重要方面固定燃烧系统源PM2.5的排放测量和表征

FuelProcessingTechnology85(2004)687–

699

http://wendang.chazidian.com/locate/fuproc

ImportantaspectsinsourcePM2.5emissions

measurementandcharacterizationfrom

stationarycombustionsystems

S.WinLee*,I.He,B.Young

CANMETEnergyTechnologyCentre,NaturalResourcesCanada,Ottawa,ON,CanadaK1A1M1

Abstract

Duringtheconstructionandevaluationofasamplerformeasurementandcharacterizationfineparticulatematter(PM)emissionsfromstationarycombustionequipment,severaltechnicalchallengeswerenoted.Thesamplerdesignincorporateddilution,coolingandmoistureadditiontothestackgasinsideaninertdilutiontunneltocloselysimulatenear-ambientconditionstopromoteatmospherictransformationofsourceparticles.Theautomatedandon-lineprocesscontrolcapabilitiesofthesystemallowsforsimulationofarangeofambient-likehumidityandtemperatureconditionsforPMsampling,whileprovidingreproducibleparticulatemassemissionresults.SubsequentanalysesofthesizesegregatedPM2.5,PM10andtotalparticulatesamplesyieldconcentrationsofparticlemass,carbons,acidicspeciesandtraceelements.Thefirst-generationsamplingsystemwasappliedona150-kWoil-firedboileranda0.7-MWthcoal-firedpilotscaleboilertoprovidesourcePMcharacteristicprofiles.Challengesnotedduringtheinitialstudiesincluded(a)difficultyinachievingoptimumdilutionandresidencetimewhilesustainingisokineticsamplingforhighfluegasvelocities;(b)inaccuraciesinmeasurementandcontrolofsamplingsystemflowratestomaintainabalancedflowsystem;and(c)particledepositionsinseveralsystemcomponents.Asecond-generationsystemwaslaterconstructedthatprovidedahigherdilutionupto80-foldandtheextendedresidencetimesupto80s.ReliablemeasurementandcontrolofthegasflowrateswereachievedusingaCO2tracertechnique.Samplingsfromthecombustionunitswithstackvelocitiesrangingfrom3to10m/sweresuccessfullyperformed.ForPMmeasurementonboilerswithastackvelocityhigherthan10m/s,afluepre-separatororsplitterisrequired.Thesampler’soveralldesignisbeingfurthermodifiedforadditionalimprovementsincludingasuitablesplitterdesign.ThispaperfocusesprimarilyonthetechnicalissuesrelatingtosourcePMsamplingequipment,

*Correspondingauthor.

0378-3820/$-seefrontmatterD2004ElsevierB.V.Allrightsreserved.doi:10.1016/j.fuproc.2003.11.014

688S.W.Leeetal./FuelProcessingTechnology85(2004)687–699

whileinitialPM2.5massemissionresultsfromthecombustionofaNo.4fueloilarereportedsimplytoillustrateitscapabilities.

D2004ElsevierB.V.Allrightsreserved.

Keywords:Particulatematter;Stationarycombustionsystem;Dilution

1.Introduction

ThenewUSNationalAmbientAirQualityStandardsof1997andtheCanadaWideStandardsof2000demonstratetheimportanceofreportedassociationsbetweenparticlepollutionandadversehealtheffects[1,2].InCanada,PM10,orparticulatematterwithanaerodynamicdiameterequaltoorlessthan10Am,hasbeendeclaredtoxicundertheCanadianEnvironmentalProtectionActin2000[3].TheCanadaWideStandardsplacePM2.5underregulatoryadvisementthatneedstobeimplementedby2010[4].Muchhasbeenreportedinrecentyearsonpolicyandregulatoryissues,exposurelevelsandhealtheffects,sourceapportionmentmodelingandmethodologydevelopmenttobridgecriticalknowledgegapsrelatedtofinePM.Opposingviewsonthenewrulescontinuetoexistamongregulatorsandindustrybutitisagreedthatmorescientificdataareneeded.Althoughitisgenerallyacceptedthatcombustiongeneratedparticleshaveagreaterimpactonhumanhealththannaturallyoccurringparticles,onlylimitedsizeandchemicalcompositiondataareavailablefortheseanthropogenicsources,especiallyforstationarycombustionsystems.Asignificantlylargeramountofresearchdataisaccessibleforthetransportationdieselenginederivedparticulates,generallyknownasdieselparticulatematterorDPM.SourcePMcontributiontotheambientfromindividualpointsourcesincludingcombustionprocessesisdifficulttoelucidatebecauseofitscomplexatmo-spherictransformationsoccurringunderdiversemeteorologicalconditions.Sourceappor-tionmentmethodsarenormallyappliedinassessingtheimpactofthetypeandquantitiesofvariousemissionsourcesonambientPMconcentrations[5–9].Thesizeandchemicalcompositionofemissionsfromeachsourcetype,knownassourcesignaturesorsourceprofiles,arerequiredinsourceapportionmentmodeling.However,commonlyavailablesource-emissioninventorydatahavebeenidentifiedasinadequatesincesampleshavebeencollectedusingconventionalhightemperaturefiltermethodsthatdonotaddressadequatelythenormaldilutionandcoolingthatoccursinaplume.OneofthesereferencemethodsistheUSEPAMethod5,whichhasbeenextensivelyusedfortotalPMmeasurementfromvariousstationarycombustionsources.Basedonthisneed,sourcedilutiontechniqueswherethehotfluegaswasdilutedwithcleanairorinertgaspriortoparticulatesamplingweredeveloped[10–14].Thesourcedilutionsamplingapproachattemptstomimicatmospherictransformationofprimaryandsecondaryparticlesinaplumeinclosevicinityanddownwindofastack.However,inadequaciesstillexistinthecurrenttechnologiesduetothecomplexitiesassociatedwithsimulationoftheambientenvironmentandthemechanicsinvolvedincontrollingisokineticsamplingandsampledilution.

Intheinterim,environmentalregulationsarealsomovingtowardsstricterengineandfuelspecificationstoreduceairpollutantemissionsfromthetransportationand

S.W.Leeetal./FuelProcessingTechnology85(2004)687–699689

industrialsectors.Quantificationandimpactassessmentofparticleemissionsfromindividualpointsourcesareessentialfortheindustryfortheirenvironmentalpolicyandeffectivemanagementofplantemissions.Theapparentneedforasamplingprocedurethatallowsforclosesimulationofnaturalcooling,http://wendang.chazidian.comboratoryexperimentswereconductedtomeasurePMemissionsforseveralresidualoilsandpulverizedcoalblendsusingthefirstgenerationprototypemeasurementsystemandtheresultshavebeenreportedelsewhere[15–22].Theimprovedperformanceofthesecond-generationsourcePMmeasurementsystemwasevaluatedandtechnicalchallengesandtheapproachestakentoaddresstheseissuesarepresented.Sincethetimeoforiginalsubmissionofthisdocument,athird-generationsamplerhasbeendesignedandiscurrentlyevaluatedforfielddemonstration.

2.Experimental

ThispapermainlydealswiththetechnicalaspectsinthedevelopmentofasourcePMmeasurementandcharacterizationtechnologyandthereforelessemphasisisgiventotheexperimentalprocedures.Althoughseveralfuelsandcombustionsystemswereusedduringmanylaboratoryexperiments,thecombustionofonlyonespecificfuelwasincludedinthisdiscussion.

2.1.Testfuel

InTable1,thepropertiesoftheNo.4typeresidualfueloilarereported.AllfuelanalyseswereperformedatCETCfollowingtheASTMtestmethodsandestablishedTable1

PropertiesofNo.4fueloil

Propertiesoffueloil

Ultimateanalysis(wt.%)

Carbon

Hydrogen

Nitrogen

Sulphur

Watercontent—KarlFisher(wt.%)

Densityat15jC(kg/m3)

Grosscalorificvalue(cal/g)

Grosscalorificvalue(MJ/kg)

Grosscalorificvalue(Btu/lb)

Kinematicviscosityat100jC(cSt)No.487.0312.200.180.730.34897.910,5824419,0512.6

690S.W.Leeetal./FuelProcessingTechnology85(2004)687–699

proceduresdevelopedfromparticipatingininternationalroundrobinstudies.Duringcombustion,thefueloilisheatedandcontinuouslyagitatedusingacirculationpumptoensuresamplehomogeneity.

http://wendang.chazidian.combustionfacility

InitiallaboratoryexperimentsfortheprototypePMsamplingsystememployeda150-kWoil-firedboilerforsimplicityandcostsavings.Theunitisasinglepass,castiron,hotwater/steamboilerdesignedfordistillateoilornaturalgasfiring.TorepresentindustrialoilfiredboilersknowntogeneratehighPMconcentrations,theexistingunitwasmodifiedtoburnNo.4typeresidualoilbyretrofittingitwithawaste-oilburner.Oilatomizationwasassistedwithafuelpre-heaterandhigh-pressureairintroducedintothenozzlealongitscentreaxis.Thedescriptionoftheunithasbeenreportedinanearlierpublication[17].

2.3.Boileroperationandemissionsmeasurement

Operatingproceduresfortheoil-firedboilerhavebeenreportedinearlierpublications

[16–18].ThecombustionsystemwasoperatedwithsomewhatatypicalboilerefficiencysettingstoallowforsuitablePMsamplingconditions.Forexample,theoilboilerwasintentionallyfiredwithNo.4fueltosimulateindustrialboileremissionsthatusuallyburnresidualoils.Inaddition,residualfuelsgeneratehigherPMconcentrationsthanNo.2oilornaturalgas,therebyshorteningthesamplingtimedrastically.Fluegasvelocityoftheunitwasapproximately3m/s.TheboilerwasallowedtoreachsteadystatecombustionconditionstoattainrepresentativeandconsistentfluegasemissionspriortoPMmeasurement.Gaseousspeciesinthedilutedfluegaswerealsodeterminedusingstandardcontinuousemissionanalyzers.Itwasalsonecessarytomonitorthefluegastemperatureandvelocitytoestablishnear-isokineticsampling.Afterdeterminingtheisokineticsamplingrateforaspecificcombustionexperiment,totalsourceparticulatesamplingwascarriedoutusingthestandardEPAMethod5procedures.PMmeasurementusingthesourcedilutionprototypeequipmentimmediatelyfollowedMethod5toensurethatthemeasurementswereperformedundersimilarboileroperatingconditions.

2.4.Fineparticulatemeasurementsystems

Twoprototypesystemshavebeenconstructedandevaluatedfordieselandlightdistillateheatingoils,residualtypeNo.4andNo.6fueloilsandpulverizedcoalblends,usingpilot-scalecombustionfacilities.TheprimaryeffortsindevelopingasourcedilutionPMmeasurementsysteminvolvedtheselectionandconstructionofsystemcomponentsandmethodprotocols.Thekeyareasarethefluegassamplingprobe,dilutionairsupplysystem,moistureintroductionsystemforhumiditycontrol,thedilutiontunnel,residencetimechamber,PMsamplingportsandtheprocesscontrolanddatamonitoringcapabilities.Inbrief,thebasicprincipleofthemethodinvolvesdilutionoffluegaswithpurifiedairby20–40timesinsideadilutiontunnelmaintainedat40%relativehumiditytoallowforcoolingandsimulationofatmospherictransformationprocesses.PortionsofthedilutedgasarewithdrawnthroughselectedcycloneandimpactorinletsandfilterpackstocollectPM2.5,

S.W.Leeetal./FuelProcessingTechnology85(2004)687–699691

PM10andtotalPMfractions.Particulatesampleswereanalyzedfortheirfilterablemass,sizedistributionandconcentrationsoftraceelements,carbonandsolublesulphatesusingappropriateanalyticaltechniques.TheanalyticalschemefortreatmentofPMsampleshasbeenreportedelsewherewithaspecialemphasisonthereceptor-comparabilityofthePMcharacteristics[18].

Thefirst-generationsystemwasconstructedbyextensivemodificationofacommercialunitmanufacturedbyURGCorporationintheUSA.Themodifiedunitusedtheoriginalsamplingprobe,dilutiontunnelandthesamplingdevices,beingPM2.5cyclone,PM10impactorandfilterpackfortotalparticulates.Theunit’stotalvolumeis0.76ft3or21.5landsampleresidencetimesof15–34sarepossibleata40-foldsampledilution[22].Majormodificationstothecommercialsystemincludedtheintroductionofmoistureinjectionmechanism,massflowcontrollersandprocesscontrolsoftware.Controlfunctionsalsoincludeafeedbackfeaturethatautomaticallyadjustsandmaintainsthedilutionratioatapre-setlevel,inresponsetothefluctuationsofstackvelocityandfluegasconditions.InitialtestssuggestedthatadditionalimprovementsaredesirabletoachieveaccurateflowmeasurementsofthefluegassamplingrateandthePMsamplestreams,increasedturbulentmixingoffluegasanddilutionairinsidethemixingchamber,longerresidencetimeandtheabilitytosampleonstackswithhigherfluegasvelocities.

Basedontheobservationsfromtheinitialwork,acustomizedsystemspecificallyintendedforindustrialandutilityboilerapplicationshasbeendesigned,asshowninFig.1.Theimproved,second-generationsystemismadeupofseverallightweightmodularpiecesforfieldportability.Thefluegasanddilutionairmixingsectionmeasures60in.inlengthwith2.5in.ID.Thetransfersectionisa32-in.long,curvedconnectorthattransfersthedilutedgastotheresidencetimechamber.Thecylindricalchamberis72in.highwithadiameterof18.5in.Theentiredilutionsystemhasavolumecapacityof11.5ft3or325l.Thesystemallowshighersampledilutionratiosofupto80-fold.AresidencetimechamberwasalsointroducedtoincreasetheresidencetimeandanewPMsamplewithdrawalconceptwasincorporatedtoensurethathomogeneoussamplesaretransferredfromtheresidencechambertothecyclonesandfilterpacks.Thehorizontalarrangementforthesamplingportsontheprevioussystemwaschangedtoawell-spaced,verticallayouttominimizeparticulatesettling.Mostimportantly,aCO2tracertechniquewasintegratedtopermitaccuratemeasurementandcontroloffluegassamplingrates.ThesamplingsystemhasbeensuccessfullyusedforPMmeasurementonboilerswithmaximumfluegasvelocitiesof10m/s.Thiscapabilityisbeingupgradedforlargercombustioninstallationshavingvelocitiesofupto30m/s,byincorporatingafluegaspre-separatororsplitteronthesamplingunit.

Althoughthesecond-generationsamplersharessomecommondesignswiththeCalTechdesign[23],suchasamixingchamberandaresidencetimechamberwithsamplingportsattachednearthebottom,theCETCsamplerhasthefollowinguniquefeatures:

Themechanismtoadjustandcontrolrelativehumidityinsidethetunnel,whichhasstrongeffectonsecondaryparticulateformation.

UsesCO2tracertechniqueinsteadoforificemetertoaccuratelymeasureandcontrolfluegassamplingrates.Thisallowssimulationofdifferentambient-likeconditionsandalsoprovidescomparabledatafordifferentsamplinglocations.