AI Solution Provides the Most Autonomous Framework of Directional Well Sections in High-DLS Well Plans

Well Construction is actively developing and deploying new technologies that provide more consistent directional drilling service delivery and support customers on their journey towards increased sustainability. This paper presents an autonomous directional drilling artificial intelligence (AI) framework. This framework was used to drill the most autonomous section to date on an offshore well in 2023. In 2024, the next level of autonomous directional drilling in a high-dogleg severity (DLS) environment was attempted. Challenges related to directional drilling include encountering constantly changing formations, steering responses, planned trajectories difficult to follow perfectly, drilling data having huge uncertainties, especially the continuous measurements being affected by shock and vibrations, and ever-changing conditions between one well to another. These challenges often require significant human skill to become an efficient directional driller; and skillful directional drillers, who make the best directional decisions, are difficult to deliver for service companies and operators. This paper presents an AI system developed to perform autonomous directional drilling. The AI system derives drilling parameters and steering recommendations to automatically send downlink commands when enabled. In addition to complex sensor fusion for the best state estimation, an optimization engine outputs the best candidate trajectory at each decision point. Several enhancements were necessary to move from advisory to control. The system has been made much more resilient to unexpected measurement-while-drilling (MWD) telemetry or downlink issues. The system's surface deviation controller actively monitors continuous data and takes the appropriate actions to rectify unplanned situations without overreacting. With respect to downlinking automation, the system is designed to use plug-ins as drivers for different equipment. This design allows for accommodating a wide array of actuators, extending from drilling rig automation systems and flow bypass downlinkers at the rig, to cloud interfaces offered by some rig contractors. Recently, the proposed AI system was tasked to perform drilling parameters control and steering control simultaneously while entirely taking command of the rig control system. The deployment was performed in a field with high-DLS requirements often exceeding 6 °/100 ft. A comparison was made with wells drilled using manual mode, those drilled in an advisory mode, and finally, those drilled in a fully autonomous mode.