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.

Spyros Voutsinas National Technical University of Athens, Greece
Co-authors:
Marinos Manolesos (1) F P Georgios Papadakis (1) Spyros Voutsinas (1)
(1) National Technical University of Athens, Athens, Greece

Printer friendly version: printer.gif Print

Presenter's biography

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

Faculty memeber at the National Technical University of Athens, Laboratory of Aerodynamics. Active in Wind Energy Research since 1990, participation in more than 40 RTD EU RTD projects on Wind Turbine, Helicopters and Aviation. Expertise in Aerodynamics, Aeroelasticity and Aeracoustics. Currently Board member of EAWE.

Abstract

Three-dimensional separation control using passive vortex generators

Introduction

Passive vortex generators (VGs) are quite popular as a means to enhance power by controlling separation. Despite that, the quantitative aspects of their function are not known in detail. Taking into account the trend towards more slender blades shaped by thicker airfoils and the increasingly important role VGs will play in design, NTUA launched a computational and experimental campaign on the effect of VGs on blades. The project aimed at: a) analyzing 3D separation on wing models, b) controlling 3D separation with VGs, and c) analyzing the flow on rotating blades with VGs. The present work focuses on b).

Approach

On the experimental side, a wing model exhibiting three-dimensional separation was tested with and without passive VGs in NTUA’s subsonic wind tunnel [1-4]. Pressure measurements on the airfoil and in the wake were combined with Stereo Particle Image Velocimetry (SPIV) recordings and flow visualizations. The wing model was placed wall to wall, while side fences were fitted in order to minimize the wind tunnel boundary layer effect and to allow adjustment of the wing Aspect Ratio (AR). The test matrix included angles of attack from -5° to 16° and the Reynolds number ranged from 0.5x10^6 to 1.5x10^6. All data reported in this paper concern a wing of AR = 2.0 at a Reynolds number of 0.87x10^6

On the computational side, simulations have been conducted using MaPflow, the in house Reynolds Averaged Navier-Stokes (RANS) flow solver [5]. MaPflow is an unsteady flow solver for the compressible Navier Stokes. In the present study the Spalart - Allmaras [6] turbulence model was used while the presence of the VGs was modelled by the BAY approximation [7], which is known to produce results comparable to simulations in which the VGs are fully resolved [8].

The present study has been conducted for a rectangular wing with an 18% thick airfoil profile designed at the NTUA [9]. The specific profile is optimized for use on variable pitch and variable speed multi MW wind turbine rotors. The profile belongs to the flat-top type experiencing trailing edge separation leading to a gradual built-up of the lift and smooth post stall behaviour. Under separated flow conditions the flow becomes highly three-dimensional and a single or more Stall Cells (SC) appear. The separated flow can be stabilized by placing a zigzag tape at x/c=0.02 and for only 10% of the span [4].

Following previous studies, e.g. [10, 11], counter rotating triangular vanes were selected as the VG basic concept. Through a computational parametric study the dimensioning of the VG was finalized.

Main body of abstract

The best performing VG configuration obtained out of the computational parametric study was experimentally tested for two different chordwise positions, namely x=0.3c and x=0.4c. In both cases the initially onset SC was suppressed for at least another 5° further than first obtained without VGs. This is clearly indicated in the lift curves (Figure 1 and Figure 2) by comparison to the case without VGs.





Up to α=12°, the CL curves are very close while beyond that the curves deviate significantly. With the VGs at x=0.4c a large SC would form at α=14° and a flow bifurcation was observed around α=13° for increasing and decreasing α. The hysteresis effect with respect to CL is shown in Figure 2. Figure 3 shows the pressure distribution along the wing chord at α=13° for increasing and decreasing α and in comparison to the uncontrolled case. It is clear that when α is increased the VG effect is strong and suppresses separation while when α is decreased, separation extends upstream of the VGs cancelling their effect. This happens because the SC formed with decreasing α was large enough to engulf the VGs into the SC pocket of separated flow. Figure 4 shows the oil flow pattern for increasing and decreasing α. Although the final pattern is slightly affected by gravity, the SC is clearly seen on the top image (decreasing α), while no such structure is observed in the lower image (increasing α). In fact the SC size was equal to the SC that would form on the wing without VGs, as the pressure distribution along the wing mid-span also confirms (Figure 3). The 3D structure in the lower image is due to the effect of the ZZ tape located at the centre of the wing span.





With the VGs at x=0.3c the Cl increased until α=16° where again flow bifurcation was observed. It was found that the flow alternated between two distinct states. One of the states, the “High Lift state”, dominated the time series, whereas the less frequent state, the “Low Lift state”, would appear for time intervals that would not exceed 3" (6% of a typical pressure signal time series, as the one given in Figure 5). The pressure distributions for the two distinct cases and for the case without VGs are shown in Figure 6. The High Lift state (red squares) follows the trend of the smaller angles, i.e. higher suction peak and limited separation compared to the uncontrolled case. On the other hand, the similarity between the uncontrolled case and the Low Lift state is clear, suggesting that the same, highly separated, state is resumed.





Initially the flow was simulated only for a very small AR wing, containing half a VG pair (AR = 0.058) with symmetry conditions at the sides of the computational domain. Such simulations have the inherent limitation that they cannot predict 3D separation. It was found that SC formation is predicted only if AR ≥ 1, in accordance with [12], and that AR is a crucial factor when it comes to computing 3D separated flows, with or without control.

The best performing case with VGs located at x=0.3c was selected for further study and the flow downstream of the VGs was investigated using SPIV. For the measurement planes shown in Figure 7, the time averaged flow characteristics were analysed in terms of velocity, vorticity and Re stress distributions. Contours of the streamwise velocity, vorticity and (v'v') ̅ (as an example) are given in Figure 8. The high concentration of (v'v') ̅ between the two vortices is indicative of the spanwise wandering of the streamwise VGs vortices. The combined conclusion (from vortex strength, vorticity contours and turbulence quantities profiles not shown here) is that at the most upstream measurement plane, turbulent transport between the VG vortices and the underlying flow is strong, while from the second plane to the third, diffusion becomes the main mechanism that governs the flow.




.

Conclusion

Control of 3D separation of the SC type by means of passive VGs was studied experimentally and computationally. An optimized VG configuration obtained through a computational parametric study was tested for two chordwise placements. Results confirm that SC formation can be effectively controlled using triangular vanes. Depending on the VGs' chordwise location flow bifurcation or hysteresis has been detected, possibly associated with the breakdown of the VG vortices and the subsequent formation of a SC. It was also confirmed that low cost simulations, carried out on wing strips with one VG and periodic conditions fail to predict the onset of SCs and therefore cannot be trusted at higher angles of attack. This constitutes a major challenge for CFD when considering the effect of VGs on wind turbine blades. In this respect of particular importance is that on rotating blades and over their inboard part where VGs are usually placed, the VGs will compete with the radial pressure gradient [13].

Detailed information has been collected with respect to the flow downstream of the VGs at α = 10° and at a Reynolds number 0.87x10^6 by means of SPIV measurements. Both the mean quantities and the turbulence characteristics of the vortical flow have been analyzed, suggesting that up to 30 VG heights downstream of the VGs vortex interaction with the underlying boundary layer is still significant, while further downstream diffusion dominates the flow. A data base has been produced providing a basis for CFD validation and turbulence modeling tuning. The latter is of particular importance as the experimental results show high anisotropy rending eddy viscosity modeling inappropriate for detailed flow simulations.


Learning objectives
1. Enhance understanding on the complex interaction between the VG flow and 3D separation.
2. Create a detailed data base for the flow downstream of the VGs, which will facilitate CFD validation both on a turbulence modeling level and as regards the mean flow characteristics.
3. Examine the limitations of eddy viscosity based RANS simulations in view of designing/optimizing VG layouts
4. Provide guidelines for CFD simulations on rotating blades



References
1. Manolesos, M., G. Papadakis, and S.G. Voutsinas, Experimental and computational analysis of stall cells on rectangular wings. Wind Energy, 2013.
2. Manolesos, M. and S.G. Voutsinas, Study of a Stall Cell using Stereo-PIV. Physics of Fluids, Submitted for publication, 2013.
3. Manolesos, M., G. Papadakis, and S. Voutsinas. A combined investigation on the formation of Stall Cells on airfoils. in EWEA Conference. 2012. Oldenburg, Germany.
4. Manolesos, M. and S.G. Voutsinas, Geometrical characterization of stall cells on rectangular wings. Wind Energy, 2013.
5. Papadakis, G., Formulation of a cell-centered (U)RANS compressible solver, in PhD Progress Report2011, National Technical University of Athens: Athens.
6. Spalart, P.R. and S.R. Allmaras. A one-equation turbulence model for aerodynamic flows. in 30th Aerospace Sciences Meeting and Exhibit. 1992. AIAA.
7. Bender, E.E., B.H. Anderson, and P.J. Yagle, Vortex generator modeling for Navier-Stokes codes. ASME Paper FEDSM99-6919, 1999.
8. Jirasek, A., Vortex-Generator Model and Its Application to Flow Control. Journal of Aircraft, 2005. 42(6): p. 1486-1491.
9. Mourikis, D., V. Riziotis, and S. Voutsinas, Optimum aerodynamic design of 52m blade for a prototype 5MW WEC, in TR-012005, National Technical University of Athens: Athens.
10. Godard, G. and M. Stanislas, Control of a decelerating boundary layer. Part 1: Optimization of passive vortex generators. Aerospace Science and Technology, 2006. 10(3): p. 181-191.
11. Ashill, P.R., J.L. Fulker, and K.C. Hackett, Research at DERA on sub-boundary layer vortexgenerator s (SBVGs). AIAA Paper, 2001. 887.
12. Taira, K. and T. Colonius, Three-dimensional flows around low-aspect-ratio flat-plate wings at low Reynolds numbers. Journal of Fluid Mechanics, 2009. 623.
13. Perivolaris, Y.G. and S.G. Voutsinas. Numerical Investigation of the Boundary Layer Control and Wind Turbine Power Production Enhancement Using Solid Vortex Generators. in Proceddings of EWEC European Wind Energy Conference, Warsaw. 2010.