Faire avancer la sûreté nucléaire

La Recherchev2


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



Email :

Mot de passe :

Boisson, JY; Bertrand, L; Heitz, JF; Moreau-Le Golvan, Y HYDROGEOLOGY JOURNAL, 9: (1) 108-123.

Type de document > *Article de revue

Mots clés > hydrogéologie, laboratoire souterrain, milieu argileux

Unité de recherche > IRSN/DEI/SARG/BEHRIG

Auteurs > BOISSON Jean-Yves

Date de publication > 01/01/2001


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 (irsn) 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 (similar to3-5%) and very low radii access porosity. Diffusion (tritiated water) coefficients (1x10(-12) to 2x10(-11) m(2)/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.