Answers are grounded in this paper and cite their sources.
Melt track defects (e.g., internal porosity, surface defects) can significantly affect printing quality. However, the correlation between these defects and molten pool variations, as well as their underlying formation mechanisms, remains unclear. This study investigates sequential molten pool variations in Laser Beam Powder Bed Fusion (PBF-LB) process. A Feature Pyramid CBAM Network (FP-CBAM Network) machine learning approach and t-SNE method are proposed to identify correlation patterns between molten pool morphological changes and the morphology of melt tracks. The overall accuracy of the proposed network for classifying melt track defects is 97.33%. Through schlieren imaging, spatter dynamics, vaporization phenomena, and molten pool are captured. Furthermore, the interplay coupling mechanism among molten pools, spatter, and vapor are analyzed, which mutually validates the feasibility of FP-CBAM Network. The quick transitions between molten pool (MP) lead mode and MP lag mode in molten pool motion induced abrupt, changes in the molten pool's overall moving velocity, which is associated with necking defects. The high-velocity aerial spatter deposition phenomenon is caused by spatters re-entering the laser beam, remelted and evaporated, which contributes to edge protrusion. The vapor jet angles near 90°, which simultaneously impacts both front and rear melt regions, generates spatters with dispersion areas significantly larger than the vapor coverage zone, resulting in increasing the probability of spatter deposition on melt track. This study provides a solution of enabling dynamic adjustment of processing parameters based on the melt surface defect prediction to improve melt quality. • A FP-CBAM Network maps the relationship between molten pool characteristics and melt track morphologies • In-situ schlieren imaging investigates the underlying mechanism • Rapid transitions in molten-pool motion are correlated to necking defects • Spatters re-entering the laser beam contributes to edge protrusion • Vapor jet angles approaching 90° increase the probability of particle adhesion
No takes yet. Share an insight, caveat, or question.
Lin et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: