11/01/92

CONTACT: Stanford University News Service (650) 723-2558

SUPERCOMPUTERS AID SEARCH FOR TURBULENCE CONTROL

STANFORD -- Using supercomputers as their microscopes, scientists at the world's largest center for the study of turbulence can now explain why the tiny ridges on a shark's skin help it swim faster.

The riblets, no deeper than the grooves on a phonograph record, are just the right height and distance apart to keep water currents from swirling down between them and pumping high-speed fluid against the bulk of the shark's body surface.

Studies like this hold the promise of massive energy savings for air, land and water transportation, said Parviz Moin, a Stanford University professor of mechanical engineering who directs the Stanford/NASA- Ames Center for Turbulence Research. Moin and his colleagues have calculated that it is possible to amplify or reduce the impact of turbulent gas and liquid flows by nearly 30 percent.

Other potential applications include improvement of indoor and outdoor environments and even the flow of blood and air inside an ailing human.

Turbulence, the complicated flows of liquids and gases through and around objects, has been called by the late Nobel physicist Richard Feynman "the last great unsolved problem of classical physics." Turbulence researchers at the Stanford/NASA center have been among the largest users of supercomputers in the world, and the turbulence problem has been a key driver in the development of supercomputers, Moin said.

Thanks to their parallel processors, supercomputers have become a microscope that Moin and his colleagues can use to see and calculate what might be called the atomic structure of simple turbulent motions, such as the water motion around a shark's riblets.

Turbulence is not the random chaos once thought, Moin said; it is a mixture of unpredictability and order, with enormous variation in the size of current patterns. In a simple flow of water through a pipe, for example, swirling currents or eddies have diameters of vastly different sizes. A small error in measurements of the pressures and velocities at one point can lead to significant changes in the prediction of the flow ahead, Moin said. That principle shows in other areas, such as the problem weather forecasters have making five-day predictions.

The motion involved even in a slow, simple flow has hundreds of active degrees of freedom, known as dimensions, Moin and postdoctoral fellow Laurence Keefe reported in the September issue of the Journal of Fluid Mechanics. This is unlike chaos in simple dynamic systems, which is usually confined to a handful of degrees of freedom, Moin said.

Nevertheless, Moin and his colleagues have been able to place over a flow a grid with about 10 million measuring points. They then numerically solve the Navier-Stokes equations, the partial differential equations that govern the motion. Because of the non-linearity of these equations and the high dimensionality of turbulence, the detailed features of turbulence could not be calculated until supercomputers.

It took about 500 supercomputer hours to numerically simulate the structure of water flowing around a shark's riblets, Moin said.

"The number of grid points that would be required to simulate the flow of air around a 747 is so large that I don't think we'll have the brute force to do it even in the next generation," he said.

Microscopic views of simple flows, however, may bring a "real pay- off in the near future," he said. He and colleagues Sandip Ghosal and Tom Lund have developed a method to combine their direct numerical simulations of the largest eddies with procedures for more accurately characterizing the effects of the smaller currents on the larger ones.

"You can't just ignore the effect of the small stuff you have missed," Moin said.

The ultimate goal is to predict flows and better manage them with active and passive controls.

Significantly more drag reduction can be obtained by active control devices - such as the shark's ridges, or "walls" that change position based on real-time readings of conditions from on-board microsensors, Moin and student Haecheon Choi have shown. Such controls aren't economically feasible yet "but at least we've demonstrated what may be possible."

"Given the elegant design of the shark unaltered by evolution," Moin said, "it's likely sharks have been moving their riblets in tune with the currents for millions of years."

-kpo/turbulence-

921101Arc2043.html


This is an archived release.

This release is not available in any other form. Images mentioned in this release are not available online.
Stanford News Service has an extensive library of images, some of which may be available to you online. Direct your request by EMail to newslibrary@stanford.edu.