Bertrand Vignal
ANDRA
France

The alternatives for deep radioactive waste disposal in France are being studied within the legal framework set in 1991. ANDRA, the National Radioactive Waste Management Agency, is in charge of the study, which demands two complementary approaches.

First, the search for potentially favorable repository sites where underground research or so-called ‘on-site’ laboratories will have qualification purposes. Initiated in France in 1994 with reconnaissance surveys, three sites were finally proposed, two in sedimentary clay formations and one in crystalline rock, the principal geological characteristics of which will be explained further on. The examination of the applications to build site specific laboratories is under way, and scheduled for completion in 1998. This will set the stage to start sinking the shafts, and the qualification will take place after the turn of the century.

Secondly, to prepare for the implementation of the qualification programs in the future ‘on-site labs’, ANDRA is conducting research in accessible underground installations, which we will refer to as ‘off-site labs’, essentially located outside France. This R&D, conducted with international cooperation in Switzerland, Belgium, Sweden and Canada is aimed at determining what happens to the waste that is disposed of, understanding and modelling the associated mechanisms, qualifying instruments and training personnel. The logic and nature of this R & D and the results obtained are presented hereafter.

Following a short presentation of the French legal framework, the complementarity of both ‘ off-site labs’ and ‘ on-site labs’ will be shown. Then the geological setup of each French specific site, as presently evaluated by surface survey, will be indicated.

The first siting operations in France in the late Eighties were thwarted by the lack of public acceptance. A Parliamentary committee took over the matter, and a law was passed in 1991 covering research into radioactive waste management. The law requires three alternatives :

• separation and transmutation,
• disposal in deep geological formations,
• long-term storage.

The CEA (the French Atomic Energy Commission) is in charge of the first and third alternatives, and ANDRA has the second. ANDRA, the National Radioactive Waste Management Agency, is a public establishment, reporting to the Ministries of Industry, Research and the Environment. Its operations are funded by the radioactive waste producers.

Siting operations were initiated by ANDRA in 1994 in several areas selected for their geological interest and their public acceptance. This led to applications in 1996 to build underground laboratories at three potentially favorable sites.

The applications were examined (public enquiries, agreement of local officials and of the National Evaluation Commission) in 1997. The governmental decisions to inaugurate sites are expected in the course of 1998.

The siting carried out in France within the aforementioned legal framework concerns sites that are potential repositories. On each of these sites, the future specific underground laboratory, the ‘on-site lab’, will qualify or disqualify the possibility of disposal.

To do this, research programs are drawn up, adapted to the host environment, clay or granite, to check the feasibility and demonstrate the safety of the disposal concepts. These programs, in accordance with the Basic Safety Rule issued by the Nuclear Installation Safety Directorate, are essentially aimed to :

• take measurements on the rock in situ and on fluids disturbed as little as possible by the experimental conditions, to obtain further knowledge of the parameters already partly evaluated by the surface survey,
• conduct more comprehensive experiments to determine the behavior of the different types of rock and fluid, allowing for natural phenomena and any changes caused by construction of the repository,
• survey the formation, particularly its spatial variability, to evaluate the capacity of the site, followed by a tentative positioning of the drifts and future repository facilities,
• determine the methods for excavating, backfilling and sealing the cavities.

Some of the tasks in these programs can be performed using conventional scientific approaches, but others require new developments. Understanding groundwater flow in a clay bed or fractured crystalline rock is a case in point, in dealing with a disposal problem where water is a potential vector for radioactivity transmission. These processes must therefore be observed in situ in deep rocks, following the logical steps identified in experimental physics: adaptation of observation tools, data acquisition and interpretation, modelling.

Because of this, most countries have decided to start research programs in mines or in specially built underground installations. In the early Eighties, methodological ‘off-site’ underground laboratories, open to international cooperation, were set up abroad, one in clay, and others in crystalline and rock salt formations.

In the context of these developments, France’s ANDRA decided to participate in the only existant laboratory in clay in Belgium and to follow the Swedish Stripa Project in granite. In the early Nineties, ANDRA then expanded its cooperation with foreign laboratories and now works in two for clay (Belgium and Switzerland) and three for granite (Sweden, Canada and Switzerland) preparatory to the forthcoming assessment of French sites.

ANDRA has been investigating three potential deep repository sites in France since 1994, one in granite and two in clay.

The granite site is located in the Vienne Department, in a granite formed 360 million years ago, with a mineralogical and chemical composition that ensures some plugging of the fractures. This granite was covered by sediments of Lias age. The 150 m thick sedimentary cover contains two aquifers located in limestone formations.

The surface geological mapping, geophysical surveys and sixteen cored boreholes helped to pinpoint the laboratory location at some distance from the major kilometric faults. While most of the metric fracturing is filled with clay or carbonate materials, the hectometric fractures could be conductive. The conductive fractures contain salt water approaching sea-water in composition, with salinity increasing with depth.

The underground laboratory should be installed at a depth of about 450 m with access by two shafts, and is expected to confirm and clarify the conceptual hydrogeological model developed by the surveys.

The second site is located in the east of the Paris Basin, at the border of the Meuse and Haute Marne Departments. The target sedimentary formation is a layer of indurated silty clay, 130 m thick and Callovo-Oxfordian in age (about 150 million years).

Several cored boreholes were drilled in the sector selected for the laboratory. The layer appears to be sub-horizontal and uniform over wide extensions, due to its deposition in a marine environment far from the shoreline. It also appears to be very impermeable (K = 10^-12 m/s) and its 45% of clay minerals are of a type imparting good radionuclide retention capacity.

The architecture of the underground laboratory will feature a single level at 500 m depth, in the middle of the layer, with access by shafts.

The third site is located on the banks of the Rhône, near Marcoule, in the Gard Department. It is also a layer of silty clay, but of Cretaceous age (about 100 million years). The layer is 300 to 400 m thick, and lies at a depth of 400 to 800 m.

Since the site is located in an area of rapid variations in sedimentary thickness, corresponding to the filling of a rift valley, potential facies changes will have to be investigated in the underground laboratory.

The laboratory, which will also have to check the absence of fracturing, will feature two levels at 500 and 700 m owing to the thickness of the layer. Access will be provided by shafts.

Assuming that a green light to build underground research laboratories is given by mid 1998, a tentative schedule of both construction and operation can be presented.

Surface borehole drilling would resume with no delay and the construction phase would actually begin one year later with the sinking of the shaft which would last about two years.

Therefore the excavation of the first drifts, giving access to the underground experimental field and fully opening the operation phase, could take place in 2001.

After some years of functioning, site evaluation reports will be issued in 2006, as required by Law, which will give Parliament the necessary elements for further decisions.

The knowledge already acquired on the French sites guides the experimental preparation carried out abroad. In fact two programmes, one in granitic rocks and one in clays have been implemented which are successively presented.

ANDRA is determined to make headway on two major topics : the understanding of groundwater flow in a fractured rock mass, and the ability to seal excavations.

To do this, the Agency is accordingly exploiting the opportunities it has been offered, and is working in three international methodological laboratories : the Swedish SKB’s HRL (Hard Rock Laboratory), the Canadian AECL’s URL (Underground Research Laboratory), and the Swiss NAGRA’s Grimsel Test Site. Practical experience with three different crystalline formations should help to adapt the knowledge gained to the specific characteristics of the French Vienne site.

Experiments in Sweden on groundwater flow and radionuclide transfer - containment capacity of the geological barrier - are being conducted on different scales. One examines an isolated fracture, while the second investigates a hectometric block. In groundwater flow modelling, an international group involving different teams and programs is addressing specific tasks, such as the study of the hydraulic drawdown caused by excavation. Geophysical work has also resumed in Switzerland, to assess the limitations of remote fracture detection.

Concerning excavation effects, an experiment has terminated in Sweden after recording the characteristics of disturbances generated by excavation. The results are being used in the gallery sealing project in Canada. Boreholes are also being sealed in Switzerland.

Before going on to review a number of ANDRA’s recent operations, carried out alone or in participation, as part of its methodological preparation for the French Vienne site, a specific feature of granitic rock should be kept in mind :

• Due to their process of formation, crystalline rocks are inherently fractured. They are made up of a very low porosity rigid matrix criss-crossed by numerous fissures and fractures. Water may circulate in some ‘conductive’ fractures, which are not totally blocked by mineral fillings.
• Geologists class the fractures in families, generated by tectonic evolution. To take this classification a step further, in order to obtain useful hydrogeological data for a safety analysis, the hydraulic characteristics of each family must be identified at various scales.
• A key element, the hydrogeological model, can be constructed using the techniques for locating the conductive fractures and determining their characteristics. These models have primarily been developed in Sweden, where ANDRA joined the HRL program at an early stage.

At the Swedish HRL (as earlier at the Canadian URL), excavation of the laboratory was a major hydraulic experiment in itself. Before excavation, the surface boreholes intersected and placed instruments in a large number of conductive fractures. The drawdown resulting from drainage at the start of underground work would thus be monitored in the fractures, providing accurate data for the hydrostructural model on a multi-hectometric scale. A similar set of instruments will be installed around the first shaft at the Vienne site.

At hectometric scale, an experiment initiated at HRL is aimed at identifying the characteristics of the groundwater flow and solute transport in a granite block with low circulation under a small hydraulic gradient. The surveys run in the boreholes are used to build a hydrostructural model: fracture mapping, seismic tests, hydraulic tests, tracer tests. These procedures will be transposed to the Vienne site.

Another study carried out by ANDRA on a decametric scale at the HRL measured the characteristics of an isolated fracture. The fracture selected was crossed by a dozen boreholes whose cores (decimetric) were used for morphological, hydromechanical and tracer studies with the aim of understanding the groundwater flow mechanisms. A comparison of results obtained from the fracture in situ and from the cores will help draw up a strategy.

While it is important to identify the characteristics of a given fracture, it is also important to find any conductive fracture some distance from the works. To do this without drilling the rock, geophysical methods are unavoidable. ANDRA has tested electromagnetic and acoustic methods on the Grimsel site. Radar developments concern both interpretation and equipment. In acoustic methods, the main efforts concern borehole transmission sources.

Closing a repository depends on expertise with seals, the ability to create them, the knowledge of their hydraulic performance and their durability. In granite, the rigidity of the rock means that the walls will not converge and the excavation-disturbed zone cannot heal over time. This is why it is important to know the characteristics of the area surrounding the gallery walls.

One experiment at the HRL was aimed to evaluate excavation effects and measurement methods. Comparisons were made between standard excavation methods, drilling/blasting and tunnelling. Away from the wall, the overall permeability of the rock is relatively unaffected by the stress and pressure distribution. In the gallery wall itself, however, changes related to the excavation method can be observed. Blasting causes several decimeters of fractures and micro-fractures, with repercussions on permeability. Tunnelling only generates micro-fractures a few centimeters long.

Studies of the disturbed zone will continue with tests of seal performance. Two full-scale gallery seals will be installed at the Canadian URL in early 1998, one made of bentonite blocks (swelling clay) and the second of concrete. ANDRA’s contribution involves mapping the effects of excavation, modelling and instrumentation.

Closure also implies sealing the boreholes. In partnership with NAGRA, ANDRA is testing several technical alternatives at Grimsel.

This experimental program in crystalline formations outside France still has many years to run. It will be extended to subjects related to radwaste packages nearfield.

ANDRA’s first experiments with clay started in 1984 in the Mol CEN/SCK’s laboratory in Belgium. The scientific approach taken by the laboratory teams was progressive. This enabled them to compile experimental data, develop laws of mechanical behavior, study the effect of heat, and evaluate hydrogeochemical investigative methods.

A desaturation experiment has just been completed at Mol, connected with the problem of sealing. The sealing of a small shaft and the instrumentation of the planned laboratory extension have been started at Mol.

R&D associated with the preparation of the experimental programs for the East and Gard sites raised the question of the weak mechanical analogy between the plastic clay of Mol (25% water content) and the stiff French clays, closer to marls (5 to 10%). A search was accordingly made for sites with characteristics closer to those of the French sites.

A marl layer was first identified in the Alsace potash mines (MDPA), where a good sized unsupported hole was drilled to gather information on the stability of such excavations.

More recently, NAGRA proposed the Mont Terri project in Switzerland. ANDRA became involved from the outset in this project, which is centered on an auxiliary gallery to a road tunnel which passes through marls. A program to measure the site’s characteristics started in 1996, and data are already available on water drainage, pore water pressure stress measurements, excavation effects and excavation methods. This partly involved running in the tools needed for larger experiments, such as tests in excavating galleries, diffusion tests, thermohydromechanical coupling etc. This program continues with a growing number of partners.

The Agency’s preparation policy was set out in detail when, from 1994 onwards, it learned more about the properties of the candidate formations in the East and Gard sites. The questions addressed were the excavation effects, viewed from the standpoint of sealing, hydraulic aspects and water chemistry.

The mechanical stability of excavations depends on the rheological properties of the host rock and the stresses to which the structure is subjected, essentially the weight of the overlying formation, the tectonic stress field and the nature of any supporting structures.

An unsupported vertical borehole, 1 m in diameter and 40 m deep, drilled at the MDPA, displayed great stability over time, with a few mm diametral convergence and some local flaking. Of all the behavior models tested, only the one which accounted for the fluid phase predicted convergences conforming to those observed.

Natural stress fields have a definite influence on cavity strength, but it is extremely difficult to measure these stresses in clay, which exhibits plastic behavior. However, they represent the boundary conditions of a mechanical problem. Hence one program at Mont Terri compares four measurement methods: overcoring, undercoring, borehole slotting, and hydraulic fracturing.

When a cavity is created, the rock is locally disturbed. This is due to stress redistribution, which leads to mechanical distortion. An area may appear in the wall with increased hydraulic transmission properties. New methods for recording the characteristics of the disturbance in forms of permeability have been validated at Mont Terri.

Excavation effects have mainly been studied with the aim of identifying their consequences on a seal. To acquire the necessary data, a small diameter shaft (1.4 m) will be sealed at Mol as part of a European program.

When a cavity is drilled in a deep clay formation, very little water drains out. The ventilation causes oxidizing conditions. The pressure drop around the excavation causes local desaturation of the clay.

A preliminary evaluation of the scale of this mechanism was made at Mol. From the phenomenological standpoint, the test showed evidence of changes in the hydraulic fields around the ventilated chamber accompanying the start of desaturation. It also revealed strong coupling between hydraulic and mechanical forces. These observations can be explained theoretically by the model of mass transfer in an unsaturated environment. One application of the model is the overall determination of hydraulic conductivity.

A major environmental parameter is the pore water pressure. This must be known at different points in order to calculate gradients, estimate convective flows, and interpret permeability tests. An improved piezometric tool was developed at Mol, but the very different hydrogeological context of the French sites demands a specific device, which was designed and tested at Mont Terri.

Samples of undisturbed pore water must be obtained to determine the water chemistry. Various sampling methods were tested in the clay at Mol, but they cannot all be transposed to the clay of the East and Gard sites. This is why further development was necessary at Mont Terri. Preliminary results indicate a good correlation in composition between water squeezed from the core and water collected on site.

At present, a new gallery is being excavated at Mont Terri, and diffusion and migration tests are scheduled in the next few years. Other projects may be undertaken linked to waste packages nearfield.

The knowledge acquired in the course of the experiments has not always been gained at the first attempt. We have had to try and try again, and have built up our knowledge simply by repeating these iterations.

Noteworthy among the major results already obtained are advances in the characterization and understanding of flows in fractured environments. Significant breakthroughs were also achieved on the problem of seals, in granite as well as clay formations, and the study of the mechanisms of formation damage by excavation.

Concerning instruments, improvements were made to geophysical methods for detecting fractures. Various instruments were developed, including those measuring the saturation of a clay or its pore water pressure. Procedures were also developed and validated, such as water sampling in clay.

Yet, all things considered, the chief advantage of international cooperation is the joint contribution of experts from different cultures on subjects of mutual interest. The result is the training of ANDRA’s multi-disciplinary teams in the practice of experimental physics, in field conditions similar to those to be encountered at the French sites.

Experimental practice also serves to identify, evaluate and involve the many contractors whose cooperation will be necessary for the implementation of the qualification programs in France.

Having analyzed the principal mechanisms occuring in a repository, and having identified the major characteristics of the French sites, the main features of the qualification program on each site are now being established.

The ideal result of the methodological preparation which continues abroad would be ultimately to have almost no R&D in ‘on-site lab’, but rather to implement a proven characterization program. The minimum expected from the preparation is an important saving in time as a result of a better understanding of experimental practice underground.

Obviously, the specificities of the sites mean that any simple transposition of methods developed ‘off-site’ cannot resolve everything. It will again be necessary to innovate, to resume R&D. ANDRA strives to be genuinely prepared for the first series of experiments it will have to conduct. After that, the results could raise new questions, which will spark further research and experiments.

Having witnessed this demonstration of the advantages of partnership projects (comparing and contrasting approaches and methods), ANDRA plans to invite the international scientific community to join in the programs of the future French research laboratories, when the time comes.

BACK