This paper presents the comprehensive analysis and optimization study of the gas lift system of 44 gas lifted wells in QQ Field in the Gulf of Suez. With the least amount of intervention, the optimized operational plan would yield an additional 7 to 10% oil gain over the current production rate, an improvement of gas utilization and operational efficiency, and operating cost reduction. This paper presents an Augmented AI approach, which is an efficient, repeatable and automated workflow to diagnose and optimize gas lift operations, reducing the time required for this process from weeks to hours. This study evaluates historical performance, identification of suboptimal performers using heuristic diagnostics, and subsequent optimization using well models coupled with advanced optimization algorithms.
The presented workflow implements a top-down approach beginning with field-level based analytics, then further delves into detailed engineering aspects of well production and individual gas lift operations. It builds from artificial intelligence and engineering-based workflows that compute smart metrics based on the historical well production and performance of gas lift systems. Heuristics and expert-based rules are also considered to perform a fast and smart diagnostics process. These analytics and diagnostics stages help build key metrics to evaluate well performance and monitor gas lift operations from both executive and asset management levels.
The study results identified suboptimal wells with inefficient gas lift performance that require detailed evaluation for improvement from both design and operational perspectives. During the optimization phase, reliable well model-based analysis coupled with physical constraints are used to provide remedial actions for the identified wells. The workflow allows automated well model construction and calibration, coupled with an unsupervised learning algorithm to detect and eliminate unreliable datasets. Afterward, specific well-level analysis is performed including gas lift performance curve while honoring existing well configuration, hypervolume-based multi-objective optimization and sensitivity analysis on key parameters for best gas lift performance.