A comparison of field-measured wind speed between a RNRG#40C Cup Anemometer and a Thies Clima 3D Ultrasonic Anemometer was performed. Differences in the horizontal wind speed indicated from both sources were analyzed under the current operational condition of the measuring station, i.e. no recalibration of the former sensor, and with the latter being assumed as reference of the true local wind. The characteristically low speed north-easterly winds yielded the strongest fluctuations of the sign of the normalized difference between anemometers, ranging from -45.18% to 15.19%. The data processing involved the synchronization of both data logging systems, the removal of measurements affected by shadowing and the restriction to a significant wind speed regime, i.e. above the cut-in wind speed of most wind turbines, 3.5 m/s. Accordingly, the error between anemometers was approximated by a simple linear regression using the horizontal turbulence intensity and wind speed increments as predictors of an independent model, whose aim is to correct the error based on cup anemometer data only.
During the evaluation of 1554 data points, 28.83% of cup anemometer measurements showed overspeeding, i.e. recorded higher velocities than those from the assumed true wind. The original magnitude of the error in the wind speed indicated by the cup anemometer was fitted linearly as a function of turbulence to yield an apparent reduction from 1.23% to 0.63%. Similarly, the apparent correction achieved by fitting the error as function of wind speed increments was 0.62%.