Modelling the permeability of random discontinuous carbon fibre preforms

Zhaofei Xiao,Nicholas A Warrior,Lee T Harper,Andreas Endruweit,Xiaoling Liu

doi:10.1177/0021998320902506

Zhaofei Xiao, Nicholas A Warrior + Show 3 more

Open Access

https://doi.org/10.1177/0021998320902506

Copy DOI

Abstract

A force-directed algorithm was developed to create representative geometrical models of fibre distributions in directed carbon fibre preforms. Local permeability values were calculated for the preform models depending on the local fibre orientation, distribution and volume fraction. The effect of binder content was incorporated by adjusting the principal permeability values of the meso-scale discontinuous fibre bundles, using corresponding experimental data obtained for unidirectional non-crimp fabrics. The model provides an upper boundary for the permeability of directed carbon fibre preform architectures, where predictions are within one standard deviation of the experimental mean for all architectures studied.

Highlights

Directed Carbon Fibre Preforming (DCFP) is an automated process for producing non-woven reinforcements for composites from chopped fibre tows [1]
The effect of binder content was incorporated by adjusting the principal permeability values of the meso-scale discontinuous fibre bundles, using corresponding experimental data obtained for unidirectional non-crimp fabrics
Experimental data showed that the presence of powder binder has a significant influence on the permeability of Directed Carbon Fibre Preforms (DCFP) preforms, when typical amounts are added for preform fixation

Summary

Introduction

Directed Carbon Fibre Preforming (DCFP) is an automated process for producing non-woven reinforcements for composites from chopped fibre tows [1]. The variability in permeability measured for a series of DCFP preforms with identical specifications was found to be as high as 34 % [5], which can cause great uncertainty in terms of developing flow patterns, affecting tooling design and injection strategy This variability is related to local variations in meso-scale fibre architecture, which depends on four main factors, i.e. the uniformity of the superficial density, the fibre orientation distribution, the fibre length distribution, and the average tow filament count. Global permeability values derived from the simulations were between 25 % and 108 % higher than corresponding experimental values, depending on the assumed through-thickness distribution of the bundles This difference was attributed to the approach for generating local permeabilities, which overlooked the non-uniformity of through-thickness fibre distributions and the effect of applied binder on effective flow channel dimensions. The effect of preform binder content is taken into account in the calculation of local permeability values, which are used as input for resin injection simulations

Objectives

Results

Discussion

Conclusion