Scale effect is a phenomenon of growing importance in rock mechanics, however large scale experiments are difficult to handle, and numerical modeling can be an easier way to study it. This paper presents some improvements made on the previously developed model for large shear displacement of rock joints, and presents some results obtained using of an originally developed finite displacement concept. The results indicate changes in the dilatation curves versus joint size.


Scale effect is a phenomenon observed in every geomaterials, very important in the case of concrete, considerable in the case of rocks, and less noticeable for sands. In rock mechanics, the notion of scale can be defined at three levels: the "micro scale" (asperities, millimeter order), the "micro scale" (laboratory sample, decimeter order) and the "macro scale" (natural scale more than one meter), thought for the differences In mechanical properties of joints at each mechanical scale. The mechanical properties of rock joints at the micro scale (asperities) are excluded from the domain studied. The mechanical parameters related to the "meso scale" are accessible to specific mechanical tests and morphologic measurements in laboratory (Hans et al. 2003, Esaki et al. 2000). However, it is much more difficult to evaluate properties at the "macro scale" because of handling difficulties. Until now, many studies have been performed.

The first practical study concerning scale effect phenomena on rock joints has been made by Pratt et al. (1974), by making direct shear tests on natural joints. As a result, a linear decrease in peak shear strength (semi logarithmic scale) as a function of the joint area (linear scale) was observed.

A more complete study of scale effect was carried out by Bandis et al. (1981). By increasing the joint size, the following phenomena were observed: a decreasing In the peak friction angle caused a decreasing in peak shear stress, and decreasing of the dilatancy angle. Some years later, another study about scale effect was carried out by Barton et al. (1982), they developed a model that is able to predict macro joints behavior, in Which the same order of results were found by Bandis et al. (1981). Pinto da Cunha et al. (1991) carried out a study on several dam foundations in which it appeared that the shear strength of joints decreased with each increase in size. Moreover, Yoshinaka et al. (1991) observed a reduction of the shear strength with increasing of specimen size. Castelli et al. (200 I) confirmed some results obtained by Bandis et al. (1981) as well as the lowering of shear strength with increasing joint size. Decreasing in the contact area with a scale increasing was also observed. Fardin et al. (2001) demonstrated scale dependency for mechanicogeometrical properties using fractal geometry.

Previous studies, and constitutive modeling have been carried out by the same authors (Vallier et al. 2004). However only the small relative displacement concept has been considered (i.e. each "meso" joint composing a macro joint evaluates without any reciprocity consideration).

This content is only available via PDF.
You can access this article if you purchase or spend a download.