For decades, the U.S. Navy and aerospace industry have struggled with the slow, fragile supply chains that keep submarines and aircraft operational.
A single replacement part often takes months to arrive, and if it contains a flaw, the wait begins all over again.
Many of these components have traditionally been made by small, specialized machine shops that are dwindling in number, leaving the defense sector searching for a faster, more reliable solution.
Researchers at Virginia Tech believe they may have found one.
A new study led by Associate Professor Prahalada Rao, published in Materials & Design, demonstrates how artificial intelligence can make additive manufacturing—essentially 3D printing with metal—faster, smarter, and far more dependable.
Rao focuses on a method called wire-arc additive manufacturing, which he describes as “welding in 3D.”
Unlike other forms of metal 3D printing that build parts slowly, wire-arc can deposit 40 to 50 kilograms of metal in a single day.
This speed makes it suitable for large, mission-critical components, but it comes with a challenge: ensuring the massive volume of molten metal is laid down flawlessly.
“Conventional machining is slow, wasteful, and often reveals defects only after weeks of work,” Rao explained. “With wire-arc additive manufacturing, we can produce large parts much faster. The key is guaranteeing quality as we go.”
That’s where AI steps in. Rao’s team has trained machine learning algorithms to monitor the “melt pool”—the pool of hot metal created during printing—in real time.
By teaching the system what a “good” melt looks like versus a “bad” one, the AI can predict defects with about 90 percent accuracy before they become serious. This allows the printer to correct errors instantly, saving time, materials, and money.
This research comes at a critical moment for the Navy, which needs quicker turnaround times to maintain its fleet. The ability to print reliable parts on demand could dramatically reduce delays and costs, while also making the system less dependent on faraway machine shops and aging workforces.
Rao’s work is part of Virginia Tech Made: Center for Advanced Manufacturing, a campus-wide effort that combines advanced materials, computational design, and data-driven process control to push Industry 4.0 forward.
Much of the testing happens in Rao’s Kelly Hall lab and the Learning Factory, the university’s hands-on hub for modern manufacturing. There, students and researchers work with the same cutting-edge 3D printers now being adopted in industry.
“Additive manufacturing is something students can pick up quickly, and having access to the latest machines prepares them for real-world jobs,” Rao said. “Our goal is simple: make parts faster, cheaper, and better. With AI, we can finally do all three.”
If this approach continues to advance, the days of waiting months for a single replacement part may soon be over—ushering in a new era of smarter, faster, and more resilient military and aerospace manufacturing.
