High Frequency Generator and Transformer for

HVDC-connected Wind Turbine Farms

Jens Due
Department of Energy Technology
Electrical Energy Engineering
Aalborg University
Email: [email protected]

AbstractThis article presents a way of modeling a system

containing a variable speed high frequency synchronous generator that supplies a HVDC connection trough a passive three phase
rectifier. The system is intended to be used in wind turbine farms,
as a new approach for enabling wind turbines to be connected
to the main grid and still be able to produce power at different
generator shaft speeds. The generator used in the test setup is aP-69
1120 W Bosch claw pole generator. The articleRectifier
describes dynamic
modeling of the generator and rectifier in separate parts as well
as combined. The generator is modeled using the qd reference
frame and the rectifier is modeled in the abc system. The output
power from the system is controlled by adjusting the field current
to the generator with a PI compensator. The developed models
are compared with experimental results.
Index Terms Claw pole synchronous generator, Dynamic
modeling, six pulse rectifier, PI compensator, wind turbine.

I. I NTRODUCTION
energy sources are attracting a continuously
rising amount of interest, due to increasing
environmental concerns. Wind power systems are currently
considered to be one of the most applicable Rectifier
natural energy
sources. State of the art wind turbines use different generator
topologies to take into account the variations in shaft speeds.
This is done either by the use of full scale converters or
by utilizing classic doubly fed induction generators along
with smaller converters. When applying doubly fed induction
generators a speed change of approximately 30 % can be
achieved [2]. This allows an increased utilization compared
to stall regulated wind turbines. The use of converters offers
an advantage because it allows full control of the output
power. Unfortunately full scale active converters has a major
drawback due to their high cost, thus a configuration without
them would be a great improvement.
Wind turbines are often placed in farms far away from
the customers, including off shore. This introduces new
challenges in transportation of energy from wind farms
into mainland. HVDC-lines from wind turbine farms to
mainland are currently being tested and it seems to be a good
solution for several reasons; for instance lower transmission
losses and easy control of the wind turbine park since the
individual wind turbine is decoupled from the common grid.
A new way of designing wind turbines is by use of direct
drive generators. This eliminates the use of gearboxes and

Rasmus . Nielsen

Christian F. Jensen

Department of Energy Technology

Department of Energy Technology
Electrical Energy Engineering
Electrical Energy Engineering
Aalborg University
Aalborg University
Synchronous
Windmill
Transformer
Email:Generator
[email protected]
Email: [email protected]

ATURAL

Rectifier

thereby reduces the maintenance. This was done in 2000

by ABB when a permanent magnet generator wound with
high voltage cables and a rectifier system was used [8]. This
project was called Windformer. The project was unsuccessful
because of lack of belief in the system from the customers [1].
igen

irec

Generator
In this article
another approachRectifier
for choosing a generator for
+
vgen Z is considered. The idea proposed
a direct drive wind turbine
is to use a high frequency synchronous generator. The output
from this generator will be controlled by controlling the field
current for the generator. This means that only the field
winding needs to be fed by a converter. Thus the converter
only needs to handle a fraction of the total power delivered
by the wind turbine. A power transformer is used in order to
Synchronous
Generatorlevel. Because of the high
transform the voltage
to a higher
frequency output from the generator, the transformer size can
be reduced and thereby
its cost.
All wind turbines
Statorreducing
winding
Rectifier
Output
in a farm needs to be coupled together. This is done by using
HVDC, and a simple diode rectifier is proposed. Because of
Field coilall generators are decoupled from their common
the rectifier
Field winding
Converter
grid. With a DC voltage as output, the wind turbine can change
shaft speed independent of the speed of the other wind turbines
in the farm. A block diagram of the system is shown onRegulator
Figure
1.

Wind turbine

Synchronous
Generator

Transformer

Wind turbine

Synchronous
Generator

Transformer

Wind turbine

Synchronous
Generator

Transformer

DC AC
Converter

Figure 1. Block diagram of the system described in this article.

Synchronous
Generator

The system contains a high frequency synchronous

generator with a controllable field. The generator voltage
will be kept low and transformed up to a higher level in a
P-83
P-86
three phase power
transformer.P-85The high three
phase voltage
Synchronous
Windturbine
Transformer
will
be rectified
by
a
three
phase
rectifier,
and
the produced
Generator
output power will be delivered to the HVDC connection, thus
P-81

Windturbine

Synchronous
Generator

P-80

P-84

Transformer
HVDC

P-79

Windturbine

Synchronous
Generator

Windturbine

Synchronous
Generator

P-89

Buck
Converter

P-82

P-87

Transformer
P-88

Transformer

P-78

Main
Converter

HVAC

Buc
Conve

eliminating the use of full scale power converters.

voltage of 11.4 V and a phase current of 56.7 A on the AC

side. This corresponds
to 14 V and 80 A on the DC side.
P-59
P-60
Vf
Vout to control the output from the system, a PI
In order
A. Objectives
P-61
P-58
igen irec
Generator
compensator
is used. The compensator is designed to work
T
Iout
Rectifier
rator
When
constructing
a
system
like
the
one
described
in
in
one
operating
point. The demand for the compensator
+
vgen Zthe introduction, it is convenient to have a complete system is a steady state error of 0, a rise time of 500 ms and a
gen
model that can predict performance under steady
Vf state and
settling vtime
of maximum 1 second. The two last demands
igen
dynamic Synchronous
conditions. It will be investigated if the model Generator
can are arbitrary
chosen.
The compensator is implemented on a
m
Pgen
Generator
predict system
performance under speed variations up to micro controller.
33.3 %. This article deals with modeling of the electrical
II. S YNCHRONOUS G ENERATOR M ODEL
part for a single wind turbine. Thus interconnections of the
Field
D nor is the
In thisVf section the
modeling of the claw pole synchronous
wind turbine park is not discussed in this article,
RPM
Synchronous Generator
PWM
PI
DCConvertergenerator will be discussed. A 3D figure of such a generator
modeling of Buck
the power transformer.
6000
Regulator
Converter
is shown on Figure 3.
Operating
Stator winding
Rectifier
Output of the parts shown on Figure
A single wind
turbine consists
area
Vref
2
D
3000
Regulator
Vout

Synchronous
Generator

ynchronous
Generator

ynchronous
Generator

P-83

ynchronous
Generator
P-81

ynchronous
Generator
P-79

ynchronous
Generator
P-89

ynchronous
Generator

Windmill

Windmill

Windmill

Windmill

Transformer

Transformer

Buck
Converter

PWM

kW

P-42

Windmill

Synchronous
Generator

Windmill

Synchronous
Generator

Windmill

Synchronous
Generator

Synchronous
Generator

Regulator

Vref

ynchronous
Generator

1120

Field winding

P-41

Load
DC Voltage

VDC

PI
Compensator

Field
supply

P-31

DC AC

Synchronous
Generator

Converter
Figure 2. Parts
in a single wind turbine system.

Figure 3. 3D figure of the claw pole generator [5].

Windmill

Transformer
The system

model will have the field coil voltage and

mechanical speed as input. No mechanical system modeling
is described. From the inputs the model must be able to
predict currents and voltages in the
system under both normal
Synchronous
Generator
circumstances and in case of irregularities.
The model must
also be able to simulate the power delivered from the generator
and rectifier.
PWM

Buck
Converter

P-86
B. P-85
System description

Transformer

Voltage
The generator DCrotor
comprise of two pieces, molded as
in Figure 3. Around the middle of one of the pieces
the field winding is placed. Thereafter the two halves are
joint together forming the rotor. When applying a current to
the rotor winding,
magnetic north and south poles will be
Synchronous
Generator
present
DC
Voltage at the top and bottom claws depending on the current
direction. The stator is constructed ofDClaminated
steel. The
Voltage
PI windings are sinusoidal distributed with 12 poles per phase.
Field
shown
supply

Regulator

In order to verify the models a test setup system is realized.

TheP-80generator P-84
for the test setup system is selected to be a
Transformer
Bosch claw pole car generator. This type has been selected
HVAC
since it has a high HVDC
number Main
of poles
and is constructed to
P-82
operate
at highP-87speed. The Converter
generator has the specifications
Transformer
presented in Table I

Buck
Converter

PWM

When modeling the generator the qd reference frame is

used. A block diagram of the generator model is shown in
Figure 4.
V rfd

v rqd
Generator
qd model

i rqd

P-78

P-88

Transformer

P-42

Synchronous
Generator

P-41

Table I
R ATED DATA FOR THE GENERATOR
Parameter
Value
Power P-32 1.12
Transformer
Voltage
11.4
Current
56.7
Speed
6000
Frequency
600
P-43

PI
Regulator

Ksr

vabc

(Ksr )-1 iabc

Load

Unit
P-40
kW
V Rectifier
A
RPM
Hz

P-33

Figure 4. Block diagram of a generator model.

HVDC
preparations
In Figure 4 the superscript

and are used to denote the

rotor reference frame and that the values are referred to the
stator respectively.

P-36

P-35

Synchronous
In order to convert the
three phase AC voltage
delivered by HVDC
The model calculates the qd currents from the field coil
Transformer
Rectifier
Generator
preparations
the generator into DC, a six pulse rectifier will be modeled voltage and input mechanical speed. The iqdHVAC
currents are
Main
and verified by means of tests in the laboratory. The rectifier transformed to the abcHVDC
system and the
load determines the abc
must be able to handle a three phase voltage with a line-line currents. These currents are used Converter
to calculate the abc voltage.
P-31

Synchronous
Generator

P-29

Synchronous
Generator

P-37

Transformer

P-34

Rectifier

P-39

HVDC
preparations

P-28

Transformer

P-30

Rectifier

P-38

HVDC
preparations

P-29

Synchronous
Generator

The abc voltage are then transformed back to the qd system

and used in the generator model.
The abc system is transformed to the qd system using the
following transformation [7, p. 111]:
f qd0s = K s f abc

= vabc iabc
3
=
(vq iq + vd id )
[W]
(12)
2
The model machine parameters are determined according to
IEEE 115 [4] and IEC 34-4 [3], where all values are referred
to the stator side.

(1)

where
f qd0s = [ fq

fd

f0 ]T

(2)

f abcs = [ fa

fb

fc ]T

(3)

Table II
M ODEL PARAMETERS EXPRESSED AS INDUCTANCES AND RESISTANCES .

and

cos( + 2
3 )

sin( + 2
3 )

cos( 2
3 )
sin( 2
3 )

cos()
2
sin()
Ks =
3
1

1
2

(4)

1
2

(5)

A graphical presentation of the transformation is given in

Figure 5

fqs

fas

vfrd

drqs
dt
drds
r
r
= rs ids + r ds +
dt
0
r
d
0 0
fd
= rr ifrd +
dt
= rs irqs r rds +

[V]

(6)

[V]

(7)

[V]

(8)

Where the flux linkages are:

rqs

= Lls irqs Lmq irqs

rds

= Lls irds Lmd (irds + ifrd )

frd

Llf d ifrd

Lmd (irds

Table III
R ESULTS FROM THE FIELD SHORT CIRCUIT TEST CONDUCTED AT 3000
RPM WITH AT RESISTIVE LOAD OF 0.175 AT RATED VOLTAGE .

(9)
0

The model is verified by utilizing two test methods. First

a field short circuit test and second an armature short circuit
test. In all tests conducted, the field voltage is adjusted so the
rated voltage can be measured at the generator terminals.
1) Field short circuit test: A set of field short circuit tests
are done with a load rage from 100-812 W. At an arbitrary
moment the field winding is short circuited and the transient
behavior of the terminal voltage is recorded. The steady state
value and the transient behavior of the terminal voltages are
compared with the results from the simulation. The steady
state value for the voltage induced due to the remanence are
also compared with the model prediction. The simulated and
measured results for the test conducted with a load of 812 w
are shown in Figure 6 and 7 respectively.
The largest deviations between the simulated and measured
results are presented in Table III

fds

The qd generator voltage equations used are:

r
vds

0.003

B. Verification of the generator model

Figure 5. Graphical presentation of the abc to qd transformation.

r
vqs

Rf d

The generator model is implemented using Matlab Simulink

where Equation 6 to 12 are used. The input for the model are
the field coil voltages and the mechanical speed. The field coil
voltage are transformed to the stator side using appropriate
turns ratio transformation. The mechanical speed is measured
by means of a tacho generator and feed to the model. It was
discovered that the synchronous generator exhibits remanence
phenomena. In order to take this into account, a offset field
coil voltage is applied the model. This voltage will have the
same property as a permanent magnet.

fcs

Unit

H
H
H
H

A. Model implementation

()d + (0)

fbs

Value
0.0709
25.06
248.5
135
441

is defined as:
=

Parameter
Rs
Lls
Lmd
Lmq
0
Lf d

ifrd )

(10)
(11)

Vpeak

Remenance
Output power

The power delivered from the generator is calculated in the

qd reference frame as:
3

Simulated
17 V
155 ms
1 Vp
824 W

Measured
16.5 V
155 ms
1 Vp
812 W

Error
2.9 %
0%
0%
1.5 %

(300 Hz) compared to the open circuit time constant of the

generator, two different time scales are used to present the
results. Only the results obtained from phase c are shown.
The two other phases have the same tendency as phase c. In
Figure 8 a high time resolution is shown. The simulated and
measured results can both be seen on the figure.

20

10
5
0
5

40

10

30

Phase c current [A]

LineLine Voltage [V]

15

15
20
0.1

0.1

0.2

0.3

0.4

0.5

0.6

Time [s]

Figure 6. Simulation of field short circuit test at 3000 RPM with at resistive
load of 0.175 at rated voltage.

20
10
0
10
20
30
40

20

15

LineLine Voltage [V]

Measured
Simulated

10

15

20

Time [ms]

10

Figure 8. Three phase short circuit on the stator side.

5

The maximum error in phase c is approximately 5 A. It is

assumed this is due to the small subtransient component which
has been neglected in the generator model. The same transient
and steady state values are obtained.
A low time resolution plot of the simulated and measured
short circuit current is shown in Figure 9 and 10 respectively.

0
5
10
15
20
0.1

0.1

0.2

0.3

0.4

0.5

0.6

40

Time [s]

30

Phase c current [A]

Figure 7. Measurement of field short circuit test at 3000 RPM with at resistive
load of 0.175 at rated voltage.

In the test the shaft speed changes from 3000 rpm to 3244
rpm.
The largest deviations from all the tests conducted are
shown in Table IV.

20
10
0
10
20
30
40

Table IV
0

D EVIATIONS BETWEEN SIMULATED

AND MEASURED RESULTS FROM THE

FIELD SHORT CIRCUIT TESTS .

Vpeak

Remanence (peak)
Output power

Load
315 W
736 W
315 W
315 W

Simulated
17.1 V
135 ms
2.4 V
351 W

Measured
15.9 V
146 ms
2.1 V
315 W

0.1

0.2

0.3

0.4

0.5

Time [s]

Error
7%
8.1 %
12.5 %
10.3 %

Figure 9. Simulated short circuit test at 3000 RPM and rated voltage.

The figures shows that the simulated and measured short

circuit currents have the same tendency. The time constant for
the simulated value is 120 ms while the time constant for the
measurement is 115 ms. This presents an error of 4.3 %.

2) Three phase short circuit test: For the three phase short
circuit test, the generator is driven at 3000 RPM and loaded
with a resistive load of 5 before the short circuit is applied.
At an arbitrary chosen moment a three phase short circuit
is applied to the stator terminals. As the frequency is high
4

As can be seen from the figure there are two stages, either
the current in the inductor is zero or larger than zero. If the
current through the inductor is larger than zero a pair of
diodes will always be conducting. If the current through the
inductor is zero no diodes will be conducting. When it has
been evaluated whether the current through the inductor is
larger than zero and this is true, the input voltage to the load
(Vrec ) will be the largest line-line voltage with two forward
voltage drops subtracted. If the current through the inductor
is zero, it is necessary to check whether the largest line-line
voltage is larger than the output voltage plus two forward
voltage drops before the diodes can start to conduct.

40

Phase c current [A]

30
20
10
0
10
20
30
40
0

0.1

0.2

0.3

0.4

0.5

Time [s]

The current in each phase must then be the same as

current flowing in the inductance, with a positive sign if
phase has the highest potential and with a negative sign if
phase has the lowest potential. The equations describing
output currents and voltages are shown in Equation 13 to

Figure 10. Measured short circuit test at 3000 RPM and rated voltage.

III. R ECTIFIER
In this section the modeling of the rectifier and the verification of this will be discussed. The model should be able
to predict which currents that are drawn from each phase at
a given three phase input voltage. The load will be modeled
using an inductor, capacitor and resistor. The rectifier and load
model is shown in Figure 11:

iL
va

D1

D2

vL

vrec

vc

iC
C

iR
R

D4

D5

iL
vL
vo

vb

iR

D3

iC

Vo
-

vf
vf

Figure 11. The rectifier and load model.

A flowchart of how the model of the rectifier works is shown

in Figure 12.

iL > 0
no

vLL > Vo+2Vf

yes

vrec = vLL - 2vf

(13)

[A]

(14)

[A]

(15)

[V]

(16)

[V]

(17)

= v0 + Rd id
[V]
= 0.28 + 0.0027 id [V]

(18)
(19)

Where id is the current through the diode. This function has

been constructed from data found in the datasheet of the diode
[6]. The current through the diode is shown as a function of
the forward voltage drop in Figure 13:
In order to verify the model, a comparison between a
commercial Spice simulator and the rectifier model loaded
with a LCR circuit is conducted. The value of the components
are L=1 mH, C=1 mF and R=1 . The results can be seen
in Figure 14 and Figure 15. As can be seen in the figures the
results are identical.
The model was verified by comparing the simulated phase
current and the output voltage with the measured phase current
and output voltage. The comparison was done with a resistive
load.
The simulation was done with sinusoidal voltages with a
amplitude and frequency as measured in the laboratory. The
results can be seen in the Figure 16 and 17. Figure 16 shows
the current in phase a, and Figure 17 shows the output voltage.

iL

yes

[A]

The principle shown in Figure 12 was implemented in

Matlab Simulink.
The diodes used are IXYS DSS 2x121-0045B schottky
diodes. The forward voltage drop is described by the following
equation:

D6

Find Largest
vLL

= iL iR
vo
=
RZ
1
vL ( ) d
=
L
= vrec vo
Z
1
=
iC ( ) d
C

the
the
the
the
17

no
vrec = Vo

Calculate iL

Figure 12. Flow chart of the function determination of diodes conducting.

5
Find Largest V

Vb

Vc

Calculate VLL

Find largest
(Va >Vb & Va>Vc)=1
(Vb >Vc & Vb>Va)=2
(Vc >Vb & Vc>Va)=3

Find smallest
(Va <Vb & Va<Vc)=1
(Vb <Vc & Vb<Va)=2
(Vc <Vb & Vc<Va)=3

15
Simulated
Measured
2

10

Phase a current [A]

Diode current [A]

10

10

5
0

10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

10

Forward Voltage Drop [V]

Figure 16. Phase a current, measured and simulated with the rectifier model.
20

30
Simulink simulation
LTspice simulation

25

15

14

Output Voltage [V]

16

10
5
0
5

12
10
8
6

10

15

2
0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

Simulated
Measured

18

20

20

0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05

Time [s]

Figure 13. Model of the diode characteristics.

Phase a current [A]

0.1

0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05

Time [s]

Time [s]

Figure 17. Rectifier output voltage, measured and simulated with unbalanced
supply .

Figure 14. Phase a current, simulated in LTspice and simulated with the
rectifier model.
20
Simulink simulation
LTspice simulation

18

is 2.6 %, while the error in the rms current is 4.3 %.

Output Voltage [V]

16

Synchronous
Generator

IV. M ERGING
OF THE GENERATOR AND RECTIFIER
Windmill
Transformer

14

MODELS

12

When combining the two models, a problem arises. This

is due to the fact that both models has voltage as input and
current as output. The problem is then to have the same voltage
on both the generator and the rectifier that results in an equal
output current for the two models.
A solution to this problem is to use a parallel impedance
between the two models. The idea is visualized in Figure 18

10
8
6
4
2
0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Time [s]

P-69

Rectifier

Figure 15. Output voltage, simulated in LTspice and simulated with the
rectifier model.

igen
Generator

vgen
-

As can be seen the deviations between the measured and the

simulated are small. The error in the average output voltage

irec
Rectifier
Z

Figure 18. Block diagram for the merging strategy for the generator and
rectifier model.

6
Synchronous Generator

Stator winding

Field coil

Rectifier

Output

As it can be seen from Figure 18 the input voltage for the

rectifier can now be calculated as:
[V]

Phase a current [A]

vgen = Z(igen irec )

(20)

The parallel impedance must be chosen so large that the

current through it is insignificant compared to the current
drawn by the rectifier. If this is done, the simulation results
will not be influenced in any significant manner.
Because of the rapid switching of the currents drawn from
the rectifier, the model gives numerical problems. To overcome
this, the slope of the phase currents are limited. This is done
by applying a first order low pass filter to the currents from
the rectifier before feeding them back to the rest of the system.
This principle is shown in Figure 19.

vabc iabc

0
-20

Line-line voltage [V]

20

10

-10

-20
20

Output voltage [V]

ifabc iabc
Filter

Figure 19. Implementation of low pass filter in rectifier currents.

The highest shaft speed is 6000 RPM. Because of the six

pulse rectifier, the highest system frequency will be 600 6 =
3.6 kHz thus the corner frequency is chosen to be 36 kHz.
This results in the following transfer function:
j
c

1
=
+1

1
j
236k

+1

v
= 0.124A
Xc

15

10

0.5

1.5

2.5

3.5

4.5

Time [ms]
Figure 20. Verification at 3000 RPM with a load resistor of 0.5 .

(21)

In order to select the parallel impedance, tests was made

with a capacitor, a resistor and a capacitor parallel with a
resistor. The solution with the capacitor gives the best results.
The size of the capacitor is set to 10 F for 3000 RPM and 5
=
F for 6000 RPM. With a phase voltage of maximum 11.4
3
6.6 V the rms value of the current trough the capacitor can be
calculated as:
ic =

20

Rectifier

Generator

H(j) =

40

-40

vabc iabc

Simulated
Measured

60

1) Verification with resistive load: Both the simulated and

measured values for the verification with an resistive load of
0.5 and a speed of 3000 RPM are presented in Figure 20.
From Figure 20 it is seen that both the simulated and
measured phase a current, line-line voltage and output voltage
share the same tendencies. The errors are presented i Table V.
Table V
E RRORS IN THE VERIFICATION AT 3000 RPM WITH A LOAD RESISTOR OF
0.5 .

(22)

Parameter
Phase a current (rms)
Line-line voltage (rms)
Output voltage (avg)

The same result is achieved if the generator runs at 6000

RPM as if the generator runs at 3000 RPM because the capacitor value is reduced by half and the frequency is doubled.

Error
6.89 %
5.2 %
4.1 %

V. S YSTEM VERIFICATION
2) Verification with LCR load: The test was conducted at
3000 RPM with a inductor of 0.6 mH, capacitor of 5.1 mF
and a load resistor of 0.5 . The results are shown in Figure
21. From Figure 21 it is seen that both the simulated and
measured phase a currents and output voltages share the same
tendencies. The simulated line-line voltage becomes unstable
when the diodes are switching.

A. Verification of the unregulated system

The verification of the unregulated system was done with
two different loads; resistive load and LCR load. The verification was at different mechanical speeds and resistive loads.
Only one test case for resistive load and LCR load is presented
in this article.
7

Measured
Simulated

18

Simulated
Measured

40

16

20

14

Output DC voltage [V]

Phase a current [A]

60

0
-20
-40

40

12
10
8
6

Line-line voltage [V]

30
2

20

10

-20
-30
20

Output voltage [V]

0.2

0.4

0.6

0.8

1.2

Figure 22. DC output voltage controlled with a PI compensator at a change

in the load resistor from 5 to 0.5 .

-10

From the above it is concluded that it is possible to design

a compensator for a given working point. Furthermore it is
suggested that an adaptive compensator can be used in order
to be able to regulate the system at different working points.

15

10

VI. C ONCLUSION
As stated in the introduction, the main goal of the article was to develop a complete model capable of predicting
both voltages and currents in the system under normal and
short circuit fault conditions. The models developed for the
synchronous generator and rectifier are capable of modeling
their individual output voltages and currents with an acceptable
degree of accuracy. Unfortunately the model for the complete
unregulated system is only capable of modeling output voltages and currents at resistive loads. With a LCR load the model
fails to modeling the line-line voltage. It is concluded that this
is due to numerical problems in the simulator.
The PI compensator can control the output voltage with an
LCR load as expected. This works, in spite of the problems
with the model merging. The reason why the PI compensator
works, is because, only a part of the synchronous generator
model and not the rectifier model is used in the designing of
the regulated system.
Conclusively, it seems to be necessary to improve the strategy for combining the models for the synchronous generator
and rectifier in order to achieve a complete model which
is capable of modeling the output voltages and current at
different loads.

Time [s]

0.5

1.5

2.5

3.5

4.5

Time [ms]
Figure 21. Verification at 3000 RPM with a inductor of 0.6 mH, capacitor of
5.1 mF and a load resistor of 0.5 .

B. Verification of the regulated system

The verification of the complete regulated system loaded
with a LCR load was done by applying a sudden change in the
DC load resistor from 5 to 0.5 . The load was connected
in accordance with Figure 2. The capacitor value was 5.1 mF
and the inductor value was 0.6 mH. The compensator used
was a PI compensator with a proportional factor of 0.13 and
a integration time of 0.077. Both the simulated and measured
output DC voltage waveforms are shown in Figure 22. As can
be seen in Figure 22 both the simulated and measured output
voltages shows the same tendencies. Still there are differences
in both rise time and settling time. All demands except rise
time are fulfilled in both the simulated and measured system.
The important data are shown in Table VI.

R EFERENCES
[1] Anders Berner. Abb melder sig ud af vindmlle-racet. Internet, Written:
15/03 2002. Shown: 05/10 2008, http://ing.dk/artikel/45650.
[2] Anca D. Hansen and Lars H. Hansen. Market penetration of wind
turbine concepts over the years.
Internet, Shown: 17/05 2008.
http://www.risoe.dk/rispubl/art/2007 136 paper.pdf.
[3] IEC34-4. Rotating electrical machines IEC 34-4. IEC, 1985. ISBN -.
[4] IEEE. IEEE guide: Test Procedures for Synchronous Machines. IEEE,
1995. ISBN - 1-55937-710-0.
[5] Infolytica. Claw-pole alternator. Internet, Written: 2008. Shown: 05/12
2008, http://www.infolytica.com/en/coolstuff/ex0070/.

Table VI
K EY

DATA FROM VERIFICATION OF THE REGULATED SYSTEM .

Parameter
Rise time
Settling time (max)
Overshoot
Steady state error

Demand
500 ms
1000 ms
0%
0%

Measured
480 ms
900 ms
0%
0%

Simulated
420 ms
640 ms
0%
0%

[6] IXYS. Power schottky rectifier. Internet, Written: 2002. Shown: 10/12
2008, http://ixdev.ixys.com/DataSheet/L227.pdf.
[7] Paule C. Krause, Oleg Wasynczuk, and Scott D. Sundhoff. Analysis of
electric machinery and drive systems. Wiley Interscience, 2002. ISBN
0-471-14326-X.
[8] Sanne Wittrup.
Afventende modtagelse af abbs nye
vindmllegenerator. Internet, Written: 16/06 2000. Shown: 05/10
2008, http://ing.dk/artikel/29471.