Page 180 - gas transport in porous media

Persoff
174
To further examine the tendency of the system for persistent pressure cycling, we
reviewed videotapes from Experiment B. A videotape record was available for the
period shown in Figure 9.4. When viewing the videotape, no motion can be seen
when gas and liquid are flowing in their own flow paths. Motion can only be seen
when some portion of the fracture changes from gas-occupied to liquid-occupied or
the reverse. Such phase-occupancy-change events (POC events) appeared to occur
instantly and generally lasted only a few seconds. Typically, a slug of water would
emerge from the liquid-occupied area and invade a gas flow path, move quickly
along the gas flow path for one or two centimeters, and then disappear back into
the liquid-occupied area. Afterward, there would often be a slight change in the
overall pattern of phase occupancy in the fracture replica. By tracing the location
of such motion on the monitor, it became clear that all such events occurred along
two paths through the fracture; comparison with the aperture map produced from
light-attenuation measurements showed, as expected, that they were in the regions of
greatest aperture. Often, an event along one part of the flow path would be followed
within a minute or so by an event farther along the same path. The occurrences of
POC events are noted in Figure 9.4. POC events along the “lower” flow path were
associated with maxima in gas inlet pressure and inlet capillary pressure; this indicates
unblocking of a gas flow path.
The smallest pore along a gas flow path is the one most likely to be invaded and
blocked by water. In these experiments gas was injected at constant flow rate, so when
this happened the gas pressure increased upstream of the point of blockage until the
pressure was sufficient to displace the water. But once gas broke through, the gas
pressure (determined by the outlet pressure and flow rate) was insufficient to keep
water out of the pore. This explains the instability for gas injected at constant flow
rate. This blockage was only observed for gas:liquid flow rate ratios less than about
20:1. At greater gas:liquid flow rate ratios, the gas flow path is wide enough that it can
contract without being shut off entirely. Another way to interpret these phenomena
is to suppose that both phases could flow continuously in stable flow paths, but this
would require a particular arrangement of gas and liquid pore occupancy. But if this
arrangement does not minimize the gas-liquid interfacial area, then surface tension
forces will tend to destabilize it. This may explain why in Experiment B, prolonged
periods without pressure cycling were observed.
In Experiment D, gas was injected at constant pressure. Pressure cycling was
therefore not observed, but changes in pore occupancy still occurred; at the upstream
sides of regions of large aperture, where only gas would be “allowed,” liquid would
accumulate. Generally liquid would accumulate along the upstream edge of such a
region, and occasionally jump across. This appeared to be an important mechanism
of liquid transport.
Thus it appears that under certain conditions one phase or the other must be inter-
mittently blocked by the other. Gas flow in a fracture is more susceptible to this kind
of instability than in porous media because (unless the adjoining matrix is signifi-
cantly permeable to gas) the two-dimensionality of the medium prevents gas from