Infinite-Acting Physically Representative Networks for Capillarity-Controlled Displacements
When immiscible fluids coexist in the pore space of granular materials, their configuration is determined by capillarity at the pore scale in many applications. In turn, macroscopic properties such as relative permeability and electrical resistivity are strongly affected by fluid configuration. For...
Main Author: | |
---|---|
Format: | Thesis |
Language: | unknown |
Published: |
The University of Texas at Austin
2008
|
Subjects: | |
Online Access: | https://dx.doi.org/10.26153/tsw/7778 https://repositories.lib.utexas.edu/handle/2152/80763 |
Summary: | When immiscible fluids coexist in the pore space of granular materials, their configuration is determined by capillarity at the pore scale in many applications. In turn, macroscopic properties such as relative permeability and electrical resistivity are strongly affected by fluid configuration. For example, disconnection of a fluid phase during displacement causes the phenomenon of residual phase saturations and changes the topology of a fluid phase. These changes have a profound influence on macroscopic properties associated with the phase. A rough measure of the fluid configuration is phase saturation, and drainage and imbibition capillary curves are the tools to describe the phase saturation as a function of capillary pressure. These curves also considerably influence several processes of environmental interest. An increasingly important example is the explanation of methane hydrate deposits in sediments below permafrost or below the seabed. The mechanism by which hydrates form is not fully understood and it would be valuable to be able to examine the drainage of gas (such as methane) into the hydrate stability zone (from a presumed accumulation below the zone) and the subsequent interaction between gas and residual brine. Predictive models of drainage and imbibition (which are the substantial phenomena of fluid configuration in pore scale) are thus of great utility in subsurface science and engineering. Drainage and imbibition models and simulations have increased their fidelity in recent years but remain prone to some deficiencies. One very important issue in drainage and imbibition modeling is the scale dependency of the simulation results. The residual phase saturations are especially affected by this scale dependency. The scale in which pore network modeling simulations are conducted is very minute compared to the real scale that fluid experiences in reservoir. This huge difference in length scale shows itself through boundary effects in the simulations. In fact, once the fluid enters the model porous media it experiences the boundaries considerably sooner than when it reaches the boundaries in a reservoir scale. The premise of this research is that reservoirs act the same way that infinite (boundary-less) media behave in pore scale modeling. Therefore, if we could model the porous media in a way that the fluid does not experience any boundary we would be able to remove the artifacts of finite networks in the simulation results. The infinite-acting network introduced in this research give a new insight into capillary dominated phenomena such as drainage. One very important finding is to distinguish infinite clusters of a phase from finite clusters which suggests an intrinsic definition of residual phase saturation, namely the volume fraction of pore space occupied by finite clusters of that phase. The key point of this new definition for residual wetting phase saturation, which is only possible in infinite-acting networks is its independence from network boundary conditions. The values predicted for Sw,irr from infinite-acting networks (14%-15%) are an upper bound for the laboratory-measured values for unconsolidated media available in literature (Morrow, 1970) and also larger than the values one gets from conventional networks. Our findings suggest that field values of Sw,irr should be larger than the values measured in core experiments in labs. This can have significant effect in original oil in place estimations. |
---|