A multivariate piecewise interpolation model for viscosity representation and its application to the numerical simulation of non-isothermal polymer melt flow

Abstract

In this study, a multivariate piecewise interpolation model is introduced to represent melt viscosity based on temperature and shear rate. The model does not require a predefined functional form for the entire data range, instead finding a suitable line between adjacent measured viscosity data points to obtain viscosity as a function of shear rate and temperature. To assess the effect of viscosity representation on polymer melt flow, a two-dimensional (2D) flow past a cylinder problem and a three-dimensional (3D) single screw extruder problem are numerically solved using different viscosity models. The pressure drop (ΔP) from the shift-factor model deviates by more than 10% at an average velocity (Uav) of 1 m/s due to overestimation of viscosity in the high shear rate region. The dimensionless thicknesses of the thermal boundary layer near the channel wall (δT/H) at wall temperature (Tw) of 210 °C are also misevaluated by more than 4% using both the shift-factor model and the Klein model, reflecting their poor viscosity evaluation. It is noteworthy that in all viscosity models, a high Tw leads to the formation of a thin thermal boundary layer, which adversely affects the heat transfer. In single screw extruder applications, viscosity misevaluation in high shear rate regions causes issues in pure drag flow cases, while misevaluation in low shear rate regions causes issues in pure back-pressure flow cases. These analyses highlight the accuracy and versatility of our multivariate piecewise interpolation model compared to existing viscosity models.