A Graph Auto-encoder Framework for Spatio-temporal Anomaly Detection of Corrosion across a Fleet of Offshore Wind Turbines Using ICCP Data
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
Abstract
Offshore wind farms are exposed to severe marine conditions, which can lead to long-term structural integrity concerns due to corrosion-induced degradation processes. Here, we propose a spatio-temporal anomaly detection methodology using the Impressed Current Cathodic Protection (ICCP) data from an offshore wind farm. First, we employ a graph autoencoder (GAE) to infer the spatial variations in the measurements. We construct a graph based on the spatial proximity between wind turbines, where nodes and edges correspond to wind turbines and distance between turbines. Then, the latent representation of the measurements obtained by the GAE, are passed to a long-short term memory (LSTM) model, which infers the temporal evolution of measured signal and predict the next state. Finally, we perform anomaly detection using a combined scoring that includes graph reconstruction errors, latent prediction errors and observation-space prediction errors. Our results highlight the potential of integrating graph‑based and sequence‑based approaches for industry‑relevant anomaly detection and demonstrate that the proposed methodology can identify turbines and corresponding time periods exhibiting deviations from fleet‑level behavior.
How to Cite
##plugins.themes.bootstrap3.article.details##
Anomaly detection, Graph Autoencoders, LSTM, Time-series analysis
Black, I. M., Yeter, B., Häckell, M. W., & Kolios, A. (2024). Assessing structural homogeneity and heterogeneity in offshore wind farms: A population-based structural health monitoring approach. Ocean Engineering, 311, 118842. doi: 10.1016/j.oceaneng.2024.118842
Brijder, R., Hagen, C. H. M., Cortés, A., Irizar, A., Thibbotuwa, U. C., Helsen, S., et al. (2022). Review of corrosion monitoring and prognostics in offshore wind turbine structures: Current status and feasible approaches. Frontiers in Energy Research, 10, 991343. doi: 10.3389/fenrg.2022.991343
Cui, Y., Bangalore, P., & Bertling Tjernberg, L. (2021). A fault detection framework using recurrent neural networks for condition monitoring of wind turbines. Wind Energy, 24(11), 1249–1262.
Erdogan, C., & Swain, G. (2021). Conceptual sacrificial anode cathodic protection design for offshore wind monopiles. Ocean Engineering, 235, 109339. doi: 10.1016/j.oceaneng.2021.109339
Erdogan, C., & Swain, G. (2022). The effect of macrogalvanic cells on corrosion and impressed current cathodic protection for offshore monopile steel structures. Ocean Engineering, 265, 112575. doi: 10.1016/j.oceaneng.2022.112575
Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780. doi: 10.1162/neco.1997.9.8.1735
Jiang, G., Tang, H., Yue, J., Ding, X., He, Q., & Xie, P. (2026). Conditional graph autoencoder with embedded prior knowledge for wind turbine fault diagnosis. IEEE Transactions on Instrumentation and Measurement.
Jin, X., Lv, S., Kong, Z., Yang, H., Zhang, Y., Guo, Y., & Xu, Z. (2024). Graph spatio-temporal networks for condition monitoring of wind turbine. IEEE Transactions on Sustainable Energy, 15(4), 2276–2286.
Kalovelonis, D. T., Gortsas, T. V., Tsinopoulos, S. V., & Polyzos, D. (2025). A novel design methodology for sacrificial anode cathodic protection systems using numerical modeling: A case study of offshore wind turbine monopile foundations. Ocean Engineering, 318, 120169. doi: 10.1016/j.oceaneng.2024.120169
Kipf, T. N., & Welling, M. (2016). Variational graph autoencoders. arXiv preprint arXiv:1611.07308.
Martin, R., Lazakis, I., Barbouchi, S., & Johanning, L. (2016). Sensitivity analysis of offshore wind farm operation and maintenance cost and availability. Renewable Energy, 85, 1226–1236. doi: 10.1016/j.renene.2015.07.078
Okenyi, V., Bodaghi, M., Mansfield, N., Afazov, S., & Siegkas, P. (2022). A review of challenges and framework development for corrosion fatigue life assessment of monopile-supported horizontal-axis offshore wind turbines. Ships and Offshore Structures, 1–15. doi: 10.1080/17445302.2022.2140531
Pinciroli, L., Baraldi, P., & Zio, E. (n.d.). Early anomaly detection in wind turbines by causality-based graph attention networks. SSRN. Retrieved from https://ssrn.com/abstract=6466344
Porchetta, S., Temel, O., Warner, J. C., Muñoz-Esparza, D., Monbaliu, J., van Beeck, J., & van Lipzig, N. (2021). Evaluation of a roughness length parametrization accounting for wind–wave alignment in a coupled atmosphere–wave model. Quarterly Journal of the Royal Meteorological Society, 147(735), 825–846.
Price, S. J., & Figueira, R. B. (2017). Corrosion protection systems and fatigue corrosion in offshore wind structures: Current status and future perspectives. Coatings, 7(2), 25. doi: 10.3390/coatings7020025
Qian, P., Tian, X., Kanfoud, J., Lee, J. L. Y., & Gan, T.-H. (2019). A novel condition monitoring method of wind turbines based on long short-term memory neural network. Energies, 12(18), 3411.
Ren, Z., Verma, A. S., Li, Y., Teuwen, J. J. E., & Jiang, Z. (2021). Offshore wind turbine operations and maintenance: A state-of-the-art review. Renewable and Sustainable Energy Reviews, 144, 110886. doi: 10.1016/j.rser.2021.110886
Rokhlin, S. I., Kim, J.-Y., Nagy, H., & Zoofan, B. (1999). Effect of pitting corrosion on fatigue crack initiation and fatigue life. Engineering Fracture Mechanics, 62(4–5), 425–444. doi: 10.1016/S0013-7944(98)00101-5
Santhakumar, S., Smart, G., Noonan, M., Meerman, H., & Faaij, A. (2022). Technological progress observed for fixed-bottom offshore wind in the EU and UK. Technological Forecasting and Social Change, 182, 121856. doi: 10.1016/j.techfore.2022.121856
Shittu, A. A., Mehmanparast, A., Shafiee, M., Kolios, A., Hart, P., & Pilario, K. (2020). Structural reliability assessment of offshore wind turbine support structures subjected to pitting corrosion-fatigue: A damage tolerance modelling approach. Wind Energy, 23, 2004–2026. doi: 10.1002/we.2542
Tremps, L., Yeter, B., & Kolios, A. (2024). Review and analysis of the failure risk mitigation via monitoring for monopile offshore wind structures. Energy Reports, 11, 5407–5420. doi: 10.1016/j.egyr.2024.05.026
Zamanzadeh Darban, M., et al. (2025). A novel anomaly detection method for multivariate time series based on spatial-temporal graph learning. Journal of King Saud University – Computer and Information Sciences. doi: 10.1007/s44443-025-00024-3
Zheng, Y., Ma, S., Chen, J., Zhang, J., et al. (2024). Spatiotemporal anomaly detection for multivariate time series based on graph neural network. Information Fusion, 102255. doi: 10.1016/j.inffus.2024.102255
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.