Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira.

ISSN 0122-701 y ISSN-e: 2344-7214 397

Design, assembly and experimental tests of a

Savonius type wind turbine
Diseño, montaje y pruebas experimentales de una turbina eólica tipo Savonius

E. A. Salazar-Marín ; A. F. Rodríguez-Valencia

 Abstract— The present research project focuses on the Savonius, se realizaron pruebas experimentales en el laboratorio
development of a Savonius type wind turbine, where a literature de fluidos y máquinas hidráulicas de la Universidad Tecnológica
review is initially carried out to recognize the appropriate de Pereira. Como resultados importantes del análisis (simulación
parameters for its design such as turbulence, wind velocity, and y experimentación) se obtuvo que el coeficiente de potencia de la
air density, followed by a design methodology to achieve the turbina eólica tiene un valor de 0,2 y la potencia mecánica en el eje
dimensioning by means of resistance calculation and selection of de la turbina varía con el cubo de la velocidad del viento. Con
component materials, after having the final concept defined. turbinas de este tipo se pueden lograr potencias mayores a 100 W
Through the use of computational tools such as Solidworks, con área de barrido de 1m2 y velocidades de viento de 10 m/s.
Inventor, Simulation Mechanical, the design stage was validated Palabras claves— Turbina eólica Savonius, energía eólica,
through simulation to study the dynamic fluid behavior and the velocidad del viento, coeficiente de potencia, potencia mecánica.
components were modeled to evaluate their response to wind
loads, verifying their resistance. Subsequently, according to the
modeling, the detailed drawings of the components of the turbine
were obtained, with which they were assembled. Finally, having I. INTRODUCTION
the physical model of the Savonius wind turbine, experimental
tests were carried out in the laboratory of Fluids and Hydraulic
Machines of the Technological University of Pereira. As an
W IND energy is a renewable energy source which can be
exploited to transform it through mechanical devices
such as a wind turbine, where it converts the kinetic energy of
important result of analysis (simulation and experimental tests) a
power coefficient of 0,2 was obtained and shaft mechanical power the wind into mechanical energy of the axis [1] [2]. There are
varies with cubic velocity wind. Power over 100 W can be reached mainly two large groups of wind turbines that have been
with sweep area 1 m2 and 10 m/s velocity turbines. designed for such process and are classified according to the
orientation of the turbine shaft. They are vertical axis turbines
Index Terms— Savonius wind turbine, wind energy, wind speed, and horizontal axis turbines [3]. The horizontal axis turbines
power coefficient, mechanical power. that are found in the majority of wind farms are constituted by:
foundation, tower, gondola with power train, rotor and paddles.
Resumen--- El presente proyecto de investigación se centra en el
According to the Guide on Mini-Wind Technology [4], in the
desarrollo de un aerogenerador de tipo Savonius, en el que
inicialmente se realiza una revisión de la literatura para reconocer case of horizontal axis wind turbines, the rotor can be
los parámetros apropiados para su diseño, como la turbulencia, la windward, that is, in the direction of wind incidence in front of
velocidad del viento y la densidad del aire, seguido de una the tower, or to leeward, in which case the rotor is located
metodología de diseño para lograr el Dimensionamiento mediante behind the tower in the dominant direction of the wind. Most of
cálculo de resistencia y selección de materiales componentes, luego the wind turbines are rotor to windward of the tower, which
de haber definido el concepto final. Mediante el uso de means that they require some guidance system. In the case of a
herramientas computacionales como Solidworks, Inventor, downwind rotor, the rotor is self-orientating, which simplifies
Simulation Mechanical, la etapa de diseño se validó a través de la
its design.
simulación para estudiar el comportamiento dinámico del fluido y
Horizontal axis turbines are more efficient than vertical axis
los componentes se modelaron para evaluar su respuesta a las
cargas de viento, verificando su resistencia. Posteriormente, de turbines, are more proven, are cheaper and there are many
acuerdo con el modelado, se obtuvieron los planos detallados de products to choose from. However, they have the difficulty of
los componentes de la turbina, con los cuales fueron ensamblados. supporting the continuous orientations and their efficiency is
Finalmente, teniendo en cuenta el modelo físico del aerogenerador reduced by operating in a turbulent regime.

This manuscript was sent on February 25, 2019 and accepted on September A. F Rodríguez-Valencia, is MSc of Technological university of Pereira
23, 2019. It was developed with own resources to build the turbine, and tests and currently he works such a Professor in Energy department of Autonomous
were carried out in fluids laboratory of university (UTP). University of Occident (Cali) (e-mail: [email protected]).
E. A. Salazar-Marín, is with Technology Department of Technological
University of Pereira, CP. 660003 (e-mail: [email protected]).
398 Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira

Vertical axis turbines are oriented in the predominant wind

direction due to their symmetry, are less sensitive to high
turbulence conditions and produce less vibration, these
conditions make them ideal for integrations in residential, urban
and even buildings [5]. On the other hand, their efficiency is
lower compared to horizontal axis and they are not very proven
since they are now in full development. There are two types of
vertical turbines, those based on drag and those based on lift.
The first one is less efficient, but usually less robust.
Within the vertical axis turbines are the Savonius type turbines
which are used to convert the wind power in torque on a rotating Fig. 1. Profile of the blade. [7]
axis. "The vertical axis turbines were invented by the Finnish
engineer Sigurd J. Savonius in 1922", they can start with a slow The overlap relation is a dimensionless parameter that relates
wind velocity, being very easy to manufacture; It has a small the perpendicular distance between the cross and the center of
turning speed and its performance is relatively low. rotation with the chord line length which is given as (1) [8]:
The research work emphasizes Designing the physical
components of a wind turbine type Savonius based on the 𝑒
𝑜𝑣𝑒𝑟𝑙𝑎𝑝 𝑟𝑒𝑙𝑎𝑡𝑖𝑜𝑛 = 𝑅𝑠 = (1)
𝑑
design parameters of this type of rotors and the parameters that
characterize the use of wind energy such as wind speed, air
Where 𝑒 is the distance between the inner tips of the blades
density of a region and turbulence. Having the selected final
that have overlap and 𝑑 corresponds to the value of the semi-
concept and the sizing of the wind turbine, modeling and
cylindrical blade diameter which in turn coincides with the
simulation of the rotor is carried out. Subsequently, the
length of the rope. Figure 2 schematizes the relationship that
construction of the wind turbine is made according to the
has been said.
obtained plans and finally experimental tests are carried out in
In this way, the overlap ratio is a design parameter of the
the laboratory.
Savonius rotor, several authors have studied the optimal overlap
ratio in rotors with 2 blades, which is the particular case of the
II. DESIGN PARAMETERS OF THE SAVONIUS TYPE WIND
present investigation. According to Menet, J et al. [7] the
TURBINE
overlap ratio in a two-bladed rotor is in the range of 0.15 to 0.3
Several investigations [6], [7], [8] have been made regarding times the length of the blade.
the study of the performance of the Savonius rotor from
different geometrical configurations of its rotor. These studies
have been made based on numerical computational analysis
taking advantage of the use of CFD (Computational Fluids
Dynamics) and experimental tests in wind tunnels. The
different configurations that have developed over time, have
looked for an improvement in the performance of the
aerodynamic behavior of the Savonius rotor. That is why the
geometric configuration is fundamental in terms of its
performance, the configuration that has been said for a
Savonius rotor is determined by the following design
parameters of the rotor:
• Rotor aspect ratio.
• Area swept by the rotor blades. Fig. 2. Schematization of the geometric parameters and the overlap ratio for a
• Geometric shape of the blades (profile, length of rope). two-bladed rotor. [8]
• Relationship of overlap and spacing between the blades.
The aspect ratio is a dimensionless parameter which involves
• Number of blades and stages.
the height of the rotor H and the diameter of the circle that is
• Other additional components to the rotor structure (axis,
formed when rotating the tips of the blades called diameter of
sail-flag effect, support structure, etc.).
the rotor D. Figure 3 shows the geometrical values of the rotor
The geometric shape of the blades is going to have a
where the aspect ratio can be seen.
semicircular profile which will be obtained from a PVC tube of
6 inches in diameter, keeping the traditional shape of the
Savonius rotor blades. Figure 1 shows the profile of the blade
with its dimensions.
Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira 399

structural parameter that conditions the performance of the

Savonius rotor, as stated above. Rotors of two blades have a
greater variation in the torque induced in the axis with respect
to those of three or more blades; this variation is presented,
since for a two-bladed rotor the angle of separation between
them is 180 °, in this way every 180 ° a maximum torque is
presented because the force received by the blades of the turbine
is maximum at that point. This can be observed in the curve of
torque against angle of attack presented by Hadi [10], which is
shown in Fig. 5.

Fig. 3. Rotor aspect ratio. [9]

The equation that relates the height of the rotor H with the
diameter of the rotor D, is given by (2) [9]:

𝐻
𝑟𝑜𝑡𝑜𝑟 𝑎𝑠𝑝𝑒𝑐𝑡 𝑟𝑎𝑡𝑖𝑜 = 𝛼𝑅𝐴 = (2)
𝐷

The swept area of a rotor is generally defined as the projected

area that is formed during the rotation of the blades. In the case
of a Savonius-type rotor, the projected area corresponds to the Fig. 5. Variation of the static torque against the angle of attack of a Savonius
turbine with two blades [10].
area of a rectangle. This can be seen in Figure 4 which
schematically represents the projection of this area. As regards a rotor with three or more blades, the variation of
the torque with respect to the angle of attack of the air flow
tends to be more constant, also allowing the handling of high-
speed ratios at the tips. In the particular case of the design of the
rotor that is going to be carried out, a two-bladed rotor was
selected, since, although the torque presents a greater variation
as a function of the angle of attack, the power coefficient curve
in function of the speed ratio at the tips presents higher values
of power coefficient. The following curve, Fig. 6, shows the
comparison of the curves of the power coefficient as a function
Fig. 4. Representation of the swept area of a Savonius rotor. [6] of the wind speed.

Thus the calculation of the swept area by the Savonius rotor

will be given by (3) [10]:

𝑠𝑤𝑒𝑝𝑡 𝑎𝑟𝑒𝑎 = 𝐴𝑠 = 𝐻𝑟 · 𝐷 (3)

The dimensioning of the final plates of the rotor is carried out

based on the criteria proposed by Akwa et al. [6], which states
that the thickness of the final plates must be minimum which is
relative to the height of the rotor. It also recommends a value of
1.1 times the value of the diameter of the rotor 𝐷 for the
calculation of the diameter of the final plates 𝐷𝑒𝑝 . This is how
the equation for the calculation of the diameter of the final
plates is given by (4) [6]:

𝐷𝑒𝑝 = 1,1 · 𝐷 (4)

The value of the final plate thickness is influenced by the Fig. 6. Power curves against wind speed of a rotor of 2, 3 and 4 blades [9].
nature of the material and the moment of inertia of the rotor.
The Savonius rotor stages represent the superimposition of
Torres, Daniela M. [11] made a decision matrix to select a
simple Savonius rotors where they are out of phase at a certain
material of the blades and end plates in which she selected as
angle [12], which is observed in Fig. 7.
best option the 16 gauge galvanized sheet that corresponds to a
thickness of 1.52 mm. The number of blades is a fundamental
400 Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira

representing the behavior of a particular type of a wind turbine,

which is applicable for the VAWT (Vertical Axis Wind
Turbine) and HAWT (Horizontal Axis Wind Turbine), the TSR
is referred to (5).

𝛺·𝑅
𝑇𝑆𝑅 = 𝜆 = (5)
𝑈∞

Where Ω is the angular velocity of the rotor, R is the distance

from the center of rotation of the rotor to the tip of the blade.
Fig. 7. Savonius rotors with different stages [12].
The Reynolds number is a dimensionless parameter used in the
For the present design a two-stage rotor is considered, since study of fluid flow through turbomachinery and flow systems,
multiple stages are considered mainly for two-bladed rotors, in which relates inertial forces to viscous forces [12]. For a
addition to the advantages of the reduction in moment Savonius type wind turbine the Reynolds number is defined in
fluctuations. equation (6).

𝜌·𝑈∞ ·𝐷
Fig. 8 shows the power coefficient as a function of wind 𝑅𝑒 = (6)
𝜇
speed for two-stage rotors with 2, 3 and 4 blades. Where ρ is the density of air, 𝑈∞ is the velocity of the free
wind current, D is the diameter of the rotor and μ is the dynamic
An initial parameter to carry out the design of the turbine is viscosity.
the speed of the free wind current. That is why the location of The Reynolds number plays a very important role in the wind
the turbine is important. According to the Atlas of the Wind tunnel experimentation of wind turbines when evaluating their
[13], the average wind speed for the municipalities of Guática aerodynamic performance, authors such as Niaz [15] conducted
and Quinchía is approximately 8 m/s. These sites are suitable a study regarding the influence of the Reynolds number on the
places within the department of Risaralda for the location of a aerodynamic performance of a three-blade Savonius turbine
Savonius wind turbine, given its proximity and which can with different overlap ratios in which it is expressed that: for
become test sites for this type of wind turbines. The average high Reynolds numbers the modeled turbine without
annual wind speed at the Technological University of Pereira overlapping radius generates better aerodynamic coefficients,
according to the meteorological unit is 0.77 m/s and according on the other hand for low Reynolds numbers the model with a
to data collected by Torres D [11] the average annual speed of radius of moderate overlap generates better results.
the meteorological unit of the Mundo Nuevo neighborhood is
3.81 m/s for the year 2014. After having made the analyzes of
the sites mentioned above, a wind speed of 𝑈∞ = 8 m/s is set III. DESIGN OF THE PHYSICAL COMPONENTS OF THE
as the design parameter value, so that the turbine may be SAVONIUS TYPE WIND TURBINE
suitable for experimental testing in any of the afore mentioned
sites.
The flow of air through the rotor induces forces on the blades
and the different elements that make up the support structure.
The determination of the loads due to air flow is based on
studying the behavior of the flow and applying the different
theories that are derived in the analysis of momentum. This is
how, the analysis of the different forces that act on a turbine is
convenient to treat separately by doing an aerodynamic analysis
and then consider the forces that are transmitted between the
different elements.

According to the established design parameters, the

component parts by which this device will be composed are
listed:
• 4 semicircular blades.
• 3 Final dishes.
• Turbine rotor shaft
Fig. 8. Power coefficient curves versus wind speed for a 2-stage rotor with 2, 3
• Support structure of the rotor.
and 4 blades [9]. • 2 roller bearings.

The Tip Speed Ratio TSR (λ) [14] is a dimensionless The following is a brief description of the equations used in
parameter that relates the tangential speed of the tips of the the design of the physical components of the turbine:
blade with the speed of the free wind current, in addition to
Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira 401

DESIGN OF THE ROTOR: The tree design includes the 𝐾𝑓𝑚 : Fatigue concentration factor at medium stress.
following stages according to Vanegas, L [16]: 𝐾𝑓𝑓 : Fatigue stress concentration factor for finite life.
• Material selection. 𝐾𝑓 : Fatigue stress concentration factor for infinite life.
• Constructive design. 𝑁: Security factor.
• Verification of resistance: Static, to fatigue and dynamic 𝜎𝑒 : Von Mises equivalent normal stress.
loads. 𝜏𝑒 : Von Mises equivalent shear stress.
• Verification of the rigidity: deflection by bending and slope
𝑆𝑛 : Resistance to corrected fatigue.
of the elastic and deformation by torsion.
• Modal analysis.
It is important to clarify that several checks of the resistance
of a shaft must be made such as resistance to vibrations, rigidity,
Selection of shaft material. The most commonly used material
modal analysis, etc., as previously established.
in the manufacture of shafts is steel with low or medium carbon
content. This material is recommended for the resistance of
Selection of wind turbine bearings: Based on the diagrams of
dynamic loads since it has fatigue limit and makes it suitable to
shear force and bending moment, it can be determined the radial
support many load cycles. If additional properties such as
loads that the supports of the shaft must bear, that´s to say the
corrosion resistance are required, a steel whose properties allow
bearings. In addition, the bearing located in the lower part must
resistance to the required characteristics must be selected.
support the axial load that corresponds to the weight of the
rotor. Therefore, the values of the reactions in the supports of
Calculation of rotor’s weight. The determination of rotor’s
the tree are:
weight is made from the sum of the individual weights of each
of the component elements.
𝐹𝑟𝐴 ≅ 15 𝑁
The blades as defined in the preliminary design part, will be
𝐹𝑟𝐷 ≅ 15𝑁
made of PVC material, the final plates that make up the rotor
𝐹𝑎𝐷 = 116 𝑁
from AISI / SAE 1020 Cold Rolled steel sheet and the panels
that are part of the rotor of acrylic material. The total weight of
Determination of the equivalent radial dynamic load: Because
the rotor is determined by (7).
of the bearings supports A and D, support radial loads and axial
𝐹𝑔,𝑝𝑙𝑎𝑡𝑜 𝑓𝑖𝑛𝑎𝑙 loads, it is necessary to convert these load values into a single
𝐹𝑔𝑇 = 4 · (𝜌𝑃𝑉𝐶 · 𝑔 · ∀𝑎𝑠𝑝𝑎,𝑃𝑉𝐶 ) + 3 · ( · radial load value that would have an effect on the life of the
𝑚2
𝐴𝑝𝑙𝑎𝑡𝑜 𝑓𝑖𝑛𝑎𝑙 ) + 4 · (𝜌𝑎𝑐𝑟í𝑙𝑖𝑐𝑜 · 𝑔 · ∀𝑎𝑐𝑟í𝑙𝑖𝑐𝑜 ) + ∀á𝑟𝑏𝑜𝑙 · (7) equivalent bearing to the actual load applied, this load is known
as "equivalent radial dynamic load 𝑃𝑟 " [17], which is
𝜌𝑎𝑐𝑒𝑟𝑜 · 𝑔 + 𝐹𝑔,𝑜𝑡𝑟𝑜𝑠 determined by (9).

Calculation of the diameter of the rotor shaft. The 𝑃𝑟 = 𝑋𝐹𝑟 + 𝑌𝐹𝑎 (9)
calculation of the diameter of the rotor shaft will be made based
on the theory of shaft design recommended by Vanegas, L [16]. 𝐹𝑟 : radial actual load [N], 𝐹𝑎 : axial actual load [N], 𝑋: radial
Since a shaft performs a rotating movement, the loads it must load factor, 𝑌: axial load factor.
support are variable over time. That is why the design equations
of the Fatigue Theory are applied. In shaft fatigue analysis there Blade resistance evaluation: As indicated in the preliminary
are three methods which allow to calculate the shaft diameter design, the blades are manufactured from a PVC pipe of 6 in.,
required to adequately resist dynamic loads. These methods are Based on these dimensions and material the blade's resistance
the following: must be evaluated, so for its analysis it is take the blade that is
in the lower stage being this the most critical to support drag
 Von Mises method. forces and weight loads. According to the loading arrangement
 Method adopted by Faires. and finding the most critical section, verification of resistance
 ASME method. to tearing and crushing was carried out:

The Von Mises method [16] is applied for this case after the  Calculation of tear stress [16]: It is determined with (10):
different aerodynamic loads in the rotor have been determined,
from which the required diameter of the shaft can be obtained 𝑆𝑠,𝑡𝑒𝑎𝑟 =
𝐹
(10)
to support dynamic loads over time, this is shown below in (8). 2·𝑏1 ·𝑙1

𝑏1 : Distance from the edge of the blade to the periphery of the

√𝜎𝑒(𝑚) 2 +3𝜏𝑒(𝑚) 2 √𝜎𝑒(𝑎)2 +3𝜏𝑒(𝑎) 2
1
= + (8) hole [m], it is obtained of the planes, 𝑙1 : Thickness of the
𝑁 𝑆𝑦 𝑆𝑛 material to be evaluated [m].
Where the stresses 𝜏 𝑦 𝜎 can be calculated like this:
𝜏𝑒(𝑚) = 𝐾𝑓𝑚(𝑇) 𝑆𝑠(𝑚)  Crushing resistance [16]: Crushing resistance is
𝜏𝑒(𝑎) = 𝐾𝑓𝑓(𝑇) 𝑆𝑠(𝑎) determined with (11):
𝜎𝑒(𝑚) = 𝐾𝑓𝑚(𝐹) 𝑆(𝐹)(𝑚) + 𝐾𝑓𝑚(𝑀) 𝑆(𝑀)(𝑚)
𝐹
𝜎𝑒(𝑎) = 𝐾𝑓𝑓(𝐹) 𝑆(𝐹)(𝑎) + 𝐾𝑓𝑓(𝑀) 𝑆(𝑀)(𝑎) 𝑆𝑐𝑟𝑢𝑠ℎ𝑖𝑛𝑔 = (11)
𝑙·𝑑𝑝
402 Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira

Design of the rotor bolts [16]: The fastening elements that are
used to join the blades with the final plates and the shaft with
the blades are bolts subjected to static loads, so design theories
for combined static loads are applied. The Octahedral shear
stress Theory is applied to one of the points subjected to normal
stress and the other subjected to shear stress by means of the
equation 12.
2 2
𝑆 𝑆𝑠 1 2
( ) +( ) =( ) (12)
𝑆𝑦 𝑆𝑠𝑦 𝑁

DESIGN OF THE STRUCTURE: The structure of the wind

turbine will be composed of two trusses and six beams of which
four are loaded beams and these supports the rotor, the other
two beams are unloaded beams as they provide stability and
rigidity to the structure. This must provide resistance to the
loads to which the turbine will be subjected (wind loads, weight Fig 10. Free body diagram of the truss.
loads of the same structure and the rotor), rigidity, low
deflection in the elements and minimum turbulence of the air
flow that enters the rotor. Since the structure will be formed by
two lateral trusses joined by beams, all its component elements
must be designed, so the design of the structure is focused on
the analysis of the trusses, analysis of the beams and later on
the design of the Connection elements.

Loads that the structure supports. The main element of the

wind turbine is the rotor to be this one, the component that
supports different types of loads that in turn transmit them to
the structure, on the other hand, the structure itself generates
loads (own weight of this), that is why all possible loads must
be considered. Initially the own weight of the structure is
determined.
Fig 11. Structural analysis of the truss using the computer tool Autodesk
Truss design. The design of the truss consists of the Simulation Mechanical
dimensioning of its structural profiles evaluating their behavior
to resistance, stability and deflection. The geometric shape of The analysis of the truss is then carried out by means of the
the truss is composed of structural profiles interconnected in computer program Autodesk Simulation Mechanical, which
nodes by means of brackets, which is shown in Fig. 9. provides the numerical values of the reactions in the supports,
the internal loads to which each element is subjected (structural
profile) and the way in which the truss may become deformed.
Fig. 11 shows the analysis of the truss, which shows the axial
internal loads of each element, likewise they are numbered and
indicate the elements and nodes by means of numbers.

Design of the elements or structural profiles of the truss.

Based on the tensile or compression forces that support the
elements of the fence, we proceed to the determination of the
stresses caused by internal loads in each element, considering
that each element that forms the structure supports axial loads
that generate traction and compression normal stresses [16].
Since the structure does not support live loads, its size is small
compared to roof structures, and since most of the loads are
dead, it is designed from the normal stress equation given by
Fig 9. Profile of the truss with nodes named according to the given
(13).
𝐹
nomenclature in Autodesk Simulation Mechanical (measurements in mm). 𝑆𝐸 = ± 𝐴,𝐼 (13)
𝐴𝐸

A free-body diagram of the structure is shown in Fig. 10.

𝑆𝐸 : Normal stress that support the element [Pa], 𝐹𝐴,𝐼 : internal
axial load in the element [N], 𝐴𝐸 : minimum area of the
transversal section of the element [m2].
Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira 403

The design equation of the elements subjected to simple static

loads is given by the equation 14:

𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑟𝑒𝑠𝑖𝑠𝑡𝑒𝑛𝑐𝑒 𝑆𝑢 𝑜 𝑆𝑦
𝑆 ≤ 𝑆𝑑 = (14)
𝑁

Where S is the maximum stress to which the element is

subjected. This equation can be used with the yield stress 𝑆𝑦 or
with the ultimate stress 𝑆𝑢 of the material, in this case the creep
resistance 𝑆𝑦 is considered in equation 14 in order to avoid
failure due to plastic deformation.

Design of the loaded beam. The design of the loaded beam is

made based on the design equation presented above, where the
loads supported by the beam produce normal bending stresses. Fig. 12. Three-dimensional view of the general assembly of the wind turbine.
Observing the bending moment diagrams, the critical point is
in the middle of the beam where the bearing is supported. Simulation of air flow through the rotor of the Savonius
Calculating the normal stress, it has: wind turbine: Initially a dynamic simulation of the wind
turbine rotor is executed, for this it is required to have the solid
𝑀𝐹 ∗𝑐 modeling of the rotor and once this is obtained, a volume must
𝑆= (15)
𝐼
be defined of control which encloses the rotor, and it is there
where certain border conditions are established as the different
The design of the connection elements of the structure is not
thermodynamic properties (Pressure, Temperature, Enthalpy,
presented, but all its dimensioning is also important for the
etc.). Another region of analysis is the rotating region in which
construction part.
the surface objectives are defined, understood as
thermodynamic properties or dynamic variables that are desired
Modeling the wind turbine by using the solidworks
to be ascertained during the development of the analysis. The
computational tool: the entire design process of the physical
ambient pressure and the average temperature of the place were
components of the wind turbine was presented based on the
respectively defined as 86 kPa and 21 ° C, in addition the wind
application of the different design theories for the different load
speed for the analysis is handled with 8 m / s, the same with
conditions of the component elements. Thanks to the design
which the design process was carried out.
process, the necessary dimensions of the component elements
The pressure gradients as shown in Fig. 13 in the different
were determined to allow an adequate mechanical resistance to
concave and convex parts of the rotor generate resultant drag
the loads, minimum deformations and stability in the support
forces in each portion of the rotor which acting on a distance
structure, as well as the selection of the relevant materials of
produce a net torque that in turn allows the rotation of the rotor.
each element, connection forms or union, functionality,
To obtain a better appreciation of the pressure distribution in
aesthetics and manufacturing costs. The dimensions thus
the rotor and a closer approximation, an external flow analysis
determined, and the shape of the elements were modeled by
with the static rotor is performed.
using the Solidworks computational tool. The modeling process
consists of obtaining 3D of each component element giving it
its corresponding shape and the required dimensions.
Subsequently, the assembly process is passed through the tools
provided by the program, such as positional relationships. The
general assembly of the wind turbine is presented in Fig. 12.

Fig 13. Pressure distribution in the rotor in external flow.

In the previous figure it can be seen that the highest pressure

(yellow lines) is given on the concave part of the rotor blade.
The distribution of pressures can also be seen in the floor plan
shown in Fig. 14.
404 Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira

Fig 16. Distribution of speeds in the rotor plane for the position φ = 90 ° lower
stage.
Fig 14. Pressure distribution in the rotor plane.
TABLE I
Analyzing the distribution of pressures, it is observed that the DYNAMIC VARIABLES AND THERMODYNAMIC PROPERTIES
DETERMINED BY DYNAMIC FLUID ANALYSIS FOR THE ROTOR
highest pressure is produced on the concave side of the blade WITHOUT PANELS.
and in the convex part a lower pressure is distributed. This Dynamic variable
differential of pressure and consequently a difference in the and Minimum Maximum Mean S.I
drag forces on the blades causes the torque generated by the thermodynamic value value value Unity
drag force on the concave side to be greater compared to the property
torque generated by the drag force on the convex side, allowing Total pressure 86036.76 86036.77 86036.77 Pa
that the rotation of the rotor is given. Also, for this case, the X- velocity 0.004 0.005 0.004 m/s
distribution of speeds in the floor plan for the static case of the y- velocity 0.018 0.019 0.018 m/s
rotor is presented in the following figures.
z- velocity 7.483 7.484 7.483 m/s
In the Fig. 15 and 16, it is observed that the geometric Normal force in Z 7.4 7.421 7.413 N
configuration of the Savonius rotor (overlap) contributes to the Density 1.02 1.02 1.02 kg/m^3
air circulation from one blade to the other one, thus increasing Static torque (axis
its induced torque, since part of the circulating air increases the 0.395 0.396 0.396 Nm
y)
drag force in each concave part. Through the analysis, other Dynamic torque
0.233 0.272 0.242 Nm
dynamic variables and thermodynamic properties were (axis y)
determined, such as the forces on the blades, the torque, and the
density of the air. The following table provides the value of the According to the results obtained presented in the table I, it
dynamic variables and thermodynamic properties. can be expressed that the air velocity in the rotor presents only
one predominant component along the coordinate axis z (wind
incidence) and in the rest the values are null, meaning this there
is a one-dimensional flow. Due to this, the force exerted by the
air flow on the rotor has a component along the coordinate axis
z, in the direction of air flow, the other force components are
zero. As for the torque, it presents a value of 𝑀𝑌 = 0,396 𝑁𝑚
in the negative direction of the coordinate axis.

V. PHYSICAL ASSEMBLY OF THE SAVONIUS WIND

TURBINE
The process of cutting with frame-saw was used to give the
required dimensions to the different structural profiles. These
dimensions were obtained from the design process, and were
reflected in the plan of the structure, where said measures were
Fig 15. Speed distribution in the rotor plane for φ = 0 ° upper stage.
guaranteed with the use of the meter and squares. In the same
way, this tool was used to cut the angles of the structure, the
rotor and the PVC tube to obtain the blades. The mechanical
shear was used throughout the process of cutting the gills to
give the required dimensions. On the other hand, the circular
shape of the final plates was obtained thanks to the oxyfuel
process. Table II presents the process of cutting metals by
oxyfuel.
Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira 405

TABLA II
MATERIALS AND MECHANICAL ELEMENTS USED IN THE
ASSEMBLY OF THE WIND TURBINE.

Type of Dimensions and Obtention

Material
element specifications process

Tubular
structural Cold rolled
26 mm x 26 mm x Steel AISI
profile of (cold
6 m length SAE 1020
square lamination)
section
Cold rolled
Steel AISI
Steel plate 3 mm thick (cold
SAE 1020
lamination)
38 mm (equal
Angle
wings) x 1 m Fig 17. Assembly of the wind turbine.
connector ASTM A36 lamination
length, thickness 2
structure
mm
Angle
VI. EXPERIMENTAL TESTS OF THE SAVONIUS WIND
20 mm (equal TURBINE
connector
wings) x 1.5 mm ASTM A36 lamination
blade end
thick Experimental tests were performed on the wind turbine
plate
without panels for which measurements were made of wind
Union bolts Nominal Diameter
of the 1/4", number of
Steel AISI Industrial speed, angular speed of the rotor, and static torque
SAE 1020 process corresponding to different wind speeds. The tests were carried
structure threads per inch 20
Steel AISI Cold out in the wind tunnel of the Laboratory of Fluids and Hydraulic
Rotor shaft Diameter 5/8 "
SAE 1020 lamination Machines of the Technological University of Pereira with the
Steel AISI Cold collaboration of Eng. Yamal Mustafá Iza M.Sc., which through
Final dishes 14 gauge
SAE 1020 lamination
the use of a frequency inverter can be varied its angular
6 "in diameter x 1 Industrial velocity, which means having a variable wind speed. To carry
blades PVC
m in length. process
out the measurement of the experimental data during the tests,
Panels of 490mm x the following measurement devices were used:
Industrial
Panel 300mm x 2.5mm Acrylic
process
thickness
Industrial  Anemometer model AM-4206, resolution for wind
bearing UCF 201-8 steel
process speed measurement of 0.01 m / s and resolution for
Industrial temperature measurement of 0.1 ℃. The wind speed
hinge Ref. 3 IMDUMA steel
process
data is measured at 70 cm from the rotor.
 LT Lutron DT 2236 tachometer, resolution 0.1 rev /
The main shaft of the rotor required certain dimensions
min.
according to its design process. The tree was then submitted to
the turning process with its initial dimensions to bring it to said
The obtained angular velocity curve is shown as a function
dimensions. During the mechanization of the tree, it became
of the wind speed in Fig. 18.
necessary to use a bezel to support the tree due to its length of
1.5 m. The use of the calibrator and micrometer as measuring
elements was essential to guarantee the required dimensions of 500
said mechanical element. 450
400
The drilling operations enabled the drilling of the holes in Angular
350
structural elements, brackets, angles, blades, tree and end velocity of
300
the shaft of
plates. These holes will allow through them pass the union bolts the wind 250
of the structure and the rotor, where the adjustment between turbine 200
150
bolt and hole is a fit with clearance. It became necessary to [min-1]
100
make different assemblies to carry out all the various drilling 50
operations required. The assembly that was carried out for the 0
drilling process in the laboratory of Machines 2 4 6 8 10
Air velocity [m/s]
After carrying out all the manufacturing operations required, Fig 18. Angular velocity as a function of wind speed.
the final assembly is carried out, which can be seen in Fig. 17.
The curve obtained from torque as a function of the wind
speed is shown in Figure 19. For this case the torque is
measured statically, with the difference that in this case the
wind affects the rotor to different speeds, achieving balance of
the rotor with the addition of different counterweights.
406 Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira

mechanical power data as a function of the wind speed, in which

a cubic tendency is observed. This is what is expected since
0.45
theoretically the mechanical power of a wind turbine varies
0.40
0.35
with the cube of the wind speed.
0.30
0.25
Torque
[Nm] 0.20
VIII. CONCLUSIONS
0.15
0.10 • The parameters of the wind energy were characterized,
0.05 especially the variability of the wind speed in the whole
0.00 Colombian national territory.
3 5 7 9 11 13
Air velocity [m/s] • The data taken from the meteorological unit of the ceiling
Fig 19. Torque as a function of the wind speed for the rotor. of the Faculty of Mechanical Engineering showed that the
average wind speed is 1 m / s for this reason this site is not
Fig. 20 shows the mechanical power as a function of the wind suitable for this type of wind turbine. Places in the coastal areas
speed, for this case the mechanical power was determined of Colombia represent a good wind potential for the use of this
theoretically from the experimental data of static torque and type of energy.
angular speed of the rotor corresponding to table 8.3. The
determination of the mechanical power is determined by (16) • The design of the wind turbine was based on the scientific
: articles investigated, theories studied and the experiences of
𝑃𝑟 = 𝑀𝑇 · 𝜔 (16) other authors.

12.0
• Applied design theories allowed to select the appropriate
materials and the required dimensions, to obtain a physical
10.0 model that guarantees the operating conditions based on which
mechanical 8.0 it was designed.
power
6.0
[W] • The modeling of the wind turbine was developed using the
4.0 Solidworks computer tool, based on the sizing obtained through
2.0 the design process.
0.0
• The analysis of the behavior of the flow allowed to know
2.0 4.0 6.0 8.0 10.0
the distribution of pressures in the blades, the effect of overlap,
Air velocity [m/s]
the velocity field on the rotor and the torque generated by the
Fig 20. Mechanical power as a function of wind speed. drag forces.

• The different manufacturing processes required were

carried out in order to carry out the physical assembly of the
VII. ANALYSIS OF EXPERIMENTAL RESULTS Savonius wind turbine.

According to the tabulated experimental data, the following • Experimental tests of the turbine with the axial fan of the wind
can be expressed: tunnel were developed in the laboratory of Fluids and Hydraulic
Machines of the Faculty of Mechanical Engineering of the
• It can be observed in the curve of figure 22 that there is a Technological University of Pereira.
linear relationship between the wind speed and the angular
speed of the rotor, thus confirming the TSR (Tip Speed Ratio).
On the other hand, the angular velocity of the rotor without ACKNOWLEDGMENT
panels is much higher compared to that obtained from the rotor
with panels, because the panels increase the inertia of the We thank and acknowledge the great work that each of the
system and also restrict movement. teachers, administrators, workers and students of the
Technological University of Pereira have done for providing us
• The behavior of the data of the curve of figure 23 of torque with their knowledge, teaching, time, service, collaboration and
as a function of the wind speed have a functional relationship friendship during the stage of academic formation and
of order 2. Professional as Mechanical Engineer.

• The graph in figure 24 shows the behavior of the

Scientia et Technica Año XXIV, Vol. 24, No. 03, septiembre de 2019. Universidad Tecnológica de Pereira 407

REFERENCES http://www.ntnamericas.com/es/website/documents/brochures-and-
literature/catalogs/bearing-units%202400-IX.pdf.
[1] N. Vaughn. “Introduction to renewable energy” United States: Boca
Raton CRC: Press Taylor & Francis group, 2011, 350p. ISBN 13:
978-1-4398-9120-9. [Online]. Available:
https://www.loot.co.za/product/vaughn-c-nelson-introduction-to-
Andrés Felipe Rodriguez Valencia was
renewable-energy/pvnj-2740-ga50. born on September 14, 1992 in La Victoria
[2] R. Hugh. “Wind power, The wind generation,” [Online]. Available: Valle (Colombia). He completed his
http://web.ing.puc.cl/~power/paperspdf/CapituloEolico.pdf. undergraduate studies in Mechanical
[3] E. H Lysen, “Introduction to wind energy,” Wind energy. Engineering (2010-2016) at the
Consultancy services wind energy developing countries. 2nd ed.
Netherlands: 1983, 311p.
Technological University of Pereira (UTP).
[4] D. B. Sanz, I. Cruz. “Guide on small wind technology,” Arias During the realization of the career, he was
Montano Graphics. Madrid: 2012. 118p. [Online]. Available: academic monitor of the department of
https://www.fenercom.com/publicacion/guia-sobre-tecnologia- mathematics and monitor of the laboratory of fluids and
minieolica-2012/ hydraulic machines. At a professional level, he worked in the
[5] Aracuan. Region of turbulence around a building [online].
maintenance area at the company Trapiche Biobando S.A.S,
Available: http://www.aracuan.com.ar/turbulencia.htlm.
[6] J. V. Akwa, H. A. Vielmo, and A. P. Petry. “Review on the where the company's assembly activities were supported, in the
performance of Savonius wind turbines,” ScienceDirect, Renewable area of extraction and processing. He worked at Accedo
and sustainable energy reviews. Pp. 3054-3064, June 2012, 16 Colombia S.A.S in the commercial sales part of the company.
(2012) 3054– 3064, [Online]. DOI: 10.1016/j.rser.2012.02.056 In 2017, he started as Professor of the Faculty of Mechanical
[7] J. L. Menet, N. Bourabaa. “Increase in the Savonius rotors Engineering, has directed the courses of Static and Hydraulic
efficiency via parametric investigation,” Ecole Nationale Superieure
d'ingenieurs en Informatique Automatique Mecanique Énergetique Machines and Fluids Laboratory. With the Department of
Électronique de Valenciennes ENSIAME, [online].Available: Mathematics, he has guided the Linear Algebra courses. He is
http://educypedia.karadimov.info/library/23_1400_jeanlucmenet_0 candidate for the master’s degree in Mechanical Engineering in
1.pdf. the line of manufacturing and design processes of the same
[8] F. Wenehenubun, A. Saputra, and H. Sutanto. “An
experimental study on the performance of Savonius wind turbines university, where he graduated from, since 2017.
related with the number of blades,” ScienceDirect, 2nd International ORCID: https://orcid.org/0000-0002-0540-8019
Conference on Sustainable Energy Engineering and Application,
ICSEEA 2014, 68 (2015) 297 – 304, [online]. DOI:
10.1016/j.egypro.2015.03.259. Edgar Alonso Salazar Marín was born on
[9] A. A. Kadam, S. S. Patil. “A review study on Savonius wind rotors June 6, 1973 in Santa Rosa de Cabal.
for accessing the power performance,” IOSR Journal of Mechanical
and Civil Engineering, pp. 18-24, Second National Conference on
(Risaralda / Colombia). He´s Mechanical
Recent Developments in Mechanical Engineering 18 | Page Engineer and did doctored studies in
M.E.Society's College of Engineering, Pun [Online]. Available: Polytechnical University of Catalonia- Spain
https://www.researchgate.net/publication/281899860_A_Review_ (2003-2007), He has master´s studies in
Study_on_Savonius_Wind_Rotors_for_Accessing_the_Power_Per mechanical engineering of Andes University-
formance.
[10] M. H. Ali. “Experimental comparison study for Savonius wind
Bogotá (1997-1998) and Master in Automatic Systems of
turbine of two and three blades at low wind speed,” International Production – UTP (2002-2004). He worked as researcher in
Journal of Modern Engineering Research IJMER, Vol. 3, Issue. 5, Coffee Research National Center (Cenicafé) from 1999 to 2000.
pp. 2978-2986, Oct. 2013, [Online]. Available: He has worked as professor in Technology Faculty on
http://www.ijmer.com/papers/Vol3_Issue5/CJ3529782986.pdf Technological University of Pereira since 2000 and he´s
[11] D. Torres. “Design of a low power vertical shaft wind generator.
developed different projects about renewable energies, like
Degree work, Faculty of Mechanical Engineering, Technological
University of Pereira, Pereira, Risaralda, 2015 [online]. Available: Solar systems (thermal and photovoltaic), solar vehicles and
repositorio.utp.edu.co/dspace/handle/11059/5145. conversion from combustion vehicles to electric vehicles.
[12] J. A. Menet. “Double step Savonius rotor for local production of ORCID: https://orcid.org/0000-0003-2873-245X
electricity: a design study,” ScienceDirect, Feb. 2004, [online].
DOI: 10.1016/j.renene.2004.02.011
[13] Colombia. energy mining planning unit EMPU. Institute of
hydrology, meteorology and environmental studies IDEAM. Atlas
of Wind and Wind Energy of Colombia, 2006 [online]. Available:
http://www.upme.gov.co/Atlas_Viento.htm.
[14] J. F. Manwell, J. G. McGowan. “Wind energy explained. Theory,
design and application,” Global energy concepts. 2nd Ed. USA:
Wiley, 2010, pp. 1-100. [online]. Available:
http://ee.tlu.edu.vn/Portals/0/2018/NLG/Sach_Tieng_Anh.pdf.
[15] K. N. Morshed, M. Rahman and G. Molina. “Wind tunnel testing
and numerical simulation on aerodynamic performance of a three-
bladed Savonius wind turbine,” International Journal of Energy and
Environmental Engineering, 2013 [online]. Available:
https://link.springer.com/article/10.1186/2251-6832-4-18.
[16] L. V. Vanegas. “Design of machine elements,” Mechanical Design,
vol. 1, 1ra Ed, Pereira, Risaralda, Colombia. UTP, 2016. [online].
Available: http://blog.utp.edu.co/lvanegas/dis1/.
[17] NTN. Bearings catalog [online]. Available: