“All models are wrong, but some are useful.”
This famous quote by George E.P. Box illustrates both the challenge and the appreciation of building models. Models are needed to predict future performance of an oil or gas field, but, at the same time, models are often biased and inaccurate.
Most investment decisions rely on our ability to predict and to plan the future, and, in that regard, nothing is more important than accurately modeling future well performance. Consequently, three of the four chosen papers address different aspects of this topic.
The first paper deals with ways to improve the accuracy of our predictions. It offers readers a rigorous checklist of questions to ask when developing reservoir models to guide them toward less-biased forecasts.
The second paper deals with how we develop reservoir-prediction tools for asset management. Active reservoir management addresses the almost impossible task of maximizing short-term production while optimizing ultimate recovery. However, to evaluate the different reservoir-drainage mechanisms, one needs good models, and the more advanced the drainage mechanism is, the more crucial the model is. As an illustration, it is much easier to model a pressure-depletion scheme than to try to predict a secondary- or tertiary-recovery process. Assessing and quantifying uncertainties as part of the modeling are becoming increasingly common, and this practice improves the ability to develop a sound decision basis.
The development of unconventional shale resources has further challenged the ability to predict performance. Prediction of such unconventional resources does not necessarily require new tools but rather new assumptions and new experience-based calibration methods. More than 40% of the papers I reviewed for this issue dealt with prediction of production and ultimate recovery of shale gas, which clearly illustrates the increasing interest in this topic and the current challenges faced by today’s petroleum engineers. The lack of history and of good analogs further adds to the uncertainty. I am sure that more research and the availability of more production data will enable us to develop better models and, hence, increase the accuracy of our predictions. The third paper provides great insight into our understanding of unconventionals.
This Month's Technical Papers
New Time-Rate Relations for Decline-Curve Analysis of Unconventional Reservoirs
Instilling Realism in Production Forecasting Decreases Chances of Underperformance
Conformance Control and Proactive Reservoir Management Improve Deepwater Production
Distributed Microchip System Records Subsurface Temperature and Pressure
Recommended Additional Reading
OTC 23949 In-Well Distributed Fiber-Optic Solutions for Reservoir Surveillance by Juun van der Horst, Shell, et al.
IPTC 16866 Uncertainty Assessment of Production Performance for Shale-Gas Reservoirs by Jiang Xie, Chevron, et al.
IPTC 16505 Joint Inversion of Time-Lapse Crosswell Seismic and Production Data for Reservoir Monitoring and Characterization by Lin Liang, Schlumberger, et al.
| Erik Vikane, SPE, is production manager for Statoil Oseberg East. He has 22 years of diverse experience in the upstream business. Starting as a well-intervention engineer, Vikane then moved to reservoir engineering and has held various positions within reservoir management, early-phase development, exploration, and business development. His areas of interest include reservoir and asset management, reservoir performance and monitoring, and integrated uncertainty studies. Vikane holds an MS degree in petroleum engineering from the Norwegian University of Science and Technology. He serves on the JPT Editorial Committee and the SPE Reservoir Management 2020 Forum Committee. |