Scientific Bulletin of the The 6th International Conference on
Politehnica University of Timisoara Hydraulic Machinery and Hydrodynamics
Transactions on Mechanics Special issue Timisoara, Romania, October 21 - 22, 2004
COMPUTATIONAL MODELLING OF GIROMILL WIND TURBINES
Francisc GYULAI, Prof. Adrian BEJ, Senior Scientific Researcher
Department of Hydraulic Machinery Wind Energy Research Center
Wind Energy Research Center “Politehnica” University of Timisoara
“Politehnica” University of Timisoara
Bv Mihai Viteazu 1, 300222, Timisoara, Romania
Tel.: (+40) 256 403689 (96), Fax: (+40) 256 403682, Email: mh@[Link]; adibej@[Link]
ABSTRACT Fx : drag force [N]
Fr : radial force [N]
Vertical axis wind turbines with straight blades, known
as Giromill or H-Darrieus type, are of present interest H : rotor height (blade length) [m]
for small-power scale applications. Their aerodynamics kT : turbine influence coefficient [-]
is more complicated than that of horizontal axis wind l : airfoil chord length [m]
turbines due to the dependence of all kinematic PT arb : wind turbine power at turbine shaft [W]
parameters on the blade polar position reported to
R : turbine radius [m]
wind direction.
S = 2 R ⋅ H : swept area [m
2]
The paper shows some aspects concerning software
for Giromill wind turbine designing as well as the v : airflow absolute velocity (wind speed) [m/s]
results developed by computational modeling for w : airflow relative velocity [m/s]
various values of wind turbine's tip speed ratio. This u : tangential velocity (blade tip speed) [m/s]
issue has particular aspects for the low-speed wind z : number of blades [-]
turbines. In the modeling algorithms is used the lifting α : angle between tangential and absolute velocity [-]
method for a wide filed of the blade's angle of attack β : angle between tangential and relative velocity [-]
and for a flow with boundary layer detachment. There ϕ : polar position angle [-]
are used recent information concerning the finite blade
span. The model uses a catalog of NACA four digit λ = u / v : tip speed ratio [-]
airfoils developed for a wide range of geometry, angles λ 0 : rated tip speed ratio (turbine rapidity) [-]
of attacks, rough-ness, Reynolds numbers and blade ρ : air density [kg/m3]
spans. The methodology is applied for small power
wind turbine designing used in small households. 1. INTRODUCTION
The results achieved through the computational In the latest years the market of wind turbines has
modeling are confronted with experimental results shown certain preferences for the horizontal axis wind
from the literature as well as with own testing done turbines (HAWT). Despite this trend, for small power
within an aerodynamic wind tunnel of "Politehnica" scale there is still of interest also the wind turbines
University of Timisoara. with vertical axis (VAWT) especially as Girromil type
(H-Darrieus, U-shaped Darrieus, and VAWT straight
KEYWORDS bladed Darrieus). The reference literature in this field
wind turbine, computational modelling, aerodynamics, shows only briefly information in which means the
vertical axis aerodynamics of their blades [1 - 5]. Some recent pa-
pers re-approach some of these issues [5] rather for
NOMENCLATURE VAWTs for household uses [7 -9]. We have taken
again the issue of these wind turbines into the frame
A: lifting area [m2] of a research project concerning a 2.5 kW wind turbine
C P arb : power coefficient corresponding to turbine meant for certain households aiming either to save part
shaft [-] of the taxes spend on electrical energy bought from
C x : drag coefficient [-] electrical grid, or as stand alone energy source. In this
C y : lifting coefficient [-] context we did some computational modeling and
testing in wind tunnel. Further there are shown the
F y : lifting force [N]
main conclusions concerning this research.
283
2. SPECIAL ASPECTS CONCERNING VAWTS
Darries wind turbines were positively evaluated
due to the fact that they do not require yaw systems
and because heavy parts of their power drive (gearbox,
electrical generator, etc) are placed close to the ground.
The main disadvantage of the classical solution is the
top bearing that is guyed by three cables and three
supplementary foundations for these. Hence a lot of
damages have emerged. Further, the curved shape of
the blades requires special technologies. These disad-
vantages are overtaken by the Girromill solution (H- Figure 1a. Dependence of kinematic measures on the
Darrieus). But still remain some disadvantages of blade position
aerodynamic nature more severe then the absence of
the yaw system, which do the VAWTs to be feasible
only for small powers. These disadvantages are hence
to the unsteady interaction between the airflow and
the blades. This operating regime loads the sustaining
structure with oscillating forces and torque which
involve some designing difficulties. This unsteady load
generates difficulties in using higher elevations (higher
towers), tendency which obviously is used to HAWTs
having certain economical benefits. Despite all these, Figure 1b. Kinematics in the case of "u > v"
Giromill solution for small power scale might have
advantages then HAWTs. Our researches aimed to fit
such of wind turbines on the top of house roofs, offer-
ing cheap solutions for reducing the household expenses
on electrical energy consume.
3. OPTIMIZATION OF GIRROMIL WIND
TURBINES THROUGH COMPUTATIONAL
MODELING
3.1. Aerodynamic bases of wind turbine models
The software set aims engineering applications asking Figure 1c. Kinematics in the case of "u < v"
thus knowledge, which are used in technical applications
with an enough precision. We used in this way the
lifting method [6], developed on the basis of some
experimental data.
In figure1 and using its related relations can be
seen the connection between the blade polar position,
the inlet velocity triangle as well as the aerodynamic
forces on the blade. In figure 1a. can be seen the used Figure 2. Velocity triangle and the directions of
notes, and in figure 1b and 1c are given the measures aerodynamic forces
in relative position to the tangential velocity ( u ), for
ratio u / v > 1 and u / v < 1 . The angle field of relative The relations for the inlet velocity triangle are:
velocity (w) on which depends directly the incidence α=ϕ ; w2 = u 2 + v 2 − 2 ⋅ u ⋅ v ⋅ cos α (1)
angle of blade airfoil is shown in figure 2. Angle β ,
v
in the first case extends on a limited field ( β max , β min ), sin β = ⋅ sin α (2)
w
and in the second case the field extends on 2π radians
w2
between +π and −π radians. The airfoil setting angle Fy = C y ⋅ ρ ⋅ ⋅A (3)
is very low, closed to zero. 2
An important remark emerges: the incidence angle w2
variation and the aerodynamic force variation depends Fx = C x ⋅ ρ ⋅ ⋅A (4)
2
on the polar blade position. Therefore must be used
an airfoils catalog extended on the entire field of the 3.2. Airfoils catalog
incidence angle up to 3600, especially when a slow We used original software developed for NACA 4
wind turbine is approached. digit series airfoils. Instead the current information
284
used for low incidence angles field, there was extended as well as the dissipations inside the near-wake. These
this field for the entire possible values of incidence influences are evaluated through an empiric constant
angles (0, 3600). ( kT ) found as result of some measurements done on
The software allows to compute the lifting, drag and real turbines.
moment coefficients for finite blade span, depending Selected from the data volume obtained during the
on: relative chamber, chamber position, relative thick- designing of the 2.5 kW wind turbine, we show further
ness, airfoil surface roughness (smooth and standard some significant examples.
roughness), Reynolds number. Details about the Figure 3 shows the relation between the dimen-
airfoils catalog can be find in reference [6]. sionless curves and the operational curves and figure
3.1. Finite blade span effect 4 gives the variation of aerodynamic radial force
depending on blade polar position.
The Giromill wind turbine design requires corrections Table 1 and 2 contain the kinematic measures and
under the data given in the airfoils catalog (infinite span) aerodynamic forces computed for an operational regime.
for the case of a finite span blade. For a wide range of
incidence angles we used distinctly relations for low
incidences (fully attached regime), and for the detached
flow in conditions of high incidences (post-stall regime).
For the first case we used the Prandtl's relations, and
for the second one the relations given by Viterna and
Corrigen which were also used by Lissaman [10, 11].
3.2. Wake effect
Wilson's paper [4] as well as Paraschivoiu's papers
[5] approach the issue of wake effect. For VAWTs the Figure 3a. Dimensionless curves for different
rapidity ( λ0 )
conditions downwind the turbine are much complex
then those of HAWTs. The optimal solutions are
usually conditioned by the wind speed field and pres-
sure field downwind the rotor. Wilson formulates the
necessity of some future researches aiming to draw
nearer the theoretical and the experimental perform-
ances. The data given in the reference literature aren't
enough certain for the engineering practice.
The new proposed models which considerate the
pressure fall effect inside the near-wake [6] offer a new
vision about the limits of extracted energy from airflow.
In our software we introduces, as a temporally solution,
a decreasing coefficient of wind speed through that we
estimated the energetic influence of the wake condi- Figure 3b. Operational curves
tions. For verifying the computational results we used
a statistical model, which reflects a synthesis of an
engineering experience, gathered in this field [12].
3.3. Some results of computational modeling
The software set achieved for Giromill wind turbines
offers several possibilities for analyses developed
during the wind turbine designing task. These analyses
can be group in two directions:
• The influence of wind turbine rapidity on dimen-
sionless curves, and the setting of main parameters
for operational curve (rotational speeds, main dimen-
sional features) Figure 4. Radial aerodynamic force dependence on
• The setting of geometrical details of wind turbine blade polar position
blades suitable with a desired operational curve, and
the calculation of aerodynamic forces used for 4. TESTING IN WIND TUNNEL
mechanical strength calculation. The control and the corrections on the designing
One of the less certain point of the software set is algorithm of software set were done using two control
the evaluation of wake effect on the velocity triangle. methods: comparations with other technical achieve-
The software identifies certainly the inlet velocity mets in VAWTs filed, and testing in wind tunell upon
triangle but less certainly the outlet velocity triangle a wind turbine model.
285
Table 1. Kinematic measures Figure 5 shows this model during the testing done
Turbine T23 ; D = 2.0 m ; H = 3 m in wind tunnel.
u = 15m/s ; v = 12 m/s; ρ = 1.225 kg/m3
l = 0.4 m; z = 3 ; kT = 0.5 ; smooth surface
ϕ v1r v1t α1 β1 w1 w1r w1t
0 0 0
[ ] [m/s] [m/s] [ ] [ ] [m/s] [m/s] [m/s]
0 0 6 0 0 9 0 9
30 3 5.2 30 17.0 10.25 3 9.8
60 5.2 3 60 23.4 13.08 5.2 12
90 6 0 90 21.8 16.16 6.0 15
120 5.2 -3 120 16.1 18.73 5.2 18
Figure 5. Small-scale model of a Giromill wind rotor
150 3 -5.2 150 8.45 20,42 3 20.2
tested in wind tunnel
180 0 -6 180 0 21 0 21
210 -3 -5.2 -150 -8.45 20.42 -3 20.2 The testing results allowed us to adjust our designing
240 -5.2 -3 -120 -16.1 18.73 -5.2 18 software set meant for Giromill wind turbines, enabling
270 -6 0 -90 -21.8 16.16 -6 15 also to assess the constant referring to the influence
300 -5.2 3 -60 -23.4 13.08 -5.2 12 of wind turbine upon the wind speed field. At the
330 -3 5.2 -30 -17.0 10.25 -3 9.8 same time during testing was confirmed a variation
360 0 6 0 0 9 0 of aerodynamic loads upon the turbine structure.
v1r: radial component of absolute inlet velocity
v1t: tangential component of absolute inlet velocity REFERENCES
w1r: radial component of relative inlet velocity 1. Hau E. (1988) Windkraftagen, Springer Verlag, Deutsch-
v1t: tangential component of relative inlet velocity land, pp.127-129.
1: index of the inlet section 2. Spera D.A. (1994) Introduction to modern wind turbines
/ Wind turbine technology, ASME Press, New York,
Table 2. Aerodynamic forces USA, pp.47
ϕ inc. Fy Fyt Fyr Fx Fxt Fxr 3. Divone L.A. (1994) Evolution of modern wind turbine /
0 0 Wind turbine technology, ASME Press, New York,
[] [ ] [N] [N] [N] [N] [N] [N]
0 0 0 0 0 0.57 -0.57 0 USA, pp.73.
30 17 90 27.47 86.5 1.27 -1.22 0.37 4. Wilson R.E. Aerodynamic behavior of wind turbines /
60 23.4 127 50.53 116.7 13.83 -12.7 5.5 Wind turbine technology, ASME Press, New York,
90 21.8 200 74.12 185.3 16.15 -15 6 USA, pp.215.
120 16.1 312 86.42 299.4 1.34 -1.29 0.37 5. Paraschivoiu I. (1988) Double-multiple stream tube
150 8.4 196 28.84 194.1 8.93 -8.84 1.31 model for studying vertical axis wind turbine, Journal
180 0 0 0 0 2.81 -2.81 0 of propulsion and power, vol.4, No.4.
210 -8.4 -196 28.84 -194.1 8.93 -8.84 -1.31 6. Bej A. (2001) Optimizarea constructiei turbinelor eoliene
240 -16.1 -312 86.42 -299.4 1.34 -1.29 -0.37 cu autoplafonare de putere si frinare aerodinamica -
270 -21.8 -200 74.12 -185.3 16.15 -15 -6 Teza de doctorat, U.P. Timisoara.
300 -23.4 -127 50.53 -116.7 13.83 -12.7 -5.5 7. Kamissoko F., De Vilde R.D.L., Hirsch Ch. (1993),
330 -17 -90 26.47 -86.5 1.27 -1.22 -0.37 Theoretical and experimental study of VAWT straight
360 0 0 0 0 0.57 -0.57 0 bladed Darrieus rotor, European Union Community
Wind Energy Conference, Lubeck-Travemunde, Ger-
z =1 PTarb =584 many, pp.725-727.
turbine power [W] z=2 PTarb =1128 8. Timmer W.A., Van Rooij R.P.J.O.M (2001) Some
z=3 PTarb =1692 aspects of high angle of attack flow on airfoils for wind
inc. : incidence angle turbine application, European Wind Energy Conference,
Fyt: tangential component of lifting force Copenhagen, Denmark.
Fyr: radial component of lifting force 9. Yasuyuki O., Yasuyuki N., Shinichi S. (2001) A study
Fxt: tangential component of drag force of U-shaped Darrieus wind turbines, European Wind
Energy Conference, Copenhagen, Denmark.
Fxr: radial component of drag force
10. Viterna L.A., Corrigan R.D. (1981) Fixed pitch rotor
The testing was done in one of the laboratory for performance of large horizontal axis wind turbines,
aerodynamics from "Politehnica" University of NASA CP-2230, USA.
Timisoara, using a small Giromill rotor having the 11. Lissaman P.B.S.(1994) Wind turbine airfoils and rotor
following features: diameter: 300 mm; height (balde wakes / chapter 6 issued in Wind turbine technology,
span): 300 mm; airfiol: NACA 0020; chord length: ASME Press, New York, USA.
12. Gyulai F. (2004) Mathematical models for dynamic
100 mm; number shaft): 130 W; rated tip speed ratio: 2;
study of wind turbines , The 6th International Confer-
rated wind speed: of baldes: 3; rated power (at turbine's ence on Hydraulic Machinery and Hydrodynamics,
20 m / s, maximum rotational speed: 2500 rpm. Timisoara, Romania.
286
