The future of unpolluted energy is hot. Temperatures hit 800 Celsius in parts of solar power plants and advanced nuclear reactors. Finding materials which will stand that type of heat is hard . So experts look to Mark Messner for answers.
A principal mechanical-engineer at the U.S. Department of Energy’s (DOE) Argonne National Laboratory, Messner is among a gaggle of engineers who are discovering better ways to predict how materials will behave under high temperatures and pressures. The present prediction methods work well, but they take time and sometimes require supercomputers, especially if you have already got a group of specific material properties—e.g., stiffness, density or strength—and want to seek out out what type of structure a material would wish to match those properties.
“You would typically need to run plenty of physics-based simulations to unravel that problem,” said Messner.
Looking for a shortcut, he found that neural networks, a kind of AI (Artificial intelligence) that uncovers patterns in huge data sets, can accurately predict what happens to a material in extreme conditions. and that they can do that much faster & easier than basic simulations can.
Messner’s new method found the properties of a material quite 2,000 times faster than the basic approach, as reported in an October 2019 Journal of Mechanical Design article. Many of the calculations, Messner realized, could run on a daily or regular use laptop with a graphics processing unit (GPU)—instead of a supercomputer, which are often inaccessible to most businesses.
This was the 1st time anyone had used a so-called convolutional neural network—a sort of neural network with a different’, simpler structure that’s ideal for recognizing patterns in photos—to accurately recognize a material’s structural properties. it’s also one among the primary steps in accelerating how researchers design and characterize materials, which could help us move toward a totally clean energy economy.
Cats on the web play a task
Messner began designing materials as a postdoctoral researcher at DOE’s Lawrence Livermore National Laboratory, where a team sought to create structures on a 3D printer at a scale of microns, or millionths of a meter. While leading edge, the research was slow. Could AI speed up results?
At the time, technology giants in Silicon Valley had started using convolutional neural networks to acknowledge faces and animals in images. This inspired Messner.
“My idea was that a material’s structure is not any different than a 3D image,” he said. “It is sensible that the 3D version of this neural network will do an honest job of recognizing the structure’s properties—just-like a neural network learns that a picture may be a cat or something else.”
To test his theory, Messner took four steps. He:
- Designed an outlined square with bricks—like pixels;
- Took random samples of that style and used a physics-based simulation to make 2 million data points. Those points linked his design to the specified properties of density and stiffness;
- Fed the two million data points into the convolutional neural network. This trained the network to seem for the right results;
- Used a genetic algorithm, another sort of AI designed to optimize results, along side the trained convolutional neural network, to seek out an overall structure that might match the properties he wanted.
The result? The new AI method found the proper structure 2,760 times faster than the basic physics-based model (0.00075 seconds vs. 0.207 seconds, respectively).
New AI tools boost nuclear innovation
This abstract idea might transform how engineers design materials—especially those meant to face up to conditions with high temperatures, pressures and corrosion.
Messner recently joined a team of engineers from Argonne and DOE’s Idaho and Los Alamos National Laboratories that’s partnering with Kairos Power, a nuclear startup. The team is creating AI-based simulation tools which will help Kairos design a molten salt nuclear-reactor , which, unlike current reactors, will use molten salt as a coolant. With those tools, the team will project how a selected sort of stainless steel , called 316H, will behave under extreme conditions for many years .
“This may be a small, but vital, a part of the work we do for Kairos Power,” said Rui Hu, a nuclear engineer who is managing Argonne’s role within the project. “Kairos Power wants very accurate models of how reactor components are getting to behave inside its reactor to support its licensing application to the Nuclear Regulatory Commission. we glance forward to providing those models.”
Another promising avenue for this type of work is 3D printing. Before 3D printing caught on, engineers struggled to truly build structures just like the one Messner found using AI in his 2019 paper. Yet making a structure layer by layer with a 3D printer allows for more flexibility than traditional manufacturing methods.
The future of engineering could also be in combining 3D printing with new AI-based techniques, said Messner. “You would give the structure—determined by a neural network—to someone with a 3D printer and that they would print it off with the properties you would like ,” he said. “We aren’t quite there yet, but that is the hope.”
The research was published on Journal of Mechanical Design.
