How good is a nacelle transfer function?
The wind speed a turbine reports is a corrected reading, not a raw one. Here is what that correction does, and how well it holds up against an independent lidar measurement.
The wind speed measured by the anemometer on a turbine nacelle is affected by the rotor, the nacelle geometry and the local flow distortion around it. To account for this, the control system applies a Nacelle Transfer Function (NTF) that corrects the raw reading to better predict the free stream wind speed. This corrected value is the SCADA wind speed normally used for analysis.
What the correction does
This plot shows the NTF for an example turbine, raw nacelle wind speed on the x-axis and corrected wind speed on the y-axis. The fit is a fifth-order polynomial.
What the R² means here. A near-perfect R² only confirms that the polynomial matches the controller's transfer function. It is not a measure of how close the corrected wind speed is to the true free stream. That needs an independent reference.
Checking it against a lidar
The second plot is a simple check of how good the correction actually is. It compares the NTF corrected wind speed against a nacelle mounted lidar measuring the free stream ahead of the rotor, an independent reference the turbine never sees.
The result is a good average fit. The corrected wind speed reads slightly high at the higher wind speeds, and there is some scatter, with an RMSE of 1.14 m/s. For a measurement the turbine uses to judge its own performance, that is a meaningful spread.
A question for commissioning
Most turbines use the manufacturer's default NTF for their whole life. Given the residual bias and scatter seen here, it is worth asking whether a site-specific NTF should be derived as part of commissioning, when a reference lidar is more likely to be on site, rather than relying on a generic OEM curve.
Charts and analysis by PowerVeritas. Where open datasets are used, sources are credited on the attributions page.