Abstract: Selection of 3D Geostatistical Reservoir Representations Based on Dynamic Production Criteria
Isabelle Morelon, Dominique Guerillot
Geostatistical models allow to take into account uncertainties in reservoir description. In this way, several possible reservoir images are generated, and must be considered when evaluating production forecasts. Nevertheless, flow simulations are too expensive to be performed on a high number of geostatistical representations; in addition, one can expect that some images have the same production behavior. Hence, reservoir images representative of the possible production behaviors of the reservoir must be selected.
In this paper, we propose a selection method based on dynamic criteria: simplified and much more less expensive 3D flow simulations are performed on many geostatistical representations. Production results are obtained and representative reservoir images are selected by considering these results. Simplified flow simulations consist in computing only one constant pressure field corresponding to a steady-state flow within the reservoir. Actual boundary conditions are kept unchanged. Saturations calculations take into account multiphase parameters. Production results obtained by simplified flow simulations are wrong, but allow to obtain a right sorting of the images for cases belonging to the validity domain of the method.
The selection method has been applied on reservoir cases corresponding to di- or triphasic fluid flow. These tests have shown that the proposed selection method is valid for diphasic water-oil fluid flow with no significant capillary and gravity effects compared to viscous effects.
The proposed selection method based on low-cost simplified fluid flow simulations is one efficient solution to the problem of dealing with multiple geostatistical representations. And this approach could be advantageously mixed with homogenization methods.
AAPG Search and Discovery Article #90956©1995 AAPG International Convention and Exposition Meeting, Nice, France