We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.NA

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Numerical Analysis

Title: A Posteriori Error Estimation for the p-curl Problem

Abstract: We derive a posteriori error estimates for a semi-discrete finite element approximation of a nonlinear eddy current problem arising from applied superconductivity, known as the $p$-curl problem. In particular, we show the reliability for non-conforming N\'{e}d\'{e}lec elements based on a residual type argument and a Helmholtz-Weyl decomposition of $W^p_0(\text{curl};\Omega)$. As a consequence, we are also able to derive an a posteriori error estimate for a quantity of interest called the AC loss. The nonlinearity for this form of Maxwell's equation is an analogue of the one found in the $p$-Laplacian. It is handled without linearizing around the approximate solution. The non-conformity is dealt by adapting error decomposition techniques of Carstensen, Hu and Orlando. Geometric non-conformities also appear because the continuous problem is defined over a bounded $C^{1,1}$ domain while the discrete problem is formulated over a weaker polyhedral domain. The semi-discrete formulation studied in this paper is often encountered in commercial codes and is shown to be well-posed. The paper concludes with numerical results confirming the reliability of the a posteriori error estimate.
Comments: 32 pages
Subjects: Numerical Analysis (math.NA)
MSC classes: 35K65, 65M60, 65M15, 78M10
Journal reference: SIAM J. Numer. Anal. 58(1) 460-491 (2020)
DOI: 10.1137/16M1075624
Cite as: arXiv:1605.06532 [math.NA]
  (or arXiv:1605.06532v2 [math.NA] for this version)

Submission history

From: Andy Wan T. S. [view email]
[v1] Fri, 20 May 2016 20:38:43 GMT (1174kb,D)
[v2] Mon, 10 Feb 2020 05:53:12 GMT (2237kb,D)

Link back to: arXiv, form interface, contact.