Page 295 - Fundamentals of Gas Shale Reservoirs

CHARACTERIZATION OF FLUID BEHAVIOR AND EQUATIONS OF STATE VALID FOR NANOPOROUS MEDIA 275
addition to other relevant factors, such as heterogeneity, pore certain improvements on these two equations by incorporating
connectivity, surface roughness, and wettability, the macro‐ a correlation of an additional volume correction for high‐
and micropore geometry of the media is a critical factor in pressure–high‐temperature as a function of the reduced tem­
determining the effective porous media modifications. perature, molecular weight, and acentric factor.
Depending on this, properties such as permeability, pressure, Previous studies have made the modifications on such
storage capacity, phase behavior, viscosity, interfacial equations to take into account the pore‐proximity effect by
tension, and capillary pressure can be altered significantly. means of the correlations of the critical properties (density,
This in turn may have a profound impact on the performance temperature, and pressure) and the acentric factor against the
of wells completed in hydrocarbon‐bearing shale reservoirs. ratio of the pore size to molecule size, and the reduction of the
An accurate understanding of the relevant processes and effective molecular volume of the adsorbate. Proving whether
their roles is instrumental in determining the composition, these are the proper and sufficient corrections for pore‐
storage, and transport of fluids. It is required for effective proximity is open for future research. Then, the pore‐proximity
reserves estimation; prediction of variations in fluid compo­ effects observed in various size pores in shale can be averaged
sition, condensate dropout, and blockage; and long‐term over the representative elementary volume of shale porous
production forecasting and planning of shale reservoirs by media considering the pore‐size distribution (Civan, 2002b).
means of a model‐based approach (Devegowda et al., 2012; Theoretically, the conventional bulk fluid properties are
Zhang et al., 2013a, b). attained as the ratio of the pore size to molecule size approaches
In this section, we discuss approaches available for esti­ infinity. Various applications have shown that the fluid prop­
mating the conditions and properties of fluids contained in erties are close to their bulk values for pore sizes greater than
nanopores. We start by pointing out the primary factors of 10 nm. Other types of properties of practical importance, such
interest which include the pore‐fluid composition (number as the existing real‐gas deviation factor and viscosity correla­
density of molecules) and state (gas, liquid, or solid), phase‐ tions, such as those given by Dranchuk and Abou‐Kassem
equilibrium, pressure and temperature, pore geometry and (1975) and Bergman and Sutton (2007) can also be modified
structure (shape, surface roughness, texture, fabric, and pore for pore‐proximity once the critical properties are correlated
size distribution), pore‐wall proximity (ratio of the pore size against the ratio of the pore size to molecule size.
to molecule sizes), pore–wall (organic or inorganic) wetta­ Examples are elaborated for single‐ and multiple‐component
bility, adsorption, and the potentials that determine the pore– mixtures concerning the gas and gas–condensate applica­
wall interactions with molecules of various types. Different tions by Michel et al. (2011a, b), Devegowda et al. (2012),
fluid pressures and compositions are attained in different and Zhang et al. (2013a, b). In Michel et al. (2011a, b), the
size pores, so the reservoir properties are an average over impact of pore proximity on gas z‐factors is shown to have
properties that vary on the nanoscale. an influence on gas formation volume factors and densities
Prediction of the phase behavior and phase‐equilibrium in (Figure 1 in Michel et al., 2012). Across a wide range of
shale is a complicated issue and molecular dynamics (MD) pressures and temperatures, significant differences in the z‐
calculations have provided considerable insight into the factors are observed implying that gas‐in‐place estimates,
nanoscale behavior. One of the key advantages of MD simu­ gas reservoir material balance calculations, and prediction of
lations is that at the nanoscale it may be prohibitively expen­ gas flow rates are likely to be erroneous if pore proximity
sive or even impossible to conduct lab‐scale measurements effects are not considered.
and MD simulations provide a virtual framework for under­ In Devegowda et al. (2012), the phase behavior of a rich
standing nanoscale phenomena. However, MD calculations gas–condensate sample was analyzed in different pore sizes,
currently are only practical on systems with very simple pore showing significant differences in the phase envelopes and
geometries, limited fluid complexities, and idealized percentages of liquid dropouts (% by volume) in pores of 2,
pore wall potentials. As such they serve very well in exploring 4, and 5 nm for the gas–condensate mixture (Figures 4 and 5
the phenomena occurring in nanoscale pores and supplying in Devegowda et al., 2012). The modified phase envelopes
estimates on changes to parameters such as the critical tem­ were calculated using the Peng‐Robinson EOS by inputting
perature and pressure, but cannot adequately substitute for the critical point data for the confined fluids (Singh et al.,
laboratory measurements. To use industry‐scale reservoir 2009) into a commercially available PVT package (CMG,
simulation software requires alterations to the algebraic 2008). Comparisons to the corresponding bulk values are
equations of state to account for the above‐mentioned unique also shown and illustrate dramatic changes in the calculated
features of hydrocarbon‐bearing shale systems. phase diagrams for the confined fluids. Specifically, the
The bulk fluid behavior has been satisfactorily described liquid‐dropout is significantly lower due to pore‐proximity
by certain empirical modifications of the van der Waals effects in comparison to the corresponding values in bulk
equation of state (VDW‐EoS). The two outstanding candi­ (Figure 5 in Devegowda et al., 2012). This has serious impli­
dates are the Soave–Redlich–Kwong (SRK‐EoS) and the cations in the calculations of well productivity and for reser­
Peng–Robinson (PR‐EoS) EoS. Baled et al. (2012) provide voir management. It is highly likely that reserves calculations