Improvement of Multiparametric Strategy in Two-Level Model Optimization of Assemblies - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Improvement of Multiparametric Strategy in Two-Level Model Optimization of Assemblies

Résumé

Designing structural assemblies led to analysis of quasi-static problems with local nonlinearities such as unilateral contacts with friction. Therefore, optimization of structural assemblies is very expensive in terms of computation time. In order to reduce this cost, a two-level model optimization strategy based on the successive optimizations of a metamodel and of the full mechanical model is used [1] [2]. In the first level, a metamodel based on a cumulative approximation [3], and a global optimizer are used to obtain approximated optima and, in the second level, mechanical model is connected directly with a local optimizer for locating the precise optimum. The main drawback of this strategy is the CPU time required to generate the metamodel. In order to reduce the computation time we use a dedicated FEM for solving assembly problems based on a mixed domain decomposition and on an iterative scheme called LaTIn [4]. The assembly is decomposed into substructures with an elastic behavior and interfaces with a nonlinear law taking into account the frictional contact conditions. In the context of parametric optimization solved problems are very similar in the sense that only parameters varies. For a new set of design parameters, the LaTIn algorithm can be reinitialized using a previous converged solution and enables us to obtain faster convergence. The reuse of converged solution called MultiParametric Strategy [5] leads a significant reduction of computational time. The main objective of this study is to improve the MultiParametric Strategy by an optimization of the reinitialisation step. We use the fact that with a previous close solution only a few iterations are needed to reach convergence in the new calculation. Different norms on the parameters' space and taking into account metamodel responses are defined and compared. The strategy allows us to significantly reduce the computation time associated to the building of the metamodel (gain greater than 20) and, consequently, of the whole optimization process. This strategy will be presented on an industrial test-case of torque transmission by friction. This work was supported by the Agence Nationale de la Recherche as part as ANR-08-COSI-007-10: Distributed Multi-Disciplinary Optimization. References [1] H. Engels, W. Becker, A. Morris, Implementation of a multi-level optimisation methodology within the e-design of a blended wing body. Aerospace Science and Technology, 8(2), 145-153, 2004. [2] B. Soulier, P.A. Boucard, A multiparametric strategy for the two step optimization of structural assemblies. Structural and Multidisciplinary Optimization, 47(4), 539-553, 2013. [3] J. Rasmussen, Nonlinear programming by cumulative approximation refinement. Structural and Multidisciplinary Optimization, 15(1), 1-7, 1998. [4] P. Ladeveze, Nonlinear computational structural mechanics: new approaches and non-incremental methods of calculation, Springer, New York, 1999. [5] P.A. Boucard, L. Champaney, A suitable computational strategy for the parametric analysis of problems with multiple contact. International Journal for Numerical Methods in Engineering, 57(9), 1259-1281, 2003.
Fichier non déposé

Dates et versions

hal-01431903 , version 1 (11-01-2017)

Identifiants

  • HAL Id : hal-01431903 , version 1

Citer

Bruno Soulier, Luc Laurent, Pierre-Alain Boucard. Improvement of Multiparametric Strategy in Two-Level Model Optimization of Assemblies. 1st International Conference on Engineering and Applied Sciences Optimization, Jun 2014, Kos, Greece. ⟨hal-01431903⟩
174 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More