Unlocking Reservoir Potential: AI-Driven Integrated Well Placement Under Uncertainty in a Giant Onshore Field

Abstract The proposed paper aims to introduce an innovative, automated workflow for identifying and drilling high-performance infill wells in the large onshore reservoir in UAE, a region challenged by its complex geological setting and reservoir dynamics. The focus is on leveraging multi-disciplinary data integration and advanced automation techniques coupled with machine learning to enhance the precision and speed of well placement decisions under complex conditions. Our methodology employs a multi-disciplinary approach that synthesizes various static and dynamic data to map drilling opportunities and risks. We utilize an automated opportunity index generation procedure to rapidly identify potential infill well locations. By integrating these maps with existing well data and "no-go zones", we develop an optimized well placement target map. The procedure also includes automated design and optimization of well trajectories that account for surface and subsurface constraints both for producers and injectors, resulting in a tailored well design and trajectory that enhances drilling accuracy and efficiency. Implementing our integrated, automated infill well placement solution in one of the UAE's largest onshore reservoirs with more than 1000 wells produced substantial results. The pilot project successfully extended the production plateau by several years, resulting in significant incremental oil production. Our approach also facilitated notable operational cost savings and improved the efficiency of well placement processes, with the time required for placement decreasing from several weeks to just days—an efficiency gain of up to 80%. These improvements were achieved by optimizing the number and placement of new infill wells at strategic locations, and by removing underperforming wells, which helped streamline operations and reduce unnecessary expenditures. This methodology not only optimizes well placement under complex geological and operational conditions but also significantly contributes in cost-effective reservoir management and enhanced production efficiency. This paper introduces a novel approach by integrating multi-disciplinary data and automated mapping techniques to optimize infill well placement in complex dynamic reservoir environments. By streamlining the identification of drilling opportunities and risks, our methodology enhances the precision and speed of decision-making processes, offering practical benefits such as reduced new infill well count, optimized oil recovery, and cost-effective reservoir management—an invaluable resource for practicing engineers navigating similar challenges in their field.