We present detailed comparisons between theoretical predictions and experiments for the initiation of radial hydraulic fractures transverse to a wellbore. Most of these experiments provide a measure of the time evolution of pressure, length, opening and flux entering the fracture during the test. We compare all these measurements simultaneously to our theoretical predictions, thoroughly testing hydraulic fracture initiation theory, which is shown to be predictive. These comparisons notably emphasize the difference between the initiation pressure (the wellbore pressure for which the fracture initiates) and the breakdown pressure (a misnomer which simply corresponds to the maximum pressure recorded during the test).
In this paper, we will compare hydraulic fracturing laboratory experiments performed in different tight materials (PMMA, cement, shale) with the predictions from a numerical solution of the governing equations of hydraulic fracture mechanics (e.g. Geerstma & de Klerk, 1969; Detournay & Peirce, 2014). We focus on transverse radial hydraulic fractures initiating from a radial notch at the wellbore wall (see Figure 1).
In particular, we will clarify the difference between the initiation pressure (the wellbore pressure at which the fracture initiates) and the breakdown pressure (a misnomer which simply corresponds to the maximum pressure recorded during a test, usually carried out at constant pump rate). This difference stems from the strong viscous effects at play during wellbore pressurization, fracture initiation and early propagation (i.e. fluid entrance in the newly created fracture). It has been observed on a number of experiments (e.g. Weijers, 1995; Zhao et al., 1996) and has a strong practical application, first because the maximum pressure is the most obvious value to report and second because that maximum value is necessary to specify injection systems.
First we will review the governing equations of hydraulic fracture mechanics, highlighting the importance of their coupling with wellbore pressurization at initiation and early fracture growth. We will restrict ourselves to solids-free Newtonian fracturing fluids (let us note that, in practice, most hydraulic fractures are initiated with such a fluid) and linear elastic fracture mechanics. We briefly recall the relevant scaling for such radial fracture geometry. Indeed, let us note that hydraulic fracturing is a multi-scale problem and that it is not possible to run laboratory experiments at a 1:1 scale to understand the problem. Therefore, understanding of the scaling laws is very important. Furthermore, validation of a numerical simulator with experiments is necessary to use it for field scale design and interpretation.