The Random Finite Element Method (RFEM) is an increasingly popular tool in geotechnical engineering, especially for analysis of spatial variation and uncertainty in slope stability. Although the method has gained prominence in recent years, topological effects of strong and weak zones and the impact of their locations remain largely unknown. Although numerous potential slip surface realisations can be generated with RFEM, probabilistic failure statistics are often governed by several representative slip surfaces (RSS). In this research, random field similarity methods and clustering techniques are coupled with RFEM slope stability simulation to determine the impact of shear strength spatial patterns on slope failure mechanisms and safety factors. Regions of significance are highlighted within a case study of a Victorian open-cutbrown coal mine, with particular attention given to the effects on the slope failure surface as well the factor of safety. Results are presented of Factor of Safety distributions when particular slip surfaces and clustering constraints are imposed, providing further understanding of the impacts of shear strength characteristics on probabilistic simulation results.
This paper considers probabilistic slope stability analysis using the Random Finite Element Method (RFEM) combined with processes to determine the level of similarity between random fields. A procedure is introduced to predict the Factor of Safety (FoS) of individual Monte Carlo Method (MCM) random field instances prior to finite element simulation, based on random field similarity measures. Previous studies of probabilistic slope stability analysis have required numerous MCM instances to reach FoS convergence. However, the methods provided in this research drastically reduce computational processing time, allowing simulations previously considered too computationally expensive for MCM analysis to be simulated without obstacle. In addition to computational efficiency, the comparison based procedure is combined with cluster analysis methods to locate random field characteristics contributing to slope failure. Comparison measures are presented for slope geometries of an Australian open pit mine to consider the impacts of associated factors such as groundwater on random field similarity predictors, while highlighting the capacity of the similarity procedure for prediction, classification and computational efficiency.