Multi-scale probabilistic simulation chain for the continuous modelling of the manufacturing process and the structural behavior of disordered fiber-reinforced injection molded components (MeproSi)

• Applicants: Dr. Carla Beckmann; Professorin Dr.-Ing. Luise Kärger
• Subject Area: Lightweight Construction, Textile Technology ; Plastics Engineering
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 464119659

Due to the increasing demand for lightweight components not only in aerospace applications but also areas dominated by industrial scale mass production such as the automotive sector, discontinuously fiber reinforced materials gain an increasing importance. Modern long fiber composites manufactured in thermoset injection molding might exhibit similar mechanical properties as aluminum. Nevertheless, distinctively lower amounts of energy are consumed during the manufacturing process, further improving the CO2-footprint. A major shortcoming is their random, disordered and process dependent microstructure, causing an inherent inevitable uncertainty in their macroscopic material response. This uncertainty is caused by the uncertainties in the local fiber content, length and orientation distributions as well as by uncertainties in the material response of the constituents themselves and the fiber matrix interfaces. Superimposed are local variations in the microstructural properties imposed by the manufacturing process. In structural application, the inevitable uncertainty in the material response results in over-conservative designs, unnecessarily large safety margins and thus a suboptimum exploitation of the materials lightweight potential.The objective of the proposed project is a prediction of the process-related aleatoric uncertainties in the material response and the resulting structural behavior of injection molded parts under in-service conditions. For this purpose, an integrated probabilistic multiscale simulation along the process chain on all relevant scales will be established. Starting point is the uncertain process, resulting in an uncertain microstructure and thus an uncertain structural response. Based on a probabilistic process simulation, the local microstructure and the probability distributions of its governing parameters will be derived. Using a probabilistic homogenization procedure, a stochastic description of the material behavior is obtained, providing the input for a random field representation of the material response. Appropriate strategies for determination of individual representations of the random fields based on spatially varying stochastic properties obtained in the process simulation have to be defined and implemented. By this means, the effect of uncertainties on the structural response of short fiber reinforced components can be predicted in an integrated manner, starting from the process conditions. By this means, the inherent uncertainties in the material and structural response become accessible and controllable by computational methods. The initial focus is on aleatoric uncertainties. In a second phase, an extension to include epistemic uncertainty, e.g. from measurement or modelling inaccuracies, is intended.

• DFG Programme: Research Grants