Machine learning methods in the development of leak detection techniques
https://doi.org/10.30724/1998-9903-2025-27-2-177-186
Abstract
The aim of the work is to explore the possibility of using artificial intelligence to find leak detection using training data obtained using vibroacoustic sensors. The paper describes methods of continuous monitoring. Methods of periodic monitoring of leak detection in the pipeline are presented. Testing work was carried out on the laboratory stand. Methods for machine learning are considered, namely boosting, k-nearest neighbors, random forest method, multi-layer neural network method. An algorithm has been developed for compiling data arrays. The results of the program are shown.
Keywords
About the Authors
T. O. Politova
Kazan State Power Engineering University
Russian Federation
Russian Federation
Tatiana O. Politova,
Kazan
A. R. Zagretdinov
Kazan State Power Engineering University
Russian Federation
Russian Federation
Airat R. Zagretdinov,
Kazan
M. V. Sidorov
Kazan State Power Engineering University
Russian Federation
Russian Federation
Mikhail V. Sidorov,
Kazan
Sh. G. Ziganshin
Kazan State Power Engineering University
Russian Federation
Russian Federation
Shamil G. Ziganshin,
Kazan
Yu. V. Vankov
Kazan State Power Engineering University
Russian Federation
Russian Federation
Yuri V. Vankov,
Kazan
References
1. Report on the state of the sphere of heat power engineering and heat supply in the Russian Federation [Electronic resource]. Available online: https://minenergo.gov.ru/viewpdf/10850/80685 (accessed January 20, 2020).
2. Avdyunin E.G. Sources and heat supply systems. Heating networks and heating points. – Moscow; Vologda: Infra-Engineering, 2019. 300c.
3. Assessment of the condition of pipelines using convolutional neural networks / Yu. Vankov, S. Ziganshin, T. Politova [et al.] // Energy. – 2020. – Volume 13, No. 3. – p. 618.
4. Marchenko, A.L. Python: a big book of examples / A.L. Marchenko. Moscow: Moscow University Press, 2023. 361, p .
5. Pavlov A.N., Sosnovtseva O.V., Ziganshin A.R. Multifractal analysis of chaotic dynamics of interacting systems // Izvestiya vuzov. Applied nonlinear dynamics. 2003. Vol. 11, No. 2. pp. 39-54.
6. Klyuev V.V., Sosnin F.R., Kovalev A.V. Non-destructive testing and diagnostics: handbook. Moscow: Mashinostroenie, 2005. 656 p.
7. Safronchik, V.I. Protection of underground pipelines with anti-corrosion coatings / V.I. Safronchik; Stroyizdat: Leningrad, Russia, 1977. (In Russian)
8. Suris M.A., Lipovskikh V.M. Protection of pipelines of thermal networks from external corrosion. Moscow: Energoatomizdat, 2003. 216 p
9. Analysis of the scale invariance of vibroacoustic signals of a pipeline with leaks / A.R. Zagretdinov [et al.] // SPbNTORES: proceedings of the annual NTC. 2023. № 1 (78). Pp. 88-90.
10. Analysis of fluctuations with trend deviation based on the best fit polynomial / Shanshan Zhao [et al.] // Environmental Science. 2022. No. 10. pp. 1-7.
11. Pavlov A.N., Pavlova O.N., Koronovsky A.A. Jr. A modified method of fluctuation analysis of unsteady processes // Letters to the Journal of Technical Physics. 2020. Vol. 46, No. 6. pp. 47-50.
12. Gaponenko S.O., Kondratiev A.E., Zagretdinov A.R. Low-frequency vibroacoustic method for determining the location of hidden channels and pipelines. // Proceedings of the 2nd International Conference on Industrial Engineering. Chelyabinsk, 2016. pp. 2321-2326.
13. G. Loizu and S. J. Maybank, “Nearest Neighbor and Bayesian error Coefficient", IEEE translation. Pattern analysis. Max. Intel., Volume 9, pp. 254-62, February 1987.
14. The K-role Game Method (KNN).[Electronic resource]// URL: https://habr.com/ru/articles/801885 /: (Accessed 05/07/2024).
15. Comparison of algorithms based on random forests: credal random forest random forest and oblique random forest random forest / K. J. Mantas, J. G. Castellano, S. Moral-Garcia, H. Abellan // Software computing - a combination of fundamentals, methodologies and applications. – 2019. – Volume 23, No. 21. – Pp. 10739-10754.
Review
For citations:
Politova T.O., Zagretdinov A.R., Sidorov M.V., Ziganshin Sh.G., Vankov Yu.V. Machine learning methods in the development of leak detection techniques. Power engineering: research, equipment, technology. 2025;27(2):177-186. (In Russ.) https://doi.org/10.30724/1998-9903-2025-27-2-177-186
Views:
71
Email this article 