Fourier-based generation method of rough discrete fracture network

Abstract

Natural fractures have been often generated with rough morphologies as a result of complex geological processes. Joint surface undulation/roughness degree dominantly affects mechanical characteristics of jointed rock masses, hence should be considered in geometrical modeling of fracture networks. Considering the intrinsic geometric characteristic of natural fractures, this paper proposes a method to generate rough discrete fracture networks (RDFN) for jointed rock mass. Fourier transform is employed to obtain statistical spectrum information of natural rough fractures and fractures in the RDFN model are individually reconstructed in spatial frequency domain based on the information. Besides, three scale-independent shape descriptors, namely Elongation, Regularity and Smoothness, are selected to characterize the geometric characteristics of fractures in different scales and to verify the rationality of the proposed method. A case study is further employed to demonstrate the accuracy of the proposed method, and its applicability to numerical realization is discussed. The proposed method is important for a wide range of applications when considering the discrete modelling of rough fractures.