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.

Laure Steer Fraunhofer IWES, Germany
Co-authors:
Laure Steer (1) F P Mumtaz Ahmad (1) Roman Braun (1) Bernhard Stoevesandt (1) Florian Sayer (1)
(1) Fraunhofer IWES, Bremerhaven, Germany

Printer friendly version: printer.gif Print

Presenter's biography

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

2013 : M.Sc. in Wind Energy, National Technical University of Athens (Greece) and Ecole des Mines Paristech (France), through the EUREC Master in Renewable Energy program.

Ms Laure Steer has been pursuing her master thesis at the Fraunhofer IWES institute. During this period, her research work was focused on airfoil design and flatback airfoil aerodynamics.
Research areas: Rotor aerodynamics, CFD

Abstract

Influence of the trailing edge angle on the aerodynamic performance of thick flatback airfoils

Introduction

In most of current rotor blade designs, thick flatback airfoils are used in the inboard part contributing to the aerodynamic and the structural performance of the blade. So far, the flatback edge is usually perpendicular to the chord line of the airfoil. However, changing the trailing edge’s direction is a potential means to influence the turbulent wake behind the airfoil and act upon the aerodynamic performance or the post stall-behavior.
This abstract presents a parameter study conducted for the DU-A 501-100 [1] and for the DU 97-W-300 [2] airfoils in order to investigate the effects of a flatback’s inclination angle.

Approach

The airfoil geometry was parameterized with Bezier curves, similarly to [3]; in total, four cubic curves were used (Figure 1).



Figure 1: Bezier parameterization of the DU-A 501-100 (left) and DU 97-W-300 (right) airfoils

For both reference airfoils, the flatback orientation was modified by introducing various inclination angles within the [-40°, +40°] range. The orientation angle is defined positively in the clockwise direction; the rotation point was taken in the middle of the trailing edge on the x-axis (Figure 2).



Figure 2: Original DU-A 501-100 airfoil and modified airfoils for inclination angles from 0° to +40° (10 ° step)

The aerodynamic performance of the profiles was assessed using two different approaches.
First, with Rfoil v1.1, an extension of the open-source solver Xfoil [4] developed in a joint research project including the Delft University of Technology, the Energy Research Center of the Netherlands, and the National Aerospace Laboratory of the Netherlands. It features in particular the ability of predicting effects of rotation on airfoil characteristics, and shows good agreement for similar profiles [3].

2-D CFD simulations were also performed using the open-source CFD software package OpenFOAM®2.1.1 [5]. Different features of OpenFOAM were tested while simulating 2-D flatback airfoils by changing numerical schemes and using different y+ values to investigate their influence on the results.
C-type structured meshes were created using the blockMesh utility in OpenFOAM, with a far field 25 meters away from the airfoil (Figure 3).



Figure 3: An overview of far field and airfoil mesh used in simulations for DU 97-W-300

Calculations rely on k-ω-SST [6] and k-kL-ω [7] turbulence models; RANS and URANS approaches were considered.

Steady-state simulations used simpleFoam, a steady-state solver for incompressible, turbulent flows, while unsteady simulations were conducted with pisoFoam, a transient solver for incompressible flows.

For the k-ω-SST turbulence model, steady and unsteady simulations were performed, with wall functions for k and ω. The total number of mesh cells is 655416.
For the k-kL-ω transition model, steady-state simulations were run without wall functions. A mesh with a total number of 711510 cells was used.

Main body of abstract

Rfoil analysis results

All simulations were performed on the DU 97-W-300 and DU-A 501-100 airfoils at a Reynolds number of 3e6 in viscous mode, for angles of attack ranging from 0° to 25° (0.5° step), with free transition.

Figure 4 shows the lift and the drag polar obtained for the DU 97-W-300 airfoil.
The airfoil has a maximum thickness of 29.94% and a trailing edge thickness of 1.74% of the chord length.
For both directions, the lift performance reduces with the increase in inclination. Modified airfoils show a similar stall behavior than the reference airfoil; no convergence issue was encountered during the computation, and all airfoils reproduced a smooth stall transition around 13°, though for different CL values.



Figure 4: Lift coefficient CL vs. angle of attack (top) and CL vs drag coefficient CD (bottom) – Rfoil results for DU 97-W-300 with positive trailing edge inclinations (left), and negative inclinations (right)

Figure 5 shows the lift and drag polars obtained for the DU-A 501-100 airfoil. The airfoil has a maximum thickness of 50.10% and a trailing edge thickness of 10.00% of the chord length.
Similarly to the previous case, the lift performance reduces with the increase in inclination in both directions. The difference of CL value between two different inclinations is much more important here; this can be explained by the larger trailing edge thickness (more than 5 times thicker than the DU 97-W-300), which amplify the aerodynamic sensitivity to flatback modifications. Modified airfoils present the same abrupt stall behavior than obtained for the reference airfoil; in this case, results show that stall tends to occur sooner for higher inclination angles.



Figure 5: Lift coefficient CL vs. angle of attack (top) and CL vs. drag coefficient CD (bottom) – Rfoil results for DU-A 501-100 for positive trailing edge inclination (left), and negative inclination (right)

CFD analysis results

All simulations were performed on the DU 97-W-300 airfoil at a Reynolds number of 3e6, resulting in an inlet speed of 45 m/s; angles of attack ranged from 0° to 21.95°; turbulence intensity at inlet was set as 0.07%.
The airfoil has a maximum thickness of 29.94% and a trailing edge thickness of 1.74% of the chord length. The resulting polars are shown in the Figures 6 and 7.



Figure 6: Comparison of the calculated lift polar with experimental results



Figure 7: Comparison of the calculated drag with experimental results

Comparing the CP curves for different turbulence models (Figure 8), it can be seen that for the k-kL-ω transition model, pressure difference between suction and pressure side of the airfoil is less than for the k-ω-SST turbulence model. For k-ω-SST steady-state CP is highly over predicted at the leading edge resulting in a very high lift coefficient at 19.18°. It seems that the k-kL-ω transition model performs better in comparison with the k-ω-SST turbulence model at higher angles of attack.



Figure 8: Pressure coefficient distribution at an angle of attack of 19.18°

Looking at velocity distributions in Figure 9, it has been observed that the k-kL-ω transition model (a) predicts boundary layer separation at 26.7 % of the chord length while the k-ω-SST unsteady-state (b) predicts this phenomenon more downstream towards the trailing edge at 29.90% of the chord length. This is the reason for higher lift coefficients with the k-ω-SST turbulence model in comparison with the k-kL-ω transition model at 19.18°.



Figure 9: Velocity distributions at 19.18° using (a) k-kL-ω steady-state (left), (b) k-ω-SST unsteady-state (center), (c) k-ω-SST steady-state (right)

For calculating L/D ratio (Figure 10), results are taken from k-ω-SST steady-state calculations until 12.37° and afterwards with k-ω-SST unsteady-state computation at higher angles of attack.



Figure 10: Lift to drag ratio for DU 97-W-300 using the k-ω-SST turbulence model


Conclusion

In this abstract, the influence of changing the trailing edge’s orientation on thick flatback airfoils has been investigated. Airfoil modifications were created by tilting the thick trailing edge of the DU 97-W-300 and DU-A 501-100 airfoils. Numerical solvers were validated and used to assess aerodynamic characteristics.

Preliminary results obtained with Rfoil show that changing the flatback’s inclination tend to diminish the aerodynamic performance, and favor stall for lower angles of attack; the sensitivity to this geometric change increases with the trailing edge thickness.

Validation against experimental results has been performed using OpenFOAM and different turbulence models.
• k-ω-SST turbulence model in steady-state mode slightly under-predicts CL and over predicts CD in the pre-stall region until 12.37°. This leads to lower L/D ratios.
• Using k-ω-SST in unsteady-state at higher angles of attacks predicts CD accurately but over estimates CL.
• k-kL-ω transition model in steady-state at higher angles of attack in the post-stall region predicts lift coefficient CL more accurately compared with the k-ω-SST turbulence model.

As models used in CFD simulations are based on different empirical assumptions, real flow behavior cannot be correctly predicted. This always results in deviations from experimental results. Also inboard regions are affected by 3D flow effects. However 2D CFD simulations can provide results close to real situations [8]. OpenFOAM can be used to predict the performance characteristics of flatback airfoils, as it provides a wide combination of numerical schemes and solvers that can be adapted together to simulate flows over airfoils.

Even though a flatback inclination does not seem to be favorable for the investigated airfoil, the effect on aerodynamic performance shall still be quantified, since an inclined flatback surface might be favorable for manufacturing purposes of a rotor blade.

2D simulations for the FX77-W500 and the DU-A 501-100 airfoils are thus being carried out to complete the validation of a CFD approach for flatback airfoils Suitable turbulence models were identified for the pre- and post-stall regions and they will be used to verify Rfoil calculations.


Learning objectives
Delegates attending this conference presentation can expect to learn more about:
- The aerodynamic performance of innovative flatback designs for thick airfoils, considering more precisely the inclination of the trailing edge
- Validation of numerical solvers for the aerodynamic analysis of thick flatback airfoils


References
[1] R. P.J.O.M. van Rooij: Thick Airfoil Design, Upwind project report D1B1.10.02-R00, obtained via personal communication to Nando Timmer and Ruud van Rooij, 2013
[2] N. Timmer, R.P.J.O.M. van Rooij: Summary of the Delft University Wind Turbine Dedicated Airfoils, AIAA-2003-0352, 41st Aerospace Sciences Meeting and Exhibit, Reno, Nevada, USA, 2003
[3] F. Grasso: Development of Thick Airfoils for Wind Turbines, AIAA-2012-0236, 50th AIAA Aerospace Sciences Meeting, Nashville, Tennessee, USA, 2012
[4] M. Drela, Harold Youngren: Xfoil, http://web.mit.edu/drela/Public/web/xfoil/, last updated 2008
[5] OpenFOAM – The open source CFD toolbox, http://www.openfoam.com/, 2013
[6] F.R. Menter, M. Kuntz, and R. Langtry. Ten Years of Industrial Experience with the SST Turbulence Model. Turbulence, Heat and Mass Transfer, 4, 2003
[7] D.K. Walters, D. Cokljat: A Three-Equation Eddy-Viscosity Model for Reynolds-Averaged Navier-Stokes Simulations of Transitional Flow, Journal of Fluids Engineering 2008 (130) 121401-1–121401-14
[8] D. Schulze, T. Rolf. Application of 2D CFD in Rotor Blade Design Potential and Challenges