Investigation of Two-fluid Models of Fluidisation Using Magnetic Resonance and Discrete Element Simulations

Abstract

The two-fluid model (TFM) coupled with the kinetic theory of granular flow (KTGF) permits detailed simulations of fluidisation at industrial scales in practicable computational time. However, the TFM requires several assumptions regarding the particle-fluid and particle-particle interaction. These models require validation using experimental measurements before the simulations can be used to design industrial fluidised beds. In this paper we present experimental measurements using Magnetic Resonance Imaging (MRI) and compare these with TFM simulations. However, from these experimental comparisons alone, it is difficult to identify whether errors in the TFM arise from the particle-fluid or particle-particle interaction, or both. We therefore explore the origins of any inaccuracies in TFM simulations of fluidised beds using a recently-developed Discrete Element Model coupled with volume averaged Computational Fluid Dynamics (DEM-CFD). The DEM-CFD model is shown to provide accurate predictions of the time averaged velocity of the particulate phase. A comparison of the TFM and DEM-CFD results demonstrates that modelling the particulate phase using an ideal KTGF does not accurately represent the fluidisation behaviour. These errors are attributed to the lack of rotational friction effects in the KTGF. However, by introducing an apparent coefficient of restitution to account for rotation, TFM simulations of the voidage and time-averaged particle velocity agree with the experimental and DEM-CFD results more closely.