Prediction of transverse permeability in representative volume elements with closely arranged fibers through the application of delaunay-triangulation and electrical-circuit analogy

Abstract

The distribution of fibers in composite manufacturing significantly affects fiber bundle permeability, a crucial factor for resin impregnation and void formation, which determines final product quality. However, as the industry increasingly seeks composite materials characterized by high fiber volume fractions, the existing numerical analysis methodologies employed for predicting permeability require a considerable computational cost. In this study, the permeability prediction model was developed using a fluid domain simplification technique based on Delaunay triangulation. The inter-fiber flow was approximated using a lubrication model, and continuity was ensured using a circuit analogy. This model can explain stochastic permeability characteristics according to fiber distribution and accurately predicts the permeability regardless of the distance between fibers. Additionally, compared to existing computational fluid dynamics simulations, it achieved an incredible accuracy of 97.74% and reduced computational costs by an average of 1,700 times.