In this study, we explore the nature of hydraulic fracture growth in the presence of pre-existing fractures. The hydraulic crack intersects a pre-existing fracture to which it is initially perpendicular and it is assumed not to immediately cross, but rather to propagate along the pre-existing fracture. A 2D boundary element model has been developed to consider interaction of a hydraulic fracture with pre-existing fractures that can result in fracture branching, division of the fluid flow, and opening or frictional sliding in pre-existing fractures. Fracture propagation in the elastic medium is driven by an incompressible, Newtonian fluid injected at a constant rate. The frictional stress on the joint surfaces is assumed to obey the Coulomb law. The governing equations for quasi-static fluid-driven fractures are given and corresponding scaling is provided to identify important parameters. The Displacement Discontinuity Method and the Finite Difference Method are employed to solve this coupled problem. A method for separately tracking the crack tip and the fluid front (accounting for the existence of fluid lag) is also included in the model. Numerical results are obtained for crack trajectory, internal pressure, frictional contact stresses, opening and shear displacements, and fluid lag size, as well as for crack re-initiation from secondary flaws. Post crack coalescence, the hydraulic fracture growth mode changes from tensile to shearing, which may contribute to abnormal fracture pressure and to a reduction in fracture width. In the presence of pre-existing fractures, the fluid-driven cracks can be arrested or retarded in growth rate as a result of diversion of fluid flow and frictional sliding along the pre-existing fractures. Fluid penetration becomes more difficult for frictionally weak preexisting fractures. A competition between crack propagation from secondary flaws and delamination along the pre-existing fractures plays an important role in crack patterns and fluid flow paths that develop.
Fluid-driven fracturing or hydraulic fracturing, is a widely used technology especially for stimulating oil and gas wells and for measuring in-situ stress. A number of less-conventional uses for the technology are becoming more common, among them stimulation of geothermal reservoirs [1], inducing caving in coal and metal mining [2,3], and disposing of wastes by injecting them into underground rock formations [4]. Efforts to model and predict hydraulic fracture growth started in the 1950s, but continue today. In fact, despite a considerable number of laboratory and field studies, of various kinds, showing branched and non-planar fracture growth is common, most models and designs made with them assume a single planar fracture geometry is formed in the process. In many applications, the fracture process is still modelled using ideal fracture geometries such as the KGD model [5,6].
Geological discontinuities such as natural joints, faults and flaws, as well as bedding planes, are present in most formations [7-9]. Artificial discontinuities can be induced by previous fracturing treatments or by stress-induced fracturing associated with, for example, mining and reservoir fluid production. The terminology ?pre-existing fractures? refers to all these types of existing discontinuities which exist prior to a current fracturing treatment. These pre-existing fractures can significantly affect the propagation of new hydraulic fractures. For instance, cases of fractures arresting at low frictional strength bedding planes due to sliding and opening on the bedding plane have been documented [10]. In general, the reduction in the growth rate or arrest of the fluiddriven fracture results from
