Modelling of normal net stress effect on two-phase relative permeability and capillary pressure of rough-walled fracture

摘要

Fluid-flow in fractured reservoirs is highly sensitive to the change of effective stress during fluid injection or production. Permeability, capillary pressure and relative permeability of rock fractures to oil and water directly impact the amount of hydrocarbons that can be ultimately recovered; however, these parameters are difficult to measure in the lab as a function of effective stress. This stimulates development of computational algorithms to predict the impact of stress changes on two-phase fluid-flow properties of fractures at depth. In this work, we developed a numerical approach for determining relationships between normal effective stress, elastic rock properties, fracture aperture distribution, aspect ratio scaling, oil/water interfacial tension, contact angle and two-phase fluid-flow characteristics of rough-walled fractures. We extended a well-established approach developed for modeling of single-phase fluid-flow in rough-walled fractures. According to this approach, the aperture distribution is replaced by a network of elliptical cavities forming connected pathway from the inlet to the outlet. The extension towards two-phase flow is based on our previous analytical model, in which a two-phase fluid-flow is calculated in a deformable elliptical cavity. The numerical algorithm developed in this work allows quick computation of the impact of the stress-change on two-phase fluid-flow properties of fractured rock. Relative permeabilities of fractures are shown to be non-linear functions of water saturation dependent on the effective normal stress. The capillary pressure-saturation curve for rough-walled fracture is shown to be a function of the effective normal stress. The dependency of fracture permeability, fracture porosity and surface area of open/closed fracture on the effective normal stress is also predicted by the model, which can be used as input parameters for reservoir simulators.