Influence of Combined Empirical Functions on Slug Flow Predictions of Pipelines With Variable Inclinations

Abstract

Prediction of internal multiphase flows in subsea pipelines is an integral part of the oil and gas production system design. High mass and pressure fluctuations are often encountered during the operation with a liquid-gas slug flow regime exhibiting a sequence of long gas bubbles and aerated liquid slugs. It is important for industry to realistically identify the slug flow occurrence and predict slug flow characteristics, depending on several multiphase flow-pipe parameters. These may be achieved using a one-dimensional, steady-state, mechanistic model accounting for a mass and momentum balance of the two liquid-gas fluids within a controlled volume often referred to as a slug unit. By reducing a 3-D flow problem to a 1-D one, several empirical or closure correlations and associated empirical coefficients have been introduced in the literature and used in commercial software predicting slug flows in subsea jumpers, pipelines and risers with variable inclinations. This study aims to investigate the influence of combined 25 closure functions on the predictions of slug flows in horizontal and inclined pipes based on a steady-state mechanistic model for a wide range of superficial liquid and gas velocities. The model with studied closures is implemented by the authors of this study as the numerical tool iSLUG. The model performance is verified with respect to the estimated film liquid holdup, film length and pressure drop per length of a slug unit for an empirically specified translational velocity, slug liquid holdup, slug liquid length and pipe wall wettability. Closure combinations are analyzed using the relative performance factors and compared against available experimental data in order to identify a set of functions suitable for upward, downward and horizontal flows, and the effect of diameter and inclination on the model prediction is considered. The present method and analysis outcomes may further contribute to the improvement of transient liquid-gas flow models to predict more practical cases.