Vision-based in-situ monitoring system for melt-pool detection in laser powder bed fusion process

Abstract

A low-cost, lateral off-axis non-IR imaging system is proposed to inspect the melt-pool in the Laser Powder Bed Fusion (L-PBF) processing of IN718 powder and stainless steel 316 L powder. In the proposed system, a perspective transformation method is applied to transform the captured side-view image into a top-view image and the melt-pool boundary is then determined based on an inspection of the gray-level intensity profile of the captured image and its first derivative. For the IN718 L-PBF process, the measurement results obtained for the melt-pool width and melt-pool length vary by no more than 13% and 3%, respectively, from the experimental results reported in the literature. For the SS316L, the measured melt-pool width is within 6% of the ex-situ measurements obtained by optical microscopy, while the melt-pool length differs by no more than 9% from the numerical predictions. Overall, the results show that the proposed system provides a viable means of detecting potential build failures during the L-PBF process and improving the yield of the fabrication process as a result.