Monte Carlo Simulation (MCS) is often applied for probabilistic slope stability analysis. However, in many situations, the computational efforts required may become prohibitive, especially when complex soil behavior necessitates non-linear finite element modelling for each simulation and/or if the failure probability is very low. This paper presents a probabilistic approach where the physical behavior is represented by a novel response surface (RS), and first-order reliability method (FORM) and MCS have been applied to calculate the failure probability. The method is demonstrated through a case study to estimate the probability of slope failure using the commercially available finite element software PLAXIS. The results of the reliability analyses with response surface were compared with direct MCS results using the full PLAXIS model. The comparison indicates that the combined RS-MCS and RS-FORM provide similar results as the direct MCS and the former have the advantage of being computationally efficient.
Aleksander S. Gundersen, D. Ryghseter, A. Worren, F. Nadim and Z. Liu
Keywords: slope stability, response surface, First-Order Reliability Method, Monte Carlo Simulation, Finite Element Method
Publisher: DFI