Shear Strength Criteria for Clean Rock Joints: Representation and Comparison of Different Failure Criteria

Shear strength along clean rock joints is fundamental to the design and safety of slopes, underground excavations, and dam foundations. However, predictions can vary considerably because available shear-strength criteria rely on different assumptions regarding roughness, dilation, scale effects, boundary conditions, and parameter determination. This review first synthesizes the main factors controlling peak shear strength, including joint characteristics, stress state, loading and boundary conditions, and hydrological effects. It then examines the principal parameters used to characterize joint roughness and the main methods available for their measurement and estimation. Building on this foundation, the reviewed shear-strength criteria are organized into seven conceptual families, including linear frictional models, 2D asperity- and dilation-based models, empirical JRC–JCS criteria, constitutive joint models, 3D morphology-based criteria, transitional/degradation-based models, and machine-learning/data-driven approaches. The review is complemented by an applied case study based on direct-shear testing from the Romaine-IV hydroelectric project. The dataset includes 20 specimens tested under multiple normal-stress stages and represents anorthositic and gabbroic rocks with varying degrees of weathering and alteration. The Mohr–Coulomb, Patton, and Barton–Bandis criteria are compared with the measured shear-strength envelopes to examine their behaviour under a common project setting. Particular attention is given to differences in parameter requirements and to the influence of JCS on the Barton–Bandis predictions. The sensitivity analysis highlights the importance of selecting JCS values that are consistent with the actual condition of the joint wall, especially where weathering or mineralogical alteration is present. The case study is intended as a site-specific engineering comparison rather than a general ranking of the three criteria. More broadly, the review shows that the practical applicability of a shear-strength criterion depends not only on its theoretical formulation, but also on the availability, quality, and consistency of the parameters required for its implementation. The proposed classification framework and applied comparison provide practical guidance for criterion selection while identifying areas where improved parameter characterization, standardized testing, and broader multi-site validation are still needed.