The transport of unprocessed gas streams in production and gathering pipelines is becoming more attractive for new developments, particularly those in less friendly environments such as deep offshore locations. Transporting gas, oil, and water together from wells in satellite fields to existing processing facilities reduces the investments required for expanding production. However, engineers often face several problems when designing these systems. These problems include reduced flow capacity, corrosion, emulsion, asphaltene or wax deposition, and hydrate formation. Engineers need a tool to understand how the fluids travel together, to quantify the flow reduction in the pipe, and to determine where, how much, and what type of liquid that would form in a pipe. The present work provides a fundamental understanding of the thermodynamics and hydrodynamic mechanisms of this type of flow. We present a model that couples complex hydrodynamic and thermodynamic models for describing the behavior of fluids traveling in near-horizontal pipes. The model presented herein focuses on gas transmission exhibiting low-liquid loading conditions. The model incorporates a hydrodynamic formulation for three-phase flow in pipes, a thermodynamic model capable of performing two-phase and three-phase flash calculations in an accurate, fast, and reliable manner, and a theoretical approach for determining flow pattern transitions in three-phase (gas-oil-water) flow and closure models that effectively handle different three-phase flow patterns and their transitions. The unified two-fluid model developed herein is demonstrated to be capable of handling three-phase systems exhibiting low-liquid loading. Model predictions were compared against field data with good agreement. The hydrodynamic model allows (1) the determination of flow reduction due to the condensation of liquid(s) in the pipe, (2) the assessment of the potential for forming substances that might affect the integrity of the pipe, and (3) the evaluation of the possible measures for improving the deliverability of the pipeline.
Gas-Liquid Two-phase Stratified Flow Interface Reconstruction with Sparse Batch Normalization Convolutional Neural Network