Unsaturated Soils: Experimental Studies

Recent theories indicate that capillary pressure is perhaps not only a function of saturation but may also depend on the time rate of change of saturation. This is known as the dynamic or non-equilibrium effect. There is compelling experimental evidence reported in the literature that the non-equilibrium effect is observable, quantifiable, and significant. However, almost all reported experiments relate to unsaturated systems. In this work, we report on a recent series of experiments involving water and PCE. Quasi-static as well as dynamic capillary pressure curves for primary drainage, main drainage, and main imbibition, are measured. The data are used to estimate the non-equilibrium capillary pressure coefficient. Finally, a continuum-scale two-phase flow model has been employed to simulate the experiments. Variations of average pressures and average water saturation with time are calculated and compared with measured curves. It is found that the displacement process takes place much faster in simulations than in experiments. This is believed to be due to presence of dynamic effects not captured with the numerical model.

In this study the behavior of an unsaturated sand is examined. The sand used is Hostun sand from France. The Soil-Water-Characteristic Curves (SWCCs) of loose and dense specimen were determined by using the suction mode and pressure mode axis translation technique under various vertical net stresses. One dimensional compression and rebound behavior was studied for specimens at different initial matric suction values. During these tests the matric suction was maintained constant during loading and unloading processes. To study the wetting induced collapse of the sand, specimens with a known matric suction and water content were prepared and then saturated at several vertical net stresses during loading. The study showed that the SWCC is not significantly influenced by vertical net stress. However there is a influence of initial suction on the compression behavior of the material which is manifested on the stillness moduli, compression index and collapse potential of the sand used in this study. In comparison to these results, results from one dimensional compression and rebound tests, where water content was kept constant, for Quartz sand Weferlingen were added. Results from collapse potential test for Quartz sand Weferlingen are given, too.

The paper presents an experimental investigation into the permeability of compacted kaolin. Tests were performed in a suction-controlled oedometer using the air overpressure technique: ambient air pressure (raised above the atmospheric pressure) was kept constant and pore water pressure was modified. Permeability was measured by monitoring water inflow in response to a suction decrease and interpolating experimental data using a simplified solution of Richards’ equation (unsteady-state method). Corrections were made to account for the impedance of the high air-entry ceramic disc, the water lost by evaporation into the air pressure line and the air diffusing through the ceramic. The experimental results have shown that axis-translation technique may be problematic at high degrees of saturation and that permeability can be significantly underestimated.

Moisture content and moisture movement are key factors in controlling the progress and rate of biodegradation within a landfill as it is the aqueous environment that facilitates the transport of nutrients and microbes. The modelling of infiltration and water movement requires information on the moisture retention and hydraulic conductivity properties, usually in the form of empirical functions such as those proposed by van Genuchten (1980). In waste however, the particle and pore size distribution, heterogeneity of waste composition and leachate chemistry complicate the determination of moisture retention and hydraulic conductivity. In this paper we describe the modification and use of a standard pressure plate apparatus to establish moisture retention properties of samples of 250 mm in diameter and the difficulties of using this method. Some initial results are presented.

A recently developed two-component model for the relationship between saturation and pore water suction (pF-curve) is used, whose frame consists of simple, physically meaningful values. The main purpose in this paper is to describe the amount of pore water in soils at high suction. Relative air humidity may be regulated by salt solutions of defined concentration or by sulfuric acid of varying concentrations, which leads to a steady-state water content in soil specimen. In the range of high suction values, data achieved in this way are fitted by the twocomponent model with high accuracy. The comprehension of many phenomena such as drying and cracking, apparent overconsolidation near the surface or strength loss of cuttings in overconsolidated soils by weathering depends on the understanding of water binding behavior in soils, which is improved by this quantitative model.

Clays are geomaterials which are used in various applications in ur society. The mor common geotechnical applicatio s are: waterproof cores of dams, barriers to contain waste, slurry walls etc. But even if clays are not used as construction materials the geotechnical engineer encounters them frequently during construction of e.g. foundations and tunnels. Within the group of clay minerals we find individual clay species that possess geotechnical properties that vary enermously. These differences in geotechnical pr perties can b related directly to the differences in microstructure of the clay minerals. In this contribution it is shown how the knowledge of clay microstructure from X ray diffraction analysis, thus knowledge of the clay mineralogy, can be used to assess the geotechnical behavior of the clay. When mineralogy alteration processes of the clay are know from appropriate geological analogues, the geotechnical characteristics of the engineered clay structure can be estimated on the basis of these processes (e.g. containment of hazardous waste, infiltration of permeates in dams etc.). This implies however that the clay used as construction material is well characterised. A method to prepare such a clay atlas and how to use it is presented in this contribution.

The paper aims at investigating suction induced by static compaction in clay specimens. To this end, kaolin powders prepared at target water contents were statically compacted in a shear box apparatus. Trento (TNT) high-suction tensiometers (0-1800 kPa) were in stalled through the loading pad to monitor suction changes during the loading-unloading paths. Specimens were compacted at water contents ranging from 0.22 to 0.30 and at vertical stresses of 300,600, and 1200 kPa. This made it possible to explore abroad spectrum of compaction conditions. Experimental results are presented and discussed in the paper. The most striking aspect of the results is that suction of unloaded specimens increased as degree of saturation increased. This can be explained by the dependency of the main wetting curve upon void ratio.

Loess is a loose, open-structured metastable soil of aeolian origin, predominantly composed of 20–60 µm quartz particles bonded together by clay particles and, in some cases, carbonate compounds. When dry it can withstand high overburden stresses, whilst upon saturation it collapses creating potentially enormous engineering problems. The mechanisms controlling this metastable behaviour involve the disintegration of inter-particle clay and chemical bonding and variations in the pore water pressures (i.e. suctions) during saturation.

The paper describes methods of creation of artificial loess samples using an airfall approach to allow the variation of bonding constituents, arrangements and formation processes. The amount of powdered clay added to primary quartz particles (ground silica) was varied and three methods for activation of clay bonding were employed, i.e. water spray, capillary wetting and steaming. The reproducibility of the synthetic loess created was determined through oedometer testing of the resulting samples, the results of which are reported together with those for undisturbed samples of mid-European loess. Similarities and differences are discussed in the context of the likely bonding mechanisms. It is concluded that creation of reproducible synthetic loess samples, while controlling its constituents, makes possible the individual examination of the different parameters that control bonding in loess.

Natural collapsible residual soils are a common occurrence in Brazilian unsaturated deposits. The work presented in this paper is part of a larger project designed to identify the most important parameters controlling the mechanical behaviour of this type of material. A laboratory testing programme has been carried out in artificially unsaturated cemented samples reproducing some of the characteristics of natural collapsible soils. A technique for sample preparation has been developed, which has resulted in specimens with high values of void ratio and various degrees of cementation. In particular, it has been decided to use soil cemented mixtures with expanded polystyrene particles which has led to light samples with low density, meta-stable mechanic structure, and good workability. These artificial samples have been subjected to direct shear and oedometer tests, including conventional and suction-controlled oedometer tests. Oedometer tests have been used to characterize the behaviour of this material and to quantify the potential collapse due to loading and wetting. From the observed behaviour it is suggested that the initial void ratio, cement agent and initial suction are all important factors influencing the potential collapse of the soil. For saturated artificial soil samples the critical state lines in the (v, ln p’)-plane and (q, p’)-plane have also been calculated by using an approximated procedure based on the results of direct shear tests.

The main goal of this paper is to present the first results of a study performed to provide insights into the relationship between the hydromechanical stress path experienced by a compacted soil and the modification of its pore space geometry. A new oedometer employing the axis translation technique was used to characterize the hydromechanical behaviour of the tested material. The fabric of the tested samples was determined using the mercury intrusion porosimetry technique under 4 stress levels and two different suctions. From this information, the macro and micropore volume variations were then determined. The results showed that mechanical loading produced a progressive reduction of the macropore volume and a significant increase of the micropore volume beyond a stress of 250 kPa. The obtained results tend to demonstrate that suction strengthens soil fabric, as the initial “double structure” of the tested material was not destroyed in the case of the unsaturated sample by the loading up to 1 000 kPa, whereas, the saturated sample exhibited a more homogenous fabric.

The paper presents an experimental determination of the osmotic component of suction by means of the squeezing technique. Soil was mixed with distilled water at different water contents and squeezed at different pressures. Osmotic suction was estimated from electrical conductivity of water extracted from the soil and two different empirical relationships were used to estimate osmotic suction. Experimental results have shown that osmotic suction depends upon the extraction pressure and the initial water content. The osmotic suction measured using the squeezing technique was then compared with the difference between total suction (measured with transistor psychrometer) and matric suction (measured with axis translation technique).

This paper is a study of the correlation between the humidity, density and matrix suction of non-plastic soil taken from a region located near the city of Rio de Janeiro. The study was conducted through laboratory tests, using minilysimeters and equipment that indirectly measured the humidity through soil suction: in effect, an automated tensiometer, an equivalence tensiometer (EQ2), a TDR and a specially adapted system, which measures the suction in situ, using filter paper. The electronic instruments were connected to a data logger and placed in three mini-lysimeters. Cycles of wetting and drying of the confined material were simulated. The soil was characterised. This characterisation enabled the determination of physical indexes, density, plasticity, grain size, porosity, and the characteristic curve of water retention, using the Richards pressure pan (cooker) test. The results obtained by the instruments, humidity and matrix suction, were compared with the characteristic curve. Overall, the equipment was satisfactory. The low presence of clay in this soil, around 7%, and its associated high porosity, were factors in limiting water retention. The equilibrium time of the soil humidity and the filter paper, obtained at intervals of fifteen days, presented better results than the weekly intervals (seven days). The equivalence tensiometer EQ2 compared to the filter paper presented quite efficient results for high suction values.

The paper presents suction measurements on pyroclastic soils using vacuum tensiometer and high-suction tensiometers. Laboratory measurements with the vacuum tensiometer aimed at assessing the performance of this instrument, which had to be used for field measurement of matric suction. These measurements were also used to develop and validate a numerical model for predicting air cavity growth in the vacuum tensiometer. The model was intended to be a tool for detecting possible malfunctioning of the tensiometers in the field. Measurement with high-suction tensiometers aimed at corroborating the results of vacuum tensiometer measurements. When using high-suction tensiometers, contact can be better controlled and suctions approaching the limit value of the vacuum tensiometers (about 80 kPa) can be measured with no risk of cavitation.

Soils in arid and semi-arid zones undergoes volume changes due to wetting. Depending upon the type of clay minerals present, degree of saturation, externally applied load and bonding, the fine grained soils either swells or compresses. One of the parameter that affects the volume change behaviour is the primary clay mineral present in their clay size fraction. A simple method of identifying the same has been presented. It has been brought out that in an expansive unsaturated undisturbed soil, the diffuse double layer repulsion, the stress state and the bonding play significant role in their volume change behaviour. In nonexpansive fine grained unsaturated undisturbed soils, the shearing resistance at particle level (including the matrix suction and bonding) and fabric play a significant role in influencing the volume change behaviour. While both the mechanism co-exist, one of them play a dominant role depending upon the primary clay mineral is swelling or non swelling.

This paper presents preliminary results of research on hydro-mechanical characterisation of a highly compacted bentonite-sand mixture for its potential use as buffer material in nuclear waste disposal system. Development of swelling pressure with change in suction is presented. Wetting curve of the material was derived from the test as suction was reduced from its initial value using vapour equilibrium technique as well as via axis-translation technique. For comparison, three-dimensional free swell tests were also performed by inducing different suction paths into several specimens such that wetting and drying curves could be obtained. No significant swelling pressure development was observed as the specimens were wetted from as-prepared and oven-dried states to about 10000 kPa suction. Comparison with the three-dimensional free swell test results leads to a conclusion that from dry state to about 10000 kPa suction, the maximum swelling pressure upon saturation is suction independent. It is also shown that swelling pressure curve, wetting-drying curves and void ratio-suction relationship generally exhibit hysteresis. Moreover, the results indicate that even at high suction, void ratio change plays an important role in determining the shape of the wetting curves.

Most countries favour compacted air-dry bentonite for engineered barriers in final repositories for nuclear waste. While the hydrophilic properties of bentonite appear to be ideally suited for minimising any contact of water with the waste the process of resaturation is not fully understood yet. In order to investigate the dynamics of water uptake several resaturation experiments with liquid water as well as with water vapour have been performed. A series of tests concerning the uptake of liquid water in compacted MX-80 bentonite samples has been completed recently. The experiments provide uptake rates and moisture distributions as a function of time. Analogous experiments with water vapour are presently running. Some first results are presented. The data gained is used to check new conceptual models which explain resaturation by flow of liquid water and by diffusion of water vapour, respectively. In the new models the effects of hydration on the local pore water content and the change of porosity corresponding to the amount of hydrated water are considered. The results strongly suggest that vapour diffusion plays a significant - if not dominant - role in the resaturation process of bentonite.

There are two widely accepted approaches to determine the shear strength of unsaturated soils: the effective stress approach (Bishop, 1959) and the independent stress state variables approach (Fredlund et al., 1978). The main difference between these two approaches lies in how to reduce the effect of increasing matric suction. Bishop’s coefficient χ can be calculated using

ϕb

(χ = tan

ϕb

/ tan

ϕ’

) and vice versa. However, Bishop’s approach with the coefficient χ “of the matric suction efficiency” seems to be closer to the reality since there is a difference between decreasing ϕb and the experimental results, which have proved that the effective friction angle slightly increases with increasing matric suction. The coefficient χ decreases with increasing suction as a result of decreasing both the total area of water-solid contacts and number of particles connected by water menisci. This explains why the parameter χ decreases with increasing porosity. Thus, the maximum effect of matric suction can be reached with higher water content as a result of two opposite influences: decreasing matric suction and increasing number of particles connected by water menisci (matric suction efficiency χ).

The analysis of this problem is based on the experimental programme of the bentonite buffer material under unsaturated conditions. Pure bentonite as well as bentonite mixtures with siliceous sand and graphite were tested. Samples with extremely different porosities were prepared under compaction pressures ranging from 300 kPa to 100 MPa and then tested in the triaxial apparatus.

Research in the field of unsaturated soil mechanics for high plastic clays is very active. One area of considerable current interest is development of general constitutive models for unsaturated clay based sealing materials in a frame work that can be implemented in numerical modeling tools (Alonso

et al

. 1990, Delage and Graham 1995, Toll 1990). In particular, more quantitative information is required to define the features of yielding, failure and strain hardening for predictive modeling applications. Soil suction must be controlled and independently measured in laboratory tests. This will allow examination of behaviour along any stress path that can be expected to occur in engineering applications which will provide the necessary material information to calibrate and validate proposed constitutive frameworks.

This paper presents details of laboratory tests in a custom triaxial system with stress path automation and independently controlled and measured suction (Blatz and Graham 2000, 2003). Details describing the equipment will be given along with selected results for the yield, strength, and strain hardening behavior of a high-plastic sand-clay material at suctions from 5 MPa to 160 MPa and isotropic pressures from 1 MPa to 6 MPa. The results demonstrate the importance of independent measurement and control of suction along well controlled loading paths for interpreting the behaviour of unsaturated high plastic clays.

Investigation into the mechanical and hydro-mechanical behaviour of unsaturated sand requires special laboratory equipment for testing and preparing the specimens. In this paper, the requirements for a triaxial testing device for unsaturated sand are presented. With respect to the requirements for testing unsaturated sand (axis-translation technique), measuring small overall and pore-water volume change, and the control or measurement of small values of suction, a new triaxial device has been developed. It consists of a double-walled triaxial cell, a modular loading frame including the axial power unit, and a pc for controlling and data logging. The device enables the determination of the shear strength, the SWCC, and the hydraulic conductivity of unsaturated sand.

This paper deals with a theoretical description of the undrained and drained behaviour of unsaturated soils under isotropic load. By taking into account the compressibility of the solid mass and the fluid/gas mixture inside the pores a theoretical relation between the pore water coefficient (B) and the saturation (S

r

) is derived, which is shown to be similar to measured data given in the literature. The approach described involves natural strains, schematizations of the solid mass and the fluid/gas mixture and the use of different boundary conditions for respectively B=1 and B<1.

One-dimensional consolidation tests are often performed without accurate control of the initial degree of saturation, since most commercially used testing apparatuses do not permit back-pressuring. As a result, the initial degree of saturation is often less than unity and, thus, the measured ‘Apparent’ Compression Index can be very different than the value corresponding to a fully saturated sample. Similar differences are caused by most oedometer tests being performed up to a maximum stress lower than the maximum pre-consolidation pressure. Oedometer tests on undisturbed samples, on initially unsaturated samples of reconstituted soil, and a data-base of commercial oedometer tests are presented in order to exhibit the effect of high maximum preconsolidation pressure, cementation and initial unsaturation. Finally, expressions for the volumetric deformation of unsaturated soils proposed by Alonso et al. (1990) are used to explain different trends exhibited by the ‘Apparent’ Compression Index for low- and high-plasticity soils.

This paper presents the experimental results from several conventional oedometer tests conducted on compacted and unsaturated soils. Soil samples of medium plasticity, but with different grain size distribution and degree of saturation were tested. Some of these tests were interrupted before consolidation was achieved, and the water content of the soil was measured both before and after each test. The results point out that the recorded time-settlement curves consist of two distinct stages that are characterized by a different settlement rate. The early stage should essentially be ascribed to the faster dissipation of air pressure, while the final stage should be caused by the gradual expulsion of water from the soil. The results are also analysed using Terzaghi’s theory [1] in order to ascertain whether this theory is suited to simulating the time-histories of settlement recorded during the tests.

Volume change as a result of drying is often neglected in soil mechanics and soil hydrology, despite the important influence it has in the change of mechanical stability and water flow. Therefore, processes which lead to volume change have to be understood. Tensile stresses as the main parameter for shrinkage are a result of hydraulic and mechanical mechanisms in unsaturated soils or soil substrates. Both mechanisms have to be recognised as dependent processes. Unsaturated soils are defined as 3-phase systems. Capillary forces in soil pores act as contractive forces of the liquid phase on the solid phase. The resulting tensile stress caused by water increases with decreasing degree of water saturation. This causes shrinkage in a given soil volume, including soils with small plasticity. Mechanical stress parameters will simultaneously be changed with shrinkage, which as a result also change the hydrological parameters altering the pore system. The separation of the mechanical from the hydraulic stressis difficult. Therefore, a method was developed, which allows the determination of tensile stress under defined boundary conditions and is based on the general stress equation. Also a method is described by which this information is used for general modeling of volume change by hydraulic stress and general empirical functions used in hydraulic modeling.

The paper presents experimental results linking matric suction and tensile strength of compacted clays. Test results from a cohesive soil are presented and discussed with respect to the soil structure and the interaction of soil and water. It is assumed that two main groups of pores can be clearly identified in compacted clays; the pores between aggregates (interaggregate pores) and pores between particles (intraaggregate pores ). Based on a description of soil-water-interaction an expected behaviour, describing tensile strength as a function of matric suction, is derived and compared with the experimental results. The laboratory test results indicate that there is a strong correlation between the pore size distribution (assessed by interpretation of the soil water characteristic curve SWCC) and the tensile strength of compacted soils. Furthermore, the test results are compared by using micro-mechanical considerations of the interaction between the skeleton of unsaturated soils (interparticle contact force) and by using numerical calculations with an elastic relationship.

The UK Environment Agency currently maintains over 35,000 km of coastal and flood defence embankments in England and Wales with an annul expenditure of approximately £400m. The majority of the flood defences are earth embankments built from locally available materials using traditional construction methods. Although the performance of the flood defence embankment can be jeopardised by several different failure modes, instability due to the fine;l,,,lay fill has been identified as one of the main causes of failure especially along the eastern coast and Tha dfg mes Estuary. At these locations medium to highly plastic clays are commonly used as fill material. The fine fissuring of these clay fills can reduce the mass permeability of the embankment, leading to a rapid ingress of floodwater with catastrophic collapse of the inward face. As part of a wider research programme into the fine fissuring of clays, preliminary laboratory tests have been carried out into the formation of fissures for a variety of clays with different plasticities. The tests investigated an empirical relationship between the geometry of thin discs of soil with the onset of fissuring for different plastic indices.

Marly formations are widespread in Greece with major works founded on, or excavated in them. Perhaps the most spectacular of them is the Corinth Canal, a ca. 7 kilometre long canal with very steep (average inclination 4.5: 1) and Corinth marl. The excellent stability of the steep canal slopes has motivated extensive laboratory investigations of the mechanical properties of the Corinth marl at the Geotechnical Department of the National Technical University of Athens since several years, the latest of which involved the drying properties of the Corinth marls. The drying portion of the soil-water characteristic curve was determined from undisturbed block samples and from reconstituted samples re-consolidated to the in-situ stresses. The airentry pressure was found to be higher in the case of the undisturbed samples compared to the reconstituted/re-consolidated samples. Furthermore, the variation of the void ratio with suction up to the air-entry pressure was found to be close to the curve of void ratio versus effective stress from one-dimensional consolidation over the same stress range. This behaviour constitutes evidence of the validity of the effective stress principle up to the air-entry pressure. Finally, the paper concludes with estimations on the apparent cohesion change of the Corinth marl with suction.

Weather-related geo-hazards are a major concern for the railway industry m Canada. The financial losses that result from derailments and delays amount to millions of dollars every year. On the other hand, the assessment and management of geo-hazards is a difficult problem that involves complex coupled phenomena and numerous soil and weather parameters. The primary goal of this paper is to illustrate the manner whereby unsaturated soil mechanics can be taken from the soil property assessment level using techniques based on the soil-water characteristic curve (SWCC), to the solution of this real and highly complex problem. First, a concise description of the weather-related geo-hazards assessment model (W-GHA model) is given. Deterministic and probabilistic aspects of the model were developed within a Decision Analysis framework. The deterministic core of the model consists of a two-dimensional stability analysis combined with the analysis of the effects of weather conditions on the pore-water pressures. According to the proposed model, weather conditions interact with the ground through the flow of liquid water, water vapour, and heat. Critical embankment stability conditions are determined using a Dynamic Programming Method (DPM) combined with Finite Element based stress fields. The soil system is ultimately represented by a series of partial differential equations (PDE’s) governing conservation of mass and momentum. A discrete stochastic analysis is implemented within the proposed framework. Several unsaturated soil property functions are required as input to the system of PDE’s. The hydraulic conductivity (i.e., coefficient of permeability), vapour diffusion coefficient, thermal conductivity, volumetric specific heat, and shear strength are all nonlinear functions physically related to the SWCC. The methodology by which these soil property functions can be inter-related to the SWCC is presented. The theoretical model illustrates the manner whereby it is possible to quantitatively assess embankment stability based on weather conditions. The methodology is feasible and yet relatively comprehensive.

A fine-grained slope that exhibits slow movement rates was investigated to understand the mechanisms which lead to a consecutive development of mass movements in the Vorarlberg Alps (Austria). For that purpose intensive hydro-meteorological, hydro geological and geotechnical observations as well as survey of surface movement rates were conducted from 1998 –2001. Subsurface water dynamics at the creeping slope turned out to be dominated by a 3 dimensional pressure system. The pressure reaction is triggered by fast infiltration of surface water and subsequent lateral water flow in the south west part of the hillslope. The related pressure signal was shown to propagate further downhill, causing fast reactions of the piezometric head in 5.5 m depth on a daily time scale. The observed pressure reactions might belong to a temporary hillslope water body that extends further downhill. The related buoyancy forces could be one of the driving forces for the mass movement. A physically based hydrological model was adopted to model simultaneously surface and subsurface water dynamics including evapotranspiration and runoff production. It was possible to reproduce subsurface pressure reactions and observed runoff in principle. However, as soil hydraulic functions were only estimated on pedotransfer functions a quantitative comparison between observed and simulated subsurface dynamics is not feasible. Nevertheless, the results suggest that similar methods of coupling surface and subsurface processes should be employed in coupled models for large mass movement and that it is possible to reconstruct important spatial structures based on sparse observations in the field.

The paper deals with the stability of slopes under drawdown conditions taking into account the effect of negative pore-water pressures, usually neglected in routine analyses. A simplified unsaturated soil profile above the water table was assumed and soil suction was estmiated using a Brooks and Corey type soil-water characteristic curve. A parametric study was carried out to evaluate the influence of geometrical, physical and mechanical parameters of a simple slope. The results for complete and rapid drawdown conditions show that the factor of safety varies less than 10% for dams or reservoirs with initial drawdown ratios less than 0.3. Soil suction becomes more significant (safety factor variation up to 50%) for railway or road embankments with the water level at the base of the slope.

The papers presents a numerical analysis of the stability of a moraine slope that reactivated after two heavy rainfall events in November 2000. To analyse the pore water pressure distribution within the slope, rainfall and evapotranspiration were applied to the slope over a period of four years that preceded the main landslide reactivation. Because of the uncertainty about the saturated hydraulic conductivities of the landslide soils, different combinations of these coefficients were analysed. The factor of safety of the slope was then calculated at the time where the slope moved, as recorded by the inclinometer measurements. This made it possible to verify which combinations of saturated hydraulic conductivities were compatible with the slope movements.

Research on a bentonite-based engineered barrier for use in the safe underground disposal of high-level radioactive waste is a special multidisciplinary issue. In order to obtain findings enabling the design of such a construction, it is necessary to employ all the available experimental tools and procedures. With respect to the extremely long-term time requirements for the rheological stability and safety of the system as a whole, the results of long-term research have fully justified the use of physical modelling. The most relevant model types applied have been found to be those made at a scale of 1:1, referred to as Mock-Up models. A vertical model of a bench-scale buffer mass test of Czech smectitic clay in the KBS-3 modification (Swedish system) has been built at the Centre of Experimental Geotechnics, CTU, Prague. The model is fully instrumented, providing the continual multiparametrical measurement of all the relevant parameters. The whole experiment including the development of all the parameters measured can be seen on free-access dynamic web-sites (http://ceg.fsv.cvut.cz).

Since 1974, Belgium investigates the design for disposal of its High Level Radioactive Waste (HLW) in a deep clay formation, the “Boom Clay”. Although the clay formation is the main (natural) barrier against the transport of the radionuclides towards the biosphere, the design also involves several engineered barriers (multi-barrier principle). In the design developed in the late 1980’s, a non-saturated bentonite based material was chosen as part of this barrier system. Prior to demonstrating this design in

in-situ

conditions, a surface mock-up test has been operated between 1997 and 2002. This test served as a preliminary test on the performance of several components of the system, such as bentonite based backfill blocks and instrumentation. With clearly defined heating and hydration conditions, it gave us the opportunity to perform a large scale simulation of the hydration/saturation of the backfill at controlled conditions. After describing the general disposal design and the experimental set-up, this paper will detail the measurements and observations obtained during operating and dismantling the mock-up. To support the interpretation of these measurements and observations, a modelling of the experimental set-up is being performed. We further detail the characterisation programme carried out to obtain the input data for the modelling. Finally, lessons learned for the development of the design for the HLW disposal will be drawn.