(133c) Continuum Simulations of Dense Granular Flow and Model Assessment | AIChE

(133c) Continuum Simulations of Dense Granular Flow and Model Assessment

Authors 

Belekar, V. V. - Presenter, Iowa State University
Passalacqua, A., Iowa State University
Heindel, T., Iowa State University
Sinha, K., AbbVie Inc.
Subramaniam, S., Iowa State University
The rheology and hydrodynamics of powder flow are important in many industrial processes such as pharmaceutical production, agricultural production, and cement mixing. Key processes in the pharmaceutical industry involve the mixing of different granular materials. Heat and mass transfer considerations are also important in the drying of powders and wet granulation. Accurate and computationally feasible prediction of granular rheology, hydrodynamics, and heat and mass transfer in production-scale devices is a challenge because of the large separation of time and length scales. While Discrete Element Model (DEM) simulations allow for a reasonable representation of particle-level physicochemical processes and particle-particle interactions using a contact force model, they are computationally expensive for systems with more than 10 million particles, thereby rendering the simulation of production-scale devices infeasible with today’s computers. Continuum simulations involve solving the averaged equations for conservation of mass, momentum, and energy by treating the granular medium as a continuum. In this study, continuum simulations of dense monodispersed granular flows are performed to study flows in canonical geometries using three different constitutive models (Chialvo-Sun-Sundaresan (CSS) model (Chialvo et al., 2012), Srivastava and Sundaresan (S&S) model (Srivastava et al., 2003), and the Schaeffer model (Schaeffer et al., 1987)). The results are compared with DEM simulations and experimental data. The mean particle velocity and the error in the prediction of solid stresses are also quantified, thereby permitting a comparative assessment of the different constitutive models.