Hydro-thermo-mechanical behaviour of the Tournemire compacted shale (SE of France)

Titre du congrès :Hydro Predict 2008 - International Interdisciplinary Conference on Predictions for Hydrology, Ecology and Water Ressources Management using Data and Models to Benefit Society Ville du congrès :Prague Date du congrès :15/09/2008

Within the framework of studies on fluid transfers in compacted argillaceous formations likely to host a radioactive waste geological repository, thermo-osmosis appears to be a process which can significantly affect fluid fluxes in this environment. However thermo-osmotic coefficients data in the literature are few and badly constrained as no in-situ experiments have been done, so far. Thermo-osmosis is a coupled transport process which consists in a water flux under a temperature gradient while chemical concentration remains constant. Proceeding in a temperature increase in a compacted clay rock, the system tends to reduce its entropy inducing a water flux to the cooler part of the shale. In the geological context, we can observe temperature gradients at the basin scale due to the regional geothermal gradient or in a nuclear waste repository after radioactive decay or, during repository galleries excavation (leading to a local modification of the regional geothermal gradient). We have carried out in-situ thermo-osmosis experiments on the Toarcian shale of Tournemire (SE of France) which presents analogies with shales likely to host a radioactive waste repository. In a packer limited test interval installed in a borehole, we have substituted the water in equilibrium with the formation by a warmer one (temperature increments of 2, 5, 10 and 20°C) in order to induce a potential thermo-osmotic response. Experiments are then interpreted with a thermo-hydro-mechanical model based on coupled flux equations combined with mass conservation laws. Once the hydraulic properties of the shale are determined, we simulate an aquathermal effect and changes in water properties (viscosity and density) with temperature. Next, we identify an eventual additional effect such as thermo-osmosis by fitting the model prediction for a range of thermo-osmotic coefficients with the measured data.