With additive manufacturing processes, new design proposals can be realized without the limitations of semi-finished products and machining processes. New design processes are required in order to utilize the improvement potential for more powerful and efficient components. Today, numerical methods based on mathematical or empirical laws are often used for this purpose. However, the trend is increasingly moving away from the direct dimensioning of components towards numerically optimized structures whose appearance is specified by the computer. With these methods, only the numerical functionality and boundary conditions are specified in the design process. The result is optimized geometries for a desired property, which then serve as the basis for the subsequent design process. When designing components with different functionalities, these methods are currently reaching their limits. As functionality increases, it becomes increasingly difficult to derive a clear geometry. This is mainly due to the fact that the mutual influence of several functionalities must be understood and described for the computer during the design process. In the case of competing functional objectives, such as creating a particularly lightweight and at the same time resilient design, these dependencies become increasingly unclear. Competing functional characteristics arise in particular through the coupling of different physical disciplines. The resulting multidisciplinary optimizations can currently only be solved to a very limited extent. The aim of this work is therefore to develop a method with which several functional characteristics from different physical disciplines can be integrated for a design. This is to be done without a direct mathematical description of the coupling theories. In the end, it should be possible to give the designer a dimensional distribution without knowing the exact dimensions at the beginning. A way is shown how an evaluation of the material in the design space can be carried out solely on the basis of the functional properties. For this purpose, an evaluation of neural networks leads to a probability distribution. This shows where material is required in the design space in order to fulfil the functional characteristics as an optimization goal. All functional characteristics are considered across three exemplary physical disciplines in a parallel approach. The entire approach is demonstrated on a motor carrier of a UAV drive unit. This is optimized multiphysically and for additive manufacturing using the approach.     «

With additive manufacturing processes, new design proposals can be realized without the limitations of semi-finished products and machining processes. New design processes are required in order to utilize the improvement potential for more powerful and efficient components. Today, numerical methods based on mathematical or empirical laws are often used for this purpose. However, the trend is increasingly moving away from the direct dimensioning of components towards numerically optimized structu...     »