Back to issue Optimization of selection and interpretation of features for predicting a well technical state
UDC: 622.276.5.054.3:004.896
DOI: -
Authors:
1,2,1,2,1,2,3,41 TatNIPIneft Institute of PJSC "TATNEFT" named after V.D. Shashin, Almetyevsk, Russia
2 Almetyevsk State Technological University "Petroleum Higher School", Almetyevsk, Russia
3 Institute of Oil and Gas Ufa State Petroleum Technological University in the City of Oktyabrsky, Oktyabrsky, Russia
4 Mechanical Engineering Institute named after A.A. Blagonravov, RAS, Moscow, Russia
Keywords: production casing leak, machine learning, statistical importance, feature optimization, water chemistry
Annotation:
The paper considers the selection of optimal number of features for prediction model training and determination of statistical feature importance using statistical and machine learning methods. The determination of the optimal feature count was conducted across 9 feature groups, categorized based on their degree of influence on the probability of leak occurrence in production wells. The F1-score on the test set served as the target metric for identifying the optimal feature group. Experiments for determination of optimal set of features suggest that the best prediction quality is achieved using 19 features, except for maximum inclination angle. However, a set of 12 features exhibits 3 % reduction in quality metrics, being an acceptable compromise for cases with lack of required additional-feature data, while labor intensity for parameter collection and processing is reduced. Additionally, feature statistical significance was evaluated using Gain-Based Importance, Bootstrap confidence intervals, and Permutation Importance methods. Analysis of statistical feature importance has revealed that such factors as sulfates content, solution supersaturation factor, Cl–(Na + K)/Mg hydrochemical coefficient, well age at the time of survey, and water cut are key contributors to casing leaks in production wells. A comparison of the different approaches confirmed the robustness of the identified factors and their critical role in determining leak risk. The obtained results justify the selection of the optimal input parameter set for the model’s field deployment and provide a foundation for further improving well integrity diagnostic methods.
Bibliography:
1. K voprosu o negermetichnostyakh ekspluatatsionnykh kolonn / L.B. Khuzina, I.G. Fattakhov, R.R. Khuzin, S.V. Lyubimova // Stroitel’stvo neftyanykh i gazovykh skvazhin na sushe i na more. – 2023. – № 4(364). – S. 35–38. – DOI: 10.33285-0130-3872-2023-4(364)-35-38
2. Nazarov V.F., Mukhutdinov V.K. Kontrol’ germetichnosti obsadnoy kolonny i NKT v nagnetatel’nykh skvazhinakh po izmereniyam kompleksnoy apparaturoy // Innovatsionnaya nauka. – 2015. – № 12-2. – S. 107–112.
3. Khamzin L.G., Vakhitova G.R. Vyyavlenie istochnikov obvodneniya achimovskikh otlozheniy neftyanogo mestorozhdeniya // Bulatovskie chteniya: sb. st. I Mezhdunar. nauch.-prakt. konf., 31 marta 2017 g. – Krasnodar: Izdatel’skiy Dom – Yug, 2017. – T. 2. – S. 289–291.
4. Vyyavlenie negermetichnosti muftovykh soedineniy nasosno-kompressornykh trub, obsadnykh i tekhnicheskikh kolonn dlya skvazhin podzemnogo khranilishcha gaza v solyanykh kavernakh metodom spektral’noy shumometrii / A.M. Aslanyan, M.V. Volkov, S.V. Soroka [i dr.] // Georesursy. – 2016. – T. 18, № 3. – S. 186–190. – DOI: 10.18599/grs.18.3.7
5. Izuchenie formirovaniya teplovoy metki v stvole skvazhiny pri induktsionnom nagreve kolonny dlya otsenki debita mezhplastovykh peretokov / I.V. Kanafin, R.F. Sharafutdinov, V.Ya. Fedotov, D.V. Kosmylin // Bulatovskie chteniya: sb. st. I Mezhdunar. nauch.-prakt. konf., 31 marta 2017 g. – Krasnodar: Izdatel’skiy Dom – Yug, 2017. – T. 1. – S. 70–72.
6. Issledovanie teplovogo polya v skvazhine pri zakolonnom dvizhenii zhidkosti v protsesse induktsionnogo vozdeystviya / F.F. Davletshin, A.Sh. Ramazanov, R.Z. Akchurin [i dr.] // Izvestiya Tomskogo politekhnicheskogo universiteta. Inzhiniring georesursov. – 2023. – T. 334, № 3. – S. 153–164. – DOI: 10.18799/24131830/2023/3/3896
7. Leonov I.V. Gidrodinamicheskiy metod vyyavleniya pritoka vody iz negermetichnosti ekspluatatsionnoy kolonny, raspolozhennoy nad UETsN // Neftepromyslovoe delo. – 2011. – № 5. – S. 35–37.
8. Negermetichnost’ skvazhin – global’naya problema, a ne lokal’naya / D.P. Anikeev, S.N. Zakirov, E.S. Anikeeva, A.D. Lysenko // Aktual’nye problemy nefti i gaza. – 2019. – № 4. – S. 15. – DOI: 10.29222/ipng.2078-5712.2019-27.art15
9. Agadulin I.I., Ignat’ev V.N., Sukhorukov R.Yu. Ekologicheskie aspekty negermetichnosti zakolonnogo prostranstva v skvazhinakh razlichnogo naznacheniya // Neftegazovoe delo. – 2011. – № 4. – S. 82–90.
10. Ekologicheskie voprosy kontrolya za ekspluatatsiey skvazhin podzemnykh khranilishch gaza / R.A. Valiullin, R.F. Sharafutdinov, A.A. Sadretdinov [i dr.] // Izvestiya Samarskogo nauchnogo tsentra Rossiyskoy akademii nauk. – 2015. – T. 17, № 5. – S. 256–262.
11. Dzyublo A.D., Ruban G.N. Nadezhnaya diagnostika i likvidatsiya zakolonnykh peretokov kak zalog ekologicheskoy bezopasnosti pri razrabotke neftegazovykh mestorozhdeniy // Aktual’nye problemy nefti i gaza. – 2018. – № 4(23). – S. 56. – DOI: 10.29222/ipng.2078-5712.2018-23.art56
12. Primenenie konteynernoy tekhnologii dlya izolyatsii pogloshchayushchikh intervalov vysokoy intensivnosti / D.A. Berdnikov, N.I. Rylov, I.G. Fattakhov [i dr.] // Stroitel’stvo neftyanykh i gazovykh skvazhin na sushe i na more. – 2014. – № 11. – S. 15–19.
13. Vakhitova G.R., Prayya E.D., Ryukov R.I. Prognoz petrofizicheskikh svoystv gornykh porod s ispol’zovaniem algoritmov mashinnogo obucheniya // Neftepromyslovoe delo. – 2024. – № 4(669). – S. 27–33.
14. Startsev V.A., Buslaev G.V., Konoplyannikov A.V. Analiz opyta primeneniya metodov mashinnogo obucheniya pri burenii neftyanykh i gazovykh skvazhin // Stroitel’stvo neftyanykh i gazovykh skvazhin na sushe i na more. – 2023. – № 9(369). – S. 15–24. – DOI: 10.33285/0130-3872-2023-9(369)-15-24
15. Ishkulov I.M., Fattakhov I.G. Interpretiruemoe mashinnoe obuchenie dlya opredeleniya negermetichnosti skvazhin // Zapiski Gornogo instituta. – 2025. – T. 275. – S. 94–109.
16. O metodakh opredeleniya negermetichnosti ekspluatatsionnoy kolonny skvazhiny / I.M. Ishkulov, R.R. Vafin, D.D. Takhauov [i dr.] // Neftyanoe khozyaystvo. – 2024. – № 7. – S. 56–60. – DOI: 10.24887/0028-2448-2024-7-56-60
17. Opredelenie negermetichnosti kolonny skvazhin s ispol’zovaniem modeley mashinnogo obucheniya / I.M. Ishkulov, D.D. Takhauov, R.R. Vafin [i dr.] // Neftegazovoe delo. – 2024. – T. 22, № 4. – S. 260–267. – DOI: 10.17122/ngdelo-2024-4-260-267
18. Fattakhov I.G., Kuleshova L.S., Musin A.I. O metode ekspress-obrabotki neogranichennogo massiva nepreryvno postupayushchikh promyslovykh dannykh // Avtomatizatsiya, telemekhanizatsiya i svyaz’ v neftyanoy promyshlennosti. – 2009. – № 3. – S. 26–28.
19. Fattakhov I.G. Klassifikatsiya ob"ektov razrabotki s ispol’zovaniem metoda glavnykh komponent // Neftepromyslovoe delo. – 2009. – № 4. – S. 6–9.
20. Afanador N.L., Tran T.N., Buydens L.M.C. Use of the Bootstrap and Permutation Methods for a More Robust Variable Importance Assessment // Analytica Chimica Acta. – 2013. – Vol. 768. – P. 49–56. – DOI: 10.1016/j.aca.2013.01.004