Looking back through previous editions of this article, I note that, in 2011, I wrote, “there’s a growing tendency in some quarters to use very simple models.” That may be true, but there is also a growing tendency among vendors to offer models with more and more features. That is how they stay in business, after all, and many of the new features solve problems we have struggled with for years. The trouble, however, is sometimes a feeling that all those features have to be used at the same time.
Let’s remind ourselves that a compositional model will never have as many components as there are distinct molecules in the reservoir fluid. The most sophisticated coupled pipeline simulator is still only an approximation of the complexity of fluid flow in pipes. Even the most advanced geomechanics simulator has a very coarse model of the over-, under- and sideburdens, and a crude approximation of often unknown boundary conditions. Building an exact model of the reservoir and production system is like seeking the end of the rainbow.
Piling on complex features can easily mask critical issues. We should always strive to use the simplest model that captures the most important physics affecting the decision we are trying to make. And judging what is “the most important” physics is the responsibility of the reservoir engineer, not the software and not the software vendor.
A simulation model is only a model. And our job is to produce the reservoir, not the model.
This Month's Technical Papers
Massively Parallel Simulation of Oceanic-Gas-Hydrate Production
Scaling Up a 900-Million-Cell Static Model to a Dynamic Model
Quality-Assured Initialization Optimizes the Value of Reservoir Simulation
Recommended Additional Reading
SPE 159274 A Novel Simulation-Model Review Process by M.R. Palke, Ryder Scott Company, et al.
SPE 159584 Dual-Continuum Modeling of Shale and Tight Gas Reservoirs by Prince N. Azom, The University of Texas at Austin, et al.
SPE 162380 Application of Streamline Simulation to Reservoir Management: Management of WAG-Injection Optimization by Mohamed Bouaouaja, Schlumberger, et al.
SPE 162045 Stochastic Uncertainty Analysis in Compositional Simulation for Giant-Gas/Condensate-Field Reservoir-Performance Prediction by E. Descubes, Schlumberger, et al.
SPE 164142 Uncertainty-Quantification Workflow for Mature Oil Fields: Combining Experimental-Design Techniques and Different Response-Surface Models by Syed Jawwad Ahmed, ADNOC, et al.
| Martin Crick, SPE, is chief petroleum engineer with Tullow Oil, responsible for all aspects of reservoir and production engineering in the group worldwide. Currently, he is on assignment in Ghana as subsurface manager of the Jubilee field. Previously a principal reservoir engineer with Schlumberger, Crick was responsible for the design of the reservoir-engineering features in Petrel and, most recently, for a review of well-test-interpretation workflows within Schlumberger. His experience over 25 years in the industry has focused on reservoir engineering, especially simulation in support of field-development planning. Crick holds a BS degree in physics from the University of Bristol. |