Production Optimization by Real-Time Modeling and Alarming: The Sendji Field Case

Abstract

Summary Digital technologies can improve oil production and reduce operating costs. In this prospect, TOTAL launched a corporate program called "Field Monitoring" to capitalize affiliate previous experiences. For gas-lifted wells, it resulted in the development of the "Well Performance Monitoring" (WPM) tool implemented in December 2006 at the Sendji field located offshore Congo. This paper presents the main workflows, functionalities, and added value of WPM. Besides multi-real-time data trends, the application offers color-coded multialarm visualization versus time. These alarms are ranked in terms of "priority" and aggregated at each well, platform, field, and asset level, which allows well engineers to easily track down underperforming wells. Alarms priority can be linked either to well integrity issues or production losses. These real-time estimated losses are computed using network and well-modeling softwares; they highlight deviation from optimum operating set points. WPM makes easier regular optimization of gas-lift optimization by feeding models inputs with user-validated real-time data. A major challenge of right-time production optimization is to sustain updated models. Two workflows address this issue: 1) automatic modeling of production well tests and 2) real-time modeling of the actual behavior of the production system. When a significant difference occurs between modeling and reality, a specific alarm is triggered. This paper shows various examples of enhanced well performance diagnosis and the corresponding production gains. Also, good support from onsite teams was achieved thanks to key technical solutions, such as user-friendly interface, possible individual tuning of each alarm, and handling of unstable or unmatched wells. Finally, WPM enables computation of KPIs that allow the evaluation of potential well interventions. This tool will be deployed at a corporate level and extended to ESP-lifted wells, gas injectors or producers, and to subsea wells in deep offshore environment. The general alarming, interface features, and data architecture will be applied to surface equipment monitoring and optimization.