Study of Evaluation Criteria for Reservoir Upscaling

The hydrocarbon reservoirs are characterized by their distribution of petrophysical properties which are obtained from well data and seismic information. The geological model of the field, called also geostatistical model or fine model, is built with respect to the peculiarities and heterogeneity of the reservoir geometry. Cell size of the geological model is generally small comparing to the field size, and the fine model contains several hundred million of gridblocks. Numerical fluid flow simulations may require a lot of CPU time and are generally impossible to achieve at the detailed scale of the geological model. To reduce the simulation cost in calculation time or to perform a simulation, reservoir engineers upscale the geological model to an upscaled mesh on which the numerical simulations of production can be performed in a satisfactory lapse of time. The equivalent properties obtained on the reservoir model loss detailed geologic information. The purpose thus is to find and study different criteria which can evaluate the characteristics of information loss in reservoir upscaling. On SPE10 test case, the authors show how this criteria can evaluate upscaling results in reservoir engineering.