Three-dimensional Voronoi-based distinct element model for simulation of hydraulic fracture propagation

Abstract

fracture propagation 300 Deep Mining 2017, Perth, Australia small apertures, and so they require either the build-up of fluid pressure or stress-induced shear dilatancy to develop into significant flow pathways. Further, natural fractures can also redistribute the in situ state of stress, particularly when deformation occurs along these features; this impact is also critical given that, ultimately, it is the local stress state rather than the far-field stress state which determines the critical minimum pressure for fracture opening and extension. The combined stress redistribution and flow redirection effects influence the mode of interaction between new and pre-existing fractures, determining whether induced fractures cross the natural fracture, become redirected, or have their propagation arrested (Gale et al. 2007). All of the stress and flow related influences of natural fractures are highly dependent on the in situ fracture characteristics, including aperture, roughness, and orientation, which are often poorly understood. The factors discussed above (in particular the effects of pre-existing natural fractures on the local stress state and rock mass permeability) increase the gap between the in situ conditions and the simplifying assumptions for analytical models for hydraulic fracturing (e.g. significant geometrical constraints and continuous isotropic and homogenous media) such as those described in Sneddon (1946), Sneddon and Elliot (1946), Khristianovich and Zheltov (1955), Geertsma and de Klerk (1969), Perkins and Kern (1961) and Nordgren (1972). This limits the validity of such conventional analytical solutions for simulation of hydraulic fracturing in naturally fractured formations. Nevertheless, analytical models provide a basis for understanding fundamental physical behaviours, and can serve as a check on more complex numerical approaches. To address this limitation, a new approach is introduced in this paper for simulation of hydraulic fracturing process based on the synthetic rock mass (SRM) concept (Pierce et al. 2007) and the application of three dimensional Voronoi tessellation for fracturing of intact rock (Ghazvinian et al. 2014). 2 Numerical simulation of hydraulic fracturing Numerical simulation of hydraulic fracturing provides a tool for prediction of ground response in terms of dimensions of induced fractures, their interaction with the pre-existing natural fractures, activation/slipping of the exiting discontinuities and the associated stress shadowing, etc. Therefore, allowing for optimisation of the design parameters and improved efficiencies during the field application of hydraulic fracture stimulations (oil and gas, e.g. Lee et al. 2016; Maxwell et al. 2016) and preconditioning (mining, e.g. Brzovic et al. 2015; Damjanac et al. 2015; Preisig et al. 2015). The multiphysics nature of hydraulic fracturing is a complex process. The hydraulic behaviour of the injection fluid as well as the mechanical response of the rock formation and the interdependence of these two factors assure the challenging task of numerical simulation of hydraulic fracturing, particularly at MMHF operation scale. In the simplest form, the propagation of hydraulic fractures (HFs) are calculated by means of analytical solutions in combination (may or may not be fully-coupled) with continuum numerical analysis of the mechanical component of the model. In a more advanced approach, which can also be referred to as the current state-of-practice in two and three-dimensional (3D), a distribution of natural fractures is introduced to the hydro-mechanically coupled distinct element model (DEM) of an intact rock, in the form of a discrete fracture network (DFN), whereby fluid flow is allowed on the fracture surfaces between the blocks (e.g. Katsaga et al. 2015; Nagel et al. 2013a; Riahi & Damjanac 2013). This approach can provide realistic information regarding fracture aperture and fluid flow within a fractured reservoir, however, it lacks the ability to propagate new hydraulic fractures through the intact rock. The emerging codes based on finite-discrete element method (FDEM) provide a solution to a fully-coupled model with the ability to simulate initiation and growth of HFs as well as their interaction with the pre-existing natural fractures (Grasselli et al. 2015). However, the application of this method for simulation of hydraulic fracturing is currently limited to two-dimensional models and has non-trivial simulation time. The 3D nature of pre-existing fractures in unconventional reservoirs require the simulation of hydraulic fracturing to be performed in 3D (Damjanac et al. 2015). With the recent advances in formulation and efficiency of DEM-based numerical codes, the SRM concept has shown promising results for simulation of fully hydro-mechanically coupled, 3D HFs (Damjanac et al. 2015; Damjanac & Cundall 2016).