ABSTRACT: In this paper, we conduct a preliminary study to assess permeability change due to construction of repository tunnel for a high level radioactive waste. This is accomplished by combining several existing numerical models developed for different purposesuch as the Micromechanics-Based Continuum (MBC) model, a joint element model, and Reynolds equation. Joint permeability is defined in a slightly different way from general porous media and is selected to evaluate the variation of the permeability. The joint permeability significantly increases in the area around the excavation tunnel, which is usually named the excavationdisturbed-zone (EDZ). The proposed method is considered to be employed in evaluating the effect of excavation in the safety assessment of the repository.
INSTRUCTION
Joints can significantly influence the hydraulic characteristics and control the mechanical behavior of rock masses. Therefore, the mechanical and hydraulic properties of joints are of considerable interest in designing projects related to jointed rock mass. Although the mechanical and hydraulic properties of rock masses have been studied for a long time with significant results, understanding of the hydraulic properties of deformed rock masses is limited.
The hydraulic properties of a deformed rock mass should be carefully considered particularly in the design of underground repositories for a high level radioactive waste, since nuclear particles can be dissolved in the groundwater and widely transported. The intact rock in the jointed rock masses has fairly low conductivity; typically in the range of 10 ']5 to 10 ']ø m/s. For this reason, fluid flow in jointed rock masses occurs mostly through the embedded joints (Hudson, 1993).
Fluid flow through rock joints is usually expressed in terms of cubic law. The cubic law is theoretically deduced from assumptions that fluid flows through apertures bounded by parallel plates under saturated, laminar, incompressible flow conditions. The parallel plates model, however, can be considered only a qualitative description of flow in joints, since they are not smooth parallel plates, but are in fact rough and contact each other at discrete points over the joint. Since lwai (1976) evaluated the validity of cubic law at rough-walled joints under low normal stress, many researchers (Witherspoon et al., 1980; Tsang and Witherspoon, 1981; Tsang, 1984) have presented the deviation from the cubic law. They have been developing more realistic models for the description of the flow in the rock joint. Brown (1987) and Zimmerman (1991) have shown that the simpler Reynolds lubrication equation, instead of the full Navier-Stokes', can be used to describe the flow in joints.
The effect of the aperture distribution within a single rock joint on its flow properties has been emphasized repeatedly. Tsang et al. (1990) divided this work into three groups according to the method of detailed measurement of the aperture distribution: surface profiling (Bandis et al., 1981; Brown and Scholz, 1984), injection method (Pyrak-Nolte et al., 1987; Gale, 1987), and resin casting technique (Hakami, 1990). These measurements have provided useful and important data for the studies of flow through joints with variable aperture distribution. However, the problem lies in the fact that the geometry of apertures will change with any load applied and it is very difficult to obtain experimentally the joint aperture distribution under various stress states.