The objective of this study is to develop an upscaling methodology for simulation of tracer injection in naturally fractured reservoirs. 2-D fine grid models that explicitly incorporate fractures are used to represent the realistic fracture distribution from geostatistical characterization. The objective is to determine equivalent properties of the fractured systems, which, as included in conventional double porosity models, will result in a behavior similar to the fine-grid response. The upscaling methodology is based on flow simulation with constant pressure and constant flux boundary conditions. The upscaling parameters of each equivalent double-porosity grid are determined by matching results of the fine grid simulation. The effects of some important factors such as coarse grid size, injection rate, and well location are examined to show the robustness and accuracy of the upscaling methodology.


Geological characterization of fractured reservoirs has progressed considerably in recent years, allowing for realistic representation of fracture networks. Despite advances in reservoir simulation and computer hardware, such detailed geostatistical results cannot be used directly in the simulation of fractured reservoirs. Up-scaling techniques are needed to translate the geostatistical data into reservoir simulation parameters.

There are several upscaling techniques (1–4) used in the simulation of single-porosity reservoirs to reproduce the details of small-scale fluid mechanics and reservoir heterogeneity in coarse-grid models. However, there is no general procedure that is widely accepted for upscaling. The upscaling methods for naturally fractured reservoirs are further behind.

Based on the type of fracture system, the upscaling study of naturally fractured reservoirs can be divided into two categories: the single continuum, and the dual continuum approach. For the single continuum approach (5–7), the fractured system is considered as a heterogeneous matrix system with discrete fractures. Various methodologies, such as the boundary element and finite volume methods, are widely used to deal with complex fracture systems. Most single continuum methods consider single-phase upscaling only, and treat the upscaled permeabilities as tensor terms. The dual continuum approach (8–11) is based on the double porosity model of Warren (12) and Root, and its representation in the simulation of fractured reservoirs (13). The dual continuum upscaling therefore requires determination of the properties required for double porosity simulation of the fractured system.

Most of upscaling research concentrates on obtaining equivalent properties for a single-porosity simulation of a fractured system. Upscaled single-porosity models do not exhibit some of the important characteristics of naturally fractured reservoirs. In particular, a single-porosity model cannot distinguish between the displacing fluid fronts in the fracture and the matrix. The dual continuum approach, however, includes the physics for modeling unsteady-state multiphase fluid exchange between the rock matrix and fractures. In this work, we will obtain the equivalent double-porosity parameters for a tracer displacement in a fractured porous medium. In the following, the upscaling framework, including fractured systems used in our study, is presented first. The upscaling methodology and results are described in detail.

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