Heterogeneous materials, such as composites and material composites, are capable of developing a high lightweight construction potential if optimally designed and utilizing multi-material design. For this purpose, the topology must be adapted with regard to the optimum position and content of reinforcing fibers or also the geometric arrangement of the constituents with regard to the subsequent load in the component. In sum, this results in a large solution space that can only be treaded with difficulty using conventional solution approaches. Generative design methods support the exploration of the search space, for example by using evolutionary algorithms for topology optimization.
Load-path-compliant design can be achieved economically using generative or additive manufacturing processes. A central aspect in the manufacture of fiber composite structures is the processing of short or continuous fibers into preforms which, taking into account interfacial properties between the fibers used and the surrounding matrix, lead to structurally optimized composite materials.

In the research focus "Generative Design Methods and Materials Development", different generative design methods are to be researched and applied to generative manufacturing processes. One example is additive manufacturing, which will be investigated for different material systems, such as cement, polymer, metal and ceramics. The orientation of reinforcing fibers during the deposition of the pultrudate is particularly important for a load-path-compliant design. In this context, process modeling for additive manufacturing processes is also being advanced in order to enable, for example, material prequalification from this. In addition, other additive processes such as stereolithography are also being investigated; here, the development of the suspensions to be processed plays an important role. The challenges are to master the diversity of materials and their impact on load-path-compatible design.

Another focus in the processing of long and continuous fibers is on manufacturing processes such as tape laying or fiber patch placement. A combination of different generative manufacturing processes is also planned. For example, fiber patch placement could be used to deposit a complex three-dimensional preform with optimized structure. In a next step, this geometry would have to be scanned in order to apply further reinforcement structures using additive processes. The surface scan and the material combination of preform and material play an important role here.

In order to be able to describe the corresponding cause-effect relationships during the processing of heterogeneous materials, cross-thematic collaboration in the AI production network is necessary.
The cause-effect relationships are difficult to describe physically, so they can only be described practicably using a basis of process and characterization data and process-property relationships with the aid of artificial intelligence methods.
All process steps must therefore be digitized and in-situ material characterizations integrated in order to be able to generate digital twins. Furthermore, it is also necessary to describe the resulting component properties in models. For the individual process, three levels of information must be combined for this purpose: the process parameters, the resulting real material states and the process simulation. In this way, it is possible to optimally influence the corresponding process parameters depending on the desired component properties, which, combined with in-situ material characterization, should lead to self-optimized material processing. The aim is to combine materials and process development to produce components that can be used under demanding conditions and fed into a suitable recycling process at the end of their useful life.