Estimation of macrodispersion in 2-D highly heterogeneous porous media using the Eulerian-Lagrangian localized adjoint method

Fanilo Ramasomanana, Anis Younes, Philippe Ackerer

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

The Eulerian-Lagrangian localized adjoint method (ELLAM) formulation of Younes et al. (2006) is applied to solve the advection-dispersion equation used to describe conservative solute transport in highly heterogeneous porous media. Heterogeneity is described by a correlated random field with an exponential covariance. Macrodispersion coefficients are calculated for a broad range of heterogeneities using Monte Carlo (MC) simulations on large size domains. ELLAM circumvents some drawbacks of usual particle-tracking and Eulerian-Lagrangian methods when local dispersion/diffusion is added. ELLAM is also highly efficient and well adapted for advective dominant transport and for MC simulations. For pure advection, first-order approximation provides good estimates of the duration of the preasymptotic regime and of the longitudinal macrodispersion coefficient for a variance of the log conductivity σ2 ≤ 2.25. Higher-order theories overestimated this coefficient for higher variances. Computed transverse macrodispersion is equal to 0 for each studied variance of log conductivity σ2 ∈ [0.25;1.0;2.25;4.0; 6.25; 9.0]. Local dispersion/diffusion affects the macrodispersion for quite low Peclet number (< 100) compared to previous work. For a Peclet number of 10, it leads to an increase of the longitudinal and transverse macrodispersion for low variances (σ2 = 0.25). For higher heterogeneity (σ2 = 9.0 for local dispersion and σ2 ≥ 4.0 for local diffusion), the longitudinal macrodispersion decreases due to local transverse mixing.

Original languageEnglish
Pages (from-to)43-53
Number of pages11
JournalWater Resources Research
Volume49
Issue number1
DOIs
Publication statusPublished - 2013
Externally publishedYes

Fingerprint

adjoint method
porous medium
advection
conductivity
solute transport
simulation

ASJC Scopus subject areas

  • Water Science and Technology

Cite this

Estimation of macrodispersion in 2-D highly heterogeneous porous media using the Eulerian-Lagrangian localized adjoint method. / Ramasomanana, Fanilo; Younes, Anis; Ackerer, Philippe.

In: Water Resources Research, Vol. 49, No. 1, 2013, p. 43-53.

Research output: Contribution to journalArticle

@article{dc46baf8b32b46949f9204371ec69aca,
title = "Estimation of macrodispersion in 2-D highly heterogeneous porous media using the Eulerian-Lagrangian localized adjoint method",
abstract = "The Eulerian-Lagrangian localized adjoint method (ELLAM) formulation of Younes et al. (2006) is applied to solve the advection-dispersion equation used to describe conservative solute transport in highly heterogeneous porous media. Heterogeneity is described by a correlated random field with an exponential covariance. Macrodispersion coefficients are calculated for a broad range of heterogeneities using Monte Carlo (MC) simulations on large size domains. ELLAM circumvents some drawbacks of usual particle-tracking and Eulerian-Lagrangian methods when local dispersion/diffusion is added. ELLAM is also highly efficient and well adapted for advective dominant transport and for MC simulations. For pure advection, first-order approximation provides good estimates of the duration of the preasymptotic regime and of the longitudinal macrodispersion coefficient for a variance of the log conductivity σ2 ≤ 2.25. Higher-order theories overestimated this coefficient for higher variances. Computed transverse macrodispersion is equal to 0 for each studied variance of log conductivity σ2 ∈ [0.25;1.0;2.25;4.0; 6.25; 9.0]. Local dispersion/diffusion affects the macrodispersion for quite low Peclet number (< 100) compared to previous work. For a Peclet number of 10, it leads to an increase of the longitudinal and transverse macrodispersion for low variances (σ2 = 0.25). For higher heterogeneity (σ2 = 9.0 for local dispersion and σ2 ≥ 4.0 for local diffusion), the longitudinal macrodispersion decreases due to local transverse mixing.",
author = "Fanilo Ramasomanana and Anis Younes and Philippe Ackerer",
year = "2013",
doi = "10.1029/2012WR012228",
language = "English",
volume = "49",
pages = "43--53",
journal = "Water Resources Research",
issn = "0043-1397",
publisher = "American Geophysical Union",
number = "1",

}

TY - JOUR

T1 - Estimation of macrodispersion in 2-D highly heterogeneous porous media using the Eulerian-Lagrangian localized adjoint method

AU - Ramasomanana, Fanilo

AU - Younes, Anis

AU - Ackerer, Philippe

PY - 2013

Y1 - 2013

N2 - The Eulerian-Lagrangian localized adjoint method (ELLAM) formulation of Younes et al. (2006) is applied to solve the advection-dispersion equation used to describe conservative solute transport in highly heterogeneous porous media. Heterogeneity is described by a correlated random field with an exponential covariance. Macrodispersion coefficients are calculated for a broad range of heterogeneities using Monte Carlo (MC) simulations on large size domains. ELLAM circumvents some drawbacks of usual particle-tracking and Eulerian-Lagrangian methods when local dispersion/diffusion is added. ELLAM is also highly efficient and well adapted for advective dominant transport and for MC simulations. For pure advection, first-order approximation provides good estimates of the duration of the preasymptotic regime and of the longitudinal macrodispersion coefficient for a variance of the log conductivity σ2 ≤ 2.25. Higher-order theories overestimated this coefficient for higher variances. Computed transverse macrodispersion is equal to 0 for each studied variance of log conductivity σ2 ∈ [0.25;1.0;2.25;4.0; 6.25; 9.0]. Local dispersion/diffusion affects the macrodispersion for quite low Peclet number (< 100) compared to previous work. For a Peclet number of 10, it leads to an increase of the longitudinal and transverse macrodispersion for low variances (σ2 = 0.25). For higher heterogeneity (σ2 = 9.0 for local dispersion and σ2 ≥ 4.0 for local diffusion), the longitudinal macrodispersion decreases due to local transverse mixing.

AB - The Eulerian-Lagrangian localized adjoint method (ELLAM) formulation of Younes et al. (2006) is applied to solve the advection-dispersion equation used to describe conservative solute transport in highly heterogeneous porous media. Heterogeneity is described by a correlated random field with an exponential covariance. Macrodispersion coefficients are calculated for a broad range of heterogeneities using Monte Carlo (MC) simulations on large size domains. ELLAM circumvents some drawbacks of usual particle-tracking and Eulerian-Lagrangian methods when local dispersion/diffusion is added. ELLAM is also highly efficient and well adapted for advective dominant transport and for MC simulations. For pure advection, first-order approximation provides good estimates of the duration of the preasymptotic regime and of the longitudinal macrodispersion coefficient for a variance of the log conductivity σ2 ≤ 2.25. Higher-order theories overestimated this coefficient for higher variances. Computed transverse macrodispersion is equal to 0 for each studied variance of log conductivity σ2 ∈ [0.25;1.0;2.25;4.0; 6.25; 9.0]. Local dispersion/diffusion affects the macrodispersion for quite low Peclet number (< 100) compared to previous work. For a Peclet number of 10, it leads to an increase of the longitudinal and transverse macrodispersion for low variances (σ2 = 0.25). For higher heterogeneity (σ2 = 9.0 for local dispersion and σ2 ≥ 4.0 for local diffusion), the longitudinal macrodispersion decreases due to local transverse mixing.

UR - http://www.scopus.com/inward/record.url?scp=84876513911&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84876513911&partnerID=8YFLogxK

U2 - 10.1029/2012WR012228

DO - 10.1029/2012WR012228

M3 - Article

AN - SCOPUS:84876513911

VL - 49

SP - 43

EP - 53

JO - Water Resources Research

JF - Water Resources Research

SN - 0043-1397

IS - 1

ER -