Synergistic Integration of Machine Learning and Causal Decision Modeling for Offshore Wind Farm Site Suitability Assessment

Abstract

Wind energy is among the most valuable natural resources globally. Offshore wind turbines have proven to be reliable and effective due to their ability to capture higher wind speeds. Furthermore, offshore wind farm site selection is critical for maximizing renewable energy production. However, conventional frameworks in the literature often focus solely on static site attributes or dynamic operational factors in isolation. This study introduces a comprehensive decision-making framework that integrates machine learning and advanced multi-criteria analysis for offshore wind site suitability assessment. The proposed methodology combines static criteria, including technical, environmental, and infrastructure parameters, with dynamic performance indicators such as lifecycle energy output, operational efficiency, downtime, and maintenance costs, predicted using the XGboost algorithm. Criteria weights are determined through the DEMATEL (Decision-Making Trial and Evaluation Laboratory) approach, based on expert pairwise comparisons, effectively capturing interdependencies among criteria. Candidate sites are then ranked using the MARCOS (Measurement of Alternatives and Ranking according to Compromise Solution) method. A multiple-stage validation strategy is then established in order to evaluate each element of the proposed hybrid architecture. The suggested framework delivers a transparent, data-driven decision-support tool, enabling stakeholders to optimize offshore wind farm investments with enhanced accuracy and sustainability.