SCAL measurements on core material from the Dunga Field in Kazakhstan clearly show two distributions of flow regimes, one with flow (good flow sand) and one without. If no precautions are taken the SOR and SWIR measurements are not representative of the reservoir. It is essential to map out the location and the nature of the high and low permeable zones in order to properly model the reservoir flow pattern. Visual inspection of the core reveals mm-scale bedding of darker and lighter facies. This paper describes an imaging study to address whether the bedding and the flow/saturation properties could be related and an attempt to quantify the good flow regime.

Scanning electron microscopy (SEM) images taken at a magnification of 10000 times and stitched together to show the area across the bedding apparent in core do not reveal anything other than homogeneously distributed grains, feldspar and clays. This result indicates that either the flow heterogeneities are not related to the laminations, or that the grain size does not play a role.

Computational tomography (CT) scanning was used in order to determine the density and spatial differences in the rock. The 3D tomograms were recorded in-house at an approximately 3 µm resolution and supplemented by experiments conducted at the European Synchrotron Radiation Facility (ESRF) where a resolution of 700 nm was achievable. Both of these tomograms clearly show structured, denser regions, which would likely be responsible for the observed flow heterogeneities. The CT results were combined with an automated mineralogy approach to determine the composition of the cores across the laminations. The primary focus of this experiment was to investigate whether the denser bands are related to certain mineralogy in the core sample. The QEMScan results do indicate that the dense bands are a result of clay rich sand laminations. This seems to constitute the regime of the porous media where the pore space cannot be resolved (the microporous regime).

The segmentation of the 3D data and following pore-connectivity analysis is used in an attempt to quantify the amount of good flow sand. The 3D data does not give rise to a successful prediction of the permeability which is indicating that even the good flow sand has permeability in the single-digit mD range. This is the case for both the 700 nm and 3 µm resolution datasets. There is a linear correlation between the connected porosity versus both the experimental results for SOR and SWIR. Thus, it may be possible to predict SOR and SWIR from CT measurements on plugs where SCAL measurements have not been conducted. It is, however, noted that the quantification of good flow sand (via the connected porosity) from the images does agree very well with our description of SWIR based on the good flow sand/poor flow sand combination. This is ascribed to the different length scales involved in the measurements and that fact that the limit is 3 µm for the images that we can readily treat. Getting the scales right will be a venue of future investigation.