Estudo do desempenho de correlações de arrasto sólido-gás na simulação numérica de um leito fluidizado borbulhante

Abstract

The burning of solids fuels in fluidized bed is one of the most advanced technologies used thermal energy power plants. Currently, there is a growing interest for a better understanding of the hydrodynamic processes of fluidization and mathematical models able to predict such phenomena effectively. In this work, Computational Fluid Dynamics (CFD) tool, the open source code MFIX (Multiphase Flow With Interphase Exchanges) is used in the numerical simulation of multiphase flows in a bubbling fluidized bed. In the mathematical modeling of the problem, the Eulerian approach is employed, i.e. the particulate solid and gas phases are considered to be interpenetrating continua. The equations of mass and momentum are solved for each phase, and the coupling between phases is achieved via a momentum transfer term, which is caused by gas-solid drag. The latter can be modeled by theoretical or empirical correlations, among which stand out the correlations of Gidaspow, Hill-Koch-Ladd (HKL), Syamlal and O’Brien and Arastoopour. In the present work the performance of these correlations in the numerical simulation of a bubbling fluidized bed with a central jet of gas is studied. The results produced using different correlations are compared to experimental results available in the literature. The model of Syamlal and O’Brien requires less processing time and presents mesh independent results even for coarse meshes, but predicts much smaller bubbles than those observed experimentally. The models of HKL and of Gidaspow give results which show good agreement with experimental results, especially with regard to the shape of bubbles. These models require high processing time and very refined meshes to produce results independent of the discretization. For the Arastoopour model, preliminary results are presented, which differ greatly from other models and require further study