Foundations of space-time finite element methods: Polytopes, interpolation, and integration

Cory V. Frontin, Gage S. Walters, Freddie D. Witherden, Carl W. Lee, David M. Williams, David L. Darmofal

Research output: Contribution to journalArticlepeer-review

11 Scopus citations

Abstract

The main purpose of this article is to facilitate the implementation of space-time finite element methods in four-dimensional space. In order to develop a finite element method in this setting, it is necessary to create a numerical foundation, or equivalently a numerical infrastructure. This foundation should include a collection of suitable elements (usually hypercubes, simplices, or closely related polytopes), numerical interpolation procedures (usually orthonormal polynomial bases), and numerical integration procedures (usually quadrature rules). It is well known that each of these areas has yet to be fully explored, and in the present article, we attempt to directly address this issue. We begin by developing a concrete, sequential procedure for constructing generic four-dimensional elements (4-polytopes). Thereafter, we review the key numerical properties of several canonical elements: the tesseract, tetrahedral prism, and pentatope. Here, we provide explicit expressions for orthonormal polynomial bases on these elements. Next, we construct fully symmetric quadrature rules with positive weights that are capable of exactly integrating high-degree polynomials, e.g. up to degree 17 on the tesseract. Finally, the quadrature rules are successfully tested using a set of canonical numerical experiments on polynomial and transcendental functions.

Original languageEnglish (US)
Pages (from-to)92-113
Number of pages22
JournalApplied Numerical Mathematics
Volume166
DOIs
StatePublished - Aug 2021

All Science Journal Classification (ASJC) codes

  • Numerical Analysis
  • Computational Mathematics
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Foundations of space-time finite element methods: Polytopes, interpolation, and integration'. Together they form a unique fingerprint.

Cite this