Experimental and numerical analysis of performance of two fluted mixer designs

Abstract

In this paper, the effect of the Maddock mixer design on its performance in the single screw extrusion of High-density Polyethylene (HDPE) melt has been investigated experimentally on a special extrusion line equipped with a barrel having several glass windows as well as theoretically using a three-dimensional finite element method simulation. The red-green-blue spectral analysis has been used for the experimental quantification of the mixing efficiency, whereas a generalised Newtonian model (with Carreau-Yasuda viscosity function) and a viscoelastic modified White-Metzner model have been utilised in flow simulations. Based on the performed experimental work, it has been found that the mixer with no undercut wiping flight ('Closed' mixer) has a faster mixing/purging but a similar final mixing performance as the mixer with undercut wiping flight ('Open' mixer). It has also been revealed that the 'Open' mixer creates an unwanted dead zone (in the form of the stagnation layer of material rotating between the mixer and the barrel) at which the polymer melt degradation may take place. It has been found that based on the stress, residence time and pathline calculation, it is possible to predict the experimentally observed tendency of the 'Closed' mixer for a faster mixing/purging in comparison with the 'Open' mixer as well as the presence of a stagnation layer in the 'Open' mixer. The performed theoretical parametric study indicates that the gradual opening of the wiping gap causes the pressure driven flow to become more dominant than the drag flow in shearing as well as in the wiping gap independently of the utilised rheological model. On the other hand, the predicted particle trajectories inside the mixer were found to be shorter for the viscoelastic model, tending to occur mainly in the middle of the channel and thus leaving the mixer a little bit faster in comparison with the purely viscous calculations. Finally, it has been revealed that for the highest tested wiping flight opening, the viscoelastic modified White-Metzner model predicts back flow over the wiping flight, whereas the purely viscous Carreau-Yasuda does not, which can be explained by the elongational viscosity, considered in the viscoelastic modified White-Metzner model. This suggests that the modified White-Metzner model should be preferred more than the purely viscous Carreau-Yasuda model in the mixing element die design optimisation.