Abstract This study addresses the complex optimization of gas lift well groups supplied by a single compressor unit under fixed total gas volume constraints. The interdependence of well performance and reservoir management requirements-including reservoir pressure limits, individual production quotas, surface facility constraints, and completion conditions-necessitates a fast-response multi-objective optimization framework. To overcome computational bottlenecks of traditional reservoir modeling, a data-driven proxy model accelerates production prediction by learning from full-field reservoir simulations segmented into well-group performance datasets. Wellbore hydraulics are modeled in Prosper while surface networks are simulated in GAP, with Python OpenServer coupling these modules to integrate real-time KPIs. The multi-objective optimization employs GPU-accelerated NSGA-II to simultaneously maximize cumulative production, extend stable production duration, and minimize total gas injection volume, subject to constraints including total gas availability, per-well production caps, reservoir pressure thresholds, and maximum injection limits. The method established in this study was used to optimize the gas injection rate of the gas lift well in one oil field in Iraq. The results show that under the same production target conditions, the required injection volume was reduced by 84.6% after optimization, greatly saving the development cost of gas injection wells.
Jiang et al. (Mon,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: