A thermo-hydro-mechanical model for fracture propagation and evolution, fluid flow, and transport in the disturbed rock zone of an argillaceous repository

Document available in extended abstract form only. A key reason for considering argillaceous rock for nuclear waste disposal is the extremely low permeability and limited fracturing generally found in this rock type. Underground research laboratories in Switzerland, Belgium, and France have been developed to investigate the potential of argillaceous rock for nuclear waste disposal. However, excavations in argillaceous rock have shown the general tendency for fracturing to occur in the immediate vicinity of the excavations, called the disturbed rock zone (DRZ). Fractures are generated as a result of changes in mechanical stress caused by the excavation and changes in saturation caused by ventilation. The attributes of fractures in the DRZ evolve in response to post-closure changes in temperature, water redistribution, and materials interactions associated with a nuclear waste repository, and there is a tendency for the DRZ fractures to close over time, called self-sealing. Here we present a new computational method applicable to three-dimensional discrete fracture networks in the DRZ of an argillaceous rock. The method is intended to predict, in a mechanistic manner, the fracture formation and evolution as well as flow and transport through the fractured porous rock in the DRZ accounting for dynamically changing thermal-hydrologic-mechanical conditions. The new computation method simulates the non-isothermal multi-phase flow and transport processes with TOUGH2, a widely used subsurface simulator developed at Lawrence Berkeley National Laboratory. We use EOS7R, one of the TOUGH2 modules for thermal-hydrologic flow processes involving brine, water, and air phases as well as transport processes involving two-component radionuclide chains (Oldenburg and Pruess, 1995). A mechanical lattice model based on the rigid body-spring concept of Kawai (1977) is coupled to the TOUGH2 model to compute elastic response and fracture propagation. These two models share the same unstructured, three-dimensional Voronoi grid and the same set of nodes, where the scalar field quantities (e.g. temperature, pressure, and saturation) and the generalized displacement are obtained by TOUGH2 and the lattice model, respectively. As shown in Figure 1, each node pair ij that is mechanically connected is composed of a zero-size spring set located at the centroid of the corresponding Voronoi facet. Fractures propagate along Voronoi cell boundaries as thermal-hydrologic-mechanical induced stresses evolve and exceed the materials strength. Local fracture permeability and porosity are assigned based on grid geometry and apertures computed by the lattice model from mechanical strain induced at fracture locations. To demonstrate the utility of the coupled approach, fracture development in a clay sample subjected to drying environment is simulated. Figure 2 shows numerical results of crack pattern demonstrated by the 3-D coupled lattice model under drying shrinkage condition with restrained boundaries except one surface. Two stages of crack opening can be observed: 1) vertical cracks due to shrinkage strain increases, and 2) lateral cracks caused by curling of slab structure. The present modeling results include complicating factors such as time-dependent crack development, and the coupling of internal flow and mass transport with crack opening. Future developments will include coupling the lattice model with TOUGHREACT (Xu et al., 2004) to incorporate chemical processes, in particular, chemical-mechanical effects (e.g., swelling caused by changes in brine composition) that are characteristic of argillaceous rock types. (authors)