DU00-W-212 wind turbine airfoil has been previously tested to determine the Reynolds number dependency characteristics of its aerodynamic polar data in the context of the AVATAR project [1]. More recently, it has also been tested in the DTU-PLCT wind tunnel as a part of a benchmark airfoil test campaign [2]. In this study, wind tunnel measurements of the same airfoil are performed 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). The dataset includes surface pressure distributions, load balance, wake rake and infrared thermography data in a Reynolds number range up to 5E6. A comparison of the measured polar data shows good agreement with previous measurements obtained in the AVATAR project (Figure 2). In this study, we specifically focus on the transition characteristics, as measured by use of the infrared thermography technique. Figure 3 shows the variation of measured transition locations with the angle of attack at a Reynolds number of 3E6. The results show reasonable agreement with XFOIL simulations (with N=11) and with previous CFD predictions [3]. Our final presentation will consist of more details including the Reynolds number dependency characteristics of the transition locations of the DU00-W-212 airfoil.

Although the dependence of lift on the laminar-turbulent transition for an airfoil at high Reynolds numbers could be small [1], lift-to-drag ratio of wind turbine airfoils is highly affected by the transition location on the surface [2] because the width of the drag bucket is directly related to the extent of the laminar flow [3]. The development of accurate transition models for the flow solvers could be quite challenging [4] and may require calibrations [5, 6]. The hybrid Large Eddy Simulation / Reynolds Averaged Navier–Stokes (LES/RANS) simulations have been shown to predict the flow around airfoils better than the RANS simulations due to more accurate results at high angles of attack [7, 8]. In addition, when an LES/RANS simulation with a transition model is performed, it can accurately predict laminar separation bubbles [9] and give better insights into the unsteady transitional boundary layers for wind turbine airfoils [10, 11].
The measurement campaign on the DU 00-W-212 wind turbine airfoil through the AVATAR project has enabled researchers to validate the computational models through comparisons with the wind tunnel data [2], especially for surface pressure distributions and aerodynamic lift and drag coefficients. This airfoil has been simulated with RANS equations and multiple transition models, including e^N method [1, 2, 6], γ-model [5, 12], and γ − Reθt -model [6, 12]. The laminar-turbulent transition on the suction side of the DU 00-W-212 airfoil swiftly shifts towards the leading edge around α = 6◦ and α = 10◦ for Re = 3 × 10^6 and around α = 4◦ and α = 8◦ for Re = 9 × 10^6 which could be difficult to predict accurately by the simulations [12].
In this study, the CFD simulations are performed in detail for the DU 00-W-212 airfoil with different turbulence models and transition models. The transition location from the leading edge is predicted for different angles of attack for the DU 00-W-212 airfoil by performing CFD simulations and compared with the recent experimental results obtained at METUWIND/RUZGEM Large Scale Wind Tunnel [13] for 3 million Reynolds number. 2-D steady-state RANS simulations and 3-D DDES simulations for DU 00-W-212 airfoil are performed for selected angles of attack and for 3 to 15 million Reynolds numbers by using SU2 CFD solver.

With increasing size of wind turbine installations, the tip speed of the turbine blades will now often exceed 110m/s. Collision with sand, rain, hail and other particulate matter has a damaging effect on the leading edge of these blades (Keegan 2014, Pugh 2018).
Erosion begins with roughness of the leading edge (Sareen et al. 2014). Aerofoils are very sensitive to flow disturbance at the leading edge, as all downstream flow across the aerofoil surface is affected (Zidane et al. 2016).
With extended wear, pitting, gouges and eventual delamination of the leading edge will occur (Sareen et al. 2014). This will have further detrimental effects on blade performance and may lead to structural issues if left unchecked.

This recently initiated research work aims to use Computational Fluid Dynamics (CFD) simulation and Particle Image Velocimetry (PIV) testing in a subscale wind tunnel, to more comprehensively characterise
aerodynamics effects of leading edge erosion. CFD modelling of the NREL 5MW research turbine blade profile using OpenFOAM is ongoing, with initial results indicating a;
-5.8% reduction in lift performance for a 1mm sand grain roughness leading edge.
-7.8% reduction in lift performance for a 2mm sand grain roughness leading edge.

New PIV equipment is also being installed in a subscale wind tunnel to examine the characteristics of airflow over rough models experimentally. This will be used both for validation of the simulations made and to gain insight into the small scale flow dynamics.
The results of this simulation and wind tunnel testing will then be integrated with full turbine models, such as OpenFAST, to examine the effects of leading edge erosion on annualised energy production and structural loads.

At wind speeds between the cut-in speed and those required for rated power, loss of aerodynamic efficiency of the blades will necessarily reduce power generated. However, by using the pitch control to reduce the angle of attack to prevent stall, some of the losses could be recouped without the need for any repair works. An additional aim of the project is to determine these pitch adjustments needed for various levels of erosion, to maintain optimum energy production.

This project is funded by the Sustainable Energy Authority of Ireland and the project is also taking part in collaboration with International Energy Agency's wind task 46. We aim to disseminate findings through several publications over the project’s life, including a forthcoming state of the art review.

In 2020, wind energy generation contributed 24% of the total electricity generation in the United Kingdom significantly contributing towards reducing greenhouse gas emissions (Thomas, 2021). However, the efficiency of wind turbines can be degraded by the presence of surface roughness. For example, dust contamination on the leading edge and over the front half of the chord can have a significant effect on the aerodynamic performance (Khalfallah and Koliub, 2007; Ren and Ou, 2009).
In the current study, the influence of roughness wavelength and amplitude are investigated systematically using roughness patterns based on the sine function, consisting of the standard, absolute, and negative sine functions (see Figure 1). The roughness pattern is applied to the suction side of a DU96-W-180 aerofoil from 10% to 50% chord (c). The amplitude (A) is varied from 2.4 × 10^(−3)c to 1 × 10^(−2)c and three different wavelengths λ, 0.05c, 0.1c, and 0.2c, are considered. The performance of the clean and rough aerofoils is investigated using 2D Reynolds-averaged Navier–Stokes (RANS) simulations at Reynolds number 1.5 × 10^6 for angle of attack ranging from −5° to 10°. The simulations are performed using the CFD package STAR-CCM+ by Siemens PLM Software. The numerical setup is validated against the experimental data by Sareen et al. (2014) for the clean aerofoil.
The influence of roughness amplitude and pattern at λ = 0.1c on the lift coefficient is presented in Figure 2, which shows that Cl is reduced in all cases compared to the clean aerofoil. There is no significant dependency on the roughness pattern for the smallest amplitude. However, the standard sine function has the most significant adverse effect on Cl at medium and high amplitudes, followed by the negative and the absolute sine function. The drag coefficient (not shown) is increased due to roughness in all cases. At the minimum amplitude, Cd values are similar for all roughness patterns, whereas for the higher amplitude cases Cd values of the absolute sine function show the lowest increase in drag, while significantly higher increases are observed for the other patterns. Overall, roughness cases based on the standard sine function most strongly influence aerodynamic performance.
In the next stage, a similar 2D investigation will be conducted for a thicker aerofoil typical for the root section of a turbine blade before moving on to 3D investigations on more complex roughness patterns.