Clara Velte has taken on a major task: She will attempt to use equations to describe what US physicist Richard Feynman called the most important unresolved problem in classical physics—turbulence. With the support of two large foundations—the European Research Council (ERC) and Poul Due Jensens Fond—she has just opened an advanced experimental laboratory in the fluid hall at DTU Mechanical Engineering. She will attempt to combine empirical data and theory to make new discoveries here.
Turbulence is a well-known phenomenon. When you stir milk into your coffee, water runs out of the bathtub, or smoke rises from a flame, a pattern of vortices quickly forms in the flow. While the liquid or gas moves as a cohesive mass at low speed, it behaves predictably—it is ‘laminar’. But as soon as it reaches a critical speed, it becomes turbulent.
Even though the turbulence pattern is unique every time, it is believed that a certain type of flow will always have the same basic characteristics and statistics. But even with statistical descriptions, we cannot currently understand the flows created in a jet engine, on the leeward side of a wind turbine, when a liquid is pumped over a wall, etc. And it is a problem in many industrial contexts that we cannot predict and model how turbulence will develop.
Better understanding of turbulence
Each area of industry has its own way of tackling this unpredictability. The methods are based on experience of how the flow usually behaves, but they could be much better if we had a deeper understanding of what happens.
With a better theoretical knowledge, it would be possible to mix more effectively or better control the process. Golf balls, for example, have small dimples on the surface to make the air flow around the ball turbulent. This increases the mixing of energy-rich air close to the ball and reduces drag, so that the ball can fly further. However, on a wind turbine blade it may be better in some cases to maintain the laminar flow and avoid turbulence. A micro-vortex generator can be used to create this effect. But a deeper understanding of the processes that lead to laminarization is lacking.
“It’s not realistic that we will ever reach the point of predicting exactly which direction the vortices of a turbulent flow will take. But we believe it’s possible to describe turbulence better using statistics. We will attempt to do this over the next five years,” says Clara Velte.
80-year-old theory
In 1941, Russian mathematician Andrey Kolmogorov presented a theory that could explain how the energy in a flow is dispersed. He made an analogy between the small and medium-sized vortices in turbulence and the molecules of a thermodynamic system, and assumed that the small vortices are in a kind of equilibrium relative to each other.
According to his theory, all vortices get their energy from slightly larger vortices, and eventually the smallest vortices turn into heat. He did not believe that the small and medium-sized vortices would be affected by the dynamics of the large vortices.
Clara Velte explains, using the cup of tea sitting in front of her:
“The tea is in equilibrium. Even though the molecules in it are moving quickly, the liquid does not jump up by itself, and the movement follows the classic theory. But there are many cases where turbulence is more unpredictable, such as when a flow accelerates, when you stir a cup of tea, when diesel is injected into an engine, or air passes the rotor plane of a wind turbine. ‘Shear layers’ then arise, where one part of the flow moves faster than another. And classical theory appears to breaks down.”
Not even the most powerful computers can solve the Navier-Stokes equations, which describe turbulence movements. People therefore have to resort to approximations and rules of thumb to explain how a smooth, neat flow can break up into chaotic vortices.
Clara Velte finds the shear layers the most interesting aspect to study, and the newly designed laboratory will give her a unique opportunity to combine her theories with experiments, and thereby hopefully find solutions to the determining equations for turbulence fields.
Knowing is better than just believing
Clara Velte studied civil engineering at Chalmers technical college in Gothenburg, and joined DTU as a PhD student after completing her master’s degree in 2005. She brought with her the seed of an idea, that her supervisor at Chalmers planted when he talked about the shortcomings of turbulence theory.
She was offered permanent employment at DTU, and the idea eventually evolved into an application to several foundations, for funds to set up a test laboratory. It was no easy task. Her applications were repeatedly rejected, but instead of giving up, she honed her arguments.
“Many believe that the classic theory has been ‘proven’ through experimentation. But that is not how it works. You can only confirm that the experiment is consistent with your observations. I can find a thousand methods to measure that the Earth is flat, while other measurement strategies can show that it curves or is almost round. The way the experiment is designed can influence the outcome. I want to push the boundaries of the assumptions in classical theory in my experiments—especially the equilibrium assumption. I will look at the actual physics and work out how it interplays with our theoretical work,” says Clara.
“Everyone knows that the old theory does not fully work and is purely based on assumptions. Yet they still considered it a law of nature. It has almost evolved into a religion. If we end up disproving the old theory, the scientific community must accept it. But we may also find that much of the old theory works. And then we hope that we can show where it fails and why.” In any case, Clara and her team hope that they end up finding a more accurate description of turbulence.
“I know—as does everyone else deep down—that there is something missing. Companies certainly have problems with it. That is why they are showing interest in our work, and I really hope we can help them.”
Read more in the article 'Wind tunnels, soap bubbles, and super-fast cameras'