Data Science-Based Methodology to Correlate Multiphase Flow Measurement with Separator Reference Data

Abstract

Accurate measurement of multiphase flow rates, comprising oil, water, and natural gas, is essential for efficient petroleum production management. This data is crucial in optimizing reservoir performance, ensuring operational safety, and maximizing economic returns. Multiphase flow meters (MPFMs) emerge as promising equipment for the industry's future due to their ability to provide real-time measurements of the combined rates of various phases. However, despite technological advancements, MPFMs have not yet achieved full reliability, presenting challenges related to the precision and consistency of measurements. Currently, three-phase separators are considered reference equipment for multiphase flow measurement, as they allow physical separation of the phases and individual measurement of oil, water, and gas with greater accuracy. In this context, it becomes fundamental to develop methodologies that enable the approximation and correlation of data obtained by MPFMs with the reference data from three-phase separators. Such an approach aims to validate and calibrate MPFMs, enhancing their reliability and promoting their integration into industrial workflows. This study proposes a methodology based on data science and machine learning to correlate multiphase flow measurements from MPFMs with reference data from three-phase separators. By establishing a reliable correlation between these systems, the goal is to improve confidence in MPFM measurements, facilitating their adoption as precise and efficient tools in monitoring and controlling processes in the petroleum industry.