Vertical Axis Wind Turbine with

Continuous Blade Angle Adjustment

by
Samuel Bruce Weiss
Submitted to the Department of Mechanical Engineering
in Partial Fulllment of the Requirements for the Degree of
Bachelor of Science
at the
Massachusetts Institute of Technology
June 2010
c 2010 Massachusetts Institute of Technology
All rights reserved
Signature of Author ................................................................................................................
Department of Mechanical Engineering
May 7, 1993
Certied by .............................................................................................................................
Sanjay E. Sarma
Associate Professor of Mechanical Engineering
Thesis Supervisor
Accepted by ............................................................................................................................
John H. Lienhard V
Collins Professor of Mechanical Engineering
Chairman, Undergraduate Thesis Committee
Vertical Axis Wind Turbine with
Continuous Blade Angle Adjustment
by
Samuel Bruce Weiss
Submitted to the Department of Mechanical Engineering
on May 10, 2010 in partial fulllment of the
Requirements for the Degree of Bachelor of Science in
Mechanical Engineering
ABSTRACT
The author presents a concept for a vertical axis wind turbine that utilizes each blades
entire rotational cycle for power generation. Each blade has its own vertical axis of rotation
and is constrained to rotate at the rate of one half of a revolution per full revolution of
the rotor. For a rotor of radius r and blades of width b, a technical analysis predicts a
theoretical maximum power coecient of C
P
=
b
2r+b
, neglecting wind ow interference by
upwind blades. This theoretical power coecient is generally greater than the eciency of
a typical Savonius wind turbine (C
P
0.15), and it reaches C
P
= 0.5 at the limiting blade
width, b = 2r. The analysis also predicts a static torque and optimal tip-speed ratio that
are both greater than those of a Savonius wind turbine with similar blade dimensions.
Design considerations for implementing the kinematic constraint and for blade adjustment to
account for changes in wind direction are discussed, and the authors prototype is presented.
Testing of the prototype demonstrated that implementation of the kinematic constraint
is feasible, and that eciencies greater than those achievable by a Savonius turbine are
plausible. In 4
m
s
wind conditions, the prototype yielded an estimated C
P
of 0.15, with much
room for improvement through design changes and blade optimization in future iterations
of this style of turbine.
Thesis Supervisor: Sanjay E. Sarma
Title: Associate Professor of Mechanical Engineering.
2
Acknowledgements
The author is grateful for the supervision, guidance, and support of Sanjay Sarma throughout
the development of this thesis project. He would also like to thank Bill Buckley, David Dow,
and Patrick McAtamney for their help in the construction of the turbine prototype.
3
Contents
1 Introduction 6
1.1 Types of Wind Turbines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 The Savonius Wind Turbine . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Literature Review 8
2.1 Swinging Blade Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Pitch Angle Adjusting Turbines . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 Similar Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 Concept 10
3.1 Blade Movement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2 Blade Angle Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.3 Advantages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4 Disadvantages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4 Theoretical Analysis 14
4.1 Lift and Drag Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2 Static Torque . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3 Dynamic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.4 Flow Interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5 Design and Prototype 20
5.1 Machine Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.2 Kinematic Constraint Implementation . . . . . . . . . . . . . . . . . . . . . 21
5.3 Blade Angle Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.4 Blade Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.5 Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
6 Conclusions 25
6.1 Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
References 26
4
List of Figures
1 Savonius Turbine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Swinging Blade Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 Blade movement for turbine with continuous blade rotation . . . . . . . . . . 11
4 Lift and drag forces acting on turbine blade . . . . . . . . . . . . . . . . . . 15
5 Blade movement in dynamic case . . . . . . . . . . . . . . . . . . . . . . . . 16
6 Torque on turbine blades during rotation . . . . . . . . . . . . . . . . . . . . 17
7 Torque on blade for dierent rotor-tip speed ratios, and Power-Speed curve . 18
8 Two-dimensional ow analysis in static case . . . . . . . . . . . . . . . . . . 20
9 Solid model views of turbine prototype . . . . . . . . . . . . . . . . . . . . . 22
10 Turbine prototype and qualitative testing . . . . . . . . . . . . . . . . . . . . 24
5
1 Introduction
Wind turbines and windmills have been in use for centuries for a variety of applications, and
there is a constant eort to modify and improve their design. This paper presents a concept,
analysis, and design of the authors idea for a drag-based vertical axis wind turbine that
utilizes each blades entire rotational cycle for power generation. This paper supplements
a mechanical prototype intended to demonstrate the kinematic constraints and mechanical
mechanisms required by this design.
1.1 Types of Wind Turbines
Most wind turbines can be classied according to two main attributes. The rst is the method
of power generation; turbines are called lift-based or drag-based depending on their primary
source of propulsion. The second is the orientation of the axis of rotation of the turbine,
which can be horizontal (almost always parallel rather than transverse to the wind ow)
or vertical (transverse to the wind ow). Each combination of axis orientation and power
generation method has its own advantages and disadvantages, which are briey discussed
here.
Lift-based wind turbines (especially horizontal turbines) are the most common turbine used
for large-scale power generation. They tend to have much-higher rotational speeds, which
reduces power loss when connecting a turbine to a generator. The relatively small surface
area of their narrow blades also gives lift-based turbines a fairly high power-to-weight ratio.
Drawbacks of lift-based turbines include the requirement of precise blade shaping and bal-
ancing, as well as (in the case of horizontal turbines) ineciencies at small scale. Drag-based
turbines, on the other hand, are often preferred for applications that require direct applica-
tion of mechanical work (e.g. milling corn, for a windmill) because of the high rotor torque,
or for fullling small-scale power requirements. [7]
Horizontal axis turbines are often easier and more ecient to scale (by extending a tall,
vertical tower) than their vertical counterparts. Horizontal axis turbines also experience
little variable interference of the wind ow between blades, whereas with a vertical axis
turbine there is direct and unavoidable interference. Vertical turbines, however, are often
more easy to install and repair since their machinery for power generation is usually located
6
at ground level, whereas horizontal axis turbines machinery is high o the ground at the
rotors center. [7]
Examples of these types of turbines include modern power-generating wind turbines (hori-
zontal, lift-based), 12th century European windmills (horizontal, drag-based), Darrieus wind
turbines (vertical, lift-based), and Savonius turbines (vertical, drag-based). A wealth of in-
formation is available for each of these types of turbines, but we will focus on discussing
some details of the vertical, drag-based, Savonius turbine as it relates most closely to the
wind turbine concept presented in this paper.
1.2 The Savonius Wind Turbine
wind ow

F
+
D
Figure 1: An S-shaped,
Savonius turbine. Positively
contributing drag forces
(F
+
D
) outweigh negatively
contributing drag forces (F

D
)
to create a net positive torque
about the rotor center.
Perhaps the most prevalent vertical axis wind turbine is the
S-shaped, Savonius wind turbine. The Savonius turbine con-
sists of cupped blades that catch the wind to generate power
(see Fig. 1). Power generation is possible because of blade
geometry the cupped blades have a higher drag coecient
when moving with the wind than when moving against the
wind. The dierence between the torque on the blades trav-
eling downwind and the blades traveling upwind gives the net
torque about the generator.
There are several important parameters that we will use to
characterize the ability of the Savonius turbine to generate
power (these will also be used to compare the new design to
the Savonius turbine). The rst is tip speed ratio, , which is
dened as the ratio between a blades speed at its tip and the
faraway wind velocity. For a turbine of radius R, spinning at
angular rate , in a wind with faraway velocity v
0
, we have
=
R
v
0
. (1)
We can also dene a turbines eciency, also known as its power coecient, C
P
, as the
ratio between the power produced by the turbine and the power contained in the wind that
passes through the reaches of the turbine blades. If we dene the swept area, A
S
as the
7
cross-sectional area of the wind ow that the turbine blades pass through, then a turbine
that produces power P has eciency
C
P
=
P
1
2
A
S
v
3
0
, (2)
where is the air density. A typical Savonius turbine achieves an eciency of C
P
= 0.12
to 0.15 at a tip speed ratio of about = 0.7, [6, 9] but an optimized Savonius turbine,
studied both theoretically and empirically by Modi and Fernando, can reach a peak eciency
of C
P
= 0.32 at a tip speed ratio of = 0.8. [6]
A turbines start-up wind velocity and static torque can also be important parameters. The
start-up wind speed is the lowest wind speed at which the turbine spins (if it is self-starting,
at all). This is typically on the order of a few meters per second, but depends heavily on
turbine design and construction. The static torque is the torque that the wind applies about
the rotor center when the rotor is xed. The static torque is often a function of rotor angle,
and a high static torque often indicates a low start-up wind speed.
2 Literature Review
Many other variations of drag-based vertical axis turbines have been designed, prototyped,
and implemented for power generation. Documented in this section is prior research on
the two types of turbines that provided the main inspiration behind the authors own de-
sign. Other vertical axis turbines that are similar in concept to the authors design are also
discussed.
2.1 Swinging Blade Design
Several turbine designs have sought to eliminate the negative drag that acts on the blades
traveling upwind in a Savonius turbine. Many of these employ slatted, or swinging blades
that are constrained to trap wind during their downwind travel, but swing freely (and thereby
reduce or eliminate drag) during their upwind travel (see Fig. 2).
8
wind ow
Figure 2: A modied depic-
tion of Reupke and Proberts
slatted blade design. Blade
segments are free to swing
to reduce drag during upwind
travel. [8]
Tabassum and Probert investigated a slatted design and found
that the design increased startup torque by as much as 35 per-
cent over a standard Savonius wind turbine, and conrmed that
the net static torque on the rotor was positive for every rotor
orientation. [10]
Further investigation into the slatted design by Reupke and
Probert conrmed that this style turbine was self-starting, and
that it had a slightly higher optimal tip speed ratio (approx-
imately 0.8) than a typical Savonius turbine (approximately
0.7). This slatted design yielded a slightly greater torque than
a Savonius turbine at low speeds, but the centrifugal forces of
the swinging blades severely hurt the turbines eciency at high
speeds. As a result, its peak power coecient only reached
0.05, compared to 0.18 for a similarly constructed Savonius
turbine. [8]
2.2 Pitch Angle Adjusting Turbines
Hwang et. al. discuss a modication of a lift-based vertical turbine that varies the pitch of
its blades slightly as the turbine rotates. By optimizing the pitch angle based on the blades
position, the lift force is increased, and power generation increased by 30 percent over a
turbine with xed pitch angles. [4]
2.3 Similar Concepts
The author has found that other individuals have attempted their own implementations of a
concept similar to his own. Unfortunately, the author did not identify these similar vertical
axis wind turbines until very late in the course of his research, largely due to their foreign
origin. Descriptions based on the authors understanding of the projects is included for
completeness, and it is likely that the authors own design could be improved upon through
feedback from these other attempts.
9
A similar turbine created by French inventor Pierre Dieudonne consists of two levels of blades
with a similar kinematic constraint, but with the two levels rotating in opposite directions.
His website and videos are in French, and so the authors understanding is almost entirely
from the images that Dieudonne provides. [2]
The author has also found limited information on a turbine recently produced by a Chinese
company, Jiangsu Wynch Corp., Ltd., that appears to employ a similar kinematic constraint.
This turbine is supported by a tall stand to reach a higher altitude, and has much narrower
blades than the authors prototype. [5]
3 Concept
The purpose of the authors drag-based vertical axis turbine design is to utilize the entire
rotational cycle of each blade for positive power contribution. Typical Savonius wind turbine
blades detract from the net torque on the rotor as they rotate into the wind; net positive
power generation is possible only because the drag coecient of the blades in their downwind
orientation exceeds that of the blades in their upwind orientation. While other designs, such
as those that employ blade aps, seek to essentially eliminate all forces on the turbine blades
during their upwind travel, the authors design is more ambitious in that it seeks to utilize
this regime for positive power generation. The potential applications of this new design
are expected to be similar to that of Savonius turbines small scale power and direct
mechanical work but with several advantages as discussed in Section 3.3.
3.1 Blade Movement
To create positively contributing torque on each blade at every point throughout the tur-
bines rotation, the angle of each blade is continuously adjusted to attain a near-optimal
combination of drag and lift. The blade passes through four main phases as the turbine
rotates (see Fig. 3):
1. Blade moving parallel to wind the blade face is roughly perpendicular to the wind
ow, and drag forces on the blade dominate.
10
2. Blade moving perpendicular to wind, behind turbine axis here, the blade is oriented
so that both lift and drag forces contribute positively to power generation.
3. Blade moving into wind in this regime the blade is oriented to minimize drag, and
lift forces dominate and contribute positively to power generation.
4. Blade moving perpendicular to wind, in front of turbine axis as in Phase 2, both
lift and drag forces act on the blade and contribute positively to power generation.
In the course of this movement, each turbine blade performs one half-rotation per full revo-
lution of the turbine. The key to this turbine design is this kinematic constraint: relative to
the inertial reference frame, each blade is constrained to rotate at a 1:2 ratio with respect
to the rotor.
Phase 1
Phase 2
Phase 3
Phase 4
wind ow
Figure 3: Depiction of several blade orientations as the turbine rotates. Drag forces (vertical force
vectors) on the blades contribute more heavily when the blade is moving with the wind, and lift
forces (horizontal force vectors) dominate when the blade is moving into the wind. The orientation
of the blade at each point results in a positively contributing torque on the rotor. (Note that small
frictional drag forces are neglected.)
11
The implementation of this kinematic constraint is discussed in Section 5, and an analysis
of the drag and lift forces on the turbine blades is presented in Section 4.
3.2 Blade Angle Control
Because of the 1:2 ratio of blade to turbine rotation, the blade orientation is not independent
of wind direction. An integrated mechanism is required to either actively or passively orient
the blades, or the entire turbine, to account for a shift in wind direction. The design
of the prototype turbine incorporates a mechanical system that adjusts all blade angles
simultaneously so that each blade continues to square up as it is moving parallel to the wind
ow (see Section 5.3). Such a mechanical system could be either passively controlled (by a
weather vane, for instance) or actively controlled (by a sensor and actuator).
3.3 Advantages
The primary advantage of this design over most drag-based vertical axis wind turbines, like
the Savonius turbine and its variants, is that, through the harvesting of both lift and drag
forces, power is generated by every blade throughout the turbines entire rotation. This
design, therefore, is expected to be more ecient than a standard Savonius turbine or its
variants, as discussed in Section 4.3.
This type of turbine also decreases the horizontal reaction force required to support the tur-
bine when compared to a Savonius turbine. For a vertical axis wind turbine, the supporting
force is approximately equal to the sum of the lift and drag force vectors on the blades. Since
a Savonius turbine incurs both positive and negatively contributing drag forces on its blades,
a portion of these forces cancel each other out with respect to power generation, but combine
to increase the required reaction force. The new design, however, eliminates negative drag
forces, and thus reduces the required reaction force to support a turbine generating the same
amount of power.
There are several other advantages that this turbine design could have over typical vertical
axis wind turbines. While some vertical turbines are either not self-starting, or can only
self-start from certain rotational positions, a turbine of this design will be able to start from
12
any position since every blade is always contributing positive torque. For the same reason,
this style of turbine is expected to have a higher static torque than a Savonius turbine (see
Section 4.2).
One drawback of most drag-based turbines is a low tip-speed. While this design does not
solve that issue by any means, this turbines theoretical optimal tip speed is slightly greater
than that of a Savonius turbine with similarly-dimensioned blades because, in this new
design, blades traveling into the wind are also generating power.
3.4 Disadvantages
As with most design alternatives, this vertical axis turbine design has drawbacks that need
to be weighed against its advantages. The most severe drawback is the added complexity
of the new design. The kinematic constraint requires more moving parts the blades are
no longer xed rigidly to the rotor and this change eliminates much of the simplicity of
a Savonius turbine. This complexity will likely add to construction and maintenance costs,
and this tradeo needs to be weighed against any gain in power-generation.
The kinematic constraint also requires a specic blade orientation at each rotational position,
meaning that, unlike a Savonius turbine, an adjustment needs to be made to account for
a change in wind direction. As previously mentioned, this is a solvable problem, but it is
another issue that increases the complexity of this style of turbine.
A danger of the added complexity is that it could actually hurt the overall performance of
this turbine relative to other vertical axis wind turbines. Additional moving parts create
additional friction surfaces that detract from power generation. The added weight from the
mechanical systems used to orient the blade could also decrease turbine eciency.
Furthermore, since each blade rotates only one half of a revolution per full revolution of
the rotor, there is less freedom in blade design and optimization than with other vertical
axis wind turbines. Each blade must be 180

rotationally symmetric so that there is no

dierence in drag or lift forces between consecutive rotations. If blades are rigid, then they
likely cannot be scooped as in an S-shaped Savonius turbine. This issue and alternative
blade designs are discussed further in Section 5.4.
13
And nally, this turbine is still a drag-based vertical axis turbine. While it may generate
more power than a Savonius turbine, it is still limited in its tip speed and power generation
when compared to lift-based, horizontal turbines. As with a Savonius turbine or any variant,
its applications are likely restricted to small-scale electric power generation or the direct
production of mechanical power.
4 Theoretical Analysis
The analysis performed in this section will assume thin, at, and rigid turbine blades. Each
blade is assumed to pivot about a point that is a distance r from the center of the rotor,
and is kinematically constrained as discussed in Section 3.1 each blade performs one half
of a rotation per full revolution of the rotor. For a given rotational position of the rotor, ,
the at blade makes an angle with respect to its initial position (or more specically, with
respect to a vector perpendicular to the wind ow vector, v
0
). Because of the kinematic
constraint imposed on the blades, we necessarily have =

2
. The majority of our analysis
will consider the blade that begins perpendicular to the wind when the rotor is at position
= 0 (see Fig. 4).
4.1 Lift and Drag Forces
Each blade has acting on it a lift force and a drag force,

F
L
and

F
D
, respectively, which
sum to a net force

F
net
=

F
L
+

F
D
(for this analysis we will neglect the smaller frictional
drag forces on the blades). We can dene lift and drag coecients, C
L
and C
D
, as
C
L
=
F
L
1
2
v
2
0
A
and C
D
=
F
D
1
2
v
2
0
A
, (3)
where is the air density, v
0
the faraway uid velocity, and A the area of the blade face.
There exist numerous complex approximating curves for these drag and lift coecients. For
this analysis, we will use a fairly simple approximation for the lift and drag coecients on a
at plate presented by Caplan and Gardner [1] due to its ease of manipulation. It is reason-
ably close enough to approximations presented by Hoerner and others for this analysis. [3]
14

F
D

F
L
r

F
net

2
v
b
v
0
v
0
x
y
=

2
Figure 4: Lift and drag forces
acting on a blade yield a net
force approximately normal to
the blade face. The blade an-
gle, is constrained to be half
the rotor angle .
Caplan and Gardner nd that, for a at plate,
C
L
= C
Lmax
sin(2) and C
D
= 2C
Lmax
cos
2
(), (4)
with C
Lmax
in the approximate range of 1.0 to 1.2 (we will
use C
Lmax
= 1 in any calculations). These approximations,
and hence the majority of this analysis, ignore any ow-path
interference between blades. Section 4.4 briey discusses this
assumption, but an actual dynamic ow path analysis is out-
side of the scope of this report. Note that with these approx-
imations for the lift and drag coecients, we have
F
L
F
D
=
C
L
C
D
=
C
Lmax
sin(2)
2C
Lmax
cos
2
()
= tan . (5)
This result (typically expressed as
F
D
F
L
= tan(

2
) for our denition of ) implies that the
net force on the blade,

F
net
, is approximately normal to the blade face.
4.2 Static Torque
When the turbine is stationary, each blade provides a net torque on the rotor equal to
= F
L
r sin + F
D
r cos . (6)
Substituting in the expressions for the lift and drag forces from Equation (3), as well as for
the lift and drag coecients of Equation (4) gives
=
1
2
v
2
0
ArC
Lmax

sin(2) sin() + 2 cos

2
cos

. (7)
Substituting =

2
into Equation (7) and simplifying gives
=
1
2
v
2
0
ArC
Lmax
(1 + cos) , (8)
which is an elegant approximation for the static torque generated by each blade at rotor
position . Noting that

n1
j=0
1+cos

+
2j
n

= n for all , for n 2 , it follows that the ap-

proximate net torque on the rotor from any number of evenly-spaced blades, N 2, is
15
= N
1
2
v
2
0
ArC
Lmax
. (9)
Without ow interference from the upwind blades, static torque would increase linearly with
the number of blades, as well as with blade area and rotor radius, and would be independent
of rotor orientation (implying that the turbine should indeed be self-starting from any rotor
orientation).
4.3 Dynamic Analysis
When the turbine is in motion, we must take into account the eects of the rotor movement
on the lift and drag coecients, as well as on the relative wind speed for each blade. As-
suming that the turbine is spinning an angular rate of

= , the center of each blade
is traveling at a speed v
b
= r (see Fig 5). The velocity vector of the blade center
is v
b
= v
b
sin x + v
b
cos y, and the relative wind velocity at the blade center, v
r
, is

F
D

F
L
r

F
net

2
v
b
v
0
v
0
x
y
=

2
Figure 5: Blade movement
and coordinate system for dy-
namic analysis. The wind ve-
locity relative to the blade is
v
r
= v
0
v
b
.
v
r
= v
0
v
b
= v
b
sin x + (v
0
y
b
cos ) y. (10)
We will assume that this is the relative wind speed across
the whole blade i.e. we will neglect the fact that the blade
is also rotating at a rate of

2
with respect to an inertial
reference frame. Although an imperfect assumption, it is
a more accurate than it would be in the case of a typical
Savonius turbine, where each blade is rotating at a rate of
with respect to an inertial reference frame.
This relative wind velocity gives both a new eective wind
speed and a new eective wind direction (and thus a new
eective for calculating the lift and drag coecients in
Equation 4). These are:
v
eff
= |v
r
| (11)
and

eff
=

2
(v
r
v
0
) , (12)
where we know v
0
=

2
.
16
0 2 4 6
Blade 1
Blade 2
Blade 3
Blade 4
Average Torque
Rotor Angle (radians)
D
y
n
a
m
i
c

T
o
r
q
u
e

o
n

E
a
c
h

B
l
a
d
e
0 2 4 6
Blade 1
Blade 2
Blade 3
Blade 4
Average Torque
Rotor Angle (radians)
D
y
n
a
m
i
c

T
o
r
q
u
e

o
n

E
a
c
h

B
l
a
d
e
0 2 4 6
Blade 1
Blade 2
Blade 3
Blade 4
Average Torque
Rotor Angle (radians)
D
y
n
a
m
i
c

T
o
r
q
u
e

o
n

E
a
c
h

B
l
a
d
e
0 2 4 6
Blade 1
Blade 2
Blade 3
Blade 4
Average Torque
Rotor Angle (radians)
D
y
n
a
m
i
c

T
o
r
q
u
e

o
n

E
a
c
h

B
l
a
d
e
Figure 6: Torque on each
of four blades during rota-
tion, neglecting ow interfer-
ence, with a rotor-tip speed
ratio of k = 0.3. The rst
blade, which is perpendicular
to the wind ow when = 0,
is highlighted for clarity.
At this point a symbolic analysis ceases to provide much insight, but plots generated by a
numerical analysis demonstrate the trends that we might expect. The torque on each of four
blades as a function of the rotor angle, , is plotted in Fig. 6 for a rotor-tip speed ratio of
k = 0.3 (explained in the next paragraph).
For this style of turbine, we can dene a rotor-tip speed ratio, k, as
k =
v
b
v
0
=
r
v
0
. (13)
The rotor-tip speed ratio diers from the more traditional tip speed ratio, (dened in
Equation 1), in that the radius used is based on the blades center, and not on the blades
tip. The rotor-tip speed ratio can be related to the tip speed ratio by determining what the
tip speed of the rotor would be if it were extended to the tip of the blade at position = 0.
For a blade width b and rotor radius r, we have
= k

1 +
b
2r

= k c (14)
where the constant c can range from 1 to 2, depending on the blade width. An intermediate
value of c = 1.5 (when b = r) is assumed for the torque and power calculations graphed
later. While is a more useful quantity for comparing this turbines tip speed ratio with
that of other turbines, k better allows us to plot torque and power curves not dependent on
blade width (when neglecting ow interference).
17
The static torque on each blade (i.e. the torque when k = 0) as well as dynamic torque on
each blade (for k = 0.3 and 0.5) are plotted as functions of rotor angle, , in Figure 7a. Figure
7b shows turbine power as a function of rotor-tip speed, k. Given our initial assumptions,
we nd that the turbine has a peak power output at k = 0.5, which corresponds to a tip
speed of = 0.75 when b = r (although this could be as large as = 1.0 for a turbine with
blades just wide enough to just reach the turbine center, b = 2r).
0 2 4 6
Static Torque on Blade (k=0)
Dynamic Torque on Blade (k=0.3)
Dynamic Torque on Blade (k=0.5)
Rotor Angle (radians)
T
o
r
q
u
e
0 0.2 0.4 0.6 0.8
k
P
o
w
e
r
0 2 4 6
Static Torque on Blade (k=0)
Dynamic Torque on Blade (k=0.3)
Dynamic Torque on Blade (k=0.5)
Rotor Angle (radians)
T
o
r
q
u
e
0 0.2 0.4 0.6 0.8
k
P
o
w
e
r
(a) Torque vs. Rotor Angle
0 2 4 6
Static Torque on Blade (k=0)
Dynamic Torque on Blade (k=0.3)
Dynamic Torque on Blade (k=0.5)
Rotor Angle (radians)
T
o
r
q
u
e
0 0.2 0.4 0.6 0.8
k
P
o
w
e
r
0 2 4 6
Static Torque on Blade (k=0)
Dynamic Torque on Blade (k=0.3)
Dynamic Torque on Blade (k=0.5)
Rotor Angle (radians)
T
o
r
q
u
e
0 0.2 0.4 0.6 0.8
k
P
o
w
e
r
(b) Power vs. Rotor-Tip Speed Ratio
Figure 7: (a) Torque as a function of rotor angle, , for several rotor-tip speed ratios, k. k = 0
represents the static torque on the rotor from each blade, and k = 0.5 is the case of optimal power
generation. (b) Power vs. rotor-tip speed ratio, with a clear peak at a rotor-tip speed ratio of
approximately k = 0.5. The relationship between k and tip speed ratio, , is given in Equation (14).
Given the assumptions above (e.g. no ow interference, and ignoring blade rotation), we can
evaluate the power coecient, C
P
(dened in Equation 2), at the optimal rotor-tip speed
ratio. For a rotor radius r and a blade width b and height h, the swept area is conservatively
given as
A
S
= (2r + b)h. (15)
(This value for A
S
assumes that the turbine width is equal to the rotor diameter plus the
blade width, even though the blade tip never actually reaches a point r +
b
2
away from the
turbines center during upwind travel.) Since the torque on each blade is proportional to its
18
area, A = bh, we expect the power coecient to be proportional to
A
A
S
=
b
2r+b
. Numerically
calculating the power produced given the dynamic movement of the blade shows that the
missing constant of proportionality happens to be 1.00, so we have
C
P
=
b
2r + b
=
b
r
2 +
b
r
. (16)
The theoretical power coecient, therefore, depends only on the ratio of blade width to
rotor radius. This ratio ranges from 0 (b = 0) to 2 (b = 2r), and so C
P
can range from 0 to
0.50, although for values of
b
r
close to 2, ow interference will likely have a signicant eect,
and would reduce C
P
from this theoretical value. For b = r, the turbine still achieves a
theoretical eciency of 0.33, on par with the most optimized Savonius turbines (see Section
1.2).
4.4 Flow Interference
It is beyond the scope of this analysis to attempt to accurately predict the dynamic inter-
ference that the blades eect on the wind ow. It is probably possible to approximate the
eect of the interference on the forces on each blade as a function of , v
0
, and k, as well as
blade geometry, and then modify the torque and power outputs accordingly. At this point,
however, any selection of such a function would be arbitrary and would lend no additional
insight.
A rudimentary 2-dimensional ow analysis in the static case at rotor position = 0 was
performed for v
0
= 4
m
s
and b = r = 0.5 m, and is shown in Fig. 8. The ow paths
qualitatively suggest that, while the wind energy is certainly reduced by the time it hits the
rear blade, it is not by any means eliminated. Interference eects might be further mitigated
when the rotor is spinning, since the relative wind speed, |v
r
|, tends to be less than v
0
for
the interfering blades.
19
Figure 8: Qualitative ow analysis for a static rotor at position = 0, with a wind speed of
v
0
= 4
m
s
and turbine geometry b = r = 0.5 m. Thick lines indicate ow paths, and blank voids
indicate blade wakes. While some ow interference does occur, signicant drag and lift forces will
still be applied to the rear blade.
5 Design and Prototype
A large component of this project was the design and construction of a prototype version of
this wind turbine. The primary purpose of the prototype was to demonstrate viable imple-
mentations of the kinematic constraint on the blade rotation and blade angle adjustment.
Future goals include optimizing the turbine design, more precisely measuring its performance,
and comparing its performance to that of a Savonius turbine.
This section explains the key design elements of the prototype, discusses challenges relating to
any implementation of this turbine concept, and gives results from preliminary testing.
20
5.1 Machine Design
The prototype design (Fig. 9a) consists of three main assemblies (Fig. 9c): a central shaft
that is used to constrain blade angles (Section 5.2); a rotor assembly that supports the central
shaft through angular contact bearings and also supports the blades; and a base assembly
that supports the rotor and that would hold the machinery and electronics required for power
generation and blade control (Section 5.3).
The prototype turbine had a rotor radius of r = 0.38 m, a blade width of b = 0.30 m, and a
blade height of h = 0.48 m.
5.2 Kinematic Constraint Implementation
Elegantly implementing the required kinematic constraint between the blade and rotor rota-
tion was the most dicult task from a design standpoint. The prototype uses a chain-driven
system, in which gears on a central shaft (concentric and interior to the rotor) are linked to
larger gears on the blade axles (see Fig. 9b). A gear ratio between the central and satellite
gears of 1:2 creates the required kinematic constraint. During normal operation, the cen-
tral shaft is held xed (relative to the ground). Each rotation of the rotor results in a half
rotation of each satellite gear, and thus a half rotation of each blade, as required.
5.3 Blade Angle Control
The 1:2 gear ratio between the central gear and satellite gears, described above, greatly
simplies the issue of adjusting the angles of the blades to account for a change in wind
direction. In fact, if the wind direction shifts by an angle , then the central gears need only
be rotated by the same angle and in the same direction as the wind change. Due to the
wind shift, the new rotor angle is

= , (17)
and due to the central gear adjustment the new blade angle is

=

2
=

2
. (18)
21
(a) Isometric View (b) Top View
Rotor Assembly
Base Assembly
Central Shaft
Chains, to Blades
(c) Cross Section
Figure 9: (a) An isometric view of the turbine solid model. (b) A top view demonstrates the
1:2 gear ratio between the central and satellite gears. (c) The three main assemblies the central
shaft (dark green), rotor (tan), and base (red) are shown in cross-section.
22
Because the new blade angle,

, is still half the new rotor angle,

, with respect to the new

wind direction, the kinematic constraint is maintained.
Control of the central shaft for adjusting the blade angle can be passive or active. In a passive
system, the central shaft could be directly driven by a wind vane. In an active system, a
sensor would detect a change in wind direction and a control system would drive a motor
to adjust the central shaft orientation. The passive system provides simplicity, while the
active system could protect against the central shaft being back driven by the geared system
(through the introduction of a worm gear or otherwise). In the authors prototype, a manual
adjustment system suced, since a fan was the only tested wind source. The prototype has
the capability to be upgraded to an actively controlled system.
5.4 Blade Design
Because each blade makes only one half of a rotation per full revolution of the rotor, there
exists an additional constraint that blades need to be symmetric about their axis under
180-degree rotation. For a rigid blade, this eectively eliminates the option of a scooped
blade design found on most Savonius turbines. Blades can be aerodynamically formed to
minimize form drag during upwind travel, as long as they maintain the required rotational
symmetry. For the authors prototype, at, reinforced, foam blades were used due to their
ease of construction.
Alternatively, blades could employ a fabric that acts as a sail, creating a shallow bucket
during downwind travel and falling slack during a portion of the upwind travel (and gener-
ating no power during this interval). Advantages of this blade type would include a reduced
blade weight and a potentially larger downwind drag coecient. If pursuing this design, care
would have to be taken to not suer the same pitfalls as slatted blade designs discussed in
Section 2.1, namely, loss of eciency due to centrifugal forces on under-constrained blade
elements.
The prototype was designed so that blades can be switched out easily for future testing
of dierent designs for blade optimization (see Section 6.1). The author was only able to
construct one set of blades during the course of this project, and was therefore unable to
test dierent blade designs.
23
5.5 Testing
Initial (and hardly comprehensive) testing was performed by placing the prototype turbine
downwind of a large fan in a hallway (Fig. 10). The turbine had a start-up wind speed of
below 1.5
m
s
. At an approximate wind speed of 4
m
s
, and with a rotor position of = 0, the
rotor provided a static torque of approximately 0.8 Nm. With no load, the turbine reached
a rotational rate of = 70 rpm.
No machinery for power generation was set up, but interpolating from static torque and
no-load speed gives a theoretical power of about 2.9 W, or an eciency of C
P
= 0.15, before
any losses due to power conversion (this is about half of the theoretical eciency of 0.29,
calculated as in Equation 16). While these measurements are very approximate (accurate
to at best 30 percent), they indicate the feasibility of achieving high performance with this
style turbine. More extensive testing remains to be performed to rene these measurements,
and design optimization performed to improve eciency.
During testing, at 70 rpm the turbine experienced no noticeable wobbling, but the blades
leaned backwards signicantly during transverse and downwind travel due to drag forces. A
combination of stier blades and a constraint on the tops of the blades is likely required to
remedy this issue.
(a) Stationary prototype turbine (b) Turbine during testing
Figure 10: Prototype turbine, stationary and during testing. In a wind speed of 4
m
s
, the static
torque was about 0.8 Nm and the no-load speed 70 rpm. At this speed, drag forces on the blades
caused them to lean backwards signicantly.
24
6 Conclusions
The technical analysis of Section 4 demonstrated that this type of turbine is capable of higher
eciencies (Equation 16), slightly higher optimal tip speed ratios (Equation 14), and high
static torque (Equation 9) than a typical Savonius Turbine. These increases in performance
come at an increase in mechanical complexity, as discussed in Section 5. These tradeos,
therefore, need to be weighed, and it is likely that this style turbine is most applicable for
small scale power generation or mechanical work production, where a small turbine footprint
(high eciency) is required.
Initial testing demonstrated that achieving an eciency relatively close to the theoretical
target presented in Section 4.3 is plausible, and extensive testing should be performed on this
prototype to verify the preliminary results presented in Section 5.5. To achieve eciencies
closer to the theoretical values, there remains much optimization to be performed on this
prototype, discussed in the next section.
6.1 Optimization
There is plenty of room for improvement through future optimization of this design. The
size and weight of the prototype rotor assembly were largely driven by economical bearing
selection, and so a second generation prototype could signicantly reduce rotor inertia and
bearing friction. Stiening and constraining the tops of the turbine blades is also expected
to improve eciency signicantly.
Blade style, shape, and dimensions should also undergo extensive testing and optimization.
Dierent blade concepts (e.g. rigid or sail-like, as discussed in Section 5.4) can be considered,
and the parameters for each design should be optimized (blade width vs. rotor radius, blade
height, etc.).
25
References
[1] Caplan, N. and Gardner, T.N. A Fluid Dynamic Investigation of the Big Blade and
Macon oar Blade Designs in Rowing Propulsion. Journal of Sports Sciences. 25, 6 (2007),
pp. 643-650.
[2] Dieudonne, P. EolProcess.com. <http://www.eolprocess.com> (accessed April 22, 2010).
[3] Hoerner, S.F. and Borst, H. Fluid-Dynamic Lift: Practical Information on Aerodynamic
and Hydrodynamic Lift. Published by Liselotte A. Hoerner, 1985.
[4] Hwang, I.S. et. al. Eciency Improvement of a New Vertical Axis Wind Turbine
by Individual Active Control of Blade Motion. Proceedings of SPIE. 6173-11 (2006),
<http://dx.doi.org/10.1117/12.658935>.
[5] Jiangsu Wynch Corp., Ltd. Vertical Axis Variable Pitch Wind-Power Turbine
<http://wynchcn.en.made-in-china.com/product/EbUmIRyCXlVN/China-Vertical-
Axis-Variable-Pitch-Wind-Power-Turbine-FDC-HV-Series-.html> and Vertical Axis
Wind Turbine (VAWT) <http://www.youtube.com/watch?v=lE6yDTVR35c> (ac-
cessed April 25, 2010)
[6] Modi, V.J. and Fernando, M.S.U.K. On the Performance of the Savonius Wind Tur-
bine. Journal of Solar Energy Engineering 111 (1989), pp. 71-81,
<http://dx.doi.org/10.1115/1.3268289>.
[7] Paraschivoiu, I. Wind Turbine Design: With Emphasis on Darrieus Concept. Montreal:
Polytechnic International Press, 2002.
[8] Reupke, P. and Probert, S.D. Slatted-Blade Savonius Wind-Rotors. Applied Energy.
40, 1 (1991), pp. 65-75,
<http://dx.doi.org/10.1016/0306-2619(91)90051-X>.
[9] M. Simonds and A. Bodek Performance Test of a Savonius Rotor. Technical report No.
T10. Brace Research Institute, McGill Unversity, Quebec, 1964.
[10] Tabassum, S.A. and Probert, S.D. Vertical-Axis Wind Turbine: A Modied Design.
Applied Energy. 28, 1 (1987), pp. 59-67,
<http://dx.doi.org/10.1016/0306-2619(87)90041-9>.
26