“`html

U-M spearheads new DOE-supported computational center concentrating on future hypersonic flight

What advancements could engineering achieve with an artificial intelligence oracle capable of addressing any physics inquiry?

This kind of machine represents the overarching goal of the newly established Center for Prediction, Reasoning, and Intelligence for Multiphysics Exploration or C-PRIME, directed by the University of Michigan and sponsored by the U.S. Department of Energy’s National Nuclear Security Administration.

Although physics is regulated by numerous established equations, transitioning from those equations to predicting how physical objects will react—such as the eddies of fuel and air within a sophisticated engine or the exact air resistance acting on a vehicle’s surface—is challenging. In theory, everything is ascertainable, building from the molecular foundation, but the calculations are too extensive to execute.

While an AI strategy cannot tackle that issue directly, an AI entity could create physics models founded on known equations, utilizing them to generate reliable data. It could further apply that data to generate simplified but precise models for specific physics challenges, which would then contribute to the engineering design of intricate devices.

“The concept is that we, as humans, should provide specific principles we trust—Newton’s laws or E=mc^2. The machine subsequently assembles more intricate ideas from these fundamental components,” stated Venkat Raman, head of C-PRIME and the James Arthur Nicholls Collegiate Professor of Engineering.

“Since we rely on these foundational components, we can—largely—trust engineering concepts constructed from them.”

Nonetheless, formally establishing this trust, termed verification and validation, presents its own intricate challenge, which lies at the project’s core. The sequences of simulations created by the AI entities will operate on some of the globe’s largest supercomputers to uncover the operational mechanisms of propulsion systems behind hypersonic flight—fivefold the speed of sound. The team will concentrate on rotating detonation combustors, which are becoming pivotal technology for hypersonic flight.

Rotating detonation combustors are applicable for propulsion—in rockets, air-breathing engines, or satellite thrusters—or energy conversion, such as in gas turbines generating electricity. They offer considerable efficiency, approximately 25% superior to traditional combustion, but maintaining their combustion requires careful handling. A series of explosions circulate around a ring, with the resulting shockwave compressing and igniting the fuel-air mixture sequentially at each injection point.

“AI and hypersonics are vital to national defense and U.S. scientific leadership, and we’re dedicated to advancing technologies and expertise to propel both fields forward,” remarked Karen A. Thole, the Robert J. Vlasic Dean of Engineering. “This federal investment allows our researchers to unite knowledge in physics, computer simulation, AI, and machine learning to extend the frontiers of what’s achievable and nurture the next generation of AI-literate professionals in the process.”

Student researchers participating in the project will leverage the University of Michigan’s Ph.D. program in Scientific Computing—the first of its kind in the nation, established in 1988—managed by the Michigan Institute for Computational Discovery and Engineering, or MICDE.

The initiative is categorized into five research areas:

Furthermore, specially designed lab experiments will evaluate the precision of the AI-driven combustor design, which will be conducted at U-M by Mirko Gamba, professor of aerospace engineering, along with Carolyn Kuranz, professor of nuclear engineering and radiological sciences.

“Through our investigation and the training of the forthcoming generation of researchers, we possess the chance to influence the discipline on a vast scale,” stated Johnsen. “Specifically, it is crucial that our trainees—undergraduate and graduate scholars, alongside postdoctoral researchers—grasp how to utilize AI resources in their inquiries, as their achievements post-Michigan will hinge on their ability to do this effectively.”

David Etim, federal program manager in the Office of Advanced Simulation and Computing at the National Nuclear Security Administration, expressed strong admiration for the new center, which is part of the fourth phase of NNSA’s Predictive Science Academic Alliance Program.

“This center, with its emphasis on AI-driven solutions for intricate physics issues, aligns seamlessly with PSAAP’s mission to enhance high-fidelity predictive simulations,” Etim remarked. “We are eagerly looking forward to the pioneering contributions C-PRIME will provide in vital areas related to national security, particularly in next-generation hypersonic travel and exascale computing, thereby reinforcing the program’s influence.”

C-PRIME capitalizes on U-M’s established leadership in computational science and engineering, fortified by MICDE. U-M also hosts a $15 million Strategic Partnership and Accelerated Research Collaboration with Los Alamos National Laboratory, coordinated by MICDE, which connects Los Alamos staff scientists with U-M researchers. Additionally, the institution is collaborating with LANL on a $1.25 billion facility for advanced computing and AI research in Michigan.

C-PRIME is composed of a total of 13 co-investigators from U-M across four departments, along with a co-investigator from Princeton University. Researchers from Sandia, Los Alamos, and Lawrence Livermore national laboratories will be partnering with the center.

Raman is also an educator in aerospace engineering and mechanical engineering. Thole also teaches mechanical engineering and aerospace engineering. Duraisamy additionally serves as a professor of mechanical engineering and nuclear engineering and radiological sciences.

“`