Effect of model formulation on flow-regime predictions for bubble columns



It is well known that bubble-column flow-regime predictions of the Eulerian two-fluid model depend strongly on the model formulation. However, there is no general consensus on what force models must be included to match experiments for even the simplest flows. The flow simulated in this work corresponds to the bubble-column experiments of Harteveld with the additional assumptions of: (i) uniform-size (non-coalescing) bubbles, (ii) spatially uniform bubble injection, and (iii) non-deforming (spherical) bubbles. As in Monahan et al.'s 2005 article, the complete set of interphase force models includes drag, added-mass, lift, rotation, and strain. The simulation results are presented in the form of flow maps parameterized by the bubble Reynolds number and the average void fraction. As expected, the flow maps show a strong dependence on the model parameters; however, qualitatively only two types of flow maps are observed: (1) flows that transition from the homogeneous to the heterogeneous regime at relatively low average void fractions and (2) flows that transition at relatively high average void fractions (0.4–0.6) as are observed in the experiments. Remarkably, type-2 flow maps are only observed for two-fluid models that include the complete set of forces. © 2006 American Institute of Chemical Engineers AIChE J, 53, 2007