Influence of Roughness on the Apparent Cohesion of Rock Joints at Low Normal Stresses

The shear strength of joints is governed by surface morphology, including roughness and interlocking conditions. Several criteria have been developed in recent decades to assess shear behavior. However, most civil engineering guidelines suggest using the Mohr-Coulomb criterion to evaluate shear strength. The Mohr-Coulomb criterion relies on a friction angle and a cohesion value determined from laboratory tests, the literature, and the engineer's experience. Two distinct cohesion concepts can be considered: true cohesion, which is a bond between the walls of the joint, and apparent cohesion. Apparent cohesion is obtained by performing a linear regression for the experimental shear strengths obtained in the laboratory for a given normal load interval. For unbounded joints, true cohesion cannot be considered because the walls are not physically bonded. However, apparent cohesion, describing joint roughness, interlocking, and the conditions of the joint walls, can be used with the Mohr-Coulomb criterion. Civil engineering standards are cautious about the use of apparent cohesion during design, and guidelines often recommend considering a null or low value. In this article, we focus on the effects of surface morphology on apparent cohesion, to better appraise the relevance of taking apparent cohesion into account in the Mohr-Coulomb criterion. Direct shear tests were performed on four different rock joint replicas tested under low normal load conditions (0.1–1.0 MPa). The results showed that apparent cohesion increased with roughness. Results also showed that the interlocking and damage state of the joint wall surfaces play a key role in the apparent cohesion value. When the sample is damaged or poorly interlocked, apparent cohesion is significantly reduced.