Extension of the multi-scale probabilistic simulation chain for the continuous modelling of the manufacturing process and the structural behavior to discontinuous long fiber reinforced plastic composites (MeproSi-2)

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

In the first phase of the MeproSi project, a probabilistic, multiscale simulation for short fiber-reinforced composites was developed, capable of capturing uncertainties along the entire process chain up to the component level. In the continuation, this methodology will be extended to long fiber-reinforced thermoplastics (LFT), which are characterized by additional uncertainties such as fiber curvature, variations in fiber volume content, and heterogeneous microstructures. The goal is to build an integrated probabilistic process chain simulation to comprehensively analyze these additional effects. LFTs exhibit significant microstructural and property scatter due to processing. The main causes are variations in fiber orientation, length, and volume content. Complex modeling approaches for fiber-matrix separation exist but are computationally too intensive and not yet validated for injection molding at the component level. The mechanical behavior of discontinuously reinforced long fiber composites is strongly influenced by process-induced microstructure, leading to local variations in material properties that must be modeled probabilistically. A consistent multiscale analysis to capture these uncertainties up to the component level is still lacking and will be further developed. In WP1, injection molding experiments with LFT model materials will be conducted, varying fiber content and process parameters. This includes material production, CT analyses, and rheological and mechanical testing. WP2 develops plastification-dependent boundary conditions for process simulation initially deterministic, then probabilistic-based on free jet experiments. WP3 models fiber volume content as a field variable in the process simulation, starting with a balance equation without empirical parameters, followed by a physically based extension for fiber-matrix separation modeling. Validation will use experimental data from WP1. WP4 builds on this to develop a model for predicting local inhomogeneities such as fiber clusters and curvatures, using simulation results and experimental data for sensitivity analysis, parameter identification, and validation. The goal is reliable prediction of local material inhomogeneities. WP5 develops a method for stochastic microstructure generation for LFT materials, accounting for fiber curvature and clustering. WP6 focuses on efficient finite element modeling of these microstructures, including beam elements and multiscale simulation. WP7 investigates how microstructural uncertainties affect macroscopic material properties, based on stochastic homogenization. WP8 transfers the results to the component level using probabilistic process and structural simulation, incorporating random fields for material parameters.

• DFG Programme: Research Grants