ABSTRACT
Scott and Schoustra 1974, applying the FE method in combination with elastic-perfectly plastic constitutive soil model with Mohr-Coulomb failure criterion, produced results contradicting both reality and experiments. In contrast, Bray et al. 1994a, b also utilizing the FE method, but with a hyperbolic non-linear elastic constitutive law achieved satisfactory agreement with small-scale tests. Most recently, Anastasopoulos et al. 2007, has employed similar kind of constitutive model with isotropic strain softening-linearly reducing the mobilized frictional angle and dilation angle with the increase of octahedral plastic shear strain. Walters & Thomas 1982 performed sandbox experiment and conducted numerical simulation of their experiment by FEM. They found that non-associated flow rule and proper strain softening were essential in localization of rupture. In their FE analyses, rupture propagated through the sand and broke the ground surface with only a fraction of the displacement observed in experiments. Tani 1994 performed sandbox tests and FE analyses. He showed the importance of proper modeling the discontinuous behavior of failure surface in analyzing the post failure process as well as the process before rupture. The previous numerical models used for the analysis of fault rupture propagation through overlaying soil beds were not accurate in modeling shear band effect, mesh size effect and confining pressure effect, which shows the necessity for a sophisticated numerical model.
