Shear Behavior of Rock Joints Filled With Unsaturated Infill: Insights from Numerical Modeling and the Development of Strength Criterion

The mechanical behavior of discontinuities is significantly affected by the characteristics of the infill present along them. However, fewer studies have been performed to assess the response to discontinuities influenced by unsaturated infill. This study performs a parametric numerical study in Fast Lagrangian Analysis of Continua (FLAC-2D) to understand the shearing behavior of discontinuities influenced by unsaturated infills. The study was conducted in two steps. The initial step involves a detailed parametric study to analyze the impact of infill and discontinuity characteristics on the observed shear behavior. The study was performed by varying the thickness-to-asperity ratio (t/a) from 0.5 to 3.0, the Joint Roughness Coefficient (JRC) from 2.4 to 18.7, the degree of infill saturation (Sr) from 0.7 to 1.0, and the normal stress (σ) from 0.1 to 1.5 MPa. In the second step, a shear strength criterion was postulated based on the numerical data generated in the parametric study. The functional form of the criterion was kept similar to the Mohr–Coulomb criterion for its easy implementation in numerical and analytical analyses. The criterion was comprehensive because it could consider the implications of the above-mentioned parameters for the shear response of discontinuities. It was seen that the shear strength of infilled rock joints is enhanced with increasing JRC and σ and decreasing t/a and Sr. The numerical model was capable of capturing the critical t/a ratio beyond which the shear strength becomes a constant value. The expected mechanisms are discussed for the parametric results. The developed strength criterion was concluded to be comprehensive and accurate for shearing discontinuities under different conditions. The practical application of the developed strength criterion is also demonstrated for a rock slope undergoing failure, and comparative analyses were made with the results estimated via existing popular criteria.