The application of neural net orks in the operation of engineering structures

The article discusses current developments in the use of artificial neural networks (ANN – Artificial Neural Network) and machine vision technologies in the operation of engineering structures, with an emphasis on the identification of defects in the examination of bridge structures. Scientific research is aimed at finding the most appropriate methodology to simplify and improve the quality of the survey process, as well as the architecture of the neural network. The analysis of domestic and foreign studies devoted to the automatic detection of defects in building structures using convolutional neural networks (CNN – Convolutional Neural Network), YOLO (You Only Look Once) algorithms, AutoML (Automated Machine Learning) methods and software and hardware complexes based on unmanned aerial vehicles. The advantages and disadvantages of existing solutions related to the dependence of recognition quality on the size of the training sample, shooting conditions, the difficulty of locating low-contrast defects and the need for additional involvement in the process of engineering interpretation of the results are considered. The existing approaches to defect identification have been systematized, and the positive and negative elements of tools in machine vision systems have been identified for their prospective consideration when creating a module for bridge inspection. The conclusion is made about the prospects of introducing neural network technologies into the practice of surveying engineering structures, provided that specialized training samples are formed and the accuracy of the models is independently verified. The availability of a powerful neural network tool for examining bridge structures, trained by engineers on specific examples of the bridge sphere, will simplify the survey and increase the number of structures evaluated per unit of time by one team. In conclusion, a multi-stage approach is proposed for further research, including the recognition of structural elements, the detection and classification of defects, as well as subsequent engineering verification of the results.

1. Decree of the President of the Russian Federation No. 490 dated 10.10.2019 ‘On the development of artificial Intelligence in the Russian Federation’. Official Internet portal of legal information dated 11.10. 2019, article 0001201910110003. (In Russ.). URL: http://pravo.gov.ru /.

2. GOST R 59276–2020. Artificial intelligence systems. Ways to ensure trust. General provisions. National Standard of the Russian Federation: introduced on the 01.03. 2021 Moscow: Standartinform; 2021. 16 p. (In Russ.).

3. GOST R 59277–2020. Artificial intelligence systems. Classification of artificial intelligence systems. National Standard of the Russian Federation: introduced on the 01.03. 2021. Moscow: Standartinform; 2021. 16 p. (In Russ.).

4. GOST R 59898-2021. Quality assessment of artificial intelligence systems. General provisions. National Standard of the Russian Federation: introduced on the 01.03. 2022. Moscow: Russian Institute of Standardization; 2021. 24 p. (In Russ.).

5. GOST R 71539–2024. Artificial intelligence. Life cycle processes of an artificial intelligence system. National Standard of the Russian Federation : introduced on the 01.01. 2025. Moscow: Russian Institute of Standardization; 2024. 40 p. (In Russ.).

6. GOST R 71484.2–2024. Artificial intelligence. Data quality for analytics and machine learning. Part 2. Data quality indicators. National Standard of the Russian Federation: introduced on the 01.01. 2025. Moscow: Russian Institute of Standardization; 2024. 40 p. (In Russ.).

7. GOST R 71484.3–2024. Artificial intelligence. Data quality for analytics and machine learning. Part 3. Requirements and recommendations for data quality management. National Standard of the Russian Federation: introduced on the 01.01. 2025. Moscow: Russian Institute of Standardization; 2024. 28 p. (In Russ.).

8. GOST R 71484.4–2024. Artificial intelligence. Data quality for analytics and machine learning. Part 4. The structure of the data quality management process. National Standard of the Russian Federation: introduced on the 01.01. 2025. Moscow: Russian Institute of Standardization; 2024. 28 p. (In Russ.).

9. GOST R 59618–2021. Public roads. Bridge structures. Examination rules and test methods. National Standard of the Russian Federation: introduced on the 01.02. 2022. Moscow: Russian Institute of Standardization; 2022. 54 p. (In Russ.).

10. Shlepneva T. O., Nikishina O. V., Dedyukhina E. S. Application of artificial intelligence in the construction industry: advantages and possible ways of further development. Izvestiya Vuzov. Investment. Construction. Realty. 2025;15(2):336–346. (In Russ.). DOI 10.21285/2227-2917-2025-2-336-346. EDN: YHWLLE.

11. A neural model for detecting deformations in concrete has been created at NSTU NEVIS. NSTU NEVIS: official website. 13.01. 2025. (In Russ.). URL: https://www.nstu.ru/interviews/news_more?idnews=162901.

12. Krylov S. A., Platunov V. Yu., Kashevarova G. G. Automation of the crack detection process during the inspection of brick buildings. Building Structures, Buildings and Structures. 2024;(9):22–31. (In Russ.). EDN: LHAHKH.

13. Shorten C., Khoshgoftaar T. M. A survey on Image Data Augmentation for Deep Learning. Journal of Big Data. 2019;6(60). DOI 10.1186/s40537-019-0197-0.

14. Philip R. E., Andrushia A. D., Nammalvar A., Gurupatham B. G. A., Roy K. A Comparative Study on Crack Detection in Concrete Walls Using Transfer Learning Techniques. Journal of Composites Science. 2023;7(4). Art. 169. DOI 10.3390/jcs7040169.

15. Krylov S. A., Kashevarova G. G. Automation of the process of assessing the technical condition of the exterior walls of brick buildings using machine learning technology. Bulletin of Dagestan State Technical University. Technical sciences. 2025;52(3):61–70. (In Russ.). DOI 10.21822/2073-6185-2025-52-3-61-70.

16. Zhang, J., Beliaeva, Z. V., Huang, Y. Accuracy–Efficiency Trade-Off: Optimizing YOLOv8 for Structural Crack Detection. Sensors. 2025;25(13):3873. DOI 10.3390/s25133873.

17. Ferreira L., Pilastri A., Martins C. M., Pires P. M., Cortez P. A comparison of AutoML tools for machine learning, deep learning and XGBoost. Proceedings of the International Conference on Hybrid Artificial Intelligence Systems (HAIS). 2021. P. 39–50. DOI 10.1007/978-3-030-86230-5_4.

18. FAI ‘ROSDORNII together with FNPC ‘NPO Mars’ carried out diagnostics of the bridge using an unmanned aerial vehicle. FAI ‘ROSDORNII’: official website. online resource. 09/24/2024. (In Russ.). URL: https://rosdornii.ru/press-center/news/novosti/fau-rosdornii-sovmestno-s-fnpts-ao-npo-mars-proveli-diagnostiku-mosta-s-pomoshchyu-bespilotnogo-leta.

19. Detection of defects and damages of metal and reinforced concrete structures. Russian Software Compatibility Catalog. Number in the Unified Register of Russian Software: 27595; date of the decision: 11.04.2025. (In Russ.). URL: https://catalog.arppsoft.ru/product/6419270.

20. NTC Transcor-K LLC: official website. (In Russ.). URL: https://www.transkorworld.com.

21. Automated photo and video processing system for the inspection of concrete and stone bridge supports. Russian Software Compatibility Catalog. Number in the Unified Register of Russian Software: 28141; date of the decision: 19.05. 2025. (In Russ.). URL: https://catalog.arppsoft.ru/product/6418023.

22. Software User's Manual “Automated photo and video processing system for the inspection of concrete and stone bridge supports”. Version 03. Document No. TSF.IT.I.01.01-25. Moscow: Transcor-Service LLC, 2025. Jan. (In Russ.).URL:https://www.transkorservice.ru/files/pdf/01_Инструкция_пользователя_Диспетчер_САО_ФиВ_О_БиКОМ_2025.pdf.

23. Structura software with a neural network module: the experience of digital transformation in the survey of buildings and structures. OSIS-VENTURE ‘Survey of buildings and structures: problems and solutions’: [official website]. (In Russ.). URL: https://www.ozis-venture.ru/xv-konferencziya.

24. Kamaeva Yu. V., Adamtsevich L. A. Prospects of using predictive analytics in construction. Construction and Architecture.2023;11(2). (In Russ.). DOI 29039/2308-0191-2023-11-2-12-12. URL: https://conarc.ru/ru/nauka/article/65013/view.

25. The concept of the use of artificial intelligence in Russian Railways. Approved by the order of Russian Railways dated 14.02. 2022 № 334/r / Russian Railways. 2026. 28 p. (In Russ.). KRIZHT IrGUPS: electronic catalog. URL: https://irbis.krsk.irgups.ru/

26. Smolin A. I. Inspection of bridge spans using a protected UAV and a machine vision system. Track and Track Facilities: Popular Science, Industrial and Technical Journal. Moscow, 2025. No. 3. P. 28–30. (In Russ.).

27. BIT.Visor. BRIDGE-IT LLC is an Accredited IT Company Developing Engineering Software for the Construction Industry. https://bridge-it.ru, BRIDGE-IT LLC. (In Russ.). URL: https://bridge-it.ru/products/bit-visor/.