Viscoelastic modeling of injection molding of thermoplastics

Numerical and experimental investigations of two-dimensional, nonisothermal injection of an amorphous viscoelastic melt (polystyrene) were performed. The numerical investigation was initiated by simulation of nonisothermal extrusion into a slit die. It was expanded by the modeling of nonisothermal flow of viscoelastic melt during filling of a center-gated disk and a strip cavity and the relaxation of the stresses during the cooling stage after the cessation of the flow into the strip cavity. These simulations were performed employing Leonov's viscoelastic constitutive equation.; Time and spatial development of velocity, shear rate, shear and normal stresses and birefringence was calculated and the thickness of the thermal boundary layer was analyzed. In particular, the thickness of this layer was determined based on the gapwise position of the birefringence maximum and the glass transition temperature. It was found, that the layer calculated based on the birefringence maximum is thicker than the one based on T{dollar}sb{lcub}rm g{rcub}{dollar}.; The viability of the numerical simulations were checked employing a set of experimental procedures based on the rheo-optical characteristics of the polymer. In these studies the dynamic field was related to the birefringence through linear stress-optical law.; In the first series of experiments, on-line birefringence was measured for various combinations of wall and melt temperatures at different flow rates. It was found that due to the mirage effect caused by the unsteady thermal diffusivity field during the flow, the experiments could not be performed at low wall temperatures. In these series of experiments, although various qualitative analyses were performed, no quantitative comparison with the theory was obtained.; In the second series of experiments, residual birefringence in the part after the nonisothermal flow and relaxation was measured. It was assumed that the residual thermal stresses could be removed without affecting the flow stresses, by a slow annealing at the glass transition temperature. At low wall temperature good agreements between the experiments and the predicted results were obtained. At the wall temperature close to the glass transition temperature the experimental and the predicted results did not compare well indicating a more complex relationship between the thermal and flow residual stresses.