CAD-based Aerodynamic Shape Optimization Using Geometry Surrogate Model And Adjoint Methods

摘要

Nowadays classical configurations of civil transportation aircraft are well optimized, with further improvements expected to be of the order of single drag counts. To achieve it, the whole aircraft configuration has to be taken into account during aerodynamic shape optimization, which require handling complex geometries with Computer Aided Design (CAD) software. So far, integration of CAD engine and aerodynamic shape optimization has been difficult, due to framework complexity and problematic gradient evaluation. In this paper we propose to substitute the CAD engine with a geometry surrogate model of the CAD surface deformation. The resulting framework does not require calls to the CAD engine during the optimization, with design variables defined in CAD during preparation step. Furthermore, a geometry surrogate model has been analytically differentiated, allowing coupling with mesh adjoint method of objective function gradient evaluation.