Author(s)

G. Mathiak & E. Schnack

Abstract

It is well known that the choice of the design variables has a strong influence on the shape of an optimized structure. In three dimensions the problem is even more severe than in 2D. Therefore a new strategy is introduced which starts with few design variables and inserts new ones in areas of significant change. Thus the solution is obtained in nested design spaces. Applications of this strategy are discussed as well as advantages and limits. Examples for super curve Coons surfaces and tensor product B-spline surfaces are given. Geometrical considerations try to find the location of an optimal insertion. 1 INTRODUCTION Despite the successful application of many shape optimization methods, it is still unknown, how to choose the design model efficiently. This paper shall offer a first step to adapt

Keywords