Kazakhstan journal for oil & gas industryKazakhstan journal for oil & gas industry2707-42262957-806XKMG Engineering10875010.54859/kjogi108750Research ArticleExploring modern methods for predicting well failures in the fields of NC «KazMunayGas» JSCUtemisovaLaura G.l.utemissova@niikmg.kzhttps://orcid.org/0000-0003-4194-6727MerembayevTimur Zh.<p>PhD</p>timur.merembayev@gmail.comhttps://orcid.org/0000-0001-8185-235XBekbauBakhbergen E.<p>PhD</p>b.bekbau@kmge.kzhttps://orcid.org/0000-0003-2410-1626KMG EngineeringInstitute of Information and Computational Technologies CS MES RoKSatbayev University151220246468772205202429112024Copyright © 2024, Utemisova L.G., Merembayev T.Z., Bekbau B.E.2024<p>In the development of brownfields, various geological and technological complications can arise. To enhance the smooth operation of downhole pumping equipment, companies implement a range of methods and techniques.</p> <p>This article analyzes the potential of using machine learning to improve the reliability of underground well equipment in the fields of NC KazMunayGas JSC. The research focuses on the development and validation of predictive models that accurately forecast potential downhole equipment failures. It thoroughly analyzes existing machine learning methods, approaches and their real-life application, highlighting key success factors and limitations. The results of the study demonstrate the significant potential for using a well failure prediction model when selecting the optimal machine learning approach to reduce unscheduled downtime and optimize well maintenance processes. The authors assessed the potential for using failure prediction techniques for downhole pumping equipment in wells that utilizes sucker rod pumps. Implementing failure prediction techniques for downhole pumping equipment can help ensure uninterrupted well operation by minimizing well failures and reducing downtime for repairs.</p>downhole pumping equipmenttime between failuresunderground well workoverwell failuresfailure predictionтерең сорғы жабдықтарыістен шығуларды атқаружер асты ұңғымаларын жөндеуұңғымалардың істен шығуыістен шығуды болжауглубинно-насосное оборудованиенаработка на отказподземный ремонт скважинотказы скважинпрогнозирование отказов1.Mihaylovich NN. Povysheniye ekspluatatsionnykh kharakteristik neftepromyslovogo oborudovaniya v territorial'no-proizvodstvennom predpriyatii [dissertation]. Moscow; 2009. Available from: https://www.dissercat.com/content/povyshenie-effektivnosti-ekspluatatsii-nasosnykh-skvazhin-optimizatsiei-raboty-shtangovykh-k. (In Russ).2.Khabibullin RA, Shabonas AR, Gurbatov NS, Timonov AV. Prediction of ESPs Failure Using ML at Western Siberia Oilfields with Large Number of Wells. Russian Petroleum Technology Conference; 2020 Oct 26–29; Virtual. Available from: https://onepetro.org/SPERPTC/proceedings-abstract/20RPTC/1-20RPTC/D013S027R001/450252.3.Merembayev T, Kurmangaliyev D, Bekbauov B, Amanbek Y. A Comparison of machine learning algorithms in predicting lithofacies: Case studies from Norway and Kazakhstan. Energies. 2021;14(7):1896. doi: 10.3390/en14071896.4.Merembayev T, Amanbek Y. Time-series event prediction for the uranium production wells using machine learning algorithms. 56th U.S. Rock Mechanics/Geomechanics; 2022 June 26; Santa Fe, United States. Available from: https://research.nu.edu.kz/en/publications/time-series-event-prediction-for-the-uranium-production-wells-usi-2.5.American Petroleum Institute. API TR 11L Design Calculations for Sucker Rod Pumping Systems (Conventional Units), 5th edition. Washington : API TR 11L; 2008.6.Abdelaziz M, Lastra R, Xiao JJ. ESP data analytics: Predicting failures for improved production performance. Abu Dhabi International Petroleum Exhibition & Conference; 2017 Nov 13–16; Abu Dhabi, UAE. Available from: https://www.scirp.org/reference/referencespapers?referenceid=2938839.7.Ivanonvskii VI. Skvazhinnye nasosnye ustnovki dlya dobychi nefti. Moscow; 2002. (In Russ).8.Seradilla O, Zugastu E, Rodriguez J, Zurutuza U. Deep learning models for predictive maintenance: a survey, comparison, challenges and prospect. Appl Intell. 2020;52:10934–10964. doi: 10.1007/s10489-021-03004-y.9.Takacs G. Sucker-Rod Pumping Handbook. Production engineering fundamentals and long-stroke rod pumping. Hungary; 2015.10.Rudakov V, Merembayev T, Amirgaliyev Y, Omarova P. Time Series Analysis of Biogas Monitoring with Deep Learning Approaches. 5th International Conference on Problems of Cybernetics and Informatics; 2023 Aug 28–30; Baku, Azerbaijan. Available from: https://ieeexplore.ieee.org/document/10325955.11.Waskom ML. Seaborn: statistical data visualization. Journal of Open Source Software. 2021;6(60):3021. doi: 10.21105/joss.03021.12.Ke G, Meng Q, Finley T, et al. Lightgbm: A highly efficient gradient boosting decision tree. Advances in neural information processing systems. 2017;30.13.Breiman L. Random forests. Machine learning. 2001;45:5–32.14.Chen T, He T, Benesty M, et al. Xgboost: extreme gradient boosting. R package version 0.4-2. 2015;1(4). doi: 10.32614/cran.package.xgboost.15.Hochreiter S, Schmidhuber J. Long Short-term Memory. Neural Computation MIT-Press. 1997;9(8):1735–1780. doi: 10.1162/neco.1997.9.8.1735.