Commingled yarn composites for rapid processing of complex shapes

Abstract

Commingled yarns of reinforcing and thermoplastic fibres offer a potential for low-cost manufacturing of complex-shaped composite parts, due to reduced impregnation times and applied pressures during processing. In order to benefit from this competitive advantage, the process parameters governing consolidation must be controlled. In this study, a consolidation model, previously validated for unidirectional commingled yarn fabrics processed isothermally in a flat matched-die mould, is applied to three other processing techniques capable of producing complex-shaped composites. Tubes of braided commingled yarns were manufactured by bladder inflation moulding. Selectively reinforced polymeric parts were processed by compression–injection moulding. Stamp forming was also used to allow high-speed processing of commingled yarn-based laminates. Besides particular stamp forming cases, for which part deconsolidation occurred, the model predictions were in good agreement with the void content values obtained from specimens consolidated under different processing conditions. This suggests that the consolidation model can be successfully applied to a wide range of yarn architectures and processing techniques.

Introduction

Among the possible routes for manufacturing fibre reinforced thermoplastic composites, consolidation of preforms based on commingled yarns has emerged as a cost-effective method for the production of complex-shaped parts. As the polymer flow distance for impregnation is reduced when the matrix is already commingled with the fibres, the time for impregnation or the need for high-applied pressure are limited, leading to a reduction in manufacturing cost. Moreover, compared to other intermediate product forms such as pre-impregnated tows or powder-impregnated fibre bundles, the flexible commingled yarns can easily be converted into a highly drapeable textile fabric, reducing the likelihood of wrinkles during the forming of complex shapes. To fully benefit from these advantages, the process parameters governing consolidation must be identified and optimised. Consolidation of commingled yarns has been studied and modelled by several authors [1], [2], [3], [4], [5], [6] and has recently been reviewed by Svensson et al. [7]. In all cases, the consolidation quality was characterised by the amount of residual porosity in the part. In general, the polymer fibres melt rapidly and form liquid pools surrounding the reinforcing fibres. The problem thus reduces to that of infiltration of dry fibre bundles surrounded by molten polymer pools. As reinforcing fibres and polymer fibres are in general not distributed homogeneously within a yarn, commingling quality becomes one of the most critical parameters.

Friedrich et al. [1], [2], [3], [4] modelled the fibre bed impregnation mechanism by describing matrix flowing in one direction, orthogonal to the fibre axis. They assumed that part of the yarn was impregnated quickly by well-distributed polymer fibres or during melting of the polymer, and also defined a commingling parameter corresponding to an adjusted thickness of the dry bundle. Van West et al. [5] proposed a more elaborate finite element model for impregnation of fibre bundles of elliptic cross-section. They considered that the fibres may move in the bundle during flow of the resin, assuming a linear distribution of fibre velocities along the radius of the bundle, but did not account for the resulting volume fraction change. They also assumed that the matrix formed a continuous network at some point. An increase in void gas pressure followed as impregnation further proceeded, leading to a value of residual porosity. Cain et al. [6] adapted the model developed by Van West for the case of non-Newtonian fluids. Limitations encountered in all these models include the assimilation of the superficial velocity reported in Darcy's law to the matrix velocity and the consideration of a Carman–Kozeny approach for the fibrous bed permeability, which adds another parameter to be determined by fitting with experimental measurements of permeability or void content. In addition, these models were all compared to experimental data obtained from compression of flat plates.

The present authors recently proposed a detailed analysis of the laws governing fibre bundle impregnation, and developed a new model for consolidation of commingled fibre yarns [8]. The analysis relied on a geometrical description of the yarn structure, taking into account a distribution of cylindrical dry fibre bundles of different sizes at the onset of impregnation, as schematically described in Fig. 1. Each population is therefore impregnated concomitantly, with radial flow orthogonal to the fibre axis, but with different kinetics. The overall void volume fraction includes the contribution of the complete distribution. The model also included a possible gas pressure increase due to closing of the void within a bundle at some point during impregnation. It was demonstrated, for unidirectional commingled yarn laminates made from Polyamide 12 (PA12) and carbon fibres (CF), that the model accurately predicts the influence of the processing variables (i.e. time, pressure and temperature) on the part quality (in terms of void content) after consolidation, if appropriate selection of a critical radius related to closing of pores is made. Nevertheless, model solutions for limiting cases, considering that pores are closed from the onset of infiltration or, on the other hand, remain open all along, provided adequate brackets for an estimation of laminate void contents with no fitting parameter.

In this paper, we propose to investigate the applicability of the model previously validated for a flat geometry, a unidirectional yarn configuration and isothermal consolidation conditions to other processing techniques and yarn assembly types, always keeping the same type of individual yarn. Three case studies are proposed: (i) manufacturing of a tube by bladder inflation moulding of braided yarns, (ii) automated yarn placement and pre-consolidation, followed by polymer over-injection for the manufacture of selectively reinforced composites, and (iii) rapid non-isothermal compression moulding of a stack of unidirectional commingled yarn fabrics, pre-consolidated or not.