Methode zur Massengenerierung von Topologieoptimierungsmodellen

Abstract

Topology optimization is a common technique to generate optimal structures with respect to given loads and constraints. In this thesis, a method for the creation of large numbers of topology optimizations is proposed.

The method variates the parameters of optimization models by assigning multiple fixed states which the parameters can take, and then automatically creating an optimization model for each possible combination of these parameter states, similar to a factorial design as it known from design of experiments. Each optimization model is then submitted to Altairs solver OptiStruct for optimization.

In this thesis, it is described which types of optimization models are supported, and which parameters are available for automatic variation. Five test runs of the program were performed, and their results analyzed with respect to their similarity to each other, which showed that the program is capable of generating optimization results that are in general not very similar to each other.