Renewable Energy 93 (2016) 536e547

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Studies of some high solidity symmetrical and unsymmetrical blade

H-Darrieus rotors with respect to starting characteristics, dynamic
performances and ow physics in low wind streams
A.R. Sengupta a, A. Biswas a, *, R. Gupta b
a
Mechanical Engg. Deptt., National Institute of Technology Silchar, Silchar, India
b
NIT Srinagar, India

a r t i c l e i n f o a b s t r a c t

Article history: Vertical axis wind turbines can be successfully installed in low wind speed conditions but its detailed
Received 17 July 2015 starting characteristics in terms of starting torque, starting time and dynamic performances have not
Received in revised form been investigated thoroughly which is important for increasing the energy yield of such turbines.
7 December 2015
Amongst their designs, H-Darrieus rotor, in spite of having good power coefcient, possesses poor self-
Accepted 4 March 2016
starting features as symmetrical blade proles are used most of the times. Instead of using symmetrical
blades if unsymmetrical or cambered blades are used with high solidity, then starting performance of H-
Darrieus rotor along with its power coefcients can be improved. Though this performance improvement
Keywords:
H-Darrieus rotor
measures are correlated with improvement in the starting characteristics, a detailed work in this di-
Self-starting rection would be useful and for this reason the present work has been carried out. Three types of blade
Starting time designs have been considered; two unsymmetrical blades namely S815 and EN0005 and one conven-
Static torque coefcient tional symmetrical NACA 0018 blade, and experiments are conducted using a centrifugal blower test rig
Dynamic torque coefcient for three-bladed H-Darrieus rotors using these three considered blades at low wind streams (4 m/s, 6 m/s
Power coefcient and 8 m/s). Considering reality, the effects of ow non-uniformity and turbulence intensity on the rotor
Flow physics performance at optimum condition as well as ow physics have also been studied. It has been found that
unsymmetrical S815 blade rotor has higher dynamic torque and higher power coefcient than unsym-
metrical EN0005 and symmetrical NACA 0018 blade H-Darrieus rotors.
2016 Elsevier Ltd. All rights reserved.

1. Introduction inessentiality of yaw mechanism etc. Generally, VAWTs can be

Global warming trends, declining fossil fuel sources, hike in oil
prices etc. have renewed the interest of many researchers to
generate energy from renewable energy sources as they are inex-
haustible and pollution free. Among various renewable energy
sources, wind energy has got increasing attention as a large scale
power production technology in urban and rural areas to meet the
energy demand of people. At present, vertical axis wind turbine
(VAWT) systems are getting prime attention for power generation
in low wind regime among all wind turbine systems because
of their structural simplicity, independency of wind direction,
cost effectiveness, easier maintenance, lower noise emission,
* Corresponding author.
E-mail addresses: analsengupta88@gmail.com (A.R. Sengupta), agnibis@yahoo.
co.in (A. Biswas), rguptanitsri@gmail.com (R. Gupta).
classied into two types: Savonius and Darrieus turbines or rotors.
Darrieus rotors are lift type machines which have higher power
efcient than the Savonius rotors. H-Darrieus rotor, a variant of the
curve bladed Darrieus rotor is becoming very popular recently.
These rotors have straight vertical blades which are of simpler
construction and design, and these blades can be connected
directly to the rotor shaft by means of struts [1]. H-Darrieus rotor
shows higher power coefcient than curve bladed Darrieus rotor
[2] and hence this rotor can be useful for generating signicant
power, crop grinding, and water pumping etc. in remote areas. But
the main disadvantage of this rotor is its poor self-starting features
because of symmetrical blade designs [3,4]. So more research work
is required for nding suitable blade designs so that the VAWT can
be self-starting at all rotor angles and its performance can also be
improved. Various solutions were presented by a number of re-
searchers to overcome the poor self-starting ability of Darrieus
rotor: a combination of Savonius and Darrieus rotor for making the

http://dx.doi.org/10.1016/j.renene.2016.03.029
0960-1481/ 2016 Elsevier Ltd. All rights reserved.
 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547
Darrieus fully self-starting and its efciency improvement
compared to Savonius rotor [5], application of guide-vane [6],
employment of mechanical system for optimizing blade pitch [7],
use of specially designed blades that transform their shape almost
like an airfoil during rotation [8,9]. All these designs had more or
less problems; either these models were too complex to manu-
facture and needed good support for blades or they showed
decrease in operating range and peak efciencies though they
resulted increment of starting torque. However, unsymmetrical or
cambered airfoil shaped blades can be a better design prospect for
obtaining good starting features [10]. Dominy et al. [11] investi-
gated the effect of number of blades on the self-starting behaviour
of H-Darrieus rotor having symmetrical NACA 0012 airfoil at 10 m/s
wind speed. They observed that though two-bladed rotor has self-
starting potential, but it is not possible for all starting position of
the rotor which is a problem for commercial viability. However,
they conrmed by model testing that irrespective of its starting
position, three-bladed H-Darrieus rotor has the capability to self-
start under steady wind conditions. Barker [12] made a theoret-
ical comparison between the performances of an unsymmetrical
NACA 0012 airfoil and a symmetrical NACA 0012 airfoil blade and
noticed that the unsymmetrical prole exhibited more tangential
thrust over a wide angle of attack range and ensured that this blade
design can be consistently self-started. Despande and Li [13] per-
formed a numerical study of a three bladed xed pitch H-Darrieus
rotor based on three different airfoils which were symmetrical
NACA 0015, NACA 0018 and unsymmetrical S1210. Considering two
different solidities like 0.2 and 0.4 for symmetrical and cambered
airfoils respectively, various advantages of cambered S1210 blades
were shown. Islam et al. [14] analysed an unsymmetrical special
purpose airfoil named MI-VAWT1 for smaller capacity three bladed
H-Darrieus rotor based on different parameters varying the solidity
range from 0.1 to 1. It was reported that this particular blade shows
better aerodynamic performance than NACA 0015 airfoil. Beri and
Yao [15] investigated a two-dimensional unsteady CFD analysis on
cambered NACA 2415 shape three bladed xed pitch H-Darrieus
rotor. The 2D simulation showed that signicant average torque is
generated which could be sufciently self-starts this H-Darrieus
rotor for this particular cambered airfoil. Sayed et al. [16] performed
a CFD simulation of different low-speed ow past 2D S-series blade
proles of wind rotor developed by The National Renewable Energy
Laboratory (NREL). It was observed that at low angle of attack the
lift coefcient increases and drag decreases by using an unsym-
metrical airfoil, and an airfoil having high lift/drag ratio would have
higher efciency than the one with lower lift/drag ratio. Gupta and
Biswas [17] analysed the performance of a twisted unsymmetrical
three bladed H-Darrieus rotor by a steady-state 2D CFD analysis. It
was noticed that when blade angle is greater than 90
having blade
twist of 30
, then the self-starting of this rotor could be expected.
Recently various high solidity (in the range of 0.4e1.0) three-bladed
H-Darrieus rotors are being investigated. Saeidi et al. [18] designed
and optimized a three bladed cambered NACA 4415 shape H-Dar-
rieus rotor to check its economic feasibility. It was found that at
solidity 0.4 and rotor aspect ratio 1 with tip speed ratio (TSR) 4, the
rotor shows maximum power coefcient 0.472 and this power
coefcient decreases when solidity is decreased and increases
when TSR is increased. Bhuyan and Biswas [4] studied experi-
mentally the self-starting and performance features of a cambered
S818 shape three bladed simple H-Darrieus rotor and a hybrid
system with Savonius rotor connected at its centre. It was found
that for aspect ratio and solidity 1, simple H-Darrieus rotor is not
fully self-starting at all rotor positions and shows maximum power
coefcient of 0.28, whereas hybrid rotor fully self-starts at all rotor
angle and its maximum power coefcient is 0.34. Mohamed [19]
investigated some techniques to improve the poor self-starting
537
ability of two-bladed H-Darrieus rotor for S-1046 symmetrical
airfoil blade shape. It was concluded that solidity increment (as
high as 0.43) improves the self-starting ability of H-Darrieus rotor
as it increases the static torque coefcient with solidities higher
than 0.25. Singh et al. [3] performed an experimental analysis for
high solidity three-bladed H-Darrieus rotor having S1210 unsym-
metrical blades to evaluate its self-starting characteristics and
performance with various rotor aspect ratios (from 0.8 to 1.2) at
various rotor angular positions and wind speeds. This rotor
exhibited self-starting for all aspect ratios considered at all rotor
positions while the highest power coefcient of 0.32 was found for
aspect ratio 1. It was noticed that this unsymmetrical blade with
high solidity has got better prospect in terms of starting torque and
power coefcient though its operating range is smaller. Nguyen
et al. [20] performed a numerical study on the blade thickness ef-
fect and starting time at different azimuthal angles for three bladed
H-Darrieus rotor at wind speeds 3e7 m/s using four symmetrical
blades namely NACA 0012, NACA 0015, NACA 0018 and NACA 0021.
They concluded that for larger thickness of the airfoil, self-starting
ability is also higher. However a detail performance study of un-
symmetrical three bladed H-Darrieus rotor for different high so-
lidities with regards to static characteristics, starting time, dynamic
performances and various unsymmetrical blade proles at low tip
speed ratios are less. So some research work is needed in this area
to show that these types of vertical axis rotors can work in low
wind speed conditions effectively.
2. Objective of the present work
From the earlier section, it has been understood that for solving
poor self-starting problem of H-Darrieus wind rotor, optimal blade
number should be three with cambered or unsymmetrical airfoil
shape. It has also been seen that high solidity three-bladed H-
Darrieus rotor has much better prospect in terms of starting torque
and performance. Most of the existing studies have been done with
low solidity and high TSR for rotor power generation, for which the
high solidity rotors might not be appropriate as suggested by
various researchers in the past [1,13,15,19]. But high solidity and
low TSR unsymmetrical H-Darrieus rotors give some promising
things like: it gives higher static and dynamic torque [3,4] than the
symmetrical blade rotors. In low TSR power coefcient of such
rotors are less, but in terms of dynamic torque these designs are
very signicant for small scale applications like pumping, grinding,
sailing boat etc. So in such environment H-Darrieus rotor perfor-
mance data is needed. In this paper, along with rotor solidity and
aspect ratio like the past researches, starting time has been
considered an important parameter for improving the perfor-
mances of cambered blade H-Darrieus rotor, which could increase
the energy yield in low wind regimes. Moreover, there is still very
less work on the self-starting evaluation, static and dynamic torque
coefcient measurement etc. of high solidity H-Darrieus rotor
having cambered blades at different rotor aspect ratios, so that
some clear perception can be gained for this type of VAWT rotor in
low wind regime. It has also been observed that unsymmetrical
airfoils can exhibit higher tangential thrust which helps them to
offer better self-starting features than symmetrical airfoils, and
thicker cambered blades are also better for good self-starting of a
rotor [3,4,21,22]. Therefore, unsymmetrical S815 airfoil prole has
been selected here as its thickness is higher than other popular
NREL S-series airfoils like S818, S809, S1210 etc. [23], and its
detailed performance of H-Darrieus rotor with S815 blades has not
been investigated comprehensively in the past. Another unsym-
metrical EN0005 blade is chosen as its performance is better than
NACA 0018 and NACA 4418 airfoil in case of self-starting [24]. So, in
this paper experimental analysis has been done on symmetrical
538 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547

Fig. 1. (a) NACA 0018 blade prole (b) S815 blade prole (c) EN0005 blade prole (d) NACA 0018 rotor blades (e) S815 rotor blades (f) EN0005 rotor blades (g) Schematic diagram of
the system set-up (h) Real image of the experimental set-up.
 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547 539

Flow rate
control
wheel

Blower casing
Convergent divergent nozzle Coupling

(a) (b)
10 10 4 m/s
4 m/s

Wind speed (m/s)

Wind speed (m/s)

8 8 6 m/s
6 m/s
6 8 m/s
8 m/s 6
Average
4 4
(4 m/s)
Average
2 2 (6 m/s)
0.8 1 1.2 0 0.08 0.16 0.24 0.32 Average
H/D Width of the rotors (m) (8 m/s)

(c) (d)
Fig. 2. (a) Schematic diagram and (b) actual diagram of centrifugal blower apparatus (c) wind speed variation for three H/D conditions (d) wind speed variation along the width of
the rotors.

NACA 0018 and unsymmetrical S815 and EN0005 three-bladed H- The solidity (s) of the rotors has been calculated using Eqn. (1)
Darrieus rotors at different rotor aspect ratios (0.9e1.1) for three [3] where the rotors diameters (D) can be measured from the three
different low wind speed conditions (4 m/s, 6 m/s and 8 m/s). NACA considered aspect ratios (H/D) namely 0.9, 1.0 and 1.1. For each of
0018 airfoil prole is chosen for comparison purpose as it is more the rotors the blade numbers (N) has been kept 3.
popularly used in case of H-Darrieus rotors and has high power
coefcient than various NACA series airfoils [1,24]. Main goal of this Nc
s (1)
paper is to determine the best H-Darrieus rotor conguration that D
has better self-starting features, lower starting time, and higher
dynamic performances at low tip speed ratios compared to existing
symmetrical and unsymmetrical blade H-Darrieus rotor designs.
4. Experimental procedure

3. Design and fabrication of H-Darrieus rotors

The H-Darrieus rotors mounted on the structure were placed in
open air at the exit of a centrifugal blower tunnel test rig whose
For fabricating a three bladed H-Darrieus rotor, rstly the
schematic and actual diagram can be seen in Fig. 2(a) and (b). An
outline of NACA 0018, S815 and EN0005 shapes were generated
electric motor having 10 hp and rated rpm 2800 ran the blower. A
which can be seen in Fig. 1(a), (b) and (c). Then following the
hand wheel is there on the blower which helps in controlling the
procedure mentioned by Singh et al. [3], three numbers of blades
wind speed as per requirement. Turbulent intensity of the blower
for all the types were made that have been pictured in Fig. 1(d), (e)
was measured to be less than 1%. Now since the momentum of the
and (f). Material is chosen to be balsa wood as it is of light weight,
ow coming out of the ducted blower is diffused, therefore the
has high strength to weight ratio and stiff enough to withstand the
position of the rotors were rstly settled in the free-stream zone for
wind thrust [3]. Several smaller and experimental research purpose
obtaining uniform wind speed across the rotors. From the blower
VAWT rotor blades are made of wood as it is readily available, easily
opening three different distances were found by adjusting the wind
shaped and contains good fatigue properties [25]. Here, each of the
ow rate where almost uniform wind stream (1%) was owing
blades has height (H) 29 cm and chord (c) 5 cm. These wooden
throughout the swept area of the rotors for all the three considered
blades were connected to a MS shaft of diameter 1.2 cm and length
H/D ratios. Finally for calculation purpose wind speed at the mid-
70 cm by means of fully threaded circular struts. Single struts were
section of the rotors was considered. Velocity of wind was
used for mounting each blade at their mid chord positions. The
measured with the help of a digital vane type anemometer of range
three numbers of struts, which were joined by arc welding to the
0e30 m/s and accuracy level of 1%. The rpm of the rotors was
central shaft, were 120
apart from each other, 19 cm long and
measured by a digital non-contact tachometer whose accuracy
4 mm in diameter. The aspect ratio (H/D) and hence solidity of the
level was also 1%. The starting time of the rotors at different
rotors were controlled by nut and bolt arrangement on the struts.
azimuthal angular positions was measured by a digital stopwatch
The three H/D ratios are 0.9, 1.0 and 1.1 for which the solidity values
whose least count was 0.01 s. It was observed that the overall un-
are calculated to be 0.46, 0.51 and 0.56 respectively that were used
certainty (j0 ) of these instruments as given by Eqn. (2) [3,4] was
for the rotors. For positioning the rotors centre i.e. the shaft, two
1.414%.
ball bearings have been used at each end of the supporting struc-
ture in which the rotors are mounted. The bearings are cleaned and h i1
j0 jAM 2 jTM 2
2
washed in petrol and lubricating oil is used periodically for mini- (2)
mizing friction. The schematic diagram of the rotors and actual
experimental set-up picture are shown here in Fig. 1(g) and (h). where jAM and jTM are the uncertainties in the Anemometer and
540 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547

S815, wind speed 4 m/s NACA 0018, wind speed 4 m/s

0.09 0.09

Static torque coeff

Static torque coeff

0.9 H/D 0.9 H/D
0.07 0.07
1.0 H/D 1.0 H/D
0.05 0.05
1.1 H/D 1.1 H/D
0.03 0.03
0.01 0.01
0 60 120 180 240 300 360 0 60 120 180 240 300 360
Azimuthal angle (degree) Azimuthal angle (degree)

(a) (b)

S815, wind speed 6 m/s NACA 0018, wind speed 6 m/s

0.07 0.9 H/D 0.07 0.9 H/D

Static torque coeff

Static torque coeff

0.05 1.0 H/D 0.05 1.0 H/D

1.1 H/D 1.1 H/D
0.03 0.03
0.01 0.01
0 60 120 180 240 300 360 0 60 120 180 240 300 360
Azimuthal angle (degree) Azimuthal angle (degree)

(c) (d)

S815, wind speed 8 m/s NACA 0018, wind speed 8 m/s

0.07 0.07
0.9 H/D 0.9 H/D
Static torque coeff

Static torque coeff

0.05 1.0 H/D 0.05 1.0 H/D

1.1 H/D 1.1 H/D
0.03 0.03
0.01 0.01
0 60 120 180 240 300 360 0 60 120 180 240 300 360
Azimuthal angle (degree) Azimuthal angle (degree)
(e) (f)
EN0005 at H/D = 1.0 EN0005 at H/D = 1.1 4 m/s
0.07 4 m/s
0.07
Static torque coeff

Static torque coeff

6 m/s 6 m/s
0.05
8 m/s 0.05 8 m/s
0.03 0.03

0.01 0.01
0 60 120 180 240 300 360 0 60 120 180 240 300 360
Azimuthal angle (degree) Azimuthal angle (degree)

(g) (h)
Fig. 3. Static torque coefcient vs. azimuthal angle positions of the rotors at three different aspect ratios and wind velocities.

Tachometer readings respectively.

 
Fig. 2(c) shows the variation of wind speed with respect to ro- T M  S  rshaft ds  g (3)
tors aspect ratio. It can be seen that almost all the data points of
upstream wind speed falls on the straight line for the selected
aspect ratios which are 0.9, 1.0 and 1.1 which means that the wind
ow is continuously repeatable at all H/D ratios. After that, the ow Table 1
uniformity is also ensured which is supported by Fig. 2(d) that Static torque coefcient vs. azimuthal angle positions of the rotors at 1.1 H/D and
4 m/s wind velocity.
represents the wind speed variation across the width of the rotors.
The deviation of the upper and lower limit values from the average Azimuthal angle (degree) S815 NACA 0018 EN0005
wind speed was within 1% and this average wind speed was taken 0 0.033 0.032 0.031
for this present experimentation purpose. 30 0.072 0.068 0.070
To calculate the torque of the rotors, a rope brake dynamometer 60 0.043 0.044 0.042
arrangement was used as shown in Fig. 1(c). A nylon string was 90 0.034 0.033 0.033
120 0.033 0.032 0.031
twined around a pulley attached to the rotors shaft through which a 150 0.072 0.068 0.069
brake load was hung and the other end of the string was tied to the 180 0.043 0.044 0.045
dynamometer spring balance to measure its reading. The me- 210 0.034 0.033 0.031
chanical torque (T) produced by the rotors shaft was determined by 240 0.033 0.032 0.031
270 0.072 0.068 0.069
noting down the weights and the spring balance readings and later
300 0.043 0.044 0.045
putting these data in the following Eqn. (3) [4]: 330 0.034 0.033 0.034
360 0.033 0.032 0.031
 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547 541

NACA 0018 4 m/s S815 4 m/s

2.6 2.6 6 m/s
6 m/s
2.2 2.2 8 m/s
8 m/s

TSR

TSR
1.8 1.8
1.4 1.4
1 1
0.8 0.9 1 1.1 1.2 0.8 0.9 1 1.1 1.2
H/D H/D

(a) (b)

EN0005
2.6 4 m/s
2.2 6 m/s
TSR 1.8 8 m/s
1.4
1
0.8 0.9 1 1.1 1.2
H/D

(c)
Fig. 4. Tip speed ratio vs. aspect ratio (H/D) for (a) NACA 0018 rotor, (b) S815 rotor and (c) EN0005 rotor.

S815 0.9 H/D

0.15 NACA
NACA0018
0018 0.9 H/D
0.15
Dynamic torque coeff
Dynamic torque coeff

1.0 H/D 1.0 H/D

0.1 0.1 1.1 H/D
1.1 H/D
0.05 0.05

0 0
0.6 1 1.4 1.8 2.2 2.6 0.4 0.8 1.2 1.6 2 2.4 2.8
TSR TSR

(a) (b)

0.15 EN0005
0.9 H/D
Dynamic torque coeff

1.0 H/D
0.1
1.1 H/D

0.05

0
0.4 0.8 1.2 1.6 2 2.4
TSR

(c)
Fig. 5. Dynamic torque coefcient vs. tip speed ratio for (a) NACA 0018 (b) S815 and (c) EN0005 rotors.

where M is the weight, S is the spring balance reading, rshaft is the

radius of the rotor shaft, ds is the nylon string diameter and g is the 4T
Ct (5)
gravitational acceleration. Tip speed ratio l, is the ratio of the blade rU 2 D2 H
speed to the free-stream wind speed and can be determined using
the following Eqn. (4) [3,4]:
Cp l  Ct (6)
uD
l (4) where r is the density of air, Ct is the torque coefcient and Cp is the
2U
power coefcient.
where u is the angular velocity, D is the rotor diameter and U is the The blower experiments were performed under controlled en-
free-stream wind velocity. The performance of rotor is determined vironments such as uniform wind across the cross-section of the
by nding the torque and power coefcients which can be obtained wind rotor, repeatable wind speed for a set upstream condition,
from the following Eqns. (5) and (6) [3,4]: negligible upstream turbulence etc. However these limits the scope
542 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547

rotor inertia in the scaled-up designs. But if the inertia can be

0.04 Bearing maintained such that it matches with that of the experimental
0.03 power model on reduced scale, like making the blade thin while main-
Power (W)

taining its thickness-to-chord ratio, camber position, percentage

0.02 Struts camber etc. xed as the experimental model, then similar results
parasitic may again be expected. Even then this means the blade might
0.01
become weaker in the radial direction due to reduced thickness
power
0 thereby making the blade failure-prone under strong radial loads.
80 160 240 320 Therefore in that event the results should be interpreted without
considering the radial load carrying ability of the rotor blades. Thus
RPM (N) this also imposes a limit on the scalability aspect.
Fig. 6. Power loss due to bearing friction and struts parasitic drag.

NACA 0018 0.9 H/D S815 0.9 H/D

0.2 0.2
1.0 H/D 1.0 H/D
0.15 1.1 H/D 0.15
1.1 H/D

CP
0.1 0.1
CP

0.05 0.05
0 0
0.4 0.8 1.2 1.6 2 2.4 0.4 0.8 1.2 1.6 2 2.4
TSR TSR

(a) (b)

0.2 EN0005 0.9 H/D

1.0 H/D
0.15
1.1 H/D
Cp

0.1
0.05
0
0.4 0.8 1.2 1.6 2 2.4
TSR

(c)
Fig. 7. Power coefcient vs. tip speed ratio for (a) NACA 0018 and (b) S815 and (c) EN0005 rotors.

of such experiments as controlled environments will not be always 4.1. Error analysis
physically possible in reality. Further the experiments have been
performed in scaled down models as the outlet diameter of the In most experimental works there exist some unavoidable un-
blower test rig was not very large. This also limits the scope of the certainties. After obtaining the uncertainty calculations of
present experiments. As far as scalability is concerned, similar Anemometer and Tachometer, now the uncertainties of power co-
performances can be expected if the rotor solidity is kept constant efcient and torque can be evaluated by Eqns. (7) and (8). These
and the ow patterns in the upstream conditions also do not vary errors can be computed by using the uncertainty method of Kline
idealistically. Another constraint with scalability is the increase of and McClintock [3,26].
s
 2  2   2  
dCp du dU dT dU 2
2 (7)
Cp u U T U

where torque uncertainty can be computed as:

dT dS
(8)
T S

5. Results and discussions

To nd out the self-starting capability of the rotors, an experi-

Fig. 8. Upper and lower limit of wind velocity variation of 6 m/s wind speed and for 4% mental investigation was performed to assess the static torque
turbulence intensity level. coefcients at different azimuthal angular positions over a

0.132
Dynamic torque coeff
0.128
0.124
0.12
5.6
Fig. 9. (a) Dynamic torque coefcient vs. wind velocity and (b) Cp vs. wind velocity for S815 rotor.
complete rotation cycle for three different rotor aspect ratios like
0.9, 1.0 and 1.1 and three different wind speeds from 4 to 8 m/s. At
rst, the graphs have been plotted for S815 and NACA 0018 in Fig. 3
by varying the H/D ratios with respect to a particular wind speed
for the rotors. From these graphs it has been observed that at 1.0
and 1.1 H/D ratio the average static torque coefcient values are
very close to each other which is somewhat higher than 0.9 H/D
case. Therefore, for EN0005 rotor the static torque coefcient
graphs have been generated only for 1.0 and 1.1 H/D conditions
varying the selected wind speeds. It can be noticed that for all the
graphs the static torque values are cyclic with a periodicity of
around 120
for these three bladed rotors. From the graphs it can
also be seen that the rotors have self-starting ability for three
different wind velocities at all the three aspect ratios as they show
positive static torque coefcients at all rotor azimuthal angle po-
sitions. It can be observed that irrespective of wind speed condi-
tion, these static torque coefcients increases as the rotors aspect
ratio or solidity increases i.e. increased solidity leads to improved
self-starting capability during starting of the rotors.
All the experiments are conducted in low wind speeds and the
models are also of small scale. Therefore much variations in static
torque can not be expected for the selected H-Darrieus rotors. But if
carefully noticed, small variations in static torque can be observed
which has been given in Table 1 for 1.1 H/D as a reference.
Fig. 4 represents the variation of rotor TSR with aspect ratios (H/
D) at zero-load condition. It can be seen that TSR for the rotors
rstly increases with the increase of H/D to a certain limit but then
it drops with any further increase of the aspect ratios for all three
wind speed environments. That means for H/D ratio 0.9 and 1.0, all
the rotors can extract more power from the owing wind than H/D
ratio 1.1. It might be due to early transition to turbulent ow with
increase of wind speed leading to early separation. So it can be
noted that for any value of wind velocity, aspect ratio 1.0 is the
optimum value for the rotors which leads to higher TSR values. It
1.4
Torque ripple factor
1.2
1 operating range or tip speed ratio and wind speed conditions.
0.8
0.6
2 4 6 8
Wind velocity
Fig. 10. Variation of torque ripple factor vs. wind velocity for NACA 0018, S815 and
EN0005 blade rotors.
A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547 543
S815 S815
1.0 H/D 0.2 1.0 H/D
0.19
Cp
0.18
0.17
5.8 6 6.2 6.4 6.6 5.6 5.8 6 6.2 6.4 6.6
Wind velocity (m/s) Wind velocity (m/s)
(a) (b)
can also be said that at H/D ratio 1.0 the rotors will extract highest
power from wind than the other two ratios in these wind speed
conditions.
Fig. 5(a) and (b) shows the variation of dynamic torque coef-
cient with tip speed ratio (TSR) for the rotors. It is observed that for
the rotors, the dynamic torque is highest for H/D 1.0 and lowest
for H/D 0.9 and S815 blade rotor has higher dynamic torque than
the other two. The highest dynamic torque can be seen in case of
S815 at H/D 1.0 at TSR 1.48 and for the rotors as the TSR value
increases the dynamic torque decreases. It can also be seen that the
operating range of S815 blade rotor has been increased compared
to NACA 0018 and EN0005 blade rotors and average value of dy-
namic torque also has been increased in case of unsymmetrical
S815 rotor. Here the power loss of the rotors due to bearing friction
and parasitic drag of struts also has been plotted with respect to
rpm of the rotors as shown in Fig. 6. This graph represents the
average range of power loss which is almost same for the rotors. It
can be noticed that bearing power loss is higher than struts para-
sitic drag loss for the rotors within the operating range considered
but both are less than 0.03 W.
In Fig. 7 three graphs have been plotted to nd the value of
power coefcient (Cp) with respect to tip speed ratio for the rotors
at three selected values of H/D i.e. 0.9, 1.0 and 1.1. It can be seen that
for all H/D, Cp values for the rotors rstly increase as TSR value
increases up to some limit, but after that fall down with further
increase of TSR. This means an optimum value of TSR can be found
for which Cp reaches its maximum value. Here the operating range
for the rotors is almost equal but the power coefcient (Cp) of un-
symmetrical S815 rotor is higher than symmetrical NACA 0018 and
unsymmetrical EN0005 rotors. In case of S815 rotor the maximum
Cp of 0.19 has been obtained at TSR value 1.48 and for H/D 1.0
whereas for NACA 0018 rotor maximum Cp of 0.16 has been found
out at TSR value 1.43 and for H/D 1.0 and for EN0005 rotor
maximum Cp of 0.175 has been found out at TSR value 1.52 and for
H/D 1.0. Most of the Cp values for three different aspect ratios are
also higher for cambered bladed rotor than symmetrical one. The
uncertainty in power coefcient as calculated using Eqn. (7) is in
NACA 0018 2.2%e5.5% range and the same using Eqn. (8) for torque lies in the
range of 0.3%e0.6%.
EN0005 The above analysis reveals that the performance of the three-
S815 bladed H-Darrieus rotors is dependent on aspect ratio or solidity,
Finally it has been found that in these selected wind speed condi-
tions, aspect ratios and TSR range, three-bladed cambered S815 H-
Darrieus rotor is more efcient than symmetrical NACA 0018 and
unsymmetrical EN0005 blade rotors.
10 For wind generated from the blower and scaled experiments in
laboratory environment, we maintain ow uniformity, turbulence
intensity etc. within acceptable limits of laboratory testing. How-
ever in reality, these might not be always true, for e.g., ow may be
544 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547

NACA 0018 at H/D = 0.9 S815 at H/D = 0.9

7
4 m/s

Starting time (second)

4 m/s

Starting time (second)

6 7
6 m/s 6 6 m/s
5
5 8 m/s
4 8 m/s
4
3 3
2 2
0 60 120 180 240 300 360 0 60 120 180 240 300 360
Azimuthal angle (degree) Azimuthal angle (degree)

(a) (b)

NACA 0018 at H/D = 1.0 S815 at H/D = 1.0

Starting time (second)

7 6 4 m/s
4 m/s

Starting time (second)

6 5 6 m/s
6 m/s
5 8 m/s
8 m/s 4
4
3 3
2 2
0 60 120 180 240 300 360 0 60 120 180 240 300 360
Azimuthal angle (degree) Azimuthal angle (degree)

(c) (d)
NACA 0018 at H/D = 1.1
7 S815 at H/D = 1.1
Starting time (second)

4 m/s 6 4 m/s
Starting time (second)

6
5 6 m/s 5 6 m/s
4 8 m/s 4 8 m/s
3 3
2 2
0 60 120 180 240 300 360 0 60 120 180 240 300 360
Azimuthal angle (degree) Azimuthal angle (degree)

(e) (f)

EN0005 at H/D = 1.0

6 4 m/s
Starting time (second)

5 6 m/s
8 m/s
4
3
2
0 60 120 180 240 300 360
Azimuthal angle (degree)

(g)
Fig. 11. Starting time vs. azimuthal angle positions of all the rotors at three different wind velocities.

non-uniform and turbulence may be induced in the ow. From the (intensity 4%), the rotor dynamic torque and power coefcient
present experimental analysis, it has been found that for 6 m/s varies in the range of 0.1242e0.13 and 0.176e0.195 respectively.
wind speed and 1.0 H/D condition the S815 rotor showed highest The standard deviations for dynamic torque coefcient and power
power coefcient of 0.19, and these optimal conditions have been coefcient have been calculated to be 0.003 and 0.007 respectively.
considered to match with the real conditions. Now to connect with A non-dimensional parameter torque ripple factor (g) is dened
reality rstly we have considered steady turbulence of turbulence as the difference between the maximum torque Tmax and the
intensity 4% as well as ow non-uniformity across the rotor cross- minimum torque Tmin compared to the average torque Tavg and can
section of 1 in 30 which is presented in Fig. 8. Considering turbu- be written as Eqn. (9):
lence intensity 4%, the wind speed ranges in between 5.76 m/s
and 6.24 m/s. Again considering ow non-uniformity of 1 in 30,
Tmax  Tmin
three more wind velocities viz. 5.9 m/s, 6.1 m/s and 6.5 m/s are also g (9)
created. Now in these ve selected wind speeds, experiments Tavg
have been conducted and the dynamic torque and power co- The parameter g, which depends on blade designs and wind
efcients have been calculated which are presented in Fig. 9(a) and velocities, stands for stableness and smoothness of drive of rotor
(b) respectively. From the diagrams it can be seen that with which is very essential for safe performance of the rotor. High value
the consideration of non-uniform ow, and steady turbulence of this factor in case of wind rotors is primarily unwanted as it
 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547
causes noise and vibration in the system which may shorten the
lifetime of the wind rotor system [27]. High torque ripple factor can
change the system's geometry which can reduce its general per-
formance. When steady torque is generated, this factor equals to 1
[4,28].
Fig. 10 shows the variation of torque ripple factor (TRF) with
respect to wind velocity considered for the rotors. It is observed
that for safe operation of the rotors, 4 m/s and 6 m/s wind velocities
display low TRF, but velocity 4 m/s is much better as for this case g
is close to 1. So velocity 4 m/s is the optimal wind speed for these
rotors for which vibrations, noise and fast wearing of the rotors can
be avoided and system's life-time also can be increased.
For the selected three different wind speeds (4 m/s, 6 m/s and
8 m/s) minimum starting time of the rotors has also been measured
and plotted in Fig. 11. For S815 and NACA 0018 the graphs have been
plotted for all three H/D ratios. Then from the diagrams it has been
found that for H/D ratio 1.0 the rotors show lowest average starting
time than the other two H/D ratios and so for EN0005 rotor, the
starting time has been plotted for only 1.0 H/D ratio. This starting
time was calculated by considering start of rotation of the blades
from different azimuthal positions. This starting time is the time
required for the rotors to start a steady rotation for the respective
wind speed conditions i.e. the minimum time when sufcient
torque is generated as the wind rotor makes one complete rotation,
and after that its blade speed gradually increases to its rated con-
dition. As the wind speed increases from 4 m/s to 8 m/s, this
starting time also gets reduced. The starting time has been
measured by keeping a particular H/D ratio constant and varying
the wind velocities. It can be noticed that from 0
rotor azimuthal
position the starting time starts increasing up to the range of
120
e150
and then at 180
angular position both the rotors show
longest starting time for any H/D conditions and for any considered
wind speed. Most of the cases it is seen that at 0
e120
range and
330
e360
range the rotors show shorter starting time.
From the graphs it is also observed that the average starting
time for EN0005 rotor is lesser than S815 and NACA 0018 rotors. As
an example for H/D 1.0, it is seen that for wind speeds 4 m/s, 6 m/
s and 8 m/s the averaged minimum starting time (for all rotor an-
gles) needed for EN0005 rotor is 4.1 s, 3.4 s and 2.7 s whereas for
S815 rotor is 4.5 s, 3.8 s and 2.8 s and the same for NACA 0018 rotor
is 4.8 s, 3.9 s and 3 s respectively. As the cambered shape blade
provides higher tangential thrust so may be for this reason the
higher cambered blade EN0005 is taking somewhat lesser time
than the other two rotors. This time difference may seem very less,
but for a greater time period considered for annual energy yield,
this may perhaps show much better results for unsymmetrical
EN0005 rotor.
While comparing the self-starting with respect to various pre-
vious works on three-bladed H-Darrieus rotors with different blade
shapes, it can be seen that the static torque coefcients of the
0.15
Static torque coeff
0.1
0.05
0 S818
0 60 120 180 240 300 360
-0.05
-0.1
Azimuthal angle (degree)
(a)
Fig. 12. (a) Comparison of static torque for different blade shapes (b) Comparison of Cp with TSR for different blade shapes.
545
present S815 rotor are higher than some well-established results of
existing H-Darrieus rotors. One such work done by Mohamed [19]
is reproduced here in Fig. 12(a), which shows that, although in all
the cases static torque coefcient values are positive, S815 H-Dar-
rieus rotor shows much higher values (around 2 times) than S1046
blade rotor. Another analysis done with S818 bladed H-Darrieus
rotor by Bhuyan and Biswas [4] is also reproduced which shows
that S818 blade rotor generates some negative torque values in
several rotor angles which is unlike the present S815 rotor i.e. S815
has better self-starting potential compared to S1046 and S818 ro-
tors. Fig. 12(b) shows the variation of Cp with respect to TSR for
three different rotors-one is the present S815 rotor and the others
are NACA 0015 [29] and NACA 0021 [30] from past researches. The
graph shows that the two literature results have been taken for
those selected ranges of TSR at which the present results are ob-
tained. Though for NACA 0015 and NACA 0021 rotors the solidity
was kept 0.1 which is lesser than the present study, then also in this
TSR range it is seen that the Cp is much higher for S815 rotor
comparatively. Therefore in this TSR range corresponding to the
selected wind speed conditions S815 blade rotor shows better
performance. This present rotor would show higher Cp if the TSR is
increased that is when the wind velocity is also increased. But in
this paper we have considered wind speed up to 8 m/s and within
this velocity and corresponding to that considered TSR range S815
rotor shows better results.
6. CFD simulation of H-Darrieus rotors
The 2D CFD simulation of the three bladed H-Darrieus rotors for
the three different airfoil shapes is performed by ANSYS 14.0 soft-
ware. The computational domain is made such that the boundaries
are far enough from the rotor diameter to reduce the boundary
effects. The upstream and side boundaries are 5 times the diameter
of rotor axis and the downstream boundary is 10 times the diam-
eter away from the same [22]. There is a circular interface that
separates the rotating inner zone to the outer stationary zone.
Various levels of renement is done to attain the Grid Indepen-
dence Limit (GIL) mesh. Unstructured triangular meshing has been
done to complement the complex blade designs. The grids around
the blade peripheries are much denser than those far away from the
turbine. The skewness is kept under 0.5 for the domain geometry.
The RNG ke turbulence model [13,15,31] with standard wall
function is used for simulation of the rotors. The different input
parameters like wind velocity, rpm, air density etc. are same for all
three cases, and the simulations are run for the optimum condition
of the rotors at which their performances are the maximum. The
SIMPLE algorithm is used for pressureevelocity coupling and the
minimum convergence criteria is set to be 1e-06.
In the above diagrams, the contour plots of static pressure have
been plotted to understand the ow physics of the H-Darrieus
S815
(1.0 0.2 S815
H/D) 0.16
S1046 NACA
0.12
0015
Cp
0.08
0.04
(1.0
H/D) 0
0.4 0.8 1.2 1.6 2 2.4 2.8
TSR
(b)
546 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547
Fig. 13. Contour of static pressure (Pascal) for (a) S815 rotor (b) EN0005 rotor and (c) NACA 0018 rotor.
rotors. For all the three rotors the blades at 8 o'clock position have
been considered as the advancing blade. The contour plots show a
decrease of static pressure form the upstream side to the down-
stream side of the rotors. In Fig. 13(a), the static pressure for S815
rotor decreases from 5.56  101 Pa in the upstream side
to 1.31  101 Pa in the downstream side across the advancing
blade. In Fig. 13(b), the static pressure for EN0005 rotor decreases
from 4.77  101 Pa in the upstream side to 8.56  100 Pa in the
downstream side across the advancing blade. Similarly in Fig. 13(c),
the static pressure for NACA 0018 rotor changes from 3.32  101 Pa
in the upstream side to 1.81  101 Pa in the downstream side
across the advancing blade. Therefore, the difference of static
pressure between upstream and downstream side is higher for
S815 H-Darrieus rotor than the other two rotors meaning greater
lift force than the rest. Therefore this ow physics gives an idea why
the performance of the S815 blade H-Darrieus rotor is better than
the remaining two H-Darrieus rotors. Further these diagrams show
another important ow physics in terms of the tendency towards
ow separation. In Fig. 13(a), for S815 blade there is a viscous shear
layer on the suction side of the advancing blade adhering to its
blade surface. Then also its performance is expected to be better as
the layer is not separated. In Fig. 13(b), for EN0005 blade, a sepa-
rated vortex is clearly seen on the suction side of the blade in 3/4th
quarter chord of the blade and so its performance is lower than
S815 blade rotor. In Fig. 13(c), for NACA 0018 blade, the ow is
separated earlier than rest of the rotors as its can be seen that there
is a diffusion of the separated vortices among all the blades, and
therefore its performance is least among all the three rotors
studied.
So it can be said that this cambered H-Darrieus rotor has high
potentiality to be used for several small-scale applications in low
wind speed environment for power generation purpose. Further
research for scaled-up systems using this cambered S815 blade
with high solidity can be done to improve its operating range which
may increase its performance and thus it can be popularly used in
remote areas where grid-connected electricity is limited.
7. Conclusions
In this paper a three-bladed H-Darrieus rotor with unsymmet-
rical S815 blades is compared with another two three-bladed H-
Darrieus rotors with symmetrical NACA 0018 blade and unsym-
metrical EN0005 blade in same operating conditions to investigate
their self-starting features and performance. Firstly to check their
self-starting ability, static torque coefcients at all rotor azimuthal
angles for three different H/D ratios (0.9, 1.0 and 1.1) are found out.
Then their power coefcients (Cp) are calculated for these three H/D
ratios at several TSR values. Further minimum starting time for
steady rotation is also measured and the ow physics have been
studied through 2D CFD simulation in terms of static pressure and
vortex formation for the rotors. All these results are analysed to
make a clear comparison based on self-starting characteristics and
 A.R. Sengupta et al. / Renewable Energy 93 (2016) 536e547
performance between symmetrical and unsymmetrical airfoils of
three-bladed H-Darrieus rotor. From the present study the
following conclusions can be summarized:
 All the rotors exhibit positive static torque coefcients for all
rotor angular positions and for all three H/D ratios considered.
These positive values indicate that the rotors possess self-
starting characteristics as there is no sign of negative torque
coefcients during their steady rotational movements.
 All the rotors show favourable dynamic torque in this TSR range
and low wind conditions but S815 blade rotor possesses higher
dynamic torque than NACA 0018 blade and EN0005 blade rotors
for the considered three H/D ratios.
 Optimum rotor H/D ratio has been found to be 1.0 for which the
rotors show highest Cp, and the maximum Cp values are 0.19 for
S815 rotor, 0.175 for EN0005 rotor and 0.16 for NACA 0018 rotor
at TSR value 1.48, 1.52 and 1.43 respectively. Therefore, 18.7% and
9.4% increase of power coefcient can be observed in case of
cambered S815 blade and EN0005 blade rotors at H/D 1 or
solidity 0.51 compared to NACA 0018 rotor. So in this TSR range
and corresponding to the selected wind speed range S815 rotor's
performance is better.
 EN0005 blade H-Darrieus rotor exhibits minimum starting time
from static position compared to S815 and NACA 0018 blade H-
Darrieus rotors which can be benecial for the former one to
show higher performance in longer time period of operating
condition.
 In low tip speed ratio and in low wind speed conditions (4e8 m/
s), the present unsymmetrical S815 blade H-Darrieus rotor
possesses higher power coefcient than EN0005 and conven-
tional NACA 0018 blades which has been found here in addition
to NACA 0015 [29] or NACA 0021 [30] blade H-Darrieus rotors
which have been reproduced from existing literature.
 Both the rotors show that at 4 m/s wind speed, torque ripple
factor is close to 1 which means that at 4 m/s wind speed these
rotors are capable of generating almost steady torque along with
lower noise level, vibrations and greater service time.
 The insights of higher performance of S815 blade rotor has been
obtained from the ow physics study of all the rotors. It has been
seen that S815 rotor shows greater pressure difference across
the advancing blade meaning better aerodynamic performance
of this rotor. Again, the viscous layer on the downstream side is
attached with the blade compared to the corresponding blade of
the remaining rotors and thereby complementing better per-
formance of S815 blade rotor.
The present research has shown that high solidity unsymmet-
rical S815 blade H-Darrieus rotor possess higher static and dynamic
torque coefcient and higher power coefcient than the symmet-
rical NACA 0018 and unsymmetrical EN0005 blade H-Darrieus ro-
tors. If this present design is scaled up then also its performance
may be investigated to compare the results with these ndings.
Further, this work might direct future researchers towards more
investigations on high solidity cambered blade rotors to improve
their starting features and power coefcients on larger scale for
increasing energy yield in low wind speed conditions.
References
[1] M.H. Mohamed, Performance investigation of H-rotor Darrieus turbine with
547
new airfoil shapes, Energy 47 (2012) 522e530.
[2] S. Errikson, H. Bernhoff, M. Leijon, Evaluation of different turbine concepts for
wind power, Renew. Sustain. Energy Rev. 12 (5) (2008) 1419e1434.
[3] M.A. Singh, A. Biswas, R.D. Misra, Investigation of self-starting and high rotor
solidity on the performance of a three S1210 blade H-type Darrieus rotor,
Renew. Energy 76 (2015) 381e387.
[4] S. Bhuyan, A. Biswas, Investigations on self-starting and performance char-
acteristics of simple H and hybrid H-Savonius vertical axis wind rotors, Energy
Convers. Manag. 87 (2014) 859e867.
[5] R. Gupta, A. Biswas, K.K. Sharma, Comparative study of a three-bucket Savo-
nius rotor with a combined three-bucket Savonius-three-bladed Darrieus
rotor, Renew. Energy 33 (9) (2008) 1974e1981.
[6] M. Taka, H. Kuma, T. Maeda, Y. Kamada, M. Oki, A. Minoda, A straight-bladed
vertical axis wind turbine with a directed guide vane row-effect of guide vane
geometry on the performance, J. Therm. Sci. 18 (2009) 54e57.
[7] I. Paraschivoiu, O. Trifu, F. Saeed, H-Darrieus wind turbine with blade pitch
control, Int. J. Rotating Mach. 2009 (2009) 1e7.
[8] J. DeCoste, A. Smith, D. White, D. Berkvens, J. Crawford, Self-starting Darrieus
Wind Turbine, Design Project Mech. 4020, Dalhousie University, Hallifax,
Canada, 2004.
[9] P. Bhatta, M.A. Paluszek, J.B. Mueller, Individual Blade Pitch and Camber
Control for Vertical Axis Wind Turbines, 2008. WWEC2008, Kingston, Canada.
[10] V.G. Dereng, Fixed geometry self-starting transverse axis wind turbine. United
States Patent 4264279. Available at: http://www.freepatentsonline.com/
4264279.html (accessed 29.05.15).
[11] R. Dominy, P. Lunt, A. Bickerdyke, J. Dominy, Self-starting capability of a
Darrieus turbine, Proc. Inst. Mech. Eng. Part A J. Power Energy 221 (2007)
111e120.
[12] J.R. Barker, Features to aid or enable self-starting of xed pitch low solidity
vertical axis wind turbines, J. Wind Eng. Ind. Aerodyn. 15 (1983) 369e380.
[13] P. Deshpande, X. Li, Numerical study of giromill-type wind turbines with
symmetrical and non-symmetrical airfoils, Eur. Int. J. Sci. Technol. 2 (8) (2013)
195e208.
[14] M. Islam, A. Fartaj, R. Carriveau, Design analysis of a smaller-capacity straight-
bladed VAWT with an asymmetric airfoil, Int. J. Sustain. Energy 30 (3) (2011)
179e192.
[15] H. Beri, Y. Yao, Effect of camber on self starting of vertical axis wind turbine,
J. Environ. Sci. Technol. 4 (3) (2011) 302e312.
[16] M.A. Sayed, H.A. Kandil, A. Shaltot, Aerodynamic analysis of different wind-
turbine-blade proles using nite-volume method, Energy Convers. Manag.
64 (2012) 541e550.
[17] R. Gupta, A. Biswas, CFD analysis of a twisted three-bladed H-Darrieus rotor,
J. Renew. Sustain Energy 2 (4) (2010) 043111e043126.
[18] D. Saeidi, A. Sedaghat, P. Alamdari, A.A. Alemrajabi, Aerodynamic design and
economical evaluation of site specic small vertical axis wind turbines, Appl.
Energy 101 (2013) 765e775.
[19] M.H. Mohamed, Impacts of solidity and hybrid system in small wind turbines
performance, Energy 57 (2013) 495e504.
[20] C. Nguyen, T. Le, P. Tran, A numerical study of thickness effect of the sym-
metric NACA 4-digit airfoils on self starting capability of a 1kW H-type vertical
axis wind turbine, Int. J. Mech. Eng. Appl. 3 (3e1) (2015) 7e16.
[21] B. Kirke, Evaluation of Self-starting Vertical Axis Wind Turbines for Stand-
alone Applications, Ph.D. thesis, Grifth University, School of Engineering,
1998.
[22] L.A. Danao, N. Qin, R. Howell, A numerical study of blade thickness and
camber effects on vertical axis wind turbines, Proc. Inst. Mech. Eng. Part A J.
Power Energy 226 (7) (2012) 867e881.
[23] D.A. Grifn, NREL Advanced Research Turbine (ART) Aerodynamic Design of
ART-2B Rotor Blades, NREL/SR-500e28473, 2000.
[24] N.C. Batista, R. Melicio, J.C.O. Matias, J.P.S. Catalao, New blade prole for
Darrieus wind turbines capable to self-start, in: IET Conference on Renewable
Power Generation, 2011, pp. 1e5.
[25] J.P. Tillman, Improvements to Vertical Axis Wind Turbine Blades to Aid in Self-
starting, 2011. Masters Theses. Paper 697, http://thekeep.eiu.edu/theses/697.
[26] S.J. Kline, F.A. McClintock, Describing uncertainties in single-sample experi-
ments, Mech. Eng. 75 (1953) 3e8.
[27] B. Yang, X.W. Shu, Hydrofoil optimization and experimental validation in
helical vertical axis turbine for power generation from marine current, Ocean.
Eng. 42 (2012) 35e46.
[28] A.R. Malipeddi, D. Chatterjee, Inuence of duct geometry on the performance
of Darrieus hydro turbine, Renew. Energy 43 (2012) 292e300.
[29] A. Rossetti, G. Pavesi, Comparison of different numerical approaches to the
study of the H-Darrieus turbines start-up, Renew. Energy 50 (2013) 7e19.
[30] M.H. Mohamed, A.M. Ali, A.A. Haz, CFD analysis for H-rotor Darrieus turbine
as a low speed wind energy converter, Eng. Sci. Tech. Int. J. 18 (2015) 1e13.
[31] Y. Chen, Y. Lian, Numerical investigation of vortex dynamics in an H-rotor
vertical axis wind turbine, Eng. Appl. Comp. Fluid Mech. 9 (1) (2015) 21e32.