CU study a step toward more-efficient wind farms
By Clint Talbott
Being first in line has its advantages, even for wind turbines, which are propelled by comparatively smooth wind flow that helps them produce near-optimal power at varying wind speeds.
But in the first row’s wake, turbulence increases and air speed drops, cutting turbine power production and even causing excessive wear on turbines. How that turbulence dissipates is not fully understood.
To start filling that gap in understanding, researchers at the University of Colorado Boulder have performed the first direct measurements of dissipation rates of wind turbines in real-world conditions. Wake turbulence might dissipate faster than commonly thought and in ways existing weather models do not reflect, their study suggests.
Julie Lundquist, CU-Boulder assistant professor of atmospheric and oceanic sciences, presented these findings during a scientific session on weather-driven renewable energy at the fall meeting of the American Geophysical Union in San Francisco.
“We care about wakes not just because they’re beautiful, but studying them will help us understand how to produce more power and minimize turbine maintenance costs,” Lundquist said.
Before this study, turbulence dissipation rates had been directly measured only in wind tunnels. Estimates of dissipation rates from scanning lidar have also been carried out at this location in a previous study in which Lundquist’s research group collaborated with CU’s Cooperative Institute for Research in Environmental Sciences and the National Oceanic and Atmospheric Administration’s Earth Systems Research Laboratory.
That study was just published in the Journal of Atmospheric and Oceanic Technology.
“Although wind-tunnel measurements of wind speed can represent what happens in the real atmosphere, scaling wind-tunnel measurements of turbulence is more difficult,” she added. “We need to compare real-world measurements to wind-tunnel measurements to understand how that scaling works for turbulence.”
At the AGU session, Lundquist outlined the design and results of the Turbine Outflow Dissipation Study, conducted in fall 2012 at the National Renewable Energy Laboratory’s National Wind Technology Center south of Boulder.
The study employed an array of sensors—measuring wind speed, wind direction, temperature and turbulence—from ground level up to 430 feet. As winds approached the turbine from the west, those measurements were taken from a tall tower designed and operated by NREL.
Direct measurements downwind of the turbine came from the CIRES “tethered lifting system”—a small blimp raised and lowered by winch and that carried a host of instruments from ground level to the top height of the turbine blades.
In front and behind that turbine, measurements were taken remotely from ground-based Doppler lidars.
This study was extensively supported by NREL and CIRES, and it required approval from the Federal Aviation Administration.
“We went to so much effort to collect these measurements because there’s a very poor understanding of turbine wake dissipation rates in the real atmosphere,” said Lundquist, also a fellow in CU’s Renewable and Sustainable Energy Institute (or RASEI).
“It’s very important to understand how that increased turbulence dissipates,” she said. “Once you understand wake evolution, you can site turbines in optimal locations.”
Wake turbulence is a type of instability in wind flow and results from wind flowing through a wind turbine’s rotor.
Smooth airflow propels wind turbines most efficiently. But in the wake of an upwind turbine, downwind turbine blades can receive turbulent flow, which can apply uneven pressure to the turbines’ mechanical parts.
“That makes them less efficient, and it can also damage them over the long term; wind turbines are designed to survive a 20-year lifetime.”
In large wind farms, wind turbines in the middle receive turbulent “wake loads” from every direction and start breaking down much earlier than the turbines on the farms’ leading edge, which enjoy “smoother, less-turbulent flow,” Lundquist said.
By better understanding turbulent wake flows across a range of conditions in a wind farm, Lundquist said, planners might design a wind farm in such a way that each turbine enjoys maximal wind speed and minimal turbulence.
Weather models typically assume that within any “grid cell,” dissipation and production of turbulence are equal: “Whatever turbulence is generated in that cell is also going to be dissipated in that cell.”
In large-scale weather models, Lundquist said, that assumption is valid. In a 10-kilometer grid, “any turbulence generated will dissipate by the time it moves out of that cell.”
“But as our models increase in fidelity, the time that it takes for turbulence to cascade down into smaller scales is longer than the time it takes for a flow to move across any individual grid cell.”
And for wind farms to be optimally laid out, flow simulation models must incorporate the wake dissipation rates measured in real, atmospheric conditions, Lundquist said.
“I think that we, as a community, are making some incorrect assumptions about how we model the way that turbulence propagates and evolves. And the only way to correct those assumptions is to collect data and to determine whether or not the production and the dissipation of turbulence finds a balance in these regions. And if not, then we need to develop a more accurate way to represent the decay of turbulence.”
“I think we can probably squeeze the turbines closer together,” Lundquist said. Data from scanning-lidar observations in an Iowa wind farm she studies indicate that the wakes are disappearing “faster than we would have thought in daytime conditions.”
The research was funded by the Center for Research and Education in Wind. She collaborated with Ludovic Bariteau of CIRES and NOAA and Leosphere, a French manufacturer of Doppler lidars.
Lundquist also credited Yannick Meillier, a former CIRES researcher, with discussing the project at length with her and helping to frame the study, as well as the efforts of NREL staff Andrew Clifton and Ryan King and CU graduate students Matthew Aitken, Brian Vanderwende, Michael Rhodes, Joshua Aikens, Clara St. Martin and Ryan King for their efforts in the field deployments.