Sub-core permeability estimation from coreflooding experiments: Numerical analysis for evaluating accuracy and improving methods

Abstract

Coreflooding experiments are one of the primary methods for reservoir rock characterization and have been developing in recent years, providing increasing detail. An advanced analysis of coreflooding experiments consists of constructing three dimensional permeability ($k$) maps of the core sample with sub-core resolution. Such detailed characterizations provide important information on the core heterogeneity and enable the construction of accurate numerical flow models that can reproduce experimental results. This work presents a method for estimating $k$, combining data from multiple coreflooding experiments, and studies the estimation accuracy considering a large number of test cases consisting of synthetic data obtained by numerical experiments. Cases include varying parameters, such as: fluid injection flow rates and types, relative permeability curves, capillary pressure functions and core heterogeneity. The new method is shown to significantly improve the estimation accuracy in comparison to methods that incorporate data only from a single experiment. Furthermore, a method for improving $k$ estimations in regions where water is trapped is presented and its implementation is shown to increase the overall accuracy of $k$ estimation. An additional method is considered, that requires only single-phase flow solutions, thus increasing the computational efficiency. However, it is shown to be applicable only for a small subset of cases.