Stochastic modeling of preform heterogeneity to address dry spots formation in the VARTM Process

Abstract

Dry spot formation during the filling process was investigated for a system consisting of a heterogeneous porous medium coupled with a homogeneous porous layer. This system is representative of a composite manufacturing process in which the resin is introduced through an injection gate into a mold containing stacks of woven fabrics and a vacuum is drawn at the downstream end to impregnate the fibers. Dry spots are the regions that are not impregnated by the resin when the resin reaches the vent. The heterogeneity of the porous medium was quantified by modeling it with by-pass paths. They represent the gaps at the intersections of fiber tows in woven fabrics and have a larger permeability than the bulk porous medium permeability. When these fabrics are stacked to form a preform before injecting the resin, it is probable that some of these gaps across the thickness for all the layers would be inline creating the by-pass path for the resin. Hence, the need for stochastic modeling and random distribution of these by-pass paths. A log–normal distribution, K bp ∼ LN ( μ ,σ K ) was chosen to describe their formation. A five-parameter space was defined to describe the stochastic system. Numerical simulations of resin flow in this space for discrete combinations of the parameters were performed and the dry spot content ( c DR ) was obtained. The average dry spot content ( c ¯ DR ) and its standard deviation ( σ DR ) were calculated by conducting sufficient number of simulations with random placement of by-pass paths for each set of parameters. It was found that ( c ¯ DR ) and ( σ DR ) depends on ( σ K ), but the influence of ( σ K ) on ( c ¯ DR ) and ( σ DR ) can be altered by adjusting other parameters.