Back to the programme printer.gif Print

Wednesday, 12 March 2014
09:00 - 10:30 Aerodynamics and rotor design
Science & Research  

Room: Llevant
Session description

The session is oriented to show recent computational and experimental findings on aerodynamic phenomena in horizontal (HAWT) and vertical access wind turbines (VAWT), as well as on new developments on system identification techniques related to the aeroelastic behaviour of wind turbine rotors and new aerodynamic design trends for very large wind turbines.

Learning objectives:

Delegates will learn about:

  • recent computational and experimental findings on aerodynamic phenomena in HAWT and VAWT
  • new developments on system identification techniques related to the aeroelastic behaviour of wind turbine rotors
  • innovative design trends for the aerodynamics of very large wind turbines
Lead Session Chair:
Alvaro Cuerva, Universidad Politécnica de Madrid, Spain

Sandrine Aubrun, Univ. Orléans, PRISME Laboratory, France
Laurent Beaudet Institut Pprime, UPR 3346 CNRS – Université de Poitiers – ENSMA, France
Laurent Beaudet (1) F P Christophe Sicot (1) Serge Huberson (1)
(1) Institut Pprime, Chasseneuil Futuroscope CEDEX, France

Printer friendly version: printer.gif Print

Presenter's biography

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

Mr. Beaudet graduated from ENSEEIHT, a French leading engineering school in Fluid Mechanics in Toulouse, after spending a semester at the Queen’s University of Belfast to get specialized in Aeronautics. Shortly after, he worked 3 years as engineer in the vertical axis wind turbine manufacturer Noveol to expertise all aerodynamic issues. While working for Noveol, he has been completing a PhD thesis on the vertical axis wind turbine aerodynamics in Institut Pprime of Poitiers about to end in early 2014. For the last 3 years, he also taught conventional wind turbine preliminary design and airfoil aerodynamics in University of Poitiers.


Pressure-velocity analysis of dynamic stall on a vertical axis wind turbine


The current interest in Darrieus vertical axis wind turbines (VAWTs) deals with their estimated better behavior than their horizontal axis counterparts in urban [1] and offshore [2] environments. Some new small urban VAWTs make use of high solidity to operate at low optimal tip speed ratio making them more silent but also prone to dynamic stall [3]. The few studies about high solidity VAWTs do not clarify the role dynamic stall plays in the energy exchange process and in the loads experienced by the blades. This study supplies keys for improved understanding of dynamic stall and tools to help optimization through numerical simulation.


To enable optimization of VAWTs efficiencies, a proper numerical model must be chosen. The vortex models represent currently the best compromise between accuracy, simplicity, cost and speed of computation. A classical panel method [4] was set up, but this model is inherently inviscid, hence it considers neither diffusion of vorticity nor blade stall. Corrections to account for these effects require suitable adjustments. In particular, some parameters of semi-empirical dynamic stall models are not directly adapted to a VAWT configuration because they were initiated for helicopters [5].
Therefore, experiments were undertaken to get a better insight into the impact of dynamic stall on the pressure distribution on the blades and on the dynamics of the vortices downstream of the rotor. The learning from this work will enable a more valuable tuning of the models and will facilitate control of dynamic stall and optimization of power production.
The experiments were carried out on a small scale VAWT in a closed loop wind tunnel. The rotor was composed of three straight blades held between two disks at each blade tip to promote 2D flow conditions (fig. 1). Solidity, defined as the ratio between cumulated blades area and swept area, equals 0.64. The tip speed ratios extended from 1 to 2, which are conditions that stimulate dynamic stall.

Fig. 1 – Overview of the experimental set up

Pressure measurements and particles image velocimetry (PIV) were performed at mid-span of the blades. Measure of pressure was selected to give relevant knowledge of the boundary layer separation and of the degree of influence of dynamic stall vortices on loads. PIV permits a tracking of the dynamic stall vortices along with an evaluation of their diffusion. The field of view for PIV was positioned in the near wake, outside and downstream of the rotor (fig. 2). Processing of the velocity fields includes proper orthogonal decomposition to filter the small scale turbulence, phase averaging over nearly 60 samples, vortex detection using a Γ2 criterion [6] and tracking algorithm and elliptic fit of the vortical coherent structures.

Fig. 2 – Position of the field of view relative to the rotor

Main body of abstract

Two singular behaviors are presented here: tip speed ratios (λ) equal to 1.9 and 1. Given the geometrical characteristics of the wind turbine, case λ = 1.9 is expected to be near the optimal tip speed ratio for power production, whereas case λ = 1 may be encountered during start-up and when the rotor is affected by gusty winds.
In the former case (λ = 1.9), pressure measurements reveal no sign of massive boundary layer separation. This can be explained by the wake and solid blockage effects of this high solidity VAWT. It results in a velocity deficiency in the rotor area and thus in lower angles of attack than usual, which are sufficiently low for the blade not to stall. Another explanation is the enhanced effect of flow curvature [7] caused by the high chord-to-radius ratio which may disturb boundary layer separation.
The comparison between numerical and experimental results indicates that pressures follow the same trends. For example, in the inner side of the blade at 20% of the chord from the leading edge (fig. 3), one can notice that there is a correct match of the pressures in the downwind half. In the other half, a gap between the results in the windward region is visible. This is partly due to the proximity with the wake of the previous blade (fig. 6) and the absence of diffusion of vorticity in the numerical model.

Fig. 3 – Comparison of pressures at 20% of chord on the inner side of the blade (λ = 1.9)

In the latter case (λ = 1), dynamic stall occurs and the blade creates a leading edge vortex that produces a powerful suction peak that travels towards the trailing edge on the inner side of the blade. On the one hand, it can be considered as beneficial for power production. Its decreasing magnitude as it moves rearward is associated with its growth of its size and the increase of distance from its center to the airfoil (fig. 4). On the other hand, the comparison with numerical simulations (fig. 5) shows another large detriment: a large region of separated boundary layer follows the vortex shedding. This is noticeable because of higher pressures in the experiments than in the numerical simulation where separation is not modeled. Consequently, after the vortex detaches from the blade, the blade does not contribute efficiently to power production because of full separation of the flow from roughly θ ≈ 120° to 210°, when the angle of attack finally changes sign.

Fig. 4 – Evolution of pressure suction peaks on the inner side of the blade (λ = 1)

Fig. 5 – Comparison of pressures at 20% of chord on the inner side of the blade (λ = 1)

Other tests from λ = 1.3 to 1.7 with different Reynolds numbers confirm the existence of dynamic stall and its delayed onset as tip speed ratio and Reynolds number rise. Hence, the balance between positive effect of suction and negative effect of separation is modified.
If we now analyze the vorticity field in the near wake, case λ = 1.9 displays one vortex, per blade and per cycle, passing through the field of view. It is a natural consequence of VAWT functioning because the angle of attack relative to the blade fluctuates. It is also computed in the numerical simulations and the results are in accordance with the experimental observations (fig. 6). The main discrepancies are a slightly different timing in the appearance of the vortex in the field of view, and a different size mainly due to the lack of diffusion. The evaluation of vortex properties reveal that there are significant enlargement of vertical structures and a rapid fall of the mean vorticity (fig. 7).

Fig. 6 – Comparison of vorticity fields (λ = 1.9)

Fig. 7 – Evolution of properties of the shed vortex (λ = 1.9)

Case λ = 1 exhibits two supplementary vortices that the numerical model does not model (vortices labeled B and C on fig. 8). They are the product of a complex interaction between the dynamic stall vortex and the blade. It is estimated that at least in the field of view, turbulence make total circulation in each vortex decline almost at the same rate (fig. 9), giving rise to weaker induced velocities. Still, it is found that the process of decay is different depending on whether the vortex is positive or negative, as the evolutions of mean vorticity and radius for positive vortices are steeper than for negative vortices.

Fig. 8 – Comparison of vorticity fields (λ = 1)

Fig. 9 – Evolution of properties of the shed vortices (λ = 1)


Results of this work contribute to both a better numerical modeling and a better understanding of the impact of dynamic stall on a VAWT to aim at optimizing the design of the rotor.
This study highlights the essential role of dynamic stall and diffusion modeling to accurately simulate high VAWTs performance. Vortex growth and speed over the blade were assessed, giving helpful data to adjust some parameters of the Leishman-Beddoes dynamic stall model [5] so it can deliver realistic simulations. The quantification of vortex decay in the near wake grants a meaningful reckoning of an effective viscosity parameter to be introduced in diffusion schemes. Besides, the calculated trajectories and deformations of the vortices represent a useful database for validation of numerical models.
Overall, this investigation also provides relevant data for understanding how to control dynamic stall in a high solidity VAWT. This control can be seen as the inquiry of the optimal operational parameters to prevent a negative balance of the dynamic stall effects. The coupled experimental/numerical study of the pressure distributions permitted a quantification of the suction effect of dynamic stall vortex. It is shown that the lowering of pressure can be locally and temporarily amplified by a factor superior than two, which may induce benefits that can be counteracted by the detrimental action of the succeeding boundary layer separation. One can also look for a functioning without any dynamic stall. For the given geometry of the rotor, the current work establishes to what extent Reynolds number and tip speed ratio delay dynamic stall onset. It can lead to significant impacts on the design of VAWTs. As an example, it is demonstrated that this specific high solidity VAWT can prevent dynamic stall from occurring at a tip speed (λ ≈ 2) for which other VAWTs suffered from it [8-9]. Control of dynamic stall can also be processed through the use of passive disturbance generators, active actuators, or variable pitch for instance. It is thought that a proper control could improve efficiency by making the dynamic stall vanish or by making it stay profitable for a longer time.

Learning objectives
Dynamic stall is a phenomenon that limits VAWT performances at low tip speed ratios. The work presented here gives valuable data to understand some aspects of dynamic stall in order to reach a better efficiency. Also, it provides some orders of magnitude for empirical parameters that are needed to construct numerical models that may be used to optimize VAWT design.

[1] S. Mertens. Wind Energy in the Built Environment: Concentrator Effects of Buildings. PhD thesis, Delft University of Technology, 2006.
[2] J. Paquette and M. Barone. Innovative Offshore Vertical-Axis Wind Turbine Rotor Project. In EWEA 2012 Annual Event, 2012.
[3] K. W. McLaren. A Numerical and Experimental Study of Unsteady Loading of High Solidity Vertical Axis Wind Turbines. PhD thesis, McMaster University, 2011.
[4] J. Katz and A. Plotkin. Low-Speed Aerodynamics: From Wing Theory to Panel Methods. McGraw-Hill, Inc., 1991.
[5] J. G. Leishman and T. S. Beddoes. A Semi-Empirical Model for Dynamic Stall. Journal of the American Helicopter Society, 34(3):3-17, 1989.
[6] L. Graftieaux, M. Michard, and N. Grosjean. Combining PIV, POD and Vortex Identification Algorithms for the Study of Unsteady Turbulent Swirling Flows. Measurement Science and Technology, 12:1422-1429, 2001.
[7] P. G. Migliore, W. P. Wolfe, and J. B. Fanucci. Flow Curvature Effects on Darrieus Turbine Blade Aerodynamics. Journal of Energy, 4(2):49-55, 1980.
[8] C. J. S. Ferreira, G. A. M. van Kuik, G. J. W. van Bussel, and F. Scarano. Visualization by PIV of Dynamic Stall on a Vertical Axis Wind Turbine. Experiments in Fluids, 46:97-108, 2009.
[9] A. Laneville and P. Vittecoq. Dynamic stall: the case of the vertical axis wind turbine. Journal of Solar Energy Engineering, 108(2):140-145, 1986.