J. Wind Eng. Ind. Aerodyn.

153 (2016) 3445

Contents lists available at ScienceDirect

Journal of Wind Engineering

and Industrial Aerodynamics
journal homepage: www.elsevier.com/locate/jweia

A new approach for optimization of Vertical Axis Wind Turbines

M. Abdul Akbar n, V. Mustafa
Department of Civil Engineering, National Institute of Technology Calicut, NIT Campus (P.O), Calicut, Kerala 673601, India

art ic l e i nf o a b s t r a c t

Article history: A new approach has been developed for the optimization of Vertical Axis Wind Turbines (VAWTs). The
Received 16 May 2015 approach is derived from double multiple streamtube theory (DMST) and adopts the concept of
Received in revised form Representative streamtube wherein the entire aerodynamic property of the VAWT is assumed to be
14 March 2016
represented by a single streamtube occupying a specic azimuthal location. Five input parameters
Accepted 14 March 2016
Available online 31 March 2016
namely; power, wind velocity, aspect ratio, air viscosity and air density and six output parameters that are
minimally required for the construction of a straight bladed, constant pitch VAWT are considered in the
Keywords: study. The discrete data of lift and drag coefcients pertaining to specic values of Reynolds number and
Vertical Axis Wind Turbine angle of attack are arranged in the order of decreasing lift to drag ratio and starting with the rst set of
Analysis
coefcients, a check is made if it is eligible to become Representative streamtube for the particular
Renewable energy
problem. The check is done through a tri-directional Demand Factor test that seeks the compatibility of
Double multiple streamtube theory
Optimization Reynolds number and angle of attack values of the aerofoil data under consideration. The rst set which
gains the eligibility to become Representative streamtube marks the solution for the particular problem.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction methods developed (Maheri et al., 2006) to aid faster convergence

A good summary of the models developed for analysis of Ver-
tical Axis Wind Turbines (VAWTs) has been presented in (Islam
et al., 2008). In general, the computational models have been
classied under three categories: Momentum model, vortex model
and cascade model. Among these models, momentum model is the
most prominent one and the one that is being researched the
most. Momentum model was rst introduced as single streamtube
theory which evolved into multiple streamtube theory which
further evolved into double multiple streamtube theory (DMST)
with increased accuracy achieved on every revision. Fig. 1 depicts
the progression of streamtube theory with each set of parallel
lines denoting a streamtube. In this gure: V u , V u 1; V u 2 and V u
3 are the upstream wind velocities in the streamtubes that are a
modication of the external wind velocity effected through the
actuator disk effect of the wind hitting the rotating VAWT. Here,
V d 1; V d 2 and V d 3 are the downstream wind velocities in the
streamtubes.
Double multiple streamtube theory (DMST) is being attributed
to Paraschivoiu who devised the theory as early as 1981 (Para-
schivoiu, 1981). The theory has thus been widely used (Homicz,
1991; Tang et al., 2011; Beri and Yao, 2011). The DMST has also
been validated experimentally (Bogeanu et al., 2010) and
n
Corresponding author.
E-mail address: maakbar83@yahoo.co.in (M. Abdul Akbar).
of DMST iterations has improved it's acceptability among
researchers. The versatility of the method has enabled its use as a
primary analysis tool to lead in to analyzing secondary effects like
the gyroscopic effect (Blusseau and Patel, 2012; Borg et al., 2013).
Streamtube theories are the most convenient method for
optimization exercises. Re-running the analysis after changing the
parameters are the quickest with regards to computational effort
and time. On the other hand, rigorous analysis such as Computa-
tional uid dynamics (CFD) involves, extra effort in re-runs
including changing the parameter for the new model, regenerat-
ing the mesh around the modied region, further renement to
meet error tolerance, etc. Also, such methods demand signicant
computational time. Table 1 outlines some of the optimization
studies carried out in the past using streamtube theories. In this
table, SST stands for single streamtube theory and MST stands
for multiple streamtube theory (the theories depicted in Fig. 1).
Despite being the simplest of methods available for the optimi-
zation of VAWTs, the formulation of streamtube theories (as will be
discussed in Section 2) is complex enough that a formula based
numerical optimization process is difcult through it. This is because,
the variables involved in analysis are so inter-connected to each other
that the seperation of their effects is difcult. It is for this reason that
most of the optimization studies using streamtube theories are based
on the traditional approach of incrementing certain analysis con-
trolling parameters and re-running the analysis for each case

http://dx.doi.org/10.1016/j.jweia.2016.03.006
0167-6105/& 2016 Elsevier Ltd. All rights reserved.
 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445 35

Nomenclature DFA aero-Induction demand factor

DFC compatibility demand factor
List of symbols DFC-L lower limit of compatibility demand factor
DFC-U upper limit of compatibility demand factor
angle of attack DFH hybrid demand factor
tip speed ratio L length of the blade
L lower bound of tip speed ratio LD the lift to drag ratio at which computation is being
opt the optimum value of tip speed ratio at which com- carried out
putation are performed LDL lower bound of optimum lift to drag ratio
U upper bound of tip speed ratio LDU upper bound of optimum lift to drag ratio
viscosity of wind N number of blades
angular velocity Ns number of streamtubes (including upstream and
density of air downstream streamtubes)
azimuthal location P power requirement (rated power)
a induction factor R radius of the rotor
AR aspect ratio Re Reynolds number
C chord length Sol solidity
CD drag coefcient Va axial ow velocity
CL lift coefcient Vd(1), Vd(2), Vd(3) downstream wind velocities in the double
Cn force coefcient along the normal direction multiple streamtube model
CP coefcient of power Ve wind velocity for the downstream half
CPL lower bound of optimum coefcient of power V1 external wind velocity
CPmax optimum coefcient of power at which calculation is VR relative velocity
being performed Vu upstream wind velocity in the single
CPU upper bound of optimum coefcient of power streamtube model
CQ torque coefcient Vu(1), Vu(2), Vu(3) upstream wind velocities in the multiple and
Ct force coefcient along the tangential direction double multiple streamtube model
CT coefcient of thrust

seperately and determining the optimum combination based on the
results of all permutations and combinations.
The input and output parameters in the optimization of VAWT
that are considered in this study are listed in Table 2. The input
parameters listed in the table covers most of the factors that are
typically considered in VAWT design and the output parameters
include all the parameters that are minimally required to com-
pletely manufacture a simple and straight bladed VAWT. Extra
parameters may be required to dene VAWTs with pitched blade
and curved shapes, which are not considered in this study. The
aerofoil shapes are classied into various families like NACA, Selig,
Eppler, Selig/Donovan, Miley, Wortmann, Clark and Gottingen.
There are also stand-alone types like Jacobs USNPS4, Bergey BW-3,
etc. The output parameters listed in the table are depicted in Fig. 2.
The optimization process developed through this paper adopts a
simplistic model of maximization of lift to drag ratio which is a direct
measure of increase in the efciency of VAWT. The methodology uses
the concept of effective lift to drag ratio discussed in (Soraghan et al.,
2013). Through the effective lift to drag ratio, the entire aerodynamic
performance of the VAWT is represented by the azimuthal location
of the streamtube that provide the value of torque equal to average
instantaneous torque for the VAWT. The ratio of lift and drag coef-
cient corresponding to that particular streamtube is called effective
lift to drag ratio. For convenience, the azimuthal location that pro-
vides the effective lift to drag ratio is referred to as Representative
streamtube in this paper. The optimization of the VAWT on the basis
of Representative streamtube provides a great advantage over the
regular approach that it obviates the need of carrying out iterations
for all azimuthal locations.
2. Basic formulations
Consider the plan view of a 3-bladed VAWT as shown in Fig. 3.
Only parameters that are relevant to the discussion in this section
are depicted. Additional features such as connecting struts, sub-
structure and machine components are not shown for clarity.
During operation, the VAWT is subjected to two types of forces
simultaneously. They are:
1) Wind force from an arbitrary direction.
2) The angular rotation of VAWT about its central pole.
Due to the omni-directionality of VAWT rotor, regardless of the
wind direction, a symmetric face is always presented to receive
the wind without the need for a yawing mechanism which is a
feature in Horizontal axis wind turbines (HAWTs). Consider a
close-up view of the VAWT geometry shown in Fig. 4. The net
effect of two forces acting together can be obtained by considering
the resultant of the forces. The angle at which the resultant force
hits the blade will also be modied due to the combined action of
the forces. Fig. 5 shows a scaled up image of the wind hitting the
rotating VAWT wherein, the actual wind velocity, V 1 gets mod-
ied to V a due to the actuator disk effect which further gets
modied to V R due to the rotating effect of the VAWT.
From basic trigonometry using Fig. 5, the angle at which the wind
hits the rotating VAWT denoted as angle of attack () is given by,
sin V a sin =V R 1
Here, V a is the modied wind velocity due to the actuator disk
effect of the wind hitting the rotating VAWT. Hence, out of the
actual wind velocity, V 1 only a portion of the same would be
effective in generating wind energy. The relationship between
these velocities is denoted by a dimensionless factor termed,
induction factor (a) given by,
a 1  V a =V 1 2
36 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445

Fig. 1. Evolution of stream tube theory: (a) Single streamtube theory, (b) Multiple streamtube theory, and (c) Double multiple streamtube theory.

Table 1
Optimization studies carried out on VAWTs using streamtube theories.

Variables considered/remarks Theory Reference

Aspect ratio, solidity, Reynolds number of operation, tip speed ratio (TSR) MST Sullivan, 1979
Optimal performance analysis in the absence of blade drag Newman, 1983
Rotor efciency, solidity, interference coefcients Loth and McCoy, 1983
Optimum value of inow factor Newman, 1986
Symmetric NACA aerofoils DMST Solum et al., 2006
Camber and aerofoil thickness DMST Claessens, 2006
Optimal positioning of the blades to maximize efciency of VGOT (variable geometry oval trajectory) DMST Ponta et al., 2007; Ponta and Lago, 2008
Optimum TSR and number of blades for a Zephyr VAWT MST Ajedegba, 2008
Discusses optimal features with regards to control parameters DMST Farthing, 2009
Optimization model for Darrieus Straight bladed VAWT based on analysis conducted using DMST DMST Ming, 2009
Solidity, TSR, blade pitch angle DMST Asher et al., 2010
Rotor area, strut design, solidity, TSR, number of blades, blade thickness, point of attachment, pitch angle, blade DMST Deglaire, 2010
camber
Blade prole, rotor solidity, Reynolds number, aspect ratio MST Samaraweera et al., 2010
Twisting of blades of an innovative VAWT DMST Sharpe and Proven, 2010
Optimization of pitch angle DMST Okpue, 2011
Optimum positioning of blades, optimum solidity, angle of attack and blade pitch of VGOT DMST Ponta et al., 2011
Aerofoil shape, TSR, radius, chord length, number of blades DMST Castillo, 2011
Camber and pitching angle of wing prole MST Bitsch et al., 2011
Optimization tool for new rotor architectures DMST Castelli et al., 2011
Radius of the rotor, number of blades, chord length, blade height DMST Tang et al., 2011
Tip speed ratio DMST Helms, 2011
Tip speed ratio MST McGowan et al., 2012
Shape optimization of a novel offshore VAWT concept called NOVA DMST Shires, 2013a
Pitch angle DMST Rathi, 2012
Optimization of initial angle of individual blades for a four bladed VAWT SST Diaz and Pinto, 2012
Optimize overall dimensions Islam et al., 2012
Optimal placement in wind farms SST Chen and Agarwal, 2012
Optimal values of Masse coefcient used in calculating dynamic stall based aerodynamic coefcients DMST Shires, 2013b
Solidity, aspect ratio MST Komerath et al., 2013
Effect of variation of ns on a savonius rotor to get the most optimized conguration SST Hasan et al., 2013
Chord optimization along the blade span DMST Bedon et al., 2014
Aspect ratio, optimal strut spacing DMST Clarke et al., 2014
Rotor solidity factor, blade aspect ratio, rotor aspect ratio DMST Ahmadi-Baloutaki et al., 2014
Chord to diameter ratio, height to diameter ratio, solidity, aspect ratio, height, diameter, chord length DMST Bianchini et al., 2015
Aerofoil shapes, number of blades, chord length DMST Bedon et al., 2015
Aerofoil thickness DMST Larrea, 2015
Pitch angle DMST Wille, 2015

From Pythagoras theorem, the relative velocity (V R ) can be the rotor. Substituting Eq. (3) in (1) and using the expressions of
written from Fig. 5 as, Eq. (2), we get Eq. (4) as,
q q
 2  2  2  2
VR V a sin V a cos R 3 sin 1  a sin = 1  a sin 1 a cos 4

Here, is the angular velocity of the VAWT which varies Here, is the azimuthal location or the instantaneous position
instantly with the change in wind velocity, and R is the radius of of the streamtube under consideration. Similarly, expressing angle
 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445 37

Table 2
Input and output parameters in optimization process.

Input parameters Output parameters

Power (P) Number of blades (N)

Wind velocity (V1) Aerofoil shape
Viscosity of the wind () Chord length (C)
Density of air () Length of blades (L)
Aspect ratio (AR) Radius of rotor (R)
Angular velocity at rated wind speed ()

Fig. 4. Free body diagram of blade section.

Fig. 2. Output parameters of VAWT as per Table 2.

Fig. 3. Schematic plan of VAWT depicting aerodynamics.

of attack in terms of cosine of the angle, we get Fig. 5. Wind hitting the VAWT.
q
cos 1  a cos = 1 a sin 2 1  a cos 2 5 The relative velocity ratio can be written by dividing Eq. (3) by
actual wind velocity, V 1 and using the expression of Eq. (2) as,
Dividing Eq. (4) by Eq. (5), we get a simpler expression for angle q
 2  2
of attack as, V R =V 1 1  a sin 1  a cos 8
 
tan 1  a sin = 1 a cos 6 Betz thrust coefcient, C T for an ideal wind turbine is given by,

In Eqs. (4)(6), is the Tip speed ratio (TSR) and is dened as C T 4a1  a 9

R=V 1 7 This is equated to thrust coefcient calculated aerodynamically

for each of the streamtubes which leads to expression given by
38 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445
Eq. (10).
  2    
4a1  a NC= R V R =V 1 C t cos = sin  C n 10
Eq. (10) which is an implicit equation is solved through trial
and error procedure for each of the streamtubes to determine its
induction factor. The normal coefcient C n and tangential coef-
cient C t can be calculated from lift (C L and drag coefcients
C D based on the geometry of Fig. 4 as,
C n C L cos C D sin 11
C t C L sin  C D cos 12
The torque coefcient C Q for the VAWT is obtained after
summation of the contributions of all the streamtubes (Eq. (13))
and coefcient of power C P ) is obtained after the multiplication of
torque coefcient by TSR (Eq. (14)),
"  #
  XNs VR 2
C Q NC=2R 1=N s Ct 13
i1 V1
i thereafter (Pratumnopharat and Leung, 2011). The resulting
CP CQ 14
The coefcient of power (C P ) is a direct measure of the ef-
ciency of VAWT in producing power. The total power generated is
given by Eq. (15).
P C P RLV 31 15
Reynolds number (Re ), which is necessary for the determina-
tion of the lift and drag coefcients is calculated for the stream-
tube under consideration using Eq. (16) which uses the value of
relative velocity, V R obtained through Eq. (3).
Re V R C= 16
Aspect ratio (AR) of the VAWT which is one of the input
parameters listed in Table 1 depends on the radius of VAWT and
the length of its blades and is dened as per Eq. (17).
AR R=L 17
As per the denition, it is evident that aspect ratio provides an
indication of the slenderness of the overall VAWT geometry. It is
evident from Eqs. (10) and (13) that the variables; number of
blades (N), chord length (C) and Radius (R) appears in a particular
format in relation to each other. This formation is denoted by the
term solidity (Sol) which is given by Eq. (18). As the name implies,
the variable is indicative of the amount of solid blade material that
is present within the VAWTs exterior dimensions.
Sol NC=R 18
The implementation of DMST approach takes into account the
variation of wind velocity as it passes through the upstream and
downstream halves of the rotor (as evident from Fig. 1(c)). The wind
velocity that reaches the downstream actuator disk is a diminished
value of the exterior wind velocity modied by the actuator disk of the
corresponding streamtube in the upstream half of the rotor. The
exterior wind velocity for the downstream half, V e is obtained by
equilibrium considerations and is given by Eq. (19).
V e 2V a V 19
The Eq. (1) through (19) are standard equations that are already
established. The modied forms of the above equations that are
necessary for implementation of the proposed optimization appr-
oach are discussed in the following section.
3. Formulaic modications and approach development
As per denition of Representative streamtube, the torque
attained for the azimuthal location should be equal to
instantaneous torque for the VAWT. Therefore, for Representative
streamtube, Eq. (13) can be re-written without the summation
(independent of the azimuthal location) as,
  2
C Q NC=2R V R =V 1 C t 20
The coefcient of power of Representative streamtube can be
written using Eqs. (14) and (19) as,
  2
C P NC=2R V R =V 1 C t 21
Considering Eq. (10) for Representative streamtube and divid-
ing it by Eq. (21) and re-arranging the terms using Eqs. (11) and
(12), we get
  !
2C P C L =C D cos  sin
4a1  a cos = sin    22
C L =C D sin  cos
For a 40.4, Eq. (22) can be modied using a correction to Eq.
(9) originally proposed by Glauert in 1926 and widely used
expression is given by Eq. (23).
0:889  0:444a 1:556a2
  !
2C P C L =C D cos  sin
cos = sin    23
C L =C D sin  cos
Considering Eq. (21), it can be seen that using the Eq. (18) for
Sol, and using the expression for C t presented in Eq. (12) and
further using the expression for relative velocity ratio of Eq. (8);
the entire expression of coefcient of power (C P for Representa-
tive streamtube can be expressed as,
h 2  i 2
C P Sol 1  a sin 1  a cos C L sin  C D cos =2
24
From Eq. (24), it can be seen that, coefcient of power (C P ),
which is a direct measure of the efciency of the VAWT, can be
calculated using the dimensionless variables, solidity and tip speed
ratio and other intrinsic properties which are specic to Repre-
sentative streamtube, namely, its induction factor, azimuthal
location, angle of attack and force coefcients (lift and drag coef-
cients). Apart from aerofoil shape used in the calculation of force
coefcients, the equation does not directly depend on the output
variables listed in Table 2. This distinct property of the equation is
exploited in the development of the optimization approach dis-
cussed in this paper.
Based on the concept of effective lift to drag ratio, an envel-
oping study was carried out by (Soraghan et al., 2013) on
NACA0012 aerofoils for the range of Reynolds numbers that are
typically experienced by VAWTs. Through this study, the para-
meters of the VAWT were varied and pertaining to different
incremental values of solidity and tip speed ratio, the case that led
to the maximum value of coefcient of power were retained
seperately along with the ratio of lift to drag ratio of Representa-
tive streamtube. It must be noted, with reference to Eq. (24) that
for a given set of C P , , Sol, C L and C D , there is also a unique set of ,
a and which dene the particular case. However, only the former
variables have been documented in the description of optimal
characteristics as the latter ones are internal properties which
cannot be used as optimization controlling parameters.
Through the study (Soraghan et al., 2013), the optimal solidity
for the VAWT was found to be 0.13 which was independent of
effective lift to drag ratio. The lower and upper bounds of optimal
characteristics identied were as follows:
Lower bound: - A maximum coefcient of power of 0.15 was
attained for an effective lift to drag ratio of 10 at an optimal tip
speed ratio of 3.85.
 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445
Upper bound: - A maximum coefcient of power of 0.47 was
attained for an effective lift to drag ratio of 140 at an optimal tip
speed ratio of 4.9.
Based on these values, a linear proportionality diagram was
created for calculating the optimal tip speed ratio and maximum
coefcient of power for a given effective lift to drag ratio (Fig. 6).
Here, opt optimum tip speed ratio for the given effective lift
to drag ratio (LD).
U uper bound of tip speed ratio4.9.
L lower bound of tip speed ratio 3.85.
CPmax maximum coefcient of power for the given effective lift
to drag ratio (LD).
CPU upper bound of maximum coefcient of power 0.47.
CPL lower bound of maximum coefcient of power0.15.
LD effective lift to drag ratio at which calculation is performed.
LDU upper bound of effective lift to drag ratio 140.
LDL lower bound of effective lift to drag ratio10.
The optimal value of solidity is constant throughout. For the
purposes of this paper, the intermediate optimal values of tip
speed ratio and solidity is assumed to be obtained through linear
interpolation of the effective lift to drag ratio at the specic point
at which analysis is carried out (Eqs. (25) and (26)).
opt L LD  LDL U  L =LDU  LDL 25
Tip Speed
Solidity
Ratio
0.13 (Sol)
3.85 (L)
opt
0.13 (Sol) 4.9 (U)
Fig. 6. Optimum parameters proportionality diagram.
Fig. 7. Schematic representation of set of eligible solutions.
39
 
C pmax CpL LD  LDL CpU  CpL =LDU LDL 26
The key aspect of the approach developed in this paper is to
determine the highest lift to drag ratio (along with its associated
properties) that is eligible to become Representative streamtube
for the problem under consideration. The maximization of effec-
tive lift to drag ratio is the optimization criteria as it is a direct
measure of efciency of the VAWT (Eq. (24)). Through this paper, it
is attempted to t a lift to drag ratio (along with its associated
properties) to be Representative streamtube for the given problem,
an approach that is not seen presented so far by works of
literature.
It is known that a set of lift and drag values is dependent on
three factors viz: aerofoil shape, Reynolds number and angle of
attack. The compatibility of a given lift to drag ratio for a particular
problem should satisfy the local compatibility of these three
parameters. Based on logical reasoning, it is concluded that a
factor based approach for ensuring compatibility should fulll the
following basic criteria's:
1) The factor should provide for a tri-directional testing as there
are three criteria's that needs to be compatible (for a given set of
lift and drag values) viz: the aerofoil shape, Reynolds number
and angle of attack.
2) The equivalence of the factor written in three forms should
include at-least one factor in terms of the Reynolds number and
at-least one factor in terms of angle of attack as these variables
Effective lift Coefficient of
to drag ratio power
10 (LDL) 0.15 (CPL)
Linear
interpolation of
coefficient of
LD CPmax power based on
intermediate
values
140 (LDU) 0.47 (CPU)
40 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445

factor and Compatibility demand factor; the expressions of which

are given by Eqs. (28)(30).
Aero-Induction demand factor,
n  2 o
DF A log Cp= 1  a sin fSolg 28

Hybrid demand factor,

( )
VR 2
sin 2 Sol
2
DF H log Cp= 29
V1

Compatibility demand factor,

h i
DF C log fARgP=V 1 N 2 2 Re2 sin2 30

As evident from the formulations, the demand factor expres-

sions contain the logarithmic function as the order of magnitude
of the expression (without them) was large making computations
difcult. The factors (inside the logarithmic function) are always
positive as the terms in the computation such as sin and sin
which have chances of attaining a negative value are used as their
squares and hence the resulting expression is invariably positive.
As logarithm of a negative value is undened, the unconditionally
Fig. 8. Legend for DF A diagrams (Not to scale). positive nature of the expressions ensures its validity everywhere.
With reference to Eq. (30), the VAWT manufacturer cannot
insist on a xed value of aspect ratio as that will make the design
Table 3 very rigid. An acceptable range of aspect ratio is specied, and any
Choosing the DF A closest to the rounded off value. design that falls in that range is taken as a solution to the problem.
Sol Cp LD DF A Before rounding Difference
With regards to the number of blades, there will be a minimum
and maximum number of VAWT blades from considerations of
 28  3.869 0.13 0.438 127 4.795 2.3 2.2735 0.0264 erection feasibility and that which the manufacturer would be
 27  3.684 0.13 0.438 127 4.795 2.3 2.3170 0.0170 comfortable to build. Therefore, DF C can have a range of values
17 3.374 0.13 0.438 127 4.795 2.3 2.3248 0.0248
depending upon the values of aspect ratio and number of blades.
18 3.547 0.13 0.438 127 4.795 2.3 2.2836 0.0163
161 5.878 0.13 0.438 127 4.795 2.3 2.2540 0.0460 The limits are identied as the minimum and maximum values of
162 5.635 0.13 0.438 127 4.795 2.3 2.2940 0.0060 DF C .
163 5.390 0.13 0.438 127 4.795 2.3 2.3361 0.0361 It can be readily seen that, substituting the expressions for
208  4.978 0.13 0.438 127 4.795 2.3 2.3038 0.0037 aspect ratio, Reynolds number, solidity and power from Eq. (17),
209  5.251 0.13 0.438 127 4.795 2.3 2.2585 0.0414
(16), (18) and (15) respectively in Eq. (30) will lead to Eq. (29) and
further substitution of the expression for angle of attack from Eq.
(1) and induction factor expression from Eq. (2) will lead to Eq.
(28). This equivalence of the factors is given by Eq. (31).
are numerically linked to other variables of analysis (Eqs.
(1) and (16)) and are thus prudent for ensuring compatibility. DF A DF H DF C 31
3) The factor should not have any output parameters listed in The expression for DF H contains the variable, angle of attack
Table 2 as they are unknowns until the completion of design. and DF C contains the variables, angle of attack and Reynolds
4) The factor should contain all the input parameters listed in number. Hence, to ensure compatibility of the expressions, we
Table 2 as the optimized solution can be developed only after need to use the values of angle of attack and Reynolds number for
their consideration. the aerofoil data under consideration and it can be regarded as
5) The factor developed should preferably be non-dimensional. Representative streamtube for the particular problem only if it
satises Eq. (31) identically. Fig. 7 diagrammatically represents this
With these criteria's in mind, considerable amount of trial and condition.
error procedure was carried out to nally arrive at the expression Further, it can be seen that DF C contains all the input para-
for the factor. The factor was given the name Demand Factor to meters that are listed in Table 2 and hence satisfying Eq. (31)
signify its role of meeting a certain power requirement subjected identically would also mean that Representative streamtube is
to the constraints of the other input parameters. With the 4th and qualied in satisfying the optimization problem at hand.
5th criteria in mind, it was found that the non-dimensionality of
the demand factor and consideration of all the input parameters
can be ensured if the input parameters listed in Table 2 are com- 4. Allowable range of Aero-Induction demand factor
bined in the following format (Eq. (27)).
Among the demand factors, the Hybrid demand factor (DFH)
fARgP=V 1 2 27 and Compatibility demand factor (DFC) can be calculated only once
we start solving a particular problem as they depend on Reynolds
With sights on 1st and 2nd criteria, the satisfaction of 3rd number and angle of attack of the aerofoil under consideration.
criteria was attempted and it was found that despite repeated However, the range of Aero-Induction demand factor (DFA) can be
trials, it was not possible to remove the number of blades (which is calculated in advance for all values of lift to drag ratios and azi-
an output parameter) from the expression. With much con- muthal locations.
sideration, the three forms of the expression were arrived at and For the construction of the Aero-Induction demand factor dia-
are named as Aero-Induction demand factor, Hybrid demand gram, the list of all combinations of lift to drag ratio from 10 to 140
 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445 41

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9
diagram ( = 0.9 to 1.5). diagram ( = 1.6 to 2.2). diagram ( = 2.3 to 2.9).

3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8
diagram ( = 3.0 to 3.6). diagram ( = 3.7 to 4.2). diagram ( = 4.3 to 4.8).

4.9 5 5.1 5.2 5.3 5.4

diagram ( = 4.9 to 5.4).

Fig. 9. (a) DF A diagram (DF A 0.91.5). (b) DF A diagram (DF A 1.62.2). (c) DF A diagram (DF A 2.32.9). (d) DF A diagram (DF A 3.03.6). (e) DF A diagram (DF A 3.74.2).
(f) DF A diagram (DF A 4.34.8). (g) DF A diagram (DF A 4.95.4).

were listed in increments of one (e.g. 10, 11, 12, 13, .). They were
then combined with azimuthal location of  90 to 270 in
increments of one degree (e.g.  90,  89,  88,  87, .). The
resulting set of 47,160 combinations (131 multiplied by 360) of
azimuthal location and lift to drag ratio were substituted in Eqs.
(22) or (23) (depending upon the value of induction factor) to see
if there is a realistic value of induction factor that could sustain
each of those combinations. In those equations, the values of and
CPmax used are those obtained using Eqs. (25) and (26) for the
given value of lift to drag ratio.
The combinations of lift to drag ratio and azimuthal locations
which yielded a converged and realistic value of induction factor
were retained separately. Some combinations of lift to drag ratio
and azimuthal location did not yield a converged value of
induction factor and were discarded. Those combinations which
yielded an unrealistic value of induction factor (values greater
than one) were also discarded. The signicance of this exercise
was to nd out whether the particular combination of lift to drag
ratio and azimuthal location is eligible to become Representative
streamtube. If it is a possibility, then the value of induction factor
obtained is the one that will sustain the particular value of lift to
drag ratio at that particular azimuthal location. On the contrary,
the combinations of lift to drag ratio and azimuthal location for
which a converged/realistic value of induction factor was not
obtained means that the particular value of lift to drag ratio could
never be Representative streamtube at that particular azimuthal
location.
42 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445

Aero-Induction demand factor given by Eq. (28) was calculated
for the combinations of lift to drag ratio and azimuthal locations
which yielded a converged and realistic value of induction factor.
The value of DF A was rounded off to the nearest 10th of a decimal.
From the resulting set, it was found out that for certain lift to drag
ratios, multiple azimuthal locations gave the same value of
demand factor. Out of those combinations, the case that led to a
140
No. of compatible combinations
DF A value closest to the rounded off value was retained and others
were discarded.
Table 3 depicts this situation where it can be seen that a total of
nine different azimuthal locations provide a DF A of 2.3 (after
rounding off to the nearest 10th of a decimal) for a lift to drag ratio
of 127. In the table, intermediate calculations are not shown for
clarity of viewing. Out of these cases, the difference between the
actual DF A and rounded off DF A was least (0.0037) for the azi-
muthal location of 208 (highlighted in grey shade) and was
selected out of all the nine combinations. Similar exercise was
performed for all combinations of lift to drag ratio and DF A which
had multiple azimuthal locations. After this process of screening, a

120 total of 3309 values remained and they are plotted in Fig. 9(a)(g).
100 This is a polar plot with the radial direction representing the lift to
drag ratio (10140) and the angular location representing the
80
actual azimuthal location (  90 to 270) plotted on plan (view) of
60
the VAWT. Each set of plot corresponds to a particular value of
40 DF A . A legend for interpreting the plots is given in Fig. 8.
20 As evident from the gures, the values follow a denite pattern
0 as they increase from the minimum value to the maximum value.
0.8 1.8 2.8 3.8 4.8 Fig. 10 shows the plot of compatible combinations versus Aero-
Aero-Induction demand factor Induction demand factor for each combination of azimuthal loca-
Fig. 10. Plot of number of compatible combinations for each value of DF A . tion and lift to drag ratio (subjected to a maximum value of 131,
which is obtained as the number of increments in one from 10 to
140). DF A less than 0.9 and greater than 5.4 did not provide
Table 4 compatible combinations of lift to drag ratio and azimuthal loca-
Permissible lift to drag ratios for different values of DF A . tion for any of the combinations (and thus were not included in
0.9 3.1 1011 128140 15 4.4 5
the gure). For DF A values between 1.4 and 2.8, all the lift to drag
1014 1020 1531 3.7 18 1832 5961 ratio values between 10 and 140 have a chance of becoming
1 2346 36 10 2327 3941 8488 Representative streamtube.
1031 4856 3946 1617 2939 6163 5.1 Table 4 shows the values of lift to drag ratios that are admis-
1.1 59140 5155 2124 5356 4.5 6264
sible for various values of DF A . In this table, the DF A values are
1054 3.2 5960 2638 4.1 2846 8990
1.2 1033 7281 4852 16 6465 5.2 marked in bold. As evident from the table, they contain the entire
1083 3544 118140 57 19 4.6 6566 range of DF A values from 0.9 to 5.4. Under each value of DF A , the
1.3 4776 3.5 3.8 2834 3644 91 values of lift to drag ratios that provided a converged solution out
10122 92140 1018 1018 4048 4758 5.3
of the range considered (10140) are noted down. For instance,
1.42.8 3.3 2026 2330 5758 4.7 6768
All 1023 3239 3946 4.2 4449 5.4 under the DF A value of 3.1, the following ranges have been listed:
2.9 2538 4752 5356 3336 5969 69 1020, 2346, 4856 and 59140. It means that lift to drag ratios
1027 40 5658 120140 4955 4.8 of 21, 22, 47, 57 and 58 which are not part of the listing are not
29140 4450 82127 3.9 59 50 54 admissible for the DF A of 3.1. The table should be read as con-
3 5458 3.6 1331 4.3 7077
1024 62 1428 4752 1019 4.9
tinuous columns. For instance, the second column ends with the
2765 77140 4047 57 3638 5558 DF A value of 3.4 and the permissible lift to drag ratio under it
68140 3.4 5356 4 5660 7883 should be read starting from the third column (e.g. 1011, 1531,
etc.). Such an arrangement was required to ensure that the entire

Table 5
Aerofoil data for problem 1.

Given data Calculated data

Re NACA CL CD C L =C D DF C  L DF C  U

10,000,000 11 0012 1.21 0.0104 116 2.7466E  13  1.4 0

10,000,000 10 0012 1.1 0.0097 113 3.3163E  13  1.3 0
10,000,000 12 0012 1.2906 0.0116 111 2.3133E  13  1.4  0.1
10,000,000 9 0012 0.99 0.0091 109 4.0863E  13  1.2 0.1
10,000,000 13 0012 1.3687 0.0127 108 1.9761E  13  1.5  0.2
5,000,000 10 0012 1.1 0.0106 104 1.3265E  12  0.7 0.6
10,000,000 8 0012 0.88 0.0086 102 5.1628E  13  1.1 0.2
5,000,000 9 0012 0.99 0.0098 101 1.6345E  12  0.6 0.7
10,000,000 14 0012 1.4171 0.0141 101 1.7086E 13  1.6  0.3
5,000,000 11 0012 1.1842 0.0118 100 1.0986E  12  0.8 0.6
5,000,000 12 0012 1.2673 0.013 97 9.2534E  13  0.8 0.5
5,000,000 8 0012 0.88 0.0092 96 2.0651E  12  0.5 0.8
10,000,000 7 0012 0.77 0.0082 94 6.7330E 13 1 0.3
5,000,000 13 0012 1.3242 0.0143 93 7.9046E  13  0.9 0.4
10,000,000 15 0012 1.4214 0.0157 91 1.4928E  13  1.6  0.3
5,000,000 7 0012 0.77 0.0086 90 2.6932E  12  0.4 0.9
2,000,000 9 0012 0.99 0.0113 88 1.0215E  11 0.2 1.5
 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445 43

table ts in one page, without affecting the readability of the
values. Since the DF A values from 1.4 to 2.8 provided converged
solution for the entire range of lift to drag ratios (also evident from
Fig. 10), they are clubbed together as a range and mentioned All
in the table.
5. Steps involved in problem solving
The optimization process of a particular problem is carried out
through the following steps:
1) The set of aerofoil data with lift and drag values (for different
values of Reynolds number and angle of attack) that are parti-
cipating in the optimization process is arranged in decreasing
order of their lift to drag ratio's.
2) Starting with the rst aerofoil, the range of compatibility
demand factor, DF C is calculated as given in Eq. (30). For this
calculation, the values of Reynolds number and angle of attack
pertaining to the aerofoil data is used. The value of power,
velocity, density and viscosity of wind pertaining to the opti-
mization problem are used. From the values of desired aspect
ratio range and a realistic minimum and maximum value for
number of blades, the permissible range for DF C is obtained.
3) From Table 4, the permissible values of DF A for the lift to drag
ratio under consideration is referenced. Those DF A values which
comes under the range of DF C calculated in step 2 is retained
separately.
4) For these common values (with DF A DF C ), DF H is calculated
using Eq. (29) and all those cases which satisfy Eq. (31)
identically becomes the set of solutions which are eligible to
become Representative streamtube for the given problem.
Since the aerofoils are arranged in the decreasing order of their
lift to drag ratios, the rst case which satises this condition
becomes the optimized solution.
5) Once the azimuthal location of Representative streamtube
corresponding to the optimized solution is found out, the ve
variables namely chord length, length of blade, number of
blades, radius of rotor and angular velocity at rated speed can be
calculated using Eqs. (16), (15), (18), (17) & (7) respectively.
Two demonstrative problems based on this approach have
been included in Appendix.
6. Limitations of the approach
The proposed approach is based on the assumption that
Representative streamtube is representative of the entire VAWT.
The lower and upper bounds of the variables as depicted in Fig. 6 is
only valid for NACA0012. For the application of this approach to
other aerofoils, the enveloping bounds of the optimum char-
acteristics should be obtained through calculations.
The approach is based on a constant external wind velocity.
However, for Representative streamtube located in the down-
stream half of the rotor, there is an additional change in the
velocity reaching the streamtube owing to its passage through the
upstream streamtube (as depicted through Eq. (19)). This change
in velocity is not accounted in the calculation of DF C using Eq. (30).
The resulting inaccuracy can be mitigated through either one of
the following ways:
1) Neglect the cases that correspond to Representative streamtube
located in the downstream half of the rotor. This can be brought
about by screening such cases during the elimination stage
discussed through Table 3.
2) Adopt a suitable correction factor.
7. Conclusion
An approach for optimizing Vertical Axis Wind Turbines
(VAWT) has been presented that uses the theory of double mul-
tiple streamtube method but eliminates the need of iterations. A
single streamtube termed Representative streamtube is assumed
to represent the property of the entire VAWT. All the participating
aerofoils are arranged in descending order of their lift to drag
ratios (for various combinations of Reynolds number and angle of
attack) and starting with the rst set of lift and drag values, a
check is made to determine if it is eligible to become Repre-
sentative streamtube. The check is carried out through a factor
that was developed through systematic modication of existing
formulation in a way that it satises the intended objective. The
rst of the lift to drags ratios (along with its associated properties)
that satises the requirement of becoming Representative
streamtube is taken as the optimized solution for the problem.

Table 6
Comparison of optimized shape parameters of the VAWT. Table 8
Comparison of optimized shape parameters of the VAWT.
1
N NACA C (m) L (m) R (m) (s )
N NACA C (m) L (m) R (m) (rpm)
This approach 2 0012 0.50 9.55 7.76 0.81
SAWT Energy, 5 Not specied Not specied 9 6.5 Not specied This approach 2 0012 2.26 71.96 34.72 17.86
2015 Marsh and Peace, 2005 3 Not specied 5 55 38.5 28

Table 7
Aerofoil data for problem 2.

Given data Calculated data

Re NACA CL CD C L =C D DF C  L DF C  U

10,000,000 11 0012 1.21 0.0104 116 2.7466E  13 0.4 1.7

10,000,000 10 0012 1.1 0.0097 113 3.3163E  13 0.5 1.8
10,000,000 12 0012 1.290 0.0116 111 2.3133E  13 0.3 1.6
10,000,000 9 0012 0.99 0.0091 109 4.0863E  13 0.6 1.9
44 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445
Appendix
To demonstrate the working of this approach, two sample
VAWT proportioning problems have been included as under.
Problem 1
The proposed VAWT needs to produce a rated power of 50 kW
for a rated wind speed of 13 m/s. The architectural requirements
and space constraints of the location restrict the aspect ratio to the
range of 0.3 to 1.0. Viscosity and density of air for the site can be
taken as 1.78E  05 kg/m-s and 1.08 kg/m3 respectively.
Solution
For this problem, the lift and drag coefcients pertaining to
NACA0012 is taken from experiments conducted by (Sheldahl and
Klimas, 1981) under the aegis of Sandia National Laboratories.
Implementing the steps discussed in Section 5 leads to the
following:
Step 1) The aerofoil data that are participating in the optimi-
zation process is arranged in decreasing order of their lift to drag
ratio and the rst 17 of them are shown in Table 5.
Step 2) The range of DF C is calculated and shown in Table 5.
The lower limit is given as DF C L and the upper limit as
DF C  U. For this calculation, the range of minimum and max-
imum number of blades is taken as 2 and 5 respectively.
Step 3) With the range of DF C values known, the values of DF A
that are acceptable for the particular value of lift to drag ratio
are obtained from Table 4. The rst case which has some values
in the acceptable range is marked with grey shading in Table 5.
The acceptable values are 1.3, 1.4 & 1.5.
Step 4) Out of the acceptable values of 1.3, 1.4 & 1.5, the DF H
gives the same value only for 1.5.
Step 5) The VAWT output parameters are calculated using Eqs.
(15)(18) and (7). The deign parameters of the optimized VAWT
is summarized in Table 6. It is compared with a VAWT designed
for a similar condition in (SAWT Energy, 2015).
Problem 2
Redesign the VAWT mentioned in problem 1 for a rated power
of 3.0 MW and for a rated wind speed of 14 m/s.
Solution
Assumptions on minimum and maximum number of blades
and other values are retained from Problem 1.
Step 1) The list of rst four aerofoil data that are participating
in the optimization process (in decreasing order of lift to drag
ratio) are shown in Table 7.
Step 2) The lower limit (DF C  L) and upper limit (DF C  U) of
DF C are shown in Table 7.
Step 3) The rst row (marked in grey shade) of Table 7 has
values in the acceptable range. The acceptable values are 1.3, 1.4,
1.5, 1.6 & 1.7.
Step 4) Out of those values, Eq. (31) is identically satised for
values of 1.4, 1.5 & 1.6.
Step 5) Taking 1.4 as the solution and based on the azimuthal
location of that particular value, the output parameters of VAWT
have been found. The deign parameters of the optimized VAWT
is summarized in Table 8. It is compared with a VAWT designed
for a similar condition in (Marsh and Peace, 2005).
References
Ahmadi-Baloutaki, M., Carriveau, R., S-K Ting, D., 2014. Straight-bladed vertical axis
wind turbine rotor design guide based on aerodynamic performance and
loading analysis. J. Power Energy 228 (7), 742759.
Ajedegba, J.O., 2008. Effects of Blade Conguration on Flow Distribution and Power
Output of a Zephyr Vertical Axis Wind Turbine (Master of Applied Science
thesis). University of Ontario Institute of Technology, Ontario.
Asher, I.M., Drela, M., Peraire, J., 2010. A Low Order Model for Vertical Axis Wind
Turbines. In: Proceedings of the 28th AIAA Applied Aerodynamics Conference,
Chicago.
Bedon, G., Paulsen, U.S., Madsen, H.A., Bellonia, F., Castelli, M.R., Benini, E., 2015.
Aerodynamic benchmarking of the deepwind design. Energy Procedia 75,
677682.
Bedon, G., Paulsen, U.S., Madsen, H.A., Belloni, F., Castelli, M.R., Benini, E., 2014.
Aerodynamic Benchmarking of the Deepwind Design. The European Wind
Energy Association, Barcelona.
Beri, H., Yao, Y., 2011. Double multiple stream tube model and numerical analysis of
vertical axis wind turbine. Energy Power Eng. 3, 262270.
Bianchini, A., Ferrara, G., Ferrari, L., 2015. Design guidelines for H-Darrieus wind
turbines: optimization of the annual energy yield. Energy Convers. Manag. 89,
690707.
Bitsch, J., Darlow, R., Nielsen, J.G., Qumneur, C., Savino, S., 2011. Modeling of a
Vertical Axis Wind Turbine with Respect to Pitching Angle and Wing Proles
(Bachelor Project). Aalborg University, Denmark.
Blusseau, P., Patel, M.H., 2012. Gyroscopic effects on a large vertical axis wind
turbine mounted on a oating structure. Renew. Energy 46, 3142. http://dx.
doi.org/10.1016/j.renene.2012.02.023.
Bogeanu, R., Dobrescu, B., Nil, I., 2010. Aerodynamic performance prediction of
Darrieus-type wind turbines. INCAS Bull. 2 (2), 2632.
Borg, M., Collu, M., Brennan, F.P., 2013. Use of a wave energy converter as a motion
suppression device for oating wind turbines. Energy Procedia 35, 223233.
http://dx.doi.org/10.1016/j.egypro.2013.07.175.
Castelli, M.R., Englaro, A., Benini, E., 2011. The Darrieus wind turbine: proposal for a
new performance prediction model based on CFD. Energy 36, 49194934.
Castillo, J., 2011. Small-Scale Vertical Axis Wind Turbine Design (Bachelors Thesis).
Tampere University of Applied Sciences, Finland.
Chen, X., Agarwal, R., 2012. Optimal placement of horizontal - and vertical-axis
wind turbines in a wind farm for maximum power generation using a genetic
algorithm. Int. J. Energy Environ. 3 (6), 927938.
Claessens, M.C., 2006. The Design and Testing of Airfoils for Application in Small
Vertical Axis Wind Turbines (Master of Science thesis). Delft University of
Technology, Netherlands.
Clarke, J., Hancox, L., MacKenzie, D., Whelan, M., 2014. Design of a Vertical-Axis
Wind Turbine Phase II. Group 11 - MUN VAWT Design.
Deglaire, P., 2010. Analytical Aerodynamic Simulation Tools for Vertical Axis Wind
Turbines (Doctor of Philosophy dissertation). Acta Universitatis, Upsaliensis,
Uppsala, Sweden.
Diaz, D.A.G., Pinto, F.A.N.C., 2012. Vertical Wind Turbine with Variable Blade
Angular Position. In: Proceedings of ABCM Symposium Series in Mechatronics.
vol. 5, pp. 474483.
Farthing, S.P., 2009. Vertical axis wind turbine induced velocity vector theory. Proc.
IMechE Part A: J. Power Energy 223, 103114.
Hasan, O.D.S., Hantoro, R., Nugroho, G., 2013. Studi eksperimental vertical axis wind
turbine tipe savonius dengan variasi jumlah n pada sudu. J. Tek. Pomits 2 (2),
B-350B-355.
Helms, S.R., 2011. Simulations of Vertical Axis Wind Turbine Blades (Master of
Science thesis). Drexel University, Philadelphia.
Homicz, G.F., 1991. Numerical Simulation of VAWT Stochastic Aerodynamic Loads
Produced by Atmospheric Turbulence: VAWT-SAL Code, Sandia report SAND91-
1124. Sandia National Laboratories, Albuquerque.
Islam, M., Carriveau, R., Fartaj, A., 2012. Performance analyses of a xed-pitch
straight bladed VAWT with selected low Reynolds number airfoils. Int. J.
Environ. Stud. 69 (2), 289298.
Islam, M., Ting, S.-K., Fartaj, A., D., 2008. Aerodynamic models for Darrieus-type
straight-bladed vertical axis wind turbines. Renew. Sustain. Energy Rev. 12,
10871109.
Komerath, N.M., Pendharkar, A.M., Raghav, V., 2013. The low-cost vertical axis wind
turbine project: an exercise in learning across disciplines. In: Proceedings of the
120th ASEE Annual Conference and Exposition, Atlanta.
Larrea, L.R., 2015. Vibration Analysis of a VAWT (Master thesis). Technical Uni-
versity of Soa, Bulgaria.
Loth, J.L., McCoy, H., 1983. Optimization of Darrieus turbines with an upwind and
downwind momentum model. J. Energy 7 (4), 313318.
Maheri, A., Noroozi, S., Toomer, C., Vinney, J., 2006. Damping the Fluctuating
Behaviour and Improving the Convergence Rate of the Axial Induction Factor in
the BEMT-Based Rotor Aerodynamic Codes. European Wind Energy Conference,
Athens.
Marsh, G., Peace, S., 2005. Tilting at windmills: utility-scale VAWTs: towards
10 MW and beyond? Refocus 6 (5), 3742. http://dx.doi.org/10.1016/
S1471-0846(05)70457-1.
McGowan, R., Morillas, K., Pendharkar. A., Pinder, M., Komerath, N., 2012. Optimi-
zation of a Vertical Axis Micro Wind Turbine for Low Tip Speed Ratio Operation.
In: Proceedings of the 10th International Energy Conversion Engineering
Conference, Atlanta.
 M. Abdul Akbar, V. Mustafa / J. Wind Eng. Ind. Aerodyn. 153 (2016) 3445
Ming, W.F., 2009. Investigation on Straight-bladed Wind Turbines (Master's thesis).
Lanzhou University of Technology, Lanzhou.
Newman, B.G., 1983. Actuator-disc theory for vertical-axis wind turbines. J. Wind
Eng. Ind. Aerodyn. 15 (13), 347355.
Newman, B.G., 1986. Multiple actuator-disc theory for wind turbines. J. Wind Eng.
Ind. Aerodyn. 24, 215225.
Okpue, A.S., 2011. Aerodynamic Analysis Of Vertical And Horizontal Axis Wind
Turbines (Master of Science thesis). Michigan State University, East Lansing.
Paraschivoiu, I., 1981. Double-multiple streamtube model for Darrieus wind tur-
bines. In: Second DOE/NASA Wind Turbines Dynamics Workshop, pp. 1925.
Pratumnopharat, P., Leung, P.S., 2011. Validation of various windmill brake state
models used by blade element momentum calculation. Renew. Energy 36,
32223227.
Ponta, F.L., Lago, L.I., 2008. Analysing the suspension system of variable-geometry
oval-trajectory (VGOT) Darrieus wind turbines. Energy Sustain. Dev. 12 (2),
516.
Ponta, F., Otero, A., Lago, L., 2011. Innovative concepts in wind-power generation:
the VGOT Darrieus. Wind Turbines, 137162.
Ponta, F.L., Seminara, J.J., Oteroa, A.D., 2007. On the aerodynamics of variable-
geometry oval-trajectory Darrieus wind turbines. Renew. Energy 32, 3556.
Rathi, D., 2012. Performance Prediction and Dynamic Model Analysis of Vertical
Axis Wind Turbine Blades with Aerodynamically Varied Blade Pitch (Master of
Science thesis). North Carolina State University, Raleigh.
Samaraweera, K.K.M.N.P., Pathirathna, K.A.B., De Silva, H.E.D., Sugathapala, A.G.T.,
2010. Development of Darrieus-Type Vertical Axis Wind Turbine for Stand-
Alone Applications. In: Proceedings of International Conference on Sustainable
Built Environment, Kandy, pp. 421430.
SAWT Energy, http://www.sawtenergy.com/pk50.html, 1st May 2015.
45
Sheldahl R.E., Klimas, P.C., 1981. Aerodynamic Characteristics of Seven Symmetrical
Airfoil Sections Through 180-Degree Angle of Attack for Use in Aerodynamic
Analysis of Vertical Axis Wind Turbines, SAND80-2114. Sandia National
Laboratories energy report, Albuquerque.
Solum, A., Deglaire, P., Eriksson, S., Stlberg, M., Leijon, M., Bernhoff, H., 2006.
Design of a 12 kW vertical axis wind turbine equipped with a direct driven PM
synchronous generator. In: Proceedings of European Wind Energy Conference &
Exhibition, Athens.
Soraghan, C.E., Jamieson, P., Leithead, W.E., 2013. Inuence of lift to drag ratio on
optimal aerodynamic performance of straight blade vertical axis wind turbines.
In: Proceedings of European Wind Energy Association Annual Conference,
Vienna.
Sharpe, T., Proven, G., 2010. Crossex: concept and early development of a true
building integrated wind turbine. Energy Build. 12, 23652375.
Shires, A., 2013a. Design optimisation of an offshore vertical axis wind turbine.
Proc. Instit. Civil Eng. Energy 166 (EN1), 718.
Shires, A., 2013b. Development and evaluation of an aerodynamic model for a novel
vertical axis wind turbine concept. Energies 6, 25012520.
Sullivan, W.N., 1979. Economic Analysis of Darrieus Vertical Axis Wind Turbine
Systems for the Generation of Utility Grid Electrical Power, Volume II The
Economic Optimization Model, SAND78-0962. Sandia Laboratories energy
report, Albuquerque.
Tang, S., Xu, Y., Zhu, G., Fang, W., 2011. Orthogonal experiment optimum design in
the power of the rotor on straight-bladed VAWT. Energy Procedia 12, 856861.
Wille, T., 2015. Development of a Calculation method for the Interpretation of pitch
regulations operated in Vertical Wind turbines by Open Source criteria,
Bachelor thesis. Beuth University of Applied Sciences, Berlin.