Modelagem da pressão de fundo de poço em sistemas de escoamento multifásico: uma abordagem utilizando programação genética

Abstract

Wellbore pressure modeling in multiphase flow systems poses a significant challenge in the oil and gas industry due to its direct impact on production efficiency and safety. Despite extensive literature, the application of machine learning techniques for this purpose remains underexplored. This study employs a Genetic Programming (GP) approach to predict wellbore pressure. Utilizing 795 data samples from well productivity tests in Middle Eastern fields, encompassing variables such as oil flow rate, gas flow rate, water flow rate, oil density, drilling depth, wellbore temperature, and wellhead pressure, the SGP-based strategy was applied to develop interpretable symbolic models. These models demonstrated the ability to comprehensibly describe the complex relationship between operational and environmental variables and wellbore pressure. The derivation of interpretable symbolic models highlights the practical applicability of the research, providing a deeper understanding of the factors influencing wellbore pressure and facilitating more informed decision-making by industry professionals.