Published September 2025, Pg. 13-19

Section: Oil and gas field development and exploitation

UOT: 622.276.658.58

DOI: 10.37474/0365-8554/2025-09-13-19

Predictive modeling of liquid loading in gas-condensate wells using deep neural networks

V.M. Fataliyev Dr. in Tech. Sc. - Azerbaijan State Oil and Industry University

K.F. Aliyev - Azerbaijan State Oil and Industry University

Liquid loading in gas-condensate wells poses a significant challenge, often leading to reduced hydrocarbon recovery and increased operational costs as reservoir pressures decline over time. Efficiently identifying and managing the onset of liquid loading is essential for sustaining production rates and prolonging well life. This study explores the application of deep learning neural networks to predict liquid loading status of the wells, crucial for mitigating costs and enhancing production efficiency. Four distinct models were developed and trained using experimental datasets sourced from various sources, incorporating parameters such as wellhead pressure, gas-rate, and tubing inside diameter as input features. These deep neural network models achieved accuracies ranging from 57 % to 80 %, demonstrating the promise of deep learning applications in predicting the liquid loading of gas-condensate wells. Through comparative analysis, the more significant influence of wellhead pressure compared to gas rate on liquid loading determination was emphasized, highlighting the critical role of wellhead pressure in the liquid loading process. Improving the model performance by around 17 % through the utilization of gas-rate and wellhead pressure as input parameters simultaneously demonstrates that further optimization can be achieved by expanding the input parameters to include additional factors such as fluid properties, well geometry, temperature, and condensate rate. Although the model development faced limitations due to the relatively small dataset size, the performance metrics exceeded initial expectations, highlighting the robustness and scalability of neural network approaches. The outcomes of this research offer valuable insights into the practical deployment of AI-based tools in reservoir management. By enabling more accurate and timely prediction of liquid loading events, the proposed models can assist engineers in making informed decisions, optimizing production strategies, and ultimately improving the economic performance of gas-condensate assets. These contributions mark a step forward in the digital transformation of the oil and gas industry.

References:

1. Hamidov N.N., Fataliyev V.M. Experimental study into the effectiveness of the partial gas cycling process in the gas-condensate reservoir development // Petroleum Scienceand Technology, 2016, 34:7, pp. 677-684.

2. Lea J.F., Jr Rowlan L. Gas Well Deliquification // Gulf Professional Publishing, third edition, 2019 pp. 9-24.

3. Tugan M.F. Deliquification techniques for conventional and unconventional gas wells: Review, field cases and lessons learned for mitigation of liquid loading // Journal of Natural Gas Science and Engineering, 2020, № 83, pp.1386-1390.

4. Waltrich P.J., Posada C., Martinez J., Falcone G., Barbosa J.R. Experimental investigation on the prediction of liquid loading initiation in gas wells using a long vertical tube // Journal of Natural Gas Science and Engineering, 2015, № 26, pp. 1515-1529.

5. Rowlan O.L., McCoy J.N., Becker D., Podio A.L. Acoustic Liquid-Level Determination of Liquid Loading in Gas Wells // In Proceedings of the SPE Production and Operations Symposium, 2003, USA.

6. Turner R.G., Hubbard M.G., Dukler A.E. Analysis and Prediction of Minimum Flow Rate for the Continuous Removal of Liquids from Gas Wells // Journal of Petroleum Technology, 1969, 21(11), pp. 1475-1482.

7. Coleman S.B., Clay H.B., McCurdy D.G., Norris III L.H., et al. A new look at predicting gas-well load-up // Journal of Petroleum Technology, 1991, 43 (03), pp. 329-333.

8. El Fadili Y., Shah S. A new model for predicting critical gas rate in horizontal and deviated wells // Journal of Petroleum Science and Engineering, 2017, 150, pp. 154-161.

9. Guo B., Ghalambor A., & Xu C.A Systematic Approach to Predicting Liquid Loading in Gas Wells // SPE Prod & Oper, 2006, 21(1), pp. 81-88. https://doi.org/10.2118/94081-PA

10. Luo S., Kelkar M., Eduardo P., Sarica C. A New Comprehensive Model for Predicting Liquid Loading in Gas Wells // SPE Production & Operation, 2014, 29 (04), pp. 337-349.

11. Nosseir M.A., Darwich T.A., Sayyouh M.H., El Sallaly M. A New Approach for Accurate Prediction of Loading in Gas Wells Under Different Flow Conditions // SPE Production & Facilities, 2000, 15(4), pp. 241-246.

12. Ghadam A., Kamali V. Prediction of Gas Critical Flow Rate for Continuous Lifting of Liquids from Gas Wells Using Comparative Neural Fuzzy Inference System // Journal of Applied Environmental and Biological Sciences, 2015, 5 (8S), pp. 196-202.

13. Khamehchi E., Yasrebi S.V., Ebrahimi A. Prediction of the influence of liquid loading on wellhead parameters // Petroleum Science Technology, 2014, 32 (14), pp. 1680-1689.

14. Joseph A., Hicks P.D. Modelling mist flow for investigating liquid loading in gas wells // Journal of Petroleum Science and Engineering, 2018, 170, pp. 476-484.

15. Aggarwal C.C. Neural Networks and Deep Learning // Springer International Publishing AG, First edition, 2018, 52 p.

16. Abad A., Ghorbani H., Mohamadian N., et. all. Robust hybrid machine learning algorithms for gas flow rates prediction through wellhead chokes in gas condensate fields // Fuel, 2022, v. 308, pp. 12-18.

17. Weidman S. Deep Learning from Scratch // O’Reilly Media. First edition, 2019, pp. 71-98.

18. Goodfellow I., Bengio Y., Courville A. Deep Learning // Adaptive Computation and Machine Learning series, The MIT Press, 2017, p. 777.

19. Ikpeka P.M., Okolo M.O. Li and Turner Modified model for Predicting Liquid Loading in Gas Wells // Journal of Petroleum Exploration and Production Technology, 2018, pp. 1971-1993.

20. Ming R., He H. A new approach for predicting critical gas rate in condensate gas wells // International Journal of Oil, Gas and Coal Technology, 2020, 23 (1), pp.76-91.

21. Xiao H., Li X-P., Tan X-H., et al. A novel model for calculating critical droplet entrainment rate of gas well considering droplet deformation and multiple parameters // Energy Science Engineering, 2021, № 9, pp. 812-827.

22. Zhou D., Yuan H.A. New Model for Predicting Gas-Well Liquid Loading // SPE Prod & Oper, 2010, 25(2),
pp. 172-181.