This conference paper was submitted for presentation at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

Finite element analysis (FEA) of additive manufacturing (AM) is a powerful tool for understanding the thermo-mechanical behaviour of materials during the manufacturing process. This understanding can help improve the process and the structural integrity of the manufactured component. However, the inherent extreme thermal gradients, rapid cycles of heating and cooling and microscale nature of the process turns achieving an accurate FEA model into a challenging task. Calibration of an FEA model is often adopted to achieve a high level of fidelity. Experimental data is usually invaluable for that purpose. The driving force behind the mechanical effects in additive manufacturing is the heat source and how it is applied. The correct amount of heat, the speed of the heat source, and the resulting thermal magnitudes and gradients are critical for producing good structural integrity, which is free of voids and poor microstructure. A popular and effective approach for the thermal calibration of additive manufacturing FEA models is to experimentally measure melt pools under varied conditions and to use the data against FEA isotherms. FEA models can then be calibrated to produce the same measured melt pools for a comprehensive range of thermal conditions. In this reported work, a set of around 50 measurements have been carried out on the size of melt pools produced by different combinations of heat source power and speed. The experimental setup comprises a single pass deposited on a small aluminium alloy cuboid by applying laser powder bed fusion (L-PBF). The produced specimens have been cut at locations where steady state conditions are expected, and the melt pool width and height are recorded for each specimen. The data is then processed to calibrate the FEA model throughout the experimental design space. Conclusions are finally made on how thermal functions within the FEA model are formulated to achieve good simulation fidelity.