M Rotors Dreamw (Jupe)
M Rotors Dreamw (Jupe)
M Rotors Dreamw (Jupe)
P.T. Smulders
Faculty of Physics
ARRAKIS
There exists an abundance of all types of machines to convert wind power into useful mechanical
power on a shaft. There is an enormous number of inventions, claiming to be superior to all other
types, many of them being trash (Figure 1.1).
Machines to capture wind power can be classified according to the following aspects (see also
Figure 1.2):
- vertical axis
- drag
* movement: - rotation
- oscillation
Figure 1.1 Types of machines to extract power from the wind. Plenty of crazy ideas!
Figure 1.2 Classification of wind rotors using lift or drag as propelling factor
As an example let us take an airfoil, such as is used on airplanes. Let W be the relative velocity to
the airfoil (Figure 2.1)
Lift perpendicular to W
Drag in direction of W
The power of the airplane motor is needed to move against the drag on the wings and of course
the body.
It can be shown that we can write the drag and lift as follows:
(2.1a)
(2.1b)
in which A is the area c x b (see Figure 2.1). c is the so-called chord of the airfoil section, α is the
angle of attack.
CL(α) and CD(α) are dimensionless; ½ρW2 has the dimension of pressure (N/m2); ρ is the
density (in our case of air); ρair 1.2 kg m-3.
Equations (1a) and (1b) are important. For one particular shape of an airfoil, we can determine CL
(α) and CD(α) from experiments in a wind tunnel. Knowing A and α and the wind tunnel
velocity W, we can determine CL(α) and CD(α) by measuring the lift L and the drag D.
A typical example of how CL and CD vary with the angle α is shown in Figure 2.2. It is useful to
plot CL versus CD. The tangent determines an important point (A), with α = α0 (its optimum
value), where CD/CL is minimum or CL/CDis maximum. This is a design point for airplanes, wind
rotors etc. and of course gliders. Why?
Values of lift and drag coefficients can be found in different handbooks. Values of CL for
different airfoils at point A are around 1
(CL 1); (CL/CD)max values range from 20 to 150 (see also Figure 2.6). The corresponding angles
of α are 5 to 100.
Figure 2.2 Lift and drag coefficient CL(α) and CD(α)
To understand how lift is generated, let us look at the flow around an airfoil. At small angles α
the flow around the airfoil is smooth. The figures show the corresponding pressure distribution
on the airfoil surface.
Figure 2.3 Flow pressure and pressure distribution around an airfoil at small angles of
incidence α
The negative pressure on the top side "pulls" the airfoil upwards, the positive pressure on the
lower side "pushes" it upwards. Note that the contribution of negative pressures to lift on the
upper-side is larger than that of positive pressures on the lower side. If angle α increases, the
pressure distribution become more pronounced and lift increases, so CL(α) increases.
At a certain point (higher α) the flow breaks away from the upper side (see Figure 2.4). This is
called STALL. The pronounced negative pressure distribution is destroyed and the lift drops.
Figure 2.4 Flow around airfoil at stall conditions
In reality it shows that the drag and lift coefficients are not only dependent on α, but also on the
so-called REYNOLDS NUMBER.
W is relative velocity (see Figure 2.1); c is the chord, υ is the kinematic viscosity.
Basically Reynolds number takes account of the viscous forces relative to other forces in the
flow.
Fig. 2.5 shows the airfoil characteristics of the well-known NACA 4412-profile for different
Reynolds numbers. Below a certain value of Re the characteristics change considerably (curve d
and e compared to a, b and c).
Figure 2.5 Lift and drag coefficient for NACA 4412 profile at different Reynolds numbers.
As we have seen the minimum value of (CD/CL) or the maximum value of CL/CD) is an important
value for designers. Figure 2.6 shows (CL/CD)max of all kinds of airfoils at different Reynolds
numbers.
Figure 2.6 Showing how (CL/CD)max varies with Re - number for a variety of profiles. (CL/CD)max
corresponds to point A in Figure 2.2. CL value at this point for different profiles is approximately:
CL 1.
Fig. 2.6 (CL/CD)max varies with Re - number for a variety of profiles.
Most airfoil data are available for high Reynolds numbers (Re > 106). Less (but still many) for
lower Re-numbers. Figure 2.6 shows that (CL/CD)max decreases with decreasing Reynolds
number. In fact at Reynolds number around 105, curved plates are quite good airfoils. For
example, the Smitz 417a (usually called Göttingen 417) is a curved plate.
2.3 Whether to use lift or drag to extract energy from the wind?
It seems clear that we might try to extract energy from the wind by using either the drag force or
the lift force. For a rotating machine this is rather complicated, so we will examine the question
by analysing the movement in a straight line of a wing or sail, which is caused by either drag or
lift.
Suppose a curved sail moves with a velocity u in the direction of the wind speed V as depicted
below in Figure 2.8.
Figure 2.8 Model of a pure drag machine.
Note that: u < V. So the maximum power is extracted at velocities of the sail lower than V. In
fact, to extract power from the wind by a drag machine, u must be smaller than V.
Fu = L sin φ - D cos φ
(2.4)
P = Fu u
The lift and drag are given by: (see Figure 2.1 and equation 2.1)
c = chord
b = length airfoil
λ = u/V (speed ratio). Now Figure 2.9 shows tg φ = V/u = 1/λ.
Verborgen tekst 9
We find:
This shows that a relatively small, fast running airfoil can extract the power from the wind over
larger areas! This is also the basic reason that a well-designed fast running one-bladed rotor can
extract the same amount of energy from the wind as a slow running multi-bladed rotor. Further it
is clear that the power we can extract with a lift machine is much higher than with a drag
machine! (compare equations 2.3 and 2.4)
In reality λ-values that can be reached are much lower. Not only the drag of the airfoil is
important: the drag of the sailing boat in water is much higher.
A "sailing boat" on ice runs very fast (up to 100 km/hr). It has a low drag compared to a ship in
water!
Conclusion:
1. Always use lift as the driving propelling force. With the same area, lift machines can extract
more power from the wind than a drag machine.
2. Drag machines operate at speeds lower than the wind speed. Lift machines can operate at
higher speeds than the wind speed.
Figure 3.1 Wind power is converted by the rotor to mechanical shaft power
Wind power is converted into mechanical shaft power by the rotor (Figure 3.1). Rotor radius is
R. Let shaft torque be equal to Q. Let n be number of revolutions per second of the shaft. Ω =
2π.n is the angular shaft speed.
P = Q.Ω (3.1)
Define:
b) torque coefficient:
c) power coefficient
A wind rotor extracts kinetic energy from the air flow. The velocity of the air behind the rotor is
therefore lower than in front of it. The reaction force of the rotor on the airflow slows down the
velocity of the air. So besides a torque, also an axial force is exerted on the rotor. In fact, this
force decreases the momentum of the airflow. We call this axial force Fax and define:
In Fig. 3.2 a plot is shown of the torque/rotational speed characteristics at different wind speeds,
as could be measured in a wind tunnel. If we now convert these graphs into a graph of the
dimensionless torque coefficient versus the tip speed ratio, we see that all the original curves
merge into one single curve. This curve is called the dimensionless rotor torque characteristic of
the wind rotor.
Figure 3.2 How different Q - Ω curves merge into one single curve by writing them
dimensionless
A similar result is found for the axial force and power curves.
1. by forgetting altogether that rotor blades are involved and replacing the rotor by a so-
called actuator disk. This is an old concept dating back to the last century in trying to
understand ship propellers. Betz used the concept to determine the ideal maximum power
extraction by a wind rotor (Betz ± 1926). It is normally referred to as “momentum
theory”.
2. by looking at the forces on the individual rotor blades: “blade element theory”.
The two methods are used in combination to determine the rotor characteristics: so called “blade
element/momentum theory”.
Substitute the rotor by an actuator disk (see Figure 3.3). Assume that the disk extracts energy (in
some way or other) from the airflow. The flow that goes through the disk looses energy and
therefore slows down. Betz found that maximum energy is extracted if:
In that case:
So approximately no more than 60% of the undisturbed wind power flowing through the
equivalent disc area A can be extracted by a “machine”.
According to (3.7a) 8/9 Betz) 8/9 of the power of the air flow that goes through the disk (or
rotor) is extracted by the disk, as the kinetic energy drops from 1/2ρV3 to 1/2 ρ Ve3 = 1/9.1/2 ρ
V3. But we see that a part of the flow through an area A upstream of the rotor is diverted
outwards and does not pass through the rotor. This explains quantitatively Betz maximum. The
power extraction is determined both by the amount of air that goes through the rotor and by the
percentage of kinetic power that is extracted from it. By extracting little power, a large mass
flows through the rotor, but with little power extracted. In the other extreme, if all the energy
were extracted from the flow through the rotor, then the velocity behind the rotor would be zero!
That means that all the flow would be diverted outside of the rotor plane. Betz optimum is a
delicate balance between a sufficient airflow through the rotor and a sufficient power extraction
from that air!
Pmax = 16/27 . 1/2 ρ V3 . A Betz
Fig. 3.3 Description of a rotor as an actuator disk. Betz used this model to determine a
theoretical limit of maximum power output of a rotor.
As a blade element rotates, the velocity relative to it is different from the wind velocity itself. The
speed of the blade element is Ωr (Fig. 3.4).
Figure 3.4.a and 3.4.b Showing a blade element and the relative flow to that element.
As Betz model showed the velocity of air at the rotor disk is lower than the wind velocity, by the
action of all the rotor blades together. So the relative velocity is slightly different. If the value of
the air velocity at the rotor disk is V(1 - a) (so smaller than V) than the relative velocity W is
shown below. (Fig. 3.5.a)
The forces on the blade element result from the relative velocity W (Fig. 3.5.b).
(ΔFtang . r) gives a contribution ΔQ to rotor torque. The axial force mainly results from the lift!
α=φ-β
Fig.3.6 φ is the angle of relative flow to rotor plane
We could do the calculation for each and every blade element and so find the torque Q and the
axial force Fax and of course P (power). Assuming we know Ω and V and R, we could calculate
CQ, Cp and CFax. But it is not so simple! As we do not exactly know the retardation of the flow at
the disk (V(1 - a)), we do not know W and φ!! And that is what complete rotor theory is about,
solving the results from momentum theory and blade element theory all in one go! This,
however, lies outside the scope of this course module.
We already explained that according to Betz 16/27 of the power of the wind, that would
otherwise pass the rotor area without being disturbed by the rotor, can be extracted by an
"actuator disk" or theoretical rotor. A real rotor however is different. Three effects are important,
which result in lower power extraction.
1. Wake rotation
The air exerts a torque on the rotor. Vice-versa the rotor exerts a torque on the air. The wake
rotates behind the rotor and so the actual velocity of the air in the wake is higher than the velocity
if only axial changes are considered. Wake rotation can be considered as a loss!
2. Drag effects
Betz assumes that all the energy extracted by the disk is useful energy. He therefore excludes
drag losses. As we can see from equation (3.7a), drag diminishes the resultant torque! The higher
CL/CD-value, the less is the influence of drag.
3. The effect of finite number of blades
At the tip of the blade, a normal two-dimensional flow cannot be maintained. There is a high
negative pressure on the upper side of the airfoil and a positive pressure at the low side (Fig. 2.3).
The air will leak around the tip of the blade from the high pressure side to the low pressure side.
The lift reduces at the tips. Several methods have been proposed to correct for these tip losses.
The effects of the three types of losses mentioned above have been calculated and are shown in
fig. 3.7,3.8 and 3.9.
Fig. 3.7 Maximum attainable power coefficient of a rotor with an infinite number of blades
designed for a given value of the tip speed ratio λ.
3. other curves show combined effect of wake rotation and drag losses.
Fig 3.9 Effect of number of blades B on rotor performance for CD/CL = 0.03
(includes losses owing to wake rotation)
Note:
These graphs are not rotor characteristics, i.e. one curve does not represent the characteristics of
one type of rotor.
The curves only indicate the maximum attainable power coefficient, that could be obtained for a
desired tip speed ratio if the rotor is designed for maximum power extraction.
Note:
The curves implicitly show that it is possible to design rotors for different values of the tip speed
ratio. That is why we distinguish:
We can now try to understand the characteristics of real rotors (Fig.4.1). On the left we have slow
running rotors with many blades. On the right we see the fast running rotor with few blades.
Slow running rotors are used for driving piston pumps, that require large torques and run at
relatively low speeds.
Fast running rotors are used for electricity generation. Generators run at high speeds and require
low torque levels. In general rotor speeds are still too low and (except for very small machines)
there is always a gearbox between the rotor and the generator.
Fig. 4.1 Examples of rotor characteristics
For small WECS, especially wind pumps, the rotor is gradually turned out of the wind at higher
wind speeds. This is done for control and safety reasons
Safety and control are treated in detail in two separate modules of this course.
In this section we will discuss a few characteristics of the forces acting on the rotor.
Consider a rotor in yaw at an angle δ to the wind direction. If a rotor is in yaw, then all the forces
acting on all the elements of the rotor blades can be combined to two forces acting in the centre
of the rotor and a moment(Fig.4.2).
These two forces are: the axial force Fax (along the rotor axis) and a side force Fs (in the rotor
plane perpendicular to Fax). The moment Mso, which tends to push the rotor back into the wind, is
sometimes called the self-orientating moment. This moment is generated by the non-symmetrical
distribution of forces on the rotor blades.
We now define dimensionless coefficients for these forces, similar to earlier definitions (equation
3.4, 3.5 and 3.6).
So all these coefficients are not only a function of λ but also of the yaw angle δ. The curves
shown in Figure 4.1 were Cp(λ,0) curves, i.e. Cp as a function of λ with δ = 0.
Measurements of forces on wind rotors, especially slow running rotors, are rather scarce.
However, the graphs shown, based on measurements at CARDC and TU Eindhoven, are
probably typical for slow running rotors.
Figure 4.3 Cp-λ curves for various yawing angles δ of the CWD 2000.
The curve clearly shows that power is reduced by yawing and also the maximum rotor speed is
reduced.
Figure 4.4 The dimensionless axial force (CFax(λ,δ)) for various yawing angles for the
CWD2000
Note 1:
CFax for δ = 0 is nearly 1 near the design point (λ = 1.6, see Fig. 4.3). Values for fast running
rotors say with two blades are more or less the same. Such a value is comparable to the drag
coefficient of a solid circular plate of the same size as the rotor! This clearly shows the necessity
of safety systems at high windspeeds! So yawing reduces the axial force on the rotor.
Fig. 4.5 Side force measurements for CWD2000[ ]
Note: Fig 4.5 shows that at about 30o to 45o the side force becomes relatively important and tends
to push the rotor out of the wind.
Exercises
1. Determine the lift and drag of a NACA 4412 for alpha = 6 degrees at
2. Determine the power that a wind rotor of 10 m diameter can extract from the wind? Assume
the power coefficient is 0.4 and take a wind speed equal to the average for a region in your
country, where the use of wind energy seems favourable. What is the power if the rotor diameter
is 1m?
3. A WECS has a rotor of 15 m diameter. Measurements at 6 m/s show that the rotor turns at 60
rpm and that the torque on the axis is
250 Nm. Determine the tip speed ratio at which the rotor is turning; also determine the power
coefficient and the torque coefficient. How many blades do you expect the rotor will have
approximately?
4. Make a sketch of the relative velocity (relative to the blade element) of a rotor running at a tip
speed ratio of 1 (e.g. wind pump) and one of 6 (e.g. electric generation). Explain why the
requirements for the aerodynamic properties of the blade to obtain an equal power coefficient, are
different.
6. A multi-bladed (3 m) rotor of a wind pump and a two bladed (5 m) rotor of a wind generator
are located near each other at a test site. The multiblade rotor operates at design ratio tip speed 1
and the rotor of the wind generator at design tip speed 6. What is the ratio of their rotational
speeds and estimate the ratio of shaft torque of both machines.
7. Seen from the front a rotor turns clockwise. In which direction does the wake rotate? If two
rotors of equal size (one slow runner and one fast runner) operate at the same site, which wake
will exhibit more rotation and why?
8. Make a sketch of the lift coefficient of an airfoil.
9. Determine the ratio CD/CL of a NACA 4412 profile for different alpha values and sketch the
result. What does it tell you?