Estimating fractured rock effective permeability using discrete fracture networks constrained by electrical resistivity data

Abstract

Although the permeability of fractured rock mass is a fundamentally important property for the safe construction of civil and mining engineering structures such as tunnels, in situ characterization of permeability without resorting to hydraulic tests is difficult. For rapid, wide-area estimation, a method that can be conducted at a field-scale using geological and geophysical investigation data is proposed. The method is not based on conventional hydraulic test results. Instead, it combines the stochastic generation of fracture networks with the crack tensor theory. The most important parameter for this method is the fracture length distribution. Although the distribution parameters in the DFN model are assigned through sampling, a bias is generally experienced because of the limited sampling area. To improve the estimation of such parameters, in-situ electrical resistivity data and a symmetric self-consistent method are used to constrain the fracture length distribution. The proposed method is applied to the fractured crystalline rock mass of the Mizunami Underground Research Laboratory (URL) in the Tono area of central Japan. Its effectiveness and correctness are demonstrated through good correspondence of the derived effective permeability with the in-situ measured permeability.