DARPA Awards $1.8M to Intact Solutions for AI-Driven, Pre-Qualified 3D Printed Components

In its latest show of support for the additive manufacturing (AM) industry, the Defense Advanced Research Projects Agency (DARPA) has awarded $1.8 million to Wisconsin software provider Intact Solutions to develop software backed by generative AI that enables pre-qualified design of metal AM parts. The company intends to make the software, Intact.Additive, applicable to many different metal AM processes, but will place an emphasis on compatibility with powder bed fusion (PBF).

Intact Solutions has previously received funding from DARPA as well as other US federal government agencies, including NASA, the National Science Foundation (NSF), and the National Institute for Standards and Technology (NIST). The company’s partners also include key players in the US AM market, such as nTop and Desktop Metal.

The idea behind Intact.Additive is to “fuse, correlate, and unify simulation, experimental, and process data,” in other words, to optimize the cost and time efficiencies of the manufacturing process by informing generative design with data from both virtual simulations and real-world experiments. Along these lines, the goal is reminiscent of a project at the center of an MIT research paper into composite AM I wrote about back in February, which demonstrated that machine learning, simulations, and physical experiments are each most effective when there’s synergy between all three.

Qualification and certification remain probably the biggest barriers to adoption of AM for end-use production, and this is truest of all, of course, when it comes to metal parts. On the other hand, organizations in both the public and private sectors continue to gradually make strides on that front. For instance, Gulf region service bureau Immensa and certification firm DNV recently released guidance concerning digitization of metal part files for the energy sector.

Pre-qualification remains something of a Holy Grail for AM design software, as it would maximize AM’s potential for enabling firms to go from design to production as quickly as possible. With that in mind, Intact Solutions’ work here could be a perfect project for the forthcoming AI for Resilient Manufacturing (AI MFG) institute.

As I noted regarding AI MFG, manufacturers are lagging other sectors in AI-deployment, in part due to hesitancy about the best way to leverage the technology for long-term return on investment. Pre-certification of metal AM parts could certainly be one critical solution to that problem.

And, the US government — and an agency like DARPA, in particular — is likely the ideal partner in working towards such an objective, given the unusual breadth and depth of AM data controlled by the Department of Defense (DoD). For all the talk about bringing in private industry to catalyze development of more dual-use technologies, the data DoD itself already controls is arguably the biggest potential catalyst for any such endeavor.

DoD has to figure out a way it’s comfortable with for sharing what it knows: a task that sounds impossible, but is in fact only almost impossible. Someone will discover how to square that circle.

Images courtesy of Intact Solutions