Differentiated Parametric CAD Used Within the Context of Automatic Aerodynamic Design Optimization

Abstract

As environmental concern grows, the need for fuel -ef ficient aircraft becomes more and more stringent in order to meet up with the rapidly evolving standards which are set to reduce noise and emissions. In order to reach that goal relying on an efficient design process defining the detailed shape of aircraft is mandatory . In that respect the use of automatic aerodynamic o ptimization has been proven to be of great benefit over the recent past. It enables significant reduction in development cost and cycle time. After a brief overview of the set -up of the overa ll optimization process , which includes a finite volume Euler flow solver in combination with a discrete adjoint approach, we will focus in the present paper more specifically on the use of a differentiated parametric CAD modeler. Thus far the gradient of the CAD model was obtained through a finite difference approach. When using different types of variables the more the process gets dependent on the choice of the different step sizes which is not straightforward to make. Therefore the CAD definition has be en recently differentiated and its impact on convergence and final result will be described. Validation and application of differentiated CAD were performed through a set of inverse problems by progressively including a growing number of modules in to the o ptimization loop . Eventually the complete loop for automatic aerodynamic optimization was used on a generic business jet in transonic cruise conditions.