In the world of academia, the field of computational fluid dynamics [CFD] allows engineers and researchers to perform predictive modeling that can lead to faster, sleeker and safer vehicles in the air or on the ground.
More significantly for motor heads, CFD is also the main science behind Formula 1 racing, a high-octane international sport that will take place in Austin Nov. 16-18.
Using CFD-based computer simulations, aerodynamicists and other researchers can model how the laws of physics, racing conditions or design will affect an F1 car's performance. For the uninitiated, an F1 race car is the equivalent of an upside down plane. Not only can CFD result in faster lap times, but it can also improve vehicle safety for drivers who routinely go from about 185 to zero miles per hour in a matter of seconds.
What is less well known to F1 fans is that CFD has advanced in sophistication in part because of the mathematical and theoretical work being done right here on the Forty Acres.
"We are building the mathematical basis on which F1 engineers can use CFD," said Jayathi Murthy, chair of the Cockrell School’s Department of Mechanical Engineering.
While UT researchers are not yet directly involved with F1, there will likely be opportunities for collaboration in the future in areas such as transportation and energy, said Bobby Epstein (B.A. Plan II), chairman of the Circuit of the Americas and lead investor in the Formula 1 United States Grand Prix.
"This is a platform that will be viewed by 500 million people each year," Epstein said. "For us to have the opportunity to showcase the university and the research — it's pretty unique."
One area worthy of showcasing is the CFD expertise of UT's engineering and science faculty, which has given the university a leading global position in the field, Murthy said.
The Cockrell School has various CFD projects taking place across the school’s departments, including mechanical, aerospace, chemical and petroleum engineering. The Institute for Computational Engineering and Science (ICES) is home to some of the world's foremost experts in numerical methods and engineering simulation. Additionally, UT's Texas Advanced Computing Center (TACC) houses some of the country's fastest supercomputers, which are equipped to perform complex CFD calculations.
UT’s expertise in CFD falls under various categories, including: the development of the equations governing the physics of turbulent air flows; mathematical algorithms to help solve these equations; and using the power of the latest supercomputers to perform very large and complex simulations efficiently.
Murthy, who leads the National Center for Prediction of Reliability, Integrity and Survivability of Microsystems (PRISM), and her research team focus on computational techniques for flow, heat and mass transfer. The computational techniques born at UT are eventually used for CFD in the aerospace and automotive industries.
Another major area of study at the university is turbulence, which today isn't well understood or simulated. Robert Moser, a professor in the Mechanical Engineering Department and expert in turbulence modeling, is making scientific strides in understanding the physics of turbulence. He is leading an $18.7 million research project, known as PECOS, to simulate re-entry of vehicles from space.
Other researchers in the Cockrell School and at ICES are developing techniques for optimizing industrial design, for understanding complex physics and for increasing the accuracy of engineering simulations.
Meanwhile, a few trends taking shape at the university are continuing to push CFD to new levels.
One trend in engineering simulation is multi-physics, which refers to the integration of multiple engineering disciplines. Aerodynamics is just one part of what lends itself to predictive modeling, Murthy said, there are also areas like crash analysis, material design, electronics cooling, in-cylinder combustion and many others.
"It's not only mechanical, it is electrical, chemical and materials. There's going to be a lot more integration," Murthy said.
At the same time, the supercomputers used to conduct CFD experiments are constantly advancing. Under development are exascale machines, which will reach performance of at least one exaflop, which is 1,000 times faster than existing machines. TACC's Ranger — one of the largest computing systems in the world — performs at 579.4 trillion operations per second (or teraflops). TACC is building one of the country's fastest civil supercomputers, Stampede, which will have a peak performance of more than 2 petaflops.
That said, this enhanced computing power will require developing ever better computer algorithms.
The exascale machines "will need a whole new classification of algorithms to properly exploit their enormous power," Murthy said. "It's extremely exciting. And there is this whole thing I call integration with decision theory. We would like to use simulations to make better decisions about what designs to pick, what experiments to do, what technologies to invest in and the risk associated with these technologies, so I think where we are headed is tying decision theory to simulation — that's definitely an area for growth for us."
