The Investigation of Workover Rig Scheduling Optimization Using Genetic Algorithms – Part I

Abstract

Abstract Well intervention for restoring or increasing production is a routine activity in any oilfield, however, it becomes an optimization opportunity for large fields with thousands of wells and high failure rate. Balancing between proactive and reactive workovers to optimize the rig schedule is a complex decision that requires a compelling objective function. Most of the time the decision is driven by economics, which are estimated from the well production potential. But there are other factors that can lead to different drivers or decisions outside of pure economics such as the duration of rig travel, uncertain duration of the job, and the "lost-oil" during the time the well is "down". The study investigates the application of evolutionary computation techniques, in particular the use of genetic algorithms for the complex task of well workover rig scheduling. It addresses the case of large fields with a high level of activity, sizable number of failures, healthy list of well backlog, and a large number of active workover rigs. This specific scenario is common in large waterflooding and heavy oil fields. The research explores the application of the "Traveling Salesman" optimization problem for scheduling using genetic algorithms and contrasts the new approach with a classic manual workflow to discuss its advantages and value creation. Lastly, it should be noted that this work is still ongoing research and while the results are promising, the complexity of the task, still demands additional study. At this point, the authors believe that there is not "one" single objective function representation for this complex task and encourage researchers to explore what is the best fit for their specific operations.