Abstract

We performed laboratory scale experiments for the study of horizontal fluid driven fracture initiation from an open-hole section. In order to clarify the role of material microstructure and fluid infiltration, we used viscous Newtonian fluids to initiate penny shaped fractures from radial notches in blocks of natural sandstone. We varied the fluid viscosity and injection rate over ranges allowing for significant fluid losses toward the dry porous rock. For low efficiency fluids, the fracture propagation was characterized by non-monotonic pressure variations. We propose a formulation of the problem of hydraulic fracture initiation in an elastic homogeneous permeable material. Our model accounts for fluid losses toward the rock and a cohesive zone at the fracture tip. We performed dimensional analysis in order to lump the physical parameters into meaningful dimensionless groups. Scale effects concerning the fluid efficiency, decompression effects and material failure could be predicted from dimensional analysis and corroborated with experimental observations. We also propose a numerical solution for this model. The numerical results compare well with experimental results. This work provides a proper scaling for hydraulic fracture initiation. This scaling is necessary for comparing laboratory scale results to field observations.

Introduction

Fracturing of horizontal wells has been pioneered in the North Sea. Accurate control of the location of each fracture required isolated intervals along the horizontal section, because of nearby gas and water bearing layers. Later, fracturing horizontal wells has been done from perforated completions, with isolation of treated zones. However, cost reduction and practical limitations have motivated some operators to try stimulating horizontal wells without zone isolation. As an extreme case, one would even attempt to fracture an open-hole section. In principle, the fracture can start anywhere along the open section, but permeability variations and initial flaws in the rock would play an important role. In that case, successful treatments must start by controlling the initiation points. If the borehole can be notched in such a way that the notch dominates any natural variations, we would know where the fractures initiate and could control the number of fractures.

Assessing the minimum notch size for controlling crack initiation requires an in-depth understanding of the different mechanisms that interact during the early stages of the fracture onset. We designed laboratory experiments in order to simulate the process of hydraulic fracture initiation under in-situ conditions. The considerable scale difference between field and laboratory conditions demands special care when one draws conclusions from small-scale tests results. This implies the design of proper scaling laws based on a set of the governing equations underlying a description of hydraulic fracture initiation. The complexity of hydraulic fracture initiation hinders the use of analytical modelling. Numerical modeling is therefore required. The validity of such a modeling is legitimized by a good agreement between experimental observations and model predictions.

Experimental set-up and method

The laboratory tests are performed using the hydraulic fracturing experimental set-up developed at the faculty of Applied Earth Science of TU Delft1 (see Figure 1).

Block geometry.

A cubical sample up to 30 cm on a side can be tested. The six faces are polished to prevent any stress inhomogeneity. Teflon sheets and a layer of Vaseline limit the development of shear stresses between the pressure plates and the block. A 23mm diameter borehole is drilled in the center of one face across the complete block. The notch is a circular cut made in the wellbore, in the middle of the injection zone, perpendicular to the borehole axis (see Figure 2). Only dry samples (or wet samples without back pressure) can be tested. More detailed descriptions can be found in the work of Weijers2.

Block geometry.

A cubical sample up to 30 cm on a side can be tested. The six faces are polished to prevent any stress inhomogeneity. Teflon sheets and a layer of Vaseline limit the development of shear stresses between the pressure plates and the block. A 23mm diameter borehole is drilled in the center of one face across the complete block. The notch is a circular cut made in the wellbore, in the middle of the injection zone, perpendicular to the borehole axis (see Figure 2). Only dry samples (or wet samples without back pressure) can be tested. More detailed descriptions can be found in the work of Weijers2.

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