Page 132 - gas transport in porous media
P. 132

125
                           Chapter 7: Gas-Phase Dispersion in Porous Media
                           and Webber, 1953). Effective diffusion coefficients can be estimated using tortuosi-
                           ties (see above) or can measured directly in the laboratory (Taylor, 1949; Bruce and
                           Webber, 1953; Currie, 1961; Shearer et al., 1973; Sallam et al., 1984; Johnson and
                           Perrott, 1991; Barone et al., 1992; Jin and Jury, 1996; Batterman et al., 1996; Schaefer
                           et al., 1997; Moldrup et al., 1998, 2000) and field (Raney, 1949; Lai et al., 1976;
                           Weeks et al., 1982; Kreamer et al., 1988).
                             The longitudinal dispersivity term scales with the degree of heterogeneity of the
                           physical system and is often measured using column-scale nonreactive tracer tests.
                           Gas-phase longitudinal dispersivities have been measured in laboratory porous media
                           systems and are found to range approximately between 0.2 and 2.9 cm (Popovi˘cová
                           and Brusseau, 1997; Ruiz et al., 1999; Garcia-Herruzo et al., 2000; Costanza-
                           Robinson and Brusseau, 2002). Dispersivities measured in the field tend to be
                           larger due to increased system heterogeneity. Ideally, dispersivities should be solute-
                           independent and insensitive to changes in carrier gas velocity or nonequilibrium
                           effects. However, if the data analysis fails to consider all relevant transport processes,
                           dispersivity can become a “lumped” solute-dependent parameter that no longer solely
                           represents physical heterogeneity of the porous medium (Costanza-Robinson and
                           Brusseau, 2002).
                             The Peclet number, P e , is a dimensionless measure of the degree of dispersion
                           experienced by a solute, defined alternatively as the ratio between the advective and
                           dispersive processes or the ratio of advective to diffusive processes (Rose, 1973). The
                           former definition is more encompassing and will be used here. The Peclet number is
                           usually obtained by fitting a solute breakthrough curve with an advective-dispersive
                           solute transport model. The magnitude of the Peclet number is inversely proportional
                           to the degree of dispersion. Thus, low Peclet numbers correspond to a large degree
                           of solute spreading. The Peclet number, also termed the Brenner number (Rose and
                           Passioura, 1971), can be related to the dispersion coefficient as follows:
                                                              vL
                                                         P e =                            (7.4)
                                                              D
                           where L is a characteristic length of the system (L). The characteristic length can
                           be defined at small- or macro-scales (e.g., grain diameter or column length) (Rose,
                           1973). The specific formulation of the Peclet number varies by application and field
                           of study. The macroscale length is used commonly in the fields of soil physics and
                           hydrology, while grain-scale lengths are often used in engineering disciplines. Thus,
                           caution should be exercised when interpreting absolute values of Peclet numbers or
                           when comparing values from different studies.


                           7.2.4  Functional Dependence of Dispersion and the Dispersion Coefficient
                           The dispersion coefficient is a function of the solute velocity, v. While the diffusion
                           term is independent of v, the mechanical mixing component of dispersion is propor-
                           tional to velocity (see Equation 7.3). Thus, D decreases with decreases in velocity.
                           However, slower velocities lead to larger residence times, thereby allowing more time
   127   128   129   130   131   132   133   134   135   136   137