Relations between the hydraulic aperture, shear dilation, closing and opening joint are proposed and considered for coupling between conductivity, shear behaviour and cyclic loading and unloading normal behaviour of a joint. General expressions for the variations of mechanical aperture and hydraulic conductivity with depth are presented. The determination of the overall permeability of jointed rock masses and its anisotropy are discussed and geothermal energy production.
The permeability of rock masses controls not only discharge into underground structures but also the safety of engineering structures. It is an important aspect of radioactive waste disposal and geothermal engergy production.
In situ tests such as injection tests have been employed to characterise the hydraulic apertures of joints in order to obtain the permeability tensor (Snow, 1966; Louis and Maini, 1970) and to estimate the Representative Element Volume (REV). Because of the heterogeneity of discontinuities, the REV may be so large that it is technically impossible or uneconomical to conduct large scale in situ tests at the REV scale (Boodt & Brown, 1985). Small scale in situ tests have to maintain the assumed laminar flow and cannot account for joint interaction and connectivity as adequately as the large scale tests. Furthermore, all injection tests characterise the permeability of rock masses by the water quantity absorbed in a unit time by known joints intersecting the boreholes or by a unit length of borehole. Also radial flow through joints is lower than longitudinal flow (Witherspoon et al., 1980), which indicates the complicated effect of roughness in radial flow. Thus, there is doubt about the utility of injection test results: If possible, longitudinal flow in situ tests should be carried out (Barton & Lingle, 1982; Neretnieks et al., 1987).
The basic cause of the problems described above is that the heterogeneity of a network. of discontinuities requires a statistical approach. The discontinuity frequency theory based on persistent joints (Hudson & Priest, 1983) can be valid for impersistent joints (priest & Samaniego, 1983). Also, the equivalent permeability tensor is applicable on an acceptable level of accuracy and precision for predicting flow (Samaniego, 19~4). The different accuracy levels correspond to the different REV sizes. For a volume below the REV, the. connectivity of discontinuities dominates. The near field mechanical and hydraulic problems are better treated by discontinuum methods than by equivalent continuum methods (Robison, 1982; Cundall, 1983; Long, 1983). Even for a volume above the REV, 'super large conduits' such as faults have to be considered separately.
Because both discrete and equivalent material models require a knowledge of the joint aperture and because the aperture depends on stress and shear displacement, it is intended to study (1) the mechanical and hydraulic behaviour of a joint and their coupling. (2) the determination of the permeability of a jointed rock. mass, (3) the directional variation of permeability. (4) the variation of permeability with depth. 2 MECHANICAL AND HYDRAUUC BERA VIOUR OFAJOINT The flow through a joint has traditionally been regarded as laminar flow.