«Automation and Informatization of the fuel and energy complex»

Building predictive models based on artificial intelligence to solve the problem of predicting defects precipitation in pipelines

UDC: 620.19:621.43
DOI: 10.33285/2782-604X-2023-10(603)-38-47

Authors:

¹ Gazprom diagnostics, St. Petersburg, Russia
² National University of Oil and Gas "Gubkin University", Moscow, Russia

Keywords: pipeline, defects, in-pipe diagnostics, machine learning, linear regression, Random Forest, mathematical modeling

Annotation:

The article is devoted to the problems of building stable models for predicting pipeline defects occurrence using artificial intelligence. When specifying the model boundary conditions, a widespread class of defects was chosen – stress-corrosion defects arising due to stress corrosion cracking. When calculating and training, the data from the "Infotech" system of the Yamburg–Elets I gas pipeline for the period of 2016–2022 were used. The linear regression and the Random Forest method were used as modeling algorithms. The convergence of results was assessed depending on the depth of training. The accuracy of prediction values for the Random Forest machine learning model reached 75 %. The prospects for further work to improve accuracy characteristics are assessed.

Bibliography:

1. Otsenka veroyatnosti razrusheniya uchastka magistral'nogo nefteprovoda po dannym vnutritrubnoy diagnostiki / D.A. Neganov, V.M. Varshitskiy, A.A. Belkin, E.N. Figarov // Neft. khoz-vo. – 2022. – № 5. – S. 108–112. – DOI: 10.24887/0028-2448-2022-5-108-112
2. Varshitskiy V.M., Kozyrev O.A. Nesushchaya sposobnost' truboprovoda s lokal'nym korrozionnym defektom // Neft. khoz-vo. – 2021. – № 8. – S. 106–109. – DOI: 10.24887/0028-2448-2021-8-106-109
3. Abakumov A.A., Abakumov A.A. (ml.). Magnitnaya diagnostika gazonefteprovodov. – M.: Energoatomizdat, 2001. – 432 s.
4. Ovodkova K.V., Zhuchkov K.N., Zav'yalov A.P. K voprosu podgotovki iskhodnogo massiva informatsii dlya obucheniya neyronnykh setey opredeleniyu parametrov defektov truboprovodov // Tr. RGU nefti i gaza im. I.M. Gubkina. – 2023. – № 2(311). – S. 85–97. – DOI: 10.33285/2073-9028-2023-2(311)-85-97
5. Zhuchkov K.N., Zav'yalov A.P. Sovershenstvovanie tekhnologii vnutritrubnoy diagnostiki truboprovodov s ispol'zovaniem algoritma avtomatizirovannoy obrabotki diagnosticheskikh dannykh // Nauka i tekhnologii truboprovod. transp. nefti i nefteproduktov. – 2022. – T. 12, № 6. – S. 540–549. – DOI: 10.28999/2541-9595-2022-12-6-540-549
6. Acoustic detection and localization of gas pipeline leak based on residual connection and one-dimensional separable convolutional neural network / Wendi Yan, Wei Liu, Hongbo Bi [et al.] // Transactions of the Institute of Measurement and Control. – 2023. – Vol. 45, Issue 14. – P. 2637–2647. – DOI: 10.1177/01423312231156264
7. Diagnosis and Recognition of Pipeline Damage Defects Based on Neural Network Algorithm / Min Zhang, Yanbao Guo, Deguo Wang, Qiang Du // Proc. of the 2022 14th Int. Pipeline Conf. Vol. 1: Pipeline Safety Management Systems; Project Management, Design, Construction, and Environmental Issues; Strain-Based Design and Assessment; Risk and Reliability; Emerging Fuels and Greenhouse Gas Emissions, Sept. 26–30, 2022, Calgary, Alberta, Canada. – DOI: 10.1115/IPC2022-87163
8. Zavyalov A., Zhuchkov K., Vasilchenko M. Process Pipeline Strength Calculation Methodology Enhancement Using Finite-Element Method // J. of Pipeline Systems Engineering and Practice. – 2023. – Vol. 14, No. 2. – DOI: 10.1061/JPSEA2.PSENG-1401
9. Improving the accuracy of estimates of the pulse sequence period using the methodology of complete sufficient statistics / K. Zhuchkov, M. Vasilchenko, A. Zagrebneva, A. Zavyalov // Scientific Reports. – 2022. – Vol. 12. – P. 19932. – DOI: 10.1038/s41598-022-24457-2
10. Analiticheskie i tekhnicheskie aspekty diagnostiki lineynoy chasti gazoprovodov / I.I. Veliyulin, V.I. Gorodnichenko, V.A. Aleksandrov [i dr.] // Territoriya Neftegaz. – 2021. – № 7-8. – S. 44–55.
11. Issledovanie osobennostey KRN magistral'nykh gazoprovodov bol'shogo diametra / V.A. Seredenok, V.L. Onatskiy, V.N. Tolkacheva, R.V. Aginey // Truboprovod. transp.: teoriya i praktika. – 2016. – № 5(57). – S. 12–16.
12. Sil'vestrov A.S., Bulkin V.A., Anvarov A.D. Termotsiklicheskie protsessy kak prichina KRN na magistral'nykh gazoprovodakh // Vestn. Kazanskogo tekhnolog. un-ta. – 2011. – № 18. – S. 168–173.
13. Obzor primeneniya neyrosetey v oblasti dobychi i transporta nefti i gaza / I.I. Khasanov, Z.R. Khasanova, R.A. Shakirov, O.I. Nedel'chenko // Transp. i khranenie nefteproduktov i uglevodorodnogo syr'ya. – 2022. – № 3-4. – S. 11–15. – DOI: 10.24412/0131-4270-2022-3-4-11-15
14. Tolstov A.G. Vvedenie v informatiku sistem tekhnicheskoy diagnostiki: monogr. – M.: IRTs Gazprom, 2007. – 487 s.
15. A comparison of random forest based algorithms: random credal random forest versus oblique random forest / C.J. Mantas, Ja.G. Castellano, S. Moral-García, J. Abellán // Soft Computing – A Fusion of Foundations, Methodologies and Applications. – 2019. – Vol. 23, No. 21. – P. 10739–10754. – DOI: 10.1007/s00500-018-3628-5
16. Ispol'zovanie funktsiy s finitnym nositelem pri otsenke vibroparametrii uzlov gazoperekachivayushchego agregata / E.N. Chepel', K.N. Zhuchkov, A.P. Zav'yalov, Yu.A. Gnezdilova // Oborudovanie i tekhnologii dlya neftegazovogo kompleksa. – 2023. – № 2(134). – S. 86–91. – DOI: 10.33285/1999-6934-2023-2(134)-86-91
17. Budzulyak B.V., Lopatin A.S., Lyapichev D.M. Tekhnicheskoe diagnostirovanie oborudovaniya i truboprovodov ob"ektov neftegazovogo kompleksa s primeneniem innovatsionnykh tekhnologiy // Avtomatizatsiya, telemekhanizatsiya i svyaz' v neftyanoy promyshlennosti. – 2019. – № 11(556). – S. 21–26. – DOI: 10.33285/0132-2222-2019-11(556)-21-26