This paper presents an analysis of a poroelastic mechanism for deviating hydraulic fractures, by modifying the stress field by fluid injection and pumping in the reservoir. This mechanism is studied within the confines of a simple model involving one injection and one pumping well and a hydraulic fracture propagating along a path initially equally distant from the two wells. Analysis of the fracture deviation from its straight-ahead path, and determination of the conditions leading to attraction of the fracture by the injection well are both based on a theoretical study of the stress trajectories. Comparison of the analytical prediction of the fracture path with the computed path using a numerical technique shows excellent agreement between the two methods.
It has recently been recognized that a hydraulic fracture could be deviated from its trivial path of propagation (which is perpendicular to the minimum far-field compressive stress) by exploiting poroelastic effects (Bouteca et al., 1983; Bruno and Nakagawa, 1991; Elbel and Mack, 1993; Mack and Elbel, 1994; Wright et al., 1994; Boone and Wawrzynek, 1994). For example, in a paper that appears to be the first to recognize the possibility of engineering deviation of hydraulic fractures using poroelastic effects, Bouteca et al. (1983) proposed to take advantage of the principal stress reorientation induced by fluid injection from a wellbore in a porous layer to alter the direction of fracture propagation. This concept was demonstrated in laboratory and field experiments in which two wells were linked by hydraulic fracture. More recently, Elbel and Mack (1993) reported evidence of stress reorientation induced by production and suggested that the stress field perturbation caused by production from a fractured well creates conditions favorable for propagating a secondary hydraulic fracture from the producing well that is orthogonal to the primary one (see, also, Wright et al., 1994; Mack and Elbel, 1994). Another proposal discussed by Boone and Wawrzynek (1994) is to take advantage of the typically large in situ permeability contrast in the horizontal and vertical directions to flip, in particular circumstances, the plane of propagation of the fracture from vertical to horizontal. The initial motivation for our paper came, however, from the work of Bruno and Nakagawa (1991), in which an experimental demonstration of the alteration in the direction of propagation of a tensile fracture by combined fluid injection and pumping was established. The experiments revealed that a fracture, initially propagating along a path equidistant from an injection and a production port, is deviated by the injection port. Increased injection pressure causes larger fracture deviation and, eventually, attraction of the fracture by the injection port. The mechanism of fracture reorientation was attributed by Bruno and Nakagawa (1991) to the local pore pressure gradient at the crack tip. This point was disputed by Detournay and Boone (1993), however, who argued that the pore pressure near the crack tip could not influence the orientation of fracture extension, because the pore pres- sure field is not singular (see, also, the author's reply by Bruno (1994)).