.Inthispaper,thestructurallayersandmethodsofmulti-sensorinformationfusiontechnologyareanalyzed,andthenitsapplicationinfaultdiagnosisofhydraulicsystemisdiscussed.2.Technologyofmulti-sensorinformationfusionMulti-sensorinformationfusionisamultilayer,allroundprocessingprocedure.Itcandetect,fuse,correlate,estimateandcombinealloftheparametersmeasuredinhydraulicsystemtoachievethestateestimation,includingsituationestimationandriskestimationofthesystemaccurately1,2.Tofaultdiagnosissystemofhydraulicsystem,multi-sensorinformationfusionconsistsofdatafusionandknowledgefusion,inadditiondata-to-knowledgefusion,thatisdatamining,isalsoincluded.2.1.Layersofmulti-sensorinformationfusionAsshowninfigure1,informationfusioncanbepartedinto3layers3.Figure1.SchematicdiagramoflayermodelofinformationfusionInformationfusionofdetectinglayerandfaultdiagnosis.Informationfusionofdetectinglayeristofuseoriginalinformationmeasuredbythesamekindsensorsbeforetheirpretreatment.Bythis,infirsttime,theoperatingconditionsofthesystemcanbemonitoredintuitivelyandperceptually.Atthesametimealloftheinformationisinputtedintodatabasetocarrydataminingout.Informationfusionoffeaturelayerandfaultdiagnosis.Informationfusionoffeaturelayeristofuseoriginalinformationmeasuredbyallkindssensorsandrelatedtheoreticalknowledge.Bythisthefaultofhydraulicsystemcanbeidentifiedandlocated,butitisall.Thespecificmethodsandtechnologyaimingatthefaultdiagnosiscannotbepresentedhere.Informationfusionofdecisionlayerandfaultdiagnosis.Thisisthefusionofthehighestlayer.Allinformationmeasuredbydifferentkindssensorsandrelatedtheoreticalknowledgearefusedandthecountermeasures,thatisthespecificmethodsandtechnologyaimingatthefaultdiagnosisincludingfaultisolation,redundancycontrollingandsoonareachieved.Andifthecountermeasuresareprovedtobeworkable,theexperienceofthistypicalcasecanalsobeinputtedintothedatabasetousesometime.2.2.Methodsofmulti-sensorinformationfusionTherearemanymethodsofmulti-sensorinformationfusion,suchasbasedonBayestheory,Demper-Shafer(D-S)theory,neuralnetworktechnologyandsomeestimationtheoryandsoon4,5.AsamodifiedtheoryofBayestheory,D-Stheory,alsocalledevidencetheory,hasawiderapplicationinmulti-sensorinformationfusiontechnology.Thismethodavoidsthesimpleassumptiontoanunbeknownprobabilityandshowsthedeterminacyandindeterminacyofinformation.ThebasicmethodofD-Stheoryisdividingtheevidencesetintomutuallyindependentparts.Eachevidenceparthasaprobabilitydistributionfunctiontotherelatedtheoreticaldiagnosis,alsocalledbelieffunction.Basedonthefusionofdifferentevidenceandtherelatedtheoreticaldiagnosis,thatistointegrateallofthebelieffunctions,thetotalbeliefdegreeofintegratedevidencebasedontherelatedtheoreticalknowledgecanbeobtained6,7.Figure2showsthecourseofreasoningofD-Stheory.Figure2.SchematicdiagramofthecourseofreasoningofD-Stheory3.Applicationofmulti-sensorinformationfusiontechnologyonfaultdiagnosisofhydraulicsystem3.1.StructureandprincipleofthefaultdiagnosissystemThecommonfailuremodesofhydraulicsystemconsistofoilleakage,abrasion,corrosion,fatigue,cavitations,hydraulicpressureseizure,andimpactbreakandsoon.Thustoahydraulicsystemthemonitoringparametersincludehydraulicpressure,flowquantity,temperature,oilleakageandsoon.Thefaultdiagnosissystemofhydraulicsystembasedonmulti-sensorinformationfusiontechnologyincludestwofunctionmodules,dataacquisitionmoduleandcentralprocessmodule.Thedataacquisitionmoduleisinstalledateachmajorcomponent,includingsensor,signalconditioningcircuit,A/Dconvertor,andbusinterfaceandsoon,toachieveeachstatussignalofthehydraulicsystemacquisitionandtransmission.ThecentralprocessmoduleconsistsofCPUandthesoftware.Consideringthatthecomplexitiesofhydraulicsystem,useIPCastheCPUofinformationfusiontoachievethedataanalysis,fusion,faultdiagnosisandgivingcountermeasure7-9.Figure3showstheblockdiagramofthemodeloffaultdiagnosissystem.Figure4showstheflowchartofthediagnosticprogram.3.2.CharacteristicsofthefaultdiagnosissystemAlloftheoperatingparameters,includinghydraulicpressure,flowquantity,temperature,oilleakageandsoon,canbemonitoredanytime.BasedonD-Stheory,byfusingalloftheoperatingparametersmeasuredbysensorsthestaterecognition,thetypicalfaultdiagnosis,andthesafetyprotectioncanbeachievedsuccessfully.3.3.KeytechnologyofthefaultdiagnosissystemInordertomakeallthesignalsaccurate,howtochooseeachsensorreasonablyandpretreatallthesignalsavailably.Inordertomakesurethatthefaultdiagnosisisright,howtochooseandimplementthemethodofinformationfusion.BecausethemethodofD-Stheorymaynotworkwellundercertaincondition,theothermethods,forexample,neuralnetworktechnologymaybeabetterchoice.Figure3.BlockdiagramofthemodeloffaultdiagnosissystemFigure4.Flowchartofthediagnosticprogram4.ConclusionsBasedonmulti-sensorinformationfusiontechnology,thefaultdiagnosissystemmakesfulluseofmultiplesignalsthatcanbemeasuredfromhydraulicsystemtorealizeconditionalarminganddiagnosisofthehydraulicsystems.Thiscanincreaseworkefficiencyandreliabilityofthehydraulicsystems.Themodeloffaultdiagnosissystemofhydraulicsystempresentedinthispaperisageneralizedmodel.Inspecificengineeringpracticethemonitoringparametersandtheactualstructureandimplementationofthefaultdiagnosissystemdependonthecorrespondinghydraulicsystem.AcknowledgmentsWewouldliketothankthesupportofScienceandTechnologyProjectofWenzhouCity(H20110007)andNaturalScienceFundofGansuProvince(1014RJZA023).References1VarshneyPK1997Multi-sensorDataFusionElectronics&CommunicationEngineeringJournalDecember2452532WaltzEandLlinasJ1991Multi-sensorDataFusionArtechHouse35423AlanNS2001DatafusionsystemengineeringIEEEAESSSystemsMagazineJune7144HalldL2000Mathematicaltechniqueinmulti-sensordatafusionArtechHouse15215HFDurrantWhyte2001Sensormodelsandmulti-sensorintegrationTheInt.J.ofRoboticsResearch7(6)87926MouradO2004SomenotesonfusionofuncertaintyinformationInt.J.ofIntelligentSystems19(6)4574717RichardT2003PrinciplesofeffectivemultisensordatafusionMilitaryTechnology27(5)29378ZhangYDandJiangXW1999Multi-sensorinformationfusiontechniqueanditsapplicationonintelligentfaultdiagnosissystemChineseJ.ofTransducerTech.18(2)18229AnFY,LuHW,LiuCJ,etal.2006ResearchandapplicationofinformationfusiontechnologyonmachineryfaultdiagnosisChineseJournalofChongqingUniversity29(1)1518