Enhanced Oil Recovery Technologies: A Review of Mechanisms, Emerging Methods, and Artificial Intelligence-Based Optimization Strategies

Abstract

Primary and secondary recovery methods typically leave 60–70% of original oil in place due to capillary trapping, adverse mobility ratios, and reservoir heterogeneity. Enhanced oil recovery (EOR) targets this residual oil through thermal, chemical, gas, microbial, hybrid, and nanofluid-based processes that alter pore-scale forces to improve displacement and sweep efficiency. Thermal methods reduce viscosity but are energy-intensive; chemical flooding enhances mobility control and lowers interfacial tension but faces instability in high-temperature, high-salinity reservoirs; gas injection achieves miscibility and enables CO2 sequestration yet suffers from early breakthrough and poor conformance; hybrid techniques combine mechanisms to yield 20–40% incremental recovery at the cost of operational complexity. Recent integration of artificial intelligence and machine learning has advanced EOR screening, performance prediction, and optimization, with deep learning models achieving up to 95% accuracy in method selection. Key challenges remain in chemical stability, economic feasibility, and field-scale implementation. Future development requires physics-informed AI frameworks, cost-effective and environmentally benign formulations, and reservoir-specific hybrid designs to maximize recovery while supporting carbon management objectives.