DATA SCIENCE-BASED METHODOLOGY TO CORRELATE OPERATIONAL DATA FROM MULTIPHASE FLOW METERS AND SEPARATOR VESSELS USING SEPARATOR DATA AS REFERENCE

In this study, a data-science-based methodology is presented for correlating Multiphase Flow Meter (MPFM) measurements with separator vessel data to enhance accuracy in quantifying oil, water, and gas-flow rates in production lines. Accurate flow rate quantification is essential for reservoir optimization, operational safety, and maximizing economic return. Although MPFMs provide real-time readings, their accuracy remains lower compared to single-phase flowmeters downstream of the separator vessel. Data from twelve wells in a Brazilian offshore field were analyzed, following a workflow comprising (i) Data Collection and Pre-processing; (ii) Correlation Analysis; and (iii) Validation of correlation results. An average delay of 85.94 seconds was identified between gas-flow rate signals; compensating for this shift aligned the time series and eliminated phase oscillations. Oil and water data did not meet minimum quality criteria and were thus excluded from the analysis. The proposed methodology improves MPFM reliability and can ultimately feed machine-learning models for real-time dynamic correction, reducing reliance on physical separators on offshore platforms.  

DOI:  10.26678/ABCM.COBEM2025.COB2025-2649
⁠Conference: 28th International Congress of Mechanical Engineering

Conference Paper
Back to Projects

Paper

Imagem da Sessão Paper

Presentation

Imagem da Sessão  Presentation