抄録
Intelligent drilling and geosteering are reshaping modern drilling operations by enabling more accurate subsurface characterization, adaptive trajectory control, and automated decision-making in complex geological settings such as deep and heterogeneous reservoirs. Recent advances in downhole sensing, artificial intelligence, digital twins, and automated control systems have driven a shift from experience-based operations to data-driven closed-loop drilling. This review summarizes key developments using a unified perception–decision–execution framework. Perception technologies (e.g., measurement-while-drilling and multi-source data fusion) are reviewed for real-time capability and accuracy. Decision methods, including physics-based models, data-driven approaches, and digital twin systems, are discussed with emphasis on uncertainty handling and prediction improvement. Execution technologies, such as rotary steerable systems and automated control, are analyzed for stability and adaptability. The integration of these components is shown to improve drilling efficiency, reduce geological uncertainty, and enhance well placement. Remaining challenges include data quality, real-time communication, model generalization, system integration, and cost constraints. Future directions focus on intelligent subsurface models, edge computing, physics-informed digital twins, and fully closed-loop autonomous drilling systems. This review provides a structured overview and practical reference for advancing intelligent drilling toward industrial applications.
