«Environmental protection in oil and gas complex»

Application of neural network approaches to improve the reliability of residual life assessment of offshore structures

UDC: 622.276.04:004.032.26
DOI: -

Authors:

¹ National University of Oil and Gas "Gubkin University", Moscow, Russia

Keywords: fatigue life, neural network model, residual service life, offshore oil and gas structures, occupational safety, S–N curve, failure prediction, machine learning, offshore platforms

Annotation:

The authors of the article note that occupational safety during the operation of offshore oil and gas facilities directly depends on the accuracy of fatigue life prediction for structural elements subjected to variable loads. Traditional calculation methods demonstrate significant variability of results, thus complicating maintenance planning and increasing the risk of failures. The adaptive neural network model based on a multilayer perceptron (MLP), trained on datasets derived from several classical S–N approaches, is considered. The developed model, implemented in the Jupyter Notebook environment, is tested on an extended set of input parameters, and shows high accuracy, robustness, and the ability to be retrained for specific operational conditions. This fact enables the integration of disparate calculation schemes into a unified, stable forecast, significantly improving the reliability of residual life estimation. The presented approach lays the foundation for the practical application of modern machine learning techniques aimed at enhancing equipment reliability and occupational safety of offshore infrastructure.

Bibliography:

1. Starokon' I.V., Kalashnikov P.K. Problemy otsenki vozdeystviya vibro-kolebatel'nykh protsessov na snizhenie dlitel'nosti bezopasnoy ekspluatatsii svarnykh soedineniy morskikh statsionarnykh platform dlya dobychi nefti i gaza na shel'fe // Zashchita okruzhayushchey sredy v neftegazovom komplekse. – 2020. – № 1(292). – S. 27–30. – DOI: 10.33285/2411-7013-2020-1(292)-27-30
2. Starokon' I.V. Rezul'taty eksperimental'no-analiticheskogo issledovaniya effektivnosti tekhnologiy remonta svarnykh soedineniy opornykh blokov morskikh statsionarnykh platform s pozitsiy prodleniya ikh resursa // Stroitel'stvo neftyanykh i gazovykh skvazhin na sushe i na more. – 2018. – № 10. – S. 59–62. – DOI: 10.30713/0130-3872-2018-10-59-62
3. Predictive Modeling and Uncertainty Quantification of Fatigue Life in Metal Alloys using Machine Learning: preprint / Chang Jiang, D. Basvoju, A. Vakanski [et al.]. – 2025. – 18 p. – DOI: 10.48550/arXiv.2501.15057
4. A generalized machine learning framework to estimate fatigue life across materials with minimal data / D.V. Srinivasan, M. Moradi, P. Komninos [et al.] // Materials & Design. – 2024. – Vol. 246. – Article No. 113355. – DOI: 10.1016/j.matdes.2024.113355
5. Jimenez-Martinez M., Alfaro-Ponce M. Fatigue Life Prediction of Aluminum Using Artificial Neural Network // Engineering Letters. – 2021. – Vol. 29, Issue 2. – P. 704–709. – URL: https://www.researchgate.net/publication/352383400_Fatigue_Life_Prediction_of_Aluminum_Using_Artificial_Neural_Network (data obrashcheniya 12.07.2025).
6. Prediction of Stiffness and Fatigue Lives of Polymer Matrix Composite Laminates Using Artificial Neural Networks: NASA TM-20230005410 / S.K. Mital, S.M. Arnold, P.L.N. Murthy, B.L. Hearley. – Cleveland: NASA Glenn Research Center, 2023. – 59 p.
7. Chen Jie, Liu Yongming. Fatigue modeling using neural networks: a comprehensive review: preprint. – Tempe: Arizona State University, 2021. – 62 p. – DOI: 10.22541/au.163254701.14733101/v1
8. Farhadi S., Tatullo S., Ferrian F. Comparative analysis of ensemble learning techniques for enhanced fatigue life prediction // Scientific Reports. – 2025. – Vol. 15. – Article No. 11136. – DOI: 10.1038/s41598-024-79476-y
9. A fast prediction method of fatigue life for crane structure based on Stacking ensemble learning model / Zhao Jincheng, Dong Qing, Xu Gening [et al.] // Journal of Engineering and Applied Science. – 2024. – Vol. 71. – Article No. 207. – DOI: 10.1186/s44147-024-00545-0
10. High-cycle fatigue S-N curve prediction of steels based on a transfer learning-guided convolutional neural network / Wei Xiaolu, Wang Chenchong, Jia Zixi, Xu Wei // Journal of Materials Informatics. – 2022. – Vol. 2, Issue 3. – Article No. 9. – DOI: 10.20517/jmi.2022.12
11. Fedotov S., Sinitsin F. Mashinnoe obuchenie: rukovodstvo. – 2024. – URL: https://education.yandex.ru/handbook/ml/article/about (data obrashcheniya 09.07.2025).
12. GOST 25.504-82. Metody rascheta kharakteristik soprotivleniya ustalosti. Raschety i ispytaniya na prochnost'. – Vved. 1983–07–01. – M.: Izd-vo standartov, 1982. – 55 s.
13. Klykov N.A. Raschet kharakteristik soprotivleniya ustalosti svarnykh soedineniy. – M.: Mashinostroenie, 1984. – 157 s.
14. Using the Smith-Watson-Topper Parameter and Its Modifications to Calculate the Fatigue Life of Metals: The State-of-the-Art / T. Łagoda, S. Vantadori, K. Głowacka [et al.] // Materials. – 2022. – Vol. 15, Issue 10. – Article No. 3481. – DOI: 10.3390/ma15103481