BUBBLE COMPUTATION, GRANULAR TEMPERATURES, AND REYNOLDS STRESSES

Abstract

Abstract Bubbles were simulated in a two-dimensional fluidized bed with a constant inlet velocity using two computer codes, the IIT code and the MFIX code. The computational results were compared to the Jung et al. (Citation2005) experiments in a thin bubbling bed of 530 μm glass beads. The use of higher order numerics produces better bubble resolution due to smaller numerical diffusion. The computed bubble sizes and their distributions agreed with the experiments. The simulations show that there is no bubble formation for sufficiently elastic particles. CFD computations and previous experiments show that in the bubbling fluidized beds there exist two random oscillations. The first kind is due to random oscillations of particles and is measured by the conventional granular temperature. The second one is due to motion of bubbles and gives rise to Reynolds type stresses. It is shown that the particle granular temperature is much smaller than the bubble-like granular temperature computed from the average of the normal Reynolds stresses, measured by Cody using a shot noise technique. Keywords: FluidizationGranular temperatureNumerical diffusionBubble sizeBubble formation criterion Acknowledgment We thank M. Syamlal and S. Benyahia of Fluent, Inc., for their technical support with the MFIX computer code. This study was supported by the University of Pittsburgh NETL student Partnership Program, partially by the National Science Foundation grant to Professor Gidaspow and by the ORISE program. Notes a The upwind scheme for the IIT code is the donor cell method. b Experimental value was measured in the center region.