A decoupled strategy to solve reduced-order multimodel problems in the PGD and Arlequin frameworks

David Néron (1) , Hachmi Ben Dhia (2) , Régis Cottereau (2)

1 LMT - Laboratoire de Mécanique et Technologie
2 MSSMat - Laboratoire de mécanique des sols, structures et matériaux

In this paper, we investigate the coupling of reduced models for the simulation of structures involving localized geometrical details. Herein, we use the Arlequin method, originally designed to deal with multimodel and multiscale analyses of mechanical problems, to mix reduced models built using the Proper Generalized Decomposition. Instead of solving the global coupled problem in a monolithic way, the LATIN strategy is used to propose a decoupled algorithm. The numerical examples demonstrate the feasibility of the approach and in particular its potentiality in terms of flexibility.

Mots clés: reduced-order modeling; PGD; multimodel; multiscale; Arlequin; LATIN

Domaines: Mécanique des structures [physics.class-ph]; Mécanique des solides [physics.class-ph]; Calcul parallèle, distribué et partagé [cs.DC]

David Néron, Hachmi Ben Dhia, Régis Cottereau. A decoupled strategy to solve reduced-order multimodel problems in the PGD and Arlequin frameworks. Computational Mechanics, 2016, 57 (4), pp.509-521. ⟨10.1007/s00466-015-1236-0⟩. ⟨hal-01292497⟩