Determining the permeability of fractured rocks is very important in different field of engineering works, such as oil reservoirs, dam constructions and underground excavations. Fractured rock mass permeability is calculated by analytical and numerical methods. Analytical solutions basically are applicable for simple or regular joint set systems. On the other hand numerical simulations demand a large computing resources and its very time consuming particularly when we have a large model size with densely populated of fractures. According to the literature; investigation of fractured rock mass permeabilities using Dual Fracture Method (DFM) approach looks an efficient method for the cases which rock matrix permeability cannot be neglected. In such rocks, the fracture system has low porosity and storage capability but dominates the traort of fluids and the matrix, has high porosity and dominates fluid storage. In fact rock mass permeability in this method is a combination of open fracture permeability and rock matrix permeability. However, evaluating of rock mass permeability for these cases in previous works was limited to regular joint sets or simple fracture geometry such as en echelon fracture pattern, and they did not consider the effect of fracture density, or the effect of matrix permeability and fracture aperture pattern on Representative Elementary Volume (REV), and therefore the permeability in such studies was not reliable. This was the motivation of this study. Using Universal Distinct Element Code (UDEC), rock matrix was generated with some kind of tessellation elements called Voronoi elements (VE). The aperture between these elements was calibrated with laboratory values of rock mass core permeability of a fractured granular rock in the field, and the rock matrix permeability could be changed with changes in aperture between Voronoi elements. Then stochastic DFN models were generated with different fracture densities and aperture patterns to fit on the matrix model. Permeability and REV size were observed in both DFN and combined model too. Keywords: Dual Porosity, Voronoi elements, Dual Fracture Method, Representative Elementary Volume
