A group of scientists, some who conduct experiments in a lab and others who use computers to create simulations, are working together to do what they call “predictive science.” Data from laboratory experiments at the University of Illinois Urbana-Champaign inform computationalists who, in turn, write complex equations that require a supercomputer to solve. The simulations they do require some of the largest computers in the world and help predict conditions that are unachievable in a lab.
“We seek to do simulations that are truly predictive to impact the design of scramjets—the technology for airbreathing hypersonics propulsion. We have a good idea of the physics required, including the high-speed flow, combustion kinetics, transport, and the degradation of the new composite materials we are considering,” said AE’s Jonathan Freund. “We have a suite of physics-targeted experiments and physics-integrated experiments to inform our integration of models. And we have our ACT II arcjet combustion tunnel against which to test our integrated physics predictions.”
Freund is the director of the Center for Exascale-enabled Scramjet Design. He leads a team of 13 faculty researchers, 7 postdocs and research staff, and 19 graduate students who are studying various aspects of this predictive science challenge.
One of the questions being tackled by experiments and simulation is about how carbon oxidizes at different pressures, temperatures, gas environments. He said Ph.D. candidates Kaan Kirmanoglu and Nicholas Anderson provide one example of the many synergetic relationships between experimentation and simulation within the center.
“I use Nick’s experimental data to build and validate my models/simulations,” said Kirmanoglu. “But sometimes simulation results lead to reevaluating experiments. Together, we learn which experimental configuration and data are most suitable for model validation and how simulations can best represent experimental conditions. We also do a lot of brainstorming and calculations together to tackle the problems we encounter. Although the end goal is experimental results informing simulations, we work two-way to build the tools needed.”
Freund said their close pairing is part of the bigger model the center is tasked with. ACT II can simulate the interior flow of a scramjet if it were flying at supersonic speeds of Mach 6 to Mach 8. Ultimately, the simulations will represent turbulent combustion, oxidation, and pyrolization of novel composite materials.
Another one of the experiments in the center is testing a thermal protective materials made with carbon fibers. It is coated with a lightweight material designed to evaporate and protect the surface of a vehicle under supersonic conditions.
“We’re creating multiphysics simulations simulations to affirm that we can indeed do simulations that are truly predictive,” Freund said. We’re developing models for things people have not represented before,” Freund said. “But the predictive science methods, the approach we are developing to handle uncertainties, is a new way of looking for physics that is missing. That part of the effort is general and can be used in many applications outside of scramjet design.”
The Center for Exascale-enabled Scramjet Design is an interdisciplinary center pulling from faculty experts in The Grainger College of Engineering and the National Center for Supercomputing Applications. CEESD is in its third year of a five-year project funded by the Department of Energy. It is supported by National Nuclear Security Administration, Advanced Simulation and Computing, and Predictive Science Academic Alliance Program III.