Euler-Euler Large Eddy Simulations for dispersed bubbly flows

A bubble column provides a good experimental system for the study of turbulent phenomena in bubbly flows. It has a wide range of length and time scales on which turbulent mixing takes place. The largest turbulence scales are comparable in size to the characteristic length of the mean flow and depend on reactor geometry and boundary conditions. The small scales depend more on the bubble dynamics and hence are proportional to the bubble diameter. In bubbly flows, the small scales are responsible for the dissipation of the turbulent kinetic energy as in single-phase flow, but the bubbles can also generate back-scatter, i.e. energy transfer from smaller to larger scales. The combination of the both effects and yields an overall enhancement or attenuation of the turbulence intensity.
In the present paper the effect of turbulence modelling is investigated. In the CFD simulations of bubble columns RANS models are used for turbulence modeling traditionally, but the turbulence is modeled only isotropic and without resolved turbulence length scales. Large Eddy Simulation (LES) offers the opportunity to resolve the large-scale anisotropic turbulence directly and to model the small scales with a Subgrid-Scale (SGS) model. The filtering is mostly based directly on the grid width.
In the present work, Euler – Euler modelling of bubbly flow is performed using two types of turbulence modelling, (unsteady) RANS and LES. The simulations are carried out for two rectangular bubble columns with different inlet designs, the ones of (Pfleger et al. 1999) and (Akbar et al. 2012) and compared with the respective experimental data and previous own results of URANS simulations. During all the calculations the bubble coalescence and breakup are neglected. For the Akbar experiment with a low gas superficial velocity, the bubble size distribution is assumed to be monodispersed. The bubbles induced turbulence is dominant in this case. For such a case, LES may not represent the best option for turbulence prediction, since the largest fluctuations are close to the bubble surface and cannot be resolved, but instead are modeled with a very simple SGS model. However, good results are obtained in the same experiment with a much higher gas superficial velocity, since large-scale turbulence is present and mostly resolved. For the other configuration, LES shows a more plausible amplitude and period in liquid velocity fluctuation in the measure point than the results of URANS. The SGS turbulent kinetic energy will also be considered using two methods of estimation for zero-equation SGS models to improve the prediction.