Authors : Boma-George Esther Daisy; Victor Joseph Aimikhe

Volume/Issue : Volume 10 - 2025, Issue 12 - December

Google Scholar : https://tinyurl.com/bdft2nyz

Scribd : https://tinyurl.com/4j2vs4m6

DOI : https://doi.org/10.38124/ijisrt/25dec1210

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Abstract : Accurate prediction of Test Rack Opening (TRO) pressure is essential for the optimal design and calibration of gas lift valves, directly affecting unloading stability, gas injection efficiency, and overall production performance. Traditional approaches relying on force-balance equations and iterative test-rack calibrations are often constrained by simplifying assumptions and sensitivity to operational variability. This study develops and benchmarks nineteen (19) machine learning models to predict TRO pressure using a field dataset comprising 328 valves from 20 wells. A rigorous workflow encompassing data cleaning, feature engineering, multicollinearity reduction, and systematic validation was implemented. Seventeen input parameters, including dome pressure, fluid gradients, and well depth measurements, were evaluated. Among the algorithms tested; including linear models, support vector regression, ensemble methods (Random Forest, Gradient Boosting, XGBoost, LightGBM, CatBoost), and neural networks. The Random Forest Regressor exhibited the best performance, achieving a test R2 of 0.7437 and an Average Absolute Relative Error (AARE) of 8.87%. Feature importance analysis revealed Measured Depth, Dome Pressure, and Unloadable Gradient as the primary predictors, consistent with the physical mechanics of gas-lift systems. Time-series models (ARIMA, Prophet) performed poorly (R2 < 0.03), confirming that TRO pressure is inherently a static design parameter rather than a dynamic variable. The proposed predictive framework minimizes reliance on repetitive physical calibration, enables rapid design iterations, and provides interpretable, data-driven insights for optimizing gas-lift systems and improving operational reliability.

Keywords : Artificial Lift; Gas Lift Valve; Test Rack Opening Pressure; Machine Learning; Predictive Modeling; Random Forest; Feature Importance; Petroleum Engineering.

References :

1. P. Movahed et al. Gas Lift Optimization in the Permian Using Machine Learning and Artificial Intelligence in SPE Artificial Lift Conference and Exhibition – Americas, 2024.
2. F. E. Ashford and K. L. Wright, Overview of gas lift; Part 2: Operational fundamentals for the performance of a gas lift well. Houston, TX: John M. Campbell & Company, 2022.
3. API RP 11V5: Recommended practices for operation, maintenance, surveillance, and troubleshooting of gas‐lift installations, American Petroleum Institute, 2008.
4. K. L. Decker, "IPO Gas Lift Design With Valve Performance," SPE Production & Operations, vol. 23, no. 04, pp. 464–467, 2008.
5. H. W. Winkler and P. T. Eads. Algorithm for more accurately predicting nitrogen-charged gas-lift valve operation at high pressures and temperatures. SPE Production Engineering, vol. 4, no. 4, pp. 413–418, 1989.
6. M. S. eed Yakoot, A. M. S. Ragab, and O. Mahmoud, "Machine Learning Application for Gas Lift Performance and Well Integrity," in SPE Europec, 2021.
7. Z. Tariq et al., "A systematic review of data science and machine learning applications to the oil and gas industry," J Petrol Explor Prod Technol, vol. 11, pp. 4339–4374, 2021.
8. M. Nait Amar and N. Zeraibi, "Applications of machine learning in petroleum engineering: A review," Journal of Petroleum Science and Engineering, vol. 195, p. 107987, 2020.
9. H. Rahmanifard and I. Gates, "A Comprehensive review of data-driven approaches for forecasting production from unconventional reservoirs: best practices and future directions," Artif Intell Rev, vol. 57, no. 8, 2024.
10. F. Ma et al., "Predictive modeling of oil rate for wells under gas lift using machine learning," Sci Rep, vol. 15, p. 27765, 2025.
11. E. J. R. Coutinho, M. Dall'Aqua, and E. Gildin, "Physics-Aware Deep-Learning-Based Proxy Reservoir Simulation Model Equipped With State and Well Output Prediction," Front. Appl. Math. Stat., vol. 7, p. 651178, 2021.
12. A. A. Elgibaly, M. Ghareeb, S. Kamel, and A. M. El-Bassiouny, "Prediction of gas‐lift performance using neural network analysis," AIMS Energy, vol. 9, no. 2, pp. 355–378, 2021.
13. C. Miao, L. Zhang, X. Wang, and H. Chen, "Machine learning-based workflow for gas lift optimization in unconventional fields," Journal of Petroleum Technology, vol. 76, no. 3, pp. 48–62, 2024.
14. N. A. Sami, "Application of machine learning algorithms to predict tubing pressure in intermittent gas lift wells," Petroleum Research, vol. 7, no. 2, pp. 246–252, 2022.
15. "Practical equations calculate gas flow rates through venturi valves," Oil & Gas Journal, Feb. 2010.
16. API RP 11V6: Recommended practice for design of continuous flow gas lift installations using injection pressure operated valves, American Petroleum Institute, 1999.
17. M. A. Shahri and H. W. Winkler, "Practical Method for Measurement of Injection-gas Throughput of Each Gas-Lift Valve before Well Installation," in SPE Production and Operations Symposium, 2011.
18. K. Altarabulsi and P. Waltrich, "An Improved Method to Calculate Gas-Lift Valve Set Pressure," SPE Journal, vol. 29, no. 10, pp. 5452–5463, 2024.
19. M. Mohammadpoor and F. Torabi, "Big data analytics in oil and gas industry: An emerging trend," Petroleum, vol. 6, no. 4, pp. 321–328, 2020.
20. H. Jo, H. Pan, and J. E. Santos, "Machine learning applications in petroleum engineering: A comprehensive review of 254 papers from 2018-2024," Journal of Petroleum Science and Engineering, vol. 232, p. 112456, 2025.
21. S. Ben Jabeur, C. Gharib, S. Mefteh-Wali, and W. Ben Arfi, "CatBoost model and artificial intelligence techniques for corporate failure prediction," Technol Forecast Soc Change, vol. 166, p. 120658, 2021.
22. M. Ghodsi, P. Vaziri, M. Kanaani, and B. Sedaee, "Optimizing oil production forecasts in Iranian oil fields: a comprehensive analysis using ensemble learning techniques," J Petrol Explor Prod Technol, vol. 15, no. 4, 2025.
23. A. M. Salem, M. S. Yakoot, and O. Mahmoud, "Addressing Diverse Petroleum Industry Problems Using Machine Learning Techniques: Literary Methodology, Spotlight on Predicting Well Integrity Failures," ACS Omega, vol. 7, no. 3, p. 2504, 2022.
24. L. Nazari, E. Camponogara, L. S. Imsland, and L. O. Seman, "Neural networks informed by physics for modeling mass flow rate in a production wellbore," Eng Appl Artif Intell, vol. 128, p. 107528, 2024.
25. L. Xue et al., "Physics-informed neural network-based petroleum reservoir simulation," Petroleum Science, vol. 21, no. 1, pp. 328–341, 2023.
26. Y. Zhu et al., "Comprehensive feature selection methodology for petroleum production prediction using hybrid statistical-ML approaches," Energy, vol. 296, p. 131234, 2025.