From wind to validation: Measurements on the AD8 turbine in Bremerhaven

As the dimensions of wind turbines continue to increase in size, so do the mechanical loads acting on their components. The incoming wind field also covers an increasingly larger area. As such, models play an important role in the development of new wind turbines, serving as the basis for new turbine designs. These models were developed on smaller turbines and are still in use today. But how valid are such models for the ever-larger turbines? Test bench measurements are also an important aspect of plant development. How well do test bench measurements reflect reality? We investigated these and other questions in the “Testfeld BHV” project.

Extensive measurements on the AD8-180 in Bremerhaven

From 2018 to 2022, we performed extensive measurements on the AD8-180 research turbine in Bremerhaven. When it was built back in 2017, the AD8-180 was the largest wind turbine in the whole world. The results help us to continually make our test bench measurements more precise and validate the simulation models with the help of the generated data. This gives us a good data basis for better understanding of the wind turbines and their development, which we can utilize to improve turbine design, especially of large turbines. In the scope of the publicly funded Testfeld BHV project and other associated projects, we installed extensive measuring systems on the AD8-180:

  • load measurements in acc. with IEC 61400-13;
  • power performance measurements in acc. with IEC 61400-12-1 and further power performance measurements;
  • additionally, the measurement of the bending line at five radii on all three rotor blades;
  • acceleration measurements in one rotor blade, including for the analysis of the torsion of the rotor blade;
  • measurement of the aerodynamic flow around the rotor blade;
  • acceleration measurements on the drive train; and
  • acoustic measurements.

In parallel, we performed extensive wind measurements – upstream of, on, and downstream of the turbine – with a total of more than 10 wind lidar (light detection and ranging) devices and an IEC-compliant wind measurement mast up to hub height. In addition, SCADA data and the manufacturer’s turbine model from the design and certification process were also made available to us. In cooperation with additional Fraunhofer institutes and other partners, we were able to employ and test innovative sensors (e.g., a vibrometer together with Fraunhofer IOSB).

Test field measurement system © Fraunhofer IWES, Photo Hauke Müller

Comparison of results: in the field and on the test bench

One of the project’s focuses was the direct comparison of field and test bench measurements at the Dynamic Nacelle Testing Laboratory (DyNaLab). We investigated whether the vibration behavior of a wind turbine drive train can be replicated on a nacelle test bench and thus also analyzed. The focus here was on the strongest excitations induced by the meshing of the two gear stages. The nacelle of the AD8-180 had already been tested on the nacelle test bench in advance. For a direct comparison, we performed the mechanical measurement of the drive train on the test bench with 27 3D acceleration sensors in exactly the same positions as in the field test. The initial results have already been published [1]. One challenge when comparing the measurements was the differences arising due to the different framework conditions: whereas we were able to recreate individual situations specifically on the test bench, the external conditions in the field test were complex and could only be recorded and replicated to some extent. The load on the drive train depends on the incoming wind field, which cannot be recorded in minute detail. Our results show that, despite these challenges, the vibrations measured in the field test resulting from the gear meshing, one of the main sources of excitation, can be reproduced well on the test bench. In addition, vibrations induced by characteristics of the test bench drive are also easily identifiable. The natural frequencies excited on the test bench and in the field are comparable to some extent in terms of frequency and mode shape. In particular, the modes assigned to the bending movements for the drive train match in the field and on the test bench. In this way, we were able to demonstrate that the results of a vibration analysis of a drive train performed on a nacelle test bench are also of value for the system in the field. Consequently, this task can be performed cost-effectively and efficiently for turbine manufacturers in the scope of a nacelle testing campaign.

Aerodynamic measurements and simulation

A further measurement presented here as an example was the aerodynamic measurement on an AD8 rotor blade. In the scope of the HighRe project, we installed a fixture equipped with pressure sensors at each of two radii on the outside of a rotor blade. In addition, we installed a five-hole probe at both radii, which measured the inflow at the rotor blade at a distance of approximately 1.5 m in front of the leading edge, thus making it possible to eliminate the local dynamic pressure [2, 3]. With the help of these data, it was then possible to measure the aerodynamic pressures depending on the inflow in certain blade segments with high temporal resolution. On this basis, it is then possible to validate aerodynamic and aeroelastic models and simulations of all kinds. [4]

Validation of the simulation models

The quality of simulation models is evaluated by comparing simulation results with measured loads. For this, it is important that the model and the real system be subject to the same environmental conditions. For every aero-servo-elastic simulation with the MoWiT tool, the model of which is based on the manufacturer’s system model, we use a wind field file as input. This wind field file must represent the wind conditions in the field at the time the actual measurement was taken as accurately as possible. This is a major challenge, but the detailed wind measurements are very helpful.

Wind measurement overview © Fraunhofer IWES

With regard to the global turbine behavior, we were able to demonstrate that the simulated loads at the blade root matched the measured loads well. [5] Although the performance still displayed deviations, we were able to illustrate that time series-related wind fields make it possible to reproduce specific 10‑minute simulations [5]. On the basis of a pitch maneuver, we showed that the measured frequencies correspond well with simulated torsional frequencies. Regarding the torsional moment during power production, the identification of the natural frequencies of the torsional modes was far less clear. This could indicate that the model should be calibrated using measured data and that the linearity assumption on which the modal analysis is based could be violated for these long rotor blades. [6] We are further expanding the methodology for the comparison of many simulations with many measurements and the resulting evaluation of the model quality. [7]

Data provide an important basis for future projects

The first publications on the analysis of the torsional deformation [6], aerodynamic measurement, validation of the MoWiT load model, and various aspects of the extensive wind field measurements [8] [9] [10] [11] have already been published. Our colleagues presented further publications this year at TORQUE 2024 in Florence and at WindEurope in Dublin. Further presentations and publications are currently in preparation.

In the scope of the Testfeld and HighRe projects, an extensive data collection of measurement and model data from wind turbines has been created which represents a unique data basis for science. IWES is already utilizing these data in ongoing projects, for example for the development of AI methods. Further projects in collaboration with industry and research partners are planned for the future.

List of references

[1] K. Eustorgi, S. Mechler and A. Wegner, “Comparison of wind turbine drive train loads measured in the field and on a nacelle test bench,” in Conference for Wind Power Drives, Aachen, Germany, 2021.
[2] M. Fredebohm, J. N. Theron, L. Höning, S. Weis, B. Stoevesandt, A. Wegner, N. Denecke, H. Rosemann and M. L. Huhn, “Development of an aerodynamic measurement system for wind turbines,” Journal of Physics: Conference Series, vol. 2507, p. 012020, 2023.
[3] A. Wegner, S. Mechler, L. Höning, N. Denecke and B. Stoevesandt, “Aerodynamic conditions measured at a rotor blade of large wind turbine prototype,” Journal of Physics: Conference Series, no. 022024, p. 2767, 2024.
[4] L. Höning, P. Meyer, M. L. Huhn, J. N. Theron, P. Thomas, A. Wegner, S. Mechler, J. Gottschall and B. Stoevesandt, “Validating low- and high-fidelity simulations of a yawed 8 MW wind turbine against measurements,” Journal of Physics: Conference Series, vol. 022038, no. 2767, 2024.
[5] M. Huhn and A. F. Gómez-Mejía, “Aeroelastic model validation with 8 MW field measurements: Influence of constrained turbulence with focus on power performance,” Journal of Physics: Conference Series, vol. 2265, p. 032058, 2022.
[6] A. Wegner, M. L. Huhn, S. Mechler and P. Thomas, “Identification of torsional frequencies of a large rotor blade based on measurement and simulation data,” Journal of Physics: Conference Series, vol. 2265, p. 032021, 2022.
[7] P. J. Meyer, M. L. Huhn and J. Gottschall, “Development of a Load Model Validation Framework Applied to Synthetic Turbulent Wind Field Evaluation,” Energies, vol. 17, no. 4, p. 797, 2024.
[8] P. Meyer and J. Gottschall, “Evaluation of the “fan scan” based on three combined nacelle lidars for advanced wind field characterisation,” Journal of Physics: Conference Series, vol. 2265, no. 022107, 2022.
[9] L.-Y. Hung, P. Santos and J. Gottschall, “A comprehensive procedure to process scanning lidar data for engineering wake model validation,” Journal of Physics: Conference Series, vol. 2265, no. 022091, 2022.[10] A. Giyanani, M. Sjöholm, G. R. Thorsen, J. Schuhmacher and J. Gottschall, “Wind speed reconstruction from three synchronized short-range WindScanner lidars in a large wind turbine inflow field campaign and the associated uncertainties,” Journal of Physics: Conference Series, vol. 022032, no. 2265, 2022.
[11] P. J. Meyer, A. Giyanani and J. Gottschall, “Constrained synthetic wind fields from high-resolution 3D WindScanner measurements,” Journal of Physics: Conference Series, vol. 042036, no. 2767, 2024.

More information:

One Year of the New European Wind Atlas
Bundled wind energy expertise for virtual turbine models
The potential of floaters for wind energy (fraunhofer.de)


Leave a Reply

Your email address will not be published. Required fields are marked *