Experimental and Numerical Studies on Flow Characterstics in Rock Fracture Intersections Considering Influence of Intersecting Angle

"In this study, laboratory flow tests were conducted on 2-D artificial rock fracture network models containing a single intersection constituted by two inlet and two outlet branches. Three models with different intersecting angles of 60°, 90° and 120° were tested using both dye solution and distilled water. Constant flow rates were imposed on one inlet, and the flow patterns in the fracture intersections were captured by using a CCD camera. The test results were compared with the predictions made by numerical simulations that solved the Navier-Stokes equations. The results revealed that when Reynolds number (Re) is as low as less than 1, the intersecting angle has negligible influence on the flow rate of each branch and the flow patterns within a fracture intersection. As the interesting angle ? and Re (e.g., from 10 to 200) change, the normalized flow rates of different outlet branches change significantly, due to the inertial effect of fluid flow. These changes could be represented by a series of mathematical expressions involving Re and ?, which could be applied to fracture network models to improve their prediction accuracy on hydraulic properties (e.g., permeability and breakthrough curves of particles) of rock masses. INTRODUCTIONCoupled stress, flow, heat and mass transport are the corner stone subjects for many geoscience and geo-engineering disciplines, such as geological disposal of radioactive waste and geothermal energy exploitation. Many outstanding issues need to be investigated to enhance the knowledge on these subjects, and specifically more in-depth research in testing and modeling of the fluid flow in fractured rock masses is needed. In recent years, discrete fracture network (DFN) models have been extensively used to study the fluid flow and mass transport behaviors of fractured rock masses (e.g., Min et al., 2004; Zhao et al., 2011). In these models, each fracture is usually represented by a parallel-plate model, in which fluid flow exactly follows the cubic law. Therefore, comparative flow rate in each outlet branch of an intersection in these models only depends on its aperture. In reality, due to the surface roughness of fractures, the magnitude of mechanical aperture is usually larger than that of the hydraulic aperture of each fracture (Li et al., 2008; Xiong et al., 2011), and abundant flow patterns could exist in an intersection, influenced by its geometry, varying with changing Reynolds number (Re) (Kosakoswski & Berkowitz, 1999). Recent studies have also revealed that different mixing patterns might exist at a fracture intersection, bounded between streamline routing model and complete mixing model, that determine the final distribution of solutes in fractured rock masses (Johnson and Brown, 2001; Park et al., 2001). Numerical simulation has been the principal method adopted in previous studies, and the studies conducting flow tests on intersected fractures are less reported (Wilson & Witherspoon, 1976; Johnson and Brown, 2001). In this study, laboratory flow tests and numerical simulations by solving the Navier-Stokes equations were conducted on 2-D models containing a single intersection constituted by two inlet and two outlet branches, intersected at different angles. The main objective is to study the influence of intersecting angle on the flow behavior of the intersected fractures and the variations of flow patterns and fluid redistribution in each model along with the changes of the Re and the intersecting angle ?. The influence of surface roughness was not considered."