In situ and laboratory investigations of fluid flow through an argillaceous formation at different scales of space and time, Tournemire tunnel, southern France

In the context of a research and development program on waste disposal, an experimental site (Tournemire tunnel, Aveyron, France) was selected by the French Institute for Nuclear Protection and Safety (IPSN) in order to undertake studies on potential fluid flow at different scales of space and time within a 250-m-thick argillaceous formation. The argillite has a low natural water content (~3–5%) and very low radii access porosity. Diffusion (tritiated water) coefficients (1×10^–12 to 2×10^–11 m²/s) and hydraulic conductivities derived from different types of laboratory tests (10^–14 to 10^–13 m/s) are characteristics of a very low-permeable rock. In situ hydraulic tests (including long-term hydraulic-head measurements) were used to obtain values for hydraulic head and hydraulic conductivity at a scale of 1–10 m (10^–13 to 10^–11 m/s). Despite uncertainties on these data (due to a scale factor, presence of fissures, and possible artefacts due to hydro-chemo-mechanical coupling), it is expected that fluid flow is essentially governed by diffusion processes. Identification of possible natural flows at larger scales of time and space was investigated using natural isotopic tracers from interstitial fluids. Modelling, based on the deuterium profile along the clay formation and assuming pure diffusion processes, provides estimations of possible flow times. However, lack of knowledge concerning the past geological evolution of the site and the possible role of a fracture network do not permit reduction of uncertainties on these estimations at this stage. Electronic Publication     

Stochastic analysis of the effect of heterogeneity and fractures on radionuclide transport in a low-permeability clay layer

Deep low-permeability clay layers are considered as safe environments for disposal of high-level radioactive waste. In Belgium, the Boom Clay is a candidate host rock for deep geological disposal. In this study, we analyze the effects of fractures and spatially variable hydraulic conductivity on radionuclide migration through the clay. Fracture geometry and properties are simulated with Monte Carlo simulation. The heterogeneity of hydraulic conductivity is simulated by direct sequential co-simulation using measurements of hydraulic conductivity and four types of secondary variables. The hydraulic conductivity and fracture simulations are used as input for a transport model. Radionuclide fluxes computed with this heterogeneous model are compared with fluxes obtained with a homogeneous model. The output fluxes of the heterogeneous model differ at most 8% from the homogeneous model. The main safety function of the Boom Clay is thus not affected by the fractures and the spatial variability of hydraulic conductivity.     

Hydrogeological modeling of radionuclide transport in low permeability media: a comparison between Boom Clay and Ypresian Clay

Deep low-permeability clay layers are considered as suitable environments for disposal of high-level radioactive waste. In Belgium, the Boom Clay is the reference host formation and the Ypresian Clay an alternative host formation for research and safety and feasibility assessment of deep disposal of nuclear waste. In this study, two hydrogeological models are built to calculate the radionuclide fluxes that would migrate from a potential repository through these two clay formations. Transport parameter heterogeneity is incorporated in the models using geostatistical co-simulations of hydraulic conductivity, diffusion coefficient and diffusion accessible porosity. The calculated radionuclide fluxes in the two clay formations are compared. The results show that in the Ypresian Clay larger differences between the fluxes through the lower and the upper clay boundary occur, larger total output radionuclide amounts are calculated and a larger effect of parameter heterogeneity on the calculated fluxes is observed, compared to the Boom Clay.     

The use of two-part Hooke’s model (TPHM) to model the mine-by test at Mont Terri Site,Switzerland

The full-scale mine-by (MB) test conducted in 2008 in the Mont Terri underground rock laboratory (Switzerland) investigated the deformation and the coupled hydro-mechanical behavior of the Opalinus Clay in response to tunnel excavations. The Opalinus Clay is currently under investigation in Switzerland as a potential host rock for geologic disposal of high-level radioactive waste. To further improve the understanding and modeling of the coupled processes and their impact on the performance of a geologic repository in Opalinus Clay, a newly developed two-part Hooke’s model (TPHM) was implemented into a geomechanical simulator. A three-dimensional simulation model based on the TPHM was then developed to predict the deformation and pore pressure responses in the near field of the MB Niche 2 test at the Mont Terri Site. The usefulness and validity of the TPHM are demonstrated by the consistency between simulation results and field observations. Simulation results show that the pore pressure disturbance becomes visible at about 11 m ahead of the mine-by excavation advancing face (along the longitudinal direction of the MB Niche). The results also demonstrate that there exists a good correlation between the excavation damage zone (EDZ) and the pore-pressure evolution, which may have important practical implications for monitoring EDZ evolution with pore-pressure sensors. The simulation results, which are sensitive to the constitutive relationships used in the model, capture both the observed displacements and the size of the damage zone, whereas the approach based on the conventional Hooke’s law underestimates both. The comparison between simulated and observed results also indicates that laboratory-measured mechanical properties can be used to accurately predict field-scale mechanical deformations, as long as valid constitutive relationships are employed.     

Experimental Study of the Brittle Behavior of Clay shale in Rapid Unconfined Compression

The mechanical behavior of clay shales is of great interest in many branches of geo-engineering, including nuclear waste disposal, underground excavations, and deep well drilling. Observations from test galleries (Mont Terri, Switzerland and Bure, France) in these materials have shown that the rock mass response near the excavation is associated with brittle failure processes combined with bedding parallel shearing. To investigate the brittle failure characteristics of the Opalinus Clay recovered from the Mont Terri Underground Research Laboratory, a series of 19 unconfined uniaxial compression tests were performed utilizing servo-controlled testing procedures. All specimens were tested at their natural water content with loading approximately normal to the bedding. Acoustic emission (AE) measurements were utilized to help quantify stress levels associated with crack initiation and propagation. The unconfined compression strength of the tested specimens averaged 6.9 MPa. The crack initiation threshold occurred at approximately 30% of the rupture stress based on analyzing both the acoustic emission measurements and the stress–strain behavior. The crack damage threshold showed large variability and occurred at approximately 70% of the rupture stress.     

Transport of U in the Opalinus Clay on centimetre to decimetre scales

The Opalinus Clay formation in North Switzerland is a potential host rock for a deep underground radioactive waste repository. The distribution of ²³⁸U, ²³⁴U and ²³⁰Th was studied in rock samples of the Opalinus Clay from an exploratory borehole at Benken (Canton of Zurich) using MC-ICP-MS. The aim was to assess the in situ, long-term migration behaviour of ²³⁴U in this rock. Very low hydraulic conductivities of the Opalinus Clay, reducing potential of the pore water and its chemical equilibrium with the host rock are expected to render both ²³⁸U and ²³⁰Th immobile. If U is heterogeneously distributed in the Opalinus Clay, gradients in the supply of ²³⁴U from the rock matrix to the pore water by the decay of ²³⁸U will be established. Diffusive redistribution separates ²³⁴U from its immobile parent ²³⁸U resulting in bulk rock ²³⁴U/²³⁸U activity disequilibria. These may provide a means of estimating the mobility of ²³⁴U in the rock if the diffusion rate of ²³⁴U is significant compared to its decay rate. Sampling was carried out on two scales. Drilling of cm-spaced samples from the drill-core was done to study mobility over short distances and elucidate possible small-scale lithological control. Homogenized 25-cm-long portions of a 2-m-long drill-core section were prepared to provide information on transport over a longer distance. Variations in U and/or Th content on the cm-scale between clays and carbonate-sandy layers are revealed by β-scanning, which shows that the (dominant) clay is richer in both elements.     

Environmental geological assessment of a solid waste disposal site: a case study in Sivas, Turkey

The selection of the disposal site is probably the most important step in the development of solid waste management. In site selection, geology plays a determining role. This study evaluates the characteristics of the environment on the basis of the geological, hydrogeological and geo-engineering properties of the solid waste site of the Sivas city, Turkey. The area is underlain by the Oligocene-Miocene rocks which have limited aquifer properties. Thin Quaternary alluvium and soil cover overlie the Oligo-Miocene rocks, which are represented as well graded sand and inorganic silt of low plasticity. The Quaternary alluvium and soil cover are classified as inorganic clays having a low plasticity and the permeability varies from 1.2×10⁻⁶ to 3.11×10⁻⁶ m/s. These values are much higher than 1×10⁻⁸ m/s, which is accepted for waste disposal standards. Seepage waters have a potential to pollute the ground water and the Kızılırmak River, which is 500 m to the southwest of the waste disposal area and because the disposal site is close to the river, the potential for flash flooding poses a high pollution risk. The waste disposal area must be covered by clay layers or an impervious artificial membrane. In addition, seepage must be controlled and removed from the site.     

Biogeochemical processes in a clay formation in situ experiment: Part E - Equilibrium controls on chemistry of pore water from the Opalinus Clay, Mont Terri Underground Research Laboratory, Switzerland

The chemistry of pore water (particularly pH and ionic strength) is an important property of clay rocks being considered as host rocks for long-term storage of radioactive waste. Pore waters in clay-rich rocks generally cannot be sampled directly. Instead, their chemistry must be found using laboratory-measured properties of core samples and geochemical modelling. Many such measurements have been made on samples from the Opalinus Clay from the Mont Terri Underground Research Laboratory (URL). Several boreholes in that URL yielded water samples against which pore water models have been calibrated. Following a first synthesis report published in 2003, this paper presents the evolution of the modelling approaches developed within Mont Terri URL scientific programs through the last decade (1997-2009). Models are compared to the composition of waters sampled during dedicated borehole experiments. Reanalysis of the models, parameters and database enabled the principal shortcomings of the previous modelling efforts to be overcome. The inability to model the K concentrations correctly with the measured cation exchange properties was found to be due to the use of an inappropriate selectivity coefficient for Na-K exchange; the inability to reproduce the measured carbonate chemistry and pH of the pore waters using mineral-water reactions alone was corrected by considering clay mineral equilibria. Re-examination of the measured Ca/Mg activity ratios and consideration of the mineralogical composition of the Opalinus Clay suggested that Ca/Mg cation exchange rather than dolomite saturation may control the ratio of these ions in solution. This re-examination also suggests that the Ca/Mg ratio decreases with increasing pore-water salinity. Several possible reasons for this are proposed. Moreover, it is demonstrated that feldspar equilibria must not be included in Opalinus Clay modelling because feldspars are present only in very small quantities in the formation and because Na/K ratios measured in pore water samples are inconsistent with feldspar saturation. The principal need to improve future modelling is additional or better data on rock properties, in particular: (i) a more detailed identification of phases in the Opalinus Clay that include redox-sensitive elements together with evaluation of their thermodynamic properties; (ii) an improved understanding of the distribution of celestite throughout the Opalinus Clay for Sr/SO₄ concentrations control; (iii) improvements in analytic and thermodynamic data for Ca-Mg rock cation exchange and mineral chemical properties and (iv) the measurement of composition and stability constants of clay minerals actually present in the formation.     

Compartmented flow at the Bátaapáti site in Hungary

Integration of hydrogeological and geological data into a conceptual model is critical for site investigation programs, since this model is the basis for hydrogeological modeling and for engineering. In the Hungarian Radioactive Waste Disposal Investigation Program, several methods have been used to characterize the potential host rock (granite) at the Bátaapáti site for a repository for low and intermediate level radioactive waste. Hydrogeological data acquisition revealed some characteristic aspects of the site. One of the most important is the presence of extensive rock deformation zones (faults) with low hydraulic conductivity, which strongly reduce the direct hydraulic communication between adjacent blocks of rock (compartmentalization). This characteristic of the rock mass results in a mosaic-like distribution of rock compartments, each with an almost constant hydraulic head. Within the compartments, hydraulic tests have shown that transmissivity is strongly scale dependent: the larger the scale, the higher the measured transmissivity. The extensive highly transmissive zones cause very low hydraulic gradients within each block, thus the transport processes are strongly influenced by the low or average transmissivity zones and the rock matrix.     

Laboratory and In Situ Simulation Tests of the Excavation Damaged Zone Around Galleries in Opalinus Clay

In the context of nuclear waste disposal in clay formations, laboratory and in situ simulation experiments were performed to study at reduced scale the excavation damaged zone (EDZ) around tunnels in the indurated Opalinus Clay at Mont Terri, Switzerland. In the laboratory, thick-walled hollow cylindrical specimens were subjected to a mechanical unloading mimicking a gallery excavation. In samples cored parallel to bedding, cracks sub-parallel to the bedding planes open and lead to a buckling failure in two regions that extend from the borehole in the direction normal to bedding. The behaviour is clearly anisotropic. On the other hand, in experiments performed on specimens cored perpendicular to bedding, there is no indication of failure around the hole and the response of the hollow cylinder sample is mainly isotropic. The in situ experiment at Mont Terri which consisted in the overcoring of a resin-injected borehole that follows the bedding strike of the Opalinus Clay showed a striking similarity between the induced damaged zone and the fracture pattern observed in the hollow cylinder tests on samples cored parallel to bedding and such a bedding controlled “Excavation” Damaged Zone is as well consistent with the distinct fracture patterns observed at Mont Terri depending on the orientation of holes/galleries with respect to the bedding planes. Interestingly, the damaged zone observed in the hollow cylinder tests on samples cored parallel to bedding and in situ around URL galleries is found to develop in reverse directions in Boom Clay (Mol) and in Opalinus Clay (Mont Terri). This most probably results from different failure mechanisms, i.e. shear failure along conjugated planes in the plastic Boom Clay, but bedding plane splitting and buckling in the indurated Opalinus Clay.     

Constitutive Relationships for Elastic Deformation of Clay Rock: Data Analysis

Geological repositories have been considered a feasible option worldwide for storing high-level nuclear waste. Clay rock is one of the rock types under consideration for such purposes, because of its favorable features to prevent radionuclide transport from the repository. Coupled hydromechanical processes have an important impact on the performance of a clay repository, and establishing constitutive relationships for modeling such processes are essential. In this study, we propose several constitutive relationships for elastic deformation in indurated clay rocks based on three recently developed concepts. First, when applying Hooke’s law in clay rocks, true strain (rock volume change divided by the current rock volume), rather than engineering strain (rock volume change divided by unstressed rock volume), should be used, except when the degree of deformation is very small. In the latter case, the two strains will be practically identical. Second, because of its inherent heterogeneity, clay rock can be divided into two parts, a hard part and a soft part, with the hard part subject to a relatively small degree of deformation compared with the soft part. Third, for swelling rock like clay, effective stress needs to be generalized to include an additional term resulting from the swelling process. To evaluate our theoretical development, we analyze uniaxial test data for core samples of Opalinus clay and laboratory measurements of single fractures within macro-cracked Callovo-Oxfordian argillite samples subject to both confinement and water reduced swelling. The results from this evaluation indicate that our constitutive relationships can adequately represent the data and explain the related observations.     

Modelling the Mont Terri HE-D experiment for the Thermal–Hydraulic–Mechanical response of a bedded argillaceous formation to heating

Coupled thermal–hydrological–mechanical (THM) processes in the near field of deep geological repositories can influence several safety features of the engineered and geological barriers. Among those features are: the possibility of damage in the host rock, the time for re-saturation of the bentonite, and the perturbations in the hydraulic regime in both the rock and engineered seals. Within the international cooperative code-validation project DECOVALEX-2015, eight research teams developed models to simulate an in situ heater experiment, called HE-D, in Opalinus Clay at the Mont Terri Underground Research Laboratory in Switzerland. The models were developed from the theory of poroelasticity in order to simulate the coupled THM processes that prevailed during the experiment and thereby to characterize the in situ THM properties of Opalinus Clay. The modelling results for the evolution of temperature, pore water pressure, and deformation at different points are consistent among the research teams and compare favourably with the experimental data in terms of trends and absolute values. The models were able to reproduce the main physical processes of the experiment. In particular, most teams simulated temperature and thermally induced pore water pressure well, including spatial variations caused by inherent anisotropy due to bedding.     

Analysis of flow path around the sealing section HG-A experiment in the Mont Terri Rock Laboratory

As part of the HG-A experiment in the Mont Terri Rock Laboratory, large-scale in situ water/gas injection experiments was conducted in a microtunnel. This research work focuses on the numerical analysis of the experimental data and the in situ observations. Concerning a temporary change of the hydromechanical properties of Opalinus Clay during experimental operations, three phases were numerically interpreted. These included the generation of excavation damaged zone during tunnel excavation, in which highly permeable flow paths around the tunnel have been formed; the self-sealing effect during water tests; and the pressure evolution during a long-term gas injection test. A coupled two-phase flow and mechanics model, taking into account the strong anisotropic properties of Opalinus Clay, was developed to interpret the measured data. The hydraulic anisotropy was described by a transversely isotropic permeability tensor. An elasto-plastic model was established to consider both stiffness anisotropy and strength anisotropy. Anisotropic plasticity was studied using the microstructure tensor method.     

Large-scale tracer profiles in a deep claystone formation (Opalinus Clay at Mont Russelin,Switzerland): Implications for solute transport processes and transport properties of the rock

Natural tracers (Cl^? and stable water isotopes) in pore water of the Opalinus Clay and adjacent formations were studied in the motorway tunnel at Mont Russelin, Switzerland. The Opalinus Clay occurs in the core of an anticline which is cut by a complex system of thrust faults. Concentration profiles of natural tracers were taken from 17 boreholes along a 363 m long section. Pore waters of drillcore samples were analysed with indirect and direct methods. The Cl^? and stable water isotope distribution in the pore water shows a regular and well defined profile, with a conspicuous decrease towards the overlying Dogger limestone aquifer. The highest Cl^? values (approximately 23,000 mg/L) are found in the core of the anticline in Liassic claystones underlying the Opalinus Clay. To quantify the large-scale transport properties of the Opalinus Clay formation, a 2D transport model was constructed and used to reproduce the observed concentration profiles. The calculations indicate that the observed tracer distributions are consistent with diffusion as the dominant transport process. Groundwater flow in the overlying Dogger aquifer was initiated about 2–4 Ma ago, which is long after the folding of the Jura Mountains and probably coincides with the exposure of the aquifer to freshwater recharge following continued erosion of the anticline. The calculations suggest that tracer distributions are controlled by 1) the timing of freshwater recharge in the overlying limestone aquifer, 2) the shape of the anticline and 3) the magnitude and the anisotropy ratio of diffusion coefficients.     

Engineering geological evaluation of geological barrier rocks at landfills and repositories

 A site specific evaluation of geological barriers is required for landfills and waste repositories. The waste-repository-rock system has to be taken into consideration for this. Since the geotechnical barrier in conjunction with geological barriers contributes considerably to long-term isolation of the harmful substances from the biosphere, it is absolutely necessary to use engineering geology and hydrogeological methods for a quantitative assessment of the barrier effect of the host rock and the geological environment. For a waste disposal mine, the load-bearing capacity of the rock, the protective properties of the surrounding rock formations and the geological stability of the area are important factors in a safety analysis, of which the geotechnical stability analysis is an important part. Such an analysis comprises an engineering geological study of the site, laboratory and in situ experiments and model calculations, long-term monitoring, and special geological and geochemical investigations. Existing geomechanical modelling techniques provide a useful tool to perform barrier integrity calculations and to design an underground disposal mine. As an example, the barrier performance of the rock salt of the Gorleben salt dome is demonstrated using various safety indicators. Received: 3 January 1997 · Accepted: 29 January 1997     

Deformation mechanisms and hydraulic properties of fault zones in unconsolidated sediments; the Roer Valley Rift System,The Netherlands

In general, faults cutting through the unconsolidated sediments of the Roer Valley Rift System (RVRS), The Netherlands, form strong barriers to horizontal groundwater flow. The relationships between deformation mechanisms along fault zones and their impact on the hydrogeological structure of the fault zone are analyzed in a shallow (0–5 m below land surface) trench over one of the faults in the study area. Recently developed digital-image-analysis techniques are used to estimate the spatial distribution of hydraulic conductivity at the millimeter-scale and to describe the micromorphologic characteristics of the fault zone. In addition, laboratory measurements of hydraulic conductivity on core-plug samples show the larger-scale distribution of hydraulic conductivity in the damage zone flanking the main fault plane. Particulate flow is the deformation mechanism at shallow depths, which causes the damage zone around the fault, in the sand-rich parts, to have a relatively enhanced hydraulic conductivity. The fault core is characterized by reduced hydraulic conductivity due to clay smearing, grain-scale mixing, and iron-oxide precipitation. Electronic Publication     

In situ and laboratory investigation of the alteration of Boom Clay (Oligocene) at the air–geological barrier interface within the Mol underground facility (Belgium): Consequences on kerogen and bitumen compositions

The Boom Clay formation (Oligocene) is studied as a reference host rock for methodological studies on deep geological disposal of radioactive waste. During excavation of galleries within the Clay formation (HADES underground research facility, Mol, Belgium), the physico-chemical conditions are significantly modified as an air–clay interface is created. In order to study the long-term impact of the air–clay contact on the organic matter contained in the Boom Clay, two types of samples were studied: (1) a reference series of clay samples having been in contact with the atmosphere of the HADES gallery for increasing times up to several years and (2) unaltered clay samples submitted to artificial oxidation in a ventilated oven at 80 °C. The evolution of geochemical data of the two series was compared using Rock-Eval pyrolysis, GC–MS and size exclusion chromatography. The organic matter of the unaltered clays sampled in the HADES galleries is dominated by type III kerogen (terrestrial) with some contribution of type II (marine) and is thermally immature. The evolution of geochemical parameters during air alteration for the two series are very similar. They show progressive oxidation of kerogen accompanied by the release of bitumen enriched in low molecular weight constituents. Molecular analysis evidences the presence of a complex mixture of aliphatic and aromatic O-bearing compounds, inherited from the degradation of kerogen as well as from the clay catalyzed oxidation of the bitumen. These results show that (1) air oxidation is a major process in the in situ alteration of the organic matter of Boom Clay within the HADES galleries, (2) laboratory oxidation experiments at 80 °C yield similar results as in situ air alteration of Boom Clay and (3) artificial air oxidation may be used to assess the long term exposure of the organic matter to air.     

Differential Hydrogeological Effects of Draining Tunnels Through the Northern Apennines,Italy

Water inflows are a major challenge in tunnelling and particularly difficult to predict in geological settings consisting of heterogeneous sedimentary rock formations with complex tectonic structure. For a high-speed railway line between Bologna and Florence (Italy), a series of seven railway tunnels was drilled through turbiditic formations, ranging from pelitic rocks with thin arenitic layers over sequences including thick-bedded sandstone to calcareous rocks showing chemical dissolution phenomena (karstification). The tunnels were built as draining tunnels and caused significant impacts, such as drying of springs and base-flow losses at mountain streams. A comprehensive hydrological monitoring programme and four multi-tracer test were done, focusing on four sections of the tunnel system. The tracer tests delivered unprecedented data on groundwater flow and transport in turbiditic aquifers and made it possible to better characterize the differential impacts of tunnel drainage along a geological gradient. The impact radius is 200 m in the thin-bedded sequences but reaches 2.3–4.0 km in calcareous and thick-bedded arenitic turbidites. Linear flow velocities, as determined from the peaks of the tracer breakthrough curves, range from 3.6 m/day in the thin-bedded turbidites to 39 m/day in the calcareous rocks (average values from the four test sites). At several places, discrete fault zones were identified as main hydraulic pathways between impacted streams and draining tunnels. This case shows that ignoring the hydrogeological conditions in construction projects can cause terrible damage, and the study presents an approach to better predict hydraulic impacts of draining tunnels in complex sedimentary rock settings.