A hybrid approach for deriving probability distributions of rock mass discontinuity parameters for slope reliability analysis

Abstract

It is widely recognized that the rock mass discontinuity properties (e.g., dip angle and dip direction, trace length and aperture, cohesion and friction) can greatly affect the stability of rock slope. The discontinuity properties are often considered to be cross-correlated random variables and are modeled by standard probability distribution families. Reliability analyses of rock slopes accounting for uncertainties in the discontinuity properties and construction of discrete fracture network do not make best use of available field survey data of rock masses. In this paper, a hybrid approach, proposed by Ching and Phoon (2019), is employed for constructing the multivariate probability distribution of rock mass discontinuity properties based on field survey data. Then the discrete fracture network models and probabilities of failure of the jointed rock slopes are obtained accordingly. Copula theory is adopted to model the prior and site-specific joint probability distributions of the discontinuity parameters. A rock slope in China is taken as an example to demonstrate the effectiveness of the proposed method. The research outcomes can provide a new perspective to understand the distribution of rock mass discontinuities and slope failure mechanism.