Semi-analytical modelling of fluid flow in unconventional fractured reservoirs including branch-fracture permeability field

Abstract

Growing demand for energy and unavailability of new viable energy resources have played a crucial role in the persistent exploitation of unconventional resources through multistage hydraulic fracturing. Currently, standard modelling approaches idealize a fractured media as an interplay of several homogeneous continuum of normal diffusive characteristics. However, evolved branch-fractures generate a space with extreme heterogeneity around primary fracture plane. The precise characterization of these branch-fractures is imperative for well performance analysis along with subdiffusive behaviour of unconventional matrices. This study presents two semi-analytical models that account for the branch-fracture permeability field and subdiffusion. The first model, Induced Branch-fracture Subdiffusive Flow model (SIBFF), accounts for exponential permeability field concept and subdiffusive transport behaviour of matrices. Compared to the earlier analytical models, the SIBFF model accounts for more comprehensive transport mechanisms and medium properties. The other model, Fractal Branch-fracture model, couples fractal porosity/permeability distribution of branch-fracture and subdiffusion to account for more detailed description of stimulated reservoir volume (SRV) and unfractured inner region. The wellbore pressure solution is derived by discretizing the reservoir into several flow regions and imposing both flux and pressure continuity at the interface between contiguous segments. The inclusion of permeability field and fractional flux law introduces important complexities to the mathematical model that are carefully resolved by implementing Bessel functions and Laplace transformation (LT). Finally, the solution is inverted to time domain using Gaver-Wynn-Rho (GWR) algorithm. This study also assessed the applicability of four numerical inversion methods and found GWR method more suitable and predictive. The sensitivity of important model parameters is presented. Results were verified analytically and validated against Niobrara and Eagleford field data. It is shown that the models could be implemented to quantify the efficiency of a stimulation job, to decide on the necessity of re-fracturing a formation and to analyze horizontal well performance with better predictive capability. The proposed models could further be employed to characterize different flow regimes for unconventional reservoirs.