Probabilistic Analysis of Rock Slope Stability Considering the Spatial Variability of Rock Strength Parameters

Abstract

Naturally characterized by inherent heterogeneity, rock masses exhibit discontinuous composition comprising rock materials, fractures, and bedding planes. The presence of these natural features challenges the adequacy of conventional analytical methods for stability assessment of practical applications. This paper proposes an alternative approach utilizing adaptive finite-element limit analysis and an efficient strength reduction method for deterministic and probabilistic analyses of rock slopes. The calculations employ the generalized Hoek–Brown and equivalent Mohr–Coulomb criteria to determine the factor of safety. To assess the impact of spatial variability, various scenarios, such as noncorrelation and correlation of rock mass parameters, are considered in estimating the failure probability for sensitivity analysis. The results indicate that the generalized Hoek–Brown and the equivalent Mohr–Coulomb criteria reveal only slight differences in the factors of safety during deterministic analysis, while they significantly influence the failure probability. Particularly, the more rock mass parameters are simulated as the spatial variability, the greater failure probabilities are produced. Conversely, the strong correlations among these parameters are modeled, resulting in a reduction in the failure probability. These findings imply that practitioners can use more flexible methods to design protection or predict the failure of rock slopes while minimizing costs.