Recent developments of numerical shape optimization for automotive aerodynamics are presented. The work is focused on sensitivity maps calculated with the adjoint method and their application to full scale production vehicles. Since adjoint based optimization was extended to the field of fluid dynamics, comprehensive research has been conducted to develop the method and mature it for industrial applications. Until now only few groups have succeeded in calculating adjoint based sensitivity maps for entire vehicles. The presented results are based on an OpenFOAM implementation of the adjoint method for shape sensitivity calculations. Sensitivity maps for the cost function drag and the recently implemented cost function rear lift have been successfully computed for production vehicles. The calculations are based on flow solutions obtained with the standard computational procedure used in the aerodynamic development process. These sensitivity maps predict how local movements in the surface normal direction will affect the overall aerodynamics of a car. This is a powerful method through which the vehicle aerodynamic development process can be accelerated. The drag sensitivity calculation method has been employed for several vehicle development projects parallel to the regular wind tunnel measurements. As a result, a high level of confidence in the sensitivity calculations has been gained, and the method has been brought to a fully operative level. By coupling the drag sensitivity calculation with a parameter-free morphing algorithm, an automatic sensitivity driven shape optimization routine has been setup. Simple test cases with automatically generated meshes have been successfully morphed with respect to drag reduction. The use of a parameter-free routine is advantageous compared to the traditional parameter based morphing, since no inherent limitations are imposed on the wall movements and the additional process step of setting up the morphing parameters becomes redundant. The application of this method during the early stages of the aerodynamic development process has a high potential to greatly reduce the time needed for costly wind tunnel experiments.
Automatic aerodynamic shape optimisation based on the adjoint method
2014
12 Seiten, Bilder, Quellen
Conference paper
English
Automatic aerodynamic shape optimisation based on the adjoint method
British Library Conference Proceedings | 2014
|Shape optimisation for aerodynamic performance using adjoint methods
Tema Archive | 2014
|Automatic aerodynamic shape optimization based on the adjoint method
Automotive engineering | 2014
|Adjoint Based Aerodynamic Optimisation of a UCAV
British Library Conference Proceedings | 2013
|Adjoint Based Aerodynamic Optimisation of a UCAV
AIAA | 2013
|