Available online at www.sciencedirect.

com
ScienceDirect

Energy Reports 6 (2020) 184–189

www.elsevier.com/locate/egyr

6th International Conference on Energy and Environment Research, ICEER 2019, 22–25 July,
University of Aveiro, Portugal

Casing optimization of a Savonius wind turbine

E. Antara , M. Elkhouryb ,∗
a Industrial and Mechanical Engineering department, Lebanese American University, Byblos, Lebanon
b School of Engineering, Lebanese American University, Byblos, Lebanon

Received 5 August 2019; accepted 22 August 2019

Abstract
In an attempt to enhance the performance of a Savonius Vertical Axis Wind Turbine (VAWT), case sizing optimization was
performed utilizing Computational Fluid Dynamics (CFD) technique. A 2D parametric optimization process was employed to
optimize a number of assigned design parameters of the casing’s geometry. The configuration of El-Askary et al. (2015) was
adopted as a starting baseline for the optimization. The turbine with the optimized casing provided a superior performance
compared to a caseless one, specifically at low Tip Speed Ratios (TSRs). A maximum increase of 42.5% in the power coefficient
(Cp) was attained at a tip speed ratio of 0.59. The obtained results demonstrate the nonexistence of a universal optimal casing
shape that maintains peak turbine performance at all tip speed ratios. However, optimized casing dimensions follow a clear
trend that is closely related to TSR values.
⃝c 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 6th International Conference on Energy and Environment Research, ICEER 2019.

Keywords: Savonius wind turbine; Optimization

1. Introduction

Drag-type Savonius Vertical Axis Wind Turbines (VAWTs) rely on the pressure difference on both sides of the
blades for torque generation. These turbines are renowned for their excellent startup characteristics, ability to operate
in omnidirectional wind flows, and ease of blade fabrication [1,2]. These unique characteristics are the motive
behind the renewed interest in optimizing the performance of these devices. Researches have considered various
design layouts to understand and improve the performance of a Savonius turbine. Kianifar, A. and Anbarsooz, M. [3]
studied the effect of the Savonius blade curvature on the turbine’s efficiency. It was concluded that having a rotor
overlap resulted in an increase in the Power Coefficient (Cp) value up to a certain extent. Frikha et al. [4] conducted
numerical investigations with experimental validations to characterize the performance of a multi-stage Savonius
VAWT. Two to five stages were analyzed where it was shown that the Torque Coefficient (CT ) and power coefficient

∗ Corresponding author.
E-mail addresses: elie.antar@lau.edu (E. Antar), mkhoury@lau.edu.lb (M. Elkhoury).

https://doi.org/10.1016/j.egyr.2019.08.040
2352-4847/⃝ c 2019 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 6th International Conference on Energy and Environment Research, ICEER
2019.
 E. Antar and M. Elkhoury / Energy Reports 6 (2020) 184–189 185

values increased with number of stages. Many more techniques have demonstrated their viability in improving the
performance of Savonius VAWTs, one of which is the utilization of rotor’s casing [5,6].
The goal of this study is to optimize the efficiency of a Savonius VAWT, with a radius of 0.08 m and an overlap
ratio of 0.15, at different Tip Speed Ratios (TSRs). This is accomplished by parametrically optimizing the shape of
a baseline rotor casing design, proposed by El-Askary et al. [7], which have shown superior performance compared
to other published configurations. Different casing geometrical parameters are assigned and optimized following a
proposed scheme. Computational Fluid Dynamics (CFD) technique is adopted utilizing the 2-equation SST k − ω
turbulence model. Different optimal casing shapes are anticipated to be generated at different TSRs, hence a trend
between the optimal casing shapes and TSR is thoroughly studied.

2. Methodology

The dimensions of the Savonius rotor is listed in Table 1. The geometry of the baseline casing design, suggested
by El-Askary et al. [7], was parametrized to yield five distinctive optimization parameters shown in Fig. 1 (blue).
Dimensions in red, like the distance between walls 1 and 3 (5.7R), were kept constant for a fair comparison between
the generated samples.

Fig. 1. Geometry of the Savonius turbine with casing, dimensions referred to in blue were varied during the optimization process. . (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Table 1. Dimensions of the Savonius rotor.

Blade radius 0.08 m
Rotor radius 0.149 m
Overlap ratio 0.15
Number of blades 2

At a given TSR, the optimization process started with radius R1, where different Cp values were generated for
different values of R1. After observing a deteriorating trend in performance, the optimal value of R1 was selected.
With the optimized value of R1, the same process was followed with subsequent parameters R2, D1, R3, and finally
D2, where the optimized value of a given parameter was adopted during the remaining process at a given TSR.
Fig. 2 depicts the order of the optimization process, highlighting also the effect on the casing’s geometry resulting
from selecting different values for the geometrical parameters.
186 E. Antar and M. Elkhoury / Energy Reports 6 (2020) 184–189

Fig. 2. The optimization order of the different casing geometrical parameters.

3. Numerical setup
The governing transient 2D flow equations were solved using ANSYS Fluent. The SST k − ω turbulence
model was adopted in the present simulations owing to its accuracy and acceptable computational demands. The
computational domain was divided into two flow zones, rotor and wind tunnel, with a sliding interface. The domain
dimensions exceeded the minimum sizing guidelines published by Rezaeiha et al. [8], which assured the omission
of blockage effects and the full development of the wake region. The generated mesh was composed of unstructured
triangular elements with inflation layers near solid surfaces in order to ensure that y + ⩽1.0. Appropriate mesh
controls were deployed in order to refine the mesh in certain areas where information was expected to vary rapidly.
Initial studies were conducted in order to generate a mesh-independent grid, and select the right time step size.
Consequently, these calculations showed that a grid with 332,768 elements and a time step size equivalent to 1◦
rotor rotation had to be utilized. The utilized boundary conditions are delineated in Table 2.

Table 2. Boundary conditions used in the numerical analyses.

Free stream velocity (U) 6 m/s
Turbulence intensity (I) 5%
Outlet gauge pressure 0 Pa
Blade and casing walls Non-slip
Upper and lower domain bounds Symmetry

4. Results
Different power coefficient curves for different turbine configurations are shown in Fig. 3 (left), including relevant
experimental data (blue) published by Fujisawa [9]. The optimized Cp values (green) were generated following the
optimization scheme described previously with more than 100 numerical simulations conducted involving different
casing shapes and TSRs. It is evident that the performance of the turbine with the optimized casing is substantially
improved compared to a caseless configuration, where Cp increases by 21%, 27%, and 15% at TSRs 0.59, 0.76, and
 E. Antar and M. Elkhoury / Energy Reports 6 (2020) 184–189 187

Fig. 3. Power coefficient values from 2D simulations (left), and the generated optimized casing shapes at different TSRs (right) . (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

1.19 respectively. At TSR = 0.42, the performance of the optimized casing slightly drops with Cp = 0.14. Relative
to the baseline casing design, the value of Cp is increased by 24%, 32%, 29%, and 48% at TSRs of 0.42, 0.59,
0.76, and 1.19, respectively. Fig. 3 (right), shows the resulting optimized casing shapes which are clearly different
at various TSRs as anticipated. A trend in the casing’s geometry and TSR can be readily pointed out, where with
increasing TSR: curvature of wall 1 decreases, distance between walls 2 and 3 decreases, and returning blade jet
angle decreases. The last two coherent observations signal that there is certain value of the returning blade angle of
attack that should be maintained for optimal performance. Inspecting the Cp values of the caseless configuration, it
is evident that 2D CFD overestimates the performance of the turbine, since the generated Cp curve from 2D CFD
is shifted vertically upwards relative to the experimental data. This shift effect is frequently reported in published
2D studies [10], and is a consequence of following a rather computationally feasible approach. Following a similar
cumbersome optimization process on a 3D model would, for sure, require substantial computational resources, as
shown by Antar and Elkhoury [11].
Torque coefficient curves at TSR = 0.76 is shown in Fig. 4, where it is clear that the torque generated by
the turbine is considerably amplified. The maximum value of CT increases by almost three times relative to the
caseless configuration, and the optimized CT never attains a negative value. In order to further investigate the issue,
instantaneous pressure contours for the caseless, baseline, and optimized configurations with CT equal to 0.36, 0.72,
and 1.16 respectively, are considered at TSR = 0.76 and θ = 145◦ (Fig. 5). It can be seen that the positive pressure
area on the concave side of the advancing blade is the greatest for the optimized configuration, followed by the
baseline configuration and finally the caseless one. The same description applies for the negative pressure area on
the convex side of the blade, which is the greatest for the optimized design. This increase in negative pressure is a
consequence of the nozzle-like effect resulting from having a curved wall 1, which is not presented in the baseline
design or the caseless configuration. Consequently, the pressure difference on the advancing blade is tremendously
increased, and is responsible for the elevated CT value. This outweighs the increase in pressure difference on the
returning blade, which generates counter-rotating torque (Fig. 5).

5. Conclusion
The geometry of a casing that engulfed the rotor of a Savonius turbine was successfully optimized for maximum
efficiency at four TSRs. Below are the main findings from the conducted parametric CFD investigations:
• Compared to a caseless Savonius turbine, Cp increased by 21%, 27%, and 15% at TSRs 0.59, 0.76, and 1.19
respectively when adopting the optimized casing designs. Cp marginally drop to 0.14 at TSR 0.42.
188 E. Antar and M. Elkhoury / Energy Reports 6 (2020) 184–189

Fig. 4. Torque coefficient curves from numerical simulations at TSR = 0.76.

Fig. 5. Instantaneous pressure (Pa) contours at TSR = 0.76 and θ = 145◦ .

• Relative to the baseline casing design, Cp increased by 24%, 32%, 29%, and 48% at TSRs of 0.42, 0.59, 0.76,
and 1.19 respectively when utilizing the optimized casing designs.
• The dimensions of the optimized casing had a clear trend with TSR, whereby with increasing TSR: The
curvature of wall 1 decreased, the distance between walls 2 and 3 decreased, and the returning blade jet angle
decreased.
• 2D Reynolds-Averaged Navier–Stokes simulations seem to be unable to replicate the complicated flow
around a Savonius VAWT. Hence, performing 3D numerical simulations relying on hybrid resolving-modeling
techniques like Detached Eddy Simulations (DES) or Partially Averaged Navier–Stokes (PANS) render the
simulation more accurate.

References
[1] Wenehenubun F, Saputra A, Sutanto H. An experimental study on the performance of Savonius wind turbines related with the number
of blades. Energy Procedia 2015;68:297–304. http://dx.doi.org/10.1016/j.egypro.2015.03.259.
[2] Elkhoury M, Kiwata T, Aoun E. Experimental and numerical investigation of a three-dimensional vertical-axis wind turbine with
variable-pitch. J Wind Eng Ind Aerodyn 2015;139:111–23. http://dx.doi.org/10.1016/j.jweia.2015.01.004.
 E. Antar and M. Elkhoury / Energy Reports 6 (2020) 184–189 189

[3] Kianifar A, Anbarsooz M. Blade curve influences on the performance of Savonius rotors : experimental and numerical. 225, 2010, p.
343–50, http://dx.doi.org/10.1177/2041296710394413.
[4] Frikha S, Driss Z, Ayadi E, Masmoudi Z, Abid MS. Numerical and experimental characterization of multi-stage Savonius rotors.
Energy 2016;114:382–404. http://dx.doi.org/10.1016/j.energy.2016.08.017.
[5] Manganhar AL, Rajpar AH, Luhur MR, Samo SR, Manganhar M. Performance analysis of a savonius vertical axis wind turbine
integrated with wind accelerating and guiding rotor house. Renew Energy 2019;136:512–20. http://dx.doi.org/10.1016/j.renene.2018.12.
124.
[6] Backman J, Hansen-haug M, Laaksonen M, Alkki M, Aura P. Experimental and numerical analysis of vaned wind turbine performance
and flow phenomena. 159, 2018, http://dx.doi.org/10.1016/j.energy.2018.06.204.
[7] El-Askary WA, Nasef MH, AbdEL-hamid AA, Gad HE. Harvesting wind energy for improving performance of savonius rotor. J Wind
Eng Ind Aerodyn 2015;139:8–15. http://dx.doi.org/10.1016/j.jweia.2015.01.003.
[8] Rezaeiha A, Kalkman I, Blocken B. CFD simulation of a vertical axis wind turbine operating at a moderate tip speed ratio: Guidelines
for minimum domain size and azimuthal increment. Renew Energy 2017;107:373–85. http://dx.doi.org/10.1016/j.renene.2017.02.006.
[9] Fujisawa N. On the torque mechanism of Savonius rotors. J Wind Eng Ind Aerodyn 1992;40:277–92. http://dx.doi.org/10.1016/0167-
6105(92)90380-S.
[10] Jin X, Zhao G, Gao K, Ju W. Darrieus vertical axis wind turbine: Basic research methods. Renew Sustain Energy Rev 2014;42:212–25.
http://dx.doi.org/10.1016/j.rser.2014.10.021.
[11] Antar E, Elkhoury M. Parametric sizing optimization process of a casing for a Savonius Vertical Axis Wind Turbine. Renew Energy
2019;136:127–38.