Ris0-R-15OO(EN)

Control design for a pitch-regulated, vari­

able speed wind turbine
Morten H. Hansen, Anca Hansen, Torben J. Larsen, Stig 0ye,
Foul Sorensen and Peter Fuglsang

Ris0 National Laboratory

Roskilde
Denmark
January 2005
Author: Ris0-R-15OO(EN)
Title: Control design for a pitch-regulated, variable speed wind tur­ January 2005
bine
Department: Wind Energy Department

Abstract (max. 2000 char.): ISSN 0106-2840

ISBN 87-550-3409-8
The three different controller designs presented herein are similar and all based on
Pi-regulation of rotor speed and power through the collective blade pitch angle and
generator moment. The aeroelastic and electrical modelling used for the time-domain
analysis of these controllers are however different, which makes a directly quantita­
tive comparison difficult. But there are some observations of similar behaviours
should be mentioned:

• Very similar step responses in rotor speed, pitch angle, and power are
Contract no.:
seen for simulations with steps in wind speed.
ENS1363/02-0017
• All controllers show a peak in power for wind speed step-up over rated
wind speed, which can be almost removed by changing the parameters of
the frequency converter. Group's own reg. no.:
1110039-00
• Responses of rotor speed, pitch angle, and power for different simulations
with turbulent inflow are similar for all three controllers. Again, there
seems to be an advantage of tuning the parameters of the frequency con­ Sponsorship:
Danish Energy Authority
verter to obtain a more constant power output.

Tire dynamic modelling of the power controller is an important result for the inclu­
sion of generator dynamics in the aeroelastic modelling of wind turbines. A reduced Cover: Nil
dynamic model of the relation between generator torque and generator speed varia­
tions is presented; where the integral term of the inner Pi-regulator of rotor current is
removed be-cause the time constant is very small compared to the important aeroe­
lastic frequencies. It is shown how the parameters of the transfer function for the
remaining control system with the outer Pi-regulator of power can be derived from
the generator data sheet.

file main results of the numerical optimisation of the control parameters in the pitch
Pi-regulator performed in Chapter 6 are the following:

• Numerical optimization can be used to tune controller parameters, espe­

cially when the optimization is used as refinement of a qualified initial
guess.

• file design model used to calculate the initial value parameters, as de­
scribed in Chapter 3, could not be refined much in terms of performance
related to the flapwise blade root moment (1-2 %) and tilt tower base
moment (2-3 %).

• Numerical optimization of control parameters is not well suited for tuning

from scratch. If the initial parameters are too far off track the simulation
might not come through, or a not representative local maximum obtained. Pages: 85
Figures: 65
file last problem could very well be related to the chosen optimization method, Tables: 1
where more future work could be done. References: 8

Riso National Laboratory

Information Service Department
P.O.Box 49
DK-4000 Roskilde
Denmark
Telephone +45 46774004
biblfgirisoe.dk
Fax +45 46774013
www.risoe.dk
Contents

Preface 5

1 Introduction 6
1.1 Test turbine 6
1.2 Summary of results 8

2 Controller by VES 9
2.1 Simulation tool: DIgSILENT 9
2.2 Turbine modelling 9
2.2.1 Wind model 9
2.2.2 Mechanical model 10
2.2.3 Aerodynamic model 10
2.2.4 Generator concept and model 11
2.3 Control concept 12
2.4 Simulation results 17

3 Controller by AED 24
3.1 General info of the HAWC code 24
3.1.1 Test of dynamic wake model in HAWC 24
3.2 Turbine modelling 31
3.2.1 Turbine natural frequencies and mode shapes 31
3.3 Controller design 32
3.4 Results 34
3.4.1 Step response 34
3.4.2 Turbulent simulations 39
4 Controller by DTU 45
4.1 Short description of the FLEX5 code 45
4.2 Turbine data 45
4.3 Pitch servo model 45
4.4 Generator model 45
4.5 Control system 46
4.6 Results 48

5 Reduced model of DFIG generator dynamics 57

5.1 Generator dynamics 57
5.2 Reduced model 61
5.3 Parameter derivation 62
6 Numerical optimization of PI-controller 64
6.1 Method 64
6.2 Result 65
6.3 Conclusion 71

References 72

Appendix A 74

Appendix B 76

Ris0-R-15OO(EN) 3
Appendix C 80

Appendix D 81

4 Ris0-R-15OO(EN)
Preface
The work presented in this report is part of the EFP project called “Aeroelastisk Integre-
ret Vindm0llestyring” partly funded by the Danish Energy Authority under contract
number 1363/02-0017.
The main objective of the project is to implement controller design of wind turbines in
the general design process, in a similar way as is used for structural dynamics and aero­
dynamics. Specific focus is on pitch-controlled turbines with variable speed. The poten­
tial of future controllers will be identified and design limits will be mapped. As part of
this, individual pitching of each blade based on load response input parameters will be
treated, aiming at optimization of the operational state, reduction of loads and enhance­
ment of stability characteristics.
The basis of the project is an existing traditional pitch controlled turbine with variable
speed. General design tools for controller design using both linear and non-linear aeroe-
lastic methods will be developed. These design tools will be verified using traditional
methods, and optimum controller strategies will be developed. As part of this, the neces­
sary changes in the structural and aerodynamic design will be identified, to obtain an
optimum regulator.
The design of controller strategies will address the following main principles:
1. Optimization of power production with simultaneous load reduction of the main
wind turbine components (gear box, tower, blades)
2. Adaptive control with respect to actual operational conditions.
Partners: IMM og MEK(DTU), VES.
Contact: Morten H. Hansen, [email protected], tel. +45 4677 5077

Ris0-R-15OO(EN) 5
1 Introduction
This report deals with the design and tuning of the control system for the traditional
pitch-regulated, variable speed wind turbine, where PI-regulators are used for regulating
rotor speed and electrical power through the collective pitch angle and generator torque.
Three different examples of controller designs for a specific test turbine are presented
based on three different models of the aeroelastic and electrical systems.
The Wind Energy Systems research programme (VES) at Ris0 presents a controller de­
sign with state-of-the-art electrical modelling and simplified aeroelastic modelling. The
model and analysis are based on the simulation tool DIgSILENT.
The Aeroelastic Design research programme (AED) at Ris0 and Stig 0ye from the De­
partment of Mechanical Engineering at DTU present two controller designs with state-
of-the-art aeroelastic modelling and simplified electrical modelling. The models and
analyses are based on the simulation tools HAWC and Flex5, respectively.
The test turbine used for these investigations is described briefly in the following section.
Sections 2, 3, and 4 contain the methods and results of the VES, AED, and DTU control­
lers, respectively. Section 5 deals with the modelling of a Doubly-Fed Induction Genera­
tor (DFIG), where also a reduced model is described which can be used for aeroelastic
analyses in the lower frequency range. In Section 6, it is shown how to use a numerical
optimization tool for improving the parameters of the PI-regulator in pitch control loop
with respect to lower flapwise root bending loads. A summary of all results is given in
the end of this introduction (Section 1.2).

1.1 Test turbine

The test turbine for which a controller must be designed is described in this section. The
data for the turbine is given in a HAWC format, see command file and input files in Ap­
pendices A - D. The rated power of the turbine is 2 MW.
Figure 1 shows the aerodynamic power coefficient (Cp) surface computed with the aero­
elastic stability tool HAWCStab based on the Blade Element Momentum method. Using
HAWCStab, the optimal pitch angles and rotor speeds for maximum Cp have been com­
puted based on this Cp-surface with a lower constrain on pitch angle and upper con­
strains on rotor speed and power. The blue curve in Figure 1 represents the Cp for these
optimal operational conditions, which are plotted in Figure 2. The lower constrain on the
collective pitch angle was about 0.5 deg, and the upper constrains on rotor speed and
power were 18.8 rpm (generator speed of 1600 rpm and gear ratio of 85) and 2 MW, re­
spectively. Note that these operational conditions do not correct the power output for
losses in the drive-train and generator, whereby the rated electrical power is assumed to
be given by the maximal aerodynamic power of 2 MW.
Figure 3 shows the aerodynamic power and thrust for the test turbine with the opera­
tional conditions shown in Figure 2. Rated power is reached at a wind speed of approxi­
mately 12 m/s, where the rated rotor speed is reached around 10 m/s.
Finally, Figure 4 shows the natural frequencies of the first ten structural mode shapes of
the test turbine computed with HAWCStab as function of the wind speed, where the ef­
fects of pitch angle and rotor speed variations can be seen as changes in frequencies.

6 Ris0-R-15OO(EN)
 Cp [-]

Figure 1:Aerodynamic power coefficient surface with a curve representing the power
coefficient at the operational conditions for optimal aerodynamic power with a lower
constrain on pitch angle and upper constrains on rotor speed and power.

2000 -

p 20 - 1800 -

1600 -

1400 -

1200 -

1000 -

800 -

600 -

400 -
Pitch angle 200 -
Aerodynamic power
Rotor speed Thrust

Wind Speed [m/s] Wind Speed [m/s]

Figure 2: Operational conditions for op­ Figure 3: Aerodynamic power and thrust
timal aerodynamic power with constrains for the test turbine with the operational
on pitch angle, rotor speed and power. conditions shown in Figure 2.

1st lat. tower

1st long, tower
1st fixed-free shaft
1 st flapwise BW
1st flapwise FW
1st sym. flap
1 st edgewise BW
1st edgewise FW
2nd flapwise BW
2nd flapwise FW

0 5 10 15 20 25
Wind Speed [m/s]

Figure 4: Natural frequencies of the first ten structural turbine modes computed with
HA WCStab as function of wind speed with the variation ofpitch angle and rotor speed
given in Figure 2.

Ris0-R-15OO(EN) 7
1.2 Summary of results
The three different controller designs presented herein are similar and all based on PI-
regulation of rotor speed and power through the collective blade pitch angle and genera­
tor moment. The aeroelastic and electrical modelling used for the time-domain analysis
of these controllers are however different, as mentioned above, which makes a directly
quantitative comparison difficult. But there are some observations of similar and minor
deviating behaviours should be mentioned:

• Very similar step responses in rotor speed, pitch angle, and power are seen for
simulations of 1 m/s wind speed steps from 4 to 12 m/s and again from 12 to 20
m/s; also the tower top response computed by AED and DTU are very similar
(cf Figures 14, 31 and 42, and Figures 15, 33 and 43).
• All controllers show a peak in power for wind speed step-up over rated wind
speed, which VES have shown can be almost removed by changing the parame­
ters of the frequency converter (cf. Figure 16).
• Similar step responses in rotor speed, pitch angle, and power are seen for simu­
lations of 2 m/s wind speed steps from 20 to 5 m/s computed by VES and DTU
(cf Figures 18 and 44), except that the VES controller shows a dip in power for
all wind step-downs unless the integral term of speed controller is increased, re­
quiring faster pitch activity.
• Responses of rotor speed, pitch angle, and power for different simulations with
turbulent inflow are similar for all three controllers. Again, there seems to be an
advantage of tuning the parameters of the frequency converter to obtain a more
constant power output.
Note that the observed similarities in responses are mainly the result of similar tuning of
the control parameters of the PI-regulator.
The dynamic modelling of the power controller is an important result for the inclusion of
generator dynamics in the aeroelastic modelling of wind turbines. A reduced dynamic
model of the relation between generator torque and generator speed variations is pre­
sented, where the integral term of the inner PI-regulator of rotor current is removed be­
cause the time constant is very small compared to the important aeroelastic frequencies.
It is shown how the parameters of the transfer function for the remaining control system
with the outer PI-regulator of power can be derived from the generator data sheet.
The main results of the numerical optimisation of the control parameters in the pitch PI-
regulator performed in Chapter 6 are the following:

• Numerical optimization can be used to tune controller parameters, especially

when the optimization is used as refinement of a qualified initial guess.
• The design model used to calculate the initial value parameters, as described in
Chapter 3, could not be refined much in terms of performance related to the
flapwise blade root moment (1-2 %) and tilt tower base moment (2-3 %).

• Numerical optimization of control parameters is not well suited for tuning from
scratch. If the initial parameters are too far off track the simulation might not
come through, or a not representative local maximum obtained.
The last problem could very well be related to the chosen optimization method, where
more future work could be done.

8 Ris0-R-15OO(EN)
2 Controller by VES
2.1 Simulation tool: DIgSILENT
DIgSILENT is a dedicated electrical power system simulation tool used for assessment
of power quality and analysis of the dynamic interaction between wind turbines/wind
farms with the grid.

DIgSILENT provides a comprehensive library of models for electrical components in

power systems e.g.: generators, motors, controllers, power plants, dynamic loads and
various passive network elements as e.g. lines, transformers, static loads and shunts. The
program also provides a dynamic simulation language DSL which makes it possible for
the users to create its own models library, as it is the case for the mechanical side of a
wind turbine. Compared to HAWC and Flex5, the mechanical models implemented in
DIgSILENT are relatively simpler, while the electrical models are very detailed.

2.2 Turbine modelling

A simplified block scheme of the wind turbine model is shown in Figure 5. The basic
block scheme of the wind turbine consists of a wind model, an aerodynamic model, a
transmission system, generator model and a control block model.

2.2.1 Wind model

The wind model used is described in detail in [1]. The main advantages of this wind
model are the fast computation, reduced memory requirements and the ease of use, either
for variable or constant speed models. This wind speed model is very suitable for simul­
taneous simulation of a large number of wind turbines, making it possible to estimate
efficiently the impact of a large wind farm on the power quality.
The structure of the wind model is shown in Figure 6. It provides an equivalent wind
speed veq to the aerodynamic model. The wind model includes two sub-models: a hub
wind model and a rotor wind model.
The hub wind model models the fixed-point wind speed at hub height for each wind tur­
bine. In this hub wind model, the park scale coherence can be taken into account in the
case when a whole wind farm is modelled. The second is the rotor wind model, which
contains an averaging of the fixed-point wind speed over the whole rotor, the rotational
turbulence and the tower shadow model. The wind shear is not included in the rotor wind
model, as it only has a small influence on the power fluctuations.

Pref
Control model

blade 'rotor
P
gen
Wind eq JAerodynamic' Mechanical DFIG
model model model model

Figure 5: Scheme of wind turbine model.

Ris0-R-15OO(EN) 9
 Hub wind model Rotor wind model
Fixed point wind speed Averaged wind over rotor

Wind turbine Wind

Rotational turbulence
Turbine
Wind farms
Tower shadow
(park scale wind model)

Figure 6: Structure of the wind model.

Notice that the rotor position 9rot is used in the wind model, and therefore this wind
model can be used both variable speed and constant speed.
2.2.2 Mechanical model
In the mechanical model, the emphasis is put only on those parts of the dynamic struc­
ture of the wind turbine that contribute to the interaction with the grid. Therefore only
the drive train is considered in the first place, because this part of the wind turbine has
the most significant influence on the power fluctuations. The other parts of the wind tur­
bine structure, e.g. tower and the flap bending modes, are neglected.
The mechanical model is shown in Figure 7. It is essentially a two mass model connected
by a flexible low-speed shaft characterized by a stiffness k and a damping c. The high­
speed shaft is assumed stiff. Moreover, an ideal gear with the ratio 1:qgear is included.
The two masses correspond to the large turbine rotor inertia Jrot, representing the blades
and hub, and to the small inertia Jgen representing the induction generator. The generator
inertia is actually included in the generator model in DIgSILENT.
2.2.3 Aerodynamic model
The aerodynamic model is based on tables with the aerodynamic power efficiency
C (6, A) or torque coefficient Cq (6, A), which depend on the pitch angle 9 and on
the tip speed ratio A. This is a quasistatic aerodynamic model which determines the out­
put aerodynamic torque directly from the input wind speed according to:

T — Prot p n R^ u^ Cp (6, A) (1.1)

2 a

or

To — p n R3 u2 Cq (6, A) (1.2)
2

II ii
J 1 ngear
T k Tsshaft
IT?
J
___

Aero Low speed Gear High speed

dynamic shaft box shaft Generator

Figure 7: Drive train model in DIgSILENT.

10 Ris0-R-15OO(EN)
The aerodynamic model, implemented in DIgSILENT, can also include a model for dy­
namic stall [2,3]. The implemented model for dynamic stall is based on 0ye's dynamic
stall model [4],

2.2.4 Generator concept and model

A doubly-fed induction generator is basically a standards wound-rotor induction machine
with the stator windings directly connected to the three-phase grid and with the rotor
windings connected to a back-to-back partial scale frequency converter, which consists
of two independent converters connected to a common de-bus. These converters are il­
lustrated in Figure 8, as rotor side converter and grid side converter.
The behaviour of the generator is governed by these converters and their controllers both
in normal and fault conditions. The converters control the rotor voltage in magnitude and
phase angle and are therefore used for active and reactive power control.
The doubly-fed induction generator (DFIG) model in DIgSILENT, illustrated in Figure
9, contains the usual induction machine model and the PWM rotor side converter in se­
ries to the rotor impedance Zrot (DIgSILENT GmbH, 2002).
The line side converter is independent component in DIgSIlENT's library and it can be
added to the machine model to form the DFIG with back-to-back converter. The model
equations of the doubly-fed machine can be derived from the normal induction machine
equations by modifying the rotor-voltage equations:

i ^
— (ZZ +
(9„ —' (9„ dt
(1.3)
i
Hr = Rr l + ./ ¥ +
—a>„ dt

where u, /, and y/ arc space vectors for the voltage, current and flux, respectively. is
the synchronous speed, cor is the angular speed of the rotor, while co„ is the nominal elec­
trical frequency of the network. Note that the equations are expressed in per unit system.
The dynamic model of the induction generator is completed by the mechanical equation:

Gridaide
coMverfer comverfer

Figure 8: Doubly-fed induction generator.

Figure 9: Doubly-fed induction generator model in DIgSILENT.

Ris0-R-15OO(EN) 11
 / d)r = % - %, (1.4)

where J is generator inertia, Te is the electrical torque, is the mechanical torque. This
induction machine model of fifth order, based on Equations (1.3) and (1.4), is able to
represent rotor and stator transients correctly. Applying the principle of neglecting stator
transients to the doubly-fed induction machine model leads to the following third order
model:

", = & +; v.
a.
(1.5)
. " #r) i A/
—r = K 1 + ./ +"
-r 0,

The mechanical equation is the same as in case of the fifth order model.

2.3 Control concept

Figure 10 sketches the control system of a variable speed DFIG turbine implemented in
Dig SILENT. Two control levels using different bandwidths can be distinguished:
- Doubly-fed induction generator (DFIG) control level
- Wind turbine control level
The DFIG control, with a fast dynamic response, contains the electrical control of the
power converters and of the doubly-fed induction generator. The DFIG control contains
two controllers:
- Rotor side converter controller - controls independently the active and reac­
tive power on the grid point M - see Figure 10.

Measurement
and point M

Rotor side drill side

converter controller converter controller

l)Fld control level

Grid
operator*
control *>*tcm

Maximum power tracking

W ind turbine control level

Figure 10: Ch’eraU control system of variable speed wind turbine with doubly-fed induc­
tion generator.

12 Ris0-R-15OO(EN)
 Grid side converter controller - controls the DC link voltage UDC and guaran­
tees unity power factor in the rotor branch. The transmission of the reactive
power from DFIG to the grid is thus only through the stator.

The wind turbine control, with slow dynamic response, provides reference signals both
to the pitch system of the wind turbine and to the DFIG control level. It contains two
controllers:
- Speed controller - has as task to control the generator speed at high wind
speeds i.e. to change the pitch angle to prevent the generator speed becoming
too. At low wind speeds, the pitch angle is kept constant to a optimal value.
- Maximum power tracking controller - generates the active power reference
signal for the active power control loop, performed by the rotor side converter
controller in DFIG control level. This reference signal is determined from the
predefined characteristic P-05 look-up table, illustrated in Figure 11, based on
filtered measured generator speed. This characteristic is based on aerodynamic
data of the wind turbine's rotor and its points correspond to the maximum aero­
dynamic efficiency.
Both the speed controller and the maximum power-tracking controller are active in the
power limitation strategy, while only maximum power tracking controller is active in the
optimization power strategy. In the case of high wind speeds there is a cross coupling
between these two controllers.
Figure 12 illustrates explicitly the speed controller, the maximum power tracking con­
troller and the rotor side converter controller. The grid side converter controller is out of
interest in the present study and therefore it is shown as a “black box”.

[rpm]

Figure 11: Power-omega characteristic used in the maximum power-tracking controller.

Ris0-R-15OO(EN) 13
 <•> Speed controller * 0

T ransmission

scheduling

Generator

Wind turbine control level DFIG control level

Maximum power tracking Rotor side converter controller-active power P

-> M

Rotor side convertor controller - reactive power Q

Rectlve power control

Grid side
,\o „ converter
controller

Figure 12: Wind turbine control level.

In the rotor side converter control, there are two independent control braches, one for the
active power control and the other for reactive power control. Each branch consists of 2
controllers in cascade. Only the active power control loops are relevant in the present
study.
Notice that all these controllers, except maximum power tracking controller, are basi­
cally PI controllers. Both the speed controller and the maximum power tracking has as
input the filtered measured generator speed. The generator speed is measured and a low-
pass filter is used to avoid that the free-free frequency in the transmission system is am­
plified through the control system.

Speed controller implementation and design

The error between the filtered measured generator speed and the rated generator speed is
thus sent to the PI speed controller. The output of this Pi-controller is used as reference
pitch signal d ref to the pitch system. To get a realistic response in the pitch angle con­
trol system, the servomechanism model accounts for a servo time constant Tsen,a and the
limitation of both the pitch angle (0 to 30 deg) and its gradient (+10 deg/s). Thus the
reference pitch angle 6 ref is compared to the actual pitch angle 9 and then the error Ad
is corrected by the servomechanism.
A gain scheduling control of the pitch angle is implemented to compensate for the exist­
ing non-linear aerodynamic characteristics.

14 Ris0-R-15OO(EN)
The total gain of the system in the speed control loop Ksystem, can be expressed as a pro­
portional gain KPI in the PI controller times the aerodynamic sensitivity of the system
dP_

-i-l

dP dP_
(1.6)

dP
The aerodynamic sensitivity---- of the system depends on the operating conditions (the
d0
set-point power value, the wind speed or the pitch angle). The more sensitive the system
is (larger pitch angles 9 / higher wind speeds) the smaller the gain for the controller
should be and vice versa.

The total gain of the system Ksystem is thus kept constant, by changing Kpi in such a way

that it counteracts the variation of the aerodynamic sensitivity by the reciprocal

dP
sensitivity function

dP
^PI ^basis (1.7)
~de

where Kbams is the constant designed proportional gain of the PI- controller. The varia­
tions of the aerodynamic sensitivity with the pitch angle, implemented in Dig SILENT
and Flex5 are illustrated in Figure 13. Note that an offset of 0.1 MW/deg is present be­
tween these two curves. The reason is that in DIgSILENT the dynamic inflow is not
taken into account.

DlgSILEI
Flex 5

Figure 13: Sensitivity function dP/d0 versus pitch angle 0.

Ris0-R-15OO(EN) 15
Assuming that the dynamic behaviour of the system can be approximated with a second
order system behaviour, then the parameter design of the PI speed controller can be
based on the transient response analysis for a second order system:

K a 2Q
(1.8)
I s + D s + K + 2 £ aQQ s + a
where I, K and D denote the system inertia, stiffness and damping, respectively. Hence

the natural frequency and the damping ration can be expressed as: a q = . and

1 D D
£= aQ, respectively.
2 Ia q 2K

In the literature, two main parameterizations of the PI controller are applied:

PI controller - parameterization 1:

y = (K„ +—) u (1.9)

where the proportional and integral gain can be expressed as (cf. Section 4):
t-i-i
K = -K °q dP_
d#
and KT 5 2 £ (1.10)
N 30 l aQ
gear ^

In this PI controller parameterization form, the integral gain Ki is proportional to the

stiffness K of the system, while the proportional gain Kp is proportional with the damp­
ing of the system D = 2 £ K / oq . Notice that when this parameterization is used, both
controller parameters contain gain scheduling, i.e. they vary with the reciprocal sensitiv-
' dPn
dP_ -1
ity function
d#

PI controller - parameterization 2:

1
y = Kp (1+—) u (1.11)

where the integral time and the proportional gain are expressed as:
-i-1
T = KP 2 £ 2 £ K^Q dP
T =K and Kp 5 d# (1.12)
aQ aQ n 3Q
gear ^

This parameterization of the PI controller is typically used in control systems. Notice that
here only one parameter, i.e. the proportional gain Kp, has a proportional variation with
dP[ '
the reciprocal sensitivity function while the integral time Ti can be directly
d#_
determined based on design parameters (damping ration £ and natural frequency a0).

16 Ris0-R-15OO(EN)
The implementation in DIgSILENT of the PI controllers is using the parameterization 2.
As design parameters in the controllers, there are used the damping ratio £ = Q.66
and the natural frequency about Q.1 Hz (i.e. a Q = Q.6). Using the expressions (1.6) and
(1.12), the proportional gain Kbasis of the PI- controller can be expressed as:

K = K ^Q
basis _ 3Q (1.13)
a
Q N —
gear n

Based on design data (£ and a0) and on Equations (1.11) and (1.13), the parameters of
the PI speed controller (SC) become: KspC = 85Q (normalized) and TiSC = 2.2.

Active power control (rotor side converter controller)

The active power reference Pggidg’ref signal, determined from the maximum power track­
ing look-up table, is further used in the active power control on the grid, performed by
the rotor side converter controller.
The aim of the rotor side converter is to control independently the active and reactive
power. As mentioned before, the attention, in this report, is mainly drawn to the active
power control. The active power is controlled indirectly by controlling the impressed
rotor current. As illustrated in Figure 12, the active power control, in DIgSILENT, con­
tains two control loops in cascade: a slower (outer) power control loop and a fast (inner)
rotor current control loop. The slower power control loop has thus as output the refer­
ence rotor current signal, which is used further by the fast current control loop. The pa­
rameters of the PI controllers in the rotor side converter are the following:

for the power controller (PC): KP

pC = Q.1 and TiPC = Q.1

for the current controller (CC): KcpC = 1 and TCC = Q.Q1

Notice that these two control loops existing in the rotor side converter controller and tak­
ing care of the active power on the grid, are not implemented in HAWC and Flex5.

2.4 Simulation results

To evaluate the performance of variable speed wind turbine controller implemented in
power system simulation DIgSILENT a set of step response simulations with determinis­
tic wind speed (no turbulence, no tower shadow) are performed. Typical quantities ver­
sus of time are shown: wind speed in [m/s], pitch angle in [deg], generator speed in
[rpm] and the generator power in [MW]. For comparison purpose, the scales of the fig­
ures have been kept similar to those used in the FLEX5 simulations. Unless it is speci­
fied, the simulations are performed using the parameters of the PI controllers, mentioned
before.
In Figure 14, Figure 15 and Figure 16, the wind speed is stepped-up with 1m/s every
each 2Q s.

Ris0-R-15OO(EN) 17
 14 00 |-
12.00 -

____r

Const wind Model: Wind speed [m/s)

2600 -

■1 000

Pitch angle control: Pitch angle |deg)

1560 -

1080 -

Maxmu power tracking

PQ Control: Grid power [MW]

Figure 14: Step in windfrom 5 m/s to 12 m/s.

In the power optimization strategy (at low wind speeds less than 12 m/s), as illustrated in
Figure 14, the speed controller is passive, keeping the pitch angle constant to the optimal
value (i.e. zero for the considered wind turbine). Meanwhile, the power controller con­
trols the active power to the active power reference signal provided by the maximum
power tracking look-up table. The generator speed is thus continuously adapted to the
wind speed, in such a way that it is extracted maximum power out of wind. Notice that
the response time in the steps is bigger at lower wind speeds than at higher wind speeds.
At higher wind speeds than 12 m/s, illustrated in Figure 15, both the speed controller and
the active power controller are active.

18 Ris0-R-15OO(EN)
 22 00 |-

Consl wind Model: Wind speed [m/s)

Pitch angle control: Pitch angle |deg)

1622 -

Maxmu power tracking

PQ Control: Grid power [MW]

Figure 15: Step in windfrom 12 m/s to 20 m/s.

The steps in the wind speed yield changes in both the pitch angle and the generator
speed. The step response of the pitch angle and generator speed does not present big
overshoots and oscillations (fast in steady state). Remark also that the response of the
generator speed is almost identical over the whole range between 12 m/s and 20 m/s, a
fact that indicates that the gain scheduling in the speed controller is working properly.
The power controller keeps the active power to 2 MW with a small deviation about 1 %.
By changing the parameters of the frequency converter controller, the performance of the
system can be improved. Figure 16 illustrates the influence of the parameter tuning in the
frequency converter controller. A similar simulation to that shown in Figure 15 is shown
in Figure 16, but with a stronger integral action of the controllers in the frequency con­
verter, i.e. Tfc = 0.01 and Tfc = 0.001. This action has as result an error in the con­
trol of the power less than 1 per thousand, while the other signals are unchanged.
Figure 17 and Figure 18 illustrate the simulations where the wind speed is stepped-down
from 20 m/s to 4 m/s with 1 m/s step and 2 m/s step, respectively. Figure 17 shows that
as long as the wind speed is higher than 12 m/s, the power is kept constant to 2 MW. The
generator speed is slightly varying around its rated value, while the pitch angle is
changed at each 1 m/s step to another appropriate steady state value. As desired, the
power is stepped out of the rated power (i.e. 2 MW) when the wind speed becomes less
than 12 m/s. The pitch angle is then kept constant to zero, while the generator speed var­
ies according to the P-oi characteristic, shown in Figure 11.

Ris0-R-15OO(EN) 19
 22 00 |-

Consl wind Model: Wind speed [m/s)

Pitch angle control: Pitch angle |deg)

Pitch angle control: _r |. con v. parameters

1622. -

Max mu power tracking model Speed |rpm)

------------- Maxmu pov/er tracking model: Speed |rpml_new_freq.conv parameters

2 010 -

PQ Control: Grid power (MVq

PQ Control: Grid power |MV^_nevi_lreq. conv parameters

Figure 16: Influence of control parameters of the frequency converter.

21.00 t_

Const wind Model: Wind speed [m/s)

Pitch angle control: Pitch angle |deg)

1200 -

Max mu pov/er tracking

0 50 -

PQ Control: Grid power [MW]

Figure 17: Im/s step-down in windfrom 20 m/s to 5 m/s.

However, as it is illustrated in Figure 18, keeping the same control parameters (i.e.
Tfc = 2.2) also for the case when there is 2 m/s step in wind, reveals that the speed
controller is too lazy/slow. Due to the cross coupling between the controllers at high
wind speeds, the power control is then not able to keep the power to the rated power.
Figure 18 also illustrates the simulation when the integral action of the speed controller
is increased (Tfc = 0.5 ). A reduced integral time of the speed controller is correspond­

20 Ris0-R-15OO(EN)
ing to a lower damping ration or a higher natural frequency for the second order system.
It means that to keep the power to the rated power, it is necessary to be able to force the
generator speed to react faster to changes in the wind.
Note that such changing of the integral time of the speed controller yields to an increased
rate of the pitch angle. This can off course imply both higher requirements to the pitch
system and higher aerodynamic loads. As always, the choice of control parameters is and
has to be a trade-off between different aspects. It seems that in the case of higher wind
steps, as it is the case of gusts, the parameters designed through the transient response
has to be slightly changed to be able to handle gusts in a satisfying way.
Figure 19 reflects the same aspects as Figure 18, but for the case of a turbulent wind with
mean speed 20 m/s and turbulence 10 %. By making the speed controller faster, the
power control manages to keep the power to the rated value, and to avoid thus the drops
in the power.
Figure 20 illustrates the influence of better tuning of the parameters in the frequency
converter controller. Notice that the frequency converter alone can improve the control
of the power, without affecting the behaviour of the other signal, as i.e. the pitch angle
and the generator speed.

21,00 t___

10 20 -

Const wind Model: Wind speed [m/s)

Pitch angle control: Pitch angle |deg). Ti*2.2

Pitch angle control: Pitch angle |deg). Ti*0.5

1200 -

- Max mu power tracking model: Speed |rpmj. Ti-2.2

- Maxmu power tracking model: Speed |rpm], Ti*0.5

1 500 -

0 50 -

PO Control: Grid power (MVty. T.-2 2

PO Control: Grid power [MVIfl. Ti=0.5

Figure 18: 2 m/s step down in windfrom 20 m/s to 5 m/s and influence of the speed con­
troller.

Ris0-R-15OO(EN) 21
 Pitch angle control: Pitch angle [deg]_Ti=0.5
Pitch angle control: Pitch angle [deg]_Ti=2.2

---------- Maximo power tracking model: Speed |rpm]_T>0.5

---- Maxima pov/er tracking model: Speed |rpm]_Ti=2.2

PO Control: Grid power |MWLTi=(i,S

PQControl: Grid power [MWl_Ti=2.2

Figure 19: Turbulent wind. Influence of integral time of the speed controller.

------------- Rotor wind Model: v/sftc

22.00
20.00 A------------------A--1---------------------------A^v — yv — A— - -A
18.00 - -c—rtAV kA~ V" Tv'-1 W” - Wr - ■> 1
— li---------- i__iL V----------- 1------------------ -W\-\rTvV 1" ~ IV 7V --/V V------i
16.00
14.00 — — — — — — fV~ A r — — ■
12.00
0.00 60.00 120.0 180.0
------------- Pitch angle control: Pitch angle Ideg]_TI=0.5_nev/_freq.conv_parameters
Pitch angle control: Pitch angle [deg]_7i=0.5_otd_freq.cony_parame1ers

0.00 60.00 120.0 180 0

----------- Maximo power tracking model: Speed |rpm]_Ti"0.5_nev/_lreq.conv_parameters
— Maximu pov/er tracking model: Speed |rpm]_T=0-5_nev/_freq.convi_paramelera

2.018
2.006
1.994
1.982
1-970 I—
60.00 120.0 180,0 240.0 Is] 300,0
PQ Conirot: Grid power (MV^_Ti=0 ,S_ncwJreq.conv_pa ramders
PQ Control: Grid power [MW]_Ti=0.5_old_freq.conv_parairelers

Figure 20: Turbulent wind. Influence offrequency converter.

Figure 21 illustrates a simulation of variable speed wind turbine with turbulent wind,
where the mean is 8 m/s and the turbulence intensity of 10 %.
The generator speed is tracking the slow variation in the wind speed. As expected for this
wind speed range, the pitch angle is passive, being kept constant to its optimal value (i.e.
zero for the considered wind turbine). The active power delivered to the grid does reflect

22 Ris0-R-15OO(EN)
the variation in the wind speed. Notice that the fast oscillations in the wind speed are
completely filtered out from the electrical power.
Figure 22 illustrates the simulation with turbulent wind, where the mean speed is 12 m/s
and the turbulence intensity of 10 %. The power is kept to the rated power as long as the
energy in the wind permits that.

Pitch angle control: Pitch angle |degl_Ti*0.5

------------- Maxmu power tracking I: Speed |rpml_Ti*0.5

PO Control: Grid power [MWLTf=0 5

Figure 21: 8 m/s turbulent wind (turbulence 10 %).

Pitch angle control: Pitch angle |degl_Ti*0.5

Maxmu power tracking I: Speed |rpml_Ti*0.5

PO Control: Grid power [MWLTi=0 5

Figure 22: 12 m/s turbulent wind (turbulence 10 %).

Ris0-R-15OO(EN) 23
3 Controller by AED
3.1 General info of the HAWC code
HAWC [5] is a computer program with the purpose of predicting load response for a
horizontal axis two and three bladed wind turbine in time domain including all aspects
that might be important for the dynamic loads. The program consists of several sub­
packages, each dealing with a specific part of the turbine and the surroundings.
The core of the program is the structural model, which is a finite element model based on
Timoshenko beam elements. The turbine is divided into three substructures: Tower, na­
celle and rotor blades. Each substructure has its own coordinate system allowing for
large rotation of the substructures. There are six degrees of freedom (DOFs) for each
element node, i.e., for at typical wind turbine the total number of DOFs is approx. 250.
The aerodynamic model is based on a modified Blade Element Momentum (BEM)
model. This model has through the years of development evolved from a static frozen
wake method to a dynamic wake formulation including corrections for yawed flow. The
local aerodynamic load is calculated at the blade sections using 2D lift, drag and moment
profile coefficients. Unsteady aerodynamic effects are modelled by a Beddoes-Leishman
type dynamic stall model.
The wind field turbulence model used for load simulations is the Mann model [6]. This
model is a full 3D turbulence field with correlation between the turbulence in the three
directions. The turbulence field is a vector field in space; a so-called frozen snapshot of
the turbulent eddies in the wind. This turbulence is transported through the wind turbine
rotor with the speed of the mean wind speed. It is also possible to use the wind format
from a FLEX wind generator, but this feature has not been used in the present simula­
tions.
The HAWC model is equipped with interface for control systems through a Dynamic
Link Library (DLL) format [7]. This interface enables control of the turbine pitch angles
and generator torque from a control program outside the HAWC core.

The generator model used for these simulations is based on a simple assumption that the
generator torque can be controlled as desired by the regulator, but with a small phase
delay corresponding to a 1st order filter with a time constant of 0.1 s.
The pitch servo is implemented as a 1st order filter with a time constant of 0.2 s and a
max. allowable speed of 20 deg/s. Some extra features regarding offshore foundations
and wave interactions are also sub-packages in the program, but not used for simulations
related to this report.
3.1.1 Test of dynamic wake model in HAWC
To evaluate the dynamic wake model that is implemented in HAWC a few simulation
without turbulence has been performed. The controller for these simulations are a bit
special since it is implemented as an asynchronous generator providing an almost con­
stant rotor speed, and a pitch regulator keeping the power at nominal by adjusting the
pitch angle. After a certain time the pitch angle is increased in one step with 0.5 deg and
back again a little bit later. The dynamic load changes have been plotted in Figure 23 to
Figure 28.

24 Ris0-R-15OO(EN)
 rwake Sa dat Channels: 3 10 4 22 20 19
•jj. 1.97E+00
S 1.97E+00
2. 1.97E+00
1.97E+00
1.97E+00
™ 1.97E+00
° 1.97E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

" 110 1^^

r l) 0.90 ■ ....... ............ ............ . i ■ ... .... ... ... .1... ............ ....7... ........ .............

0.50........ -..........................*. ................1......................... ...............i..... ...... .......................^............................... ..■».....................

J 0.30....................i.................. '.................. ;.................. ................... i..................
o. : 1 1
g ______ ; _________j______ ;____________

10.00 28.33 46.67 65.00 83.33 1 01.67 120.00

167.67
137.18
106.70
76.21
45.73
15.24
-15.24
10.00 28.33 46.67 65.00 83.33 101.67 120.00

-620.39
-709.19
-797.99
-886.79
-975.59
-1064.39
-1153.19
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ -3.35E+00
v, -3.50E+00
£ -3.64E+00
-3.79E+00
I -3.93E+00
m -4.08E+00
>' -4.22E+00
10.00 28.33 46.67 65.00 83.33 1 01.67 120.00

1.26E-01
1.06E-01
8.63E-02
6.66E-02
E 4.70E-02
2.73E-02
7.65E-03
10.00 28.33 46.67 65.00 83.33 1 01.67 120.00
Time [sec]

Figure 23: Simulation at 5 m/s. Step change in pitch angle at constant rotor speed. The
induction changes sloM’ly from one level to another causing momentary load changes
that flatten off when the induction reaches a new equilibrium. From top: Rotational
speed, pitch angle (opposite sign than normal convention), electrical power, blade flap
root moment, induced wind speed longitudinal, induced wind speed tangential.

Ris0-R-15OO(EN) 25
 rwake 7a.dat Channels: 3 10 4 22 20 19
1.98E+00
1.98E+00
1.98E+00
1.97E+00
1.97E+00
1.97E+00
1.97E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

1.10
0.90
0.70
0.50
0.30

c§ -0.10
10.00 28.33 46.67 65.00 83.33 101.67 120.00

612.47
570.38
528.28
486.18
444.09

359.89
10.00 28.33 46.67 65.00 83.33 101.67 120.00

178.83
261.41

426.57
509.15
591.73
674.31
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ -3.43E+00
v, -3.55E+00
M. -3.67E+00
-3.79E+00
1 -3.90E+00
^ -4.02E+00
>' -4.14E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ 5.82E-03
„ -2.34E-02
k -5.27E-02
-8.19E-02
5 -1.11E-01
S' -1.40E-01
>' -1.70E-01
10.00 28.33 46.67 65.00 83.33 101.67 120.00
Time [sec]

Figure 24: Simulation at 7 m/s. Step change in pitch angle at constant rotor speed. The
induction changes slowly from one level to another causing momentary load changes
that flatten off when the induction reaches a new equilibrium. From top: Rotational
speed, pitch angle (opposite sign than normal convention), electrical power, blade flap
root moment, induced wind speed longitudinal, induced wind speed tangential.

26 Ris0-R-15OO(EN)
 rwake 10a.dat Channels: 310 4 222019
1.98E+00
1.98E+00
1.98E+00
1.98E+00
1.98E+00
1.98E+00
1.98E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

1.10
0.90
0.70
0.50
0.30

« -0.10
10.00 28.33 46.67 65.00 83.33 101.67 120.00

1675.56
1622.60
B 1569.64
1516.67
1463.71
°l 1410.75
1357.79
10.00 28.33 46.67 65.00 83.33 101.67 120.00

_ -2047.34
-2113.60
g -2179.86
“ -2246.11
-2312.37
x, -2378.63
S -2444.89
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ -3.35E+00
-3.43E+00
S. -3.51 E+00
-3.59E+00
I -3.67E+00
ig- -3.75E+00
>' -3.83E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ -2.56E-01
„ -2.90E-01 ..................;................... ;.... ,.... ,... :................
!. -3.23E-01 ................. :................... i....
-3.57E-01
= -3.90E-01
S' -4.24E-01
........:................... :...................f................... =...................
>' -4.58E-01
10.00 28.33 46.67 65.00 83.33 101.67 120.00
Time [sec]

Figure 25: Simulation at 10 m/s. Step change in pitch angle at constant rotor speed. The
induction changes slowly from one level to another causing momentary load changes
that flatten off when the induction reaches a new equilibrium. From top: Rotational
speed, pitch angle (opposite sign than normal convention), electrical power, blade flap
root moment, induced wind speed longitudinal, induced wind speed tangential.

Ris0-R-15OO(EN) 27
 rwake 12a.dat Channels: 310 4 222019
1.99E+00
1.99E+00
1.99E+00
1.99E+00
1.99E+00
1.99E+00
1.98E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

-3.75

-4.15

-4.55
-4.75
10.00 28.33 46.67 65.00 83.33 101.67 120.00

2277.01
2208.47
2139.93
2071.39
2002.85
1934.31
1865.77
10.00 28.33 46.67 65.00 83.33 101.67 120.00

-1750.15
-1813.57
-1876.'
-1940.41
-2003.83
-2067.25
-2130.67
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ -1.93E+00
v, -1.98E+00
M. -2.03E+00

-2.09E+00
I -2.14E+00
^ -2.19E+00
>' -2.25E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

-1.21E-01
-1.43E-01
£ -1.64E-01
-1.85E-01
.E -2.06E-01
-2.28E-01
> -2.49E-01
10.00 28.33 46.67 65.00 83.33 101.67 120.00
Time [sec]

Figure 26: Simulation at 12 m/s. Step change in pitch angle at constant rotor speed. The
induction changes slowly from one level to another causing momentary load changes
that flatten off when the induction reaches a new equilibrium. From top: Rotational
speed, pitch angle (opposite sign than normal convention), electrical power, blade flap
root moment, induced wind speed longitudinal, induced wind speed tangential.

28 Ris0-R-15OO(EN)
 rwake 14a.dat Channels: 310 4 222019
•jj. 1.99E+00
J> 1.99E+00
S 1.99E+00
1.99E+00
1.99E+00
S’ 1.99E+00
5 1.99E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

-8.08
-8.28
-8.48
-8.68
-8.88

« -9.28 ^^^^
10.00 28.33 46.67 65.00 83.33 101.67 120.00

2389.79
2302.87
2215.95
2129.02
2042.10
1955.18
1868.26
10.00 28.33 46.67 65.00 83.33 101.67 120.00

-1290.93
-1354.69
-1418.46
-1482.22
-1545.98
-1609.75
-1673.51
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ -1.07E+00
-1.11E+00
S. -1.15E+00
-1.19E+00
I -1.23E+00
S' -1.27E+00
>' -1.32E+00

10.00 28.33 46.67 65.00 83.33 101.67 120.00

-1.34E-02
-2.58E-02
£ -3.82E-02
-5.06E-02
.E -6.30E-02
-7.53E-02
> -8.77E-02
10.00 28.33 46.67 65.00 83.33 101.67 120.00
Time [sec]

Figure 27: Simulation at 14 m/s. Step change in pitch angle at constant rotor speed. The
induction changes slowly from one level to another causing momentary load changes
that flatten off when the induction reaches a new equilibrium. From top: Rotational
speed, pitch angle (opposite sign than normal convention), electrical power, blade flap
root moment, induced wind speed longitudinal, induced wind speed tangential.

Ris0-R-15OO(EN) 29
 rwake 22a.dat Channels: 310 4 222019
1.99E+00
1.99E+00
1.99E+00
1.99E+00
1.99E+00
1.99E+00
1.98E+00
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ -19.14
o) -19.35
2- -19.55
-19.75
J -19.95

ffl -20.36
10.00 28.33 46.67 65.00 83.33 101.67 120.00

2663.30
2530.20
2397.10
2264.01
2130.91
1997.81
1864.71
10.00 28.33 46.67 65.00 83.33 101.67 120.00

-560.08
-620.53
-680.97
-741.41
-801.85
-862.29
-922.73
10.00 28.33 46.67 65.00 83.33 101.67 120.00

-1.90E-01
-2.17E-01
-2.44E-01
-2.71E-01
-2.98E-01
-3.26E-01
-3.53E-01
10.00 28.33 46.67 65.00 83.33 101.67 120.00

„ 3.22E-02
3.06E-02
k 2.90E-02
2.75E-02
5 2.59E-02
S' 2.43E-02
>' 2.27E-02
10.00 28.33 46.67 65.00 83.33 101.67 120.00
Time [sec]

Figure 28: Simulation at 22 m/s. Step change in pitch angle at constant rotor speed. The
induction changes slowly from one level to another causing momentary load changes
that flatten off when the induction reaches a new equilibrium. From top: Rotational
speed, pitch angle (opposite sign than normal convention), electrical power, blade flap
root moment, induced wind speed longitudinal, induced wind speed tangential.

30 Ris0-R-15OO(EN)
3.2 Turbine modelling
3.2.1 Turbine natural frequencies and mode shapes
The turbine has been modelled based on data delivered by Rise. Natural frequencies and
corresponding damping levels are shown in Figure 29 and Table 1. A vibration mode of
utmost importance for control design is the free-free vibration in the drive train, which is
not shown in the Figures since these calculations are performed with a fixed generator
(corresponding to a clamped brake situation). This vibration mode is primarily the rotor
and generator vibrating in counter phase. A simplified expression to calculate this fre­
quency is:

1 J
t+mg
1
Ffree- free
2n 1
2nFL^^ Kshf

where Ir is the rotational inertia of the rotor, n is the gear exchange ratio, Ig is the rota­
tional inertia of the generator at the high speed shaft, Fedge is the 1st edgewise blade bend­
ing frequency and Kshaft is the torsional stiffness of the drive train system. For the current
turbine this frequency is calculated to 1.67 Hz (corresponding to 5.3P).

Turbine mode Frequency [Hz] Damping [log. decr. %]

1st tower long. bending 0.34 2.94
1st lateral tower bending 0.35 2.93
1st shaft torsion 0.61 3.58
1 yaw 0.94 2.47
1st tilt 1.01 2.74
1st symmetric flap 1.21 3.53
1st horizontal edge 1.68 3.03
1st vertical edge 1.77 3.36
2nd yaw 2.39 6.51
2nd tilt 2.47 8.18
Table 1: Frequency and corresponding damping at standstill calculated with
a4%raab.

Ris0-R-15OO(EN) 31
Figure 29: Mode shapes at stand still for 2 MW test turbine calculated with HA WC. See
list offrequencies and corresponding damping levels in Table 1.

3.3 Controller design

The turbine controller is basically a PI-regulator that adjusts the pitch angles and genera­
tor on basis of measured rotational speed. The controller diagram can be seen in Figure
30 In this Figure different parts of the regulation can be seen. In basic the regulator con­
sist of a power controller and a pitch controller. The power controller is adjusting the
reference power of the generator based on a table look up (2) with input of the rotor
speed. This table has been created on basis on quasi-static power calculations. In this

32 Ris0-R-15OO(EN)
table the overall characteristics of the turbine control regarding variable speed, close to
rated power operation and power limitation operation are given.
In the power regulation it is important to avoid too much input of especially the free-free
drive train vibration. If not taken into consideration, this vibration could very well be
amplified through the power control creating very high torque oscillations.
Therefore a special band stop filter (1) has been applied to filter out this vibration fre­
quency. For the current turbine the filter is implemented as a 2nd order Butterworth filter
with center frequency at the free-free eigenfrequency.
To limit the aerodynamic power to the turbine a Pi-regulator (3) is applied that adjusts
the reference pitch angle based on the error between rotational speed and rated speed.
The Pi-regulator has a minimum setting of zero deg, which makes the turbine operate
with zero pitch angles at low wind speeds. The proportional and integral constants Kp
and Ki are set to respectively 1.33 and 0.58, which corresponds to a frequency of 0.1 Hz
and a damping ratio of 0.6.
A special gain scheduling (5) to adjust for increased effect of pitch variations at high
wind speed compared to lower speeds is applied. This gain scheduling handles the linear
increasing effect of dP/d 9 with increasing wind speed and pitch angle. This gain func­
tion is implemented following the expression gain(Q) = 1/(1+6 IKK), where the value KK
is the pitch angle where the gain function shall be 0.5 (cf. Section 4).
To increase the gain when large rotor speed error occurs another gain function (4) is ap­
plied to the PI-regulator. This gain function is simply 1.0 when the rotor speed is within
10 % of rated speed and 2.0 when the error exceeds 10 %. This very simple gain function
seems to limit large variations of the rotational speed at high wind speed.
The pitch servo is modelled as a 1st order system with a time constant of 0.2 s. The
maximum speed of the pitch movement is set to 20 deg/s.
The generator is modelled as a 1st order system with a time constant of 0.1 s, which
means that the reference torque demand from the regulator has a little phase shift before
executed.

torot Rs.ref

'error 0 col, ref

Gain 1

Gain 2
^col

(5)

Figure 30: Control diagram of regulator. Input is measured rotational speed and outpnit
is reference electrical power to generator and reference collective pitch angle to pitch
sen’o.

Ris0-R-15OO(EN) 33
A drive train filter has been included in the generator model to decrease response at the
free-free drive train frequency. This filter makes sure that any torque demand at the free-
free frequency is in counter phase before entering the structural model.

3.4 Results
3.4.1 Step response
To evaluate the behaviour of the controller a step response simulation has been per­
formed. The wind speed is increased for every 20 s with 1 m/s. There are no turbulence,
tower shadow or wind shear. Only the tilt angle of 6 deg contributes to varying aerody­
namic during a rotor revolution.
In Figure 31 to Figure 34 the results of the step simulation are shown. At low wind speed
the controller operates with constant pitch angle of zero degree and varying rotational
speed. It takes quite some time (app. 30 s.) for the rotational speed to change from level
to another. The result of this is that in a turbulent situation the rotor speed will never be
optimal. Since the consequence of a too low rotor speed is a much low power output than
if the rotor speed was a little too high it might be a good idea to change the rotor power-
speed characteristic in the controller to a bit lower values of demanded power than the
optimal values calculated from quasi static analysis. This will make the rotor speed a
little higher than the optimal value and the average power output in a turbulent situation
will increase a little.
At high wind speeds the rotational speed vary around the rated speed. The response of
the rotational speed is almost identical whether the wind speed is stepping from 12 m/s
or 20 m/s which indicates that the gain scheduling in the controller is good. The response
of the tower in the longitudinal is highly damped which is not the case for the lateral di­
rection. The loads are however largest in the longitudinal direction. The cyclic loads on
the rotor originating from the 6 deg tilt angle can mainly be seen in the flapwise bending
moment, the rotational bending at hub, and the tower top tilt moment. The effects of
these increase with wind speed. For the yaw moment there is a sign change in the mo­
ment direction close to rated power. At low wind speed the yaw moment is very small
though negative where the yaw moment increased as the pitch angles increases. The rea­
son for the significant level of yaw moment at high wind speed is the lift forces on the
pitched blades have a moment arm around the tower centerline, which increases with
pitch angle.

34 Ris0-R-15OO(EN)
 rpe5step.dat Channels: 6 3 4 10 49 50
12.70
11.30

25.00 50.00 75.00 100.00 125.00 150.00

„ 2.03E+00
jn 1.91E+00
1 1.80E+00
1.69E+00
1.57E+00
CD
S’ 1.46E+00
o 1.34E+00
0.00 25.00 50.00 75.00 100.00 125.00 150.00

2185.15
1787.85
1390.55
993.25
595.95
198.65
-198.65
25.00 50.00 75.00 100.00 125.00 150.00

„ 1.00E-01 [
HI B.67E ”0 2.........................-............-i.......................................i....................................... f...................................... i.......................................i.......................................
(D : : : : :
2. 3.33E-02........................ :........................:........................ :........................ :........................:........................
^ o.ooe+oo ------------------ i-------------------i------------------- 1------------------- i------------------ i------------------
/V -3.33E-02........................ :........................:........................ ;........................ :........................:........................
J -6.67E-02........................ |........................ |........................ |........................ I........................ |........................
™ -1.00E-01 I ; ; ; ; ;
0.00 25.00 50.00 75.00 100.00 125.00 150.00

2.95E-03
-2.21 E-03
-7.37E-03
-1.25E-02
-1.77E-02
-2.29E-02
-2.80E-02
25.00 50.00 75.00 100.00 125.00 150.00

3.72E-01
3.03E-01
2.34E-01
1.65E-01
9.58E-02
2.68E-02
-4.21 E-02
25.00 50.00 75.00 100.00 125.00 150.00

Figure 31: Step response with steps from 5 m/s to 11 m/s. From top: wind speed, rota­
tional speed rotor, electrical power, pitch angle (opposite sign than normal convention),
toM’er top deflection lateral, toM’er top deflection longitudinal.

Ris0-R-15OO(EN) 35
 rpe5step.dat Channels: 18 27 28 42 44 17
-560.61
-922.68
-1284.74
-1646.80
-2008.87
-2370.93
-2732.!
25.00 50.00 75.00 100.00 125.00 150.00

_ 167.28
110.09
g 52.89
-4.30
-61.49
x, -118.69
S -175.88
0.00 25.00 50.00 75.00 100.00 125.00 150.00

150.00

150.00

0.00 25.00 50.00 75.00 100.00 125.00 150.00

25.00 50.00 75.00 100.00 125.00 150.00

Figure 32: Step response with steps from 5 m/s to 11 m/s. From top: Flap root moment,
rot. bending moment at hub, driving torque at hub, tilt moment at toM’er top, yaw moment
at toM’er top, angle of attack in 3/4 radius.

36 Ris0-R-15OO(EN)
 rpe5step.dat Channels: 6 3 4 10 49 50
26.30
23.70
1 21.10
18.50
15.90
=-| 13.30
10.70
150.00 200.00 250.00 300.00 350.00 400.00 450.00

„ 2.01 E+00
.y 2.01 E+00
E_ 2.00E+00
1.99E+00

CO 1.98E+00
™ 1.97E+00
o 1.96E+00
150.00 200.00 250.00 300.00 350.00 400.00 450.00

2021.05
2007.27
1993.50
1979.72
1965.94
1952.17
1938.39
150.00 200.00 250.00 300.00 350.00 400.00 450.00

-1.99E-02
-2.10E-02
-2.20E-02
-2.31E-02
-2.42E-02
-2.53E-02
-2.64E-02
150.00 200.00 250.00 300.00 350.00 400.00 450.00

Figure 33: Step response with steps from 11 m/s to 25 m/s. From top: wind speed, rota­
tional speed rotor, electrical power, pitch angle (opposite sign than normal convention),
toM’er top deflection lateral, toM’er top deflection longitudinal.

Ris0-R-15OO(EN) 37
 rpe5step.dat Channels: 18 27 28 42 44 17
-51.16
-545.43
-1039.69
-1533.96
-2028.22
-2522.49
-3016.75
150.00 200.00 250.00 300.00 350.00 400.00 450.00

150.00 200.00 250.00 300.00 350.00 400.00 450.00

089.21
069.53
049.85
030.18
010.50
990.82
971.14
150.00 200.00 250.00 300.00 350.00 400.00 450.00

_ -1513.12
-1536.08
Z -1559.03
-1581.98
-1604.93
x, -1627.88
5 -1650.84
150.00 200.00 250.00 300.00 350.00 400.00 450.00

537.39
429.31
321.23
213.15
105.08
-3.00
-111.08
150.00 200.00 250.00 300.00 350.00 400.00 450.00

-1.92

150.00 200.00 250.00 300.00 350.00 400.00 450.00

Figure 34: Step response with steps from 11 m/s to 25 m/s. From top: Flap root moment,
rot. bending moment at hub, driving torque at hub, tilt moment at toM’er top, yaw moment
at toM’er top, angle of attack in 3/4 radius.

38 Ris0-R-15OO(EN)
3.4.2 Turbulent simulations
rt5a.dat Channels: 6 3 4 10 49 50
8.02
7.08
1 6.13
5.19
4.24
>] 3.30
2.36
0.00 50.00 100.00 150.00 200.00 250.00 300.00

„ 1.39E+00
1.33E+00
1 1.28E+00
1.22E+00
1.17E+00
S’ 1.11E+00
o 1.06E+00
0.00 50.00 100.00 150.00 200.00 250.00 300.00

274.15
249.16
224.17
199.18
174.19
149.20
124.21
50.00 100.00 150.00 200.00 250.00 300.00

— 1.00E-01
□, 6.67E-02
(D
2. 3.33E-02
^ O.OOE+OO
s' -3.33E-02
J -6.67E-02
ro -1.00E-01
I 50.00 100.00 150.00 200.00 250.00

6.26E-03
" 2.18E-03
E, -1.89E-03
-5.96E-03
-1.00E-02
*l -1.41E-02
-1.82E-02
0.00 50.00 100.00 150.00 200.00 250.00 300.00

1.07E-01
" 9.01 E-02
7.28E-02
5.55E-02
3.82E-02
*1 2.09E-02
3.57E-03
0.00 50.00 100.00 150.00 200.00 250.00 300.00

Figure 35: Simulation at 5 m/s at 10 % turb. intensity. From top: wind speed, rotational
speed rotor, electrical power, pitch angle (opposite sign than normal convention), toM’er
top deflection lateral, toM’er top deflection longitudinal.

Ris0-R-15OO(EN) 39
 rt5a.dat Channels: 18 27 28 42 44 17
-134.80

-1602.05
0.00 50.00 100.00 150.00 200.00 250.00 300.00

998.31

-905.01
0.00 50.00 100.00 150.00 200.00 250.00 300.00

95.95
0.00 50.00 100.00 150.00 200.00 250.00 300.00

-2205.96
0.00 50.00 100.00 150.00 200.00 250.00 300.00

0.00 50.00 100.00 150.00 200.00 250.00 300.00

0.00 50.00 100.00 150.00 200.00 250.00 300.00

Figure 36: Simulation at 5 m/s at 10 % turb. intensity. From top: Flap root moment, rot.
bending moment at hub, driving torque at hub, tilt moment at toM’er top, yaw moment at
toM’er top, angle of attack in 3/4 radius.

40 Ris0-R-15OO(EN)
 rt12a.dat Channels: 6 3 410 49 50
15.65
14.32
12.98
11.65

50.00 100.00 150.00 200.00 250.00 300.00

„ 2.01E+00
jn 2.00E+00
1 1.99E+00
1.97E+00
1.96E+00
CD
S’ 1.95E+00
o 1.93E+00
0.00 50.00 100.00 150.00 200.00 250.00 300.00

2026.28
1999.76
I 1973.25
1946.74
1920.23
tD 1893.72
CL
1867.20
0.00 50.00 100.00 150.00 200.00 250.00 300.00

-2.96
-4.52

-7.64

50.00 100.00 150.00 200.00 250.00 300.00

-2.16E-03
-1.12E-02
-2.03E-02
-2.93E-02
-3.84E-02
-4.74E-02
-5.65E-02
50.00 100.00 150.00 200.00 250.00 300.00

3.32E-01
3.03E-01
2.73E-01
2.44E-01
2.14E-01
1.85E-01
1.55E-01
50.00 100.00 150.00 200.00 250.00 300.00

Figure 37: Simulation at 12 m/s at 10 % turb. intensity. From top: wind speed, rota­
tional speed rotor, electrical power, pitch angle (opposite sign than normal convention),
toM’er top deflection lateral, toM’er top deflection longitudinal.

Ris0-R-15OO(EN) 41
 rt12a.dat Channels: 18 27 28 42 44 17
_ 486.57
-965.44
Z -1444.31
-1923.18
-2402.05
x, -2880.92
5 -3359.78
0.00 50.00 100.00 150.00 200.00 250.00 300.00

_ 1586.43
1084.79
g 583.14
81.50
420.14
*, -921.79
5 -1423.43
0.00 50.00 100.00 150.00 200.00 250.00 300.00

_ 1080.89
1057.87
g 1034.85
“ 1011.84
988.82
>i 965.80
5 942.78
0.00 50.00 100.00 150.00 200.00 250.00 300.00

-90.92
469.01
-847.09
-1225.18
-1603.26
-1981.35
-2359.43
0.00 50.00 100.00 150.00 200.00 250.00 300.00

1602.24
1124.70
647.17
169.64
-307.89
-785.42
-1262.96
0.00 50.00 100.00 150.00 200.00 250.00 300.00

Figure 38: Simulation at 12 m/s at 10 % turb. intensity. From top: Flap root moment,
rot. bending moment at hub, driving torque at hub, tilt moment at tower top, yaw moment
at tower top, angle of attack in 3/4 radius.

42 Ris0-R-15OO(EN)
 rt22a.dat Channels: 6 3 410 49 50
29.64
26.99
24.33
21.68
19.02
16.37
13.71
50.00 100.00 150.00 200.00 250.00 300.00

2.08E+00
2.05E+00
2.01E+00
1.97E+00
1.93E+00
1.89E+00
1.85E+00
0.00 50.00 100.00 150.00 200.00 250.00 300.00

2053.35
2033.93
2014.50
1995.08
1975.66
1956.23
1936.81
50.00 100.00 150.00 200.00 250.00 300.00

-15.73
-17.15
-18.57
-19.99
-21.41
-22.83
-24.25
50.00 100.00 150.00 200.00 250.00 300.00

4.85E-02
" 2.11 E-02
E, -6.25E-03
-3.36E-02
-6.10E-02
*l -8.83E-02
-1.15E-01
0.00 50.00 100.00 150.00 200.00 250.00 300.00

2.64E-01
" 2.20E-01
£ 1.75E-01
1.31E-01
8.63E-02
*\ 4.19E-02
-2.55E-03
0.00 50.00 100.00 150.00 200.00 250.00 300.00

Figure 39: Simulation at 22 m/s at 10 % tnrb. intensity. From top: wind speed, rota­
tional speed rotor, electrical power, pitch angle (opposite sign than normal convention),
toM’er top deflection lateral, toM’er top deflection longitudinal.

Ris0-R-15OO(EN) 43
 rt22a.dat Channels: 18 27 28 42 44 17
1596.88

-3144.98
0.00 50.00 100.00 150.00 200.00 250.00 300.00

3065.67

-2917.18
0.00 50.00 100.00 150.00 200.00 250.00 300.00

730.30
0.00 50.00 100.00 150.00 200.00 250.00 300.00

0.00 50.00 100.00 150.00 200.00 250.00 300.00

2878.61
2051.73
1224.85
397.96
-428.92
-1255.80
-2082.68
0.00 50.00 100.00 150.00 200.00 250.00 300.00

0.00 50.00 100.00 150.00 200.00 250.00 300.00

Figure 40: Simulation at 22 m/s at 10 % turb. intensity. From top: Flap root moment,
rot. bending moment at hub, driving torque at hub, tilt moment at toM’er top, yaw moment
at toM’er top, angle of attack in 3/4 radius.

44 Ris0-R-15OO(EN)
4 Controller by DTU
4.1 Short description of the FLEX5 code
FLEX5 is a computer program designed to model the dynamic behaviour of horizontal
axis wind turbines operating in specified wind conditions including simulated turbulent
wind. The program operates in the time domain and produces time-series of simulated
loads and deflections.
The structural deflections of the turbine are modelled using relatively few but carefully
selected degrees of freedom using shape functions for the deflections of the tower and
the blades while using stiff bodies connected by flexible hinges to model the nacelle,
rotor shaft and hub.
The aerodynamic loads on the blades are calculated by the normal Blade-Element-
Momentum method modified to include the dynamics of the induced velocities of the
wake. This dynamic wake model is important for a realistic prediction of the response
due to pitch angle changes and hence for the design of the control system of a pitch con­
trolled wind turbine.
The properties of the pitch servo, the generator and the control system are modelled in
separate modules of the program and can be selected from a library of standard models
or can be programmed by the user. These models will be described in more detail below.

4.2 Turbine data

The structural and aerodynamic data for the generic 2 MW variable speed, pitch con­
trolled wind turbine have been provided by Rise in the format designed for the HAWC
code, see Appendices A to D.
While the airfoil data can be used directly, the structural data have been used to produce
an equivalent FLEX5 input file with properties that result in a turbine model with the
same primary natural frequencies as the HAWC model. These frequencies were provided
by Rise in addition to the HAWC data files.

4.3 Pitch servo model

The pitch servo is modelled to be following a setpoint angle signal from the control sys­
tem with a simple 2.order response with specified resonance frequency and damping.
This will simulate a hydraulic servo system with position feedback with sufficient accu­
racy for the present purpose. The frequency and damping ratio have been chosen as 0.8
Hz and 0.8 (relative to critical). These values produce a good fit to the actual pitch servo
response of the old 2 MW Tj$reborg turbine.

4.4 Generator model

The FLEX5 generator model is used to prescribe the generator air gap torque while the
inertia of the generator rotor is part of the turbine structural model. Any losses (mechani­
cal and electrical) are simulated by adding an equivalent torque to the simulated airgap
torque.
The generator model used for this case is a simplified model of a variable speed genera­
tor where the electrical output is assumed to be controlled by some unspecified generator
controller in direct response to a power setpoint specified by the wind turbine control
system. No attempt has been made to simulate the actual electrical properties of the gen­

Ris0-R-15OO(EN) 45
erator or the grid connection. The generator torque is simply calculated as the setpoint
power + losses divided by the generator rotational speed. However, to avoid instability
of the free-free torsional mode of the rotor shaft when operating at constant power, the
rotational speed signal is put through a 2.order low pass filter before the division is car­
ried out. The filter frequency is chosen as 0.8 Hz with a damping ratio of 0.7, which at
the free-free shaft frequency of 1.65 Hz provides sufficient attenuation and phase lag to
avoid the instability.
Also, the setpoint power signal is put through a l.order filter with a time constant of 0.5
s. This is done to avoid too abrupt power changes due to steps in the power setpoint.
The combined losses (mechanical and electrical) are assumed to be 2%, 4% and 8% at a
load of 0%, 50% and 100% respectively (parabolic interpolation).

4.5 Control system

The FLEX5 control system model for the variable speed, variable pitch turbine simulates
two simple controllers: a generator power versus rotor speed controller and a PI-pitch
angle control of the rotor speed. The two controllers are designed to work independently
in the below rated and the above rated wind speed range respectively.
The generator power controller specifies the power setpoint for the generator as a
tabulated function of the generator speed. The speed signal input is the lowpass-filtered
RPM-signal used in the generator model. The tabulated power vs. generator speed func­
tion for the 2 MW turbine is shown in Figure 41.
The data for the lower part of the curve up to 800 kW / 1500 RPM has been chosen in
such a way that the generator torque controls the rotor speed to be approximately propor­
tional to the wind speed up to about 8.5 m/s with a tip speed ratio of 8.7 (optimum).
Above 9 m/s the rotor speed would become too large if the constant tip speed ratio were
continued. Therefore a much steeper, linear slope is introduced up to the rated power of
2 MW at 1580 RPM. This part will control the rotor speed for wind speeds between 8.5
and 12 m/s (rated). The control mode corresponds to the operation of a constant speed
turbine with a large slip asynchronous generator. If the turbine accelerates to above 1580
RPM due to wind speed above 12 m/s (rated) the generator power is kept at a constant 2
MW in which case there is no longer any control of the RPM from the generator torque.

1000 1400

Figure 41: Generator power setpoint as function afgenerator speed.

46 Ris0-R-15OO(EN)
This is where the pitch control takes over.
The pitch angle controller is a simple Pi-controller with the generator speed (low pas
filtered) as input and the pitch servo set-point as output. The reference speed of the con­
troller is 1600 RPM, which is sufficiently above the speed for maximum generator power
to allow for fluctuations around the reference (target) speed without interference from
changes in the generator power. This makes the design of the Pi-controller very simple.
If we assume that the turbine is stiff so that shaft rotation as the only degree of freedom it
is possible to derive the following 2nd order differential equation for the shaft rotation
angle when the PI-control is active and the generator operates at constant power:

Ip + Dp + Kp = 0 (3.1)

where p = Q-Q0 is angular velocity of the rotor speed difference between actual and
reference speed. The parameters of the 2nd order differential equation are

rotor + N gear I generator

1 ( dP 30
D= KpNgear
P gear (3.2)
n« { " dd n
1( ) ir 30
K= KINgear
I gear
Q01 " dd n
Within the limits of the simplifying assumptions it can be seen that the PI-control system
will have a 2nd order response with the following resonance frequency and damping ra­
tio:

(o0 =JK and D D

Z (3.3)
2I#o 24ki

Experience shows that a satisfactory response is obtained by choosing the frequency and
damping as follows:

(90 = 0.6 rad/s and Z = 0.6 - 0.7 (3.4)

The corresponding values of the constants KP and KI of the PI-controller can now be
calculated when the pitch sensitivity dP/dd is known. This quantity must be calculated
from the aerodynamic properties of the rotor. For a number of wind speeds above rated
ordinary static BEM-calculations are performed to determine the pitch angle that pro­
duces the rated mechanical power at each wind speed at the reference RPM. At each of
these operational situations the derivative dP/dd is calculated by doing two new calcula­
tions with small changes of the pitch angle to each side but still using the same induced
velocities as the central calculation (frozen wake assumption).
The calculated pitch sensitivity will show a large variation with wind speed (numerically
increasing with increasing wind) so constant values of KP and KI will not give the de­
sired result because of the change in total gain. However, if we plot the pitch sensitivity
as a function of the pitch angle, we will find a nearly linear relation. This implies that a
gain correction factor GK(d) = 1/(1+d /KK) multiplied with both KP and KI will solve
the problem and produce nearly constant gain over the range of relevant wind speeds.
The constants KP and KI are now calculated using the value of dP/dd corresponding to
0 deg. pitch while the parameter KK is determined as the pitch angle where dP/dd has

Ris0-R-15OO(EN) 47
increased by a factor of 2. This type of gain correction factor has been used with success
for both the old Nibe B and Tj$reborg turbines.
Regarding the present 2 MW turbine the following values have been found according to
the method above. These values are used in the following FLEX5 simulations.
KP = 0.14 deg/RPM KI = 0.06 deg/s/RPM KK = 6.2 deg
The pitch setpoint provided by the PI-controller is the sum of the proportional term and
the integral term. The values of the integral term alone and the total sum must be limited
to a certain minimum angle to prevent the integral term from saturating and to prevent
the pitch angle from hitting the physical stops during operation in wind speeds below
rated. The FLEX5 implementation allows this minimum angle to be prescribed as a func­
tion of the wind speed at the hub averaged over time by a 1 .order filter with a specified
time constant. For the test examples below the time constant is only 6 s. It would nor­
mally be 30 to 60 s. This minimum angle will be the angle at which the turbine operates
at winds below rated. The minimum angel is specified as a tabulated function of the wind
speed. Below 10 m/s the angle is 0 deg. From 10 to 12 m/s the angle increases to +1 deg.
and further to +6 deg. at 16 m/s and above. The idea behind this is to limit the maximum
blade load and prevent potential stalling of the blades when operating in turbulence
around rated wind speed.
One further feature has been implemented: When the turbine operates in turbulence
above rated wind speed the rotor speed will vary somewhat around the reference RPM
and might sometimes even deviate below the RPM for max generator power. This will
result in a short, but possibly deep reduction of the generator power even if there is suffi­
cient wind available to produce max. power continuously. This is avoided to a large de­
gree by including a conditional statement for the power controller: if the pitch angle is
more than 1 deg. above the specified minimum angle at the current wind speed the power
setpoint should be kept at max. power independent of the instantaneous RPM. The result
is a much improved power quality (fewer dips) at the expense of some short term over­
loading of the generator and gearbox (larger torque due to lower RPM).
Finally, please note that the inevitable sampling and processing delays are not simulated.
This could lead to somewhat optimistic results regarding the controller performance.

4.6 Results
Eight test examples have been simulated with the FLEX5 code to illustrate the perform­
ance of the control system. The results are shown as time series plot of selected “sen­
sors” in the figures on the following pages. The two time-series plotted at the bottom of
each figure show the tower top deflections in the longitudinal (wind) direction (L) and in
the transverse direction (T)
The first 3 simulations show the response to idealised steps in the wind input with no
turbulence. The following five figures show the results from simulations with turbulent
wind at 5 different mean wind speeds with a turbulence intensity of 10%.
Even though the control system is only a first try with no detailed optimisation done, the
results show a very well behaved response.

48 Ris0-R-15OO(EN)
 80m 2MW VS, testO, V=4 -> 12 22.10.03 17:34

Vhub (m/s) Wind speed

Tetap(deg) Pitch angle (B1 )

2.0--
1.5--

1.0--
0.5--

OmGen (RPM) Generator speed

1800
1600--
1400-
1200-
1000-'
800 — -

Power (el.)
2000

1500-

1000-'
500-'

Tower mode L

0.4-'

0.3-

0 2-
0.1-'

Tower mode T
o.oo
-0.01-
-0.02-
-0.03-

-0.04-

-0.05

Figure 42: Step responses for wind speed steps from 4 - 12 m/s.

Ris0-R-15OO(EN) 49
Figure 43: Step responses for wind speed steps from 12 - 20 m/s.

50 Ris0-R-15OO(EN)
 22.10.03 17:38

Vhub (m/s) Wind speed

Tetap(deg) Pitch angle (B1 )

OmGen (RPM) Generator speed

1800

1600

1400--

1200-

1000-

Power (el.)
2500

2000

1500-

1000-

500-

Tower mode L

0.4--

0.2-
0.0-'
-0.2

Tower mode T
o.oo
-0.01-

-0.02-

-0.03-
-0.04-
-0.05
-0.06

Figure 44: Step responses for wind speed steps from 20 - 4 m/s.

Ris0-R-15OO(EN) 51
 23.10.03 16:38

Vhub (m/s) Wind speed

Tetap(deg) Pitch an
o.oio
0.005--

0.000
-0.005--

-0.010
OmGen (RPM) Generator speed
1550

1500--

1450-'

1400-

1350-

1300

Power (el.)

900--

800-

700-

600-

GX7 ( m ) 1 Tower mode L

0.30
0.28-
0.26-
0.24-
0.22-
0.20--
0.18--
0.16

GX9 ( m ) 1 Tower mode T

-0.005

-0.010-
-0.015

-0.020-

-0.025-

-0.030

Figure 45: Responses at mean wind speed of 8 m/s and turbulence intensity of 10 %.

52 Ris0-R-15OO(EN)
 23.10.03 15:35

Vhub (m/s) Wind speed

Tetap(deg) Pitch angle (B1 )

0.6-
0.4 —

0.2-

OmGen (RPM) Generator speed

1580

1560-

1540-

1520-

1500

Power (el.)

1600 —

1400-

1200-

1000-

GX7 ( m ) 1 Tower mode L

0.40

0.35-

0.30-

0.25

GX9 ( m ) 1 Tower mode T

-0.01

-0.02-

-0.03-

-0.04-

-0.05

Figure 46: Responses at mean wind speed of 10 m/s and turbulence intensity of 10 %.

Ris0-R-15OO(EN) 53
Figure 47: Responses at mean wind speed of 12 m/s and turbulence intensity of 10 %.

54 Ris0-R-15OO(EN)
 80m 2MW VS, V=15, ti = 0.1 23.10.03 15:50

Vhub (m/s) Wind speed

Tetap(deg) Pitch angle (B1 )

OmGen(RPM) Generator speed

1640

1620-

1600

1580-

1560

Power (el.)
2020

2010-

2000

1990-'

1980

GX7 ( m ) 1 Tower mode L

0.40
0.35-
0.30-
0.25
0.20-'
0.15-
0.10

GX9 ( m ) 1 Tower mode T

o.oo
-0.02-
-0.04-

-0.06-

”0.08-

-0.10

Figure 48: Responses at mean wind speed of 15 m/s and turbulence intensity of 10 %.

Ris0-R-15OO(EN) 55
Figure 49: Responses at mean wind speed of 20 m/s and turbulence intensity of 10 %.

56 Ris0-R-15OO(EN)
5 Reduced model of DFIG generator dynamics
This chapter describes the dynamics of the doubly-fed induction generator, derives a
strongly reduced model of the generator including converter and control, and finally de­
rives the parameters of the reduced model based on generator data sheet information, and
compares the parameters to estimates based on simulations with the detailed DIgSILENT
model described in Chapter 2.
The dynamics included in the reduced model represents the power controller of the gen­
erator, whereas the generator itself including the rotor current control is modelled in
steady state. This is possible because the response from rotor current reference as input
to torque as output is very fast. The rotor speed is assumed constant in the first place, but
it is straightforward to extend the model with the dynamics due to the generator inertia,
because the generator torque is output of the model.
The idea is to provide a simple generator model, which is useful in design models for
wind turbine controllers. The model can also be applied in aeroelastic models of wind
turbines.

5.1 Generator dynamics

The generator dynamics dealt with in this chapter includes the DFIG control level shown
in Figure 12 in addition to the generator. Only the active power dynamics are dealt with
here, because reactive power only has marginal influence on the mechanical part.
Figure 50 shows the power control of the doubly fed induction generator with a few ad­
ditions to Figure 12. Some of the variables are in p.u. values, which is often used in elec­
trical systems. The p.u. value is defined as the ratio between the physical value and a
certain base, which is often a rated value. Thus, it is of course essential to use the correct
base.
As mentioned in Chapter 2, the controller is a cascade of two PI controllers. The first PI-
controller (the power controller) is the slowest, with an integration time TiP = 10 ms in
the DIgSILENT model. It controls the power Pgrid to the reference value Pgridref set by the
wind turbine controller. The integration time can be changed like the rest of the control­
ler dynamics, because this part of the model is completely open to the DIgSILENT user,
but the PI controller with TiP = 10 ms has shown to give stable response.
The output of the power controller is the reference of the q component iqref of the rotor
current, which is given in p.u. in the stator flux coordinate system. The stator flux coor-
Mechanical
system

p-u-

PI PI Gen Grid

Figure 50. Block diagram for power control of doubly-fed induction generator.

Ris0-R-15OO(EN) 57
dinate system is defined with d-axis collinear with the stator flux linkage, which rotates
synchronously with the power system frequencyf0=50 Hz.
The reason to chose the stator flux coordinate system is that the cyclic (sinusoidal) varia­
tions of the electric variables (voltages, currents, flux) are removed. Also, in the stator
flux coordinate system, which is very close to the stator voltage coordinate system, a
change of the current in the q direction mainly results in changed active power, while a
change in the current in the d-direction mainly causes changed reactive power.
The second PI-controller is the rotor current controller. It controls the rotor current iqmeas
to the reference value iqref with an integration time Tii = 1 ms. The output of the rotor
current controller is the modulation factor Pmq, given also in the stator flux coordinate
system. The modulation factors Pmd and Pmq determine the rotor side converter voltage,
and they are defined as the fraction of the dc voltage Udc (see Figure 10), e.g. Uq = Pmq
Udc.
Summarising the control cascade, a power reference is set by the first controller. The
output of that controller is a reference value of the generator rotor current. The second
controller has the rotor current as input and the voltage for the rotor side converter as
output.
The generator in Figure 50 is shown as a black box. Note that the input to the build-in
DIgSILENT generator model is the turbine power pt. The generator inertia J is also built
into the generator model according to (2.4). Thus, the used output form the generator
model is the rotor speed, and not the generator torque.
A “p.u. conv” block is given between the generator block and the mechanical system
model. The purpose of the “p.u. conv” block is to convert the physical unit values in the
mechanical model to the p.u. values in the build-in generator model, and to be able to
keep the mechanical torque Tmec constant rather than the mechanical power pt.

Finally a grid block is shown in Figure 50 to indicate the feedback from the generator to
the measured power. The generator produces stator power Ps as well as rotor power Pr,
and the total grid power Pgrid is measured on the other side of the transformer (see Figure
10).
Figure 51 shows the dynamic response to a stepwise change in power reference, simu­
lated with the DIgSILENT model. The aerodynamic torque is kept constant in the simu­
lation. The response is simulated with the original integration time TiP = 10 ms and with
a ten times slower integration time TiP = 100 ms, to illustrate that the dynamics of the
step response is mainly due to this integration time.

58 Ris0-R-15OO(EN)
 100ms

10ms

Time [s]

Time [s]

Figure 51. Dynamic response to step in power reference.

Figure 52 shows the dynamic response to a stepwise change in the rotor current refer­
ence. In this case, the turbine power pt is kept constant. It is seen that the power response
to the change in rotor current reference is very fast.

Figure 52. Dynamic response to step in rotor current reference.

Ris0-R-15OO(EN) 59
For aeroelastic models, it is of interest how the generator torque responds to changes in
generator speed. For frequencies around 1 Hz and below, the directly connected squirrel
cage induction generator responds approximately to the well-known steady-state torque
characteristics, which increases the torque when the speed is increased, and consequently
damps the free-free oscillations of the generator inertia against the turbine rotor inertia.
Since it is not possible to change the speed as input to the Dig SILENT model, the me­
chanical torque is changed sinusoidal instead, and the torque is plotted vs. the speed as
shown in Figure 53.
Finally, the response to step changes in the modulation factor has been simulated with
the detailed Dig SILENT model. The simulation results are shown in Figure 54. In this
case, the inertia is multiplied by 1000 in the simulations to reduce the influence of speed
changes and focus on the electrical response.

Generator speed [pu]

Figure 53. Generator torque vs. speed for 1 Hz and 5 Hz variations in mechanical
torque.

3 090

Figure 54. Dynamic response to step in modulation factor.

60 Ris0-R-15OO(EN)
There is a coupling between id and uq, which is very weak in Figure 54. Lindholm [9] has
shown that this coupling is proportional to the slip. Lindholm also suggested an algo­
rithm to eliminate the cross coupling, but this has not been implemented in the DIgSI-
LENT model. To minimise the coupling in the example, a working point with a very low
slip (0.05%) has been selected for the simulations in Figure 54.
According to Lindholms results, when the slip is zero, the dynamic equation is of the
rotor in stator flux dq-coordinates is
diq
uq = ^ +04 — (5.1)

where ois the leakage factor

2
L
o = 1- (5.2)
LsLr

With the parameters for the present generator, (5.1) is a first order system with the tran­
sient rotor time constant

Tr —r- = 110 ms (5.3)

Rr

The transient rotor time constant can be seen in Figure 54 in the response of iq and the
power immediately after the step in Pmq. It is observed that the power and torque re­
sponse to a step in the rotor voltage (implemented as a step in the modulation factor) is
much slower than the response to a step in the rotor current reference. Thus, the current
controller speeds up the response considerably. As a result of this speed up, the rotor
voltage has to respond with a transient, which is also seen as a spike in Pmq in Figure 52.
In other words, in Figure 52, we are trying to make a stepwise change in iq, which causes
a transient according to (5.1).
There is plenty of reserve to provide this voltage transient from the converter in the pre­
sent simulation, where the voltage requirement is very low because of the low slip. Op­
eration with higher slips would bring the voltage to the limit of the capacity of the power
converter, but for normal operation of the wind turbine, it is not expected to be a need for
very fast torque response. With the present controller, the slow power controller ensures
that the current reference does not change fast. However, it is an issue to ensure that the
necessary voltage reserves are available in the converter. More simulations with the de­
tailed DIgSILENT model could provide more knowledge on this issue.

5.2 Reduced model

Based on the results in section 5.1, the reduced model neglects the dynamics of the rotor
current control, and consequently assumes a steady state transfer Kgen of the rotor current
Iq = Iqeto the generator torque Tgen as illustrated in Figure 55. For simplicity, the model
does not include the losses. Note that the generator speed, which can be derived from
equation (2.4), is included in the model as a constant in the first place.

Figure 55. Reduced model.

Ris0-R-15OO(EN) 61
Including dynamics only from the PI controller, the system in Figure 51 gets the transfer
function
1 + s I®z
# (s)
1 + s I mp
(5.4)

which has the asymptotic Bode plot shown in Figure 56.

In Figure 56, K is the steady state open loop amplification
K = Kff ' Kgen ' ^gen (5.5)

This dynamic response derived in (5.4) and Figure 56 corresponds very well to the simu­
lated power response in Figure 51. Form Figure 51, the “instantaneous” response has
been estimated to 18.6 % of the DC response. This result together with the amplifications
in Figure 56 is used to estimate the open loop amplification K according to

—— = 0.186 ^ K = 0.23 (5.6)

K 1 + v '

Using this, the time constant of the pole in Figure 56 is estimated to

^ ^ ^^ = 54 _c M

which agrees quite well with the dynamic response in Figure 51.

5.3 Parameter derivation

All parameters in the reduced model in Figure 55 are known except the generator con­
stant Kgen. Kgen can be calculated from (5.6) and (5.5), but the value of Kfound in (5.6) is
based on simulation results with the detailed Dig SILENT model. In this section, the
steady state equations of the generator are used to derive Kgen, and as a result all parame­
ters for the reduced model can be obtained from the generator data sheet and the control­
ler parameters, which should all be available in wind turbine documentation.
According to Leonhard [8], the generator torque Tgen is given as
T-gen = -3L„ Im(f/) (5.8)

where Lm is the mutual inductance, Im(x) denotes the imaginary part of x, and x* denotes
complex conjugated x. The minus sign in (5.8) is not in [8] because motor sign is used in
[8], whereas we use generator sign. In generator operation, the rotor current must be
ahead of the stator current to transfer the torque, and in that case, Im(/s/r*) becomes nega­
tive. The stator flux % is given as

K
K +1

K
T„

Figure 56. Asymptotic Bode-plot of the closed loop transfer from power setpoint to
power output

62 Ris0-R-15OO(EN)
 ^ s = Ls1s + LmK (5.9)

where Ls is the total stator inductance. Inserting Is from (5.9), (5.8) can be modified to
Tgen =-3 LL Im^jI*)
(5.10)

In the stator flux coordinate system, /s is per definition real / = i//sd, while Ir is complex
I = Id + j Iq. Thus, in the stator flux coordinate system, the torque is given as

Tgen = 3^ ^dIq (5.11)

The stator line-to-zero voltage Usl0 is given by

Usl0 = RsIs + ^s + jOe^s (5.12)

where Oe = 2n-50 radls is the angular frequency of the electrical variables. Since the sta­
tor resistance is relatively small, the first term in (5.12) is neglected. Moreover, in steady
state, which is obtained as soon as the rotor current has reached the reference value, the
second term in (5.12) becomes zero. Thus

= |U„„| = Oe N = ^e^,d (5.13)

where Usn is the rated voltage, assuming that the grid is ideal. Inserting /sd from (5.13)
in (5.11) yields

T„, ^ (5.14)

For the sake of completeness, the inductances Lm and Ls are given by the main reactance
Xm and the stator leakage reactance Xs in the generator data sheet according to
Xm = jOeLm
(5.15)
Xj + Xm = j%

Inserting (5.15) in (5.14) yields

K
(5.16)
'^r/q Oe X. + X.

Thus, the steady state transfer Kgen from rotor current q-component Iq in the stator flux
coordinate system to generator torque Tgen, can be assumed independent on the operation
point, and it can be obtained from generator data sheet parameters according to (5.16).
Calculating according to (5.16) yields Kgen = 3.72 kNmIA. Using the estimate in (5.6)
together with (5.5), and applying Kff and Ogen in physical units, Kgen = 3.70 kNmIA is
obtained.
The conclusion of this chapter is that the reduced model seems to represent the dynamics
of the generator very well with a bandwidth of more than 20 Hz, which is a fully suffi­
cient bandwidth of typical aeroelastic design models.

Ris0-R-15OO(EN) 63
6 Numerical optimization of Pl-controller
The numerical optimization code HAWTOPT has been used to investigate whether im­
provement in regulation performance can be obtained using numerical tuning of the con­
trol parameters Kp and Ki in the control algorithm described in Chapter 3. This optimi­
zation code is a multipurpose optimization code that previously has been used for etc.
optimization of aerodynamic blade layout, redistribution of blade stiffness and inertia for
minimization of risk of edgewise blade vibrations, but until now pitch control parameters
have not been included as optimization parameters, which is the new aspect in this inves­
tigation. Using the optimization code the main control parameters Ki and Kp in the PI-
regulation has been found based on an optimization strategy of minimization of the stan­
dard deviation of the blade flapwise load signal.
The optimization is based on a 300 s time simulation at various wind speeds of power
production with IEC turbulence. The simulation platform is HAWC even though Flex or
other aeroelastic codes could have been used as well. At each wind speed the optimal
parameters of Ki and Kp have been found. The objective function is minimization of the
standard deviation of the blade root flapwise moment. Constraints of the design variables
were some min. and max. values, but convergence was obtained after within 25 iterations
without reaching these limits. Time consumption for optimisation of Kp and Ki at a sin­
gle wind speed was app. 20 hours on a 3.0 GHz Pentium PC

6.1 Method
HAWTOPT is a general multipurpose optimization code which handles execution of
external programs, changes in input parameters through text files, execution of post
processing programs and evaluation of the results in term of minimization or maximiza­
tion of an object function. Several optimization methods can be chosen with this code.
One method is the simplex optimization, which finds the nearest maxima or minima of
an object function using gradients (the “hill-climbing” strategy). The method is rather
fast but suffers from local maxima problems - the optimizations end when a local optima
is found. Another method is the genetic algorithm [8], which is a completely different
technique, based on ideas of biological gene and chromosome selection. The advantage
of this method is especially that it is not restricted to local maxima, but on the cost of
computational time. The method is computational more expensive than the simplex
method.

64 Ris0-R-15OO(EN)
 Optimisation
algorithm

Update control
settings /

HAWC

processing

Figure 57: Route diagram of the optimization of control parameters using HA WTOPT.

6.2 Result
Before optimization of parameters for all wind speeds some initial exercises were carried
out at a single wind speed. The wind speed chosen was 15 m/s.

Initial exercises at 15 m/s

The first objective was to ensure that the optimization converges towards a solution, as
shown in Figure 58, and that the results of the new parameters seemed reasonable (see
Figure 59 and Figure 60).
The importance of simulation length is investigated in Figure 61, where significantly
different parameters were obtained if only 60 s was used instead of 300 s. The conclu­
sion of this was to use the full 300 s simulation.
In Figure 62 the importance of good initial value guesses was investigated. Optimization
performance in term of iteration numbers was found with initial value parameters using
the design method used in Chapter 3 (Kp=1.33, Ki=0.58), and rather arbitrary values of
Kp=1.0 and Ki=1.0. It seems that both optimizations end up with the same results if
many iterations were carried out, but the number of iterations before convergence was
much less with the design method parameters.
In Figure 63 the choice of optimization method was investigated. This was done using 60
s simulations, which is the reason that the resulting parameters is not identical to the pa­
rameters seen in the previous investigations. The genetic algorithm shows fastest con­
vergence in term of iteration numbers, but as this method in itself is more time consum­
ing per step the simplex algorithm is still the fastest. The converged result of the genetic
algorithm was slightly different than the simplex algorithm, which indicates that the op­
timization problem might be more non-linear than a simplex algorithm can handle.
Never the less the simplex algorithm has still been used for the full 300 s simulation at
every wind speed.

Ris0-R-15OO(EN) 65
 # Iterations

Figure 58.'Convergence reached at app. 20 iterations.

20 40 60 80 100 120 140 160 180 200 220 240 260 280 300
t sec

500 v

-1,500
5 -2,000

-3,000

Figure 59: Time-series offlapwise blade root bending moment with initial control pa­
rameters (top) and with control parameters after minimization of the standard dexiation
of this load (bottom).

66 Ris0-R-15OO(EN)
 2.1

Si 2.05

1.95 I

S 19

o 2.05

1.95

S 19

Figure 60: Time-series of rotor speed with initial control parameters (top) and optimized
control parameters (bottom).

Time series 60 s
Time series 300 s

'A A A A A A * A A A A

Ki = 0.486

Ki = 0.296

# Iterations

Figure 6F Illustration of importance of simulations length in the optimization. At only

60 s simulation the optimized values are significantly different that for 300 s.

Ris0-R-15OO(EN) 67
 # Iterations

Figure 62 Illustration of importance of good initial value guesses. It seems that the two
simulations reaches the same result if more iterations were carried out, but the conver­
gence is obtained much sooner for the good initial values guess.

Figure 63 difference in results as function of different optimization methods. The genetic

algorithm shows best performance for this problem in terms op iterations steps. However
since this method is rather time consuming the simplex algorithm is still the fastest.

68 Ris0-R-15OO(EN)
Optimization at several wind speeds using 300s. simulations.

Based on the initial investigations at 15 m/s a full range of optimizations covering the
above rated wind speeds interval has been performed. The simulation length is 300 s and
the optimization method is the simplex algorithm.
In Figure 64 the resulting Pi-parameters can be seen as well as the object function and in
Figure 65 performance of other sensors than the object function can be seen. What is
interesting to notice is that the tendency in parameter is slightly higher proportional gains
and slightly lower integral gains than what is the result of the design model. These
changes lead to a bit “softer' regulation with higher standard deviation in the rotational
speed. The tower base loading is also slightly reduced.

Initial values
Optimum

V (m/s)

0.800

0.600

5 0.400

0.200 ^ -■-Initial value

Optimum
0.000

V (m/s)

Objf. function

1.000
0.990 - Objf. function

0.980
0.970
0.960
0.950

V (m/s)

Figure 64. Object function and initial and optimized values of Kp and Ki at the different
wind speeds. The maximum improvement is seen to be at 14 m/s with a load reduction of
4 % in the standard dexiation offlapwise blade root moment.

Ris0-R-15OO(EN) 69
 M_f lap
M_towerbase
M towertop

V (m/s)

P elec -

10.96

Omega

□mega

V (m/s)

Pitch

V (ms)

Figure 65: Performance of other sensors than the object function. Values normalized by
the standard de\’iation from the initial value simulations.

70 Ris0-R-15OO(EN)
6.3 Conclusion
The main result of the optimization investigation is that numerical optimisation can be
used to tune controller parameters, especially when the optimization is used as refine­
ment of a qualified initial guess. Another result is that the design model used to calculate
the initial value parameters, as described in Chapter 3, could not be refined much in
terms of performance related to the flapwise blade moment. Numerical optimization of
control parameters is not well suited for tuning from scratch. If the initial parameters are
too far off track the simulation might not come through, or a not representative local
maximum obtained. The last problem could very well be related to the chosen optimiza­
tion method, where more future work could be done.
The improvement in flapwise load was app. 1-2% reduction in standard deviation. As an
extra profit the tower base tilt moment was reduced with 2-3% in standard deviation.
These improvements do not seem large, which is only an indication that the starting
point was close to optimal.
Since the numerical optimisation uses many simulations and thereby large computational
time it is essential to have fast, but still accurate, simulation codes. The present used ver­
sion of HAWC might be in the upper end regarding time consumption for such use.

Ris0-R-15OO(EN) 71
References
[1] Sorensen, P., Hansen, A. D., & Rosas, P. A. C, 2001, “Wind models for prediction
of power fluctuations from wind farms”, Proc. of ‘The Fifth Asia-Pacific Confer­
ence on Wind Engineering (APCWEV)’, Kyoto, Japan, Journal of Wind Engineer­
ing, no. 89, pp. 9-18.
[2] Sorensen, P., Hansen, A. D., Janosi, L., Bech, J., & Bak-Jensen, B., 2001, Simula­
tion of interaction between wind farm and power system, Riso-R-1281, Riso Na­
tional Laboratory.
[3] Hansen, A. D., Sorensen, P., Blaabjerg, F., & Bech, J., 2002, “Dynamic modelling
of wind farm grid connection”, Wind Engineering, 26(4), pp. 191-208.
[4] 0ye, S, 1991, “Dynamic stall - simulated as time lag of separation”, In Proceed­
ings of the 4th IEA Symposium on the Aerodynamics of Wind Turbines, McAnulty,
K. F (Ed.), Rome, Italy.
[5] Petersen, J. T., ”The Aeroelastic Code HawC - Model and Comparisons”, In 28th
IEA Experts Meeting: State of the Art ofAeroelastic Codes, DTU, Lyngby, 1996.
[6] Mann, J., “Wind Field Simulation”, Prob. Eng. Mech., Elsevier Science, Vol.
13(4), pp. 269--283, 1998.
[7] Larsen, T. J., Description of the DLL Regulation Interface in HAWC, Riso-R-
1290(EN), Riso National Laboratory, September 2001.
[8] Goldberg, David E., Genetic Algorithms in search, optimization & machine learn­
ing, Addison-Wesley Publishing Co., Reading, Mass., USA, 1989.

72 Ris0-R-15OO(EN)
Appendix A
HEAD HAWC input for "Aeroelastisk int. styring" /*
data baseret diverse m0ller /*
80m tarnh0jde. kt/mh 23.09.2003 ;
DEFINE_STRING CASE_IDENTIFICATION m01
DEFINE_STRING STRUCTURE_FILE_EXT 001
DEFINE_STRING AERODYN_FILE_EXT 001
DEFINE_STRING PROFILE_FILE_EXT 001
DEFINE_STRING GENERATOR_EFFICIENCY_FILE_EXT 001
DEFINE_STRING GEAR_BOX_EFFICIENCY_FILE_EXT 001
PARAMETERS TURBINE 3 2 1 2 ;
PARAMETERS GEAR_BOX 85.0 2 1 ;
PARAMETERS GENERATOR 150.0 9.99E3 2000.0 0.75 1500.0 1
PARAMETERS TILT 6.00 ;
PARAMETERS HAM DYNAMIC STALL 1.50 6.00 0.75 0.14 *
0.53 0.30 0.70 0.95 1 1
ROTOR BASIC LAYOUT 3 2 0.00 0.00 0.00 *
0.00 0.00 0.00 *
0.00 0.00 0.00 0.0 ;
DAMPING TOWER 0.0 0.0 0.0 3.06E-3 3.06E-3 1. 45E-3 ;
DAMPING SHAFT 0.0 0.0 0.0 1.00E-3 1.00E-3 3. 20E-3 ;
DAMPING BLADE 0.0 0.0 0.0 1.39E-3 9.00E-4 1. 39E-3 ;
DEFINITION AEROCAL_ INDUCTION 32 1. 0 1.0 1.0 0.1 1. 0 *
4. 0 4.0 4.0 0.4 4. 0 *
2. 0 2.0 2.0 2.0 1. 0
; DEFINITION AEROCAL_MODIFICATIONS 1 1 15 ;
DEFINITION WIND_FIELD 1.25 9.00 80.00 0.01 1 ;
DEFINITION TOWER AERODYNAMIC LOAD 0.6;
DEFINITION TOWER_SHADOW 1.200 2.150 1.5 1
; Blade nodes
NODE BLADE 1 1 0.000 0.000 0.000
NODE BLADE 1 2 0.000 0.000 2.000
NODE BLADE 1 3 0.000 0.000 7.000
NODE BLADE 1 4 0.000 0.000 12.000
NODE BLADE 1 5 0.000 0.000 20.000
NODE BLADE 1 6 0.000 0.000 25.000
NODE BLADE 1 7 0.000 0.000 28.000
NODE BLADE 1 8 0.000 0.000 31.000
NODE BLADE 1 9 0.000 0.000 34.000
NODE BLADE 1 10 0.000 0.000 36.000
NODE BLADE 1 11 0.000 0.000 38.000
NODE BLADE 1 12 0.000 0.000 40.000
NODE BLADE 2 1 0.000 0.000 0.000
NODE BLADE 2 2 0.000 0.000 2.000
NODE BLADE 2 3 0.000 0.000 7.000
NODE BLADE 2 4 0.000 0.000 12.000
NODE BLADE 2 5 0.000 0.000 20.000
NODE BLADE 2 6 0.000 0.000 25.000
NODE BLADE 2 7 0.000 0.000 28.000
NODE BLADE 2 8 0.000 0.000 31.000
NODE BLADE 2 9 0.000 0.000 34.000
NODE BLADE 2 10 0.000 0.000 36.000
NODE BLADE 2 11 0.000 0.000 38.000
NODE BLADE 2 12 0.000 0.000 40.000
NODE BLADE 3 1 0.000 0.000 0.000
NODE BLADE 3 2 0.000 0.000 2.000
NODE BLADE 3 3 0.000 0.000 7.000
NODE BLADE 3 4 0.000 0.000 12.000
NODE BLADE 3 5 0.000 0.000 20.000
NODE BLADE 3 6 0.000 0.000 25.000
NODE BLADE 3 7 0.000 0.000 28.000
NODE BLADE 3 8 0.000 0.000 31.000
NODE BLADE 3 9 0.000 0.000 34.000
NODE BLADE 3 10 0.000 0.000 36.000
NODE BLADE 3 11 0.000 0.000 38.000
NODE BLADE 3 12 0.000 0.000 40.000
Nacelle nodes
NODE NACELLE 1 0. 0 0.000 0.000 ; Tower top
NODE NACELLE 2 0. 0 -0.600 0.000 ; Gearbox
NODE NACELLE 3 0. 0 -2.186 0.000 ; Main shaft flange
NODE NACELLE 4 0. 0 -3.486 0.000 ; Rotor centre
Tower nodes
NODE TOWER 1 0 000 0. 000 0. 000 ;
NODE TOWER 2 0 000 0. 000 -20 000 ;
NODE TOWER 3 0 000 0. 000 -40 000 ;
NODE TOWER 4 0 000 0. 000 -60 000 ;
NODE TOWER 5 0 000 0. 000 -76 000 ;
NODE TOWER 6 0 000 0. 000 -80 000 ;
DEFINITION CONCENTRATED MASS 1 5 2 75.3E+3 409.4E+3 20.0E+3 367.7E+3
TYPES BLADE_AERODYN_3 1 1 1 ;
TYPES BLADE_STRUCTURE_3 1 1 1 ;
TYPES SHAFT_STRUCTURE 1 ;

74 Ris0-R-15OO(EN)
 TYPES TOWER_STRUCTURE 1 ;
TYPES USE_CALCULATED_BEAMS ;
INITIAL AZIMUTH 0.000 1.85 0.000 ;

WRITE BEAM_DATA_TO_RDA_DIRECTORY ;
RESULTS BLADE_CO_SYS_FORCE 1 ;
RESULTS BLADE_CO_SYS_DEFORM 1 ;
RESULTS FORCE_MOMENT_SIGN 4 ;
RESULTS RESPONSE_TIME_OFFSET 10.0 ;
RESULTS TIME ;
RESULTS AZIMUTH 0.0 ;
RESULTS ROTOR_SPEED ;
RESULTS ELECTRICAL_POWER ;
RESULTS WIND_SPEED_AT_HUB 1 1 1 1 1 ;
RESULTS PITCH_BEARING_ANGLE 1 1 1
RESULTS BLADE_FORCE_MOMENT 1 1 0 0 0 1 1 0 Rotor centre
RESULTS BLADE_FORCE_MOMENT 1 2 0 0 0 1 1 0 R=1.2m
RESULTS BLADE_FORCE_MOMENT 1 5 0 0 0 1 1 0 r=2 0m
RESULTS BLADE_FORCE_MOMENT 1 9 0 0 0 1 1 0 r=34m
RESULTS SHAFT_FORCE_MOMENT 4 0 1 0 1 1 1 Rotor centre
RESULTS TOWER_FORCE_MOMENT 1 1 1 0 1 1 1 Tower base
RESULTS TOWER_FORCE_MOMENT 2 1 1 0 1 1 1 Tower h=25m
RESULTS TOWER_FORCE_MOMENT 4 1 1 0 1 1 1 Tower top
RESULTS BLADE_DEFORMATION 1 12 1 1 0 0 0 0 Tip
RESULTS SHAFT_DEFORMATION 4 1 0 1 0 0 0 Rotor centre
RESULTS TOWER_DEFORMATION 4 1 1 0 0 0 0 Tower top
DEFINITION RESPONSE_CALC_LIMITS 1 0E- 3 1 .0E -3 0.0 5. 0E-1
WRITE POWER 7.0 25.0 1.0 4 1 0.10 0.10 0.10
; WRITE RESPONSE 30.0 0.025 1 1 0 1 2 1 5 1 ;
STOP ;

Ris0-R-15OO(EN) 75
Appendix B
1 1 1
1 BLADES
1 24 R m xcg ycg ri_x ri_y x_sh y_sh E G I_x I_y I_P k x k_y A Str. pitch
0.000 2.940E+03 0.0 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 9.790E+01 9.790E+01 9.790E+01 0.5 0.5 0.8830 0.00
2.000 2.940E+03 0.0 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 9.790E+01 9.790E+01 9.790E+01 0.5 0.5 0.8830 0.00
2.001 1.066E+03 0.0 0.0 2.430E-01 2.180E-01 0.000 0.000 4.00E+10 1.50E+10 5.860E-02 5.909E-02 1.174E-01 0.5 0.5 0.5960 16.00
3.200 1.166E+03 0.0 0.0 3.430E-01 3.180E-01 0.000 0.000 4.00E+10 1.50E+10 6.860E-02 5.909E-02 1.274E-01 0.5 0.5 0.5960 26.00
4.200 4.626E+02 0.0 0.0 5.350E-01 5.130E-01 0.000 0.000 4.00E+10 1.50E+10 6.615E-02 6.076E-02 1.274E-01 0.5 0.5 0.2360 11.00
5.700 4.635E+02 5.210E-01 5.230E-01 0.000 0.000 4.00E+10 1.50E+10 6.301E-02 6.350E-02 1.264E-01 0.5 0.5 0.2360 14.10
7.700 4.371E+02 4.870E-01 5.920E-01 0.000 0.000 4.00E+10 1.50E+10 5.194E-02 7.673E-02 1.284E-01 0.5 0.5 0.2230 17.57
10.200 3.646E+02 4.070E-01 6.320E-01 0.236 0.024 4.00E+10 1.50E+10 3.018E-02 7.281E-02 1.029E-01 0.8 0.4 0.1860 19.92
12.200 3.146E+02 3.770E-01 5.910E-01 0.274 0.010 4.00E+10 1.50E+10 2.244E-02 5.488E-02 7.732E-02 0.9 0.4 0.1600 17.36
14.200 2.822E+02 3.570E-01 5.740E-01 0.268 0.007 4.00E+10 1.50E+10 1.793E-02 4.655E-02 6.448E-02 1.0 0.4 0.1440 14.79
16.200 2.724E+02 3.220E-01 5.520E-01 0.259 0.006 4.00E+10 1.50E+10 1.411E-02 4.145E-02 5.557E-02 0.8 0.4 0.1390 11.89
18.200 2.587E+02 2.670E-01 5.300E-01 0.248 0.006 4.00E+10 1.50E+10 9.241E-03 3.626E-02 4.547E-02 0.8 0.4 0.1320 9.30
20.200 2.391E+02 2.190E-01 4.750E-01 0.236 0.005 4.00E+10 1.50E+10 5.713E-03 2.695E-02 3.263E-02 0.8 0.4 0.1220 7.24
22.200 2.185E+02 1.810E-01 4.030E-01 0.223 0.004 4.00E+10 1.50E+10 3.577E-03 1.774E-02 2.136E-02 0.8 0.4 0.1110 5.59
24.200 2.019E+02 1.520E-01 3.370E-01 0.210 0.002 4.00E+10 1.50E+10 2.332E-03 1.146E-02 1.381E-02 0.8 0.4 0.1030 4.28
26.200 1.793E+02 1.310E-01 3.040E-01 0.197 0.002 4.00E+10 1.50E+10 1.528E-03 8.281E-03 9.800E-03 0.8 0.4 0.0913 3.26
28.200 1.597E+02 1.100E-01 2.720E-01 0.184 0.001 4.00E+10 1.50E+10 9.633E-04 5.909E-03 6.870E-03 0.8 0.4 0.0817 2.40
30.200 1.421E+02 9.900E-02 2.440E-01 0.170 0.001 4.00E+10 1.50E+10 7.007E-04 4.243E-03 4.939E-03 0.8 0.4 0.0727 1.66
32.200 1.185E+02 8.400E-02 2.300E-01 0.155 0.001 4.00E+10 1.50E+10 4.214E-04 3.136E-03 3.557E-03 0.8 0.4 0.0605 1.04
34.200 8.859E+01 7.200E-02 2.090E-01 0.142 0.001 4.00E+10 1.50E+10 2.274E-04 1.940E-03 2.166E-03 0.8 0.4 0.0452 0.63
36.200 6.282E+01 5.800E-02 1.980E-01 0.080 0.000 4.00E+10 1.50E+10 1.048E-04 1.234E-03 1.342E-03 0.8 0.4 0.0321 0.34
38.200 3.822E+01 3.800E-02 1.260E-01 0.030 0.000 4.00E+10 1.50E+10 2.724E-05 3.018E-04 3.283E-04 0.8 0.4 0.0195 0.18
39.200 2.842E+01 2.100E-02 1.040E-01 0.010 0.000 4.00E+10 1.50E+10 6.007E-06 1.548E-04 1.607E-04 0.8 0.4 0.0145 0.82
40.000 2.234E+01 5.000E-03 2.400E-02 0.000 0.000 4.00E+10 1.50E+10 2.548E-07 6.223E-06 6.478E-06 0.8 0.4 0.0114 6.34
1 SHAFT
1 6 R m xcg ycg ri_x ri_y x_sh y_sh E G I_x I_y I_P k_x k_y A Str. pitch
0.000 9.800E-01 0.0 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 1.960E-02 1.960E-02 1.136E-03 0.5 0.5 0.2870 0.00
0.600 9.800E-01 0.0 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 1.960E-02 1.960E-02 1.136E-03 0.5 0.5 0.2870 0.00
0.600 9.800E-01 0.0 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 6.791E-03 6.791E-03 1.362E-02 0.5 0.5 0.2870 0.00
2.186 9.800E-01 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 6.791E-03 6.791E-03 1.362E-02 0.5 0.5 0.2870 0.00
2.186 9.800E-01 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 9.790E+01 9.790E+01 9.790E+01 0.5 0.5 0.9990 0.00
3.486 9.800E-01 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 9.790E+01 9.790E+01 9.790E+01 0.5 0.5 0.9990 0.00
1 TOWER
1 83 R m xcg ycg ri_x ri_y x_sh y_sh E G I_x I_y I_P k_x k_y A Str. pitch
0.000 6.203E+03 0.0 0.0 1.066E+00 1.066E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.8070 0.00
0.982 3.097E+03 0.0 0.0 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
1.964 3.097E+03 0.0 0.0 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
2.945 3.097E+03 0.0 0.0 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
3.927 3.097E+03 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
4.909 3.097E+03 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
5.891 3.097E+03 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00

76 Ris0-R-15OO(EN)
 6.872 3.097E+03 0. 0 0. 0 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
7.854 3.097E+03 0. 0 0. 0 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
8.836 3.097E+03 0. 0 0. 0 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
9.818 3.097E+03 0. 0 0. 0 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.987E-01 8.987E-01 1.793E+00 0.5 0.5 0.4020 0.00
10.790 2.891E+03 0. 0 0. 0 1.510E+00 1.510E+00 0.000 0.000 2.10E+11 8.10E+10 8.399E-01 8.399E-01 1.676E+00 0.5 0.5 0.3760 0.00
11.780 2.685E+03 0. 0 0. 0 1.511E+00 1.511E+00 0.000 0.000 2.10E+11 8.10E+10 7.811E-01 7.811E-01 1.558E+00 0.5 0.5 0.3490 0.00
12.760 2.479E+03 0. 0 0. 0 1.512E+00 1.512E+00 0.000 0.000 2.10E+11 8.10E+10 7.223E-01 7.223E-01 1.441E+00 0.5 0.5 0.3220 0.00
13.740 2.479E+03 0. 0 0. 0 1.512E+00 1.512E+00 0.000 0.000 2.10E+11 8.10E+10 7.223E-01 7.223E-01 1.441E+00 0.5 0.5 0.3220 0.00
14.720 2.479E+03 0. 0 0. 0 1.512E+00 1.512E+00 0.000 0.000 2.10E+11 8.10E+10 7.223E-01 7.223E-01 1.441E+00 0.5 0.5 0.3220 0.00
15.700 2.479E+03 0. 0 0. 0 1.512E+00 1.512E+00 0.000 0.000 2.10E+11 8.10E+10 7.223E-01 7.223E-01 1.441E+00 0.5 0.5 0.3220 0.00
16.690 2.274E+03 0. 0 0. 0 1.513E+00 1.513E+00 0.000 0.000 2.10E+11 8.10E+10 6.625E-01 6.625E-01 1.323E+00 0.5 0.5 0.2960 0.00
17.670 2.274E+03 0. 0 0. 0 1.513E+00 1.513E+00 0.000 0.000 2.10E+11 8.10E+10 6.625E-01 6.625E-01 1.323E+00 0.5 0.5 0.2960 0.00
18.650 5.390E+03 0. 0 0. 0 9.830E-01 9.830E-01 0.000 0.000 2.10E+11 8.10E+10 6.625E-01 6.625E-01 1.323E+00 0.5 0.5 0.7000 0.00
19.640 2.264E+03 0. 0 0. 0 1.505E+00 1.505E+00 0.000 0.000 2.10E+11 8.10E+10 6.527E-01 6.527E-01 1.303E+00 0.5 0.5 0.2940 0.00
20.620 2.244E+03 0. 0 0. 0 1.493E+00 1.493E+00 0.000 0.000 2.10E+11 8.10E+10 6.380E-01 6.380E-01 1.274E+00 0.5 0.5 0.2920 0.00
21.600 2.225E+03 0. 0 0. 0 1.481E+00 1.481E+00 0.000 0.000 2.10E+11 8.10E+10 6.223E-01 6.223E-01 1.245E+00 0.5 0.5 0.2900 0.00
22.580 2.215E+03 0. 0 0. 0 1.470E+00 1.470E+00 0.000 0.000 2.10E+11 8.10E+10 6.076E-01 6.076E-01 1.215E+00 0.5 0.5 0.2870 0.00
23.560 2.195E+03 0. 0 0. 0 1.458E+00 1.458E+00 0.000 0.000 2.10E+11 8.10E+10 5.939E-01 5.939E-01 1.186E+00 0.5 0.5 0.2850 0.00
24.540 2.176E+03 0. 0 0. 0 1.446E+00 1.446E+00 0.000 0.000 2.10E+11 8.10E+10 5.792E-01 5.792E-01 1.156E+00 0.5 0.5 0.2830 0.00
25.530 2.156E+03 0. 0 0. 0 1.434E+00 1.434E+00 0.000 0.000 2.10E+11 8.10E+10 5.655E-01 5.655E-01 1.127E+00 0.5 0.5 0.2800 0.00
26.510 2.136E+03 0. 0 0. 0 1.422E+00 1.422E+00 0.000 0.000 2.10E+11 8.10E+10 5.508E-01 5.508E-01 1.098E+00 0.5 0.5 0.2780 0.00
27.490 2.117E+03 0. 0 0. 0 1.411E+00 1.411E+00 0.000 0.000 2.10E+11 8.10E+10 5.380E-01 5.380E-01 1.078E+00 0.5 0.5 0.2760 0.00
28.470 2.107E+03 0. 0 0. 0 1.399E+00 1.399E+00 0.000 0.000 2.10E+11 8.10E+10 5.243E-01 5.243E-01 1.049E+00 0.5 0.5 0.2730 0.00
29.450 2.087E+03 0. 0 0. 0 1.387E+00 1.387E+00 0.000 0.000 2.10E+11 8.10E+10 5.116E-01 5.116E-01 1.019E+00 0.5 0.5 0.2710 0.00
30.430 2.068E+03 0. 0 0. 0 1.375E+00 1.375E+00 0.000 0.000 2.10E+11 8.10E+10 4.978E-01 4.978E-01 9.996E-01 0.5 0.5 0.2690 0.00
31.420 2.048E+03 0. 0 0. 0 1.363E+00 1.363E+00 0.000 0.000 2.10E+11 8.10E+10 4.851E-01 4.851E-01 9.712E-01 0.5 0.5 0.2670 0.00
32.400 2.029E+03 0. 0 0. 0 1.352E+00 1.352E+00 0.000 0.000 2.10E+11 8.10E+10 4.733E-01 4.733E-01 9.457E-01 0.5 0.5 0.2640 0.00
33.380 2.019E+03 0. 0 0. 0 1.340E+00 1.340E+00 0.000 0.000 2.10E+11 8.10E+10 4.606E-01 4.606E-01 9.212E-01 0.5 0.5 0.2620 0.00
34.360 1.813E+03 0. 0 0. 0 1.329E+00 1.329E+00 0.000 0.000 2.10E+11 8.10E+10 4.087E-01 4.087E-01 8.173E-01 0.5 0.5 0.2360 0.00
35.340 1.803E+03 0. 0 0. 0 1.317E+00 1.317E+00 0.000 0.000 2.10E+11 8.10E+10 3.979E-01 3.979E-01 7.958E-01 0.5 0.5 0.2340 0.00
36.330 1.784E+03 0. 0 0. 0 1.305E+00 1.305E+00 0.000 0.000 2.10E+11 8.10E+10 3.871E-01 3.871E-01 7.742E-01 0.5 0.5 0.2320 0.00
37.310 1.764E+03 0. 0 0. 0 1.293E+00 1.293E+00 0.000 0.000 2.10E+11 8.10E+10 3.763E-01 3.763E-01 7.536E-01 0.5 0.5 0.2300 0.00
38.290 1.754E+03 0. 0 0. 0 1.281E+00 1.281E+00 0.000 0.000 2.10E+11 8.10E+10 3.665E-01 3.665E-01 7.330E-01 0.5 0.5 0.2280 0.00
39.270 1.735E+03 0. 0 0. 0 1.270E+00 1.270E+00 0.000 0.000 2.10E+11 8.10E+10 3.567E-01 3.567E-01 7.125E-01 0.5 0.5 0.2260 0.00
40.250 1.715E+03 0. 0 0. 0 1.258E+00 1.258E+00 0.000 0.000 2.10E+11 8.10E+10 3.469E-01 3.469E-01 6.929E-01 0.5 0.5 0.2240 0.00
41.230 1.705E+03 0. 0 0. 0 1.246E+00 1.246E+00 0.000 0.000 2.10E+11 8.10E+10 3.371E-01 3.371E-01 6.742E-01 0.5 0.5 0.2210 0.00
42.220 3.793E+03 0. 0 0. 0 8.230E-01 8.230E-01 0.000 0.000 2.10E+11 8.10E+10 3.273E-01 3.273E-01 6.546E-01 0.5 0.5 0.4930 0.00
43.200 1.676E+03 0. 0 0. 0 1.223E+00 1.223E+00 0.000 0.000 2.10E+11 8.10E+10 3.185E-01 3.185E-01 6.370E-01 0.5 0.5 0.2170 0.00
44.180 1.656E+03 0. 0 0. 0 1.212E+00 1.212E+00 0.000 0.000 2.10E+11 8.10E+10 3.107E-01 3.107E-01 6.203E-01 0.5 0.5 0.2150 0.00
45.160 1.646E+03 0. 0 0. 0 1.202E+00 1.202E+00 0.000 0.000 2.10E+11 8.10E+10 3.018E-01 3.018E-01 6.047E-01 0.5 0.5 0.2140 0.00
46.140 1.627E+03 0. 0 0. 0 1.191E+00 1.191E+00 0.000 0.000 2.10E+11 8.10E+10 2.940E-01 2.940E-01 5.880E-01 0.5 0.5 0.2120 0.00
47.130 1.617E+03 0. 0 0. 0 1.180E+00 1.180E+00 0.000 0.000 2.10E+11 8.10E+10 2.862E-01 2.862E-01 5.723E-01 0.5 0.5 0.2100 0.00
48.110 1.597E+03 0. 0 0. 0 1.169E+00 1.169E+00 0.000 0.000 2.10E+11 8.10E+10 2.783E-01 2.783E-01 5.566E-01 0.5 0.5 0.2080 0.00
49.090 1.431E+03 0. 0 0. 0 1.159E+00 1.159E+00 0.000 0.000 2.10E+11 8.10E+10 2.440E-01 2.440E-01 4.880E-01 0.5 0.5 0.1850 0.00
50.070 1.411E+03 0. 0 0. 0 1.148E+00 1.148E+00 0.000 0.000 2.10E+11 8.10E+10 2.372E-01 2.372E-01 4.743E-01 0.5 0.5 0.1840 0.00
51.050 1.401E+03 0. 0 0. 0 1.137E+00 1.137E+00 0.000 0.000 2.10E+11 8.10E+10 2.303E-01 2.303E-01 4.616E-01 0.5 0.5 0.1820 0.00
52.030 1.382E+03 0. 0 0. 0 1.127E+00 1.127E+00 0.000 0.000 2.10E+11 8.10E+10 2.244E-01 2.244E-01 4.488E-01 0.5 0.5 0.1800 0.00
53.010 1.372E+03 0. 0 0. 0 1.116E+00 1.116E+00 0.000 0.000 2.10E+11 8.10E+10 2.176E-01 2.176E-01 4.351E-01 0.5 0.5 0.1780 0.00
54.000 1.362E+03 0. 0 0. 0 1.105E+00 1.105E+00 0.000 0.000 2.10E+11 8.10E+10 2.117E-01 2.117E-01 4.234E-01 0.5 0.5 0.1770 0.00
54.980 1.343E+03 0. 0 0. 0 1.094E+00 1.094E+00 0.000 0.000 2.10E+11 8.10E+10 2.058E-01 2.058E-01 4.106E-01 0.5 0.5 0.1750 0.00
55.960 1.333E+03 0. 0 0. 0 1.083E+00 1.083E+00 0.000 0.000 2.10E+11 8.10E+10 1.989E-01 1.989E-01 3.989E-01 0.5 0.5 0.1730 0.00
56.940 1.323E+03 0. 0 0. 0 1.073E+00 1.073E+00 0.000 0.000 2.10E+11 8.10E+10 1.931E-01 1.931E-01 3.871E-01 0.5 0.5 0.1720 0.00

Ris0-R-15OO(EN) 77
 57.920 1.166E+03 0.0 0.0 1.063E+00 1.063E+00 0.000 0.000 2.10E+11 8.10E+10 1.676E-01 1.676E-01 3.342E-01 0.5 0.5 0.1510 0.00
58.910 1.147E+03 0.0 0.0 1.052E+00 1.052E+00 0.000 0.000 2.10E+11 8.10E+10 1.617E-01 1.617E-01 3.244E-01 0.5 0.5 0.1500 0.00
59.890 1.137E+03 0.0 0.0 1.041E+00 1.041E+00 0.000 0.000 2.10E+11 8.10E+10 1.568E-01 1.568E-01 3.146E-01 0.5 0.5 0.1480 0.00
60.870 1.127E+03 0.0 0.0 1.030E+00 1.030E+00 0.000 0.000 2.10E+11 8.10E+10 1.519E-01 1.519E-01 3.048E-01 0.5 0.5 0.1460 0.00
61.850 1.117E+03 0.0 0.0 1.019E+00 1.019E+00 0.000 0.000 2.10E+11 8.10E+10 1.480E-01 1.480E-01 2.950E-01 0.5 0.5 0.1450 0.00
62.830 1.107E+03 0.0 0.0 1.009E+00 1.009E+00 0.000 0.000 2.10E+11 8.10E+10 1.431E-01 1.431E-01 2.862E-01 0.5 0.5 0.1430 0.00
63.810 1.088E+03 0.0 0.0 9.980E-01 9.980E-01 0.000 0.000 2.10E+11 8.10E+10 1.382E-01 1.382E-01 2.764E-01 0.5 0.5 0.1420 0.00
64.800 1.078E+03 0.0 0.0 9.870E-01 9.870E-01 0.000 0.000 2.10E+11 8.10E+10 1.343E-01 1.343E-01 2.675E-01 0.5 0.5 0.1400 0.00
65.780 1.068E+03 0.0 0.0 9.760E-01 9.760E-01 0.000 0.000 2.10E+11 8.10E+10 1.294E-01 1.294E-01 2.597E-01 0.5 0.5 0.1390 0.00
66.760 9.241E+02 0.0 0.0 9.660E-01 9.660E-01 0.000 0.000 2.10E+11 8.10E+10 1.097E-01 1.097E-01 2.195E-01 0.5 0.5 0.1200 0.00
67.740 9.143E+02 0.0 0.0 9.550E-01 9.550E-01 0.000 0.000 2.10E+11 8.10E+10 1.058E-01 1.058E-01 2.127E-01 0.5 0.5 0.1190 0.00
68.720 9.036E+02 0.0 0.0 9.450E-01 9.450E-01 0.000 0.000 2.10E+11 8.10E+10 1.029E-01 1.029E-01 2.058E-01 0.5 0.5 0.1170 0.00
69.710 8.938E+02 0.0 0.0 9.340E-01 9.340E-01 0.000 0.000 2.10E+11 8.10E+10 9.898E-02 9.898E-02 1.989E-01 0.5 0.5 0.1160 0.00
70.690 8.830E+02 0.0 0.0 9.230E-01 9.230E-01 0.000 0.000 2.10E+11 8.10E+10 9.584E-02 9.584E-02 1.921E-01 0.5 0.5 0.1150 0.00
71.670 8.732E+02 0.0 0.0 9.120E-01 9.120E-01 0.000 0.000 2.10E+11 8.10E+10 9.251E-02 9.251E-02 1.852E-01 0.5 0.5 0.1130 0.00
72.650 8.624E+02 0.0 0.0 9.010E-01 9.010E-01 0.000 0.000 2.10E+11 8.10E+10 8.928E-02 8.928E-02 1.784E-01 0.5 0.5 0.1120 0.00
73.630 8.526E+02 0.0 0.0 8.910E-01 8.910E-01 0.000 0.000 2.10E+11 8.10E+10 8.604E-02 8.604E-02 1.725E-01 0.5 0.5 0.1110 0.00
74.610 9.614E+02 0.0 0.0 8.790E-01 8.790E-01 0.000 0.000 2.10E+11 8.10E+10 9.467E-02 9.467E-02 1.891E-01 0.5 0.5 0.1250 0.00
75.600 9.496E+02 0.0 0.0 8.680E-01 8.680E-01 0.000 0.000 2.10E+11 8.10E+10 9.114E-02 9.114E-02 1.823E-01 0.5 0.5 0.1230 0.00
76.580 1.058E+03 0.0 0.0 8.570E-01 8.570E-01 0.000 0.000 2.10E+11 8.10E+10 9.898E-02 9.898E-02 1.970E-01 0.5 0.5 0.1370 0.00
77.560 1.039E+03 0.0 0.0 8.460E-01 8.460E-01 0.000 0.000 2.10E+11 8.10E+10 9.486E-02 9.486E-02 1.901E-01 0.5 0.5 0.1350 0.00
78.540 1.666E+03 0.0 0.0 6.550E-01 6.550E-01 0.000 0.000 2.10E+11 8.10E+10 9.134E-02 9.134E-02 1.823E-01 0.5 0.5 0.2170 0.00
78.540 9.800E-01 0.0 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 9.790E+05 9.790E+05 9.790E+05 0.5 0.5 9.9900 0.00
80.100 9.800E-01 0.0 0.0 1.000E-03 1.000E-03 0.000 0.000 2.10E+11 8.10E+10 9.790E+05 9.790E+05 9.790E+05 0.5 0.5 9.9900 0.00

78 Ris0-R-15OO(EN)
Appendix C
1 40 m rotorradius
1 14 ch th twist x _ae y_ae PC set
0.000 0.001 100.000 0.000 0 000 0.000 1
1.200 2.450 100.000 5.100 0 000 0.000 1
5.200 2.650 80.000 6.700 0 000 0.000 1
9.200 3.000 52.000 9.400 0 000 0.000 1
13.200 3.200 33.000 9.400 0 000 0.000 1
17.200 2.800 26.000 5.400 0 000 0.000 1
21.200 2.400 22.200 3.500 0 000 0.000 1
25.200 2.000 20.000 2.300 0 000 0.000 1
29.200 1.800 19.000 1.300 0 000 0.000 1
33.200 1.550 18.500 0.500 0 000 0.000 1
37.200 1.180 18.000 0.000 0 000 0.000 1
39.200 0.620 16.300 1.000 0 000 0.000 1
39.600 0.400 15.700 2.400 0 000 0.000 1
40.000 0.010 15.000 6.000 0 000 0.000 1

80 Ris0-R-15OO(EN)
Appendix D
1 Converted format
7 NACA 634xx
1 35 15.00000 NACA 63415, Modified to Re = 6 million
-180.00000 0.00000 0.10000 0 .00000
-170.00000 0.80000 0.24000 0 .40000
-150.00000 0.80000 0.58000 0 .32500
-120.00000 0.49000 1.06000 0 .32500
-90.00000 0.11400 1.30000 0 .32500
-60.00000 0.75500 1.06000 0 .32500
-30.00000 1.20000 0.45000 0 .22000
-20.00000 1.20000 0.20000 0 .05000
-15.00000 1.10000 0.10000 0 .00000
-10.00000 0.82000 0.01200 0 .00000
-6.00000 0.36500 0.00900 -0 .03750
-2.00000 0.08570 0.00600 -0 .06600
2.00000 0.54610 0.00600 -0 .07800
4.00000 0.77630 0.00650 -0 .08400
6.00000 0.99500 0.00960 -0 .09000
8.00000 17000 0.01320 -0 .09000
10.00000 31400 0.01690 -0 .09000
12.00000 44500 0.02350 -0 .09000
14.00000 45500 0.04700 -0 .09000
16.00000 44830 0.08200 -0 .09000
18.00000 36660 0.12424 -0 .09670
20.00000 32500 0.17700 -0 .10330
25.00000 25906 0.32423 -0 .12000
30.00000 19000 0.47536 -0 .14000
40.00000 06000 0.69000 -0 .18000
50.00000 0.93000 0.90000 -0 .22000
60.00000 0.75500 1.06000 -0 .26000
70.00000 0.55500 1.18000 -0 .29250
80.00000 0.34000 1.26000 -0 .32500
90.00000 0.11400 1.30000 -0 .32500
100.00000 0.11000 1.27000 -0 .32500
120.00000 0.49000 1.06000 -0 .32500
150.00000 0.80000 0.58000 -0 .32500
170.00000 0.80000 0.24000 -0 .40000
180.00000 0.00000 0.10000 0 .00000
2 35 18.00000 NACA 63418, Modified to Re = 6 million
-180.00000 0.00000 0.10000 0 .00000
-170.00000 0.80000 0.24000 0 .40000
-150.00000 0.80000 0.58000 0 .32500
-120.00000 0.49000 1.06000 0 .32500
-90.00000 0.11400 1.30000 0 .32500
-60.00000 0.75500 1.06000 0 .32500
-30.00000 1.20000 0.45000 0 .22000
-20.00000 1.20000 0.20000 0 .05000
-15.00000 1.10000 0.10000 0 .00000
-10.00000 0.82000 0.01200 0 .00000
-6.00000 0.36500 0.00900 -0 .03750
-2.00000 0.08140 0.00600 -0 .06600
2.00000 0.54710 0.00600 -0 .07800
4.00000 0.78000 0.00700 -0 .08400
6.00000 0.99000 0.00900 -0 .09000
8.00000 17700 0.01200 -0 .09000
10.00000 36000 0.01500 -0 .09000
12.00000 45000 0.02400 -0 .09000
14.00000 48000 0.04800 -0 .09000
16.00000 46670 0.09032 -0 .09000
18.00000 44330 0.13352 -0 .09670
20.00000 40500 0.18528 -0 .10330
25.00000 32500 0.33655 -0 .12000
30.00000 23000 0.47691 -0 .14000
40.00000 06000 0.69000 -0 .18000
50.00000 0.93000 0.90000 -0 .22000
60.00000 0.75500 1.06000 -0 .26000
70.00000 0.55500 1.18000 -0 .29250
80.00000 0.34000 1.26000 -0 .32500
90.00000 0.11400 1.30000 -0 .32500
100.00000 0.11000 1.27000 -0 .32500
120.00000 0.49000 1.06000 -0 .32500
150.00000 0.80000 0.58000 -0 .32500
170.00000 0.80000 0.24000 -0 .40000
180.00000 0.00000 0.10000 0 .00000
3 35 21.00000 NACA 63-421 , , Modified to Re = 6 million
-180.00000 0.00000 0.10000 0 .00000
-170.00000 0.80000 0.24000 0 .40000
-150.00000 0.80000 0.58000 0 .32500
-120.00000 0.49000 1.06000 0 .32500

Ris0-R-15OO(EN) 81
 -90.00000 -0.11400 1.30000 0 32500
-60.00000 -0.75500 1.06000 0 32500
-30.00000 -1.20000 0.45000 0 22000
-20.00000 -1.20000 0.20000 0 05000
-15.00000 -1.10000 0.10000 0 00000
-10.00000 -0.82000 0.01200 0 00000
-6.00000 -0.36500 0.00900 -0 03750
-2.00000 0.08140 0.00600 -0 06600
2.00000 0.54710 0.00600 -0 07800
4.00000 0.78000 0.00700 -0 08400
6.00000 0.99000 0.00900 -0 09000
8.00000 1.16000 0.01200 -0 09000
10.00000 1.29000 0.01500 -0 09000
12.00000 1.33000 0.03400 -0 09000
14.00000 1.33330 0.06900 -0 09000
16.00000 1.31670 0.10000 -0 09000
18.00000 1.30000 0.14000 -0 09670
20.00000 1.26670 0.18000 -0 10330
25.00000 1.21683 0.31564 -0 12000
30.00000 1.16000 0.46268 -0 14000
40.00000 1.06000 0.69000 -0 18000
50.00000 0.93000 0.90000 -0 22000
60.00000 0.75500 1.06000 -0 26000
70.00000 0.55500 1.18000 -0 29250
80.00000 0.34000 1.26000 -0 32500
90.00000 0.11400 1.30000 -0 32500
100.00000 -0.11000 1.27000 -0 32500
120.00000 -0.49000 1.06000 -0 32500
150.00000 -0.80000 0.58000 -0 32500
170.00000 -0.80000 0.24000 -0 40000
180.00000 0.00000 0.10000 0 00000
4 35 24.00000 NACA 63- 424 , Modified to Re = 6 million
-180.00000 0.00000 0.10000 0 00000
-170.00000 0.80000 0.24000 0 40000
-150.00000 0.80000 0.58000 0 32500
-120.00000 0.49000 1.06000 0 32500
-90.00000 -0.11400 1.30000 0 32500
-60.00000 -0.75500 1.06000 0 32500
-30.00000 -1.20000 0.45000 0 22000
-20.00000 -1.20000 0.20000 0 05000
-15.00000 -1.10000 0.10000 0 00000
-10.00000 -0.82000 0.01200 0 00000
-6.00000 -0.36500 0.00900 -0 03750
-2.00000 0.08140 0.00600 -0 06600
2.00000 0.54710 0.00600 -0 07800
4.00000 0.78000 0.00700 -0 08400
6.00000 0.99000 0.00900 -0 09000
8.00000 1.12000 0.01600 -0 09000
10.00000 1.18000 0.02600 -0 09000
12.00000 1.20000 0.05400 -0 09000
14.00000 1.20000 0.08700 -0 09000
16.00000 1.20000 0.12000 -0 09000
18.00000 1.19715 0.15907 -0 09670
20.00000 1.18800 0.20171 -0 10330
25.00000 1.16170 0.31655 -0 12000
30.00000 1.14000 0.45000 -0 14000
40.00000 1.06000 0.69000 -0 18000
50.00000 0.93000 0.90000 -0 22000
60.00000 0.75500 1.06000 -0 26000
70.00000 0.55500 1.18000 -0 29250
80.00000 0.34000 1.26000 -0 32500
90.00000 0.11400 1.30000 -0 32500
100.00000 -0.11000 1.27000 -0 32500
120.00000 -0.49000 1.06000 -0 32500
150.00000 -0.80000 0.58000 -0 32500
170.00000 -0.80000 0.24000 -0 40000
180.00000 0.00000 0.10000 0 00000
5 35 30.00000 NACA 63- 430 VELUX measurements, Re = 1.6*10A6
-180.00000 0.00000 0.10000 0 00000
-170.00000 0.80000 0.24000 0 40000
-150.00000 0.80000 0.58000 0 32500
-120.00000 0.49000 1.06000 0 32500
-90.00000 -0.11400 1.30000 0 32500
-60.00000 -0.75500 1.06000 0 32500
-30.00000 -1.20000 0.45000 0 22000
-20.00000 -1.20000 0.20000 0 05000
-15.00000 -1.10000 0.10000 0 00000
-10.00000 -0.67590 0.02340 0 00740
-6.00000 -0.33490 0.01260 -0 00970
-2.00000 0.04730 0.01110 -0 02620
2.00000 0.45190 0.01260 -0 04150
4.00000 0.61180 0.01460 -0 04700
6.00000 0.75000 0.01610 -0 05280
8.00000 0.88530 0.01790 -0 06220
10.00000 0.96230 0.02340 -0 07320
12.00000 1.01450 0.03630 -0 08470
14.00000 1.04300 0.06270 -0 09400

82 Ris0-R-15OO(EN)
 16.00000 1.07160 0.09410 -0.09760
18.00000 1.10010 0.14025 -0.09310
20.00000 1.12870 0.19270 -0.10000
25.00000 1.15000 0.32000 -0.12210
30.00000 1.13539 0.44713 -0.14420
40.00000 1.07000 0.69828 -0.18280
50.00000 0.93000 0.90022 -0.22140
60.00000 0.75500 1.06000 -0.26000
70.00000 0.55500 1.18000 -0.29250
80.00000 0.34000 1.26000 -0.32500
90.00000 0.11400 1.30000 -0.32500
100.00000 -0.11000 1.27000 -0.32500
120.00000 -0.49000 1.06000 -0.32500
150.00000 -0.80000 0.58000 -0.32500
170.00000 -0.80000 0.24000 -0.40000
180.00000 0.00000 0.10000 0.00000
6 35 75.00000 NACA 63- 430, VELUX meat 1.6*10a6 (used at
75%)
-180.00000 0.00000 0.10000 0.00000
-170.00000 0.80000 0.24000 0.40000
-150.00000 0.80000 0.58000 0.32500
-120.00000 0.49000 1.06000 0.32500
-90.00000 -0.11400 1.30000 0.32500
-60.00000 -0.75500 1.06000 0.32500
-30.00000 -1.20000 0.45000 0.22000
-20.00000 -1.20000 0.20000 0.05000
-15.00000 -1.10000 0.10000 0.00000
-10.00000 -0.67590 0.02340 0.00740
-6.00000 -0.33490 0.01260 -0.00970
-2.00000 0.04730 0.01110 -0.02620
2.00000 0.45190 0.01260 -0.04150
4.00000 0.61180 0.01460 -0.04700
6.00000 0.75000 0.01610 -0.05280
8.00000 0.88530 0.01790 -0.06220
10.00000 0.96230 0.02340 -0.07320
12.00000 1.01450 0.03630 -0.08470
14.00000 1.04300 0.06270 -0.09400
16.00000 1.07160 0.09410 -0.09760
18.00000 1.10010 0.14025 -0.09310
20.00000 1.12870 0.19270 -0.10000
25.00000 1.15000 0.32000 -0.12210
30.00000 1.13539 0.44713 -0.14420
40.00000 1.07000 0.69828 -0.18280
50.00000 0.93000 0.90022 -0.22140
60.00000 0.75500 1.06000 -0.26000
70.00000 0.55500 1.18000 -0.29250
80.00000 0.34000 1.26000 -0.32500
90.00000 0.11400 1.30000 -0.32500
100.00000 -0.11000 1.27000 -0.32500
120.00000 -0.49000 1.06000 -0.32500
150.00000 -0.80000 0.58000 -0.32500
170.00000 -0.80000 0.24000 -0.40000
180.00000 0.00000 0.10000 0.00000
7 2 100.00000 Cylinder (used at 100%)
-180.00000 0.00000 0.60000 0.00000
180.00000 0.00000 0.60000 0.00000

Ris0-R-15OO(EN) 83
Mission
To promote an innovative and environmentally sustainable
technological development within the areas of energy, industrial
technology and bioproduction through research, innovation and
advisory services.

Vision
Rise’s research shall extend the boundaries for the under­
standing of nature’s processes and interactions right down to
the molecular nanoscale.

The results obtained shall set new trends for the development
of sustainable technologies within the fields of energy, industrial
technology and biotechnology.

The efforts made shall benefit Danish society and lead to the
development of new multi-billion industries.

www.risoe.dk