Simulation of the Grenoble bubble-column experiment
Buoyancy-driven bubbly flows exhibit complex multi-scale behavior that is difficult to capture in numerical simulations. For example, the flow in bubble columns transitions from the homogeneous to the heterogeneous regime as the superficial gas velocity increases. However, the transition point depends on several parameters such as the bubble size distribution, the sparger design, and whether bubble coalescence is present. Thus, unless the experiments are carefully designed and executed, it is difficult to use data from the literature to validate numerical simulations that are sensitive to these parameters. The Grenoble bubble- column experiments were performed under well-controlled conditions where bubble coalescence is absent and the bubble size in nearly constant. Moreover, data were taken with an optimized Doppler optical probe that simultaneously measured bubble volume fraction, bubble velocity and size, as well as the liquid velocity at a high sampling frequency. Thus, in addition the time-averaged statistics, for the first-time joint statistics such as the probability density function (pdf) of volume fraction and velocity and the related conditional velocity given volume fraction are available for validation. In this seminar, I will report on a validation study for a Eulerian two-fluid model implemented in OpenFOAM. Direct comparisons are made between the Grenoble data and the model predictions in both the homogeneous and heterogeneous regimes. Overall, the model predictions are in good agreement with the experiments, including the shape of the joint bubble-velocity, volume-fraction pdf and its dependence on the flow regime. These results are very encouraging and suggest the model can be used reliably for scale-up studies of industrial bubble columns.