Dynamic probabilistic analysis of non-homogeneous slopes based on a simplified deterministic model

Abstract

This article proposes an efficient procedure for the probabilistic seismic analysis of slopes and applies it to two slope examples including a real engineering project. The proposed procedure evaluates the stability condition of a non-homogeneous slope by using a discretization kinematic approach (DKA), and accounts for the time and space variations of a seismic loading by employing pseudo-dynamic method (PDM). Three uncertainty quantification techniques (Subset Simulation, Global Sensitivity Analysis and First Order Reliability Method), are implemented into the procedure in order to provide multiple probabilistic results (e.g. failure probability, reliability index, design point and sensitivity index) for slope seismic reliability problems. The introduced DKA is validated by comparing with two previous studies. Additionally, a variety of hypothetical cases were analyzed in order to provide some insights into the following two issues about slope seismic analyses: 1) the comparison of two seismic methods (PDM and pseudo-static method) in a probabilistic framework; 2) the accuracy of FORM for estimating failure probability. By benefiting from the high computational efficiency of the introduced DKA, two further studies related to sensitivity analysis and fragility curves are also carried out. The obtained results show that the proposed procedure is effective for slope probabilistic seismic analyses and is able to efficiently provide a variety of useful results.