Turbulence unraveled

© 2012 Yu-Ting Wu

© 2012 Yu-Ting Wu

Researchers working with Fernando Porté-Agel are using lasers to make invisibles turbulence visible.

Clear air turbulence has long been a nuisance and a hazard to planes and passengers. It is invisible, unpredictable, and tends to strike when it is least welcome, for example during the inflight chicken salad lunch. But turbulence also occurs closer to ground, and being able to see it, or simply getting a better grasp of it, would be an important step forward for aviation and other industries.

Wind power generation is one such industry. Although wind may flow smoothly into a wind turbine, the rotating blades that slice through the moving air stir it up into a chaotic collage of whorls – or turbulence. Since it takes time for this turbulence to die down, a second turbine further downwind will be driven by wind that hasn’t yet had time to settle fully. This process can perpetuate itself through an entire wind farm and can lead to the often observed loss of productivity for downwind turbines; losses of up to 40 per cent in the worst known cases.

The trouble with turbulence is that it is difficult to measure. In wind farms, researchers usually rely on sensors mounted on poles or towers. But you just can’t set up enough sensors at enough locations within and surrounding a wind farm to get an accurate picture of where the main whorls tend to be, how far they extend, and how long they persist.

Using beams of bundled light – lasers – researchers at ENAC have found a way to solve this conundrum. By shooting a laser beam into the air and measuring the light that is reflected back by humidity or tiny particles that are naturally present in the atmosphere, a lidar – the laser equivalent of a radar – can provide a detailed profile of the air it traverses – almost in real time.

“Last year, for the first time, we used multiple lidars to measure the actual turbulent wake behind a wind turbine,” says Fernando Porté-Agel, from the Wind Energy and Renewable Energy Laboratory. These field measurements, performed on a wind turbine near Martigny, in the canton Valais, can be used to confirm predictions made using computer simulations of wind blowing through wind farms. And they also provide data that are otherwise difficult, or even impossible, to obtain.

Turbulent phenomena are complex, and it often takes more than one single experimental method to identify and understand them. “In our wind tunnel experiments, we found that the turbulence created by a wind turbine settled faster when the atmosphere was most turbulent, for example on a hot day, when the warm ground heats the air just above it, and sets it in motion,” says Porté-Agel. “Our computer simulations under the same conditions show the same behavior. And now, we would like to study the turbulence behind the wind turbines in Valais, Switzerland under these conditions to see if we can confirm this phenomenon outdoors and understand why it occurs.”

In the long run, a better understanding of these phenomena should lead to improved design of wind farms, taking into account not only the arrangement of the individual turbines within them, but also the surrounding topography, climate, and dominant winds, ultimately leading to more efficient exploitation of these expensive infrastructure projects, says Porté-Agel.

Their ability to visualize turbulence has already found applications in other fields. Solar Impulse, the lightweight solar airplane on a mission to fly around the globe without a drop of fuel, is so light that it gets bounced around by turbulence like a bird in a storm. At high altitudes the risk can be managed, but closer to ground, especially during take off and landing, a bounce in the wrong direction could be dangerous. Addressing this risk, members of Porté- Agel’s group collaborated with Solar Impulse to develop a portable lidar-based solution to scan the air around the airfield for dangerous whorls. Once a green light is given, the pilot can be assured that the plane will safely glide down to the runway for touchdown.