This paper gives an overview of the design and wind tunnel test results of the wind turbine dedicated airfoils developed by Delft University of Technology (DUT). The DU-airfoils range in maximum relative thickness from 15% to 40% chord. The first designs were made with the XFOIL code. The computer program RFOIL, which is a modified version of XFOIL featuring an improved prediction around the maximum lift coefficient and the capability of predicting the effect of rotation on airfoil characteristics, has been used to design the airfoils since 1995. The measured effect of Gurney flaps, trailing edge wedges, vortex generators (vg) and trip wires on the airfoil characteristics of various DU-airfoils is presented. Furthermore, a relation between the thickness of the airfoil leading edge and the angle-of-attack for leading edge separation is given.

No abstract

The two-dimensional characteristics of airfoil NACA 0018 have been measured for Reynolds numbers between 0.15x10 6 and 1.0x10 6 to establish the lift, drag and moment curves that serve as input to performance calculations of vertical axis wind turbines. At the lower surface laminar separation occurs at low to medium angles of attack, which is of significant influence on the characteristics and the radiated noise. For the situation with a lower surface laminar separation bubble, span wise wake rake traverse measurements showed an irregular three-dimensional pattern. Noise reduction could be achieved with zigzag tape at the 70% to 80% lower surface chord station. Significant post-stall hysteresis loops occurred showing a high loss in lift.

In modern wind turbine blades, airfoils of more than 25% thickness can be found at mid-span and inboard locations. At mid-span, aerodynamic requirements dominate, demanding a high lift-to-drag ratio, moderate to high lift and low roughness sensitivity. Towards the root, structural requirements become more important. In this paper, the performance for the airfoil series DU FFA, S8xx, AH, Risø and NACA are reviewed. For the 25% and 30% thick airfoils, the best performing airfoils can be recognized by a restricted upper-surface thickness and an S-shaped lower surface for aft-loading. Differences in performance of the DU 91-W2-250 (25%), S814 (24%) and Risø-A1-24 (24%) airfoils are small. For a 30% thickness, the DU 97-W-300 meets the requirements best. Reduction of roughness sensitivity can be achieved both by proper design and by application of vortex generators on the upper surface of the airfoil. Maximum lift and lift-to-drag ratio are, in general, enhanced for the rough configuration when vortex generators are used. At inboard locations, 2-D wind tunnel tests do not represent the performance characteristics well because the influence of rotation is not included. The RFOIL code is believed to be capable of approximating the rotational effect. Results from this code indicate that rotational effects dramatically reduce roughness sensitivity effects at inboard locations. In particular, the change in lift characteristics in the case of leading edge roughness for the 35% and 40% thick DU airfoils, DU 00-W-350 and DU 00-W-401, respectively, is remarkable. As a result of the strong reduction of roughness sensitivity, the design for inboard airfoils can primarily focus on high lift and structural demands.

This paper gives an overview of the design and wind tunnel test results of the wind turbine dedicated airfoils developed by Delft University of Technology (DUT). The DU-airfoils range in maximum relative thickness from 15% to 40% chord. The first designs were made with XFOIL. Since 1995 RFOIL was used, a modified version of XFOIL, featuring an improved prediction around the maximum lift coefficient and capabilities of predicting the effect of rotation on airfoil characteristics. The measured effect of Gurney flaps, trailing edge wedges, vortex generators and trip wires on the airfoil characteristics of various DU-airfoils is presented. Furthermore, a relation between the thickness of the airfoil leading edge and the angle-of-attack for leading edge separation is given.