Dynamic Behavior of Local Solids Concentration in Fluidized Beds: Experimental Validation of an Eulerian-Eulerian Model

Recent years have witnessed significant improvements in the computer modeling of fluidized particle systems. Such models hold the potential to simulate complex systems and thus greatly reduce design efforts. As we approach the goal of simulating arbitrary systems, there exists a greater need to challenge the ability of the models to predict key features of operating beds. The simplest challenge is to predict global time average properties such as the pressure gradient and bed expansion. The next level of detail is prediction of local time average characteristics such as the solids fraction at a specified location. A third level is the prediction of dynamic characteristics of local behavior, which has important implications for heat and mass transfer processes, reaction kinetics, and mixing. Our aim is to investigate the ability of an Eulerian-Eulerian (two-fluid) model to predict local fluctuations in the solids phase concentration. Experimentally we utilize needle-capacitance probes to measure time series transients of local solids fraction. The Eulerian-Eulerian model selected for this validation test is the MFIX CFD package (www.mfix.org). Two closures for the interphase momentum transfer coefficient are adopted, the first empirical (a combination of Ergun (1), and Wen and Yu (2)) and the second due to Koch and Hill (3) estimated from lattice-Boltzmann simulations. For validation purposes we look at the dynamic behavior of the experimental and model solids fraction transients. While statistical and spectral features of the artificial signals compare favorably with experiment, local dynamic features are not captured very well. Three reasons account for this: 1) the model does not predict expansion in the emulsion, 2) the model does not capture small-scale density fluctuations in the emulsion, and 3) the model does not correctly predict the bubble to wake transition.