Simple Upscaling of Oil-Water Relative-Permeability Models

Abstract

A good description of relative permeabilities (relperms) is crucial to predict the performance of petroleum reservoirs. They are important input for evaluating fluid flow, and establishing the most suitable strategies for reservoir development and management. A poor relperm description is a potential threat to business decisions. The applicability of a relperm dataset is highly sensitive to the scale at which it was acquired. For example, while special-core analysis (SCAL) data, which are often obtained from laboratory experiments may be adequate for core-scale simulation studies, they often fail to provide reliable description of fluid flow at the field scale. Although several upscaling techniques are available, they vary in their complexity and reliability. One area of improvement is to simplify the upscaling process. Extending and improving the approach introduced by Alpak et al. (2010), this paper uses the idea of global connectivity factor and a linear scaling algorithm to develop a simple technique to upscale oil-water relperms. Compared to Alpak et al. (2010), whose work was limited to two parameters, the new technique includes mathematical functions to extrapolate five of the Corey parameters (exponents and endpoints) from SCAL to their respective field-scale equivalents. With some oil-water SCAL datasets from three plugs in a Niger-Delta oil reservoir, the practical applicability of the proposed workflow is demonstrated. In addition, this paper covers the following areas: (i) upscaling individual SCAL plugs to preserve field heterogeneities; (ii) generating an upscaled model that aggregates all the SCAL plugs as an approximate homogeneous system; and (iii) constructing the full range of upscaled relperm models that reflect uncertainties for realistic reservoir simulations.