One of the most common industrial polymer processes is fiber spinning, converting polymer into fiber. In the melt-spinning process, which is relatively simple and economical compared to wet- and dry-spinning, the polymer melt is extruded through small orifices in the spinneret and drawn into thin fibers by a uniaxial drawing process. During the growth of the fiber industry since the early 1960's, numerous experimental and theoretical efforts have been made in order to better understand the spinning process for a wide range of process conditions. For semicrystalline polymer melt-spinning, crystallization during solidification is closely tied to final fiber properties. Particularly for high-speed melt spinning, the so-called flow-induced crystallization (FIC) occurs as a result of high tensile stresses in the fibers. Crystallization and orientation also depend heavily on fiber temperature, so that the quench process plays a significant role in controlling fiber properties such as orientation, degree of crystallinity, mechanical properties, etc. Several models have been developed to describe crystallization during the spinning process, including models which combine crystallization kinetics with the Newtonian constitutive equation for the polymer melt. Other models incorporate crystallization with strain-softening and strain-hardening associated with nonlinear viscoelastic effects which produce a necking phenomenon. Most models to date lack sufficient details to adequately describe the melt-spinning process at industrial scales. Many models focus only on a single filament, and multifilament models often fail to capture complexities in the spinning process such as nonlinear constitutive behavior. More accurate and flexible models are required due to the growing diversity of spinning materials and the need for fibers with desired properties. Therefore, the purpose of this paper is to develop a multifilament spinning model which describes semicrystalline behavior in a nonisothermal multifilament setting. A secondary goal is to develop a model which will run on a desktop computer in reasonable time (e.g. 1 to 2 hours). The current model combines the flow-enhanced crystallization (FEC) model of McHugh et al. (Journal of Applied Polymer Science, 100, 3240–3254, 2006) with a generalization of Dutta's multifilament model (Polymer Engineering and Science, 27, 1050–1058, 1985). In particular, we include a more comprehensive formulation of air drag effects than is found in Dutta's model. Many spinning models have ignored air drag due to the complexity of the model. Air drag can be ignored near the spinneret because the fiber velocity is small in that region. However, since air drag increases quadratically with spinning velocity, it should not be ignored at higher fiber speeds. The proposed model, with air drag included, calculates the interdependent quench air velocity and temperature distributions among fibers modeled by the 1-D FEC model, which describes crystallization using Avrami crystallization kinetics. The crystalline phase is treated as growing and orientating rigid rods, and the amorphous phase is described by a modified Giesekus fluid model, a realistic viscoelastic model with finite chain extensibility. To test the model, we make comparisons with the Dutta's model prediction for low-speed spinning of PET (Polyethylene Terephthalate). The results show a qualitative match with Dutta's multifilament calculation, with differences due to our use of more accurate quench terms and the FEC spinning model. Developing a model with high accuracy and flexibility to describe and predict the real process is an essential component of this effort. The long-term goal is to provide an advanced framework for modeling of the fiber spinning process, with ease of extension to more complex (e.g. higher-dimensional) fiber spinning models.

Extended Abstract Status: Not Uploaded