Deep geological disposal remains the preferred option at present for the management of long-living and heat-emitting radioactive waste, which consists of confining the waste during a very long period (several hundreds of thousands of years) by placing them in a deep geological formation. Therefore, the understanding of the long-term behaviour of formations is becoming a key issue to ensure the feasibility of the geological disposal facilities, particularly regarding the generation and migration of gases. The present PhD work aims at better understanding the complex hydro-mechanical response of different argillaceous formations to gas migration process. To this end, gas flow through Boom Clay (one of the potential candidate plastic Paleogene clay formations to host nuclear waste in Belgium) has been deeply investigated on the basis of laboratory experiments at different scales and their numerical modelling. This main study has been complemented by presenting tests on two indurated and deeper claystone Mesozoic formations, considered as candidate host rocks in the Swiss programme for deep geological disposal, namely Opalinus Clay and "Brauner Dogger".

The different materials have been firstly characterised to evaluate mechanical (compressibility on loading) and two phase flow properties (water retention and permeability). Gas injection tests under oedometer and isotropic conditions have been performed following different testing protocols, in which boundary conditions have been were carefully controlled. Major relevance has been given to restore the in situ stress state and to ensure full saturation conditions before the gas tests. Special emphasis has been placed in measuring sample deformation along the gas injection and dissipation process. The anisotropy of Boom Clay has been studied by carrying out tests with bedding planes parallel and normal to flow. Air injections have been performed at three different controlled-volume rates. The dissipation stages after shut-off have been also analysed to study air intrinsic permeability changes. Microstructure of samples before and after air injection tests have been evaluated by different techniques: mercury intrusion porosimetry, field-emission scanning electron microscopy and micro-focus X-ray computed tomography. Gas migration turned out to be a fully coupled hydro-mechanical process. Air injection at constant stress induced expansion of the samples during pressure front propagation and compression during air pressure dissipation.

The deformational behaviour was dependent on the injection rate. At slower injection rates expansion occurred during the injection while at higher rates it was delayed in time. Air intrinsic permeability resulted higher than water permeability suggesting that air flow took place along preferential pathways. Evaluation of the microstructural changes induced by air migration revealed the opening of fissures and allowed quantifying their apertures and separation, as well as their volume and connectivity. Air intrinsic permeability was found to be dependent on the fissured volume. To complete and better understand the gas transport mechanisms, numerical simulations of the experimental results have been performed using a fully coupled hydro-mechanical finite element code, which incorporates an embedded fracture permeability model to account for the correct simulation of the gas flow along preferential pathways. Clay intrinsic permeability and its retention curve have been made depend on strains through fracture aperture changes. Numerical results not only accounted for the correct simulation of the recorded upstream pressures and outflow volumes and pressures, but also on the volume change behaviour. The experimental and numerical information provided a good insight into the mechanisms of gas transport in deep clay formations and highlighted the role played by the deformational response on the air transport properties of argillaceous rock formations.