From Manual to Autonomous: Field Experience and Lessons Learned from Implementing the First AI-APC Autonomous Gas-Lift Optimization Solution

Abstract

Abstract Gas-lift optimization in large oil fields is often constrained by manual well reviews, intermittent well-test data, limited visibility of bottomhole conditions, and difficulty of implementing new setpoints across a large well stock. This document summarizes the field experience from implementing an AI-APC Autonomous solution for gas-lifted oil producers. The solution combines artificial-intelligence assisted inferential models, digital well-performance calculations and Advanced Process Control (APC) to autonomously adjust gas-lift and production-control settings while respecting reservoir-management, well-integrity and surface-facility constraints. This integration allows seamless and precise well adjustments, optimizing an entire portfolio of wells simultaneously with minimal operational footprint and human intervention. The system utilizes up-to-date well models to calculate in real time the optimal gas lift and downhole flowing pressure based on actual flowing conditions, thereby continuously achieving production targets. The solution promotes more efficient energy usage by automatically managing all producing wells while enforcing reservoir management strategy, respecting surface facility constraints and safeguarding well integrity. Ultimately, this results in maximized performance, enhanced efficiency, and reduced exposure to adverse environmental and process conditions, supporting more sustainable operations by reducing greenhouse gas emissions compared with previous practices. The system was deployed across hundreds of strings in a major smart field, leveraging existing instrumentation and requiring only a single additional APC server in the control room. Multiple autonomous scenarios were tested for maximum efficiency, optimal energy utilization, and maximum production. Results included in this work demonstrate the solution's benefits, with potential gains of 2–5% in additional production. This centralized approach enables rapid and large-scale deployments with a minimal footprint and no production loss during the deployment stage. In practice, a significant reduction in effort was observed when updating and executing gas-lift optimization across numerous wells. The world-first AI-APC solution for gas-lifted wells enabled effective and reliable autonomous operations. This represents an important operating-philosophy shift, moving from manual well adjustments toward autonomous and precise control by virtual agents, unlocking the Digital Barrel and stepping into a new era of smarter operations.