AAPG ACE 2018

Datapages, Inc.Print this page

Capturing Reservoir Heterogeneity in Reservoir Models – How Much is Enough?

Abstract

As the cost of computing decreased and the sophistication of reservoir modeling software increased over the past decade or two, the ability to capture reservoir heterogeneity in reservoir models has increased significantly. This increase was augmented by the development of a variety of algorithms that enabled increased efficient use of data obtained at a variety of scales, from core plug to 3D seismic, to be used to constrain reservoir models. While these important advances clearly enabled very large and very detailed reservoir models to be constructed relatively simply, a robust debate within the industry as to the overall value of highly detailed reservoir models as the basis for dynamic models that are in turn used to produce the volumetric forecasts required to evaluate and ultimately justify project development decisions continues. This “debate” has been referred to by some as the “Frankenstein vs. Gilligan” model debate. Studies completed by the author and others over the past two decades suggest that a better frame for this discussion and evaluation of model-based workflows is “fit-for-purpose” or “fit-for decision” modeling. In other words, reservoir models should capture the level of reservoir heterogeneity needed to make project development decisions. Often times, the level of geologically heterogeneity that needs to be incorporated in dynamic models used to justify project decisions is actually quite low. For example, once the reservoir net to gross ratio exceeds roughly 15-20%, the reservoir is essentially connected and increased model heterogeneity adds little value. This applies to both carbonate and clastic reservoirs. The impact of sparse data and the dynamic model size (as measured by number and/or dimension of grid cells) has greater influence on project decision making than fully capturing detailed reservoir stratigraphy and/or reservoir property heterogeneity. Results of various studies on Permian Basin (USA) and large Middle East carbonate reservoirs as well as clastic reservoirs in the San Joaquin Basin in California and the Maracaibo Basin in Venezuela are used to support these conclusions.