Large-Scale Geological Carbon Sequestration - Introduction to Carbon Capture and Sequestration - page 462

contains a mixture of mineral phases and pore space, the pore space
being fi lled either by brine, CO 2 , or a two-phase mixture of brine and CO 2 ,
i.e., for each grid block our model must keep track of the porosity

Φ and

the water and CO 2 saturations S w and S g .
The discretization of space into an ensemble of simulation grid
blocks implies several approximations. First, fl uid and rock properties are
not homogeneous at the grid-block scale. This means that the effective
grid-block scale properties of the porous media of interest cannot neces-
sarily be determined from measurements made on a scale smaller than
the grid blocks (for example, from experiments on 5-10 cm core samples)
because many properties are infl uenced by the heterogeneity of matter
within the scale of the grid blocks. This is illustrated by the permeability
anisotropy of natural formations: at the grid-block scale, horizontal per-
meability tends to be higher than vertical permeability, because discrete
regions of higher or lower permeability within grid blocks are oriented in
the direction of the bedding ( Figure 10.2.4 ). A single core sample may
not well represent a region of the formation large enough to capture this
effect. This effect is often modeled by assuming that permeability in the
direction parallel to the bedding is greater (for example, by a factor of 10)
than in the direction normal to the bedding.

Rock properties are often detemined from core
samples and well logs (on a ~10 -1 m length scale)

Significant heterogeneity (e.g., lower permeability
zones) may exist within grid blocks.

Figure 10.2.4 Sub-grid-block scale permeability features
Schematic figure illustrating the tendency of high- or low-permeability zones on a
sub-grid-block scale (such as sand or shale lenses) to be oriented in the direction
of the bedding.