Optimizing Simulation Models by Upscaling from Integrated Reservoirs Models; A Case History

Abstract

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractThis paper presents a case study of using integrated models in improving upscale accuracy that increases the model speed. This has a great impact on CPU time and therefore the cost of the study. A unique process that integrates data from all possible sources (open hole logs, RFT and production data, seismic and geologic interpretations) was used in a geologic model. This static geologic model was then used to build a DYNAMIC simulation model. We superimposed water encroachment movement over the geologic model that in turn was recycled back in developing the simulation (engineering) model. Up-scaling the resultant model was done using a multi step technique that captured all the varying details of the complex reservoir such as: a) A tilt in the free water level coupled with the effect of capillary forces on reservoir fluid distribution; b) Reservoir stratification which has the effect of limiting bottom water support and causing the injected water to move rapidly through the high permeability layers ; c) While the effects of stratification and faulting are apparent, communication is observed across these features.The study was done on a sector model extracted from the full field geological model as a pilot testing of the method. The same procedure was then applied to the full field model. As a result of using this technique, pressure and water saturation match was obtained and prediction cases were submitted to predict reservoir performance for designing operational and production strategies for the field.

Synthetic Fractal Modelling of Heterogeneous and Anisotropic Reservoirs for Use in Simulation Studies: Implications on Their Hydrocarbon Recovery Prediction

“…Unfortunately, such data are only available at some locations, usually wells, and a complete reservoir description at this scale cannot be obtained. Recently developed statistical techniques can go some way towards helping to construct a reasonable reservoir model, but all of them rely on significant variability in the reservoir only occurring at a scale at or larger than the inter-well spacing Al-Ali and Al-Qassab 2000;Alqassab and Heine 1998). Because such data are impossible to obtain for the whole reservoir, it was decided to construct a synthetic property distribution mimicking real reservoirs and representing the scale necessary for reservoir simulation and representing the macroscopic laboratory measurements at the same time.…”