Measurements and CFD simulations of multiphase flows provide significant insight and details that can be used by engineers to design multiphase flow pipelines. This study uses experimental results obtained with a Wire Mesh Sensor (WMS) to validate CFD results for multiphase flow in larger diameter pipes including slug and churn flows in horizontal and vertical pipes. WMS is a state-of-the-art device which yields instantaneous void fraction distributions at pipe cross-sections. Several case studies in slug and churn flow regimes, which are widely observed in the oil and gas industries, were examined in the present work. Comparison of CFD results for average void fraction, temporal phase evolution, and spatial phase distribution show that results are qualitatively and quantitatively in a good agreement with the experimental data. Liquid film behaviour within the elbow for both orientations was investigated. The results from the study have been applied to elbows to examine erosion occurring in multiphase flow through elbows. Dry out was simulated along the outer radius of the bend in the vertical orientation demonstrating the risk of high erosion rates in that area.
Multiphase flow is encountered in many applications in the petroleum and chemical industries. Understanding phase distributions and velocities can aid in the design of process equipment, separators and slug catchers, and prediction of erosion/corrosion and sand transport. Understanding details of multiphase flow is also very important to build models for flow assurance that can be used for equipment design.
Computational Fluid Dynamics (CFD) is a powerful tool that can offer insight into flow behaviour. However, the use of CFD in the simulation of multiphase flow has been mostly implemented for small diameter pipe applications. For example, Ratkovich et al. (1) used Volume of Fluid (VOF) model to compare average void fraction predictions obtained from CFD in a 19 mm vertical pipe with those of empirical correlations. In the study, the flow regime was slug flow and it was found that there are errors of 48 and 25% in the CFD predictions for Newtonian and non-Newtonian liquids, respectively.
Another drawback that can be inferred from the review of available studies is that the range of velocities employed is limited to lower superficial velocities. For instance, using the VOF model, Da Riva and Del Col (2) studied the behaviour of huge waves of churn flow for four different pipe diameters of 10, 20, 30, and 32 mm, while the superficial gas velocity (VSG) ranged from 0.8 to 6 m/s.
