Physicists Reconsider The Nature Of Turbulence

The findings were reported in the March 25 issue of Nature.

According to James A. Glazier, professor of physics at Notre Dame and principal investigator, it's important to know if turbulence is the same under all situations or whether there are changes in extreme situations.

"People have been beating their heads against the concept of turbulence for some time," he says. Russian scientist Andrey Kolmogorov first proposed a theory of turbulence in the 1940s that was "pretty good," a theory scientists today refer to as the classical theory of turbulence.

Up until Komogorov attempted to describe turbulence, physicists recognized four patterns of flow: still, steady convection, periodic convection, and turbulent convection. Each of these can be pictured by imagining a pan of water on a stove. When the burner is lit beneath the pan, the flow of the water will go through each of these transitions as it is heated. Among the factors that define the transitions are abrubt changes in the efficiency of heat distribution. Another defining factor is the complexity of the flow, which increases and becomes more and more disordered in each subsequent phase.

Essentially what Kolmogorov did was explain the flow pattern and transport properties of turbulence and distinguish turbulence from chaotic flow patterns, resulting in five recognized states.

Kolmogorov's theory was based on the idea of a self-similar energy cascade. That is, the energy supplied by the flow formed eddies, which in turn formed smaller eddies, and so on, and that each layer related to the one immediately above it in the same way.

Kolmogorov's theory was so convincing that for many years people believed that it provided an essentially complete picture of turbulent flow, despite a gradually increasing catalogue of small discrepancies. However, experiments by Albert Libchaber in the 1980s turned the experimental study of turbulence theory upside down and let Itamar Procaccia of the Weizmann Institute in Israel to theorize that Kolmogorov turbulence, defined by classical theory, could be divided further into "soft" turbulence and "hard" turbulence, thus making six recognized patterns of flow.

And the very theory that predicted hard trubulence also led to predictions by Eric Siggia of Cornell University of a seventh final and universal pattern of flow, that of ultra-hard turbulence. If the existence of such a transition could be proved, it would have important consequences. For example, says Glazier, all calculations of thermal transport in stars would have to be reworked. Its very idea helped initiate the National Turbulence Center proposed for Brookhaven National Laboratory, and research funds already have been spent studying the theoretical phenomenon.

While seveal scientists have claimed to have found evidence for ultra-hard turbulence in previous experiments using helium, Glazier believes these studies to be flawed and the Nature report offers explanations for at least a few of the claimed sightings. In the Notre Dame/Tohoku University experiment, the physicists used mercury, which is special in that it has a high thermoconductivity relative to its viscosity. The transition point in mercury predicted by ultra-hard turbulence believers is much lower in temperature difference, making it easier to find. "If you're going to find ultra-hard turbulence," Glazier says, "the easiest place to find it is in mercury."

But it isn't found, despite that the experiments were carried out at temperatures 100 times higher than the predicted point of transition to ultra-hard turbulence. "It's not a matter of not pushing the experiment hard enough," explains Glazier. "The theory very precisely predicts where ultra-hard turbulence would be expected. It isn't there. And if it isn't there, something in the theory isn't quite right."

Most recent experiments on turbulence have looked at the statistics of complex structures in the flow using multiple temperature or velocity sensors within the fluid. The Notre Dame/Tohoku University experiment went back to a more traditional, even old-fashioned technique, measuring the total heat transport of the fluid, which provides a quick aggregate view of the phenomenon.

This simplified the experimental design and allowed for a practical experiment on a larger scale than had been attempted previously. The liquid consisted of almost one ton of mercury driven by a heating coil that supplied up to 20,000 watts into a one square foot area.

The hard part was extracting the heat at the top in a controlled fashion, said Glazier. The research team used a cooling coil design borrowed loosely from nuclear power plant design, with more than 200 individual cooling loops and high pressure, high velocity cooling water.

"To get rid of the heat, we used two large air conditioning units designed to cool entire buildings," Glazier said. "Despite the rather brute-force technique, our control of the temperature was better than 1 percent, essential if we were going to obtain an unambiguous result."

In addition to Glazier, other scientists on the research team were Takehiko Segawa, Antoine Naert and Masaki Sano, all at Tohoku University.