Methodology for predicting electric submersible pumps failure in case of telemetry system sensor failure

UDK: 622.276.53.054.23:621.67-83

DOI: 10.24887/0028-2448-2025-8-64-68

Key words: electric submersible pump (ESP), failure prediction, operation in conditions of reduced insulation, Cox model

Authors: I.A. Lakman (Ufa University of Science and Technology, RF, Ufa); A.A. Agapitov (INTAS-Company LLC, RF, Ufa); L.F. Sadikova (INTAS-Company LLC, RF, Ufa); S.M. Gumerov (INTAS-Company LLC, RF, Ufa); V.G. Prytkov (Zarubezhneft JSC, RF, Moscow); A.L. Tistol (Zarubezhneft JSC, RF, Moscow); D.A. Chernov (ZARUBEZHNEFT-Dobycha Kharyaga LLC, RF, Moscow); S.V. Blagorodov (ZARUBEZHNEFT-Dobycha Kharyaga LLC, RF, Moscow)

One of the problems in obtaining accurate predictions of submersible pump failures is the lack of data due to submersible sensor failure. The purpose of the study is to create a reliable model for predicting electric submersible pump (ESP) failures due to insulation reduction when the telemetry sensor fails. Data for training and validation of the model were collected from 76 wells during 2018-2024. The analysis showed that the probability of ESP failure after sensor failure is highest in the first 30 days, it then decreases, and after that it starts to increase again after six months. The average values and standard deviations of the motor operating current and active power for 30 days and their differences with the current daily average values were used as factors influencing the failure. The prediction model was based on Cox proportional hazards regression with time since sensor failure. The failure prediction horizon was 20 days. As a result of testing, the model predicted 4 out of 6 failuers of ESPs operating with reduced insulation. Industrial operation of the developed module of the ARM Technologist system at the field of ZARUBEZHNEFT-Dobycha Kharyaga LLC showed an acceptable accuracy of 3 out of 5 within 5 months

References

1. Sabirov A.A., Degovtsov A.V., Kuznetsov I.V. et al., Forecasting the operating time to failure, selection of design and optimization of procurement of electric centrifugal pump installations for complicated wells stock (In Russ.), Territoriya Neftegaz, 2019, no. 7–8, pp. 44–48.

2. Danilko A.I., Stukach O.V., The software module for evaluation of the failure probability of the well electric submersible pump installation (In Russ.), Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta, 2020, no. 1–2 (97), pp. 55–66, DOI: https://doi.org/10.17212/2307-6879-2020-1-2-55-66

3. Shportko A.A., Kulaev E.G., Complex analysis of operation and failures of electric submersible pumping unit (ESP) (In Russ.), Oborudovanie i tekhnologii dlya neftegazovogo kompleksa, 2013, no. 6, pp. 25–29.

4. Dyshin O.A., Abasova S.M., Reliability assessment of electric submersible pumps using censored samples (In Russ.), Nauchnye trudy NIPI Neftegaz GNKAR = SOCAR Proceedings, 2013, no. 4, pp. 61–68, DOI: https://doi.org/10.5510/OGP20130400179

5. Solov’ev I.G., Govorkov D.A., Konstantinov I.V., Estimation of ESP operational resource dynamics using factor-based model (In Russ.), Izvestiya Tomskogo politekhnicheskogo universiteta. Inzhiniring georesursov, 2022, V. 333, no. 9, pp. 158–167, DOI: https://doi.org/10.18799/24131830/2022/9/3541

6. Shabonas A.R., Electric submersible pump operation mode optimization to increase the run-to-failure time (In Russ.), Neftepromyslovoe delo, 2021, no. 8(632),

pp. 30–36, DOI: https://doi.org/10.33285/0207-2351-2021-8(632)-30-36

7. Enikeev R.M., Penzin A.V., Latypov B.M. et al., Increasing the efficiency of operation of complicated oil wells using intelligent algorithms (In Russ.), Stroitel’stvo neftyanykh i gazovykh skvazhin na sushe i na more, 2023, no. 8(368), pp. 50–58, DOI: https://doi.org/10.33285/0130-3872-2023-8(368)-50-58

8. Brasil J., Maitelli S., Nascimento J. et al., Diagnosis of operating conditions of the electrical submersible pump via machine learning, Sensors, 2023, V. 23,

DOI: https://doi.org/10.3390/s23010279

9. Youngsoo Song, Sungjun Jun, Tan C. Nguyen et al., Experimental data-driven model development for ESP failure diagnosis based on the principal component analysis, Journal of Petroleum Exploration and Production Technology, 2024, V. 14, pp. 1521–1537, DOI: https://doi.org/10.1007/s13202-024-01777-9

10. Lakman I.A., Agapitov A.A., Sadikova L.F. et al., Analyzing the possibility of applying machine learning methods in predictive analytics to determine the probability of failures of electric submersible pump assembly (In Russ.), Neftyanoe khozyaystvo = Oil Industry, 2024, no. 8, pp. 132–136, DOI: https://doi.org/10.24887/0028-2448-2024-9-132-136