AI makes 3D printing more robust for use outside traditional factories

Industrial 3D printing is considered flexible, but it quickly reaches its limits outside controlled production environments. Researchers at the Rutgers School of Engineering now show how additive manufacturing can be stabilized with the help of artificial intelligence. Two recent studies led by Rajiv Malhotra address key weaknesses of modern manufacturing processes: susceptibility to disturbances and the high experimental effort required for process development.

The first study focuses on so-called expeditionary additive manufacturing. In this approach, components are produced in environments where vibrations, temperature fluctuations, or untrained personnel can impair print quality. Such scenarios are relevant, for example, in spaceflight, military operations, or disaster relief.

“We were trying to understand how we can make expeditionary additive manufacturing robust to such unknown and disruptive disturbances,” Malhotra said. “Whether or not your part will turn out or not,” Malhotra said, “can have missions fail completely. You could have people die.”

“We created a tool which addresses that issue,” Malhotra said. “We don’t have to anticipate anymore. Whatever disturbances come, we can deal with it without throwing away the part or stopping failure, both of which are bad for mission assurance.”

To achieve this, Malhotra’s team developed an AI-based control system using conditional reinforcement learning. A camera continuously monitors the printing process, while the AI detects deviations and adjusts printing parameters in real time.

“We trained the AI to expect the unexpected, rather than expect the expected,” Malhotra said. “We have created a new AI technique that ‘robustifies’ expeditionary manufacturing beyond the reach of literature. It reduces defects by 10 times or more, increasing quality by similar amounts even when the disturbances are not known in advance.”

Given that expeditionary manufacturing is critical for defense, space and disaster recovery applications, he said, “the resilience we achieved is both critical and hitherto unrealized in the state of the art.”

The second study, published in the Journal of Intelligent Manufacturing, deals with accelerating innovation cycles in manufacturing. Instead of relying exclusively on physical models or extensive experimental series, the developed system combines small datasets from real experiments with knowledge from scientific literature. Large language models extract relevant relationships from publications and link them with measurement data to generate reliable predictions. According to the researchers, around 30 samples were sufficient to achieve results that would otherwise require several hundred experiments.

“We cut short the samples that you have to make,” Malhotra said. “That means you’re doing things much faster. Our job really is to take an existing AI system and not spend $5 billion, but spend really next to nothing to say, ‘Find me a hypothesis that works for my case,’” Malhotra said.

This innovation could accelerate development in industries including aerospace, automotive, electronics, and defense. “This method reduces the need for human interpretation and large experiments, speeding up innovation for new or complex manufacturing processes,” he said.

Both approaches demonstrate how AI can make 3D printing more technically robust and economically efficient. Especially for applications outside traditional production halls, such systems could prove crucial in keeping additive manufacturing reliably usable in the future.