Prediction of the Flow Regime in Liquid-Gas Flows Through Straight Pipes Using Computational Fluid Dynamics

Abstract

The simulation of multiphase flows generally relies on the a priori knowledge of the flow regime that develops in the problem of interest. In the case of two-phase flow in pipes, this knowledge comes from flow regime maps, which were constructed in the classical literature using both theoretical and experimental basis. Some efforts aiming the construction of those kind of maps were done applying CFD (Computational Fluid Dynamics) tools but to a limited extent. The objective of this study is to use CFD tools in order to examine the behavior of two-phase liquid-gas flow in straight pipes and to capture the corresponding flow regime according to the inlet conditions. In particular, oil-gas flows were considered. These flows were assumed as incompressible and isothermal. The computations were performed for two-dimensional (2D) channels and three-dimensional (3D) pipes, and the results were compared to the flow regime maps available in the literature. A detailed comparison between three- and two-dimensional solutions obtained for the same set of parameters was performed in order to validate the use of 2D simulations for the prediction of the flow regime. The method used to determine the phase boundaries is Volume of Fluid (VOF) and turbulence was treated with the k-epsilon model, as incorporated in the open-source toolkit OpenFOAM(R). Adaptive refinement was applied in order to sharpen the liquid-gas interphase with the aim to reduce the computational cost of the simulations maintaining the total number of cells in a tractable amount. Due to limitations in the mesh size, only a portion of the flow regime map could be assessed, including those flow regimes with features of the interphase captured by the mesh. These results were compared with the flow regime map of Taitel and Dukler and other maps constructed from experimental results, where the emphasis was put in the limits defining regions inside the map.