Experimentally Validated Numerical Modeling of Polymer Melting and Flow in Material Extrusion-based Additive Manufacturing

Material extrusion-based additive manufacturing, strictly speaking of the Fused Filament Fabrication (FFF), is characterized by using a thermoplastic polymer in form of a solid filament as a built material. The filament partially melts inside of a heated hotend and is subsequently extruded under the pressure generated by the filament feeding force of the feeding rollers. The processes in the heated extrusion channel are complex and difficult to model, as the melt properties are a nonlinear function of temperature and shear rate. Moreover, there is a phase transition from solid filament to liquid melt. In this paper, the required feeding force and resulting extrudate temperature at the nozzle outlet at different feeding rates, liquefier temperatures and nozzle geometries of two acrylonitrile-butadiene styrene (ABS) and polypropylene (PP) materials were investigated. For a better understanding of the melting process in the hotend, the degree of melting of the strands was determined via dead-stop experiments at selected operating points. Two process windows could be identified: At low feeding rates, the feeding force increases linearly with increasing velocity, and at a certain point, the force increases rapidly and fluctuates. The melting investigations showed that this is related to the increasingly unmolten material reaching the nozzle outlet. The semi-crystalline PP showed a smaller processing window for stable flow compared to ABS and lower extrudate temperatures due to its differentiated cp progression at the phase transition. Based on the experimental studies, Computational Fluid Dynamics (CFD) simulations were performed to predict the pressure and temperature distributions inside the channel. For modelling the shear-dependent viscosity the Carreau model was used, while the temperature dependency was described by a jump function with a solid viscosity of 10(6) Pas. It is shown that the numerical model can predict the feeding force with good accuracy and represent the change between process windows at increasing speeds.