Numerical analysis of uncertainty propagation in short fiber-reinforced composites: From injection molding to material testing

The present study investigates a novel methodology for the numerical assessment of uncertainties and their propagation in injection-molded fiber-reinforced polymers (FRPs). Focusing on two primary sources of uncertainty, which are the microstructural variability due to injection molding process parameters and inherent material scatter, the research examines their individual contributions to the scattering of effective properties, specifically the Young’s modulus of the composite. A random vector model was used to describe the orientation states across the structure, derived from multiple injection molding simulations with varying input parameter distributions. The scatter of structural material parameters is further built upon a joint distribution between orientation concentration and constitutive parameters. The results reveal that while the overall orientation states exhibited less scatter than expected, a clear relationship emerged between the concentration of injection molding parameters and the cumulative distribution functions (CDFs) of effective modulus, indicating non-linear interactions between orientation and material scatter. Additionally, the analysis highlighted the increased sensitivity of scattering behavior based on sample orientation, emphasizing the effect of geometry on flow properties. This research underscores the complex interplay of uncertainties in determining effective material behavior, suggesting that future studies should explore a broader range of input parameters and refine distribution assumptions. The findings provide valuable insights for advancing the design and manufacturing processes of polymer composites, establishing a foundation for more comprehensive analyses of uncertainty in material properties. • Holistic probabilistic simulation strategy from process to structural level established. • Uncertainty propagation from processing conditions to material response quantified. • Uncertain morphology identified as most important driver of material uncertainty. • Full account made to correlations using concept of conditional probability.