Two-Phase Flow Regime Identification with a Multiclassification Support Vector Machine (SVM) Model

This paper presents a novel method for the classification of vertical and horizontal two-phase flow regimes in pipes based on a multiclass support vector machine model. Using previously published experimental data for gas-liquid vertical and horizontal two-phase flows, the goal of the model is to predict the transition region between the flow regimes. The transition region is determined with respect to pipe diameter, superficial gas velocity, and superficial liquid velocity. The support vectors of these data are identified and used to determine the transition zone between the multiphase flow patterns. The model proved to be an accurate classification tool for the identification of two-phase flow regimes in pipes. Our computational results show that flow regime predictions from the MSVM models are generally more accurate than predictions based on theoretical correlations.