.PowderTechnology11220007986Studiesontheeffectofswirlnumbersonstronglyswirlingturbulentgas-particleflowsusingaphase-DopplerparticleanemometerL.X.Zhou),Y.Li,T.Chen,Y.XuDepartmentofEngineeringMechanics,TsinghuaUniersity,Beijing100084,PeoplesRepublicofChinaAbstractTheeffectofswirlnumbersontheflowbehaviorofstronglyswirlingturbulentgas-particleflowswithswirlnumbersofss0.47,1.0,.1.5and2.1insudden-expansionandcyclonechambersisstudiedusinga2-Danda3-DphaseDopplerparticleanemometersPDPA.Theaxialandtangentialtime-averagedandtherootmeansquareRMSfluctuationvelocitiesofgasandparticlephasesandparticleconcentrationaremeasured.Theresultshowsthattheswirlnumberhasanobviouseffectontheaxialvelocityprofiles,theRankinevortexstructureoftangentialvelocityprofiles,therelationshipbetweentwo-phasevelocities,theturbulenceintensitylevelandtheanisotropyofturbulence.Keywords:Gas-particleflows；Stronglyswirlingflows；PDPAmeasurement1.IntroductionSwirlinggas-particleflowsareencounteredincycloneseparators,cyclonecombustors,hydrocyclonesandswirlwxburnerrcombustors1,2.However,mostofthepresentwxstudiesarelimitedtoasinglegas-phaseflowfield35.wxYuuetal.3studiedthegasflowfieldinacycloneseparatorusingprobeandhot-wiresystems,andtheresultsshowareductionofgasaxialandtangentialvelocitiesinthepresenceofparticles.TheprobemeasurementsmadewxbySilvaandNebra4givesimilarresults.ThestudiesbywxOgawaetal.5demonstratedthattherewasnodistinctdifferencebetweenthedust-ladengasflowandthepureairflowinacyclonechamber.Obviously,probemeasure-mentscannotgivegas-particletwo-phaseflowfield.Stud-iesusingconventionalandmodifiedLaserDopplerVe-.wxlocimeterLDV2,6showsimilarbehaviorofgasandparticleflowfieldse.g.,Rankinevortexstructuresof.tangentialvelocityprofilesandslipvelocitybetweengasandparticlephasesinswirlingflows.However,eventhemodifiedLDVmeasurementscannotgivereliableresults,becauseitisunabletoidentifyparticlesizesandtheeffectofparticlesizeisincludedintheparticleturbulentfluctuation.ThephaseDopplerparticleanemometer)Correspondingauthor.Tel.:q86-10-6278-2231；fax:q86-10-6278-5569.E-mailaddress:L.X.Zhou.PDPAmeasurementsallowobtainingthedetailedinfor-mationofparticlevelocity,sizeandconcentration.ThePDPAwasfirstusedtostudysudden-expansionswirlinggas-particleflowwithswirlingnumberss0.47bySom-wxmerfeldandQiu7.Theirresultsdemonstratethedifferentbehaviorofparticlesofdifferentsizesinswirlingflows.However,noPDPAmeasurementsofstronglyswirlinggas-particleflowsarereported.Inthispaper,thePDPAsystemisusedtostudystronglyswirlinggas-particleflowsinsudden-expansionandcy-clonechamberswithtangentialinlets.Theswirlnumberswerevariedfrom1.0to1.5and2.1.Theresultsarecomparedwiththoseforthecaseofss0.47andana-lyzed.Thepurposeofthisstudyistoclarifytheeffectofswirlnumbersonthebehaviorofswirlinggas-particleflows.2.Experimentalset-upandmeasurementmethodsTheexperimentalset-upisshowninFig.1aandb.Itconsistsofatestsection,afeedingsystemandaPDPA.system.ThetestsectionisaplexiglasschamberFig.1b.withoneaxialinletf60mmplustwosymmetric.rectangulartangentialinlets68=32mm.Thechamberisatubeof812-mmlengthandaninnerdiameterof120mm.Ononeside,aslotisopenedandapieceofopticalglassismounted.Theexittubeisacylinderof1500mm()L.X.Zhouetal.rPowderTechnology1122000798680.Fig.1.a1:Dataprocessor；2:signalprocessor；3:converter；4:transformer；5:powersupply；6:waterpurifier；7:laser；8:lasercon-troller；9:splittinggroup；10:emitterunit；11:receiver；12:testsection；13:flowstabilizer；14:cycloneseparator；15:powderfeeder；16:flow.meter；17:valve；18and19:compressor.bExperimentalset-up.lengthand96mmdiameter,towhichalargersudden-ex-pansionchamberisconnectedforprotectingthetestsec-tionfromthedisturbanceofdownstreamflows.Acycloneseparatorisusedtocollecttheparticles.Twoblowerswithvariableflowratessupplytheaxialandtangentialflowsviathevalves,wherebytheflowratemaybeadjusted.Theflowratesweremeasuredbythreeseparateflowmeters.Sphericalglassbeadsofsizesrangingfrom0to150mmwereusedinmeasurements.TheparticlesizedistributionisshowninFig.2.Particlesbelow10mmweretakenasthegastracer.Thisavoidedmismatchedrefractiveindicesbetweenseedinganddispersed-phaseparticlesandalsoeliminatedinterferencebetweenparticles.Thewidesizedistributionallowsobtainingthedifferentbehaviorofpar-Fig.2.Particlesizedistribution.Fig.3.PDPAinstrumentation:1,laser；2,splitter；3and4,lens；5,measuringvolume；6,receiver；7,filter；8,detector.ticleswithdifferentsizes.Particlesareintroducedfromtheaxialinletusingascrewfeeder.A3-DPDPA,madebyDantecInc.,anda2-DPDPAmadebyAerometrics,wereusedtomeasuresudden-ex-pansiongas-particleflows.Asforthemeasurementtechnique,the3-DPDPAwithabackwardscatteringarrangementisshowninFig.3.Itisawell-knownandwidelyusedtechnique.TheparticlevelocityismeasuredbasedontheDopplerfrequencyshift.TheparticlesizeandconcentrationaremeasuredbasedonthephasedifferencecausedbyMiescatteringandthenumberofparticlespassingthemeasuringvolumeduringacertaintimeinterval.TheparametersoftheopticalsystemareshowninTables1and2.Themeasurementswereperformedatfivecross-sec-tionsinthechamber,andforeachsection,thereareabout2535measuringlocationsalongthediameter.Inthesingle-phaseflowmeasurement,morethan1000samplesweretakenateachmeasuringlocation,while20005000samplesweretakenfortwo-phaseflowmeasurement.TheinletflowconditionsaregiveninTables35.Theswirlnumberisdefinedas:D3r222rwurdrH0Ss,D3r22DrurdrH40whereDistheinletdiameterandDisthechamber34diameter.Duringtheexperiments,thetangentialinletflowveloc-itywasconstant,andtheaxialinletflowvelocitywasTable1TransmittingopticsUUUxyz.Wavelengthofthelasermm514488476.FrequencyshiftMHz404040.Beamseparationmm747474.Diameterofmeasuringvolumemm1.351.351.35.Focallengthofthelasermm500500500.Fringeseparationmm3.53.313.23Fringenumber363636()L.X.Zhouetal.rPowderTechnology1122000798681Table2ReceivingopticsThemaximumdetecteddiameter260.24mm53Themaximumdetectedconcentration2=10rcmOff-axialangle1478Focallengthofreceivinglength600mmThevaluablesignlevely6dBThemaximumphaseerror258Themaximumsphericalerror15%.increasedfrom0case1to12mrscase3,andinturn,decreasedtheswirlnumberfrom2.1downto1.0.3.ExperimentalresultsanddiscussionFigs.412showthemeasurementresultsforcyclonic.gas-particleflowswithss2.1case1.Thecomparisonbetweentheparticle-ladengasandthepuregasflowfieldisgiveninFigs.47.Figs.4and5showthegeneralbehaviorofcyclonictwo-phaseflowfieldw-shapedaxialvelocityprofileswithanannularreverseflowzoneandRankine-vortextangentialvelocityprofilesoftwophases.Itcanbeseenthatthepresenceofparticlesreducesthegastangentialvelocityeverywhere.Thegasaxialvelocitydecreasesinthenear-wallzone,butincreasesinthenear-axiszone.ThemeasuredfluctuationvelocitiesFigs.6and7showthecentralpeakandtheincreasedvaluenearthewall,becauseofthelargevelocitygradientinthesolid-bodyrotationregionandtheregionnearthewall.Thecomparisonbetweenthesetwofiguresshowsthattheaxialfluctuationvelocitiesaremuchsmallerthanthetangentialones,demonstratingthehighanisotropyofturbulentfluc-tuations.Moreover,thesefiguresindicatethatthepresence.ofparticlesreducesgasrootmeansquareRMStangentialfluctuationvelocityalmosteverywhereandaxialfluctua-tionvelocityinmostregionsbutincreasesthelatteroneinreverseflowzones.Forparticleconcentration,alargeamountofparticles.almostconcentratesinthenear-wallregionFig.12duetothestrongcentrifugalforce.Figs.811givethecomparisonbetweenthetwo-phasetangential,axialvelocitiestogetherwiththeRMSfluctua-.tionvelocitiesfortwo-particlesizes60and100mm.Table3.Inletflowconditioncase1Axialinletvelocity0mrsTangentialinletvelocity10mrsSwirlnumber2.1Particlemeandiameter76.3mm3Particlematerialdensity2400kgrmParticlemassloading0.01kg-solidrkg-gasTable4.Inletflowconditioncase2Axialinletvelocity5mrsTangentialinletvelocity10mrsSwirlnumber1.5Particlemassloading0.01kg-solidrkg-gasFigs.8and9showthattheparticletangentialvelocitylagsbehindthegastangentialvelocityalmosteverywherewiththelargerparticleshavinglowervelocitiesduetotheirhigherinertia.Theparticleaxialvelocitylagsbehindthegasaxialvelocityonlyinthenear-wallregion.ItcanbeseenfromFigs.10and11thattheparticleturbulentfluctuationislowerthanthegasfluctuationinmostre-gionsbothinaxialandtangentialdirections,andthelargertheparticlesize,thesmallertheparticlefluctuation.Figs.13,1518givethemeasurementresultsforthe.sudden-expansionflowswithss1.5case2.AscanbeseenfromFig.13,inthiscasetheaxialvelocitiesoftwophaseshavenow-shapeddistributionsasthatforthecases.ofss2.1Fig.4and0.47Fig.14,butarehigherneartheaxisandlowerwithasmallpeaknearthewall.Theparticleaxialvelocitylagsbehindthegasoneinmostregions,andthevelocityslipbetweentwophasesincreaseswiththeincreaseinparticlesize,buttherelativeslipisnotverylarge.Inthenear-wallregion,theparticleaxialvelocityisslightlylargerthanthegasvelocity.Theparti-cletangentialvelocitylagsbehindthegastangentialveloc-.ityFig.15,andthelargertheparticlesize,themorethelag.Inthefirstandsecondsections,thestructureofrigidbodyrotationintangentialvelocityprofilesisnotobviousbecauseoftheinjectionofaxialnon-swirlingflows.Inthesubsequentsections,duetotheincreasingeffectofcentrifugalforce,thetangentialvelocitydistributionoftwophasesshowsanobviousRankine-vortexstructurewiththegraduallyvanishingeffectofinletconditions.Fig.16givestheparticleandthegasaxialfluctuationvelocitieswithhigherv