NARA Discovery
Article Details
← Back to Search Results
Journal Article

Postprocessing of standard finite element velocity fields for accurate particle tracking applied to groundwater flow

Philipp Selzer; Olaf A. Cirpka
Computational Geosciences · Vol. 24, Issue 4 · pp. 1605-1624 · 2020

Abstract

Particle tracking is a computationally advantageous and fast scheme to determine travel times and trajectories in subsurface hydrology. Accurate particle tracking requires element-wise mass-conservative, conforming velocity fields. This condition is not fulfilled by the standard linear Galerkin finite element method (FEM). We present a projection, which maps a non-conforming, element-wise given velocity field, computed on triangles and tetrahedra, onto a conforming velocity field in lowest-order Raviart-Thomas-Nédélec ( $\mathcal {RTN}_{0}$ R T N 0 ) space, which meets the requirements of accurate particle tracking. The projection is based on minimizing the difference in the hydraulic gradients at the element centroids between the standard FEM solution and the hydraulic gradients consistent with the $\mathcal {RTN}_{0}$ R T N 0 velocity field imposing element-wise mass conservation. Using the conforming velocity field in $\mathcal {RTN}_{0}$ R T N 0 space on triangles and tetrahedra, we present semi-analytical particle tracking methods for divergent and non-divergent flow. We compare the results with those obtained by a cell-centered finite volume method defined for the same elements, and a test case considering hydraulic anisotropy to an analytical solution. The velocity fields and associated particle trajectories based on the projection of the standard FEM solution are comparable to those resulting from the finite volume method, but the projected fields are smoother within zones of piecewise uniform hydraulic conductivity. While the $\mathcal {RTN}_{0}$ R T N 0 -projected standard FEM solution is thus more accurate, the computational costs of the cell-centered finite volume approach are considerably smaller.

Bibliographic Information

JournalComputational Geosciences
PublisherSpringer
Publication Date2020-08-01
Publication Year2020
Volume24
Issue4
Pages1605-1624
Document TypeJournal Article
Print ISSN1420-0597
eISSN1573-1499
DOI10.1007/s10596-020-09969-y

Access Information

NARA Access Coverage1997-01-01~Current
Journal Homepagehttps://www.springer.com/journal/10596
Publisher PageOpen Publisher Page
Full-text access depends on NARA's subscribed coverage and institutional access.