ML17055C375
| ML17055C375 | |
| Person / Time | |
|---|---|
| Site: | Nine Mile Point  |
| Issue date: | 09/05/1986 |
| From: | HALITSKY J JAMES HALITSKY |
| To: | |
| Shared Package | |
| ML17055C371 | List: |
| References | |
| NUDOCS 8609120406 | |
| Download: ML17055C375 (32) | |
Text
ATTACH?IE'l<T3COttCEttTRATIONCOEFFICIENTSINATMOSPHERICDISPERSIONCALCULATIOttSByJamesHalitsky,Ph.D.~Theconcentrationcoefficienttechniqueisamethodforcalculatingthe.concentrationfieldintheatmosphereinthevicinityofabuildingwhenanairbornesubstanceisreleasedfromornearthebuilding.TheoperativeparameterintheconcentrationcoefficientKwhich,inthema)orityofapplications,isindependentofsubstancereleaserate,windspeed,andbuildingsize,butisdependentonwinddirection,buildingshape,sourceconfiguration,andreceptorlocation.Thepredictionofrealconcentrationsdependsonone'sabilitytomakeanappropri-ateestimateofKandtotransformitintoaconcentrationestimate.Thispaperdescribesavailable.datasourcesforK,examinesitsnature,andillustratesitsusebyacasestudy.Atmosphericdispersionisamixingprocesswherebyairbornematterisspreadoveranever-increasingvolumeofairspacebytheturbulentmotionoftheatmosphere.Acontinuousreleaseofmatterintoasteadywindproducesastationary(time-independent)plume,characterizedbynonzeroconcertrationsofthedispersedmatter.Thispaperdealswiththecalculationofsuchconcentratiorsinplumescreatedfromsourcesnearbuildingsurfaces.PlumesofthistypehaveconcentrationdistributionsdifferentfromtheGaussiandistributionthatexistsinfree-streamplumeslyingwellabovetheregionofwinddisturbancecreatedbythebuilding.Themostaccu-ratemethodforestimatingconcentrationsinstationaryplumesfrombuildingsourcesistheconcentrationccefficienttechnique.Aconcentrationcoefficientisanondimensionalrepresentationofarealconcentrationinthesamesersethatapressurecoefficientisanondimensionalrepresentationofarealpres-sure.Inbothcases,thecoefficientisfoundbydividingameasuredquantitybyanartificialreferencequantityconstructedfromthefieldboundaryconditions.Forapressurefield,thereferencequantityisthedynamicpressure.Foraconcentrationfield,thereferencequantityisanartificialconcentrationCf(amount/volume)createdfromthereleaserateQ(amount/time)ofpurematter,themganwindvelocityU(length/time)atadesignatedlocation,refandacharacteristioareaA(length),producingCf=Q/AUForaconcentrationC(amount/volume)ataspecifiedpointx,y,z,thecorrespondingconcentra-tioncoefficientK(dimensionless)isKaC/Crf=CAU/Qo(2)~J.HalitskyisaconsultantinEnvironmentalMeteorology,122N.HighlandPlace,Croton-on-Hudson,NXt0520.PSHtt~ftAvD411C'one.19$J860'F1204068b0905PDR*DOCK050004fO(APDR
TheusefulnessofKderivesfromthefactthat,inmanycases,itremainssubstantiallyconstantforawiderangeofmagnitudesofQ,A,andU.ThisrakesitpossibletoestimateCinthefull-scaleatmospherefromKobtairedinawindtunnelrodeltest,simplybyusingEqua-tion2withfull-scalevaluesofQyAyandU.ThefeasibilityofthetechniquerestsontheavailabilityofpublishedKdataandthevalidityofapplyingthedatatoaPull-scaleconfi-gurationwhichdoesnotresemblethemodelconfLgurationindetail.Thispapercontains(1)areviewoftheavailabledatabaseforestimatingK,(2)asec-tiononthenatureofK,(3)acasestudytoillustratetheapplicationoftheconcentrationcoefficienttechniquetoacomplicatedbuilding,and(4)anevaluationoftheaccuracyofthetechniquebycomparisonoftheestiratedconcentrationswithconcentrationsmeasuredinawindtunnelmodeltestofthestudyprototype.KLsusuallyreportedintheliteratureasKisoplethsintheairspacesurroundingabuildingorgroupofbuildings.Onlythreefairlycomprehensivestudieshavebeenmadeonsimplegeometricshapes,buttheseareimportantbecausesuchshapescanbeconstruedtoresemblepor-tionsoflarger,morecomplicatedstructures.TheexperimentsthatproducedtheKisoplethswereconductedwithwindtunnelmodels;thefullreportsontestingprocedures,datareduction,anddatainterpretationaregiveninlaboratoryprospectreportsReferences1,2and3.AbridgementsofReferences1and2appearLntheopenliteratureasReferences4and5.AnabrLdgementofReference3appearsinReference6.References7and8presentanup-to-date(1980)compilationofresearchdataonflowanddiffusLonnearbuildingsandcontainsomeoftheKisoplethdrawingsLn'References1and3.Lrmu~aAnexampleofaKLsoplethdrawing,abridgedfromReference1,isshownLnFigure1.TheLsoplethsaredrawnintheplaneofeachvisiblebuildingsurfaceandLntheairspaceinsec-tionsaboveandtothesideof'hebuilding.EachisoplethLsidentifiedbyavalueofKrang-ingfromzeroatsomedistancefromthebuildingtoamaximumattheexhaustport.Eachiso-plethlineis,infact,theintersectionoiaconstantKsurfacewithabuildingsurfaceorasectioninspace.AmentalreconstructionwillshowthatthespacearoundthebuildingisoccupiedbyacontinuousfieldofK,madevisiblebydiscreteconstantKsurfaces.TheKLsoplethsurfacesandlineswerecreatedbyinterpolatingcurvesthroughanarrayofKdatapointswhichhadbeenobtainedbytransformingmeasurementsofCLnawindtunneltesttocorrespondingvaluesofKbyEquation2andplottingthemLnaspacecreatedbydividingallreallengthsbythebuildiqgheightH.Thereforeeachdatapointrepresentedanondimensional-izedconcentrationatanondimensionalixedlocation,i.e.,C(x,y,z,)-K(x/Hyy/Hyx/H)g(3)andtheconstantofproportionalityisCfofEquation1.refInestablishingCefitisnecessarytoadoptsomeconventionforthedesigationsorQiA,andU.Invariably,thesourcestrengthQistheflowrateofpurecontaminantpassingthroughtheexhaustportcrossseotion.HoreflexibilityisavailableforAandU~ItLspreferable,butnotmandatory,thatAbeassociatedwithadistinctivefeatureoitheflowfieldaroundthebuilding.ThisfeatureLsalocalzoneoftoroidalcirculation,calledacavity,lyingwithinalargedisturbedflowxonecalledawake.Thecavityandwakeoriginate,andarecoincident,attheupwindedgesofthebuilding,butthecavityisfiniteinlengthandmaxLmumcrosssection,whilethewakecrosssectiongrowscontinuouslywithdistancedownwinduntilthewakedisappearsasfreestreamkineticenergydiffusesintoit.
CavityflowcontrolsdispersionnearthebuildingbecausethecontaminantisusuallydischargedwithinthecavityorsufficientlycloseoutsidetodisperseinaflowfLeldthatmustconformwiththecavityshape.Thebuildingfeaturethatcontributesmosttothecreationofthecavityflowisthebuildingfrontalareaprogectedonaplanenormaltothewind.References1and3bothemploythefrontalareaforAbutwithadifferencethatmaybesigni-ficantifthebuildingislongandnarrow.'TheAinReference1isthefrontalareaofthelargestsideofthebuilding;itisinvariantwithwinddirection.TheAinReference3isthefrontalareapro)ectedonaplanenormaltothewind;itvarieswithwinddirection.Thelatterismorecloselyrelatedtocavitysixe,buttheformerLssimplertouseincalcula-tions.Bothareacceptable,buttheLsoplethsineachreferencewerederivedwiththedesig-natedAandshouldremainassociatedwithitinsubsequentapplLcations.TwocandidatesforreferencevelocityaretheexhauststreamvelocityVandthewindvelocityU.Both,actingtogether,defirethetotalflowfield,buttheformerdominatesintheregionneartheexhaustportwhilethelatterismoreimportantelsewhereinthecavity.Thelatterismorecommonlyusedbecausereceptorsnearbuildingswillbeinregionsoflowconcentrationawayfromtheportvicinity.References1and3bothusewindvelocityatroofheightforU.TheheLghtspecifLcationLsnotimportantLnReference1becausethewindhadauniformmeanvelocityprofile.ZtissignificantinReference3becausetheprofilewasoftheboundarylayertype,L.e.,velocityincreasingwithheightfrommeroatthegroundasinthenaturalatmosphere.UseoffrontalareaforAandroofwindvelocityforUissuitableforisolatedstructuresasinReferences1and3.Practicalitydictatesotherchoicesinothersituations.Forexam-ple,inreportingfull-scaletestsofdispersionatanuclearreactorcomplexinReference10,Awasdefinedasthefrontalareofthereactorcontair~entstructurealone,althoughmanyotherlargebuildingsLnthecomplexalsocontributedtocavityformation,andUwasdefiredasthewindvelocityatthe6m(19.7ft)levelonatowerlocated600m(1,969ft)upwindofthecoplex.Kvaluesderivedfromthisreferencemustbead)ustedforusewithAandUspecifiedLnotherbuildingarrangements.Theonlyotherrefererceparameteristhelengthusedfornondimensionalixingrealdis-ances.BuildingheightHwasusedfortheblockbuildingsinReferences1and3.Thebuild-inginReference2wasahalf-sphereatopaverticalcylinder;thereferencelengthwasselectedtobethediameterratherthantheheight.Thisconvention,also,mustberetairedinlocatingKvaluesfromReference2.TheconstancyofKoverarangeofscalesdependsoninvarianceofthenormalizedflowfLeldinwhichthedispersiontakesplace.Suchinvarianceoccurs,accordingtothehydro-dynamioequationsofmotionanddispersion,whentheconfigurationisinvariantandaminimumReynoldsNumberisobserved.AconfigurationLsastatementofnormalizedboundaryconditionsforaspecificflowanddispersionfield.ThefLeldunderconsLderationhereistheatrosphereinturbulentmotionoverthebuildingandsurroundingterrain.AlthoughthefieldLseffectivelyinfiniteinextentupwardandhorixontallyoutwardfromthebuilding,itisconvenLentandsufficientlyaccuratetoconsideronlythatportioninanimaginaryboxonthegroundenclosir~thebuild-ing,Thewallsandroofoftheboxaresetonlyfarenoughawayfromthebuildingtoprovidewindpropertiesintheirplanesessentiallythesameasifthebuildingwereabsent.Adis-tanceofthreebuildingheightsLsusuallyrequiredinthelateral,vertical,andupwinddirec-tions;alargerdistanceisrequiredinthedownwinddirectiontoaccommodatetheslowdecayofwinddisturbancescreatedbythebuilding.Theconfigurationhasthreecomponents.ThegeometricconfLgurationLstheshapeofthefacesofthebox,fLveofwhichareorthogonalimaginaryplanesandone,thebottom,isanirregularsolidsurfaceconformingtotheterrainandbuildingexteriorcontours,continuousexceptfortheexhaustandintakeports.Thedynamicconfigurationisthedistributionofvelocitiesalongtheboxfaces.(DensityandtemperatureareignoredinthepresentcontextbecausedifferentialsbetweenbuildingandatmosphericairaretoosmalltomakesignLficantchangesfromtheflowpatternsthatwouldexistunderisothermalconditions.)Thesourcecon-figurationisthedistributionofconcentrationacrosstheexhaustport.TwoconfigurationsaresaidtobethesameLfthepropertiesofoneconvertintothepro-pertiesoftheotherwhenmultipliedbyasingleconstant.Theconstantistheratioofmagni-tudesofacharaoteristioproperty.Forgeometricsimilarity,itistheratioofreference
lengths.Fordyr~icsimilarity,itistheratioofreferencewindvelocities.ForsourcesLmilarity,itistheratioofreferencesourceconcentrations,whichisproportLonaltotheratioofsourcestrengthswhengeometricanddynamLcsimilarityarepresent.TheminimumReynoldsNumber,formedfromthereferencelength,thereferencevelocity,andthedensity,restrictstheamountofturbulentenergyconvertedtoheatbymolecularinterac-tiontoaverysmallvalue,therebypreservingtheturbulentcharacterofthefield.Thiscri-terionisimportantinsettingwindtunneltestconditionsbecausethesmallmodelsizecreatesasmallReynoldsNumber.Itisautomaticallyobservedwhenmodel-generatedKisoplethsareusedforestimatingfull-scaleconcentrationsbecausethefull-scaleflowfieldwillhaveahigherReynoldsNumberthanthati,nthetest.Inpractice,exactsimilarityofconfigurationsbetweenmodelandfullscaleisnotpossi-bleornecessary,providedthatsubstantialsimilarityisachievedinmaJorfeatures.Forexample,inmodeltestingofagivenprototypebuilding,mirorsurfaceirregularitiesareomit-tedandlargerprotrusions,recesses,andevenotherbuildingslocatedoutsidetheexhaust-intakepatharereplicatedonlyincrudeblockform.1ihenusingmodel-derivedisoplethstoestimatefull-scaleconcentrationsforanotherbuildingconfiguration,greaterdeviationfromsimilarityistobeexpected;however,considerabledeviationcanbetoleratedwithouttoogreatapenaltyinreducedaccuracy.AsanaidinestimatingKwhenexactsimilarityisabsent,itisusefultobeawareoftworegionswhereKtakesoncharacteristicvalues(apartfromK=0outsidetheplume).fIOnesuchregionisthevicinityoftheexhaustport.TheexhaustmixturecrossestheplaneoftheportthroughexhaustareaAwithuniformvelocityVandconcentrationC,carry-ingcontaminantflowQ,allrelatedbyeeeC=Q/AV(4)ThecorrespondingKvalueLsfoundbysettingCinEquation2equaltoCinEquati,on4toobtaineK=(A/A)(U/V).(5)Forexample,inFigure1themodelbuildingwasa0.38m(15in)cubewithA~0.145m22(1.56ft)2theexhaustportwasa0.0127m(0.5in)demetercirclewithA=0.000127m(0.00136ft),and0andVwereeachequalto1.22m-s(4fps),yieldingK=1,146.ThesameconfigurationforaPull-scalebuildingintheatmospherewouldhavethRsameK,sinceAwouldbescaledbythesamefactorasAandUbythesamefactorasV.ItshouldbenotedthatKisafunctionofthebuilding/portarearatioandwind/exhaustvelocityratLo.IfeitheroPtheseshouldchangersKwouldalsochangeandtheFLgure1iso-plethswouldnolonger,beexactlytransferabletothePull-scalebuilding.However,ifthechangestillresultedincompletecaptureofthecontaminantinthecavity,thechangesLntheisoplethpatternwouldbelocalizedtotheexhaustvicinity.AnotherregionwhereKtakesonaspecialvalueisatthedownwindendofthecavity.Despiteasmallvari.ationofconcentrationwithheight,whichisresponsivetoexhaustportconditions,afairlyuniformaverageconcentrationCiscreatedintheswirly,hLghly-turbulentcavityflow.CisgivenbywC=Q/A0(6)whereAisthemaximumcross-sectionareaofthewakenormaltothewinddirectionandUistheaveragewindvelocityinthecontaminatedregionJustdownwindofthecavity,andKfol-lowsbycombiningEquation6withEquation2:Ka(A/A)(U/U),(7)
HeasurementsshowthatA2AandanaverageUU/3;Equation7thenyieldsK=1.5.ThatthisisarealisticviuemaybeseeninFigure1;Kisabout1.5to2overthedownwindwwtaceofthecubewheretheexposureismainlytowell-mixedcavityconcentrationsinthereturnflow.AlthoughFigure1isspecifictoacubicalbuilding,similarKpatternsappearwithblockbuildingshavingdifferentproportionsbutallhavingroofexhausts.Forexample,threesetsofKisoplethsforablockbuildinghavingsidesintheratio1:3:3andwithdifterentfacespresentedtothewindareshowninReference4,Figures20,21,and22.AllhaveKvaluesofabout1.5attheleeface,althoughsmallervaluesappearinthelowerhalfofthetallnarrowbuildingbecauseofstronghorirontalinfusionoffreshairneartheground.(NotethattheK(=K)valuesinFigures20and21areincorrect;theyshouldbeinterchanged.)emaxInReference6,forblockbuildingsinaboundarylayer,isoplethsatthedownwindwallhaveaboutthesame'veragevalueof1.5asinReference4,althoughtherange(tromlargeattherooftosmallattheground)isgreater.Thedifferenceisduetothedifferenceinconfi-gurations.TheReterence6buildingshadverysmallexhaustportsandwereimmersedinadeepboundarylayer.ThewindstreamseparatedattheupstreambuildingedgesbutreattachedtotheroofandwallsandseparatedagainatthedownwindedgestoformaleecavitythatwassmallerthantheReference4cavity,withA"A.ThisalonewoulddoubleCandK.However,becausethereleasewasinthesmoothflow,someofthecontaminantdiffusedupwardbeforereachingthewwwcavityandwasnotrecirculatedtotheleewall.ThelossofcontaminantandthereducedAcontributedinoppositewaystofortuitouslyproducinganaverageKthatwasaboutthesameinwbothtests.Roundbuildings,suchasnuclearreactorcontairmentstructures,producesmallercavitiesthansharp-edgedbuildings;therefore,Kshouldbelarger,accordingtoEquation7.Reference5,Figure5.29c,showsKaveragingabout3attheendotthecavity(about2.25diametersdownwindofthebuildingcenter.TheprevalentappearanceotanaverageKot1.5attheleefaceofasharp-edgedbuildingisapowerfulgeneralixationthatmaybeusedindesignofwallandgroundintakeswhentheexhaustisontheroof,butafewwordsofcautionarewarranted.TheexistenceofK1.5>stemmingfromEquation7,impliesthatallofthereleased.contaminantistrappedinthewakeandtlowsdownwindthroughA.Thisistrueforexhaustswhosejetvelocity,diameter,andelevationareinsufficienttothrustthecontaminantthroughthecavityboundary.Figure1isanexampleofthiscondition.However,large-diameterhigh-velocityjetsfromstubstacksonroof-rountedfansoftenhavesufficientmomentumtopenetratetheboundary,allowingsomeofthecontaminanttoescapeandleavingthebalancetocreateareducedQ,saytQ,whichcreatesTheestimationoffisbeyondthescopeofthispapersinceitrequiresfamiliaritywithteinteractionofjetplumeswithroofcavityflow.AppendixAprovidessomecommentsonthissubject.Roofdispersionpatternsarecontrolledbytwofactors;thepresenceorabsenceofaroofcavityattheexhaustportandthestrengthoftheexhaustjet.Ifacavityispresentandthejetisweak,thepatternissimilartoFigurel.Ifacavityisabsent,osifthereisacav-ityandthejetisstrongenoughtopenetratetheboundary,thepatternisthatofaplumefromashortstack.Aroofcavitywillbecreatedwheneverasharprootedgeispresentedtothewind.Ittheedgeisnormaltothewind,thecavitywillextendtheentirewidthottherootandpartorallotitslength,dependinguponthebuildingproportionsandtheapproachwindvelocityprofile.Ittheedgeisatanangletothewind,thecavitywillcoveraportionottheroofcontiguoustotheedge,thesireofthecoveragedecreasingwithgreaterdepartureottheedgeanglefromnormal.Theportionofaroofnotcoveredbyacavitymaybeconsideredasaregionotsmoothflowinthedirection,andatthevelocityot,theapproachwind.Suchregionsoccurinnormalorientationdownwindotthecavitywhenthebuildingproportionsandapproachwindvelocityprofilearesuchastocreateflowreattachment.GuidelinestorestimatingthereattachmentregionmaybetoundinReferences7and9.Smoothflowregionsalsooccuratthecenteroftherootinorientationsotherthannormal.Figure2(abridgedfromReference4)showsKisoplethstoracubein45orientation;thesmoothflowregionliesbetweentheKs0isopleths,andthecavitiesareatthelateralcorners.Theplume,whichiswellformedoverthebuilding,des-cendsrapidlyatterpassingthedownwindcornerandiswhollycapturedinthecavity.The
averageKontheleefaceLsabout3,owingtolocaldownwarddiffusionofhighplumeconcen-trations,whichaugmentthemoredLffuseconcentrationsLnthereturnflow.Inevaluati,ngaproposeddesign,thewinddirectionshouldbeallowedtorotate'through360andthelocationsofexhaustportandintakeobservedwithrespecttoroofcavities.Inanyonedirection,ifbothexhaustplumeandintakeareLnthecavity,aKvalueattheLntakeshouldbeselectedfrcmamongthepublishedKisoplethsandadJustedforthedifferencebetweenthedesignKandthepublishedK.NotethatthisadJustmentLsgreatestnearthesource,ewhereKisdependentonAand5,andwillnotbenecessaryattheleewallunlessthefrac-tionof5retairedinthecavitycPianges.Iftheexhaustplumeandtheintakeareoutsidethecavity,dispersionshouldbecalculatedasintheopenatmosphere.AnambiguityintheforegoingdiscussionisthecriterionforestablishirgwhenanexhaustplumeLsfullytrappedinthecavityandwhenitcanbeconsideredtohaveescapedintothefreestream.AppendixAofferssomesuggestionsinthisregard.AnestimatewasmaderecentlyofdLspersionofcontaminatedairreleasedinseveralpossiblemodesnearthesurfaceofthereactorenclosurebuildingofanuclearpowerplant.Theesti>>matewasfollowedbyawindtunneltestofthesameplant.TheestimateandthetestresultsprovideanopportunitytoillustratehowtheKconceptisappliedLnpractice,toevaluatethepredictivetechniqueingereral,andtodemonstratehowseeminglysmalldeviationsbetweentheconceptualandrealconfLgurationscanproducesignificantchangesintheconcentrationfield.Thefacilityhastworeactors,designatedUnits1and2,eachinitsownenclosurebuild-ingbutservedbyacommoncontrolbuildingandservingacommonturbinebuilding,allJoiredtoformtheirregularbuildinginthecenterofFigure3.Twolargenaturaldraftcoolingtowers(N),twomechanicaldraftcoolingtowers(M),anelectricalswitchyard,andvariousbuildingssurroundthecentralstructure.Theventilationexhaustsystemforeachunithastwointernalpathways,onlyoneofwhichwillbeLnuseatagiventime.Eachpathwayterminatesinasmalllouveredpenthouse,desig-natedinboardvent(IV)oroutboardvent(OV),ontheroofoftheunit'sauxiliarybuilding.AdesignalternativeunderconsiderationwasastackreleasewhoseportwasattheelevatLonoftheroofoftheenclosurebuilding.Afourth(accidental)releasemodewasseepagethroughtheexteriorwallsoftheenclosurebuilding.Thereceptorwasconsideredtobeinthecontrolroomintheinteriorofthecontrolbuilding.Thecontaminantcouldenterthecontrolroomviafreshairintakesspanningthewestwallofthecontrolbuildingatanelevationof13.7m(45ft)aboveground,or,iftheintakesareclosed,byinfiltrationthroughtheroofofthecontrolbuilding.Anestimateofconcen-trationatthecenteroithewallintake(B)andatthecenteroftheporousroofarea(D)wasdesired.TheprototypewindconditionwasanaturalboundarylayerLnneutralstabilitywithU1.52m-s(5fps)atanemometerheightof58.5m(192ft).Thiscorrespondstoawindspeedof1ms(3.3fps)atthe10m(32.8'ft)elevation,aconventionalassumptionfornuclearreactoraccidentdispersioncalculations.(TheatmosphericstabLlityusuallyassumedinsuchcalculationsisstronglystable,withaconsequentdifferenceintheapproachwindturbulenceandmeanvelocityprofile.However,thewinddisturbancecreatedbythebuildingoverwhelmstheapproachwindcharacteristics,makingtheassumptionofneutralstabilityvalid,)InselectingtheappropriatesetofKisopleths,itisnecessarytoevaluatethecontrol-lingfeaturesof'heflowbetweensourceandreceptor.BothIVandOVforeachunitarelocatedatthesurfaceofalargecompositestructureconsistingoftheunit'scombiredenclo-sureandauxiliarybuildings.Thesmalllouveredpenthousesovertheexhaustportsdestroyany'pwardmomentumLntheexhaustJets,ensuringthatthereleaseisatthesurfaceofthestruc-ture.ReceptorsBandDarelocatedaboutonestructurediameterfromitscenter.By Lll(0l' inspection,thecriticalwinddirections(producinghighestreceptorconcentrations)appeartobeMSMforUnit1andNNEforUnit2.Thereceptorsareintheleeofthestructureinthesewinddirections.Theeft'ectofthefourcoolingtowers,whosewakesandinternally-generatedexternalaircirculationsaltertheapproachwindcharacteristicsinacomplicatedmanner,arenotconsideredintheestimate.TheconfLguratLonsuggeststhat,Lneachwinddirection,thereceptorsareatthebottomofalargebuildingcavitywhichiscontaminatedbyreleasesfromsmallsurfaceports.TheisolatedcontainmentstructuretestsofReference2fulfillthesecriteria.Horeover,Refer-ence2providestheonlyavailablesetofisoplethsinspacedownwindofabuilding.Figure6isareproducti.onofoneofseveralK-LsoplethdrawingsinRefererce2;itrepresentsanupwindmid-heightreleaselocation.TheReference2building,identifiedastheEBR-ZZcontairwentstructure,isacircleinplanview.CircularapproximationstotheUnit1andUnit2compositestructuresareshownasdashedcirclesof61.0m(200ft)diameterinFigures4and5.(TheorLentationofthesefig-uresissuchthatthewindpassesfromlefttorightacrossthepage.)Theupperpartsofthefiguresareelevationsectionsthroughthecentersofthecircles,withtheEBR-IZbuildingshowninitscorrectproportions.TheplacementoftheEBR-IZbuildingwassuchastomatchthema)ordimensionsofthecompositestructuresLnplanviewandatthetopinelevation.ThisresultedinanEBR-IZbaseat16.8m(55ft)belowplantgrade.TheKLsoplethsinReference2arepresentedassevendrawings,eachcorrespondingtoasmallsurfacesourceatadifferentlocation:top;midheightupwind,side,anddownwind.Theappropriatesetforusehereinwaschosentoprovidematchi.ngsourcelocations.ForUnit1LnaMSWwind(Figure4),theventsclearlyareattheupwindlocationbuttheheightisuncer-tain.ForUnit2inaNNEwind(Figure5)theventsaremidwaybetweenthesideanddowrwindlocationsand,again,theheightsareuncertain.ToestablishtheeffectiveventheightLnproportiontothestructureheight,LtLsimpor-tanttoconsidertheeffectivegradeintheregionofthecavity.Thegradeestablishesthefrontalcross-sectionareaA,whichcreatesthecavitycross-sectionareaA.BoththeseareascontributetotheestablishmentoftheaverageKatthedownwindendoft5ecavityfromwhichthereturncavityflewtowardreceptorsBandDarises(seediscussioninconnectionwi.thEqua-tions6and7).TheeffectivecavitygradeisshowninFigures4and5.Itwascalculatedbyanaveragigalgorithmthattookintoaccountroofelevationsalongthesectionlineaswellasat~22.5inaxLmuthfromthesection1Lneasrepresentativeoftheaverageheightofthecavitybase.ForUnit1-MSWtheeffectivecavitygradewas1).3m(60ft)aboveplantgrade,andtheEBR-IInetAabovecavitygradewas22017m(21,$13ft).ForUnit2-NNE,thecorrespondingvalues2were7.6m(25ft)and2,667m(28,710ft).TheheightoftheventsinproportiontothebuildingheightabovecavitygradeLnUnit1-MSMconfLgurationisshowninFLgure6.ZtcorrespondswelltothemidheightEBR-IZsourceheight.SimilaragreementLsobtainedintheUnit2-NNEconfiguration(notshown),althoughtheventsareslightlyabovethesourceduetothelowereffectivegrade.ItremainstobeestablishedthattheKisoplethsformidheightsourcesinReference2canbeusedinthepresentcaseinviewoftheviolationoftherequirementofgeometricsimilarLty,i.e.,thebuildLngheight/buildingdiameterratiosaredifferent.TheonlyargumentIcanofferistheobservationpreviouslymadethattheaverageKvalueofthelcewallforblockbuildingshaveheight/widthratiosof1and1/3isabout1.5forboth,andthedistributionsoverthewallfromtoptobottomaresimilar(higheratthetop,loweratthebottom;seeReference4,Figures20and21).Byanalogy,averticalcontractionoftheEBR-II.LsoplethsinthesameproportLonastheheightcontractionshouldbepermitted.Itwasrotnecessarytore-drawtheisoplethstoacontractedverticalscaleaslongasthereceptorswereplacedinthecorrectverticalrelationinFigure6.TheplacementofBandDhorixontallyinFigure6wasdonestraight-forwardlybydividingrealdownwindandcrosswinddistancesbythescaledEBR-IZdiametertoobtainthex/Dandy/Dcoordinates.Vertically,bothreceptorswereplacedLnthebaseplane,eventhoughthepropertionateheightprocedureputBbelowgradeandDaboveintheUnit1-MSMconfigurationandbothaboveintheUnit2-NNEconfiguration.TherationaleforthisinthecaseofDwasthatiso-plethsinthegroundplanedonotchangeduringverticalcontractionofthefield,and,sinceD
wasapointintheroofplane,itwasmorereasonabletoleaveitinthegroundplarethantosetitaboveamathematicalrigidity.ZnthecaseofB,alocationbelowgradeismeaninglesssincenoisoplethsexistthere.ThepresenceofUnit1inthecavityofUnit2intheNHEwinddirectioncreatesacomplexinterferenceflowpatternwhoseeffectontheisoplethpatterncouldnotbepredicted;itwasignoredfortheestimate.TheinterpolatedvaluesofKatBandDforIVandOVsourcesaregivenintheupperpartofTable1.TheUnit1-WSWvalueswereobtainedfromFigure6.TheUnit2-HNEvalueswereobtairedfromsimilarfiguresbasedonReference2,Figures11and14'tthewall(B),Krangedfrom2.5to3.4.Attheroof(D),Krangedfrom2.3to2.8.'valuesforthestackreleaseswerenotestimatedbecauseofuncertaintyastohcwmuchofthereleasedQwouldenterthecavity.Fractionalcapturewasthoughttobequitelikelysincethestackreleaseelevationwas12.5m(41ft)higherthanthevents,andthestackswereuncapped.KvaluesfortheseepagereleasearegiveninTable2.TheywereobtainedbyaveragingtheKvaluesatBandDfromallsevenofthedrawingsofReference2,withthesidereleasedrawingsyieldingtwovalues,oneforthereceptorsonthesamesideasthesourceandtheotherforthereceptorsontheoppositeside.Therationaleforthisprocedureisthataseepagereleasewouldoccurovertheentireexteriorsurfaceoftheenclosurebuilding.Theaveragingprocedureoverallsourceswasconsideredtobethebestrepresentation.TheaverageKrangedfrom5.9to10.1.Theestimatingproceduresdescribedabovedifferfromthoseinthepre-testestimateintworespects.OneisintheplacementofDatthebaseinsteadofattheproportionaleleva-tion;thisproducedonlyaminorchangeintheestimate.Theotheristheuseofanire-pointaveragefortheseepagereleaseinsteadofasinglereleaseatthedownwindmidheightsource.ThischangewasmadebecausethewindtunneltestshowedhighervaluesofKbyafactorof2to3thanwereoriginallypredicted.Hindsightsparkedtherealisationthatuseofamidheightsourcealoneomittedtheimportantbasereleasecontribution.Thetestwasconductedinneutralstabilityona1/240scalemodelwithtunnelwindvelocityUa3.1m-s(10fps)atthe585m(192ft)elevationandallothervelocities(sourceandcoolingtowers)doubledtomaintaintheircorrectrelationtothewindvelocity.TheNtowersweremadeoperationalwithinternalaxialflowfans,drawingtunnelairinattheirbasesanddischargingitthroughtheirtops.TheMtowersweremadeoperationalbyanexternalcompres-sorthatprovidedthecorrectoutflow,buttheairsourcewasoutsidethetunnel.Thetestprogramprovidedconcentrationmeasurementsatatotalof42tapsinthewestwallandroofofthecontrolbuildingforeightreleaseconfigurations(IV,OV,stack,andseepageontwounits)and16winddirectionswithMandNonandoff(allcombinationswerenottested).SometestswerealsodonewithUnit2removed.Thetestresultswerereporte),inaccordancewithconventionalpracticeinnuclearplantevaluation,asfull-scaleCU/Q(m).KwasfoundbyapplyingEquation2toobtainKA(CU/Q)tt(8)withAasgivenpreviouslyfortheWSWandNNEdirections.ToprovideKvaluesforcomparisonwithestimates,fourwalltaps(Nos.15-18)wereaver-agedtorepresentB,and16rooftaps(Hos.27-42)wereaveragedtorepresentD.ThelowerpartsofTables1and2showtestvaluesforBandDi,ntheUnit1-WSWandUnit2-HHEconfi-gurations.
IntheUnit2-NNEconfiguration,themaximumtestRforaventreleaseoccurredatDwithNandMoff;itsvaluewas2.6.TheestimateforthesereconfLgurationwas2.3.FortheseepagereleasewithNandMoff,thetestvaluewas11.8;thecomparableestimatewas9.7.Theagreementisgood(althoughtheestimatingprocedurewaschangedinretrospect,asdis-cussedpreviously).Ontheotherhand,verypooragreementwasfoundintheUnit1-NSNconfiguration.ThemaximumtestKforaventreleasewas0.1,whereastheestimatewas2.8.ThemaximumtestKfortheseepagereleasewas16.6atDwithNonandMoff;theestimatewas5.9.Themagni-tudesofthediscrepanciesandthefactthattheyoccurredinoppositedirectionswarrantsanattemptatexplanation.IntheUnit2-NNEconfiguration,theventswereatthepointofflowseparationatthesouthendoftheauxiliarybuildingroof.Thisensuredcompletedescentofthereleaseintothecavity,andtherearpresenceofUnit1intheleeofUnit2createdablockagethatstrengthenedthecavitycirculation.Thesetwofactorsprovidedforcompletecaptureofthereleaseinthecavityflow,producingtheexpectedvalueofK.IntheUnit1-NSNconfiguratLon,theventswereattheupwindcornerofadLagonally-orientedbuildingandtheplumedevelopedverticallyupwardLntraversingtheroof,therebyplacingsomeoftheeffluentabovethecapturezonedownwindoftheenclosurebuildingroof.Second,Unit2wasLnapositiontodeflectsomeof'hewindpassirgaroundthewestsideofUnit1intothecavityregion.Third,theturbinebuLldinginterceptedthedownflowattheendofthecavityandturnedLtdownwindinsteadofbacktoUnit1.Thecombinationoffrac>>tionalescapeofthedevelopingplume,windinfectionbyUnit2,anddownflowinterceptionbytheturbinebuildingresultedinalmostnoportionofthereleasereachingthereceptors.Fortheseepagerelease,thelow-levelwLndflowpatterninthecavitywasalteredbythepresenceoftheleeunitandtheturbineroomtoproduceaflushingactionatDfortheUnit1-NSNrelease,therbycreatinglargervaluesLnthelatterconfiguration.TheIVreleasesproducedhigherKvaluesthandidtheOVreleases,indicatingsomelossofQpriortoentryoftheplumeintothecavityinthelattercase.Therefore,theOVlocationispreferable.ThestackproducedsignLfLcantlylove!Kvalues.Themaximumwas0.7atDforUnit2-NNE,comparedto2.6fortheIVrelease.ThisLsattributabletothelargerescapefractionwiththehLgherreleasepoint.OfconsiderableinterestisthelowerKvaluesatwallintakeB.Thetestmaximumwas1.0comparedto2.6attheroof.ThisisattributabletothelateralLnflowofuncontaminatedwindalongtheground,strikingthewallandflowingupward,therebypreventingthecontaminatedcavityflowatDfromdescendingtoB.TheeffectofthecoolingtowersappearedintheNNEorientation,asexpectedfromtheirupwindplacement.OperationofNreducedthemaximumKatDfrom2.6to1.8,withsimilarreductionsfortheotherventandthestack.OperationofMreducedKfrom2.6to0.2.Thesereductionsareattributabletowinddisturbancescreatedbythecoolingtowerplumes.ThecirculationLnatransverseJetplumeisapairofcounter<<rotatinghelicalvortices,upalongtheplumecenterlLneanddernoneachside,andlongitudinallydownwind.SinceNandMstraddleUnit2intheNNEorientation,eachplumeproducesadownflowattheunit.InadditLon,Npro-ducesaflowtowardtheeastneartheground,whileMproducesaflowtowardthewest.ThehelicalcirculationsalteranddisplacetheUnit2cavitysothatthereceptorsareinregionsoflowerconcentrationthaninanundisturbedNNEwind.TheeffectLsmorepronouncedwhenthewindvelocityislow,asinthetest.TheinfluenceofMisstrongerthanthatofNbecause,theMplumeisclosertotheground(exitportsatelevation17.4m[57ft])whiletheNplumeishigh(originatingaselevation121.9m[400ft]).TheconcentrationcoefficienttechniqueisgenerallybelLevedtobequitoaccuratef'rpredict-ingfull-scaleprototypeconcentrationsfromscale-modeltestmeasurementssincesuchtestsaredesignedtoprovideexactsimilarityintheimportantshapeandflowcharacteristics.ItalsoperformswellLnprovidingestimatesofKincasesofnonexactsimilarity,provLdedthatthereleasecanbeassociatedwithanisolated,cavity-producingstructureforwhLchKisoplethsareavailable.
Themethodislessaccuratewhenthereleasepointisnearthecavityboundary,sincesmallchangesindistancefromthebuildingsurfacemaycreatelargechangesinthefractionoftheplumecapturedbythecavity.Itisalsolessaccuratewhenwindflowaberrationsareintroducedbynearbyportionsofthestructureorbymoredistantstructures,suchascoolingtowers,whichgeneratepersistenthelicalcirculationsforlongdistances.ItseemsunlikelythatKisoplethsforconfigurationshavingthesevariationswillbeavailableinthenearfuture.Thealternativesforthedesigneraretoimproveone'sabilitytoextrapolatebeyondtheavailabledatabystudyofthereferences,ortobeconsoledbytheobservationthatinthepresentteststhemaximumincreaseabovetheestimatedvaluewas41$(K=11.8increasingto16.6fortheseepagereleaseintheUnit2-NNEconfiguration).Indispersioncalculations,afactorof<2isconsideredacceptable.1~Halitsky,J.1963."Gasdiffusionnearbuildings."NewYorkUniversity,DepartmentofMeteor.andOcean.GSLRep.63-3,Feb.1963.(Loancopyavailablefromauthorfordupli-cation.)2.Halitsky,J.;Golden,J.;Halpern,P.;andWu,P.1963.'Windtunreltestsofgasdiffu-sionfromaleakintheshellofanuclearpowerreactorandfromanearbystack."NewYorkUniversity,DepartmentofMeteor.andOcean.GSLRep.63-2,April1963.(Loancopyavailablefromauthorforduplication.)3.Wilson,D.J.1976."Contaminationofbuildingairintakesfromnearbyvents,"Universi-tyofAlberta,DepartmentofMechanicalEngineering,Rep.1,May,1976,Edmonton,Al'berta,Canada.4.Halitsky,J.1963."Gasdiffusionnearbuildings."~~~~ig,Vol.69,pp.464-485.5~Halistky,J.1968."Gasdiffusionnearbuildings."Qt)JU1KGlggy~~EtlmLY39K.iD.H.Slade,ed.,U.S.AtomicEnergyComm.Div.ofTec.Inf.,pp.221-255.(AvailableasTID-24190fromNTIS,U.S.Dept.ofComm.,Springfield,VA22161.)6.Wilson,D.J.1976."Contaminationofairintakesfromroofexhaustvents."JZR1UU~NUA,Vol.82,No.1,pp1028-1038.7.Hosker,R.P.,Jr.1984."Flowanddiffusionnearobstacles."U.S.DepartmentofEnergy,OfficeofSci.andTech.Inf.,Chapter7.(AvailableasDOE/TIC-27601fromNTIS,U.S.Dept.ofComm.,Springfield,VA22161.)8.Hosker,R.P.,Jr.1981."Methodsforestimatingwakeflowandeffluentdispersionnearsimple,block-likebuildi.ngs."NOAAEnv.Res.Labs.,AirResourcesLabs.,SilverSpring,MD,TechMem.ERLARL>>108.9.ASHRAE.1985.fiat~~~~ujnm~,Chapter14.NewYork:AmericanSocietyofHeat-ing,Refrigerating,andAir-ConditioningEngineers,Inc.10.Halitsky,J.1966."AmethodforestimatingconcentrationsintransverseJetplumes,"Vol.10,pp.821-843~ThesponsorofthecasestudyinvestigationhasmadetheKestimatesandwindtunneltestresultsavailableforthispresentationintheinterestof'isseminationofscientificdata,buthasdeclinedtobeidentified.ThewindtunneltestswereconductedattheFluidDynamicsandDiffusionLaboratory,ColoradoStateUniversity,FortCollins,CO80523.
44 APPEHDIZAPenetrationofaJetPlumeThroughaCavityBoundary10ThefollowingdiscussionisbasedongetplumepropertiespresentedindetailinReferencePeretrationofacavityboundarybyagetplumeisamuchunder-investigatedsub]cot.PlumeriseincreaseswithemissionvelocityratioV/UanddecreaseswithrateofinfusionofcavityairintotheexhaustJet.IncomparisonwithPreestreamconditions,thelocalcavitywindisslowerandmoreturbulent.TheslowervelocitycreatesahigherV/Uandinducesgreaterplumerise.Thehigherturbulenceandwindshearproducemorerapidinfusionofcavity8airintotheJet,producinglessrise.Thetwoeffectstendtocanceleachother.Iknowofnoanalyticalprocedureorexperimentaldatatoquantitythesetrends.Isuggestthattheplumerisebecalculatedasifthecavitywerenonexistentandaseparateestimatemadeofthecavityboundary.Thoplumewouldthenbeconsideredtohaveescapedthecavityifthecalculatedplumecenterlineliesatleastoabovethecavityboun<<dary,andtohavebeencompletelycapturedifthecenterlineliesbelow,atalldistancesuptothereceptor.Thisisacrudecompromisebasedontheconceptthatmostoftheplumewilldes-cendintothecavityduetothehighcavityturbulencewhenthecenterlineliesalongthecav-ityboundary,andmostoftheplumewillescapeifthecenterlineis2aabovetheboundary.Z
TABLE1KValuesforIVandOVReleases~MZL~~&KMallRoof'allRoofupwindmidheightdownwindmidheightopp.sidemidheightfAverage911142.52.83.82.52.3inboardventIVinboardventIVinboardventIVinboardventIVoutboardventOVoutboardventOVoutboardventOVoutboardventOVstackstackstackonoff0offoffononoffononoff0offoff0ononoffononoff0offoffonon0.10.10.10.10.31.81.02'0.60.20.50.I0~3130.82.10.20.400.20.IOejTABLE2KValuesforSeepageReleases~~2LMnQ~MallRoofMallRoofSuucmupwindPasoupwindmidheighttopdownwindmidheightdownwindbasesourcesidebaseopp.sidebasesourcesidemidheightopp.sidemidheightAverage891011121313141420672033322552344219142517351015355543~34333?10.15.97.99.7onoffoffoffonon016.60.112.81+711.82.75~9
0n1100522K~i.15V/U10/H~1/30K~1146Figure).Kisoplethsforacubeinnormalorientationtothewind.(AbridgedfromReference4,Figure16.)MIND01021K~V/U~10/H~1/30a1146Figure2.Kisoplethsforacubeatb5orientationtothewind.(AbridgedfromReferenceb,Figure19.)
'NATURALDRAFTCOOLINGTOWER2NMECHANICALDRAFTCOOLINGTOWERSM8SCALE-FTO2SOSnoAUXILIARYBUILDINGENCLOSUREBUILDINGTURBINEBUILDINGELECTRICALSWITCHYARDUNIT2UNITISTACKIVOVPLANTTRUE'NWINDNATURALDRAFTCOOLINGTOWERINNNEWINDPiguce3.Planviewofbuildingatcasestudysite.
STACKVENTS-WSMMIND~EBRIIEBRIIANEMON.41190+137B45345~PLANTANEMONE19260IPLANTGRADE100R+AVERAGECAVITYGRADEEVATIONSECTIONA-AELUNIT2UNITIro)0LENGTHSINFTFigure4.LocalconfigurationforreleasesfromUnit1inaMSMwind.
t0C' EBRII-NNEMIND~STACKVENTSAVERAGECAVITYGRADEIPLANTGRADEVERTICALSECTIONB-8UNIT2UNITILENGTHSINFTFigure5.LocalconfigurationforreleaaesfromUnit2inaNNEwind.
=