In 2D airfoil polar measurements in wind tunnels, airfoil lift coefficient variations with angle of attack are generally obtained through the integration of surface pressure distributions. For this purpose, first surface pressures are recorded with a selected sampling frequency for each angle of attack setting and for a given measurement duration. Then for each pressure tap, data are averaged to get average pressure coefficient distribution around the airfoil, which is in turn integrated to find the average lift coefficient value at that angle of attack. This standard approach, however, does not give any information regarding the instantaneous lift fluctuations such as the standard deviation of lift, which can be of significance especially at very high angle of attack tests. In this study, we performed polar measurements for DU00-W-212 airfoil in the new large-scale wind tunnel of the Center for Wind Energy Research (RÜZGEM) at the Middle East Technical University (METU). The tests are performed in the 2.5 m x 2.5 m cross-section test section of the METU-RÜZGEM wind tunnel using a 0.9 m chord and 2.5 m span model in an angle of attack range of -20 to 30 degrees (Figure 1). Surface pressures are measured at each angle of attack using 500 Hz sampling frequency and for 30 seconds. By integrating instantaneous surface pressure distributions, we created instantaneous lift coefficient data and its time history during the measurement duration of 30 s. Using the time history, we calculated average Cl value and its standard deviation for a given angle of attack. Figure 2 shows Cl-alpha variation for the DU00-W-212 airfoil at Re=3E6, comparing the average Cl value obtained from average Cp data and the one obtained from the time history of the Cl for that angle of attack. The error bars show the standard deviation of the lift fluctuation. The results show that in the linear range the standard deviation levels are very low and they get higher in the positive/negative post-stall regions. We also observe a mismatch between the two average values, especially in the negative stall region. In our final presentation we will show analysis results attempting to address these differences and show data in a wider angle of attack range.

A boundary layer suction (BLS) device removes the low-energy flow within the boundary layer. This type of flow control device could remarkably increase the aerodynamic performance of the airfoils by increasing the lift and decreasing the drag force on the airfoils. In addition, it may delay or eliminate the laminar-turbulent transition and flow separation on aerodynamic surfaces. BLS flow control has been studied for relatively thin wind turbine airfoils [1-5]. Nonetheless, a large and finely-resolved design space of BLS parameters on wind turbines with a large rotor diameter has not been explored yet. Moreover, the literature lacks how the variation of the values of BLS parameters affects the flow around very thick wind turbine airfoils at operational Reynolds numbers.

In this work, multiple BLS parameters are optimised on FFA-W3-241 and FFA-W3-480GF airfoil profiles of the DTU 10-MW RWT [6] by a design of experiment approach to enhance their aerodynamic performance. The optimisation framework is illustrated in Figure 1. A design of experiment method, Latin Hypercube Method, determines the design cases with different BLS parameter values, which will be simulated with CFD. SU2 flow solver is used to solving URANS equations along with the k-ω SST turbulence model. Polynomial regression is implemented to construct a design space from the simulation outputs. Thus, a relationship between the design and airfoil performance parameters could be founded. The optimisation framework is run for different angles of attack and Reynolds numbers to analyse the effects of BLS parameters on the aerodynamic characteristics of the airfoils, such as the flow separation location and aerodynamic force coefficients, at both on-design and off-design turbine operating conditions. Then, airfoil polars are generated with optimised BLS parameters for the maximum lift coefficient and maximum lift-to-drag ratio, separately for different Reynolds numbers. These polars will be fed into a Blade Element Momentum (BEM) method code to investigate the change in turbine power generation at different operational conditions, which will be presented in another presentation [7].

The computational setup has been validated against experimental data in the literature. Preliminary results showed that such a flow control could increase the lift-to-drag ratio of a thin wind turbine airfoil [8] and remarkably delay or completely eliminate flow separation on a thick wind turbine airfoil [9]. Figure 2 shows how BLS affects the flow around the 48%-thick FFA-W3-480GF airfoil and eliminates the flow separation on the suction surface.

Airfoil design is typically focused on performance characteristics of geometries with a non-constant thickness from leading to trailing edge. Past studies, however, have indicated that curved plate airfoils can produce a greater L/D than these conventional airfoils at low Reynolds numbers [1,2]. A performance comparison of traditional, flat plate, and curved plate airfoils is shown in Figure 1. Previous studies at Clarkson University have shown that the constant thickness GOE417a airfoil, developed by Schmitz [1], can be effective for use on small wind turbine blades [3-5]. The simpler means of construction can also reduce manufacturing costs substantially [4].

An investigation was conducted to improve the performance of the GOE417a airfoil throughout the Re regime of the rotor. Airfoils were developed using a geometric computational strategy based on parabolic and circular arcs as well as a tangent arc geometry derived from the GOE417a itself. Preliminary analysis in XFOIL indicated that the parabolic and circular arc airfoils underperformed, when compared to the baseline GOE417a. Airfoils designed with the tangent arcs were observed to have comparable performance to the baseline throughout the Re range. Figure 2 illustrates the preliminary XFOIL results for the most optimal airfoil. The improved design outperformed the GOE417a by 18% at a Re = 400,000. Experimental data for the GOE417a, however, illustrates a discrepancy from the XFOIL results as noted in Figure 2 [6]. Despite this, the new geometry indicated an overall potential to outperform the GOE417a.

To validate these results, a Navier Stokes analysis using ANSYS FLUENT was initiated and is currently underway. Figure 3 depicts an initial velocity field at 6° angle of attack. Both the angle of attack and Re will be varied to determine the maximum L/D performance. Results from the new airfoils will be compared to the GOE417a and the most effective airfoil will be used to construct an improved blade geometry for the Clarkson ducted turbine, the goal being a reduction in cost per unit energy, $/kWh, of the design.

As the power generation capacity and the size of the wind turbines are increasing, research on loads and aerodynamics at high Reynolds numbers gain importance. Modern wind turbines operating under high Reynolds numbers are exposed to higher loads, especially the root sections of the blades.
These sections transmit loads to the hub, and due to the square-cube law, higher stiffness and thereby thicker airfoils are needed as the blades increase in size. Prediction and validation of the aerodynamic performance of thick airfoils are inherently challenging because separation often appears somewhere on either the suction side or the pressure side. In addition to this, the Coriolis forces and centrifugal forces will affect the flow in the separated zones and thereby influence the aerodynamic performance.
The aim of this study is to show an insight into thick airfoil aerodynamics by wind tunnel experiments and to introduce the experimental methodology and processing tools developed for the measurements. This is to obtain a better understanding of the 2D aerodynamics of thick airfoils. For this purpose, the infrared thermography (IR) technique is used, where an infrared camera is placed in a wind tunnel and used for measurements on a thick airfoil at high flow speeds. In this way, boundary layer transition from laminar to turbulent flow and separation can be determined by the infrared camera for predicting the aerodynamic performance in comparison with pressure measurements. In addition to pressure and IR measurements, tufts are also used for comparison and highlighting the flow behavior around the model. The model investigated in this study is the DU00-W2-350 airfoil with a chord of 1 meter and a span of 2 meters. The tests are carried out in the Poul la Cour tunnel (PLCT) of DTU with a cross-sectional area of 2 x 3 meters. The airfoil contains 97 pressure taps and is tested for Reynolds numbers up to 7 million. In order to process the images obtained from the infrared camera, image distortion correction and calibration tools are developed. Furthermore, transition and separation detection tools are also developed with the purpose of establishing a real-time visualization of the flow and the aerodynamic properties. In this way, the challenges related to the flow behavior and aerodynamic performance of a thick airfoil are highlighted.