Unified Flow Pattern Prediction of Gas/Viscous-Liquid Two-Phase Flow using Machine Learning

Accurate prediction of flow patterns for two-phase flows in pipes is vital for optimum design and flow assurance of production systems. Currently, mechanistic models are widely used to predict flow patterns in pipes at different operational, geometrical, and fluid properties conditions. However, if a two-phase flow system is highly complex in geometry and fluid properties, it is very challenging to develop a reliable physics-based mechanistic model. The predictions of such models often result in high uncertainty of flow patterns, especially in the condition of viscous liquid two-phase flows in vertical, horizontal, and deviated conduits. Alternatively, data-driven models, which learn and adapt to complex systems, are powerful tools to predict flow behavior and characteristics. This article aims to utilize machine learning (ML) to predict flow patterns for viscous liquid two-phase flow systems in horizontal, inclined, and vertical pipes. A large database of two-phase flow patterns was collected from open literature that covers a wide range of inclination angles and liquid viscosity (2-to-2000 mPa.s). Several ML models are evaluated in this study, namely logistic regression, random forests, support vector machine, K-nearest neighbor, and XGBoost. The ML predictions show that even the simplest machine-learning algorithm, i.e. logistic regression, gives more accurate predictions than the mechanistic model. Results revealed that trained machine learning models applied on test samples are able to predict the flow pattern with a high degree of accuracy of >90% outperforming the mechanistic model, which predicts the data with 60% accuracy. The analysis of the relative importance of the input variables reveals that liquid and gas superficial velocities, and pipe inclination angle are the most important factors for correct classification.