(37a) CFD Modeling of Mixed Flow Regimes and Regime Transitions in Gas-Liquid Systems

CFD modeling of mixed flow regimes and regime transitions in gas-liquid systems

Gas liquid flows in the oil and gas industries, and nuclear industries often involve transient flow patterns such as slug flow, annular flow etc. These flows are often associated with the existence of more than one flow topology such as droplet, bubbly and separated flows. Mixed flow regimes such as these are more pronounced when the flow path involves bends, junctions and inclines. These flows can often involve regime transitions as well. The ability to predict these flows reliably is crucial for flow assurance and nuclear safety. In this paper we will introduce CFD based multiphase models for describing such mixed flow regimes, as well as the ability to predict regime transitions.

The Multi-Fluid Volume of Fluid method (Multi-Fluid VOF) in ANSYS Fluent implemented in the scope of the Eulerian Framework of Two-phase flow modelling is used for this numerical study. The compressive scheme as second order reconstruction method for the free surface is used for solving the volume fraction equations. The applied compression level is based on gradient normalization approach in order to handle intermittent flow regime behavior covering disperse as well as sharp interfaces. A turbulence damping approach is applied to reduce the turbulent fluctuations in the vicinity of the free surface and to handle the non-physical turbulent overproduction of kinetic energy due to the high velocity gradient across the interphase. This paves the way to capture the congestion of the cross sectional area above the free surface and the acceleration of the gaseous phase and growth of interfacial waves.  

The model is applied to several problems motivated by flow assurance concerns such as slug flows in horizontal channels, and several flow regimes (slug, annular and separated flow regimes) in pipe bends. Flow parameters such as slug frequencies, pressure drops and gas holdups are compared with experimental data. Model results compare favorably with experimental observations. We also apply this model to the problem of regime transitions in a vertical gas liquid riser, where the flow transitions from a bubbly regime to a slug regime. The model is shown to predict better results than models that do not account for such transitions.