**Abstract**

Powder Bed Fusion (PBF) is one of the most popular additive manufacturing methods employed extensively to fabricate complex parts especially in industries with stringent standard criteria, including aerospace, medical, and defense. DMLS/PBF fabrication of parts that is free of defects represents major challenges. A comprehensive study of thermal defects, contributing parameters, and their correlation is necessary to better understand how process specifications initiate these defects. Monitoring & controlling temperature and its distribution throughout a layer under fabrication is an effective and efficient proxy to controlling process thermal evolution, which is a completely experimental technique. This being highly costly specifically for metal printing, computer-based numerical simulation can significantly help the identification of temperature distribution during the printing process. In this paper, a multiscale modeling technique is demonstrated with commercially available software tools to correlate the defect generation in metal PBF process and significant process parameters. This technique can help efficiently design the process setting in addition to or even absence of experimental monitoring data. This research work is a part of a larger project of closed-loop control strategy development using physics-based modeling and graph-based artificial neural network implementation for reducing thermally induced part defects in metal 3D printed process.

**Keywords:** powder bed fusion, process parameters, defect generation, thermal anomalies, artificial neural network, in-situ monitoring, feedback control
