TRR402-C05

Generative design for actively controlled mold systems with high system complexity and functional density

The production of hybrid lightweight structures with graded material transitions requires new manufacturing and mold technologies. Electromagnetic and ultrasonic methods for orienting and positioning reinforcing fibers across material boundaries are particularly promising. To enable this, the materials must be kept in a molten state until the desired fiber structure is established. At the same time, mold are required that allow spatially resolved thermal control and precise regulation without impairing essential functions such as melt guidance, temperature control, and part demolding. The resulting system complexity is difficult to manage using current development methods. The objective of this subproject is therefore to develop a methodology that enables the automated generation, arrangement, and holistic optimization of mechanical, thermal, and actively controlled mold elements by means of generative artificial intelligence. To this end, 3D CAD data, formalized input conditions, and suitable artificial neural network architectures will be investigated and evaluated. Based on reference processes in compression molding and injection molding, data sets will be created to train the neural networks and assess them using quality metrics and expert knowledge. The methodology is intended to be integrated into conventional mold development workflows and designed to be adaptable to emerging technologies for influencing fiber orientation. In this way, complex mold can be developed more rapidly while meeting requirements for component quality, process efficiency, and sustainability.

Period

01.04.2025–31.12.2028

Project partners