Characterising the Pore Space of Selected Sandstone Samples using Multiple Approaches

A comprehensive knowledge of the porosity and pore size distribution (PSD) of hydrocarbon reservoirs is vital to several petroleum engineering disciplines including reserve estimation, reservoir characterisation, drilling operations and reservoir
development planning. This work examines the three methods of Mercury Injection Capillary Pressure (MICP) Testing, Pore Network Modelling (PNM) and Nuclear Magnetic Resonance (NMR) which are currently used within the petroleum industry to determine representative measures of porosity and PSD.
Although MICP is a common method used within the petroleum industry, several factors impact its suitability for determining porosity and PSD. These are related to the destructive nature of the test which is challenging when samples are limited in quantity as well as the limitations of MICP to provide robust results for certain kinds of reservoir material, particularly for those that are unconsolidated and unconventional. Recent advances in PNM and NMR have made these approaches attractive alternatives for pore evaluation studies which can enhance, supplement or replace the information derived from MICP testing. To examine the applicability of PNM and NMR methods to determine porosity and PSD, three sandstone core samples were used throughout this study. These were the Berea and Bentheimer core samples, which are consolidated and homogenous in nature allowing an opportunity for the benchmark testing of the PNM and NMR approaches and an Athabasca Oil Sand (AOS) sample, which is a prime example of unconsolidated material containing a very viscous in-situ fluid.
During the PNM process, micro-computed tomography (micro-CT) was used to obtain 2D contiguous images of a sample which were then compiled to produce a 3D representation of the pore space. Based on the literature, a 12-step comprehensive PNM approach was developed in this work and applied to the benchmark Berea and Bentheimer core samples to derive their porosity and PSD. This had a substantial processing time of over 100 hours (> 4 days) for each sample. The key findings from this comprehensive approach formed the basis of a simplified recommended PNM practice having only 9 steps and an anticipated processing time of 7 hours and 26 hours for homogenous and heterogeneous samples respectively. This simplified recommended PNM practice was then applied to the AOS sample with the porosity and PSD results showing a good agreement to that from the MICP and NMR testing.
The determination of porosity and PSD from NMR testing requires specific fluids to be contained in the pore space. This generally involves the removal and replacement of all original fluids with water (or brine) since the response of the low-viscosity water correlates well with surface measurements of the pore space. This typically precludes the testing of samples imbued with their original fluids which poses several restrictions for the NMR testing of unconsolidated and partially consolidated material.
The development of techniques which allow for the robust pore space testing of these kinds of materials without the cleaning or removal of their native fluids is therefore valuable to the petroleum industry. Based on these ideas, a novel empirical transform was developed which could allow the NMR testing of samples containing viscous fluids. This transform used the NMR response of glycerol (which is 1,412 times more viscous than water at 20oC) to develop a transform based on viscosity. The use of this transform showed great success in obtaining a robust PSD for the AOS sample containing its native bitumen which is comparable to the PSDs from the MICP and PNM approaches.
These results indicate that the PNM and NMR approaches can provide comparable results to conventional MICP testing, that they can be used as independent techniques for evaluating the pore space and that they can provide a robust measurement of the
porosity and PSD for samples imbued with their native hydrocarbon fluids. When compared to MICP testing, these approaches might be preferred when testing a limited quantity of core samples, partially consolidated and unconsolidated samples and samples containing their original fluids.
Future work to strengthen these results include using a wider range of sandstone samples to test the developed recommended PNM practice and using a wider range of samples containing a greater variety of fluids to test the empirical transform.