Coarse-grained dynamics of gas-fluidized beds

Sung-Joon Moon, Sankaran Sundaresan, Ioannis G.Kevrekidis

Princeton University, USA

Particulate flows involving granular materials, powders, and gas-particle mixtures are widespread in nature and industry. Although kinetic theory-based continuum models for rapid granular flows (of large particles) have been developed in the literature, these models have not been tested against flows where the particle assemblies undergo rapid compaction or dilation; most gas-particle flows occurring in industrial practice do manifest such inhomogeneities.

There are also a number of particulate systems, involving fine particles, where cohesive forces become important, and for which continuum models are simply unavailable; either validated continuum models or equation-free multiscale computational schemes, which do not require explicit closures are needed to analyze large scale flows involving such particles.

We describe in this paper a computational study of model problems addressing both of these issues. We model gas-fluidized beds of grains and powders using a hybrid approach by combining dissipative particle dynamics with volume-averaged gas hydrodynamics.

First, we simulate planar traveling waves in gas-fluidized beds and by ensemble averaging the data in a co-traveling frame we determine variables that typically appear in continuum models for gas-particle mixtures. These results are then used to interrogate the adequacy of constitutive models for the stresses in the granular phase which are commonly used in the literature. We find that the particle phase stresses exhibit pronounced compaction/dilation rate dependence, which is not captured by the commonly used constitutive models.

Second, we study the mixing/de-mixing phenomena in fluidized beds of binary mixtures, using descriptions at different levels of coarse-graining. The coarse-grained dynamics of this model problem are explored without explicit derivation of the corresponding governing equations, following an equation-free multiscale computational approach. Through this example, we demonstrate how such a multiscale approach can be applied to gas-particle flow problems for which reasonable continuum models are unavailable.