NASA is turning to artificial intelligence to design spacecraft and mission hardware. Researchers have dubbed the specialized, one-off parts produced by this process “evolved structures,” which can save up to two-thirds of the weight of traditional components and tolerate higher structural loads while requiring only a fraction of the time needed to develop parts designed by humans. McClelland, a research engineer who pioneered the design, uses commercially available AI software to create these parts at NASA’s Goddard Space Flight Center.
To create these parts, a computer-assisted design (CAD) specialist begins by drawing in the surfaces where the part connects to the instrument or spacecraft, as well as any bolts and fittings for electronics and other hardware. The designer may also need to block out a path so that the algorithm doesn’t block a laser beam or optical sensor. More complex builds might require spaces for technicians’ hands to maneuver for assembly and alignment. Once all off-limits areas are defined, the AI connects the dots, producing complex structure designs in as little as an hour or two.
“The algorithms do need a human eye,” McClelland said. “Human intuition knows what looks right, but left to itself, the algorithm can sometimes make structures too thin.”
However, evolved structures also offer benefits beyond just their fast production time. Parts are analyzed using NASA-standard validation software and processes to identify potential points of failure. McClelland said, “We found it actually lowers risk. After these stress analyses, we find the parts generated by the algorithm don’t have the stress concentrations that you have with human designs. The stress factors are almost ten times lower than parts produced by an expert human.”
NASA has already adopted McClelland’s evolved components for missions at different stages of design and construction, including astrophysics balloon observatories, Earth-atmosphere scanners, planetary instruments, space weather monitors, space telescopes, and even the Mars Sample Return mission. The AI-designed parts have proven particularly useful in areas with tricky design requirements, where combinations of specific interfaces and exacting load specifications were proving to be a challenge for designers.
The EXoplanet Climate Infrared TElescope (EXCITE) mission, for example, turned to evolved structures to help develop a titanium scaffold for the back of its telescope. Here, an IR receiver housed inside an aluminum cryogenic chamber connects to a carbon fiber plate supporting the primary mirror. “These materials have very different thermal expansion properties,” said Goddard physicist Peter Nagler. “We had to have an interface between them that won’t stress either material.”
The EXCITE mission plans to use a near-infrared spectrograph to perform continuous observations of hot Jupiter-type exoplanets orbiting other stars, with a long-duration NASA Super-Pressure Balloon lofting the mission’s SUV-sized payload. An engineering test flight is planned as early as the fall of 2023.
AI-assisted design is a growing industry, with everything from equipment parts to entire car and motorcycle chassis being developed by computers. The use case for NASA, however, is particularly strong. “Here at NASA, we make thousands of bespoke parts every year,” McClelland said.
The future of AI-assisted design lies in 3D printing with resins and metals, which will enable larger components such as structural trusses, complex systems that move or unfold, or advanced precision optics. “These techniques could enable NASA and commercial partners to build larger components in orbit that would not otherwise fit in a standard launch vehicle,” McClelland said. They could even facilitate construction on the Moon or Mars using materials found in those locations.