Ocean Engineering 260 (2022) 112054

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Performance investigation of a Savonius rotor by varying the blade

arc angles
Khaled R. Abdelaziz a, Mohamed A.A. Nawar a, Ahamed Ramadan b, Youssef A. Attai a, Mohamed
H. Mohamed a, c, *
a
Mechanical Power Engineering Department, Faculty of Engineering EL-Mattaria, Helwan University, P.O.Box 11718, Cairo, Egypt
b
College of Engineering (Cairo Branch), Arab Academy for Science Technology and Maritime Transport (AASTMT), P.O.Box 2033, Elhorria, Cairo, Egypt
c
Mechanical Engineering Dept, College of Engineering and Islamic Architecture, Umm Al-Qura University, Makkah, Saudi Arabia

A R T I C L E I N F O A B S T R A C T

Keywords: Wind energy is a promising source of renewable energy due to its availability. Savonius turbine is a vertical axis
Wind energy wind turbine that converts kinetic energy from the wind to mechanical energy. The current research investigates
Savonius turbine the performance of the conventional Savonius turbine by varying the position of the endpoints based on varying
Conventional rotor
the blade arc angle. In addition, the outer blade arc angle of a modified configuration, which generated by
Modified Savonius rotor
varying the gap ratio, is also investigated. The ANSYS-Fluent package is used to perform the numerical simu­
Gap ratio
Blade arc angle lation of the turbine to solve the equations of the unsteady Reynolds Averaged Navier–Stokes (RANS). The flow
characteristics through the turbine are resolved by using the SST k-ω turbulence model. The numerical result of
the conventional configuration is validated by using previous numerical and experimental research works. The
current investigation proved that varying the outer and inner blade arc angles to 160◦ and 20◦ enhanced the
maximum power coefficient by 4.5% and 12.9%, respectively. Furthermore, the maximum power coefficient of
the modified Savonius rotor at a gap ratio of 0.1333 and an outer blade angle of 180◦ increased by 13.6%
compared to the conventional profile.

1. Introduction consequences of structural failure (Johari et al., 2018). Vertical axis

wind turbines are optimum for low wind speeds in turbulent environ­
Since the industrial development in Europe in the 19th century, fossil ments, such as urban sites (Sharma and Sharma, 2017).
fuels have played a significant role in energy production. Conventional Vertical axis wind turbines, such as Savonius and Darrieus turbines,
energy sources are not the optimum choice as they have several de­ have acquired an increasing interest in energy production in recent
merits, such as global climate change, carbon emissions, and a high cost decades (Bianchini et al., 2015). The merits of VAWTs include simple
of production (Bonalumi and Giuffrida, 2016). In recent centuries, the design, inexpensive maintenance and manufacturing costs, and unidi­
population growth rate and economic development have caused a rectional rotation (Rezaeiha et al., 2017). The low output power, oper­
massive demand for energy (Edelenbosch et al., 2017). Renewable en­ ating range, and efficiency are the significant drawbacks of VAWTs
ergy sources are considered the best alternative for energy generation. (Minetto and Paraschivoiu, 2020). Savonius turbine is a VAWT that
Wind energy has demonstrated its importance among other renewable produces a large starting torque compared to the Darrieus turbine at low
energy sources (Yannopoulos et al., 2015). Wind turbines are classified wind velocities (Goodarzi and Keimanesh, 2015). The conventional
into two categories based on their rotating axis: horizontal axis wind design of a Savonius rotor involves two opposite semi-circular blades,
turbines (HAWTs) and vertical axis wind turbines (VAWTs) (Wang such as S-shape, with an overlap distance between the blades (Toja-Silva
et al., 2016). For large-scale energy production, HAWTs are used et al., 2013). Two circular endplates govern the flow stream on the
because they generate a huge amount of output power. Rural areas are blades. The design preserves unidirectional rotation depending on the
the preferred locations for HAWTs due to the significant noise level initial position of the rotor (Chen et al., 2017), as shown in Fig. 1. The
generated from the operation process and the potentially fatal initial position of the Savonius rotor defines the direction of rotation.

* Corresponding author. Mechanical Power Engineering Department, Faculty of Engineering EL-Mattaria, Helwan University, P.O.Box 11718, Cairo, Egypt.
E-mail addresses: moh75202@yahoo.de, mhmohamed@uqu.edu.sa (M.H. Mohamed).

https://doi.org/10.1016/j.oceaneng.2022.112054
Received 19 March 2022; Received in revised form 28 May 2022; Accepted 16 July 2022
Available online 31 July 2022
0029-8018/© 2022 Elsevier Ltd. All rights reserved.
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

Savonius rotors produced a higher power coefficient compared to

three-bladed rotors. Table 1 shows the design parameters for the con­
ventional Savonius profile, which are recommended by previous studies.
The second research area was interested in enhancing the perfor­
mance of the Savonius turbine using various designs of stator and rotor.
Stators are used to guide the incoming flow to the advancing bucket to
increase the advancing drag and direct the flow away from the returning
bucket to reduce the inverse drag. Saad et al. (Saad El-Deen et al., 2019,
2020), Ramadan et al. (2021), and Mohamed et al., 2011a, 2021
investigated various designs of stators. Some of the modified rotor
profiles resulted from parametric studies of the rotor parameters, as
reported by Shashikumar et al. (2021a), Shashikumar et al. (2021b),
Meri et al. (2019), and Roy and Saha (2015). Optimization software also
aided in varying the profile of the rotor to enhance its performance, as
reported by Mohamed et al. (2011b), Kerikous et al. (Kerikous and
Thévenin, 2019), and Chan et al. (Chan, Bai, He). Sharma et al. (Sharma
and Sharma, 2016) investigated using two-quarter auxiliary blades in
front of the original blade of a conventional Savonius turbine. Al-Gh­
riybah et al. (2019) studied the effect of varying the angle of an inner
auxiliary blade parallel to the original blade.

1.1. Present work novelty

The blade arc angle affects the performance of the Savonius rotor.
The initial design of the conventional Savonius used a 180◦ semi-circular
profile. Mao et al. (Mao and Tian, 2015) investigated the effect of
varying the blade arc angle of the conventional profile from 150◦ to 200◦
by varying the position of the outer endpoint (P1) and fixing the position
of the inner endpoint (P2) (see Fig. 1). Also, the authors changed the
position of the outer endpoint (P1) by trimming the constant bucket
diameter (D) to get a specific blade arc angle, which neglected the
variation in Dr. The maximum performance was achieved at 160◦ , which
enhanced the best Cp by 8.37%. Driss et al. (2015) investigated varying
the blade arc angle of unconventional Savonius rotors. The authors were
interested in demonstrating the total variation of pressure and velocity
Fig. 1. Conventional Savonius rotor. without showing the change in the output power. The optimum blade
arc angle for a Bach-type Savonius rotor is 124◦ , according to Saad
The working mechanism relies on the pressure difference between the El-Deen et al. (Saad et al., 2020). Several previous research works
concave and convex sides of each blade, which is converted into drag optimized the rotor profile by fixing the position of the inner and outer
force. The variation between the advancing drag and the returning drag endpoints and changing the position of the points that govern the rotor
allows the rotation of the Savonius rotor (Saad et al., 2020). The per­ profile, such as (Mohamed et al., 2011b; Kerikous and Thévenin, 2019;
formance of a Savonius turbine is monitored by the power coefficient Kerikous and Thévenin, 2019; Chan, Bai, He).
(Cp), which is the ratio between the power extracted by the rotor and the The present research aims to enhance the performance of the con­
available wind power. ventional Savonius rotor by varying the position of the endpoints of the
Since the initial design developed by S.J. Savonius (1931), several blades. Two methods were used in the current study using three con­
previous research works have been conducted to evaluate the perfor­ figurations. The first method involved varying the blade arc angles of the
mance characteristics of the Savonius turbine. Two major areas were conventional Savonius rotor by dividing the bucket profile into two
studied in previous investigations. Firstly, the geometrical parameters of angles, as shown in Fig. 1. The outer blade arc angle (ψ ) is measured
the conventional design were investigated. The aspect ratio (AR) is the from the rotor center to the position of the outer endpoint (P1). The
ratio between rotor height (H) to rotor diameter (Dr). According to inner blade arc angle (θ) is measured from the bucket center to the
Sheldahl et al. (1978) and Ferrari et al. (2017), the best aspect ratio for position of the inner endpoint (P2). The second method involves varying
the conventional rotor is 1. The overlap distance (e) is the horizontal the position of the endpoint by changing the gap ratio (δ) and changing
distance between the inner endpoints (P2) of the Savonius rotor. Varying the outer blade arc angle (ψ ) for the best gap ratio. The main concept
the overlap distance is an effective method to improve the performance behind changing the outer blade arc angle is to minimize the inverse
of Savonius rotor. For the semi-circular Savonius rotor, an optimum drag on the returning blade by varying the position of the outer endpoint
overlap ratio (e/Dr) equal to 0.15 was recommended by Fujisawa (1992)
and Akwa et al. (2012). Using circular endplates is one of the most Table 1
effective methods to enhance the performance of a Savonius turbine. Rotor design parameters.
The endplate size ratio is the ratio between the diameter of the endplate
Number Parameter Value Dimension
(Do) to the rotor diameter (Dr). The best endplate size ratio is 1.1, ac­
1 Rotor diameter (Dr) 300 mm
cording to Saad et al. (2020) and Jeon et al. (2015). The gap ratio (δ) is
2 Endplate diameter (Do) 330 mm
the ratio between the vertical distance between the buckets (a) to the 3 bucket diameter (D) 180 mm
rotor diameter (Dr). The best gap ratio conventional Savonius rotor is 4 Shaft diameter (Dsh) 6.75 mm
0.133, according to Abdelaziz et al. (2022). Saha et al. (2018) investi­ 5 Overlap ratio (e/Dr) 0.2 –
gated the number of blades for the conventional rotor. Two-bladed 6 Aspect ratio (H/Dr) 1 –
7 Rotor thickness 1 mm

2
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

(P1). Changing the position of the inner endpoint (P2) improves the flow source terms.
characteristics in the overlap region, which enhances the advancing ANSYS fluent calculates the moment by integrating the pressure
drag. Varying the gap ratio simulates the effect of changing the inner readings over the rotor surface. Then, each force value (integral pres­
endpoint (P2). The bucket diameter (D) is changed by varying the po­ sure) is multiplied by the arm length. The torque coefficient (Cm) was
sition of the endpoints. The rotor diameter (Dr) and Reynolds number calculated by dividing the moment by the theoretical torque.
(Re) must be preserved at constant values. For the fully turbulent flow,
T
the dimensionless parameters (Cp, Cm, λ) that control the aerodynamic Cm = 1 (5)
ρARV 2
performance of the Savonius rotor provide a distinctive fingerprint 2

curve. This fingerprint is the result of the full similarity principle (Kamal A represents the turbine’s frontal area, A = (Dr H), and R represents
et al., 2022). With changing Reynolds number value, the performance the radius of the rotor. V is the wind speed. The power coefficient (Cp) is
curve of the Savonius rotor is changed, according to Kamoji et al. (2008) calculated as:
and Hayashi et al. (2005). To validate the results of a certain study, the
Tω
Reynolds number must be preserved similar to that study. CP = λ C m = 1
(6)
ρ A V3
In this research, a three-dimensional numerical analysis of a con­ 2

ventional Savonius turbine was performed to study the impact of vary­ The tip speed ratio (λ) is calculated as:
ing the position of the endpoints of the rotor profile (P1) and (P2). The
ωR
outer blade arc angle (ψ ) is studied by varying the position of the outer λ= (7)
V
endpoint (P1) to decrease the inverse torque. The inner blade arc angle
(θ) was investigated by varying the inner endpoint (P2) to enhance the For the validation process, Re must be preserved constantly to the
flow characteristics in the overlap region, as shown in Fig. 1. The effect previous experimental and numerical results.
of varying the outer blade arc angle (ψ ) was also investigated for a
ρ V Dr
modified Savonius rotor, where the best gap ratio (δ) was obtained by Re = (8)
μ
Abdelaziz et al. (2022). The SST k-ω turbulence model is used to
perform the numerical simulation using ANSYS fluent. Three modified The time step size is determined by the rotational speed and time step
configurations were studied to perform higher than the conventional angle.
Savonius rotor. 2λv
ω= [rad / s] (9)
Dr
2. Modelling and validation
π time step angle
time step size = [s] (10)
2.1. CFD modelling 180 ω
The time step angle defines the number of time steps required to
Unsteady Reynolds Average Navier-Stokes equation (URANS) is the complete one rotor cycle. If the time step angle is one, the rotor requires
most suitable method for the current analysis, according to Saad et al.
360-time steps to complete a cycle. The best time step angle from pre­
(Saad El-Deen et al., 2019). Menter (1994) developed The SST k-ω vious studies was 0.5 according to (Saad et al., 2020; Baz et al., 2015).
turbulence model, which shares the merits of both k-ω and k-ε models
The numerical value of Cm at a certain λ changes with each rotor cycle
(ANSYS, 2013). The mathematical model for the present study depends until it settles with a small variation after a certain cycle. The final so­
on three governing equations of continuity, momentum, and turbulence.
lution is accepted for a certain case after reaching a difference of 1 ×
The governing equations for unsteady Newtonian incompressible tur­ 10− 4 between the average Cm for the last two cycles.
bulent flow are presented in tensor form. The continuity equation is The height of the first cell (y) from the rotor’s surface is a critical
described as parameter that must be estimated correctly before starting the solution.
∂ui According to Saad et al. (Saad El-Deen et al., 2019), the dimensionless
=0 (1) distance between the first layer and the wall is specified as yþ. Three
∂xi
regions could be found near the wall surface, the viscous sub-layer (yþ
The momentum equation is described as
≤ 5), the buffer layer (5 < yþ ≤ 30), and the fully turbulent layer (yþ >
∂ui ∂ui 1 ∂p μ ∂2 ui ∂ ( ′ ′) 30) (Saad et al., 2020). The inflation option is used to define the value of
+ uj = − + + uu (2) the height of the first layer. For the SST k-ω turbulence model, y þ
∂t ∂xj ρ ∂xi ρ ∂xi xj ∂xj i j
should be less than 1 (Shashikumar et al., 2021a). yþ is calculated as
xi and xj represent the flow direction; i and j = 1, 2, 3 are the Car­ follows:
tesian axes; ui represents the flow mean velocity; ui is the velocity
′

ρut y
fluctuation. The instantaneous velocity (ui ) is the sum of the velocity y+ = (11)
μ
fluctuation and the mean velocity, ui = ui + ui . ρ represents the air
′

density; p represents the fluid pressure; t represents the time. The term ut is the friction velocity, which is calculated as:
√̅̅̅̅̅
(ui uj ) is known by Reynolds stresses.
′ ′
τw
ut = (12)
The governing equations for the SST k-ω turbulence model are ρ
described as
[ ]
τ w is shear stress of the wall, which is calculated as:
∂(ρk) ∂(ρkui ) ∂ ∂k
+ = Γk + Gk − Yk + Sk (3) 1
∂t ∂xi ∂xj ∂xj τw = Ćf ρV 2 (13)
2
( ) [ ]
∂(ρω) ∂ ρωuj ∂ ∂ω Ćf is the local friction coefficient.
+ = Γω + Gω − Yω + Dω + Sω (4)
∂t ∂xj ∂xj ∂xj A SIMPLE algorithm was selected for pressure-velocity coupling,
according to (Saad et al., 2020). For the gradient spatial discretization,
Gk represents the generation of turbulence kinetic energy. Gω rep­
the least square cell-based option was utilized. The spatial pressure
resents the generation of ω. Y k and Y ω represents the dissipation of k and
discretization is set to second order, which improves the accuracy over
ω due to turbulence. Γ k and Γ ω are the effective diffusivity of k and ω.
the standard and linear schemes (Shashikumar et al., 2021b). According
Dω represents the term cross-diffusion. Sk and Sω represent user-defined

3
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

to (Saad et al., 2020; Saad et al., 2020; Saad El-Deen et al., 2020, 2020; various previous studies, such as (Saad et al., 2020). The exit surfaces
Baz et al., 2015), the second-order upwind option is enabled for the are set to a pressure outlet with an atmospheric pressure gauge value.
spatial discretization of momentum, turbulent kinetic energy, and tur­ The side surfaces are set to symmetrical sides.
bulent dissipation rate. The second-order implicit option is enabled for
transient formulation. For the transient model, the Sliding Mesh Model 2.3. Numerical independence studies
(SMM) was enabled for the rotating region (Saad et al., 2020). Before
the transient model, a steady model was used as an initialization step to Two independent investigations were performed in series to obtain
decrease the computation time required for the solution. The moving the optimum numerical solution for the current rotor geometry and
reference frame model was used for the steady model before starting the turbulence model. Domain size and the number of cells were the inde­
SMM model. pendent studies conducted in the current work. Increasing the size of the
computational domain and the number of cells will increase the accu­
2.2. Design parameters for turbine racy of the results and the solving time required for the solution. The
solving time was the main issue for previous numerical research because
The design parameters for the conventional Savonius rotor are unsteady models require advanced technologies, especially for 3-D
similar to the parameters used in the previous studies (Abdelaziz et al., simulations. The huge solving time for unsteady models forced most
2022; Baz et al., 2015; Baz et al., 2015, 2015; Ramadan et al., 2018), as previous researchers to conduct their investigations using 2-D models.
shown in Table 1. These values preserve the Reynolds number with a However, unsteady 3-D models are common nowadays. These two
value of Re = 184,838, which is very similar to the Reynolds value of the studies will help in finding the optimum solving time and accuracy of
experimental study by Sugiharto et al. (2020). Reynolds number is results. The effects of domain size and the number of cells were studied
calculated according to Equation (8). For wind rotors, the Reynolds separately. However, both studies have a mutual influence on the re­
number governs the relationship between rotor diameter (Dr) and wind sults. For studying the domain size, an assumption of the cell size of 8
speed (V). Defining the value of Reynolds number is vital to obtain the mm within the rotating region has been made with a constant growth
unique performance curve for Sugiharto et al. (2020). The computa­ rate of 1.2 to comply with the previous recommendations of (Saad et al.,
tional domain is divided into two main regions: the stationary and the 2020; Abdelaziz et al., 2022; Abdelaziz et al., 2022, 2022; Shashikumar
rotating regions, as shown in Fig. 2. A circular interface is used to et al., 2021b).
separate the two regions. The size of the interface defines the rotating
zone. The diameter of the circular interface should be very close to the 2.3.1. Domain size independence study
diameter of the endplate (Do). According to Abdelaziz et al. (2022), the The domain size study aims to reach the best size for the stationary
difference between the diameter of the interface and Do is 5 mm. region that preserves the required accuracy while increasing the solving
To avoid the huge number of cells in 3-D simulation and control the rate. For the current 3-D numerical model, four parameters describe the
cell distribution, the stationary region is divided into 27 parts. This total dimensions of the stationary domain, as shown in Fig. 2. The error
technique increases the number of cells near the rotating region and percentage between the numerical data and the experimental mea­
decreases the number of cells away from the rotating region. This surements of Sugihatro et al. (Sugiharto et al., 2020) was also moni­
technique improves the mesh quality and decreases the number of un­ tored. For the X-axis, the length of the stationary region is divided into
structured mesh due to the presence of the rotor, according to (Abdelaziz two regions: the upstream and the downstream Xup and Xdown,
et al., 2022). The boundary settings of the computational domain are respectively.
shown in Fig. 2. The inlet faces are set to velocity inlet with an incoming Xup is the upstream length taken from the wind source to the center
air velocity of 9 m/s, which was taken as an average velocity from of the rotor. The value of Xup is taken similar to (Abdelaziz et al., 2022;
Baz et al., 2015; Baz et al., 2015, 2015; Ramadan et al., 2018), which
equals Xup ¼ 6.5 Dr. Xdown is the distance from the rotor center to the
outlet faces. Y and Z are the width and height of the stationary region,
respectively. Xdown, Y, and Z parameters are studied in the present work.
The best downstream length is Xdown ¼ 20 Dr, as shown in Fig. 3. Fig. 4
shows that the best width of the stationary region should be Y ¼ 10 Dr.
The best height of the stationary region in the Z-axis with the rotor in the
middle of this distance should be Z ¼ 10 Dr, as shown in Fig. 5. The error
percentage for the optimum size of the stationary domain is 1.62%. The

Fig. 2. Computational domain and boundary conditions. Fig. 3. Effect of varying downstream length (Xdown) on Cm.

4
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

(Saad et al., 2020). Using only tetrahedral cells will increase the total
number of cells required to capture the small variation in performance
(Shashikumar et al., 2021b). The used mesh combination limited the
distribution of unstructured cells in the rotating region and the small
stationary region around the rotating zone, as shown in Fig. 6. The mesh
distribution increased the efficiency of the present numerical model,
which could be noticed from the low relative errors for various mesh
models. The relative error between the Cp of the present numerical
model and the experimental study of Sugiharto et al. (2020) measures
the effectiveness of these models. As the number of cells increases, the Cp
increases. After a certain number of cells (model 5), a very small vari­
ation in the value of Cp equals 1 × 10− 4 can be accepted.
The optimum number of cells at that point is 2,620,790. The relative
error for the best model is 1.62%. The results of the current indepen­
dence study are similar to the number obtained in (Saad et al., 2020;
Abdelaziz et al., 2022).
Despite the lowest relative error noticed for the third and fourth
Fig. 4. Effect of varying domain width (Y) on Cm. models, a large variation in the numerical Cp with increasing the number
of cells could be noticed. Increasing the number of cells more than the
fifth model caused a slight change in the numerical Cp. Fig. 6 (a) shows
the mesh distribution in the computational domain. A sectional X–Y
plane shows the macroscopic view of the mesh in a layer of the domain,
as shown in Fig. 6 (b). Fig. 6 (c) shows the distribution of the unstruc­
tured mesh in the rotating zone. The inflation properties are dependent
on the incoming flow velocity, characteristic length, fluid density, and
the value of yþ. For the employed turbulence model, y þ should be less
than one (Abdelaziz et al., 2022). Substituting y þ equals 0.5, Dr equals
300 mm, ρ equals 1.225 (kg/m3), and μ equals 1.789 × 10− 5 (kg/m.s) to
Equations (11)–(13) give the height of the first layer from the rotor’s
wall (y) equals 0.017 mm 15 inflation layers for the inflation option
were used, according to (Saad et al., 2020; Abdelaziz et al., 2022). The
value was also verified using online calculators.

2.4. Validation

The validation step is essential to check the accuracy of the present

numerical model. Validation has been performed for the conventional
Fig. 5. Effect of varying domain height (Z) on Cm. Savonius rotor using previous experimental and numerical studies. The
study by Sugiharto et al. (2020) was the main experimental study for the
size of the stationary domain is similar to the size of the computational validation process. The Reynolds number used in the current study
domain in (Saad et al., 2020; Saad et al., 2020; Abdelaziz et al., 2022). equals 184,838, which is close to the value used in (Sugiharto et al.,
2020). The experimental investigations of Hayashi et al. (2005) (Re =
2.3.2. Number of cells independence study 180,000) and Kamoji et al. (2008) (Re = 155,000) were used as other
An independent analysis of the number of cells is required after references to ensure the credibility and suitability of the present nu­
reaching the optimum domain size to determine the best number of cells merical model. Besides, the 2-D numerical studies of Mohamed et al.
to increase the solving rate and achieve the current model’s maximum (2011b) (Re = 210000) and Baz et al. (2015) (Re = 180000) were also
accuracy. Table 2 shows the seven models of cell numbers that were used for the validation process.
studied in the current work. The rate of growth of the cell size was kept The comparison between the Cm and λ for the current study with
constant at 1.2. The cell size of the rotor governs the number of cells previous numerical and experimental references is shown in Fig. 7.
preserving a constant growth rate for the computational domain. The Using 3-D simulation with the SST k-ω turbulence model for the present
cell size of the rotor for the first model was 20 mm, while for the seventh study made the results very close to the experimental measurements of
model it was 8 mm (Abdelaziz et al., 2022). Dividing the stationary (Kamoji et al., 2008; Hayashi et al., 2005; Hayashi et al., 2005, 2005;
region into 27 parts allows using a combination of structured (hexahe­ Sugiharto et al., 2020) than the 2-D numerical studies of (Mohamed
dral) cells and unstructured (tetrahedral) cells for mesh distribution et al., 2011b; Baz et al., 2015). The high accuracy of the present results
with the experimental studies could be explained as the 3-D numerical
simulation by applying the SST k-ω turbulence model simulates the
Table 2
Number of cells independence study. performance of the actual rotor by calculating specific errors, which
were not considered for the 2-D investigations of (Mohamed et al.,
Model Number of Cp Cp Experimental ( Relative
2011b; Baz et al., 2015).
cells Numerical ANSYS, 2013) error %
Fig. 8 shows the comparison between the Cp with λ for the current
1 745,445 0.1553 0.1728 +10.127
study with previous numerical and experimental research works. The
2 1,299,889 0.1674 +3.125
3 1,815,502 0.1708 − 1.157 maximum Cp for the present model for the conventional Savonius tur­
4 2,292,090 0.1748 +1.157 bine is 0.1756 at λ = 0.8, giving a relative error of 1.62% compared with
5 2,620,790 0.1756 + 1.620 (Sugiharto et al., 2020). The low error percentage for the present model
6 3,309,304 0.1755 +1.562 makes it a trusted model for future investigations into the Savonius
7 4,034,730 0.1754
turbine. The average value of Cm changes with increasing the number of
+1.504

5
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

Fig. 6. Mesh propagation in the computational domain.

rotor cycles until it settles at a specific cycle. For the present work, the
solution reached the required accuracy within the 5th cycle, which is
similar to (Saad et al., 2020; Abdelaziz et al., 2022). Fig. 9 shows the
variation in the instantaneous values of Cm according to the 5th and 6th
cycles.

3. Results and discussion

The current study aims to increase the power coefficient of the

conventional Savonius rotor by reducing the inverse drag on the
returning blade and increasing the advancing drag on the advancing
blade. The inverse drag is reduced by reducing the position of the outer
endpoint (P1) and the advancing drag is enhanced by varying the po­
sition of the inner endpoint (P2). The 1st configuration involves varying
the position of the outer endpoint (P1) of the rotor bucket to move along
the circumference of the rotor diameter (Dr) to change the outer blade
arc angle (ψ ) with preserving the inner endpoint (P2) at a fixed position,
as shown in Fig. 10 (a), which is conducted with a different method than
Mao et al. (Chen et al., 2017).
Fig. 7. Variation of Cm with λ of the present work with previous numerical and
The 2nd configuration involves varying the position of the inner
experimental references.
endpoint (P2) on the circumference of the bucket (r) by varying the inner
blade arc angle (θ) for the best ψ of the first configuration, as shown in
Fig. 10 (b). The 3rd configuration investigates the effect of varying the
gap ratio (δ) and the outer blade arc angle (ψ ), as shown in Fig. 10 (c).
The 3rd configuration aims to combine the effects of the 1st and 2nd

Fig. 8. Variation of CP with λ of the present work with previous numerical and
experimental references.

Fig. 9. Variation of the instantaneous Cm with Φ for the 5th and 6th cycles.

6
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

Fig. 10. Modified Savonius rotors.

configurations. The instantaneous CP is drawn with the variation of the

azimuth angle (Φ) for several cases in each configuration. These curves
give a microscopic view of the variation that occurred for each
configuration.

3.1. Performance of the 1st configuration

The first configuration involves varying the outer blade arc angle (ψ )
from 140◦ to 200◦ at λ = 0.8, as shown in Table 3. Accordingly, the
location of the outer endpoint (P1) is moved along the circumference of
the rotor diameter (Dr). Consequently, the bucket radius (r) is slightly
changed to preserve the location on the inner endpoint (P2). Fig. 11
demonstrates the variation of Cp with ψ for the 1st configuration at λ =
0.8. As ψ increases from 140◦ to 160◦ , Cp enhances until it reaches the

Table 3
Design parameters of the 1st configuration.
Case Blade arc angle (ψ ) Bucket radius (r) Rotor diameter (Dr) Cp
Fig. 11. Variation of Cp with ψ for the 1st configuration at λ = 0.8.
(◦ ) (m) (m)

1 140 0.09462 0.3 0.1745

2 150 0.0925 0.1782
maximum value of 0.1836 at ψ = 160◦ . However, at ψ > 160◦ , Cp de­
3 160 0.09111 0.1836 clines. For ψ > 180◦ , the extra angle delays the flow entrance to the
4 170 0.09027 0.1794 concave side of the advancing and returning blades, which reduces the
5 180 0.09 0.1756 average value of the Cp. The maximum value of the Cp for the 1st
6 190 0.09 0.1602
configuration is 0.1836 compared with that of 0.1756 for the
7 200 0.09 0.1406

7
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

conventional Savonius turbine. The percentage enhancement in the Cp

for the 1st configuration compared with the conventional Savonius
turbine equals 4.55% at λ = 0.8. According to Table 4, the maximum
improvement in the power performance was achieved by increasing the
bucket radius with a ratio of (r3/r5) equals 1.012.
Fig. 12 shows the variation of the instantaneous Cp with the azimuth
angle (Φ) for each case of the 1st configuration at λ = 0.8. For the values
of ψ equal to 140◦ and 200◦ , there is a negative value for the instanta­
neous Cp that reduces the average Cp generated by the rotor. At ψ =
190◦ , the modified bucket shifts up the profile of the instantaneous Cp to
avoid the negative values of Cp. For ψ = 150◦ , 160◦ , and 170◦ , the
modified buckets increase the amplitude of the instantaneous Cp, which
increases the average Cp generated.

3.2. Performance of the 2nd configuration

After reaching the optimum blade arc angle from the 1st configura­
tion, the position of the inner endpoint (P2) of the arc bucket is changed
for the 2nd configuration. For this modification, the inner blade arc
angle (θ) is varied while preserving other parameters. This design is a
method to apply the second concept to increase the flow usage in the Fig. 12. Variation of instantaneous Cp with Φ for the 1st configuration.
overlap region. At λ = 0.8, θ is studied with values ranging from 0◦ to
30◦ . At 0◦ < θ ≤ 20◦ , the Cp improves until it reaches the maximum value
of 0.1983 at θ = 20◦ , as shown in Fig. 13. For θ > 20◦ , the Cp declines.
The percentage enhancement in the Cp at λ = 0.8 equals 12.9% and 8.0%
compared with the conventional and the 1st configuration Savonius
rotors, respectively.
The instantaneous effect of varying the inner blade arc angle (θ) on
Cp for a full cycle of rotation for the 2nd configuration is shown in
Fig. 14. At 0◦ < θ ≤ 20◦ , the maximum instantaneous Cp increases with a
slight variation in the minimum instantaneous Cp, which increases the
average Cp. At θ = 30◦ , the modified bucket blocks part of the advanced
blade and decreases the flow passage through the overlap region
resulting in a decrease in the instantaneous value of the Cp. Conse­
quently, the minimum instantaneous value of the Cp of the modified
bucket reaches zero. The best value of θ equals 20◦ .

3.3. Performance of the 3rd configuration

The 3rd configuration aims to investigate the performance of a

modified Savonius rotor, where the gap ratio (δ) and the outer blade arc
angle (ψ ) are the parameters that affect the performance. The 3rd Fig. 13. Variation of Cp with θ for the 2nd configuration at λ = 0.8.
configuration is a trial to combine the effect of the variations obtained in
1st and 2nd configurations.

3.3.1. The effect of varying the gap ratio (δ)

The gap ratio is studied with values from 0 to 0.2667 at λ = 0.8 at an
outer blade arc angle (ψ ) of 180◦ . Fig. 15 shows the effect of δ compared
with the conventional Savonius rotor. As δ increases, the Cp rises until a
maximum value of 0.1996 is achieved at δ = 0.133. At δ > 0.133, the
value of Cp decreases. The percentage improvement in the Cp at λ = 0.8
equals 13.6%, 8.7%, and 0.65% compared with the conventional, 1st,
and 2nd configurations of Savonius rotors, respectively. Fig. 16 illus­
trates the effect of δ on the variation of the instantaneous Cp at λ = 0.8.
At δ > 0.166, there is a negative value of the instantaneous Cp that

Table 4
Design parameters of the 3rd configuration at δ ¼ 0.1332.
Case Blade arc angle (ψ ) Bucket radius (r) Rotor diameter (Dr) Cp
(◦ ) (m) (m)

1 130 0.10236 0.3 0.182

2 140 0.098324 0.1885
3 150 0.09509 0.1936
4 160 0.09257 0.1952
Fig. 14. Variation of instantaneous Cp with Φ for the 2nd configuration.
5 170 0.09066 0.1965
6 180 0.08933 0.1996

8
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

Fig. 17. Variation of Cp with ψ for the 3rd configuration at λ = 0.8.

Fig. 15. Variation of Cp with δ for the 3rd configuration at λ = 0.8.

Fig. 16. Variation of instantaneous Cp with Φ for the 3rd configuration. Fig. 18. Variation of instantaneous Cp with Φ for the 3rd configuration.

reduces the average Cp generated by the rotor. But, at δ = 0.166, the reduces the average Cp. The best value of ψ equals 180◦ at δ = 0.1332.
modified bucket shifts up the profile of the instantaneous Cp to avoid the
negative values of Cp. In addition, at δ < 0.166, the modified bucket 3.4. Instantaneous Cp for various configurations
increases the amplitude of the instantaneous Cp, which increases the
average Cp generated. The azimuth angles (Φ), where the maximum and minimum values
of the instantaneous Cp occur, are essential to display the pressure and
3.3.2. The effect of varying the outer blade arc angle (ψ ) velocity contours for each configuration. Fig. 19 shows the instanta­
This section investigates the effect of varying ψ with the best gap neous Cp with the variation of the azimuth angle (Φ) for the 5th cycle of
ratio for the third configuration on the turbine performance. According the conventional and modified Savonius rotors. All the modified rotors
to the variation between the axis of the bucket center and the center of produce higher values of the maximum instantaneous Cp than the con­
the rotor diameter, the values of ψ > 180◦ should not be used. The ventional Savonius rotor. The maximum instantaneous Cp for conven­
modification involves varying the outer blade arc angle (ψ ) from 130◦ to tional Savonius rotor occurs at Φ = 21◦ and 201◦ . For the 1st, 2nd, and
180◦ at λ = 0.8 and δ = 0.1332, as shown in Table 4. The value of the 3rd configurations, the azimuth angle of the maximum instantaneous Cp
bucket radius (r) should be changed to adjust the value of ψ . As shown in equals (Φ = 12◦ , 192◦ ), (Φ = 14.5◦ , 194.5◦ ), and (Φ = 33◦ , 213◦ ),
Fig. 17, at ψ < 180◦ , the Cp decreases. The reduction in performance respectively.
could be explained by the increase in the bucket ratio (r). The maximum
value of Cp equals 0.1952 at ψ = 180◦ with δ = 0.1332.
3.5. Pressure contours
Fig. 18 shows the effect of ψ on the instantaneous Cp at δ = 0.1332.
At ψ < 180◦ , there is a reduction in the lowest value of the instantaneous
To understand the variation resulting from each configuration,
Cp that reduces the average Cp generated by the rotor. Also, the
contours of pressure and velocity are used at Φ = 33◦ and 213◦ , where
maximum of the instantaneous Cp increased, which increased the
the best instantaneous Cp is reached for the 3rd configuration at λ = 0.8,
average Cp generated. However, the reduction in the instantaneous Cp
see Fig. 20. For the conventional Savonius rotor, the pressure distribu­
value is larger than the increase in the maximum Cp value, which
tion on the buckets is like (Abdelaziz et al., 2022; Baz et al., 2015; Baz

9
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

distributions on the concave surface of the advancing bucket and the

convex surface of the returning bucket, which increased the advancing
drag, as shown in Fig. 20 (c). For the 3rd configuration, the maximum
pressure scale has the lowest value compared to other configurations, as
shown in Fig. 20 (d).

3.6. Velocity contours

The velocity distribution for various rotors is presented at Φ = 33◦

and 213◦ , where the best instantaneous Cp is reached for the 3rd
configuration at λ = 0.8, as shown in Fig. 21. The velocity distribution
for the conventional rotor is very close to (Abdelaziz et al., 2022; Baz
et al., 2015; Baz et al., 2015, 2015; Ramadan et al., 2018). In the 1st
configuration, the location of the high-velocity region was moved to the
tip of the concave surface of the advancing bucket, as shown in Fig. 21
(b). Any minor changes in the size and location of the high velocity and
stagnant air regions in the vicinity of the advanced and returning
buckets, respectively, cause a change in the performance of such turbine
configurations.

Fig. 19. Variation of instantaneous Cp with Φ for conventional and modified 3.7. Performance of modified rotors
Savonius configurations.
Fig. 22 shows the variation in the Cp for the tip speed ratio (λ) for
et al., 2015, 2015; Ramadan et al., 2018). For the 1st configuration, conventional and modified rotors. The maximum value of the Cp for the
reducing the value of ψ decreases the area of the returning bucket conventional Savonius rotor at λ = 0.8 equals 0.1756. On the other
affected by the high-pressure region, which in turn aids in reducing the hand, the maximum values of the Cp for the 1st, 2nd, and 3rd configu­
inverse drag, as shown in Fig. 20 (b). The 2nd and 3rd configurations rations are 0.1836, 0.1983, and 0.1996, respectively. The percentage
involved using two different methods to improve the flow characteristics increase in the maximum Cp for the three configurations compared to
in the overlap region. These methods improved the pressure the conventional design are 4.5%, 12.9%, and 13.6%, respectively.

Fig. 20. Pressure distribution for various configurations at λ = 0.8, and Φ = 33◦ and 213◦ .

10
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

Fig. 21. Velocity distribution for various configurations at λ = 0.8, and Φ = 33◦ and 213◦ .

zone. Three configurations were investigated in the present work to

achieve the two concepts. The 1st configuration involved varying the
position of the outer endpoint (P1) by an angle (ψ ) and fixing the inner
endpoint (P2). The 2nd configuration involved changing the position of
the inner endpoint (P2) by an angle (θ) for the best model from the first
configuration. The 3rd configuration involved varying the gap ratio (δ)
to employ a similar concept to the 2nd configuration. Also, the 3rd
configuration involved decreasing the outer blade arc angle of the best
gap ratio. The modified configurations resulted in an enhancement in
the Savonius turbine performance. The performance variation for the
used configurations of Savonius rotors revealed that:

• For the 1st configuration, a maximum Cp of 0.1836 was obtained at

ψ = 160◦ and λ = 0.8. The increase in the maximum Cp is 4.5%
higher than the conventional Savonius rotor.
• For the 2nd configuration, a maximum Cp of 0.1983 at θ = 20◦ and λ
= 0.8. The increase in the maximum Cp is 12.9% higher than the
conventional Savonius rotor.
• For the 3rd configuration, a maximum Cp of 0.1996 was obtained at
Fig. 22. Variation of Cp with λ for various configurations.
δ = 0.1333, ψ = 180◦ and λ = 0.8. The increase in the maximum Cp is
13.6% higher than the conventional Savonius rotor.
Hence the 3rd configuration has the privilege over the other configu­
• Changing the gap ratio (δ) and varying the position of the endpoints
rations in enhancing the performance of such a turbine.
(P1, P2) improved the performance of the semi-circular Savonius
rotor.
4. Conclusion • Varying the gap ratio (δ) is one of the most effective methods of
enhancing the performance of a conventional Savonius rotor.
Two main concepts were investigated in the present work to improve • Changing the blade arc angles should be carefully studied for newly
the performance of the Savonius turbine. The 1st concept involved developed rotors, as it may cause an enhancement or reduction in
changing the outer blade arc angle (ψ ) to reduce the inverse drag and performance.
improve the advancing drag by changing the bucket radius (r). The 2nd
concept involved improving the characteristics of the flow in the overlap

11
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

• Further automatic optimization is still needed in the future using Meri, S.A.R., Salleh, H.B.B., Nemah, M.N., Al-Quraishi, B.A., Asmuin, N.Z.B., 2019.
Performance Evaluation of Savonius wind turbine based on a new design of blade
optimization algorithms to optimize the performance under these
shape. Int. J. Mech. Eng. Technol. 10 (Issue 1), 837–846.
design parameters. Minetto, R.A.L., Paraschivoiu, M., 2020. Simulation based analysis of morphing blades
applied to a vertical axis wind turbine. Energy 202, 117705.
Mohamed, M.H., Janiga, G., Pap, E., Thevenin, D., 2011a. Optimal blade shape of a
CRediT authorship contribution statement modified Savonius turbine using an obstacle shielding the returning blade. Energy
Convers. Manag. 52, 236–242.
Khaled R. Abdelaziz: Conceptualization, Methodology, Writing – Mohamed, M.H., Janiga, G., Pap, E., Thevenin, D., 2011b. Optimal blade shape of a
modified Savonius turbine using an obstacle shielding the returning blade. Energy
original draft. Mohamed A.A. Nawar: Conceptualization, Methodol­ Convers. Manag. 52, 236–242.
ogy, Validation, Software. Ahamed Ramadan: Data curation, Writing – Mohamed, M.H., Alqurashi, F., Thevenin, D., 2021. Performance enhancement of a
original draft. Youssef A. Attai: Conceptualization, Methodology, Savonius turbine under the effect of frontal guiding plates. Energy Rep. 7,
6069–6076.
Writing – original draft. Mohamed H. Mohamed: Supervision,
Ramadan, A., Yousef, K., Said, M., Mohamed, M.H., 2018. Shape optimization and
Conceptualization, Software, Writing – review & editing. experimental validation of a drag vertical axis wind turbine. Energy 151, 839–853.
Ramadan, A., Hemida, M., Abdel-Fadeel, W.A., Aissa, W.A., Mohamed, M.H., 2021.
Comprehensive experimental and numerical assessment of a drag turbine for river
Declaration of competing interest hydrokinetic energy conversion. Ocean Eng. 223, 108587.
Rezaeiha, A., Kalkman, I., Blocken, B., 2017. Effect of pitch angle on power performance
and aerodynamics of a vertical axis wind turbine. Appl. Energy 197, 132–150.
The authors declare that they have no known competing financial Roy, S., Saha, U.K., 2015. Wind tunnel experiments of a newly developed two-bladed
interests or personal relationships that could have appeared to influence Savonius-style wind turbine. Appl. Energy 137, 117–125.
Saad El-Deen, A.E., Abd El-Maksoud, R.M., Nawar, M.A.A., Attai, Y.A., 2019.
the work reported in this paper.
A numerical study of the effect of the stator on Savonius bach-type rotor
performance. In: Thirteen International Conference of Fluid Dynamics, pp. 21–22.
References Cairo, Egypt.
Saad El-Deen, A.E., Nawar, M.A.A., Attai, Y.A., Abd El-Maksoud, R.M., 2020. On the
enhancement of Savonius Bach-type rotor performance by studying the optimum
Abdelaziz, K.R., Nawar, M.A.A., Ramadan, A., Attai, Y.A., Mohamed, M.H., 2022.
stator configuration. Ocean Eng. 217, 107954.
Performance improvement of a Savonius turbine by using auxiliary blades. Energy
Saad, A.S., El-Sharkawy II, Ookawara, S., Ahmed, M., 2020. Performance enhancement
244 (A), 122575.
of twisted-bladed Savonius vertical axis wind turbines. Energy Convers. Manag. 209,
Akwa, J.V., Da Silva Júnior, G.A., Petry, A.P., 2012. Discussion on the verification of the
112673.
overlap ratio influence on performance coefficients of a Savonius wind rotor using
Saha, U.K., Thotal, S., Maity, D., 2018. Optimum design configuration of Savonius rotor
computational fluid dynamics. Renew. Energy 38 (1), 141–149.
through wind tunnel experiment. J. Wind Eng. Ind. Aerod. 96, 1359–1375.
Al-Ghriybah, M., Zulkaflia, M.F., Didane, D.H., Mohd, S., 2019. The effect of inner blade
Savonius, S.J., 1931. The S-rotor and its applications. Mech. Eng. 53, 333–341.
position on the performance of the Savonius rotor. Sustain. Energy Technol.
Sharma, S., Sharma, R.K., 2016. Performance improvement of Savonius rotor using
Assessments 2019 (36), 100534.
multiple quarter blades – a CFD investigation. Energy Convers. Manag. 127, 43–54.
ANSYS, 2013. Fluent Theory Guide Release 15.0. ANSYS, Inc.
Sharma, S., Sharma, R.K., 2017. CFD investigation to quantify the effect of layered
Baz, A.R., Mahmoud, N.A., Hamed, A.M., Youssef, K.M., 2015. Optimization of two and
multipleminiature blades on the performance of Savonius rotor. Energy Convers.
three rotor Savonius wind turbine. Proceedings of ASME Turbo Expo. In: Turbine
Manag. 144, 275–285.
Technical Conference and Exposition.
Shashikumar, C.M., Honnasiddaiah, R., Hindasageri, V., Madav, V., 2021a. Experimental
Bianchini, A., Ferrara, G., Ferrari, L., 2015. Design guidelines for H-Darrieus wind
and numerical investigation of novel V-shaped rotor for hydropower utilization.
turbines: optimization of the annual energy yield. Energy Convers. Manag. 89,
Ocean Eng. 224, 108689.
690–707.
Shashikumar, C.M., Hindasageri, V., Madav, V., 2021b. Numerical investigation of
Bonalumi, D., Giuffrida, A., 2016. Investigations of an air-blown integrated gasification
conventional and tapered Savonius hydrokinetic turbine for low-velocity
combined-cycle fired with high-sulphur coal with post-combustion carbon capture
hydropower application in an irrigation channel. Sustain. Energy Technol.
by aqueous ammonia. Energy 117, 439–449.
Assessments 43, 100871.
Chan CM, Bai HL, He DQ. Blade Shape Optimization of the Savonius Wind Turbine Using
Sheldahl, R.E., Feltz, L.V., Blackwell, B.F., 1978. Wind tunnel performance data for two
a Genetic Algorithm.
and three bucket Savonius rotors. J. Energy 2 (3), 160–164.
Chen, W.H., Chen, C.Y., Huang, C.Y., Hwang, C.J., 2017. Power output analysis and
Sugiharto, B., Soeparman, S., Widhiyanuriyawan, D., Wahyudi, S., 2020. Characteristics
optimization of two straight-bladed vertical-axis wind turbines. Appl. Energy 185,
of the Savonius wind turbine using multiple guide vanes. Int. J. Fluid Mach. Syst. 13
223–232.
(No. 3). July-September.
Driss, Z., Mlayeh, O., Driss, S., Driss, D., Maaloul, M., Abid, A.S., 2015. Study of the
Toja-Silva, F., Colmenar-Santos, A., Castro-Gil, M., 2013. Urban wind energy exploitation
bucket design effect on the turbulent flow around unconventional Savonius wind
systems: behavior under multidirectional flow conditions - opportunities and
rotors. Energy 89, 708–729.
challenges. Renew. Sustain. Energy Rev. 24, 364–378.
Edelenbosch, O.Y., Kermeli, K., Crijns-Graus, W., Worrell, E., Bibas, R., Fais, B.,
Wang, Y., Sun, X., Dong, X., Zhu, B., Huang, D., Zheng, Z., 2016. Numerical investigation
Fujimori, S., Kyle, P., Sano, F., van Vuuren, D.P., 2017. Comparing projections of
on aerodynamic performance of a novel vertical axis wind turbine with adaptive
industrial energy demand and greenhouse gas emissions in long-term energy models.
blades. Energy Convers. Manag. 108, 275–286.
Energy 122, 701–710.
Yannopoulos, S.I., Lyberatos, G., Theodossiou, N., et al., 2015. Evolution of water lifting
Ferrari, G., Federici, D., Schito, P., Inzoli, F., Mereu, R., 2017. CFD study of Savonius
devices (Pumps) over the centuries worldwide. Water 7 (9), 5031–5060.
wind turbine: 3D model validation and parametric analysis. Renew. Energy 105,
722–734.
Fujisawa, N., 1992. On the torque mechanism of Savonius rotors. J. Wind Eng. Ind.
Nomenclature
Aerod. 40 (3), 277–292.
Goodarzi, M., Keimanesh, R., 2015. Numerical analysis on the overall performance of
Savonius turbines adjacent to a natural draft cooling tower. Energy Convers. Manag. A: Rotor area projected to the wind (m2)
99, 41–49. AR: Aspect ratio (Hr/Dr) (− )
Hayashi, T., Li, Y., Hara, Y., 2005. Wind tunnel tests on a different three-stage Savonius a: Gap distance (m)
rotor. JSME International Journal. Ser. Bibliogr. 48, 9–16. Cm: Rotor torque coefficient (− )
Jeon, K.S., Jeong, J.I., Pan, J.K., Ryu, K.W., 2015. Effects of end plates with various Cp: Rotor power coefficient (− )
shapes and sizes on helical Savonius wind turbines. Renew. Energy 79, 167–176. D: Bucket diameter (m)
Johari, M.K., Jalil, M.A.A., Shaiff, M.F.M., 2018. Comparison of horizontal axis wind Dr: Rotor diameter (m)
turbine (HAWT) and vertical axis wind turbine (VAWT). Int. J. Eng. Technol. 7, Dsh: Shaft diameter (m)
74–80. e: Overlap distance (m)
Kamal, A.M., Nawar, M.A.A., Attai, Y.A., Mohamed, M.H., 2022. Blade design effect on H: Rotor height (m)
Archimedes spiral wind turbine performance: experimental and numerical P: Pressure (Pa)
evaluations. Energy 250, 123892. P1: The outer endpoint of the bucket (− )
Kamoji, M.A., Kedare, S.B., Prabhu, S.V., 2008. Experimental investigations on single P2: The inner endpoint of the bucket (− )
stage, two stage and three stage conventional Savonius rotor. Int. J. Energy Res. 32, R: Rotor radius (m)
877–895. r: Bucket radius (m)
Kerikous, E., Thévenin, D., 2019. Optimal shape of thick blades for a hydraulic Savonius Re: Reynolds number (ρ.V.Dr/μ)
turbine. Renew. Energy 134, 629–638. T: Rotor output torque (N.m)
Mao, Z., Tian, W., 2015. Effect of the blade arc angle on the performance of a Savonius t: Time (s)
wind turbine. Adv. Mech. Eng. 7 (5), 1–10. V: Wind speed (m/s)
Menter, F.R., 1994. Two equations eddy-viscosity turbulence models for engineering Xdown: Downstream domain length (m)
applications. AIAA J. 32, 1598–1605. Xup: Upstream domain length (m)

12
K.R. Abdelaziz et al. Ocean Engineering 260 (2022) 112054

Y: Width of the domain (m) λ: Tip speed ratio (− )

y: Height of the first layer (m) τw : Wall shear stress (Pa)
y+ : Non-dimensional wall distance (− ) δ: Gap ratio (− )
Z: Height of the domain (m)

Abbreviations
Greek symbols CFD: Computational fluid dynamics
Φ: Azimuth angle (◦ ) HAWT: Horizontal-axis wind turbine
θ: Inner blade arc angle (◦ ) k-ε: Turbulence kinetic energy - turbulent dissipation
ψ : Outer blade arc angle (◦ ) RANS: Reynolds-Averaged Navier-Stokes
μ: Dynamic viscosity of the air (kg/m.s) SMM: Sliding mesh model
ut: Friction velocity (m/s) SST kω: Shear stress transport - Turbulence kinetic energy - turbulent dissipation rate
ρ: The density of the air (kg/m3) VAWT: Vertical axis wind turbine
ω: Rotational speed (rad/s)

13