Abstract In order to effectively solve modern automotive design problems including the results of nonlinear FEA and multi-body dynamics, a progressive meta-model based design optimization is presented. To reduce the number of initial sample points, two sampling methods are introduced. Then, for efficient and stable construction of meta-models, three metamodel methods are newly introduced which are numerically based on the singular value decomposition technique. To design a practical system considering manufacturing tolerances and optimizing multiple performances, a robust design optimization, 6-sigma constraints and multi-objective strategies are implemented when solving the approximate optimization problem constructed from the meta-models. Until the convergence criteria are satisfied, the initially developed meta-models are progressively improved by adding only one point that minimizes the approximate Lagrangian in the consecutive optimization iterations. Finally, one validation sample and four automotive applications are solved to show the effectiveness of the proposed approach.
Innovative design optimization strategy for the automotive industry
International Journal of Automotive Technology ; 15 , 2 ; 291-301
2014-03-01
11 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
RDO (Robust Design Optimization) , RBF (Radial Basis Function) , RSM (Response Surface Method) , MDO (Multi-Disciplinary Optimization) , ALM (Augmented Lagrange Multiplier) , DOE (Design of Experiments) , SAO (Sequential Approximation Optimization) , PMM (Progressive Meta-Model) Engineering , Automotive Engineering
Innovative design optimization strategy for the automotive industry
British Library Online Contents | 2014
|Innovative design optimization strategy for the automotive industry
Online Contents | 2014
|Innovative welding technologies for the automotive industry
Tema Archiv | 2000
|SYSTEMS & SOFTWARE - Innovative hints from the automotive industry.
Online Contents | 2002
Advancing Automotive Design With Innovative Collaboration
Online Contents | 2010