14.11.2013 Views

download pdf version of PhD book - Universiteit Utrecht

download pdf version of PhD book - Universiteit Utrecht

download pdf version of PhD book - Universiteit Utrecht

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

5.1 Introduction<br />

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .<br />

ber <strong>of</strong> six.<br />

There exist many PNM studies investigating various geometrical and topological<br />

properties <strong>of</strong> porous media and their effect on flow and transport in porous<br />

media. They include: the effect <strong>of</strong> pore size distribution [Lindquist et al., 2000,<br />

Jerauld and Salter, 1990], network structure and coordination number [Al-<br />

Kharusi and Blunt, 2008, Sok et al., 2002, Mahmud et al., 2007, Arns et al.,<br />

2004, Øren and Bakke, 2003a, Mogensen and Stenby, 1998], pore angularity<br />

and shape factors [Sholokhova et al., 2009, Zhang et al., 2010, Patzek and Kristensen,<br />

2001, Øren et al., 1998a], and correlation functions [Arns et al., 2003b,<br />

Rajaram et al., Ferrand et al., 1994, Lymberopoulos and Payatakes, 1992, Jerauld<br />

and Salter, 1990, Tsakiroglou and Payatakes, 1991, Renault, 1991, Ioannidis<br />

et al., 1993]. In almost all <strong>of</strong> these studies, the selected pore size distributions<br />

are such that most <strong>of</strong> the pore space is assigned to the pore bodies. In<br />

such cases, the modeling <strong>of</strong> flow within pore bodies will be <strong>of</strong> major influence<br />

on hydrodynamic properties <strong>of</strong> the network, certainly in the case <strong>of</strong> flow on<br />

partially drained pore spaces. This issue, however, unlike the above-mentioned<br />

properties <strong>of</strong> porous media has not been addressed in any pore-network study.<br />

We will address this particular issue in this study.<br />

5.1.3 Objectives and approach<br />

Under partially-saturated conditions, much <strong>of</strong> a pore body is occupied by the<br />

non-wetting phase, such that the wetting phase can only flow through the<br />

edges, similar to the flow conditions within drained pore throats. Under these<br />

conditions, the effect <strong>of</strong> resistance within the pore-bodies become important<br />

and comparable to pore throats; taking this effect into account can improve<br />

the result <strong>of</strong> k r − S calculations. This is, however, not possible within current<br />

pore-network modeling approaches. In fact, this is one <strong>of</strong> the shortcomings<br />

<strong>of</strong> pore-network modeling compared to other pore-scale modeling approaches,<br />

such as lattice-Boltzmann method (LBM). LBM has the advantage that it fully<br />

discretizes the space within each pore space. In PNM, however, the smallest<br />

discretization unit is normally one pore body or pore throat. That is, in PNM,<br />

only one (average) pressure (or one average concentration) is assigned to each<br />

pore body. Thus, using pore body or pore throat discretization, one needs<br />

to apply so-called effective parameters [Meile and Tuncay, 2006, Rao<strong>of</strong> and<br />

Hassanizadeh, 2008, 2010b]. While this assumption may be acceptable for a<br />

fully saturated pore body (considering the large size <strong>of</strong> a pore body compared<br />

to pore throats connected to it), it is a crude approximation when the water<br />

95

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

Saved successfully!

Ooh no, something went wrong!