Non-Uniformity of Stress States within a Dense Sand Specimen

Non-uniformity of deformations has always been a concern in interpretation of laboratory triaxial test results due to boundary effects. This paper describes the application of a novel inverse analysis framework, SelfSim (Self learning simulations) which permits the extraction of relevant soil (or material) behavior using boundary measurements of load and displacement, to the interpretation of a triaxial test of a dense sand specimen. The specimen is sheared in axial compression with frictional platens, and experiences large dilation and global softening behavior during shearing. SelfSim extracts the non-uniform stress and strain field within the specimen. Soil experiences diverse stress paths arid exhibits a range of mobilized secant friction angles under a variety of loading modes. The approach has the potential to provide greater insights into the sand behavior under previously unexplored complex shear modes than the simple interpretation of behavior using global measurements of force and displacements.