Modelling variation of textile fabric permeability at mesoscopic scale

Abstract

Textile permeability in general shows a high variance. The present work describes a method to model textile variability at the mesoscopic scale based on a generalised textile model. Inhomogeneities were introduced into the textile structure by randomly moving the tow paths at the cross-overs according to a given normal distribution. The effects of various factors on the evaluated permeability variation were explored and demonstrated, using non-crimp fabric and plain weave models. Fabric architecture was shown to be important in that it imposed a limit to the degree of variation of the tows. A linear relationship can be deduced for the relative variation of permeability as a function of the relative variation of fibre volume fraction, which links the predicted and measured values. When utilising predicted mesoscopic permeability values in macroscopic analyses, the domain size of the mesoscale model has to be related to the element size of the macroscopic analyses.

A multi-physics process model for simulating the manufacture of resin-infused composite aerostructures

“…Additionally, since fabric deformation changes the internal architecture of a fabric weave, permeability properties also need to be characterised in relation to the shear deformation. This effect has been investigated both experimentally [18,19] and with predictive modelling [19][20][21].…”

Uncertainty in the manufacturing of fibrous thermosetting composites: A review

“…Other sources of uncertainty in the impregnation stage can be resin viscosity variations, foreign material inclusion during impregnation, preform volume fraction variations, as well as accidental misplacement of the preform in the mould [5,29]. Experimental and simulations results using non -crimp fabrics have shown that there is a linear relationship between the coefficient of variation of permeability and the coefficient of variation of fibre volume fraction [13,30]. Several experimental and simulation studies have outlined the stochastic nature of permeability.…”

“…According to Hoes [27] global permeability values can be described by a normal distribution [13][14][15][16]. However, simulation results indicated that permeability at the mesoscale (unit cell size) cannot be represented by a normal distribution [30]. This was also observed in the case of a random mat [40].…”

“…In the case of smaller moulds the effect of fibre misalignment at the mesoscale, whether large or small is dominant. On the other hand, as the mould dimensions increase, these local inhomogeneities tend to cancel each other out and the textile appears to be more uniform leading to lower global permeability variations [13,30]. This suggests that the global permeability distribution is governed by the mould dimensions; larger moulds lead to lower global permeability variations [30].…”

“…In general, nesting affects the laminate thickness, the fibre volume fraction and the pore pattern. Therefore nesting can introduce significant spatial scatter in laminate properties at different locations over a composite part, as well as considerable batch to batch variability [30][31][32][33][34].…”