Domain decomposition for 3D nonlinear magnetostatic problems: Newton-Krylov-Schur vs. Schur-Newton-Krylov methods - CNRS - Centre national de la recherche scientifique Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Domain decomposition for 3D nonlinear magnetostatic problems: Newton-Krylov-Schur vs. Schur-Newton-Krylov methods

Résumé

Domain decomposition is a strategy designed to be used on parallel machines. This strategy leads to hybrid methods between direct and iterative solvers and allows users to benefit from the advantages of both. Lately, the growing size of simulations in electromagnetics brought to light the interest of using domain decomposition. Nonlinearity is also one of the problems specificities where the need for an efficient solver is high. This paper provides a comparison between two techniques of domain decomposition for solving 3D nonlinear magnetostatic problems. A test case illustrates the results that can be expected.
Fichier principal
Vignette du fichier
Digest.pdf (577.44 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04115574 , version 1 (02-06-2023)

Identifiants

  • HAL Id : hal-04115574 , version 1

Citer

Mohamed Ghenai, Ronan Perrussel, Olivier Chadebec, Frédéric Vi, Jean-Michel Guichon, et al.. Domain decomposition for 3D nonlinear magnetostatic problems: Newton-Krylov-Schur vs. Schur-Newton-Krylov methods. 24th International Conference on the Computation of Electromagnetic Fields (Compumag'2023), May 2023, Kyoto, Japan. ⟨hal-04115574⟩
37 Consultations
38 Téléchargements

Partager

Gmail Facebook X LinkedIn More