Almost all natural soil deposits are anisotropic and non-homogeneous. The anisotropy is pro-duced from a combination of particle placement during deposition/formation (also called geomet-rical or inherent anisotropy) or/and from overbur-den pressures (stress-induced anisotropy) that is caused by the initial state of stress at the end of consolidation (Casagrande and Carrillo 1944). This causes strength and stiffness to be orientation dependent, which is important in geotechnical analysis of problems involving large rotation of principal stress directions, such as embankments, tunnels, deep excavations and slopes. However, in practice the design of such structures is still based on isotropic models in most cases. These parame-ters are usually obtained from some standard laboratory devices such as triaxial or direct shear test.
In order to consider anisotropy in the analysis, some models have been developed such as kine-matic hardening models and micro-structure mod-els. Kinematic hardening models (e.g. Whittle and Kavvadas, 1994, Pestana and Whittle, 1999) are complicated and they require additional parame-ters that cannot be obtained easily by conven-tional laboratory tests. In micro-structure models there is no need of introducing additional material parameters because the strength of soil in any di-rection can be modelled by a so-called micro-structure tensor (Pietruszczak and Mroz 2000). In this study a micro-structure model is implemented into the multilaminate framework, introduced for soils by Pande and Sharma (1983).
Conceptually both strength and stiffness ani-sotropy can be modelled by this approach, but be-cause bonding of natural clays first of all causes a variation of yield surface in different directions only strength anisotropy has been considered in this study. Finite element stability analyses of a slope are performed employing the above men-tioned model. The effect of soil strength anisot-ropy is studied and results are compared to calcu-lations employing isotropic models.