Global enhanced oil recovery assessment: A comprehensive data-driven analysis using self-organizing maps

Abstract

This study presents a comprehensive analysis of Enhanced Oil Recovery (EOR) methods' application in the global oil industry, identifying trends and main factors influencing their adoption. A wealth of unstructured and scattered data from public documents covering traditional EOR methods and newer technologies was processed to create a reliable global database, encompassing 1237 EOR projects in reservoirs around the world. Data science techniques and self-organizing maps (SOM) were employed for data analysis, dimensionality reduction, and visualization of complex problems, proving suitable for high-dimensionality and incomplete data. The research integrates various data modalities and data science visualization techniques, enabling a high-level representation of EOR projects' self-contained knowledge, and providing a comprehensive view of EOR application worldwide. The study reveals implicit interrelationships between multiple variables and EOR methods, identifying trends, research and development directions, and better-informed decisions. In-depth analysis showed that factors impacting the choice, application, and success of EOR methods include technical properties of reservoirs, local, commercial, and temporal aspects. Despite the diffusion of secondary and tertiary recovery technologies, a concentration of EOR methods exists in specific regions. Factors influencing EOR projects' implementation, maintenance, and expansion range from technical reasons, deselected reserves, and infrastructure to incentive policies and regulatory aspects of each country, business structure, and global economic scenario. Both petrophysical and non-petrophysical variables influence the application of EOR methods. Availability and costs of purchasing and transporting injection fluids greatly impact decision-making processes. The analysis contributes to understanding the complex network of factors influencing EOR, and this work can aid in making more informed decisions in the industry and research for EOR applications.