Design of Experiments Methods Applied to CFD Simulations in Vehicle Aerodynamics

2004-01-0443

03/08/2004

Event
SAE 2004 World Congress & Exhibition
Authors Abstract
Content
A study of design of experiments (DOE) methods applied to vehicle external aerodynamics numerical simulations is presented. Three different methods are tested: the L32 Taguchi design, the Central Composite design and the Optimal Latin Hypercube sampling design. CAD surfaces for all three sets of vehicle configurations are generated thanks to a deformation module developed in-house at PSA Peugeot Citroën in CATIA V5. Variations of 6 aerodynamically relevant geometric parameters were considered. A validated methodology for prediction of aerodynamic coefficients based on the ANSA, GAMBIT, TGRID and FLUENT software is used. Post-processing of the DOEs in iSIGHT permits study of first order and quadratic effects as well as their interactions. It also creates analytical models, which can be used for optimizing the set of parameters within the considered design space. Limitations of the models are discussed.
Meta TagsDetails
DOI
https://doi.org/10.4271/2004-01-0443
Pages
12
Citation
d'Apollonia, A., Granier, B., and Debaty, P., "Design of Experiments Methods Applied to CFD Simulations in Vehicle Aerodynamics," SAE Technical Paper 2004-01-0443, 2004, https://doi.org/10.4271/2004-01-0443.
Additional Details
Publisher
Published
Mar 8, 2004
Product Code
2004-01-0443
Content Type
Technical Paper
Language
English