Numerical solution of gas–solid flow in fluidised bed at sub-atmospheric pressures

Fluidised beds are characterised by excellent thermal and chemical uniformity and have a wide application range including heat and surface treatment, ore roasting and catalyst production. However, compared to other gas-based systems, to fluidise a particulate mass, a significant quantity of gas is required. To conserve gas there is potential to operate the fluid bed under low-pressure conditions. It is also observed that heat transfer remains constant with reduction in pressure. The present work has numerically studied the nature of hydrodynamics in fluidised bed at sub-atmospheric conditions and a new drag law is proposed to account for the increased mean free path of the fluid. A wide range of sub-atmospheric pressures were considered such that slip flow regime, which is characterised with Kn ∼ 1, is applicable. An open source code (MFIX) is used to numerically solve the multiphase problem of a jet in the fluidised bed column with an immersed surface at vacuum pressure conditions. Bubbling fluidisation in shallow and deep beds are also solved. The new drag model takes into consideration the effect of slip flow to model drag force on the particles and the results of velocity distributions in the column and around the submerged surface is presented. The results of velocity distributions from the slip flow model are compared with the existing Gidaspow’s model. Significant differences were observed in the simulation results of velocity distributions and flow structure in the fluidised bed under vacuum conditions.

Figure optionsDownload as PowerPoint slideHighlights
► A new drag model to account for vacuum pressures (slip-flow) in fluidised beds.
► New model compared with Gidaspow’s drag model for performance in vacuum.
► Significant difference in drag force prediction by the two models at low pressures.
► Numerical solution of deep and shallow beds in addition to an immersed surface.
► Significant variation in flow structure, bubble size and velocity distributions found.