Novel non-destructive evaluation techniques in Metal Additive manufacturing: Leveraging AI and high-resolution optical imaging for fine feature detection and dimensional qualification

The growing adoption of metal additive manufacturing (AM), particularly powder bed fusion (LPBF), has transformed the manufacturing landscape by offering unique design freedom, reduced material waste, and improved supply chain bottlenecks. The widespread acceptance of metal AM in the Air Force and the commercial industry is hindered by challenges associated with ensuring the dimensional accuracy and integrity of printed parts. Current methods for dimensional verification often rely on post-build inspections, which can be time-consuming, costly, and impractical for complex geometries. In response to these challenges, we introduce Addiguru ZOOM: Zone-Oriented Optical Monitoring, an innovative methodology designed to revolutionize dimensional verification in metal AM. Leveraging advancements in optical sensing, computer vision (CV), and artificial intelligence (AI), Addiguru ZOOM aims to provide real-time insights into part dimensions and geometries during the metal LPBF process. By integrating high-resolution optical sensors into the manufacturing environment, Addiguru ZOOM enables manufacturers to monitor and control the printing process with improved accuracy and efficiency. Addiguru ZOOM is the Zone-Oriented Optical Monitoring approach, which focuses on capturing fine feature measurements and delivering dimensional data in real-time. By dividing the build area into distinct zones and strategically placing optical sensors, Addiguru ZOOM can monitor the formation of printed features with high resolution and detail. This granular level of monitoring enables early detection of defects, deviations, and anomalies, allowing operators to intervene promptly and optimize process parameters for improved part quality. To validate the efficacy of Addiguru ZOOM, we will conduct experiments on an EOS-M290 or equivalent LPBF machine. The experimental setup will involve the installation of high-resolution optical sensors equipped with advanced AI and CV algorithms for feature detection and dimensional analysis. Throughout the printing process, the optical sensors will capture real-time data on part geometry and feature dimension, later verified by actual measurements. By providing accurate and comprehensive data during the printing process, Addiguru ZOOM reduces the reliance on post-build inspections, such as computed tomography (CT) scanning, which can be costly and time-consuming. The successful implementation of Addiguru ZOOM has the potential change the game of metal additive manufacturing by providing manufacturers with insights into the printing process and their part geometries.