CFD Modelling of Commercial Gas/Liquid/Liquid Separators

Abstract

This thesis examines the use of multiphase modeling in Computational Fluid Dynamics (CFD) commercial code ANSYS FLUENT to study the flow in a three-phase separator. Also, this thesis presents a framework in which CFD data is used to build an Artificial Neural Network (ANN) model to help predict water outlet pressure in a three-phase separator. First, a 2D CFD model is built and validated against a reference numerical model found in the literature. The validation procedure is based on the results generated for the volume fraction. Since the process of finding the water outlet pressure is iterative and can be only executed with a trial-and-error procedure, a Proportional Pressure Controller (PPC) is utilized to find the suitable pressure for different inlet conditions. The PPC findings are compared to the pressure found by the trial-and-error procedure for the 2D model and compared to the pressure of the reference model. Pressure data are extracted from the 2D model and used to examine the ability of the ANN model to predict the pressure relative to mass flowrate. Also, this thesis discusses the "partial" results of a 3D CFD model of the three-phase separator. Flow data extracted from the latter model, upon convergence, should help develop a more complete ANN model with the capability of analyzing flow within a three-phase separator without resorting to the computationally extensive 3D CFD calculations.

Date of Award	6 May 2024
Original language	American English
Supervisor	Abdallah Berrouk (Supervisor)