Vol. 21, No. 4, pp. 357-366 (2025)
FROM CLASSICAL MODELS TO NEURAL OPERATORS: A STUDY ON MOORING
TENSION PREDICTION FOR FLOATING OFFSHORE WIND TURBINE
Lang-Wei Zhong 1, 2, Yu-Ping Lan 1, 2, Xi-Wei Tang 1, Wei Huang 1, 2, ∗, Jian-Wen He 3 and Si-Wei Liu 4
1 School of Civil Engineering, Sun Yat-Sen University, Guangzhou, China
2 State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian, China
3 Research Center of Submarine Pipeline and Cable,
Research Institute of Tsinghua University in Shenzhen, Shenzhen, China
4 Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University,
Hung Hom, Kowloon, Hong Kong, China
*(Corresponding author: E-mail:This email address is being protected from spambots. You need JavaScript enabled to view it.)
Received: 16 February 2025; Revised: 7 August 2025; Accepted: 7 August 2025
DOI:10.18057/IJASC.2025.21.4.8
 View Article |
Export Citation: Plain Text | RIS | Endnote |
ABSTRACT
Offshore wind technology has emerged as a promising solution for harnessing wind resources in both shallow and deep waters. For floating offshore wind turbines, mooring systems play a critical role in maintaining station-keeping functions. Effective monitoring of mooring loads is essential to ensure safe and cost-efficient operation and maintenance. However, direct measurement of dynamic mooring line tensions is often costly and impractical. To address this challenge, this paper focuses on predicting mooring tensions using accessible motion data from the floating platform. We propose the use of deep learning algorithms, including deep neural networks and deep operator networks, to predict the mooring tension of floating offshore wind turbines based on six-degree-of-freedom (6-DOF) platform motions from OpenFAST or OrcaFlex under extreme sea conditions as input. We compare the prediction accuracy and generalization performance of various models, with the Transformer model under a local attention mechanism (LA-Transformer) and the multi-input deep operator network with an attention mechanism (MIONet-ATT) standing out as the top performers. Specifically, LA-Transformer inherits the precise and efficient foundational architecture of Transformer while pruning timing-independent components, achieving an optimal balance between computational complexity and accuracy. It is a novel model with outstanding precision and a moderate parameter size. MIONet-ATT, designed based on the concept of operator networks, offers better interpretability and generalization. Additionally, it demonstrates promising potential in non-extreme sea state applications.
KEYWORDS
Neural networks, Deep operator networks, Semi-submersible floating offshore wind turbine, Moooring tension, Transformer
REFERENCES
[1] Hall M, Goupee A. Validation of a lumped-mass mooring line model with DeepCwind semisubmersible model test data. Ocean Engineering. 2015;104:590-603.
[2] Chen L, Basu B, Nielsen SR. A coupled finite difference mooring dynamics model for floating offshore wind turbine analysis. Ocean Engineering. 2018;162:304-15.
[3] Zhang Y, Shi W, Li D, Li X, Duan Y. Development of a numerical mooring line model for a floating wind turbine based on the vector form intrinsic finite element method. Ocean Engineering. 2022;253:111354.
[4] Sidarta DE, Kyoung J, O’Sullivan J, Lambrakos KF. Prediction of offshore platform mooring line tensions using artificial neural network. International Conference on Offshore Mechanics and Arctic Engineering: American Society of Mechanical Engineers; 2017. p. V001T01A79.
[5] Qiao D, Li P, Ma G, Qi X, Yan J, Ning D, et al. Realtime prediction of dynamic mooring lines responses with LSTM neural network model. Ocean Engineering. 2021;219:108368.
[6] Wang Z, Qiao D, Yan J, Tang G, Li B, Ning D. A new approach to predict dynamic mooring tension using LSTM neural network based on responses of floating structure. Ocean engineering. 2022;249:110905.
[7] Zheng Z, Hu Z, Xie X, Huang W. Local scour around the monopile: A microscopic perspective using CFD-DEM. Ocean Engineering. 2024;299:117318.
[8] Ma H, Zhang S, Li B, Huang W. Local scour around the monopile based on the CFD-DEM method: Experimental and numerical study. Computers and Geotechnics. 2024;168:106117.
[9] Li B, Huang W, Chen X. Tendon fatigue analysis of tension leg platform using a new time-domain quasi-dynamic method. Marine Structures. 2025;99:103701.
[10] Huang W, Li B, Chen X. A new quasi-dynamic method for the prediction of tendon tension of TLP platform. Ocean Engineering. 2023;270:113590.
[11] Ning J-H, Liu S-W, Wan J-H, Huang W. Line-element formulation for upheaval buckling analysis of buried subsea pipelines due to thermal expansion. Adv Steel Constr. 2021;17:210-20.
[12] Ning J, Lin T, Liu S, Bai R, Huang W. Three-dimensional pipeline element formulation for global buckling analysis of submarine pipelines with sleeper. Marine Structures. 2023;90:103428.
[13] Wang C, Liu J, Li B, Huang W. The absolute nodal coordinate formulation in the analysis of offshore floating operations, Part II: Code validation and case study. Ocean Engineering. 2023;281:114650.
[14] Li B, Wang C. The absolute nodal coordinate formulation in the analysis of offshore floating operations Part I: Theory and modeling. Ocean Engineering. 2023;281:114645.
[15] Huang W, Tang X. Prediction of the dynamic response on the spread catenary mooring system for FPSO by long short-term memory method. Ocean Engineering. 2025;340:122113.
[16] Tang X, Huang W, Li X, Ma G. Prediction of mooring tension of floating offshore wind turbines by CNN-LSTM-ATT and Chebyshev polynomials. Ocean Engineering. 2025;331:121327.
[17] Payenda MA, Wang S, Jiang Z, Prinz A. Prediction of mooring dynamics for a semi-submersible floating wind turbine with recurrent neural network models. Ocean Engineering. 2024;313:119490.
[18] Das S, Saha P. A review of some advanced sensors used for health diagnosis of civil engineering structures. Measurement. 2018;129:68-90.
[19] Wan A, Chang Q, Khalil A-B, He J. Short-term power load forecasting for combined heat and power using CNN-LSTM enhanced by attention mechanism. Energy. 2023;282:128274.
[20] Li Y, Xiao L, Wei H, Kou Y, Yang L, Li D. A time–frequency physics-informed model for real-time motion prediction of semi-submersibles. Ocean Engineering. 2024;299:117379.
[21] Cao Q, Goswami S, Tripura T, Chakraborty S, Karniadakis GE. Deep neural operators can predict the real-time response of floating offshore structures under irregular waves. Computers & Structures. 2024;291:107228.
[22] Zhao J, Zhao Y, Zou L. A fully adaptive time–frequency coupling model using self-attention mechanism based on deep operator network for very short-term forecasting of ship motion. Physics of Fluids. 2024;36.
[23] Zhang Q, Zhang H, Zhao X, Ding J, Xu D. Multiple-input operator network prediction method for nonlinear wave energy converter. Ocean Engineering. 2025;317:120106.
[24] Thomsen JB, Bergua R, Jonkman J, Robertson A, Mendoza N, Brown C, et al. Modeling the TetraSpar floating offshore wind turbine foundation as a flexible structure in OrcaFlex and OpenFAST. Energies. 2021;14:7866.
[25] Salehinejad H, Sankar S, Barfett J, Colak E, Valaee S. Recent advances in recurrent neural networks. arXiv preprint arXiv:180101078. 2017.
[26] Graves A. Long short-term memory. Supervised sequence labelling with recurrent neural networks. 2012:37-45.
[27] Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems. 2012;25.
[28] Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, et al. Attention is all you need. Advances in neural information processing systems. 2017;30.
[29] Lu L, Jin P, Karniadakis GE. Deeponet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators. arXiv preprint arXiv:191003193. 2019.
[30] Khanam R, YOLOv11 MH. an overview of the key architectural enhancements. Published online October 23, 2024. 2024. doi: 10.48550. arXiv preprint arXiv241017725.
[31] Tian Y, Ye Q, Doermann D. Yolov12: Attention-centric real-time object detectors. arXiv preprint arXiv:250212524. 2025.
[32] Chen Y, Yuan L, Qin L, Zhang N, Li L, Wu K, et al. A forecasting model with hybrid bidirectional long short-term memory for mooring line responses of semi-submersible offshore platforms. Applied Ocean Research. 2024;150:104145.
[33] Lee S, Shin Y. On the training and generalization of deep operator networks. SIAM Journal on Scientific Computing. 2024;46:C273-C96.