download pdf version of PhD book - Universiteit Utrecht

5.2 Network Generation
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
lated to the pore body radii. Thus, the three networks have different aspect
ratio’s (ratio of pore body to pore throat radius). Properties of the two porousmedium-specific
networks will be given later.
Figure 5.1: The distribution of pore-body radii (columns) together with three
distributions of pore-throat radii (R 1, R 2, and R 3) (lines). The mean radii of
the pore-body and pore-throat distributions are shown above each distribution.
5.2.2 Determination of the pore cross section and corner
half angles
A key characteristic of real porous media is the angular form of pores. It has
been demonstrated that having pores with a circular cross section, and thus
single-phase occupancy, causes insufficient connectivity of the wetting phase
and as a result poor representation of experimental data [Zhou et al., 2000b].
Angular cross sections retain the wetting fluid in their corner and allow two or
more fluids to flow simultaneously through the same pore. Pores which are angular
in cross section are thus a much more realistic model than the commonly
employed cylindrical shape. In the present work, pore bodies are considered to
be cubic in shape, whereas, pore throats are assigned a variety of cross sectional
shapes including circular, rectangular, and scalene triangular.
The shape of an angular pore cross section is prescribed in terms of a dimensionless
shape factor, G, [Mason and Morrow, 1991] which is defined as
G = A P 2 (5.2)
97