Generalized bounding surface model for cohesive soils: a novel formulation for monotonic and cyclic loading
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Constitutive models have been developed to simulate the soil response under different loading conditions. Within the wide spectrum of these conditions, earthquakes have been of interest to the research community worldwide. Traditionally, researchers have largely focused on simulating the response of cohesionless soils. By contrast, far less research has been done on simulating the response of cohesive soils subjected to cyclic loading. Consequently, certain issues associated with the simulation of such response are still being addressed.
A review of experimental results has identified several key characteristics of cyclically loaded cohesive soils that any rational mathematical simulation must account for. Accordingly, the scope of this dissertation is the development of an improved and generalized elastoplastic bounding surface model that is suitable for simulating the general response of cohesive soils, with emphasis on cyclic loading.
In its most general form, the Generalized Bounding Surface Model (GBS-Model) for cohesive soils is a fully three-dimensional, time-dependent model that accounts for both inherent and stress induced anisotropy. The rotational hardening law and the shape hardening function were chosen after a thorough review of past modeling practices; in both cases, the selected functional form simplified earlier versions of the bounding surface model without compromising the GBS-Model's predictive capabilities. In addition, to better simulate the behavior of cohesive soils exhibiting softening, the model employs a non-associative flow rule. To more accurately simulate the behavior of cohesive soils subjected to cyclic loading, the model uses a robust general methodology for locating the projection center at any point in stress invariant space, within or on the bounding surface, associated with the radial mapping rule that results in a Lode angle dependency not only of the plastic potential surface but of the bounding surface as well.
If all of the aforementioned features of the GBS-Model are not deemed necessary for the simulation of a particular soil and boundary conditions, then the model can be suitably simplified making it appropriate for adaptive model simulations that can be tailored (through suitable simplifications) to the specific nature of a boundary value problem. For instance, the GBS-Model can be adaptively changed from a complex and more accurate version (e.g., anisotropic with non-associative flow rule) to a simpler one (e.g., isotropic with associative flow rule); thus, minimizing errors and increasing numerical efficiency, thereby reducing computational cost. Finally, a wide range of cohesive soils is used to compare experimental results
with the numerical simulations. Besides, a sensitivity analysis, showing how the new
parameters influence the simulations, is also presented.