Machine Learning-Driven Subsurface Zonation and Connectivity Mapping for Sustainable Reservoir Management

This study integrates advanced data analytics and machine learning techniques to optimize reservoir characterization and production strategies in the Gabo Field, Niger Delta. Utilizing k-means clustering, decline curve analysis (DCA), decision trees, and Deepseek-R1 large language model, we conducted comprehensive reservoir connectivity and AI zonation analyses, well placement optimization, and enhanced oil recovery (EOR) potential assessment. The methodology combines petrophysical data, production histories, and fluid contact depths to delineate reservoir zones, identify production regimes, and evaluate EOR candidates. Results reveal three distinct reservoir zones and significant heterogeneity in reservoir quality, with Zone C exhibiting superior porosity and permeability. Connectivity analysis identifies three compartments, highlighting the isolated nature of high-performing wells. The integration of machine learning models achieves 87% accuracy in EOR candidate identification and provides insights into optimal intervention strategies. This research demonstrates the potential of AI-driven approaches in reservoir management, offering a 40% reduction in analysis time and identifying bypassed zones with 40% higher recovery potential compared to traditional methods. The findings have implications for field development planning, production optimization, and the broader application of data analytics in petroleum systems.