The Representative Elementary Volume (REV) of fractured rock masses plays an important role on determining the rock mechanical characteristics. Firstly studying the general methodology of rock REV are summarized, and the modelling of Discrete Fracture Networks (DFN) is proposed as the investigation foundation of the REV of fractured rock masses, and then the numerical method to determine the REV is analysed and used in the case study. Combined modelling technique of random fractured network with discrete numerical method, the size effect of rock parameters has been analysed under single fluid flow condition and hydro-mechanical coupling condition. The calculation results of examples mentioned above indicate that the permeability tensor will decrease with the increasing stress and gradually tend to smooth while the studying scale increasing. The above achievements show that the proposed numerical method is an efficient approach to determine the rock mechanical parameters and REV, and also should be extended to complex three-dimensional problem in the geotechnical field in future.
The natural rock masses has experienced a long period of complex and unbalanced geological process, the rock masses has been cut by the fractures, and the randomness of the fractures distribution has induced the rock masses to present obvious discontinuous, heterogeneous and anisotropic characteristics (Harrison et al. 2000, Jing 2003). And the geological investigation has stated that fractures in disorder are obeying some probability distribution, and the geometric characteristics of above factures such as fracture width, dip, length and density determine the formation of the fractured network, so the mechanical property of rock masses implies some uncertainties. According to current research investigations, the value of rock mechanical parameters obtained by the field tests or laboratory tests appear obviously different due to different rock sample size, so the rock size effect is widely used to research on the rock strength, deformation, permeability and stress measure.