Modelling Processing of Unfilled and Long-Glass Fibre Reinforced Thermoplastics in a Screw-Barrel Unit

Abstract

In injection moulding, long glass fibre reinforced thermoplastics (LGFT) are an attractive way to produce large parts at low cost. The strength of the part depends chiefly on the average fibre length, fibres which are subjected to considerable attrition during processing in conventional three stage screws. First of all, in this study we have coupled a melting analysis in a conventional screw to a model of fibre breakage whereby a fibre anchored at one end in the solid bed is submitted, at its other end, to the intense shear stress of the molten polymer flowing in the film close to the barrel. As the melting of the solid bed progresses, more fibres are unlayered and submitted to bending which intensity is depending on both the fibre length and orientation. When the bending is too high, the fibre breaks. Bimodal fibre length distribution are obtained and compared to existing data. The sensibility of the model to main processing parameters such as screw rotation, initial fibre length, viscosity, barrel temperature and screw geometry are also investigated. Next, we present a new analytical solution for flow of a viscous fluid in a single screw channel that takes into account the torsion and curvature of the channel. Contrary to common knowledge in polymer processing based on the Parallel Plate Model, we found that, in the case of cross-sections with large aspect ratio, torsion effects can be significant. The implication of the model on velocity field, residence time and mixing efficiency is investigated and compared to the predictions of the classical Parallel Plate Model, to finite elements calculations, and to 3D experimental measurements. Indeed, an innovating device has been developed in our laboratory to visualize the flow of a viscous fluid in the channel of a screw. It consists of a transparent barrel and of a rotating screw, pumping a transparent viscous fluid at room temperature. A particle plunged in the flow is constantly monitored by four video-cameras placed around the barrel and recording its position in a frame. The 3D path lines are then computed.