• Aucun résultat trouvé

Mass transfer and hydrodynamic characteristics of new carbon carbon packing: Application to CO2 post-combustion capture

N/A
N/A
Protected

Academic year: 2021

Partager "Mass transfer and hydrodynamic characteristics of new carbon carbon packing: Application to CO2 post-combustion capture"

Copied!
12
0
0

Texte intégral

(1)

This is an author-deposited version published in:

http://oatao.univ-toulouse.fr/

Eprints ID: 6133

To link to this article

:

DOI:10.1016/J.CHERD.2010.09.023

URL: http://dx.doi.org/10.1016/J.CHERD.2010.09.023

To cite this version

:

Alix, Pascal and Raynal, Ludovic and Abbé, François and

Meyer, Michel and Prevost, Michel and Rouzineau, David (2011) Mass transfer

and hydrodynamic characteristics of new carbon carbon packing: Application

to CO2 post-combustion capture. Chemical Engineering Research and Design,

vol. 89 (n°9). pp. 1658-1668. ISSN 0263-8762



O

pen

A

rchive

T

oulouse

A

rchive

O

uverte (

OATAO

)

OATAO is an open access repository that collects the work of Toulouse researchers and

makes it freely available over the web where possible.

Any correspondence concerning this service should be sent to the repository

administrator:

staff-oatao@listes.diff.inp-toulouse.fr



(2)

Mass

transfer

and

hydrodynamic

characteristics

of

new

carbon

carbon

packing:

Application

to

CO2

post-combustion

capture

Pascal

Alix

a

,

Ludovic

Raynal

a

,

Franc¸ois

Abbe

b

,

Michel

Meyer

c

,

Michel

Prevost

c

,

David

Rouzineau

c,∗

aIFP,BP.3,69360Solaize,France

bSnecmaPropulsionSolide,GroupeSafran,Lescinqchemins33187LeHaillan,France

cUniversitédeToulouse,INPT,ENSIACET,LaboratoiredeGénieChimique(UMR5503),4alléeEmileMonso,BP84234,31432Toulouse, France

a

b

s

t

r

a

c

t

Anovelstructuredpacking,the4Dpacking,hasbeencharacterizedintermsofhydrodynamics,effectiveareaand gassidemasstransfercoefficient.Theincreaseofthe4Dopeningfractionallowstoreducepressuredropandto getabettercapacitythanMellapak500Yand750Y,forwhichthegeometricareasaresimilar.The50%open4D packing,4D-50%,leadstoeffectiveareaswhicharehigherthanMellapak500Yones,anddoubledcomparedwith MellapakPlus252Yones.Effectiveareasforthe4Ddonotdecreasewhentheopeningfractionincreasesfrom30to 50%,thisindicatesthatanon-negligibleamountofdropletsisgeneratedat50%.Gassidemasstransfercoefficient hadbeenmeasuredwithanoriginalexperimentalmethod:waterevaporation.Correspondingresultsseemtobe inagreementwiththeliterature,andwiththefactthatalargeamountofdropletsisgenerated.Correlationsare proposedforbotheffectiveareaandgassidemasstransfercoefficientforthe4D-50%.

The4D-50%packingcouldbeveryinterestingforpost-combustionCO2capturesinceitgenerateslowpressure

dropandaveryhighinterfacialarea.Thiswillbefurtherconfirmedbyaneconomicstudyforwhichtheabsorber plantwillbedesignedwitharatebasedmodel.

Keywords: Newpacking;Post-combustioncapture;Absorption;Pressuredrop;Masstransferefficiency

1.

Introduction

The captureand geologicalstorage of the CO2 emitted by

power plants is one important way to reduce greenhouse

gasesemissions.Hugegasflowratesmustbetreatedfor post-combustioncaptureofCO2,whichleadtoverylargecapture

amineplants.Theoptimizationofsuchhighvolumereactor designisthusofgreatimportanceinordertoreduceinvest costs.Sincecaptureprocessoperatesdownstreamthepower plant,itrequires verylowpressuredropinordertoreduce theboosterfanelectricconsumption,operatingcostsand effi-ciencylost.

Tominimizevolumereactorandpressuredrop,very effi-cientgasliquidcontactorsareneeded.Intheframeworkof

Correspondingauthor.

E-mailaddress:david.rouzineau@ensiacet.fr(D.Rouzineau).

GASCOGNE project, supported by ANR, a novel structured

packinghasbeenconsideredforchemicalengineering stud-ies,the“4Dpacking”.Thelatterhadbeeninitiallydeveloped and ismanufactured bySPS. Tobuild-upmodels, testsare highlyneededtocharacterize4Dpackingsintermsof hydro-dynamicsandmasstransfer.Theaimofthepresentstudyisto determinethepressuredrop,theeffectivearea,ae,andthegas

sidemasstransfercoefficient,kG,forthreegeometric

config-urations.SinceoneassumesthatMEA30wt%isthebasecase fortheprocess(Knudsenetal.,2006;Feronetal.,2007),onecan considerthatfastreactionswilloccurincaptureplants.aeand

kGbecomethemainparameterstoestimatetheefficiencyof

anabsorber(Danckwerts,1970).Ithastobenoticedthat exper-imentshavebeenconductedwithbothIFPandLGCfacilities.

(3)

Nomenclature

AB distancebetweentwocrossingsoffibres

Ac columncross-section(m2)

ae effectiveoreffectiveareaofthepacking(m−1)

ae,4D effective oreffective areaofthe 4Dpackings

(m−1)

ae,void effectiveoreffectiveareaoftheIFPempty

col-umn(m−1)

ae,global effective or effective area ofcolumn section

(m−1)

ag geometricarea(m−1)

C0

CO2 CO2 molar concentration in the liquid bulk

(molm−3)

Cg concentration of gas phase needed for kgae

measurement(moll−1)

Cl concentrationofliquidphaseneededforkgae

measurement(moll−1)

C0

OH OH−ionsmolarconcentrationintheliquidbulk

(molm−3)

C0

Na Na+ionsmolarconcentrationintheliquidbulk

(molm−3)

d columninnerdiameter(m)

DCO2 CO2 diffusion coefficient in the liquid phase

(m2s−1)

Dm diameterofthemandrel

dp drippointdensityoftheliquiddistributor(m−2)

dz heightofanelementofthecolumn(m)

E enhancementfactor

FG

CO2 CO2gasmolarrate(mols −1)

FL

CO2 CO2liquidmolarrate(mols −1) FS gasF-factor=√G×VsG(Pa0.5)

FS|fl gasF-factoratthefloodingpoint(Pa0.5) G gasflowrate(kgh−1m−2)

G0 airflow(ms−1)

h,hg,CO2,etc. contributionsofacation,ananion,anda

gas,respectively(lmol−1)

H column’sheight(m)

He Henryconstant(Pam3mol−1)

Hw Henryconstantforwater(Pam3mol−1)

I ionicstrengthofsolution(moll−1) k2 kineticconstant(m3mol−1s−1)

k∞

2 infinitely diluted solution kinetic constant

(m3mol−1s−1)

kg gassidemasstransfercoefficient(ms−1)

KG overall gas side mass transfer coefficient

(molPa−1m−2s−1)

kL liquidsidemasstransfercoefficient(ms−1)

Lf widthofcarbonfabric

Nf numberofspindle

P pressure,neededforkgaemeasurement(Pa)

PCO2 CO2partialpressure(Pa) Ptot totalpressureofthecolumn(Pa)

QL liquidload(m3m−2s−1)

S channelsideofthestructuredpacking(m)

T temperature(K)

TAS temperatureatadiabaticconditionsof

satura-tion(◦C)

TGin,TGout inletandoutletgastemperatures(◦C)

TLin,TLout inletandoutletliquidtemperatures(◦C)

Tws waterstoragetemperature

VsG superficialgasvelocity(ms−1)

Y absolutehumidity, needed forkgae

measure-ment

YAS absolutehumidityatadiabatictemperatureof

saturation

Yin,Yout absoluteinletand outlethumiditiesofairin

column

yx molarfractionofthecomponentxinthegas

phase(mol/mol)

z axiscolumnposition(m)

1P/L pressuredropperunitlengthofthepackedbed (Pa/m)

1Ptot totalpressuredropofthecolumn(Pa)

Non-dimensionalterms

Ha Hattanumber

Greek

 braidangle(◦)

ε voidfractionofthebed

fCO2 specific absorbed molar flux of CO2

(molm−3s−1)

wCO2 absorbedmolarfluxofCO2(molm−2s−1)

G gasdensity(kgm−3)

Inthefollowing,theexperimentalset-upsandmethodsare firstdescribed.Second,resultsareshownanddiscussed.Last, correlationstopredictpressure drop,effectiveareaandgas sidemasstransfercoefficientareproposed.

2.

Methods

and

materials

2.1. Packingstructure

Thesubjectofthisstudyfocusesonanewstructuredpacking technique(SPSPatent2005)developedbySnecmaPropulsion Solide(theSAFRANgroup).Itisconstructedwithinterwoven carbonfibres(Fig.1).Thetubesareformedwithcarbonfabrics whicharewovenonamandrelaccordingtoaparticularbraid angle(Fig.2a).Thebraidangle()correspondstotheangle formedbetweenabraidthreadandthebraidaxis(Fig.2b).

(4)

Fig.2–(a)Carbonfabricswovenonamandrel;(b)braid angle;(c)crossingoffibres.

Thedistancebetweentwofibrecrossings(Fig.2c)alongthe circumferenceisgivenby

AB= 2Dm Nf

, (1)

whereDmisthediameterofthemandrel(equivalenttothe

diameterofatube)andNfisthenumberofspindles(Fig.2a,

16spindles).If

AB≤cosLf, (2)

Fig.3–Onetubewithholes.

whereLfisthewidthofacarbonfabric,thenthereisnofree

spacebetweenthefibres(thereisnohole);butif AB> Lf

cos, (3)

then ahole isformed. Fig. 3representsatube withholes. Therefore,thevalueofthebraidangledeterminesthetube holesizes,andthelowerthebraidanglethebiggertheholes.

The diameter of the mandrel can vary from 4.5mm to

20mm,andthebraidanglecanvaryfrom15◦to45.The

open-ingsthereforeswingfrom0%to85%,correspondingtoahole surfaceareafrom0mm2to735mm2.Thetubesarethenfitted

togetheraccordingtothefourdiagonalsofacubeasshown inFig.4a,whichdemonstrateswhythispackingiscalled“4D packing”. Finally,thelayout isrepeated inthe threespatial directions(Fig.4b)toobtainthefinalstructure(Fig.4c).

Twopackingsstructuresweremadewith10mmdiameter

tubes.Thefirstgenerationwasmadewithabraidangleof30◦,

soaholesizeof7.4mm2 (correspondingtoanapproximate

openingof30%;named4D30%,Fig.5).Thispackingpossesses avoidfractionofapproximately94% andageometricarea,

ag,of420m2m−3(thesurfaceareaisevaluatedby

geometri-calcalculation,knowingthesurfaceoftubes,thediameterof holeandthenumberoftubeperpacking’svolumeinm3).The

second generationisrelativelysimilar tothefirstonewith only8spindles(Nf)insteadof12,sotheholesizechangesand

becomes26.6mm2(relativetoanapproximateopeningof50%,

named4D-50%).Thispackingmaintainsageometricarea,ag,

of330m2m−3.Table1resumesthecharacteristicsofthesetwo

packing.

Thisstructureisadvantageousbecausemanyparameters canbemodifiedatwilltooptimizetheperformancesofthe structuredpacking.Inparticular,itispossibletochangethe tubediameters,theholesizesoftubes(openings),thesizes ofcarbonfabric(numberoffibres),andthetubeangles. More-over,thispackingpossessesotherinterestingpropertiessuch asasmalltubethickness(0.2mm),assuggestedduringthe evolutionofstructuredpacking,and asignificantstructural cohesion(mechanicalstrength)duetogeometryofthe struc-ture(usingthefourdiagonalsofacube).

2.2. IFPfacilities

Experimentshavebeencarriedoutina0.73mheight,150mm internaldiametercolumn.Thepressureisclosetothe

(5)

atmo-Table1–Characteristicsofthetwogenerationofpacking.

Packing Voidfraction(tube) Dm Nf Lf ag Voidfraction

Sepcarb®

4D-30% 31.24% 10mm 12 2mm 420m2m−3 94%

Sepcarb®

4D-50% 49.84% 10mm 8 2mm 327m2m−3 95%

Fig.4–(a)Tubesfittingaccordingtothefourdiagonalsofa cube;(b)reproductionofthelayoutinonedirection;(c)final structure.

Fig.5–4Dpackingelements.

sphericpressure,thetemperatureistheroomtemperature. Liquidload,QL,variesfrom10to60m3m−2h−1,gasvelocity

variesfrom0to2m/swhichleadstoaF-factor,Fs,between0

and2Pa0.5.Thedrippointdensityoftheliquiddistributor,dp,

whichisthenumberofliquidinjectorsbysurfacearea,isclose to3350m−2.AccordingtoFairandBravo(1990)orAroonwilas

etal.(2001),itishighenoughtoensurethatthedistributor doesnotinfluencetheresults.

Table2givespackedbedcharacteristicsforthethreetested geometries.Ithastobenoticedthatthereisanon-negligible voidzoneforallexperiments.Forthe4D-50%,a100mmheight bloc had beentested. Elements of4Dpackingare oriented at45◦ fromeach other,and thebed heightsare comprised

between0.2and0.6mapproximately.

2.3. LGCfacilities

Theexperimentalhydrodynamicsand masstransfer

coeffi-cientsetupforthisstudyisaglasscolumnwithaninternal diameterof150mmandheightof1m.Countercurrent

oper-ation with an air–water system was used and all studies

were carried out atroom temperatureand under standard

atmosphericpressure.Forthe pressuredropmeasurement,

the packedbed height is 0.9m, and forthe masstransfer measurementthreeconfigurationsareusedwithbedheights comprisedbetween0.1and0.3mapproximately(seeSection

2.6).

Theliquidflowsfromatankthroughapumpandflowmeter (withameasurementprecisionof±2.5%)andwassuppliedat thetopofthecolumnviathesameplatedistributorprovided byIFP.Theliquidisagaincollectedintothetankafterhaving passedthroughthepacking,withsuperficialliquidvelocities intherangefrom1to30m3m−2h−1.Thegasflowissupplied

atthebottomofthecolumnandwasmeasuredbytwo

dif-ferentflowmeters(withaprecision±1.6%)forsuperficialgas velocitiesfrom0to2ms−1foranemptycolumn.The

pres-suredroppermeterwasmeasuredusinganinclinedU-tube filledwithwater,whichyieldedpressuremeasurementswith aprecisionof0.05mbar.

2.4. PressuredropmeasurementsatLGC

Theexperimental procedure usedtomeasure thepressure

dropconsistsofaperiodicincreaseofgasflowforaconstant

Table2–Geometriccharacteristicsof4Dbeds.

4D-30% 4D-50%

Configuration 1 1 2

Packedbedheight(m) 0.441 0.20 0.315

Voidzoneheight(m) 0.289 0.53 0.415

Numberofblocs 9 4 5a

Geometricarea(m2m−3) 420 327

Voidfraction(%) 94 95

Material Carbon

Stateofsurface Smooth

(6)

liquidflowuntilfloodingisreached.Thefloodingpointcanbe definedasthepointwhereareversalliquidflowappears.At thismoment,theliquidisunabletoflowdownwardthrough thepacking,thepressuredropincreasesdrastically,andan

accurate pressure measurement is impossible due to the

instabilityofthesystem.Beforeeachtest,ahighliquidflow issuppliedandpassesthroughthebedfor30mintofullywet thepackingandavoiddryzones.

2.5. EffectiveareameasurementsatIFP

Inthepresentstudytheair/NaOHsystemhasbeenchosen.

Withinthepackedcolumn,hydroxidesareconsumedbythe

absorbedCO2accordingtothefollowingreactions(Pohorecki

andMoniuk,1988):

CO2(G)↔CO2(L) (4)

CO2(L)+OH−↔HCO3− (5)

HCO3−+OH−↔CO32−+H2O (6)

Reaction(4)correspondstogastoliquidabsorption; reac-tion(5)isthereactiontoconsiderforkineticssincereaction(6)

hasamuchhigherreactionratethanreaction(5).An enhance-ment factor, E, can be used to describe the impact ofthe chemicalreactionontheabsorptionrateofCO2(Danckwerts,

1970): CO2=ϕCO2.ae,global=



1 kG + He E.kL

−1

.ae,global.

!

PCO2−He.C0CO2



CO2=KG.ae,global.

!

PCO2−He.CCO2 0



(7)

Atcolumninlet,theCO2molarfractioninthegasphaseis

closeto400ppmv,whiletheconcentrationofNaOHinthe liq-uidphaseisequalto0.1moll−1andinlargeexcesscompared

totheabsorbedCO2.Duringatest,a1m3storagetankensures

thatthesodiumhydroxideandbicarbonatesconcentrations areconstantandnegligible,respectively(Raynaletal.,2004). Thisleadstoapseudo-firstorderreactionandfastreaction regime(Seibertetal.,2005;AlixandRaynal,2008)forwhich (Danckwerts,1970):

E=Ha=

p

DCO2.k2.C 0 OH kL C0CO 2=0mol m −3 (8)

Itcanalsobeassumedthatthegassideresistanceis negli-gible.Withoutgasresistance,thecombinationofrelations(7) and(8)gives: CO2=

p

DCO2.k2.C 0 OH He .PCO2.ae,global (9)

Withinthepackedcolumn,one-dimensionaland station-aryplugflowsofliquidandgasareassumed.Then,thepacked bedcanbesimulatedbyasuccessionofsingleelements(Fig.6) forwhichthemassbalanceleadstothefollowingsetof

equa-Fig.6–Singleelementtomodelpackedbedof

experimentalcolumns,andestimatemasstransfer

parameters. tions:

dFGCO 2 dz =CO2.Ac FG CO2= yCO2 1−yCO2 ×FG N2 yCO2= PCO2 Ptot (10)

TheCO2gasmolarfractionismeasuredattheinletand

attheoutletofthecolumnviainfraredmeasurements.The effectiveareaofthecolumnsection,ae,global,isassumedtobe

constant.Thecolumnisassumedtobeisothermandisobar, thisisjustifiedbythesmallamountofCO2whichisabsorbed,

the negligible correspondingtemperature increase,and the very littlepressure drop.Diffusioncoefficient, DCO2, kinetic

constant,k2, andHenryconstant,He,havebeen calculated withrelationsgivenbyPohoreckiandMoniuk(1988).Forthe presentchemicalsystem,thesodiumhydroxidesolutioncan beconsideredunloaded,andtheliquidviscosityisclosetothe waterviscosity.Thisleadsto:

log



H w He



=−I×h=−20001

!

C0 Na+C0OH



×(0.091+0.066+hg,CO2) log



100 Hw



=9.1229−5.9044×10−2×T+7.8857×10−5×T2 log



k 2 k2∞



=0.0221×I−0.016×I2 log(k∞ 2)=11.895− 2382 T log(DCO2)=−8.1764+712.5T −2.594×10 5 T2 (11)

FromtheinletCO2molarfraction,yCO2,in,relations(9)and

(10)showthatae,globalistheonlyparametertoadjustinorderto

fittheCO2outletmolarfraction.ThentheCO2profiledirectly

givesae,global.

Liquidsamplesaretakenatinletandatoutletofthe col-umn.CO2contentismeasuredbyHCltitration.Massbalance

betweenthegasandtheliquidphasehasthusbeenchecked.

2.6. Gassidemasstransfercoefficientmeasurements atLGC

Generallyusedmethodsofkgaedeterminationcanbe

classi-fiedintothreemaincategories:

(7)

Table3–Differentmethodsusedintheliteratureforkgaedetermination.

Absorption Evaporationofpurliquid Absorptionwith

chemicalreaction

Examples Air-CO2aorSO2aorNH3aorO2b/water Air/waterormethanolcorMEK... NH3d/H2SO4

SO2a/NaOH

Obligation Humidificationofgas Humidificationofgas

Measurements Cg,Cl,T,P,Ye,f,g T,P,Ye,f,h Cl,Cg,T,P,Ye,f,g

Difficulties Cg,Cl,noisotherm Tinterfaceh,i,j Cg,Cl,noisotherm

a Billet(1995).

b PuranikandVogelpohl(1974). c SuroskyandDodge(1950). d NakovandKolev(1994). e VidwansandSharma(1967). f SharmaandDankwerts(1970). g Hüpenetal.(2006).

h Treybal(1965). i Ondaetal.(1968).

j KawasakiandHayakawa(1972).

- theevaporationofapureliquidbyaninertgas, - theabsorptionwithchemicalreaction.

Table3givesthedifferentsystemsusedintheliteraturefor thedeterminationofkgae,andtheneedofmeasurementand

thedifficulties ofeach methodare presented.Followingthe analysisofthistable,themethodofevaporationofpure liq-uidisappliedheretowaterasthepureliquidandtoairasthe inertgas;despitethefactthatthemethodislessused nowa-days,itpresentshoweversomeadvantages.Indeed,usingthe methodofevaporationofpureliquid,weensurethatthe trans-ferresistanceistotallylimitedtogas-phase.

However,itisrecognizedthatdependingonoperating con-ditionsandincasesofabsorptionwithchemicalreactionmass transfer resistanceispractically confinedtothe gasphase. Theprocessbyphysicalabsorptionisnotsatisfying,becauseit doesnotallowtoreducesufficientlytheliquidfilmresistance inordertoobtainkgae.However,thenumberofmeasurement

carriedoutforthemethodofevaporationofapureliquidis lowcomparedtoothertwomethods,whichlimitsapriorithe uncertaintiesevaluationofkgae.

Regardingthequalityofresultsonecanexpectfromthis method,thearticleofSuroskyandDodge(1950)canserveas areferencebecauseitiscomprehensive,fullofuseful

infor-mation and details a clear methodology treatment of raw

experimentalvalues.Theresultsshowthatitispossibleto haveaprecisiononthekgaeof10%.Moreover,the

measure-mentofairhumidityisnowadaysachievedwithmoreefficient equipment.

Theoperatingmode istoobtain adiabatic conditionsof saturationbytakingintocontacttheairwithwaterat counter-currentundersteady-stateconditions.Theairhumidification ascoolingprocedureleadsthesystemtowardsastable oper-atingconditionsintermsofwatertemperatureatadiabatic conditions of saturation (TAS). The gathered experimental

data’satstableconditionoffunctioningpermitthe determi-nationofkgae,usingmassbalanceforgas-phase:

LnYAS−Yin YAS−Yout =

kgaeH

G0 (12)

WithYinandYout asabsoluteinletandoutlethumidities

ofair incolumn, G0 asairflow, Z as column’s heightand

YASasabsolutehumidityatadiabatictemperatureof

satura-tion.Inordertooverwhelmtheimpactofextremityeffects,

authors(SuroskyandDodge,1950)recommendrealisingthe

same measurements withatleast twodifferent heightsof

packing.So, ourpilotcontainsacolumnof150mm diame-ter,withwaterandairatcounter-current(Fig.7).Dryairflow ismeasured;thenitishumidifiedandheated(byan evapora-torfedbyameasuringmicro-pump)inordertoobtaindesired conditionsofbottom-column.

Temperatureanddewpointaremeasuredatinletand out-letsectionsbyahygrometer(hygrometerwithcooledmirror permittinganaccuracyof±0.1◦C ofthe dewpoint). Water

circulatesinclosedloopthroughstorage,up-column,liquid distributor(withadjustableheighttodistribute5cmontopthe packing),gasdistributortocomebackintostorage.Flows,inlet andoutlettemperaturesofcolumn(TLin andTLoutforliquid

temperature and TGin and TGout forgastemperature),

abso-luteinletand outlethumiditiesofair(YinandYout),aswell

asstoragetemperature(Tws)aremeasured(Fig.7).Theliquid

storagehastobesolimitedintermsofvolume(10l)inorder toreducethetimeofstabilisation.Theoperatingconditions arepredictedtoobtainanadiabatictemperatureofsaturation closetothesurroundingtemperature.

Foreach test,thesteadystateisreached(observedafter 30minwhenallmeasurementsarestable).Inthiscase,the temperaturesTLin,TLoutandTwsareequal,andadiabatic

tem-peratureofsaturationisreached.Thistemperaturepermits tocalculateYAS,theabsolutehumidityatadiabatic

tempera-tureofsaturation.Andthefinalresultkgaeiscalculatedbythe

relation(12).

For thepresent study six4D-30%and4D-50%blocshad

beentested.

3.

Results

and

discussion

3.1. Pressuredropcurve

Fig.8givesthedrypressuredropof4Dpackingsasa func-tion oftheF-factor.Logarithmicscalesare usedinorderto checkthepressuredroppowerlawcoefficientwhichislinked

to the flow regime. Present measurements are compared

withcalculatedpressuredropforthreeMellapakstructured

packingscommercializedbySulzerChemtech:M252Y,M500Y

andM750Y.Ithastobenoticedthatthegeometricareasof selectedcommercialpackingsareclosedtothe4Dones(see

Tables2and4).Calculationshavebeencarriedoutwiththe manufacturersoftwareSulcol2.0.First,theexperimental

(8)

pres-Fig.7–Experimentalsetupforkgaemeasurement.

Table4–GeometriccharacteristicsoftestedMellapakpackingsforeffectiveareameasurement.

MellapakPlus252Y Mellapak250Ya Mellapak

500Ya

Packedbedheight(m) 2 3 3

Columndiameter(m) 0.15 0.43 0.43 Geometricarea(m2m−3) 250 250 500 Voidfraction(%) 98 98 97.5 Corrugationlength(mm) 19 19 9 Channelangle(◦) 45 45 45 Material 316L

Stateofsurface Textureonthewall,perforated

a Tsaietal.(2008).

suredroponacolumnequippedwith4D-30%issimilartothe

onecalculatedonacolumnequippedwithMellapak750.Y.

Second,theuseofthe4D-50%allowstoreducethepressure

dropby50%approximately.Thepressuredroponacolumn

equippedwith4D-50%becomessimilartotheonecalculated

0.1 1.0 10.0 100.0 10.0 1.0 0.1 Fs (Pa0.5)P/L (mbar/m) 4D-30% 4D-50% M500Y, Sulcol 2.0 M750Y, Sulcol 2.0 M252Y, Sulcol 2.0

Fig.8–DrypressuredropasafunctionofFsfor4Dopen packings.Comparisonwithcalculatedpressuredropfor

Mellapak750Y,Mellapak500YandMellapakPlus252Y

(Sulcol2.0).

onacolumnequippedwithMellapak500Y.Ithastobenoticed thattheslopeofthecurveiscloseto1.9forboth4D-30%and 4D-50%,thisresultissimilartotheresultsobtainedbySpiegel andMeier(1992).Last,thecalculatedpressuredropona col-umnequippedwithMellapakPlus252.Y,whichisthereference caseforthepresentwork,is50%lowerthantheonemeasured

onacolumnequippedwith4D-50%.

Fig.9givesthewettedpressuredropof4Dpackingsasa functionoftheF-factorandforQL=28m3m−2h−1.Such

liq-uidloadisexpectedforCO2absorbers.Presentmeasurements

arecomparedwithcalculatedpressuredropforMellapak252Y, 500Yand750Y.Inthecaseof4Dpackings,thepressuredrop

is reduced by30% whilethe capacity is increased by 25%

whentheopeningfractionincreasesfrom30to50%.Below theloadingpoint,theexperimentalpressuredropona col-umnequippedwith4D-30%(respectively4D-50%)issimilarto theonecalculatedonacolumnequippedwithMellapak750.Y (respectively500Y).Thefloodinglimitofthe4D-30%issimilar totheonegivenforMellapak750Y,andthefloodinglimitof the4D-50%ishigherthanthosegivenforMellapak500Yand 750Y.Foracolumnequippedwith4D-50%whichoperatesat afloodingpercentageequals 70%,thepackedbedpressure dropwillbecloseto3mbar/m.Last,itshouldbenoticedthat thecapacityoftheMellapakPlus252Yis40%higherthanthe

4D-50%one.

3.2. Effectivearea

First,thevoidzoneimpactshouldbecharacterizedforopen 4Dpacking(cf.Table2).Fig.10giveseffectiveareawhichhas

(9)

0.0 0.1 1.0 10.0 100.0 10.0 1.0 0.1 Fs (Pa0.5)P/L (mbar/m) 4D-50% 4D-30% M500Y, Sulcol 2.0 M750Y, Sulcol 2.0 M252Y, Sulcol 2.0 slope=1.9

Fig.9–WettedpressuredropasafunctionofFsfor4D openpackings,QL=28m3m−2h−1.Comparisonwith

calculatedpressuredropforMellapak750Y,Mellapak500Y andMellapakPlus252Y(Sulcol2.0).

beenmeasuredwithanemptycolumn,ae,void,asafunction

ofQLandFs.First,itappearsthatae,voiddoesnotdependon

thegasflowrateforthetestedrange.Second,ae,voidislowand

comprisedbetween20and60m2m−3.Last,asimpleand

accu-ratecorrelationcanbeusedtoestimateae,voidforthepresent

work:

ae,void=7.66×(3600×QL)0.492 (13)

CO2absorbedratemeasurementsgivetheaveraged

effec-tiveareafortheentirecolumn,ae,global(seeSection2.5).

Forthe4D-30%and4D-50%packings,thevoidzonecannot beneglected(seeTable2)andthepackedbedeffectivearea

y = 7.6558x0.492 0 10 20 30 40 50 60 70 60 50 40 30 20 10 0 QL (m3/m2/h) ae,void (m 2 /m 3 ) Fs = 0.85 Pa0.5 Fs = 1.2 Pa0.5 Fs = 1.5 Pa0.5

Fig.10–Effectiveareaforthevoidzone,asafunctionofQL

andFs.

willbegivenby:

ae,4D=ae,global×H−aHe,void×(H−H4D)

4D (14)

Twoconfigurationshavebeentestedwiththe4D-50%(see

Table2),thenrelation(14)leadstotwodifferentvaluesfor

ae,4D.Forthiswork,anarithmeticaverageofthesetwovalues

isretained.Fig.11givestheeffectiveareameasuredwiththe 4DpackingsasafunctionofFsandQL.Presentexperiments

arecomparedwitheffectiveareasmeasuredfordifferent com-mercialpackings:MellapakPlus252.Y(AlixandRaynal,2009), Mellapak250Yand500Y(Tsaietal.,2008).Ithastobenoticed thatMellapakPlus252.Yhadbeencharacterizedwiththesame

experimental columnas 4D packings, however 2%CO2/1N

sodiumhydroxidesystemwasused.Thelattercouldleadto underestimatetheeffectiveareaabout15%becauseofgas

lim-Fig.11–Effectiveareafor4DpackingsasafunctionofQL.ComparisonwithdifferentMellapakcommercialpackings(Alix andRaynal,2009;Tsaietal.,2008).

(10)

itationandnon-pseudo-firstorderreaction(AlixandRaynal, 2008).IthastobealsonoticedthatMellapak250.Yand500.Y hadbeencharacterizedbySRPwitha430mmdiameter

col-umn. Theuse ofdifferent columndiameters could impact

experimentalresults(Olujic,1999),however,sincethecolumn

diametersare always muchhigherthan channelsizes(see

Table4),thediameterimpactshouldbemoderate(Henriques de Britoet al., 1994). First,onecan observe that Mellapak-Plus252.YandMellapak250.Yeffectiveareasaresimilar,this wasexpectedsincebothpackingshaveverysimilargeometric characteristics(seeTable4).Thisresultindicatesthat chem-icalsystemanddiametereffectsshouldbemoderateforthis packing,thenitisreasonabletocomparepresentresultswith thoseobtainedwithMellapak250.YandMellapakPlus252.Y. SinceMellapak500Ychannel sizeislower than the Mella-pak250.Yone(seeTable4),scaleeffectshouldbelowerfor the500.Y.Thisalsoindicatesthatitisreasonabletocompare presentresultswiththoseobtainedwithMellapak500.Y.

Second,Fig.11showsthat,whatevertheopeningfraction, 4Deffectiveareasaresimilarwhileagdecreasesstronglywhen

theopeningfractionincreases.Thisresultcouldbeexplained bythefactthatmoredropletsaregeneratedwhenopening fractionincreases (Alix and Raynal, 2008). Then,the effec-tiveareacouldexceedthegeometriconefor4D-50%(ratioup to1.4).Third,experimentalvaluesaremorescatteredfor4D packingthanthoseobtainedwithMellapakpackings.Thisis explainedbythefactthattheeffectiveareaislesssensitiveto thegasflowrateforMellapakthanfor4Dpackings.Inthecase ofopen4Dpackings,thegasflowrateimpactislinkedtothe amountofdropletslikeforrandompackings.Last,4D interfa-cialareasaremuchhigherthanthoseobtainedwithMellapak packingsinspiteofthefactthattheMellapakgeometricareas couldexceed4Dones.Thisresultisveryinterestingsincethe absorberheightwillbedirectlylinkedtothe packing effec-tivearea,andthisefficiencygaincouldcompensatethelower capacityof4Dpackings(seeSection3.1).The4D-50% struc-tureisthemoreinterestinggeometrysinceitgivesthebest compromisebetweenefficiencyandcapacity.

Toestimatethegainrelativesto4D-50%packingforCO2

captureplants,oneshouldmodeltheoverallabsorber (includ-inghydrodynamics,massandheattransfer,thermodynamics, and kinetic).Then,acorrelation toestimatethe interfacial areaishighlyneeded.Fig.12givestheeffectiveareaforthe 4D-50%packingasafunctionofFsandQL.Thefollowingrelation

canbegivenwithanaccuracyof±5%:

ae,4D-50%=A×(3600×QL)B A=137.77×Fs+134.39 B=0.085 (15)

3.3. Gassidemasstransfercoefficient

kgae measurementsare realisedwith 3differentheightsof

packinginordertodeducetheextremityeffects,i.e.1blocs, 2blocsand3blocsofpacking(0.1m,0.2m,and0.3m, respec-tively).First,thepacking4D-50%hadbeentestedforaliquid loadscaleof7–23m3m−2h−1,andforagasflowratescaleof

3500–7200kgm−2h−1.

Thevaluesofkgaearepresentedinlogarithmicscaleasa

functiongasflowrate’slogarithminFig.13. Regardingthis figure,twostraightlines(dotted)areaddedwhichcorrespond at+10%and−10%oftheaveragevalue,andpointsareincluded betweenthesetwoboundaries.

y = 246.03x0.0929 y = 283.07x0.0879 y = 328.69x0.0745 200 250 300 350 400 450 500 550 600 80 70 60 50 40 30 20 10 0 QL (m3/m2/h) a e,4D (m 2 /m 3 ) 4D_50%, Fs=0.8, IFP 4D_50%, Fs=1.1, IFP 4D_50%, Fs=1.4, IFP

Fig.12–Effectiveareafor4D-50%asafunctionofFsandQL. Therefore,underfunctionalconditionsofthispacking,the resultspresentaquasi-independencyonliquidflowrate,and animportantdependencyongasflowrate.Thefollowing rela-tionshipisthensuggested:

kg∗ae=3.7488×10−3∗G0.843 (16)

withGinkgh−1m−2andkg*aeins−1.

ThepowerofGisaround0.8against0.7formostof tradi-tionalpacking(Ondaetal.,1968).Evenifitisnottraditional, ithasbeenalreadyobservedforsomeotherinternalsof col-umn.AccordingtoDwyerandDodge(1941),thedependency ongasflowrateforringswith25mmdiameterisapowerof 0.77andfor12.5mmringdiameteris0.9.Forthe4Dpacking, thetubediameteris10mmwhichisclosedtothe12.5mmring diameter.Thatissuggestingthatforthe4D-50%the character-isticdimensionofflowislowandthetestsin150mmcolumn diameterarerepresentative.

However,itispossiblethatthisvalueislinkedtothefact thataeisverysensitivetothegasflowrate.Regardingits

par-ticularstructure,thispackinggenerates: - liquiddropletsthroughlargeopenings.

- liquid film which follows the assembly angle ofpacking tubesandthereforeanincreaseinmovementwithrespect toaverticalflow. 0,4 0,45 0,5 0,55 0,6 0,65 0,7 0,75 0,8 0,85 0,9 0,95 1 3,9 3,85 3,8 3,75 3,7 3,65 3,6 3,55 3,5 log(G) log(k g ae ) +10% -10%

Fig.13–Dependencyoflog(kgae)asafunctionoflog(G)for

(11)

0,6 0,65 0,7 0,75 0,8 0,85 0,9 0,95 1 1,05 1,1 1,15 1,2 1,25 1,3 3,9 3,85 3,8 3,75 3,7 3,65 3,6 log(G) lo g (kg ae ) +10% -10%

Fig.14–Dependencyoflog(kgae)onlog(G)forthe4D-30%.

- thepossibilityofcreatingtheequivalentofabubblingin themiddleoftubesofpackingstructureathighgasflow rates,withmaybeapossibleliquidupstreaminthechannels formedbytubes.

Second,the4D-30%packing istested withaliquidload equals16m3m−2h−1andthepreviousrangeofgasflowrate.

TheresultsareillustratedinlogarithmicscaleinFig.14.One canobserveahighdependencyonLOG(G)(1.54powerofG),as forthe4D-50%.Thiscouldbealsoexplainedbythefactthat

aeisverysensitivetothegasflowrate.

kg∗ae=1.4612×10−5∗G1.54 (17)

withGinkgh−1m−2andkg*aeins−1.

4.

Conclusions

and

perspectives

Anovelstructuredpacking,the4Dpacking,hadbeen char-acterizedintermsofhydrodynamicsandmasstransfer.The

4Dpackingismadeofcarbontubeswhichcouldbeopened

byadjustingmanufacturingparameters(suchasbraidangle). Withanopeningfractionof50%,thecapacityofthe4D

pack-ing is maximum and becomesbetter than those given for

Mellapak500Yand750Ywhichhavesimilargeometricareas. Howeverthecapacityofthe4D-50%is40%lowerthantheone givenfortheMellapakPlus252.Ywhichisourreferencecase. Theeffectiveareaofthe4Dpacking,ae,doesnotdecrease

whentheopeningfractionincreasesfrom30to50%.This indi-catesthatmoredropletsaregeneratedwiththe4D-50%,this isconfirmedbythefact thatae becomesmoresensitive to

thegasflowrate4D-50%.Theeffectiveareaismuchhigher thanthosemeasuredwithMellapakpackings.Inparticular, theeffectiveareaisdoubledcomparedtothe MellapakPlus 252Yone.Thenthe4D-50%packingshouldbeveryefficient forCO2capture.Thisshouldatleastcompensatethelower

capacityofthe4D-50%.

Thegassidemasstransfercoefficient,kG,hadbeen

mea-sured for 4D packing via water evaporation. This original methodgivesveryaccurateresults.Correlationsaregivento estimateae andkG*ae.Thesecorrelationswillbeused

after-wardstoprovidearate-basedmodelfortheabsorber(Aspen

ratesep). Themodelshows theperformance of4D packing

inCO2 capturewith 30wt%MEA. Experimentswill bealso

conductedwitha400mmdiametercolumninordertovalid presentresultsintermsofpressuredropcurvesandeffective area.

Acknowledgments

This work is supported byANR (French National Research

Agency)throughtheGASCOGNEproject.

TheauthorswouldliketothanktheANRforitsfinancial

supportundertheGASCOGNEproject.

References

Alix,P.,Raynal,L.,2008.Pressuredropandmasstransferofa highcapacityrandompacking.ApplicationtoCO2

post-combustioncapture.In:GHGT-9Congress,Washington, DC.

Alix,P.,Raynal,L.,2009.Characterizationofahighcapacity structuredpackingforCO2capture.In:WCCE8Congress,

Montreal.

Aroonwilas,A.,Tontiwachwuthikul,P.,Chakma,A.,2001.Effects ofoperatinganddesignparametersonCO2absorptionin

columnswithstructuredpackings.Separationand PurificationTechnology24,403–411.

Billet,R.,1995.PackedTowers,Weinheim.

Danckwerts,P.V.,1970.GasLiquidReaction.McGraw-Hill,New York.

Dwyer,O.E.,Dodge,B.F.,1941.Rateofabsorptionofammoniaby waterinapackedtower.IndustrialandEngineering Chemistry33,485.

Fair,J.R.,Bravo,J.L.,1990.Distillationcolumnscontaining structuredpacking.ChemicalEngineeringProgress86(1), 19–29.

Feron,P.H.M.,Abu-Zahra,M.,Alix,P.,Biede,O.,Broutin,P.,de Jong,H.,Kittel,J.,Knudsen,J.,Raynal,L.,Vilhelmsen,P.J.,2007. Developmentofpost-combustioncaptureofCO2withinthe

CASTORIntegratedProject:resultsfromthepilotplant operationusingMEA.In:3thInternationalConferenceon CleanCoalTechnologiesforourFutur,Cagliari,Italy. HenriquesdeBrito,M.,vonStockar,U.,MenendezBangerter,A.,

Bomio,P.,Laso,M.,1994.Effectivemass-transferareaina pilotplantcolumnequippedwithstructuredpackingsand withceramicrings.IndustrialandEngineeringChemistry Research33,647–656.

Hüpen,B.,Hoffmann,A.,Gorak,A.,Löning,J.-M.,Haas,M., Runowski,T.,Hallenberger,K.,2006.InstitutionofChemical Engineers,SymposiumSeriesNo.152,523.

Kawasaki,J.,Hayakawa,T.,1972.Directcontactmassandheat transferbetweenvaporandliquidwithchangeofphase. JournalofChemicalEngineeringofJapan5(2),119. Knudsen,J.N.,Vilhelmsen,P.J.,Jensen,J.N.,Biede,O.,2006.First

yearoperationwith1t/hCO2absorptionpilotplantatEsbjerg

coal-firedpowerplant.In:VGBConference,Chemieim Kraftwerk,11–12October,BadNeuenahr,Germany. Nakov,S.,Kolev,N.,1994.Performancecharacteristicsofa

packingwithboundarylayerturbulizers.IV.Gasfilm controlledmasstransfer.ChemicalEngineeringand Processing33,437.

Olujic,Z.,1999.Effectofcolumndiameteronpressuredropofa corrugatedsheetstructuredpacking.TransICheme77A, 505–510.

Onda,K.,Takeuki,H.,Okumoto,Y.,1968.Masstransfer coefficientsbetweengasandliquidinpackedcolumns. JournalofChemicalEngineeringofJapan1(1),56. Pohorecki,R.,Moniuk,W.,1988.Kineticsofreactionbetween

carbondioxideandhydroxylionsinaqueouselectrolyte solutions.ChemicalEngineeringScience43(7),1677–1684. Puranik,S.S.,Vogelpohl,A.,1974.Effectiveinterfacialareain

irrigatedpackedcolumns.ChemicalEngineeringScience29 (2),501–507.

Raynal,L.,Ballaguet,J.P.,Barrere-Tricca,C.,2004.Determination ofmasstransfercharacteristicsofco-currenttwophaseflow withinstructuredpacking.ChemicalEngineeringScience59, 5395–5402.

Seibert,F.,Wilson,I.,Lewis,C.,Rochelle,G.,2005.Effective gas/liquidcontactareaofpackingforCO2

(12)

absorption/stripping.GreenhouseGasControlTechnologiesII, 1925–1928.

Sharma,M.M.,Dankwerts,P.V.,1970.Chemicalmethodsof measuringinterfacialareaandmasstransfercoefficientsin twofluidssystems.BritishChemicalEngineering15(4),522. Spiegel,L.,Meier,W.,1992.Ageneralizedpressuredropmodelfor

structuredpackings.IChemeSymposiumSeries,no.128, B85–B91.

SPS:patentFR0511051(2005).

Surosky,A.E.,Dodge,B.F.,1950.Effectofdiffusivityongas-film absorptioncoefficientsinpackedtowers.Industrialand EngineeringChemistry42(6),1112.

Treybal,R.E.,1965.MassTransferOperations.McGraw-Hill Company.

Tsai,R.,Schultheiss,P.,Kettner,A.,Lewis,C.,Seibert,F.,Eldridge, B.,Rochelle,G.,2008.Influenceofsurfacetensiononeffective packingarea.IndustrialandEngineeringChemistryResearch 47(4),1253–1260.

Vidwans,A.D.,Sharma,M.M.,1967.Gas-sidemasstransfer coefficientinpackedcolumns.ChemicalEngineeringScience 22,673.

Figure

Fig. 1 – Carbon fibres woven.
Fig. 3 – One tube with holes.
Fig. 4 – (a) Tubes fitting according to the four diagonals of a cube; (b) reproduction of the layout in one direction; (c) final structure.
Fig. 6 – Single element to model packed bed of experimental columns, and estimate mass transfer parameters
+6

Références

Documents relatifs

We relate R-equivalence on tori with Voevodsky’s the- ory of homotopy invariant Nisnevich sheaves with transfers and effective motivic

Since the binding sites gathered by MED-SMA share binding modes to ligands, this type of cluster could be used to search for specific drugs; here, drugs to

The results show that despite same-content comparisons play a major role in the accuracy of psychometric scaling, the use of a small portion of cross-content comparison pairs is

Zouaimia Rachid, dérèglementation et ineffectivité des normes en droit économique algérien, revue idara n° 01, annee 2001. instruction n° 07-11 du 23/12/2007, fixant les condition

El primer objetivo de este libro es, pues, caracterizar los sistemas de innovación agropecuarios a escala de los países de América Latina y, por lo tanto, las

If the literature review shows that very few methods and information systems exist to support decision-making in such a situation, experience shows that most decisions are based on

Une collaboration avec le laboratoire LP3 (Université de Marseille) nous a permis d'estimer les paramètres caractéristiques du plasma (température, densité électronique,

Elevation-induced climate change as a dominant factor causing the late Miocene C 4 plant expansion in the Himalayan