Experimental and numerical investigation of three-dimensional stability of slopes

Slope stability problems are among the most commonly addressed problems in geotechnical engineering. Since Fellenius (1927) suggested the method of slices, several slice methods have been proposed attempting to permit a quantitative assessment of the stability of slopes. These conventional methods of slope stability analysis are mainly based on the concept of limiting equilibrium. The limiting equilibrium concepts have been expanded to accommodate three-dimensional features of slopes and several 3-D slope stability methods have been proposed in the literature, ranging from a rough approximation method (weight average method suggested by Sherard, 1963) to a rigorous method (generalized 3-D stability method using columns by Fredlund and Lain, 1992).; The main problem in using 3-D methods is that although these methods satisfy 3-D boundary conditions, they generally predict excessively high safety factors, which are far different from the real field situation. In this study, the reason behind a high unreasonable value of factor of safety from the 3-D calculations was explained in terms of strength anisotropy and intermediate stress parameter. It was also shown in this study that the safety factor using isotropic strength can be an unreasonable prediction when the effect of anisotropy on shear strength of soil becomes significant in 2-D or 3-D slopes.; The effect of strength anisotropy to slope stability was investigated by conducting numerical analyses using finite element method (CRISP, 1987). Anisotropic bounding surface model (Crouch & Wolf, 1992) and cubic element were incorporated into the program to simulate anisotropic behavior of soils and consider 3-D boundary condition of slopes. A verification study was made for the anisotropic model and cubic element by comparing numerical predictions with centrifuge model tests. Based on the numerical analyses, it was concluded that extension area on potential slip surface of slopes exists as long as deformation of slopes is allowed and the extension area is more significant in 3-D slope than in 2-D slope.; A centrifuge model test was conducted to verify the numerical results, in which 3-D failure of slope was imposed by local surcharge load. The geometry of 3-D slip surface was assumed based on the deformations observed in the model and 3-D factor of safety for the slip surfaces was evaluated from moment equilibrium condition. Based on the stability analysis, it was found that the 3-D factor of safety is realistically predicted only when the numerically determined extension area is adequately incorporated in the stability analysis.