H HM THM C P ocesses Na Ba - andra.fr · the possible swelling of soil during the re-saturation phase, a specific saturation procedure was proposed and applied (Figure 1). Consolidation

Poster [THMN]

H,HM, and THM-C Processesin Natural Barriers

r

P/THMN/1

Page 593INTERNATIONAL MEETING, SEPTEMBER 17...>...18, 2007, LILLE, FRANCECLAYS IN NATURAL & ENGINEERED BARRIERSFOR RADIOACTIVE WASTE CONFINEMENT

THREE DIMENSIONAL ANALYSESOF COMBINED GAS, HEAT AND NUCLIDE

TRANSPORT IN A REPOSITORYCONSIDERING COUPLED THERMO-HYDRO

GEOMECHANICAL PROCESSES

V. Javeri1

1. Gesellschaft fr Anlagen- und Reaktorsicherheit (GRS) mbH, Schwertnergasse 1, 50667 Kln,Germany ([email protected])

INTRODUCTION

To assess the long term safety of a repository for the heat generating nuclear waste in a deep geologicalformation, generally groundwater flow into the repository is postulated. The water can react with theradioactive waste or with its containers and can gradually disassemble them. The radioactive substancesafter being dissolved in water can be transported out of the repository. The transport of heat andcontaminated fluid can be enhanced by the gas generation in the repository building. The gas, chieflyhydrogen, is produced due to the corrosion of metallic materials. The temperature distribution and thepressure build-up in the repository as well as the subsequent transport of the contaminated two phase fluidcan influence the geomechanical behaviour of the host rock. The transient stress situation can in turn affectthe hydrogeological properties of the host rock and thus the two phase flow behaviour.

ANALYSES

To analyse the coupled thermo-hydro-geomechanical processes and their influence on gas and nuclidetransport in a non-isothermal two phase flow configuration, simplified two and three dimensional modelsof a repository for nuclear waste in a deep host rock are considered. The system consists of three materialdomains: elastic barrier rock on the top, host rock (plastic clay) at the bottom, repository backfilled withan elastic material containing nuclide, heat and gas sources embedded in host rock. The analyses of highlynon-linear processes are performed for the post closure phase up to 104 years by using a modified versionof computer code TOUGH2/EOS7 to simulate the two phase flow, nuclide and heat transport and the finitedifference code FLAC3D to determine the geomechanical behaviour. The essential assumption for thecoupling of these codes is an identical numerical mesh grid for two codes. Within sequential coupling ofTOUGH2 and FLAC3D, the codes are executed one after another at each time step. The transienttemperature and the two phase pressure distributions computed by TOUGH2 are transferred to FLAC3Dto determine the three dimensional stress situation, which is then transferred back to TOUGH2 to continuethe two phase flow and heat and nuclide transport calculation including hydrogeological properties, suchas porosity, permeability and capillary pressure, depending upon mean effective stress. Thus, the main statevariables are exchanged between the codes at each time step.

CONCLUSIONS

The scoping coupled three dimensional thermo-hydro-geomechanical analyses, which indicate thattransport behaviour of the contaminated two phase fluid in a non-isothermal configuration is influencedsignificantly by the transient stress conditions, can quantify the safety margin related to hydro-fracturingand dilatancy in the host rock due to fluid pressure build-up in the repository, if the hydrogeologicalproperties, such as porosity, permeability and capillary pressure, depending upon the mean effective stressare employed. The procedure proposed for the coupling of the thermo-hydro-geological code TOUGH2 andthe thermo-mechanical code FLAC3D can be employed to estimate the impacts of the coupled processesreasonably and can also be used to define bounding analyses.

P/THMN/2


EXPERIMENTAL STUDYON THE THERMO-HYDRO-MECHANICAL

BEHAVIOUR OF BOOM CLAYTrung Tinh Le1, Yu-Jun Cui1, Pierre Delage1 and Xiang-Ling Li2

1. ENPC/CERMES, 6 et 8, av. Blaise Pascal, Cit Descartes, Champs-sur-Marne, 77455 Marne laVale cedex 2, France, ([email protected], [email protected], [email protected])

2. ESV EURIDICE GIE, c/o SCK-CEN, Boeretang 200, BE-2400 Mol ([email protected])

INTRODUCTIONTo study the thermal-hydro-mechanical properties of Boom clay, a stiff clay from the UndergroundResearch Laboratory (URL) at Mol (Belgium), an experimental investigation was carried out. To minimisethe possible swelling of soil during the re-saturation phase, a specific saturation procedure was proposedand applied (Figure 1). Consolidation results were analysed to clarify the influence of the saturation phaseon soil characteristics. A compression velocity much lower than that in the literature seemed to benecessary to guarantee the full consolidation of the soil samples. Non-isothermal tests showed a heatingeffect on the creep behaviour of soil.

EXPERIMENTAL METHODOLOGYIn-situ measurements during the excavation of the URL (223 m underground) gave an average naturalwater content comprised between 24.3% and 25.9%. However the water content of the soil samples whenopening the package in the laboratory was around 23%, showing the necessity to re-saturate the soil samplebefore each experiment.

Based on the system reported by Delage et al. (2000), two experimental systems permitting high pressuretriaxial tests at controlled temperature (up to 100C) have been developed. The confining pressure, backpressure, and deviator pressure are applied by three volume/pressure controllers (pmax=32 MPa). During theexperiment, volume changes are measured by monitoring the volume changes of the cell water or porewater.

A specific saturation procedure was applied before each experiment (Figure 1). At the beginning, the celltemperature was regulated at 25 C and a confining pressure of 100 kPa was applied to establish the initial

Figure 1: Stress path followed during the saturation phase.

0

500

1000

1500

2000

2500

3000

3500

4000

0 2 4 6 8 10 12 14 16 18 20 22 24

Time (day)

P(kPa)

Confining pressure

Back pressure (GDS)

Back pressure (transducer)

Effective pressure

P/THMN/2

Page 596 INTERNATIONAL MEETING, SEPTEMBER 17...>...18, 2007, LILLE, FRANCECLAYS IN NATURAL & ENGINEERED BARRIERS

FOR RADIOACTIVE WASTE CONFINEMENT

system stabilisation. After that, confining pressure was increased in slope of 0.5 kPa/mn to 2500 kPa. Thispractice aims to bring the soil back to its in situ condition.When the soil volume was stabilised at 2500 kPaof confining pressure, Boom Clay Synthetic Water (BCWS) was injected at a pressure of 100 kPa by theback pressure controller from the samples bottom. The soil swelling was observed by the volume of theconfining pressure controller, while water entering the soil sample was measured by the back pressurecontroller. After around 8 days, BCSW was injected from the head of the soil sample. The sample wasafterward saturated by increasing simultaneously in steps of 250 kPa the confining pressure and backpressure to respectively 3500 kPa and 1000 kPa. During this phase, the effective pressures was always2500 kPa and Skempton B value was measured.

Two experiments were however carried out following the saturation procedure as suggested by Sultan(1997), i.e. saturation under low effective pressure that equals to 100 kPa.

After the saturation phase, soil samples were then compressed in a isotropic fashion under different rates(0.1, 0.5 kPa/mn, or in steps) at two constant temperatures of 25C and 80C. The heating of soil samplewas also carried out under different heating rates (0.1, 0.3 C/h, or in steps).

RESULTS AND INTERPRETATION

Soil swelling during the saturationDuring water injection, it was observed that the volume of water adsorbed in the soil samples was fivetimes more than the soil swelling. This difference shows that water entering the sample on the one handswells the clay pellets, on the other hand fills the existing voids or dissolves the remaining air bulbs, thussaturates the samples. The application of the proposed saturation procedure gave very small soil swelling:0.8 % to 1.8%, as compared to the swelling of 4% - 5% obtained under low effective pressure. Theseswelling values are comparable to those obtained by Coll (2005) who followed the similar saturationprocedure. It shows that the soil volumetric strain depends significantly on the experiment boundaryconditions.

Mechanical and thermal consolidationThe elastic limit values obtained from mechanical consolidation tests have been found to be stronglydependent of saturation stress. It is in the order of 0.8 MPa under low stress (as Sultan 1997) and of 2.5MPa under high stress (as Coll 2005). The results presented a good repeatability though experiments werecarried out on soil samples of different ages.

The volume change was time-dependant at a constant effective pressure. The applied compression velocityof 0.5 kPa/mn, which was reported to be slow enough to guarantee a complete consolidation by Sultan(1997), has been found to be too high. This remark again confirms the influence of saturation procedureon the soil behaviour. A consolidation velocity of 0.1 kPa/mn was then applied in order to guarantee fullsoil consolidation. Interestingly, when compressing initially unsaturated soil samples 100 to 2500 kPaunder a velocity of 0.5 kPa/mn, no time-dependant volume change behaviour was observed, showing asignificant effect of water saturation on the soil volumetric behaviour.

Heating tests presented an important creep behaviour of natural Boom clay. The volume change at aconstant temperature higher than 30C showed to be relatively significant (10-8 s-1). Interestingly, the higherthe temperature, the more significant the creep is.

ReferencesColl, C. (2005). Endommagement des roches argileuses et permabilit induite au voisinage douvrages

souterrains. Ph.D. thesis, lUniversit Joseph Fourier-Grenoble 1, France.

Delage, P., Sultan, N. and Cui, Y.J. (2000). On the thermal consolidation of Boom clay. CanadianGeotechnical Journal 37, 343-354.

Sultan, N. (1997). Etude du comportement thermo-mcanique de l'argile de Boom: expriences etmodlisation. Ph.D. thesis, Ecole Nationale des Ponts et Chausses, Paris.

P/THMN/3


HYDRAULIC IN-SITU MEASUREMENTSIN THE OPALINUS CLAY IN THE FRAME

OF A HEATER TEST PERFORMEDAT THE MONT TERRI URL

Klaus Wieczorek, Norbert Jockwer

Gesellschaft fr Anlagen- und Reaktorsicherheit (GRS) mbH, Theodor-Heuss-Strasse 4,38122 Braunschweig, Germany ([email protected], [email protected])

INTRODUCTIONFrom October 2003 to December 2005, the HE-D heating experiment was performed at the Mont TerriRock Laboratory by ANDRA and GRS in order to investigate thermal effects on clay formations hostinga repository for high-level radioactive waste (Wileveau and Rothfuchs, 2003). The horizontal heaterborehole was drilled in March 2004. Heating started on April 6, 2004 with a total power of 650W. OnJuly 6, 2004, heater power was increased to 1950W. The maximum temperature of 100 C was reachedat the heater/rock interface. On March 13, 2005 the heaters were shut down and the cool-down phasebegan.

GRS task in the project was, beside laboratory investigations and numerical modelling performed byseveral teams in parallel, the measurement of rock temperature, pore water pressure, and gas migration.

EXPERIMENTAL CONCEPTTwo arrays of boreholes of 20 mm diameter for pore-water pressure (PWP) and temperature measurementswere drilled perpendicular to the heater borehole from the MI Niche into the test region in October/November 2003, prior to drilling of the heater borehole. The PWP boreholes BHE-D07 to BHE-D11(Array 1) are positioned above the axis of heater borehole, whereas the boreholes BHE-D12 to BHE-D17(Array 2) are located below the heater. The boreholes were instrumented with minipackers installed at thebottom to give a chamber with a minimized volume and Pt-100 temperature sensors. The voids remainingin the borehole between the packer and the niche wall were sealed with resin two weeks after installation.The test intervals were flushed with synthetic pore water and the air was removed.

After cool-down, when a steady temperature and pore pressure state had been reached, the minipackerintervals were used for measurement of rock permeability to water and of gas entry pressure.

For investigating the gas pressure development in sealed boreholes and determining the gas permeabilityof the host rock in the different stages of the experiment, two sets of boreholes (BHE-D18 to BHE-D23)with a diameter of 76 mm and a length of about 10 m were drilled from the MI Niche perpendicular to theheater borehole into the test region. Each set consists of three boreholes 30 to 40 cm distant to each other.They pass the heater borehole at a distance of 0.70 m and 1.55 m, respectively. All boreholes were sealedwith quadruple packer systems for testing in separate intervals.

RESULTS AND INTERPRETATIONPrior to heating, the rock temperature at 8 m from the MI-Niche ranged between 14C and 15C. Close tothe MI-Niche the temperature rose up to 19C. During the first heating phase (heater power 650W; April6, 2004 - July 6, 2004) the temperature increased from 14 to 21C above the heater and 25C below theheater. Above the heater the temperature increased to 41C in phase 2 (heater power 1950W), while belowthe heater the temperature reached 53C at the end of the heating phase. The difference in the temperaturelevels between Array 1 and Array 2 hints to an anisotropy of the thermal conductivity. With heater shut-down on March 13, 2005 temperature started to decrease at all measurement locations previously affected

P/THMN/3



by heating. By mid September 2005, tem-perature ranged between 17 C and 19 C atboth arrays 1 and 2.

The pore pressure data recorded throughoutthe experiment by Array 2 are shown inFigure 1. Before start of the heating phase 1the pore-water pressure in the test intervalswas almost stationary. The test intervalsclose to the heater reacted to both heatingphases with a marked pressure increase anda subsequent partial relaxation (D14, D15,D16). Interval D17 which is far from theheater showed a slower response. InDecember 2004, pore pressure wasdisturbed by a heater failure. During thecool-down phase starting in March 2005the pore pressure returned smoothly to lowvalues; no hints to fracturing were detected.

After cool-down, water injection testsyielded permeability values between 10-20m2 and 10-19m2. Gas entry pressuremeasurements resulted in typical valuesbetween 0.8MPa and 1.8MPa.

The central interval of two of the boreholesfor gas migration measurements wereinflated with nitrogen to 0.9 MPa prior tostart-up of heating. The pressures increasedslightly as a result of thermal expansion and

gas release during the first heating phase. Increase of the heater power during the second phase led to anacceleration of the pressure increase. No significant gas flow into the rock was detected during heating.After cool-down of the rock, gas injection tests were performed by stepwise increase of gas pressure. Afirst gas flow was detected at pressures above 1.5MPa, and above 3 to 3.5MPa led to gas fracturing (seeFigure 2). The pressure response in the surrounding test intervals showed that the fracture was orientedparallel to the bedding plane of the clay formation.

ACKNOWLEDGEMENTSThe GRS work was funded by the German Federal Ministry of Economics and Technology (BMWi) undercontract No. 02E9773. The French Agence Nationale Pour la Gestion de Dchets Radioactifs (ANDRA)coordinated the experiment. The authors thank for the support.

ReferencesWileveau, Y., Rothfuchs, T. (2003): HE-D Experiment: Test Plan. Mont Terri Project, Technical Note2004-20.

Figure 1: Evolution of pore-water pressure in Array 2(below the heater).

Figure 2: Gas injection tests in interval 3 of borehole BHE-D21.

0

500

1000

1500

2000

2500

3000

3500

4000

27.09.2005 12:00 28.09.2005 00:00 28.09.2005 12:00 29.09.2005 00:00 29.09.2005 12:00 30.09.2005 00:00

Pressure/kPa

D21P3

Gas flow afterexceeding entrypressure

Gas fracextension

Constant headinjection

Gas frac

P/THMN/4


HYDRO-MECHANICAL BEHAVIOUROF NATURAL BOOM CLAY

IN CONTROLLED-SUCTION TESTS

E. Romero1*, A. Lima1, A. Gens1 & X.L. Li4

1. Department of Geotechnical Engineering and Geosciences, Universitat Politcnica de Catalunya,c/ Jordi Girona 1-3, Campus Nord, Mdulo D2, 08034 Barcelona, Spain([email protected]*; [email protected]; [email protected])

* Corresponding author2. EIG, EURIDICE, Boeretang 200, B-2400 Mol, Belgium ([email protected])

INTRODUCTIONSince 1974, Belgium investigates the design for disposal of its High Level Radioactive Waste in a deepclay formation, the Boom clay. This Tertiary clay formation is the subject of extensive research dealingwith all phenomena that may possibly affect the performance of this potential disposal site during galleryconstruction and final operation. Specifically, during excavation of the deep underground facility some de-saturation can be induced on the formation by ventilation in the galleries. In addition, the initial suction ofthe engineered barrier surrounding the waste may also induce some drying to the clay formation. Regardingthese aspects, the study of the unsaturated state becomes mandatory to understand the response of theinitially saturated clay under excavation and its sensitivity to de-saturation effects and suction cycling (re-saturation by water inflow from the clay formation).

There are a number of laboratory results concerning the saturated behaviour of natural Boom clay andrecent works on this area are described in Sultan (1997) and Coll (2005). In contrast, experimentalinformation concerning partially saturated states on natural Boom clay is very limited and concentrated onthe characterisation of compacted states (Romero et al. 1999). There is a need, therefore, to examine moresystematically partial saturation consequences on the hydro-mechanical response of natural Boom clayunder controlled-suction conditions. The paper explores these consequences by presenting the results of acomprehensive test programme carried out under oedometer and isotropic stress conditions usingcontrolled-suction techniques. Hydro-mechanical test results were interpreted and discussed within theframework of an elastoplastic model (Alonso et al. 1990).

EXPERIMENTAL PROGRAMME AND SELECTED RESULTSThe paper describes and presents the relevant results of an experimental programme aimed at studying thecoupled hydro-mechanical response of natural Boom clay under partially saturated states. The initial stateof the sample matched the in situ natural conditions, which implied following a rigid preconditioningprotocol prior to testing. As preliminary work, the retention curve of the material was determined and themicrostructure of the material characterised by mercury intrusion porosimetry. Different stress paths underisotropic and oedometer conditions were followed to cover a wide range of partially saturated features(loading / unloading paths at constant suction and drying / wetting paths at constant stress). Vapour andliquid transfer techniques were applied to control total and matric suction, respectively. Figure 1(a) showsthe shrinkage undergone by the natural material on drying under oedometer conditions at a constant verticalnet stress of 2.5 MPa (equivalent to the in situ effective stress). In this case, the saturated sample was letto equalise to a total suction of 20MPa higher than the air-entry value of the material estimated in 5MPa,which was applied through vapour equilibrium technique. As complementary information, Figure 1(b)shows the shrinkage undergone by the material under isotropic stress conditions (constant mean net stressof 2.0 MPa) when applying a drying path with axis translation (a liquid phase transfer technique). In thiscase, a fully instrumented triaxial cell with local axial and radial transducers was used to monitor theevolution of shear and volumetric strains undergone by the material when subject to matric suction increase

P/THMN/4



at controlled rate (matric suction increased from 10 kPa to 500 kPa in 16 days). As observed, the shrinkageevolution was completely isotropic. Despite the air-entry value was not exceeded, the shrinkage of thematerial was only slightly lower to that previously presented, stating the efficiency of the liquid phasetransfer technique.

ReferencesAlonso, E.E., Gens, A. & Josa, A. (1990). A constitutive model for partially saturated soils. Getechnique40(3): 405-430.

Coll, C. (2005). Endommagement des roches argileuses et permabilit induite au voisinage douvragessouterrains. Thse de doctorat de lUniversite Joseph Fourier, Grenoble.

Romero, E., Gens, A. & Lloret, A. (1999). Water permeability, water retention and microstructure ofunsaturated Boom clay. Engineering Geology 54: 117-127.

Sultan N. (1997). Etude du comportement thermo-mcanique de largile de Boom: exprience etmodlisation. Thse de doctorat de lENPC, Paris.

Figure 1: (a) Time evolution of shrinkage (volumetric strain) at a constant vertical net stress of 2.5 MPa(oedometer conditions). The saturated Boom clay was allowed to equalise to a total suction of 20 MPausing vapour equilibrium technique. (b) Time evolution of shear and volumetric strains along a drying pathunder isotropic stress conditions (constant mean net stress of 2.0 MPa). Starting from natural conditions,matric suction was increased at controlled rate to 0.5 MPa using axis translation technique.

0.01 0.1 1 10 100 1000 10000 100000

Time (min)

0.4

0.3

0.2

0.1

0

22C

Total suction:Applied:20 MPa with 210.2 mS/cm of NaCl solutionFinal measurement:17.7 MPa with 201.3 mS/cm of NaCl solution

(a)

Volumetricstrain, v(%)

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

10 100 1000

ua - uw (kPa)

Shear strain

Volumetric strain (b)

vand s(%)

P/THMN/5


A CONSTITUTIVE APPROACH TO ADDRESSTHE THERMAL AND HYDRIC IMPACTS

IN THE CONCEPT OF DEEP RADIOACTIVEWASTE REPOSITORIES

B. Franois1, M. Nuth1, L. Laloui1

1. Soil Mechanics Laboratory, Ecole Polytechnique Fdrale de Lausanne (EPFL), Station 18, CH-1015 Lausanne, Switzerland([email protected], [email protected], [email protected])

INTRODUCTIONPerformance assessment of deep repositories for heat-generating radioactive waste (e.g. vitrified high-levelradioactive waste - HLW) requires the capability of predicting reliably the thermo-hydro-mechanicalbehaviour of both clay barrier and host rock/clay. Due to the complexity and highly coupled nature ofphenomena taking place in a waste repository, an adequate understanding of the constitutive behaviour ofclays and the capability to model their evolutions are challenging tasks. Within this context, a mechanicalstress-strain constitutive framework taking into consideration the non-isothermal and unsaturatedconditions is proposed to model the behaviour of the aforementioned confining materials. Thermo-plasticity of saturated and unsaturated materials is featured within the framework which also takesadvantage of a convenient generalized effective stress (Nuth & Laloui, 2007). The proposed constitutivecontext allows a performing prediction of experimental features of behaviour by comprehending complexthermo-hydro-mechanical coupling effects.

THERMO-HYDRO-MECHANICAL PROCESSESIn the years following the construction of the underground disposal, the near-field which can be definedas the zone altered by the presence of radioactive waste undergoes complex mechanical, hydric and thermalinterdependent loads (THM couplings). Indeed, between the times of excavation and HLW placing,drainage and consolidation processes occur in the surrounding host rock so that a decrease in effectivestress is assorted with swelling of the host rock and likely initiates desaturation. Once the canister andengineered (initially unsaturated) barrier are in place, the heating from the canister and the hydration fromthe surrounding rock become the two principal environmental loads. According to the stress state in theclayey barrier, the soil may be subject to thermal and hydric swelling or undesired plastic shrinkage dueto thermo-plasticity and wetting collapse.

ACMEG-TS: A CONSTITUTIVE FRAMEWORKBased on experimental evidence on the relevant THM features of behaviour of fine grain soils, a newunified thermo-mechanical constitutive model for unsaturated soils (ACMEG-TS, for AdvancedConstitutive Model in Environmental Geomechanics Thermal and Suction effects) has been developedin the framework of elasto-thermoplasticity of unsaturated soils. The model uses the concept of generalizedeffective stress adapted from Bishop (1959):

where ij, pa, pw, Sr are the mechanical external load, the pore air and pore water pressures and the degreeof saturation, respectively. This generalized stress evidences a dependency on the thermal, hydric andmechanical histories of the material. Indeed, the retention capacity of the soil (i.e. the degree of saturationat a given suction) depends on its dry density, suction level and temperature. Therefore, this single stressapproach converts a complex multi-phase and multi-stress medium in which multi-physics processes occurinto a single equivalent mechanical state using several coupling equations.

= ( ) + ( ) ij ij a ij r a w ijp S p p

P/THMN/5



In the context of elasto-plasticity, the total strain increment tensor, due to THM loading is decomposed intonon-linear thermo-elastic and thermo-plastic components. Material plasticity is induced by two coupledhardening processes: an isotropic and a deviatoric mechanism. The yield functions of the two mechanicalthermo-plastic mechanisms have the following expressions (Figure 1) (Franois and Laloui, 2007), riso, rdev,d and b being material parameters:

Preconsolidation pressure, pc is known to depend on temperature and suction conditions so that it is notonly the hardening variable during mechanical loading but also the one variable piloting the shape of theyield limit enabling to model thermal and hydric collapse.

In terms of hydric response, desaturation is also believed to raise elasto-plastic phenomena. Upon adrying process, suction increase is associated with a reduction in the degree of saturation Sr that fallssignificantly once the air entry suction se is overpassed. Figure 2 gives a qualitative representation of thistemperature and material dry density dependent hydric limit. Under re-wetting, an hysteretic hydricphenomenon occurs which is also represented by a yielding process (Figure 2). Further wetting-dryingcycles activate successively two hydric yield limits (Franois et al., 2007).

NUMERICAL SIMULATIONSFinally, the presented advanced constitutive model has been implemented in the LAGAMINE finiteelement code (Collin 2003), providing a performing numerical tool to address, to model and most of all tounderstand the highly coupled mechanical, hydric and thermal effects governing the THM behaviour ofgeological and engineered barriers in a HLW deep repository.

ReferencesBishop, A.W. 1959. The principle of effective stress. Tecnisk Ukeblad 39, pp 859-863

Collin, F. 2003. Couplages thermo-hydro-mcaniques dans les sols et les roches tendres partiellementsaturs, PhD Thesis, Universit de lige, Lige.

Francois, B. & Laloui, L. 2007. A stress-strain framework for modeling the behaviour of unsaturated soilsunder non-isothermal conditions. Springer Proceedings in Physics 113, Theoretical and numericalunsaturated soils mechanics, pp. 119-126.

Franois B., Nuth M. & Laloui L. 2007. Mechanical constitutive framework for thermal effects onunsaturated soils, 10th International Symposium on Numerical Models in Geomechanics, NUMOG X,Rhodes (Greece).

Nuth, M. & Laloui, L. 2007. Implications of a generalized effective stress on the constitutive modelling ofunsaturated soils. Springer Proceedings in Physics 113, Theoretical and numerical unsaturated soilsmechanics, pp. 75-82.

Figure 1: Yield limits for the THM elasto-plastic framework. Figure 2: Schematic represen-tation of the water retentioncurve model.

f p p r f q Mpiso c iso dev= ( ) = = ' 's,T ;0 1 bb Log d pp rc devs,T'

( )

=0

P/THMN/6


NUMERICAL ANALYSIS ON STABILITYOF BOOM CLAY TUNNEL BY SHIELD

CONSTRUCTION

Chen Weizhong1, Jia Shanpo1, Yu Hongdan1, Wu Guojun1, Li Xiangling2, Bernier Frdric2

1. Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan 430071([email protected], [email protected], [email protected], [email protected])

2. Belgian Nuclear Research Centre, Mol 2400 ([email protected], [email protected])

NUMERICAL SIMULATION OF SHIELD TUNNEL BY 3D NON-LINEAR FEM

The CLIPEX project was realized in the frame of extension of the underground research facility (HADES)in Boom clay formation at a depth of about 223 m at Mol site. This project aimed to provide theexperimental data to characterize the shortterm hydro-mechanical behavior of the clay, induced by theexcavation of a gallery (Bernier et al, 2003). The construction of connecting gallery has been realized witha Tunnel Boring Machine composed of a 2.7m long shield. The wedge block system was used for lining.

Three dimensional finite elements model has been set up in this paper, which symmetry in a vertical planealong the gallery has been considered (Fig.1). Mohr-Coulomb model is applied to describe the mechanicalbehavior of Boom clay and boundary conditions of the numerical model are as following: (1) The initialin-situ stress and pore pressure depends on the depth; (2) Displacement in the normal direction of thevertical planes is fixed, and the bottom boundary of the model displacements in the z direction is fixed ;(3) The upper boundary of the model applies the weight of the upper ground; (4) Owing to the very lowpermeability of the Boom Clay, a 2m/day advancing rate can been assumed as very fast, and consequentlythe response of the medium to the excavation can be idealized as undrained.

To model adequately this soil-lining interaction with a large-strain mode, the contact between the clay andthe lining is modeled using the contact elements. As for the connecting gallery, the excavation sequenceby shield tunneling followed as: (1) Excavation of a drilling length of 1m (the length of lining) under theprotection of the shield, which is 3m long; (2) Activation of the lining segment 4m behind the frontconsidering of the initial ground stress release. The lining segments are placed in contact with the soilconsidering the presence of gap.

Figure 1: Finite element model used in calculation.

P/THMN/6



Fig.2 shows the distribution of pore pressure of surrounding rock during shield tunneling. Excess poreliquid pressure of the surrounding rock is induced by the shield tunneling, and the negative pore pressureof surrounding rock appears in the forward region of shield. The change of pore pressure is very obviousin the earlier stage of shield tunneling and the excess pore is gradual dissipated with the advancing ofshield. The maximal pore pressure of excavated region is 3.134MPa when the shield drills 9m in the Boomclay, while the maximal pore pressure is 2.585MPa when the shield drills from 13m to 21m in the Boomclay.

The deformation of surrounding rock during shield tunneling (fig.3) shows the induced radial convergence.The surrounding rock which is supported by shield becomes construction gap during the course of shieldadvancing before supported by concrete lining, which lead to a very large deformation. The maximaldeformation of surrounding rock is up to 9.7cm and the disturbed region of surrounding rock is expandedgradually with the shield tunneling.

CONCLUSIONSBased on the comprehensive analysis on the primary components of surrounding rock associated withshield tunneling, a flexible and plastic seepage flow-stress coupling analysis model is set up in accordancewith the effective stress analysis and the model change method is put forward to imitate moving-forwardshield process by means of FEM program. Through the result of the hydro-mechanical disturbancesinduced by the excavation of the connecting gallery, the distributions of the plastic zone extent, pressureexerted on the lining, effective stress, pore pressure and radical displacement range around the tunnelassociated with the shield tunneling operation are calculated, which furnishes a important basis for theassessment of the stability of the connecting gallery. Also, the results in this paper are useful for the shieldtunnel construction and the field measurement.

Reference[1] Bernier F, Li X L, Verstricht J, et al. CLIPEX: clay instrumentation programme for the extension of anunderground research laboratory (Final Report) [R]. [s.l]: EUR 20619 EN, 2003.

Figure 2: Distribution of pore pressure of surround-ing tunnel during construction (9m)

Figure 3: Distribution of deformation of surround-ing rock during shield tunneling

P/THMN/7


A THERMO-HYDRO-MECHANICAL MODELFOR CLAY IN UNSATURATED CONDITIONS

CHEN Weizhong1, TAN Xianjun 2, WU Guojun 3, JIA Shanpo 4

1. Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan 430071([email protected], [email protected], [email protected], [email protected])

INTRODUCTION

Two different kinds of approaches can be identified to tackle the complex thermo-hydro-mechanicalcoupling problems. The first one is based on the extension of Biots theory for saturated porous media tounsaturated media. It consists of finding an equivalent interstitial pressure for the multiphase media as afunction of pore pressure and saturation degree of each phase (Coussy, 2004). In the second approaches,the concept of net stress (total mean stress plus gas pressure) is introduced into equations of unsaturatedmaterials (Alonso et al., 1990). The capillary pressure is used as a complementary force variable. Elasticproperties and plastic functions depend explicitly on the capillary pressure. In this work, the secondapproach based on the concept of net stress is preferred. Displacement, capillary pressure, air pressure andtemperature are chosen as the basic variables in the governing equations. The present mathematical modelcan be considered as a fairly complete treatment of significant coupling effects in the unsaturated mediaunder thermal, hydraulic and mechanical loading.

THERMO-HYDRO-MECHANICAL COUPLED CONSTITUTIVE MODEL OF CLAY

Based on the assumption of clay as a porous media, considering liquid flow on the temperature gradient(thermo-osmosis); heat flow on water potential gradient (thermal-filtration), and heat flow due to thermalconvection; temperature and stress effects on the retention curve; temperature effect on void ratio; heat sinkdue to thermal deformation; compressibility of water, air and the medium. The equilibrium equation, liquidwater mass balance equation, air mass balance equation, and heat energy balance equation are established.The Thermo-hydro-mechanical coupled constitutive model of clay is expressed as:

The flow model is based on two interacting continua: one representing the liquid flow and the other thegas flow. Thus, at every point in space, two pressures are introduced: one denoting the average liquidpressure and the other the average gas pressure. The mass balance equations for both water and air asfollows:

The heat transfer model is based on the assumption that energy is transferred through all three phases andthat the phases are at thermal equilibrium. The energy balance equation as follow:

d d d d d ij ijkloe

kl klp

ij c ij aC KB p p K= ( ) 1 BB Tij2 d

LtL

ptL

ptL

Tt

kV c a l li11 12 13 14

+

+

+

+ jj

l

l

jl j l lTij

l

j

px

g kTx

+

== lq

L VtL

pctL

patL

Tt

akai21 22 23 24

+

+

+

+ jj

a

pax j

akaTijTx j

ql a

=

LtL

ptL

ptL

Tt

vV c a li31 32 33 34

+

+

+

+ ( ) ccT v cT

en

l

sij l lij

( ) + ( )( ) =

+

+ +

ai a

11

ee e n

eT

eQl aij s

+( )+

+

++

1

111

nn Qe e n

eQl lla+

+( )+1

1

P/THMN/7



Based on the unsaturated experimental results ofclay in laboratory, the developed models areapplied to simulate the evolution law of theexcavation and ventilation of a deep cavity, the2D-axisymetric finite element model (fig.1).

CONCLUSIONSBased on the developed complete T-H-M coupledconstitutive model of unsaturated clay, numericalsimulation of a underground clay tunnel duringexcavation has been done to obtain variation lawof saturation degree and temperature with timeand deformation induced by coupling effects.From the theoretical and numerical research work,it can be concluded that:

(1) The developed thermo-hydro-mechanicalmodel of unsaturated clay could simulate the flowprocess of clay from state of saturated tounsaturated state and then to resaturated state, which can be used to practical tunnel excavation in clay.

(2) A small value of the (thermo-osmosis) has a significant effect on capillary pressure and displacementdistributions. Both capillary pressure and displacements are significantly increased due to the presence offully coupled heat-water flow.

(3) The drying and resaturation processes of clay are mainly controlled by the intrinsic permeability ofclay. Plastic deformation and damage can be induced by the excavation and the ventilation phases.

ReferencesCoussy, O., 2004. Poromechanics. JohnWiley & Sons.

Alonso, E.E., Gens, A., Josa, A., 1990. A constitutive model for partially saturated soils. Geotechnique 40(3), 405430.

Figure 1: Geometry model.

60 m

60m

symmetry

Ux=0symmetry

R = 2.95 m

Uy = 0 symmetry

pw = p0 , pg = pg, ij = ijw 0 0

p w=p0,pg=p g,ij=ij

w0

0

P/THMN/8


3D MODELLING OF TER EXPERIMENTACCOUNTING FOR FULLY ANISOTROPICTHERMO-PORO-ELASTIC BEHAVIOUR

R. Giot1, D. Hoxha1, A. Giraud1, F. Homand1, K. Su2, C. Chavant3

1. Laboratoire Environnement Gomcanique et Ouvrages, LaEGO-ENSG-Nancy Universit, Ruedu Doyen Marcel Roubault, B.P. 40, 54501 Vanduvre-ls-Nancy, France

2. ANDRA Agence Nationale pour la Gestion des Dchets Radioactifs, Parc de la Croix Blanche,1-7 rue Jean Monnet, 92298 Chtenay-Malabry, France

3. EDF R&D 1 avenue du Gnral de Gaulle, 92141 Clamart Cedex

INTRODUCTIONThe TER experiment is an in situ experiment conducted in the ANDRA Meuse / Haute-Marne laboratory,at the depth of -490 m, aiming at characterizing both the thermal properties of the argillite and itsmechanical and hydraulic response to a thermal loading. It consists in imposing a given heat flow on alength of three meters on the wall of a borehole of 101 mm diameter, through an instrumented heaterdeveloped for the experiment. Strains, displacements, pore pressure and temperature sensors are set in theadjacent rock mass to record its evolution.

ANISOTROPIC COUPLED THERMO-HYDRO-MECHANICAL BEHAVIOURThe full thermo-poro-elastic anisotropic behaviour of the argillite has been accounted for. It concerns thepermeability, the thermal conductivity, the elastic properties and the coupling parameters. The media issupposed to be fully saturated, and we considered the Biot theory (Coussy, 2004). Under these hypotheses,the variation of pore pressure and stresses in the rock mass due to the change of temperature can beexpressed through the equations:

In those equations, is the Biot second order tensor relative to the transverse isotropic porous medium.The effect of anisotropy is then taken into account in the hydro-mechanical coupling parameters. Theequations of thermo-poro-elasticity are implemented for the case of transverse isotropy. Those equationsare based on the relations between macroscopic and microscopic properties of the constituents.

THREE DIMENSIONAL MODELLING OF TER EXPERIMENTThe different thermal loading phases of the TER experiment have been modelled with a numerical code.The mesh used for the 3D calculations is given on figure 1. It consists of approximately 30000 nodes.

The first calculations, without considering the anisotropy on coupled HM parameters, show a good overallprediction of temperature evolution in both heating phases when optimal thermal values are considered anda good prediction of pressure evolution (figure 2) during heating. This result reduces the uncertainty aboutthe thermo-hydro-mechanical coupling parameters measured on the argillites samples. Nevertheless, thevalues considered for the parameters for the heating phase does not allow to explain the evolution of porepressure during the drilling of heating borehole. Moreover, the overall anisotropy must be taken intoaccount to explain the differences in the evolutions of sensors in symmetrical positions with respect to theheater. Optimised values of the wide coupling parameters should lead to explain pore pressure evolutionduring drilling and heating at the same time. A coupling of the 3D modelling with an inversion code canbe used to identify those wide parameters and their optimal values. So, the calculations with the fully

p p -MB: Mm3M0

0fl m

= + +

= ( ) 0 O 0 OC : B p p A

B

P/THMN/8



anisotropic thermoporoelastic behaviour should improve the knowledge on the values of the parameters ofthe hydromechanical coupling, such as Biot coefficients. Moreover, additional information should beinferred on the microscopic properties of the constituents.

Additional phenomena can explain the discrepancies between measures and modelling, such as plasticityand viscoplasticity or damage around the borehole. Concerning the hydraulic conductivity, the calculationsseem to indicate that smaller values than those initially supposed, should be considered. Some questionsremains on the hydraulic conductivity, regarding its potential evolution with temperature, pore pressure orporosity.

References[1] COUSSY O. (2004). Poromechanics, JohnWiley & Sons, Chichester, England.

Figure 1: Mesh used for the 3D modelling of the TER experiment.

Figure 2: Comparison between measured and calculated pore pressure.

0

1

2

3

4

5

6

0 10 20 30 40 50

TER1403_PRE_01

model

model with offset

Pore Pressure

(MPa)

Time (days)

P/THMN/9


HYDROMECHANICAL MODELINGOF SHAFT EXCAVATIONIN M/HM LABORATORY

G. Duveau1, Y. Jia1, K. Su2 and J. F. Shao1*

1. Laboratoire of Mechanics of Lille, UMR8107 CNRS, USTL, France

2. Andra, Chatenay Malabry, France

* Corresponding author: [email protected], Polytech-Lille, cit scientifique, 59655Villeneuve dAscq, France

In the framework of feasibility study for geological storage of nuclear waste, a series of in situ experimentsare being performed by Andra in the Meuse/Haute Marne underground research laboratory at BURE. Themain purpose of these experiments is to evaluate thermal and hydromechanical responses of engineered andgeological barriers subjected to excavation, drying, heating and chemical degradation. The REP-Experiment is one of the first ones performed in this context. It consists to determine hydromechanicalresponses in rock formation during progressive excavation of the main access shaft. Variations of porepressure, radial displacement and relative displacement have been measured at different locations. In thispaper, we present hydromechanical modeling of the REP-experiment using fully coupled finite elementmethod.

In the first part, the general layout of the REP-Experiment including the excavation sequences, locationsof measuring points and boundary conditions is given. A synthetic analysis of the observed results isprovided in order to capture various coupling phenomena to be taken into account. Then a constitutivemodel is proposed for the description of coupled poro-mechanical behavior of argillite in saturated andunsaturated conditions. It is an elastoplastic model coupled with induced damage by microcracks. Theconcept of general effective stress is used to take into account the influence of pore pressure and capillaryeffect on mechanical behavior. The degradation of elastic behavior due to damage is considered. A non

Figure 1: Variation of pore pressure during excavation at the point REP2102_PRE_03.

0

1

2

3

4

5

6

7

0 50 100 150 200

P(MPa)

time (day)

day of excavation

r = 4,85 m

REP2102_PRE_03

P/THMN/9



associated plastic flow rule is used to describe the transition from plastic compressibility to dilatancy. Thematerial softening behavior is attributed to progressive growth of microcracks. The models parametershave been determined from representative laboratory tests. The numerical simulations are compared withtest data for various loading paths on partially saturated samples. In the third part, we present the numericalmodeling of REP-experiment. In this paper, only the results from axi-symmetric modeling are presented.A fully coupled hydromechanical code is used. A series of sensitivity study is performed in order todetermine influences of main parameters controlling hydromechanical responses; such as in situ stresses,mechanical properties, ventilation condition and permeability variation. It is found that the evolution ofdamaged zone is strongly influenced by initial stress state and mechanical strength. The variation ofpermeability with damage plays an important role in the evolution of over pore pressure and affectsmechanical response. It is also important to take into account the ventilation condition on the shaft wall,generating an unsaturated zone around the shaft. Further, material anisotropy and initial stress anisotropyhave to be taken into account in 3D modeling. In Figures 1 and 2, we show some representative results.Numerical modeling is compared with in situ measurement for pore pressure and radial displacement attwo measuring points.

ACKNOWLEDGEMENTThe present work is jointly supported by the European Project FIKW-CT2000-00029 and Andra (Frenchnational agency for management of radioactive wastes).

Figure 2: Variation of radial displacement at the point PPA0031_DFO_01.

0

1

2

3

4

5

0 50 100 150 200

simulation

PPA0031_DFO_01

radial displacement (mm)

jours

r=4,44m

PPA0031_DFO_0

day of

excavation

time (day)

P/THMN/10


MODELLING OF DRYING DAMAGEIN ENGINEERED AND NATURAL CLAY

BARRIERS FOR NUCLEAR WASTE DISPOSAL

H. Peron1, L. Laloui1, T. Hueckel2, L.B. Hu2

1. Swiss Federal Institute of Technology Lausanne, EPFL - Soil Mechanics Laboratory,CH 1015 Lausanne, Switzerland ([email protected], [email protected])

2. Duke University, Department of Civil and Environmental Engineering, Durham, N.C. 27706,U.S.A. ([email protected], [email protected])

INTRODUCTION

Damage induced by desiccation (i.e. drying and cracking) is a ubiquitous issue in nuclear waste storageengineering. Desiccation cracks can affect engineered clay barriers when submitted to hydro-thermalcoupled solicitations induced by the confined waste (Graham et al. 1997). Cracks are also a potential threatto the EDZ: a cracked dried zone can form in the host clay of the underground gallery due to ventilationeffects (Ramambasoa, 2000). Dramatic increase of the permeability and modifications of the mechanicalproperties of the confining barrier could ensue. The fundamental mechanisms of the phenomenon ofcracking during drying and the ways to control or avoid it are still elusive. This paper presents first theresults of an experimental program aiming at clarifying the various phenomena leading to desiccationcracking as well as the parameters which control initiation and propagation of cracks. A comprehensivemodel is thereafter presented, which unifies the various features of desiccation. Desiccation cracks areassumed to be the consequence of tensile stress generation, until tensile strength is reached. The presentednon conventional model is capable of reproducing desiccation induced stresses formation until a crackingcriterion is met. The model has been validated by implementing it into a finite element scheme andsimulating real case desiccation cracking.

EXPERIMENTAL CHARACTERISATION OF DESICCATION

Desiccation tests were carried out on initially saturated virgin clayey soils. Three kinds of tests wereperformed: free desiccation tests, constrained desiccation tests and crack pattern tests (Fig. 1), all consistingin air drying of soil cakes in controlled humidity and temperature conditions while imposing variousmechanical boundary conditions. Strain evolution, water content, and suction were measured for the wholebody, as well as locally in the vicinity of newly formed cracks (Peron et al. 2006). To investigate inparticular the strains induced by the fluid removal, unconstrained drying tests were performed with variousfluids (Peron et al. 2007). Experimental results highlighted the following fundamentals of desiccationmechanisms:

i) Unconstrained shrinkage gives rise to irreversible drying deformations occurring in saturated conditionsuntil the air entry value is reached. This drying phase is characterized by a surface evaporative flux. In

Figure 1: Typical crack pattern, air drying with 2D bottom restraint.

P/THMN/10



this case, shrinkage is presumably driven by perimeter capillary forces while its amount is controlled bythe skeleton compressibility. Thereafter, upon further drying and de-saturation strains are mainly elastic.

ii) Tensile stresses arise due to constrained shrinkage, via a strain compensation mechanism. Dryingdamage is thus the result of restraining boundary conditions or moisture gradients which do not respectcompatibility conditions. Any soil structure may also be responsible for self-stress generation leading todamage.

iii) The critical processes leading to cracking, as well as cracking initiation are quasi-saturated ones. Theyoccur for saturation ratios very close to unity, before suction overcomes the air entry value and air inva-des the pore space.

MODELLING CONCEPTDesiccation cracks are likely to occur if shrinkage (i.e. strain due to fluid removal) is constrained. In thissituation, tensile stresses are generated, which reach the soil tensile strength. The Bishop-Schrefflereffective stress is used to describe strain evolution. It is assumed that the desiccation induced effectivestress increment may be expressed by:

d = dnet + d(Srs)I (1)

where dnet is a net stress increment, Sr is the saturation ratio, s is the suction and I the identity matrix. Inthis scheme, the net stress increment is the results of a strain compensation mechanism and it isproportional to the difference between the strain increment induced by the fluid removal (the only oneoccurring in case of unconstrained shrinkage) and the actual strain increment. Stresses in a soil undergoingdesiccation depend on two factors: strains induced by the fluid removal, and change in soil stiffness duringthe process (given that it strongly depends on soil hydration state).

A hardening multi-mechanism plasticity framework is used in the model. A tensile criterion is integrated,based on Griffith criterion. Cracking initiation is supposed to be reached when the minor principal effectivestress becomes equal to the tensile strength of the soil. This enables the prediction of tensile (mode I) crackoccurrences such as desiccation cracks. The constitutive model also reproduces the plastic drying strainsarising during the saturated stage of drying.

VALIDATION, MODELLING OF DESICCATION CRACKSA coupled thermo-hydro-mechanical formulation of the problem is given. Once the tensile strength isreached, finite elements are used to tackle the process of crack formation. The point where cracking mayoccur is set as two separate nodes stuck together but belonging to neighbouring elements. When theeffective stress reaches the cracking criterion, the nodes separate. Some essential features are captured bythe model, such as the creation of crack characteristic length (i.e. crack spacing). The crack spacing (whichgreatly determines the value of permeability of the cracked body) appears to be controlled by tensilestrength, soil stiffness and moisture gradients as well as the strains induced by fluid removal.

References:Graham, J., Chandler, N.A., Dixon, D.A., Roach, P.J., To, T., and Wan, A.W.-L. 1997. The Buffer/Container Experiment: results, synthesis, issues. Atomic Energy of Canada Limited Report, AECL-11746, COG-9746-I. Chalk River, ON.

Peron, H., Laloui, L., Hueckel, T. and Hu, L. 2006. Experimental study of desiccation of soil. UNSAT2006, ASCE Geotechnical Special Publication, Miller et al. eds., 147: 1073-1084.

Peron H., Hu L., Hueckel T., Laloui L. 2007. The influence of the pore fluid on desiccation of a deformableporous material. Springer Proceedings in Physics, Experimental unsaturated soil mechanics, 413-420.

Ramambasoa N., Nguyen-Minh D., Rejeb A. 2000. Hydromechanical behavior of a shale. Application tothe Tournemire site. 4th NARMS, Pacific rocks: Rock around the Rim, 1067-1072.

P/THMN/11


ELASTOVISCOPLASTIC BEHAVIOUROF MEUSE/HAUTE-MARNE ARGILLITE:LABORATORY TESTS AND MODELLING

D. Hoxha1, C. Auvray1, A. Giraud1, F. Homand1, K. Su2

Laboratoire Environnement Gomcanique et Ouvrages, LaEGO-ENSG-Nancy Universit, Ruedu Doyen Marcel Roubault, B.P. 40, 54501 Vanduvre-ls-Nancy, France

ANDRA Agence Nationale pour la Gestion des Dchets Radioactifs, Parc de la Croix Blanche,1- rue Jean Monnet, 92298 Chtenay-Malabry, France

INTRODUCTIONLong-term hydromechanical behaviour of geological barriers for nuclear waste deep disposal is of a greatimportance for the safety of the disposal. This behaviour is highly influenced by the state of saturation ofthe rock masses and their creep aptitude. In this paper we show the new laboratory results obtained oncreep tests on fixed partially saturated conditions. Then an elastoviscoplastic model describing short andlong term behaviour of this rock on saturated and partially saturated conditions is proposed.

CREEP TESTS ON CONTROLLED HR CELLSTypical results from creep tests on constant relative humidity atmosphere and temperature are presentedon the figure 1. A step-wise uniaxial mechanical loading has been applied on the samples, that wereinitially equilibrated in the constant relative humidity cell. During a step the applied stress is kept constantand creep strains are recorded. On the same time at periodic intervals the ultrasonic wave velocities aremeasured in two normal to the applied stress direction in order to follow an eventual differed damage.

The creep strain rates are strongly influenced by the level of the desaturation with the more saturatedsample showing a higher rate. Besides, all samples show a volume evolution during the creep which signis function of applied stress and relative humidity. Finally, when while keeping a constant axial stress thehygrometry has changed an evolution of the strains has been recorded with amplitudes being function ofthe hygrometry change.

Figure 1: Typical results from relative humidity controlled creep test.

EST16350-1

-0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0 100 200 300 400 500

t (jours)

dformations

diffres

axiale

latrale

distorsion

volumique

P/THMN/11



ELASTOVISCOPLASTIC MODELLING OF M/HM ARGILLITE BEHAVIOURThe model proposed here is an extension of a previous model developed for the short-term behaviour ofM/HM argillite on saturated and partially saturated conditions (Hoxha et al., 2007). Details of this modelare not described here (see the above mentioned reference).

The principal idea of this extension consists in coupling of the viscoplastic behaviour with the plastic one.The short-term failure surface is identified as the envelope of macroscopic failure of samples tested in thelaboratory and is given by:

(1)

Following overstress approach the rate of viscoplastic strain is formally written as:

(2)

In this equation indicates the overstress function that in general case could be function of the

stress and a set of internal variable Vp. The viscoplastic yield surface is supposed of the same form as the

instantaneous failure surface, i.e.:

(3)

Similarly to the plastic behavior we define a long-term-final surface, which is the final position of theviscoplastic yield function:

(4)

In order to account for the volumetric strain evolution, the direction of the viscoplastic flow is defined fromthe same potential as the plastic flow that leads to:

(5)

Since the pure deviatoric long-term behavior of this rock could be reasonably well described by aLemaitres like model, the overstress function is chosen to be:

(6)

where

Finally, the rate of viscoplastic strain is written:

(7)

The influences of the viscoplastic behavior on the short term response and vice versa are taken into accountthrough two hardening rules:

(8.a)

and (8.b)

These last equations have some important consequences on the rock behavior since firstly they give to thesurface (4) the status of the long-term failure (which explains also its nomination), and secondly theyintroduce an evolution (decreasing) of the instantaneous failure surface because of the viscoplastic flowyields) which is consistent with general observations on rocks.

The model parameters are identified by means of the creep tests and numerical results are compared withlaboratory results with a good agreement

References[1] Hoxha D., Homand F., Giraud A., Auvray C (2007) - Saturated and unsaturated behaviour modelling ofMeuse-Haute/Marne argillite, International Journal of Plasticity, 23, pp. 733-766

R m eq eq m = ( , , , )@ @

W W vp vp pF V= ( )( ),F Vp ,( )

fdvp

m eq eqvp

m( , , , )@ @ =

R m eq eqvp

m0 0( , , , )@ @ =

W vp

vpp

p pF Vg g

= ( )

, . /

F V Af

vpp

invp vp

nvp

invp

m vp

, .,( ) ( )

0

p

m eq

m=

@

W

vp invp vp

nvp

invp

m vpAf

=( )

.

,

0

pp pg g

/

( ) = ( )vp in vpExp b* *. 0vp k= .

P/THMN/12


LONG-TERM BEHAVIOUROF A LIME TREATED SOIL

UNDER PERCOLATION CONDITIONS

B. Lequiller1, O. Cuisinier1, V. Ferber1, Y.J. Cui2, D. Deneele1

1. Laboratoire Central des Ponts et Chausses, Route de Bouaye, BP 4129, F-44341 BouguenaisCedex ([email protected])

2. Soil Mechanics Research Center (CERMES), Institut Navier, cole Nationale des Ponts etChausses, 6-8 av. B. Pascal, F-77455 Marnes-la-Valle Cedex ([email protected])

INTRODUCTIONLime has been added to clayey soils in order to improve their physical and mechanical properties forcenturies. Much research has been conducted over the past 40 years in the field of geochemistry tounderstand the complex interactions between lime and lime-reactive soils. However, few studies have beencarried out to determine the coupling between these physico-chemical processes and the soil mechanicalbehaviour. Furthermore, it is of interest to estimate the life time or durability of the constructions whenexposed to environmental stress factors, such as water circulation over long term periods. Hence, the maingoal of this study is to evaluate the long-term behaviour of lime stabilized soils under percolationconditions.

EXPERIMENTAL PROCEDUREPercolation equipment that could simultaneously test specimens under accelerated leach conditions, givethe permeability changes of specimens over the percolation time, and allow leachates chemical analyses ofspecies wash-out as well as mechanical testing has been designed. Twelve flexible wall column leachcylinders were used. Distilled water having a pH equal to 7 was leached trough the specimens under apressure of 80 kPa. The leachates were continuously collected and different chemical analyses were doneat various percolation times (pH, exchangeable pore cations concentration, silicon and aluminiumconcentrations). The specimens were also subjected to several analyses after diverse percolation times (pHof the soil, cation exchange capacity, titration to determine the percentage of free lime remaining in thesoil, etc.). Finally, after leaching, some of the specimens were subjected to unconfined compression teststo estimate the effects of percolation on the strength of the treated soil.

The soil chosen for the testing is a clayey silt from Jossigny which is located near Paris. The limemodification optimum (LMO) of the soil was determined using the Eades and Grim pH test and was foundto be equal to 1 % CaO. Three different percentages of lime, 0 %, 1 % and 3 %, as well as two differentcuring times, 7 and 25 days, were selected for this study. The compaction parameters were chosen equalto the optimal parameters of each material, the dry densities going from 1.55 to 1.80Mg/m3.

RESULTS AND INTERPRETATIONThe given data focus on the results obtained for the natural specimen and specimens treated at 1 % CaOafter 25 days of curing time. Among all the parameters available, only three parameters, thought to be themost interesting and providing a global idea of the results are presented in this paper: that is to sayleachates pH, calcium concentrations and permeability changes. Figure 1-(a) and 1-(b) show respectivelythe changes in pH, calcium concentration and permeability of specimens under percolation. It was chosento chart the different parameters as a function of the number of pore volumes passed through thespecimens: one pore volume corresponding respectively to about 65 ml and 80 ml for natural and treatedspecimens. It can be observed that the pH of the natural specimen remains constant, whereas the pH of thetreated specimen varies with the number of pore volumes passed through the cell. In fact, the pH of thetreated specimen remains relatively constant to 11.5 until 40 pore volumes and drops thereafter to reach a

P/THMN/12



pH of 8.5 at 70 pore volumes. The leachates calcium concentration of both materials is decreasing withthe number of pore volumes passed through the cell and goes respectively for the treated and untreatedspecimens from 180 mg/L and 120 mg/L at the beginning of the percolation to 25 mg/L and 50 mg/L atthe end of the percolation. The permeability of the treated specimen is 10 fold higher than that of thenatural specimen. Two phases can be distinguished for the treated specimen: the first one going from 0 porevolume to 20 pore volumes where the permeability remains relatively constant and the second one goingfrom 20 pore volumes to 70 pore volumes where the permeability decreases. In the case of the naturalspecimen, the permeability is globally decreasing with the number of pore volumes.

From these results different hypothesis that need to be confirmed by additional testing can be done. Thefirst phase observed on Figure 1-(b) for the treated specimen, as well as the decrease recorded for theuntreated specimen can probably be attributed to a phase of saturation and rearrangement of the soil fabriccaused by the flow and confining pressures applied. The second phase observed for the treated specimencould be explained by voids clogging due to the formation of cementation products. This phenomenonseems supported by the decrease of calcium concentrations recorded for the treated specimen, which couldreflect the catching up of calcium for the formation of the cementation products. Finally, the decrease ofpH in the case of the treated specimen could indicate a carbonation of the system, entailing penalizingconditions for the formation of cementation products.

MAIN CONCLUSIONS AND STUDY OUTLOOKSWater circulation leads rapidly to a decrease of pH that could be penalizing for the specimens treatedat their LMO. The decrease in permeability observed for these specimens could be due to theformation of cementation products. However, additional testing such as mercury intrusion porosimetrytests, thin sections and scanning electron microscope observations, as well as, longer periods ofpercolation are necessary to validate the different hypothesis.

Figure 1: Changes in pH, calcium concentration (a), and permeability (b) observed for the naturalspecimen and specimen treated at 1 % CaO with a cure of 25 days during percolation.

0

4

8

12

0 20 40 60 80 100

Number of pore volumes

pH

0

50

100

150

200

250

Natural specimenTreated specimenpH[Ca

2+] concentrations

[Ca2+](mg/L)

(a)

1.0E-10

1.0E-09

1.0E-080 20 40 60 80 100

Number of pore volumes

Permeability(m/s)

Natural specimenTreated specimen

Saturation ?

Formation of

cementation products ?

(b)

P/THMN/13


STRESS REDISTRIBUTIONAND HYDROMECHANICAL EFFECTSDUE TO EXCAVATION AND DRILLINGOPERATIONS IN TER EXPERIMENT

L. Uhlig1, M. Jobmann1, M. Polster1, J. Vaunat2, B. Garrite2, Y. Wileveau3

1. DBE TECHNOLOGY GmbH, Eschenstrae 55, D-31224 Peine ([email protected])2. UPC, Edificio C-1, Campus Norte UPC, 08034 Barcelona ([email protected])3. ANDRA, 1/7 rue Jean Monnet, 92298 Chtenay-Malabry cedex, France

([email protected])

INTRODUCTIONIn the Meuse Haute Marne Underground Research Laboratory (URL-Bure), a thermal experiment (namedTER) is performed for evaluation of the behavior of the Callovo-Oxfordian clay formation due to a thermalimpact. This in-situ test is providing the necessary data for assessing the feasibility of a deep geologicalwaste repository for heat generating radioactive waste. The gained in-situ data is analyzed and in particularcompared to accompanying numerical modeling performed by various modeling teams. This paper dealswith results of the hydromechanical modeling of the host rock behavior due to the excavation work at the490 m level of the URL-Bure prior to heating and in addition with the effects due to the drilling processof TER 1101 heater borehole. The paper presents a comparative study on the modeling results obtained bythe teams of DBE Technology by using the Finite Difference program FLAC and UPC by using the FiniteElement code Code_bright.

DRIFT EXCAVATIONS AT 490 M LEVELBoth numerical models used for calculation of stress field are given in figure 1 and 2. The DBETecmodeling took a simplified time schedule for simulation of the drift excavations into account (from Feb/05 to Sept/05), whereas UPC simulation considered instantaneous excavation.

Two host rock behavior approaches were used. UPC assumed linear elastic behavior whereas DBETecapplied an elastoplastic Hoek-Brown model with hydraulic coupling. Starting from the initial conditions(v=12.0 to 12.7 MPa, h=12.0 to 12.4 MPa, H=16.0 to 16.1 MPa) both teams calculated the stress redis-tribution in the TER pillar. In spite of the use of different material models nearly the same results were

Figure 1: Part of DBETec model, drifts and heaterbore hole, view from south.

Figure 2: UPC model of drift excavations.

Heater borehole

TER1101

189000 zones

Model dimension / surrounding rock volume:

120 m x 150 m x 100 m

GMR

GKE

GEX

GLE

GEX drift

Heater

borehole

57600 zones

v0= -12 MPa

h0 = -12 MPa

H0= -16 MPa

GKE

GMR

P/THMN/13



obtained. The calculations show anincrease of vertical stress with simulta-neous decrease of the higher horizon-tal stress due to the pillar effect. Theaverage calculated stress field nearTER borehole changes to: v = 14.5 -15.0 MPa; h = 11.5 - 12.0 MPa; H =13.8 - 14.3 MPa.

DRILLING OF TER 1101HEATER BOREHOLEDuring drilling of the heater boreholepore pressure evolution was measuredby the sensors next to the borehole(figure 3). The sensor in closesthorizontal distance (0.5 m) recorded apore pressure increase of 20 bars, theclosest sensor in vertical direction (0.7m) showed a pore pressure decrease of17 bars. These pore pressure changesare enormous regarding the distance of3 times of borehole diameter to the closest sensor.

Aimed at determining the degree of HM-coupling and its parametersa back analysis of these pore pressure evolution had been performedby DBETec and UPC. Both teams applied HM-coupled 2D models(via Biots Theory) with reference values for hydraulic andmechanical behavior but used different continuum codes and differentmechanical approaches.

The UPC modeling for analysis of the pore pressure effects(Code_bright, elastic approach) showed a calculated maximum porepressure reaction of up to 3 bars at the nearest sensors (c.f. figure 4).The DBETec model (FLAC3D, elastoplastic approach) also highlyunderestimates the pore pressure changes. A variation of strengthparameters between peak and residual strength of the Callovo-Oxfordian resulted in calculated pore pressure changes between 0.3and 3 bars at the sensor locations.

It can be concluded, that both applied models showed only 1/10 1/6 of the measured pore pressure changes and therefore highly underestimate the hydro-mechanical (MH)host rock reaction due to the drilling. Similar observations have been gained at other URLs, too (c.f. URLHades, hydromechanical response of Boom clay during connecting gallery excavation). Those resultsunderline clearly the necessity of a better understanding and refinement of the hydraulic coupling lawapplied to argillaceous rock.

ReferencesVaunat, J., Garitte, B. (2006): Interprtation par modlisations thermohydromcaniques delexprimentation TER, Rapport intermdiaire UPC, Barcelona, Spain.

Polster, M., Jobmann, M., Uhlig, L. (2006): Results on stress, strain and pore pressure distribution in thepillars & heater borehole effect, Technical Meeting Nr. 3, Bure, France

Wileveau, Y. (2006): Premiers resultats de lexperimetation TER, Periode de juillet 2005 a septembre2006, Laboratoire de recherche souterrain de Meuse/Haute-Marne

Figure 3: Measured pore pressure evolution during TER 1101drilling, ANDRA.

Figure 4: Calculated porepressures due to the drilling, UPC.

P/THMN/14


THERMODYNAMIC MODEL ON SWELLINGOF BENTONITE BUFFER

AND BACKFILL MATERIALSHaruo SATO

Japan Atomic Energy Agency (JAEA), 432-2, Hokushin, Horonobe-cho, Hokkaido 098-3224,Japan

INTRODUCTIONIn the safety assessment of the geological disposal for high-level radioactive waste, bentonite is used asone of the engineered barriers composing multi-barrier system and of the backfill materials, and its physicaland chemical functions are expected. Smectite (montmorillonite) is a major clay mineral constituent of thebentonite and backfill materials. Bentonite swells by contacting with groundwater and controls thegroundwater flow. Since this swelling property is caused by the hydration of smectite, understanding thethermodynamic properties of water near the surface of smectite are important.

With respect to the thermodynamic properties of water near the surface of bentonite, the activity, therelative partial molar Gibbs free energy (GH2O), the relative partial molar enthalpy (HH2O) and entropy(SH2O) of water have been reported and it has been reported that the activity of water near the surface ofmontmorillonite as well as its chemical potential are lower than those of free water (Torikai et al., 1995,1996). Moreover, it has been reported that the water near the surface of montmorillonite has a high-moisture potential and a high suction (Suzuki et al., 1992; Kanno and Wakamatsu, 1993). The moisturepotential of bentonite is closely related to swelling pressure, one of the fundamental properties. Kahr et al.(1989) have predicted the swelling pressure for two kinds of bentonites (Montigel, MX-80) versus drydensity based on water vapor adsorption-desorption isotherms and Kanno and Wakamatsu (1993) have alsopredicted the swelling pressure based on moisture potential for a Japanese bentonite (Kunigel-V1).However, such conventional models to predict swelling pressure are dependent on the kind of bentoniteand smectite content and are quite limited from the viewpoint of generalizability.

In this study, swelling pressure of bentonite occurring in the permeation process of moisture was estimatedbased on thermodynamic theory and the thermodynamic data of water on the surface of smectite, and ageneralizable model was developed.

MEASUREMENTS OF THERMODYNAMIC DATA OF WATERThe vapor pressure method was utilized to determine the thermodynamic data of water on the surface ofsmectite. The activity of water on the surface of smectite was measured as a function of water content (0-83 %)and temperature (15-40 C) in a dry density range of 0.6-0.9 Mg/m3, and GH2O was determined at 25 C.

PurifiedNa-smectite ofwhich all interlayer cationswere completely exchangedwithNa+ ions and soluble saltswere removed,was used in this study.As the starting material, Kunipia-P bentonite (provided fromKunimineIndustries Co. Ltd.) of which smectite content is approximately 100 wt.%, was used. The cation exchangecapacity (CEC)was alsomeasured by the ammonium ion exchangemethod andwas 122.8 meq/100 g.

ACTIVITY AND GH2O OF WATERFigure 1 shows the activities of water obtained in this study and those of water on montmorillonite (Kunipia-F bentonite) (Torikai et al., 1996) at 298.15 K versus water content. The activity and GH2O of waterdecreased with a decrease of water content in the range of water content lower than about 40 %. Similartrends were obtained to data of Kunipia-F bentonite (Na-bentonite) of which smectite content was approxi-mately 100 wt.% (Torikai et al., 1996). From the specific surface area of smectite (about 800 m2/g) and thecorrelation between GH2O and water content, water affected from the surface of smectite was estimated to beup to approximately 2 water layers.

P/THMN/14



MODELLING OF SWELLINGPRESSURESwelling pressure against smectite partialdensity (= montmorillonite partial density)was estimated from the difference (GH2O)in the chemical potential balance of water inequilibrium state between free water andmoisturized smectite, and compared to datameasured for various kinds of bentonites(Kunigel-V1, MX-80, Kunipia-F, Kunipia-F with silica sand, Kunigel-V1 with silicasand) of which smectite contents wererespectively different (Suzuki et al., 1992;JAEA, 2006). Figure 2 shows the compari-son of the calculated and measured swellingpressures. The calculated swelling pressureswere in good agreement with the measureddata over the range of smectite partial den-sity. This indicates that swelling pressure of

compacted bentonite can be quantitatively calculated based on the thermodynamic data of water on thesurface of smectite for an arbitrary bentonite and dry density if smectite content and silica sand content ofthe bentonite are known.

ReferencesJAEA, JAEA Database on Basic Characteristics of Buffer Material 2006, JAEA Website (2006).

Kahr, G.. et al., Mater. Res. Soc. Symp. Proc., Vol.127, pp.683-689 (1989).

Kanno, T. andWakamatsu, H., Mater. Res. Soc. Symp. Proc., Vol.294, pp.425-431 (1993).

Suzuki, S. et al., Tech. Rep., PNC TN8410 92-057 (1992).

Torikai, Y. et al., Mater. Res. Soc. Symp. Proc., Vol.353, pp.321-328 (1995).

Torikai, Y. et al., Nucl. Technol., 115, pp.73-80 (1996).

Figure 2: Comparison of calculated (in thiswork) and measured (Suzuki et al., 1992;JAEA, 2006) swelling pressures. Theauthor also calculated swelling pressuresbased on the activity data of water onKunipia-F bentonite (Torikai et al., 1996).

Figure 1: Activities of water on smectite obtained in thisstudy and those of water on montmorillonite (Kunipia-Fbentonite) (Torikai et al., 1996).

P/THMN/15


3D ANALYSIS OF A HEATING TESTIN THE OPALINUS CLAY

A. Gens1, B. Garitte1, J. Vaunat1, Y. Wileveau2

1. Technical University of Catalonia (UPC), Building D2, Campus Nord, c/ J. Girona, 1-3, 08034Barcelona, Spain ([email protected], [email protected], [email protected])

2. ANDRA Agence Nationale pour la Gestion des Dchets Radioactifs, RD 960 - BP 9, 55290Bure, France ([email protected])

INTRODUCTIONThe HE-D experiment is a heating experiment which has been run within the Opalinus clay formation atthe Mont-Terri underground rock laboratory (Switzerland) between March 2004 and June 2005 (Wileveau,2005). In this experiment, two power-controlled heaters have been placed in a horizontal borehole and twoheating phases were realized, at respectively, 33% and 100% of nominal power (1950 W for both heaters).Around this heating borehole, about 30 temperature sensors, 10 water pressure sensors and 3 extensometerswere placed (Fig. 1.a). This paper presents the interpretation of field measurements by means of a 3D FiniteElement analysis with special attention devoted to interplay between the thermal, hydraulic and mechanicalresponse of the host rock.

FEATURES OF THE ANALYSISThe theoretical THM formulation used herein allows for solving simultaneously equations of stressequilibrium, water mass balance, solid mass balance and energy balance (Olivella et al., 1996). A coupleddamage/elastoplastic law developed for modelling the mechanical response of soft rocks has been used.Because the rock is anisotropic due to intense bedding, a true 3D analysis was conducted consideringvalues for the thermal conductivity, stiffness and strength along and across the bedding. Mesh used for theanalysis is shown in Fig 1.c.

RESULTS AND INTERPRETATIONIsochrones of temperature and pore pressure field computed at the end of the second heating phase areshown in Figure 2.a and 2.d. In Fig. 2.b, 2.c, 2.e and 2.f., evolution of temperature and pore pressure during

Figure 1: Problem characteristics: a) Setup of the field experiment; b) spatial distribution of sensors aroundthe heater; c) mesh used for the analysis.

Mesh

a)

HDB08

HDB17 HDB16

HDB15

HDB14

HDB13HDB11

HDB03

HDB09HDB10

H3

H2H1

-4

-3

-2

-1

0

1

2

3

4

-4 -3 -2 -1 0 1 2 3 4

y=8my=9m

y=8m

y=12.2m y=11.2m

y=11.8m

y=12.3m

y=8m

y=8.5my=11.3m

H1, H2 and H3:

y=9m and

y=11.8m. r=0.15m

r=1.11m

r=2.96m r=1.42m

r=0.952m

r=0.775m

r=3.34mr=2.95m

r=1.27m

r=0.85m

r=1.27m

b)

B03B14

Bedding orientation

c)

P/THMN/15



the test is compared with experimental results at sensors B03 (along bedding) and B14 (perpendicular tobedding, see Fig. 1.b for locations). Numerical analysis evidences the development of an anisotropic fieldof temperature that causes in turn an anisotropic distribution of pore pressure build-up as a result of rockand water thermal expansion. Pore pressure increase reaches at some points 3 MPa. Because of thedissipation of pore pressure under the gradients generated by temperature increase, maximum in porepressure is reached well before the end of heating.

Strain evolution is compared in Fig. 2.g at a point in the near field (BHE D4 4-5) and in the far field (BHED4 4-5). Model captures well the respective extension/compression cycle and pure compression phase thatoccurs at both points. As a result of the increase in stresses due to rock expansion, model predicts thedevelopment of an anisotropic damage zone around the heaters (Fig. 2.h). Comparison between computedand measured damaged Young moduli allows for a validation of model prediction (Fig. 2.i).

References:Olivella, S., Carrera, J., Gens A. & Alonso, E.E. (1994): Non-isothermal Multiphase Flow of Brine andGas through Saline media, Transport in Porous Media, vol. 15, pp. 271-293.

Wileveau, Y. (2005). THM behaviour of host rock (HE-D) experiment: Progress report. Part 1, TechnicalReport TR 2005-03. Mont Terri Project.

Figure 2: Comparison between model prediction and measurements.

a)

0

10

20

30

40

50

60

0 100 200 300 400 500

Time from heating [days]

Temperature[C]

Temp D03

3D

b)

0

10

20

30

40

50

60

0 100 200 300 400 500


Temperature[C]

Temp D14

3D

c)

d)

-2000

-1500

-1000

-500

0

500

1000

1500

2000

2500

3000

0 100 200 300 400 500


Pressure[kPa]

Pres D03

3D

e)

-2000

-1500

-1000

-500

0

500

1000

1500

2000

2500

3000

0 100 200 300 400 500


Pressure[kPa]

Pres D14

3D

f)

-0.045

-0.035

-0.025

-0.015

-0.005

0.005

0 100 200 300 400 500

Time [days]

Strains[%]

BHE-D4 4-5

3D

BHE-D5 6-7

3D

g) h)

0

2

4

6

8

10

12

0 2 4 6

Distance to bh0 [m]Unloadingmodulus[Gpa]

0 days (1)

90 days (2)

110 days (3)

338 days (4)

E measurements

1 2

34

i)

P/THMN/16


THM ANALYSIS OF A GMOCK-UPLABORATORY EXPERIMENT USING

A DOUBLE-STRUCTURE EXPANSIVE MODEL

A. Jacinto1, M. Snchez2, A. Ledesma1

1. Tech. University of Catalunya UPC, Jordi Girona, 1-3 (D2), 08034 Barcelona, Spain,([email protected]; [email protected])

2. University of Strathclyde, Department of Civil Engineering, 107 Rottenrow, Glasgow G4 0NG,U.K. ([email protected])

INTRODUCTIONMany laboratory and in situ experiments have been carried out simulating the effect of hydration andtemperature on bentonite, simulating repository conditions. Generally, those experiments are performedunder confined conditions, and temperature, relative humidity (or suction), and stresses at differentbentonite locations are measured during the test. In addition to that experimental effort, several numericalcodes have been developed with the aim of simulating those conditions. These codes consider the mainmechanisms involved: thermal conduction, advective and diffusive flows of liquid water and vapour,swelling, among others. Most of these processes are coupled and the problem becomes highly nonlinear.Despite those difficulties most of the simulations show in general a good comparison between T-Hmeasured and computed variables (i.e., temperature and relative humidity). However, Mechanicalpredictions are still far from being satisfactory. Stresses are rather difficult to predict when the samebentonite block develops swelling due to hydration (i.e. on the side in contact with the host rock) andshrinkage due to loss of water (i.e., on the side where a heater or the waste is placed). The paper presentsa numerical THM simulation of a laboratory mock-up test analysed by using a double-structureexpansive model, focussing on the comparison between measured and computed stresses.

MECHANICAL CONSTITUTIVE MODELThe model used in this work was developed by Sanchez et al (2005), and is based on a framework definedby Gens and Alonso (1992) in order to improve the prediction capabilities of the so called BarcelonaBasic Model. The model explicitly considers the two pore levels that actually exist in expansive clays: amacrostructural level modelled by means of BBM, and a micr

Documents

H HM THM C P ocesses Na Ba - andra.fr · the possible swelling of soil during the re-saturation phase, a specific saturation procedure was proposed and applied (Figure 1). Consolidation