Pore no more: US scientists develop real-time defect detection for 3D metal printing

Yahoo

9 hours ago

Scientists from the federally funded Argonne National Laboratory in Illinois and the University of Virginia have developed a new approach for detecting defects in metal parts produced by 3D printing.
By combining artificial intelligence (AI), X-ray imaging, and thermal imaging, this approach could pave the way for real-time self-correcting systems in the future. 3D printing of metals involves a method called laser powder bed fusion, where you build objects layer by layer by melting metal powder with a laser.
However, a big issue in this process is defects, especially keyhole pores, which are tiny holes that form when the laser melts too deeply. These pores weaken the final object, which is a significant concern when printing high-performance parts, such as rocket nozzles or surgical implants.
Keyhole pores also pose a significant challenge as they compromise the structural integrity of printed parts. These tiny voids form when excessive laser energy creates deep, narrow holes that trap gas, leading to internal cavities as the metal solidifies.
Recurrent microscopic keyhole pores can act as stress concentrators, increasing the risk of cracks or failure under pressure. This is particularly hazardous in critical applications, such as aerospace, automotive and medical devices, where part reliability is crucial.
Detecting and preventing keyhole pores is therefore vital in ensuring the performance, safety, and durability of additively manufactured metal components.
https://www.youtube.com/shorts/Lg_fVCxXhgY
To help achieve this, researchers developed a method to identify and predict these pores real-time using a combination of thermal imaging, X-ray imaging, and machine learning.
This new process utilizes powerful X-rays (from a government laboratory) to capture snapshots of what was happening inside the metal as it was being printed. A camera also recorded thermal images (from the surface) at the same time.
Then, a trained AI model was used to teach it how specific surface heat patterns predict pore formation. Once trained, the model could detect when a pore was forming just from the thermal image, with extremely high accuracy and within milliseconds.
Thermal cameras are already installed on several 3D printers. However, until now, they couldn't reliably spot internal defects. The new method developed by the collaborative team of scientists utilizes existing cameras and AI to instantly detect flaws, eliminating the need for expensive X-rays every time.
​"Our approach can readily be implemented in commercial systems. ​With only a thermal camera, the machines should be able to detect when and where pores are generated during the printing process and adjust their parameters accordingly," said Kamel Fezzaa, a physicist at Argonne and a member of the scientific team.
Eventually, this technology could be paired with automatic corrections, such as adjusting the laser or reprinting a layer, to fix problems as they occur. This makes 3D printing more reliable for mission-critical parts.
It could reduce waste, save money, and increase safety. "Our X-ray beams are so intense that we can image more than a million frames per second," added Samuel Clark, another physicist at Argonne.
"Next, the researchers will develop sensing technologies that can detect other types of defects that occur during the additive manufacturing process. The goal is to create a system that not only detects defects but can enable repairs during 3D printing," a release by the U.S. Department of Energy stated.
The study has been published on the website DOE Pages.