Welcome to the next installment of our search for an accurate, cost effective and reproducible method to assess the energy improvement from wind turbine upgrades.
Here, we'll take a look at techniques that attempt to measure this free stream wind flow, mainly using wind sensors like ground based LiDAR or anemometers mounted to a met tower. These are typically used in the context of IEC 61400-12-1 power curve test to certify a turbine produces its claimed power in an absolute sense.
But, it is plausible that they might be able to look at relative changes due to turbine upgrades as well. These techniques also attempt to overcome the high uncertainty or insufficient accuracy of the nacelle anemometer being a point measurement. In other words, ground based LiDAR and met tower sensors can usually measure some variability of the wind with height that is shear and veer.
But there are critical downsides that preclude their use in many situations:
These sensors can be expensive and potentially temperamental, reducing the cost effectiveness and reproducibility of these tests.
They are labor intensive. The equipment may need to be moved around the plant to perform the measurements for different turbines. This is because the sensors need to be relatively close to the turbines of interest.
These tests are time-intensive. A measurement campaign must collect sufficient data in the baseline state to enable future comparison with the optimized state. This means the upgrades need to be delayed until the baseline data is collected, which may result in lost energy.
These tests are also limited in terms of wind direction. Since we cannot use the measurements if the sensors are in the wake of a turbine.
Lastly, results from these tests may still have high uncertainty or low accuracy. This can be due to noisy sensors, drifting calibrations or the inability to capture differences in, for example, turbulence levels before and after an upgrade. High uncertainty means you will only be able to assess relatively large changes in turbine performance.
The IEC 61400-12-1 standard even allows for about a 1% drift in wind speed sensor calibration over the span of a test. This is about the same order of magnitude of the improvement that we would be looking to detect from a wind turbine upgrade.
So, if we can't use nacelle power curves, and we don't want to perform expensive power curve tests, where can we go next?