Applied Energy 137 (2015) 117–125

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Wind tunnel experiments of a newly developed two-bladed

Savonius-style wind turbine
Sukanta Roy, Ujjwal K. Saha ⇑
Department of Mechanical Engineering, Indian Institute of Technology Guwahati, Guwahati 781039, India

h i g h l i g h t s

 Benchmark wind tunnel experiments on a novel Savonius-style wind turbine (SSWT).

 Analysis on the performance and starting characteristics of the newly developed wind turbine.
 Comparative examination of the newly developed turbine with other tested models.
 Discussion on the influences of Reynolds number and tip speed ratio.
 Estimation of overall efficiency and payback period for the wind energy conversion system (WECS).

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

Article history: Wind tunnel experiments have been conducted with a newly developed two-bladed Savonius-style wind
Received 17 May 2014 turbine specifically meant for a small-scale energy conversion. This novel shape of the turbine blade is
Received in revised form 5 October 2014 evolved from a series of experiments with different types of blades in the recent past. The developed
Accepted 6 October 2014
two-bladed turbine is tested in an open type test section and its performance is assessed in terms of
power and torque coefficients. Experiments have also been conducted with other standard blades such
as semi-circular, semi-elliptic, Benesh and Bach types in order to have a direct comparison. In this study,
Keywords:
all the reported experimental data are inclusive of wind tunnel blockage corrections. Further, the effects
Savonius-style wind turbines
Power coefficient
of Reynolds number on the dynamic and static characteristics are also discussed. The present investiga-
Torque coefficient tion demonstrates a gain of 34.8% in maximum power coefficient with the newly developed two-bladed
Static torque coefficient turbine.
Tip speed ratio Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction however, it possesses a lower efficiency as compared to its coun-

terpart. This class of wind turbines patented by Savonius S.J. in
Recent instabilities in the world energy market due to depletion 1920s is a modification of Flettner’s cylinder used to propel the
of fossil fuel sources, global warming threats and increasing price ships [11]. The maximum efficiency reported by Savonius was
of fossil derivatives necessitate the need of harnessing the clean 31%, though this claim was not agreed by subsequent investigators
and renewable sources of energy. In this context, research in the [12–14]. For a conventional SSWT, the maximum conversion effi-
field of wind energy is becoming particularly important [1–6]. ciency as reported by Sandia laboratory was found to be 24.4% [15].
Although considerable progress has been achieved in the area of Apart from this efficiency factor, SSWTs deliver a number of
horizontal axis wind turbines, the available technical designs of advantages over other wind turbines targeting small-scale energy
vertical axis wind turbines are not yet satisfactory for the develop- conversion such as design simplicity, ease of fabrication, ease of
ment of low cost reliable wind energy converters for conditions installation in confined spaces like rooftops, buildings or above
corresponding to the off-grid power generation at low wind the communication towers. Further, they can operate at low wind
speeds. Thus, attention is being paid to the small-scale wind tur- speeds causing lesser wear and tear, and their operation is inde-
bines for distributed energy systems [4–10]. The Savonius-style pendent of wind direction thereby avoiding the need of a yaw
wind turbine (SSWT) appears to be promising for such conditions, mechanism. [13,14,16–21]. Looking at these advantages, it can
clearly be said that if a higher efficiency can be achieved, these
⇑ Corresponding author. Tel.: +91 361 2582663; fax: +91 361 2690762. SSWTs would make an attractive source of power generation.
E-mail address: [email protected] (U.K. Saha).

http://dx.doi.org/10.1016/j.apenergy.2014.10.022
0306-2619/Ó 2014 Elsevier Ltd. All rights reserved.
118 S. Roy, U.K. Saha / Applied Energy 137 (2015) 117–125

Nomenclature

q density of air (kg/m3) DO end plate diameter (m)

x rotational speed of the turbine (rad/s) f blockage correction factor
h turbine angle to the wind direction (°) F mechanical load applied (N)
gG generator efficiency H height of the turbine (m)
gO overall efficiency of WECS Pavailable power available in the wind (W)
gT turbine efficiency Pturbine power produced by the turbine (W)
gTS transmission and storage efficiency R radius of rotation of the turbine (m)
A swept area of the turbine (m2) Re Reynolds number
BR blockage ratio SSWT Savonius-style wind turbine
CP power coefficient T dynamic torque on the turbine [Nm]
CT torque coefficient TS static torque on the turbine [Nm]
CTS static torque coefficient TSR tip speed ratio
d chord length of the blades (m) V free stream wind speed (m/s)
D outer diameter of the turbine (m) WECS wind energy conversion system

During the last 40 years, a host of researchers aimed to improve shown an improvement in terms of performance (power and tor-
the efficiency and starting performance of SSWTs not only through que coefficients) and aerodynamic characteristics (lift and drag)
parametric analysis, but also with multi-staging, and using various over the conventional SSWTs [28]. A comparison of the modified
augmentation techniques [14,21]. Most of these studies revolved Bach type over classical Bach type is discussed in the later sections.
around conventional semi-circular blades. A few investigators car- Fig. 2 shows the newly developed blade profile that consists of
ried out the performance analysis with helical and twisted blades. several individual arcs, where the dimensions are given with
Although noticeable performance gains were achieved, however, respect to chord length (d) of the blade. This novel blade profile
the designs reported seemed to have been more complex and was generated from a series of experimental studies on Bach and
expensive [22–25]. Improvement in the performance coefficients Benesh type turbines by altering the geometric arcs, overlap dis-
and static torque characteristics were also reported with the use tances and dimensions of blade profiles [30]. The impression of
of deflectors, guide vanes, nozzles or curtains placed upstream to the final blade shape was dictated by the experimental investiga-
the turbine blades. However, these made the turbine system more tions and the dimensions obtained were measured accurately.
complex and direction dependent [14]. Since one of the major ben- Based on this blade geometry, a pattern was fabricated which
efits of SSWT is its design simplicity, low cost and independency to was then used for manufacturing the turbine blades.
wind direction, a design modification having more complexity According to literatures, the use of end plates has shown an
should probably not be encouraging. In this aspect, some studies improvement in the performance of SSWTs [14,20,31]. It is mainly
with modified blade shapes were attempted reporting encouraging caused by the improved pressure difference between the concave
results. This has brought a renewed interest among the researchers and convex surfaces of SSWT blades. In the present work, the
to pursue further and detailed assessment of the blade design end plates are used at the top and at the bottom of turbine models.
[12,26–29]. Thus, in order to boost the applicability of these tur- For all the test cases, the height of the turbine (H) to end plate
bines, a novel but simple SSWT is developed and tested in a low diameter (Do = 1.1 D) ratio is maintained at 1 keeping
speed wind tunnel. The objective of the study is to improve the effi- H = Do = 230 mm. All the blades are fabricated from galvanized iron
ciency as well starting characteristics of SSWTs with this novel sheets of thickness 0.63 mm.
blade shape. Tests with other standard blades such as semi-circular,
semi-elliptic, Benesh and Bach types have also been conducted
under an identical platform in order to have a direct comparison. 3. Experimental setup

The experiments were carried out in a low speed wind tunnel

2. Description of blade profiles under open type test section as shown in Fig. 3. This experimental
facility mainly consists of: (i) a fan section followed by an inlet sec-
Figs. 1 and 2 show the dimensions of various blade profiles tion, (ii) diffuser section, (iii) settling chamber that contains the
tested in this experimental study. As seen from Fig. 1(a) and (b), coarse screen, a honeycomb structure and a fine screen to make
the overlap distance between the semi-circular and semi-elliptic the wind flow streamlined, (iv) a nozzle section to accelerate the
blade profiles are kept as 20% of the blade chord length [13,14]. flow, and (v) an open type test section which includes the turbine
The Benesh type blade profile (Fig. 1c) claimed to be one of the model and the torque measurement assembly. The wind tunnel
most efficient SSWT blade profiles [15] is also tested. Since long, structure is fabricated from galvanized iron sheets, whereas the
Bach type blade profiles are among the interests of investigators test section and tunnel stands are made of mild steel angles. The
for SSWTs [26–28]. Thus, a modified Bach type SSWT is selected dimensions of each section are demonstrated in Fig. 3. The frontal
from the recent investigations of Roy and Saha [28]. This blade pro- view of the open type test section is a square cross-section of the
file has evolved through a series of unsteady simulations on classi- size 500 mm  500 mm. A variable voltage regulator drives the
cal Bach type SSWT by varying the overlap distance from 0 to 0.5d fan and regulates the wind speed in the range of 0–10 m/s. The tur-
(d is the blade chord) and blade arc angle from 90° to 165°. The bine models are fixed at a distance of 250 mm from the exit of the
simulation studies show an improved design of the Bach type wind tunnel.
SSWT at an overlap distance of 0.4d and a blade arc angle of For measurements of mechanical torque and power, gradual
135°. This blade profile as depicted in Fig. 1(d) is named as modi- loads are applied to the turbine shaft through a pulley connected
fied Bach type blade profile. This modified Bach type blade has to the spring balance dynamometer arrangement. The dynamic
 S. Roy, U.K. Saha / Applied Energy 137 (2015) 117–125 119

1.1 D 1.1 D
D D

d d
0.2 d 0.2 d

(a) Convenonal semi-circular (b) Semi-ellipc

1.1 D 1.1 D
D D

0.172 D
d
0.075 D
0.1 d
0.32 D 0.3 D 0.4 d
135° 0.42 d

(c) Benesh type (d) Modified Bach type

Fig. 1. Dimensions of various blade profiles.

1.1 D 0.563 d
D
0.281 d 0.332 d
0.616 d

0.098 d d
0.474 d
0.302 d 0.1 d 0.007 d 0.421 d
0.183 d
0.68 d
0.746 d 0.098 d
0.969 d

27°

Fig. 2. Dimensions of the newly developed blade profile.

and static torques are measured within the uncertainty level of and static torque coefficient, respectively [33]. Whereas, uncer-
±2%. The rotational speed of the turbine is recorded by a digital tainty in the tip speed ratio calculation is found to be ±2.3%.
tachometer having an accuracy of ±1%, whereas the wind speed
is measured with the help of a hot wire anemometer having an 4. Experimental validation and blockage correction
accuracy of ±2%.
The uncertainties of the experimental data are calculated by The performance of the turbine is often measured in terms of tor-
using the sequential perturbation technique, which are found to que coefficient (CT), power coefficient (CP), and static torque coeffi-
be ±5.1%, ±4.6% and ±4.6% for power coefficient, torque coefficient cient (CTS), which are expressed by the following equations [34]:
120 S. Roy, U.K. Saha / Applied Energy 137 (2015) 117–125

Coarse screen Honeycomb structure All dimensions are in mm

Diffuser
Fine screen
secon SSWT model
Seling chamber
AC motor Bearing house
Fan Nozzle secon

375
500
Wind

Pulley

450 770 750 910

2880 700 500

(a) Side view of the wind tunnel with open type test sec on (b) Front view of the test secon

Fig. 3. Schematic diagram of the experimental setup [32].

T
CT ¼ ð1Þ 0.60
0:5qAV 2 R Convenonal SSWT
0.50
V = 6 m/s
Pturbine Tx T Rx 0.40
CP ¼ ¼ ¼ ¼ C T  TSR ð2Þ
Pavailable 0:5qAV 3 0:5qAV 2 R V
CT

0.30

TS 0.20 Present Data

C TS ¼ ð3Þ
0:5qAV 2 R 0.10
Blockage corrected data
Irabu and Roy [35]
where q is the density of air [kg/m3], V is the free stream wind 0.00
speed [m/s], A is the swept area of the turbine [m2], TSR is the tip 0 0.2 0.4 0.6 0.8 1 1.2
speed ratio, R is the radius of rotation of the turbine [m], x is the TSR
rotational speed of the turbine [rad/s], Pturbine is the power pro-
duced by the turbine [W], Pavailable is the power available in the Fig. 4. Validation and correction for CT calculation.

wind [W], T and TS are the dynamic and static torque on the turbine
[Nm], respectively.
Blockage in the wind tunnel experiments is one of the unavoid-
0.30
able phenomena and thereby this must be taken into account Convenonal SSWT
while presenting the experimental results [20,31,32,35,36]. The 0.25
V = 6 m/s
blockage correction factor (f) is largely dependent on the type of
0.20
wind tunnel test section used (open or closed type), and the block-
CP

age ratio (BR = area of turbine model to the area of test section). As 0.15
found in open literature, the investigations related to blockage cor- 0.10 Present data
rection are mostly carried out in closed type test sections [36–39]. Blockage corrected data
However, the correlations used for a closed type test section may 0.05
Irabu and Roy [35]
not be appropriate for an open type test section under dynamic 0.00
loading conditions. In this regard, Roy and Saha [32] have devel- 0 0.2 0.4 0.6 0.8 1 1.2
oped new correlations for blockage corrections in the wind tunnel TSR
experiments of SSWTs under an open test section facility, where Fig. 5. Validation and correction for CP calculation.
the blockage correction factor (f) is used to correct the measured
parameters such as wind speed (V), mechanical load applied to
the turbine (F), and rotational speed of the turbine (N). The
obtained correlations are as follows [32]: It was reported that the blockage correction factor is a function of
  the tip speed ratio (TSR) and blockage ratio (BR). Based on the above
1 mentioned correlations, suitable blockage correction factors of
V ¼ V 1  f ð4Þ
5 f = 4–9% are incorporated corresponding to various TSRs and BR of
21.16%. As shown in Figs. 4 and 5, the experimental data are com-
F  ¼ Fð1  f Þ ð5Þ pared and validated with the results of Irabu and Roy [38]. The pres-
ent data are obtained at a wind speed of 6 m/s (Re = 9.5  104).
N ¼ Nð1  f Þ ð6Þ Having included the blockage correction, the maximum power coef-
 S. Roy, U.K. Saha / Applied Energy 137 (2015) 117–125 121

ficient (CPmax) of a conventional SSWT (with overlap dis- 0.35

tance = 0.2d) is found to be 0.23 at a TSR = 0.7. Whereas, without Blockage corrected data
0.30
any blockage correction, the CPmax is obtained as 0.25 at a Re = 9.8x104
0.25
TSR = 0.74.
0.20

CP
0.15
5. Results and discussion
0.10
0.05 Classical Bach type
Experimental models are tested for Reynolds numbers of
Modified Bach type
Re = 6.0  104–1.5  105 to determine the torque coefficient (CT), 0.00
power coefficient (CP), and static torque coefficient (CTS). With 0 0.2 0.4 0.6 0.8 1 1.2
the change of load at rotating conditions, the CT and CP are calcu- TSR
lated with respect to TSR. Whereas, the CTS are obtained at various
Fig. 7. Variation of CP for classical and modified Bach type SSWTs.
turbine angular positions (h = 0–360° in steps of 10°) with respect
to wind direction. The performance of the newly developed two-
bladed turbine is compared with the conventional semi-circular, 0.70
semi-elliptic, Benesh, and modified Bach type turbines. All the Blockage corrected data
0.60
experimental data reported in this study are inclusive of blockage Re = 6.0x104
corrections. However, the optimum performance is discussed tak- 0.50
ing into account the corrected and uncorrected data. 0.40

CT
0.30 Convenonal
Semi-ellipc
5.1. Dynamic performance study 0.20 Benesh type
0.10 Modified Bach type
Initially, a comparative study was carried out for modified and Newly developed
0.00
classical Bach type SSWTs at Re = 9.8  104 (Figs. 6 and 7). The 0 0.2 0.4 0.6 0.8 1 1.2
blade arc angle of the classical Bach type was reported to be TSR
135° without any overlap between the blades [40], which had
shown a CPmax of 0.26 at TSR = 0.73. The modified Bach type, on Fig. 8. Variations of CT at Re = 6.0  104.
the other hand, had shown a CPmax of 0.30 at TSR = 0.81. Thus, a
performance gain of 15.4% was obtained with the modified Bach
type over the former one. This was basically due to the flow accel-
0.35
eration through the passage of the overlap, which enabled an addi-
Blockage corrected data
tional force to the concave side of the returning blade. 0.30
Re = 6.0x104
Figs. 8 and 9 show the torque and power coefficients of newly 0.25
developed two-bladed SSWT at Re = 6.0  104. These performance 0.20
CP

indices are also presented for other tested turbines. It may be

0.15 Convenonal
observed that the torque coefficient values decrease with the Semi-ellipc
increase of TSR. This is mainly caused by the gradual loads applied 0.10 Benesh type
to the turbine shaft, which in turn, reduces the turbine rotational 0.05 Modified Bach type
speed. On the other hand, the power coefficient increases with Newly developed
0.00
an increase of TSR up to a certain maximum value, beyond which 0 0.2 0.4 0.6 0.8 1 1.2
it decreases with further increase in TSR. The newly developed TSR
blade gives a CPmax of 0.27 at TSR = 0.77. At this low Reynolds num-
Fig. 9. Variations of CP at Re = 6.0  104.
ber (Re = 6.0  104), the modified Bach type has shown a nearly
similar performance to the newly developed SSWT; whereas, the
conventional semicircular blade has displayed the lowest power
0.70
and torque coefficients. The performance gains of the newly devel- Blockage corrected data
oped SSWT over conventional, semi-elliptic and Benesh type tur- 0.60
Re = 8.3x104
bines are found to be 28.6%, 17.4% and 3.8%, respectively. 0.50
Similarly, Figs. 10–17 demonstrates the variations of torque and 0.40
power coefficients at Re = 8.3  104, 9.8  104, 1.2  105 and
CT

Convenonal
0.30
1.5  105, respectively. In all the cases, the newly developed two- Semi-ellipc
0.20 Benesh type
0.10 Modified Bach type
0.70
Blockage corrected data Newly developed
0.60 0.00
Re = 9.8x104 0 0.2 0.4 0.6 0.8 1 1.2
0.50 TSR
0.40
Fig. 10. Variations of CT at Re = 8.3  104.
CT

0.30
0.20
Classical Bach type bladed turbine shows an improvement in the power and torque
0.10
Modified Bach type coefficients as compared to other tested turbines. With blockage
0.00 correction, a CPmax of 0.31 is obtained for the newly developed tur-
0 0.2 0.4 0.6 0.8 1 1.2
bine at Re = 1.2  105 and TSR = 0.82. In contrary, the CPmax for
TSR
modified Bach type, Benesh type, semi-elliptic and conventional
Fig. 6. Variation of CT for classical and modified Bach type SSWTs. blade profiles are found to be 0.30, 0.29, 0.26 and 0.23, respectively
122 S. Roy, U.K. Saha / Applied Energy 137 (2015) 117–125

0.35 0.35
Blockage corrected data Blockage corrected data
0.30 0.30
Re = 8.3x104 Re = 1.2x105
0.25 0.25
0.20 0.20
CP

CP
Convenonal Convenonal
0.15 0.15
Semi-ellipc Semi-ellipc
0.10 Benesh type 0.10 Benesh type
0.05 Modified Bach type Modified Bach type
0.05
Newly developed Newly developed
0.00 0.00
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2
TSR TSR
Fig. 11. Variations of CP at Re = 8.3  104. Fig. 15. Variations of CP at Re = 1.2  105.

0.70
Blockage corrected data
0.60
Re = 9.8x104 0.70
0.50 Blockage corrected data
0.60
Re = 1.5x105
0.40 0.50
CT

0.30 Convenonal
0.40
Semi-ellipc

CT
0.20 0.30 Convenonal
Benesh type
Modified Bach type Semi-ellipc
0.10 0.20
Newly developed Benesh type
0.00 0.10 Modified Bach type
0 0.2 0.4 0.6 0.8 1 1.2 Newly developed
TSR 0.00
0 0.2 0.4 0.6 0.8 1 1.2
Fig. 12. Variations of CT at Re = 9.8  104. TSR

Fig. 16. Variations of CT at Re = 1.5  105.

0.35
Blockage corrected data
0.30
Re = 9.8x104
0.25
0.35
0.20 Blockage corrected data
0.30
CP

Convenonal Re = 1.5x105
0.15
Semi-ellipc 0.25
0.10 Benesh type 0.20
CP

0.05 Modified Bach type Convenonal

Newly developed 0.15
0.00 Semi-ellipc
0 0.2 0.4 0.6 0.8 1 1.2 0.10 Benesh type
Modified Bach type
TSR 0.05
Newly developed
0.00
Fig. 13. Variations of CP at Re = 9.8  104. 0 0.2 0.4 0.6 0.8 1 1.2
TSR

Fig. 17. Variations of CP at Re = 1.5  105.

0.70
Blockage corrected data
0.60
Re = 1.2x105
0.50 Fig. 18 indicates the effect of Reynolds number on dynamic per-
formances of the newly developed blade profile. It is seen that with
0.40
the increase of Reynolds number up to Re = 1.2  105, the CPmax
CT

0.30 Convenonal
Semi-ellipc
increases, beyond which it again decreases.
0.20 Benesh type
0.10 Modified Bach type
Newly developed 5.2. Static performance study
0.00
0 0.2 0.4 0.6 0.8 1 1.2
The starting characteristic of SSWTs is often described through
TSR
the analysis of static torque coefficients. Fig. 19 demonstrates the
Fig. 14. Variations of CT at Re = 1.2  105. static torque coefficients for various tested blade profiles at
Re = 6.0  104–1.5  105. The rotational angle (h) is the angle
at Re = 1.2  105. Table 1 shows the values of maximum power between the chord of the advancing blade to the direction of wind
coefficients corresponding to various TSRs at Re = 6.0  104– flow [28]. At Re = 6.0  104, the conventional semi-circular blades
1.5  105. It is observed that for the new developed blade profile, show a very low static torque coefficients in the range of h = 110–
the maximum power coefficients are obtained in the range of 170° and 290–350°. During the experimental process, in this range,
TSR = 0.74–0.83. Whereas, for modified Bach, Benesh, semi-elliptic the conventional SSWT does not start by itself. More precisely, in
and conventional type turbines, the CPmax values fall in the range of the range of h = 140–170° and 320–350°, the negative pressure on
0.66–0.73, 0.69–0.75, 0.71–0.81 and 0.72–0.81, respectively. Thus, the returning blade caused the negative static torque coefficients.
in each case, there is a shift in the range of optimum TSR. With the further increase of Reynolds number, similar starting
 S. Roy, U.K. Saha / Applied Energy 137 (2015) 117–125 123

Table 1
Maximum power coefficients corresponding to tip speed ratios (blockage corrected data).

Blade profile Re = 6.0  104 Re = 8.3  104 Re = 9.8  104 Re = 1.2  105 Re = 1.5  105
CPmax TSR CPmax TSR CPmax TSR CPmax TSR CPmax TSR
New SSWT 0.27 0.77 0.28 0.75 0.30 0.81 0.31 0.82 0.29 0.79
Modified bach 0.27 0.75 0.27 0.72 0.29 0.78 0.30 0.80 0.28 0.77
Benesh 0.26 0.73 0.26 0.72 0.28 0.77 0.29 0.80 0.27 0.77
Semi-elliptic 0.23 0.73 0.23 0.70 0.25 0.73 0.26 0.75 0.24 0.72
Conventional 0.21 0.66 0.22 0.67 0.23 0.71 0.23 0.73 0.22 0.69

0.34
Blockage corrected data
0.30 CSSWT profile MSSWT profile
Semi-ellipc profile Benesh profile
0.26 Convenonal profile
CPmax

0.22 0.60
Convenonal
0.18 Semi-ellipc 0.40
Benesh type
0.14 Modified Bach type

CTs
0.20
Newly developed
0.10 Re = 6.0x104
0.00
50000 70000 90000 110000 130000 150000
Blockage corrected data
Re - 0.20
0 60 120 180 240 300 360
Fig. 18. Effect of Re on the variations of CPmax. Rotaonal angle (θ)

0.60
problems are observed for conventional SSWT; however, the values
of static torque coefficients are slightly increased. The peak starting 0.40
performance is seen at angular positions of h = 30–60° and 210–
CTs

0.20
240° and at Re = 1.5  105, the CTSmax is obtained as 0.39. At
Re = 6.0  104, the semi-elliptic type shows poor starting perfor- 0.00 Re = 8.3x104

mances in the range of h = 120–170° and 300–350°, while the neg- Blockage corrected data
- 0.20
ative static torque coefficients are seen at h = 150–170° and 330– 0 60 120 180 240 300 360
350°. However, with the increase of Reynolds number, the starting Rotaonal angle (θ)
performance got improved and the negative range got reduced to
h = 150–160° and 330–340° at Re = 1.2  105. This has been further 0.60
amended at Re = 1.5  105, where the negative static torque coeffi-
cients were found only at h = 160° and 360°. With the semi-ellipti- 0.40
cal blade, the highest static torque coefficient of CTSmax = 0.41 is
CTs

0.20
obtained at Re = 1.5  105.
The Benesh type turbine has shown an improvement in terms of 0.00 Re = 9.8x104
starting characteristics over semi-circular and semi-elliptical type Blockage corrected data
turbines. The Benesh type, unlike the above two turbines, has com- -0.20
0 60 120 180 240 300 360
pletely overcome the negative values of CTS. However, lower positive
Rotaonal angle (θ)
values are observed in the range of h = 140–170° and 320–350°. At
these angular positions, the positive values of CTS obtained are insuf- 0.60
ficient for self-starting at lower Reynolds numbers.
The modified Bach type has displayed improved CTS over the 0.40
Benesh type. At low Reynolds numbers, the modified Bach type
CTs

0.20
shows its self-starting incapability in the range of h = 150–170°
and 330–350°. Like previous models, this turbine also indicates 0.00 Re = 1.2x105

high CTS in the range of h = 30–60° and 210–240°. For the Benesh Blockage corrected data
-0.20
and the modified Bach types, the CTSmax values are found to be 0 60 120 180 240 300 360
0.46 and 0.49 at Re = 1.5  105, respectively.
Rotaonal angle (θ)
Among all the test cases, the newly developed SSWT is found to
be most effective considering the self-starting criterion. With this 0.60
turbine, the problems of self-starting can be substantially reduced,
however, at low Reynolds number, this problem predominates in 0.40
the range of h = 160–170° and 340–350°. With the increase of Rey-
CTs

0.20
nolds number, the starting problem at these angular positions gets
significantly reduced. At Re = 1.5  105, the maximum and mini- 0.00 Re = 1.5x105

mum static torque coefficients are found to be 0.51 and 0.08, Blockage corrected data
- 0.20
respectively. A clear observation of maximum and minimum static 0 60 120 180 240 300 360
torque coefficients for various turbine models is depicted in Table 2. Rotaonal angle (θ)
It is noted that unlike the dynamic characteristics (Fig. 18), the sta-
tic characteristics increases with an increase of Reynolds number. Fig. 19. Starting characteristics of various blade profiles.
124 S. Roy, U.K. Saha / Applied Energy 137 (2015) 117–125

Table 2
Maximum and minimum coefficients of static torque at various Reynolds numbers (blockage corrected data).

Blade profile Re = 6.0  104 Re = 8.3  104 Re = 9.8  104 Re = 1.2  105 Re = 1.5  105
CTSmax CTSmin CTSmax CTSmin CTSmax CTSmin CTSmax CTSmin CTSmax CTSmin
New SSWT 0.48 0.07 0.49 0.07 0.50 0.08 0.50 0.08 0.51 0.08
Modified bach 0.46 0.03 0.47 0.03 0.48 0.04 0.48 0.04 0.49 0.05
Benesh 0.44 0.01 0.44 0.01 0.45 0.02 0.45 0.02 0.46 0.02
Semi-elliptic 0.37 0.06 0.39 0.06 0.39 0.03 0.41 0.01 0.41 0.01
Conventional 0.35 0.13 0.37 0.12 0.37 0.10 0.38 0.05 0.39 0.02

Table 3
Overall efficiency and payback period for all the tested SSWTs (H = DO = 1 m, V = 7.8 m/s).

Parameters Newly developed Modified Bach Benesh Semi-elliptic Semi-circular

Turbine efficiency, gT 31% 30% 29% 26% 23%
Generator efficiency, gG 80% 80% 80% 80% 80%
Transmission-Storage efficiency, gTS 70% 70% 70% 70% 70%
Overall efficiency of WECS, gO 17.4% 16.8% 16.2% 14.6% 12.9%
Available power in the wind 264.2 W 264.2 W 264.2 W 264.2 W 264.2 W
Power output 46.0 W 44.4 W 42.8 W 38.6 W 34.1 W
Power output in kW h per day 1.10 kW h 1.07 kW h 1.03 kW h 0.93 kW h 0.82 kW h
Power cost per kW h 0.20 USD 0.20 USD 0.20 USD 0.20 USD 0.20 USD
Turbine capital cost 400 USD 400 USD 400 USD 400 USD 400 USD
Maintenance cost 100 USD 100 USD 100 USD 100 USD 100 USD
Payback period 6.2 years 6.4 years 6.6 years 7.4 years 8.4 years

5.3. Economic aspects Table 3 shows the details of the assessment on the payback periods
for all the tested SSWTs using Eqs. (7)–(9) [41]. It has been observed
A complete wind energy conversion system (WECS) includes that the payback period for a conventional SSWT is 8.4 years, which
the wind turbine, a support structure, an electrical generator, an is much higher than the newly developed blade design. However,
energy storage device and the transmission auxiliaries. To esti- the payback periods for modified Bach, Benesh, and semi-elliptic
mate the power output of a WECS, the losses experienced in type are calculated as 6.4, 6.6 and 7.4 years, respectively.
the turbine section, electricity generation, transmission and stor-
age have to be taken into account. Thus, the overall efficiency (gO) 6. Conclusions
and the power output of a WECS are calculated by the following
equations [41]: The performance of the newly developed two bladed SSWT is
gO ¼ gT  gG  gTS ð7Þ reported and the results are compared and discussed with modi-
fied Bach, Benesh, semi-elliptical and conventional semi-circular
Power output of WECS ¼ Power available in the wind bladed SSWTs. The main observations are summarized as follows:

 overall efficiency (a) With the newly developed SSWT, a noticeable improvement
1 in the maximum power coefficient is observed over other
¼ qAV 3  gO ð8Þ
2 tested models. Performance gains of 3.3%, 6.9%, 19.2% and
34.8% are achieved over modified Bach, Benesh, semi-ellipti-
where, gT, gG, and gTS are the efficiencies corresponding to wind
cal, and conventional SSWTs, respectively.
turbine, electrical generator, and transmission and storage systems,
(b) The newly developed SSWT not only can increase the magni-
respectively.
tude of the static torque coefficient, but also overcome the
Now, considering CPmax of SSWTs to be the turbine efficiency
effects of negative torque due to change in the blade arc
(gT), gG = 80% (a more commonly found efficiency), and gTS = 70%
geometry. In terms of self-starting prospect, the results indi-
(considering a maximum loss), the overall efficiency of the WECS
cate that the CTSmax of the newly developed blade profile is
with the newly developed blade is estimated to be 17.4%. Whereas,
higher by 31.6%, 22.0% and 11.1% than the conventional,
the overall efficiencies of the WECS with modified Bach, Benesh,
semi-elliptical and Benesh blade profiles, respectively. How-
semi-elliptic and semi-circular turbines are estimated to be
ever, a marginal improvement of 4.2% is recorded over mod-
16.8%, 16.2%, 14.6%, and 12.9%, respectively. Using these values
ified Bach type blade profile.
of gO in Eq. (8), the power outputs of the WECS with the newly
(c) With the increase of Reynolds number, the dynamic perfor-
developed, modified Bach, Benesh, semi-elliptic and semi-circular
mance index i.e. the power coefficient increases up to a cer-
turbines (H = DO = 1 m) are estimated to be 46.0 W, 44.4 W,
tain limit of Re = 1.2  105, beyond which it again decreases.
42.8 W, 38.6 W and 34.1 W, respectively.
In contrary, with the increase of Reynolds number, the start-
Assuming the turbine capital cost and maintenance cost to be
ing performance index i.e. the static torque coefficient
400 USD and 100 USD, the payback period for WECS involving the
increases irrespective of the magnitude of Reynolds number.
newly developed SSWT is calculated by the following equation [41]:
(d) For each of the tested blade profiles, the CPmax values are
Total cost experienced at an optimum range of TSR such as 0.75–
Pay back period ¼
Power cost per kWh  Power output in kWh=day 0.82, 0.72–0.80, 0.72–0.80, 0.70–0.75 and 0.66–0.73 for the
ð400 þ 100Þ newly developed, modified Bach, Benesh, semi-elliptical
¼ ¼ 2273days ¼ 6:2 years ð9Þ
ð1:10  0:20Þ and conventional blades, respectively.
 S. Roy, U.K. Saha / Applied Energy 137 (2015) 117–125 125

(e) For the newly developed SSWT, the maximum power coeffi- [15] Benesh AH. Wind turbine with Savonius-type rotor. US Patent; 1996, No.
5494407.
cient recorded is 0.31 with wind tunnel blockage correction.
[16] Golecha K, Eldho TI, Prabhu SV. Influence of the deflector plate on the
(f) The overall efficiency and the payback periods for the wind performance of modified Savonius water turbine. Appl Energy
energy conversion systems are estimated for all the test 2011;88(9):3207–17.
cases. The payback periods for the newly developed, modi- [17] Abraham JP, Plourde BD, Mowry GS, Minkowycz WJ, Sparrow EM. Summary of
Savonius wind turbine development and future applications for small-scale
fied Bach, Benesh, semi-elliptical, and conventional SSWTs power generation. J Renew Sustain Energy 2012;4:1–21.
are found to be 6.2, 6.4, 6.6, 7.4 and 8.4 years, respectively. [18] Plourde BD, Abraham JP, Mowry GS, Minkowycz WJ. Use of small-scale wind
energy to power cellular communication equipment. Sens Transducers
2011;13:53–61.
7. Future work [19] Plourde BD, Abraham JP, Mowry GS, Minkowycz WJ. Simulations of three-
dimensional vertical-axis turbines for communications applications. Wind Eng
In this study, the improvement of newly developed SSWT blade 2012;36:443–54.
[20] Abraham JP, Plourde BD, Mowry GS, Sparrow EM, Minkowycz WJ. Numerical
profile is explored through experimental analysis of power coeffi- simulation of fluid Flow around a vertical-axis turbine. J Renew Sustain Energy
cient, torque coefficient and static torque coefficient. A better per- 2011;3(033109):1–13.
formance is observed with the newly developed blade profile. [21] Akwa JV, Vielmo HA, Petry AP. A review on the performance of Savonius wind
turbines. Renew Sustain Energy Rev 2012;16(5):3054–64.
However, the flow physics and drag-lift characteristics are yet a [22] Kamoji MA, Kedare SB, Prabhu SV. Performance tests on helical Savonius
matter of research. The performance improvement may be due to rotors. Renew Energy 2009;34(3):521–9.
more contribution of lift in the new design as can be seen in the [23] Damak A, Driss Z, Abid MS. Experimental investigation of helical Savonius
rotor with a twist of 180°. Renew Energy 2013;52:136–42.
case of modified Bach type SSWT [28]. Thus, as a future work,
[24] Saha UK, Rajkumar MJ. On the performance analysis of Savonius rotor with
three-dimensional unsteady simulation study has to be pursued twisted blades. Renew Energy 2006;31(11):1776–88.
to obtain an insight on the aerodynamic drag-lift relationships [25] Grinspan AS, Saha UK, Mahanta P. Experimental investigation of twisted
and flow properties around this newly developed SSWT. bladed Savonius wind turbine rotor. Int Energy J 2004;5(1):1–9.
[26] Modi VJ, Roth NJ, Fernando MSUK. Optimum-configuration studies and
prototype design of a wind-energy-operated irrigation system. J Wind Eng
References Ind Aerodynamics 1984;16:85–96.
[27] Modi VJ, Fernando MSUK. On the performance of the Savonius wind turbine.
[1] Rolland S, Newton W, Williams AJ, Croft TN, Gethin DT, Cross M. Simulations ASME J Solar Energy Eng 1989;111:71–81.
technique for the design of a vertical axis wind turbine device with [28] Roy S, Saha UK. Numerical investigation to assess an optimal blade profile for
experimental validation. Appl Energy 2013;111:1195–203. the drag based vertical axis wind turbine. In: Proc. ASME 2013 International
[2] Chong WT, Poh SC, Fazlizan A, Yip SY, Chang CK, Hew WP. Early development Mechanical Engineering Congress & Exposition, San Diego, USA; 15–21
of an energy recovery wind turbine generator for exhaust air system. Appl November 2013, Paper no:IMECE2013-64001.
Energy 2013;112:568–75. [29] Tabassum SA, Probert SD. Vertical-axis wind turbine: a modified design. Appl
[3] Obara S, Morizane Y, Morel J. Study on method of electricity and heat storage Energy 1987;28(1):59–67.
planning based on energy demand and tidal flow velocity forecasts for a tidal [30] Roy S. Aerodynamic performance evaluation of a novel Savonius-style wind
microgrid. Appl Energy 2013;111:358–73. turbine through unsteady simulations and wind tunnel experiments. Ph.D.
[4] Saeidi D, Sedaghat A, Alamdari P, Alemrajabi AA. Aerodynamic design and Thesis 2014, Department of Mechanical Engineering, IIT Guwahati, India.
economical evaluation of site specific small vertical axis wind turbines. Appl [31] Plourde BD, Abraham JP, Mowry GS, Minkowycz WJ. An experimental
Energy 2013;101:765–75. investigation of a large, vertical-axis wind turbine: effects of venting and
[5] Danao LA, Eboibi O, Howell R. An experimental investigation into the influence capping. Wind Eng 2011;35:213–20.
of unsteady wind on the performance of a vertical axis wind turbine. Appl [32] Roy S, Saha UK. An adapted blockage factor correlation approach in wind
Energy 2013;107:403–11. tunnel experiments of a Savonius-style wind turbine. Energy Convers Manage
[6] Morbiato T, Borri C, Vitaliani R. Wind energy harvesting from transport 2014;86:418–27.
systems: a resource estimation assessment. Appl Energy 2014;133:152–68. [33] Moffat RJ. Contributions to the theory of single-sample uncertainty analysis.
[7] Danao LA, Edwards J, Eboibi O, Howell R. A numerical investigation into the ASME J Fluids Eng 1982;104:250–64.
influence of unsteady wind on the performance and aerodynamics of a vertical [34] Shaughnessy BM, Probert SD. Partially-blocked Savonius rotor. Appl Energy
axis wind turbine. Appl Energy 2014;116(1):111–24. 1992;43(4):239–49.
[8] Rolland SA, Thatcher M, Newton W, Williams AJ, Croft TN, Gethin DT, et al. [35] Barlow JB, Rae WH, Pope A. Low speed wind tunnel testing. 3rd ed. New York,
Benchmark experiments for simulations of a vertical axis wind turbine. Appl USA: John Wiley & Sons; 1999.
Energy 2013;111:1183–94. [36] Chen TY, Liou LR. Blockage corrections in wind tunnel tests of small
[9] Chong WT, Fazlizan A, Poh SC, Pan KC, Hew WP, Hsiao FB. The design, horizontal-axis wind turbines. Exp Therm Fluid Sci 2011;35(3):565–9.
simulation and testing of an urban vertical axis wind turbine with the omni- [37] Biswas A, Gupta R, Sharma KK. Experimental investigation of overlap and
direction-guide-vane. Appl Energy 2013;112:601–9. blockage effects on three-bucket Savonius rotors. Wind Eng
[10] Balduzzi F, Bianchini A, Carnevale EA, Ferrari L, Magnani S. Feasibility analysis 2007;31(5):363–8.
of a Darrieus vertical-axis wind turbine installation in the rooftop of a [38] Irabu K, Roy JN. Characteristics of wind power on Savonius rotor using a guide-
building. Appl Energy 2012;97:921–9. box tunnel. Exp Therm Fluid Sci 2007;32(2):580–6.
[11] Fleming PD, Probert SD, Tanton D. Designs and performances of flexible and [39] Ross I, Altman A. Wind tunnel blockage corrections: review and application to
taut sail Savonius-type wind-turbines. Appl Energy 1985;19(2):97–110. Savonius vertical-axis wind turbines. J Wind Eng Ind Aerodynamics
[12] Kamoji MA, Kedare SB, Prabhu SV. Experimental investigations on single stage 2011;99(5):523–38.
modified Savonius rotor. Appl Energy 2009;86(7–8):1061–73. [40] Kacprzak K, Liskiewicz G, Sobczak K. Numerical investigation of conventional
[13] Roy S, Saha UK. Review on the numerical investigations into the design and and modified Savonius wind turbines. Renew Energy 2013;60:578–85.
development of Savonius wind rotors. Renew Sustain Energy Rev [41] Al-Bahadly I. Building a wind turbine for rural home. Energy Sustain Develop
2013;24:73–83. 2009;13(3):159–65.
[14] Roy S, Saha UK. Review of experimental investigations into the design,
performance and optimization of the Savonius rotor. Proc IMechE Part A: J
Power Energy 2013;227(4):528–42.