A mathematical and computational study of the dispersivity tensor in anisotropic porous media
Introduction
Dispersion in porous media is among the most significant topics in the study of solute transport. The classical Advection–Dispersion model assumes that dispersion follows Fick’s Law at the macroscopic level, such that dispersive flux is proportional to the concentration gradient [1]. The properties of dispersion coefficients, especially the longitudinal dispersion coefficient, have been studied extensively and at many scales. Representative literature includes: [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18]. At the Darcy or laboratory scale, the most widely applied parameterization of dispersion in porous media was proposed by Scheidegger [19]. In its simplest form, Scheidegger’s parameterization [19] states that the dispersion coefficient is the sum of a mechanical-dispersion term and a pore-diffusion term, and that the mechanical dispersion coefficient is a linear function of mean fluid velocity through the definition of dispersivity.
Even though most early work on dispersion focused on longitudinal dispersion, recent studies (e.g., [20], [21]) have revealed that transverse dispersion and the mixing it describes can be more important to chemical reactions and biodegradation of groundwater contaminants than longitudinal dispersion. In practical applications, the transverse dispersion coefficient is also assumed to be a linear function of mean fluid velocity following Scheidegger’s parameterization [1], [22]. A widely used estimation of transverse dispersivity is (where is the pore length [L]) [23]. However, transverse dispersion may not necessarily follow the same properties as the longitudinal dispersion. For example, according to [7], field-scale transverse dispersivity diminishes in proportion to instead of solely depending on pore geometry as generally applied.
Previous studies also hypothesized that the dispersivity of an anisotropic porous medium should be a higher-order tensor (see Section 2 for more details). The higher-order tensorial form of dispersivity has been subjected to very few studies. We are not aware of any study that provides a solid proof (starting with well-defined assumptions) to the tensorial form of dispersivity. Although the complexity of a higher-order tensor may impede its application in practical groundwater modeling, a comprehensive study will allow us to improve our understanding of the basic properties of dispersion and to develop better parameterizations, particularly since new upscaling theories and computational tools may allow us to revisit this important topic.
The objective of this paper is to revisit dispersion in anisotropic porous media and to explore the tensorial properties of dispersivity from a scientific standpoint. This study utilizes the method of moments [24], a rigorous approach based on a few well-defined assumptions. In this approach, large-scale heterogeneity is neglected through the discretization of the spectrum of medium properties, which is mathematically equivalent to using a periodic approximation of heterogeneity. We focus on dispersion at large time, when the system is at local equilibrium and changes in the system are gradual. Local equilibrium indicates a status that all the velocity fields, both the high and low velocities, are equally sampled by solute mass in a statistical sense (refer to [9] for more information of local equilibrium). Our mathematical analysis demonstrates that in an anisotropic medium with a mean flow that has an angle with the coordinates, the mechanical dispersion tensor can be expressed through a fourth-order dispersivity tensor. A constructive derivation in Section 4 yields a mathematical formula of the dispersivity tensor in periodic porous media starting from first principles under well-defined assumptions. The formula indicates the dependency of the dispersivity on flow conditions as well as pore geometry. Furthermore, we design a numerical experiment to explicitly calculate the fourth-order dispersivity tensor based on pore-scale simulations of the flow and transport, and to evaluate the effects of fluid velocity and molecular diffusion. The results provide a comprehensive understanding of the physical basis and properties of the dispersivity tensor; to the best of our knowledge, this is the first of its kind. Finally, through the comparison of numerical results between 2-D and 3-D models, and between orderly and randomly packed porous media, we demonstrate the effect of pore geometry on dispersivity, providing insights for future pore-scale simulations of dispersion in porous media.
Section snippets
Background of dispersivity tensor
Bear [25] studied the variance of mass in a 2-D isotropic porous medium, and arrived at the relationship between the variance and displacement of the mass,where is the variance of concentration distribution [L2], and is the mass displacement [L], is the Kronecker symbol [–], and and are the longitudinal and transverse dispersivities of the isotropic medium [L]. When the mean flow has an angle β with the Cartesian coordinate system, the
Method of moments
The mathematical analysis and numerical simulations in this paper are based on the method of moments [24]. A curcial advantage of this method is that it is a rigorous and exact approach to calculate large-time dispersion in porous media that are extended in infinite domains and are spatially periodic. The dispersion coefficient is calculated explicitly based on a volume averaging approach. In the case of homogeneous and isotropic molecular diffusion, the mechanical dispersion tensor [L2/T]
A mathematical formula of the fourth-order dispersivity tensor
Although the tensorial form of dispersivity was hypothesized over 50 years ago [25], a solid proof and general agreement of why a higher-order tensor is necessary and how it explicitly relates to medium geometry, velocity magnitude, and molecular diffusion have yet to be achieved. In this section, a constructive derivation based on Brenner’s method of moments (Eq. (11)) and Fourier expansions shows that the dispersivity has a fourth-order tensorial form.
First, it is recognized that velocity
Pore-scale simulation approach
In order to gain a deeper understanding of the properties of dispersivity tensor, we design a pore-scale numerical experiment to explicitly calculate the dispersivity tensor through the method of moments (Eqs. (11), (12)). The finite element code COMSOL Multiphysics 4.2a (Comsol Inc., Palo Alto, California) is used. Fig. 2 shows the porous structures of the studied unit cells.
This study focuses on flows with low Reynolds number, which can be described by the steady-state linear Stokes equation
2-D isotropic case
In a 2-D isotropic medium, the dispersivity tensor contains only two major components, the longitudinal and transverse dispersivities. When the Cartesian coordinates align with the principle directions of the porous medium and the flow is in the direction, the mechanical dispersion coefficient tenor is,It can be noted that is the longitudinal dispersivity [L] and is the transverse dispersivity [L]. Fig. 3 shows the change of the longitudinal and transverse
Conclusions
From a scientific focus, this study uses the method of moments [24] to investigate dispersion in anisotropic porous media with flows that may have angles with the principal directions of the media. The first contribution of this work is the solid mathematical proof to the fourth-order dispersivity tensor, which was hypothesized by Bear [25]. The mathematical analysis indicates that the dispersivity does not solely depend on the pore geometry as generally assumed in practical applications, but
Acknowledgement
This material is based upon work supported by the National Science Foundation under Grant No. 0738772. Additional funding was provided by a Stanford Graduate Fellowship. The MATLAB code provided by Dr. David L. Hochstetler to generate the randomly packed medium is gratefully acknowledged. The authors thank the four anonymous reviewers for their helpful comments.
References (56)
Prediction by the method of moments of transport in a heterogeneous formation
J Hydrol
(1988)Choice of dispersion coefficients in reactive transport calculations on smoothed fields
J Contam Hydrol
(2002)- et al.
Numerical simulation of biodegradation controlled by transverse mixing
J Contam Hydrol
(1999) - et al.
Calculation of the effective properties describing active dispersion in porous media: from simple to complex unit cells
Adv Water Resour
(2001) - et al.
Comparison of theory and experiments for dispersion in homogeneous porous media
Adv Water Resour
(2010) On the tensorial nature of advective porosity
Adv Water Resour
(2005)- et al.
Effects of compound-specific transverse mixing on steady-state reactive plumes: Insights from pore-scale simulations and darcy-scale experiments
Adv Water Resour
(2013) - Bear J. Dynamics of fluids in porous media. Dover Publications;...
A mathematical model for dispersion in the direction of flow in porous media
Old SPE J
(1963)Diffusion and dispersion in porous media
AIChE J
(1967)
Three-dimensional stochastic analysis of macrodispersion in aquifers
Water Resour Res
Solute transport in heterogeneous porous formations
J Fluid Mech
Stochastic theory of field-scale fickian dispersion in anisotropic porous media
Water Resour Res
Analysis of macrodispersion through volume-averaging: moment equations
Stoch Hydrol Hydraul
The concept of Block-Effective macrodispersivity and a unified approach for Grid-Scale- and Plume-Scale-Dependent transport
J Fluid Mech
On the use of block-effective macrodispersion for numerical simulations of transport in heterogeneous formations
Water Resour Res
Volume averaging for determining the effective dispersion tensor: Closure using periodic unit cells and comparison with ensemble averaging
Water Resour Res
A critical review of dispersion in packed beds
Heat Mass Transfer
Impact of upscaling on solute transport: Traveltimes, scale dependence of dispersivity, and propagation of uncertainty
Water Resour Res
Modelling of block-scale macrodispersion as a random function
J Fluid Mech
On the theory of flow of miscible phases in porous media
I.A.H. Assembly of Toronto
Pore-scale simulation of dispersion and reaction along a transverse mixing zone in two-dimensional porous media
Water Resour Res
Contaminant hydrogeology
A theory of dispersion in a porous medium
J Fluid Mech
A general theory of taylor dispersion phenomena
Physico-Chem Hydrodyn
Cited by (9)
Effect of distance-dependent dispersivity on density-driven flow in porous media
2020, Journal of HydrologyCitation Excerpt :The simulation of DDF problems is based on coupling Darcy’s groundwater flow equation to the solute transport equation via a state relation expressing the density as a function of solute concentration. Transport of solute in the aquifer is ruled by advection, representing the solute displacement by the mean fluid flow, and by dispersion, which accounts for solute spreading caused by velocity variations due to the heterogeneity of the porous medium at different scales (Liu and Kitanidis, 2013; Kitanidis, 2017; Dai et al., 2020). Dispersion processes have been found to play a major role in DDF problems as they cause mixing between different fluids.
Teaching and communicating dispersion in hydrogeology, with emphasis on the applicability of the Fickian model
2017, Advances in Water ResourcesCitation Excerpt :The issue of whether the tensor follows Scheidegger’s parameterization is a more complex one. To begin with, the usual Scheidegger’s parameterization disregards effects of anisotropy in the porous medium itself (more on this issue in Liu and Kitanidis, 2013). Anisotropy in the Scheidegger dispersion matrix is strictly related to the direction of flow, which is a principal direction.
Editorial: A tribute to Stephen Whitaker
2013, Advances in Water ResourcesIntrapore Geometry and Flow Rate Controls on the Transition of Non-Fickian to Fickian Dispersion
2023, Water Resources ResearchCalculating Trajectories Associated With Solute Transport in a Heterogeneous Medium
2018, Water Resources Research