Leveraging Reservoir Knowledge, Multi-Scenario Modelling and Artificial Intelligence to Robustly Screen Development Options on a Click of a Button

Abstract During the past decade, Petroleum Development Oman's Hydrocarbon Maturation Centre (HMC) has established thematic catalogues with reservoir knowledge for various formations in PDO's acreage. The objective of this multidisciplinary collection of data, experience and insights is to reduce FDP cycle times by providing users web-based access to standardized datasets, workflows and past development decisions and thereby leveraging the use of analogues for development decision making. However, despite the vast amount of analogue information specific development decisions such as optimal pattern type, well spacing, water throughput, etc. and corresponding recovery factors and production profiles are still evaluated using conventional forecasting tools, which are seemingly time and resources consuming. In this project, PDO has developed an innovative web-based tool that leverages the extensive analogue knowledge, multi-scenario modelling, and Artificial Intelligence (AI) to eliminate or significantly reduce the requirement for such reservoir simulation studies. For several reservoirs formations, 10,000's of sector simulation models were built, covering all possible combination of static and dynamic parameters with the uncertainty ranges quantified in the thematic catalogues and representing the most common development concepts. The entire dataset is then used to train AI models and build the model library. The generated sector model library is searchable and can be used to identify reservoir analogues. The AI webtool then can predict recovery factors, forecast profile ranges etc. for any analogue or new sector model for the available development concepts at a click of a button. This allows a fast & robust evaluation of subsurface development options with clarity on trade off's, eliminates repetitive modelling & simulation requirements, and significantly accelerates simulation-based forecasting by 50% or more. Unlike most traditional forecast prediction tools, available in the company, the FDIC-AI dashboard is based on extensive multi-scenario modelling and simulation results and use artificial intelligence to create predictions for any reservoir models with parameter combination within the defined ranges. This means that the tool can 1) generate valuable insights on production behaviour based on geological characteristics and development concept, II) be used for predictive analysis and benchmarking of reservoir expected performance, III) improve the efficiency in making early decisions on development concept decisions by quantifying trade-offs, VI) support the aspiration of the company towards leveraging analogue-based developments and agility in delivering projects.