Determination of large-scale reservoir properties using global and local upscaling approaches

Abstract

Fine-scale reservoir properties obtained from Geostatistics generally cannot be handled by dynamic reservoir simulator. Reservoir properties for larger grid blocks are therefore. These larger grid blocks must produce similar results to those obtained from fine scale simulation. Conventional methods used to compute properties of coarse blocks are performed for the entire field and take a lot of computational time. This study introduces a new method to determine properties of large grid blocks. Local upscaling is performed at data locations, and then upscaled properties are populated for the entire reservoir using Geostatistics. In this study, permeability was chosen as a parameter of interest. Three reservoir models were generated using Geostatistics. The permeability fields of Reservoir I and II were generated based on Krigging estimation while that of Reservoir III was generated based on Sequential Gaussian Simulation. The permeabilities of these reservoirs were upscaled by using harmonic-arithmetic averaging method. The coarse-scale permeability distributions based on global and local upscaling were compared suing values of permeability. In addition, results of coarse-scale simulations based on global and local upscaling were compared with results from fine-scale simulation. The results of the study show that the new algorithm to determine large-scale permeability distribution based on local upscaling is as effective as the procedure based on global upscaling.