Assessment of Slope Stability through RFEM under Probabilistic Framework with Material and Loading Uncertainties
摘要
Conventional deterministic techniques, which compute a single Factor of Safety (FS), are simple but fail to consider uncertainties in soil parameters. Probabilistic methods like First-order approximations address this but are inefficient at modeling spatial correlations. The Random Finite Element Method (RFEM) merges probabilistic analysis with finite element modeling to overcome these issues; however, it typically does not provide a direct FS, reducing its practicality for engineers. This research introduces a modified RFEM that incorporates explicit FS calculations, applying it to a case study of a railway embankment. The method employs Monte Carlo simulations to perform a stochastic analysis, treating cohesion, internal friction angle, unit weight, and seismic coefficient as random variables and accounting for their spatial and cross-correlations. It visualizes failures through deformed meshes and strain contours and quantifies stability via Complementary Cumulative Distribution Function (CCDF) plots of the reciprocal of FS. This enhanced framework integrates crucial probabilistic insights into the familiar FS, making modified RFEM more directly applicable to slope design.