Abstract:

To quantitatively characterize the scale effect of hydraulic properties in fractured rock masses and determine the representative elementary volume (REV) under anisotropic conditions, this study proposes a novel method for estimating the REV size based on permeability analysis. This study built a three-dimensional discrete fracture network (DFN) model based on joint data obtained from a dam site in southwestern China. From this DFN model, 504 two-dimensional sub-models were extracted, with sizes ranging from 1 m×1 m to 42 m×42 m. A series of steady-state seepage numerical simulations were conducted on these models using the finite element method to obtain the equivalent permeability. This study proposes a new method for determining the representative elementary volume (REV) size of fractured rock masses based on permeability. This method provides a reliable estimate of the REV size by analyzing the tensor characteristic of the directional permeability, as well as its constant characteristic beyond the REV size. The results show significant hydraulic property differences in different directions, with the equivalent average permeability ranging from 4.71×10-16 to 2.55×10-15 m2. The hydraulic REV sizes range from 6 to 36 m, and the REV size is negatively correlated with the permeability of the rock mass, significantly reflecting the anisotropy of the fractured rock mass. The maximum REV size occurs in the direction of the average angle of each fracture set, while the minimum size occurs along the bisector of the angle between two adjacent fracture sets. REV size is also negatively correlated with the average trace length of intersecting fracture sets. These findings suggest that the angle distribution of the fracture group reasonably explains the distribution of the direction REV extreme value, and the proposed method for estimating the directional REV size is effective.

Key words: discrete fracture network (DFN), anisotropy, scale dependency, representative elementary volume (REV)