The mechanical behavior of rough rock joints can vary as a function of scale. Characterization of this scale effect can be problematic due to the difficulties in performing shear tests on larger rock joint specimens. This paper investigates the use of the 2D Combined Finite-Discrete Element Method (FEM/DEM) as tool to perform numerical direct shear tests. Direct shear models of a 400 mm long joint and 200, 100, and 50 mm subdivisions thereof are constructed and numerical direct shear testing is performed. The results show that the relative influence of scale on shear stiffness is captured. In contrast, a decrease in the average shear strength as a function of scale is not observed. The variation in the stress and dilation between specimens decreased as longer specimen lengths were considered and the upper bound shear strength showed a clear, albeit minor, scale effect.
It is widely recognized that the mechanical behavior of rough rock joints can vary as a function of scale, with shear strength decreasing as the nominal area of the sheared discontinuity plane increases (negative scaleeffect). The severity of this scale-dependency and the mechanism by which it occurs remains arguable. This behavior is partly attributed to the variation of intact asperity strength and partly to the variation in surface geometry (roughness) with increasing scale . While the decrease of intact rock strength with specimen size is well-established, contradictory findings regarding roughness scale-dependency are common . Hence, further work aimed at explaining the shearing process as a function of scale is necessary. The conventional approach to studying the scaledependency of discontinuity shear strength involves large-scale in situ direct shear testing. The frequency of such tests is limited by the associated complexity and cost and is limited to discontinuities of a few square meters.