You are on page 1of 8

228

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 6, NO. 1, JANUARY 2015

2-D-CFD Analysis of the Effect of Trailing Edge


Shape on the Performance of a Straight-Blade
Vertical Axis Wind Turbine
Khaled M. Almohammadi, Derek B. Ingham, Lin Ma, and Mohamed Pourkashanian

AbstractThis paper numerically investigates the effect of the


trailing edge profile on the performance of the straight-blade
vertical axis wind turbine (SB-VAWT). In a 2-D cross-section of
the SB-VAWT model, four trailing edge profiles are investigated,
namely sharp, rounded, S-blunt, and R-blunt. The numerical
investigation is based on the unsteady Reynolds averaged NavierStokes (URANS) equations combined with the transition shear
stress transport (SST) model in order to account for the transition in the boundary layer in the vicinity of the airfoils. It has
been found that the trailing edge profile may play a significant role
in improving the turbine performance and should be accurately
accounted for in the design process of the SB-VAWT.
Index TermsComputational fluid dynamics (CFD), straightblade vertical axis wind turbine (SB-VAWT), trailing edge profile,
transition, unsteady Reynolds averaged Navier-Stokes (URANS).

I. I NTRODUCTION

N THE LAST few years, harnessing wind energy using


wind turbines has gained a remarkable consideration in the
scientific community due to the rapid growth in the population
and the depletion of fossil fuels. Two turbine designs, namely
horizontal axis wind turbine (HAWT) and vertical axis wind
turbine (VAWT), have been developed to efficiently convert the
kinetic energy in the wind into mechanical energy which is used
for electrical generation. However, the HAWTs are more efficient when the turbulence level is low and the wind flow is
relatively uniform provided that the flow is aligned to the wind
which is maintained using a yawing controlling mechanism
[1], [2].
In urban regions, the flow is complex due to the obstacles that
disturb the wind flow and this results in a continuous change in
the wind direction. The VAWTs operate independently of the
wind direction and this makes it more efficient in urban regions
compared to HAWTs [3]. Therefore, improving the efficiency
of this kind of turbine is essential to maximize the harnessed
power of the wind.
Manuscript received April 23, 2014; revised August 22, 2014 and October
12, 2014; accepted October 20, 2014. Date of publication November 24, 2014;
date of current version December 12, 2014. Paper no. TSTE-00174-2014.
K. M. Almohammadi is with the Department of Mechanical Engineering,
Taibah University, Madinah, Saudi Arabia.
D. B. Ingham, L. Ma, and M. Pourkashanian are with the Energy Technology
and Innovation Initiative (ETII), University of Leeds, Leeds LS2 9JT, U.K.
(e-mail: M.Pourkashanian@leeds.ac.uk).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TSTE.2014.2365474

On the basis of the underlying physics, the VAWT could be


classified into lift-based turbines, such as the Darrieus turbine,
and drag-based turbine, such as the Savonius turbine. In the
early 1920s, Betz [4] has reported that no more than 59.3% of
the theoretical wind power can be harnessed from the wind.
This is much higher than the maximum theoretical efficiency
that can be obtained from the drag-based turbines, namely
29.6% [5] and therefore the lifting-based turbine is desirable
in urban regions.
Several researchers have focused on improving the airfoil
profile [6][9] in order to maximize the turbine performance
but less attention has been paid to the significance of the
trailing edge in the context of VAWTs. Straight and curved
blade shapes are commonly employed in VAWTs. However,
the straight blade shape is more common due to its simple
design and its low manufacturing and maintenance cost [10] in
contrast to the curved blade shape. Further, Hamid et al. [11]
compared the performance of straight and curved blades for
VAWTs and concluded that the straight bladed performed better
than curved blades at low wind speeds. The trailing edge profile is traditionally, sharp, rounded, or blunt. The separation of
the boundary layer and the vortex formation and shedding near
the trailing edge is mainly influenced by the trailing edge shape.
From a physical point of view, the trailing edge profile plays
an extremely important role on the flow development on the
airfoil surface during the turbine operation. In 1950, Summers
and Page [12] performed one of the early investigations on the
trailing edge profile effect on the aerodynamics of the airfoil
and they concluded that increasing the trailing edge thickness
results in an increase in the lift coefficient and the slope of the
lift curve. In the same year, Smith and Schaefer [13] investigated the airfoil performance by fixing the airfoil trailing edge
thickness and cutting away 1.5% of the chord length from the
rear of the trailing edge and they concluded that the increase in
the trailing edge caused by cutting off the chord has a minor
effect on the lift coefficient and this agrees with the recently
published research by Gmez and Pinilla [14]. A year later, a
similar conclusion was reported by employing a rounded trailing edge with a radius of about 15% of the maximum airfoil
thickness [15]. All these researches have mainly focused on a
static single airfoil and this was continued to the 1970s by the
work performed by NASA [16] which was followed by several
similar investigations [14], [17], [18].
In the context of wind turbines, the analysis of the trailing
edge profile is also performed by many researchers on a single

1949-3029 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

ALMOHAMMADI et al.: 2-D-CFD ANALYSIS OF THE EFFECT OF TRAILING EDGE SHAPE ON THE PERFORMANCE OF SB-VAWT

static airfoil rather than on a full rotating turbine model. The


analysis is generally aimed at increasing the maximum lift but
limited attention has been paid on the trailing edge effect on the
full turbine model where the stall develops in a real situation.
Standish and Van Dam [19] analyzed the blunt trailing edge
effect on the airfoil performance but the employed turbulence
model, namely the SpalartAllmaras model, failed to capture
the stall, which significantly affects the turbine performance.
Further, they have experimentally investigated the effect of the
blunt trailing edge by increasing the trailing edge thickness
while maintaining the chord length. The outcome of the experimental investigation is that increasing the trailing edge thickness reduces the transition sensitivity over the airfoil surface
and particularly the formation of the leading edge vortex [18].
The purpose of this paper is to numerically investigate the
effect of the trailing edge profile of a 2-D cross-section model
of the SB-VAWT on the performance of the straight-blade
VAWT (SB-VAWT). Four trailing edge profiles are investigated, namely sharp, rounded, S-blunt and R-blunt. The
numerical investigation is based on the unsteady Reynolds averaged Navier-Stokes (URANS) equations combined with the
Transition SST model in order to account for the transition in
the boundary layer in the vicinity of the airfoils. The Transition
SST model has been investigated by many researches. For
an oscillating airfoil, Wang et al. [20] found that this model
could accurately produce the experimental data, including the
laminar-turbulent bubbles and the complex flow structure near
the airfoil at a high angle of attack, with reasonable accuracy.
Further, for VAWTs, Danao et al. [21] have concluded that the
Transition SST model is capable of capturing the flow structure developed by the VAWT. Therefore, we concluded that the
Transition SST model is suitable for the purpose of this paper.
II. SB-VAWT M ODELS
The simplicity and the insensitivity of the SB-VAWT to the
wind flow direction make these kinds of turbines a viable candidate for urban regions. However, the flow in the vicinity of
the turbine airfoils and across the turbine is relatively complex and this is mainly driven by the continuous separation and
reattachment of the airfoil boundary layer as the airfoil rotates
during the turbine operation. Therefore, several models have
been developed in order to estimate the efficiency of the SBVAWT and consequently optimize turbine performance, and
these models are the stream tubes model (single, double, and
double multiple), cascade model, vortex model, and computational fluid dynamics. Stream tube models are based on the
momentum theory where the momentum rate of change of the
air due to the stream wise forces caused by the flow velocity
exerted on the blades are equal to the total change of the velocity multiplied by the mass flow rate and to the total pressure
average throughout the turbine. The models of this class are
the single stream tube model [22], multiple stream tube model,
and double multiple stream tube model [23], [24]. These models are limited to low tip speed ratios, which is the ratio of the
rotational speed of the turbine to the undisturbed wind speed
and solidities [1], and extensive modifications are required to
be implemented in order to employ these models in complex

229

regimes, such as the flow in urban regions. Alternatively, the


vortex model, which is based on the influence of the vorticity
on the velocity field, may be employed. The vortex model
was introduced by Larsen [25] in 1975 but it was Fanucci
and Walters [26] who employed this model in an SB-VAWT.
However, far too little attention has been paid to the vortex
model despite the improvements made by many researchers
[27][29] since this model is considered to be computationally
expensive compared to the stream tubes model (single, double, and double multiple), cascade model [10] for SB-VAWTs.
However, the vortex model requires previous knowledge of the
lift coefficient of the investigated airfoil, either theoretically or
numerically which is not the case when employing the URANS
equations, and therefore, the URANS equations are desirable
for relatively accurate computations of SB-VAWTs. Another
method for modeling the flow of wind turbines is the cascade
model which was first employed in the context of the VAWT by
Hirsch and Mandal [30]. The idea of the cascade model is that
the airfoils are equally spaced on a surface with a length equal
to the circumference of the turbine; then the Bernoulli equation
is employed to find the relationship between the wind velocity
and the wake velocity. These models do not sufficiently model
the nonlinearity of the flow in the viscous region near the airfoils where a complex nonlinear phenomenon develops and this
is known as dynamic stall. In 1990, Leishman [31] concluded
that this nonlinear phenomenon, which is present in all SBVAWTs, can only be correctly modeled by employing the full
Navier-Stokes equations with a suitable turbulence model, and
therefore the full Navier-Stokes equations are the basis of this
investigation where a transitional model, namely the Transition
SST model is employed to capture the flow development in the
boundary layer of the airfoils.
III. N UMERICAL S CHEME
A. Computational Domain
The geometry of the investigated SB-VAWT is created based
on the experimental setup produced by Bravo and McLaren
[32][34]. The turbine is constructed of three NACA 0015 airfoils (straight bladed) equally spaced by 120 with a chord
length of 0.4 m, a radius of 3.125c, and the blade heights are
7.5c. Therefore, the turbine reduced frequency K = c/2R is
0.16 which indicates a considerable level of unsteadiness. This
turbine is mainly designed for the use in urban regions and
therefore it is expected that the power coefficient peaks are to
be at a tip speed ratio less than 2, as experimentally shown in
Fig. 1. The experiment is performed in a wind tunnel of size
9 m 9 m wind tunnel and the torque is measured by averaging
over 120 s at a wind speed of 10 m/s [32][34].
Therefore, a 2-D full model of the turbine is created to
perform the computations by employing the sliding mesh technique with an interface boundary condition between a circular
rotating region and a rectangular stationary region.
The left face of the rectangle is set to the velocity inlet boundary condition and the right face is set to the pressure outlet
boundary condition, whereas the upper and the lower faces are
set to a solid wall condition. The length of the rectangular stationary region is 87.5c, whereas the width is 50c. The radius of

230

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 6, NO. 1, JANUARY 2015

Fig. 1. Experimental setup and SB-VAWT performance by Bravo and McLaren


[32][34].

Fig. 3. Schematic of trailing edge profile for NACA 0015.

Fig. 2. Schematic of the 2-D computational domain for the investigated


SB-VAWT.

the circular rotating region is 8.75c. The domain size sensitivity has been investigated by Mclaren [35], and therefore these
dimensions have been fixed for all the computations presented
in this paper and are shown in Fig. 2.
Four trailing edge shapes, as seen in Fig. 3, have been implemented in this investigation for the traditional NACA 0015
where the sharp trailing edge is set as the base airfoil from
which the trailing edge is modified into rounded and two blunt
trailing edges (one created from the sharp trailing edge, named
the S-Blunt T.E., and one created from the rounded trailing
edge, named the R-Blunt T.E.). The rounded edge is created
with a radius of 10% of the chord length, namely 0.04 m, and
this resulted in the thicker trailing edge in order to smooth the
airfoil surface. Because two different thicknesses are employed,
the blunt trailing edge has to be created and investigated in both
thicknesses at a distance of 10% of the chord length from the
end of the sharp and rounded trailing edge in order to analyze
the effect of the trailing edge thickness, and then the rounded
and sharp trailing edge may be compared. The geometry of
the turbine is created using GAMBIT and the simulation is
performed using the commercial software ANSYS FLUENT.
The computational domain is meshed carefully by employing
structured quadrilateral cells with approximately 2000 nodes on
the airfoil surface with the relative distance of the first node
from the surface of the airfoil, namely y + , of about 1.
The structured mesh thickness is estimated based on
flow Reynolds number RN = c U / = 2.83 105 , where
U = 10 m/s and c = 0.4 m, to be about 0.752 mm. This value
is considered to be conservative because the magnitude of
the velocity near the airfoil during turbine operation is about
17 m/s, therefore the actual boundary layer thickness is much
less than 0.752 mm and is resolved within the structured mesh.

The rest of the domain is created with unstructured quadrilateral cells to maintain computational stability and to minimize
the numerical diffusion effect [36], [37]. This resulted in four
meshes with the order of about 500 000 cells for the use of computations, which is sufficiently dense to describe the physics
involved in the SB-VAWT. Based on [38], this mesh size is satisfactory for the purpose of the investigation of the trailing edge
effect on the SB-VAWT analyzed in this paper. It is important
to note that the 2-D CFD model of the SB-VAWT may substantially over predict the peak in the power coefficient of the
turbine at high tip speed ratios as reviewed in [38].
However, selecting the appropriate turbulence model is
essential in order to describe the physics involved in these kinds
of turbines. In the region near the boundary layer of the airfoil, the flow is in transitional state due to the effect of the
low Reynolds number, and as a result, separation bubbles are
formed [39], [40]. These bubbles significantly affect the turbine
performance and the accuracy of the numerical prediction, and
may only be predicted when a transitional model is employed.
Therefore the Transitional SST model is employed for all the
computations presented in this paper.
B. Computational Simulation
The simulation begins by implementing a first-order scheme
for the temporal and spatial discretization in order to ensure
the stability of the simulations. The time step is sufficiently
small, of the order of 105 , where the turbine completes one
full revolution (360 ) in 5000 time steps. After a few revolutions, higher discretization schemes are employed, namely
second-order schemes, in order to improve the accuracy of the
predicted turbine performance. Therefore, the turbine convergence is achieved in about 610 turbine rotations depending on
the operating tip speed ratio and it takes about 4050 iterations
per time step. It must be highlighted that the operating tip speed
ratio is varied by changing the rotational speed of the turbine
since the undisturbed wind is fixed at 10 m/s.
The convergence assessment is an essential step in the decision of achieving converged numerical computations. In the
context of VAWTs, the convergence is usually considered to
be achieved when the periodic torque, or instantaneous power

ALMOHAMMADI et al.: 2-D-CFD ANALYSIS OF THE EFFECT OF TRAILING EDGE SHAPE ON THE PERFORMANCE OF SB-VAWT

coefficient, for the turbine is observed. This occurs when the


instantaneous torque profile achieves a quasi-steady state, i.e.,
where at each time step the flow field is observed to be visually the same as that at the previous time the blade was at
that location and the averaged torque does not change by more
than 2%. However, this may not be sufficient, and therefore a
more sensitive method for examining the torque convergence is
necessary.
A recent method in assessing the convergence [38], namely
the rolling average convergence (RAC), is employed in all the
analyses in this paper. Instead of calculating the instantaneous
power coefficient at each time step and creating the periodic
power coefficient, the average power coefficient is computed
at each time step. The computed averaged power coefficient at
each time step is compared to the rolling averaged power coefficient computed directly from the periodic power coefficient
and when this does not change by more than 2%, the solution is
considered to be converged. Applying this method ensures that
the periodic power coefficient does not vary by more than 2%
in a minimum of two completed revolutions, i.e.,
RAC =

Rolling avraged power coefficient at each time step


 2%.
Avraged power coefficient per one revolution
(1)

Comparing the rolling average power coefficient to the averaged power coefficient per one revolution is considered to be an
efficient way of deciding when to stop the simulation and consider that the convergence of the torque is obtained. However,
the averaged power coefficient may oscillate before the solution
is completely converged, and this may result in a misleading convergence. Therefore the rolling averaged convergence
is employed in the convergence assessment in this paper.
IV. F LOW V ISUALIZATION AND A NALYSIS
Three positions have been selected to visualize and analyze
the flow around the airfoils, namely 0 , 120 , and 240 , because
the main flow features are present at these azimuthal positions.
The azimuthal angle of 0 is located at the 12 oclock position and the turbine rotates counter clockwise. At an azimuthal
angle of 0 , the flow is mainly attached and the separation may
only be present due to the transition of the flow. At an azimuthal
angle of 120 , the flow is under dynamic stall conditions where
a strong separation and vortex shedding of the boundary layer
in experienced. At an azimuthal angle of 240 , the flow is
recovering the complexity in the dynamic stall regime and the
boundary layer separation and shedding is in the process of
settling down. It is essential to visualize and analyze the role
that the trailing edge contributes in these flow regimes for SBVAWTs since the dynamic stall process is initiated within the
boundary layer and this may be affected by the trailing edge
shape.
The visualization of the flow in the vicinity of the airfoils is
presented in Fig. 4 which compares the pressure contours for
the sharp, rounded S-blunt and R-blunt trailing edge shape at
azimuthal angles of 0 , 120 , and 240 . At an azimuthal angle
0 , the pressure contours are similar but with a slight difference in the rounded trailing edge profile. At this position, the

231

Fig. 4. Total pressure contours for the sharp, rounded S-blunt and R-blunt
trailing edge shape of the NACA 0015 at TSR 1.75 for an SB-VAWT.

flow near the airfoil is attached to the airfoil, and the adverse
pressure and the laminar separation bubbles do not significantly
affect the flow. However, as the turbine blade rotates during the
turbine operation, the flow becomes relatively more complex.
At the azimuthal angles of 120 and 240 , the flow is no longer
similar due to the effect of the trailing edge which changes
the adverse pressure progression toward the leading edge. It is
clear from Fig. 4 that the leading edge vortex at an azimuthal
angle of 120 is affected by the trailing edge shape, and at an
azimuthal angle of 240 this effect is relatively lower but it is
still observed. Therefore, it is important to account for the trailing edge profile in the design process of the full turbine model
instead of optimizing a single rotating airfoil.
However, the difference in the development of the flow
appears more clearly in the vorticity contours in the vicinity
of the airfoils, as shown in Fig. 5. At an azimuthal angle of 0 ,
the vorticity contours are similar. Therefore, at this azimuthal
angle, there is no effect of the trailing edge profile on the flow
around the blade. However, the vortex location and strength
over the airfoil at an azimuthal angle of 120 are clearly not
similar and also the shedding of the vortex into the wake is
not similar. This suggests that the trailing edge profile plays
an important role on the development of the flow near the turbine airfoils as the airfoil is in the dynamic stall regime. This
affects the blade aerodynamics and the wake which interacts
with the downstream blade during the turbine operation. Also,
the flow around the turbine airfoil at an azimuthal angle of
240 is not similar. However, it could be claimed that this is
caused by the flow interaction with the upstream airfoil on the
airfoil at this azimuthal angle. Despite this, it is evident from
the visualization of the flow at an azimuthal angle of 120 that
the flow around the airfoil is significantly affected due to the
trailing edge profile as the complexity of the flow increases.
Therefore, the flow around the airfoil at an azimuthal angle of
240 is partially affected by the trailing edge as well as the
vortex convected from the upstream airfoil.

232

Fig. 5. Vorticity around the airfoils with the sharp, rounded S-blunt and R-blunt
trailing edge shape of NACA 0015 at TSR 1.75 for an SB-VAWT.

Fig. 6. Instantaneous lift (Cl) and drag (Cd) coefficients for the sharp, rounded
S-blunt and R-blunt trailing edge shape verses angle of attack of NACA 0015
for an SB-VAWT.

It is clear from the visualization of the flow that the trailing


edge profile plays an important role in the flow development
during the turbine operation.
The effect of the trailing edge may be clearly seen in Fig. 6,
which compares the instantaneous lift and drag coefficients for
the four investigated trailing edges. Further, two tip speed ratios
have been selected for the comparison and these are 1.75 (peak
of the experimental data) and 2.5 (peak of the numerical computations). The instantaneous lift coefficient at the tip speed ratio
of 1.75 does not change until the airfoil approaches the dynamic
stall regime at an angle of attack higher than about 30 . Beyond
this value of the angle of attack, the effect of the trailing edge
is clearly seen. It is observed that at this tip speed ratio, namely
1.75, the downstream instantaneous lift coefficient is significantly higher than for the sharp, S-blunt and R-blunt trailing
edges. Also, the instantaneous drag coefficient for the rounded
trailing edge is less than for the sharp, S-blunt and R-blunt
trailing edges.

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 6, NO. 1, JANUARY 2015

Fig. 7. Torque profile at TSR 05-4 for the sharp, rounded S-blunt and R-blunt
trailing edge shape of NACA 0015 for an SB-VAWT.

At the tip speed ratio of 2.5, the airfoil produced a considerably more instantaneous lift coefficient at the peak at an angle
of attack of about 15 compared to the sharp, S-blunt and
R-blunt trailing edges. However, a slight increase in the drag
may be observed on the downstream side of the turbine when
the rounded trailing edge profile is employed. This does not
significantly affect the turbine performance and this is compensated by the increased lift on the upstream side of the turbine.
However, it must be highlighted, from this comparison, that
thick trailing edges are more desirable for increasing the lift
generation for SB-VAWTs. However, a superior lift generation
may be obtained if a thick trailing edge is rounded.
However, it is necessary to investigate the effect of the trailing edge profile on the torque generated by the turbine as the
turbine airfoils rotate at several tip speed ratios. The converged
torque profiles during one turbine rotation at several tip speed
ratios are compared in Fig. 7.
At tip speed ratios 0.5 and 1, the torque profiles for all the
investigated trailing edge profiles are similar and this is mainly
due to the slow rotation of the turbine which results in a small
disturbance to the flow and therefore the flow is gradually developed. However, the effect of the trailing edge at low tip speed
ratios is not clearly observed despite it being present. However,
between the tip speed ratios 1 and 1.75, the thickness of the
trailing edge profile plays an important role as the thicker airfoil produces a higher torque. It should be highlighted that in
this range of values of the tip speed ratios, the blunt airfoil contribution to the torque profile is minor and this agrees with the
findings discussed in the literature [13], [14]. As the tip speed
ratio increases to values greater than 2, the effect of the trailing
edge thickness becomes insignificant and the rounded trailing
edge becomes more efficient in producing torque. The rounded
and blunt shape of the trailing edge induces two vortices behind
the trailing edge and this reduces the adverse pressure and
results in an attached and more stable boundary layer, and consequently higher torque. Due to the sharp corners of the blunt
trailing edge, this advantage vanishes as the rotational speed
increases because the corner of the blunt trailing edge acts as

ALMOHAMMADI et al.: 2-D-CFD ANALYSIS OF THE EFFECT OF TRAILING EDGE SHAPE ON THE PERFORMANCE OF SB-VAWT

233

It is clear that the trailing edge profile significantly affects the


SB-VAWT performance.
For tip speed ratios less than 1.75, the increase in the thickness of the trailing edge significantly increases the turbine
performance. However, the sharp and rounded edge do not
significantly contribute to the turbine performance despite the
sharp trailing edge being more favorable compared to the blunt
trailing edge.
For tip speed ratios greater than 1.75, the rounded trailing
edge, as well as the trailing edge thickness, becomes both significant in improving the SB-VAWT. It should be noted that at
high tip speed ratios, the sharp and blunt trailing edges do not
play a significant role in improving the turbine performance.
Therefore, thick and rounded trailing edges are desirable for
SB-VAWTs.
Fig. 8. Power curve at TSR 05-4 for the sharp, rounded S-blunt and R-blunt
trailing edge shape of NACA 0015 for a SB-VAWT.

a sharp trailing edge. The rounded shape of the trailing edge


ensures that the induced vortices move around the trailing edge
in order to efficiently stabilize the flow development in the
boundary layer.

V. SB-VAWT P ERFORMANCE
The effect of the trailing edge profile and thickness is compared visually and on the torque profiles produced at several tip
speed ratios. However, it is important to investigate the effect
of the trailing edge profile on the SB-VAWT performance at
several tip speed ratios. Fig. 8 compares the investigated turbine performance for the sharp, rounded, S-blunt and R-blunt
trailing edge shape at several tip speed ratios. It has been
consistently reported in the literature that a satisfactory mesh
independent solution for 2-D URANS of an SB-VAWT may
considerably over predict the experimental power coefficient
at high tip speed ratios [41][43]. This is mainly due to the
omission of the 3-D effects and the tip losses which occur in
reality and these are not accounted for in the 2-D computations.
However, the over prediction of the power coefficient presented
in this paper is in the order of the over predictions reported
in the literature [41][43]. Therefore, the present predictions
may be considered to be accurate within the 2-D assumption.
Further, it should be noted that in order to obtain satisfactory
mesh independent solutions in 3-D for such turbine designs and
operating conditions is not as yet possible. Therefore, in order
to suggest further design improvements of the performance of
the SB-VAWTs, then this is usually achieved by using 2-D
computations either by analyzing a single stationary 2-D airfoil [44], [45] or three rotating 2-D airfoils as in the real case
of a working turbine [46]. Further, it is essential to consider
improving the turbine performance on the full model, namely
3 rotating airfoils, in order to account for all the interactions of
the flow around the airfoils. Of course, these improvements can
only be qualitative and not quantitative. Thus, the aim of this
paper is to suggest modifications in the design of the trailing
edge of an operating SB-VAWT in order to improve the power
generated by employing 2-D URANS.

VI. C ONCLUSION
In this paper, four trailing edge profiles have been investigated, namely the sharp, rounded, S-blunt and R-blunt trailing
edge shapes, at several tip speed ratios. The analyses are performed by employing the URANS equation combined with the
Transition SST model in order to account for the flow transition. The analyses are performed on a full turbine model and the
convergence is assessed by employing the rolling convergence
average criteria.
It has been found that the trailing edge profile plays an important role on improving the performance of the SB-VAWT. At
relatively low tip speed ratios, the thickness of the airfoil may
significantly improves the turbine performance. However, at
high tip speed ratios, namely higher than 1.75, the rounded
trailing edge becomes significantly important in improving the
turbine performance. This is evident in the instantaneous lift
and drag coefficients.
These findings are essential in designing the SB-VAWT and
should be considered in the optimization process of the SBVAWT. Implementing the rounded and thick trailing edge may
be attractive in improving the turbine performance for the
turbines that are already in operation.
ACKNOWLEDGMENT
K. M. Almohammadi would like to express his gratitude to
Taibah University, Kingdom of Saudi Arabia for supporting him
to perform his Ph.D. study at the University of Leeds.
R EFERENCES
[1] T. Burton et al., Wind Energy Handbook, 2nd ed. Hoboken, NJ, USA:
Wiley, 2001, p. 642.
[2] I. Paraschivoiu, Wind Turbine Design: With Emphasis on Darrieus
Concept. Montral, Canada: Polytechnic International Press, 2002.
[3] S. Eriksson, H. Bernhoff, and M. Leijon, Evaluation of different turbine
concepts for wind power, Renew. Sustain. Energy Rev., vol. 12, no. 5,
pp. 14191434, 2008.
[4] K. Bergey, The Lanchester-Betz limit, J. Energy, vol. 3, pp. 382384,
1979.
[5] S. Mathew, Wind Energy: Fundamentals, Resource Analysis and
Economics. New York, NY, USA: Springer, 2006.
[6] M. C. Claessens, The design and testing of airfoils for application in
small vertical axis wind turbines, M.S. thesis, Faculty of Aerospace
Engineering, Delft Univ. Technology, Delft, The Netherlands, 2006.

234

[7] M. Islam, A. Fartaj, and R. Carriveau, Analysis of the design parameters


related to a fixed-pitch straight-bladed vertical axis wind turbine, Wind
Eng., vol. 32, no. 5, pp. 491507, 2008.
[8] M. Islam, D. Ting, and A. Fartaj, Design of a special-purpose airfoil
for smaller-capacity straight-bladed VAWT, Wind Eng., vol. 31, no. 6,
pp. 401424, 2007.
[9] F. Saeed et al., Inverse airfoil design method for low-speed straightbladed Darrieus-type VAWT applications, in Proc. 7th World Wind
Energy Conf., Ontario, Canada, 2008, pp. 111.
[10] M. Islam, D. Ting, and A. Fartaj, Aerodynamic models for Darrieus-type
straight-bladed vertical axis wind turbines, Renew. Sustain. Energy Rev.,
vol. 12, no. 4, pp. 10871109, 2008.
[11] A. H. A. Hamid et al., Comparative analysis of straight-bladed and
curved-bladed vertical axis wind turbine, Adv. Math. Comput. Methods,
vol. 2, no. 1, pp. 18, 2012.
[12] J. L. Summers and W. A. Page, Lift and Moment Characteristics
at Subsonic Mach Numbers of Four 10-Percent-Thick Airfoil Sections
of Varying Trailing-Edge Thickness. Washington, DC, USA: National
Advisory Committee for Aeronautics, 1950.
[13] H. A. Smith and R. F. Schaefer, Aerodynamic Characteristics at Reynolds
Numbers of 3.0 x 10 (6) and 6.0 x 10 (6) of Three Airfoil Sections Formed
by Cutting Off Various Amounts From the Rear Portion of the NACA
0012 Airfoil Section. Langley, VA, USA: National Aeronautics and Space
Administration, 1950.
[14] A. Gmez and . Pinilla, Aerodynamic characteristics of airfoils with
blunt trailing edge, Rev. Ing., vol. 24, pp. 2333, 2006.
[15] L. J. Herrig, J. C. Emery, and J. R. Erwin, Effect of Section Thickness and
Trailing-Edge Radius on the Performance of NACA 65-Series Compressor
Blades in Cascade at Low Speeds. Langley Field, VA, USA: National
Advisory Committee for Aeronautics, 1951, vol. 29.
[16] C. D. Harris, Wind-Tunnel Investigation of Effects of Trailing-Edge
Geometry on a NASA Supercritical Airfoil Section. Washington, DC,
USA: National Aeronautics and Space Administration, Langley Research
Center Langley Station, 1971.
[17] C. J. Doolan, Numerical simulation of a blunt airfoil wake using a twodimensional URANS approach, in Proc. 16th Australas. Fluid Mech.
Conf., 2007, pp. 10951101.
[18] J. P. Baker, E. A. Mayda, and C. P. Van Dam, Experimental analysis
of thick blunt trailing-edge wind turbine airfoils, J. Solar Energy Eng.,
vol. 128, no. 4, p. 422, 2006.
[19] K. J. Standish and C. P. Van Dam, Aerodynamic analysis of blunt trailing edge airfoils, J. Solar Energy Eng., vol. 125, no. 4, pp. 479487,
2003.
[20] S. Wang et al., Numerical investigations on dynamic stall of low
Reynolds number flow around oscillating airfoils, Comput. Fluids,
vol. 39, no. 9, pp. 15291541, 2010.
[21] L. A. Danao, J. Edwards, O. Eboibi, and R. Howell, The performance of
a vertical axis wind turbine in fluctuating windA numerical study, in
Proc. World Congr. Eng., vol. 3, 2013, pp. 35.
[22] R. J. Templin, Aerodynamic Performance Theory for the NRC VerticalAxis Wind Turbine. Ottawa, ON, Canada: National Aeronautical
Establishment, 1974.
[23] I. Paraschivoiu, Double-multiple streamtube model for studying verticalaxis wind turbines, J. Propul. Power, vol. 4, no. 4, pp. 370377,
1988.
[24] I. Paraschivoiu and F. Delclaux, Double multiple streamtube model with
recent improvements, J. Energy, vol. 7, no. 3, pp. 250255, 1983.
[25] H. C. Larsen, Summary of a vortex theory of the Cyclogiro, in Proc.
2nd U.S. Natl. conf. Wind Eng. Res., Jun. 1975, pp. V-8-1-3.
[26] J. B. Fanucci and R. E. Walters, Innovative wind machines: The theoretical performance of a vertical axis wind turbine, in Proc. Vertical
Axis Wind Turbine Technol. Workshop, Sandia Laboratory Report SAND
76-5586, May 1976, pp. III-61III-93.
[27] O. Holme, A contribution to the aerodynamic theory of the vertical-axis
wind turbine, in Proc. Int. Symp. Wind Energy Syst., Cambridge, U.K.,
Sep. 79, vol. 1, 1977, p. 4.
[28] J. H. Strickland, B. T. Webster, and T. Nguyen, Vortex Model of the
Darrieus Turbine: An Analytical and Experimental Study. Lubbock, TX,
USA: National Aeronautics and Space Administration, Texas Tech Univ.,
1980.
[29] J. L. Cardona, Flow curvature and dynamic stall simulated with an
aerodynamic free-vortex model for VAWT, Wind Eng., vol. 8, no. 3,
pp. 135143, 1984.
[30] H. Hirsch and A. C. Mandal, A cascade theory for the aerodynamic
performance of Darrieus wind turbines, Wind Eng., vol. 11, no. 3,
pp. 164175, 1987.

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 6, NO. 1, JANUARY 2015

[31] J. G. Leishman, Dynamic stall experiments on the NACA 23012 aerofoil, Exp. Fluids, vol. 9, no. 1, pp. 4958, 1990.
[32] K. McLaren, S. Tullis, and S. Ziada, Computational fluid dynamics simulation of the aerodynamics of a high solidity, small-scale vertical axis
wind turbine, Wind Energy, vol. 15, no. 3, pp. 349361, 2011.
[33] R. Bravo, S. Tullis, and S. Ziada, Performance testing of a small verticalaxis wind turbine, in Proc. CANCAM, 2007, pp. 470471.
[34] S. Kooiman and S. Tullis, Response of a vertical axis wind turbine to
time varying wind conditions found within the urban environment, Wind
Eng., vol. 34, no. 4, pp. 389401, 2010.
[35] K. W. McLaren, A numerical and experimental study of unsteady
loading of high solidity vertical axis wind turbines, Ph.D. dissertation, Mechanical Engineering, McMaster Univ., Hamilton, ON, Canada,
2011.
[36] K. M. Almohammadi et al., CFD sensitivity analysis of a straight-blade
vertical axis wind turbine, Wind Eng., vol. 36, no. 5, pp. 571588, 2012.
[37] K. M. Almohammadi, D. Ingham, L. Ma, and M. Pourkashanian, CFD
modelling investigation of a straight-blade vertical axis wind turbine,
presented at the 13th Int. Conf. Wind Eng., Amsterdam, Netherland,
2011, pp. 18.
[38] K. Almohammadi et al., Computational fluid dynamics (CFD) mesh
independency techniques for a straight blade vertical axis wind turbine,
Energy, vol. 58, pp. 483493, 2013.
[39] E. A. Mayda and C. P. Van Dam, Bubble-induced unsteadiness on a
wind turbine airfoil, J. Solar Energy Eng., vol. 124, no. 4, pp. 335345,
2002.
[40] U. B. Mehta and Z. Lavan, Starting vortex, separation bubbles and
stallA numerical study of laminar unsteady flow around an airfoil,
J. Fluid Mech., vol. 67, no. 2, pp. 227256, 1975.
[41] C. Li et al., 2.5 D large eddy simulation of vertical axis wind turbine
in consideration of high angle of attack flow, Renew. Energy, vol. 51,
pp. 317330, 2013.
[42] M. R. Castelli, A. Englaro, and E. Benini, The Darrieus wind turbine: Proposal for a new performance prediction model based on CFD,
Energy, vol. 36, no. 8, pp. 49194934, 2011.
[43] R. Howell et al., Wind tunnel and numerical study of a small vertical
axis wind turbine, Renew. Energy, vol. 35, no. 2, pp. 412422, 2010.
[44] M. Mirzaei, S. N. Hosseini, and J. Roshanian, Single and multi-point
optimization of an airfoil using gradient method, Aircr. Eng. Aerosp.
Technol., vol. 79, no. 6, pp. 611620, 2007.
[45] J. Y. Li et al., Aerodynamic optimization of wind turbine airfoils using
response surface techniques, Proc. Inst. Mech. Eng. A. J. Power Energy,
vol. 224, no. 6, 2010, pp. 827838.
[46] T. J. Carrigan et al., Aerodynamic shape optimization of a verticalaxis wind turbine using differential evolution, ISRN Renew. Energy,
vol. 2012, p. 16, 2012.

Khaled M. Almohammadi received the B.Sc.


degree in mechanical engineering from King Fahd
University of Petroleum and Minerals (KFUPM),
Dhahran, Saudi Arabia, in 2004 and the M.Sc. degree
in computational fluid dynamics from the University
of Leeds, Leeds, U.K., in 2009.
He is currently an Assistant Professor in the
Mechanical Engineering Department, Taibah
University, Madina, KSA. From 2004 to 2006,
he was with the SWCC power plant and worked
in power generation and water desalination. His
research interests include modeling and optimizing straight blade vertical axis
wind turbines.

Derek B. Ingham, Professor of Applied


Mathematics, has worked for over 40 years on
a wide variety of engineering and industrial mathematical problems. He is on the Editorial Boards of
numerous research journals and has been the Director
of the Energy Technology and Innovation Initiative
(ETII) and CFD Centre, Faculty of Engineering,
University of Leeds, Leeds, U.K. He has written 16
books, over 750 research papers in referred journals,
and over 40 confidential industrial reports. He has
received funding from 49 different organizations and
in excess of 6 million pounds.

ALMOHAMMADI et al.: 2-D-CFD ANALYSIS OF THE EFFECT OF TRAILING EDGE SHAPE ON THE PERFORMANCE OF SB-VAWT

Lin Ma is an Associate Professor with the Energy


Technology and Innovation Initiative (ETII) and CFD
Centre, Faculty of Engineering of the University of
Leeds, Leeds, U.K. He has supervised 14 Ph.D. students, published over 150 research papers, and has
contributed to 6 book chapters. His research interests include wind energy, coal/biomass combustion
for power generation, and carbon capture and sequestration technologies.

235

Mohamed Pourkashanian is the Head of the


Leeds University Energy and Technology Innovation
Initiative (ETII) and the Director of the Pilot-Scale
Advanced Capture Technology (PACT) national
facilities. He holds a Chair in High Temperature
Combustion Technology and has completed numerous major research projects on clean energy technology. He has published over 400 refereed research
papers and has coauthored books on coal combustion. He played a leading role in developing
the NOx postprocessing computer codes and subsequently soot/NOx models that were later employed in the commercial CFD
software. He is a member of numerous international and national scientific bodies, including a Member of EERA Implementation Plan 20132015
(contribution to CCS-EII Team, SET-PLAN), a Member of the Coordinating
Group of UKCCSRC, an Invited Member of the All Party Parliamentary
Renewable Transport Fuels Group, Member of Technical Working Group for
the Department of Energy and Climate Change (CCS Roadmap UK2050), and
Expert-Member in EU-GCC Clean Gas Energy Network.
Dr. Pourkashanian has received a substantial number of grants from RCUKEPSRC, the EU, NATO, and industry.

You might also like