A new paper* by Shahab Mohaghegh, Professor of Petroleum and Natural Gas Engineering at West Virginia University and president of Intelligent Solutions (IS),
provides an update on the application of artificial intelligence and data mining to real time reservoir modeling and management. IS’ surrogate reservoir model (SRM) is a lightweight characterization of a field that enables compute-intensive techniques such as neural networks and fuzzy logic to be performed in real time.
SRM has been successfully field-tested on a Middle East giant oil field. Prior to SRM, a single simulator run on a million cell full field model took some 10 hours on a cluster of 12 parallel CPUs. The SRM approach sped up the modeling to the extent that it became possible to perform tens of millions of meaningful numerical experiments to ‘comprehensively explore’ the reservoir model’s solution space resulting in the development of a new field development strategy. The field’s operator has now decided to apply the technique on other assets.
Currently, SRM is a service offering but a new ‘Intelligent Surrogate Modeling and Analysis (ISMA)’ software package is scheduled for release in 2010.
The paper also outlines ‘top-down reservoir modeling,’ (TDRM) which Mohaghegh describes as ‘the hottest of our current workflows.’ TDRM builds its model of the reservoir from production data rather than the convention ‘bottom-up’ approach of building a geological model. More from www.IntelligentSolutionsInc.com.
* www.oilit.com/papers/IntelligentSolutions.pdf.
© Oil IT Journal - all rights reserved.