Effective characterization of unconventional resources is essential for economic feasibility studies in competitive market conditions. Characterizing these reservoirs presents significant challenges because of inherent complexity and requires methods capable of mitigating the uncertainty. The presence of natural fractures is a key indicator of production potential of unconventional reserves. Predictive simulation methods using complex fracture networks (natural and induced) rely on physical properties of the fractured system to estimate flow rates and production trends. Reasonable estimates of the matrix rock and fracture properties are extremely important for reliable predictions. However, limited data availability and technical knowledge increase the challenges of characterizing fracture properties, making a random sampling approach an appealing option.
Moreover, random sampling underestimates permeability values by an order of magnitude compared to those generated from actual fracture-length aperture correlations. These underestimates affect predicted flow rates and production estimates. Further, model fracture shapes are too simple (e.g., rectangular, circular, or elliptical) and do not account for the anisotropic nature of source rocks, which, more than likely, cause asymmetrically shaped fractures. Physical properties of large fractures can exhibit significant spatial variation rather than a constant value across the fracture, a common assumption in current modeling practice. This study focuses on generating physical properties (aperture, permeability, and porosity) of natural fractures in which aperture sizes are correlated to fracture lengths for more realistic fracture geometries. Localized aperture values are generated by considering a known aperture opening shape profile. Permeability and porosity values are calculated based on these physically and spatially correlated aperture values. This method incorporates physical correlations beyond assumptions of regular fracture shape and single physical property across the fracture. The new method is expected to improve efficacy of predictive simulation tools used to develop and plan production operations for unconventional reservoirs and works well with either structured or unstructured grids.
