6 Loads and Structural Stresses - Removed

Chapter 6
Loads and Structural Stresses
Wind turbines are subjected to very specific loads and stresses. At a first glance, the
main problem appears to be the stability in heavy storms and hurricanes. However, the
continuous alternating loads - even under normal wind conditions - are just as problem-
atic. Alternating loads are more difficult to cope with than static loads because the mate-
rial becomes "fatigued".
The dimensions of the components present a further problem. As a working medium,
the air is of low density so that the surfaces required for converting energy must be
large. If the dimensions of the rotor increase, the dimensions of other components must
also increase, for example the tower height. Large structures are inevitably elastic and
the changing loads thus create a complex aeroelastic interplay which induces vibrations
and resonances and can produce high dynamic load components.
The structural dimensioning of a wind turbine must be considered under three differ-
ent aspects:
Firstly, attention must be paid to ensuring, from the point of view of breaking
strength, that the components are designed for the extreme loads encountered. This
means in real terms that the turbine and its essential components must be able to with-
stand the highest wind speeds which may occur.
The second requirement is that the fatigue life of the components must be guaranteed
for their service life, as a rule 20 to 30 years. While the stresses with respect to extreme
loads can be estimated relatively easily, the problem of "fatigue life" is virtually the key
issue with wind turbines. Wind turbines are the perfect "fatigue machines"!
The third aspect concerns the stiffness of components. On the one hand, structures
with elastic properties reduce fatigue, but on the other hand, external excitations pro-
duce vibrations in elastic components. The vibration behaviour of a wind turbine can be
kept under control only when the stiffness parameters of all its components are carefully
matched in order to avoid hazardous resonances and additional dynamic loads. Dynamic
loading on the wind turbine unavoidably does exist due to the changing external loads
but it should not be amplified by a critical vibrational behaviour of the components.
An important set of problems, even before the structural design loads are calcu-
lated, concerns the loads to be stipulated and the situations in which the loads occur
168 CHAPTER 6 LOADS AND STRUCTURAL STRESSES
which determine the dimensions of the structure. This requires a complete overview
of all external operating conditions and of possible malfunctions of the turbine. On
the basis of this, the so-called load cases can be defined. However, the real loads to
which the wind turbine is subjected can never be covered in their entire complexity
which is why they can always be stipulated only in an approximated, idealized form
as design loads. The load assumptions, i.e. the load cases with the loads form an
important basis in the design process. The mathematical methods needed for calcu-
lating structural loads and material stresses include some of the most complex theo-
retical tools required for developing wind turbines. The models are basically not
different from those used in other fields of technology. Nonetheless, the course of
action to be taken in relation to the structural design of a wind turbine is governed
by its own set of problems.
The starting point for the entire load spectrum of a wind turbine are the loads act-
ing on the rotor. The loads on the rotor blades are passed on to the other components
and to a great extent determine their loading. Compared to these loads, the loads
originating directly from downstream components are less significant. Discussions
of the loads acting on a wind turbine can, therefore, be concentrated on the rotor and
deal with it as being representative of all parts.
6.1 Loads on the Wind Turbine

The causes of all forces acting on the rotor are attributable to the effects of aerody-
namic, gravitational and inertial forces. The different loads and stresses can be classi-
fied according to their effect with time on the rotating rotor (Fig. 6.1):
- Aerodynamic loads with a uniform, steady wind speed, and centrifugal forces,
generate time-independent, steady-state loads as long as the rotor is running at a
constant speed.
- An air flow which is steady, but spatially non-uniform over the rotor-swept area
causes cyclic load changes on the rotating rotor. This includes, in particular, the
uneven flow towards the rotor due to the increase in wind speed with height, a
cross-flow towards the rotor and interference due to flow around the tower.
- The inertia forces due to the dead weight of the rotor blades also cause loads
which are periodic and thus unsteady. Moreover, the gyroscopic forces produced
when the rotor is yawed must also be included among those which increase or al-
ternate with each revolution of the rotor.
- In addition to the steady-state and cyclically changing loads, the rotor is subjected
to non-periodic, stochastic loads caused by wind turbulence.
For an investigation of structural stresses, it is important to consider the effects of
load variations with time. Fluctuating and alternating loads must be recognized,
especially with respect to the fatigue life of the structure. Extreme loads in specific
situations have to known to ensure the survival of the structure with respect to ulti-
mate stress.
6.1 LOADS ON THE WIND TURBINE 169
Fig. 6.1. Effect of aerodynamic, gravitational and inertial loads of the rotor of a horizon-
tal-axis wind turbine
It is not possible to recognize beforehand which of the loads within the entire
range of loads will be dominant. As is the case in all structures, the larger the tur-
bine, the greater is the significance of the gravitational loads. Moreover, the elastic-
ity of the structure plays an increasing role with respect to the extent to which the
external loads are transformed into structural stresses. For example, the speed vari-
ability of the rotor or the elasticity of the rotor blades are of significance for the
extent to which the external loads are converted into structural stresses.
In other words: apart from the external loads, the load level is also determined by
the design of the wind turbine. In general it holds true that the more elastic the
structures are, the better alternating loads can be absorbed and thus material fatigue
can be reduced, but on the extreme end, the flexibility of the structure causes vibra-
tion problems, and not lastly the mathematical effort increases with increasing elas-
ticity in the dimensional design of the structure.
6.2 Coordinate Systems and Terminology

Unfortunately, there is to the present day no legally binding standard for the position
and orientation of the coordinates in which the load parameters are represented. The
same also applies to the terms to be used for the parameters. In English usage, the des-
ignations used in IEC Standard 61400-1 have gained currency without being legally
binding in the sense of a Standard [1]. In German usage, many other terms are also
commonly used and in the German edition of this book older German expressions are
also used and their English equivalents are referred to only occasionally.
It would be confusing if one would try to put all the parameters and figures in one
common co-ordinate system. Already the very different order of magnitude of the vari-
ous values speaks against it. Therefore to represent the loads on the rotor and the struc-
tural stresses, three co-ordinate systems are suitably used (Fig.6.2).
The forces and moments acting on the rotor blades are resolved in a rotating co-
ordinate system with respect to the local rotor blade cross-section. In the direction of the
airfoil chord, the "chordwise" component is obtained and perpendicularly to the airfoil
chord it is the "flapwise" component. This approach is practical when the loads on the
rotor blades themselves are considered.
The breakdown with respect to the plane of rotor rotation provides the "tangential
force components" in the plane of rotation and the "thrust components" perpendicularly
to the plane of rotation. These co-ordinates express the total forces and moments on the
rotor when they are passed on to the remaining parts of the turbine in the form of loads.
At the transition from the chordwise and flapwise directions of the blade to the tangen-
tial and thrust directions of the rotor, the local twist angle and the blade pitch angle must
be taken into consideration.
A third system of co-ordinates has its origin at the tower base, in the case the wind
turbine is considered to be a “building”. In this coordinate system, the forces and mo-
ments acting on the tower and the foundation are shown, at least in Germany, in accor-
dance with the guidelines of the Deutsches Institut für Bautechnik (German Institute for
Structural Engineering) (DIBt) [2]. The DIBt is attempting to introduce an adaptation to
the IEC Standard in the more recent editions of its "Guidelines".
6.3 SOURCES OF LOADING 171
Fig. 6.2. Co-ordinates and techn

nical terms for representing loads and stresses on the rotor
6.3 Sources of Loadin

ng
The sources of loading to be b taken into account are aerodynamic, gravitational and
inertial loads. There are also
o loads arising from operational actions and different oppera-
tional states of the wind turbine. In the worst case, many of these sources produce looads
simultaneously resulting in cumulative effects.
The complex load spectru um of the rotor and the entire wind turbine becomes ccom-
prehensible only when the total
t loading is mentally resolved into components whhose
origins are independent of on ne other. This applies both to the loads due to aerodynaamic
forces and to those resulting from gravitational and inertial forces. As regards the aaero-
dynamic loads, the load situaation is determined by the varying flow conditions acting on
the rotor.
6.3.1 Gravity and Inertial Loads

Whereas the aerodynamic loading can only be calculated with difficulty, loads
caused by the dead weight of the components and by centrifugal and gyroscopic
forces are relatively simple to calculate. The only difficulty is that, at the beginning
of the design phase, the masses of the components are not known. As mass can only
be calculated as a consequence of the complete load spectrum, including the dead
weight, several "iteration loops" are unavoidable when dimensioning the structure.
First assumptions as to the weight are best taken from empirical data, prepared sta-
tistically from existing turbines.
Gravity loads
Loads resulting from the dead weight of the components must naturally be taken
into consideration for all components of the turbine. In a wind turbine, the rotor
blade weight is of special significance for the blades themselves, as well as for the
"downstream" components.
The rotor blade weight generates alternating tensile and pressure forces along the
length of the blade and large alternating bending moments around the chordwise
and flapwise axes in the blades over one rotor revolution. The significance of this
gravitational loading increases from the blade tip to the root, i.e. in the opposite
direction from the influence of the aerodynamic loads. This cyclic loading and par-
ticularly the cyclic bending moments around the blade's chord axis, occur with 107
to 108 cycles during the life of a turbine, assuming a rotor speed of 20 to 50 rpm and
a service life of 20 to 30 years. A number of 106 load cycles is reached after only
approximately 1000 hours operating time. From this number of cycles onward,
steel, for example, may only be stressed with its allowable fatigue stress.
Thus, together with wind turbulence, the influence of the gravitational forces be-
comes the dominant factor for the fatigue strength of the rotor blades. The larger the
rotors, the greater these influences will be. As is the case with any other structure, as
dimensions grow, it is ultimately the structure's weight which becomes the main prob-
lem with respect to strength. For horizontal-axis rotors, the situation is aggravated by
the fact that the dead weight causes alternating loads. Proponents of the vertical-axis
concept thus rightly point out that for this particular reason, the vertical-axis rotor is
more suitable for extreme dimensions, as the alternating loads caused by the dead
weight of the rotor blades are avoided.
In the past, some designers of horizontal-axis rotors attempted to compensate for
these alternating bending moments by installing lead-lag hinges at the rotor blade
roots. These did not, however, prove successful in practice. For one, the compli-
cated mechanisms involved are too expensive and, for another, they are associated
with additional problems of dynamics. If the rotor diameter is very large, this under-
taking would prove to be pointless, in any case. The better approach, or rather the
only possible approach, is to reduce the natural weight of the rotor blades. A light-
weight type of construction is almost mandatory for very large rotor blades even if
expensive materials such as carbon fibre have to be used.
Centrifugal loads
Centrifugal forces are not very significant in wind rotors, due to their comparatively
low rotational speed. This is in contrast to helicopter rotors, where blade strength and
dynamic behaviour are determined by the centrifugal forces.
With a special trick, centrifugal forces can even be used to relieve the load on the
rotor blades. On some rotors, the rotor blades are inclined downwind out of the
plane of rotation, in a slight V-shaped form. This so-called cone angle of the rotor
blades has the effect that the centrifugal forces, in addition to the tensile forces,
create a bending moment distribution along the blade length which counters the
bending moments created by the aerodynamic thrust. However, complete compen-
sation can only be achieved for one rotor speed and one wind speed.
If the rotor is subjected to other flow conditions, the effect of the cone angle can
be reversed. When the aerodynamic angles of attack are negative, for example with
a sudden drop in wind speed, or fast pitching of the blades (rotor emergency stop),
the direction of thrust can be reversed for a short time so that the bending moments
from the aerodynamic forces and the centrifugal force combine. Whether or not a
cone angle of the rotor blade makes sense technically must, therefore, be decided
after having taken several aspects into consideration. In more recent turbines there
is a tendency to have rotors without cone angle.
Gyroscopic loads
Loads caused by gyroscopic effects occur when the rotating rotor is yawed into the
wind. A fast yawing rate leads to large gyroscopic moments, which manifest themselves
as pitching moments on the rotor axis. However, as yawing rates are normally low, the
practical effects are very slight, or, in other words, the yawing rate must be so slow
(approx. 0.5 degrees/sec) that gyroscopic moments do not play a role. It would be
uneconomical to have to dimension the structure to the gyroscopic forces (s. Chapt.
10.2). It may occur, that the torsional eigenfrequency of the tower will be excited
caused by the yawing motion. This effect has to be considered in the design of the yaw-
ing system to avoid additional loading on the tower and on the coupled rotor-tower
system.
The attempts to build wind turbines with passive yawing have shown that the gyro-
scopic forces become a serious problem for these turbines. When wind directions change
rapidly, it is unavoidable that the rotor will also be yawed very quickly. Under these
conditions, rotor blades, in particular, are subjected to extraordinary bending loads due to
the gyroscopic forces involved. Abrupt changes of wind direction are to be expected
above all during low wind speeds. This is another reason why passive yawing, which, in
any case, can only be implemented on downwind rotors which are no longer being built,
is more than problematic (s. Chapt. 10.2).
6.3.2 Uniform and Steady-State Air Flow

Assuming a uniform, steady wind flow is, of course, an idealization which does not
exist in the open atmosphere. For practical purposes, this concept is nevertheless
useful to calculate the mean load level occurring over a relatively long period of
time. If a steady, symmetrical flow entering the area swept by the rotor is assumed,
the rotor blades of a horizontal-axis rotor are subjected to steady-state aerodynamic
forces. This characteristic distinguishes the horizontal-axis rotor from the rotors
with a vertical axis of rotation. Darrieus rotors or similar types are already subject
to time-variant loads due to aerodynamic forces under these conditions (Chapt. 5.8).
The wind loads on the rotor blades during steady and symmetrical flows are largely
determined by the effective wind speed varying from the blade root to the tip. In
addition, the geometrical shape of the rotor blades influences the load distribution
over the length of the blade. Diagrams 6.3 and 6.4 provide an impression of the
aerodynamic load distribution on the rotor blades.
The bending moments on the rotor blades in the chordwise direction are the result of
the tangential force distribution, whereas the thrust distribution is responsible for the
blade bending moments in the flapwise direction. Owing to the rotor blade twist, in
particular, the distribution profile changes distinctly from the start-up wind speed to the
shut-down wind speed. The twist is optimized for a nominal wind speed only so that the
distribution of aerodynamic loads corresponds approximately to the theoretical
optimum only for this windspeed. At other wind speeds, especially higher ones, the
flow separates in the blade sections near the hub. This causes the distribution of the
aerodynamic loads to change considerably.
Integrating load distributions over the length of the rotor blade yields the overall
rotor loads and moments. The tangential loading provides the rotor torque, and the
thrust load distribution provides the total rotor thrust. These two parameters essen-
tially determine the static load level for the entire turbine. In rotors with blade pitch
control, rotor torque and rotor thrust increase continuously up to the point where the
control system limits the power to the rated power (Fig. 6.5). Rotor thrust is greatest
at the rated power point, then drops off again.
In the case of rotors not incorporating pitch control, where power capture is re-
stricted merely by aerodynamic stall, rotor thrust continues to increase, or remains
at an approximately constant level, after having reached rated power. For this rea-
son, among several others, turbines without pitch control are subjected to higher
steady-state loads (Chapt. 5.3.3).
The previous examination of the loads on rotor blades only relates to the distribu-
tion of loads in the direction along the blade. This two-dimensional load picture in
reality masks a "mountain range" of loads which also extends in the direction of the
blade chord. Information about load distribution over the blade chord is usually of
minor significance but is, nevertheless, necessary for dealing with some problems
concerning torsional stiffness of the blade. Furthermore, this load distribution must
be taken into consideration when dimensioning the skin and ribs of the rotor blade,
at least when dealing with large rotor blades with a correspondingly deep chord.
The chord-wise load distribution is usually derived from pressure distribution meas-
urements, carried out on model airfoils in the wind tunnel. Airfoil catalogues contain
information on these pressure distributions. They are characteristic of each airfoil and
vary with the aerodynamic angle of attack (Fig. 6.6). Moreover, like the shape of the
airfoil lift and drag characteristics, they are affected by the Reynolds number. Hence,
applying them to the airfoil cross-sections of the original rotor blade must be done with
some care.
Fig. 6.3. Tangential load distribu

ution over the blade length of the experimental WKA-60 windd
turbine
Fig. 6.4. Thrust load distribution

n over the blade length of the WKA-60
176 CHAPTER 6 LOADS AND STRUCTURAL STRESSSES
Fig. 6.5. Torque and rotor thrustt with a steady air flow on the WKA-60
Fig. 6.6. Aerodynamic pressure distribution for the NACA 4412 airfoil [3]
6.3.3 Vertical Wind Sheear and Cross Winds

The wind flow produces un nsteady, cyclically varying loads as soon as it strikess the
rotor asymmetrically. One unavoidable asymmetry of the oncoming wind flow w is
caused by the increase in wind
w speed with height. During each revolution, the rrotor
blades are subjected to hig gher wind speeds in the upper rotational sector andd are
thus subjected to higher loaads than in the sector nearer the ground. A similar assym-
metry of flow at the rotor is caused by the largely unavoidable crosswinds whhich
occur with fast changes in wind
w direction.
The vertical wind shear and
a crosswinds on the rotor lead to a cyclically increaasing
and decreasing aerodynamicc load distribution over the rotor blades. Compared too the
basic loading with a steady y, symmetrical wind, there are considerable variationns in
load (Fig. 6.7). The linearly
y asymmetrical wind stream assumed here in this exam mple
qualitatively stands for the vertical wind speed profile or also for an asymmetrrical
wind flow due to a change in n wind direction.
The changing aerodynam mic loading on the rotor blades during one rotor revvolu-
tion, of course, also means varying total rotor loads and hence varying loads forr the
remaining parts of the turb bine. The cyclically changing pitching and yawing mo-
ments, in particular, repressent considerable fatigue loads for the mechanical ccom-
ponents of the yaw drive. This
T applies especially to hingeless two-bladed rotors. For
this reason, large wind turbbines with two-bladed rotors are usually built with a tee-
tering hub, which more or less
l compensates for these changing loads.
Fig. 6.7. Cyclically changing flaapwise bending moment at the blade root as a consequence oof the
wind shear on the WKA-60 as example
e
6.3.4 Cross Wind on the Rotor

A similar asymmetry of rotor cross wind as that due to the increase in wind speed with
height is produced by the cross wind on the rotor as a consequence of rapid changes in
the wind direction. The relatively inert yaw drive can follow these only with a consider-
able time delay so that the incident airflow exhibits a yaw angle with respect to the rotor
axis at times. In addition, asymmetric inflow conditions for the rotor can be caused by
diverted wind streams in the case of topographically complex terrain or also due to the
rotor design in the case of an inclined rotor axis. This is another reason for keeping the
rotor axis inclination as small as possible although it is required with very flexible rotor
blades in order to guarantee that sufficient free space is maintained between the blade
tips and the tower when maximum deflection occurs.
The mathematical treatment of the rotor in a cross wind with respect to the aerody-
namic forces is not simple. Some useful approaches in this respect are provided by the
blade element theory but these remain valid only up to certain yaw angles. Some possi-
bilities which are better in part are given with the vortex model of the rotor [4]. Figure
6.8 shows the result of a measurement on an experimental installation of the Dutch ECN
Institute. The authors point out that the agreement with the mathematical results is still
acceptable at a yaw angle of up to 200 whilst the measured effect is clearly underesti-
mated by the theory at an angle of 300.
Fig. 6.8. Measured yaw moment of the rotor at various crosswind angles on a Dutch 2-bladed test
rotor in the wind tunnel [5]
6.3.5 Tower Interferencce

The rotor of a horizontal-axiis wind turbine necessarily rotates in close proximity too the
tower. The clearance betweeen the rotor rotational plane and the tower is generally kept
as small as possible in order to limit the length of the nacelle. A nacelle which protruudes
very far causes the rotor forcces to act with great leverage with respect to the tower aaxis.
In any case, however, the disstance between rotor and tower is so small that the aeroody-
namic flow around the towerr influences the rotor.
The influence of the aerod dynamic flow around the tower on the rotor is at a minim mum
when the rotor is mounted in n the traditional position upwind of the tower. The upw wind
rotor is affected merely by a retardation of the flow in front of the tower, the so-caalled
“bow wave” or tower dam effect.
e This tower dam effect was still a considerable faactor
with the old-style windmills and their mill houses, but with today's slender towers it is
only slight. Its effect is stilll perceptible, but the practical effects on rotor loadingg are
slight as long as a minimum clearance between rotor blade and tower of approximaately
one tower diameter is maintaained (Fig. 6.9). However, the tower dam is a possible haz-
ard with respect to the excitaation of tower vibration if the rotor speed remains withinn the
range of the natural bending frequency of the tower for any length of time (Chapt. 111.4).
A completely different problem arises when the rotor is mounted on the down-w wind
side of the tower. This type off design used to be considered to be advantageous in connnec-
tion with the slender towers in n the large first-generation wind turbines.
Fig. 6.9. Flow field due to the tower

t dam ahead of a cylindrical tower with diameter D (poteential
flow theory)
A reduction in flow velocity on the down-wind side of the tower is still percepti-
ble even at a relatively large distance. The rotor blades must pass through this
wind-sheltered area with each revolution. This tower shadow effect represents a
serious problem for the wind turbine in several respects it must be discussed more
extensively.
The aerodynamic influence of the tower has to be considered even in the case of an
upwind rotor. As almost all towers of modem turbines have a circular cross-section,
only the flow around a circular cylinder has to be considered. The internal friction of
the flowing medium and the surface friction (boundary layer) of the body encountered
cause an area of detached flow behind the body, the so-called wake area (Chapt. 5.4).
The wake in the flow behind a circular cylinder consists of a more or less extensive
area of increased turbulence with a considerably decreased mean flow velocity. An-
other typical characteristic of the wake behind a body with a circular cross-section are
the alternating vortices on both sides, occurring with a defined frequency (Karman
vortices). Depending on the Reynolds number of the flow, which is referred to the
cylinder diameter, three characteristic regions can be observed (Figs. 6.10 and 6.11).
Subcritical region
When the Reynolds number is below approximately 3 to 4x105, i.e., at a slow flow ve-
locity, the boundary layer remains laminar. Flow separation takes place ahead of the
widest point of the cylinder cross-section. The flow wake is relatively wide and distinct
Karman vortices occur periodically. Under these conditions, the air drag coefficient of
the circular cylinder is relatively high and equals approximately 1.0.
Supercritical region
At a certain flow velocity, characterised by the so-called "critical Reynolds num-
ber", the boundary-layer flow at the cylinder surface shifts from a laminar to a tur-
bulent condition. This effect influences the shape of the wake considerably. The
high-energy, turbulent boundary layer causes the flow around the body to persist, so
that the flow wake is narrowed. The periodic Karman vortices disappear almost
completely. The drag coefficient is drastically reduced to values of between 0.25
and 0.35. As it is a boundary-layer effect, the point of change is influenced by the
surface roughness of the object.
Transcritical region
Above the critical Reynolds number there follows a "transitional region" where the
flow wake starts to become wider again. In the transcritical region, the drag coeffi-
cient rises to values of approximately 0.5. The Karman vortices again occur peri-
odically, but somewhat more weakly.
A brief estimate of the flow around a tower of a large wind turbine shows that with
tower diameters of several meters and wind speeds of between 5 and 25 m/s, the Rey-
nolds numbers are so high that a turbulent flow can always be expected.
Fig. 6.10. Flow around a circularr cylinder in dependence on the Reynolds number
Fig. 6.11. Air drag coefficient of a circular cylinder in dependence on the Reynolds number [66]
In this region, the maximuum wind speed reduction in the flow wake can be eesti-
mated by the following form
mula:
Δ
1 1
How does the tower shad dow affect the aerodynamics of the rotor? First of all,, the
reduced flow velocity to thee rotor blades as they pass through the tower wake is an
important factor. Reduced wind
w speed goes hand in hand with a change in the effecctive
aerodynamic angle of attack k. Both lead to a sudden decrease in the lift of the rrotor
blade. It affects both the aero
odynamic loading and the torque generated.
This process is of very sh hort duration, corresponding to the rotor speed, and reepre-
sents an impulse-like disturbance at the rotor blade. From the aerodynamic poinnt of
view, this means that transieent aerodynamic effects can play a role which means,, for
example, that the temporal gradient
g of the change in the angle of attack can have a sig-
nificant influence on aerodyn namic forces and moments. While the treatment of transsient
aerodynamic problems is diffficult, it can, however, become necessary when attemppting
a theoretical treatment of the tower shadow problems of a down-wind rotor.
On the other hand, the dissturbance due to the tower wake continues for long enoough
for elastic yielding of the rotor blades to have a damping effect. Thus, the tower wakke is
also a problem of aeroelastiics, i.e. the dynamic response of the rotor blades. Figgures
6.11 and 6.12 show two examples of the effects of the tower wake on a down-w wind
rotor.
Fig. 6.12. Calculated increase in

n the flapwise bending moment at the blade root due to the toower
shadow, using Growian as an ex xample
In the case of a downwin nd rotor the tower shadow contributes to a considerablee ex-
tent to the load spectrum. Th he flapwise bending moment is an important parameterr for
rotor blade dimensioning. Th he influence of the tower shadow is considerable, especiially
considering the high numberr of load cycles of 107 to 108 during the life of the turrbine
(Fig. 6.12). The tower shado ow effect thus becomes a factor which cannot be ignoored
with regard to the fatigue lifee of the rotor blades.
The electric power outputt of down-wind rotors is a clear indicator of the influencce of
tower shadow interference. In I extreme cases, power losses of up to 30 or 40 % beelow
the average output were meaasured (Fig. 6.13). At the usual rotor speeds, the frequeency
of tower shadow interference falls within the range of some of the critical natural fre-
quencies of the turbine, in paarticular that of the drive train (Chapt. 11.2.4).
Last but not least, the inffluence of the tower shadow on the noise generated byy the
wind turbine must also be pointed
p out (Chapt. 15.2.2). This effect turned out to bbe of
such importance that it causeed the virtually complete disappearance of downwind rootors
among today's wind turbines.
Fig. 6.13. Influence of tower shadow

s on the rotor torque at the example of the experim
mental
MOD-0 wind turbine [7]
6.3.6 Wind Turbulence and Gusts

While power output and enerrgy yield of a wind turbine are determined by the long-tterm
variations of the mean wind speed, the non-cyclic fluctuating loads on the wind turrbine
are determined by the short-term fluctuations of the wind speed, the wind turbuleence
and the gusts. The ever-preesent wind turbulence contributes considerably to mateerial
fatigue, particularly of the rotor blades. Extreme wind speeds, though far more rare,
must also be taken into consideration when designing for fatigue strength. Moreover,
they can increase loads up to the point of fracture. The most serious problems as far as
loading is concerned are presented by the stochastic fluctuations of the wind. There is a
number of different "turbulence models" which can be traced back to two fundamental
approaches (see Chapt. 13).
Spectral model of turbulence

The continuous nature of the wind speed fluctuations is reproduced best by means of a
statistical approach in the form of a turbulence spectrum. The spectral representation of
the wind is also used in general meteorology where the most varied turbulence spectra
have been developed. Common turbulence spectra have been produced by Davenport,
Kaimal or von Karman and can be found in the relevant literature (see Chapt. 13.3.4).
In the load calculations it is generally assumed that the turbulence is a one-
dimensional fluctuation of wind speed in the longitudinal direction. In reality, wind
speed fluctuations naturally also have lateral components. Mathematical treatment of a
two-dimensional turbulence model is very difficult, however, and generally not
necessary when dealing with wind turbines. More important, as far as loading on the
wind rotor is concerned, is the spatial distribution of longitudinal turbulence over the
rotor-swept area.
Gust model
While the spectral model of turbulence is of a statistical nature, a deterministic ap-
proach can also be used. The basic idea is to define discrete idealised gust shapes,
represented by increasing and decreasing wind speed over time. These gusts are
then assumed to be discrete isolated events for the calculation of loads. It is obvious
that in the process, the continuous nature of the turbulence is lost. The response of
the structure shows only the reaction to an isolated gust, without taking into considera-
tion the situation before and immediately after the event. For this reason, the gust model
is unsuitable for calculating fatigue strength. Gust models were used only in the initial
phase of wind energy technology, in which, provided with a probability of occurrence, a
load spectrum was assembled from various forms of gusts and used as input for the
fatigue calculation. Present-day calculation methods are based without exception on the
spectral model of turbulence
The significance of these discrete gusts for load calculation primarily lies in the de-
termination of extreme loads. For this purpose, the characteristic properties of the gusts
must be known. Meteorological research has not given much attention to this special
problem so far, so that no adequate data on gust factors, rise and fall times, spatial
extent and similar parameters are available. Attempts to compile data usable for wind
energy technology today were carried out, in particular, by Frost [8]. From such
data, idealised gust shapes have been derived for calculating loads on wind turbines
(Fig. 6.14).
Frost specified gust factorrs as a function of gust duration (Fig. 6.15). They alsoo de-
pend on the level of the meaan wind speed. The higher this is, the smaller are the gust
factors to be expected. The frequency of occurrence is also to be seen in connecction
with the mean wind speed an nd the gust factor (Fig. 6.16).
Figure 6.17 shows the effeect of wind turbulence on the specific dynamic load situaation
of a wind turbine. Bending deflection
d in the rotor blades was initially calculated taaking
into account only the influennce of the cyclic disturbances in the flow caused by w wind
shear, tower influence and simmilar parameters, but ignoring turbulence. Including thee tur-
bulence spectrum, the deflectiion values are almost doubled.
Fig. 6.14. Idealised gust shapes [8]

Fig. 6.15. Gust factors in depend

dence on gust duration [8]
Fig. 6.16. Gust factors in depend

dence on the mean wind speed and occurrence probability [8]
6.4 DESIGN LOAD ASSUMPTIONS 187
Fig. 6.17. Rotor blade bending deflection in flapwise direction, at the HWP-300 wind turbine [9]
6.4 Design Load Assumptions
If the reasons for loads are known, the task is to recognise the conditions during which
the wind turbine is subjected to the essential loads. These conditions and the loads oc-
curring in them are recorded in the form of so-called load assumptions in which both
the external conditions for the causes of the loading, for example the wind speeds and
the turbulence, and the corresponding parameters of the state of the turbine such as rotor
speed or rotor blade pitch angle are established.
It is not apparent from the outset in which load case the dimensioning loading of
different types such as fracture, fatigue or failure of stability will occur. The definition
and systematic arrangement of the load cases must, therefore, be comprehensive
enough to include all operating conditions and additionally also other critical states in
the life cycle of the system which can be expected with a certain probability. This
opens up another problem of the system of load cases: the probability of occurrence. If
one were to define as a load case all conceivable combinations of states and external
conditions, i.e. of the hurricane of the century with simultaneously occurring especially
bad operating conditions, the design process would be carried ad absurdum. The load
cases on which the design is based, and thus, naturally, also the failure of the system,
are turned into a question of statistically based probabilities. Naturally, this applies not
only to wind turbines but to all technical systems and structures.
A further problem consists in that the specification of load cases and the loads to be
stipulated are always associated with a certain idealisation and simplification of a real
state. The loads determined by calculations in the defined load cases are therefore load
assumptions which deviate from the real loads to a certain degree. However, this devia-
tion from reality must always be to "the safe side". In other words, the load assumptions
used for the design must always be somewhat higher than the loads actually to be ex-
pected in operation.
According to their intended purpose, load assumptions should have a general validity
as far as possible, so that they form a generally accepted basis for the design of a sys-
tem, claims which form the basis for the existence of the "Norms" and "Standards"
described in the next chapter. On the other hand, the technical concept of the turbine
influences the nature and extent of the loading to a certain degree so that certain load
cases occur only with certain technical concepts. A mixture of general validity and
individual significance is thus largely unavoidable.
It should also be pointed out that, as a rule, the terms "load cases" and "load assump-
tions" are not clearly delimited with respect to one another. Strictly speaking, the term
"load case" designates the situation in which loads occur. This situation is defined by
the external conditions, on the one hand, and, on the other hand by the machine status
(see Chapt. 6.5). The load assumptions per se are the idealised loads to be stipulated in
the load cases. However, the term load assumptions is often also used as the generic
term for the totality of design loads which are then classified as individual load cases.
6.4.1 International and National Design Standards

First attempts at a systematic definition of load assumptions and load cases for modern
wind turbines had been undertaken in the eighties in connection with the development
of the large experimental Growian and WKA-60 wind turbines [10]. The standards were
initially developed on a national basis but information was exchanged especially under
the auspices of the International Energy Agency (IEA) as early as the beginning of the
eighties [11]. Germanischer Lloyd used this as a basis for building up their set of regu-
lations which represents an important basis for many wind turbine developments until
the present day.
At the same time, similar standards were also produced in the US, in Denmark and in
Sweden. In 1988, the International Electrotechnical Commission (IEC) took over this
task on an international basis.. Today the national rules and regulations have largely
been replaced by the IEC standards, although the national standards are still in existence
since the European standards still do not have unrestricted legal force. Against this
background, a "sideways glance" into the national building regulations is sometimes
still necessary.
Up to now the following regulations have been published by the IEC:
IEC 61400-1 Safety Requirements

IEC 61400-2 Safety Requirements of Small Wind Turbines
IEC 61400-3 Design Requirements for Offshore Wind Turbines
IEC 61400-11 Acoustic Noise Measurement Techniques
IEC 61400-12 Wind Turbine Performance Testing

IEC 61400-121 Power Performance Measurements of Grid Connected Wind Turbines
IEC 61400-13 Measurement of Mechanical Loads
IEC 61400-21 Measurement and Assessment of Power Quality Characteristics of
Grid Connected Wind Turbines
IEC 61400-23 Full Scale Structural Testing of Rotor Blades
IEC 61400-24 Lightning Protection
IEC 61400-25 Communication Standard of Control and Monitoring of Wind Power
Plants
The decisive document with regard to the load assumptions is the IEC 61400-1 standard
"Safety Requirements". The statements contained in the following chapters are, there-
fore, an excerpt from IEC 61400-1.
On the basis of the IEC standards numerous so-called “certification organizations”
offer an independent certification service for wind turbines. They also consider to a
certain extent inhouse or national standards for some aspects of the design or for the
load calculations. The most important classification organizations, which grant certifica-
tions for wind turbines are:
Germanischer Lloyd
The guidelines of the Germanischer Lloyd (GL) are of particular importance in Ger-
many but some other European countries are also working with these guidelines or are
using the certification by the GL. The rules are more detailed than those of the IEC in
some aspects and also comprise rules for the mathematical methods to be used. From
the point of view of content, there are some differences to be noted. For example, the
intensity of turbulence to be assumed, which forms the basis of the load cases, is differ-
ent. Whilst the IEC is assuming turbulence intensities of 15-18 %, the GL guidelines
require 20 % overall. However, these and some other small deviations are about to be
corrected.
Det Norske Veritas

Next to the Germanischer Lloyd, Det Norske Veritas (DNV) is the second international
classification company which has its origin in shipbuilding. The DNV has also issued
"Guidelines for the Design of Wind Turbines". In principle, the same applies as for the
GL guidelines, i.e. there are still some deviations from IEC 61400-1 in detail.
Besides the IEC-standards some national standards are still in force, particularly the
standards in the Netherlands, Denmark and Germany. They play a role in the national
building permissions. Some regulations of the national standards are not included in the
IEC standards.
Dutch NVN Guideline 1400-0

In the Netherlands, there is the NVN guideline for the construction of wind turbines. It
is largely related to IEC 61400-1. There are only a few deviations in the safety factors to
be applied.
Danish Standard DS 472

In Denmark, the national body of rules "Loads and Safety of Wind Turbine Construc-
tion" (DS 472) applies. This, too, largely matches IEC 61400-1. There are still differ-
ences in the definition of wind speed classes (see Chapt. 6.4.2). There are also special,
simplified regulations for the construction of stall-controlled turbines with a rotor di-
ameter of up to 25 m.
German DIBt Guidelines

In Germany, wind turbines are legally graded as "buildings". The technical type approval
and the planning permission are therefore under the jurisdiction of the competent building
authorities. These use the certification companies, e.g. the GL, as expert assessors for the
"technical part" of the building. This procedure, which appears to be somewhat whimsical,
has led to the "Deutsches Institut für Bautechnik (German Institute for Structural Engineer-
ing) (DIBt)" issuing their own guidelines for the construction of wind turbine installations.
Although these relate primarily to the so-called "Proof of Stability for Tower and Founda-
tion", they contain many specifications which relate to the wind turbine as a whole. By
now, these relate largely to the IEC Standard but there are still differences, primarily the so-
called "wind zones" which are not identical with the wind speed values in the "wind turbine
classes" according to IEC (see Chapt. 6.4.2).
Certification business
The "certification" of wind turbines has become an extensive branch of the trade which
requires some relevant critical remarks with which not everyone may agree. The
independent verification, i.e. testing of the structure with respect to safety, is quite indis-
putably in the public's interest and is therefore a "must". Checking the performance
characteristic by manufacturer-independent neutral institutes or expert assessors also
makes sense. In Germany, this check is not a subject for the type approval. The purchaser
should be aware of this fact and in his own interest request to see the certificates for the
performance characteristic.
The newly developed fields of wind energy utilization such as, for example, the off-
shore siting of wind turbines, make completely new demands on the design and the
technical equipment. The development of the design standards and the certification must
keep up with these developments which makes it unavoidable that the latest and more
extensive sets of rules must always be available.
On the other hand, the certification activities are increasingly spreading to almost all
aspects of wind energy utilisation, following a trend to have every characteristic of a
product or even of an associated action like the investment in a wind farm authenticated
and certified by independent experts. Against this background, extensive test procedures
were developed for wind turbines which are reflected in the most varied test certificates.
From the environmental conditions through the load assumptions, from the construction
to the surveying of wind turbine installations, all these fields are the subject of certifica-
tions. Without any claim for completeness,
these are:
- Production methods,
- Quality assurance,
- Test procedures,
- Transportation,
- Erection and assembly,
- Commissioning,
- Maintenance regulations,
- Operational sequences,
- Quality of power output,
- Electrical characteristics and grid compatibility et al.
Most of the organisations offering certifications are profit-oriented commercial under-

takings. For this reason, they will attempt to extend their services to all possible areas.
However, in many areas it is more than doubtful whether a "certificate", for example for
"production" or for "transportation", has any objective use. Neither is the situation im-
proved by the fact that the organisations have for some years been advertising so-called
"accreditations" which, in turn, are issued by private commercial organisations. This
practice, too, and the standards of the authorities which extend over ever more new
fields, serve the general need of the consumers for ever more safeguards and less per-
sonal responsibility.
As ever, the decisive criterion for the quality of a product is the technical competence
and financial capacity of its manufacturer. It is only the manufacturer who really bears
responsibility for his product through his warranty. If something goes wrong, the finan-
cial consequences will hit the manufacturer and his customer, never the certification
organisations. Certificates, however attractive-sounding, are therefore no substitute for
trust in the manufacturers and their product.
6.4.2 IEC Classes of Wind Turbines and German Wind Zones

The structural dimensioning of a wind turbine is determined to a considerable extent by
the wind conditions at the intended site. The demands on structural strength are much
lower on sites with relatively low wind speeds and little turbulence than on sites with
high average wind speeds and corresponding air turbulence. The turbulence can also be
increased by particular forms of terrain (see Chapt. 13.4). For this reason, the load as-
sumptions are adapted to the different wind conditions. It would not make any economic
sense to design the structural strength sweepingly for the highest wind speeds and on-site
conditions occurring.
IEC 64100-1 defines fourr different classes of wind conditions called "Wind Turrbine
Generator System Classes" (WTGS ( Classes). The wind conditions are defined byy the
extreme wind speeds to be assumed
a and the average wind speed at the site, and byy the
turbulence intensity. The extrreme wind speeds are aimed at ensuring that adequate brreak-
ing strength and stability is provided
p whilst the average wind speed is of importancee for
the service strength, i.e. the fatigue
f strength, of the material. In this context, a designn life
of at least twenty years must be
b used as a basis.
A so-called reference wind d velocity is used for identifying the different classes. Thhis is
the maximum wind speed wh hich, seen statistically, is exceeded only once in 50 yearss and
is measured as a ten-minute mean value. The short-term extreme wind speeds withhin a
3-second period, i.e. the maximum gusts to be expected, are derived from this value. The
wind turbulence is assumed to t be characterised in two categories A and B. The stanndard
deviation of the longitudinal wind velocity change in the 10-minute mean values is sppeci-
fied by the parameter α (Tablle 6.18). Apart from these four classes, there is also a speecial
class S for special site condittions. The values to be assumed in this class must be aggreed
individually with the licencing g authorities. In addition, the Germanische Lloyd has deffined
a new "Off-shore" class which h has been adopted by the IEC [12].
Table 6.18. WTGS Classes to IE

EC 64100-1
It should be pointed out th hat in conjunction with the IEC Standard, the indices deffined
there are used for identifying the parameters. Otherwise, the older indices still in com
mmon
use in Germany and used in the other chapters would have to be continuously comppared
with the designations of the IE
EC Standard and confusion is preprogrammed.
There is another special feature
f to be noted in Germany which, although it is oof no
importance internationally, will
w still play a role in Germany for the foreseeable futture.
The Deutsches Institut für Bautechnik
B has also introduced a classification in its Guuide-
lines for Wind Turbines whicch contain the so-called Proof of Safety for tower and fooun-
dation [2]. The wind condittions are graded into so-called Wind Zones. The prooof of
stability is, therefore, checkeed in accordance with the defined wind zones as part off the
building permit procedure. On O the other hand, the manufacturers of wind turbiness al-
most exclusively use the IEC C classes as a basis for the technical engineering part. For
this reason, there is frequently a problem of compatibility of the DIBt wind zones with
the wind turbine classes according to IEC.
The DIBt defines the following wind zones and specifies the reference wind speed at
10 m height for these (not at hub height as with the IEC):
- Wind Zone I: Weak-wind regions (24.3 m/s)

- Wind Zone II: Normal German inland site (27.6 m/s)
- Wind Zone III: Typical coastal site (32.0 m/s)
- Wind Zone IV: North Frisian islands (36.8 m/s)
The IEC class 1 forming the basis for the design of wind turbines corresponds approxi-
mately to Wind Zone IV, IEC Class 2 corresponds to Wind Zone III and IEC Class 3
can be equated with Wind Zone II. The precise differences are deliberately not taken
into consideration here since it is hoped that they will soon be a thing of the past in the
course of the on-going harmonisation of standards in Europe as well as internationally.
6.4.3 Normal Wind Conditions

The so-called "normal wind conditions" reflect the wind conditions occurring "fre-
quently", i.e. frequently in the course of a year, during the operation of a wind turbine.
They are characterized by:
Mean annual wind speed

The mean annual wind speed at rotor hub height, measured as a 10-minute mean value,
is the most important parameter for characterizing the wind conditions. With respect to
the fatigue strength, it is not only the fluctuation of the instantaneous wind speed about
the annual mean value which plays a role but also the longer-term variations of the
mean wind speed. Although their associated frequency of load cycles is lower by sev-
eral orders of magnitude, they still have a certain influence on the fatigue strength. Seen
from the point of view of the total load spectrum, they represent transitions from one
wind speed class to another. They can be interpreted as long-wave, periodic oscillations
with large amplitudes.
Wind speed frequency distributions

The wind speed frequency distribution is assumed to be a Rayleigh distribution
(Weibull distribution with a form factor of k = 2).
Vertical wind shear

The vertical wind shear indicates the average change in the mean wind speed with
height. It is accounted for by Hellmann's power law with the exponent used being α =
0.20. The vertical profile of the wind speed is of considerable significance, for example,
for the flapwise bending moment if the rotor blades. The high numbers of load cycles
are the result of the speed of rotation of the rotor and are correspondingly high with a
stipulated rotor design life of 20 years.
Wind turbulence
Apart from the cyclic loads resulting from the dead weight of the components and the
asymmetrical flows on the rotor, wind turbulence is the second decisive factor for the
fatigue strength. According to IEC 64100-1 the "normal turbulence" of the wind is de-
scribed by the standard deviation:
· 15 ⁄ 1
The parameters depending on the wind turbine class are contained in Table 6.18.
6.4.4 Extreme Wind Conditions

To determine the maximum loads on the wind turbine, "extreme wind conditions" must
be stipulated. These comprise short-time extreme wind speeds and also the loads result-
ing from extreme changes in the wind direction and the vertical shear in the wind speed,
and certain combinations of these influences. The continual nature of the turbulence is
covered with the statistical approach to turbulence during normal wind conditions so
that the extreme events can be assumed deterministically to be single events.
Extreme wind speeds and gusts

The short-term extreme wind speeds averaged over 3 seconds (gusts) are derived from
the reference wind speed. The so-called 50-year gust, the highest wind speed, which is
exceeded only once within a period of 50 years, is obtained from:
.
1.4 ·
The so-called annual gust is:

0.75 ·
where z is the hub height.

It is assumed that the annual gust only takes place in combination with a limited devia-
tion of the wind speed direction, this means short-term deviations of of ±l50.
Gusts occurring with increased frequency in operation are called "extreme operating
gusts". They are calculated from the turbulence model in dependence on the rotor diame-
ter according to IEC.
To account for the turbulence distributed over the rotor-swept area and unevenly, a
so-called “coherent turbulence function” is also stipulated according to the IEC stan-
dard. The extreme coherent gust is to be assumed to have a speed of 15 m/s. Beyond
that, it is to be combined with a uniform change in wind direction within 10 seconds.
In some older national standards such as, e.g. at the Germanischer Lloyd, the gusts
are also still calculated with a gust factor. The positive gust is defined as:
Whereas the negative gust is:
The gust factor is taken into consideration by:
The gust amplitudes exceeded with a probability of once annually ("normal operating
gust") are defined as 9 m/s. Gusts which are exceeded only once in 50 years ("extreme
operating gust") are defined to have an amplitude of 13 m/s.
Extreme changes in wind direction

Extreme changes in wind direction within a range of seconds can cause extraordinary
loads. According to IEC 64100-1, extreme changes in wind direction within 6 seconds
are calculated in dependence on the standard deviation of the turbulence and on the
rotor diameter. They are to be specified as annual and 50-year events.
Extreme crosswind
The load assumptions must take into account an extreme crosswind profile as a 50-year
event. This requires the assumption of an asymmetric rotor inflow produced within a
period of 12 seconds both in the vertical direction (vertical wind shear) and in the hori-
zontal direction. The shear to be assumed is calculated in dependence on the turbulence,
the rotor diameter and the rotor hub height.
6.4.5 Other Climatic and Environmental lnfluences

Climatic parameters and environmental influences other than wind can affect the loads
acting on a wind turbine. The limits are specified in the IEC 64100-1. The essential
impacts are:
Temperature range
The verifications of strength should be carried out for a temperature range of -10°C to
+40°C, In the case of special operating conditions (e.g. "Arctic climate"), the appropri-
ate individual verifications must be made.
Air density
The calculation of aerodynamic loads is based on the assumption of the air density of
the standard atmosphere (at sea level):
1.225 /
Relative humidity
The humidity of the air is of less influence under normal atmospheric conditions. Ac-
cording to IEC a value of 95% is acceptable.
Solar radiation
The solar radiation is assumed to be 1000 W/m² (Central European conditions).
Ice accretion
One of the environmental factors which may contribute to extraordinary loads is the
build-up of ice on the rotor blades. But even thick ice formations on the rotor blades do
not cause the danger of severe failure loads. Similar to aircraft wings, aerodynamic lift
is reduced, with the consequence that rotor performance is reduced, and with it the
aerodynamic loading (see Chapt. 18.8.2). The load assumptions distinguish between
rotating parts (rotor) and non-rotating parts. For the non-rotating parts, an ice accretion
of 30 mm is assumed. For the rotor blades, a varying mass distribution of the accreted
ice from the root of the blade to its tip is assumed, and a difference in ice accretion
between the individual blades [2].
Salt content of the air

The salt content of the air requires special design features in the area of cooling and
ventilation and, naturally, special methods for the surface treatment of steel compo-
nents. This is particularly true in the off-shore region.
Bird strike
One load case which is fortunately very rare can be caused by a large bird colliding with
the rotating rotor. To take this hazard into consideration, the earlier Swedish load as-
sumptions suggested some assumptions about impact velocity and bird weight [13]. The
resulting impact may be of significance for the dimensioning of the rotor blade shell.
Orographie influences
The influence of the orographic situation on the wind speeds (wind flow over hills and
mountains) must be taken into consideration above a certain, predetermined influencing
quantity.
Earthquakes
For installations in hazardous regions with a risk of earthquakes, the local building
regulations concerning earthquake protection must be consulted.
6.4.6 Other External Conditions

In order to provide a complete overview of all conceivable loads, the load assumptions
require the consideration of other external conditions. In addition, the entire life cycle
from the assembly of the plant through its erection at the site up to its operating and re-
pair conditions must be taken into account. With respect to its operation, in particular, the
following external conditions must be noted:
Influences from the grid

IEC 64100-1 mentions the most important electrical parameters such as voltage, fre-
quency and the shut-down characteristic of the turbine with the associated tolerances
with regard to the presence of any loads from the grid. If these tolerances are exceeded,
the presence of special loads having their causes in the electricity grid cannot be ruled
out. At this point, reference should also be made to the grid connection regulations of
the power companies (see Chapt. 10.5.2). Variable-speed installations with their "soft"
grid coupling are largely protected against loads from the electricity grid.
Influence of adjacent wind turbines

The turbulence intensity is assumed to have values of 16 and 18% at rotor hub height in
the IEC load assumptions. But it must be taken into consideration in this context that
when wind turbines are erected in close proximity to one another, i.e. in a wind farm,
the turbulence intensity is increased in the field (see Chapt. 18.3). Although the 02/1999
issue of IEC 64100-1 does not yet contain any specifications for taking these influences
into account, an appropriate approach is proposed in the DIBt Guideline according to
which no verification is required if the turbines are spaced apart by more than eight
times the rotor diameter. If the distance is less, as it is in many wind farm installations, a
method for calculating the increased turbulence intensity is specified [2]. It is common
practice to perform an individual turbulence study for larger wind park installations, in
some countries it is required in the building permission.
Other influences
IEC 64100-1 and the DIBt Guideline both contain a number of further notes regarding
any additional loads to be taken into consideration under certain circumstances. Without
claiming completeness, examples of these are:
- Wind loads during assembly and repair

- Loads from uneven mass distribution (rotor)
- Pressure of soil and grounnd water on the foundation

- Structural inaccuracies in tower and foundation.
These items are not covered d by any generally valid standards, reference being m
made
instead to the need for individ
dual consideration.
6.4.7 Safety Factors

Safety factors have the purpo ose of compensating for inaccuracies in the load assumpttions
and calculating methods, strructural inaccuracies and, not lastly, for deviations inn the
actual strength values from th he specified material characteristics. The safety factor iss the
ratio of the design value to the
t calculated or specified value. With respect to the looads,
the design value is the result of multiplying the calculated value by the safety factor tto be
assumed. With regard to thee material parameters, the design value is formed from m the
specified value divided by th he safety factor. Defining meaningful safety factors initiially
requires a classification of th
he consequences of the failure of a component. Accordinng to
IEC 64100-1, the componen nts are graded in two classes with respect to their "paartial
safety factors" or "potential safety
s factors":
- "Fail-safe" components, the failure of which is absorbed by a safety system aand
does not lead to any severre damage on the wind turbine,
- "Non-fail-safe" componen
nts, the failure of which leads to severe damage.
In addition, higher safety facctors are defined for the limit loads (fracture/failure off sta-
bility, critical deformations) than for the fatigue loads. Table 6.19 contains the saafety
factors for the design loads, to
t be specified according to IEC 64100-1:
Table 6.19. Partial safety factorrs for design loads according to lEC 64100-1 (simplified)
The safety factors for the material parameters depend on the type of material, onn the
one hand, and, on the otherr hand, on the type of loading. They must be specifieed in
close relationship with the material
m standards. It is necessary to observe the natiional
standards such as, e.g. DIN 18 800 for structural steel or DIN 1045-1 for reinfoorced
concrete. IEC 64100-1 generrally recommends a value of not less than 1.1 for the m main
strength values but values of 1.0 are also permissible for the fatigue strength.
When defining the safety factors
f it may be tempting to compensate for a less accuurate
calculation by applying high h safety factors. According to all previous experience, this
strategy will lead to success only in small installations, if at all. The larger plants m
must
6.5 OPERATIONAL STATUS AND
A LOAD CASES 199
be loaded up to the limits of the strength of the material or of the component stiffneess if
it is intended to keep the buiilding masses within tolerable limits. High building maasses
decrease the stiffnesses and increase the forces of gravity. The attempt to compennsate
for the associated increased loads by using higher safety factors, in turn creates a vi-
cious circle at the end of wh hich the structural safety is lower instead of higher. Inn the
case of dynamically highly sttressed systems, which undoubtedly includes wind turbiines,
the structural safety can lasttly be achieved only by greater understanding of the looads
involved and of the structural dynamics (s.a. Chapt. 6.7.2).
6.5 Operational Statu

us and Load Cases
For a machine with moving parts,
p the condition of the loaded system is a decisive ccrite-
rion for the stresses occurrinng during the loads acting on it from the outside. The load
cases in the narrower sense are, therefore, produced by connecting the external coondi-
tions with the operating statees of the wind turbine. In this context, it is not only the ooper-
ating states per se which musst be considered but also other states which may occur wiithin
the entire life cycle of the pro
oduct. States which become special "load cases" may ooccur
during transport and assembly, and possibly also during large repair processes. Onn the
other hand it would be nonseensical, however, to produce such states which then beccome
the dimension-determining lo oad case for the components. For economic reasons, the load
cases should be restricted to those which are absolutely necessary. These are the
operating conditions and the abnormal occurrences which can be expected with a cerrtain
probability.
The basic structure of thee link between external loads and operating conditions and
the structural stresses to be ex
xpected in it becomes clear in Table 6.20.
Table 6.20. Basic load case systtem and types of stress in wind turbines
6.5.1 Normal Operation

n
The loads to which the wind d turbine is subjected under "normal" operating condittions
are mainly relevant to fatigu
ue life. The main causes are the high load cycle numberrs of
the alternating bending mom ment resulting from the dead weight of the rotor blaades,
occurring with each rotation n of the rotor, the asymmetric aerodynamic forces accting
cyclically through the rotation of the rotor and the ever-present wind turbulence. The
loads are therefore largely covered with respect to fatigue strength by the load cases in
normal operation. The individual load cases are distinguished by the operating cycle of
the turbine. The starting point for defining the load cases is the frequency distribution of
the wind speed on which the design is based, and the operating cycle of the turbine.
Power production
The range of wind speeds within which the turbine is operated is divided into "classes",
each of which is characterised by a characteristic wind speed:
- cut-in wind speed

- partial-load wind speed
- rated wind speed
- full-load wind speed
- cut-out wind speed.
For each of these characteristic wind speeds a load case group is formed. The associated
number of load alternations is derived from the fraction of time each wind speed class
occupies within the wind speed distribution and from the number of rotor revolutions in
these time segments. Considering the turbine's design life, this implies, for example, 107
to 108 load cycles for the bending stress of the rotor blades.
The asymmetrical flow conditions for the rotor and the random wind speed fluctua-
tions and the loads resulting from any malfunctions are added to this "basic load spec-
trum". Having regard to the fatigue strength, the influence of the flow around the tower
must not be forgotten. The loading by the flow around the tower occurs with the number
of load cycles of the rotor revolutions during the life of the turbine for each individual
rotor blade. For the total rotor force, this number is multiplied by the number of rotor
blades.
Start-up and shut-down of the rotor

Rotor start-up and shut-down involve special load cases and load changes. These events
occur so frequently during the life of the turbine that it must be assumed that they have
an influence on fatigue life. They do in fact also represent a group of load cases, as
different starting conditions with regard to wind speed, rotor speed or even blade pitch
angle must be considered. When the rotor in wind turbines with pitch control starts up,
the rotor blade pitch angle is either in the feathered position or in the starting position.In
both cases, a more or less large component of the bending moment acts around the
softer flapwise axis due to the inherent weight of the blades. If it occurs often enough,
this special load case can be significant to the overall fatigue loading.
In larger wind turbines, the normal shut-down of the rotor is controlled by means of
blade pitch control as the rotational speed varies, so that no special loads are involved.
One exception is fast braking, the "emergency shut-down", where the reversed aerody-
namic thrust can cause increased loads.
6.5 OPERATIONAL STATUS AND LOAD CASES 201
Parked rotor at extreme wind speeds

It is generally when the rotor is parked that the wind turbine has to cope with the highest
wind speed, the so-called survival wind speed. For turbines with pitch control it is as-
sumed that the rotor blades are in the feathered position and that the rotor is aligned
with the wind. Under these conditions, the load level is much lower than under cross-
wind conditions. Naturally, the precondition for this is that the yawing system and the
blade pitch control are functional when the survival wind speed occurs.
In the case of fixed-blade small turbines, this problem does not occur. The strength of
the turbine must be verified with rotor blades under cross-wind conditions. The assump-
tion of a correct drag coefficient is of essential significance here. For blades subjected to
a cross-wind the cD-values range from 1.3 to 1.8.
Some wind turbines, e.g. Enercon turbines, do not stop the rotor at extreme wind
speeds. The turbine continues to run with a particular blade pitch angle and at a reduced
speed. In this state the loads on the turbine are no greater than with the rotor blocked. In
addition, it considered to be an advantage that when the wind drops, the turbine can
continue to run again at full power within a very short time without loss of energy due
to cutting-in and cutting-out processes.
6.5.2 Technical Faults

Technical faults and defects can subject the wind turbine to additional loads not covered
by the other load cases. It can be assumed that most technical defects, in as much as
they are relevant to operational reliability, lead to an emergency stop of the rotor via a
safety system, so that these types of defects do not result in any "extraordinary" loads.
On the other hand, some malfunctions are possible which cause abnormal loads on the
structure before the rotor shuts down. These events must be recognised and included in
the definition of load cases. Thus, a theoretical failure mode and effects analysis should
be carried out for reliability-related areas of operation, such as blade pitch control and
the rotor brake systems, at least in larger turbines.
Rotor emergency stop

Most technical failures will trigger the rotor emergency stop via the safety circuit, but
the abrupt deceleration of the rotor results in an extraordinary loading situation for the
wind turbine. With large rotors and under certain circumstances, this situation can in-
crease the bending stress on the rotor blades up to the strength limit. In the case of a
defect, for example a loss of the electrical system (generator release) or a fault in the
control system, the rotor blades must be pitched very rapidly towards feathering to pre-
vent rotor "runaway". For this, the blade is pitched so fast that, for a short period of
time, the rotor blades are subjected to negative aerodynamic angles of attack. The aero-
dynamic thrust then acts in the opposite direction. If the rotor blades are positioned at a
cone angle to each other, the bending moments from thrust and centrifugal force are
superimposed in the same direction. Instead of compensating for each other as in normal
operation, they add up with the consequence of an extreme bending moment on the
rotor blades. A very careful analysis and optimisation of the emergency shut-
down procedure is required in order to remain within the given load limits under these
conditions.
Control system fault

In turbines with blade pitch control, a failure in the control system can lead to a pitch
angle which is inappropriate for the operating condition. This is directly associated with
special aerodynamic loads and indirectly with other consequences, such as rotor over-
speed.
A fault in the control system or in the yaw drive can result in extreme cross-winds
acting on the rotor. Loads due to extreme cross-wind angles or yaw angles must, there-
fore, be seen not only against the background of extreme meteorological conditions but
also with a view to technical defects.
Generator short-circuit
A short circuit in the electrical generator causes an extreme load on the drive train. The
generator short-circuit torque can amount to up to seven times the value of the rated
torque (Chapt. 9.1). However, the maximum torque is limited to a low value by the
coupling existing between the gearbox and the generator, usually to 3- to 4-times the
nominal torque (s.a. Chapt. 8.9).
Rotor overspeed
Rotor overspeed in most cases is a consequence of other faults. Defects in the blade
pitch control or a sudden loss of the electrical load, for example in case of a power sys-
tem shut-down, can cause the operating speed of the rotor to be exceeded. Rotor "run-
away" is basically the most severe safety hazard in a wind turbine (Chapt. 14.7).
A large enough safety margin between the permissible operating speed and the
"maximum speed before fracture" is therefore required. For large wind turbines this
margin is relatively small, in the order of 50 % of the rated speed. A greater safety mar-
gin would increase the mass of the rotor blades to a non-economical limit. Therefore the
reliability of the rotor blade feathering system is very important to ensure the stop of the
rotor in a case of emergency.
Rotor unbalance
The mass balance of the rotor blades has to be checked very carefully in the manufac-
turing process. However defects during operation for example in the case of damage to
the rotor blades, loss of a structural part or the formation of ice on the rotor blades, an
unbalance of the running rotor must be expected until the rotor stops.
Therefore, a certain unbalance mass must be assumed, the magnitude of which must
be related to the size of the rotor. The resultant load case must be verified with respect
to strength as well as any vibration problems which may be present.
6.6 STRUCTURAL STRESSES IN THE WIND TURBINE 203
Other situations
Besides the technical faults mentioned other situations can occur, which can cause extra
loads. The IEC standards and the regulations of DIBt include further references to those
events.
6.6 Structural Stresses in the Wind Turbine

The external loads acting on a system, and possibly also internal states of stresses, are
transformed into material stresses through the dimensioning of the components. The
required, load-bearing material cross-sections are then decided by the permissible mate-
rial parameters. In the previous chapters the external loads and the circumstances where
the loads act on the wind turbine were outlined. The logical next question is what
stresses, deformations and stability problems the loads cause in the structure and the
components of the wind turbine.
6.6.1 Kind of Stressing

In principle the structure and the components of a wind turbine must be dimensioned
with regard to their loading with the following criteria:
- Ultimate limit states of carrying of strength and stability

These are understood to be limit loads which lead to failure of the strength of the
structure. They are primarily loads which trigger tensile stress, compressive strain and
bending stress. Apart from this, there is also the so-called failure of stability in the
form of buckling, for instance in thin-walled tubular steel towers, and kink in long,
slender mechanical transmission elements, and, not lastly, the tilting of the entire sys-
tem.
- Material fatigue
Persistent, alternating stresses will lead to a fracture of material only after a certain
period of time. Apart from the magnitude of the load, the number of load cycles
within a given period, the design life, is of decisive importance here. The so-called
Proof of Service Strength must be provided for this period.
- Limit states of operating capability
Even if there is no structural failure, unacceptably large deformations of the compo-
nents may restrict, or render impossible, the operating capability of the system. De-
formation of the rotor blades is a typical example: it can become so great that the
blade tips have no more clearance from the tower.
There is no a priori answer to the question of which of these requirements determines
the dimensioning of the various components of a wind turbine. It depends on the con-
ceptual design of the wind turbine and the design of its components. Nevertheless, for a
common design there are experiences which can serve as reference values (Table 6.21).
Table 6.21. Design drivers for th

he main wind turbine components [14]
In principle, the structuraal dimensioning of wind turbines is treated mathematiccally

by the same methods as thosse known also from other fields of technology. There arre no
special features with regard tot the materials used. The treatment of highly loaded, fiibre-
reinforced composite materiaals as used for rotor blades, for example, is "state of thee art"
today.
Calculating the breaking strength or the stability limits in the case of single, nnon-
recurring loading is a traditioonal task related to strength. They are considered as quuasi-
static load cases and can therrefore be treated by using comparatively simple calculaation
methods. They do not, thereffore, require any further explanations in this book. The iissue
of fatigue strength representts another problem entirely, which requires "wind-turbbine-
specific" knowledge and willl, therefore, be discussed in more detail in the next chapter.
The material fatigue due to persistent, alternating loads is determined by two deciisive
factors: the external loads accting on the system within the predetermined life cycle and
the so-called dynamic respon nse characteristic of the structure with respect to the aalter-
T resultant structural dynamics have a great influencee on
nating external conditions. The
the fatigue strength. Basicallly, the following applies: in a stiff system, all unsteeady
loads, i.e. the loads changing
g with time, in particular, must be sustained by the structture,
with the consequence of rapiid material fatigue. In a soft, elastic system, the alternaating
loads cause movements of th he components and are thus absorbed by the inertial foorces
of the accelerated masses, th he material fatigue being much less under these conditiions.
Calculating fatigue strength in wind turbines presents special requirements for two rea-
sons:
With a design life of 20 years,
y the alternating loads, for example the cyclic bendding
stress in the rotor blades, reesulting from their dead weight during the rotation off the
rotor, but also alternating sto
ochastic loads from the turbulence of the wind, lead too ex-
traordinarily high cyclic load d numbers. In comparison with other systems, these arre at
their upper limit (Fig. 6.22).. For some materials such as, for example, fibre-reinforced
composite materials, but also for the welded seams of steel structures, cyclic load
numbers of 107 to 108 are at the limit of current experience. These uncertainties must be
compensated for by particularly fatigue-resistant designs or by higher safety factors.
Fig. 6.22. Permissible alternating stresses and load cycle numbers of various systems [14]
The second issue is the alternating loads resulting from the wind turbulence. Map-
ping these loads correctly by means of a suitable turbulence model requires a profound
understanding of the characteristics of wind. There are no relevant suitably comprehen-
sive models from other fields of technology in existence. Although the influence of
wind turbulence on material fatigue is known also in aeronautical engineering or in the
aerodynamics of tall, slender buildings, the conditions are quite different in both fields.
For this reason, the methods developed here can be translated to wind power technology
only with regard to the basic assumptions.
6.6.2 Load Spectra

In simple stress situations with static loads, it is sufficient to calculate the structural
strength separately for individual load situations or load cases. If a safe fatigue life is
required with alternating loads, elementary fatigue strength theory assumes that stress
fluctuations occur with constant amplitudes within the lifetime of a component. If the
stress amplitudes are below the fatigue strength of the material, then the number of load
cycles no longer plays a role, i.e. changes in load can be endured any number of times.
If the stress amplitudes are higher than the fatigue strength allows, only a certain num-
ber of load fluctuations can be sustained, i.e. the material is only "fatigue-limited". In
the case of steel, this mechanism is represented by the well-known "Wöhler-line". This
fatigue model has been found useful for "normal" engineering problems.
The elementary theory is no longer adequate for designing for the fatigue strength of
dynamically highly stressed systems such as aircraft, automobiles and wind turbines.
The load spectrum with regard to material fatigue consists of periodic and stochastic
stress fluctuations, with varying mean values and fluctuations. The single stress situa-
tions can no longer be considered independently, but must be assessed in their totality,
as a load spectrum. Against this background, calculating the endurance strength re-
quires more complex models which can also be summarised under the title of damage
accumulation.
The load spectrum summarises the stress situation of a component over its entire life
in an idealised form. The load sequence within an operating cycle of the wind turbine,
which the component has passed through a certain number of times within its life, forms
the basis for the load spectrum. The progression of the cyclic bending moment experi-
enced by the rotor blades of a wind turbine in the individual load cases serves as an
example (Fig. 6.23).
Fig. 6.23. Idealised fatigue load sequence of bending stress in chordwise direction of a rotor blade
According to the load case definition, operation under load is composed of five load
cases, each characterised by a certain wind speed. During the start-up and shut-down
process, the rotor blades are subjected to a higher load level as the blade pitch angle is
set such that the dead weight bends around the flapwise axis. This condition has been
assumed as occurring with a certain frequency in the lifetime of the wind turbine.
In full-load operation the amplitudes of the bending moment around the chordwise
axis are determined primarily by the dead weight of the rotor blades. The influence of
wind shear can also be recognised clearly but it must be noted that an extreme wind
shear has been assumed. The effects of wind turbulence on the chordwise bending
moment are only slight. It is the flapwise bending moment that is primarily affected.
Naturally, the weight of the rotor blades in relation to the aerodynamic forces plays a
decisive role in this.
Apart from the alternating amplitudes in the single load cases, the transitions from
one load case to the next also play a role. Seen from the collective load point of view,
these transitions result in ad

dditional stress amplitudes. In this example the maxim mum
amplitude excursion is evideently created by the transition from operation at rated poower
to shut-down. To each load case, load cycle numbers are assigned which are deduuced
from assumptions about the frequency of occurrence of the operating cycle withinn the
life span and from the propo ortion of time occupied by each load case in the operaating
sequence. They range from around
a 104 for the occurrence of rare events such as extrreme
8
wind shear, up to around 10 for the alternating bending load at partial-load wind speeed.
Fig. 6.24. Measured stress ampllitudes (without mean value) against load cycle number meassured
on the rotor blades of the WKA--60 [15]
Figure 6.24 shows an exam mple of a measured load or stress spectrum as it is normmally
represented. The material streess measured at the rotor blades of the WKA-60 turbinee has
been plotted against load cy ycle numbers. In the form shown, the load spectrum iss re-
ferred to the period of one ho
our and, therefore, be projected to a service life of 30 yeears,
corresponding to about 108 load cycles, for assessing the fatigue situation. The two
curves for the flapwise and chordwise components of the bending stress show deearly
that the chordwise bending component
c is almost exclusively determined by the consstant
amplitude of the gravitationaal loading, whereas the flapwise component is determ mined
by the aerodynamic loading anda hence has a varying amplitude.
Such load or stress spectrra must be prepared theoretically for every structural ccom-
ponent subjected to dynamicc stress, but in any case for the rotor blades and hub,, the
main shaft and gear box and possibly also for the highly stressed components of the yaw
system. On the other hand, the
t computation work must be restricted to an amount that
makes economic sense. In th he case of small systems, therefore, simplified methods will
be used.
6.7 Mathematical Models of Structural Dynamics
The elastic components of the structure can be excited into oscillations by external
forces. If the excitation frequency is the same as the natural frequency of the compo-
nents, resonances will occur. The oscillation is continuously supplied with energy which
leads to it building up, an effect which can lead to the complete destruction of the struc-
ture. The oscillations are only counteracted by the so-called structural damping which,
as a rule, is very small, and possibly by some external damping such as, e.g. the aerody-
namic damping with rotor blade oscillations in the flapwise direction. Freedom from
resonances is thus an important criterion in the design of the structures and components.
This set of problems will be discussed in greater detail in Chapter 8 "Vibration Prob-
lems".
As far as the fatigue strength is concerned, the predominant factors are the absorption
of the external forces by a "yielding" of the structures and the conversion of energy into
the inert masses of the components moved. This phenomenon considerably reduces the
stress levels of the structure, i.e. the material stresses.
6.7.1 Functional and Structural Modelling of the Wind Turbine

One problem of structural dynamics consists in the fact that stress can be calculated
only in a coherent mathematical model on the basis of the excitation, e.g. the turbulence,
via the aerodynamic behaviour of the rotor and the power and speed control of the wind
turbine and including the elastic properties of the stressed components. Developing such
a complex model composed of several part-models is an important step in the calcula-
tion of fatigue strength. However, this also poses the first risks. The quality of the re-
sults is decided to a large extent by the simplifications and assumptions which must
necessarily be made in this context.
Aerodynamic rotor model

The calculation of aerodynamic loading, both due to the steady-state flow against the rotor
as well as from wind turbulence, requires an aerodynamic rotor model. Blade element
theory, as outlined in Chapter 5, is a suitable instrument for aerodynamic loads from a
steady-state wind flow. Dynamic loads caused by wind turbulence and the elastic response
of the structure can be calculated by means of a simplified aerodynamic model. A linear
analytical approach for the dependence of the aerodynamic force coefficients on the angle
of attack is often sufficient (Chapt. 6.5.2).
Elastic structure model

Theoretical tools for calculating elastic structures are currently in use in many areas of
mechanical engineering. They are based almost without exception on the finite-element
model, with the aid of which the natural frequencies and properties of the structural
components can be calculated. Knowing the natural frequencies, the dynamic responses
6.7 MATHEMATICAL MODELS OF STRUCTURAL DYNAMICS 209
(deformations, accelerations, stresses) under the influence of external forces can then be
calculated. The computer programs based on this can also be applied to the components
of wind turbines.
It is of little help to proceed on the basically correct assumption, that the dynamic re-
sponse, and with it the stresses, can only be calculated correctly if the elastic character-
istics of the entire turbine are taken into consideration. An elastic structural model of the
entire turbine would inevitably require an enormous computational effort resulting in a
corresponding amount of data, with the associated risk of missing the critical points.
It is therefore important to use a good measure of feel for how the components are dy-
namically coupled, in order to define subsystems with the aid of which the significant
loads can be calculated. In most cases, for example, the rotor blade vibration can be con-
sidered in isolation, possibly coupled with the behaviour of the drive train in the case of
the lead-lag vibration. The rotor/tower system can also be generally considered in isola-
tion with regard to the bending vibration of the tower.
Functional model of the power control

If the wind turbine is equipped with blade pitch control and a variable-speed rotor, its
functional behaviour has an influence on the loading. An algorithm for the blade pitch
and speed control of the rotor is thus necessary. In most cases a linear model is suffi-
cient for considering the influence of the control characteristics.
It is not necessary to proceed on the basically correct assumption, that the dynamic
response, and with it the stresses, can only be calculated correctly if the functional and
elastic characteristics of the entire turbine are taken into consideration. An elastic struc-
tural model of the entire turbine inevitably requires an enormous computational effort
resulting in a corresponding amount of data, with the associated risk of missing the
critical points. It is therefore important to use a good measure of feel for how the com-
ponents are dynamically coupled, in order to define subsystems with the aid of which
the significant loads can be calculated. In most cases, for example, the rotor blade vibra-
tion can be considered in isolation, possibly coupled with the behaviour of the drive
train in the case of the lead-lag vibration. The rotor/tower system can also be generally
considered in isolation with regard to the bending vibration of the tower.
6.7.2 Representation of the Wind Turbulence

There are basically two methods for determining wind turbulence by theoretical means.
One is via the energy spectrum of the turbulence, the other is by means of an actual
wind speed time history (s. Chapt. 13).
Independently of the chosen method, one phenomenon affecting the reaction of the
wind rotor to turbulence must not be overlooked. In the open atmosphere, wind speed
and turbulence are always unevenly distributed in space over the rotor-swept area. Many
gusts strike the rotor not as a whole, but only on one side or only partially. This fact is
significant for the response of the structure as regards the rotating rotor. The rotor blades
"beat" into the gusts, i.e. the local wind speed changes, at their tangential speed.
An observer travelling with thhe rotor blade experiences these speed changes considerrably
more strongly than he woulld in the steady-state system. Moreover, depending onn the
duration of the gust and the speed of the rotor, the rotor blade can encounter the ssame
gust several times (Fig. 6.25).
Fig. 6.25. Effect of an uneven wind-speed distribution over the swept rotor area on the upw
wind
velocity of the rotating rotor blaade
6.7 MATHEMATICAL MODELLS OF STRUCTURAL DYNAMICS 211
This process of rotationall sampling is of considerable significance with respect too the
effect of wind turbulence on the rotor blades, especially with large rotors. The fatiguuing
effect on the structure can increase by up to 50 % compared to a merely time-depenndent
approach to turbulences in a non-rotating, stationary reference system.
The stochastic changing loads, i.e. the turbulence of the wind, can be represeented
on two different ways. Acccording to this also the methods of calculation and the
presentation of the results arre different. The most important models are statisticcally
based and known as „time history or time domain method“ and „spectral methhod“
(Fig. 6.26).
Fig. 6.26. Mathematical models for calculating the structural dynamic response to wind tuurbu-
lence: time domain and spectral approach
Time history method

If the time history of the active force is known, for example the variation of wind speed
with time, the resultant response of the structure versus time can be calculated. This
requires an aerodynamic model of the rotor, so that the variation of the aerodynamic
force can be determined from that of the wind speed. Using the elastic structure model,
the response of the structure over time is obtained.
The advantage of this method is that all parameters are time-dependent, a form of
presentation which is advantageous for several purposes. Moreover, functional algo-
rithms, for example for the influence of the control system, can be taken into considera-
tion. The influence of periodic forces, for example from the shear wind gradient or
tower interference, can also be determined well by means of the time history approach.
The serious disadvantage of this method is the more or less random "segment" of wind
turbulence used as a basis. This does not lead to a comprehensive picture. If this were
attempted, the calculation effort would become extremely high. Hence, this method is
more suitable for a selective "check", rather than for comprehensive structural dimen-
sioning with respect to fatigue life.
Spectral method
In the so-called “spectral method”, frequency-dependent representations (spectra) of
forces and responses are processed instead of their progression over time. This method
uses a statistical turbulence spectrum of the wind as the load input (Chapt. 13).
It must be possible to represent the structure in the form of linear or linearised equa-
tions (linear systems theory). The excitation spectrum causes excessive dynamic peaks
of response in the regions of the natural frequencies of the structure. The extreme values
of the required parameters (deformations, forces etc.) which are decisive for the dimen-
sioning of the structure can be represented as follows:
where is the quasi-statically calculated mean value, σx is the standard deviation of the
dynamic excursions about the mean value and K is the so-called “peak factor” based on
statistical reliability calculations.
The link between the excitation spectrum and the spectra of the response reaction is
established via so-called “transfer functions”. The "aerodynamic admittance" leads from
the wind spectrum to the aerodynamic force parameters, "mechanical admittance"
represents the link between the active forces and the deformations or stresses of the
structure.
The decisive advantage of the spectral method is the reliable acquisition of the entire,
real load spectrum caused by the wind turbulence. This method is thus predestined for
calculating structural fatigue. The fact that the required deformation and stress parame-
ters are only available as frequency-dependent spectra, and not as plots against time is,
6.7 MATHEMATICAL MODELS OF STRUCTURAL DYNAMICS 213
admittedly, a disadvantage in view of some of the technical problems at hand. For exam-
ple, it is difficult to process the functional characteristics of a wind turbine methodically,
with respect to the influence of the control system on the loads (functional model).
Deterministic approach
In contrast to the statistical methods described above, it is also possible to follow a deter-
ministic approach for calculating the dynamic structure responses. As in the example of
the time history method, one single event, for example a discrete gust, can be used as load
input, rather than the continuous progression of wind speed (Chapt. 6.2.4). The structural
response derived from this provides information on the dynamic load magnifications to be
expected. From the results, all-inclusive "dynamic magnification factors" for the quasi-
statically calculated stress can be derived.
The continuous nature of wind turbulence and of the response of the structure is, of
course, lost in the process. It is also not possible to cover all of the load inputs with
respect to the overall load spectrum by this method. Up to a certain point, one can get
by with assuming a certain frequency of the various discrete events (gusts), but
the validity of the results with respect to the structure's fatigue nevertheless remains
questionable.
6.7.3 Analytical Approaches and Numerical Computer Codes

In the procedure for structural dimensioning, the mathematical models outlined are com-
bined with one another (Fig. 6.26) [9]. This provides the structural loads in the form of so-
called "stress resultants" at pre-defined points of intersection of the structural components
as time or as frequency spectra (Chapt. 6.6.2). The combined stresses of all given load
cases represent the load spectra for the individual components of the wind turbine.
The calculated material stresses are compared, as usual, with the permissible stress val-
ues. To be able to determine the permissible values, material properties and the design
standards to be used, for example for the welded seams, are needed. With the considera-
tion of safety factors to guard against fracture, or for some other defined threshold value,
the structure can be dimensioned, or, if dimensioning has been determined, design verifi-
cation can be carried out (Chapt. 6.8).
In principle analytical approaches are possible to calculate the structural dynamic
stresses at least for geometrically simple structures. For example for the rotor blades,
because they can be represented by a simple beam model. The first step is the modal
analysis to determine the natural frequencies and mode shapes (s. Chapt. 7). The modal
analysis can be based on the well known finite element calculations. The next step is
establishing of the equations of motion for each element and to combine them with the
boundary conditions matched at the interfaces. The equations can be solved in prede-
termined time steps. The result is the distribution of the loading and stress parameters
over the blade length. Under simplified conditions the analysis can be carried out in an
analytical form [14].
With the availability of increasing computer power numerical simulation codes for
structural dynamics are widely in use. In most cases they will be based on the time do-
main method. In this way also non–linear effects, for example at large deformations,
and non-stationary effects, fo or example in aerodynamics, can be included. Furtherm more

a great number of componen nts can be modelled in a detailed finite-element represeenta-
tion (multi-body simulation)). By numerical integration of the equations over the ttime,
subdivided into short time steps,
s the problems can be dealt with any desired leveel of
accuracy. A great variety of commercial available computer codes are offered for sttruc-
tural dynamics. They are useed in nearly all fields of technology for dynamically loaaded
systems. In wind energy tecchnology structural dynamic simulations become morre or
less mandatory in the certificcation process.
A critical remark should be allowed about the extensive use of structural dynaamic
simulations, regardingless th here undisputed importance. The very complex simulaation
codes are not suited for find ding an optimal conceptual design of a system. This task
remains to the design engineer or to the design team. Finding an optimal conceptt in-
cludes much more aspects, likel costs, manufacturing capabilities or operational isssues.
An optimum concept only caan be based on “experience” and “creativity”. The first one
is as important as the second one.
Fig. 6.27. Flow chart for calculatting the loads and for dimensioning the structure
6.8 CONCEPTUAL DESIGN FEATURES AND STRUCTURAL STRESSES 215
6.8 Conceptual Design Features and Structural Stresses

The design engineer, with his choice of conceptual design features, determines the load-
ing on the components within certain limits. The general aim must be to reduce the
loads to a level which is the unavoidable minimum. The loading on the rotor and the
turbine in a steady mean wind and that resulting from the weight of the components is
unavoidable. Loads resulting from the turbulence of the wind are avoidable to a certain
degree, however. Damping these dynamic loads by suitable design features, allowing
the turbine to exhibit a "softer" dynamic response, is a key problem in the design.
As a first step, it will be attempted to reduce the high alternating loads on the rotor
blades. The dynamic response of the rotor to the wind gusts affects, in particular, the
flapwise bending moment in the root area of the blades. This load is of decisive signifi-
cance for the fatigue loading on the rotor blades.
In addition - and this aspect is just as important - the cyclically changing total rotor
forces and moments must be evened out. These loads are passed on to the other turbine
components and determine the dynamic load level for the mechanical drive train, the
yawing system and the tower of the turbine.
The most important system features determining the dynamic load level of the wind
turbine are the number of rotor blades, the function of the rotor hub in the case of two-
bladed rotors, the type and quality of power control, and, not least, the stiffness of the
electrical coupling to the fixed-frequency grid.
Out of these possibilities open to the design engineer, two basic philosophies emerge.
On the one hand, there is the school still adhering to the old English motto: "Make it
stiff and strong and you will never be wrong". The older, stall-controlled Danish wind
turbines followed this principle. On the other hand, there is the endeavour to keep the
dynamic response of the design and structure as soft as possible so as to reduce material
stress. It goes without saying that this approach is the more promising one for large
turbines even if it is associated with more development work.
6.8.1 Number of Rotor Blades

Considering the sum total of the rotor forces during a steady-state but asymmetrical
wind flow, serious differences become apparent which depend on the number of rotor
blades. This is clearly illustrated by the example of the aerodynamic yaw moment and
the driving torque. While one- and two-bladed rotors generate considerable alternating
loads with respect to the yaw moment and a pulsating drive torque, the rotor moments
almost completely balance out overall during a revolution in rotors having more than
two blades (Fig. 6.28). One-bladed rotors behave quite unfavourably in this respect.
Their geometric asymmetry, and with it their aerodynamic asymmetry, causes extreme,
alternating rotor forces and moments even with a symmetrical wind flow.
The critical influence of the number of blades becomes even clearer if the dynamic
response of the elastic rotor is also considered. This is especially true of rotors with less
than three blades, a fact which has long been known empirically. The deformations
experienced by a rotor under the influence of external forces, primarily bending of the
blades, produce inertia forces due to the structural masses being accelerated. The mo-
ment of inertia of the rotor around its instantaneous axis of motion is of importance to
the dynamic response to thesse external loads. When the rotor is rotating, the momennt of
inertia changes during a roto or revolution in relation to a fixed axis when the rotor oonly
has two blades, i.e. behaves like a rotating rod. Whereas rotors with three or m more
blades behave like a disk as far
f as the moment of inertia is concerned, i.e. are symm metri-
cal in terms of mass, the maass of the rod-shaped two-bladed rotor is asymmetrical and
its moment of inertia has a pulsating
p profile during a revolution. Depending on wheether
the rotor blades are perpend dicular or parallel to the axis under consideration, the m mass
moment of inertia varies from a maximum to a minimum value. This phenomenonn has
serious consequences with respect
r to the dynamic re action of the rotor during exxcur-
sions from its normal position.
If an asymmetrical wind flow,
f for example in a horizontal rotor position, causes a de-
flection of the rotor blades, th
he resultant angular velocity around the vertical axis is ccom-
paratively small, as the momeent of inertia of the rotor about its vertical axis is large inn this
position. If the rotor continuees to turn towards the perpendicular, the moment of innertia
about the vertical axis reducees. Since, for physical reasons, the rotational momentum m is
maintained, the angular velo ocity around the vertical axis becomes all the greater. The
consequence is a dynamically y-caused yaw moment about the vertical axis. This dynaamic
moment of reaction reinforcees the aerodynamic yaw moment already existing from m the
asymmetrical flow. The pitching moment, triggered, for example, by the asymmettrical
flow to the rotor due to shear wind, is thus reinforced in a two-bladed rotor.
Fig. 6.28. Aerodynamic yaw mom

ment of a rotor with different numbers of blades with an asym
mmet-
rical wind flow, calculated using
g the WKA-60 as an example
6.8 CONCEPTUAL DESIGN FEATURES
E AND STRUCTURAL STRESSES 217
Hence, wind turbines witth two-bladed rotors are subjected to particularly highh dy-
namic loads if the rotor bladees are joined rigidly to the rotor shaft. In order to reducee the
negative consequences for th he total system, either the strength and stiffness of the tur-
bine components must be diimensioned to accommodate this increased load, or thee de-
sign concept of the two-blad ded rotor must be selected such that it can largely redduce
these dynamic loads itself byy means of an appropriate controlled compliance.
6.8.2 Rotor Hub Hingess in Two-Blade Rotors

In order to reduce the poor dynamic response of the two-bladed rotor to asymmetrrical
flow conditions, a series of design ideas have been proposed and to a large part also
been realised, at least in expeerimental wind turbines. The preferred solution is the inntro-
duction of hinges, providing g the rotor blades with additional degrees of freedom m of
movement, so that the dynamic alternating loads can be reduced in the rotor itsellf by
"limited yielding" due to th he acceleration of its own masses. The simplest wayy of
achieving this compliance is by installing hinges between the rotor blades and the rrotor
shaft, i.e. in the rotor hub. Figure 6.29 shows the basic possibilities in the design of ttwo-
bladed rotors.
Fig. 6.29. Hingeless rotor and ro

otor hub hinges in two-bladed wind rotors
Hingeless rotor
The hingeless rotor, i.e. with
h the rotor blades joined rigidly to the rotor shaft, repressents
the traditional design. The old-fashioned
o windmill rotor has always been a hingeeless
rotor. This simple type is com
mpletely adequate for rotors with three or more rotor blaades,
even today. Two-bladed roto ors, too, were built with rigid hubs for reasons of simpliicity
(WTS-75, AEOLUS II). Thee advantage is the simple construction of the rotor hub and
the disadvantage is the fact that wind turbulence, in combination with the dynamicc re-
sponse of the two-bladed ro otor which tends to amplify the asymmetrical and cyyclic
alternating loads, must be sustained fully by the structure. This requires a stiff design
with corresponding expenditure on materials. This primarily affects the rotor blades, but
also the loads on the mechanical drive train and the yaw mechanism and the tower. If
the rotor is additionally subjected to tower shadow effects, the load situation becomes
even more unfavourable. Hingeless two-bladed rotors should, therefore, not be set up in
the downwind position.
Blade-flap hinges
Hinges permitting a limited flapping motion of the rotor blades were introduced as early
as 1940 in Smith-Putnam's wind turbine (Chapt. 2.3). The main advantage of a rotor
with individual blade-flap hinges is that it can evade symmetrical gusts, i.e. those strik-
ing the entire rotor area, as well as asymmetrical gusts.
One disadvantage of the flapping movement already became apparent in the turbine
mentioned above. The relatively large flapping movement of the blades shifted the cen-
tre of gravity closer to the rotor axis. The conservation of rotational momentum forced
the blade, which was closer to the rotational axis, to accelerate its rotational movement
about the rotor axis. The consequences were dynamically produced lateral forces and
torques acting on the rotor shaft. In operation, a rotor with individual flapping blade
movement would, therefore, be found to be running relatively roughly. In more recent
wind turbines, a blade-flap hinge has only been used on experimental one-bladed rotors.
Teetering rotor
The mechanical complexity associated with individual blade-flap hinges can be reduced
by connecting the entire rotor to the rotor shaft by means of a single hinge. The rotor is
thus able to perform teetering movements about the rotor shaft. A teetering hub of this
type was used for the first time in 1959 by Ulrich Hütter.
The teetering rotor responds to a symmetrical loading on the rotor in the same way as
a hingeless rotor. Asymmetrical loads, however, can be balanced out. The teetering
rotor results in considerable improvement, particularly as far as the cyclic loads caused
by the vertical windspeed profile are concerned. The unfavourable yaw and pitch mo-
ments of the rotor disappear almost entirely. Installing a teetering hinge on a two-bladed
rotor achieves dynamic characteristics comparable to those of a three-bladed rotor.
Hence, two-bladed rotors with teetering hubs and three-bladed rotors with hingeless
hubs can be considered to be genuine alternative concepts. The teetering rotor of the
former large experimental turbines was the preferred design for the large two-bladed
rotors. However, the mechanical elements to enable the rotor to teeter proofed to be
unreliable. In the future the load balancing of two blade rotors will be achieved by more
sophisticated blade pitch control systems (see Chapt.10).
Teetered rotors with blade pitch coupling

An elegant method of restricting the flapping or teetering movements of the rotor blades
while reinforcing their load-compensating effect, is to couple the teeter movement to an
adjustment of blade pitch angle (Fig. 6.30). This load compensation system was used
for the historical W34 wind turbine of U. Hütter in the Fifties (s. Chapt. 2.4). Couppling
teetering and blade pitch moovements is achieved either by means of a mechanical llink-
age or by suitably tilting thee teetering axis with respect to the rotor shaft. This laatter
method is called "δ3-couplingg", a term adopted from helicopter technology.
However, the effect of blaade pitch coupling depends to a great extent on the aeroody-
namic sensitivity of the roto
or. For heavy rotors, the effect was considered too weaak to
justify the mechanical comp plexity (for example MOD-2). But also some smaller tur-
bines with teetering rotors, for example the earlier American ESI turbines, manaaged
without blade pitch coupling.
Fig. 6.30. Teetered rotor with bllade pitch coupling

a) by means of a mechanical liinkage
b) by tilting the teetering axis (δ
( 3-angle)
Blade lead-lag hinges

The theoretically greatest dy ynamic compliance of the rotor can be achieved by proovid-
ing the rotor blades with ad dditional lead-lag freedom. Helicopter rotors have flappping
and lead-lag hinges, as is generally
g known. However, the mechanical complexity of
these is enormous. There is the additional hazard of high degrees of instability, so that
flapping and lead-lag hinges are not found in large wind rotors. An attempt in this diirec-
tion was made by John Brow wn in 1955 with his 100 kW wind turbine, which wass un-
necessarily applied to a threee-bladed rotor. The project turned out to be a failuree for
other reasons, too. It is easierr to provide the rotor blades with lead-lag freedom by uusing
a variable rotor speed, which h can then be considered as a "collective lead-lag motionn" of
the rotor blades.
Considered all in all, mecchanical hinges between rotor and rotor shaft have not bbeen
successful in large wind turbiines. Apart from other problems, all systems have displaayed
great wear. Improvements haave been achieved today due to progress in control enginneer-
ing. Individual control of thee blade pitch angle has a similar effect to load-compensaating
hinges between the rotor blades and the rotor hub. This makes it possible to control the
pitch angle of each individual rotor blades periodically via the rotational cycle or, even
better, in dependence on loading, in such a manner that the alternating loading can be
corrected by the amplitude profile of the wind velocity or any other asymmetric rotor
loading (s. Chapt. 6.8.4).
6.8.3 Stiffness of the Rotor Blades

It is obvious that the existing flexural elasticity of the rotor blades can be used to reduce
the symmetrical and asymmetrical external loads. This method has been applied suc-
cessfully with helicopter rotors where the introduction of elastic rotor blade root hinges
allows the rotor blades to perform a flapping motion. Generally, an appropriately tuned
rotor blade bending elasticity over the entire blade length can have the same effect.
On wind rotors, the practical implementation of this solution is not easy. It is difficult
to achieve a high bending elasticity of the blades without coupling together several
degrees of elastic freedom, including undesirable ones. The aeroelastic behaviour is
then difficult to control, above all with respect to blade pitch control. In addition, the
full-load deflection can become so great that the free space between the blade tips and
the tower becomes a critical design criterion.
In principle, using the bending elasticity of the rotor blades specifically as a means
for reducing the dynamic loading is independent of the number of rotor blades. It can
also be used for reducing the level of dynamic loading for three-bladed rotors. This
effect is also increasingly taken into account in the design of more recent turbines. For
example, the rotor blades of the large Vestas turbines are relatively flexible. This design
absorbs the loads more "softly" and saves weight in the rotor blades (Fig. 6.31).
Fig. 6.31. Vestas V80 at full power with greatly deflected rotor blades
6.8.4 Power Control System

Obviously, the aerodynamic power control of the rotor must have some influence on the
loading. The first, more basic question is the question about the differences between
blade pitch control and power limitation due to aerodynamic stall. In earlier years this
almost amounted to a question of faith in wind energy technology but in the meantime,
this conflict has lost its meaning since almost all the larger turbines have blade pitch
control with a variable-speed mode of operation.
Stall-type systems with fixed rotor blades must doubtlessly withstand higher ultimate
loads. As explained in Chapter 5.3.2, the rotor thrust does not drop off after the rated
power level has been reached and the associated stall has begun as in the case of a blade
pitch control system. This high shearing force of the rotor represents a comparatively
high load both for the tower and the foundation, but also for the rotor blades themselves.
An even more significant fact is that at extreme wind velocities, there is no possibility
of bringing the rotor blades into a favourable feathered position and the loads become
extremely high when parked. Since the dimensions of the tower and its foundations are
determined by the maximum loads this will lead to greater structural masses and thus
also to higher manufacturing costs. However, these parked loads can be avoided by
using an active stall control system (s.a. Chapt. 5.3.3) which, on the other hand, makes
it more difficult to obtain a fundamental comparison between pitch control and stall.
The situation becomes more complicated when fatigue strength is considered. The
reaction to a fluctuation in the wind velocity differs greatly in the two control systems.
At higher wind velocities within the range of its power rating and above, a stall-
controlled turbine will always run “close to stalling”. Any short-term increase in wind
velocity will immediately lead deeper into the stalling range, with the consequence that
the lift coefficient will drop, and thus also the loading. For this reason, the fatigue load
spectrum resulting from the wind turbulence is more favourable here than in pitch-
controlled units.
Due to the inertia of blade pitching, rotors with pitch control are not capable of re-
sponding to the short-term fluctuations in wind energy which is why the fatigue loads
resulting from the wind turbulence become higher. The blade pitch control can only be
used for responding to relatively long-lasting fluctuations in the wind velocity (Fig.
6.32).
It must be pointed out here, however, that the comparison with stall with blade pitch
control at a fixed rotor speed has become more or less academic today. In almost all of
the more recent plants, the blade pitch control is associated with a variable-speed oper-
ating mode and the conditions are completely different (s.a. Fig. 6.35). Combining blade
pitch control with a variable mode of operating the rotor has the effect of almost com-
pletely smoothing out the power delivery and thus also making the loading more uni-
form with respect to the torque acting on the mechanical drive train (gearbox).
In addition, the blade pitch control has a further option which can contribute to further
lowering of the load level in future. In most cases, an individual electric pitch actuator is
used today for each rotor blade. This opens up the opportunity of adjusting the pitch angle of
the rotor blades individually, i.e. independently of one another. For the reasons outlined in
Chapter 5.3.1 the introduction of a load-depending individual pitch control has not been
introduced in series-producedd wind turbines. There are ongoing attempts but the technnical
complexity is very high. A satiisfactory solution remains a challenge for the designers.
Fig. 6.32. Influence of blade piitch control on the smoothing of the electric power output w
with a
fixed rotor speed
Using such a control metthod makes it possible to compensate for asymmetric rrotor
loads. For example, the perioodically cyclic alternating loading resulting from the vverti-
cal wind shear could be com mpensated for by applying a cyclic change in pitch aangle
superimposed on the normall control function. However, a cyclic change in pitch aangle
would not compensate for otther asymmetric rotor inflows occurring stochastically. T This
could only be achieved by using
u a load-dependent and completely individual conntrol
system for each single rotor blade.
b
In recent years, the leadin
ng manufacturers such as Vestas or General Electric hhave
undertaken trials with an individual blade pitch control system. This type of control was
provided for the Vestas V90 0 unit but has not yet been adopted in series productioon to
the present day. As explainedd later an individual blade pitch control would be of speecial
significance especially for tw
wo-bladed rotors. Large two-bladed rotors such as could be
found in the early, large experimental units could be provided with such a control sys-
tem and could experience a renaissance for special applications, especially for very
large off-shore installations. Having this aim in mind, the basic groundwork for individ-
ual blade pitch control is also being carried out in various research institutes [16].
However, one must not overlook the difficulties of implementing a reliable individ-
ual blade pitch control arrangement. The essential problem is the availability of the
correct signal as reference variable for the control. It is possible to measure the flexural
loading of the rotor blades directly by means of strain gauges and to use these signals. It
is also possible to use acceleration values from the tower head movement as input sig-
nals. But the practical difficulties lie in the fact that, apart from the reliability of the
sensors, there are many details to be optimized. Even the location of the measurement
already becomes problematic due to the continuously changing asymmetric wind load-
ing. In addition, the rotor pitch actuator must be capable of providing an adequate pitch-
ing rate for this task. It is certain that the blade pitch drives needed for this purpose will
become heavier and more expensive. In view of these manifold development tasks,
individual blade pitch control will probably remain to be an option for future wind tur-
bines used commercially.
6.8.5 Rotor Speed Flexibility and Variable-Speed Operation
One of the most important design features of an increasing number of wind turbines is
the variable speed control of the rotor. This makes it possible to achieve two aims at the
same time.
Firstly, if the bandwidth of the available rotor speed is great enough, the rotor speed
can be adapted to the wind speed and it can be operated with its optimum tip-speed ratio,
i.e. with its maximum power coefficient. This results in an increased energy supply in
comparison with an operating mode with fixed rotor speed (s.a. Chapt. 14.6.3).
The second advantage lies in the fact that the rotor can store or deliver short-term
power changes by deviations in its speed, i.e. by increasing or decreasing its kinetic
energy like a flywheel. Having this capability makes it possible to very effectively
smooth out the delivery of the electrical power and the dynamic load changes. In com-
parison with the aerodynamically wind-controlled operation requiring a relatively large
speed range, a comparably low speed elasticity of only a few percent is sufficient to
achieve a noticeable reduction in the dynamic loads. The mechanical and electrical
options for a flexible-speed or speed-controlled operating mode are explained in Chap-
ters 8.10 and 9.4. The available technical solutions behave differently with respect to the
loads.
Torsional compliance in the mechanical drive train

Wind turbines equipped with synchronous generators directly coupled to the grid must
have a minimum of torsional compliance and damping in the mechanical drive train.
There are either torsionally elastic components which are built into the low-speed or
high-speed shaft, or the gearbox must have a torsion-elastic suspension. Naturally, the
effect of such measures depends greatly on the actual design adopted. Apart from torsional
elasticity, adequate damping is required to keep the vibrational behaviour under conntrol.
Torsionally elastic gearbox suspensions
s were to be found in some large, experimeental
first-generation wind turbiness (Chapt. 8.9). The gearbox was able to respond to an insstan-
taneous torque peak with a toorsional freedom of about 20 to 30 degrees to smooth out the
load peak.
The gearboxes in today’s wind turbines are mounted much more simply and inexppen-
sively in elastic rubber bodiies (s.a. Chapt. 8.8.4 and Fig. 8.44). Although this typpe of
mounting is primarily used fo or preventing the transmission of structure-borne soundd and
for decoupling the vibration characteristics, dynamic load peaks are also removed, eeven
if only to a limited extent.
Hydrodynamic slip in the drive

d train
An even more effective torsiional compliance, being more strongly damped, is achieeved
by installing a hydraulic cou
upling in the mechanical drive train. Couplings of this type
have a rotational slip of aro
ound 2 to 3 %, as a rule. A combination of a synchronnous
generator coupled directly too the grid and hydraulic coupling in the mechanical ddrive
train has been used in the passt in several types of turbines, for example the earlier H
How-
den HWP-330 and the Westtinghouse WWG-0600 or the MOD-0 turbine (Chapt. 88.9).
Using the MOD-0 turbine as an example, Figure 6.33 shows the effect of a hydraaulic
coupling with respect to po ower output smoothing. Drive trains including a hydraaulic
coupling are also found in soome recent designs (General Electric). Obviously the poossi-
bility of using a synchronouss standard generator is considered as an advantage.
Fig. 6.33. Smoothing of the poower output due to the installation of a hydraulic coupling inn the
mechanical drive train of an MO
OD-0 [17]
Electrical slip of an inductiion generator

In wind turbines with inducttion generators, loading peaks can be smoothed out viaa the
electrical slip of the generator (Fig. 6.34). However, large induction generators oonly
have low slip values in theirr standard production models (Chapt. 9.1). It is only wiith a
slip of at least 1 to 2 % thatt the dynamic load level is reduced perceptibly and, att the
same time, unwanted drive trrain vibrations are avoided.
Fig. 6.34. Smoothing of power and torque with an induction generator with a nominal slip oof 0.5
and 5.5 % [18]
Controlled variable-speed operation

True smoothing of the power captured by the rotor is only achieved by controlled variaable-
speed operation of the rotor. The
T generator can be operated in variable-speed mode if a fre-
quency converter is connected d (Chapt. 9.5). With this arrangement, the generator torquee can
be controlled to a constant vaalue, independently of the speed, within a given speed raange.
The result is a complete smootthing out of the power transferred, and thus also of the loadding,
within the predetermined speed limits (Fig. 6.35).
The capability of a variaable speed is, however, limited by the speed range im mple-
mented technically, so that laarger fluctuations of wind speed cannot be compensatedd for.
Effective smoothing of heav vier wind gusts and their associated power and load peeaks
can only be achieved with the t aid of blade pitch control. Rotor speed variability and
blade pitch control should, therefore, always be considered together, as their efffects
complement each other. Forr this reason, nearly all recent turbines are equipped with
both of these system featuress.
Fig. 6.35. Control of the poweer output with a variable-speed synchronous generator withh fre-
quency converter, using the Gro
owian turbine as an example [18]
6.9 Measuring the Strructural Stresses

The theoretical determination of the structural stresses to which a wind turbine is sub-
jected under the different loaading conditions and in the various operating modes stilll has
its limits in spite of elaboratee and complex mathematical methods. Although signifiicant
advances have been achieveed in this field in recent years, a refinement of experrtise,
especially in the area of fatig
gue loads occurring in long-term operation, is essential iff the
weights of the components, and thus ultimately also the manufacturing costs, are tto be
reduced further. Apart from m the development of mathematical models, therefore,, the
measurement of stresses acctually occurring has occupied a predominant positionn in
numerous research and deveelopment projects for many years. Some of the earlier large
experimental wind turbines had h literally been designed as test-beds for the investigaation
of loads. Such measurements are still being carried out in newly developed turbbines
even today.
6.9 MEASURING THE STRUCTURAL STRESSES 227
Naturally, the experimental load investigations also include measurements and tests
which can be carried out on test stands for individual components. Tests carried out on test
stands have the invaluable advantage of conveying the correlation between set loads and
the responses of the test objects under reproducible conditions. They are appropriate when-
ever unknown material properties, the interaction of different materials in a specific design,
uncertainties concerning manufacturing techniques or even the verification of calculated
results are to be investigated. The loads themselves must, however, be specified, i.e. their
correctness must be assumed. Apart from test-bed trials with gearboxes, or total mechanical
drive trains, the rotor blades commonly are subject of testing on test beds.
6.9.1 Rotor Blade Testing

To verify their mechanical properties and dynamic characteristics, newly developed rotor
blades are tested on special test stands (Fig. 6.36). The static loading capacity of the blades
is first verified experimentally and the precalculated stresses in the load-bearing structural
elements are determined experimentally with the aid of strain gauges. The deflections
measured are an additional criterion for checking the constructional design assumptions.
Dynamic fatigue load spectra can only be simulated to a limited extent. The very high load
cycle numbers in the lifetime of a wind turbine, combined with the associated amplitudes,
can only be represented in elaborate long-term test programs. Only the critical elements of
the design are, therefore, tested with a dynamic load spectrum, in the form of smaller test
objects, for example the load-transferring elements between the rotor blade structure and
the hub. This makes it possible to test the overall design for its fatigue life at least in critical
sections.
Fig. 6.36. Rotor blade test stand at LM Glasfiber

Another important task is th he determination of the most important natural frequenciess. On

a test stand, where the cantilev vered rotor blade is fixed at the root and is induced to vibbrate,
both natural frequencies and vibration modes can be measured with high precision. Al-
though the natural frequenciess determined under these conditions do not correspond exaactly
to the natural frequencies of th he rotating rotor, the predicted stiffness parameters of the non-
rotating blade can still be veriffied in this way.
6.9.2 Data Acquisition Systems

S and Field Measurements
Inquiring into the loads and structural stresses actually existing is, of course, posssible
only on the wind turbine itseelf. The usual approach is to measure the deflections off the
selected components by meaans of strain gauges and then to deduce the material sttress
values. However, to obtain a complete overview of the entire load spectrum requuires
arduous, long-term measurem ment campaigns the results of which only become meanning-
ful after a large amount of daata has been statistically processed. Analysing the results in
detail can be very difficult. Correlating
C the results with the causes of the loads can oonly
be carried out to a limited exxtent, as the structural deformations only reflect the summ of
all loads. To isolate aerodynaamic loads, for example, is extraordinarily difficult. Simmilar
problems are presented when n particular load states are to be correlated with the evvents
triggering them, for example individual gusts.
One of the first systematiic measuring programmes was carried out by NASA inn the
years 1977 to 1979 during a Swedish-American test programme using the earlier Daanish
wind turbine at Gedser [19]]. From 1976 to 1986 comprehensive measurements w were
taken out in the American wind
w energy programme with the experimental MOD-0 [[20].
This was followed by the pu ublication of measurements made on the large experimeental
turbines such as MOD-1, MOD-2, WTS-4 and Growian [21, 22]. One of the resullts is
shown in Figure 6.37.
Fig. 6.37. Calculated and measu

ured bending moments of the rotor blades of a MOD-2 [21]
6.9 MEASURING THE STRUCTTURAL STRESSES 229
The measured cyclic dynaamic loads on the rotor blades of the MOD-2 correlate rrela-
tively well in the statistical mean with the values predicted by the computer progrrams
MOSTAB and GEM. Obtain ning such measurement data and the associated refinem ment
of the computer programs are a decisive prerequisites for the reliable structural dim
men-
sioning of progressive lightwweight design concepts.
Fig. 6.38. Data acquisition systeem of an ENERCON E-40 prototype

Some comments must be made regarding data acquisition and evaluation in connec-
tion with load measurements taken at wind turbines. As test objects, all newly devel-
oped prototypes are equipped with elaborate measuring and data acquisition systems. In
the period of test operation, setting up and operating this measuring equipment takes up
a large part of the development work. This applies both to assembling the hardware and
to developing the software for data editing and evaluation. Figure 6.38 shows the basic
set-up of a data acquisition and evaluation system used for an early prototype of the
ENERCON E-40 wind turbine.
Using various sensors and transducers such as strain gauges, accelerometers, force
and displacement sensors, anemometers or instruments for measuring electric parame-
ters, the data acquisition system can record approximately 200 test points. The meas-
urement signals are amplified and sampled at a predetermined rate by a multiplexer. The
analogue signals are then digitised and subsequently converted into a serial data stream
by using PCM (pulse code modulation). PCM technology is generally required so that
the data arriving in parallel from a large number of measuring points can be transferred
using only one signal line. The time-consuming statistical processing of the load data
acquired requires extensive computer programs. A numerical method known as the
“rainflow method” has proved to be especially helpful for this purpose.
An important point must also be raised with regard to the design of the data recording
and processing system. If possible, this system, used only for the test purposes, should
be completely independent of the data processing system of the operational control
system of the wind turbine. A functional link would give cause for grave concern for
safety reasons.
References
1. International Electrotechnical Commission (IEC): Wind Turbine Generator Systems, Part 1
Safety Requirements, IEC 61400-1, 2nd edn. (1999)
2. Deutsches Institut für Bautechnik (DIBt): Richtlinie für Windenergieanlagen: Einwirkungen
und Standsicherheitsnachweise für Turm und Gründung, Berlin, vol. 8 (March 2004)
3. Abbott, H., Von Doenhoff, A.E.: Theory of Wing Sections. Dover Publications Inc., New
York (1958)
4. Burton, T., et al.: Wind Energy Handbook. Wiley, Chichester (2001)
5. Schepers, J.G., Snel, H.: Investigation of dynamic inflow effects and implementation of an
engineering method, ECN-Report ECN-C-94-107, Petten (1995)
6. Schewe, G.: Untersuchung der aerodynamischen Kräfte, die auf stumpfe Profile bei großen
Reynolds-Zahlen wirken. DFVLR Report 84-19 (1980)
7. Snyder, M.H.: Wakes Produced by a Single Element and Multiple Element Wind Turbine
Towers. Wind Energy Conversion Devices, Von-Karman-Institut (1981)
8. Frost, W., Long, B.H., Turner, R.E.: Engineering Handbook on the Atmospheric Environ-
mental Guidelines for Use in Wind Turbine Generator Development. NASA Technical Paper
1359 (1978)
9. Garrad, A.D., Hassan, D.: Taking the Guesswork out of Wind Turbine Design. In: AWEA
National Conference, Cambridge, Massachusetts, USA (1986)
10. Huß, G., Hau, E.: Lastfalldefinition für WKA-60. MAN-Report (1986)
REFERENCES 231
11. IEA Expert Group Study: International Recommended Practices for Wind Energy Conver-
sion Systems Testing 3. Fatigue Characteristics. 1st edn. (1984)
12. Lloyd, G.: Rules and Regulations IV - Non Marine Technology, Part 1 Wind Energy, Regu-
lations for the Certification of Wind Energy Conversion Systems, Clegster 1-10 (1993), N1
(1994), N2 2 (1998)
13. Dahlroth, D.: Load Cases for Medium-Sized Wind Power Plants, Greenford, 9. IEA Meeting
of Experts, Structural Design Criteria for LS-WECS (1983)
14. Spera, D.A. (ed.): Wind Turbine Technology. ASME Press, New York (1994)
15. Langenbrinck, J.: Verifikation der Rotorlasten für die Windkraftanlage WKA-60 auf
Helgoland, master thesis, TU München, Lehrstuhl für Leichtbau (1991)
16. Caselitz, P.: Individual blade pitch control design for load reduction on large wind turbines.
In: EWEC, Milano (2007)
17. Thomas, R., Richards T.R.: ERDA/NASA 100 kW MOD-0 Wind Turbine Cleveland, Ohio,
NASA Lewis Research Center (1973)
18. Cramer, G.: Entwicklung der Regelungsstruktur und Betriebsführung der Windener-
gieanlage AWEC-60, SMA Report (September 1987)
19. Lundsager, P., Christensen, C.J.; Fraudsen, S.: The Measurements on the Gedser Wind Mill
1977-79, Risø National Laboratory (1979)
20. Linscott, B.S., et al.: Experimental Data and Theoretical Analysis of an Operating 100-kW
Wind Turbine (MOD-0), DOE/NASA TM-73883 (1977)
21. MOD-2 Wind Turbine System Development, Boeing Construction, Final Report, NASA CR
N2 168007 (1982)
22. Huß, G.: Messprogramm an der Versuchsanlage WKA-60 auf Helgoland, BMFT- Research
Project 0328508D, Final Report (1994)

6 Loads and Structural Stresses - Removed

Uploaded by

Copyright:

Available Formats

6 Loads and Structural Stresses - Removed

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

6 Loads and Structural Stresses - Removed

Uploaded by

Copyright:

Available Formats

Chapter 6

Loads and Structural Stresses

6.1 Loads on the Wind Turbine

6.2 Coordinate Systems and Terminology

Fig. 6.2. Co-ordinates and techn

6.3 Sources of Loadin

6.3.1 Gravity and Inertial Loads

6.3.2 Uniform and Steady-State Air Flow

Fig. 6.3. Tangential load distribu

Fig. 6.4. Thrust load distribution

6.3.3 Vertical Wind Sheear and Cross Winds

6.3.4 Cross Wind on the Rotor

6.3.5 Tower Interferencce

Fig. 6.9. Flow field due to the tower

Fig. 6.12. Calculated increase in

Fig. 6.13. Influence of tower shadow

6.3.6 Wind Turbulence and Gusts

Spectral model of turbulence

Fig. 6.14. Idealised gust shapes [8]

Fig. 6.15. Gust factors in depend

Fig. 6.16. Gust factors in depend

6.4 Design Load Assumptions

6.4.1 International and National Design Standards

IEC 61400-1 Safety Requirements

IEC 61400-12 Wind Turbine Performance Testing

Det Norske Veritas

Dutch NVN Guideline 1400-0

Danish Standard DS 472

German DIBt Guidelines

Most of the organisations offering certifications are profit-oriented commercial under-

6.4.2 IEC Classes of Wind Turbines and German Wind Zones

Table 6.18. WTGS Classes to IE

- Wind Zone I: Weak-wind regions (24.3 m/s)

6.4.3 Normal Wind Conditions

Mean annual wind speed

Wind speed frequency distributions

Vertical wind shear

6.4.4 Extreme Wind Conditions

Extreme wind speeds and gusts

The so-called annual gust is:

where z is the hub height.

Whereas the negative gust is:

The gust factor is taken into consideration by:

Extreme changes in wind direction

6.4.5 Other Climatic and Environmental lnfluences

Salt content of the air

6.4.6 Other External Conditions

Influences from the grid

Influence of adjacent wind turbines

- Wind loads during assembly and repair

- Pressure of soil and grounnd water on the foundation

6.4.7 Safety Factors

6.5 Operational Statu

6.5.1 Normal Operation

- cut-in wind speed

Start-up and shut-down of the rotor

Parked rotor at extreme wind speeds

6.5.2 Technical Faults

Rotor emergency stop