Improving sub-grid scale accuracy of boundary features in regional finite-difference models

Advances in Water Resources
By:  and 

Links

Abstract

As an alternative to grid refinement, the concept of a ghost node, which was developed for nested grid applications, has been extended towards improving sub-grid scale accuracy of flow to conduits, wells, rivers or other boundary features that interact with a finite-difference groundwater flow model. The formulation is presented for correcting the regular finite-difference groundwater flow equations for confined and unconfined cases, with or without Newton Raphson linearization of the nonlinearities, to include the Ghost Node Correction (GNC) for location displacement. The correction may be applied on the right-hand side vector for a symmetric finite-difference Picard implementation, or on the left-hand side matrix for an implicit but asymmetric implementation. The finite-difference matrix connectivity structure may be maintained for an implicit implementation by only selecting contributing nodes that are a part of the finite-difference connectivity. Proof of concept example problems are provided to demonstrate the improved accuracy that may be achieved through sub-grid scale corrections using the GNC schemes.
Publication type Article
Publication Subtype Journal Article
Title Improving sub-grid scale accuracy of boundary features in regional finite-difference models
Series title Advances in Water Resources
DOI 10.1016/j.advwatres.2012.02.011
Volume 41
Year Published 2012
Language English
Publisher Elsevier
Publisher location Amsterdam, Netherlands
Contributing office(s) Office of Groundwater
Description 11 p.
Larger Work Type Article
Larger Work Subtype Journal Article
Larger Work Title Advances in Water Resources
First page 65
Last page 75
Google Analytic Metrics Metrics page
Additional publication details