Anda di halaman 1dari 16

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/224600599

Research laboratory for power Electronic


generator systems in wind turbines comprising
converters, generators, interaction...

Conference Paper October 2009


Source: IEEE Xplore

CITATIONS READS

8 118

9 authors, including:

Joerg Dannehl Ralf Lohde


Danfoss A/S Christian-Albrechts-Universitt zu Kiel
20 PUBLICATIONS 1,414 CITATIONS 12 PUBLICATIONS 161 CITATIONS

SEE PROFILE SEE PROFILE

C. Wessels
Nordex SE
18 PUBLICATIONS 756 CITATIONS

SEE PROFILE

All content following this page was uploaded by Joerg Dannehl on 31 May 2014.

The user has requested enhancement of the downloaded file.


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

Research Laboratory for Power Electronic Generator Systems in Wind


Turbines Comprising Converters, Generators, Interaction and Grid
Interaction

F.W. Fuchs, J. Dannehl, R. Lohde, V. Dinkhauser, S. Jensen, A. Knop,


K. Rothenhagen, S. Thomsen, C. Wessels
Institute for Power Electronics and Electrical Drives,
Christian-Albrechts-University of Kiel, D-24143 Kiel, Germany
Member of CEwind, Center of Excellence for Wind Energy of Universities in Schleswig-
Holstein, Germany
Phone: +49 (0) 431-880-6100, Email: fwf@tf.uni-kiel.de

Acknowledgement
Parts of this research work have been funded by the European Union, the State of Schleswig Holstein,
Germany, the German Research Foundation (DFG) and industry.

Keywords
Electrical drive, Test bench, Wind generator systems, Measurements, Simulation,

Abstract
Power generated from wind is actually a very great and a very fast growing renewable source of the
world energy consumption. Optimizing performance and costs is an important issue for wind turbine
generators. A laboratory for comprehensive research on the four types of power electronic generator
systems for wind turbine generators and for their grid integration is presented. It consists of drives in a
downscaled power range, special developed devices for generating test conditions and for
measurement as well as of numerical simulations. The laboratory is presented and research including
measurements are shown.

Introduction
Energy generation by means of wind turbine generators today is a must on the way to reduce carbon
dioxide production considerably. Converter fed, variable speed generators are state of the art in wind
turbine generators. The power is harvested by the wind rotor and fed to the electrical grid via gear,
generator and converter. Installations worldwide grow rapidly. Their performance still can and has to
be optimized as there are the control, the grid side performance, the losses and the costs of wind
turbines including their power electronic generator systems. Another high requirement is energy feed
in to the grid by the wind turbine conforming to standards. The grid codes define the conditions at the
point of coupling, for enabling safe and controllable grids. These codes define steady state as well as
dynamic performance and are getting continuously stronger because of the higher density of these
decentralized power sources. Both fields require intensive research and development of wind turbines
and their power electronic generator systems.
Research and predevelopment in this field is carried out by analysis and simulation, by
implementation and measurement at test set-ups in the laboratories and finally also at the real wind
turbine generators. In general, in many laboratories worldwide research on power electronic generator
systems for wind turbines is being carried out. Results are shown in the publications covering this
research, for example in [1, 2]. The research includes new power electronic circuits, new control
methods for better performance and as well as methods and hardware to comply with the grid codes in
the way of delivering reactive power or sustaining during voltage sags or swells [3]. Some
publications present laboratories for power electronic systems for renewable energies. General reports
on research laboratories for power electronic systems in wind turbines have barely been found. In [4] a

EPE 2009 - Barcelona ISBN: 9789075815009 P.1


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

green lab is presented, for research work on power electronic systems for renewables, intended for
work of graduate students and for PhD students, in [5] only for graduate students.
At the Kiel University a Research Laboratory for power electronic generator systems, their converters,
generators, interaction and grid integration of wind turbines has been established. It is part of the
power electronics laboratory and intended for research work in PhD, in industrial as well as in mixed
projects. It comprises test set-ups for this work, devices to produce the test conditions, special
developed measuring equipment for example for the grid impedance as well as simulation models. The
laboratory is in parts integrated in CEwind Competence Centre of Universities for Wind Energy in
Schleswig-Holstein, which covers moreover research on mechanical, other electrical and economical
subjects.
Comprehensive research on converters, generators and grid integration of wind turbines is possible
and has been done in this laboratory. Comparative analysis between generator variants can be done,
too. Many investigations have been carried out, some at the test benches, more in simulation. These
are low voltage ride through, high voltage ride through with FACTS, wind park energy feed in via
HVDC, wind fluctuation and energy gain, grid side converter control optimization, control of the
torsional drive train, grid fault propagation, sensor fault tolerant operation as well as electrical
condition monitoring of doubly fed induction generators.
The laboratory will be presented and typical research results, combined from simulation and
measurements at the test drive. An overview of the laboratory is given first, followed by several
research results. A conclusion closes the publication.

Overview on the Laboratory


In wind turbines with variable speed today four generator types can be found: Doubly Fed Induction
Generator, Squirrel Cage Induction Generator, Synchronous Generator with Field Excitation and
Synchronous Generator with Permanent Magnets. The laboratory comprises as a central facility
converter fed generator test systems of this four different types with a nominal power of 22 kW for
research. The basic structure is shown in fig. 1. Table I gives an overview to all basic structure drives
in the laboratory suited for wind energy research.

Figure 1: Wind turbine generator and equivalent set-up in the laboratory, electromechanical part, basic structure
Top: real wind turbine system; bottom: typical equivalent laboratory setup for wind turbine system

For investigations in the laboratory, special equipment is available. A grid emulator for delivering high
and low voltages as well as harmonics and asymmetries to these generators have been built. A special
test setup enables to analyze the drive control with oscillating, unknown loads with gear, as the drive
train in wind turbines, another research on winding fault monitoring. A grid impedance measurement
unit for a high frequency range, to identify the impedance for better energy feed in to the grid, has

EPE 2009 - Barcelona ISBN: 9789075815009 P.2


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

Table I: Power Electronic Generator Systems installed in the laboratory and research work
(all generators driven either by converter fed dc or ac machine)

Type of Drive Power Research Work

Generator Speed Sensorless Control


Doubly Fed Induction Generator, 22 kW, 400 V
Generator Sensor Fault Tolerant Control
Back-to-Back Converter at Rotor 1500 1/min
Extreme Voltage Ride Through
Doubly Fed Induction Generator, 2.2 kW, 400 V Condition Monitoring for Generator
Back-to-Back Converter at Rotor 1500 1/min Faults
Grid Side Control at
Induction Generator with
22 kW, 400 V Harmonics/Unsymmetries
Short Circuit Rotor,
1500 1/min Nonlinear Control
Back-to-Back Converter at Stator
LCL filter Control
Induction Generator with Grid Side Control at
Short Circuit Rotor, 5,5 kW, 400 V Harmonics/Unsymmetries
Back-to-Back Converter at Stator, 1500 1/min Nonlinear Control
Oscillating Load with Backlash LCL filter Control
Synchronous Generator
22 kW, 400 V
with Field Excitation, State Space and Nonlinear Load control
1500 1/min
Back-to-Back Converter at Stator
Synchronous Generator
22 kW, 400 V Speed Sensorless Control
with Permanent Magnets,
1000 1/min Extreme Voltage Ride Through
Back-to-Back Converter at Stator

been developed and is being used for analysis. Simulation models and programs for the four variants
of generators as well as the other subjects have been developed, including grid integration, wind
model and AC as well as VSC-HVDC interconnection to a wind park. The performance of the
laboratory systems and simulations has been verified by measurements results at real wind turbine
generators from publications. The research results of course have to be scaled up to the power range of
real wind turbines. Fig. 2 shows the laboratory with 12 workbenches. In the middle on the left as right
side the four main generators are installed, visible by the cubicles containing the converters.
Analyzing the power electronic generator components, systems, their interaction and their grid
interaction is done to a great deal by means of simulation. Here the simulation tools
Matlab/Simulink/Plecs, Simplorer, PSCAD, Spice and INCA3D are used.

Figure 2: Photo of the laboratory with major part for wind power research

EPE 2009 - Barcelona ISBN: 9789075815009 P.3


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

The Converters in the drive test set-ups are built in the institute. Control is realized either by
dSpace-systems, Tricore microcontrollers, FPGAs or mixed solutions. Electrical machines are
standard machines from industrial manufacturers, as shown in fig. 3.
For measurement, in general standard equipment is used as digital oscilloscopes, fourier analyzers and
power meters. Special equipment is a fast data acquisition system as used in wind industry.

Figure 3: Laboratory drive test bench for analysis and the verification of theoretical results for the behavior of
machines and converters

Research in the Laboratory


Results from several research projects for wind turbine generators are presented.
Laboratory Setup for Low Voltage Ride Through of Wind Turbines
Constantly stepped up grid codes require an adapted performance of the grid side converter or
generator performance and thus control methods. Laboratory test benches with down scaled power are
necessary to analyze these advanced control methods not only by simulation during the stage of
development or finally on the multi mega watt wind energy plant [6]. For generating grid faults and
unbalanced voltages a laboratory scaled, converter based grid emulator with a continuous apparent
power of 30 kVA has been developed. The emulator is able to provide different time based behaviors
of the three phase grid voltage with a high dynamic performance. Demands to the dynamic of the
emulator system are derived from an analysis of the failure propagation effects in large wind farms [7]
and literature [8]. Figure 4 shows the structure of the grid emulator laboratory setup, where the line

Figure 4: Structure of the laboratory set-up of the grid emulator and the line side converter

EPE 2009 - Barcelona ISBN: 9789075815009 P.4


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

side converter is investigated. The design and control structure of the grid emulator is explained in [9].
To give an example, fig. 5 shows the measured states of the grid emulator controlled to a voltage dip
of 50% of grid voltage.
The emulator can be connected to the four types of power electronic generator system of wind power
plants, which are available in the laboratory. The behavior of grid connected devices can be tested and
analyzed during grid faults, harmonic distortion unbalanced voltages, and over voltages. Thus, control
strategies for generator systems like doubly fed induction generators under unbalanced voltage dips
[10] are able to be investigated by experimental results. Other analyses [11] were led out by means of
these laboratory set-ups.

Emulator voltage during 50 ms voltage dip to 50% zoom in: fault occurs zoom in: fault cleared
ua
ub
uc 200 200 200
uRMS

100 100 100


Voltage [V]

0 0 0

-100 -100 -100

-200 -200 -200

0.23 0.24 0.25 0.26 0.27 0.28 0.29 0.3 0.31 0.32 0,25 0,255 0,26 0,3 0,305 0,31

Emulator active power during 50 ms voltage dip to 50% zoom in: fault occurs zoom in: fault cleared

30 30 30
active power [kW]

20 20 20

10 10 10

0 0 0
0.23 0.24 0.25 0.26 0.27 0.28 0.29 0.3 0.31 0.32 0,25 0,255 0,26 0,3 0,305 0,31
time [s] time [s] time [s]

Figure 5: Output voltages and power of the grid emulator with resistive load at a symmetrical voltage dip to 50%
of grid voltage (measurement)
Fault Ride Through in Wind Farms with FACTS
The amount of installed wind power connected to the grid worldwide is increasing notably. This
development has forced the grid operators to tighten their grid connection rules in order to limit the
effects of wind power parks on network quality and stability. Important issues are the steady state
active and reactive power feed in capability, continuously acting voltage control and fault ride through
behavior like low and high voltage ride through (LVRT, HVRT). This means, that grid codes demand
wind farms to stay connected to the grid and stabilize the grid voltage in case of grid faults.
Among the grid faults the voltage swell is a critical event that can be caused by switching off large
loads or switching on capacitor banks. When using wind turbines with doubly-fed induction generators
(DFIG) as shown in fig. 6 the operation during a grid fault is complex because the stator is directly
connected to the grid while the rotor is connected via converter, which enables control. With no
mitigation of the voltage fault the danger of destruction is imminent.
The application of Flexible AC Transmission Systems (FACTS) is analyzed to mitigate voltage swells
at DFIG based wind farms theoretically and by means of simulation results [12] and [13].

EPE 2009 - Barcelona ISBN: 9789075815009 P.5


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

Figure 6: Wind energy system with doubly fed induction generator

In order to avoid problems concerning grid faults most of the wind turbines installed are automatically
disconnected from the grid and reconnected when the fault is cleared. This behavior does not fulfill the
new grid code requirements in some countries concerning HVRT capability and thus wind farms are
not allowed to be connected to the grid. This is the reason why the HVRT capability of wind turbines
with doubly fed induction generators must be implemented. There are different ways to mitigate
voltage swells in distribution systems. Instead of designing all system components to be tolerant
against voltage swells or changing the conventional control methods, FACTS devices can be used to
lower the voltage at the wind power generator.
FACTS can be classified by series and shunt compensation devices. Among these the Dynamic
Voltage Restorer (DVR) [14] and the Static Compensator (StatCom) [15] are the most effective
devices, both based on the voltage source converter concept. A DVR is a power electronic apparatus
that protects a sensitive load from disturbances in the supply voltage by injecting voltage in series with
the load. The StatCom injects a reactive shunt current into the grid system to correct the voltage swell.
Fig. 7 shows simulation results for the mitigation of a 30 % symmetrical voltage swell without phase
angle variation by a DVR and StatCom device. Both devices can mitigate the voltage swell and
minimize the power oscillations in the generator.

Voltage (swell 30%) Voltage (swell 30%)


UWT UWT
U in [p.u.]
U in [p.u.]

1.2 Ug 1.2 Ug

1 1

0.8 0.8
0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8

WT power WT power
1 1
P,Q in [p.u.]
P,Q in [p.u.]

0 0

-1 -1
P P
-2 Q -2 Q

0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8

Figure 7: Simulation results of line voltage (upper) and power (lower) at the wind turbine for mitigation of a
30 % symmetrical voltage swell with Dynamic Voltage Restorer (left) and StatCom (right) (UWT voltage wind
turbine, Ug voltage grid)
Load-Frequency Control of Synchronous Areas Using a Wind Farm Connected via
HVDC-VSC
To connect offshore wind farms to the electrical grid onshore flexible, reliable and effective
transmission systems are needed. If the distance between the wind farm to the point of common
coupling is long, i.e. more than 100 km, connection via High Voltage DC (HVDC) transmission
systems makes sense due to reduced losses in transmission cables. The HVDC system decouples the
wind farm grid from the electrical grid. Decoupling effects of a HVDC system in terms of frequency
behavior on both converter stations lead to problems in primary reserve recall procedure. Load-
frequency control of synchronous areas using HVDC systems is investigated in this work.

EPE 2009 - Barcelona ISBN: 9789075815009 P.6


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

Figure 8: Overview of a HVDC-VSC connected windfarm with possible flow of active and reactive power
between farm and the utility gid

The investigated system consisting of a wind farm connected to the grid by a HVDC based on Voltage
Source Converters (VSC), shown in fig. 8, can be notional split in half at the DC-link of the HVDC
connection. According to the DC characteristic of the link, only active power can be transmitted
between both converters. To keep a constant DC-link voltage, active power injected into the DC-link
by one converter is fed into the grid by the second converter. As a result the HVDC-VSC connection
has a pass through character related to active power.
The frequency of the voltage within the windfarm is given by the wind farm HVDC converter. The
frequency generation at the HVDC converter of the wind farm is independent from the grid side
converter. Full power converter WPS have no rotational masses directly connected to the grid. DFIG
based WPS are also equipped with a converter, that will cover the machine characteristic from a PCC
point of view. In sum, no load-frequency dependent rotational generation is existing in the windfarm
grid.
The problem to adopt load-frequency control
to an HVDC-VSC connected wind farm can
finally be split up into two main sections.
First, the primary reserve has to be supplied
solely by the WPS. Therefore the wind farm
may be treated separately from the HVDC
connection. The HVDC system takes no
contribution to this process. Second, the
recall process of the reserve has to be
modified in contrast to standard applications
and has to be observed for the system
presented by wind farm and HVDC
connection.
Load frequency control of synchronous
areas can be adopted to a wind farm that is
connected to the grid via HVDC-VSC link.
For good performance a procedure should be
selected that ensures fast communication
between both HVDC converters. Stability
analysis and performance simulation on this
subject have been carried out. Fig. 9 shows
the power reserve recall of a wind turbine
generator using rotational speed control.
Figure 9: Time domain simulation of reserve recall (5%) [16].
of a 3.6 MW full power converter WPS using rotational
speed control

EPE 2009 - Barcelona ISBN: 9789075815009 P.7


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

Reality like Performance of Laboratory Drives / Wind Power Capture Optimization


(PMSM)
Downscaled power generator systems for wind turbines in the laboratory should for some experiments
show a similar dynamic behavior as the real application. A drive with permanent magnet synchronous
machine and voltage source inverter for wind power application has been modeled and a 15 kW test
sample has been built up in the laboratory. The behavior of the wind turbine has been emulated. This
has been done by implementing the behavior of wind velocity, rotor properties and its inertia to the
controlled dc machine driving the test generator [17]. A maximum power control is added. The drive
performance has been analyzed by means of simulation and measurements. Main subjects of the
analysis are the control performance and the dynamic power flow of the drive when driven by the
wind turbine. Fig. 10 shows the wind emulated, wind power and mechanical power versus time for a
wind turbine generator in the laboratory. Beyond former analysis [18] energy gain dependence on
control parameter selection has been analyzed [19, 20].

Figure 10: Wind velocity, wind power and generator mechanical power of a wind turbine generator in the
laboratory with low speed controller dynamics; wind power input realized by a controlled motor (measurement)
Control of grid-connected PWM converters with LCL filters and small DC link
capacitors
Nowadays, grid-connected PWM converters with DC link capacitors (see fig. 11) are more and more
used. Typical applications can be found in regenerative energy systems and regenerative drives, for
example in wind turbines. Optimization of the converter system for these applications is still in
progress. For instance, replacing the electrolytic DC link capacitors by film capacitors improves the
system life time and use of LCL filters instead of conventional L filters offers cost reduction as
smaller filter elements can be used [21, 22, 23]. For wind power applications use of LCL filters is
more important. For such high power applications the LCL filter is mandatory due to very limited
switching frequencies. In order to attenuate the switching current harmonics with reasonable filter
elements a sharp filtering characteristic is required.
The improvements with respect to filtering characteristic or life time come along with new challenges
the control has to cope with. LCL filter resonance causes unwanted oscillations or even instability.
Thus, active resonance damping by control is necessary since passive damping by resistors is
inefficient and inflexible. Reduced DC link capacitance requires fast voltage and current control in
order to maintain stability and to keep the voltage within a proper range. Moreover the background
line voltage distortions are to be damped in order to fulfill the standards.
The research of various control approaches is subject of this project. Cascaded control based on
conventional methods known from field-oriented motor control is state of the art in controlling the
grid-connected PWM converters as well. Various dedicated extensions to the control system are
necessary in order to solve the mentioned issues. Limitations of the conventional approaches are
clearly pointed out in this project [24]. Fig. 12 shows the need for active damping in case of PI
converter current control. Moreover, research emphasizes the application of modern control methods

EPE 2009 - Barcelona ISBN: 9789075815009 P.8


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

for handling the problems related to the LCL filter, the small DC link capacitors, and rejection of line
voltage distortions. Namely, PI state space control, flatness-based control [25], passivity-based
control, and Sliding Mode control [26] are studied. Promising results have been obtained. However,
further analyses and optimizations are in progress.
Experimental tests on a 22 kW laboratory test set-up with squirrel-cage motor (fig. 11) verify
theoretical results. The power converter configuration is a back-to-back converter with film capacitors.
By paralleling electrolytic capacitors higher, commonly used capacitances can be obtained. The
converter is connected directly or via transformer to the 400 V grid via L or LCL filter. It is a flexible
system with various possible filter element values. The induction generator is driven by a DC
machine, whereas generating and motoring operation is possible. The control is implemented on a
DSP that is part of a dSPACE 1006 system. For the measurements 16 parallel fast AD converters are
available. Control algorithms are written in C programming language and integrated into a Simulink
model. Real Time Workshop makes the interface to the hardware.

Figure 11: Induction motor drive system with grid-connected Figure 12: Measurement result: PI converter
PWM converter with LCL filter and DC capacitors. IM drive: current control without active damping (left)
inverter-fed squirrel cage induction motor/generator and with active damping (right) [24]

Fast Switching Converter for Grid Feeding for Impedance Measurement


The impedance of the electrical a.c. grid is an important information for feeding in high amount of
electrical power for example from wind farms. Research has been done in this field for measuring the
grid impedance in a broad frequency range up to 10 kHz, including the range of the switching
frequencies of converters in wind turbine systems.
For a three phase high frequency grid analysis (50 Hz ... 10 kHz) a fast switching converter to feed the
measuring currents to the grid with a fast and robust control is very important. In this project, a three-
phase converter with high output frequencies (switching frequency 50 kHz) for measuring the grid
impedance has been designed and developed [27]. The work comprises the complete development of
this converter with the power stack, the driver for the semiconductors and the control, realized as a
FPGA-based control structure with the hardware. A FPGA-based tolerance band or hysteresis control
meets the given control requirements. The converter has an output power of 15 kVA and a
measurement current in a frequency range from 50 Hz up to 10 kHz. The converter set-up is shown in
fig. 13.

Figure 13: Prototype of a measurement current injection converter for grid impedance measurement

EPE 2009 - Barcelona ISBN: 9789075815009 P.9


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

The selected method to measure the grid impedance is injecting a current signal at non characteristic
frequencies [28, 29]. By this method a variable frequency current source injects a current into the grid.
With the following voltage response at the same frequency it is possible to determine the magnitude
and phase of the impedance. A measurement of the various grid impedance magnitudes over a 10-hour
period in the laboratory is shown in fig. 14. The method was experimentally verified on many supply
impedance configurations. Fig. 15 shows measurements in the laboratory with two harmonic shunt
filters (series-resonant circuit) in series connected to the grid, close to the grid impedance analyzer.
The first harmonic filter (L = 0.5mH, C = 16 F) has a resonance frequency at fR1 = 1454Hz and the
second filter (L = 5.58mH, C = 240 nF) has a resonance frequency at fR2 = 4350Hz. The measurement
result corresponds relatively good with the calculated frequency. With this test set-up several
measurements in public grids at different places have been done together with Prof. Hinrichs of
university of applied science in Kiel.

Figure 14: Grid impedance measurement in the Figure 15: Grid impedance with and without two serial
laboratory within 10 hours harmonic filters (fR1 =1454Hz, fR2 = 4350Hz)

Control of Drive Systems with Resonant Load and Backlash


Torsional vibrations in electrical drive systems with elastically coupled loads are a well known
problem. The natural damping of such systems is very low and yields to a slow decay characteristic of
torsional oscillations.
Backlash is present in many mechanical systems, such as wind energy plants. If a motor is not directly
connected to the load, backlash effects can disturb the system. They yield to high torque impulses
which can excite torsional vibrations and reduce lifetime of the system significantly.
Conventional speed controlled drive systems include an inverter-fed motor which powers a load via
gear and shaft, as shown in fig. 16.
The aim of this research work is a high dynamic speed control with active damping of torsional
vibrations, limiting the influence of backlash and the adaption of unknown parameters. Active
damping of torsional vibrations in drive systems with elastic shafts yields to high requirements to the
controller. Conventionally, control structures with PI-controller are used for speed control of torsional
drive systems [30]. However, PI control methods without additional feedback are not appropriate to
damp torsional vibrations effectively. Therefore, modern control methods promise better results.
Varying or unknown system parameters which have an influence on the control performance, such as
the inertia of the load, should be estimated during the controlled operation. The influence of nonlinear
disturbances, such as backlash and friction, should be reduced as well.
A PI-state-space-controller for drive systems with elastic coupled loads has been designed analyzed
and compared with a conventional PI controller and has been optimized in [31, 32]. The state space
control yields a better control performance concerning dynamic and active damping of oscillations. As
well, state space control is appropriate to reduce the backlash effect significantly, as can be seen in fig.
18 and 19.

EPE 2009 - Barcelona ISBN: 9789075815009 P.10


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

Measurements were taken on the drive system shown in fig. 17. A 5.5 kW induction motor is
connected via backlash clutch, long shaft, torque sensor and flywheel to a 6.4 kW servo induction
machine which can induce a disturbance torque. For tests under different conditions, the system is
variable. The Flywheel and the shaft are exchangeable and the backlash gap is adjustable. The control
algorithm is implemented on a dSPACE DS1103 PPC controller board.

Figure 16: Overview of the drive system Figure 17: Laboratory setup

Figure 18: Measurement result: Motor and load speeds Figure 19: Measurement result: Motor and load speeds
of the PI-controlled drive system with backlash of the PI-state-space-controlled drive system with
backlash
Fault-tolerant Control of a Doubly-Fed Induction Machine
Different sensor signals as of current, voltage and rotor position are needed for the control of
adjustable speed drives. Operation of the drive is not possible anymore, if one sensor fails. Focus of
this work was the development of methods which allow an operation tolerant to sensor faults [33, 34]
for doubly-fed induction generators, mostly used for wind turbine generators. Literature on this subject
is rare, so results from squirrel-cage motors have to be the base [35].
It is possible to detect failures of rotor current, stator current, stator voltage and speed sensors and to
reconfigure the control at this fault. The control can operate continuously, because switching times to
fault tolerant control are sufficiently short. A time diagram of the used procedure is given in Fig. 20.
Fig. 21 shows the used observer bank. It comprises all three types of sensors. A parity equation, which
was derived from a stator flux model, has been added to each observer. It serves for the computation
of a residual, which ensures the fault isolation, if more than one observer is used. This means an exact
differentiation of a fault in the stator current, the rotor current, the stator voltage and the positioning
sensor.

Fault detected, Fault isolated,


switch to switch to
Fault occurs Open Loop Observer

Fault-free Observer
Closed Loop Faulty Open Loop
Measurement Based
Operation Operation
Closed Loop

Time
Figure 20: Temporal switching of the control to a open-loop control without measurement feedback

EPE 2009 - Barcelona ISBN: 9789075815009 P.11


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

Fig. 22 shows the schematic construction of the test rig with a 22 kW doubly-fed induction machine,
the frequency converter and the sensors needed for control purposes. The control is implemented on a
DSP of the company dSPACE. The DSP contains the algorithms for fault detection and isolation, too.

Stator Current
Observer
Stator Voltage
Terminal Rotor Current Inverter Stator Current
Measurements Observer
Inverter Inverter
Stator Voltage Current
Group 2 Observer FDI Crowbar M
M 3~
Rotor Angle Speed Grid
Voltage
Estimator Obs. 1 PWM PWM
Rotor
Position
Rotor
Position
Speed DC-Link Rotor
PWM Voltage Current
Group 1 Obs. 2
Terminal Reference
Measurements Values Field Oriented Control
Observed
Signals
Control
Fault Detection Observer

Figure 21: Observer bank Figure 22: Test-rig and used sensors

Fig. 23 and fig. 24 show successful reconfiguration processes of the positioning sensor by the means
of a short-term switching of the closed loop to an open loop control, like it is shown in fig.20. The
oscillogram of fig. 23 shows the measured rotor position (CH1), which is measured deficient, as well
as the stator current (CH2). The fault is detected and the control is switched to open-loop. In this case
the time of open-loop operation is displayed by the "open-loop flag" (CH4). Additionally the rotor
current is shown (CH3), which shows no disturbances. As can be seen, the machine stays in operation
and even at the switching of the control no transients or current peaks are visible.

Figure 23: Reconfiguration of the rotor positioning Figure 24: Reconfiguration of a positioning sensor:
sensor Progress of the internal residuals during the fault
isolation process

Fig. 24 shows the internal computation of the residuals, which allow the isolation of the positioning
sensor defect. The result of the fault isolation has to be continuously active for at least 4 ms for a as far
as possible reliable isolation. By this a good rejection of faults of primary nature (detection of a fault
without an existing fault) is reached.
Condition Monitoring of the DFIG via Electrical Signals
The focus of the project is put on analyzing the possibility for detection of slip-ring defects and
turn-to-turn faults at the rotor and the stator of doubly fed induction machines (DFIG). The
investigated slip-ring defect was misplaced brushes resulting in different resistances of the rotor
phases. They cause control problems because boundary conditions are not fulfilled anymore. They
also increase the losses at the rotor side. Turn-to-turn faults are a result of a degradation of the
insulation caused by over currents or high switching voltages of converters. At this a new phase is
generated and a high current is induced to it [36]. This fault initiates phase-to-phase or phase-to-earth
shorts, which can partly be avoided by a fault tolerant operation.
The frequency based Motor Current Signature Analysis (MCSA) and the state space based Luenberger
observer are investigated in detail. The fault indicators of the MCSA are additional sidebands in the

EPE 2009 - Barcelona ISBN: 9789075815009 P.12


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

motor current. A lot of publications exist about this method especially for squirrel-cage induction
machines and for permanent magnet synchronous machines. The fault frequencies are:
f f ,stator = f stator 2 f rotor (in the stator current for rotor faults) (1)

f f ,rotor = 2 f stator f rotor (in the rotor current for stator faults) (2)

A simulation result of the MCSA on the rotor turn-to-turn fault is shown in fig. 25. A problem about
this method is to distinguish slip-ring and rotor turn-to-turn faults because they have the same
indicators. Also the response time is high and for slip-ring and rotor faults detection is problematic
close to synchronous speed where (1) becomes fstator.
The Luenberger observer has a different approach. It works with a model parallel to the system and a
fitting feedback of the system and model outputs. The differences are called residuals. In case of
consistency of model and system the residuals would always be zero. But at different initial values or
small differences of the parameters a feedback has to be implemented. Normally it is done by placing
the observer poles to the left s-half plane. So the residuals should converge to zero.
The model has a wrong structure in case of turn-to-turn faults. This initiates a permanent difference of
the residuals, which can be used for detection and identification of the fault and the severity of the
fault. An example also of the rotor turn-to-turn fault is given in fig. 26. In contrast to MCSA the time
delay of the fault indicator is much shorter. Furthermore it is possible to distinguish slip-ring and rotor
faults by fitting fault observers.
All simulation results [37] will be validated in laboratory this year. For this work are three 2.2 kW
slip-ring induction machines, one B6C inverter, a radio transmission system based on aTmega 8
microcontrollers and two TC1796 microcontrollers for control and detection algorithms is available at
the institute. The algorithms will be programmed in C and used on a TC1796 microcontroller. This
will show the functionality and the computation demand of each algorithm.
Further methods as negative sequence current or impedance, neural network based methods will be
investigated after the verification of the used machine model.

Figure 25: Hamming filtered spectrogram with 4kHz Figure 26: Observer residuals at a rotor turn-to-turn
sampling rate and a time shift of 0.25 sec per 0.5 sec fault at 1.1 sec versus time
Spectrum at a rotor turn-to-turn fault; fault indication
at 30 Hz

Conclusion
Power generated from wind is actually the greatest and a very fast growing renewable source of the
world energy consumption. Optimizing performance and costs is an important issue for wind turbine
generators. A research laboratory for power electronic generator systems in wind turbines, converters,
generators, their interaction and grid integration has been presented enabling the possibility to
investigate and compare the four types of variable speed generators used today. Time and cost
consuming analysis on real wind turbines thus can be minimized. Special developed devices to
produce test conditions and for measuring are part of the lab. Actual research results have been
presented and show their great scope. These are low voltage ride through, high voltage ride through
with FACTS, wind park energy feed in via HVDC, wind fluctuation and energy gain, grid side

EPE 2009 - Barcelona ISBN: 9789075815009 P.13


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

converter control optimization, grid fault propagation, grid impedance measurement, control of the
torsional drive train, sensor fault tolerant operation as well as electrical condition monitoring of
doubly fed induction generators. Further investigations are actually carried out, and will be done in
future.

References
[1] Special Section on Renewable Energy And Distributed Generated Systems, Industrial Electronics IEEE
Transactions on, Volume 53, Number 4, August 2006
[2] IEEE Transactions on Power Systems, Volume 22, Number 3, August 2007
[3] Lie Xu; Liangzhong Yao; Sasse, C.: Grid Integration of Large DFIG-Based Wind Farms Using VSC
Transmission, Power Systems, IEEE Transactions on; Volume: 22, Issue: 3, pp.: 976 - 984
[4] Blaabjerg, F.; Zhe Chen; Teodorescu, R.: Renewable energy systems in the power electronics curriculum,
Power Electronics Education, 2005. IEEE Workshop; June 16-17, 2005; pp.: 58 - 68
[5] Emadi, A.: Grainger power electronics and motor drives laboratories at Illinois Institute of Technology,
Power Engineering Society General Meeting, 2005. IEEE, 12-16 June 2005 pp.: 1168 - 1175 Vol. 2
[6] Lohde, R.; Jensen, S.; Knop, A.; Fuchs, F. W.: Analysis of Three Phase Grid Failure and Doubly Fed
Induction Generator Ride-through using Crowbars; EPE 07 European Conference on Power Electronics
and Drives, Aalborg, Denmark, 2007
[7] Lohde, R; Fuchs, F. W.: Analysis of High and Low Voltage Grid Failure Propagation in large Wind Farms
considering Transformers, Cables and VAR-Compensators, IEEE 39th Power Electronics Specialists
Conference, Rhodes, Greece, 2008
[8] S. Richter, J. von Bloh, C. Dick, D. Hirschmann, R. De Doncker: Control of a Medium-Voltage Test
Generator, 39th IEEE Power Electronics Specialist Conference, Rhodes, Greece, 2008
[9] Lohde, R.; Fuchs, F. W.: Laboratory Type PWM Grid Emulator for Testing Grid Connected PWM
Converters of Faults and Unbalances, EPE 09 European Conference on Power Electronics and Drives,
Barcelona, Spain 2009, accepted for publication
[10] Gomis-Bellmunt, O.; Junyent-Ferr, A.; Sumper, A.; Bergas-Jan, J.: Ride-Through Control of Doubly Fed
Induction Generator Under Unbalanced Voltage Sags, IEEE Transactions on Energy Conversion, Dec08
Vol. 23 Nr 4 pp 1036 - 1045, 2008
[11] Lohde, R.; Fuchs, F. W.: Comparative Analysis of Standard Speed Control and State Space Control of
Permanent Magnet Synchronous Generator in Wind Power Plants, EPE Wind Energy Chapter 2nd
Seminar, Stockholm, Sweden, 2009
[12] Wessels, C.; Fuchs, F:W.: Concept and Performance of Voltage Swell Mitigation in Wind Farms with
FACTS; EWEC 2009 Marseille; 16-19 March 2009
[13] Wessels, C.; Fuchs, F.W.: High Voltage Ride-Through with FACTS for DFIG Based Wind Turbines;
accepted for publication at European Conference on Power Electronics and Applications 2009, 8-10 Sept.
2009
[14] Awad, H. and Blaabjerg, F.; Mitigation of voltage swells by static series compensator, Power Electronics
Specialists Conference, June 2004.
[15] Wei Qiao, Venayagamoorthy, G.K. and Harley, R.G.. Real-Time Implementation of a STATCOM on a
Wind Farm Equipped With Doubly Fed Induction Generators. Industry Applications, IEEE Transactions
on, 45(1):98-107, Jan.-feb. 2009.
[16] Jensen, S.; Fuchs, F. W.: Load-Frequency Control of Synchronous Areas Using a Wind Farm Connection
via HVDC-VSC, EPE Wind Energy Chapter, 2nd Seminar, Stockholm, Sweden, April, 2009
[17] Hnemrder, S., Bierhoff, M., Fuchs, F. W: Drive with Permanent Magnet Synchronous Machine and
Voltage Source Inverter for Wind Power Application, NORPIE 2002,Nordic Workshop on Power and
Industrial Electronics, Stockholm, Sweden, 2002
[18] McIver, A., Holmes, D.G. and Freere, P.: Optimal control of a variable speed wind turbine under dynamic
wind conditions, Thirty-First IAS Annual Meeting, Conference Record, 3, pp. 1692-1698, Oct 1996.
[19] Jensen, S.; Fuchs, F. W.: Dynamic Operation and Energy Gain of a Wind Power Station with Converter
Fed Permanent Magnet Synchronous Machine, NORPIE 2006; Lund, Schweden
[20] Jensen, S., Fuchs, F. W: Detailed Analysis of the Wind Related Power Fluctuations an Energy Gain of a
PMSM Wind Power Station, DEWEK 2006; Bremen, Germany
[21] M. Liserre, F. Blaabjerg, and S. Hansen: Design and control of an LCL-filter-based three-phase active
rectifier, IEEE Trans. on Industry Applications, vol. 41, no. 5, pp. 1281-91, 2005.

EPE 2009 - Barcelona ISBN: 9789075815009 P.14


Research Laboratory for Power Electronic Generator Systems in Wind Turbines FUCHS Friedrich W.
Comprising Converters, Generators, Interaction and Grid Interaction

[22] Klumpner, C. Liserre, M. Blaabjerg, F. : Improved control of an active-front-end adjustable speed drive
with a small de-link capacitor under real grid conditions, Power Electronics Specialists Conference, 2004.
PESC 04. 2004, 1156 1162.
[23] Bierhoff, M.H.; Fuchs, F. W.: Active Damping for Three-Phase PWM Rectifiers With High-Order Line-
Side Filters, Industrial Electronics, IEEE Transactions on, volume 56, Issue 2, Feb. 2009 pp.: 371 - 379.
[24] Dannehl, J.; Wessels, C.; Fuchs, F. W.: Limitations of Voltage-Oriented PI Current Control of Grid-
Connected PWM Rectifiers With LCL Filters, Industrial Electronics, IEEE Transactions on, volume 56,
Issue 2, Feb. 2009 pp.:380 388.
[25] Dannehl, J.; Fuchs, F.W.: Flatness-based voltage-oriented control of three-phase PWM rectifiers; Power
Electronics and Motion Control Conference, 2008. EPE-PEMC 2008. 13th,
1-3 Sept. 2008, pp.:444 450.
[26] Dannehl, J.; Fuchs, F.W.: Discrete Sliding-Mode Current Control of Grid-Connected PWM Converters,
accepted for publication at Power Electronics and Applications, 13th European Conference on, Barcelona,
Spain, September, 2009
[27] A. Knop and F. W. Fuchs: High frequency grid impedance analysis with three-phase converter and fpga
based tolerance band controller, in CPE, 2009.
[28] J. P. Rhode, A. W. Kelley, and M. E. Baran: Complete characterization of utilization-voltage power system
impedance using wideband measurement, IEEE Transactions on Industry Applications, vol. 33, no. 6, pp.
14721479, 1997.
[29] Knop and F. W. Fuchs: High Frequency Grid Impedance Analysis with Current Injection, in IECON 2009,
submitted for publication
[30] Szabat, K., Orlowska, K. T.: Vibration suppression in a two-mass drive system using PI speed controller
and additional feedbacks-comparative study, IEEE Transactions on Industrial Electronics, vol. 54, no. 2,
pp.1193-1206, 2007.
[31] Thomsen, S. , Fuchs, F.W.: Speed Control of Torsional Drive Systems with Backlash, accepted for
publication at Power Electronics and Applications, 13th European Conference on, Barcelona, Spain,
September, 2009
[32] Thomsen, S., Fuchs, F.W.: Conventional and State Space Control for Active Damping of Mechanical
Vibrations in Speed Drive Systems with Backlash, VDE/VDI-Fachtagung - Elektrisch-mechanische
Antriebssysteme, Bblingen - Deutschland, 2008.
[33] Rothenhagen, K.; Fuchs, F. W.; Doubly-fed Induction Generator Model-based Sensor Fault Detection and
Control Loop Reconfiguration, Industrial Electronics IEEE Transactions on, accepted for future publication
in february 2009, published in the internet
[34] Baghli, L., Poure, P. and Rezzoug, A.: Sensor fault detection for fault tolerant vector controlled induction
machine, Power Electronics and Applications, 11th European Conference on, Dresden, Germany,
September, 2005. pp.10-P.10,
[35] Bennett, S.M., Patton, R.J., Daley, S., and Newton, D.A.: Model based intermittent fault tolerance in an
induction motor drive, Proc IMACS Multiconference: CESA'96, Lille, 9-12 July, Vol. 1, pp. 678-683.
[36] Thomson, WT.: On-lineMCSA to diagnose shorted turns in low voltage stator windings of 3-phase
induction motors prior to failure. in IEMDC 2001. IEEE International Electric Machines and Drives
Conference Cat. No.01EX485, 2001, p. 891.
[37] V. Dinkhauser and F. W. Fuchs: Rotor Turn-to-Turn Faults of doubly-fed Induction Generators in Wind
Energy Plants - Modelling, Simulation and Detection, 13th International Power Electronics and Motion
Control Conference (EPE-PEMC 2008),Poznan, Poland 2008, pp. 1842-1849.

EPE 2009 - Barcelona ISBN: 9789075815009 P.15

View publication stats