Summary
Peregrine is transforming additive manufacturing by giving industry real-time, artificial intelligence (AI)-powered quality control that reduces waste, accelerates certification, and enables the production of born-qualified parts.
Problem
Additive powder-bed fusion processes work by moving a laser or an electron beam over a bed of powder to melt tiny particles of material together. As the powder combines to form a solid object, a new layer is deposited, and the process repeats until the desired structure is formed.
Powder-bed printing can create components with complex geometries that other methods cannot achieve. Each factor, such as how well the powder is spread and the rate at which the material cools, must be carefully monitored and controlled to produce a high-quality part. Absent this monitoring, it is difficult to control component quality and qualify powder-bed-produced parts for use in high-risk applications like aerospace or nuclear power generation.
In powder-bed additive manufacturing, material quality failures can occur because of system processing errors or because of how the standard machine program or user decide to control the laser.
Scientific challenge
At the time of the research, component quality was primarily determined with an X-ray computed tomography (CT) scan after printing was complete. CT scanners were limited in the scenarios in which they could detect flaws, and they could not tell manufacturers what conditions led to the flaw. This limitation slowed innovation and increased the cost of both research and mass production.
Industry needed a new quality-monitoring tool that could evaluate the 3D-printing process in real time, while being versatile enough to function on industry-standard equipment.
Innovation
Leveraging the powder-bed expertise and equipment available at the US Department of Energy’s (DOE) Manufacturing Demonstration Facility (MDF) at Oak Ridge National Laboratory, researchers developed an AI software tool, known as Peregrine, that automatically detects material quality issues caused by machine processing errors or laser and electron beam control.
Peregrine uses a custom algorithm to examine the pixel values of camera images—searching for features such as edges, lines, corners, colors, and textures. Using this data, the software continuously checks for problems including uneven distribution of the powder or binding agent, spatters, over- or under-heating, part distortion, and porosity.
Engineers can then match flaws found during post-production analysis to the process data collected by Peregrine, thus providing crucial insights for improvement and quality control. Additionally, Peregrine can automatically alert operators of anomalies during the printing process so that real-time adjustments can be made.
Results
Peregrine can be used on nearly any powder-bed system and requires little more than a high-powered desktop computer and an off-the-shelf camera. MDF and its partners have successfully demonstrated the software on 23 powder-bed printers that use electron-beam melting, laser-powder bed, and binder jetting technologies.
Peregrine was first demonstrated in a field setting through the DOE’s Transformational Challenge Reactor program, which pursued the first 3D-printed fuel assembly rods for use inside a functioning nuclear reactor. The project was a landmark demonstration of digital manufacturing’s capability to improve US nuclear energy production by assisting in quality control, making components easier to qualify and certify for use in harsh environments.
Impact
Peregrine is an important step toward the “factory of the future,” in which custom parts are conceived using computer aided design and produced by self-correcting, automated manufacturing systems connected via a digital thread. Under this concept, products can be manufactured with less time, energy, capital, and material compared to conventional techniques.
The growing number of organizations with licenses for Peregrine include multiple DOE national laboratories, other federal agencies, academic institutions, and companies.
Additional research at MDF is correlating detected anomalies with critical material flaws and quantitatively estimating the quality of these 3D-printed components. As Peregrine matures and more data is collected, the tool is enabling the production of born-qualified components.
Support
This project was funded by DOE’s Advanced Materials and Manufacturing Technologies Office. Additional funding to further develop the technology was provided by the Office of Nuclear Energy Advanced Materials and Manufacturing Technologies program. — Logan Korn
Contact
Luke Scime, lead scientist and developer, scimelr@ornl.gov
Bob Slattery, industrial collaborations manager, slatteryrs@ornl.gov
Alex DeTrana, licensing, detranaag@ornl.gov, 865-341-0423
