Numerická studie adaptivní FFD v aerodynamické tvarové optimalizaci

A numerical study of adaptive FFD in aerodynamic shape optimization

conference paper

en

This work presents a Free Form Deformation (FFD) method for shape parameterization and applications on aerodynamic aircraft design. Our objective is to accurately resolve problems of shape optimization including geometric constraints. Here we couple an FFD based on NURBS with a Radial Basis Function (RBF) in order to improve the control of the deformations and enable a more accurate resolution of constraints even when using a coarse lattice. Our investigations also focus on efficiency, showing for instance that the highest NURBS degree improve the resolution of the optimization problem when the solution is a smooth shape in 2D without influencing the cost of optimization. This indicates a mean to avoid wiggles in shapes deformed by FFD without deteriorating the condition of the problem. Other applications in 2D and 3D also illustrate the behaviour of the FFD parameterization when refining or adapting the position of the lattice vertices.

V této práci je prezentována metoda tvarové parametrizace Free Form Deformation (FFD) a její aplikace v aerodynamickém návrhu letadel. Našim cílem je přesně řešit aerodynamickou tvarovou optimalizaci s geometrickými omezeními. Za účelem zlepšení kontroly nad deformacemi a dosažení přesnějšího dodržení omezení (nezávisle na hustotě FFD mřížky) párujeme FFD založenou na NURBS s Radial Basis Function (RBF). Naše zkoumání je také zaměřeno na efektivnost, například je zde ukázáno, že vyšší stupeň NURBS zlepší výsledek optimalizace, tj. výsledkem je hladký tvar geometrie bez oscilací, aniž by došlo k ovlivnění ceny optimalizace. Další aplikace ve 2D a 3D ukazují chování FFD parametrizace při zhušťování mřížky nebo adaptaci poloh bodů mřížky.

This work presents a Free Form Deformation (FFD) method for shape parameterization and applications on aerodynamic aircraft design. Our objective is to accurately resolve problems of shape optimization including geometric constraints. Here we couple an FFD based on NURBS with a Radial Basis Function (RBF) in order to improve the control of the deformations and enable a more accurate resolution of constraints even when using a coarse lattice. Our investigations also focus on efficiency, showing for instance that the highest NURBS degree improve the resolution of the optimization problem when the solution is a smooth shape in 2D without influencing the cost of optimization. This indicates a mean to avoid wiggles in shapes deformed by FFD without deteriorating the condition of the problem. Other applications in 2D and 3D also illustrate the behaviour of the FFD parameterization when refining or adapting the position of the lattice vertices.

numerická, FFD, aerodynamická, tvarová, optimalizace

numerical, FFD, aerodynamic, shape optimization

2014

17.01.2014

AIAA

978-1-62410-256-1

52nd AIAA Aerospace Sciences Meeting

1–11

11