A micro-scale model for fiber tow characterization under nondeterministic assumption: longitudinal and transverse permeability

摘要

Liquid Composite Molding processes are characterized by the impregnation of a dry fibrous perform by means of injection or infusion of a catalyzed resin. In recent years computational flow and cure models allowed for a remarkable time and cost compression in process planning with respect to trial and error procedures. In this contest multi-scale simulative approaches are gaining considerable attention and intriguing results have been recently presented. Most of the proposed models, however, rely on deterministic hypothesis, assuming perfect fiber packing and neglecting dimensional variations between fibers, in strong contrast with experimental observations. In this paper the influence of the stochastic variability of the fiber packing on tow permeability has been investigated by means of a CFD micro scale model. The variability of the microstructure defining the Representative Volume Element has been considered introducing random perturbations of the fiber packing. The components of the permeability tensor, in each case, have then been derived applying the Darcy model to flow simulations through the representative cell.