|About this Abstract
||7th World Congress on Integrated Computational Materials Engineering (ICME 2023)
||Automated Characterization of Generated Meltpool from High Speed Camera in Advanced Manufacturing
||Kristen Jo Hernandez, John Lewandowski, Roger French, Laura Bruckman, Jayvic Cristian Jimenez, Thomas Ciardi, Sameera Nalin Venkat
|On-Site Speaker (Planned)
||Kristen Jo Hernandez
Metal-based additive manufacturing success requires active monitoring solution for assessing quality and reliability of components. Monitors such as high speed cameras are a common and well-studied capture device for in-situ welding. The scale and time required to fabricate parts increases exponentially the number of relevant frames for feature detection, with higher resolution increases the number of feature occurrences. Techniques for automated feature extraction and quantification are required in order to perform active sensing so that errors or problems in the print process can be addressed. Laser-solid interactions present in meltpools, such as the uniformity of melt, are known predictors of build quality. Automatically determining major, minor axis, melt coordinate center, and overall area offers the first steps of using in situ measurements in real-time and obtaining characterization information to inform on print quality with minimal human intervention through the use of U-Net feature extraction and YOLO image segmentation.