From 8 to 10 July 2026, WILLOW research was presented at PHME26, the 9th Annual Conference of the Prognostics and Health Management (PHM) Society, held in Oslo, Norway. The event brought together researchers and industry experts to discuss the latest developments in asset health management, predictive maintenance and advanced monitoring technologies.
Representing the project, Flanders Make showcased research focused on applying artificial intelligence to fleet-wide corrosion monitoring in offshore wind farms. The work, presented by Saeid Bafandeh, was entitled “A Graph Auto-encoder Framework for Spatio-temporal Anomaly Detection of Corrosion across a Fleet of Offshore Wind Turbines Using ICCP Data.”
Developed within the WILLOW project, the research introduces a novel spatio-temporal anomaly detection framework that combines Graph Autoencoders (GAEs) and Long Short-Term Memory (LSTM) networks to analyse fleet-wide measurements from impressed current cathodic protection (ICCP) systems. By jointly learning the spatial relationships between wind turbines and their evolution over time, the approach enables the identification of abnormal corrosion-related behaviour across entire offshore wind farms.
The methodology demonstrates the potential of Graph AI to support scalable, fleet-level structural health monitoring, helping operators detect potential issues earlier and make more informed maintenance decisions. Such advances contribute to WILLOW’s broader objective of improving asset reliability, extending service life and reducing operation and maintenance costs in offshore wind energy.
The presentation at PHME26 highlights the project’s continued contribution to advancing data-driven solutions for offshore wind asset management and demonstrates how emerging AI techniques can support the future of predictive maintenance in the sector.
Further information
- Conference website: https://phm-europe.org/
- Full paper: A Graph Auto-encoder Framework for Spatio-temporal Anomaly Detection of Corrosion across a Fleet of Offshore Wind Turbines Using ICCP Data | PHM Society European Conference
