Stochastic Analysis of Pore Pressure Uncertainty for the Probabilistic Assessment of the Safety of Earth Slopes

Abstract

A probabilistic slope stability model was developed that includes soil shear strength and pore pressure as random variables. A local estimation technique called the nearest-neighbor method was utilized for simulation of the spatial variability of saturated soil permeability values. The spatial variation in the soil permeability causes a variation in the position of the phreatic surfaces during steady-state seepage flow and, hence, pore pressures. A number of permeability sets were generated and the resulting series of phreatic surfaces were calculated. A series of slope stability computations was made combining the spatial variability of pore pressures and soil strength yielding a probability density function of the factor of safety. The probability of failure was computed from the mean and standard deviation of the factor of safety. The resulting probabilistic model was applied to a hypothetical embankment. In addition, a sensitivity study was performed on the influence of the variability of soil permeability. The results indicated that the variability of soil strength and soil permeability has a large influence on the outcome of the probabilistic analysis and that there is a large discrepancy in the magnitude of the probability of failure if the pore’ pressure uncertainty is excluded from the probabilistic slope stability analysis.