REV (REPRESENTATIVE ELEMENTARY VOLUME) for Cores: How Small is Too Small?
In a previous articles and presentations, various authors have reported on experiments where smaller core samples were taken from a larger “parent” sample, and directly compared back to the original (longer) sample in conventional reservoirs. These are here referenced as “scale down” experiments. Literature siting scale down experiments for areas with small amounts of data in dolomite, limestone, and sandstone reservoirs are reviewed. Additionally, large amounts of data previously reported for Pennsylvanian Cisco and Canyon limestone cores of the SACROC Unit in Scurry County, Texas and Pennsylvanian Morrow sandstones of BreitBurn Energy Partners' West Hough Morrow Unit in Texas County, Oklahoma will be recapped. Newer Permian Tubb sandstone results from BreitBurn Energy Partners Bravo Dome area of Harding County, New Mexico will be shown, along with not-yet-reported comparisons from a dolomite in the Permian Basin.
Cisco/Canyon, and Morrow formation samples analyzed as twelve inch samples, were cut into six inch whole core samples and re-analyzed. Those pieces were further cut into two inch whole core samples and analyzed again. Finally, a few additional samples were plugged (1-½″ diameter) and analyzed one last time. Experiments indicated that limestone plugs and small (less than 6″ in length) whole core samples did not accurately predict the vertical permeability of full length (12″) cores. Smaller samples who's parent sample was less than 2 mD typically displayed variability of 3 Orders of Magnitude (OoM), over-estimating by 2 OoM or under-estimating by 1 OoM. Samples from higher quality reservoir intervals (>10mD) were more accurately predicted by smaller subsets. Small samples typically over-estimate their parents by 1 OoM or under-estimated them by ½ OoM, with the range decreasing as parent sample quality increases. Horizontal permeability predictions from small subsets (2″ tall whole-core samples and 1-½″ diameter plugs) fare better when compared to 6″ length parent samples (the maximum size available). These datasets fall within +/− 1 OoM of parent samples.
Ehrenberg's (2007) dolomite data set from a very high quality reservoir (>10 mD) and newly released dolomite data from a relatively lower quality portion of the Seminole San Andres Unit (<100 mD) are more uniform. Data scatter is generally contained within +/− 1 OoM for lower quality samples and + ½ to − 1 OoM for higher quality samples.
Small samples in sandstone examples tend to be optimistic compared to parent sets. For both areas studied in this examination, samples tended to err on the high side, with most samples falling within 0 to + 1 OoM. A minor amount of samples fall just below unity, and rare examples are seen below − ½ OoM.
Understanding how small sample sizes impact core analysis may lead to insights into better methods to estimate the larger scale permeability of our cores. This, in turn, may help us determine the most helpful methods to upscale to a Representative Elementary Volume (REV) for reservoir characterization of flow units in the reservoir.
AAPG Datapages/Search and Discovery Article #90250 © 2016 Southwest Section AAPG Annual Convention, Abilene, Texas, April 9-12, 2016