Multiscale modeling of combined deterministic and stochastic fabric non-uniformity for realistic resin injection simulation

Abstract

The local fiber arrangement in a bi-directional fabric formed to a complex shape was modeled considering the stochastic arrangement of filaments within yarns, which determines axial and transverse yarn permeabilities, and the stochastic arrangement of yarns in a fabric, which determines the dimensions of interyarn gap spaces locally. To mimic the uncertainty in fabric forming, drape simulation was randomized in terms of start point and yarn start orientations. From yarn permeabilities and simulated local yarn spacing distributions, local fiber volume fractions and fabric permeabilities were approximated. This allowed resin injection into a deformed fabric to be simulated for different drape scenarios with different probabilities and different degrees of fabric randomness. The results indicate that variability in fabric properties and the forming process affects flow front shapes and times for complete impregnation of the reinforcement.