Predicting permeability from other petrophysical properties
MetadataShow full item record
Understanding pore network structure of a porous medium and fluid flow in the pore network has been an interest to researchers for decades. This study focuses on the characterization and simulation of the pore networks in petroleum reservoir rocks using conventional characterization techniques. A Representative Elemental Volume (REV) model is developed which simulates the pore network as a series of non-interconnected capillary tubes of varying sizes. The model implements mercury porosimetry (MP) results and capillary pressure principles to calculate the size of each bundle of capillary tubes based on a pore throat size distribution produced by the MP experiment. It also implements electrical properties of the rocks to estimate the average length of the capillary tubes. To verify the validity of the simulated network, permeability is calculated for the simulated network using Poiseuille’s flow principles for capillary tubes. Preliminary work showed that the model is capable of simulating the pore network reasonably well because permeability estimations for the simulated network matched measurements. In this study, MP and nuclear magnetic resonance (NMR) tests as well as centrifuge and permeability tests are performed on a suite of 11 sandstone and carbonate rock samples. Because electrical tests were not available, average length of flow paths is calculated with an alternative method that uses porosity to calculate tortuosity. Permeability estimations of the simulated network are compared with measurements. Estimations are also compared to other predictions using methods that implement MP and NMR data to simulate the pore network and the results show that the developed REV model out performs all the other techniques.