Abstract
We report on numerical simulations of extended leak-off tests (XLOT) in a low-permeability rock formation and how they are impacted by the presence of a natural fracture. The simulations are carried out by using a newly developed modified discrete-element model (MDEM). The main features of XLOT are reproduced by the MDEM simulations. In the absence of a natural fracture, the hydraulic fractures created during XLOT grow perpendicular to the minimum principal stress. If a conductive natural fracture intersects the wellbore, fracture twisting is observed, and the XLOT pressure curves are markedly different from the pressure curves recorded in the absence of a natural fracture.
1. INTRODUCTION
An extended leak-off test (XLOT) provides information on hydraulic-fracture initiation and propagation in a rock formation at a given depth, as well as the magnitude of the minimum principal stress [1-4]. XLOT is usually carried out in a few-meter long open-hole section below a casing shoe. In this way, the test also provides information on the cement integrity. The test consists of typically two to three injection cycles during which the well is pressurized until a hydraulic fracture is initiated (or re-opened) and propagated a few meters into the formation. After injection is stopped, the pressure decay due to leak-off and fracture closure is recorded. In low-permeability formations, leak-off is small so that flow-back is needed to induce fracture closure. Closure of the fracture results in an inflexion in the pressure vs. time curve, and this fracture-closure pressure directly corresponds to the minimum principal stress.
For vertical wells and simple stress and geological settings, analytical models can be applied to the description and interpretation of XLOT data, allowing for an accurate determination of the minimum principal stress. However, for deviated wells, rotated stress field, heterogeneous rock formations, or in the presence of natural fractures, the hydraulic fracturing process during XLOT can be rather complex (including fracture twisting or fracture snap-off), and conventional analytical models may not be applicable any longer [5].
