Comparison of Smart Rotor Blade Concepts for Large Offshore Wind Turbines

dinnerworkableΠολεοδομικά Έργα

16 Νοε 2013 (πριν από 3 χρόνια και 6 μήνες)

94 εμφανίσεις

Comparison of

Smart Rotor Blade

Concepts

for Large Offshore Wind Turbines

B.A.H. Marrant and Th. van Holten

Delft University of Technology, Faculty of Aerospace Engineering

Kluyverweg 1, 2629 HS Delft, The Netherlands

Tel.: +31 (0)15 27 85171

Fax.: +31 (0)
15 27 83444

E
-
mail:
B.Marrant@lr.tudelft.nl

Keywords: offshore, wind turbines, fatigue

loads,
smart materials, active rotor control


Abstract

This paper provides the results of a comparison of four different
smart s
tructure
concepts to obtain active rotor control on
large offshore wind turbines.
The concepts are active trailing
-
edge flap

control
, micro
-
electro
-
mechanical tab

control
,
camber control with inflatable structures and active
blade
twist control.
The differ
ent
smart rotor blade
concepts are
compared with each other based on their potential to reduce
fatigue

loads

for particular dimensions
, their aerodynamic
efficiency, bandwidth and complexity.

Introduction

The aim of the
research project

on smart dynamic ro
tor control for large offshore wind turbines is to develop new
technology capable of considerably reducing the extreme and fatigue loads on wind turbines


in particular on very large
wind turbines for offshore application


and thereby to reduce the costs

of wind turbines. A second aim is to reduce
maintenance requirements and improve reliability by applying condition
-
monitoring techniques.


The way to achieve this is to implement recent advances in control theory, sensor
-

and actuator technology, smart
st
ructures, etc. taking into account the special requirements and conditions of offshore wind turbines. The FLEXHAT
-
program performed in the Netherlands some years ago has shown that “smart” control methods may have a significant effect
on the various loads
[1]. The purpose of the former project was to decrease loads in the drive train and the rotor blades by
using passive tip control and flexibilities in the rotor system. Unfortunately, the techniques used within the FLEXHAT
configuration were not suitable f
or incorporation into very large wind turbines. Therefore, different solutions need to be
developed.

In a preliminary study [2,3] on smart dynamic rotor control for large offshore wind turbines different concepts have already
been investigated in wind ener
gy and helicopter literature. Helicopter literature has also been studied because of its close
relation to the field of wind turbines. Moreover, smart dynamic rotor control for the purpose of e.g. vibration reduction is
a
relatively new concept in wind ene
rgy whereas this topic has been the subject of study for many years in the field of
helicopters [4]
,
[10]
.

The two approaches which will be used to reduce the structural loads on wind turbines are the reduction of the fluctuations
of aerodynamic loads and t
he active/passive damping of structural modes [5]. In the preliminary study a list of control
devices for rotor blades was made which could be used
to
control the
extreme and
fatigue loads. This list included trailing
-
edge flap control

(possibly combined w
ith leading
-
edge flap control)
, Micro
-
Electro
-
Mechanical tab control, active twist
control, part
-
span and full
-
span pitch control and camber control. This study focused mainly on the feasibility of smart
materials as a means to actuate the devices. These c
oncepts have already been presented in reference [6].

This paper will continue the previous work with the purpose to make a ranking of devices for the purpose of smart dynamic
rotor control. T
he most promising devices are considered to be
trailing
-
edge fla
ps

and
Micro
-
Electro
-
Mechanical tabs

because of their relative simplicity and their potential. On top of these two devices

active twist and variable camber with the
use of inflatable structures

are also included
. The actuators which are considered
in this
paper are mainly
smart material
actuators based on piezoelectric materials. The different concepts that result from the devices and actuators are compared
with each other based on
their potential to reduce aerodynamic loads, their a
erodynamic efficiency, b
andwidth

and

complexity
.

Approach to compare the smart blade concepts


The four smart blade concepts
which were mentioned before
will be compared with each other
based on
their ability to
reduce fatigue loads
during normal operation of the wind turbine. Th
e fatigue loads are used
as a basis to compare

the
concepts

because

wind turbine
designs are

often
governed by fatigue [12]

and because
none

of the smart blade concepts
mentioned previously will have the power to
reduce

the
extreme loads completely.
T
he fa
tigue
load case during normal
power production (DLC 1.2), as described in the IEC standard [7], has been used
as a basis for the comparison
because this
will be the phase
during which the smart blade will be operating most of its time.

This means that norm
al power production
with

the occurrence of an emergency, start up, normal shut down and stand still load cases have not been considered.


The
fatigue load
calculations
for the conventional blade and the smart blade concepts
have been
performed
for wind tur
bine
class I B

because this involves a high reference wind speed average over 10 minutes (V
ref

= 50 m/s) and a medium
turbulence intensity at 15 m/s (I
ref
= 0.14)

which is considered
to be
representative for offshore wind conditions
.



The turbulence model

which has been used is
a
three
-
dimensional
, one component model,

which

means that the wind speed
varies

over the rotor disc

area

in time.

The method to simulate turbulence makes use of Fourier series to create a number of
correlated time series from the
l
ongitudinal
velocity component spectrum

(
S
1
(f)
)

and a coherence function.

The turbulence
spectrum used for the analysis is the Kaimal spectrum
:






3
5
1
1
2
1
1
6
1
4
hub
hub
V
L
f
V
L
f
S













(1)


where


1
=I
ref
(0.75V
hub
+5.6)
:


is the longitudinal turbulence standard d
eviation with
I
ref
=0.14

L
1
:




is the longitudinal velocity integral scale, where
L
1
=8.1

1

V
hub
:




is the wind speed at hub height

f
:





is the frequency in Hertz










m
z
m
m
z
z
60
42
60
7
.
0
1



is the longitudinal turbulence scale parameter


The following exponential coherence model

(
Coh
(r,f)
)

is

used in conjunction with the Kaimal autospectrum to account for
the spatial correlation of the longitudinal
wind speed

component:


















2
1
2
12
.
0
12
exp
,
L
r
V
r
f
f
r
Coh
hub







(2)


where
r

is the magnitude of the pr
ojection of the separation vector between the two points on to a plane normal to the
average wind direction.

-60
-40
-20
0
20
40
60
20
40
60
80
100
120
140
160
160
8
10
12
14
16
18
20
y [m]
z [m]
V [m/s]

Figure
1
: Turbulence for a wind speed of 13 m/s

over the height (z) and the width (y)

An example of the turbulence at a
wind speed of 1
3

m/s at

time

t=0 can be seen in figure

1

where the total grid spans 60 m x
60 m
,
the grid interspacing is 4 m

and the hub height is 91.4 m
. The turbulence
was

rotationally sampled by
selecting wind
speeds out of the complete wind field at p
oints in space and time corresponding to positions of the rotating
blade

of

a
horizontal axis wind turbine.
C
orrelated time series were generated in 60 equally spaced points on the
rotor
blade
and
li
near
interpolation

has been used

when the point on the bl
ade was
in between the grid points.


The deterministic part of the wind field
only
consisted of wind shear
where the longitudinal
wind speed

is

given by the
following power law:






2
,
0
hub
hub
z
z
V
z
V













(3)

T
he

fatigue loads for the different
sm
art rotor blade

concepts
are compared making use of
a
benchmark

wind turbine. For
this purpose an offshore wind

turbine derived from the DOWEC

(Dutch Offshore Wind Energy Converter)
wind turbine
study

[9]

is used
since this
involves a horizontal axis (HAWT
),

upwind, pitch
-
regulated, variable speed
turbine
which
has
the size
and characteristics of a wind turbine
this research is intended for.

A drawing of the original DOWEC
design

can be
seen in figure 2 and s
ome characteristic data of th
is

benchmark turbine

are presented in table 1.




Figure
2
: DOWEC wind turbine




Table
1
: Characteristic data of the benchmark wind
turbine


For the calculations

of the loads

the Wind
s
im

[18]

package has been used with a few alterations in order to be able to
calculate t
he aerodynamic loading due to a distributed wind field.
This package uses BEM theory for the calculations of the
aerodynamic loads
in which the
wake
is approximated as constant over the rotor disc area
.

The smart rotor blade concepts
were compared by calcu
lating their fatigue damage relative to the conventional blade. As a first approximation the dynamics
of the blade are neglected and onl
y the variable

aerodynamic loading
due to the variable wind field is considered
.

The
maximum load alleviation capacity o
f the smart structures has been used in the analysis where it has been assumed that the
smart rotor blade knows exactly what the wind field looks like at every time step
, moreover as a first approximation the
smart blade is assumed to react instantaneously

to the load change.

Making these assumptions
has the
inherent
advantage
that the controller can be left out of the analysis which leads to a more straightforward comparison of the smart rotor blade

concepts.
This way a first estimate can be made of the mi
nimum required dimensions, deflections and deflection rates for
the different smart structure concepts.

The smart rotor blade concepts which are able to alter the aerodynamic blade loads can be divided into two categories:
blades which are able to actively

change the airfoil’s camber thereby shifting the c
l
-


curve up
-

or downward and blades
which are able to change the local angle
-
of
-
attack in order to obtain changes in lift coefficient. Trailing
-
edge flaps, MEM
-
tabs and variable camber control belong to t
he first category whereas active twist control
belongs to the second
.

420
425
430
435
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
x 10
7
V = 13 m/s
t [s]
M
uu
[Nm]
M
uu
mean

Figure
3
: Variation of aerodynamic blade root
bending moment

due to a varying wind field and steady aerodynamic
blade root bending moment

at a wind speed of 13 m/s

Cut
-
in wind speed V
in

3.0 m/s

Cut
-
out wind speed V
out

25 m/s

Rated wind speed V
r

12 m/s

Tip speed ratio

r

7.4

Hub height z
hub

(above the water)

91.4 m

Rotor diameter D

120 m

Rated power P
r


6.0 MW

The steady condition

of the wind turbine rotor for a uniform wind speed and uniform wake was determined for each wind
speed in the operational envelope after which the variations in the aerodynamic blade root loads due to the var
ying wind
field were calculated
.

The variations

of the aerodynamic blade root bending moment for the benchmark turbine for a wind
speed at the hub of 13 m/s can be seen in figure
3

as well as the steady aerodynamic blade root moment which was
determined for uniform wind flow conditions. This is the bla
de root bending moment with respect to the blade chord
reference system.

In the ideal case the smart blade should return the blade root bending moments to the mean value. Since
the potential of the smart blade is depending on the concept and smart structur
e length the aerodynamic bending moments
will be
returned more or less
to the mean bending moment.
Figure
4

gives the aerodynamic blade root bending moment for
the conventional blade and the reduced blade root bending moments when a smart blade is used. In

this example the smart
blade is able to change the camber of the blade with maximum changes in the lift coefficient of

c
l
= ± 0.4. The length of the
smart structure is 33% of the total blade and is located outer part of the blade (near the blade tip).
As

can be seen from the
figure the amplitudes of the aerodynamic blade root loads have been reduced significantly for the smart rotor blade
compared to the conventional blade. The calculation of the blade loads have been done for time intervals exceeding 10
minutes.
The same calculations have been done for different changes in lift coefficient
and twist
and percentages of
the
rotor
blade
containing
the

smart
structure
.

420
425
430
435
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
x 10
7
V = 13 m/s, smart length = 33%,

c
l
= 0.4
t [s]
M
uu
[Nm]
M
uu
M
uu,smart
, f
max
=


Figure
4
: Variation of the aerodynamic blade root bending moment

for the conventional and

the smart blade

due to
a varying wind field at a wind speed of 13 m/s

A smart blade which has the ability to react instantaneously to load changes because of its infinite bandwidth is not realist
ic.
A way to account for this limit
ed

bandwidth of the smart

rotor
blade

concepts
is by

cut
ting
-
off of the
maximum frequency
of the
Fast Fourier Transform
of the blade root bending moments of this smart
rotor
blade

in figure
4
.

420
425
430
435
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
x 10
7
V = 13 m/s, smart length = 33%,

c
l
= 0.4
t [s]
M
uu
[Nm]
M
uu
M
uu,smart
, f
max
=

M
uu,smart
, f
max
= 1Hz

Figure
5
:
Variation of the aerodynamic blade root bending moment for the conventional
and the smart blade

with
and without limited bandwidth

due to a varying wind field at a wind speed of 13 m/s

The result can be seen in figure
5

where the smart rotor blade concept is now limited to a bandwidth of 1Hz. It can be seen
that the fluctuations
in aerodynamic blade root bending of the smart rotor blade with limited bandwidth are still much
smaller than that of the conventional blade but there are more fluctuations with respect to the smart rotor blade with infini
te
bandwidth.

In figure
6

the powe
r
spectrum
of the aerodynamic
blade root bending
moment
s

for a wind speed of 13 m/s
can
be seen for a) the conventional blade and b) the smart rotor blade
.

The dashed lines represent the 1P, 2P, 3P and 4P
frequencies.

It can
clearly
be seen that there is
a

large
reduction in fatigue loads when
a

smart
rotor

blade is used

at its
maximum

performance
.

0.2356
0.4711
0.7067
0.9422
10
8
10
9
10
10
10
11
10
12
10
13
10
14
10
15
10
16
V = 13 m/s
f [Hz]
a)
PSD [(Nm)
2
]
0.2356
0.4711
0.7067
0.9422
10
8
10
9
10
10
10
11
10
12
10
13
10
14
10
15
10
16
V= 13m/s, smart length = 33 %,

c
l
= 0.4
f [Hz]
b)
PSD [(Nm)
2
]

Figure
6
: Power spectral density of the a) aerodynamic moment of the blade, b) aerodynamic and smart structure
moment

Once the aerodynamic

blade

root
bending
m
oments of the
conventional and smart structure
blade
were determined
for the
varying wind conditions

they
were

converted to both compressive and tensile stresses in the blade root structure. Thereafter

the

rainflow counting method

[11],

was

used

to determ
ine
the number of cycles at each
stress
amplitude and mean
stress
value

combination
.
T
he fatigue damage

at each wind speed
was then determined with the Palmgren
-
Miner rule
for the
conventional
rotor
blade
and the different
smart
rotor blade concepts with v
arying smart structure lengths
:

j
i
j
i
j
i
N
n
d
,
,
,













(4)

w
here
d
i
,j

i
s the
fatigue
damage,
n
i
,j

is the number of cycles at a particular stress amplitude and mean stress
combination
as
determined with the Rainflow counting method and
N
i
,j

is t
he number of allowable cycles before the structure will fail

for
each stress amplitude
i

and mean stress
j

combination
.

When the damage
d

is equal to 1, the structure is at the end of its life.

The number of allowable cycles

N

for
the case of
mean tensile
stress
[12]
for composites
can be calculated with:

10
1

























tu
mean
tu
N















(5)

where

tu

is the ultimate tensile stress,


is the stress amplitude and

mean


is the mean stress.
In the case of mean
compressive stress

tu

is replaced with
-

cu

in
the

mean

term.

The values of

tu

and

cu

for a glass/polyester ply are used with
50% fibre volume fraction unidirectional fibres running in the longitudinal direction, because this is the most popular
composite material to build blades with

[12]
.
Then a
ll

the damages
for each wind speed have been

summed

up
. In order to
make a direct comparison
with the conventional blade
the
ratio of the
damage for the

smart rotor
blade
and the

conventional
blade
is used
at each wind speed.
In order to obtain
an overall re
lative damage factor
,

the relative damages
for each smart
blade concept
at each wind speed
are summed up
taking into account the amount of time the wind turbine is operating at that

particular wind speed:




k
conv
k
smart
k
k
d
d
K
,
,













(6)

where
K

is the overall relative damage

factor
,

k

is the wind speed,

d
k,smart

is the total damage of the smart blade
,

d
k,conv

is the
damage of the conventional blade

and

k

is the fraction of time the wind turbine is operating at

a particular

wind speed
k

as
determ
ined with a Rayleigh wind
speed
distribution
.
Therefore t
he different smart blade concepts can be compared
using
their overall relative damage factor

for both
tension
and
compression (see equation (5) and its counterpart for compression)
.

Since
the structu
re gets more damage in case of a mean compressive stress
, only the overall relative damage factors for
mean compressive stress
will be considered. However, the
overall relative
damage

factors
for compressive and tensile mean
stress are very close to each o
ther

for the material under consideration
.

Smart rotor blade c
oncepts

T
he
concepts considered for active load control on wind turbines
are

active
trailing
-
edge flaps, MEM
-
tabs, inflatable
structures

for active camber control

and active
blade
twist.
As men
tioned before, t
hey are able to control the effective airfoil
camber or the local blade pitch. Since these devices mainly influence the lift forces and not as much the drag, their strengt
h
lies in the possibility to change the out
-
of
-
plane bending moments
of the blades. Devices which can have an influence on the
in
-
plane blade bending moments will probably need to be of non
-
aerodynamic origin.

The actuators which have been
considered for the previous devices are piezoelectric actuators.
P
iezo
-
electric mater
ials are able to provide excitations at
very high frequencies, much higher than
required for load alleviation purposes on large wind turbines
,

however because of
the difference in dimensions (and inertia) of the different smart blade concepts

there will be

a difference in
maximum
frequency which can be obtained with the piezoelectric actuators
.
Piezoelectric materials are able to
exhibit maximum
strains around
2000


[25].

When active trailing
-
edge flaps
, see figure 8,

are considered for active load control
two possible concepts can be mentioned:
active trailing
-
edge flaps
with a physical hinge

to make flap deflections possible
and smart structure trailing
-
edge flaps

based on bender elements
.


Figure
7
: Trailing
-
edge flap concepts

A
ctive trailing
-
edge flaps
with a hinge

are a classic approach and are commonly used both on helicopters and airplanes.
They can be actuated using a piezo
electric
stack actuator in combination with an LL
-
amplifier

[10]

or with

classic
electro
-
actuators.

Piezoelectric stack actuators

consist of several layers of piezoelectric sheet material where normal strain is used as
driver. Stacks are capable of delivering large forces but only small displacements. Therefore they need to be amplified with
an LL
-
amplifier.



Figure
8
:
Stack actua
tor with

LL
-
amplification

Trailing
-
edge flap control with piezoelectric stack actuators has been proven on a full scale helicopter rotor with flap
deflections of
-
10


to +8


.

The smart structure trailing
-
edge flap can be thought of as a unimorph or bimorp
h actuator, see
figure 9, which could be used as a retrofit on existing rotor blades without the need to make many changes to the structure o
f
the blade.

Unimorphs and bimorphs are bender actuators. A bender actuator is obtained when two layers are operate
d such
that one extends and the other contracts.

These actuators can give
large displacements but only
relatively
small forces.

The
Riso
-
B1
-
18 airfoil makes use of such a trailing
-
edge flap to obtain lift control [14].

The performance of these actuators ca
n
be improved when holding them under compression
as has been shown on UAV’s
[26].

In reference [
4,
6]
lift control and
moment control
were considered for trailing
-
edge flaps. With direct lift control the change in effective camber leads to a
change in lift
. In the case of moment control, the blade root is considered to be soft. Therefore the change in effective
camber results in a twist of the blade which indirectly leads to a change in lift. Since wind turbine rotor blades exhibit
relatively large torsion
stiffness, direct lift control is considered to be the
logic

option for wind turbines.
Typical possible
changes in lift coefficient are in the order of

c
l
=

0.4

and even maxima of

c
l
=

0.5

have been found in literature
[14,15
,
17
].

The aerodynamic efficienc
y which is characterised by the lift over drag ratio is hardly influenced at small
angles
-
of
-
attack

when the trailing
-
edge flap is deflected
.

Active trailing
-
edge flaps have relatively fast response times,
especially when the total trailing
-
edge flap is di
vided into separate segments, therefore reducing its inertia and making it
more fail
-
safe. The response times
of trailing
-
edge flaps
will be
slower than MEM
-
tabs but faster than active blade twist
.

A
disadvantage of a
ctive trailing
-
edge flaps

is that they

usually give rise to acoustic
problems
.





Figure
3
: Unimorph and bimorph bender elements

Micro
-
electro
-
mechanical (MEM) translational tabs were proposed for active load control by D. Yen and C.P. van Dam
[8
,20
]
, see figure 10
.

MEM
-
tabs are d
erived from Guerney flaps which were introduced by Liebeck in 1978

[21]
.

They

are
installed near the trailing
-
edge of the rotor blade

for maximum effect on the Kutta condition
.
When moving the location of
the MEM
-
tab in the direction of the
leading
-
edge it becomes less effective but
for
tab
-
locations up

to

9
5
% chord its
effectiveness can be guaranteed

[16]
.

When moving the tab
-
location further in the direction of the leading
-
edge, the tab will
lose its effect
due to reattachment of the flow b
ehind the tab
.

The MEM
-
tab deploys approximately normal to the surface
and has a maximum translation in the order of the boundary layer thickness
, 1
-
2% of the chord
.
Figure 10

only shows a
MEM
-
tab which is deployed on the upper surface in order to reduce t
he lift, but the same device can also be placed at the
trailing
-
edge of the bottom surface to increase the lift.

Another possibility to reduce the lift is to locate the tab

near the
onset
of pressure recovery in orde
r to induce flow separation
.

When the up
per and lower tab are
both
installed near the trailing
-
edge, typical increases in lift coefficient
which have been found in literature
[13,15,16
,20
]

are

c
l
=

0.3

for a tab height
-
to
-
chord ratio
h
t
/c=1%

with maximum increases of

c
l
=+0.
4

for a
h
t
/c=2%
and

w
hen the upper tab is located at
the onset of
pressure recovery even

values of


c
l
=
-
0.55

have been found.

The lift over drag ratio
s

of airfoils with
MEM
-
tabs

decrease
with respect to the base airfoil for lift coefficients up to the base airfoil maximum lift

coefficient
, above this maximum lift
coefficient it may exceed the lift over drag ratios of the base airfoil
.

Therefore smart blades with MEM tabs will have a
negative influence on the power production since
lower

lift over drag values are
encountered

[22
].

For the current generation
of airfoils it will

not

be
easy

to
implement

these devices

at the trailing
-
edge

because the

volume

available
to install the
MEM
tab
s is limited
.

This means that
when MEM
-
tabs are to be used in smart rotor blades,
airfoils with

a
n increased
trailing
-
edge thic
k
ness

should be designed to include these tabs.

Another possibility to overcome this problem would be the
use of rotational tabs instead of translational tabs.

In case of a rotational tab, the tab would be hinged at the airf
oil skin
making it possible to take a position perpendicular to the surface as well as
tangential to the surface when it is not deployed
.
and from the

Because of the minute size of these devices, much faster response times will be achieved than
with
any ot
her
concept

mentioned in this paper
.

According to the specifications in reference [20] these MEM
-
tabs could even have a
maximum frequency in the order of 70 Hz when magnetic actuation is used. However, magnetic actuation has only been
proven in a laborator
y environment which means that the maximum obtainable frequency will be much lower.

MEM
-
tabs
can be actuated using classic electro
-
actuators or smart material actuators

based on piezoelectric materials
.


Figure
4
: MEM
-
tabs

Anothe
r means of controlling the airfoil lift is
the use of

camber control, see figure 11.
Changing the airfoil camber with the
use of smart material fibres embedded in the skin is not possible as has been shown in reference [
4,
6] since there are no
smart materi
als on the market which are able to exhibit such large strains

(
in the order of
98000

)
. Therefore another set
-
up
is proposed being active camber control with the use of inflatables.
Inflatables have found an increasing amount of
applications, e.g. in the
military

and

space

applications where these structures

need to stay

operat
ional

in a harsh
environment.
In this case several separate air chambers on the “top” or the “bottom” surface of the airfoil are inflated which
push a flexible rubber skin more outwa
rd
,

thus increasing/decreasing the camber of the airfoil. Using many air chambers
provides the stiffness of the flexible outer skin.
The air chambers can be inflated or deflated with a piston like arrangement
as in a car engine or with an
air
pressure vess
el which is held under pressure with an air
-
pump.
This might involve a
complex system and leakage may be one of the biggest problems.
Instead of air, viscid fluids might also be used to ensure
maximum rigidity but this has a negative effect on the weight.

The changes

in lift coefficient

as well as the maximum
frequency

which can be obtained with this concept are assumed to be
comparable with

trailing
-
edge flaps

as given above
.

In
the case that the outer skin does not get any bubbles, a clear advantage would

be the smoother surface skin which would
reduce the acoustical problems encountered with trailing
-
edge flaps.



Figure
5
: Camber control with inflatable structures

Active blade twist
, see figure 12,

can be accomplished with the u
se of piezo
-
fibre composites

when t
he orientation of the
smart fibres is 45


with

respect to the direction of

the main spar of the blade

such that actuation of the active fibres produces
a shear deformation of the blade
.

Active fibre composites are develop
ed at

the Active Materials and Structures Lab at

MIT
.
This type of device has uniaxially aligned piezoceramic fibres surrounded by a polymer matrix and can include inactive
glass fibres for increased structural strength.

Interdigitated electrodes deliver t
he electric field required to activate the
piezoelectric effect in the fibres as seen in figure 13.


Figure
6
: Active blade twist

Therefore they are called Inter
-
Digitated Electrode Fibre Composites (IDEPFC) [23]. Active fibre com
posites offer many
advantages with respect to their monolithic counterparts [24]. The multiphase construction yields a more robust actuator
which can be added to a lay
-
up as an active layer along with the conventional fibre
-
reinforced laminae. Also the fle
xible
nature of the polymer matrix allows the material to more easily conform to the curved surfaces such as in airfoils. Moreover
the polymer matrix protects the active piezoelectric fibres which have a brittle nature.
The strain required to obtain a twis
t of
4


at the tip

has been determined to be

in the order of 800


for a
wind turbine
blade of 60 meters with
a

chord of 3.4 meters
and

a

t/c ratio of 28%.
This
amount of
strain

is still feasible for piezoelectric
materials
. However this

was determined for

a
straight blade, for a tapered blade the strains will consequently be higher.
T
he inherent

problem with
active blade twist

is
that the
complete blade must be
made with active

fibre

composite
s. This will lead to
huge cost
s

because

of the
large amount
of c
ostly piezoelectric materials

and to
an increase in blade weight

since led is one of the components in piezoelectric
material
s
.

The same concept has already applied on helicopter rotor blades [10]. However, the control forces to obtain active
twist will be

much higher for wind turbine blades with respect to the helicopter rotor blades because of the much larger t/c
ratio of typical wind turbine airfoils.

Because of this and the relatively large inertia active blade twist

will show the slowest
response times

with respect to the other concepts mentioned in the paper.

A clear advantage of this concept is that a smooth
rotor blade is obtained which does not change the aerodynamic behaviour of the original blade design.


Figure
7
: Inter
-
Digitated

Electrode
Piezoelectric

Fibre

Composite

Comparison of smart rotor blade concepts

The

smart rotor blade concepts
can

be
com
pared with each other based on their
overall relative damage factor
K
.
This factor
K

gives the f
atigue damage of a smart rotor blade concept as a fraction of the fatigue damage of the conventional blade

without smart structures

during normal power production.

When the smart
structure length
of the rotor blade is 0% the blade
becomes

the

conventional

blade

and

K

goes to
1
.

The
smart rotor blade concepts
are

compared
in figure 14

for actuator
s

with an infinite bandwidth
,

i.e. they are able to react instantaneously to load changes
,

and for different
s
mart structure
lengths in span wise direction

where th
e smart structure always starts from the blade tip
.
Before attaching to
o

much
importance to the

overall relative damage factor
K it must be
emphasised

that the smart

rotor
blade concepts
in the analysis
are able to
react infinitely fast to the load changes

and

know

these

load

changes

before they even occur
. Therefore the
reduction in damage
with respect to the conventional blade
has hardly any significance
, only the
relative differences
between the concepts
are

important
.

0
20
40
60
80
100
10
-5
10
-4
10
-3
10
-2
10
-1
10
0
Smart length [%]
Overall relative damage factor [-]

c
l
=

0.4


=

2deg


=

4deg

c
l
= -0.55 - 0.3

c
l
=

0.5

c
l
= -0.55 - 0.4

c
l
=

0.3

Figure
8
: Comparison of smart rotor blade concepts

with infinite bandwidth

From figure 14 it can be seen that the fatigue loads can be reduced more when the smart structure length is increasing but
smart structure lengths exceeding 30% become less effective in red
ucing fatigue loads.
T
railing
-
edge flap

control

or camber
control (
both

c
l
=

0.4
)

reduce the

fatigue
loads more than MEM
-
tab
control when

the
upper and lower
tab are both located
near

the trailing
-
edge (

c
l
=

0.3
)

for a tab height over chord ratio:

h
t
/c
=1%
.

In case that the upper tab is

located

near

the

onset of

pressure recovery
to induce flow separation downstream of the tab

(

c
l
=
-
0.55
-

+0.3
)

nearly
the

same
reductions in
fatigue

are

obtained as for the trailing
-
edge flap

and

camber control
concept up to
30% smart structure length
.
For larger
smart structure lengths the

trailing
-
edge flap and camber control

concept

become more
effective in reducing the loads.
When a larger MEM
-
tab (
h
t
/c
=
2%
) at the lower

surface is used (

c
l
=
-
0.55


+0.4
) the
reduction in
loads

are the same as for
the smaller
pressure tab (
h
t
/c=1%
and

c
l
=
-
0.55
-

+0.3
)
for all smart structure lengths
but the larger tab will lead to larger
drag
and thereby
reducing the aerodynamic efficiency. When the trailing
-
edge flap control and camber co
ntrol are pushed to
their extremes (

c
l
=

0.5
) the largest load reductions can be seen. However this will also lead to larger drags

due to flow
separation
.

In case of active twist control, the complete blade is twisted which means
that the

smart structure l
ength
is equal
to

100%.

From

figure 14 it can be
concluded

that
active twist

control (
tip

angles
:


=

2


and


=

4

)

is not an efficient way to
reduce loads since the same load reduction can be obtained

with
trailing
-
edge flap control
/
camber control (

c
l
=

0.4
) and
MEM
-
tab control (

c
l
=
-
0.55
-

+0.3
)

but this with only 17% smart structure
length
instead of 100% for the


=

2


and 35%
smart structure

length

instead of 100% for the


=

4


active twist concept.

The smart rotor blade concepts
are

compared
in fig
ure 1
5

for actuators
with a finite bandwidth,
i.e. they know

the load
changes

before they occur

but cannot react infinitely fast to these changes, and for different smart structure lengths in span
wise direction. It can be seen in the figure the damage red
uctions by the smart rotor blades become less than for the case
where the smart rotor blades have infinite bandwidth

leading to more realistic reductions in fatigue damage
.

Taking time
delays due to the controller and the unsteady aerodynamics will improve

the predictions of the fatigue reduction.

Since the
bandwidth of the different smart rotor blade concepts are not very well known, some numbers have been assumed while
keeping their relative reaction quickness into account.

MEM
-
tabs are considered to be t
wice as fast as active trailing
-
edge
flaps/active camber control and four times as fast as active twist. Although active trailing
-
edge flaps/active camber control

(

c
l
=

0.4
, f
max
=1Hz
)

is slower than MEM
-
tab control

(

c
l
=

0.3
,f
max
=2Hz
)
, it gives the larges
t reduction in fatigue loads.
Only MEM
-
tabs with a larger tab at the lower surface
(

c
l
=
-
0.55
-

+0.3
,f
max
=2Hz
)

can keep up with the active trailing
-
edge
flap/active camber control concept up to 15% smart structure length.

0
10
20
30
40
50
60
70
80
90
100
10
-2
10
-1
10
0
Smart length [%]
Overall relative damage factor [-]

c
l
=

0.3, f
max
= 2Hz

c
l
=

0.4, f
max
= 1Hz

c
l
= -0.55 - 0.3, f
max
= 2 Hz


=

2deg, f
max
= 0.5 Hz

Figure
9
: Comparison of smart rotor blade concepts with limited bandwidth

Conclusions

From the preceding analysis it
can be stated that
active trailing
-
edge flap control/active camber control
(

c
l
=

0.4
)

is about
twice as effective as MEM
-
tab control

(

c
l
=

0.3
)
.

Only MEM
-
tabs with a larger tab at the lower surface
(

c
l
=
-
0.55
-

+0.3
,f
max
=2Hz
)

can keep up with the active trailing
-
edge flap/active camber control concept up to 15% smart structure
length.
For active trailing
-
edge flaps, active camber control and MEM t
abs, s
mart structure lengths of 30% are most
efficient for the reduction of fatigue loads.

Although ac
tive twist control is

feasible, it is expensive, results in heavier blades
and is a
very inefficient way to reduce fatigue loads.

Since active camber cont
rol with inflatable structures is a relatively
new concept in wind energy this will need to be investigated
further
.

Acknowledgements

This research is supported by the Technology Foundation STW, applied science division of NWO and the technology
programme

of the Ministry of Economic Affairs.
I
also

want
to
thank
ir.
W.A.A.M. Bierbooms for the use of the Windsim
package.

References


[1]

G.A.M. van Kuik, J.W.M. Dekker, The FLEXHAT program, technology development and testing of flexible rotor
systems with fa
st passive pitch control, Journal of Wind Engineering and Industrial Aerodynamics, 39 (1992) 435
-
448

[2]

B.A.H. Marrant, K. Hinnen, Th. Van Holten, G.A.M. van Kuik, Smart Dynamic Rotor Control of Large Offshore
Wind Turbines, Inventory of Present Techniques
, Duwind 2002.011, May 2002

[3]

B.A.H. Marrant, Th. Van Holten, G.A.M. van Kuik, Smart Dynamic Rotor Control of Large Offshore Wind
Turbines, Inventory of Rotor Design Options and Possible Load Reductions, Duwind 2002.012, May 2002

[4]

F.K. Straub, A feasibi
lity study of using smart materials for rotor control, Smart Material Structures 5, 1
-
10, 1996

[5]

M. Geyler, Advanced Pitch Control for Wind Turbines, Duwind 2001.001, January 2001

[6]

B.A.H. Marrant and Th. van Holten, Smart Dynamic Rotor Control on Large Offsho
re Wind Turbines, Napels, April,
2003, Naples, Campania, Italy

[7]

IEC 61400
-
1 Ed.3, Wind Turbines, PART 1: Design Requirements
, 25 May 2006

[8]

D.T. Yen, C.P. van Dam, R.L. Smith and S.D. Collins, Active load control for wind turbine blades using mem
transl
ational tabs, AIAA 2001
-
0031, Wind Energy 2001, January 2001, Reno CA

[9]

H.J.T.Kooijman, C. Lindenburg, D. Winkelaar, E.L. van der Hooft, DOWEC 6 MW pre
-
design: Aeroelastic
modelling of the DOWEC 6 MW pre
-
design in PHATAS, DOWEC
-
F
1W2
-
HJK
-
01
-
046/9, September 2
000

[10]

I. Chopra, A. Gessow, Recent Progress on the Development of a Smart Rotor System,
26
th

European Rotorcraft
Forum, September 26
-
29 2000, The Hague, The Netherlands

[11]

International Energy Agency, International Recommended Practices for Wind Energy Conversi
on Systems Testing:
(3) Fatigue Characteristics, 1984

[12]

T. Burton, D. Sharpe, N. Jenkins, E. Bossanyi, Wind Energy handbook, ISBN 0 471 48997 2, John Wiley and Sons,
Ltd, 2001

[13]

J.R. Zayas, Active Aerodynamic Load Control For Wind Turbine Blades, EWEC, Athens,

Greece, 2006

[14]

N. Troldborg, Computational Study of the Riso
-
B1
-
18 airfoil with a hinged flap providing variable trailing
-
edge
geometry, Wind Engineering, Vol. 29, No. 2, 2005

[15]

W.A. Timmer, R.P.J.O.M. Rooij, Summary of the Delft University

Wind Turbine Dedic
ated Airfoils, AIAA
-
2003
-
0352

[16]

K.J. Standish, C.P. van Dam
, Computational Analysis of a Microtab
-
Based Aerodynamic Load Control System for
Rotor Blad
es, Journal of the American Helicopter Society

[17]

I. Abott, A. Von Doenhoff, Theory of Wing Sections, Dover pub
lications, Inc. New York, 1958

[18]

WINDSIM:
http://www.duwind.tudelft.nl/Website%20WINDSIM/main.htm

[19]

D.T. Yen, C.P. van Dam, R.L. Smith and S.D. Collins, Active load control for wind turbi
ne blades using mem
translational tabs, AIAA 2001
-
0031, Wind Energy 2001, January 2001, Reno CA

[20]

D.T. Yen, C.P. van Dam,
F. Bräuchle,
R.L. Smith and S.D. Collins, Active
L
oad
C
ontrol
and
L
ift
E
nhancement
U
sing
MEM

T
ranslational
T
abs, AIAA 200
0
-
2242
,
Fluids
200
0
,
19
-
22 June 2000
,
Denver CO

[21]

R.H. Liebeck, Design of Subsonic Airfoils for High Lift, Journal of Aircraft, Vol. 15, (9), September 1978, pp. 547
-
561

[22]

W.A. Timmer, R.P.J.O.M. Rooij,
Design of Airfoils for Wind Turbine Blades
,
Delft University of Technolo
gy, The
Netherlands, 3 May, 2004

[23]

R.C. Derham, N.W. Hagood, Rotor Design Using Smart Materials to Actively Twist Blades, AHS 52
nd

Annual
Forum, Washington D.C., June 4
-
6, 1996

[24]

R.B. Williams, D.J. Innman, An Overview of Composite Actuators with Piezoceramic
Fibres, Centre for Intelligent
Material Systems and Structures, Department of Mechanical Engineeering, Virginia

[25]

B. Enenkel, V. Klöppel, D. Preißler
, Full Scale Rotor with Piezoelectric Actuated Blade Flaps, 28th European
Rotorc
raft Forum, September 17
-
19,
2002, Bristol United Kingdom

[26]

R. Vos,
“Post
-
Buckled Precompressed elements: a new class of control actuators for morphing wing UAV’s”
, Ma
ster
thesis, TU Delft,

2005