Transforming Well Planning through an AI-Assisted Well Trajectory Optimization Approach Applied to an Offshore Field in the Middle East

Abstract

This paper describes the use of Artificial Intelligence (AI) to support well planning in an Abu Dhabi offshore field. In this application, AI has been used for automated and unbiased evaluation of well trajectories, with the objective to optimize the cost, risk versus value trade-offs while considering complex issues such as anti-collision with existing wells. A Rapid Random Tree (RRT) algorithm, well known for applications in robotics, has been used to generate well trajectories for 2 actual drilling projects. The algorithm creates a full and unbiased option space of feasible well trajectories, presented in a custom-built and collaborative digital solution. Results demonstrate that AI-generated well trajectories were 2-5% shorter than manually planned and/or actual drilled wells. This use case also shows that an AI can design thousands of possible well trajectories in only a few hours, adhering to well design rules and anti-collision constraints. This would lead to significant time savings, and possibly material drilling cost reductions, in even more congested brownfield assets. This paper describes a real application of AI-assisted well trajectory planning in an operational setting, with a comparison to manually planned and actual drilled wells. As such, this provides a rather unique insight into the business value-adding potential of Artificial Intelligence in traditionally manual work processes.