Publication

On numerical quadrature for C1 quadratic Powell–Sabin 6-split macro-triangles

Bartoň, M. & Kosinka, J., Mar-2019, In : Journal of Computational and Applied Mathematics. 349, p. 239-250

Research output: Contribution to journalArticleAcademicpeer-review

Copy link to clipboard

Documents

  • On numerical quadrature for C1 quadratic Powell-Sabin 6-split macro-triangles

    Final author's version, 3 MB, PDF-document

    Embargo ends: 20/08/2020

    Request copy

  • On numerical quadrature for C1 quadratic Powell–Sabin6-splitmacro-triangles

    Final publisher's version, 1 MB, PDF-document

    Request copy

DOI

The quadrature rule of Hammer and Stroud (1956) for cubic polynomials has been shown to be exact for a larger space of functions, namely the C1 cubic Clough–Tocher spline space over a macro-triangle if and only if the split-point is the barycentre of the macro-triangle Kosinka and Bartoň (0000). We continue the study of quadrature rules for spline spaces over macro-triangles, now focusing on the case of C1 quadratic Powell–Sabin 6-split macro-triangles. We show that the 3-node Gaussian quadrature(s) for quadratics can be generalised to the C1 quadratic Powell–Sabin 6-split spline space over a macro-triangle for a two-parameter family of inner split-points, not just the barycentre as in Kosinka and Bartoň (0000). The choice of the inner split-point uniquely determines the positions of the edge split-points such that the whole spline space is integrated exactly by a corresponding polynomial quadrature. Consequently, the number of quadrature points needed to exactly integrate this special spline space reduces from twelve to three. For the inner split-point at the barycentre, we prove that the two 3-node quadratic polynomial quadratures of Hammer and Stroud exactly integrate also the C1 quadratic Powell–Sabin spline space if and only if the edge split-points are at their respective edge midpoints. For other positions of the inner and edge split-points we provide numerical examples showing that three nodes suffice to integrate the space exactly, but a full classification and a closed-form solution in the generic case remain elusive.
Original languageEnglish
Pages (from-to)239-250
JournalJournal of Computational and Applied Mathematics
Volume349
Early online date20-Aug-2018
Publication statusPublished - Mar-2019

    Keywords

  • Numerical integration, Powell–Sabin spline space, Gaussian quadrature rules

View graph of relations

ID: 65502646