• Aucun résultat trouvé

Transport of water and ions in partially water-saturated porous media. Part 3. Electrical conductivity

N/A
N/A
Protected

Academic year: 2021

Partager "Transport of water and ions in partially water-saturated porous media. Part 3. Electrical conductivity"

Copied!
16
0
0

Texte intégral

(1)

HAL Id: hal-02324271

https://hal.archives-ouvertes.fr/hal-02324271

Submitted on 23 Nov 2020

HAL

is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire

HAL, est

destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Transport of water and ions in partially water-saturated porous media. Part 3. Electrical conductivity

André Revil, Abdellahi Soueid Ahmed, Stephan Matthai

To cite this version:

André Revil, Abdellahi Soueid Ahmed, Stephan Matthai. Transport of water and ions in partially

water-saturated porous media. Part 3. Electrical conductivity. Advances in Water Resources, Elsevier,

2018, 121, pp.97-111. �10.1016/j.advwatres.2018.08.007�. �hal-02324271�

(2)

ContentslistsavailableatScienceDirect

Advances in Water Resources

journalhomepage:www.elsevier.com/locate/advwatres

Transport of water and ions in partially water-saturated porous media. Part 3. Electrical conductivity

André Revila,,AbdellahiSoueidAhmeda,StephanMatthaib

aUniversité Grenoble Alpes, Université Savoie Mont Blanc, CNRS, IRD, IFSTTAR, ISTerre, 38000 Grenoble, France

bDepartment of Infrastructure Engineering, University of Melbourne at Parkville, VIC 3010, Australia

a b s t r a c t

Inhydrogeophysics,weneedareliablepetrophysicalmodelconnecting(non-linearly)theconductivityofaporousmaterial(likeasoil)totheconductivityofthe porewaterandthewatersaturation.Classicalmodelsaretoosimplisticespeciallyatlowsalinities.Theconvexityoftheelectricalconductivityofaporousmaterial asafunctionoftheporewaterconductivityisduetobothatexturaleffectandthedependenceofthespecificsurfaceconductivityonsalinity.Thetexturaleffect arisesbecauseofachangeinthedistributionofporenetworkconductanceswithsalinity.Fromvolumeaveragingarguments,itispossibletoprovideageneral equationfortheconductivityofporousmaterials.ThisapproximationisbasedonaPadé approximant,connectinglowandhighsalinityasymptoticlimitsforwhich arigorousanalysiscanbemadebasedonfourfundamentaltexturalparameters.Wediscusstheconnectionbetweenthisvolumeaveragingmodelandempirical modelsaswellaswiththedifferentialeffectivemedium(DEM)solutionforgranularmedia.TheDEMcapturesthenon-linearbehavioroftheconductivitycurve withonlytwoparametersbutitisstrictlyvalidforgranularmaterialsonly.Wecomparethemodelswithfiniteelementcomputationsusingtwothree-dimensional poregeometrieswithcontinuousanddiscontinuoussolidsurfaces,respectively.Finallythemodelsarecomparedtoexperimentaldata.

1. Introduction

Seventysixyearsago,Archie(1942)developedasimpleequation tointerpretresistivitywelllogsinclean(clay-free)formations.Inhis model,theelectricalconductivityofaporousmedium𝜎(inSm1)is writtenas

𝜎= 1

𝐹𝜎𝑤, (1)

whereF=𝜙m(dimensionless)describesthe(resistivity)electricalfor- mationfactor,𝜙denotestheconnectedporosity,and𝜎w(inSm1)de- notestheconductivityoftheporewater,whichdependsinturnonthe salinityoftheporewaterandtemperature.Theexponentm(dimension- less)iscalledthecementationexponentinthepetroleumengineering community.Forasetofsphericalparticles,mincreaseswithcementa- tionfrom1.1–1.3forcolloidalsuspensionsandunconsolidated sands to1.7–2.1forconsolidatedsandstoneswherethegrainsarebondedto- getherbycements(seeFriedman,2005).However,thevalueofmalso varieswithgrainshape(e.g.,Jacksonetal.,1978)andthisterminol- ogyis therefore confusing. Theterm“Archie porosity exponent” (or firstArchieexponent) shouldbe preferred. Notethatthepower law relationshipbetweenelectricalconductivityandporositywasalready knownmuchbeforeArchie(1942),seeforinstanceBruggeman(1935). ThisversionofArchie’slawF=𝜙mshouldalsonotbeconfusedwith Archie’slawwrittenasF=a𝜙m(developedlaterinthe50s)andusu- allyusedtofitaformationfactor/porositydatasetinwhicheachpoint

IntendedforpublicationinAdvancesinWaterResources.

Correspondingauthorat:Université SavoieMontBlanc,ISTerre,Technolac,Bat.Belledonnes,ruedulacd’Annecy,73370LeBourget-du-Lac,France.

E-mailaddresses:andre.revil@univ-smb.fr(A.Revil),abdellahi.soueid-ahmed@univ-smb.fr(A.SoueidAhmed),stephan.matthai@unimelb.edu.au(S.Matthai).

correspondstoadistinctsample,allthesamplebeingfromthesamefa- cies(e.g.,Winsauer,1952).

Electricalconductivitytomographyisakeytechniqueinhydrogeo- physics(Binleyetal.,2015).Unfortunately,Eq.(1)remainsthebasis fortheinterpretationofresistivitydatainalargenumberofpublica- tionsinhydrogeophysics.Backinthe50sand60s(e.g.,Patnodeand Wyllie,1950,andWyllieandSouthwick,1954),petroleumengineers recognizedthatasecondcontribution,calledsurfaceconductivity,was atplay,especiallyinshalyformations.WaxmanandSmits(1968)de- velopedasimpleequationwrittenatthispointas(seealsoCremersand Laudelout, 1965,Rhoadesetal.,1976, MualemandFriedman 1991, Revil,2017a,b),

𝜎= 1

𝐹𝜎𝑤+𝜎𝑆. (2)

InEq.(2),𝜎S(Sm1)denotesthisextraconductivitytermcalledsur- faceconductivityorsurfaceconductanceoftheclay(e.g.,Cremersetal., 1966).Theform ofEq. (2)can be tracedbacktoPatnodeandWyl- lie(1950)andWyllieandSouthwick(1954).Intheseminalmodelof WaxmanandSmits(1968),thesurfaceconductivityiswrittenas𝜎𝑆= (̂𝐵𝑄𝑉)∕𝐹 where ̂𝐵(intherange2.0to4.8×108m2V1s1,Na+,at 25°C)denotestheapparentmobilityforthechargecarriers(calledcoun- terions)responsibleforthesurfaceconductivity(seealsodiscussionsin Reviletal.,1998).ThetermQV(inCm3)denotesanexcessofcharge (ofthecounterions)perunitporevolumeduetothecationexchange capacityofthesurfaceofminerals.Claymineralsarecharacterizedby

https://doi.org/10.1016/j.advwatres.2018.08.007

Received11February2018;Receivedinrevisedform13August2018;Accepted14August2018 Availableonline18August2018

0309-1708/© 2018ElsevierLtd.Allrightsreserved.

(3)

Fig.1.Anatomyoftheelectricalconductivitycurveofaporousbodyasafunc- tionoftheporewaterconductivity.Thisshowsanexampleofnon-linearbehav- iorbetweentheconductivitydataoftheporousmaterialandtheconductivity oftheporewater.DatafromShainbergetal.(1980).

strongCECvalues,butpuresilicaischaracterizedbyareactivemineral surfaceresponsibleforsurfaceconductivity(seeReviletal.,2014).The surfaceconductivityisduetoconductionintheelectricaldoublelayer (Sternanddiffuselayers)coatingthesurfaceoftheseminerals.

AlthoughEq. (2)suggestsa linearrelationship between thecon- ductivity ofthe porousbody 𝜎 andthepore waterconductivity 𝜎w, Waxman and Smits (1968) recognized that, at low salinity, theac- tualbehaviorof𝜎isnon-linear.Theyobservedthatthisnon-linearbe- haviorisespeciallypronouncedformaterialscharacterizedbyahigh cationexchangecapacity(CEC).Toaccountforthisobservation,Wax- manandSmitsintroducedasalinitydependentcationmobility ̂𝐵(𝜎𝑤)=

̂𝐵[10.6exp(−𝜎𝑤∕0.013)](with𝜎w expressedinSm1).However,this explanationisunphysical sincethemobilityofthecationsis notex- pectedtodecreasewithdecreasingsalinity.

Manyerrorscanbefoundintheliteratureregardingtheimproper useofelectricalconductivityequations.Notably,whenEq.(1)isused whileEq.(2)shouldbeappliedinstead,thisresultsinunphysicalvalues anddependenciesofArchie’sporosityexponent.Forinstance,Salemand Chilingarian(1999)observed(erroneously)thatthisexponentissalin- itydependentalthoughitispurelyatexturalparameter.Thelistofsuch errorsis,unfortunately,verylongbothinpetroleumengineeringandin hydrogeophysics.Thatsaid,whileEq.(2)hasincreasinglybeenusedin hydrogeophysics,wecanwonderhowwellthisequationmodelsfresh- waterenvironmentsbecauseitdoesnotcapturethelow-salinitynon- linearbehavior(e.g.Shainbergetal.,1980,forsoils)-atleastnotina physicallymeaningfulway.

Intheprevioustwopapersofthisseries(Revil,2017a,b),wede- velopatheoryofionictransportinporousmediainunsaturatedcondi- tions.Wekeptthetheorytothehighsalinityasymptoticbehaviorfor whichsurfaceeffectsaretreatedasaperturbationofthetransportinthe connectedporespace.Inthissituation,thetransportofionsismostly affectedbythetortuosityofthebulkporespace.Inthepresentpaper, weshowthatthetortuosityaffectingthetransportoftheionschanges withthesalinity.Themaingoalofthepresentpaperistodemonstrateto thehydrogeophysicalcommunitythatbeyondEq.(2),thereisanother realmworthexploringforfreshwaterenvironments.Itmaybeimpor- tantforsoils(Rhoadesetal.,1976;Nadler,1982;1991;Binleyetal., 2015),geosyntheticclayliners(Abuel-NagaandBouazza,2016),clay suspensions(vanOlphen,1957;vanOlphenandWaxman,1958),the geophysicalmonitoringofCO2sequestration(AlHagrey,2012;Börner etal.,2013)andreactivetransportmodeling(Day-Lewisetal.,2017), andthestudyofgashydratesandpermafrost(Spangenberg,2001;Prieg- nitzetal.,2015)tociteafewexamples.Wereviewknowledgeabout thenon-linearbehaviorofelectricalconductivityatlowsalinities(see

forinstanceBruggeman,1935,WyllieandSouthwick,1954,Bussian, 1983,LimaandSharma,1990,Schwartzetal.,1989a,b),startingwitha volumeaveragingapproachtoanon-linearconductivityequationvalid foranytypeofporousmaterials.Ouranalysiswillstartwiththeworks of Johnsonetal.(1986)andJohnsonandSen(1988).Highandlow salinityasymptoticlimitscanberigorouslyderivedandtiedtogether usingaPadé approximant(i.e.,aratioofpolynomials).Then,wewill explorehowthisequationcanbeusedtoexplaintheWyllieandSouth- wick(1954),WaxmanandSmits(1968),anddualwater(Clavieretal., 1984) models.Wewillalsocomparethemodeltothedifferentialef- fectivemediumsolutionforapackofspherescoatedbyanelectrical doublelayerandimmersedin“background” (pore)water(seeBussian, 1983,LimaandSharma,1990).Finally,thesemodelswillbecompared withnumericalsimulationsattheporescaleatsaturatedandunsatu- ratedconditions.Wewillalsoderivenewexpressionsfortheconduc- tivityandnewrelationshipsbetweenthemodelsandcheckhowthese modelscomparewithexperimentaldata.

2. Volumeaveragingapproach

Fig.1showsoneexampleofanon-linearrelationshipbetweenthe totalelectricalandtheporewaterconductivity.Weseethattheconduc- tivitycurveischaracterizedbyanisoconductivitypointforwhichthe conductivityoftheporousmediumisequaltothatoftheporewater, apropertybroadlyanalyzedinthecolloidalscienceofclaysuspensions (e.g.,Street,1963,CremersandLaudelout,1965,ShainbergandLevy, 1975)andporousmedia(Bussian,1983;LimaandSharma,1990;Revil etal.,1998).Inthissection,wesummarizethefindingsregardingthe electricalconductivityofporousmediaandespeciallythosefromthe originalworksofJohnsonetal.(1986)andJohnsonandSen(1988)in anattempttoexplainthenon-linearbehaviorshowninFig.1.Westart byconsideringtwoasymptoticlimitsfortheconductivityequationcor- respondingathighsalinities(i.e.,higherthanthesalinitycorrespond- ingtotheisoconductivitypoint)andatlowsalinities(i.e.,forsalini- tiesmuchsmallerthanthesalinitycorrespondingtotheisoconductivity point).

2.1. Theformationfactor

Westartouranalysisbyconsideringthelocalconductivityproblem intheabsenceofanelectricaldoublelayeraroundthesolidphase.The solidphaseisinsulating.Theporespaceisfilledbyanelectrolyteofcon- ductivity𝜎w.Theconstitutive(localOhm’slaw)andcontinuityequation forthecurrentdensityaregivenbyj=𝜎web inVp,(i.e.,inthepore space)and·j=0onS(i.e.,atthemineralwaterinterface)where eb=𝜓b denotesthelocalelectricalfield(Vm1),and𝜓bthelocal electricalpotential(inV),jdenotesthelocalcurrentdensity(Am2).

Thesubscript“b” referstothefactthatinthissituation,theelectrical fieldisgovernedbythedistributionofthebulkconductances(Bernabé andRevil,1995).Intheabsenceofsurfaceconductivity,theconduction problemreducesto:

2𝜓𝑏=0in𝑉𝑝 (3)

̂𝐧𝐞𝑏=0on𝑆 (4)

𝜓𝑏=

{ΔΨ at𝑧=𝐿

0at𝑧=0 (5)

whereVp andSdenotetheporevolumeand(specific)interfacearea betweenthesolidandthefluidrespectively;L(inm)denotesthelength ofthecylindricalrepresentativevolumeinthedirectionoftheapplied macroscopicelectrical field𝐄=−(ΔΨ∕𝐿)̃𝐳(inVm1), ̃𝐳denotesthe unitvectorinthedirectionoftheelectricalfield,ΔΨcorrespondstothe differenceofelectricalpotentialbetweentheend-facesoftherepresen- tativevolume,and̂𝐧denotestheunitvectornormaltothesurfaceof

(4)

thegrains.Thisboundaryvalueproblemcanbewrittenintermsofa normalizedelectricalpotentialΓb foracylindricalrepresentativeele- mentaryvolumeofporousmaterial(Pride,1994),asfollows:

2Γ𝑏=0in𝑉𝑝 (6)

̂𝐧Γ𝑏=0on𝑆 (7)

Γ𝑏(ΔΨ 𝐿

)−1

𝜓𝑏=

{𝐿at𝑧=𝐿

0at𝑧=0 (8)

Inabsenceofsurfaceconductivity,theformationfactorF=𝜎w/𝜎(see Eq.(1))isobtainedbysummingupthelocalJouledissipationofenergy (JohnsonandSen(1988);RevilandCathles,1999).Themacroscopic dissipationofenergycanbewrittenasDJEwhilethelocaldissipation ofenergyisdjeb whereJandjdenotesthemacroscopicandlocal currentdensities,respectively.Thereforewehave

𝐷= 1

𝑉 𝑉𝑝𝑑(𝐞𝑏)𝑑𝑉𝑝, (9)

𝜎(ΔΨ 𝐿

)2

= 1

𝑉 𝑉𝑝𝜎𝑤||𝐞𝑏||2𝑑𝑉𝑝, (10) 1

𝐹 = (ΔΨ

𝐿 )−21

𝑉 𝑉𝑝||𝐞𝑏||2𝑑𝑉𝑝, (11) 1

𝐹 = 1

𝑉 𝑉𝑝||Γ𝑏||2𝑑𝑉𝑝, (12)

whereVisthetotalvolumeoftheconsideredrepresentativeelementary volumeandwhere𝜎wisconsideredconstantovertheporespace.This principleisalwaysvalidsincethemacroscopicdissipationofenergyat thescaleofarepresentativeelementaryvolumeisthesumofthelocal dissipations(herelimitedtotheporespace).FromEq.(12),theinverse oftheformationfactorappearstocorrespondtoaneffectiveporosity (theconnectedporosity,apurelygeometricalquantity,is𝜙=Vp/V).

Foranytypeofporousmaterial,thiseffectiveporosityisbuiltbygiving someweighttothethroatsandnoweighttodead-endsofpores.This canbeshownforinstanceforthematerialsketchedinFig.2forwhich thenumericalsimulationisperformedathighsalinity.Thedistribution oftheΓb-fieldisshowninFig.3.This∇Γb-fieldisstronginporethroats andnullindead-ends(Section6belowexplainsindetailhowthesecom- putationsareperformed).Thiswouldremaintrueforanytypeofpore network,idealizedornot.Theformationfactorgoestoinfinityaspores becomedisconnected(andverylargenearthepercolationthreshold).

Thispaperdoesnotfocusonthespecificphysicsofelectricalconduc- tionclosetopercolation.

2.2. Highsalinityasymptoticlimit

Nowwediscusstheimpactofanelectricaldoublelayer,coatingthe surfaceofthegrains(Fig.4a),contributingextrasurfaceconductivity totheporousmedium.WeintroduceΣS(inS),thespecificsurfacecon- ductivity/conductanceoftheelectricaldoublelayer(inS)(seeFig.4, ande.g.,vanOlphen,1957,JohnsonandSen,1988),

Σ𝑆=

0

(𝜎(𝑥)𝜎𝑤)

𝑑𝑥, (13)

where𝜎(x) denotesthelocalconductivityin thevicinityof themin- eralsurface/porewaterinterface(seedetailsinFig.4b).Outsidethe electricaldoublelayer,𝜎(x)=𝜎wwherexdenotesthelocalcoordinates normaltotheporewater/solidinterface.Inthethindoublelayeras- sumption,thelocalelectricalconductivitydistributioniswrittenas𝜎(x)

=𝜎w+ΣS𝛿(x)where𝛿(x)denotesthedeltafunctioncharacterizingthe positionofthemineralsurface.Toobtainthehighsalinityasymptotic behavioroftheconductivity,wereplacetheconductivityofthepore

Fig.2. Sketchofthefirst3Dporousmaterialusedtoillustratetheconductivity problem.a.Geometrywithboundaryconditions.Thematerialismadeofan insulatingmineral(nullconductivity)coatedbyaconductiveelectricaldouble layer(conductivityof1Sm1).The(connected)porespacecontainsondead- endandonethroat.TheelectrodesAandBareusedtoinjectthecurrentI (inA)whilealltheotherboundariesareinsulating.Thevectorndenotesthe normalunitvectortotheexternalboundariesoftheporousbodyand𝜓the electricalpotential.b.MeshusedforthefiniteelementcalculationswithComsol MultiphysicsoftheOhmicconductivityproblem.

water𝜎wbythelocalconductivity𝜎(x)assumingthatinthishighsalin- itylimit,theelectricalfieldremainsroughlythesame.Thisyields 𝜎(ΔΨ

𝐿 )2

= 1

𝑉 𝑉𝑝𝜎(𝑥)||𝐞𝑏||2𝑑𝑉𝑝. (14) Athighsalinity,theelectricalfieldiscontrolledbythedistribution ofbulkconductancesandwereplacethelocalconductivitybyitsex- pressionobtainedwiththethin doublelayerassumption,i.e.,𝜎(x)= 𝜎w+ΣS𝛿(x).Thisyields(Johnsonetal.,1986),

𝜎(ΔΨ 𝐿

)2

= 1 𝑉 𝑉𝑝

[𝜎𝑤+Σ𝑆𝛿(𝑥)]||𝐞𝑏||2𝑑𝑉𝑝, (15)

𝜎(ΔΨ 𝐿

)2

= 𝜎𝑤

𝑉 𝑉𝑝||𝐞𝑏||2𝑑𝑉𝑝+Σ𝑆

𝑉 𝑉𝑝𝛿(𝑥)||𝐞𝑏||2𝑑𝑉𝑝, (16)

Références

Documents relatifs

Symbols represent stochastic simulations based on the Gillespie algorithm (10 7 initial mobile particles and 5 · 10 6 immobile particles per reactive region, averaged over 10

We report an experimental study of the thermal conductivity of a granular material when water is added to the system. Because of the formation of liquid bridges at the contacts

The higher the completion temperature of the heating scans, the larger the conductivity increase and the lower the activation energy (curves 2 and 3 in Fig. EPR

This paper describes the application of electrical measurements to monitor the extraction (movement of water from the mortar) of water from calcium lime, natural hydraulic lime

Lower values of the hydraulic conductivity were detected for CaCl 2 treated suspensions compared to CaO treated suspensions, as a consequence of the lower degree of aggregation

Using observed values of population structure and estimated population size from VPA and individual growth per year and year class for small type and large type whitefish

Alexandre, Bertrand, Charles et Denis sont quatre amis. Nous avons sur eux les renseignements suivants : Alexandre rencontre souvent le professeur

The different water contents investigated in this study must show the effect of water on both Na diffusion and electrical conductivity (Fig. Our findings concerning sodium as