• DocumentCode
    1936696
  • Title

    Recovery-based a posteriori error estimation for the charge in the method of moments

  • Author

    Strydom, Willem J. ; Botha, Matthys M.

  • Author_Institution
    Dept. of Electr. & Electron. Eng., Stellenbosch Univ., Stellenbosch, South Africa
  • fYear
    2015
  • fDate
    13-17 April 2015
  • Firstpage
    1
  • Lastpage
    4
  • Abstract
    Error estimation for the method of moments (MoM) is a challenging task, due to the non-local nature of the integral operators involved. In this paper, various local recovery procedures are applied to the solution charge in a Rao-Wilton-Glisson (RWG) basis function-based MoM solution on a triangle element mesh, to obtain a higher-order charge representation. The MoM with RWG basis functions is routinely used for the numerical solution of electromagnetic surface integral equations. These recovered charge distributions are then used for a posteriori error estimation of the charge solution and their performance is compared. Such estimates could be incorporated into an adaptive solver scheme to optimize solution accuracy relative to computational cost. It is found that an RWG-based charge recovery procedure that was recently proposed by the present authors, yields the best performance.
  • Keywords
    computational electromagnetics; electric charge; error analysis; integral equations; method of moments; Rao-Wilton-Glisson basis function; electromagnetic surface integral equation; higher order charge; local recovery procedures; method of moments; recovery based a-posteriori error estimation; triangle element mesh; Context; Electromagnetics; Error analysis; Finite element analysis; Integral equations; Junctions; Method of moments; Gradient recovery; boundary element method; computational electromagnetics; electromagnetic surface integral equation; error estimator; post-processing; smoothing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Antennas and Propagation (EuCAP), 2015 9th European Conference on
  • Conference_Location
    Lisbon
  • Type

    conf

  • Filename
    7228871