Sensitivity analysis of nacelle lidar free stream wind speed measurements to wind-induction reconstruction model and lidar range configuration

The sensitivity of nacelle lidar wind speed measurements to wind-induction models and lidar range configurations is studied using experimental data from the Nørrekær Enge (NKE) measurement campaign and simulated lidar data from Reynold-Averaged Navier Stokes (RANS) aerodynamic computational fluid dynamics (CFD) simulations.

In both approaches, the data correspond to measurements (or simulations) from a five-beam Demonstrator (5B-Demo) unit developed by Avent Lidar Technology and a ZephIR Dual-Mode (ZDM) unit developed by Zephir Lidar. The 5B-Demo was configured to measure at ten distances while the ZDM was configured to measure at five distances.

From the configured distances, a large number of range configurations were created and systematically tested to determine the sensitivity of the reconstructed wind speeds to the number of ranges, minimum range and maximum range in the range configurations. The wind speeds were reconstructed using both a onedimensional and two-dimensional induction model to test the sensitivity towards the wind-induction model.

In all cases, the sensitivity of the reconstructed wind speed was determined from the wind speed error and root mean square error (RMSE) of the fitting residuals. The results demonstrate that it is not possible to use RANS CFD simulated lidar data to determine optimal range configurations for real-time nacelle lidars due to their perfect (unrealistic) representation of the simulated flow field.

The recommended range configurations are therefore based on the NKE sensitivity analysis results. Based on these results, it is recommended to configure nacelle lidars to measure at approximately 3-5 ranges. The minimum distance should be configured to roughly 0.5 rotor diameters (Drot) while it is recommended that the maximum range lay within 1-1.5Drot.

In addition, the results show that the reconstructed wind speeds are insensitive to the wind-induction reconstruction model.