Abstract:
A significant challenge in horizontal well stimulation is the ability to simultaneously generate multiple hydraulic fractures (HF) with roughly uniform dimensions within a single stage. When the fractures are relatively close, the outer fractures in a stage exert a strong confining stress on the inner fractures, inhibiting their development and the desired uniform growth of fractures in the stage. There is thus a need to develop computational tools to analyze this mutual interaction between propagating HF to seek designs that can mitigate this inhibition phenomenon which is known as stress shadowing. In this paper, we report the development of an axisymmetric extended finite element method (XFEM) that can model multiple simultaneously propagating HF that are able to curve (forming bowl-shaped fractures) and which is able to autonomously partition the flux of fluid among the HF in the stage. To test the flux partitioning algorithm, we compare the XFEM code to a parallel planar displacement discontinuity (DD) code. To test the HF curving due to mutual interaction, we compare the XFEM to an axisymmetric DD code, which can model two symmetric fractures that can curve due to mutual interaction. Finally, we describe an experiment in which five mutually interacting HF with flux partitioning are allowed to curve, and we compare the results to the case in which the five HF are constrained to develop in distinct parallel planes.
Introduction
Horizontal wells frequently have a significant number of nonproducing perforation clusters in a given stage due to variations in the reservoir properties and the phenomenon of stress shadowing, in which the confining stress induced by the outer fractures in a stage serves to inhibit the growth of the inner fractures. It is thus desirable to develop computational tools that can analyze such situations to determine the optimal choice of engineering parameters to mitigate the effect of stress shadowing and reservoir heterogeneity. A complete model of this situation requires a fully coupled 3D simulator that can model curving HF that simultaneously propagate in a heterogeneous solid medium. Thus far, models of this situation have been restricted to homogeneous media involving pseudo-3D approximations that can admit curving cracks (Kresse et al., 2013), 2D-axisymmetric HF that are assumed to grow in parallel planes (Lecampion and Desroches, 2015) or arbitrarily shaped HF that are assumed to grow in parallel planes (Bunger and Peirce, 2014; Peirce and Bunger, 2015).
