Baseline Model for Bubbly Flows in OpenFoam: Current Implementation and Future Needs

Many technical processes in industries such as chemical or energy but also numerous natural phenomena involve multiphase flow. Due to the complex physics and the broad range of relevant length scales involved, it is a formidable task to achieve a better understanding of such flows. A detailed insight into the local flow field can be obtained from multiphase computational fluid dynamics, which therefore is a potentially valuable optimization and design tool. Such simulations are feasible within the framework the so-called multi-fluid modelling, in which the different phases are described as interpenetrating continua. Within this framework closure models are necessary to describe the interaction of the phases and a large body of work using different closure relations of varying degree of sophistication exists, but no complete, reliable, and robust formulation has been achieved so far. Accurate and trustworthy predictions with multiphase CFD are only possible if a fixed set of closures is available that has been validated for a wide range of flow conditions and can therefore reliably be used also for unknown flow problems.
In this work a set of closure relations has been implemented into OpenFOAM, which represents the best available knowledge and is applicable for monodisperse turbulent adiabatic bubbly flow. Specifically we select closures for the exchange of momentum, i.e. drag and non-drag forces, and the effect of the disperse phase on the carrier-phase turbulence, the so-called bubble-induced turbulence. This set of closure models may serve as a baseline for further improvements and extensions to more general situations.
Experimental data for bubbly air-water flows, available from literature and from in-house experiments, are used to validate the model. Overall all three experimental data sets are reasonably well reproduced by the simulation results, in particular in the bulk of the flow. Larger discrepancies show up especially in the near-wall region and for the turbulent quantities. Also for bubble sizes around 4.5 to 5 mm, the reproduction of the experimental data with a monodisperse model becomes challenging. From the validation cases the following needs for further developments are identified:

• improved modelling of the near-wall region
• refined description of the bubble-induced turbulence
• modelling of the bubble-size distribution coupled with multiple momentum equations for the disperse phase.

The authors acknowledge the Helmholtz Association for support of the research within the frame of the Helmholtz Energy Alliance 'Energy Efficient Chemical Multiphase Processes'.