A machine learning approach to Intelligent Artificial lift method selection

Shedrach Igemhokhai 1, *, Kelani Bello 1 and Abayomi Adejumo 2

1 Department of Petroleum Engineering, Faculty of Engineering, University of Benin, Nigeria.
2 Oriental Energy Resources Limited, Lagos, Nigeria.
 
Research Article
International Journal of Frontiers in Engineering and Technology Research, 2025, 08(02), 001-015.
Article DOI: 10.53294/ijfetr.2025.8.2.0030
Publication history: 
Received on 23 December 2024; revised on 31 March 2025; accepted on 02 April 2025
 
 
Abstract: 
Artificial lift (AL) methods are crucial for optimizing well performance and sustaining hydrocarbon production in oil and gas operations. Traditional AL selection relies on conventional methodologies and human expertise, which may be inadequate for handling complex reservoir dynamics and varying operating conditions. As the industry seeks more efficient, data-driven solutions, machine learning (ML) presents an opportunity to enhance AL selection. This study develops an ML-based stack framework of Random Forest (RF), Extreme Gradient Boosting (XGB) and Decision Tree (DT) to predict optimal AL methods. The models are trained and validated on a comprehensive dataset incorporating well particulars, production parameters, reservoir properties, and operational conditions. Performance evaluation demonstrates that the ML models achieve up to 95% accuracy in AL selection, significantly improving on traditional methods. The findings highlight the potential of ML-driven AL selection to enhance production efficiency, reduce operational costs, and optimize field performance. This study provides a foundation for integrating AI-based decision-making into artificial lift optimization, offering a more adaptive and precise approach to production engineering.
 
Keywords: 
Artificial Lift; Machine Learning; Stack Model; Production; SHAP
 
Full text article in PDF: