Machine Learning-Augmented Multi-Criteria Decision Analysis for Optimizing The Decommissioning of Subsea Assets

Abstract The decommissioning of subsea assets, including pipelines, equipment, and infrastructure, presents a critical and multi-faceted challenge for the offshore energy industry. This process requires evaluating environmental, social, technical, safety and economic factors, making decision making complicated and resource-intensive. Traditional multi-criterion decision analysis (MCDA) frameworks, while effective, often demand extensive time and resources due to their reliance on exhaustive data collection and expert judgment. This study introduces an innovative integration of machine learning (ML) with MCDA to automate the ranking of decommissioning alternatives and streamline decision making. Using data from multiple case studies, the ML model incorporates variables such as field characteristics (e.g., water depth, coastline distance), asset characteristics (e.g., pipeline length, equipment types), and required resources (e.g., ships, port facilities). The proposed approach prioritizes high-potential alternatives, reducing the focus on low-ranked options. The results demonstrate high accuracy in replicating expert decisions, significantly reducing costs by minimizing reliance on costly tools such as remotely operated vehicles (ROVs), and accelerating the decision process. By combining ML and MCDA, this framework provides a scalable, efficient, and cost-effective solution to the offshore industry’s decommissioning challenges, setting a foundation for broader AI-driven innovations in decision-making.