Abstract

Upscaling technique has become an increasingly important tool in recent years for converting highly detailed geological models of reservoir to simulation grids. However, how to qualify the upscaling results, i.e. whether an upscaled results provides a good or bad approximation, is one of the outstanding problems remained in this research field.

Comparing the simulation results for the reservoir before and after scale-up is direct and adaptable. But the shortcoming is the simulation results of the original fine model is too costly even can not be obtained. This paper evaluates the characteristics of information loss in reservoir upscaling and proposes two new indicators of information loss, ILIa and ILIb. Based on information loss, a method-independent technique for evaluating upscaling results in reservoir engineering is proposed. The new validation method can be used to evaluate the upscaling results before simulation stage.

Introduction

Upscaling is a process that scales properties of a fine grid to a coarse grid such that the fluid flow behavior in the two systems is similar. It has become an increasingly important tool in recent years for converting highly detailed geological models of reservoir to simulation grids. In the past two decades, many upscaling approaches have been proposed. However, no measure of quality is available for any upscaling routine (Christie 1996). So, how to qualify the upscaling results, i.e. whether an upscaled results provides a good or bad approximation, is one of the outstanding problems which remains in research.

Currently, the method used universally for evaluating upscaling is comparison of the simulation results for the reservoir before and after scale-up. If both simulations produce the same flow performance, the upscaling results may be taken to be a perfect approximation of the original model, and the upscaling method, which was used, may be considered ideal. Theoretically, this evaluation method is direct and adaptable. However, it is actually unfeasible because obtaining simulation results for the original model with a fine grid is too costly even can not be obtained. Some researchers use artificially small data set or employ different simulators for this kind of comparison, which should consequently be regarded as trustworthy.

Evaluation upscaling needs to be based on mathematical and geological consideration. This paper evaluates the characteristics of information loss in reservoir upscaling and proposes two new indicators of information loss, ILIa and ILIb. Based on information loss, a method-independent technique for evaluating upscaling results in reservoir engineering is proposed. The new method can be used to evaluate the upscaling results before the simulation stage.

This content is only available via PDF.
You can access this article if you purchase or spend a download.