Type:

Year:

Editor:

Author:

Garrelts, Enno; Roth, Daniel; Binz, Hansgeorg

Series:

Institution:

Section:

DOI number:

ISSN:

Abstract:

The design of topologically optimized structures is straightforward, although the main problem is really to derive the correct structure in each instance. During the development of structures for additive manufacturing in particular, saving weight is crucial because weight is proportional to cost.

In this contribution firstly, different approaches to topological optimization are presented and discussed. While computer-assisted tools provide high accuracy and demand defined conditions, approaches utilizing a pen and paper can be conducted relatively quickly, although these only provide little guidance.

Secondly, a new approach is presented which is advantageous with regard to effort and affordability, yet which maintains an accuracy of results. To support the designer, an artificial neural network is trained to adapt suitable Michell structures within a given design area. These structures depict optimal paths to conduct the loads through a component and provide guidance in designing an appropriate topology.

Evaluation has demonstrated that this new approach is capable of supporting designers in achieving lightweight structures.

Keywords:

Additive Manufacturing, Lightweight design, App, Early design phases, Artificial intelligence