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.

Susanne Lynum Kongsberg Maritime AS, Norway
Co-authors:
Susanne Lynum (1) F P Per-Åge Krogstad (2)
(1) Kongsberg Maritime AS, Trondheim, Norway (2) NTNU, Trondheim, Norway

Printer friendly version: printer.gif Print

Presenter's biography

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

Susanne Lynum is currently working at Kongsberg Maritime AS in the Wind Farm Management department. She finished her Master's degree, with specialisation within fluid mechanics and wind energy at the Norwegian University of Science and Technology, NTNU, spring 2013. During her final year of her Master’s degree she has studied the performance and wake structure of a model wind turbine through wind tunnel experiments at NTNU.

Abstract

"Wind turbine wake meandering an experimental study of tip vortex motion"

Introduction

In this research the wake meandering phenomenon is studied based on the wake of a three bladed horizontal axis model wind turbine using a hot-wire array with four hot-wire probes. Previously there have been conducted several experiments on the wake meandering phenomenon, however most of them are based on single point measurements. When measuring turbulence in a fixed point in space, the meandering of the wake will lead to a smearing out of the steep gradients found in the wake. This will result in a measured turbulence level that appears lower that what is really the case.

Approach

The fact that a wind turbine blade sheds tip vortices, a total of three for a three bladed horizontal axis turbine, which follow the wake of the wind turbine downstream, makes it possible to state if the wake meanders based on the location of the tip vortices. By using an array of four hot-wire probes, simultaneous measurements were conducted at multiple nearby points in the wake, and thus the location of the tip vortices could be obtained based on the peak level of normal stress in the wake.

The experiments were conducted in the wind tunnel at NTNU at the Department Energy and Process Engineering. The diameter on the model wind turbine rotor is 0.894 m and the lengths of the blades are 0.45 m. The blades of the turbine use the NREL S826 airfoil along the entire span and are machined from aluminium. The design point for the turbine blades is an angle of attack of 7.0° at a tip speed ratio 5, with corresponding lift and drag coefficients of 1.28 and 1.35E-2 respectively. [1, 2]

Studies of the wake meandering phenomenon have resulted in two main possible explanations of its formulation. One deals with the intrinsic instabilities in the wake characterized by a periodic vortex shedding appearing in the wake, and the other is based on how large scale turbulent eddies in the atmospheric boundary layer affect the wake. [3] Thus the aim of the study is to see the effect on the meandering of the wake of the model turbine when placed in an incoming flow with turbulence intensity typical for atmospheric turbulence, compared to a low turbulence intensity case. To generate the right turbulence intensity, a grid was inserted upstream of the turbine. The integral length scale in both the streamwise and the spanwise directions were found using auto- and cross correlation between two measurement points. This was done to be able to scale and compare the incoming flow to a realistic situation for a real size wind turbine.


Main body of abstract

The background turbulence intensity in the empty wind tunnel was about 0.3 %. The incoming flow generated when inserting the grid gave a turbulence intensity of around 5.5 % and integral length scales of Luuz/R = 6.89E-2 and Luux/R = 1.44E-1 at the position of the model wind turbine.

The torque and thrust acting on the model turbine were logged, and the free stream velocity was set to 10 m/s when the experiments were performed. Based on these values the calculated Cp,max was 0.455 at λ = 5.42 without grid generated turbulence, and 0.446 at λ = 5.72 with grid turbulence. The greatest deviation in the performance curve was found at the top of the curve; however the difference between the two cases was minor. The thrust coefficient at the tip speed ratio giving the highest power coefficient without grid turbulence was calculated to be 0.86 and dropped to 0.66 in the case with grid turbulence.

To decide where to conduct the multiple hot-wire measurements, initial measurements with a single hot-wire probe were conducted at X/D = 1 downstream in the wake of the turbine to locate the area of tip vortices. The tip speed ratio of the model turbine was set to round 6 during the measurements in the wake.

Next, the position of the three tip vortices were located at X/D = 1 downstream of the turbine using the hot-wire array. The turbine rotor was equipped with a position sensor system which allowed conditioning of the recorded hot wire signals with respect to the rotor angle. Using this system conditional averaging revealed three distinct peaks in the normal stress within each revolution.

After the position of the tip vortices had been located, new measurements were performed with the hot-wire array at X/D = 1, 3 and 5 downstream of the turbine over a wide range in the radial direction. The signals from the hot wire probes were conditioned into bins representing 2 degree increments of rotor rotation which gave 4-6 samples in each bin for each rotation. This gave a phase-averaging of the measurements and the possibility to see the “mean” flow field in the wake, see Figure 1 to 6.

Using a criterion that the centre of a tip vortex is located where the normal stress is higher than the mean value for each accumulation of normal stress, a study on the size and position of the tip vortices was conducted.

The only clear indication of three distinct tip vortices were in the case with a low turbulence incoming flow at X/D = 1, as seen en Figure 1. However, there were indications of merged tip vortices at X/D = 1 in the high turbulence intensity case and at X/D = 3 in the case with low turbulence intensity, represented in Figure 2 and 3 respectively. At X/D = 5 the individual vortices seem to be fully dissolved in both cases, see Figure 5 and 6.

The time series from the experiments were also analyzed with respect to the power spectra and by making a histogram of when and where the maximum fluctuating velocity, believed to represent the center of the tip vortices, occurred. This gave the displacement of the vortices in the streamwise direction.

In Figure 7 the power spectra from position z/R= 1.20 at X/D = 1 is given. This clearly shows the 3P frequency representing the three tip vortices in the case with low turbulence intensity. There is also a clear peak at the 2P frequency, indicating a pulsation motion in the wake of the model turbine.


Conclusion

Well-defined tip vortices were only found at X/D = 1 downstream of the model wind turbine, when the incoming flow had low turbulence intensity, based on the increased level of normal stress in the wake, see Figure 1. The presence of the tip vortices were also confirmed by strong peaks in the power spectra at the 3P frequency, as seen in Figure 7.

At the other measurement locations downstream of the turbine, phase averaging showed only broad fields with increased normal stress, indicating that the tip vortices had either merged together or were broken up, see Figure 2-6. This was also confirmed by the power spectra which lacked the 3P frequency. In the case with an incoming flow with turbulence intensity typical for atmospheric turbulence, the tip vortices seemed to already have merged together at X/D = 1. Thus, the increased turbulence intensity in the incoming flow caused the tip vortices to break up at a much earlier stage than when the background turbulence intensity is low.

Even though the turbine blades are positioned symmetrically at 120 degrees intervals, the tip vortices found at X/D = 1 were not equally distributed within the wake and were located and 30°, 128° and 224°, with centers located at the radial positions z/R =1.12, 1.15 and 1.20 respectively (Figure 1). This may have been caused by different loading and/or different pitch angles of the turbine blades. The vortex diameters, scaled by the radius, were found to be 4.00E-2, 3.11E-2, and 6.00E-2 in radial direction. The location of the peak in the normal stress tended to meander a bit back and forth radially, with a distance ranging from 1.00E-2 to 4.00E-2. At the same time the vortex centers were found to shift in streamwise direction over a distance of 1.38E-1. The tip vortices seem to meander individually within the wake, and not with the same distance, and where not of the same strength when considering the level of normal stress. A distinct 2P frequency in the power spectra, see Figure 7, indicated a pulsating motion between the tip vortices within the rotations.



Learning objectives
Due to the significant change in the wake structure at X/D = 1 when the model wind turbine was placed in an incoming flow with a turbulence intensity typical for atmospheric turbulence, new measurements should be conducted closer to the rotor of the turbine in order to learn more about the initial merging and break up of the vortices.

A study on the pulsation motion, or the 2P frequency found, should also be studied further.



References
[1] Sæta, E., The Blade Element Momentum Method, 2008: NTNU.
[2] Krogstad, P.-Å. and P.E. Eriksen, “Blind test” calculations of the performance and wake development for a model wind turbine. Renewable Energy, 2013. 50(0): p. 325-333.
[3] España, G., et al., Wind tunnel study of the wake meandering downstream of a modelled wind turbine as an effect of large scale turbulent eddies. Journal of Wind Engineering and Industrial Aerodynamics, 2012. 101(0): p. 24-33.