Yes / An approximation to the Phan-Thien Tanner (PTT) constitutive model is developed with the aim of giving low-cost simulation of Gas Assisted Injection Moulding (GAIM) while incorporating important viscoelastic characteristics. It is shown that the developed model gives a response typical of full viscoelastic models in transient and steady state uniaxial and constant shear rate deformations. The model is incorporated into a 3D finite element GAIM simulation which uses the ‘pseudo-concentration’ method to predict residual polymer, and applied to published experimental results for a Boger fluid and a shear-thinning polystyrene melt. It is shown that the simulation gives a very good match to published results for the Boger fluid which show increasing Residual Wall Thickness (RWT) with increasing Deborah number. Against the shear-thinning polymer, the quality of match depends upon which of two ‘plausible’ relaxation times is chosen; qualitatively different results arise from two different means of estimating a single relaxation time. A ‘multi-mode’ approach is developed to avoid this uncertainty. It is shown that the multi-mode approach gives decreasing RWT with increasing Deborah number in agreement with the published experimental results, and avoids the issues that arise from estimating a single relaxation time for a molten polymer.

Yes / A methodical 'design of experiment' approach is used to assess the effect of key control parameters on residual wall thickness (RWT) in Gas Assisted Injection Moulding. An empirical model is produced from which the experimental RWT can be determined at any interpolated point. This model includes only those terms with proven statistical significance. The 'true' thermal boundary conditions are determined for a 1-D approximation to the system, this is sufficient to determine the error in a simulation method that enforces coolant temperature as the mould boundary condition for temperature. It is shown that errors in heat-flux and wall temperature are small. A 3-D finite element, pseudo-concentration implementation is presented, with a novel method for simulation of internal gas injection. The simulation is shown to give good agreement with the experimental rate of growth of wall thickness as gas delay is increased; good qualitative agreement is shown for other control parameters.