Toogle Left

BATir banner

 

logo SMC logo SMC  OPTI logo SMC  OPTI  MDO Model reduction in computational mechanics
   
Contact Arnaud Robert, Rajan Filomeno Coelho
   
Keywords multidisciplinary design optimization (MDO), model reduction, metamodels, proper orthogonal decomposition (POD), adaptive learning
   
Collaborations
  • Prof. P. Breitkopf, Université de Technologie de Compiègne, France
  • Dr. I. Lepot, Dr. C. Sainvitu, Cenaero, Belgium
   
Model reduction in computational mechanics Multidisciplinary optimization (MDO) is a growing field in engineering, with various applications in aerospace, aeronautics, car industry, etc. However, the presence of multiple disciplines leads to specific issues, which prevent MDO to be fully integrated in industrial design methodology. In practice, the key issues in MDO lie in the management of the interconnections between disciplines, along with the high number of simulations required to find a feasible multidisciplinary (optimal) solution.

III.3 Fig 1

Our current research axes, investigated in collaboration with the Roberval Laboratory at the University of Technology of Compiègne (UTC-CNRS, UMR 6253) and with Cenaero research center in aeronautics, focus on the development of surrogate models (or metamodels) based on original variants of the Proper Orthogonal Decomposition (POD) method, with applications in aeronautics and automotive engineering.

     
Support
  • This project is supported through a FIRST-DOCA PhD funding (Walloon Region).
     
Selected publications
  • [1] P. Breitkopf, R. Filomeno Coelho, editors. Multidisciplinary Design Optimization in Computational Mechanics, Wiley/ISTE, 549 pages, 2010.
  • [2] M Guénot, I Lepot, C Sainvitu, J Goblet, R Filomeno Coelho. Adaptive sampling strategies for non-intrusive POD-based surrogates. Engineering Computations, 30(4):521–547, 2013.