Simulation of melt spinning including flow-induced crystallization - Part II. Quantitative comparisons with industrial spinline data

The mathematical model for melt spinning of Doufas et al. [A.K. Doufas, A.J . McHugh, C. Miller, J. Non-Newtonian Fluid Mechanics, 1999] coupling fiber microstructure (molecular orientation and crystallinity) with the macrosco pic velocity/stress and temperature fields, is tested extensively against i ndustrial spinline data for several nylon melts. Model fits and predictions are shown to be in very good quantitative agreement with spinline data for the fiber velocity and temperature fields at both low and high-speed condi tions, and, with birefringence data available for high speeds. The effects of processing parameters: quench air velocity, capillary diameter and mass throughput, as well as material characteristics: molecular weight (RV) and polymer type (i.e., homopolymers with or without additives, and copolymers) , on the spinline dynamics are accurately predicted. Under high-speed condi tions, strain softening occurs and the tensile stress at the freeze point i s predicted to be essentially independent of the processing parameters inve stigated, in agreement with experimental observations. Birefringence data a nd model predictions show that crystallization occurs mostly after the free ze point, under the locked-in tensile stress. Under low-speed conditions, t he velocity and crystallization profiles (experimental and predicted) are s hown to evolve smoothly towards a plateau value and strain hardening behavi or is predicted throughout the spinline. The ability to quantitatively desc ribe spinline data over a wide range of conditions and material characteris tics, renders the model a useful tool for optimization of melt spinning pro cesses as well as a framework for simulation of other polymer processes inv olving flow-induced crystallization. (C) 2000 Elsevier Science B.V. All rig hts reserved.