Fracture Aperture in Flow Models: to Average, or not to Average?

Summary Fracture networks are heterogeneous systems creating anisotropic flow patterns, though field-scale flow models of fractured reservoirs use up-scaling of the fracture network to effective properties to model flow. We study the impact of upscaling aperture from a realistic aperture model, where aperture is defined by a stress-aperture relation that considers aperture variations within single fractures, to an arithmetically averaged distribution. This analysis is done through calculation of the equivalent permeability tensor of a discrete fracture and matrix flow model, taking into account different contrasts between fracture and matrix permeability. Results show that a strong equivalent permeability anisotropy emerges for a low matrix permeability, and that accurate representation of fracture aperture is as important as the fracture topology, particularly when the fracture matrix permeability ratio is large. The orientation of the permeability tensor varies with varying rock matrix permeability, reflecting anisotropy of the permeability that is impossible to extract from the fracture patterns itself. The results also highlight that utilizing a constant aperture for the fracture network may not reproduce the true anisotropy of the permeability. Utilizing an equivalent aperture in flow simulations introduces errors in both the direction and magnitude of principal permeabilities and the equivalent permeability anisotropy.