Title: Critical Review of the Percentage of Cumulative Oil Production with Artificial Intelligence Methods for Gas lifted Wells
Authors: Isaac Eze Ihua-Maduenyi, Sunday Igbani, Zekieni Robert Yelebe
Volume: 10
Issue: 3
Pages: 206-227
Publication Date: 2026/03/28
Abstract:
Gas lift optimisation is critical to sustaining production and economic viability in mature oilfields, particularly brownfields where reservoir pressure decline impairs natural hydrocarbon flow. As conventional assets age, the complexity of maintaining production intensifies, with gas lift systems offering a viable solution despite their highly nonlinear dynamics, inter-well dependencies, and operational uncertainties. Traditional optimisation methods often underperform under such high-dimensional, real-time constraints. In response, recent advances have embraced artificial intelligence (AI) as a transformative paradigm for full-field gas lift optimisation. This review presents a technical evaluation of AI-driven strategies, focusing on machine learning (ML), meta-heuristic algorithms, and fuzzy logic systems. ML models, especially artificial neural networks (ANNs), capture nonlinear input-output relationships from historical data, enabling fast production forecasting and control, though they are limited by data quality and computational demands. Meta-heuristic techniques, including genetic algorithms, particle swarm optimisation, and ant colony optimisation, excel in exploring complex, multimodal search spaces but face challenges with scalability and convergence efficiency. Fuzzy logic systems offer resilience in uncertain, data-sparse environments by integrating heuristic knowledge, yet suffer from subjectivity and limited scalability. Recognising the limitations of individual approaches, the study underscores the emergence of hybrid AI frameworks that synergise multiple AI paradigms or integrate AI with physics-based and statistical models. These composite systems enhance robustness, scalability, and interpretability while dynamically adapting to evolving reservoir conditions, injection constraints, and economic objectives. The review concludes that the future of gas lift optimisation lies in intelligent, adaptive hybrid AI systems capable of real-time field-wide optimisation, offering a pathway to improved efficiency, extended field life, and maximised economic returns in the upstream oil and gas sector.