AI-Enhanced Detection of Untrained Compound Damage and Scour in Wind Turbines
DOI:
https://doi.org/10.21467/proceedings.8.2.15Abstract
Monopiles are the most used foundation type in Offshore Wind Turbines (OWTs) and are constantly subjected to environmental forces. Structural Health Monitoring (SHM) of OWTs faces many challenges, such as the identification of compound damage and scouring damage. To address these issues, this study proposes an integrated AI-Enhanced Digital Twin Framework. The framework consists of two parts: (1) It employs a few-shot learning approach that combines wavelet transform and Multiple Interference Deep Convolutional Neural Network (MIDCNN). By analyzing the probability superposition of features, the model identifies multi-fault conditions without using compound samples during training. (2) a physics-informed Digital Twin (DT) is established through Finite Element Model Updating (FEMU) to generate high-fidelity synthetic datasets. Subsequently, a novel Graph Neural Network architecture, GAT_GATv2 is introduced, which is based on Proper Orthogonal Decomposition (POD) combined with dynamic sensor topology, for classifying scour depths. The validation demonstrates that the MIDCNN achieves 97% precision for compound states without training, while the GAT_GATv2 achieve an accuracy 83.3% for scour detection. This approach provides an efficient solution for OWT monitoring.
References
f monopile supported offshore wind turbines in deeper waters incorporating stiffness of substructure and SSI. Soil Dynamics and Earthquake Engineering 83, 18–32. https://doi.org/10.1016/j.soildyn.2015.12.011
Baquerizo, J., Tutivén, C., Puruncajas, B., Vidal, Y., Sampietro, J., 2022. Siamese neural networks for damage detection and diagnosis of jacket-type offshore wind turbine platforms. Mathematics 10, 1131.
Bisong, E., 2019. The Multilayer Perceptron (MLP), in: Building Machine Learning and Deep Learning Models on Google Cloud Platform. Apress, Berkeley, CA, pp. 401–405. https://doi.org/10.1007/978-1-4842-4470-8_31
Brody, S., Alon, U., Yahav, E., 2022. How Attentive are Graph Attention Networks? https://doi.org/10.48550/arXiv.2105.14491
Corso, G., Stark, H., Jegelka, S., Jaakkola, T., Barzilay, R., 2024. Graph neural networks. Nat Rev Methods Primers 4, 17. https://doi.org/10.1038/s43586-024-00294-7
Dibaj, A., Ettefagh, M.M., Hassannejad, R., Ehghaghi, M.B., 2021. A hybrid fine-tuned VMD and CNN scheme for untrained compound fault diagnosis of rotating machinery with unequal-severity faults. Expert Systems with Applications 167, 114094.
Faraonis, P., Sextos, A., Papadimitriou, C., Chatzi, E., Panetsos, P., 2019. Implications of subsoil-foundation modelling on the dynamic characteristics of a monitored bridge. Structure and Infrastructure Engineering 15, 180–192. https://doi.org/10.1080/15732479.2018.1503689
Friswell, M.I., Mottershead, J.E., 1995. Finite Element Model Updating in Structural Dynamics, Solid Mechanics and its Applications. Springer Netherlands, Dordrecht. https://doi.org/10.1007/978-94-015-8508-8
Huynh, T.-C., Lee, S.-Y., Dang, N.-L., Kim, J.-T., 2019. Vibration-based structural identification of caisson-foundation system via in situ measurement and simplified model. Struct Control Health Monit 26, e2315. https://doi.org/10.1002/stc.2315
Jawalageri, S., Ghiasi, R., Jalilvand, S., Prendergast, L.J., Malekjafarian, A., 2024. A data-driven approach for scour detection around monopile-supported offshore wind turbines using Naive Bayes classification. Marine Structures 95, 103565. https://doi.org/10.1016/j.marstruc.2023.103565
Kim, M., Song, J., Kim, C.-W., 2025. Near-real-time damage identification under vehicle loads using dynamic graph neural network based on proper orthogonal decomposition. Mechanical Systems and Signal Processing 224, 112175. https://doi.org/10.1016/j.ymssp.2024.112175
Kipf, T.N., Welling, M., 2016. Semi-Supervised Classification with Graph Convolutional Networks. https://doi.org/10.48550/ARXIV.1609.02907
Monti, F., Boscaini, D., Masci, J., Rodolà, E., Svoboda, J., Bronstein, M.M., 2016. Geometric deep learning on graphs and manifolds using mixture model CNNs. https://doi.org/10.48550/arXiv.1611.08402
Mousavi, Z., Varahram, S., Ettefagh, M.M., Sadeghi, M.H., Feng, W.-Q., Bayat, M., 2024. A digital twin-based framework for damage detection of a floating wind turbine structure under various loading conditions based on deep learning approach. Ocean Engineering 292, 116563. https://doi.org/10.1016/j.oceaneng.2023.116563
Niu, X., Lu, G., Chen, X., Wang, H., 2023. A method for structural damage detection considering scour depth under the pile-soil interaction. Mar. struct. 88, 103365. https://doi.org/10.1016/j.marstruc.2022.103365
Prendergast, L.J., Wu, W.H., Gavin, K., 2019. Experimental application of FRF-based model updating approach to estimate soil mass and stiffness mobilised under pile impact tests. Soil Dynamics and Earthquake Engineering 123, 1–15. https://doi.org/10.1016/j.soildyn.2019.04.027
Scarselli, F., Gori, M., Ah Chung Tsoi, Hagenbuchner, M., Monfardini, G., 2009. The Graph Neural Network Model. IEEE Trans. Neural Netw. 20, 61–80. https://doi.org/10.1109/TNN.2008.2005605
Veličković, P., Cucurull, G., Casanova, A., Romero, A., Liò, P., Bengio, Y., 2018. Graph Attention Networks. https://doi.org/10.48550/arXiv.1710.10903
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
