Fracture closure and conductivity decline modeling – Application in unpropped and acid etched fractures

Abstract

Natural fractures, hydraulically generated fractures, and acid etched fractures have some degree of fracture face roughness that generates flow conductivity. While it has been proven both theoretically and experimentally that fracture conductivity depends on fracture face roughness, there are limited models that can predict fracture conductivity at different closure stresses for these various fracture roughness patterns. In addition, some of the models require detailed statistical and topographical surface profile parameters, which limit their field application.

A numerical model is developed to study the closure of rough surfaces in contact. Both asperities and semi-infinite half-spaces are assumed to be deformable. The mechanical interaction among asperities is accounted for and its effect on the fracture closure is investigated. Asperity failure is also considered in the model and the results are compared to that of perfectly elastic contact. Aperture profiles that are the output of the closure model are used to solve the fluid flow problem and study the effect of closure stress on fracture conductivity.

It is evident in our results that the closure behavior depends on the etching pattern as well as the elastic properties of the surface. The performance of a rough fracture depends on its initial aperture, asperity height distribution, roughness pattern, and the closure stress range. Certain fracture roughness patterns were able to withstand the closure stress while undergoing lower amounts of closure. Our model tends to predict fracture closure and conductivity behavior better than widely used correlations.

This paper discusses the closure of fractures and attempts to shed more light on the performance of such a stimulation technique by comparing the closure behavior of some particular surface patterns. Our model can be used to determine the most optimum fracture system for a given reservoir condition.