«Proceedings of Gubkin University»

Machine learning in diagnostic model of structural health monitoring system of process pipelines

UDC: 622.691.4
DOI: -

Authors:

¹ National University of Oil and Gas “Gubkin University”, Moscow, Russian Federation
² Gazenergoexpertiza, St. Petersburg, Russian Federation

Keywords: technical condition monitoring, gas treatment plant, gradient boosting, ARIMA, sea pipeline, process pipelines, machine learning, time series, monitoring system

Annotation:

The paper provides an assessment of the applicability of Seasonal Autoregressive Integrated Moving Average model and Regression Gradient Boosting model for predicting the load on the support of the collector of process pipelines of the gas treatment plant, the increase and decrease of which is caused by linear expansion of the metal of pipelines as a result of cycles of heating to 260–300 °C and cooling to the ambient air temperature of the treated natural gas. The values of the time series of forces on the support preceding the predicted value are used as the main features (variables) for prediction. As an additional feature for prediction, measurements of the surface temperature of the process pipelines are transferred into the regression model of gradient boosting. The results show that Regression Gradient Boosting model significantly outperforms Seasonal Autoregressive Integrated Moving Average model and shows satisfactory accuracy of prediction for practical applications. At the same time, the gradient boosting model, which is given additional temperature features, shows a mean absolute error 31 % lower than that without temperature features.

Bibliography:

1. Lyapichev D.M., Lopatin A.S. Monitoring tekhnicheskogo sostoyaniya gazoprovodov. – M.: RGU nefti i gaza (NIU) imeni I.M. Gubkina, 2021. – 216 s.
2. Avtomatizirovannaya sistema monitoringa napryazhenno-deformirovannogo sostoyaniya kriticheskih uchastkov truboprovodnyh sistem / B.V. Budzulyak, A.P. Gol’dzon, A.P. Zav’yalov [i dr.] // Avtomatizaciya, telemekhanizaciya i svyaz’ v neftyanoj promyshlennosti. – 2018. – № 10. – S. 12–16.
3. Lyapichev D.M., Nikulina D.P., Fil’chenkova A.A. Prakticheskie aspekty primeneniya sistem monitoringa i prediktivnoj analitiki gazoperekachivayushchih agregatov // Trudy Rossijskogo gosudarstvennogo universiteta nefti i gaza imeni I.M. Gubkina. – 2023. – № 2 (311). – S. 98–108.
4. Pipeline inspection and health monitoring technology. The key to integrity management / H. Lu, Z.-D. Xu, T. Isley [et al.]. – Singapore: Springer Nature Singapore Pte Ltd., 2023. – 285 p.
5. Zav’yalov A.P. Analiz sovremennyh tendencij razvitiya sistem remontno-tekhnicheskogo i diagnosticheskogo obsluzhivaniya neftegazovyh proizvodstv // Oborudovanie i tekhnologii dlya neftegazovogo kompleksa. – 2018. – № 6. – S. 67–72.
6. Lyapichev D.M. Sintez sistem monitoringa tekhnicheskogo sostoyaniya tekhnologicheskogo oborudovaniya i truboprovodov neftegazovyh ob’’ektov // Gazovaya promyshlennost’. – 2020. – № 10 (807). – S. 26–31.
7. Osnovy tekhnicheskoj diagnostiki. V 2-h knigah. Kn. I. Modeli ob’’ektov, metody i algoritmy diagnoza / pod red. P.P. Parhomenko. – M.: Energiya, 1976. – 462 s.
8. Entezami A. Structural Health Monitoring by time series Analysis and Statistical Distance Measures. – Milan: Springer, 2021. – 145 p.
9. Mustafina S.I., Zhilyakov S.A. Primenenie nejronnyh setej dlya diagnostiki i prognozirovaniya tekhnicheskogo sostoyaniya oborudovaniya // Avtomatizaciya i informatizaciya TEK. – 2023. – № 7 (600). – S. 13–21. – DOI: 10.33285/2782-604X-2023-7(600)-13-21
10. Farrar C.R., Worden K. Structural health monitoring: a machine learning perspective. – Chichester: Wiley, 2013. – 631 p.
11. Analiz rezul’tatov opytno-promyshlennoj ekspluatacii sistemy monitoringa tekhnicheskogo sostoyaniya zdanij, sooruzhenij, tekhnologicheskogo oborudovaniya i truboprovodov kompressornoj stancii / D.M. Lyapichev, M.M. Admakin, S.V. Romanov [i dr.] // Oborudovanie i tekhnologii dlya neftegazovogo kompleksa. – 2019. – № 2 (110). – S. 54–58.
12. Admakin M.M., Poletaev M.G., Romanov S.V. Opyt ekspluatacii kompleksnoj sistemy monitoringa tekhnicheskogo sostoyaniya kompressornoj stancii // Vesti gazovoj nauki. – 2020. – № 2 (44). – S. 15–21.
13. Perov S.L., Sorokin A.V. Tekhnicheskie resheniya po povysheniyu nadezhnosti tekhnologicheskih truboprovodov UPGT // Gazovaya promyshlennost’. – 2017. – № 1 (747). – S. 68–73.
14. Osobennosti proektnogo ispolneniya tekhnologicheskih truboprovodnyh obvyazok linij adsorbcii UPGT KS “Portovaya” / V.A. Seredenok, V.N. Sivokon’, S.I. Sajchenko [i dr.] // Gazovaya promyshlennost’. – 2016. – № 12 (746). – S. 58–64.
15. Issledovanie prichin neproektnyh vertikal’nyh peremeshchenij uchastkov kollektora gaza regeneracii linii adsorbcii kompressornoj stancii “Portovaya” / S.A. Shurashov, A.V. Sorokin, D.A. Voshchenkov [i dr.] // Oborudovanie i tekhnologii dlya neftegazovogo kompleksa. – 2019. – № 2 (110). – S. 69–74.
16. Vvedenie v statisticheskoe obuchenie s primerami na yazyk Python / G. Dzhejms, D. Uitten, T. Hasti [i dr.] / per. s angl. A.Yu. Gin’ko. – M.: DMK Press, 2024. – 846 s.