| dc.description.abstract | Fluid flow through rock masses is a critical issue in rock engineering and geoscience in general, from shallow to great depths. The hydraulic aperture (e) of a single joint under stress is a key variable for simulating the fluid flow through stressed rock masses. Because prediction models for the mechanical aperture (E) of different rock joints under different normal stresses are available, the hydraulic aperture at different depths can be readily estimated using numerous E＂-＂ e relations. However, experimental hydromechanical studies frequently use the normal displacement method (ND method) for determining the mechanical aperture, for which the initial aperture needs to be “estimated”. This uncertainty can propagate to the E＂-＂ e relation. In this study, we conducted high-quality laboratory measurements of mechanical/hydraulic apertures of saw-cut, polished aluminum samples, saw-cut and foliated slate samples which are comparable to smooth rock joints, using the YOKO 2 system under confining stresses from 3 to 120 MPa. Moreover, a novel pulse-decay-balance (PDB) method was proposed to simultaneously measure e and E under stress. The YOKO2 system was modified into a closed system to measure the e and E of saw-cut and foliated slate samples under a confining (normal) stress of up to 60 MPa. The hydraulic aperture was determined based on analysis of the pressure decay curve, and the mechanical aperture was determined based on the joint volume measurement (JV method) by determining the balance pressure.
This study is divided into different parts: (1) e and E measurements of saw-cut smooth aluminum samples with an identical joint surface using steady-state and joint volume method, stress-dependent e and E fit well using a hyperbolic function; (2) A proposed novel PDB method is to measure simultaneously e and E of saw-cut and foliated slate samples, stress-dependent e and E fit well using a hyperbolic function; (3) Influences of measurement and analysis methods on rock apertures determination. We measured the mechanical aperture directly from the joint volume (JV method); this method differs from the widely adopted ND method for measuring joint closure. The experimental results show that the estimation uncertainty of initial aperture, which is required for converting the joint closure curves into mechanical aperture, reduces the reliability of the existing E＂-＂ e relations. (4) When the JV method is used to reduce the influence of initial aperture estimation, it is important to note that the contribution of the matrix pore volume of intact rock to the mechanical aperture measurement is not negligible even when the porosity of the matrix is less than 0.1%. Contrary, the matrix permeability of impermeable rock is negligible for laboratory measurements; (5) Neglecting the quadratic Forchheimer term leads to an underestimation of hydraulic aperture; (6) The hydraulic aperture estimated using the proposed method was comparable to that measured in steady-state tests; (7) The hydraulic apertures of the saw-cut and foliated slate samples estimated using the simplified and general analytical solutions were in good agreement, especially for smooth joints under high normal stress. However, the error in the hydraulic aperture estimation, induced by neglecting the specific storage of the rock joint, is non-negligible for rough rock joints under low normal stress; (8) The measurable hydraulic aperture e via the close system—YOKO2 is several to several tens microns. Although the measurable mechanical aperture E is below 1 micron, the best measurement range of E is roughly 1,000 to 10,000 microns. To harmonize the optimum measurement range of hydraulic and mechanical apertures, using a typical sample size for a conventional triaxial test is suggested; (9) The effect of anisotropic behavior and open foliated joint on the matrix porosity and permeability of intact slate samples is significant; (10) A regression function based on the stress-dependent E and e fitted well with the experimental data to describe the E-e relationship. Our testing results, which show a decreasing trend in the E/e ratio with decreasing e, contradict the existing (E/e)-e relation, at least for smooth joints. Using ∆E＂-＂ e relations for hydromechanical coupling analysis is suggested to avoid errors due to the initial aperture estimation based on the ND method. | en_US |