Back to the programme printer.gif Print




Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'Aerodynamics and rotor design' taking place on Wednesday, 12 March 2014 at 09:00-10:30. The meet-the-authors will take place in the poster area.

Jan Winstroth Leibniz Universität Hannover, Germany
Co-authors:
Jan Winstroth (1) F P Jörg Seume (1)
(1) Leibniz Universität Hannover, Hannover, Germany

Printer friendly version: printer.gif Print

Presenter's biography

Biographies are supplied directly by presenters at EWEA 2014 and are published here unedited

Mr. Winstroth has been doing research in the wind industry for 2 years. He is currently a research associate at the Institute of Turbomachinery and Fluid-Dynamics at the Leibniz Universität Hannover. He studied mechanical engineering at the Universität Hannover. His research is focused on optical measurement techniques applied to full scale wind turbines and structural dynamics.

Abstract

Wind turbine rotor blade monitoring using digital image correlation: 3D simulation of the experimental setup

Introduction

Optical full-field measurement techniques such as Digital Image Correlation (DIC) provide a new scope for measuring deformations and vibrations with high spatial and temporal resolution. However, application to full scale wind turbines is not trivial [1-4]. In order to assess DIC performance with regard to wind turbines, a series of experiments is conducted and evaluated inside a 3D modeling environment. The purpose of these experiments is to demonstrate feasibility, assess accuracy, gain experience, and to identify possible perturbations before applying the proposed technique on a full scale wind turbine. The proposed setup is shown in Fig. 1.

Approach


Fig. 1: Proposed setup for application on a full scale wind turbine.

The field experiments will be conducted on 3.2 MW wind turbine with a rotor diameter of 114 m and a hub height of 93 m. All data presented in this paper is based on a 3D model of this wind turbine which is created and animated inside 3ds Max. Airfoil geometries were taken from the NREL 5MW reference wind turbine since no real airfoil geometries were available [5]. However, rotor diameter, chord length and twist angle of the blades are based on the genuine geometry of the wind turbine. DIC is performed with the commercial software Vic-3D by Correlated Solutions.
Four textures with a random black and white dot pattern are applied on each blade of the wind turbine. A detailed view of a rotor blade is presented in Fig. 2. The four textured areas on the blade represent the areas of investigation. Each textured area provides 500+ measuring points evenly scattered across the textured region. The quantity to be measured for each of these measuring points is the absolute 3D position (x-, y-, z-coordinates) and the relative displacement (u- , v- , w-coordinates) between a reference image and the current image in an arbitrary fixed coordinate system. Therefore, the surface and deformation of the blade can be reconstructed for every frame of stereoscopic images. The entire rotor is continuously sampled with 30 Hz during operation which corresponds to approximately 150 stereoscopic images for one revolution.


Fig. 2: Detailed view of the textured areas on the rotor blade.

For this paper different deformations/vibrations of the rotor (1. flapwise, 1. torsion and a combination of flapwise bending and torsion) are simulated inside the 3D environment. Each animation is recorded/rendered by two simulated cameras which are placed in front of the turbine (cf. Fig. 1). These images are later used as input data for DIC. Since out-of-plane deflection shapes and torsion angles of the blade can be precisely set for each simulation, a direct comparison between simulation and analysis with DIC is possible. Therefore, accuracy and feasibility assessment of DIC when applied to wind turbines becomes possible.


Main body of abstract

Fig. 3 shows a set of stereoscopic images rendered from 3ds Max. On the left-hand side of the figure the view from the left camera is shown and on the right-hand side the view from the right camera is shown. The corresponding 3D model calculated by DIC is presented in Fig. 4. The contours qualitatively indicate the surface position in z-direction. For all displayed 3D models the orientation of the coordinate system remains unchanged. The z-axis corresponds to the rotational axis of the rotor and the xy-plane lies within the rotor plane. A close up view of one of the blades from a steeper viewing angle is presented in Fig 5. From this angle it is possible to see that the spatial resolution of the measuring system is sufficient to partially reconstruct the airfoil shapes of the blade.


Fig. 3: Field of view of left and right camera.


Fig. 4: Reconstructed 3D model from Vic-3D.


Fig. 5: Reconstructed airfoil shapes.

The out-of-plane motion of one point close to the tip of the rotor is compared against the input data from the simulation in order to perform a quantitative evaluation of the DIC algorithm. For this experiment the rotor is turning at 12 rpm and one of the blades is performing a sinusoidal oscillation out of the rotor plane with a frequency of 0.66 Hz. This oscillation is consistent with the first flapwise mode of the blade. For this experiment one complete revolution of the rotor is simulated inside 3ds max and a total of 150 sets of stereoscopic images are rendered from the left and right camera.
Fig. 6 presents the results from the out-of-plane deflection experiment. A very good agreement between simulation and DIC reconstruction can be observed. For a better understanding of the inaccuracy associated with the DIC reconstruction the absolute error for each analyzed image is plotted in Fig. 7. The average absolute error of all 150 measurements is 0.784 mm and the maximum absolute error is 2.8 mm.
Another virtual experiment involves the torsion of the blade. For this experiment the rotor is again turning at 12 rpm and one of the blades is undergoing continuous sinusoidal torsional oscillations with a frequency of 0.66 Hz. The maximum torsion between blade root and blade tip is 1 degree. The actual torsion mode of the blade is at a higher frequency. However, a lower frequency was chosen for this experiment in order to sample one oscillation with more data points and thus be able to identify the resolution of the measuring system with respect to blade torsion. Comparison of simulation and DIC is done at r = 35 m radial position. This location is located within the third textured region when counted from the blade root. The maximum torsion angle is 0.64 degrees at this location.


Fig. 6: Results from the out-of-plane deflection experiment.


Fig. 7: Absolute error between simulation and DIC reconstruction for out-of-plane deflection experiment.

Fig. 8 presents the results from the torsion experiment. Both curves resemble the same characteristic and even the absolute error is still within reasonable limits (cf. Fig. 9). The average absolute error is 0.0466 degree for all 150 measurements and the maximum absolute error is 0.2332 degree. Even though the agreement between simulation and DIC is not as good as for the out-of-plane experiment, the authors still believe that these results are quiet impressive especially when taking into account the larger scales involved.
The final paper will also present results of a combined blade deformation (out-of-plane deflection and torsion). Furthermore, the influence of image noise and image distortion will be analyzed. Another interesting section of the final paper will deal with the yaw motion of the turbine. Here the focus will be on how much change in yaw angle is tolerable until the view angle of one of the cameras becomes too steep for DIC analysis.


Fig. 8: Results from the torsion experiment.


Fig. 9: Absolute error between simulation and DIC reconstruction for torsion experiment.


Conclusion

Full scale wind turbine testing is still a very costly and challenging task. However, data from full scale tests is of utmost importance for validation of numeric design codes and further improvement of today's multi-megawatt wind turbines. With a properly validate design code, safety factors can be decreased and costs lowered. In addition proper full scale testing setups can provide vital information for structural heath monitoring.
The potential of optical measurement techniques such as Digital Image Correlation for the purpose of full scale testing is analyzed in this work. Results indicate that vibration/deformation measurements can be recorded with a high spatial and temporal resolution. In fact the spatial resolution is high enough to reproduce the airfoil. Out-of-plane deflection of the rotor can be detected with an accuracy in the range of millimeters, enabling this technique to measure torsion angles of the blade well below 1 degree. The temporal resolution of the system is currently limited to 30 Hz which is assumed to be fast enough to capture most of the important vibrational modes.
At the moment, these finding are based only upon the results from computer simulations and a preliminary assessment on a scaled model test bench with a rotor diameter of 2 m, see [6-7]. Nevertheless, the simulations take into account many of the difficulties associated with a full scale test on a wind turbine. The described optical system is already in operation at the Institute of Turbomachinery and Fluid-Dynamics and the first full scale test is in preparation. Thank you for considering my work for presentation at EWEA 2014.



Learning objectives
New optical measurement technique for full scale monitoring of wind turbines.
New capabilities for design code validation.
Possible application for structural heath monitoring of wind turbines.



References
[1] Paulsen, U. S., Erne, O., Moeller, T., Sanow, G., and Schmidt, T., "Wind Turbine Operational and Emergency Stop Measurements Using Point Tracking Videogrammetry", In Proceedings of the 2009 SEM Annual Conference & Exposition on Experimental & Applied Mechanics, 2009.
[2] Ozbek, M., Rixen, D. J., Erne, O., and Sanow, G., "Feasibility of Monitoring Large Wind Turbines using Photogrammetry", Energy, Vol. 35, No. 12, 2010, pp. 4802 - 4811.
[3] Ozbek, M. and Rixen, D. J., "Operational Modal Analysis of a 2.5 MW Wind Turbine using Optical Measurement Techniques and Strain Gauges", Wind Energy, Vol. 16, No. 3, 2013, pp. 367 - 381.
[4] Niezrecki, C., Avitabile, P., Warren, C., Pingle, P., and Helfrick, M., "A Review of Digital Image Correlation Applied to Structura Dynamics", AIP Conference Proceedings, Vol. 1253, 2010, pp. 219 - 232.
[5] Jonkman, J., Butterfield, S., Musial, W., and Scott, G., "Definition of a 5-MW reference wind turbine for onshore system development", Tech. Rep. NREL-TP-500-38060, National Renewable Energy Laboratory, 2009.
[6] Winstroth, J. and Seume, J. R., "New Test Bench for Optical Measurments of Rotor Blade Deformations on Wind Turbines", Proceedings of 8th PhD Seminar on Wind Energy in Europe, ETH Zurich, September 2012.
[7] Winstroth, J. and Seume, J. R., "Wind Turbine Rotor Blade Monitoring using Digital Image Correlation: Preliminary Assessment on a Scaled Model" In Proceedings of the 32nd ASME Wind Energy Symposium, 2014. <--- (Not published yet)