A Study of the Behavior and Micromechanical Modelling of Granular Soil. Volume 3. A Numerical Investigation of the Behavior of Granular Media Using Nonlinear Discrete Element Simulation
Abstract
Nonlinear discrete element simulations were used to provide an insight on the nonlinear modelling of granular soil. These simulations were based on an incremental solution to the nonlinear problem of two spheres in contact, incorporated into discrete element TRUBAL which was further optimized for vector and parallel processing on the IBM 3090 supercomputer. It was found that this approach not only interprets successfully nonlinear behavior of soil, but also provides a wealth of information on the fabric changes during the loading. The yield surface of a granular medium, needed for defining the constitutive relation of such a medium, distorts by forming an apex in the direction of loading while becoming flatter in the opposite direction. This contrary to the practice followed in modelling granular media where the yield surface of soils are typically assumed to retain the same shape. Origin of this distortion phenomenon lies in the texturing (or fabric anisotropy) which occurs in the direction of prestraining, as well as in the redistribution of interparticle contact forces in the absence of significant particle movement during the small strain probes needed to define the yield surface. These phenomena cause certain slip systems to be activated which produce the characteristic apex which appear in the yield surface in the loading direction. Therefore, distribution and magnitude of the contact forces are critical for a good understanding of the macroscopic response of the medium. Accurate modelling of the contact force distribution can be achieved only if the behavior at the contact is fully understood and rigorously modelled.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 22, 1991
- Accession Number
- ADA238158
Entities
People
- Emmanuel Petrakis
- Li Ping
- Ricardo Dobry
- Tang-tag Ng
Organizations
- Rensselaer Polytechnic Institute