3D printing techniques have become well-established ways to produce custom parts and components, but there’s always room for improvement as the technology advances toward even more compelling applications.
Researchers with the University of Virginia (UVA) and Argonne set out to apply the latest artificial intelligence (AI) technologies to quality control of the metals used in 3D printing applications.
According to a report in Engineering & Technology, it can be a challenge to hunt down the occasional structural defects that form during the process of printing. Specifically, they refer to the process of laser powder bed fusion, where they contend that pores can form, compromising the part’s performance. In fact, they blame this as “one of the reasons why this approach has not become more widely adopted.”
Many 3D printing machines have sensors in place that monitor during the process, but the researchers have been able to pinpoint ways to take this inspection one step further — instead of inspecting the outside of the metal, they believe they can now inspect the inside too.
The process pairs x-ray beams and thermal imaging to see the thermal signatures created internally by pores, visible on the surface of a part. Then, according to the report, the researchers “trained a machine-learning model to predict the formation of pores within 3D metals using only thermal images.”
The result is not just hyper accuracy but also the ability for the defect-detecting AI tools to predict pore generation. Not to mention, the method is reportedly ready to be applied to existing commercial systems, where it can automatically stop the printing process when a defect is revealed.
Overall, the team hopes to ultimately create a process where defects can be detected — and even fixed — all during the manufacturing process.