Journal article 233 views
Parametric solutions involving geometry: A step towards efficient shape optimization
Computer Methods in Applied Mechanics and Engineering, Volume: 268, Pages: 178 - 193
Swansea University Author: Antonio Huerta Cerezuela
Full text not available from this repository: check for access using links below.
DOI (Published version): 10.1016/j.cma.2013.09.003
Abstract
In simulation-based engineering sciences geometry is usually data. Every time a new geometry is tested, the complete analysis is repeated: pre-process, solve and post-process. This is a major overhead in engineering design. This paper is the first contribution where the parameters defining the geome...
| Published in: | Computer Methods in Applied Mechanics and Engineering |
|---|---|
| Published: |
2014
|
| URI: | https://cronfa.swan.ac.uk/Record/cronfa20037 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract: |
In simulation-based engineering sciences geometry is usually data. Every time a new geometry is tested, the complete analysis is repeated: pre-process, solve and post-process. This is a major overhead in engineering design. This paper is the first contribution where the parameters defining the geometry are seen as coordinates and, consequently, a high-dimensional problem is solved to obtain a generalized solution: the temperature field, for instance, is known at any point in space and for any value of the parameters describing the physical domain. To overcome the "curse of dimensionality" the Proper Generalized Decomposition is employed. The generalized solution is obtained on the fly, sensitivities to parameters variations are also evaluated analytically and consequently the applicability of this approach to classical optimization or multi criteria optimization is demonstrated on academic examples.As a result of this research A. Huerta has been invited to give a plenary talk in Japan and the Forschung Fahrzeugtechnik VOLKSWAGEN (the joint research centre of the Volkswagen group) is funding a project to reduce design cycle costs for automotive aerodynamics. |
|---|---|
| College: |
Faculty of Science and Engineering |
| Start Page: |
178 |
| End Page: |
193 |