This thesis is concerned with the development of methods and models to aid in optimization and development of new materials for Fused Filament Fabrication (FFF). We demonstrate a novel FFF nozzle design to enable the first measurements of in-situ rheology inside FFF nozzles, which is critical for part performance by ensuring that the polymer extrudate is flowing at an appropriate temperature and flow rate during the part build process. Testing was performed using Acrylonitrile butadiene styrene filament and a modified Monoprice Maker Select 3D printer. Tests using the default temperature control settings of the printer showed an 11 °C drop in temperature and significant fluctuations in pressure, during printing and while idle, of ± 2 °C and +/-14 kPa. These deviations were eliminated at lower flow rates with a properly calibrated proportional–integral–derivative (PID) system. At high flow rates, drops in temperature as high as 6.5 °C were observed even with a properly calibrated PID, providing critical input to the impact of flow rate and PID calibration on polymer melt temperature inside FFF nozzles. Pressure readings ranging from 140-6900 kPa were measured over the range of filament feed rates and corresponding extrusion flow rates. Theoretical predictions of pressure profiles, assuming a powerlaw fluid model, matched well with experimental results. Our nozzle prototype succeeded in measuring internal conditions of FFF nozzles for the first time, thereby providing several important insights into the printing process which are vital for monitoring and improving FFF printed parts.

Furthermore, finite difference simulations based on first principles analysis are presented which are capable of quantifying the effect of processing conditions on the properties of semicrystalline parts made by FFF. Each layer was modelled as a rectangular cross section which was broken down into smaller elements for modelling. Crystallinity of each element was calculated using a parallel Avrami model which accounts for changes in crystallization rate due to temperature and multiple crystallization mechanisms. The amount of polymer diffusion, also referred to as the degree of healing, was calculated using a novel incremental diffusion model which accounted for not only changes in reptation time due to temperature but also restrictions to healing due to crystallinity. To the authors knowledge, this is the first healing model capable of accounting for the effect of crystallinity on healing and is relevant to any process involving healing of crystalline interfaces; not just FFF. Cumulative shear stresses between each layer and at the bottom of the part were also calculated for the first time using a force balance model by assuming constant shear strain throughout each layer. Simulations were performed using typical printing conditions for polyether ether ketone. In the first layer of a 24 layer part, the average degree of crystallinity, healing, and shear stress were 25.0%, 53.8% and 19.4 MPa respectively. The degree of crystallinity and healing at layer 22 (which represented the steady state values) were 18.4-25.0% and 51.4% respectively. When crystallinity was not accounted for, varying the printing parameters and material properties supported the use of high temperatures and specific heat in addition to a low printing speed, heat transfer coefficient, and thermal conductivity to maximize part properties. These conditions also supported crystallization, however, which led to a simultaneous reduction in the part properties when crystallinity was taken into account. These contradictory effects will need to be considered when optimizing the printing parameters, though the optimal balance will be highly dependent on the material used and the limitations of the printer.

Experimental validation of the accuracy of the heat transfer and polymer diffusion models was performed using an amorphous polymer (polyether imide). Single road wide parts were printed at various nozzle temperatures, bed temperatures, and printing speeds and the results were compared to the simulated results. The predicted shear stress in the bottom of the part ranged from 2.3-3.8 MPa and correlated to warpages at the corners of each part of 1.2-2.4 mm. A linear increase in warpage with predicted shear stress was observed supporting the shear stress model. Predicted degrees of healing ranged from 2-90% but the experimental results ranged from 15-36%. Results of the healing model underpredicted strength at low printing speeds and over predicted strength at high printing speeds. The experimental validations showed the capabilities of the models, but the effect of printing speed will need to be investigated further to improve the accuracy of the healing model.