A probabilistic framework for the stability analysis of slopes considering the coupling influence of random ground motion and conditional random field

The uncertainty of the input parameters can have a significant influence on the evaluation of the slope stability. A simple probabilistic framework is presented in this paper to quantify the influence of the uncertainties of the stochastic ground motion coupled with the spatial variability of soil parameters. The robust quasi-Monte Carlo simulation in the proposed framework is employed to estimate the failure probability of slopes. The conditional random field (CRF) is manipulated to simulate the spatial variability of soil properties with the sample points where the soil parameters are determined, and the spectral representation of stochastic process is applied to generate the horizontal and vertical stochastic ground motions. In each run of the Quasi-Monte Carlo simulation, a modified Newmark method based on an infinite slope failure model is applied to determine the permanent displacement of slopes under horizontal and vertical ground motions. The originality and significance of this proposed probabilistic framework are validated by a real slope case. The results of the case show that the un-certainty affecting the stability of the slope can decrease when CRF is used in a random field simulation to incorporate the real values of soil parameters at sample points, and the failure probability of slopes can be underestimated without considering stochastic vertical ground motion. Moreover, the correlated distance and critical permanent displacement are significant in predicting the failure probability of slopes.