Mission accomplished

In 2020, when Carnegie Mellon University entered into a five-year cooperative agreement with the U.S. Army Combat Capabilities Development Command’s Army Research Laboratory (ARL), their mission was to develop artificial intelligence approaches to enable the use of metal additive manufacturing (AM) in varied locations by a wide range of users.

Since then, more than 45 research projects have been undertaken by Carnegie Mellon researchers from across the university. Nearly all of the projects included multiple investigators, and most teams were interdisciplinary, bringing together experts from the Robotics Institute, School of Computer Science, and College of Engineering, including the Departments of Chemical, Mechanical, Electrical and Computer Engineering, and Materials Science and Engineering.

In October of 2024, the final AI Enabled Additive Manufacturing Workshop and Demonstrations were held on the CMU campus and at Mill 19 to showcase results of the most recent research. Like most of the projects funded by the program, the showcased projects delivered more than research results; they produced software products and techniques that were demonstrated to the Army on a yearly basis.

Given the multiple parameters and vast number of combinations of materials, designs, and processes that can be used to optimize the 3D printing, the program focused on the application of artificial intelligence and machine learning techniques and tools to nearly every aspect of additive manufacturing.

“AM is ripe for the application of AI, and Carnegie Mellon is deeply immersed in the AI that can deliver real solutions” said Jack Beuth, in his opening remarks. Beuth, the project’s principal investigator, co-director of Carnegie Mellon’s Next Manufacturing Center, and professor of mechanical engineering, also discussed project research he led on multi-sensor process monitoring, where data is acquired simultaneously from sensors such as high-speed cameras, infrared cameras, acoustic microphones, and photodiodes. AI algorithms are used to relate these data streams to enhance part quality. Laboratory-based demonstrations of these capabilities were given to the visiting ARL team.

Additive manufacturing is ripe for the application of AI, and Carnegie Mellon is deeply immersed in the AI that can deliver real solutions.

Jack Beuth, Faculty Co-Director, Next Manufacturing Center

Other demonstrations of software, hardware, methods, and techniques were presented throughout the two-day meeting.

Amir Barati Farimani, associate professor of mechanical engineering, showed how high-resolution simulations, deep learning, and experiments were used to characterize spatter particle dynamics. The software can enable individuals without prior technical expertise to effortlessly upload 3D models and configure the necessary processing conditions for 3D printing builds. Utilizing surrogate models and analytical solutions, the software can automatically determine the optimal printing parameters for processes such as Laser Powder Bed Fusion (LPBF) and effectively mitigate potential sources of defects like keyholing porosity and poor powder fusion.

Tony Rollett, co-director of Carnegie Mellon’s Next Manufacturing Center and professor of materials science and engineering, demonstrated software that is used to compile and consolidate large multi-modal data sets. By addressing data storage capacity; file naming conventions; and data retrieval standards, researchers will have better access to data needed for applying machine learning tools to advancing AM technology.

Mohadeseh Taheri-Mousavi, who is designing compositional alloys for powder bed fusion, demonstrated how combining Integrated Computational Materials Engineering (ICME), machine learning, and inverse design techniques can enable the discovery of target combinations of material properties. The assistant professor of materials science and engineering professor will continue that work to create advanced additively manufactured structural alloys that can sustain extreme environments with collaborators from the chemical engineering in research that is being supported by the Naval Nuclear Laboratory.

The extensive development of AI-based software for AM was reinforced by the ARL-funded acquisition of AM equipment, particularly of machines with open control systems that allow the exploration of manufacturing approaches beyond what current industry-targeted machines allow. These machines, coupled with existing industry standard machines on campus and at Mill 19 are allowing CMU Next Manufacturing Center researchers to apply AI methods to future processing scenarios while relating them to machines currently used in AM labs across the country. This full range of 3D printing equipment makes the Carnegie Mellon AM program one of the most comprehensive and well-equipped in the country.

Support for the many research projects gave students at all levels—undergraduates, master’s, and doctoral students—outstanding experience that prepares them for careers in the field.

Carnegie Mellon graduates are in high demand for AM jobs in industry, government, and academia thanks to the training and experience they receive here.

Jack Beuth, Faculty Co-Director, Next Manufacturing Center

“Carnegie Mellon graduates are in high demand for AM jobs in industry, government, and academia thanks to the training and experience they receive here,” said Beuth, who explained that nearly all graduate engineering students are also being trained in AI and machine learning, which further enhances their career prospects.

The ARL program also helped to forge a working relationship with Barnes Global Advisors, a leading firm in advanced manufacturing consulting and training. The group is eager to continue partnering with the university to create training programs that will further advance the adoption of AI-enabled additive manufacturing and expand its usability to more users at varying skill levels.

Pictured, top: Ph.D. students conduct research on the TRUMPF TruLaser Cell 3000 in the additive manufacturing lab at Mill 19.