Slug flow characteristics of air-liquid two-phase flow in horizontal pipes over a wide range of liquid viscosities

Abstract

A large number of studies have been conducted to understand the slug flow characteristics of air-liquid two-phase flow in horizontal pipes and to capture parametric trends of critical flow parameters associated with slug flow, such as the liquid slug length, elongated bubble length, and elongated bubble velocity. Based on these findings, many attempts have been made to develop empirical correlations to predict the slug flow parameters. However, the validity of these correlations is limited to a specific working fluid and narrow ranges of geometrical or operating conditions, thereby reducing their effectiveness as useful general predictive tools. This limitation has created a need to develop empirical correlations that offer high accuracy over a broad application range, i.e., a wide range of liquid viscosities, superficial liquid and gas velocities, and pipe diameters. In this study, two-phase slug flow experiments were performed to examine the hydrodynamic features of slug flow under two different pipe diameters (2 and 4 cm) and two different liquid temperatures (25 and 45°C). In addition, a new consolidated database consisting of 1563 slug flow data points was amassed from 16 sources, including the present experimental data, covering a wide range of operating conditions and fluid properties. In particular, this study includes various liquids with a wide range of viscosities, such as heavy oil, light oil, and water. The parametric study of air-liquid slug flow at various viscosity levels revealed that as the mixture velocity increases, the liquid slug length tends to decrease when the liquid viscosity is high and increase when the liquid viscosity is low. On the other hand, the elongated bubble length and elongated bubble velocity increase with increasing superficial gas velocity, regardless of the liquid viscosity. Based on these findings, three empirical correlations were developed to predict the liquid slug length, elongated bubble length, and elongated bubble velocity. These correlations provide excellent predictive capability against the consolidated database, with overall MAE values of 18.1%, 20.6%, and 9%, respectively.