Today, consistent digitalization often forms the basis for efficient product development, especially in the field of composite materials: Manufacturing processes are digitalized in the form of process simulations. The input data relating to the processing properties of the semi-finished product are in turn derived from digital models of the microscopic material structure. And while state-of-the-art methods of process data acquisition allow the validation of the simulations, machine learning approaches can be used to identify process-structure-property relationships - a simple example of the immense potential of these methods, which at IVW have long ceased to be mere tools, but in many cases represent the research object itself, for example when predictions by neural networks are to replace computationally intensive calculations in order to accelerate simulations.
Against this background, the Leibniz-Institut für Verbundwerkstoffe GmbH (IVW) has set itself the goal of researching digitalized methods even more intensively in the future, developing them further and ultimately transferring them to industry, science and teaching - always with a focus on composite materials, of course. Therefore, a new, fourth program area was founded on January 1, 2023, which is to synergetically complement the existing program areas of Component Development, Materials Science and Manufacturing Science. Under the leadership of PD Dr.-Ing. habil. David May, important future areas will be researched in the two competence areas of the program area, "Process Simulation" and " Digitalized Process and Material Development".
One focus is on end-to-end simulation chains from fiber to component, led by IVW simulation expert Dr. Miro Duhovic. To realize such multi-scale approaches, upscaling methods between the different scale levels have to be developed as well as experimental methods for input generation and digital models on which the simulation can take place. Another focus is on process data acquisition. Following the premise of "the right sensor at the right place", holistic concepts are being developed, whereby virtual studies with process simulations make it possible to minimize the data acquisition effort and at the same time maximize the informative value. In order to prepare a targeted use of the data, the program area will also address the basic topic of ontology, i.e. the semantic description of composite materials, processes and simulations as well as their relations. Closely related to this is the issue of data exploitation through data-driven methods that can be used in a variety of ways. In the program area, application scenarios are conceptualized (e.g. property prediction based on process IST data). These form the basis for the design of process data acquisition on the one hand and for the selection and implementation of the most effective machine learning methods on the other hand.
With this orientation, IVW consistently develops its previous research and closely interlocks with local research partners such as the RPTU Kaiserlautern-Landau, the Fraunhofer- Institute for Industrial Mathematics ITWM ITWM and the German Research Center for Artificial Intelligence DFKI as well as sister institutes in the Leibniz Association.