Title: Numerical Optimization Methods for Enhancing Oil Production in Gas Lift Systems: A Review
Authors: Isaac Eze Ihua-Maduenyi and Eniye Oguta
Volume: 9
Issue: 3
Pages: 81-102
Publication Date: 2025/03/28
Abstract:
Optimizing gas lift is essential for improving oil extraction from wells that have reduced reservoir pressures. Various factors, such as natural production decline, increasing water cuts, and fluctuating gas-oil ratios, necessitate advanced optimization methods to maintain and boost profitability. Successful production optimization depends on thoughtfully crafted models and simulations that apply controlled optimization methods grounded in detailed system representations. This study reviews the numerical methods developed for gas-lift optimization problems, ranging from isolated single-well analyses to field-based optimization schemes encompassing all wells in a field. Evaluation of the numerical optimization methods used shows that Sequential Quadratic Programming (SQP) was 43%, sequential linear programming (SLP) was 7%, gradient descent was11%, simplex optimization method was 4%, revised simplex optimization method was 4%, Nelder-Mead method was 7%, Newton Rapson method was 7% and trust region method was 7%. SQP is identified as one of the most effective and widely studied methods for constrained non-linear programming, successfully applied to gas allocation, production maximization, and net present value (NPV) maximization within field networks. However, SQP has limitations, including computational intensity, sensitivity to initial solutions, and challenges in updating the Hessian matrix for large-scale problems. Developing a fully automated model for gas lift and general production optimization that adapts to changing conditions and robustly predicts future development needs remains a valuable yet challenging goal. Integrating numerical models with artificial intelligence could enhance accuracy and effectively handle higher dimensions. This paper provides insights into the various approaches developed and highlights the ongoing challenges in this field.