Abstract
The fabrication of support-free structures in pellet additive manufacturing (PAM) is severely limited by gravity-induced sagging, a phenomenon lacking predictive, physics-based models. This study introduces and validates a numerical model for the thermofluid dynamics of sagging, aiming to correlate process parameters with filament deflection. A predictive finite element (FE) model incorporating temperature-dependent non-Newtonian material properties and heat transfer dynamics has been developed. This was validated via a systematic experimental study on a desktop-scale PAM 3D printer investigating nozzle temperature, printhead speed, screw speed and fan cooling, using polylactic acid (PLA) as a printing material. Findings show that process parameter optimization can reduce bridge deflection by 64.91%, with active fan cooling being the most dominant factor due to accelerated solidification. Increased printhead speed reduced sagging, whereas higher screw speeds and extrusion temperature showed the opposite effect. The FE model accurately replicated these results and further revealed that sagging ceases once the filament cools below its minimum flow temperature (approximately 150-160 °C for PLA). This validated model provides a robust foundation for tuning process parameters, unlocking effective support-free 3D printing in PAM.