Numerical simulation of fluid flow and sensitivity analysis in rough-wall fractures

Abstract

Activated natural fractures, poorly-propped and un-propped secondary hydraulic fractures may serve as the connection between reservoir fluids and the main hydraulic fractures (i.e., well-propped fractures). These branched fractures also contribute considerably to well after-fracturing productivity in unconventional tight/shale reservoirs. Rough fracture surface and its effects on fluid flow behavior in such fractures with very limited width is very important in understanding total mass transport in fractured rocks. In this paper, the effect of rough surface properties on fracture conductivity is studied through 3D numerical simulation of fluid flow behavior within the fractures. First, a tensile fracture was created on a core sample retrieved from the Montney formation, and the rough fracture surface topography was scanned via a high-resolution optical profilometer. The surface roughness characterizations were then used as reference to numerically reconstruct multiple three-dimensional fracture models. A finite element method was adopted to simulate fluid flow in the rough-wall fractures, using Navier-Stokes equation, for various fracture surface properties such as mean aperture, root mean square (RMS) asperity height, correlation length, anisotropy and shear displacement distance. Results showed that cubic law tends to overestimate the conductivity of the fractures with rough-wall surfaces, especially when the fracture aperture is small. Other surface properties, including correlation length, anisotropy and Hurst exponent also exert considerable impacts on fracture hydraulic properties. As the shear displacement distance increases, deviation of the fracture conductivity from cubic law first increases and then maintains at a constant level. The results of this paper have advanced the understanding of fracture conductivity and its dependence on fracture surface properties and will aid in improving production prediction and optimization in fractured tight reservoirs.