One of the biggest challenges in additive manufacturing today is the need for greater precision and reproducibility in 3D-printed parts. A team of Purdue University researchers may have found an intriguing solution: artificial intelligence.
“We have developed automated machine learning technology to help improve additive manufacturing,” said Arman Sabbaghi, assistant professor of statistics. Sabbaghi and his team developed a new model-building algorithm and computer application for geometric accuracy control that could potentially improve part printing.
The machine learning technology quickly corrects computer-aided design models and produces parts with improved geometric accuracy. This method helps to ensure that the produced printed parts conform more closely to the design and remain within needed tolerances. Machine learning also leads to improved consistency, assuring that the part will perform the same way even if it’s printed on a different machine some time in the future.
This new technique will be useful for many industries—such as the aerospace industry, where exact geometric dimensions are crucial to keeping hundred-million-dollar aircraft operating safely and reliably.
Additive manufacturing has become an essential tool because of the advantages it brings to the manufacturing process: the ability to create more complex shapes, reduce waste, speed up production, and reduce costs. It has grown into a $7.3 billion industry.
But 3D printing has some significant constraints that limit its potential. Printing processes still need to be overseen by humans, who need to monitor them constantly to ensure they run smoothly. Time and resources can be wasted when creating complex parts because of the quality testing that technicians need to perform when a print is complete. And often parts need to be designed to compensate for inherently weaker sections of the printed product. In addition, 3D printing also has a scaling problem: while small components are simple enough to print, larger ones can be prohibitively complicated and expensive to produce.
With this new development, part consistency and accuracy could be substantially improved. Sabbaghi said the technology also allows users to create complex designs that would not be possible with traditional manufacturing processes.
“We’re really taking a giant leap and working on the future of manufacturing,” said Sabbaghi. “This kind of innovation is heading on the path to essentially allowing everyone to be a manufacturer.”
Read more about new uses for AI at Hot New Research Uses Machine Learning to Help Buildings Save Energy.