SGER: Experimental Investigation of Stress Rotation Effects in Soils

Catholic University Of America, Washington DC

Investigators

Abstract

Modeling soil behavior during large stress reversals and modeling cross-anisotropic soil behavior have been done in the past by models formulated in terms of stress invariants and incorporating combinations of isotropic and rotational kinematic hardening. This technique seems to capture the soil behavior observed in triaxial compression and cubical triaxial tests with good accuracy. However, incorporation of measured behavior during large stress reversals in torsion shear tests into the kinematic hardening model has proved to be problematic. In fact, the attempt to include effects of stress rotation in the rotational kinematic model has illuminated why this type of model does not capture all aspects of soil behavior, and why the concepts behind the kinematic hardening models, whether translational or rotational, will not be able to correctly predict soil behavior under general stress conditions, as required in a finite element or finite difference setting. In the experimental investigation, the behavior of granular materials during pure stress rotation will be studied. This condition will highlight the shortcomings of existing isotropic and kinematic hardening models. The research will primarily be performed to (a) demonstrate the shortcoming of existing constitutive models, (b) evaluate the magnitude of the missing plastic strains and consequently the severity of the limitations of the constitutive models, (c) initiate data collection that will indicate the direction of new modeling efforts required to capture the missing capability of predicting effects of stress rotation correctly, and (d) formulate a plan for a more extensive experimental program on the basis of the findings from the exploratory experiments proposed here. Experimental results from torsion shear tests with stress rotation will demonstrate and highlight the extent of the limitations of currently existing isotropic, as well as kinematic hardening constitutive models. These limitations consist of their failure to predict soil behavior correctly in most situations that occur in realistic boundary value problems, and for which they purportedly were developed. This work will have the broader impact that renewed efforts will be forthcoming to update or reformulate constitutive models that will be able to include the effects of stress rotation. The proposed research will therefore have major impact on the future of constitutive modeling and it will impact the FEM and the FDM calculations of boundary value problems involving engineering materials.