Enhanced Bounding Surface Modeling of Sand Behaviour under Multiple Loading Cycles

Abstract

Cyclic behaviour of soil has stimulated researchers for several decades to develop constitutive models that will be able make accurate predictions of soil behaviour and at the same time they will be simple to use. Constitutive models are considered to be more valid compared to empirical models, because they comprise by equations that have actual physical meaning. Bounding surface plasticity theory is widely used by researchers and one of the most well-known models is SANISAND, established by Dafalias and Manzari (2004), which is the base for the model presented in this work.
The development of offshore industry during last years, have created the need for models that will be able to simulate soil behaviour under high-cyclic loading, since models like SANISAND were not able to capture the actual number of cycles until liquefaction. This can be done with the incorporation of an additional surface in multi-axial space; the memory surface, which keeps track of the stress history of soil, following the work of Corti (2016).
In this thesis the validation of a new, multiple-surface model, that combines the aforementioned theories, was done. This was achieved by using the data base of Wichtmann and Triantafyllidis (2016), who conducted extensive laboratory testing for sand. This database includes monotonic and cyclic tests for sand under a variety of loading conditions and relative densities. Furthermore, this model was used in order to produce cyclic contour diagrams. These diagrams are practical tools that are often used in practice from offshore foundation designers and they can be constructed using a collection of undrained cyclic tests, either triaxial or DSS.