06.08.2013 Views

structural geology, propagation mechanics and - Stanford School of ...

structural geology, propagation mechanics and - Stanford School of ...

structural geology, propagation mechanics and - Stanford School of ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

4. Application to the Aztec s<strong>and</strong>stone<br />

In order to test this computational estimation method under real-world conditions <strong>of</strong><br />

limited sample supply, we applied the algorithm to a single, representative thin section <strong>of</strong><br />

Aztec s<strong>and</strong>stone showing a cm-thick CB surrounded by relatively undeformed matrix<br />

(Figure 5.4). As is typical with foreset deposition in æolian s<strong>and</strong>stones, bedding consists<br />

<strong>of</strong> alternating, well-sorted coarse-grain (~0.4 mm) <strong>and</strong> fine-grain (~0.1 mm) layers up to<br />

several millimeters thick. Poorly sorted mixed-grain layers are also common. Within each<br />

<strong>of</strong> the three distinct bed types, the distribution <strong>of</strong> grains <strong>and</strong> pores is relatively<br />

homogeneous, both in the matrix <strong>and</strong> in the CB.<br />

Digital images were collected in BEC mode on a st<strong>and</strong>ard SEM at 15 kV <strong>and</strong> a<br />

magnification <strong>of</strong> 80x. Two images were collected from each bed type, with mosaic<br />

composites used as necessary to ensure representative coverage in coarse-grain beds. A<br />

consistent intensity threshold <strong>of</strong> 72 out <strong>of</strong> 256 was used for the binary conversions, <strong>and</strong><br />

ten 3-D pore-structure realizations were computed for each <strong>of</strong> the 12 binary images. Flow<br />

simulations were then conducted on each pore-structure realization to yield 120 total<br />

estimates <strong>of</strong> permeability relative to porosity. Figure 5.6 shows an example BEC image<br />

with its binary conversion for each bed type from both the matrix <strong>and</strong> the CB.<br />

4.1. Simulation results<br />

The resulting estimates <strong>of</strong> porosity <strong>and</strong> permeability are plotted in Figure 5.7; the<br />

averaged permeability results are summarized in Table 5.1.<br />

Porosity <strong>and</strong> permeability estimates for the matrix range from about 20-27% <strong>and</strong> 200-<br />

3,500 millidarcys (mD). Not surprisingly, average porosity for the pore-structure<br />

realizations derived from the mixed-grain bed images is lowest (22%), while average<br />

permeability estimated for the coarse-grain pore-structure realizations is highest (1,392<br />

mD). The fine-grain beds, which exhibit the tightest grain-size distribution, produced the<br />

highest average porosity estimate (25.25%) <strong>and</strong> the lowest average permeability estimate<br />

(406 mD). The combined average estimates for the matrix are 23.6% <strong>and</strong> 776 mD. This<br />

simple average bulk permeability value, however, does not account for the inherently<br />

anisotropic, layered nature <strong>of</strong> the s<strong>and</strong>stone. Figure 5.8 presents a more realistic approach<br />

to estimating directional permeability in the matrix, yielding best-estimate values <strong>of</strong><br />

1,053 mD parallel to bedding <strong>and</strong> 804 mD normal to bedding.<br />

134

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!