Abstract

Recent research on naturally fractured reservoirs suggests that microfractures and megafractures are all part of a single population that shows a power-law relation between frequency of occurrence and fracture aperture or length. The objective of this work is to investigate the effective permeability of fracture networks whose fractures obey power-law scaling relations.

For the simple case where each fracture extends through the region of interest, we present a formula that relates effective fracture-network permeability to the distribution of fracture apertures. In this simple case, the effective permeability of the region is dominated by the single largest fracture in the region.

In reality, of course, flow through fractures depends on whether fractures interconnect. Connectivity at one length scale does not relate simply with connectivity at another scale, if fracture properties obey power-law scaling relations. Fracture populations based on frequency/length scaling exponents derived from field data are not guaranteed to interconnect on either the microscopic or megascopic scales. Computer-capacity limitations make it difficult to include enough fractures to model a region even the size of one reservoir grid block. Clustering of fractures into swarms appears to be the key to fracture interconnectivity.

Introduction

Significant oil and gas have been produced from the various types of fractured reservoirs across the world.1 The behavior of naturally fractured reservoirs is very different from that of conventional reservoirs.1–3 The primary cause of this difference is the inherent character of naturally fractured reservoirs: most hydrocarbon resides in the pore space of the matrix whereas the flow of hydrocarbon towards wells is dominated by flow through networks of fractures. Consequently, the behavior of naturally fractured reservoirs is dominated by the properties of the individual fractures and the networks formed by the fractures.

One of the biggest difficulties in studying naturally fractured reservoirs is that data are limited, usually to one spatial direction (i.e., along a wellbore). The spatial variation of fracture features is complicated and irregular, and, as a result, characterization of a fractured reservoir is substantially more difficult than that of a non-fractured reservoir.

Most fractured-reservoir simulations are based on simplified idealized models.1,4 The assumptions of these models are sometimes clearly different from the conditions of underground reservoirs. Therefore, it is desirable to find an accurate, efficient, practicable simulation method that is based on data for the fractures in a given field.

Individual fractures are characterized by fracture aperture, size and orientation. Fracture aperture is the gap between fracture walls. If a fracture is defined as a disk in space, the radius of the disk quantifies the size of the fracture. Fracture orientation gives the direction and tilt of the fracture. Fractures with similar orientation can be grouped together as a fracture set.

A population of fractures is characterized by its distributions of aperture, size, density and orientation, as well as matrix block size and shape. Fracture density expresses the extent of rock fracturing, the number of fractures per unit volume or per unit area, depending on the study.

Making use of the latest findings in structural geology, this study attempts to relate, through numerical simulation, certain properties of fractures and their statistical distributions to the flow properties of naturally fractured reservoirs. With this information, the performance of commonly-used simulators for fractured reservoirs can be improved.

Simulation of Naturally Fractured Reservoirs.

Current simulation technology for naturally fractured reservoirs is based on either continuum or discrete-fracture models. By representing each fracture individually, discrete-fracture models can incorporate many of the characteristics of real fracture systems.5–10 However, their use is limited by the capacity of simulators and computer resources if there is a large number of fractures present. In a fractured reservoir, numerous fractures connect to one another to form complicated fracture networks. Available geological and engineering data, often limited to a single spatial direction (for instance, in a wellbore) or at scattered locations (coring in different wells), give little information on these networks.

Simulation of Naturally Fractured Reservoirs.

Current simulation technology for naturally fractured reservoirs is based on either continuum or discrete-fracture models. By representing each fracture individually, discrete-fracture models can incorporate many of the characteristics of real fracture systems.5–10 However, their use is limited by the capacity of simulators and computer resources if there is a large number of fractures present. In a fractured reservoir, numerous fractures connect to one another to form complicated fracture networks. Available geological and engineering data, often limited to a single spatial direction (for instance, in a wellbore) or at scattered locations (coring in different wells), give little information on these networks.

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