Recently, the effect of bubbles on the generation and destruction of turbulence in the liquid phase, the so-called Bubble-Induced Turbulence (BIT), is getting more and more attention in the numerical simulation of bubbly flows. There are several theories and models available in the literature, which helps much to understand the inherent characteristics of BIT. However, a systematic validation of these models is still missing. In the current work, popular models considering the additional BIT are implemented into a 1D Test Solver, where the standard k-e model for traditional Shear-Induced Turbulence (SIT) is available. The Test Solver was developed specially for the case of vertical pipe flows by Lucas et al. (2001) and for the purpose of an efficient pre-test of closure models for CFD codes. Its general applicability has been tested in an amount of previous work such as Lucas et al. (2005; 2007).
In the current work, turbulence parameters (k, μt) as well as liquid velocity profiles predicted by the modified k-ε model with the consideration of BIT are compared with available experimental data published by different investigators.
The contribution of BIT and the effect of various models are investigated for mono-dispersed bubbly flows. The flow is assumed to be fully-developed and moreover, the radial gas volume fraction profile is taken from the measurement directly. The results prove that for the test cases with high gas volume fractions (high superficial gas velocity or low superficial liquid velocity) the neglecting of BIT will lead to an obvious underestimation of turbulence parameters. Furthermore, noticeable inconsistency can be observed in the results delivered by different BIT models, which is mainly caused by the time scale assumed by these models for the destruction of the pseudo turbulence. In a word, further effects are needed to be invested in this aspect.