The Random Finite Element Method (RFEM) is a powerful technique for incorporating spatially variable shear strength parameters with slope stability numerical simulations. In this research, two-dimensional probabilistic analyses of a large open-cut brown coal mine are presented with particular consideration given to slope Factors of Safety (FoS), when faced with highly anisotropic cohesion and friction angle shear strength parameters. Bayesian methods are implemented to determine updated shear strength parameters based on Factors of Safety and Representative Slip Surfaces (RSS) categorizations. By this method, the impact of observed slip surface depths and safety factors is further investigated. Monte Carlo simulation is implemented in the Finite Element environment Abaqus, with an optimised Strength Reduction Method to determine Factors of Safety. Comparisons of conditional shear strength distributions are made for associated slope safety factors and shallow slip surfaces from a cross-section of the Yallourn open-cut brown coal mine, in Victoria, Australia. The updated shear strength distributions provide a greater understanding of the necessary conditions of particular slope failure mechanisms, contributing further understanding of the stability of Victorian brown coal mines.
Probabilistic methods of analysis have gained attention in recent years and are amenable to a wide range of geotechnical applications (Griffiths and Fenton 2001, Chok, Jaksa et al. 2015). When considering the stability of slopes and embankments with high levels of soil heterogeneity and anisotropy, probabilistic methods are often necessary to identify appropriate slope failure mechanisms (Griffiths and Fenton 2004, Griffiths and Marquez 2007). While homogeneous slopes often exhibit well-defined behaviour governing the failure mode and slope Factor of Safety (FoS), probabilistic simulations containing heterogeneous parameters may exhibit a wide range of failure mechanisms and FoS distributions. Matters are further complicated when soils exhibit strong spatial variability. Neglecting the spatial correlation of soil properties can significantly exaggerate the factor of safety and the probability of slope failure (Mostyn and Soo 1992).
The Random Finite Element Method (RFEM) combines random field generation with the Finite Element Method (FEM) to model spatially varying parameters with Monte Carlo simulation. The method allows slope safety factors to be determined without a predefined critical failure surface, as required for Limit Equilibrium Methods. Instead, RFEM allows for the simulation of the progressive failure of a slope until the final slope failure surface is achieved.