Dernière mise à jour : 2 oct. 2020
CAE Wissen by courtesy of Dr. Kambiz Kayvantash, CADLM (pages 76-79)
Model order reduction (ROM) techniques are interpolation methods exploiting exiting data sets (input and output) derived from an existing model or experimental setup. The starting point is a DOE-type design which covers as best as possible the design space (space filling property). Contrary to response surface (polynomial based) designs where the selection of design points at particular positions is important due to a-priori properties of the fitted surface, for ROM techniques, the most important issue is the space-filling capacity and sufficient “modal” representation of the response. In some cases ROM can be considered as algebraic solutions, exploiting decomposition of timedependent phenomena into special and temporal components, for reducing the volume of a data set while preserving the most important parts of the information contained within the data which is necessary for retrieving all or the most essential part of the data when needed. Else we can also consider ROM as clustering or other “lossy” efficient data compression techniques, allowing for the reduction of the amount of data required for the reconstruction of the complete data set. Both such techniques allow for creating on-board and real-time applications based on voluminous experimental or simulation results (ex. Finite element). In this paper we shall present the major idea behind the reduction (or fusion) methods as well as providing three potential applications for crash and safety simulations. The results are obtained by ODYSSEE (Lunar) software  and compared to LS-DYNA FEM results.
KEYWORDS: Model Reduction, ROM, POD, CLUSTERING, SVD, On-board computing, Real-time modeling, Crash, Safety, Parametric design