Incorporation of linear multipoint constraints in substructure based iterative solvers - PART I: A numerically scalable algorithm

Catherine Lacour (1, 2) , Charbel Farhat (3, 4) , Daniel J. Rixen (5)

1 ACSIOM - Analyse, Calcul Scientifique Industriel et Optimisation de Montpellier
2 I3M - Institut de Mathématiques et de Modélisation de Montpellier
3 CAS - Center for Aerospace Structures
4 Department of Aerospace Engineering Sciences
5 T.U. Delft - T.U. Delft, Faculty of Design, Engineering and Production, Engineering Dynamics

We consider the iterative solution by a class of substructuring methods of the large- scale systems of equations arising from the finite element discretization of structural models with an arbitrary set of linear multipoint constraints. We present a methodology for generalizing to such problems numerically scalable substructure based iterative solvers, without interfering with their formulations and their well-established local and global preconditioners. We apply this methodology to the FETI method, and show that the resulting algorithm is numerically scalable with respect to both the substructure and problem sizes.

Mots clés: domain decomposition; multipoint constraints; numerical scalability

Domaines: Analyse numérique [math.NA]

Catherine Lacour, Charbel Farhat, Daniel J. Rixen. Incorporation of linear multipoint constraints in substructure based iterative solvers - PART I: A numerically scalable algorithm. International Journal for Numerical Methods in Engineering, 1998, 43 (6), pp.997--1016. ⟨10.1002/(SICI)1097-0207(19981130)43:6<997::AID-NME455>3.0.CO;2-B⟩. ⟨hal-00369536⟩