Dispersion of CO2 in the water column

Dispersion of CO2 in the watercolumnDr. Baixin ChenDept. of Mechanical EngineeringHeriot-Watt UniversityEdinburgh, UKb.chen@hw.ac.ukUK CCS Community Network Meeting, 5 – 6 April, 2011l, University of Exeter

CO2 under seabed storage and challenge:Leakage prediction and risk managementCO2 leakage mechanisms (IPCC SRP on CCS, Chapter 5, 2005)control +500ppm +1000ppm +2000ppm+5000ppm +10000ppmMorphology of Sea Urchin Larva (Kurihara & Shirayama, 2004)

Model Synthesis (WP2 + WP6.3 of QICS)physicsWP 1RISCSWP 2.3L-BsedimentdispersionWP 2.1 and2.2: Bubble /plumedispersionWP 2.5 FinescalehydrodynamicsFSPOLCOMS /FVCOMPOLCOMS1.8kmresolutionshelf modelCO2 propagationSet of exposure scenarios at epicentre, locality and region8.5pH87.576.5scenario 1scenario 26scenario 35.50 10 20 30 40 50time, daysUse these to drive 1D ERSEMSystem impact

Outline of the presentationI. BackgroundII. Modelling of CO2 dispersion in asmall-scale oceana) Dynamics of CO2 drop/bubble inseawater• Numerical models developed• Model calibration with observation datafrom Lab. and small scale field Exp.b) Model application: Dispersion ofLeaked CO2 from North seabedIII. Summary and suggestions

BackgroundSea surfaceDepthSediments layer(hydrate form ableCap rockGeo-formationsNatural gas or oil (EOR)Geo-formationsCO 2 solutionplumesCO2 plumeunder seabedUnder seabed:• Multi-fluid/phase• Porous media, stratification• Interface actions (fluids and fluid/solids)Sediments:• CO2 hydrate formationCO 2 leaked tooceanOcean:Saline water reservoirs•Turbulent boundary layer•Multi-fluid plume•Stratification, free surface•Bubble/drop dynamics•Chemical/Bio-impacts

Scales and the roles of ocean turbulence ondispersions of leaked CO2 in the oceanOGCMs T ~ centuriesTurbulent Kinetic Energies Log (Ek)Regional oceanmodel T ~ yearsMeso-scale oceanmodel & Engineeringfield Exp. T ~ weeksSmall-scale two-fluidflows model & smallfield Exp. T ~ hoursThis presentationDroplet dynamics &Lab. Exp. T ~ hrs100km 1.0km 1.0mScalesBackground energy spectra from Wood, Science 1999

Theory and methodology for two-phasesmall-scale turbulent ocean modelLarge-scale informationfrom Boundaries :Mean properties (X,t)Turbulent propertiesat K > KfData analysisField Obs. DataInside of small-scalescaleocean: N-S S based 3-D 3 D unsteadyGoverning Eq.s for CO2 &seawater Forced-dissipative dissipative Energycascade theories Adjusted by observationspectrum

Governing Eq.s of two-fluid smallscaleturbulent ocean∂Governing Eq.s of Seawater∂ρ∂ρû+∂t∂xi∂ρuˆ∂tρφk∂tˆi∂ρuuˆiˆj+∂xji=&w dco∂pˆ∂Dij= − + + ( ρ − ρ 0)g∂xi∂xj+ ( F&d)+ Ffδij∗∂pˆ∂P= − ρ 0gi∂xi∂xi∂ρφˆ kûj∂ ∂ ˆ φk∂ρqˆ+ = ( ρDk) +∂xj∂xj∂xj∂xjk+ w&idcoδklGoverning Eq.s of CO2 plume∂nˆ∂td∂nˆdû+∂xjdj∂ˆα ∂ˆûα+∂t∂xjρd∂ û∂tdidj=∂ρ+∂xjqˆdn= qˆdûûdi djdα−w&∂ ˆ=∂xjρτ d dijdco/ρd+ ˆ( αρd−ρw)g i −(F&dco) iCoupled by source terms highlighted by underlines, which arethe models of CO2 droplet/bubbles, LES turbulent model anddensity change model of CO2 solution.

Turbulent kinetic energy spectrum(cm 2 s -2 )Turbulent kinetic energy spectrum(m 2 s -2 )10 010 -110 -210 -310 -4EXP data fromKeahole Pt. OffingN-S DirectionModeling10 -510 -4 10 -3 10 -210 010 -110 -210 -310 -4EXP data fromKeahole Pt. offingE-WExample of reconstruction of smallscaleturbulent oceanHzModeling10 -510 -4 10 -3 10 -2Hz(Chen et al, Direct and Large-Eddy Simulation, 2005)W-E (km)N-S (km)Modeling prediction of turbulent kinetic energy(TKE in m 2 s -2 ) distribution on a horizontalsection at depth of 750 m (Keahole case)Experimental data obtained by CREIPI at Keahole Pt. Offing(19 o 43.324N; 156 o 04.806W)

CO2 droplet/bubble ModelsAn integrated model of CO2droplet/bubble leaked fromseabeda) Sub-model of CO2 properties• CO2 solubility• CO2 phase diagramb) Sub-model of individual CO2 droplet/bubble• Momentum exchange sub-model• Mass transfer sub-model

CO2 Phase diagram

Governing Eq.s of a Drop/Bubble inseawaterMomentum exchange Eq.dudtr 2ρsρc3urd ln( mc= ((1.0 − ) g − Cd) − u)rρcρs4DdtMass Eq.dddte=−1ρcd(3ur: slip velocityedρcdt+ 2k( Cde: equivalent diameter of bubbleCs: solubilityρ : densitys−C 0))DissolutionCO2BuoyancyDragKey Parameters:• k : mass transfer coefficient•Cs: CO2 solubility•Cd: drag coefficientShape (spherical ordeformed) & interfaces(solid, liquid, and gas towater)

Model validation: CO2 droplet indeep oceanDroplet Diameter (cm)Droplet velocity modeling vs data(Nikolaus et al. EST, 2008)1.000.800.600.400.200.00:Observation Data of Droplet A(P. Brewer et al):Observation Data of Droplet B(P. Brewer et al)10 25 40 55 70Elapsed Time (min)Modeling Prediction (Droplet A)Modeling Prediction (Droplet B)Droplet dissolution modeling vs data(Chen et al, JGR 2005 and field Exp data by P. Brewer et al, EST, 2003)

Model validation: CO2 bubble inshallow oceanADCP backscatter showing bottom“flares” (red signal) at Salt Dome Juist,north sea. CO2 bubble size (de) 4 – 8mm. (F. Daniel et al, University of Kiel, to bepublished by JGR)Numerical simulation of CO2bubbles leaked from seabed(Chen et al, GHGT-9, 2008)

Model validation: Field Exp of directinjection of LCO2 into the oceanLocation : Monterey Bay(Tubeworm Slump(36.6417 N, 122.4158W) Motivations:ooTest new AcousticSonar system formonitoringCO2 plumedynamics Time: 17 ~ 21, Nov. 2005

CO2 droplet plume monitored (right top) and (right-bottom)(Brewer & Chen et al, GRL 2006)Western FlyerOne Sonar fromsurface ship forvertical detective800m900m10000 5 10 15 20 25 30Time since CO 2 release (minutes)ROVAnother Sonar inROV for horizontaldetective

Acoustic Detection (top) and modeling (bottom) ofCO2 plume at horizontal (Brewer & Chen et al, GRL 2006)

Application example: Dispersion ofleaked CO2 from North seabedTerrametrics, 2011Npd.no•CO2 leakage parameters:• Leakage rates: 0.5 kg/s/m²• Ocean current: 0.5, 0.05 m/s• Initial bubble sizes: 10, 15 30 mm•Leakage site:• North sea (58.53N 0.21E)• Leakage depths:30m, 82m; 150m

CO2 bubble (left) and CO2 enrichedwater plumes at T=1.0hrLeakage depth: 150m; de=10 mm; Uc = 0.5m/s; CO2leakage rate: 0.5kg/m2/sFrom 150m depth (m)Along the current (km)Along the current (km)

Volume of CO2 enriched plumewith pH changesCase123456Leakagedepth(m)150M150M80M80M30M30MSeasonSummerWinterSummerWinterSummerWinterCO2 enriched plum e volum e (m ^3)1.0E+071.0E+061.0E+051.0E+041.0E+031.0E+021.0E+01Volume of CO2 enriched water plumewith pH change >0.01Volume of CO2 enriched water plumewith pH chnage > 0.5Volume of CO2 enriched water plumewith pH change > 1.0Volume of CO2 enriched water plumewith pH change > 1.5All cases: De = 15 mm;Uc = 0.05 m/s; T=1.0 hr;1.0E+001 2 3 4 5 6Case No

Effects of current, leakage depth,rate and bubble sizeMax height ofseawater plume (m)Effective volume perunit width of pHDistribution (m²)Maximum pHchangeBase Case 13.35 24033 1.571Ocean current: 0.0m/s15.82 4270 2.832Leakage depth:150mCO2 leakage rate:1.0 kg/m2/sInitial bubble size:30 mm15.41 27730 1.57214.99 26991 1.85427.37 49272 1.101

• Summary:Summary and suggestions– A two-phase model of leaked CO2 dispersion insmall scale turbulent ocean is developed– Model is validated by available Lab. and filedobservation data from shallow and deep ocean• Improvements:– More field exp data of CO2 bubble/drop plume– Model of CO2 phase transition (depth 400 – 300m)– Model of bubble/drop interactions• Nesting to the mesoscale model (QICS)• Development and nesting to the model of CO2dispersion through sediments (QICS)

Acknowledgements• Dr. Peter Brewer of MBARI, USA, forcollaboration on CO2 field observations.• Mr. Jerry Blackford, , PML, project leader andmodelling collaboration.• Mr. Marius Dewar, student of Heriot-WattUniversity, performed the Case studiesFinancial supporting from projects of QICS,NERC and ECO2, EU-F7.

WP 2.5 Physical coupling of fine scale and regional hydrodynamicmodelsPreviously we have resolved ‘events’at rather coarse scales based on thePOLCOMS – ERSEM platform.Huge leakThe challenge is to resolve the localscale:Two possible approaches:Nesting either200m scale POLCOMS (tried and trusted)FVCOM model which resolves down to 5m (new technology)

Lab. Exp on CO2/CO/CO2 solution atcapillary-scale poursp outε : Porosity ofgeo-formationsρ wρ Cρ sρ Cb oyChemicalreactionswith solidgrainsµρ sρ wρ C < ρ w < ρ sp inGeo fluids: saltwater and oilCO 2 (Liquid or supercritical)CO 2 water/oil solutionsor CO 2 hydrateMicroscope + High speed CCDcameraInte raction a mon g CO 2, CO 2 solutio ns, ge o-flu id s, geo-gra ins

Preliminary ExperimentsT= 10 sec T= 120 minsCO2 bubble shrinking in a chancel (din =2.8mm) at atmosphere pressure andtemperatureDemonstrating exp set up:micro-pump testingChallenge: (RA’s work)• Suitable exp set up withMicroscopic system and capillarypressure• Observation of micro-dissolution(Micro-PIV?)

LBM model of CO2 dispersion anddissolution in pore scaleAdvantage:• First principle from Boltzmanndistribution• Direct treatment of interfaceactionsSpecifics:• Physic space – Dimensionless ---LBM spaceDistribution fonction: fa( x,ea,t)Macroscopic density:Macroscopic velocity:ρ =8∑a = 0f a1∑8a=0u = f . e aρa

LBM - testingT= 0; 900 1700 2500 3500Bubble rising in a channel (LBM)Challenge: (RA + PhD)• CO2/Water system (EOS)• Dissolution model (no one available)

• Sept-10– Sept. 11:Preliminary Schedule– Task 2.1– Task 2.2 (Small-scale ocean model)– Task 2.3 (Preliminary Lab exp: dispersion rate)– Task 2.3 LBM: CO2 EOS, interface tension• Sept 11 – Sept. 12:– Task 2.2: Coupling Small-scale ocean and bubble model– Task 2.3 Lab exp: dissolution rate– Task 2.3 LBM: Dissolution model• Sept, 12 – Sept. 13– Task 2.2: Modelling simulations and nesting to meso-scale scale model(WP2.5)– Task 2.3 Lab exp data analysis and model development– Task 2.3 LBM: Coupling dispersion and dissolution and casesimulations

a. Forced-dissipative system of small-scaleocean:∂ ρû∂tIII-2. Governing equations for reconstructing asmall-scale turbulent ocean(Chen et al, Tellus, 2003)i∂ρûiûj+∂xjForced term:Ff∂pˆ= −∂xiDissipative term:Dik′∂D+∂xjij+ ( ρ − ρ 0 )g= ρεu0(k,t)/ ∑ uk ∈ k 0(k,t)u 0= 2ρνktSi+ ( F&d) +Ffδijb. Structure function turbulent viscosity model∑ikk2F = × − − ∆20.25( uk(xk)uk(xkxk))−3/ 2k( xk , ∆xk) = 0.15C∆xk[ F ( xk,∆xk)]kνtk2f′′(k,t)k∈kf

Bubble dynamics in sedimentsAn X-Ray CT image of a sediment with interbandedclay and carbonate sand layers andcontaining (post-collection?) bubbles in bothtypes of sediments (black circles and ellipses).Bubbles in the sands are spherical away frommud contacts, e.g., as indicated by the redarrow labelled “a”whereas the bubbles in the muds are oblatespheroids, e.g. as indicated by the yellowarrow labelled “b’.Bernard. P. Boudreau, et al, GRL, 2004

LBM- CO2 dispersion in a channel and aporous mediaT= 0; 900 1700 2500 3500Bubble rising in a channel (LBM)Wie, W & Chen; to be submitted

I-3. Characteristics of CO2 droplet &bubbleRe & dv/dhσ =σ =Eo=∆ ρ gdσ2e

Challenge: Marine environmentalimpacts, risks and risk management For CO2 under seabed storage, there exists anpotential possibility that stored CO2 could migrate andleak into water column at an engineering scales, whichwill alter the local chemical environment. Lab. experiments have shown that sustained highconcentrations of CO2 would cause mortality of oceanorganisms. The effects of leaked CO2 on marineorganisms will have ecosystem consequences. The chronic effects of leaked CO2 at an engineeringscale on ecosystems over large spatial and long timescales have not yet been understood (studied?).

Natural CO2 plumes in the oceanCO2 bubble plume nearPanarea, Italy. (Photo by GiorgioCaramanna, Nottingham University, UK)B: ‘‘Lion chimney,’ active black smokervents . E: Liquid CO2 droplets from the‘CO2 lake’ in Southern Okinawa. (FumioInagaki et al, PNAS, 2006)

Bio-impact impact modeling test (deep ocean)(Activity index of zooplankton without recovering,Chen et al, JO, 2005pH plumeMc=0.6kg/s; D0 =8.0mmBio-impact indicated by an activity index ofzooplankton; 1.0: normal; 0.0: the worstT=100.3 min; Target ocean: Okinawa

Bio-impact impact modeling test (deep ocean)(Activity index of zooplankton without recovering)Chen et al, JO, 2005pH plumeBio-impact indicated by an activity index ofzooplankton; 1.0: normal; 0.0: the worstMc=0.1kg/s; D0 =8.0mmT=100.3 min

I-1.What does the leaked CO2 look like?,(Dr. Nakajima, NSRI, Japan)

II-a-2. developed andundeveloped Drop/Bubble modelsDepthGas CO2+ HydrateG-H CO2-seawater(Chen et al, Tellus, 2003)Gas CO2G CO2-seawater(Clift & Crace, 1978;Bozzano et al 2000 )L/G CO2 phase change ??-500mL-CO2LCO2-seawater (Clift & Crace,1978; Chen et al 2008 )-800mL-CO2+ Hydrate9 C 12 CL-H CO2-seawater(Chen et al, Tellus, 2003)Temperature

II-a-8: ROV observation of CO2droplet/bubble transitionMonterey Bay, CA (36.7ºN 122.1ºW)Depths: 510m to 380mThe ROV: Ventana

II-a-9. Observation data of CO2 phasetransition and dissolutionRayleigh-Taylor instability:db=4.0 – 5.0 mma ) b) c) d) e) f) g)Droplet/bubble transition (Chen, B & P. Brewer; GRL to be submitted )a) Droplet at 435 m.b) A small bubble appeared at 416m and insitutime of 12:40:11;c) & d) Liquid part shaking at in situ time of12:59:02 and 12:59:29;e) First breakup of liquid part and formed acoalesce part (top) and a droplet at13:01:09.f) Two separated droplet/bubble departed at13:02:22.g) Second break up occurred from the liquidpart of the coalesce droplet.

II-a-3. Sub-model of CO2 solubility &phase diagramHydrateLiquidGas

C[dd e]II-a-4.Cd for droplet/bubble (Chen et al,= f ( M,Eo,Re,γρ) ⋅[f (Re) = 24(1 + 0.125Re20.72dd)/ Re= 1.0 + (5.6419−8.3484×102]de1). Droplet/Bubble with hydrate film covered:2). Droplet/bubble:f ( M , Eo,Rff[12dde]2e, γ ) =10(1 + 1.3M=10(1 + 1.3Mρρ24Re= 1.5617 − 0.8405γf( R= 1+γ (1.45 −1.079×1011/ 6ρe) + 3.1Eo) + Eo1/ 6γRe−31+12M, γρ)[1+36M−3ρ)=ρCρw+ 1.4596×101/31/3] +f2−6(Chen et al, Tellus,2003 and GHGT9,2008)Re2) × 10−4Eo( γρ)1.4(1 + 30M1.51/ 6Re) + Eo1.5

II-a-5: Sub-model of CO2 drop/bubbledissolutionDroplet with hydrate (Chen et al, 2003):1/ 2 3Sh = ( 2 + 0.69 Re Sc1/ )( A / A )e( ADroplet (Clift & Crace, 1978):eff/Aeq)sheff+ 1.4766×10-5eqRe2sh= 1.0+Re(4.67071 .1871×10) × 10-4-3ReSh=2 2.15 λ(1.0 −(2.89+0.5πRe0.64)0.5(Re⋅Sc)0.5µcλ =µwBubble (Clift & Crace, 1978):K=udDf0.0113(0.45 + 0.2d0.50.065 Df0.0694 d−1/4eDe0.5f)0.5d.5

II-a-10.10. Terminal distance and time of LeakedCO2 droplet/bubbleTerminal distance: the distance rising fromleaked depthTerminal time: The time till to completelydissolved

II-b-3. Density change of CO2 solution(Song Y, B.Chen, et al. Energy, 2002)1.0121.013 Saewater30 Underground Water (This Study)40 Underground Water (This Study)50 Underground Water (This Study)Density Ratio1.0081.0061.0041.00210 0.01 0.02 0.03 0.04 0.05CO2 mass fractionThe effects of temperature on density ratio

Tasks of WP 2 of QUICS2.1 Experimental and theoretical study of CO2dynamics in seawater (HW).2.2 Development of small-scale scale two-phaseturbulent ocean modelling (HW, PML)2.3 Parameterisation and modelling of CO2dispersion in sediments. (HW)2.4 Biochemical & ecological models of impacts.(PML, NOCS, SAMS)2.5 Physical coupling of fine scale and regionalhydrodynamic models (POL, HW)