AI-Based Maintenance Optimization Improves Offshore Wind Operations

As part of InnoWind's WP3, researchers Muhammed Hassan and Hossein Naderian, under Professor Mazaher Karimi, have built an end-to-end AI system that unites fault prediction with maintenance scheduling for offshore wind turbines. Reconstructed from scratch after earlier source code was lost, the system's AI models detect early warning signs of failure—often days in advance—directly from SCADA sensor data such as vibration, temperature and power output, while a "Smart Scheduler" translates these risk levels into realistic maintenance plans that respect vessel and crew constraints, consistently prioritizing the highest-risk turbines. The results demonstrate a fully validated, risk-based and resource-aware predictive-prescriptive maintenance system that reduces downtime, improves resource use and strengthens the operational and cost case for offshore wind farm management.