Themes: Geological Sciences, Earth, Universe
location: Unité d'expertise et de modélisation des installations de stockage (UEMIS) - Fontenay-aux-Roses (92)
Master's Degree in Fluid Dynamics Research, Energetics and Transfer - skills in transfer in porous media, scientific computation and computer programming of codes.
Age limit: 26 years old unless otherwise stated.
The aim of this thesis is to study uncertainties resulting from models of gas migration at the scale of a disposal facility of HLW and ILW-LL during the period of its closure for many thousands of years. These uncertainties can be of different kinds (phenomenological, parametrical, or numerical) and their consideration need a good understanding of physical phenomena and their description by the adequate models through analysis of experimental data, and it is convenient to study their impact on hydrogen overpressure produced in waste cells, its transfer and its influence on the host rock and engineered barriers properties of the disposal. In order to evaluate these different uncertainties, the IRSN proposed to carry out a study around the following steps: i. Study of the combined effects of the new phenomena Hydo-Gas (gas entry pressure, hysteresis) implemented in the massively parallel code TOUGH2-MP (LBNL, Berkeley, USA) on hydrogen migration at the scale of the repository. This study will be applied to the generic test case proposed in European EURAD – GAZ project. During this study, technological voids (centimetric interfaces) are modelled as porous materials, and can be explicitly be integrated in the mesh or homogenized with the adjacent materials. This step is necessary in order to compare our results to those from HM-Gas models developed by different teams involved in that project and to validate the different mesh scripts developed in our unit (UEMIS). ii. Coupling a linear elastic deformation model to the model H-Gas in order to study geo-mechanics of interfaces. This coupling needs an important code development by adding new stress-strain laws in the code TOUGH2-MP. A numerical verification for this development will be carried out through analytical solutions or benchmarking with the existing codes; iii. Integration of the coupled model in step (ii) in the study carried out in step (i); iv. Study of the uncertainty on the response of the models described above and their sensitivity to one or more uncertain parameters, assumed as random variables, such as gas entry pressure, intrinsic permeability or hydrogen source term. This study will be carried out by different statistical distributions of uncertain parameters and by different sampling technics based on direct methods (e.g., First-Order-Second-Moment; Monte-Carlo – HLS ou EOF, Importance Sampling) and indirect methods (meta-models construction, quasi-analytical methods), through their implementation in the code TOUGH2-MP. This study will be conducted in collaboration with IMFT institute (Toulouse University) which possess an important background on the development of these methods.