Simulation of dense granular and fluid-solid flow

The first objective of the present work is to develop a computational framework for the use of hypoplasticity theory to predict dense granular flow regimes for fluidized bed applications. The work focuses on the theory of Wu and Bauer and includes (i) development of a finite volume framework for the computation of hypoplastic stress tensor, and integration of this framework with a scheme for the computation of the solid flow and solid fraction field, (ii) implementation of the computational scheme in the MFIX software, (iii) application and testing of the theory and the developed numerical schemes for a range of operating conditions, and demonstration of its capabilities and limitations.; The case of flow during the filling of a rectangular domain is computed using the implementation. The findings to date indicate that the theory and the numerical scheme perform reasonably over a range of operating conditions. However, difficulties are encountered in the maximum packing limit. The hypoplasticity theory being considered allows the solid to pack beyond the maximum packing limit. Computations have been made employing modifications to the Wu and Bauer theory by employing additive terms to the stress tensor from plastic theory. This approach is shown to be successful in enforcing the maximum packing limit, but improvements are required to ensure robust numerical performance when the granular flow becomes fully packed.; The second objective of the thesis is to develop a discrete element simulation (DES) to simulating both dense granular flows and fluid-solid flows for general ellipsoidal particles. However, a variety of particles in industrial practice involve particulates of other shapes, particularly ellipsoids. These can be used to model flows of grains and beans, medical tablets and capsules, and with some modifications, in emerging applications such as nanotube/fluid mixtures. The present work involves developing efficient DES methods for ellipsoidal particles and coupling this simulation to continuum simulations of the fluid phase flow.; A soft sphere model for DES is extended for use with ellipsoidal particles. An elliptical particle overlap detection method is developed for use in contact force determination. An efficient quadtree/octree search algorithm is used for neighbor detection. Three-dimensional chute flow is computed and comparisons with previously published simulations and experiments are presented. The comparisons suggest the correct imposition of surface boundary conditions is critical in making accurate predictions. The model is also used to simulate pure granular flows in vibrating beds. Flow regime maps are developed and differences between the behavior of spherical and elliptical particles are presented. Finally, the discrete element simulation is coupled to a simulation of fluid flow using the MFIX code. Fluid-solid coupling is achieved through the use empirical correlations for non-spherical particles in packed beds. (Abstract shortened by UMI.)...