This action might not be possible to undo. Are you sure you want to continue?

# THESIS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

**Analysis, Modeling and Control of Doubly-Fed
**

Induction Generators for Wind Turbines

ANDREAS PETERSSON

Division of Electric Power Engineering

Department of Energy and Environment

CHALMERS UNIVERSITY OF TECHNOLOGY

G¨ oteborg, Sweden 2005

Analysis, Modeling and Control of Doubly-Fed Induction

Generators for Wind Turbines

ANDREAS PETERSSON

ISBN 91-7291-600-1

c ( ANDREAS PETERSSON, 2005.

Doktorsavhandlingar vid Chalmers tekniska h¨ ogskola

Ny serie nr. 2282

ISSN 0346-718x

Division of Electric Power Engineering

Department of Energy and Environment

Chalmers University of Technology

SE-412 96 G¨ oteborg

Sweden

Telephone + 46 (0)31-772 1000

Chalmers Bibliotek, Reproservice

G¨ oteborg, Sweden 2005

Analysis, Modeling and Control of Doubly-Fed Induction Generators for Wind Turbines

ANDREAS PETERSSON

Division of Electric Power Engineering

Department of Energy and Environment

Chalmers University of Technology

Abstract

This thesis deals with the analysis, modeling, and control of the doubly-fed induction gener-

ator (DFIG) for wind turbines. Different rotor current control methods are investigated with

the objective of eliminating the inﬂuence of the back electromotive force (EMF), which is

that of, in control terminology, a load disturbance, on the rotor current. It is found that the

method that utilizes both feed forward of the back EMF and so-called “active resistance”

manages best to suppress the inﬂuence of the back EMF on the rotor current, particularly

when voltage sags occur, of the investigated methods. This method also has the best stability

properties. In addition it is found that this method also has the best robustness to parameter

deviations.

The response of the DFIG wind turbine system to grid disturbances is simulated and ver-

iﬁed experimentally. A voltage sag to 80% (80% remaining voltage) is handled very well.

Moreover, a second-order model for prediction of the response of small voltage sags of the

DFIG wind turbines is derived, and its simulated performance is successfully veriﬁed exper-

imentally.

The energy production of the DFIG wind turbine is investigated and compared to that of

other wind turbine systems. The result found is that the energy capture of the DFIG wind tur-

bine is almost the same as for an active stall-controlled ﬁxed-speed (using two ﬁxed speeds)

wind turbine. Compared to a full-power-converter wind turbine the DFIG wind turbine can

deliver a couple of percentage units more energy to the grid.

Voltage sag ride-through capabilities of some different variable-speed wind turbines has

been investigated. It has been found that the energy production cost of the investigated wind

turbines with voltage sag ride-through capabilities is between 1–3 percentage units higher

than that of the ordinary DFIG wind turbine without the ride-through capability.

Finally, a ﬂicker reduction control law for stall-controlled wind turbines with induction

generators, using variable rotor resistance, is derived. The ﬁnding is that it is possible to

reduce the ﬂicker contribution by utilizing the derived rotor resistance control law with 40–

80% depending on the operating condition.

Keywords: Doubly-fed induction generator, wind turbine, wind energy, current control,

voltage sag, power quality.

iii

iv

Acknowledgements

This research project has been carried out at the Department of Energy and Environment

(and the former Department of Electric Power Engineering) at Chalmers University of Tech-

nology. The ﬁnancial support provided by the Swedish National Energy Agency is gratefully

acknowledged.

I would like to thank my supervisors Dr. Torbj¨ orn Thiringer and Prof. Lennart Harnefors

for help, inspiration, and encouragement. I would also like to thank my examiner Prof. Tore

Undeland for valuable comments and encouragement. Thanks goes to my fellow Ph.D. stu-

dents who have assisted me: Stefan Lundberg for a pleasant collaboration with the efﬁciency

calculations, Dr. Rolf Ottersten for many interesting discussions and a nice cooperation, es-

pecially with the analysis of the full-power converter, Dr. Tom´ aˇ s Petr˚ u for valuable and time

saving collaboration with practical ﬁeld measurement set-ups, and Oskar Wallmark for a

good companionship and valuable discussions.

Many thanks go to the colleagues at the Division of Electric Power Engineering and the

former Department of Electric Power Engineering, who have assisted me during the work of

this Ph.D. thesis.

Finally, I would like to thank my family for their love and support.

v

vi

Table of Contents

Abstract iii

Acknowledgements v

Table of Contents vii

1 Introduction 1

1.1 Review of Related Research . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2 Purpose and Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3 List of Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Wind Energy Systems 7

2.1 Wind Energy Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.1.1 Wind Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.1.2 Aerodynamic Power Control . . . . . . . . . . . . . . . . . . . . . 8

2.1.3 Aerodynamic Conversion . . . . . . . . . . . . . . . . . . . . . . . 8

2.2 Wind Turbine Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.2.1 Fixed-Speed Wind Turbine . . . . . . . . . . . . . . . . . . . . . . 11

2.2.2 Variable-Speed Wind Turbine . . . . . . . . . . . . . . . . . . . . 11

2.2.3 Variable-Speed Wind Turbine with Doubly-Fed Induction Generator 12

2.3 Doubly-Fed Induction Generator Systems for Wind Turbines . . . . . . . . 13

2.3.1 Equivalent Circuit of the Doubly-Fed Induction Generator . . . . . 14

2.3.2 Power Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.3.3 Stator-to-Rotor Turns Ratio . . . . . . . . . . . . . . . . . . . . . 17

2.3.4 Lowering Magnetizing Losses . . . . . . . . . . . . . . . . . . . . 18

2.3.5 Other Types of Doubly-Fed Machines . . . . . . . . . . . . . . . . 19

3 Energy Efﬁciency of Wind Turbines 23

3.1 Determination of Power Losses . . . . . . . . . . . . . . . . . . . . . . . . 23

3.1.1 Aerodynamic Losses . . . . . . . . . . . . . . . . . . . . . . . . . 23

3.1.2 Gearbox Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3.1.3 Induction Generator Losses . . . . . . . . . . . . . . . . . . . . . 24

3.1.4 Converter Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.1.5 Total Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

3.2 Energy Production of the DFIG System . . . . . . . . . . . . . . . . . . . 29

3.2.1 Investigation of the Inﬂuence of the Converter’s Size on the Energy

Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

vii

3.2.2 Reduction of Magnetizing Losses . . . . . . . . . . . . . . . . . . 31

3.3 Comparison to Other Wind Turbine Systems . . . . . . . . . . . . . . . . . 31

3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

4 Control of Doubly-Fed Induction Generator System 35

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.1.1 Space Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.1.2 Power and Reactive Power in Terms of Space Vectors . . . . . . . . 36

4.1.3 Phase-Locked Loop (PLL)-Type Estimator . . . . . . . . . . . . . 36

4.1.4 Internal Model Control (IMC) . . . . . . . . . . . . . . . . . . . . 37

4.1.5 “Active Damping” . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.1.6 Saturation and Integration Anti-Windup . . . . . . . . . . . . . . . 40

4.1.7 Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.2 Mathematical Models of the DFIG System . . . . . . . . . . . . . . . . . . 41

4.2.1 Machine Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.2.2 Grid-Filter Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.2.3 DC-Link Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.3 Field Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.3.1 Stator-Flux Orientation . . . . . . . . . . . . . . . . . . . . . . . . 45

4.3.2 Grid-Flux Orientation . . . . . . . . . . . . . . . . . . . . . . . . 46

4.4 Control of Machine-Side Converter . . . . . . . . . . . . . . . . . . . . . . 47

4.4.1 Current Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4.4.2 Torque Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

4.4.3 Speed Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

4.4.4 Reactive Power Control . . . . . . . . . . . . . . . . . . . . . . . 52

4.4.5 Sensorless Operation . . . . . . . . . . . . . . . . . . . . . . . . . 54

4.5 Control of Grid-Side Converter . . . . . . . . . . . . . . . . . . . . . . . . 55

4.5.1 Current Control of Grid Filter . . . . . . . . . . . . . . . . . . . . 56

4.5.2 DC-Link Voltage Control . . . . . . . . . . . . . . . . . . . . . . . 56

5 Evaluation of the Current Control of Doubly-Fed Induction Generators 59

5.1 Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

5.1.1 Stator-Flux-Oriented System . . . . . . . . . . . . . . . . . . . . . 59

5.1.2 Grid-Flux-Oriented System . . . . . . . . . . . . . . . . . . . . . 64

5.1.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.2 Inﬂuence of Erroneous Parameters on Stability . . . . . . . . . . . . . . . 66

5.2.1 Leakage Inductance, L

σ

. . . . . . . . . . . . . . . . . . . . . . . 67

5.2.2 Stator and Rotor Resistances, R

s

and R

R

. . . . . . . . . . . . . . 67

5.3 Experimental Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

5.3.1 Comparison Between Stator-Flux and Grid-Flux-Oriented System . 71

5.4 Impact of Stator Voltage Sags on the Current Control Loop . . . . . . . . . 71

5.4.1 Inﬂuence of Erroneous Parameters . . . . . . . . . . . . . . . . . . 73

5.4.2 Generation Capability During Voltage Sags . . . . . . . . . . . . . 74

5.5 Flux Damping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

5.5.1 Stator-Flux Orientation . . . . . . . . . . . . . . . . . . . . . . . . 76

viii

5.5.2 Grid-Flux Orientation . . . . . . . . . . . . . . . . . . . . . . . . 76

5.5.3 Parameter Selection . . . . . . . . . . . . . . . . . . . . . . . . . 76

5.5.4 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

5.5.5 Response to Symmetrical Voltage Sags . . . . . . . . . . . . . . . 77

5.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

6 Evaluation of Doubly-Fed Induction Generator Systems 81

6.1 Reduced-Order Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

6.2 Discretization of the Doubly-Fed Induction Generator . . . . . . . . . . . . 81

6.2.1 Stator-Flux Orientation . . . . . . . . . . . . . . . . . . . . . . . . 82

6.2.2 Grid-Flux Orientation . . . . . . . . . . . . . . . . . . . . . . . . 82

6.3 Response to Grid Disturbances . . . . . . . . . . . . . . . . . . . . . . . . 83

6.4 Implementation in Grid Simulation Programs . . . . . . . . . . . . . . . . 87

6.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

7 Voltage Sag Ride-Through of Variable-Speed Wind Turbines 89

7.1 Voltage Sags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

7.1.1 Symmetrical Voltage Sags . . . . . . . . . . . . . . . . . . . . . . 90

7.1.2 Unsymmetrical Voltage Sags . . . . . . . . . . . . . . . . . . . . . 90

7.2 Full-Power Converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

7.2.1 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

7.2.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

7.2.3 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

7.2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

7.3 Doubly-Fed Induction Generator with Shunt Converter . . . . . . . . . . . 102

7.3.1 Response to Small Voltage Sags . . . . . . . . . . . . . . . . . . . 103

7.3.2 Response to Large Voltage Sags . . . . . . . . . . . . . . . . . . . 110

7.3.3 Candidate Ride-Through System . . . . . . . . . . . . . . . . . . . 111

7.3.4 Evaluation of the Ride-Through System . . . . . . . . . . . . . . . 114

7.4 Doubly-Fed Induction Generator with Series Converter . . . . . . . . . . . 118

7.4.1 Possible System Conﬁgurations . . . . . . . . . . . . . . . . . . . 118

7.4.2 System Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

7.4.3 Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

7.4.4 Speed Control Operation . . . . . . . . . . . . . . . . . . . . . . . 125

7.4.5 Response to Voltage Sags . . . . . . . . . . . . . . . . . . . . . . 126

7.4.6 Steady-State Performance . . . . . . . . . . . . . . . . . . . . . . 127

7.4.7 Discussion and Conclusion . . . . . . . . . . . . . . . . . . . . . . 131

7.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

8 Flicker Reduction of Stalled-Controlled Wind Turbines using Variable Rotor

Resistances 133

8.1 Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

8.1.1 Reduced-Order Model . . . . . . . . . . . . . . . . . . . . . . . . 134

8.2 Current Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

8.2.1 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

8.3 Reference Value Selection . . . . . . . . . . . . . . . . . . . . . . . . . . 139

8.3.1 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

ix

8.4 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

8.4.1 Flicker Contribution . . . . . . . . . . . . . . . . . . . . . . . . . 143

8.4.2 Flicker Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

8.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

9 Conclusion 147

9.1 Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

References 149

A Nomenclature 159

B Data and Experimental Setup 163

B.1 Data of the DFIG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

B.2 Laboratory Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

B.2.1 Data of the Induction Generator . . . . . . . . . . . . . . . . . . . 164

B.3 Jung Data Acquisition Setup . . . . . . . . . . . . . . . . . . . . . . . . . 165

x

Chapter 1

Introduction

The Swedish Parliament adopted new energy guidelines in 1997 following the trend of mov-

ing towards an ecologically sustainable society. The energy policy decision states that the

objective is to facilitate a change to an ecologically sustainable energy production system.

The decision also conﬁrmed that the 1980 and 1991 guidelines still apply, i.e., that the nu-

clear power production is to be phased out at a slow rate so that the need for electrical energy

can be met without risking employment and welfare. The ﬁrst nuclear reactor of Barseb¨ ack

was shut down 30th of November 1999. Nuclear power production shall be replaced by im-

proving the efﬁciency of electricity use, conversion to renewable forms of energy and other

environmentally acceptable electricity production technologies [97]. According to [97] wind

power can contribute to fulﬁlling several of the national environmental quality objectives de-

cided by Parliament in 1991. Continued expansion of wind power is therefore of strategic

importance. The Swedish National Energy Agency suggest that the planning objectives for

the expansion of wind power should be 10 TWh/year within the next 10–15 years [97]. In

Sweden, by the end of 2004, there was 442 MW of installed wind power, corresponding to

1% of the total installed electric power in the Swedish grid [23, 98]. These wind turbines

produced 0.8 TWh of electrical energy in 2004, corresponding to approximately 0.5% of the

total generated and imported electrical energy [23, 98].

Wind turbines (WTs) can either operate at ﬁxed speed or variable speed. For a ﬁxed-

speed wind turbine the generator is directly connected to the electrical grid. For a variable-

speed wind turbine the generator is controlled by power electronic equipment. There are

several reasons for using variable-speed operation of wind turbines; among those are pos-

sibilities to reduce stresses of the mechanical structure, acoustic noise reduction and the

possibility to control active and reactive power [11]. Most of the major wind turbine man-

ufactures are developing new larger wind turbines in the 3-to-5-MW range [3]. These large

wind turbines are all based on variable-speed operation with pitch control using a direct-

driven synchronous generator (without gearbox) or a doubly-fed induction generator (DFIG).

Fixed-speed induction generators with stall control are regarded as unfeasible [3] for these

large wind turbines. Today, doubly-fed induction generators are commonly used by the wind

turbine industry (year 2005) for larger wind turbines [19, 29, 73, 105].

The major advantage of the doubly-fed induction generator, which has made it popular,

is that the power electronic equipment only has to handle a fraction (20–30%) of the total

system power [36, 68, 110]. This means that the losses in the power electronic equipment can

1

be reduced in comparison to power electronic equipment that has to handle the total system

power as for a direct-driven synchronous generator, apart from the cost saving of using a

smaller converter.

1.1 Review of Related Research

According to [12] the energy production can be increased by 2–6% for a variable-speed wind

turbine in comparison to a ﬁxed-speed wind turbine, while in [112] it is stated that the in-

crease in energy can be 39%. In [69] it is shown that the gain in energy generation of the

variable-speed wind turbine compared to the most simple ﬁxed-speed wind turbine can vary

between 3–28% depending on the site conditions and design parameters. Efﬁciency calcu-

lations of the DFIG system have been presented in several papers, for instance [52, 86, 99].

A comparison to other electrical systems for wind turbines are, however, harder to ﬁnd. One

exception is in [16], where Datta et al. have made a comparison of the energy capture for

various WT systems. According to [16] the energy capture can be signiﬁcantly increased by

using a DFIG. They state an increased energy capture of a DFIG by over 20% with respect to

a variable-speed system using a cage-bar induction machine and by over 60% in comparison

to a ﬁxed-speed system. One of the reasons for the various results is that the assumptions

used vary from investigation to investigation. Factors such as speed control of variable-speed

WTs, blade design, what kind of power that should be used as a common basis for compari-

son, selection of maximum speed of the WT, selected blade proﬁle, missing facts regarding

the base assumptions etc, affect the outcome of the investigations. There is thus a need to

clarify what kind of energy capture gain there could be when using a DFIG WT, both com-

pared to another variable-speed WT and towards a traditional ﬁxed-speed WT.

Control of the DFIG is more complicated than the control of a standard induction ma-

chine. In order to control the DFIG the rotor current is controlled by a power electronic

converter. One common way of controlling the rotor current is by means of ﬁeld-oriented

(vector) control. Several vector control schemes for the DFIG have been proposed. One

common way is to control the rotor current with stator-ﬂux orientation [46, 61, 80, 99], or

with air-gap-ﬂux orientation [107, 110]. If the stator resistance can be considered small,

stator-ﬂux orientation gives in principle orientation also with the stator voltage (grid-ﬂux

orientation) [17, 61, 68]. Wang et al. [107] have by simulations found that the ﬂux is in-

ﬂuenced both by load changes and stator power supply variations. The ﬂux response to a

disturbance is a damped oscillation. Heller et al. [43] and Congwei et al. [13] have inves-

tigated the stability of the DFIG analytically, showing that the dynamics of the DFIG have

poorly damped eigenvalues (poles) with a corresponding natural frequency near the line fre-

quency, and, also, that the system is unstable for certain operating conditions, at least for a

stator-ﬂux-oriented system. These poorly damped poles inﬂuence the rotor current dynamics

through the back electromotive force (EMF). The author has, however, not found in the lit-

erature any evaluation of the performance of different rotor current control laws with respect

to eliminating the inﬂuence of the back EMF, which is dependent on the stator voltage, rotor

speed, and stator ﬂux, in the rotor current.

The ﬂux oscillations can be damped in some different ways. One method is to reduce the

bandwidth of the current controllers [43]. Wang et al. [107] have introduced a ﬂux differ-

entiation compensation that improves the damping of the ﬂux. Kelber et al. [54] have used

2

another possibility; to use an extra (third) converter that substitutes the Y point of the stator

winding, i.e., an extra degree of freedom is introduced that can be used to actively damp the

ﬂux oscillations. Kelber has in [55] made a comparison of different methods of damping the

ﬂux oscillations. It was found that the methods with a ﬂux differentiation compensation and

the method with an extra converter manage to damp the oscillations best.

The response of wind turbines to grid disturbances is an important issue, especially since

the rated power of wind-turbine installations steadily increases. Therefore, it is important

for utilities to be able to study the effects of various voltage sags and, for instance, the cor-

responding wind turbine response. For calculations made using grid simulation programs, it

is of importance to have as simple models as possible that still manage to model the dynam-

ics of interest. In [22, 26, 60, 84], a third-order model has been proposed that neglects the

stator-ﬂux dynamics of the DFIG. This model gives a correct mean value [22] but a draw-

back is that some of the main dynamics of the DFIG system are also neglected. In order

to preserve the dynamic behavior of the DFIG system, a slightly different model approach

must be made. As described earlier a dominating feature of the DFIG system is the natural

frequency of the ﬂux dynamics, which is close to the line frequency. Since the dynamics

of the DFIG are inﬂuenced by two poorly damped eigenvalues (poles) it would be natural

to reduce the model of the DFIG to the ﬂux dynamics described by a second-order model.

This is a common way to reduce the DFIG model in classical control theory stability analy-

sis [13, 43]. The possibility to use it as simulation model remains to be shown. In order to

preserve the behavior of an oscillatory response, it is obvious that a second-order model is

the simplest that can be used.

New grid codes will require WTs and wind farms to ride through voltage sags, meaning

that normal power production should be re-initiated once the nominal grid voltage has been

recovered. Such codes are in progress both in Sweden [96] and in several other countries

[8]. These grid codes will inﬂuence the choice of electrical system in future WTs, which has

initiated industrial research efforts [8, 20, 28, 30, 42, 72] in order to comply. Today, the DFIG

WT will be disconnected from the grid when large voltage sags appear in the grid. After the

DFIG WT has been disconnected, it takes some time before the turbine is reconnected to the

grid. This means that new WTs have to ride through these voltage sags. The DFIG system,

of today, has a crowbar in the rotor circuit, which at large grid disturbances has to short

circuit the rotor circuit in order to protect the converter. This leads to that the turbine must

be disconnected from the grid, after a large voltage sag.

In the literature there are some different methods to modify the DFIG system in order to

accomplish voltage sag ride-through proposed. In [20] anti-parallel thyristors is used in the

stator circuit in order to achieve a quick (within 10 ms) disconnection of the stator circuit,

and thereby be able to remagnetize the generator and reconnect the stator to the grid as fast

as possible. Another option proposed in [72] is to use an “active” crowbar, which can break

the short circuit current in the crowbar. A third method, that has been mentioned earlier,

is to use an additional converter to substitute the Y point of the stator circuit [54, 55]. In

[55], Kelber has shown that such a system can effectively damp the ﬂux oscillations caused

by voltage sags. All of these systems have different dynamical performance. Moreover,

the efﬁciency and cost of the different voltage sag ride-through system might also inﬂuence

the choice of system. Therefore, when modifying the DFIG system for voltage sag ride-

3

through it is necessary to evaluate consequences for cost and efﬁciency. Any evaluation of

different voltage sag ride-through methods for DFIG wind turbines and how they affect the

efﬁciency is hard to ﬁnd in the literature. Consequences for the efﬁciency is an important

issue since, as mentioned earlier, one of the main advantage with the DFIG system was

that losses of the power electronic equipment is reduced in comparison to a system where

the power electronic equipment has to handle the total power. Moreover, it is necessary to

compare the ride-through system with a system that utilizes a full-power converter, since

such a system can be considered to have excellent voltage sag ride-through performance (as

also will be shown in Chapter 7) [74].

1.2 Purpose and Contributions

The main purpose of this thesis is the analysis of the DFIG for a WT application both during

steady-state operation and transient operation. In order to analyze the DFIG during transient

operation both the control and the modeling of the system is of importance. Hence, the

control and the modeling are also important parts of the thesis. The main contribution of this

thesis is dynamic and steady-state analysis of the DFIG, with details being as follows.

• In Chapter 3 an investigation of the inﬂuence of the converter’s size on the energy

production for a DFIG system is analyzed. A smaller converter implies that the con-

verter losses will be lower. On the other hand it also implies a smaller variable-speed

range, which inﬂuences the aerodynamical efﬁciency. Further, in Chapter 3, a com-

parison of the energy efﬁciency of DFIG system to other electrical systems is pre-

sented. The investigated systems are two ﬁxed-speed induction generator systems and

three variable-speed systems. The variable-speed systems are: a doubly-fed induc-

tion generator, an induction generator (with a full-power converter) and a direct-driven

permanent-magnet synchronous generator system. Important electrical and mechani-

cal losses of the systems are included in the study. In order to make the comparison as

fair as possible the base assumption used in this work is that the maximum (average)

shaft torque of the wind turbine systems used should be the same. Finally, two different

methods of reducing the magnetizing losses of the DFIG system are compared.

• In Chapter 4 a general rotor current control law is derived for the DFIG system. Terms

are introduced in order to allow the possibility to include feed-forward compensation

of the back EMF and/or “active resistance.” “Active resistance” has been used for the

squirrel-cage induction machines to damp disturbances, such as varying back EMF

[18, 41]. The main contribution of Chapter 5 is an evaluation of different rotor cur-

rent control laws with respect to eliminating the inﬂuence of the back EMF. Stability

analysis of the system is performed for different combinations of the terms introduced

in the current control law, in both the stator-ﬂux-oriented and the grid-ﬂux-oriented

reference frames, for both correctly and erroneously known parameters.

• In Chapter 6, the grid-fault response of a DFIG wind turbine system is studied. Sim-

ulations are veriﬁed with experimental results. Moreover, another objective is also to

study how a reduced-order (second-order) model manages to predict the response of

the DFIG system.

4

• The contribution of Chapter 7 is to analyze, dynamically and in the steady state, two

different voltage sag ride-through systems for the DFIG. Moreover, these two methods

are also compared to a system that utilizes a full-power converter. The reason for

comparing these two systems with a system that utilizes a full-power converter is that

the latter system is capable of voltage sag ride-through.

• Finally, in Chapter 8, a rotor resistance control law for a stall-controlled wind tur-

bine is derived and analyzed. The objective of the control law is to minimize torque

ﬂuctuations and ﬂicker.

1.3 List of Publications

Some of the results presented in this thesis have been published in the following publications.

1. A. Petersson and S. Lundberg, “Energy efﬁciency comparison of electrical systems for

wind turbines,” in Proc. IEEE Nordic Workshop on Power and Industrial Electronics

(NORpie/2002), Stockholm, Sweden, Aug. 12–14, 2002.

The efﬁciency of some different electrical systems for wind turbines are compared.

This paper is an early version of the material presented in Chapter 3.

2. T. Thiringer, A. Petersson, and T. Petr˚ u, “Grid Disturbance Response of Wind Turbines

Equipped with Induction Generator and Doubly-Fed Induction Generator,” in Proc.

IEEE Power Engineering Society General Meeting, vol. 3, Toronto, Canada, July 13–

17, 2003, pp. 1542–1547.

The grid disturbance response to ﬁxed-speed wind turbines and wind turbines with

DFIG were presented.

3. A. Petersson, S. Lundberg, and T. Thiringer, “A DFIG Wind-Turbine Ride-Through

System Inﬂuence on the Energy Production,” in Proc. Nordic Wind Power Conference,

G¨ oteborg, Sweden, Mar. 1–2, 2004.

In this paper a voltage sag ride-through system for a DFIG WT based on increased

rating of the valves of the power electronic converter was investigated. This paper

presents one of the voltage sag ride-through system for a DFIG wind turbine that is

compared in Chapter 7.

The organizing committee of the conference recommended submission of this paper

to Wind Energy. The paper has been accepted for publication.

4. A. Petersson, T. Thiringer, and L. Harnefors, “Flicker Reduction of Stall-Controlled

Wind Turbines using Variable Rotor Resistances,” in Proc. Nordic Wind Power Con-

ference, G¨ oteborg, Sweden, Mar. 1–2, 2004.

In this paper a rotor resistance control law is derived for a stall-controlled wind turbine.

The objective of the control law is to minimize the ﬂicker (or voltage ﬂuctuations) con-

tribution. This study is presented in Chapter 8.

5

5. T. Thiringer and A. Petersson, “Grid Integration of Wind Turbines,” Przeglad Elek-

trotechniczny, no. 5, pp. 470–475, 2004.

This paper gives an overview of the three most common wind turbine systems, their

power quality impact, and its the response to grid disturbances.

6. R. Ottersten, A. Petersson, and K. Pietil¨ ainen, “Voltage Sag Response of PWM Rec-

tiﬁers for Variable-Speed Wind Turbines,” in Proc. IEEE Nordic Workshop on Power

and Industrial Electronics (NORpie/2004), Trondheim, Norway, June 14–16, 2004.

The voltage sag response of a PWMrectiﬁer for wind turbines that utilizes a full-power

converter were studied. This paper serves as a basis for the comparison of ride-through

systems of wind turbines in Chapter 7.

The organizing committee of the conference recommended submission of this paper

to EPE Journal. The paper has been accepted for publication.

7. A. Petersson, L. Harnefors, and T. Thiringer, “Comparison Between Stator-Flux and

Grid-Flux Oriented Rotor-Current Control of Doubly-Fed Induction Generators,” in

Proc. IEEE Power Electronics Specialists Conference (PESC’04), vol. 1, Aachen,

Germany, June 20–25, 2004, pp. 482–486.

The comparison between grid-ﬂux and stator-ﬂux-oriented current control of the DFIG

presented in Chapter 5 were studied in this paper.

8. A. Petersson, L. Harnefors, and T. Thiringer, “Evaluation of Current Control Meth-

ods for Wind Turbines Using Doubly-Fed Induction Machines,” IEEE Trans. Power

Electron., vol. 20, no. 1, pp. 227–235, Jan. 2005.

In this paper the analysis of the stator-ﬂux oriented current control of the DFIG pre-

sented in Chapter 5 was studied.

9. A. Petersson, T. Thiringer, L. Harnefors, and T. Petr˚ u, “Modeling and Experimental

Veriﬁcation of Grid Interaction of a DFIG Wind Turbine,” IEEE Trans. Energy Con-

version (accepted for publication)

Here a full-order model and a reduced-order model of the DFIG is compared during

grid disturbances. The models are experimentally veriﬁed with an 850 kW DFIG wind

turbine. These results are also presented in Chapter 6.

6

Chapter 2

Wind Energy Systems

2.1 Wind Energy Conversion

In this section, properties of the wind, which are of interest in this thesis, will be described.

First the wind distribution, i.e., the probability of a certain average wind speed, will be

presented. The wind distribution can be used to determine the expected value of certain

quantities, e.g. produced power. Then different methods to control the aerodynamic power

will be described. Finally, the aerodynamic conversion, i.e., the so-called C

p

(λ, β)-curve,

will be presented. The interested reader can ﬁnd more information in, for example, [11, 53].

2.1.1 Wind Distribution

The most commonly used probability density function to describe the wind speed is the

Weibull functions [53]. The Weibull distribution is described by the following probability

density function

f(w) =

k

c

w

c

k−1

e

−(w/c)

k

(2.1)

where k is a shape parameter, c is a scale parameter and w is the wind speed. Thus, the

average wind speed (or the expected wind speed), w, can be calculated from

w =

∞

0

wf(w)dw =

c

k

Γ

1

k

(2.2)

where Γ is Euler’s gamma function, i.e.,

Γ(z) =

∞

0

t

z−1

e

−t

dt. (2.3)

If the shape parameter equals 2, the Weibull distribution is known as the Rayleigh distribu-

tion. For the Rayleigh distribution the scale factor, c, given the average wind speed can be

found from (k=2, and Γ(

1

2

) =

√

π)

c =

2

√

π

w. (2.4)

In Fig. 2.1, the wind speed probability density function of the Rayleigh distribution is plotted.

The average wind speeds in the ﬁgure are 5.4 m/s, 6.8 m/s, and 8.2 m/s. A wind speed of

5.4 m/s correspond to a medium wind speed site in Sweden [100], while 8–9 m/s are wind

speeds available at sites located outside the Danish west coast [24].

7

0 5 10 15 20 25

0

0.05

0.1

0.15

Wind speed [m/s]

P

r

o

b

a

b

i

l

i

t

y

d

e

n

s

i

t

y

Fig. 2.1. Probability density of the Rayleigh distribution. The average wind speeds are 5.4 m/s

(solid), 6.8 m/s (dashed) and 8.2 m/s (dotted).

2.1.2 Aerodynamic Power Control

At high wind speeds it is necessary to limit the input power to the wind turbine, i.e., aero-

dynamic power control. There are three major ways of performing the aerodynamic power

control, i.e., by stall, pitch, or active stall control. Stall control implies that the blades are

designed to stall in high wind speeds and no pitch mechanism is thus required [11].

Pitch control is the most common method of controlling the aerodynamic power gen-

erated by a turbine rotor, for newer larger wind turbines. Almost all variable-speed wind

turbines use pitch control. Below rated wind speed the turbine should produce as much

power as possible, i.e., using a pitch angle that maximizes the energy capture. Above rated

wind speed the pitch angle is controlled in such a way that the aerodynamic power is at its

rated [11]. In order to limit the aerodynamic power, at high wind speeds, the pitch angle is

controlled to decrease the angle of attack, i.e., the angle between the chord line of the blade

and the relative wind direction [53]. It is also possible to increase the angle of attack towards

stall in order to limit the aerodynamic power. This method can be used to ﬁne-tune the power

level at high wind speeds for ﬁxed-speed wind turbines. This control method is known as

active stall or combi stall [11].

2.1.3 Aerodynamic Conversion

Some of the available power in the wind is converted by the rotor blades to mechanical power

acting on the rotor shaft of the WT. For steady-state calculations of the mechanical power

from a wind turbine, the so called C

p

(λ, β)-curve can be used. The mechanical power, P

mech

,

can be determined by [53]

P

mech

=

1

2

ρA

r

C

p

(λ, β)w

3

(2.5)

λ =

Ω

r

r

r

w

(2.6)

8

where C

p

is the power coefﬁcient, β is the pitch angle, λ is the tip speed ratio, w is the wind

speed, Ω

r

is the rotor speed (on the low-speed side of the gearbox), r

r

is the rotor-plane

radius, ρ is the air density and A

r

is the area swept by the rotor. In Fig. 2.2, an example of a

C

p

(λ, β) curve and the shaft power as a function of the wind speed for rated rotor speed, i.e.,

a ﬁxed-speed wind turbine, can be seen. In Fig. 2.2b) the solid line corresponds to a ﬁxed

0 5 10 15 20

0

0.1

0.2

0.3

0.4

0.5

5 10 15 20 25

0

20

40

60

80

100

120

Tip speed ratio

C

p

(

λ

)

a)

Wind speed [m/s]

P

o

w

e

r

[

%

]

b)

Fig. 2.2. a) The power coefﬁcient, C

p

, as a function of the tip speed ratio, λ. b) Mechanical power as

a function of wind speed at rated rotor speed (solid line is ﬁxed pitch angle, i.e., stall control

and dashed line is active stall).

pitch angle, β, while dashed line corresponds to a varying β (active stall).

Fig. 2.3 shows an example of how the mechanical power, derived from the C

p

(λ, β)

curve, and the rotor speed vary with the wind speed for a variable-speed wind turbine. The

rotor speed in the variable-speed area is controlled in order to keep the optimal tip speed

ratio, λ, i.e., C

p

is kept at maximum as long as the power or rotor speed is below its rated

values. As mentioned before, the pitch angle is at higher wind speeds controlled in order

to limit the input power to the wind turbine, when the turbine has reached the rated power.

As seen in Fig. 2.3b) the turbine in this example reaches the rated power, 1 p.u., at a wind

speed of approximately 13 m/s. Note that there is a possibility to optimize the radius of the

wind turbines rotor to suit sites with different average wind speeds. For example, if the rotor

radius, r

r

, is increased, the output power of the turbine is also increased, according to (2.5).

This implies that the nominal power will be reached for a lower wind speed, referred to

Fig. 2.3b). However, increasing the rotor radius implies that for higher wind speed the output

power must be even more limited, e.g., by pitch control, so that the nominal power of the

generator is not exceeded. Therefore, there is a trade-off between the rotor radius and the

nominal power of the generator. This choice is to a high extent dependent on the average

wind speed of the site.

2.2 Wind Turbine Systems

Wind turbines can operate with either ﬁxed speed (actually within a speed range about 1 %)

or variable speed. For ﬁxed-speed wind turbines, the generator (induction generator) is di-

9

5 10 15 20 25

10

15

20

25

5 10 15 20 25

0

20

40

60

80

100

R

o

t

o

r

s

p

e

e

d

[

r

p

m

]

a)

Wind speed [m/s] Wind speed [m/s]

P

o

w

e

r

[

%

]

b)

Fig. 2.3. Typical characteristic for a variable-speed wind turbine. a) Rotor speed as a function of

wind speed. b) Mechanical power as a function of wind speed.

rectly connected to the grid. Since the speed is almost ﬁxed to the grid frequency, and most

certainly not controllable, it is not possible to store the turbulence of the wind in form of

rotational energy. Therefore, for a ﬁxed-speed system the turbulence of the wind will result

in power variations, and thus affect the power quality of the grid [77]. For a variable-speed

wind turbine the generator is controlled by power electronic equipment, which makes it pos-

sible to control the rotor speed. In this way the power ﬂuctuations caused by wind variations

can be more or less absorbed by changing the rotor speed [82] and thus power variations

originating from the wind conversion and the drive train can be reduced. Hence, the power

quality impact caused by the wind turbine can be improved compared to a ﬁxed-speed tur-

bine [58].

The rotational speed of a wind turbine is fairly low and must therefore be adjusted to

the electrical frequency. This can be done in two ways: with a gearbox or with the number

of pole pairs of the generator. The number of pole pairs sets the mechanical speed of the

generator with respect to the electrical frequency and the gearbox adjusts the rotor speed of

the turbine to the mechanical speed of the generator.

In this section the following wind turbine systems will be presented:

1. Fixed-speed wind turbine with an induction generator.

2. Variable-speed wind turbine equipped with a cage-bar induction generator or synchro-

nous generator.

3. Variable-speed wind turbine equipped with multiple-pole synchronous generator or

multiple-pole permanent-magnet synchronous generator.

4. Variable-speed wind turbine equipped with a doubly-fed induction generator.

There are also other existing wind turbine concepts; a description of some of these systems

can be found in [36].

10

2.2.1 Fixed-Speed Wind Turbine

For the ﬁxed-speed wind turbine the induction generator is directly connected to the electrical

grid according to Fig. 2.4. The rotor speed of the ﬁxed-speed wind turbine is in principle

IG

Soft

starter

Gear-

box

Transformer

Capacitor bank

Fig. 2.4. Fixed-speed wind turbine with an induction generator.

determined by a gearbox and the pole-pair number of the generator. The ﬁxed-speed wind

turbine system has often two ﬁxed speeds. This is accomplished by using two generators

with different ratings and pole pairs, or it can be a generator with two windings having

different ratings and pole pairs. This leads to increased aerodynamic capture as well as

reduced magnetizing losses at low wind speeds. This system (one or two-speed) was the

“conventional” concept used by many Danish manufacturers in the 1980s and 1990s [36].

2.2.2 Variable-Speed Wind Turbine

The system presented in Fig. 2.5 consists of a wind turbine equipped with a converter con-

nected to the stator of the generator. The generator could either be a cage-bar induction

Power electronic

converter

G

Transformer

Gear-

box

=

= ≈

≈

Fig. 2.5. Variable-speed wind turbine with a synchronous/induction generator.

generator or a synchronous generator. The gearbox is designed so that maximum rotor speed

corresponds to rated speed of the generator. Synchronous generators or permanent-magnet

synchronous generators can be designed with multiple poles which implies that there is no

need for a gearbox, see Fig. 2.6. Since this “full-power” converter/generator system is com-

monly used for other applications, one advantage with this system is its well-developed and

robust control [7, 39, 61]. A synchronous generator with multiple poles as a wind turbine

generator is successfully manufactured by Enercon [25].

11

Power electronic

converter

SG

Transformer

=

= ≈

≈

Fig. 2.6. Variable-speed direct-driven (gear-less) wind turbine with a synchronous generator (SG).

2.2.3 Variable-Speed Wind Turbine with Doubly-Fed Induction Gener-

ator

This system, see Fig. 2.7, consists of a wind turbine with doubly-fed induction generator.

This means that the stator is directly connected to the grid while the rotor winding is con-

nected via slip rings to a converter. This system have recently become very popular as gen-

Power electronic

converter

DFIG

Transformer

Gear-

box

=

= ≈

≈

Fig. 2.7. Variable-speed wind turbine with a doubly-fed induction generator (DFIG).

erators for variable-speed wind turbines [36]. This is mainly due to the fact that the power

electronic converter only has to handle a fraction (20–30%) of the total power [36, 110].

Therefore, the losses in the power electronic converter can be reduced, compared to a system

where the converter has to handle the total power, see Chapter 3. In addition, the cost of the

converter becomes lower.

There exists a variant of the DFIG method that uses controllable external rotor resistances

(compare to slip power recovery). Some of the drawbacks of this method are that energy is

unnecessary dissipated in the external rotor resistances and that it is not possible to control

the reactive power.

Manufacturers, that produce wind turbines with the doubly-fed induction machine as

generator are, for example, DeWind, GE Wind Energy, Nordex, and Vestas [19, 29, 73, 105].

12

2.3 Doubly-Fed Induction Generator Systems for Wind Tur-

bines

For variable-speed systems with limited variable-speed range, e.g. ±30% of synchronous

speed, the DFIG can be an interesting solution [61]. As mentioned earlier the reason for

this is that power electronic converter only has to handle a fraction (20–30%) of the total

power [36, 110]. This means that the losses in the power electronic converter can be reduced

compared to a system where the converter has to handle the total power. In addition, the cost

of the converter becomes lower. The stator circuit of the DFIG is connected to the grid while

the rotor circuit is connected to a converter via slip rings, see Fig. 2.8. Amore detailed picture

Converter

Fig. 2.8. Principle of the doubly-fed induction generator.

of the DFIG system with a back-to-back converter can be seen in Fig. 2.9. The back-to-back

converter consists of two converters, i.e., machine-side converter and grid-side converter,

that are connected “back-to-back.” Between the two converters a dc-link capacitor is placed,

as energy storage, in order to keep the voltage variations (or ripple) in the dc-link voltage

small. With the machine-side converter it is possible to control the torque or the speed of

DFIG Grid

converter converter

Grid-side Machine-side

dc link

≈

≈ =

=

Fig. 2.9. DFIG system with a back-to-back converter.

the DFIG and also the power factor at the stator terminals, while the main objective for the

grid-side converter is to keep the dc-link voltage constant. The speed–torque characteristics

of the DFIG system can be seen in Fig. 2.10 [61]. As also seen in the ﬁgure, the DFIG can

operate both in motor and generator operation with a rotor-speed range of ±Δω

max

r

around

the synchronous speed, ω

1

.

13

Motor

Generator

T

ω

r

ω

1

2Δω

max

r

Fig. 2.10. Speed–torque characteristics of a DFIG.

A typical application, as mentioned earlier, for DFIG is wind turbines, since they operate

in a limited speed range of approximately ±30%. Other applications, besides wind turbines,

for the DFIG systems are, for example, ﬂywheel energy storage system [4], stand-alone

diesel systems [78], pumped storage power plants [6, 43], or rotating converters feeding a

railway grid from a constant frequency public grid [61].

2.3.1 Equivalent Circuit of the Doubly-Fed Induction Generator

The equivalent circuit of the doubly-fed induction generator, with inclusion of the magnetiz-

ing losses, can be seen in Fig. 2.11. This equivalent circuit is valid for one equivalent Y phase

and for steady-state calculations. In the case that the DFIG is Δ-connected the machine can

still be represented by this equivalent Y representation. In this section the jω-method is

adopted for calculations. Note, that if the rotor voltage, V

r

, in Fig. 2.11, is short circuited

+ +

− −

R

s

jω

1

L

sλ

jω

1

L

m

R

m

R

r

/s jω

1

L

rλ

I

s

I

r

V

s

I

R

m

V

r

s

Fig. 2.11. Equivalent circuit of the DFIG.

the equivalent circuit for the DFIG becomes the ordinary equivalent circuit for a cage-bar

induction machine. Applying Kirchhoff’s voltage law to the circuit in Fig. 2.11 yields [87]

V

s

= R

s

I

s

+ jω

1

L

sλ

I

s

+ jω

1

L

m

(I

s

+I

r

+I

R

m

) (2.7)

V

r

s

=

R

r

s

I

r

+ jω

1

L

rλ

I

r

+ jω

1

L

m

(I

s

+I

r

+I

R

m

) (2.8)

0 = R

m

I

R

m

+ jω

1

L

m

(I

s

+I

r

+I

R

m

) (2.9)

14

where the following notation is used.

V

s

stator voltage; R

s

stator resistance;

V

r

rotor voltage; R

r

rotor resistance;

I

s

stator current; R

m

magnetizing resistance;

I

r

rotor current; L

sλ

stator leakage inductance;

I

R

m

magnetizing resistance current; L

rλ

rotor leakage inductance;

ω

1

stator frequency; L

m

magnetizing inductance;

s slip.

The slip, s, equals

s =

ω

1

−ω

r

ω

1

=

ω

2

ω

1

(2.10)

where ω

r

is the rotor speed and ω

2

is the slip frequency. Moreover, if the air-gap ﬂux, stator

ﬂux and rotor ﬂux are deﬁned as

Ψ

m

= L

m

(I

s

+I

r

+I

R

m

) (2.11)

Ψ

s

= L

sλ

I

s

+ L

m

(I

s

+I

r

+I

R

m

) = L

sλ

I

s

+ Ψ

m

(2.12)

Ψ

r

= L

rλ

I

r

+ L

m

(I

s

+I

r

+I

R

m

) = L

rλ

I

r

+ Ψ

m

(2.13)

the equations describing the equivalent circuit, i.e., (2.7)–(2.9), can be rewritten as

V

s

= R

s

I

s

+ jω

1

Ψ

s

(2.14)

V

r

s

=

R

r

s

I

r

+ jω

1

Ψ

r

(2.15)

0 = R

m

I

R

m

+ jω

1

Ψ

m

. (2.16)

The resistive losses of the induction generator are

P

loss

= 3

R

s

[I

s

[

2

+ R

r

[I

r

[

2

+ R

m

[I

R

m

[

2

(2.17)

and it is possible to express the electro-mechanical torque, T

e

, as

T

e

= 3n

p

Im

Ψ

m

I

∗

r

= 3n

p

Im

Ψ

r

I

∗

r

(2.18)

where n

p

is the number of pole pairs. Table 2.1 shows some typical parameters of the induc-

tion machine in per unit (p.u.).

TABLE 2.1. TYPICAL PARAMETERS OF THE INDUCTION MACHINE IN P.U., [101].

Small Medium Large

Machine Machine Machine

4 kW 100 kW 800 kW

Stator and rotor resistance R

s

and R

r

0.04 0.01 0.01

Leakage inductance L

sλ

+ L

rλ

≈ L

σ

0.2 0.3 0.3

Magnetizing inductance L

m

≈ L

M

2.0 3.5 4.0

15

2.3.2 Power Flow

In order to investigate the power ﬂow of the DFIG system the apparent power that is fed to

the DFIG via the stator and rotor circuit has to be determined. The stator apparent power S

s

and rotor apparent power S

r

can be found as

S

s

= 3V

s

I

∗

s

= 3R

s

[I

s

[

2

+ j3ω

1

L

sλ

[I

s

[

2

+ j3ω

1

Ψ

m

I

∗

s

(2.19)

S

r

= 3V

r

I

∗

r

= 3R

r

[I

r

[

2

+ j3ω

1

sL

rλ

[I

r

[

2

+ j3ω

1

sΨ

m

I

∗

r

(2.20)

which can be rewritten, using the expressions in the previous section, as

S

s

= 3R

s

[I

s

[

2

+ j3ω

1

L

sλ

[I

s

[

2

+ j3ω

1

[Ψ

m

[

2

L

m

+ 3R

m

[I

R

m

[

2

−j3ω

1

Ψ

m

I

∗

r

(2.21)

S

r

= 3R

r

[I

r

[

2

+ j3ω

1

sL

rλ

[I

r

[

2

+ j3ω

1

sΨ

m

I

∗

r

. (2.22)

Now the stator and rotor power can be determined as

P

s

= Re [S

s

] = 3R

s

[I

s

[

2

+ 3R

m

[I

R

m

[

2

+ 3ω

1

Im[Ψ

m

I

∗

r

] ≈ 3ω

1

Im[Ψ

m

I

∗

r

] (2.23)

P

r

= Re [S

r

] = 3R

r

[I

r

[

2

−3ω

1

sIm[Ψ

m

I

∗

r

] ≈ −3ω

1

sIm[Ψ

m

I

∗

r

] (2.24)

where the approximations are because the resistive losses and the magnetizing losses have

been neglected. From the above equations the mechanical power produced by the DFIG can

be determined as the sum of the stator and rotor power as

P

mech

= 3ω

1

Im[Ψ

m

I

∗

r

] −3ω

1

sIm[Ψ

m

I

∗

r

] = 3ω

r

Im[Ψ

m

I

∗

r

] . (2.25)

Then, by dividing P

mech

with mechanical rotor speed, ω

m

= ω

r

/n

p

, the produced electro-

mechanical torque, as given in (2.18), can be found. Moreover, this means that P

s

≈

P

mech

/(1 − s) and P

r

≈ −sP

mech

/(1 − s). In Fig. 2.12 the power ﬂow of a “lossless”

DFIG system can be seen. In the ﬁgure it can be seen how the mechanical power divides

P

mech

P

mech

P

mech

/(1 −s)

sP

mech

/(1 −s)

DFIG

Converter

Grid

Fig. 2.12. Power ﬂow of a “lossless” DFIG system.

between the stator and rotor circuits and that it is dependent on the slip. Moreover, the rotor

power is approximately minus the stator power times the slip: P

r

≈ −sP

s

. Therefore, as

mentioned earlier, the rotor converter can be rated as a fraction of the rated power of the

DFIG if the maximum slip is low.

An example of how the stator and rotor powers depend on the slip is shown in Table 2.2.

It can be seen in the table that the power through the converter, given the mechanical power,

16

TABLE 2.2. EXAMPLE OF THE POWER FLOW FOR DIFFERENT SLIPS OF THE DFIG SYSTEM.

slip, s, [%] rotor speed, ω

r

, [p.u.] rotor power, P

r

stator power, P

s

0.3 0.7 −0.43 P

mech.

1.43 P

mech.

0 1.0 0 P

mech.

−0.3 1.3 0.23 P

mech.

0.77 P

mech.

is higher for positive slips (ω

r

< ω

1

). This is due to the factor 1/(1 − s) in the expressions

for the rotor power. However, for a wind turbine, the case is not as shown in Table 2.2. For a

wind turbine, in general, at low mechanical power the slip is positive and for high mechanical

power the slip is negative, as seen in Fig. 2.13. The ﬁgure is actually the same as Fig. 2.3, but

the stator and rotor power of the DFIG system is also shown and instead of the rotor speed

the slip is shown. In the ﬁgure it is assumed that the gearbox ratio is set in such a way that

5 10 15 20 25

−50

−25

0

25

50

5 10 15 20 25

0

20

40

60

80

100

S

l

i

p

[

%

]

a)

Wind speed [m/s] Wind speed [m/s]

P

o

w

e

r

[

%

]

b)

Fig. 2.13. Typical characteristic for a variable speed DFIG wind turbine. a) Slip as a function of wind

speed. b) Mechanical power (dotted), rotor power (solid) and stator power (dashed) as a

function of wind speed.

the average value of the rotor-speed range corresponds to synchronous speed of the DFIG.

Moreover, for the wind turbine in Fig. 2.13 the stator power is at maximum only 0.7 times

the rated power.

2.3.3 Stator-to-Rotor Turns Ratio

Since the losses in the power electronic converter depend on the current through the valves, it

is important to have a stator-to-rotor turns ratio of the DFIG that minimizes the rotor current

without exceeding the maximum available rotor voltage. In Fig. 2.14 a transformer is placed

between the rotor circuit and the converter. The transformer is to highlight and indicate the

stator-to-rotor turns ratio, but it does not exist in reality.

For example, if the stator-to-rotor turns ratio, n

s

/n

r

, is 0.4, the rotor current is ap-

proximately 0.4 times smaller than the stator current, if the magnetizing current is ne-

glected. Moreover, if the slip s of the DFIG is 30%, the rotor voltage will approximately be

V

rotor

R

= s/(n

s

/n

r

)V

s

= 0.3/0.4V

s

= 0.75V

s

, i.e., 75% of the stator voltage, which leaves

room for a dynamic control reserve. Note that V

rotor

R

= (n

r

/n

s

)V

R

is the actual (physical)

17

Converter

DFIG

n

s

/n

r

Fig. 2.14. Stator-to-rotor turns ratio indicated with a “virtual” transformer.

rotor voltage, while V

R

is rotor voltage referred to the stator circuit. In this thesis, all rotor

variables and parameters are referred to the stator circuit if not otherwise stated.

2.3.4 Lowering Magnetizing Losses

In an ordinary induction machine drive the stator is fed by a converter, which means that it is

possible to reduce the losses in the machine by using an appropriate ﬂux level. At low loads

it is possible to reduce the ﬂux level, which means that the magnetizing losses are lowered,

leading to a better efﬁciency. However, in the DFIG system the stator is connected to the

grid, and accordingly the ﬂux level is closely linked to the stator voltage. Still, for the DFIG

system there are, at least, two methods to lower the magnetizing losses of the DFIG. This

can be done by:

1. short-circuiting the stator of the induction generator at low wind speeds, and trans-

mitting all the turbine power through the converter. This set-up is referred to as the

short-circuited DFIG.

2. having the stator Δ-connected at high wind speeds and Y-connected at low wind

speeds; referred to as the Y-Δ-connected DFIG.

The inﬂuence that these two methods have on the overall efﬁciency of a DFIG system will

be further analyzed in Chapter 3. A brief description of these two systems follows:

“Short-Circuited DFIG”

Fig. 2.15 shows a diagram of the “short-circuited DFIG.” In the ﬁgure two switches can be

seen. Switch S2 is used to disconnect the turbine from the grid and switch S1 is then used

to short-circuit the stator of the DFIG. Now the turbine is operated as a cage-bar induction

machine, except that the converter is connected to the rotor circuit instead of the stator circuit.

This means, that in this operating condition, the DFIG can be controlled in a similar way as

an ordinary cage-bar induction generator. For instance, at low wind speeds the ﬂux level in

the generator can be lowered.

Y-Δ-connected DFIG

Fig. 2.16 presents a set-up of the Y-Δ-connected DFIG. As shown in the ﬁgure, a device

for changing between Y and Δ connection has been inserted in the stator circuit. Before a

18

Power electronic

converter

DFIG

Transformer

S1

S2

=

= ≈

≈

Fig. 2.15. Principle of the “short-circuited DFIG.”

Power electronic

converter

DFIG

Transformer

Y/Δ

S1

=

= ≈

≈

Fig. 2.16. Principle of the Y-Δ-connected DFIG.

change from Y to Δ connection (or vice versa) the power of the turbine is reduced to zero and

the switch S1 disconnects the stator circuit from the grid. Then the stator circuit is connected

in Δ (or vice versa) and the turbine is synchronized to the grid.

2.3.5 Other Types of Doubly-Fed Machines

In this section a short presentation of other kinds of doubly-fed machines is made: a cascaded

doubly-fed induction machine, a single-frame cascaded doubly-fed induction machine, a

brushless doubly-fed induction machine, and a doubly-fed reluctance machine.

Cascaded Doubly-Fed Induction Machine

The cascaded doubly-fed induction machine consists of two doubly-fed induction machines

with wound rotors that are connected mechanically through the rotor and electrically through

the rotor circuits. See Fig. 2.17 for a principle diagram. The stator circuit of one of the ma-

chines is directly connected to the grid while the other machine’s stator is connected via a

19

Converter

Fig. 2.17. Principle of cascaded doubly-fed induction machine.

converter to the grid. Since the rotor voltages of both machines are equal, it is possible to

control the induction machine that is directly connected to the grid with the other induction

machine.

It is possible to achieve decoupled control of active and reactive power of the cascaded

doubly-fed induction machine in a manner similar to the doubly-fed induction machine [47].

It is doubtful whether it is practical to combine two individual machines to form a cas-

caded doubly-fed induction machine, even though it is the basic conﬁguration of doubly-fed

induction machine arrangement. Due to a large amount of windings, the losses are expected

to be higher than for a standard doubly-fed induction machine of a comparable rating [48].

Single-Frame Cascaded Doubly-Fed Induction Machine

The single-frame cascaded doubly-fed induction machine is a cascaded doubly-fed induction

machine, but with the two induction machines in one common frame. Although this machine

is mechanically more robust than the cascaded doubly-fed induction machine, it suffers from

comparatively low efﬁciency [48].

Brushless Doubly-Fed Induction Machine

This is an induction machine with two stator windings in the same slot. That is, one winding

for the power and one winding for the control. See Fig. 2.18 for a principle sketch. To

avoid a direct transformer coupling between the two-stator windings, they can not have the

same number of pole pairs. Furthermore, to avoid unbalanced magnetic pull on the rotor the

difference between the pole pairs must be greater than one [106]. The number of poles in

the rotor must equal the sum of the number of poles in the two stator windings [106]. For

further information and more details, see [106, 108, 111].

Doubly-Fed Reluctance Machine

The stator of the doubly-fed reluctance machine is identical to the brushless doubly-fed in-

duction machine, while the rotor is based on the principle of reluctance. An equivalent circuit

with constant parameters can be obtained for the doubly-fed reluctance machine, in spite the

20

Converter

Fig. 2.18. Principle of the brushless doubly-fed induction machine.

fact that the machine is characterized by a pulsating air-gap ﬂux. It has almost the same

equivalent circuit as the standard doubly-fed induction generator [109].

21

22

Chapter 3

Energy Efﬁciency of Wind Turbines

The purpose of this chapter is to investigate the energy efﬁciency of the DFIG system and to

relate this study to other types of WTs with various electrical systems. This study focuses on

1) reducing the magnetizing losses of the DFIG system, 2) inﬂuence of the converter’s size

on the energy production (i.e., smaller converter implies a smaller variable-speed range for

the DFIG system) and ﬁnally 3) comparison of the DFIG system to other electrical systems.

In order to make the comparison as fair as possible the base assumption used in this work

is that the maximum (average) shaft torque of the wind turbine systems used should be the

same. Moreover, the rated WT power used in this chapter is 2 MW.

3.1 Determination of Power Losses

Steady-state calculations are carried out in this section in order to determine the power losses

of the DFIG system. Moreover, in order to compare the performance of the DFIG system,

the power losses of other systems with induction generators will also be presented. The

following systems are included in this study:

• FSIG 1 system — Fixed-speed system, as described in Section 2.2.1, with one genera-

tor.

• FSIG 2 system — Fixed-speed system, as described in Section 2.2.1, with two gener-

ators or a pole-pair changing mechanism.

• VSIG system — Variable-speed system with an induction generator and a full-power

converter, as described in Section 2.2.2.

• DFIG system — Variable-speed system with a DFIG, as described in Section 2.2.3.

The following losses are taken into account: aerodynamic losses, gearbox losses, gener-

ator losses and converter losses.

3.1.1 Aerodynamic Losses

Fig. 3.1 shows the turbine power as a function of wind speed both for the ﬁxed-speed and

variable-speed systems. In the ﬁgure it is seen that the ﬁxed-speed system with only one

generator has a lower input power at low wind speeds. The other systems produce almost

23

5 10 15 20 25

0

20

40

60

80

100

T

u

r

b

i

n

e

p

o

w

e

r

[

%

]

Wind speed [m/s]

Fig. 3.1. Turbine power. The power is given in percent of maximum shaft power. The solid line

corresponds to the variable-speed systems (VSIG and DFIG) and the two-speed system

(FSIG 2). The dotted line corresponds to a ﬁxed-speed system (FSIG 1).

identical results. In order to calculate the power in Fig. 3.1, a so-called C

p

(λ, β)-curve, as

described in Section 2.1.3, derived using blade-element theory has been used used.

In order to avoid making the results dependent on the torque, speed and pitch control

strategy, that vary from turbine to turbine, and anyway the settings used by the manufacturers

are not in detail known by the authors, only the average wind speed is used in the calculations,

i.e., the inﬂuence of the turbulence is ignored. The interested reader can ﬁnd information of

the inﬂuence of the turbulence on the energy production in [69].

3.1.2 Gearbox Losses

One way to estimate the gearbox losses, P

loss,GB

, is, [33],

P

loss,GB

= ηP

lowspeed

+ ξP

nom

Ω

r

Ω

r,nom

(3.1)

where η is the gear-mesh losses constant and ξ is a friction constant. According to [34], for

a 2-MW gearbox, the constants η = 0.02 and ξ = 0.005 are reasonable. In Fig. 3.2 the

gearbox losses are shown for the investigated systems.

3.1.3 Induction Generator Losses

In order to calculate the losses of the generator, the equivalent circuit of the induction gener-

ator, with inclusion of magnetizing losses, has been used, see Section 2.3.1.

For the DFIG system, the voltage drop across the slip rings has been neglected. More-

over, the stator-to-rotor turns ratio for the DFIG is adjusted so that maximum rotor voltage

is 75% of the rated grid voltage. This is done in order to have safety margin, i.e., a dynamic

reserve to handle, for instance, a wind gust. Observe that instead of using a varying turns

ratio, the same effect can also be obtained by using different rated voltages on the rotor and

stator [81].

24

4 6 8 10 12 14 16

0

0.5

1

1.5

2

2.5

3

G

e

a

r

b

o

x

l

o

s

s

e

s

[

%

]

Wind speed [m/s]

FSIG 2

FSIG 1

VSIG

Fig. 3.2. Gearbox losses. The losses are given in percent of maximum shaft power. The solid line

corresponds to the variable-speed systems (VSIG and DFIG). The dotted lines correspond

to ﬁxed-speed systems, i.e., FSIG 1 and 2 (both one-speed and two-speed generators).

In Fig. 3.3 the induction generator losses of the DFIG system are shown. The reason that

the generator losses are larger for high wind speeds for the VSIG system compared to the

DFIG system is that the gearbox ratio is different between the two systems. This implies

that the shaft torque of the generators will be different for the two systems, given the same

input power. It can also be noted that the losses of the DFIG are higher than those of the

5 10 15 20 25

0

0.5

1

1.5

2

2.5

3

G

e

n

e

r

a

t

o

r

l

o

s

s

e

s

[

%

]

Wind speed [m/s]

DFIG

VSIG

FSIG 1 & 2

Fig. 3.3. Induction generator losses. The losses are given in percent of maximum shaft power. DFIG

is solid, dashed is the variable-speed system (VSIG) and dotted are the ﬁxed-speed systems

(FSIG 1 and 2).

VSIG for low wind speeds. The reason for this is that the ﬂux level of the VSIG system has

been optimized from an efﬁciency point of view while for the DFIG system the ﬂux level is

almost ﬁxed to the stator voltage. This means that for the VSIG system a lower ﬂux level is

25

used for low wind speeds, i.e., the magnetizing losses are reduced.

For the IGs used in this chapter operated at 690 V 50 Hz and with a rated current of 1900

A and 390 A, respectively, the following parameters are used:

2-MW power: See Appendix B.1.

0.4-MW power: R

s

= 0.04 p.u., R

r

= 0.01 p.u., R

m

= 192 p.u., L

sλ

= 0.12 p.u.,

L

rλ

= 0.04 p.u., L

m

= 3.7 p.u. and n

p

= 3.

3.1.4 Converter Losses

In order to be able to feed the IG with a variable voltage and frequency source, the IG can

be connected to a pulse-width modulated (PWM) converter. In Fig. 3.4, an equivalent circuit

of the converter is drawn, where each transistor, T1 to T6, is equipped with a reverse diode.

A PWM circuit switches the transistors to on and off states. The duty cycle of the transistor

and the diode determines whether the transistor or a diode is conducting in a transistor leg

(e.g., T1 and T4).

T1 T2 T3

T4 T5 T6

V

CE0

r

CE

V

T0

r

T

⇔

⇔

Fig. 3.4. Converter scheme.

The losses of the converter can be divided into switching losses and conducting losses.

The switching losses of the transistors are the turn-on and turn-off losses. For the diode the

switching losses mainly consist of turn-off losses [103], i.e., reverse-recovery energy. The

turn-on and turn-off losses for the transistor and the reverse-recovery energy loss for a diode

can be found from data sheets. The conducting losses arise from the current through the

transistors and diodes. The transistor and the diode can be modeled as constant voltage drops,

V

CE0

and V

T0

, and a resistance in series, r

CE

and r

T

, see Fig. 3.4. Simpliﬁed expressions of

the transistor’s and diode’s conducting losses, for a transistor leg, are (with a third harmonic

voltage injection) [2]

P

c,T

=

V

CE0

I

rms

√

2

π

+

I

rms

V

CE0

m

i

cos(φ)

√

6

+

r

CE

I

2

rms

2

+

r

CE

I

2

rms

m

i

√

3 cos(φ)6π

−

4r

CE

I

2

rms

m

i

cos(φ)

45π

√

3

(3.2)

P

c,D

=

V

T0

I

rms

√

2

π

−

I

rms

V

T0

m

i

cos(φ)

√

6

+

r

T

I

2

rms

2

−

r

T

I

2

rms

m

i

√

3 cos(φ)6π

+

4r

T

I

2

rms

m

i

cos(φ)

45π

√

3

(3.3)

26

where I

rms

is the root mean square (RMS) value of the (sinusoidal) current to the grid or the

generator, m

i

is the modulation index, and φ is the phase shift between the voltage and the

current.

Since, for the values in this chapter, which are based on [89, 90, 91, 92] (see Table 3.1

for actual values), r

IGBT

= r

CE

≈ r

T

and V

IGBT

= V

CEO

≈ V

TO

. Hence, it is possible

to reduce the loss model of the transistor and the diode to the same model. The conduction

TABLE 3.1. CONVERTER CHARACTERISTIC DATA (IGBT AND INVERSE DIODE).

Nominal current I

C,nom

500 A 1200 A 1800 A 2400 A

Operating dc-link voltage V

CC

1200 V 1200 V 1200 V 1200 V

V

CEO

1.0 V 1.0 V 1.0 V 1.0 V

Lead resistance (IGBT) r

CE

3 mΩ 1.5 mΩ 1 mΩ 0.8 mΩ

Turn-on and turn-off

E

on

+ E

oﬀ

288 mJ 575 mJ 863 mJ 1150 mJ

energy (IGBT)

V

TO

1.1 V 1.1 V 1.1 V 1.1 V

Lead resistance (diode) r

T

2.6 mΩ 1.5 mΩ 1.0 mΩ 0.8 mΩ

Reverse recovery

E

rr

43 mJ 86 mJ 128 mJ 171 mJ

energy (diode)

losses can, with the above-mentioned approximation, be written as

P

c

= P

c,T

+ P

c,D

= V

IGBT

2

√

2

π

I

rms

+ r

IGBT

I

2

rms

. (3.4)

The switching losses of the transistor can be considered to be proportional to the current, for

a given dc-link voltage, as is assumed here [2]. This implies that the switching losses from

the transistor and the inverse diode can be expressed as

P

s,T

= (E

on

+ E

oﬀ

)

2

√

2

π

I

rms

I

C,nom

f

sw

≈ V

sw,T

2

√

2

π

I

rms

(3.5)

P

s,D

= E

rr

2

√

2

π

I

rms

I

C,nom

f

sw

≈ V

sw,D

2

√

2

π

I

rms

(3.6)

where E

on

and E

oﬀ

are the turn-on and turn-off energy losses, respectively, for the transistor,

E

rr

is the reverse recovery energy for the diode and I

C,nom

is the nominal current through the

transistor. In the equations above, two voltage drops, V

sw,T

and V

sw,D

, have been introduced.

This is possible since the ratios (E

on

+E

oﬀ

)/I

C,nom

and E

rr

/I

C,nom

are practically constant

for all the valves in Table 3.1. This means that for a given dc-link voltage and switching

frequency (which both are assumed in this thesis), the switching losses of the IGBT and

diode can be modeled as a constant voltage drop that is independent of the current rating of

the valves. The switching frequency used in this thesis is 5 kHz. Moreover, since the products

r

CE

I

C,nom

and r

T

I

C,nom

also are practically constant and equal to each other, it is possible

to determine a resistance, r

IGBT,1 A

, that is valid for a nominal current I

C,nom

= 1 A. Then,

the resistance of a speciﬁc valve can be determined from r

IGBT

= r

IGBT,1 A

/I

C,nom

, where

I

C,nom

is the nominal current of the valve. In this thesis, I

C,nom

is chosen as I

C,nom

= 2I

max

rms

where I

max

rms

is the maximum RMS value of the current in the valve. By performing the above

simpliﬁcation the model of the IGBT and valve can be scaled to an arbitrary rating. Using

27

the values given in Table 3.1 it is possible to determine the voltage drops V

sw,T

= 2.5 V and

V

sw,D

= 0.38 V, assuming a switching frequency of 5 kHz and the resistance r

IGBT,1 A

=

1.76 Ω. When determining V

sw,T

, V

sw,D

, and r

IGBT,1 A

the average values of all of the valves

in Table 3.1 has been used. Now, the total losses fromthe three transistor legs of the converter

become

P

loss

= 3(P

c

+ P

s,T

+ P

s,D

)

= 3

(V

IGBT

+ V

sw,T

+ V

sw,D

)

2

√

2

π

I

rms

+ r

IGBT

I

2

rms

.

(3.7)

The back-to-back converter can be seen as two converters which are connected together: the

machine-side converter (MSC) and the grid-side converter (GSC). For the MSC, the current

through the valves, I

rms

, is the stator current for the VSIG system or the rotor current for the

DFIG system. One way of calculating I

rms

for the GSC is by using the active current that is

produced by the machine, adjusted with the ratio between machine-side voltage and the grid

voltage. The reactive current can be freely chosen. Thus it is now possible to calculate the

losses of the back-to-back converter as

P

loss,converter

= P

loss,GSC

+ P

loss,MSC

. (3.8)

The total converter losses are now presented as a function of wind speed in Fig. 3.5.

From the ﬁgure it can, as expected, be noted that the converter losses in the DFIG system are

5 10 15 20 25

0

0.5

1

1.5

2

2.5

3

C

o

n

v

e

r

t

e

r

l

o

s

s

e

s

[

%

]

Wind speed [m/s]

DFIG

VSIG

Fig. 3.5. Converter losses. The losses are given in percent of maximum shaft power. DFIG is solid

and VSIG is dashed.

much lower compared to the full-power converter system.

3.1.5 Total Losses

The total losses (aerodynamic, generator, converter, gearbox) are presented in Fig. 3.6. From

the ﬁgure it can be noted that the DFIG system and the two-speed system (FSIG 2) has

roughly the same total losses while the full-power converter system has higher total losses.

28

5 10 15 20 25

0

1

2

3

4

5

6

7

8

T

o

t

a

l

l

o

s

s

e

s

[

%

]

Wind speed [m/s]

DFIG

VSIG

FSIG 1 & 2

Fig. 3.6. Total losses. The losses are given in percent of maximum shaft power. DFIG is solid, FSIG

1 and 2 is dotted and VSIG is dashed.

3.2 Energy Production of the DFIG System

In the previous section, the power loss as a function of transmitted power (or wind speed)

was determined. However, for the wind-turbine application, the most important quantity is

the energy delivered to the grid (electric energy capture). Accordingly, in this section the

results in the previous section have been used to determine the energy capture (or energy

efﬁciency) of the various systems.

In order to do this, the distribution of wind speeds must be known. As mentioned earlier

one commonly used probability density functions to describe the wind speed is the Rayleigh

function [53]. Given a probability density functions, f(w), the average (or expected) value

of the power, P(w), can be found as

P

avg

=

∞

0

P(w)f(w)dw (3.9)

where w is the wind speed.

3.2.1 Investigation of the Inﬂuence of the Converter’s Size on the En-

ergy Production

As was mentioned earlier, it is not possible to obtain a full speed range with the DFIG system

if the converter is smaller than the rated power of the turbine. This means that the smaller

the converter is, the more the WT will operate at a non-ideal tip-speed ratio, λ, for low

wind speeds. Fig. 3.7 illustrates the impact of having a smaller converter and thus a smaller

rotor-speed range, i.e., the aerodynamic losses become higher.

In Fig. 3.8 the converter losses are presented for different designs of the rotor-speed

range, i.e., a smaller rotor-speed range implies smaller ratings of the converter. It can be seen

in the ﬁgure that the converter losses are lower for smaller rotor-speed ranges (or smaller

converter ratings). Note, as mentioned earlier, that the stator-to-rotor turns ratio has to be

29

4 6 8 10

10

14

18

22

26

4 6 8 10

0

5

10

15

20

R

o

t

o

r

s

p

e

e

d

[

r

p

m

]

T

u

r

b

i

n

e

p

o

w

e

r

[

%

]

Wind speed [m/s] Wind speed [m/s]

Fig. 3.7. Rotor speed and the corresponding turbine power.

5 10 15 20 25

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

C

o

n

v

e

r

t

e

r

l

o

s

s

e

s

[

%

]

Average wind speed [m/s]

12-25 rpm

15-25 rpm

18-25 rpm

21-25 rpm

Fig. 3.8. Converter losses for some different rotor-speed ranges as a function of the wind speed.

designed according to desired variable-speed range in order to minimize the converter losses.

However, the most interesting information is the total energy efﬁciency. In Fig. 3.9 the energy

efﬁciency of the DFIG for different rotor-speed ranges (or converter sizes) can be seen. It can

be seen in the ﬁgure that the gain in energy increases with the rotor-speed range (converter

size), even though the converter losses of the DFIG system increase with the rotor-speed

range (converter size), as shown in Fig. 3.8. The increased aerodynamic capture has thus a

larger impact than the increased converter losses. If the rotor-speed range is set to 12–25, it

is possible to run at optimal tip-speed ratio in the whole variable-speed area. It can be seen in

the ﬁgure, as expected, that the rotor-speed range is of greater importance for a low average

wind-speed compared to a high average wind speed.

30

24−25 20−25 16−25 12−25

86

88

90

92

94

96

E

f

ﬁ

c

i

e

n

c

y

[

%

]

Rotor-speed range [rpm]

Fig. 3.9. Efﬁciency, for average wind speeds of 5.4 m/s (solid), 6.8 m/s (dashed) and 8.2 m/s (dotted),

of the DFIG system as a function of the rotor-speed range. Note that the aerodynamic

efﬁciency is also taken into account.

3.2.2 Reduction of Magnetizing Losses

As presented in Section 2.3.4 there are at least two ways of lowering the magnetizing losses,

i.e., this can be done by:

1. short-circuiting the stator of the induction generator at low wind speeds, and trans-

mitting all the turbine power through the converter. This set-up is referred to as the

short-circuited DFIG.

2. having the stator Δ-connected at high wind speeds and Y-connected at low wind

speeds; referred to as the Y-Δ-connected DFIG.

The break-even point of the total losses or the rated values of the equipment determines the

switch-over point for the doubly-fed generators, i.e., the Y-Δ coupling or the synchronization

of the stator voltage to the grid.

In Fig. 3.10 the energy gain using the two methods are presented. It can be seen in the

ﬁgure that the Y-Δ-connected DFIG system produces approximately 0.2 percentage units

more energy than the short-circuited DFIG system, at least for low average wind speeds.

Since the Y-Δ-connected DFIG system performs better than the short-circuited DFIG

system the Y-Δ-connected DFIG system will henceforth be referred to as the DFIG system,

and the other variants will not be subjected to any further studies.

3.3 Comparison to Other Wind Turbine Systems

The base assumption made here is that all wind turbine systems have the same average max-

imum shaft torque as well as the same mean upper rotor speed. In Fig. 3.11 the produced

grid power together with the various loss components for an average wind speed of 6 m/s are

presented for the various systems. The systems are the DFIG system, the full variable-speed

31

5 6 7 8 9 10

0

0.2

0.4

0.6

0.8

1

1.2

1.4

G

a

i

n

i

n

e

n

e

r

g

y

[

%

]

Average wind speed [m/s]

Fig. 3.10. Gain in energy production by lowering the magnetizing losses for a DFIG system as a

function of the average wind speed. Solid line is the Y-Δ-connected DFIG and dashed line

is short-circuited DFIG.

system (VSIG), one-speed system (FSIG 1), two-speed system (FSIG 2), and, a variable-

speed system equipped with a permanent magnet synchronous generator (PMSG). The aver-

age efﬁciency for the PMSG is taken from [34]. The converter losses of the PMSG system

are assumed equal to that of the VSIG system. It would also be possible to have the PMSG

connected to a diode rectiﬁer with series or shunt compensating capacitors, which may give

a possibility to reduce the converter losses [32]. However, a transistor rectiﬁer has the poten-

tial to utilize the generator best [32]. In the ﬁgure it can be seen that the one-speed system

FSIG 1 FSIG 2 VSIG PMSG DFIG

80

85

90

95

100

Gearbox losses

Generator losses

Converter losses

Grid power

A

v

e

r

a

g

e

p

o

w

e

r

[

%

]

Fig. 3.11. The produced average grid power and generator, converter and gearbox losses for an av-

erage wind speed of 6 m/s. 100% correspond to the input turbine power at optimal, with

respect to the rotor speed, aerodynamic efﬁciency.

32

(FSIG 1) has the disadvantage of poor aerodynamic efﬁciency. However, with the two-speed

system (FSIG 2) the aerodynamic efﬁciency is improved and close to the variable speed

systems (VSIG, PMSG and DFIG).

In Fig. 3.12 the produced energy of the different systems, for various average wind

speeds, are presented. In the ﬁgure, the DFIG is operated with a rotor-speed range set to

12–25 rpm.

5 6 7 8 9 10

84

86

88

90

92

94

96

PMSG

FSIG 2

FSIG 1

VSIG

DFIG

P

r

o

d

u

c

e

d

e

n

e

r

g

y

[

%

]

Average wind speed [m/s]

Fig. 3.12. Energy efﬁciency of the FSIG 1, FSIG 2, VSIG, PMSG and the DFIG system as a function

of the average wind speed.

Detailed information of the gearless electrically magnetized generator system was not

available. However, it is reasonable to assume that the losses in the diode rectiﬁer connected

to the stator, the boost converter on the dc-link, transistor converter towards the grid and the

magnetizing system of the generator are in the same range as the PMSG system. In [35], a

lower ﬁxed-speed IG WT efﬁciency was reported than in this study. The reason for this is

that the IG in this study has two generators and lower iron losses.

3.4 Discussion

In Fig. 3.12 it can be seen that the two-generator system (FSIG 2), the DFIG system, and the

PMSG system have almost the same efﬁciency. In [16] it was found that the DFIG system

produced 60% more energy compared to a ﬁxed-speed system. However, in this study the

produced energy of the systems was found to be similar. The difference between the result

here and in [16] is due to the different base assumptions used. Further, it was found in this

investigation that there is a possibility to gain a few percentage units (approximately 2%)

in energy using the DFIG system compared to the full variable-speed system. This can be

compared to a gain of 20% for the DFIG system compared to the variable-speed system

reported in [16]. The reason for the difference is again, that the base assumptions differ.

Reference [16] sets the rating of the stator windings equal while we choose the shaft power

and maximum speed instead.

33

The focus in this chapter is on the electrical energy efﬁciency of the DFIG-system in

relation to other systems. However, aerodynamics must be accounted for when ﬁxed-speed

and variable-speed turbines are compared. In order to reduce the number of uncertainties,

only the average wind speed has been used and the inﬂuence of turbulence has not been

treated. Reference [69] showed that a two-speed active stall turbine and a variable-speed

pitch turbine is fairly unaffected by the turbulence intensity for turbulence intensities up to

15%. For the more unusual turbulence intensities of 20–25% the variable-speed turbines

gained a couple of percentage units in energy production compared to the two-speed active

stall-regulated system.

Of great importance to point out is that when comparing the DFIG system to the full

variable system, the turbulence intensity, regardless of value plays an unimportant role since

the torque and speed control of the turbines are in principle the same. (The rotor-speed

range of the DFIG system is assumed to be almost the same as for the full-variable speed

system). Another problem when incorporating the effect of the turbulence intensity is that

the selection of torque, speed and pitch control inﬂuences the result. Also, among other

factors, the time delay between generator switchings for the ﬁxed-speed systems, start and

stop, Δ-Y-reconnections for the DFIG-systems must be known, in order to perform a detailed

energy capture calculation. So, in order not to include uncertainties that might not be the best

chosen, the ambition has instead been to make the comparison as clean as possible, using

only the facts that can be presented clearly and with best certainty.

3.5 Conclusion

In this chapter, it has been found that there is a possibility to gain a few percentage units

in energy efﬁciency for a doubly-fed induction generator system compared to a cage-bar

induction generator, controlled by a full-power converter. In comparison to a direct-driven

permanent-magnet synchronous generator, controlled by a converter or a two-speed genera-

tor system the difference in energy efﬁciency was found to be small.

Moreover, two methods to reduce the magnetizing losses (and thereby increase the gain in

energy) of the DFIG system, have been investigated. It was found that the method utilizing a

Y-Δswitch in the stator circuit had the largest gain in energy of the two investigated methods.

Finally, it was found that the converter losses of the DFIG can be reduced if the available

rotor-speed range is made smaller. However, the aerodynamic capture of the wind turbine is

reduced with a smaller rotor-speed range. This means that the increased aerodynamic capture

that can be achieved by a larger converter has, thus, a greater impact than the increased

converter losses.

Worth stressing is that the main reason for using a variable-speed turbine instead of a

ﬁxed-speed turbine is not the energy efﬁciency, instead it is the possibility of lowering the

mechanical stresses [53] and also improving the power quality [58].

34

Chapter 4

Control of Doubly-Fed Induction

Generator System

4.1 Introduction

In this section, different aspects of designing and implementing control systems for doubly-

fed induction generators (DFIGs) are treated.

4.1.1 Space Vectors

The idea behind space vectors is to describe the induction machine with two phases instead

of three. Space vectors were originally invented to describe the spatial ﬂux of an ac machine

[39]. A three-phase stator winding, which is supplied with three-phase currents, forms a

rotating ﬂux in the air gap. The same rotating ﬂux could also be formed with only two

phases, as seen in Fig. 4.1. This is the principle of space vectors. In order to determine the

⇐⇒

Im Im

Re Re

ψ ψ

Fig. 4.1: Principle of space vectors.

space vector, s

s

, of the three-phase quantities, s

a

, s

b

, and s

c

, the following transformation is

applied [39]

s

s

= s

α

+ js

β

=

2K

3

s

a

+as

b

+a

2

s

c

(4.1)

35

where K is the space-vector scaling constant and a = e

j2π/3

. Superscript “s” indicates that

the space vectors are referred to the reference frame of the stator of the induction machine.

The constant K can be chosen arbitrary, though if it is chosen as

K =

1

√

2

(4.2)

the space vector will be scaled according to the RMS value of the three-phase quantities.

This choice of K will be used throughout this thesis.

A general space vector can be expressed as

s

s

= se

j(θ

1

+φ)

(4.3)

where φ is a phase shift and θ

1

is the synchronous angle corresponding to the synchronous

frequency, ω

1

, as dω

1

/dt = θ

1

. It is possible to transform the vector to synchronous coordi-

nates (dq coordinates) as

s = s

d

+ js

q

= e

−jθ

1

s

s

= se

jφ

. (4.4)

The synchronous coordinate system is not indicated by a superscript. The synchronous co-

ordinate system has to be aligned with a quantity, normally the stator or rotor ﬂux of an

induction machine. However, it is also possible to align the synchronous coordinate system

with, for example, the grid voltage. Space vectors in synchronous coordinates will be dc

quantities in the steady state.

4.1.2 Power and Reactive Power in Terms of Space Vectors

The instantaneous power, P, in a three-phase system is given by

P = v

a

i

a

+ v

b

i

b

+ v

c

i

c

=

3

2K

2

Re [v

s

(i

s

)

∗

] =

3

2K

2

Re [vi

∗

] . (4.5)

The above-mentioned scaling, i.e. K = 1/

√

2, yields

P = 3Re [vi

∗

] . (4.6)

In (4.6) the instantaneous power is the real part of voltage times the complex conjugate of

the current, i.e., the same as active power in terms of phasors. It is also possible to deﬁne

a quantity the instantaneous reactive power, Q, as the corresponding imaginary part of the

above equation [5]:

Q = 3Im[vi

∗

] . (4.7)

4.1.3 Phase-Locked Loop (PLL)-Type Estimator

A PLL-type estimator can be used for estimation of the angle and frequency of a signal, e.g.,

the synchronous frequency, ω

1

, and its corresponding angle, θ

1

. The PLL-type estimator

used in this thesis is described by [37]

dˆ ω

1

dt

= γ

1

ε (4.8)

d

ˆ

θ

1

dt

= ˆ ω

1

+ γ

2

ε (4.9)

36

where ε = sin(θ

1

−

ˆ

θ

1

) and θ

1

−

ˆ

θ

1

is the error in the estimated angle. In the above equations

γ

1

and γ

2

are gain parameters. The notation “ˆ” indicates an estimated variable or parameter.

If the true frequency and position are given by dω

1

/dt = 0 and dθ

1

/dt = ω

1

, then it is

shown in [37] that the estimation error equations for ˜ ω

1

= ω

1

− ˆ ω

1

and

˜

θ

1

= θ

1

−

ˆ

θ

1

are

asymptotically stable if ¦γ

1

, γ

2

¦ > 0. This implies that ˆ ω

1

and

ˆ

θ

1

will converge to ω

1

and θ

1

respectively, asymptotically. If the difference θ

1

−

ˆ

θ

1

is small, it is possible to approximate

sin(θ

1

−

ˆ

θ

1

) ≈ θ

1

−

ˆ

θ

1

, and the following characteristic polynomial of the system described

by (4.8) and (4.9) can be found:

p

2

+ γ

2

p + γ

1

(4.10)

where p = d/dt. If the parameters are chosen as

γ

1

= ρ

2

γ

2

= 2ρ (4.11)

then ρ can be adjusted to the desired bandwidth of the PLL-type estimator.

Modiﬁed PLL-Type Estimator

If the PLL-type estimator should synchronize to a constant (or at least close to constant)

frequency, such as the grid frequency, it is possible to simplify the PLL-type estimator in

(4.8)–(4.9). This is done by neglecting (4.8); then the modiﬁed PLL-type estimator becomes

[76]

d

ˆ

θ

1

dt

= ω

1

= ω

g

+ ρε. (4.12)

In [76] it is shown that the modiﬁed PLL-type estimator rejects better voltage harmonics than

the PLL-type estimator in (4.8)–(4.9). For small bandwidths the rejection is twice as good.

4.1.4 Internal Model Control (IMC)

Due to the simplicity of IMC for designing controllers, this method will be used throughout

this thesis. IMC can, for instance, be used for designing current or speed control laws of any

ac machine [40, 44, 102]. The idea behind IMC is to augment the error between the process,

G(p) and a process model,

ˆ

G(p), by a transfer function C(p), see Fig. 4.2. Controller design

is then just a matter of choosing the “right” transfer function C(p). One common way is [31]

C(p) =

α

p + α

n

ˆ

G

−1

(p) (4.13)

where n is chosen so that C(p) become proper, i.e., the order of the denominator is equal to

or greater than that of the numerator. The closed-loop system will be

G

cl

(p) = G(p)

1 + C(p)[G(p) −

ˆ

G(p)]

−1

C(p) (4.14)

which simpliﬁes to

G

cl

(p) = G(p)C(p) =

α

p + α

n

(4.15)

37

¸

¸

G(p)

ˆ

G(p)

C(p)

i

ref

i

v

F(p)

+

+

−

−

Fig. 4.2. Principle of IMC.

when G(p) =

ˆ

G(p). The parameter α is a design parameter, which for n = 1, is set to the

desired bandwidth of the closed-loop system. The relationship between the bandwidth and

the rise time (10%–90%) is α = ln 9/t

rise

. The controller, F(p), (inside the dashed area in

Fig. 4.2) becomes

F(p) =

1 −C(p)

ˆ

G(p)

−1

C(p). (4.16)

For a ﬁrst-order system, n = 1 is sufﬁcient. The controller then typically becomes an ordi-

nary PI controller:

F(p) =

α

p

ˆ

G

−1

(p) = k

p

+

k

i

p

. (4.17)

4.1.5 “Active Damping”

For a ﬁrst-order system designed with IMC, the transfer function from the load disturbance

E to output signal i is given by

G

Ei

(p) =

G(p)

1 + G(p)F(p)

=

p

p + α

G(p) (4.18)

if all parameters are assumed to be known. If the dynamics of G(p) are fast, the load dis-

turbance rejection should be sufﬁcient. However, as the dynamics of the process, G(p), are

normally much slower than the dynamics of the closed-loop system, the disturbance rejection

is to a large extent determined by the process [76]. Therefore, addition of an inner feed-back

loop, see Fig. 4.3, can improve the disturbance rejection. Then, the transfer function in (4.18)

is changed to

G

Ei

(p) =

p

p + α

G(p)

1 + G(p)R

=

p

p + α

1

G

−1

(p) + R

. (4.19)

For a ﬁrst-order system it is possible to choose R so that the above transfer function can be

reduced to

G

Ei

(p) = K

p

(p + α)

2

(4.20)

38

¸ ¸ ¸

G(p)

R

F(p)

i

ref

i

v

v

E(p)

+

+ + +

− −

Fig. 4.3. Principle of “active damping.”

where K is a constant. This means that a load disturbance E is damped with the same

time constant as the control loop. This will be refereed to as “active damping” or “active

resistance.” “Active damping” has been used for the cage-bar induction machine to damp

disturbances, such as varying back EMF [18, 41].

For example, consider the following ﬁrst-order system (e.g., a dc machine)

L

di

dt

= v −Ri + e (4.21)

where i is the state (current), v is the input signal (applied voltage), and e is a load disturbance

(back emf). Then the “active damping” can be introduced by letting v = v

−R

a

i, where the

term R

a

i is the “active damping” term. Then, the system can be rewritten as

L

di

dt

= v

−(R + R

a

)i + e (4.22)

which has the following transfer function

G(p) =

i(p)

v

(p)

=

1

Lp + R + R

a

. (4.23)

By using IMC, the following PI controller can be found:

F(p) =

α

p

G

−1

(p) = αL + α

R + R

a

p

. (4.24)

Then, from (4.19), the transfer function from the load disturbance e to the output i can be

determined as

G

ei

(p) =

p

p + α

1

G

−1

(p) + R

a

. (4.25)

By choosing R

a

= αL −R, the transfer function is reduced to

G

ei

(p) =

p

L(p + α)

2

. (4.26)

This means that the disturbance is damped with the same time constant as the dynamics of

the control loop.

39

4.1.6 Saturation and Integration Anti-Windup

When designing control laws, the control signal cannot be arbitrary large due to design lim-

itations of the converter or the machine. Therefore, the control signal must be limited (satu-

rated). This causes the integral part of the PI controller to accumulate the control error during

the saturation, so called integrator wind-up. This might cause overshoots in the controlled

variable since the integration part of control law will keep the ideal control signal high even

when the controlled variable is getting closer to the reference value [39].

One method to avoid integration wind-up is to use the “back-calculation” method [39].

The idea behind the back-calculation method is to modify the reference value in case of

saturation, so that the ideal control signal, u, does not exceed the maximum value, i.e.,

[u[ = u

max

. The algorithm can be described as [39]

u = k

p

e + k

i

I (4.27)

u

sat

= sat(u) (4.28)

dI

dt

= e +

u

sat

−u

k

p

(4.29)

where e is the control error and I is the integral of the control error.

4.1.7 Discretization

Throughout the thesis, differential equations and control laws will be described in continuous

time. However, when implementing control laws in computers, they have to be discretized.

The forward Euler method will be used, i.e., a derivative is approximated as

dx

dt

≈

x(n + 1) −x(n)

T

sample

(4.30)

where n indicates the sample number, at time t = nT

sample

. For a continuous system given

as

˙ x(t) = Ax(t) + Bu(t) (4.31)

y(t) = Cx(t) (4.32)

the discrete equivalent using the forward Euler method becomes

x(n + 1) = (I + AT

sample

)x(n) + T

sample

Bu(n) (4.33)

y(n) = Cx(n). (4.34)

The forward Euler discretization can also be written as

p −→

q −1

T

sample

(4.35)

where q is the forward shift operator. Stability of a linear time-invariant continuous system

requires that the poles are in the left half plane. For a linear time-invariant discrete system

the corresponding stability region is inside the unit circle [88]. Mapping the unit circle

40

Im

Re

p plane

1/T

sample

Fig. 4.4: Region of stability.

onto the continuous p plane using (4.35) gives the region in the p plane where the poles

of the continuous system must be located in order to get a stable discretization [38]. Fig.

4.4 shows where the poles of continuous system must be located so that the forward Euler

discretization, in (4.35), becomes stable. As seen in the ﬁgure the poles must be inside a

circle with the radius of 1/T

sample

with the center point located at (−1/T

sample

, 0) in order

for the forward Euler discretization to be stable.

4.2 Mathematical Models of the DFIG System

In Fig. 4.5 an equivalent circuit of the DFIG system can be seen. As mentioned earlier, the

system consists of a DFIG and a back-to-back voltage source converter with a dc link. The

back-to-back converter consists of a grid-side converter (GSC) and a machine-side converter

(MSC). Moreover, a grid ﬁlter is placed in between the GSC and the grid, since both the

grid and the voltage source converter are voltage stiff and to reduce the harmonics caused

by the converter. For voltage source converters the grid ﬁlter used is mainly an L-ﬁlter or

an LCL-ﬁlter [62]. However, in this thesis the L-grid ﬁlter will be used, as shown in Fig.

4.5. More detailed description of the models of the components of the DFIG system will be

performed in the following sections. In addition, the variables and the parameters in Fig. 4.5

will also be explained.

4.2.1 Machine Model

Due to its simplicity for deriving control laws for the DFIG, the Γ representation of the IG

model will be used. Note, that from a dynamic point of view, the rotor and the stator leakage

inductance have the same effect. Therefore, it is possible to use a different representation of

the Park model in which the leakage inductance is placed in the rotor circuit, the so-called

Γ representation of the induction machine [94]. The name is due to the formation of a “Γ”

of the inductances; see Fig. 4.6. This model is described by the following space-vector

41

+ +

+ + +

− −

− − −

E

s

g

i

s

g

i

s

f

v

s

s

i

s

s

R

f

L

f

Grid ﬁlter

R

s

L

M

v

s

f

L

σ

≈

≈

=

=

DFIG

GSC

R

R

C

dc

jω

r

Ψ

s

R i

s

R

v

s

R

v

dc

dc-link

MSC

Fig. 4.5. Equivalent circuit of the DFIG system.

+

+

+

−

−

−

R

s

L

σ

L

M

R

R i

s

s

i

s

R

jω

r

Ψ

s

R

v

s

s

v

s

R

Fig. 4.6. Γ representation of the IG in stator coordinates. Superscript “s” indicates that the space

vectors are referred to the reference frame of the stator of the DFIG.

equations in stator coordinates [94]:

v

s

s

= R

s

i

s

s

+

dΨ

s

s

dt

(4.36)

v

s

R

= R

R

i

s

R

+

dΨ

s

R

dt

−jω

r

Ψ

s

R

(4.37)

where superscript s indicates stator coordinates. The model can also be described in syn-

chronous coordinates as

v

s

= R

s

i

s

+

dΨ

s

dt

+ jω

1

Ψ

s

(4.38)

v

R

= R

R

i

R

+

dΨ

R

dt

+ jω

2

Ψ

R

(4.39)

where the following notation is used:

42

v

s

stator voltage; Ψ

s

stator ﬂux;

v

R

rotor voltage; Ψ

R

rotor ﬂux;

i

s

stator current; R

s

stator resistance;

i

R

rotor current; R

R

rotor resistance;

ω

1

synchronous frequency; ω

2

slip frequency.

The stator ﬂux, rotor ﬂux, and electromechanical torque are given by

Ψ

s

= L

M

(i

s

+i

R

) (4.40)

Ψ

R

= (L

M

+ L

σ

)i

R

+ L

M

i

s

= Ψ

s

+ L

σ

i

R

(4.41)

T

e

= 3n

p

Im

Ψ

s

i

∗

R

(4.42)

where L

M

is the magnetizing inductance, L

σ

is the leakage inductance, and n

p

is the number

of pole pairs. Finally, the mechanical dynamics of the induction machine are described by

J

n

p

dω

r

dt

= T

e

−T

s

(4.43)

where J is the inertia and T

s

is the shaft torque. The quantities and parameters of the Γ

model relate to the Park model (or the T representation) as follows:

v

R

= γv

R

i

R

=

i

r

γ

Ψ

R

= γΨ

r

γ =

L

sλ

+ L

m

L

m

R

R

= γ

2

R

r

L

σ

= γL

sλ

+ γ

2

L

rλ

L

M

= γL

m

.

4.2.2 Grid-Filter Model

In Fig. 4.7 the equivalent circuit of the grid ﬁlter in stator coordinates can be seen. The ﬁlter

consists of an inductance L

f

and its resistance R

f

. Applying Kirchhoffs voltage law to the

+ +

− −

L

f

R

f

E

s

g

v

s

f

i

s

f

Fig. 4.7. Grid-ﬁlter model in stator coordinates.

circuit in the ﬁgure the following model in synchronous coordinates can be found:

E

g

= −(R

f

+ jω

1

L

f

) i

f

−L

f

di

f

dt

+v

f

(4.44)

where E

g

is the grid voltage, i

f

is the grid-ﬁlter current, and v

f

is the grid-ﬁlter voltage

supplied from the grid-side converter.

43

Harmonics

The transfer function, G

f

(p), of the grid ﬁlter can be expressed as

G

f

(p) =

i

s

f

(p)

v

s

f

(p)

=

1

L

f

p + R

f

. (4.45)

This means that the damping of the grid ﬁlter is given by

[G

f

(jω)[ =

1

L

2

f

ω

2

+ R

2

f

. (4.46)

If L

f

ω R

f

, the gain can be approximated as [G

f

(jω)[ ≈ 1/(L

f

ω). For example, if the

switching frequency of the converter is ω = 100 p.u. and L

f

= 0.2 p.u., then the gain of

the grid ﬁlter is [G

f

(j100)[ ≈ 0.05 p.u. (corresponding to a damping of 26 dB) if R

f

can be

neglected.

4.2.3 DC-Link Model

The energy, W

dc

, stored in the dc-link capacitor, C

dc

, is given by

W

dc

=

1

2

C

dc

v

2

dc

(4.47)

where v

dc

is the dc-link voltage. In Fig. 4.8 an equivalent circuit of the dc-link model, where

the deﬁnition of the power ﬂow through the grid-side converter (GSC) and the machine-

side converter (MSC, can be seen. Moreover, if the losses in the actual converter can be

≈

≈ =

=

+

−

P

r

P

f

C

dc

v

dc

MSC GSC

Fig. 4.8. DC-link model.

considered small and thereby be neglected, the energy in the dc-link capacitor is dependent

on the power delivered to the grid ﬁlter, P

f

, and the power delivered to the rotor circuit of

the DFIG, P

r

, as [76]

dW

dc

dt

=

1

2

C

dc

d

dt

v

2

dc

= −P

f

−P

r

. (4.48)

This means that the dc-link voltage will vary as

C

dc

v

dc

dv

dc

dt

= −P

f

−P

r

(4.49)

which means that P

f

= −P

r

for a constant dc-link voltage.

44

4.2.4 Summary

The total model of the DFIG system, presented in Fig. 4.5 can now be summarized in syn-

chronous coordinate, as

dΨ

s

dt

= E

g

−R

s

i

s

−jω

1

Ψ

s

(4.50)

dΨ

R

dt

= v

R

−R

R

i

R

−jω

2

Ψ

R

(4.51)

L

f

di

f

dt

= v

f

−(R

f

+ jω

1

L

f

) i

f

−E

g

(4.52)

C

dc

v

dc

dv

dc

dt

= −P

f

−P

r

(4.53)

J

n

p

dω

r

dt

= T

e

−T

s

(4.54)

where

Ψ

s

= L

M

(i

s

+i

R

) (4.55)

Ψ

R

= L

σ

i

R

+ L

M

(i

s

+i

R

) (4.56)

T

e

= 3n

p

Im

Ψ

s

i

∗

R

(4.57)

P

r

= 3Re [v

R

i

∗

R

] (4.58)

P

f

= 3Re

v

f

i

∗

f

. (4.59)

Note that in (4.50) that the stator voltage, v

s

, has been changed to the grid voltage, E

g

.

4.3 Field Orientation

In order to control the rotor current of a DFIG by means of vector control, the reference

frame has to be aligned with a ﬂux linkage. One common way is to control the rotor currents

with stator-ﬂux orientation [46, 61, 80, 99], or with air-gap-ﬂux orientation [107, 110]. If the

stator resistance is considered to be small, stator-ﬂux orientation gives orientation also with

the stator voltage [17, 61, 68]. According to [17], pure stator-voltage orientation can be done

without any signiﬁcant error. Note that in this thesis stator-voltage orientation will be, from

now on, referred to as grid-ﬂux orientation [21], i.e., the machine is aligned with a virtual

grid ﬂux.

Fig. 4.9 shows an example of the space vectors of the grid voltage and the stator ﬂux.

As illustrated by the ﬁgure there is only a small angular difference between the grid-voltage

and stator-ﬂux space vectors in the stator-ﬂux reference frame compared to the grid-ﬂux

reference frame.

4.3.1 Stator-Flux Orientation

For a stator-ﬂux-oriented system the synchronous angle θ

1

is deﬁned as

θ

1

= ∠Ψ

s

s

(4.60)

45

a)

b)

d d

q q

E

g

E

g

Ψ

s

Ψ

s

Ψ

g

Fig. 4.9. Space-vector diagram of grid voltage and stator ﬂux. a) Stator-ﬂux orientation. b) Grid-ﬂux

orientation.

where Ψ

s

s

is the stator ﬂux in stator coordinates. Then the stator ﬂux can be transformed to

synchronous coordinates as

Ψ

s

= Ψ

s

s

e

−j

ˆ

θ

1

= ψ

s

e

j

˜

θ

1

(4.61)

where

ˆ

θ

1

is the estimate of θ

1

, ψ

s

is the stator ﬂux magnitude, and

˜

θ

1

= θ

1

−

ˆ

θ

1

is the error

between the synchronous angle and its estimate. This means that for perfect ﬁeld orientation,

i.e. θ

1

=

ˆ

θ

1

, that Ψ

s

= ψ

s

, i.e., the space vector of the ﬂux is real valued. Moreover, if

the stator current, rotor current and the rotor position are measured, the stator ﬂux can be

calculated and thus the transformation angle can be found.

4.3.2 Grid-Flux Orientation

The basic idea behind grid-ﬂux orientation is to deﬁne a virtual grid ﬂux, Ψ

s

g

, as [21, 76]

Ψ

s

g

=

E

s

g

jω

g

= −

jE

g

e

jθ

g

ω

g

(4.62)

where ω

g

is the frequency of the grid voltage and θ

g

is the corresponding angle. Since ω

g

is close to constant, the virtual grid ﬂux is linked to the grid voltage. This means that the

relationship between the synchronous angle, θ

1

, and the grid voltage angle, θ

g

, for a grid-ﬂux

oriented (or stator-voltage oriented) system, is

θ

1

= ∠Ψ

s

g

= ∠ −jE

s

g

= θ

g

−

π

2

. (4.63)

Then, the grid voltage can be transformed to synchronous coordinates as

E

g

= E

s

g

e

−j

ˆ

θ

1

= jE

g

e

j

˜

θ

1

(4.64)

where E

g

is the grid voltage magnitude. This means that for perfect ﬁeld orientation, i.e.,

θ

1

=

ˆ

θ

1

, that E

g

= jE

g

, i.e., the space vector of the grid voltage is imaginary. Note that

the grid-ﬂux orientation is equal to the stator-ﬂux orientation in the steady state, if the stator

resistance can be neglected, since then

v

s

= E

g

= R

s

i

s

+

dΨ

s

dt

+ jω

1

Ψ

s

≈ jω

1

Ψ

s

(4.65)

46

and ω

1

= ω

g

. The transformation angle for a grid-ﬂux oriented system can be found directly

from measurements of the stator voltage. However, in order to have some ﬁltering effect a

PLL-estimator, as described in Section 4.1.3, can be used to track the grid frequency and its

corresponding angle.

4.4 Control of Machine-Side Converter

The main task of the machine-side converter is, of course, to control the machine. This is

done by having an inner fast ﬁeld-oriented current control loop that controls the rotor current.

The ﬁeld orientation could, for example, either be aligned with the stator ﬂux of the DFIG

or the grid ﬂux. For both reference frames the q component of the rotor current largely

determines the produced torque while the d component can be used to control, for instance,

the reactive power at the stator terminals.

4.4.1 Current Control

As mentioned earlier, it is common to control the rotor current with either stator-ﬂux orien-

tation or grid-ﬂux orientation. In order to derive the rotor-current control law, it is advanta-

geous to eliminate i

s

and Ψ

R

from (4.38) and (4.39), which yields

v

s

= −R

s

i

R

+

dΨ

s

dt

+

R

s

L

M

+ jω

1

Ψ

s

(4.66)

v

R

= (R

R

+ jω

2

L

σ

)i

R

+ L

σ

di

R

dt

+

dΨ

s

dt

+ jω

2

Ψ

s

= (R

R

+ R

s

+ jω

2

L

σ

)i

R

+ L

σ

di

R

dt

+E

(4.67)

E = v

s

−

R

s

L

M

+ jω

r

Ψ

s

(4.68)

where E is the back EMF. It is possible to decouple the cross coupling between the d and q

components of the rotor current—jω

2

L

σ

i

R

in (4.67)—in the control law [17, 80]. Further, it

is possible to include a feed-forward compensating term in the control law that will compen-

sate for the tracking error caused by variations in the back EMF. In [46, 61, 80] this is done

by feed forward of the term jω

2

Ψ

s

and neglecting the derivative of the ﬂux in (4.67). Here,

an estimate of the whole back EMF,

ˆ

E, will be used:

v

R

= v

R

+ (j ˆ ω

2

ˆ

L

σ

−R

a

)i

R

+ k

E

ˆ

E

= k

p

e + k

i

e dt + (j ˆ ω

2

ˆ

L

σ

−R

a

)i

R

+ k

E

ˆ

E. (4.69)

where “ˆ” indicates an estimated quantity. A coefﬁcient k

E

is introduced in order to make

the control law more general and to simplify the analysis in Chapter 5:

k

E

=

**0 for control without feed forward of E
**

1 for control with feed forward of E.

(4.70)

Furthermore, in (4.69), an “active resistance,” R

a

, has been introduced. The “active re-

sistance” is used to increase the damping of disturbances and variations in the back EMF.

47

Similar approaches have been used for the squirrel-cage IG [18, 41]. How to choose the

“active resistance” will be shown in next section. If the estimate of the slip frequency, ˆ ω

2

, is

put to zero in (4.69), the d and the q components of the rotor current will not be decoupled.

In [46] it is stated that the inﬂuence of the decoupling term jω

2

L

σ

i

R

is of minor importance,

since it is an order of magnitude smaller than the term jω

2

Ψ

s

. Nevertheless, here the d and q

components of the rotor current will be decoupled, since for a DSP-based digital controller

it is easy to implement.

Substituting (4.69) in (4.67), the rotor current dynamics formed by the inner loop in

Fig. 4.10 are now given by

L

σ

di

R

dt

= v

R

−(R

R

+ R

s

+ R

a

)i

R

(4.71)

where the estimated parameters in the control law are assumed to have the correct values. If

the back EMF is not compensated for, i.e., k

E

= 0 in (4.69), it is treated as a disturbance to

the rotor current dynamics. The transfer function from v

R

to i

R

is

G(p) =

1

pL

σ

+ R

R

+ R

s

+ R

a

which via (4.17) yields the following controller parameters

k

p

= α

c

ˆ

L

σ

k

i

= α

c

(

ˆ

R

R

+

ˆ

R

s

+ R

a

) (4.72)

where α

c

is closed-loop bandwidth of the current dynamics, giving

G

cl

(p) =

p

p + α

c

. (4.73)

+

+

+ +

−

− −

¸ ¸ ¸

G(p)

i

R i

ref

R

k

p

+

k

i

p

R

a

−j ˆ ω

2

ˆ

L

σ

v

R v

R

E

ˆ

E

DFIG

Fig. 4.10. Block diagram of the current control system. The dashed box is the model for the doubly-

fed induction generator.

48

Selection of the “Active Resistance”

If the “active resistance” is set to R

a

= k

R

(α

c

ˆ

L

σ

−

ˆ

R

R

−

ˆ

R

s

), the transfer function from the

back EMF, E, to the current, i

R

, cf. Fig. 4.10, is given as

G

Ei

(p) =

−p/(p + α

c

)

pL

σ

+ L

σ

α

c

k

R

+ (1 −k

R

)(R

r

+ R

s

)

(4.74)

if all model parameters are assumed to be accurate. A parameter k

R

is introduced in a fashion

similar to (4.70):

k

R

=

**0 for control without “active resistance”
**

1 for control with “active resistance.”

(4.75)

This yields

G

Ei

(p) =

⎧

⎪

⎨

⎪

⎩

−p

(p + α

c

)(pL

σ

+ R

R

+ R

s

)

k

R

= 0

−p

L

σ

(p + α

c

)

2

k

R

= 1

(4.76)

This means that the above choice of R

a

will force a change in the back EMF to be damped

with the same bandwidth as the closed-loop current dynamics. Since R

a

should be greater

than zero, the minimum bandwidth of the current control loop when using “active resistance”

becomes

α

c,min

= (R

R

+ R

s

)/L

σ

. (4.77)

For the investigated system, α

c,min

equals 0.08 p.u., which can be considered as a low value

since for modern drive, a current control loop bandwidth of 7 p.u. is reasonable [40], corre-

sponding to a rise time of 1 ms at a base frequency of 50 Hz.

In order to investigate the performance of the “active resistance” with regards to damping

of disturbances, we study the ratio between the moduli of the frequency function correspond-

ing to (4.74), with and without “active resistance”:

G

r

(ω) =

[G

Ei

(jω)[

k

R

=1

[G

Ei

(jω)[

k

R

=0

=

(R

R

+ R

s

)

2

+ ω

2

L

2

σ

(α

2

c

+ ω

2

)L

2

σ

. (4.78)

The following two extreme values of the above ratio are worth noting:

G

r

(ω) =

⎧

⎨

⎩

R

R

+ R

s

α

c

L

σ

when ω −→0

1 when ω −→∞

(4.79)

which shows that while the “active resistance” has little impact on high-frequency distur-

bances, the damping of low-frequency disturbances is signiﬁcantly improved, since, typi-

cally, R

R

+R

s

<α

c

L

σ

. In Fig. 4.11, G

r

(ω), is depicted for a current control loop bandwidth

of 7 p.u. It can be seen that when using “active resistance,” the damping of low-frequency

disturbances has been signiﬁcantly improved.

49

0 5 10 15 20 25

−100

−50

0

ω [p.u.]

G

r

(

ω

)

[

d

B

]

Fig. 4.11. Ratio of the damping improvement when using “active resistance” as a function of the

frequency, ω, of the back EMF. The bandwidth of the current control loop is set to 7 p.u.

4.4.2 Torque Control

The electromechanical torque can be found from (4.42) as

T

e

= 3n

p

Im

Ψ

s

i

∗

R

≈ −3n

p

ψ

s

i

Rq

. (4.80)

For a stator-ﬂux-oriented system the above approximation is actually an equality. Since the

stator ﬂux, ψ

s

, is almost ﬁxed to the stator voltage, the torque can be controlled by the q

component of the rotor current, i

Rq

. Since it is difﬁcult to measure the torque, it is most

often controlled in an open-loop manner. Therefore, the q component reference current, i

ref

Rq

,

can be determined from the reference torque, T

ref

e

, as

i

ref

Rq

= −

T

ref

e

3n

p

ˆ

ψ

s

. (4.81)

Instead of using the actual ﬂux in (4.81), the approximation

ˆ

ψ

s

≈ E

g,nom

/ω

1

can be used.

Fig. 4.12 shows a block diagram of the open-loop torque control scheme.

T

ref

e

i

ref

Rq

−

1

3n

p

ˆ

ψ

s

Fig. 4.12: Block diagram of the open-loop torque control.

4.4.3 Speed Control

Since the current dynamics, i.e., with the bandwidth α

c

, should be set much faster than the

speed dynamics, the speed can be controlled in cascade with the current. The mechanical

dynamics are described by

J

n

p

dω

r

dt

= T

e

−T

s

(4.82)

50

where T

e

is the electromechanical torque and T

s

is the shaft torque. The electromechanical

torque can be expressed, under the assumption that the current dynamics are much faster

than the speed dynamics, as

T

e

= T

ref

e

(4.83)

where the reference torque is set to

T

ref

e

= T

ref

e

−B

a

ω

r

(4.84)

where an “active damping” term, B

a

, is introduced. This is, as mentioned earlier, an inner

feedback loop [41], that can be used to improve the damping of disturbances. How to de-

termine the “active damping” will be shown in the next section. The transfer function from

T

ref

e

to ω

r

, treating the shaft torque, T

s

, as a disturbance, then becomes

G(p) =

ω

r

(p)

T

ref

e

(p)

=

1

J

n

p

p + B

a

. (4.85)

Using IMC, as described in Section 4.1.4, the following PI controller can be found:

F(p) =

α

s

p

G

−1

(p) = k

p

+

k

i

p

=

ˆ

Jα

s

n

p

+

B

a

α

s

p

(4.86)

where α

s

is the desired closed-loop bandwidth of the speed-control loop and the notation

“ˆ” indicates an estimated quantity. Fig. 4.13 shows a block diagram of the speed control

system.

¸ ¸ ¸ n

p

Jp

B

a

F(p)

ω

ref

r

ω

r

T

ref

e

T

ref

e

T

s

+ + +

−

− −

Fig. 4.13. Speed control loop.

Choosing the “Active Damping”

The transfer function from the shaft torque, T

s

, to the rotational speed can be described by,

see Fig. 4.13,

ω

r

(p)

T

s

=

p

ˆ

J

n

p

p

2

+ (B

a

+ k

p

)p + k

i

(4.87)

51

if the active damping term is chosen as B

a

= α

s

ˆ

J/n

p

, (4.87), becomes

ω

r

(p)

T

s

=

p

J

n

p

(p + α

s

)

2

(4.88)

i.e., change in the shaft torque, T

s

, is damped with the same time constant as the bandwidth

of the speed-control loop. In (4.88) all parameters are assumed to be ideal.

Evaluation

Fig. 4.14 shows a simulation of the speed control loop with rated driving torque. The band-

width of the current control loop is set to 1.4 p.u. and the bandwidth of the speed control

loop, α

s

, is set to 0.014 p.u. A bandwidth of 1.4 p.u. corresponds to a rise time of 5 ms and

0.014 p.u. corresponds to 0.5 s. Initially the speed reference is set to 1.25 p.u., after 1 s it is

changed to 0.75 p.u. After 4 s it is changed back to 1.25 p.u., and after 7 s the reference is

ramped down during 3 s to 1 p.u. Finally, after 13 s the driving torque is changed to 50%

of its rated value. The simulations shows that the speed-control loop behaves as expected.

Moreover, it can be seen in the simulation that the speed reference step at 1 s forces limita-

tion of the rotor current, since the maximum rotor current has been reached. This causes the

rise time of the rotor speed to be longer than the ideal.

0 5 10 15

0.5

1

1.5

0 5 10 15

0

0.5

1

1.5

R

o

t

o

r

S

p

e

e

d

[

p

.

u

.

] a)

Time [s]

R

o

t

o

r

C

u

r

r

e

n

t

[

p

.

u

.

]

b)

Fig. 4.14. Simulation of the speed-control loop. a) Rotor speed. b) Rotor current (q component).

4.4.4 Reactive Power Control

The instantaneous apparent power at the stator terminals, S

s

= P

s

+jQ

s

, can now be found

as

S

s

= 3v

s

i

∗

s

= 3

R

s

i

s

+

dΨ

s

dt

+ jω

1

Ψ

s

i

∗

s

(4.89)

52

Then the active and reactive power, neglecting the derivative of the stator ﬂux, can thus be

written as

P

s

= 3R

s

[i

s

[

2

+ 3ω

1

(ψ

sd

i

sq

−ψ

sq

i

sd

) (4.90)

Q

s

= 3ω

1

(ψ

sd

i

sd

+ ψ

sq

i

sq

) . (4.91)

For a stator-ﬂux-oriented system, i.e., ψ

sd

= ψ

s

and ψ

sq

= 0, the above is reduced to

P

s

= 3R

s

[i

s

[

2

+ 3ω

1

ψ

s

i

sq

= 3R

s

[i

R

[

2

−2

ψ

s

L

M

i

Rd

+

ψ

2

s

L

2

M

−3ω

1

ψ

s

i

Rq

(4.92)

Q

s

= 3ω

1

ψ

s

i

sd

= 3ω

1

ψ

s

ψ

s

L

M

−i

Rd

. (4.93)

For a grid-ﬂux-oriented system, where the voltage is aligned with the q axis, the expression

in (4.92) and (4.93), still holds approximately since R

s

can be considered as small. From

(4.93) it can be seen that if

i

ref

Rd

=

ˆ

ψ

s

ˆ

L

M

(4.94)

the DFIG is operated at unity power factor.

Closed-Loop Reactive Power Control

Since the ﬂux for a DFIG system can be considered as constant, there will be a static rela-

tionship between the reactive power and the d component of the rotor current, G

Qi

Rd

. This

means that IMC yields in an I controller, as

F

Q

=

α

Q

p

G

−1

Qi

Rd

= −

α

Q

3ω

1

ˆ

ψ

s

1

p

(4.95)

where α

Q

is the bandwidth of the reactive power control loop. Moreover, since, in the steady

state,

ˆ

ψ

s

≈ E

g,nom

/ω

1

, the controller reduces to

F

Q

= −

α

Q

3E

g,nom

1

p

(4.96)

or as

i

ref

Rd

= −

α

Q

3E

g,nom

Q

ref

s

−Q

s

dt. (4.97)

Of course, it would be possible to add a feed-forward term in order to compensate for the

magnetizing current, i.e., ψ

s

/L

M

, in the above control law. However, since ψ

s

/L

M

is close

to constant, the integration of the controller will compensate for the magnetizing current.

Therefore, feed-forward compensation has not been considered for the reactive power control

loop.

53

4.4.5 Sensorless Operation

“Sensorless” operation implies in this thesis that neither the rotor position nor the rotor speed

is measured. This means that the stator frequency, ω

1

, and the slip frequency, ω

2

, and their

corresponding angles, θ

1

and θ

2

, must be estimated. The purpose of this section is to give an

overview of some different estimation techniques that are available in literature. Note that if

no stator variables exist in the control law, it might be unnecessary to estimate ω

1

.

Estimation of Synchronous Frequency Angle

For a system which is oriented with the grid ﬂux, or the voltage drop across the stator resis-

tance is negligible, the angle

ˆ

θ

1

can easily be found by using a PLL, see Section 4.1.3, on the

measured grid (stator) voltage. For a stator-ﬂux-oriented control of the doubly-fed induction

generator, where the voltage drop across the stator resistance can not be neglected, the stator

ﬂux can be estimated in stator coordinates using (4.36) as [46, 56]

ˆ

Ψ

s

s

=

(v

s

s

−

ˆ

R

s

i

s

s

)dt (4.98)

and the estimate of the transformation angle, θ

1

, can then be found from

ˆ

θ

1

= ∠

ˆ

Ψ

s

s

. The

notation “ ˆ ” is used for estimated variables and parameters. Since the estimator in (4.98) is

an open-loop integration, it is marginally stable. Thus, it has to be modiﬁed in order to gain

stability. This could be done by replacing the open-loop integration with a low-pass ﬁlter

[39]. It is also possible to estimate the transformation angle in synchronous coordinates.

Starting with the stator voltage equation in stator coordinates and taking into account that for

a stator-ﬂux-oriented system, Ψ

s

s

= ψ

s

e

jθ

1

, yields

v

s

s

= R

s

i

s

s

+

dΨ

s

s

dt

= R

s

i

s

s

+

dψ

s

dt

e

jθ

1

+ jω

1

ψ

s

e

jθ

1

. (4.99)

If v

s

s

= v

s

e

j

ˆ

θ

1

and i

s

s

= i

s

e

j

ˆ

θ

1

, the above equation can be rewritten in synchronous coordi-

nates as

v

s

= R

s

i

s

+

dψ

s

dt

e

j

˜

θ

1

+ jω

1

ψ

s

e

j

˜

θ

1

(4.100)

where

˜

θ

1

= θ

1

−

ˆ

θ

1

is the angular estimation error. Taking the real part of the above equation

and neglecting the ﬂux dynamics yield

v

sd

= R

s

i

sd

−ω

1

ψ

s

sin(

˜

θ

1

). (4.101)

Now, it is possible to form an error signal suitable for the PLL-type estimator, described in

Section 4.1.3, as

ε = sin(θ

1

−

ˆ

θ

1

) = sin(

˜

θ

1

) = −

v

sd

−

ˆ

R

s

i

sd

ω

1

ψ

s

≈ −

v

sd

−

ˆ

R

s

i

sd

v

s

(4.102)

where the approximation is due to the fact that the stator is directly connected to the grid, so

ω

1

ψ

s

≈ v

s

.

54

Estimation of Slip-Frequency Angle

In the literature there are at least two methods to perform sensorless operation. In the ﬁrst

method, a set of variables is estimated or measured in one reference frame and then the

variables are used in another reference frame to estimate the slip angle θ

2

. Estimating the

rotor currents from the ﬂux and the stator currents can do this. In [15] the estimation of the

rotor currents has been carried out in stator coordinates, while in [46, 68] it has been done

in synchronous coordinates. The method will here be described in synchronous coordinates.

Starting with the stator ﬂux, which, in synchronous coordinates, is given by

Ψ

s

= ψ

s

= L

M

(i

s

+i

R

) (4.103)

and since the stator ﬂux is known, i.e., it is to a great extent determined by the stator voltage,

it is possible to use the above-mentioned equation to estimate the rotor current as follows:

ˆ

i

R

=

ˆ

ψ

s

ˆ

L

M

−i

s

(4.104)

where the stator current has been measured and transformed with the transformation angle

θ

1

; see previous section for determination of this angle. The magnitude of the stator ﬂux can

be estimated as

ˆ

ψ

s

= v

s

/ω

1

[46]. Then, if the rotor current is measured in rotor coordinates

the estimate of the slip angle can be found as

ˆ

θ

2

= ∠i

r

R

−∠

ˆ

i

R

. (4.105)

The second method is based on determining the slip frequency by the rotor circuit equation.

In [56] a stator-ﬂux-oriented sensorless control using the rotor voltage circuit equation is

proposed. The rotor voltage equation is given by

v

R

= R

R

i

R

+

dΨ

R

dt

+ jω

2

Ψ

R

. (4.106)

Neglecting the derivative of the ﬂux, the slip frequency, ω

2

, can be estimated from the imag-

inary part of the above equation as

ˆ ω

2

=

v

Rq

−

ˆ

R

R

i

Rq

ψ

Rd

=

v

Rq

−

ˆ

R

R

i

Rq

ψ

s

+

ˆ

L

σ

i

sd

. (4.107)

Then, the estimate of the slip angle,

ˆ

θ

2

, can be found from integration of the estimate of the

slip frequency, ˆ ω

2

, as

ˆ

θ

2

=

ˆ ω

2

dt. (4.108)

4.5 Control of Grid-Side Converter

The main objective of the grid-side converter is to control the dc-link voltage. The control of

the grid-side converter consists of a fast inner current control loop, which controls the current

through the grid ﬁlter, and an outer slower control loop that controls the dc-link voltage. The

reference frame of the inner current control loop will be aligned with the grid ﬂux. This

means that the q component of the grid-ﬁlter current will control the active power delivered

from the converter and the d component of the ﬁlter current will, accordingly, control the

reactive power. This implies that the outer dc-link voltage control loop has to act on the q

component of the grid-ﬁlter current.

55

4.5.1 Current Control of Grid Filter

In (4.44) the dynamics of the grid ﬁlter are described:

L

f

di

f

dt

= v

f

−(R

f

+ jω

1

L

f

) i

f

−E

g

. (4.109)

In order to introduce “active damping” and decouple the d and the q components of the

grid-ﬁlter current, the applied grid-ﬁlter voltage, v

f

, is chosen as

v

f

= v

f

−(R

af

−jω

1

ˆ

L

f

)i

f

. (4.110)

This means that the inner closed-loop transfer function, assuming ideal parameters, becomes

G(p) =

i

f

(p)

v

f

(p)

=

1

L

f

p + R

f

+ R

af

(4.111)

and, hence, by using IMC a PI controller can be determined with the bandwidth α

f

. By

choosing the active damping according to Section 4.1.5, i.e., R

af

= α

f

ˆ

L

f

−

ˆ

R

f

, the transfer

function from grid voltage (“back emf”), E

g

, to the grid-ﬁlter current with ideal parameters

then becomes

G

E

g

i

f

(p) =

p

L

f

(p + α

f

)

2

. (4.112)

Finally, the grid-ﬁlter current control law can now be written as

v

f

=

k

pf

+

k

if

p

(i

ref

f

−i

f

) −(R

af

−jω

1

ˆ

L

f

)i

f

(4.113)

where

k

pf

= α

f

ˆ

L

f

k

if

= α

f

(

ˆ

R

f

+ R

af

) = α

2

f

ˆ

L

f

R

af

= α

f

ˆ

L

f

−

ˆ

R

f

. (4.114)

4.5.2 DC-Link Voltage Control

The dc-link voltage control in this thesis is essentially following [76]. One way of simplify-

ing the control of the dc-link voltage is by utilizing feedback linearization, i.e., the nonlinear

dynamics of the dc link are transformed into an equivalent linear system where linear control

techniques can be applied [95]. This can be done by letting W = v

2

dc

[50, 76, 79]. The

dc-link dynamics (4.48) are, thus, reduced to the following linear system

1

2

C

dc

dW

dt

= −P

f

−P

r

(4.115)

where, as mentioned earlier, P

f

is the power delivered to the grid ﬁlter and P

r

is the power

delivered to the rotor circuit of the DFIG. If the power losses of the grid ﬁlter are small and

the current control of the grid ﬁlter is aligned with the grid ﬂux, the power delivered to the

grid ﬁlter can be approximated as P

f

≈ 3E

gq

i

fq

. Moreover, by assuming the current control

loop to be fast, i.e., i

fq

= i

ref

fq

, and adding an “active damping” term as

i

ref

fq

= i

ref

fq

+ G

a

W (4.116)

56

where G

a

is the gain of the “active damping,” it is possible to write the dc-link dynamics as

1

2

C

dc

dW

dt

= −3E

g

i

ref

fq

−3E

g

G

a

W −P

r

. (4.117)

The inner closed-loop transfer function becomes

G

(p) =

W(p)

i

ref

fq

(p)

=

−6E

g

pC

dc

+ 6E

g

G

a

. (4.118)

Then, by utilizing IMC, the following PI controller is obtained

F(p) =

α

w

p

G

−1

(p) = −

α

w

ˆ

C

dc

6E

g,nom

−

α

w

G

a

p

(4.119)

where the magnitude of the grid voltage, E

g

, is put to its nominal value, E

g,nom

, and α

w

is the bandwidth of the dc-link voltage control loop. If the active damping is chosen as

G

a

= α

w

ˆ

C

dc

/(6E

g,nom

), a disturbance, i.e., P

r

, will be damped as

G

PW

(p) =

−2p

C

dc

(p

2

+ 2α

w

ξp + α

2

w

ξ)

(4.120)

where ξ = E

gq

/E

g,nom

and

ˆ

C

dc

= C

dc

. With E

gq

= E

g,nom

, i.e., ξ = 1, G

PW

(p) is reduced

to

G

PW

(p) =

−2α

w

p

C

dc

(p + α

w

)

2

(4.121)

which means that a disturbance is damped with the same bandwidth as the dc-link voltage

control loop. A block diagram of the dc-link voltage controller is depicted in Fig. 4.15.

¸ ¸ ¸ 2

pC

dc

G

(p)

G

a

F(p)

W

ref

W = v

2

dc

i

fq

i

fq

P

r

3E

gq

+

+

+

−

−

−

Fig. 4.15. DC-link voltage control loop.

57

58

Chapter 5

Evaluation of the Current Control of

Doubly-Fed Induction Generators

In this chapter the current control law derived for the DFIG in the previous chapter is ana-

lyzed with respect to eliminating the inﬂuence of the back EMF, which is dependent on the

stator voltage, rotor speed, and stator ﬂux, in the rotor current. Further, stability analysis of

the system is performed for different combinations of these terms in both a stator-ﬂux and

grid-ﬂux-oriented reference frame, for both correctly known and erroneously parameters.

5.1 Stability Analysis

In order to investigate the inﬂuence of the feed-forward compensation of the back EMF and

the inﬂuence of the “active resistance” on the stability of the system, an analysis is performed

in this section. The analysis will be performed both for a stator-ﬂux-oriented system and for

a grid-ﬂux-oriented system. In this section a full-order analysis of the system is performed,

since one of the objectives is to study the impact of the current control law derived in the

previous chapter.

5.1.1 Stator-Flux-Oriented System

Consider the system described by (4.66)–(4.68). Splitting (4.66) into real and imaginary

parts, assuming stator-ﬂux orientation, i.e., Ψ

s

s

= ψ

s

e

jθ

1

, the stator voltage equals the grid

voltage, i.e., v

s

= jE

g

e

j(θ

g

−θ

1

)

. Making the variable substitution Δθ = θ

g

− θ

1

, the system

model can be rearranged as

dI

dt

= e (5.1)

dψ

s

dt

= −

R

s

L

M

ψ

s

−E

g

sin(Δθ) + R

s

i

Rd

(5.2)

dΔθ

dt

=

dθ

g

dt

−

dθ

1

dt

= ω

g

−

E

g

cos(Δθ) + R

s

i

Rq

ψ

s

(5.3)

di

R

dt

=

k

p

e + k

i

I + (j ˆ ω

2

ˆ

L

σ

−R

a

)i

R

+ k

E

ˆ

E

L

σ

−

(R

R

+ R

s

+ jω

2

L

σ

)i

R

+E

L

σ

(5.4)

59

where

E = E

g

−

R

s

L

M

+ jω

r

Ψ

s

= jE

g

e

jΔθ

−(

R

s

L

M

+ jω

r

)ψ

s

. (5.5)

In (5.1)–(5.4), the termI is the integration variable of the control error and e = i

ref

R

−i

R

is the

control error. Note that (5.1) and (5.4) are complex-valued equations while (5.2) and (5.3)

are real-valued equations. In the following analysis, the rotational speed ω

r

will be assumed

to be varying slowly, and is, therefore, treated as a parameter. Throughout this section, the

machine model parameters will be assumed to be ideal and known.

If, for example, the rotor current is controlled by a high-gain feedback, it is possible

to force the system to have both slow and fast time scales, i.e., the system behaves like a

singularly perturbed system [57]. This means, that if the bandwidth of the current control

loop is high enough, it is sufﬁcient to study the system described by (5.2) and (5.3) in order

to analyze the dynamic behavior of the DFIG. A stability analysis, assuming fast current

dynamics, can be found in [13, 43]. Later on, analysis not neglecting the current dynamics

will be compared to analysis neglecting the current dynamics; therefore a short summary

will be presented. By linearization of the nonlinear system described by (5.2) and (5.3),

the characteristic polynomial can be found. A ﬁrst-order Taylor series expansion of the

characteristic polynomial around R

s

= 0 (as R

s

is small, typically less than 0.1 p.u.) yields

p

2

+

R

s

L

M

2 −

ω

g

L

M

i

ref

Rd

E

g

p +

1 −

R

s

i

ref

Rq

E

g

ω

2

g

(5.6)

where i

ref

Rq

is the active current reference and i

ref

Rd

is the magnetization current supplied from

the rotor converter. Since R

s

is small (< 0.1 p.u.), i

ref

Rq

will only have a minor inﬂuence on

the dynamics. However, i

ref

Rd

will inﬂuence the dynamic performance. It is required that

i

ref

Rd

<

2E

g

ω

g

L

M

(5.7)

in order to maintain stability. A similar constraint can be found in [13, 43]. In order to

operate the DFIG with unity power factor, one should select [99]

i

ref

Rd

=

ψ

s

L

M

≈

E

g

ω

g

L

M

(5.8)

which value is half of the value in the condition in (5.7).

For the case when it is not possible to separate the time scales by a high-gain feedback in

the current control loop, a full-order analysis should be performed. By linearizing the non-

linear system described by (5.1)–(5.4) in a similar manner as previously, the characteristic

polynomial for the complete system can be found. A ﬁrst-order Taylor series expansion of

the characteristic polynomial around R

s

= 0 yields

(p + α

c

)

p + k

R

α

c

+ (1 −k

R

)

R

R

L

σ

(p

4

+ a

3

p

3

+ a

2

p

2

+ a

1

p + a

0

). (5.9)

60

where expressions for the coefﬁcients a

3

to a

0

become

a

3

= α

c

(1 + k

R

) +

2R

s

L

M

−

R

s

ω

g

i

ref

Rd

E

g

+ (1 −k

R

)

R

R

+ 2R

s

L

σ

(5.10)

a

2

= α

2

c

k

R

+ ω

2

g

−

i

ref

Rq

R

s

ω

2

g

E

g

+ (1 + k

R

)α

c

R

s

2

L

M

−

i

ref

Rd

ω

g

E

g

+

1 −k

R

L

σ

α

c

R

R

+ 2R

s

+

1 −k

R

L

σ

R

R

R

s

2

L

M

−

i

ref

Rd

ω

g

E

g

(5.11)

a

1

=

2α

2

c

k

R

R

s

L

M

+ (1 −k

R

)

2α

c

R

R

R

s

L

M

L

σ

−

α

2

c

k

R

R

s

i

ref

Rd

ω

g

E

g

−(1 −k

R

)

α

c

i

ref

Rd

R

R

R

s

ω

g

L

σ

E

g

−(1 −k

E

)

R

s

ω

r

ω

g

L

σ

+ (1 + k

R

)α

c

ω

2

g

1 −

i

ref

Rq

R

s

E

g

+ (1 −k

R

)ω

2

g

R

R

L

σ

1 −

i

ref

Rq

R

s

E

g

+

R

s

L

σ

(1 −2k

R

+ k

E

)ω

2

g

(5.12)

a

0

= α

2

c

ω

2

g

k

R

−

α

2

c

i

ref

Rq

R

s

ω

2

g

k

R

E

g

+ α

c

ω

2

g

1 −k

R

L

σ

R

R

+ 2R

s

−

i

ref

Rq

R

R

R

s

E

g

. (5.13)

The parameters k

E

and k

R

affect the roots of (5.9), directly and via a

0

–a

3

. The four different

combinations of k

E

and k

R

available are, according to Table 5.1, termed Methods I–IV.

Below, the characteristic polynomial (5.9) is investigated for the four different options.

TABLE 5.1. INVESTIGATED CURRENT CONTROL METHODS.

k

E

k

R

Method I 0 0

Method II 0 1

Method III 1 0

Method IV 1 1

Methods I and II

Both methods give two real-valued poles (at −α

c

, −R

R

/L

M

for Method I and two at −α

c

for Method II) and four poles given by the fourth-degree factor. In Fig. 5.1, it is shown how

one of the complex-conjugated poles given by the fourth-degree factor move with increasing

bandwidth of the current control loop, α

c

. The other complex-conjugated poles given by the

fourth-degree factor are well damped and are therefore not shown in the ﬁgure. The IM is

running as a generator at half of the rated torque, synchronous speed, and is magnetized from

the rotor circuit. It can be seen in the ﬁgure that the poorly damped poles of Method II move

with increasing bandwidth of the current control loop from stable to unstable and back to be

stable again, while for Method I the poles are stable. Method II is unstable for bandwidths

of the current control loop between 1.0–5.6 p.u. for the above mentioned operating point.

Moreover, as shown in Fig. 5.1, the real part of the poorly damped pole is very small. This

means that, when approximating fast current dynamics (marked with “x” in the ﬁgure), even

a small error (due to the approximation of a fast current dynamics) may play a signiﬁcant

61

−0.05 −0.025 0 0.025 0.05

0.97

0.975

0.98

0.985

0.99

0.995

1

Re

I

m

Fig. 5.1. Root loci of one of the poorly damped poles of the doubly-fed induction generator using

current control methods without feed forward of the back EMF without “active resistance”

(Method I, solid) and with “active resistance” (Method II, dashed). The arrow shows how

the poles move with increasing bandwidth (0.5–15 p.u.) of the current control loop. The

symbol “x” indicates the pole location when the current dynamics are neglected.

role for the result of the stability analysis. Hence, it is necessary to make a careful stability

analysis, at least when using Methods I or II.

A similar approach, as will be performed in the next section, with Routh’s table pro-

duces very large expressions of which it is difﬁcult to determine any constrains for stability.

Therefore, the approach with Routh’s table is not carried out for Methods I and II.

Method III

When using Method III, i.e., feed forward of the back EMF, the characteristic polynomial

in (5.9) is reduced to

(p + α

c

)

2

p +

R

R

L

σ

(p

3

+ b

2

p

2

+ b

1

p + b

0

). (5.14)

The system has at least three real-valued poles, two located at −α

c

and one at −R

R

/L

σ

. The

coefﬁcients in the third-degree factor become

b

2

=

2R

s

L

M

+

R

R

+ 2R

s

L

σ

−

i

ref

Rd

R

s

ω

g

E

g

(5.15)

b

1

=

(E

g

−i

ref

Rq

R

s

)ω

2

g

E

g

+

R

R

R

s

(2E

g

−i

ref

Rd

L

M

ω

g

)

L

M

L

σ

E

g

(5.16)

b

0

=

−i

ref

Rq

R

R

R

s

+ (R

R

+ 2R

s

)E

g

ω

2

g

L

σ

E

g

. (5.17)

As can be seen, the coefﬁcients are not dependent on α

c

for Method III.

62

TABLE 5.2. ROUTH’S TABLE.

p

3

1 b

1

p

2

b

2

b

0

p

1

B =

b

2

b

1

−b

0

b

2

0

p

0

Bb

0

−0

B

= b

0

0

To investigate the stability of the system, Routh’s table can be used [14], see Table 5.2.

In order for the system to be stable, the coefﬁcients in the ﬁrst column must not change

sign. Since the ﬁrst coefﬁcient in Routh’s table is 1, all other coefﬁcients must be positive in

order to maintain stability. The expression for the coefﬁcient B becomes quite complex; an

approximation is

B ≈

R

s

2E

g

−ω

g

L

M

i

ref

Rd

(R

2

R

+ ω

2

g

L

2

σ

)

L

M

L

σ

R

R

E

g

(5.18)

where a ﬁrst-order Taylor series expansion of the coefﬁcient B with respect to the stator

resistance, R

s

(around R

s

= 0), has been carried out. The following constraint can be set on

i

ref

Rd

in order to keep the coefﬁcient b

2

positive:

i

ref

Rd

<

E

g

R

s

ω

g

2R

s

L

M

+

R

R

+ 2R

s

L

σ

. (5.19)

For keeping the coefﬁcient B positive, the following constraint has to be set

i

ref

Rd

<

2E

g

ω

g

L

M

. (5.20)

Since the term−i

ref

Rq

R

R

R

s

in b

0

is at least one order of magnitude lower than the term (R

R

+

2R

s

)E

g

, b

0

can be considered to be positive. The constraint in (5.20) is “harder” than the

constraint in (5.19). The constraint in (5.20) is identical to the constraint in (5.7) where

the stability analysis was performed assuming fast current dynamics. The system has two

poorly damped poles, caused by the ﬂux dynamics, and the constraint on i

ref

Rd

relates to the

ﬂux dynamics. Therefore, the constraint on i

ref

Rd

, which relates to the ﬂux dynamics, can be

found more easily assuming fast current dynamics. Generally, a full-order analysis is still

valuable, if the current dynamics are not fast, since other parameters also may inﬂuence the

stability (for stability analysis assuming fast current dynamics).

Method IV

For Method IV, i.e., with feed forward of the back EMF and “active resistance,” the charac-

teristic polynomial in (5.9) is reduced to

(p + α

c

)

4

p

2

+

R

s

L

M

2 −

ω

g

L

M

i

ref

Rd

E

g

p +

1 −

R

s

i

ref

Rq

E

g

ω

2

g

. (5.21)

The characteristic polynomial has four real roots located at −α

c

. The second-degree factor

is identical to (5.6), where the current dynamics were neglected. Therefore, for Method IV,

the same analysis as for the case with the assumption of fast current dynamics can be used.

63

5.1.2 Grid-Flux-Oriented System

The corresponding dynamics for the grid-ﬂux-oriented system become

dI

dt

= e (5.22)

dΨ

s

dt

= E

g

−

R

s

L

M

+ jω

g

Ψ

s

+ R

s

i

R

(5.23)

di

R

dt

=

k

p

e + k

i

I + (j ˆ ω

2

ˆ

L

σ

−R

a

)i

R

+ k

E

ˆ

E

L

σ

−

(R

R

+ R

s

+ jω

2

L

σ

)i

R

+E

L

σ

. (5.24)

Note that (5.22)–(5.24) are complex-valued equations. As for the case with stator-ﬂux ori-

ented analysis, the rotational speed ω

r

will be assumed to be varying slowly and is therefore

treated as a parameter. Throughout this section, parameters will exactly as in the previous

section be assumed to be ideal and known.

If, as for the stator-ﬂux orientation, the rotor current is controlled by a high-gain feed-

back, it is sufﬁcient to study the dynamics described by (5.23), which have the following

equilibrium points:

ψ

sd0

=

L

M

i

ref

Rd

R

2

s

+ L

M

i

ref

Rq

R

s

+ E

g

ω

g

R

2

s

+ L

2

M

ω

2

g

≈

E

g

+ R

s

i

ref

Rq

ω

g

(5.25)

ψ

qd0

=

L

M

R

s

i

ref

Rq

R

s

+ E

g

−i

ref

Rd

L

M

ω

g

R

2

s

+ L

2

M

ω

2

g

≈

R

s

(E

g

−i

ref

Rd

L

M

ω

g

)

L

M

ω

2

g

(5.26)

where the approximation is due to a ﬁrst-order Taylor series expansion of R

s

around R

s

= 0.

Then, the following characteristic polynomial can be determined:

p

2

+ 2

R

s

L

M

p + ω

2

g

+

R

2

s

L

2

M

. (5.27)

In (5.27) it can be seen that the DFIG is poorly damped, and that the damping is only depen-

dent of R

s

and L

M

. Moreover if the PLL-type estimator, described in Section 4.1.3 is used

to track the grid voltage, the dynamics of the PLL will be separated from the ﬂux dynamics

in (5.27).

If the rotor currents cannot be neglected, a full-order analysis has to be performed. As

in the previous section, the dynamic systems described by (5.22)–(5.24) consists of two

parameters k

E

and k

R

that could be either set to zero or unity. This yields, in the same way

as for the stator-ﬂux-oriented system, four different options, Method I to Method IV, for the

current control law, see Table 5.1.

Methods I and II

Linearizing of the non-linear systemdescribed by (5.22)–(5.24), its characteristic polynomial

can be found. A ﬁrst-order Taylor series expansion of the characteristic polynomial with

respect to the stator resistance, R

s

(around R

s

= 0) yields

(p + α

c

)

p + k

R

α

c

+ (1 −k

R

)

R

R

L

σ

(p

4

+ a

3

p

3

+ a

2

p

2

+ a

1

p + a

0

). (5.28)

64

where the coefﬁcients a

3

to a

0

become

a

3

= 2

R

s

L

M

+ α

c

(1 + k

R

) −(k

R

−1)

R

R

+ 2R

s

L

σ

(5.29)

a

2

= α

2

c

k

R

+

2(1 + k

R

)α

c

R

s

L

σ

−

(k

R

−1) (2R

R

R

s

+ α

c

L

M

(R

R

+ 2R

s

))

L

M

L

σ

+ ω

2

g

(5.30)

a

1

= 2

α

2

c

k

R

R

s

L

M

+ α

c

(1 + k

R

)ω

2

g

−(k

R

−1)

2α

c

R

R

R

s

+ (R

R

+ 2R

s

)L

M

ω

2

g

L

M

L

σ

−2

R

s

ω

g

ω

r

L

σ

(5.31)

a

0

= α

2

c

k

R

ω

2

g

−

(k

R

−1)(R

R

+ 2R

s

)α

c

ω

2

g

L

σ

. (5.32)

In Fig. 5.2 it is shown how one of the complex-conjugated poles, as given by the fourth-

order characteristic polynomial, move with increasing bandwidth of the current control loop,

α

c

. The second-complex conjugated poles are well damped and are therefore not shown

in the ﬁgure. The operating condition is as in Fig. 5.1. It can be seen in the ﬁgure that

−0.05 −0.025 0 0.025 0.05

0.97

0.975

0.98

0.985

0.99

0.995

1

Re

I

m

Fig. 5.2. Root loci of one of the poorly damped poles of the doubly-fed induction generator using

current control methods without feed forward of the back EMF without “active resistance”

(Method I, solid) and with “active resistance” (Method II, dashed). The arrow shows how

the poles move with increasing bandwidth (0.5–15 p.u.) of the current control loop. The

symbol “x” indicates the pole location when the current dynamics are neglected.

for Method I the poorly damped pole is stable and for Method II the poorly damped pole

moves with increasing bandwidth of the current control loop from stable to unstable and

back to be stable again. For the in the ﬁgure investigated case, the system is unstable for

bandwidths between 1.1 p.u. and 3.8 p.u. for Method II. Of course, since the root loci are

plotted with numerical values the result are only valid for the given operation conditions

and for the investigated machine. As could also be seen for the stator-ﬂux oriented case, a

current control loop bandwidth of approximately 15 p.u. might not be high enough in order

to be able to make the assumption of a fast current dynamics (marked with “x” in the ﬁgure),

65

at least for Method I, since the error is in the same order of magnitude as the real part of the

pole, see Fig. 5.2.

Method III

When using Method III, i.e., feed forward of the back EMF, the characteristic polynomial

can be found from the system (5.22)–(5.24) as

(p + α

c

)

2

p +

R

R

+ R

s

L

σ

2

p

2

+ 2

R

s

L

M

p +

R

2

s

L

2

M

+ ω

2

g

. (5.33)

Note that the above characteristic polynomial has not been expanded by a Taylor series. The

system has at least four real-valued poles, two located at −α

c

and two at −(R

R

+ R

s

)/L

σ

.

Method IV

For Method IV, i.e., with feed forward of the back EMF and “active resistance,” the charac-

teristic polynomial becomes

(p + α

c

)

4

p

2

+ 2

R

s

L

M

p +

R

2

s

L

2

M

+ ω

2

g

. (5.34)

Note that the above characteristic polynomial has not been expanded by a Taylor series.

The characteristic polynomial has four real roots located at −α

c

. The second-degree factor

is identical to the characteristic polynomial in (5.27) where the current dynamics were ne-

glected, i.e., assumed to be much faster than the ﬂux dynamics. Therefore, for Method IV,

the same analysis as for the case with the assumption of fast current dynamics can be used.

5.1.3 Conclusion

It has been shown that by using grid-ﬂux orientation the stability and the damping of the

system is independent of the rotor current, in contrast to stator-ﬂux orientation. This implies

that for a grid-ﬂux-oriented system, it is possible to magnetize the DFIG entirely from the ro-

tor circuit without reducing the damping of the system. Moreover, for the grid-ﬂux-oriented

system, it is possible to produce as much reactive power as possible and still have a stable

system with the same damping from a stability point of view.

By utilizing the feed-forward compensation, stability of the derived current control law

is independent of the bandwidth of the current control loop and the order of the system to

analyze is reduced. Further, as shown in Section 4.4.1, the inclusion of the “active resistance”

improves signiﬁcantly the damping of low-frequency disturbances, for higher bandwidths of

the current control loop. Therefore, Method IV with both feed-forward compensation and

“active resistance” can be assumed to be the best one of the investigated methods.

5.2 Inﬂuence of Erroneous Parameters on Stability

We now study how the closed-loop current-control transfer function, G

cl

(p), given by (4.73)

and the transfer function from a disturbance to the rotor current, G

Ei

(p), given by (4.74) are

66

inﬂuenced by non-ideal parameters. For ideal parameters the rotor current is determined by

i

R

= G

cl

(p)i

ref

R

+ G

Ei

(p)E. (5.35)

The methods where the back EMF is compensated for using feed forward (Methods III and

IV), the back EMF will not be totally compensated for due to non-ideal parameters. This

means that the conditions for impact of parameter variations also hold for the methods with

feed forward of the back EMF, even though the effect might be less severe. Note that (5.35)

is independent of the ﬁeld orientation.

In the analysis below, the error in a parameter is denoted with the symbol ˜, e.g.

˜

L

σ

=

L

σ

−

ˆ

L

σ

. The parameters to be studied in the following are L

σ

, R

s

, and R

R

. Since L

M

is

only included in the feed-forward compensation, it has no impact in the following analysis,

and, hence, it is not included.

5.2.1 Leakage Inductance, L

σ

For errors in

ˆ

L

σ

, the rotor current, given by (5.35) for ideal parameters, is given by

i

R

≈

α

c

p + α

c

1 + j

˜

L

σ

ω

2

G

Ei

(p)

i

ref

R

+

1 −j

˜

L

σ

ω

2

α

c

p + α

c

G

Ei

(p)E (5.36)

where the approximation is due to a ﬁrst-order Taylor series expansion of

˜

L

σ

around

˜

L

σ

= 0

and L

σ

˜

L

σ

. From (5.36) it can be seen that small values of

˜

L

σ

do not signiﬁcantly

inﬂuence the dynamic performance. A similar analytical expression for larger errors in

ˆ

L

σ

is

difﬁcult to derive. In order to study the behavior for larger

˜

L

σ

, root loci are shown in Fig. 5.3

for Method I with three different values of

ˆ

L

σ

. The operating condition corresponds to that

of Fig. 5.1; however, the rotor speed is set to 1.3 p.u. so that the effect of the cross coupling

between the d and the q components is included. It can be seen in Fig. 5.3 that the inﬂuence

of errors in

ˆ

L

σ

is small for the investigated 2-MW DFIG. However, for smaller DFIGs such

as the 22-kW laboratory DFIG, the difference is larger. This is shown in Fig. 5.4. Clearly,

it is preferable to overestimate

ˆ

L

σ

. One reason for this is that the proportional part of the

controller will be increased, see (4.72). Hence, the bandwidth of the current control loop is

increased if

ˆ

L

σ

is overestimated.

5.2.2 Stator and Rotor Resistances, R

s

and R

R

Since errors in R

s

and R

R

inﬂuence the performance in the same way, we will study the sum

of the errors in the resistances:

˜

R =

˜

R

s

+

˜

R

R

.

For Methods II and IV where “active resistance” is used, the rotor current is given by

(5.35) if 2α

c

L

σ

˜

R. This means that when using “active resistance,” the system is not

dependent on errors in R

s

and R

R

. For Methods I and III, the rotor current is found to be

i

R

≈

α

c

(L

σ

p + R

s

+ R

R

−

˜

R)

L

σ

p

2

+ α

c

L

σ

p + (R

s

+ R

R

−

˜

R)α

c

i

ref

R

−

p

L

σ

p

2

+ α

c

L

σ

p + (R

s

+ R

R

−

˜

R)α

c

E

(5.37)

67

−0.05 −0.025 0 0.025

0.95

0.96

0.97

0.98

0.99

1

−0.05 −0.025 0 0.025

0.95

0.96

0.97

0.98

0.99

1

a)

Re Re

I

m

I

m

b)

Fig. 5.3. Root loci of one of the poorly damped poles of the doubly-fed induction generator using

Method I for three different errors in the leakage inductance parameter

˜

L

σ

. Solid is

˜

L

σ

= 0,

dashed

˜

L

σ

= −0.5L

σ

, and dotted

˜

L

σ

= 0.5L

σ

. The arrow shows how the poles move with

increasing bandwidth (0.5–15 p.u.) of the current control loop. a) Stator-ﬂux orientation.

b) Grid-ﬂux orientation.

−0.15 −0.1 −0.05 0 0.05

0.85

0.9

0.95

1

−0.15 −0.1 −0.05 0 0.05

0.85

0.9

0.95

1

a)

Re Re

I

m

I

m

b)

Fig. 5.4. Root loci of one of the poorly damped poles of the laboratory 22 kW doubly-fed induction

generator using Method I for three different errors in the leakage inductance parameter

˜

L

σ

.

Solid is

˜

L

σ

= 0, dashed

˜

L

σ

= −0.5L

σ

, and dotted

˜

L

σ

= 0.5L

σ

. The arrow shows how the

poles move with increasing bandwidth (0.5–15 p.u.) of the current control loop. a) Stator-

ﬂux orientation. b) Grid-ﬂux orientation.

the approximation assuming α

c

L

σ

R

s

+R

R

. In (5.37) it can be seen that if the resistances

are overestimated, i.e.,

˜

R < 0, the damping of the current dynamics are actually improved,

i.e., the same phenomenon as using “active resistance.” Fig. 5.5 shows the root loci for

Method I of the investigated 2-MW DFIG. In the ﬁgure it can be seen that the inﬂuence of

errors in the resistance is small. However, as for the case with errors in

ˆ

L

σ

, the difference

is larger for smaller DFIGs, such as the 22-kW laboratory DFIG. This is shown in Fig. 5.4.

68

−0.05 −0.025 0 0.025

0.97

0.975

0.98

0.985

0.99

0.995

1

−0.05 −0.025 0 0.025

0.97

0.975

0.98

0.985

0.99

0.995

1

a)

Re Re

I

m

I

m

b)

Fig. 5.5. Root loci of one of the poorly damped poles of the doubly-fed induction generator using

Method I for three different errors in the stator and rotor resistances parameters

˜

R. Solid

is

˜

R = 0, dashed

˜

R = −0.5R, and dotted

˜

R = 0.5R. The arrow shows how the poles

move with increasing bandwidth (0.5–15 p.u.) of the current control loop. a) Stator-ﬂux

orientation. b) Grid-ﬂux orientation.

Moreover, as shown previously when only using “active resistance” (Method II), the poorly

damped poles (corresponding to the ﬂux dynamics) could be unstable for certain operating

conditions. Therefore, especially for Method I and smaller machines, the system can become

−0.15 −0.1 −0.05 0 0.05

0.85

0.9

0.95

1

−0.15 −0.1 −0.05 0 0.05

0.85

0.9

0.95

1

a)

Re Re

I

m

I

m

b)

Fig. 5.6. Root loci of one of the poorly damped poles of the laboratory 22 kW doubly-fed induction

generator using Method I for three different errors in the stator and rotor resistances para-

meters

˜

R. Solid is

˜

R = 0, dashed

˜

R = −0.5R, and dotted

˜

R = 0.5R. The arrow shows

how the poles move with increasing bandwidth (0.5–15 p.u.) of the current control loop.

a) Stator-ﬂux orientation. b) Grid-ﬂux orientation.

unstable if the resistances are overestimated, as illustrated in Fig. 5.6. It can also be seen in

the ﬁgures that the grid-ﬂux-oriented system seems, even though the difference is small, to

69

be less sensitive to overestimated R = R

s

+ R

R

in comparison to the stator-ﬂux-oriented

system.

5.3 Experimental Evaluation

The performance of the various current control methods are evaluated by reference step

responses, see Fig. 5.7. See Appendix B.2 for data and parameters of the laboratory setup.

This has been done by letting i

ref

Rq

change from −0.25 p.u. to 0.25 p.u. when the rotor speed,

ω

r

, reaches 0.32 p.u., and vice versa when the rotor speed reaches 0.16 p.u. The DFIG is

magnetized entirely from the stator, i.e., i

ref

Rd

= 0, and is operated under no-load conditions.

Further, the stator voltage of the DFIG was 230 V. Data have been sampled with 10 kHz and

low-pass ﬁltered with a cut-off frequency set to 5 kHz. In the measurements the bandwidth

of the current control was set to 1.4 p.u. Offsets in the stator voltage measurements caused

0 0.25 0.5 0.75 1

−0.5

0

0.5

0 0.25 0.5 0.75 1

−0.5

0

0.5

0 0.25 0.5 0.75 1

−0.5

0

0.5

0 0.25 0.5 0.75 1

−0.5

0

0.5

M

e

t

h

o

d

I

M

e

t

h

o

d

I

I

M

e

t

h

o

d

I

I

I

Time [s]

M

e

t

h

o

d

I

V

d

d

d

d

q

q

q

q

Fig. 5.7. Experiment of the stator-ﬂux oriented current control step responses of the q component of

the rotor current.

a 100-Hz frequency component in the stator voltage, which inﬂuenced the performances of

the current control Methods III and VI, since the stator voltage is included in the control law.

However, a notch ﬁlter limited the inﬂuence of the 100-Hz frequency component. A scrutiny

investigation of Fig. 5.7 shows that Method II gives a 50-Hz ripple. The reason for this is

that by using only “active resistance” to damp the back EMF, the system might be degraded,

i.e., unstable, depending on the bandwidth of the current control loop, as shown earlier. Even

70

though the difference is fairly small, it can be seen in Fig. 5.7 that Method IV managed best

to follow its reference values in this comparison.

5.3.1 Comparison Between Stator-Flux and Grid-Flux-Oriented Sys-

tem

The aim of this section is to experimentally verify the analytical result obtained in Section

5.1, that by using grid-ﬂux orientation the stability and the damping of the system is inde-

pendent of the rotor current, in contrast to stator-ﬂux orientation

In Fig. 5.8 shows an experimental case of a stator-ﬂux-oriented and a grid-ﬂux-oriented

rotor current control. In the ﬁgures the d component of the rotor current is increased from 0

p.u. to 1 p.u. after 0.1 s. The q component of the rotor current is set to 0.5 p.u. When i

Rd

is

0 0.1 0.2 0.3 0.4 0.5

−1

0

1

2

0 0.1 0.2 0.3 0.4 0.5

0

0.5

1

1.5

0 0.1 0.2 0.3 0.4 0.5

−1

0

1

2

0 0.1 0.2 0.3 0.4 0.5

0

0.5

1

1.5

a) b)

C

u

r

r

e

n

t

[

p

.

u

.

]

C

u

r

r

e

n

t

[

p

.

u

.

]

c)

d

d

q

q

Time [s] Time [s]

F

l

u

x

[

p

.

u

.

]

F

l

u

x

[

p

.

u

.

]

d)

Fig. 5.8. Experimental comparison between stator-ﬂux-oriented and grid-ﬂux-oriented systems.

a) Rotor current (stator-ﬂux orientation). b) Stator-ﬂux magnitude (stator-ﬂux orientation).

c) Rotor current (grid-ﬂux orientation). d) Stator-ﬂux magnitude (grid-ﬂux orientation).

increased to 1 p.u. it can be seen that the stator-ﬂux-oriented system becomes unstable with

an increasing amplitude of the ﬂux oscillations. After 0.32 s the rotor current is put to zero in

order to put back the system into a stable operating condition. As expected from the analyt-

ical results, the grid-ﬂux-oriented system remains stable throughout the whole experiment.

During this evaluation, the bandwidth of the current control loop was set to 2.3 p.u. and the

rotor speed, ω

r

, was controlled by a d.c. machine to be 1 p.u.

5.4 Impact of Stator Voltage Sags on the Current Control

Loop

Due to the poorly damped poles, in case of a voltage sag, the ﬂux will enter a damped

oscillation. It is essential that the magnitude of the rotor current is below the rated value of

71

the converter in order not to force the crowbar to go into action, and thereby lose control of

the rotor currents and thus the power production.

Neglecting the current dynamics, the rotor voltage as given in (4.67) can be expressed as

v

R

= (R

r

+ R

s

+ jω

2

L

σ

)i

R

+v

s

−

R

s

L

M

+ jω

r

ψ

s

≈ v

s

−jω

r

ψ

s

= jE

g

e

j(θ

g

−θ

1

)

−jω

r

ψ

s

.

(5.38)

From this equation it can be noted that (since v

s

≈ jω

g

ψ

s

)

Im[v

R

] > 0 if ω

r

< ω

g

Im[v

R

] < 0 if ω

r

> ω

g

Im[v

R

] ≈ 0 if ω

r

= ω

g

.

(5.39)

If the rotor voltage is v

R,0

before the voltage sag, then the change in the rotor voltage will be

Δv

R

= v

R,0

− v

R

. Assuming that the grid voltage (or stator voltage) drops from E

g,nom

to

E

g

at t

sag

, then, at the time instant t

sag

, the rotor voltage will drop

Δv

R

(t = t

sag

) = jE

g,nom

e

jΔθ

0

−jE

g

e

jΔθ

≈ j(E

g,nom

−E

g

) = jΔE

g

(5.40)

since the stator ﬂux and the rotor speed will not change instantaneously. From (5.39) and

(5.40) it can be seen that for ω

r

> ω

g

, the magnitude of the rotor voltage will be instan-

taneously increased with ΔE

g

. If ω

r

< ω

g

, then the value of the rotor voltage magni-

tude will, accordingly, be instantaneously decreased. This implies that the worst case oc-

curs for ω

r

> ω

g

according to (5.39) and (5.40). For example, if ω

r

= 1.3, implying

v

R

≈ −j0.3 before the voltage sag, then, according to (5.40), the rotor voltage will be

v

R

(t = t

sag

) = v

R,0

+jΔE

g

= −j0.3 −j0.4 = −j0.7 for a grid voltage drop ΔE

g

=0.4 p.u.

In Fig. 5.9, the maximum rotor voltage needed due to a symmetrical voltage sag for

current control Methods I and IV can be seen. Method II is not considered, since it is actually

unstable for certain operating conditions as indicated by Fig. 5.1 and Method III due that the

results are relatively similar to those of Method IV. The DFIG is running as a generator

at rated torque and is fully magnetized from the rotor circuit. The rotor speed is 1.3 p.u.

This implies that the rotor voltage is approximately 0.3 p.u. immediately before the voltage

sag occurs. For a wind turbine, this operating condition is disadvantageous since a rotor

voltage of 0.3 p.u. is close to the maximum value needed in order to achieve the desired

variable-speed range for a wind turbine. It can be seen that the maximum rotor voltage will

increase with the size of the voltage sag. Further, the maximum rotor voltage is relatively

independent of the bandwidth of the current control loop for Method IV. It can also be noted

that, generally, Method I requires slightly more rotor voltage than Method IV, especially for

low bandwidths. Further, for higher bandwidths of the current control loop, it can be seen

that the increase in rotor voltage due to a voltage sag follows (5.40).

In Fig. 5.10, the corresponding maximum rotor current needed due to the voltage sag for

Method I can be seen. Method IV is not shown in the ﬁgure, since it manages to keep the

rotor current unaffected during the voltage sag, with known parameters. It can be seen in the

ﬁgure that the maximum rotor current increases with the size of the voltage sag, especially

for low bandwidths of the current control loop. For higher bandwidths, it can be seen that

72

0

0.1

0.2

0.3

0.4

0

5

10

0

0.5

1

1.5

0

0.1

0.2

0.3

0.4

0

5

10

0

0.5

1

1.5

a)

ΔE

g

[p.u.] ΔE

g

[p.u.] α

c

[p.u.] α

c

[p.u.]

b)

v

m

a

x

R

[

p

.

u

.

]

v

m

a

x

R

[

p

.

u

.

]

Fig. 5.9. Maximum rotor voltage, v

max

R

, due to a symmetrical voltage sag as a function of the sag

size, ΔE

g

, and the current control bandwidth, α

c

. a) Method I (stator-ﬂux-oriented system).

b) Method IV (stator-ﬂux-oriented system).

0

0.1

0.2

0.3

0.4

0

5

10

0

1

2

3

4

ΔE

g

[p.u.] α

c

[p.u.]

i

m

a

x

R

[

p

.

u

.

]

Fig. 5.10. Maximum rotor current, i

max

R

, for Method I, due to a symmetrical voltage sag as a function

of the sag size, ΔE

g

, and the current control bandwidth, α

c

(stator-ﬂux-oriented system).

the maximum rotor current is practically constant, independent of the voltage sag magnitude.

The reason is that when the bandwidth is increased, the “need” for compensating the back

EMF vanishes, see (4.74).

It is, thus, not only necessary to design the converter according to the desired variable-

speed range, but also according to a certain voltage sag to withstand.

5.4.1 Inﬂuence of Erroneous Parameters

As mentioned earlier, the methods are mostly sensitive to an underestimated L

σ

, mainly since

the bandwidth of the current control loop then becomes lower than the desired. Simulations

with

˜

L

σ

= 0.5L

σ

shows that Method I is very sensitive to an underestimated L

σ

during

voltage sags, especially for low bandwidths of the current control loop, see Fig. 5.11. By

using Method IV, the inﬂuence of an erroneous value of L

σ

is, in principle, removed. If the

current control loop bandwidth is below 2 p.u., the difference in the maximum rotor current

73

0

0.1

0.2

0.3

0.4

0

5

10

−0.5

0

0.5

1

0

0.1

0.2

0.3

0.4

0

5

10

0

0.02

0.04

0.06

0.08

a)

ΔE

g

[p.u.] ΔE

g

[p.u.] α

c

[p.u.] α

c

[p.u.]

b)

Δ

i

m

a

x

R

[

p

.

u

.

]

Δ

i

m

a

x

R

[

p

.

u

.

]

Fig. 5.11. Increased maximum rotor current, Δi

max

R

, for Method I (a) and Method IV (b), when the

leakage inductance is underestimated,

˜

L

σ

= 0.5L

σ

, due to a symmetrical voltage sag as a

function of the sag size, ΔE

g

, and the current control bandwidth, α

c

(stator-ﬂux-oriented

system).

is below 0.02 p.u., as can be seen in Fig. 5.11.

For Method IV and variations in R

s

, R

R

, and L

M

with ±50%, the difference in maximum

rotor current is insigniﬁcant; while for Method I, R

s

and R

R

have small impacts for smaller

α

c

. However, for higher values of α

c

, this impact is also insigniﬁcant.

5.4.2 Generation Capability During Voltage Sags

As an example of this, Fig. 5.12 shows the minimum remaining grid voltage that can be

handled without triggering the crowbar as a function of the power. The maximum rotor

voltage is limited to 0.4 p.u. and the crowbar short circuits the rotor circuit when the rotor

current is above 1.25 p.u. This means that when the current controller needs to put out

a higher rotor voltage in order to compensate for the sag, it will lose control of the rotor

current, and the crowbar may be triggered if the rotor current becomes too high. From the

ﬁgure it can be seen that for low bandwidths of the current control loop (allowing a lower

switching frequency), Method IVmanages to survive deeper sags than Method I. However, as

indicated by the ﬁgure, for higher bandwidths, the difference between the methods vanishes.

A bandwidth of 7 p.u. for Method IV produces very similar results as a bandwidth of 1 p.u.,

and is therefore not shown in the ﬁgure.

5.5 Flux Damping

As previously mentioned there are different methods of damping the ﬂux oscillations. As

mentioned before, one method is to reduce the bandwidth of the current control loop [43]. In

[107], a feedback of the derivative of ﬂux was introduced in order to improve the damping

of the ﬂux. Another possibility is to use a converter to substitute the Y point of the stator

winding, i.e., an extra degree of freedom is introduced that can be used to actively damp out

the ﬂux oscillations, [54].

74

0 20 40 60 80 100

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

α

c

= 1,

˜

L

σ

= 0.5L

σ

α

c

= 1,

˜

L

σ

= 0

α

c

= 1,

˜

L

σ

= 0

α

c

= 7,

˜

L

σ

= 0

R

e

m

a

i

n

i

n

g

v

o

l

t

a

g

e

[

p

.

u

.

]

Power before voltage sag [%]

Fig. 5.12. Minimum remaining voltage without triggering the crowbar of a voltage sag, for Methods

I (solid) and IV (dashed) as a function of the power. The maximum rotor voltage is set

to 0.4 p.u. and a rotor current above 1.25 p.u. triggers the crowbar (stator-ﬂux-oriented

system).

Kelber made a comparison of different methods of damping the ﬂux oscillations in [55].

The methods are 1) reducing the bandwidth of the current control loop, 2) compensation of

the transformation angle (to synchronous coordinates), 3) feedback of the derivative of the

ﬂux, and 4) the method with a converter substituting the star point in the stator winding. It

is concluded in [55] that the method of reducing the bandwidth works quite well, although it

has the disadvantage of slowly damping of a grid disturbances. Compensation of the trans-

formation angle method improves the damping only slightly. Feedback of the ﬂux derivative

method performs well and has a low cost; the disadvantage of this method is that the method

cause relatively high rotor currents. The method with a converter in the star point of the stator

winding performs very well, but the disadvantage of this method is the required addition in

hardware and software. Since there is a need for another converter, the cost is also increased.

In this section, the ﬂux oscillations will be damped by feedback of the derivative of the

ﬂux. The reason that this method is chosen is that it has low cost (i.e., no extra hardware), is

easy to implement, and can damp the ﬂux oscillations well. Due to the fact that the method

with an extra converter connected to the Y point of the stator winding has to handle the sta-

tor current, implying an increase of the losses, and the increased cost for an extra converter

this method, is not considered in this section since some of the beneﬁts and reasons for the

doubly-fed induction generator, e.g., smaller (cheaper) converter and lower losses, vanishes.

However, later on in Chapter 7 where different methods for voltage sag ride-through are dis-

cussed and compared, the system with the converter in Y point becomes very interesting and

is accordingly further investigated.

The q component of the rotor current is used for controlling the torque, but the d compo-

nent of the current can be used to damp the oscillations and improve stability. If we add a

75

component Δi

ref

Rd

to the d component of the rotor current reference, which we control as

Δi

ref

Rd

(p) = −

p

p + α

co

α

d

ˆ

R

s

ψ

s

= −

1 −

α

co

p + α

co

α

d

ˆ

R

s

ψ

s

(5.41)

then, a ﬂux differentiation compensation term has been introduced, that will improve the

damping of the system. In the above equation, a high-pass ﬁlter is used since a pure differ-

entiation is not implementable. This means that i

ref

Rd

is set to

i

ref

Rd

= i

ref

Rd,0

+ Δi

ref

Rd

(5.42)

where i

ref

Rd,0

is used to control the reactive power as discussed in a previous chapter.

5.5.1 Stator-Flux Orientation

Under the assumption that the current dynamics are set much faster than the ﬂux dynamics

and α

co

is small, the characteristic polynomial in (5.6) can be rewritten as (with correctly

known parameters)

p

2

+

α

d

+

R

s

L

M

2 −

ω

g

L

M

i

ref

Rd,0

E

g

p +

1 −

R

s

i

ref

Rq

E

g

+ α

d

R

s

E

g

−i

ref

Rd,0

L

M

ω

g

E

g

L

M

ω

2

g

ω

2

g

.

(5.43)

With the inclusion of a ﬂux damping, the constraint on the d component becomes

i

Rd,0

<

2 + α

d

L

M

R

s

E

g

ω

g

L

M

(5.44)

in order to guarantee stability. Comparing to (5.7), it is seen that the constraint on the d

component rotor current has increased 1 + α

d

L

M

/(2R

s

) times.

5.5.2 Grid-Flux Orientation

For a grid-ﬂux-oriented system the characteristic polynomial in (5.27) is changed to (with

correctly known parameters)

p

2

+

α

d

+ 2

R

s

L

M

p +

α

d

R

s

L

M

+

R

2

s

L

2

M

+ ω

2

g

(5.45)

if α

co

is small. Moreover, since R

s

is small and L

M

is large, see Table 2.1 for typical

parameters, it is possible to approximate the above characteristic polynomial as

p

2

+ α

d

p + ω

2

g

. (5.46)

5.5.3 Parameter Selection

As can be seen in (5.41), the ﬂux damping uses two parameters, α

d

and α

co

, that have to be

determined. Obviously, the cut-off frequency, α

co

, of the low-pass ﬁlter must be set lower

than the oscillating frequency in order to be able to damp the oscillation at all. The damping

term, α

d

, must be chosen smaller than the bandwidth of the current control loop, α

c

, so that

the ﬂux damper becomes slower than the current dynamics. Of course, if a ﬂux estimator

is used to determine the ﬂux, the bandwidth of the damper, α

d

, must be smaller than the

bandwidth of the ﬂux estimator.

76

5.5.4 Evaluation

Fig. 5.13 shows a simulation of a vector-controlled doubly-fed induction generator, accord-

ing to Section 4.4.1 (k

E

= 1 and k

R

= 1), with and without ﬂux damping. The reference

frame is aligned with the stator ﬂux. The reference value i

ref

Rd,0

is initially zero and is at 0.4 s

changed to 0.5 p.u. The reference value of i

ref

Rq

is initially zero and is at 0.1 s changed to

0.5 p.u., and at 0.7 s to −0.5 p.u. The bandwidth of the system, α

c

, is set to 4.7 p.u., while

0 0.2 0.4 0.6 0.8 1

−0.5

0

0.5

1

1.5

0 0.2 0.4 0.6 0.8 1

−1

0

1

0 0.2 0.4 0.6 0.8 1

0.95

1

1.05

a)

C

u

r

r

e

n

t

[

p

.

u

.

]

C

u

r

r

e

n

t

[

p

.

u

.

]

b)

Time [s]

F

l

u

x

[

p

.

u

.

]

c)

Fig. 5.13. Simulation of current control using a stator-ﬂux oriented reference frame with (solid) and

without (dashed) damping of the ﬂux oscillations. a) i

Rd

. b) i

Rq

. c) ψ

s

.

α

d

is set to 0.7 p.u., and α

co

is set to 0.05 p.u. In the simulation it is assumed that the ﬂux

can be determined from measurements of the stator and the rotor currents. The ﬁgure shows

that the oscillations in the ﬂux has been damped with the ﬂux damper. Since it is difﬁcult to

see the effect of the ﬂux damper in a measured time series, due to noise, a frequency spectra

of the ﬂux magnitude has been plotted instead in Fig. 5.14. In the ﬁgure the current control

method with feed forward of the back EMF and with “active resistance” has been used, with

and without ﬂux damping. The frequency spectra is based on a 6 s long measurement on the

laboratory DFIG setup described in Appendix B.2. The DFIG is operated as in Section 5.3.

The bandwidth of the current control loop, α

c

, was set to 2.3 p.u., the damping term, α

d

, was

set to 0.7 p.u. and, the cut-off frequency term, α

co

, was set to 0.05 p.u. It can be seen in the

ﬁgure that the 50-Hz component has been to a large extent damped, i.e., a factor of ten, by

the ﬂux damper.

5.5.5 Response to Symmetrical Voltage Sags

In this section the ﬂux damper’s response will be analyzed with respect to symmetrical volt-

age sags. It is assumed that before and directly after the voltage sag, the magnitude of the

77

10

0

10

1

10

2

10

−5

10

−4

10

−3

10

−2

10

−1

10

0

10

1

10

2

10

−5

10

−4

10

−3

10

−2

10

−1 a)

Frequency [Hz] Frequency [Hz]

F

l

u

x

a

m

p

l

i

t

u

d

e

[

p

.

u

.

]

F

l

u

x

a

m

p

l

i

t

u

d

e

[

p

.

u

.

]

b)

Fig. 5.14. Frequency spectra of the ﬂux (data from measurements). The reference frame is aligned

with the stator ﬂux. a) Without ﬂux damping. b) With ﬂux damping.

stator ﬂux can be expressed as

t < 0 : ψ

s

(t) = ψ

s0

(5.47)

t ≥ 0 : ψ

s

(t) ≈ ψ

s0

V

E

g,nom

+

1 −

V

E

g,nom

ψ

s0

e

−α

d

t/2

cos(ω

g

t) (5.48)

where ψ

s0

is the steady-state stator ﬂux prior the voltage sag and V is the remaining voltage

after the voltage sag. This means that the term 1 − V/E

g,nom

corresponds to the magnitude

of the sag. Then, from (5.41) the response in Δi

Rd

is estimated as

Δi

ref

Rd

(t) = L

−1

−

1 −

α

co

p + α

co

α

d

R

s

L ¦ψ

s

(t)¦

(5.49)

or as

Δi

ref

Rd

(t) = L

−1

−

α

d

R

s

L ¦ψ

s

(t)¦ −

α

co

p + α

co

L ¦ψ

s

(t)¦

. (5.50)

If α

co

is considered small, i.e., the low-pass ﬁlter α

co

/(p + α

co

) has low bandwidth, it is

possible to describe Δi

ref

Rd

(t) after the voltage sag as

t ≥ 0 : Δi

ref

Rd

(t) ≈ −

α

d

R

s

Δψ

s

(t) = −

α

d

R

s

(ψ

s

(t) −ψ

s0

) (5.51)

which can be written as

t ≥ 0 : Δi

ref

Rd

(t) ≈

α

d

R

s

1 −

V

E

g,nom

ψ

s0

1 −e

−α

d

t/2

cos(ω

g

t)

. (5.52)

The above expression has a local maximum for t = arccos

−2ω

g

/

α

2

d

+ 4ω

2

g

/ω

g

. How-

ever, if α

2

d

<4ω

2

g

it is possible to approximate t as t ≈ arccos(−1)/ω

g

= π/ω

g

. This means

that the extreme value of Δi

ref

Rd

(t) due to a symmetrical voltage sag can be expressed as

Δi

ref

Rd

(t = π/ω

g

) ≈

α

d

R

s

ψ

s0

1 + e

−α

d

π/(2ω

g

)

1 −

V

E

g,nom

. (5.53)

78

Consider the following values: V = 0.9 p.u., α

d

= 0.7 p.u., ψ

s0

= 1 p.u., and R

s

= 0.01 p.u.

for a numerical example. This means that the maximum value of Δi

ref

Rd

due to the voltage

sag is Δi

ref

Rd

= 0.7/0.01 1

1 + e

−0.7π/(2·1)

**(1 −0.9/1) = 9.3 p.u. This value is, of course,
**

an unrealistically high value. However, it indicates that the ﬂux damper is very sensitive to

voltage sags. In Fig. 5.15 the maximum value of Δi

ref

Rd

due to a voltage sag as a function

of the bandwidth α

d

of the ﬂux damper can be seen. The results are presented for four

different symmetrical voltage sags between V = 0.8 to V = 0.95 p.u. (note that V is the

remaining voltage). In the ﬁgure both simulated results (using stator-ﬂux orientation) as well

as analytically results from (5.53) is shown. Both methods produce similar results, although

the analytical results are generally slightly higher. The results in the ﬁgure shows that the

ﬂux damper is very sensitive to voltage sags. This means that if the ﬂux damper should work

during (small) voltage sags, the bandwidth, α

d

, of the ﬂux damper should be small. However,

then some of the advantage of the ﬂux damper is lost.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

0

5

10

15

20

M

a

x

.

Δ

i

r

e

f

R

d

[

p

.

u

.

]

α

d

[p.u.]

V = 0.95

V = 0.9

V = 0.85

V = 0.8

Fig. 5.15. Maximum of Δi

ref

Rd

due to a voltage sag as a function of α

d

. Solid lines correspond to

simulation and dashed lines correspond to results from an analytical expression.

5.6 Conclusion

In this chapter, the general rotor current control law derived in Chapter 4, with the option of

including feed-forward compensation of the back EMF and “active resistance,” in order to

eliminate the inﬂuence of the back EMF on the rotor current, has been analyzed. It was found

that the method that combines both the feed-forward compensation of the back EMF and the

“active resistance” manages best to suppress the inﬂuence of the back EMF on the rotor cur-

rent. Moreover, this method was found to be the least sensitive one to erroneous parameters

and it manages to keep the rotor current close to unaffected, even with erroneous parameters,

during a voltage sag. The choice of current control method is of greater importance if the

bandwidth of the current control loop is low.

It has been shown that by using grid-ﬂux orientation, the stability and the damping of

the system is independent of the rotor current, in contrast to the stator-ﬂux-oriented system.

79

This implies that for a grid-ﬂux-oriented system it is possible to magnetize the DFIG entirely

from the rotor circuit without reducing the damping of the system.

By utilizing feed-forward compensation, stability of the system resulting from the pro-

posed current controller was found independent of the bandwidth of the current control loop,

and the order of the system to analyze could be reduced. The introduction of an “active

resistance” in the current control law improves the damping of low-frequency disturbances

signiﬁcantly.

Finally, it is shown that the design of the converter for a doubly-fed induction generator

should also take into account a certain voltage sag to withstand and not only the desired

variable-speed range.

80

Chapter 6

Evaluation of Doubly-Fed Induction

Generator Systems

6.1 Reduced-Order Model

If, for example, the rotor current dynamics and the grid-ﬁlter current dynamics are controlled

by a high-gain feedback, it is possible to force the system to have both slow and fast time

scales, i.e., the system behaves like a singularly perturbed system [57]. This means, that the

rotor and grid-ﬁlter current can be assumed to follow their reference values accurately.

As pointed out in the Introduction, the ﬂux dynamics of the DFIG are strongly inﬂuenced

by a pair of poorly damped poles, with an oscillating frequency close to 1 p.u., i.e. close to

the line frequency. If the current control loop is much faster than the ﬂux dynamics, it is

sufﬁcient to study only the ﬂux dynamics and put the rotor current to its reference value, i.e.,

dΨ

s

dt

= E

g

−

R

s

L

M

+ jω

1

Ψ

s

+ R

s

i

ref

R

(6.1)

where the stator voltage has been put equal to the grid voltage. For a stator-ﬂux-oriented sys-

tem the above equation can be reduced to (5.2) and (5.3), where the equation is in polar form.

While for a grid-ﬂux oriented system the above equation can be used directly. However, the

synchronous frequency, ω

1

, must be determined. Either a PLL-estimator, as described in

Section 4.1.3, can be used to track the frequency of the grid voltage, or, if the frequency of

the grid is constant (or at least close to constant), the synchronous frequency can be put equal

to the grid frequency, i.e. ω

1

= ω

g

.

6.2 Discretization of the Doubly-Fed Induction Generator

If the simple-to-use forward Euler method, see Section 4.1.7, is used to simulate the sys-

tem, care must be taken not to use a too long time step or sampling period, T

sample

. For

instance, in PSCAD/EMTDC [66], when writing user-deﬁned modules, the module must be

discretizised, and this often due to its simplicity results in using the forward Euler method.

The forward Euler discretization is given by (4.35). As mentioned in Section 4.1.7, the

poles must be inside a circle with a radius of 1/T

sample

and the center point located at

(−1/T

sample

, 0) in order for the forward Euler discretization to be stable.

81

It should be pointed out that in some other programs, for instance Simpow [1] and PSS/E

[85], user-deﬁned modules return expressions for the derivatives and advanced integration

algorithms are used. In this case, the allowed time step can be made longer.

6.2.1 Stator-Flux Orientation

The solution to the characteristic polynomial for a stator-ﬂux oriented system in (5.6) is

found as

p

1,2

= −

R

s

2L

M

2 −

ω

g

L

M

i

ref

Rd

E

g

±

R

2

s

4L

2

M

2 −

ω

g

L

M

i

ref

Rd

E

g

2

−

1 −

R

s

i

ref

Rq

E

g

ω

2

g

≈ −

R

s

2L

M

2 −

ω

g

L

M

i

ref

Rd

E

g

±jω

g

. (6.2)

In order for the discretization to be stable, the above-mentioned poles should be located

inside the circle, i.e.,

−

1

T

sample

, 0

−

−

R

s

2L

M

2 −

ω

g

L

M

i

ref

Rd

E

g

, ±ω

g

<

1

T

sample

(6.3)

which yields

T

sample

<

R

s

L

M

2 −

ω

g

L

M

i

ref

Rd

E

g

R

2

s

L

2

M

2 −

ω

g

L

M

i

ref

Rd

E

g

2

4

+ ω

2

g

<

R

s

ω

2

g

L

M

2 −

ω

g

L

M

i

ref

Rd

E

g

. (6.4)

For unity power factor, i.e. i

ref

Rd

= ψ

s

/L

M

≈ E

g

/(ω

1

L

M

), the above expression is reduced to

T

sample

<

R

s

ω

2

g

L

M

. (6.5)

For the system investigated later on in this chapter and using the forward Euler method, the

sampling period should be T

s

< 4.6 µs.

6.2.2 Grid-Flux Orientation

The solution to the characteristic polynomial in (5.27), corresponding to grid-ﬂux-oriented

system, is found as

p

1,2

=

R

s

L

M

±jω

g

(6.6)

In order for the discretization to be stable, the above-mentioned poles should be located

inside the circle, i.e.,

−

1

T

sample

, 0

−

−

R

s

L

M

, ±ω

g

<

1

T

sample

. (6.7)

82

The solution to the above equation becomes

T

sample

<

2R

s

L

M

1

R

2

s

L

2

M

+ ω

2

g

≈

2R

s

L

M

ω

2

g

(6.8)

which is twice the value obtained by (6.5). Moreover, the minimum sample time for the

grid-ﬂux-oriented system is independent, in contrast to a stator-ﬂux-oriented system, of the

d component of the rotor current.

6.3 Response to Grid Disturbances

In this section, simulations and experimental results of the response of a DFIG wind turbine

to voltage sags are presented. The experiments were made on a VESTAS V-52 850 kW

WT and in Appendix B.3 a short description of the used data acquisition setup is presented.

Moreover, the simulations presented are carried out on a ﬁctitious 850-kW DFIG WT. The

following parameters were used in the simulations: R

s

= 0.0071 p.u., R

R

= 0.01 p.u.,

L

M

= 4.9 p.u., and L

σ

= 0.21 p.u. The grid ﬁlter for the grid-side converter R

i

= 0.01 p.u.

and L

i

= 0.07 p.u. The dc-link capacitance is set to C

dc

= 2.8 p.u. The simulations have

been carried out both with a “full-order” model and a second-order model.

Fig. 6.1 shows experimental results of the response of a DFIG wind turbine to a voltage

sag. The voltage drops down approximately 25%, i.e., a 75% sag, at t=0.1 s, and after 0.1 s

the fault causing the voltage sag on the grid is cleared, and the voltage starts to recover. The

wind turbine produces about 20% of the nominal power. The oscillation close to 50 Hz,

caused by the poorly damped poles due to the voltage sag, is clearly seen. In Fig. 6.2, a

simulation of the response to the same voltage sag, as shown in Fig. 6.1, is presented, for

the full-order model. Fig. 6.3 shows the corresponding simulation with the reduced-order

model of the system. It can be seen in the ﬁgure that the full-order model and the reduced-

order model produce almost the same results. One reason for this is that the bandwidth of

the current control loops (of the machine and grid-side converter) are set to 7 p.u., which is

sufﬁciently higher than the eigenfrequency of the ﬂux dynamics (close to 1 p.u.), shown in

Section 5.1. Comparing the two ﬁgures it is seen that the agreement between the experiment

and simulation is quite satisfactory. An exact agreement is not to be expected, since real

machine parameters were unknown.

In Fig. 6.4, experimental results of the response due to an unsymmetrical voltage sag are

presented. The WT now produces approximately 10% of its nominal power. Fig. 6.5 shows

a simulation of the response to the same voltage sag as in Fig. 6.4, for the full-order model

and Fig. 6.6 shows the corresponding simulation for the reduced-order model. Again, it is

seen that the agreement is quite satisfactory.

In Fig. 6.7, a severe voltage disturbance is presented. In this case the disturbance is so

large that the over voltage protection short-circuits the rotor and, after 40 ms, the breaker dis-

connects the stator from the grid. Before the disturbance the WT is producing approximately

half of its rated power.

As mentioned earlier, the simulations shown in this section are carried out for a stator-

ﬂux-oriented system. However, similar results from simulations can also be found from a

83

0 0.1 0.2 0.3 0.4

0.7

0.8

0.9

1

1.1

0 0.1 0.2 0.3 0.4

0

0.1

0.2

0.3

0.4

0 0.1 0.2 0.3 0.4

0

10

20

30

0 0.1 0.2 0.3 0.4

−10

0

10

V

o

l

t

a

g

e

[

p

.

u

.

]

a)

C

u

r

r

e

n

t

[

p

.

u

.

]

b)

P

o

w

e

r

[

%

]

c)

Time [s] Time [s]

Time [s] Time [s]

R

e

a

c

t

i

v

e

p

o

w

e

r

[

%

]

d)

Fig. 6.1. Experiment of the response to a voltage sag. a) Grid-voltage magnitude. b) Grid-current

magnitude. c) Active power. d) Reactive power.

0 0.1 0.2 0.3 0.4

0.7

0.8

0.9

1

1.1

0 0.1 0.2 0.3 0.4

0

0.1

0.2

0.3

0.4

0 0.1 0.2 0.3 0.4

0

10

20

30

0 0.1 0.2 0.3 0.4

−10

0

10

V

o

l

t

a

g

e

[

p

.

u

.

]

a)

C

u

r

r

e

n

t

[

p

.

u

.

]

b)

P

o

w

e

r

[

%

]

c)

Time [s] Time [s]

Time [s] Time [s]

R

e

a

c

t

i

v

e

p

o

w

e

r

[

%

]

d)

Fig. 6.2. Simulation of the response to a voltage sag with the full-order model. a) Grid-voltage

magnitude. b) Grid-current magnitude. c) Active power. d) Reactive power.

stator-voltage oriented (or grid-ﬂux oriented) system.

84

0 0.1 0.2 0.3 0.4

0.7

0.8

0.9

1

1.1

0 0.1 0.2 0.3 0.4

0

0.1

0.2

0.3

0.4

0 0.1 0.2 0.3 0.4

0

10

20

30

0 0.1 0.2 0.3 0.4

−10

0

10

V

o

l

t

a

g

e

[

p

.

u

.

]

a)

C

u

r

r

e

n

t

[

p

.

u

.

]

b)

P

o

w

e

r

[

%

]

c)

Time [s] Time [s]

Time [s] Time [s]

R

e

a

c

t

i

v

e

p

o

w

e

r

[

%

]

d)

Fig. 6.3. Simulation of the response to a voltage sag with the reduced-order model. a) Grid-voltage

magnitude. b) Grid-current magnitude. c) Active power. d) Reactive power.

0 0.1 0.2 0.3 0.4

0.7

0.8

0.9

1

1.1

0 0.1 0.2 0.3 0.4

0

0.05

0.1

0.15

0.2

0.25

0 0.1 0.2 0.3 0.4

−10

0

10

20

30

0 0.1 0.2 0.3 0.4

−10

−5

0

5

10

V

o

l

t

a

g

e

[

p

.

u

.

]

a)

C

u

r

r

e

n

t

[

p

.

u

.

]

b)

P

o

w

e

r

[

%

]

c)

Time [s] Time [s]

Time [s] Time [s]

R

e

a

c

t

i

v

e

p

o

w

e

r

[

%

]

d)

Fig. 6.4. Experiment of the response to a unsymmetrical voltage sag. a) Grid-voltage magnitude.

b) Grid-current magnitude. c) Active power. d) Reactive power.

85

0 0.1 0.2 0.3 0.4

0.7

0.8

0.9

1

1.1

0 0.1 0.2 0.3 0.4

0

0.05

0.1

0.15

0.2

0.25

0 0.1 0.2 0.3 0.4

−10

0

10

20

30

0 0.1 0.2 0.3 0.4

−10

−5

0

5

10

V

o

l

t

a

g

e

[

p

.

u

.

]

a)

C

u

r

r

e

n

t

[

p

.

u

.

]

b)

P

o

w

e

r

[

%

]

c)

Time [s] Time [s]

Time [s] Time [s]

R

e

a

c

t

i

v

e

p

o

w

e

r

[

%

]

d)

Fig. 6.5. Simulation of the response to an unsymmetrical voltage sag. The simulation has been per-

formed with the full-order model. a) Grid-voltage magnitude. b) Grid-current magnitude.

c) Active power. d) Reactive power.

0 0.1 0.2 0.3 0.4

0.7

0.8

0.9

1

1.1

0 0.1 0.2 0.3 0.4

0

0.05

0.1

0.15

0.2

0.25

0 0.1 0.2 0.3 0.4

−10

0

10

20

30

0 0.1 0.2 0.3 0.4

−10

−5

0

5

10

V

o

l

t

a

g

e

[

p

.

u

.

]

a)

C

u

r

r

e

n

t

[

p

.

u

.

]

b)

P

o

w

e

r

[

%

]

c)

Time [s] Time [s]

Time [s] Time [s]

R

e

a

c

t

i

v

e

p

o

w

e

r

[

%

]

d)

Fig. 6.6. Simulation of the response to an unsymmetrical voltage sag. The simulation has been per-

formed with the reduced-order model. a) Grid-voltage magnitude. b) Grid-current magni-

tude. c) Active power. d) Reactive power.

86

0 0.02 0.04 0.06 0.08 0.1

0

0.5

1

1.5

0 0.02 0.04 0.06 0.08 0.1

0

2

4

0 0.02 0.04 0.06 0.08 0.1

0

2

4

V

o

l

t

a

g

e

[

p

.

u

.

]

a)

C

u

r

r

e

n

t

(

d

)

[

p

.

u

.

]

b)

Time [s]

C

u

r

r

e

n

t

(

q

)

[

p

.

u

.

]

c)

Fig. 6.7. Severe voltage disturbance. a) Grid voltage. b) d component of the grid current. c) q

component of the grid current.

6.4 Implementation in Grid Simulation Programs

Some grid simulation programs can handle three-phase instantaneous quantities. Examples

are EMTDC and Simpow. Other programs are designed to handle the voltages as phasors,

and for these programs, 50-Hz oscillations in the output quantities cannot be captured, since

the time step is often too large for these oscillations; an example is PSS/E. However, when

handling simulations of large systems, it may not be possible to use such a short time step

(about 5 µs) as is required in order to simulate the control of the DFIGsystem. The suggested

approach is to simply ignore the 50-Hz oscillations when the DFIG system is implemented

in simulations with long time steps, as long as the disturbances are small enough not to cause

the rotor to be short-circuited. For this case, a steady-state model of the DFIG is sufﬁcient.

However, if a disturbance is large enough to cause the rotor to be short-circuited, the machine

will act as a standard squirrel-cage induction machine which can be adequately modeled with

a ﬁfth-order model of the induction machine [83].

As pointed out in [84, 60], the stator ﬂux transients may be negligible from the power

system stability analysis point of view. This means that if stator ﬂux transients are negligible

a steady-state model of the DFIG dynamics are sufﬁcient as long as the rotor circuit is not

short-circuited due to a too large grid disturbance.

87

6.5 Summary

In this chapter, simulations and experimental veriﬁcation of the dynamic response to voltage

sags of a DFIG wind turbine were presented. Simulations were carried out using a full-

order model and a reduced-order model. Both models produced acceptable results. Perfect

correspondence with experiments were not expected since the simulations were carried out

on a ﬁctitious DFIG wind turbine. The response to symmetrical as well as unsymmetrical

voltage sags was veriﬁed.

88

Chapter 7

Voltage Sag Ride-Through of

Variable-Speed Wind Turbines

As mentioned in the Introduction, new grid codes are in progress both in Sweden and other

countries. This means that new wind turbine installations have to stay connected to the grid

for voltage sags above a certain reference sag, i.e., WTs have to ride through these voltage

sags. In Fig. 7.1, the proposed Swedish requirements for voltage sags is depicted.

−0.2 0 0.2 0.4 0.6 0.8 1 1.2

0

0.2

0.4

0.6

0.8

1

1.2

V

o

l

t

a

g

e

[

p

.

u

.

]

Time [s]

Fig. 7.1. Proposed regulations from the Swedish national grid company, Svenska Kraftn¨ at [96]. Solid

line is the requirement for wind parks with a rated power larger than 100 MW. Dashed line

is the requirement for wind turbines and wind parks with a rated power between 0.3–100

MW.

First, simple space vector models will be presented for some common voltage sags that

will be used in this chapter. Then, the voltage sag response of a WT that utilizes a full-power

converter is investigated. This investigation will the serve as a basis for the comparison of

DFIG ride-through systems. In the next sections the voltage sag response of the DFIG will

be further analyzed, and systems for voltage sag ride-through will be investigated. Finally,

these systems will be compared dynamically as well as for steady-state operation.

89

7.1 Voltage Sags

With the expression “voltage sag,” it is normally implied that the grid rms voltage drops from

1 p.u. to 0.1–0.9 p.u. for a short period of time, i.e., 0.5–30 cycles. The duration of voltage

sags is mainly determined by the clearing time of the protection used in the grid [9]. The

fault clearing time for protective relays varies from 50 ms up to 2000 ms [9]. There are other

protection devices, e.g., current-limiting fuses, that might have a shorter fault clearing time

(less than one cycle). Voltage sags caused by these fuses are short and deep if the fault is in

the local distribution network but if the fault is in a remote distribution network the sag is

short and shallow [9]. The origin and classiﬁcation of voltage sags are well explained in [9].

In this section, simple space vector models will be presented for some common voltage sags.

These models are developed in [74] and the aim of the models are to estimate the moduli of

the positive- and negative-sequence voltage vectors for different types of sags.

7.1.1 Symmetrical Voltage Sags

Symmetrical (or balanced) voltage sags implies an equally reduction of the rms voltage and,

possibly, a “phase-angle jump” in all three phases [9]. Directly after a symmetrical voltage

sag, the grid voltage vector can be expressed in the synchronous reference frame as

E

g

(t = 0

+

) = jV e

j

˜

θ

0

= jV e

jφ

(7.1)

where V is the remaining rms voltage in the faulted phases,

˜

θ

0

is the initial error angle, and

φ is the “phase-angle jump.” The majority of all “phase-angle jumps” are smaller than 45

◦

[9], and the remaining rms voltage can be as low as V = 0 for a direct-to-ground fault.

7.1.2 Unsymmetrical Voltage Sags

Unsymmetrical (or unbalanced) voltage sags are more difﬁcult to model since, for instance,

the impedance of each symmetrical component can be hard to derive. However, in order

to simplify the derivation of models suitable for unsymmetrical voltage sags, the positive-,

negative-, and zero-sequence impedance are assumed to be equal. For ground faults, it is

assumed that the source and feeder impedance are much larger compared to the line-to-

ground impedance. The impedance between the two faulted lines for a line-to-line fault is

neglected. Zero sequences are not critical for a PWM rectiﬁer since such sequences ideally

disappear from the phase currents due to the absence of a neutral conductor.

Single-Line-to-Ground Fault

After a single-line-to-ground fault (SLGF) in the ﬁrst phase the grid phase voltages can be

expressed as

E

1

(t = 0

+

) =

√

2V cos(θ

g

+ π/2 + φ) (7.2)

E

2

(t = 0

+

) =

√

2E

g,nom

cos(θ

g

+ π/2 −2π/3) (7.3)

E

3

(t = 0

+

) =

√

2E

g,nom

cos(θ

g

+ π/2 + 2π/3) (7.4)

90

where E

1

, E

2

, and E

3

are the grid phase voltages directly after the sag, and V and φ are the

remaining rms voltage and “phase-angle jump” in the ﬁrst phase, respectively. The space

vector in a stationary reference frame that corresponds to (7.2)–(7.4) is then found as

E

s

g0

= j

E

p0

e

jθ

g

+E

n0

e

−jθ

g

(7.5)

where

E

p0

=

1

3

2E

nom

+ V e

jφ

, E

n0

=

1

3

E

nom

−V e

−jφ

(7.6)

are the stationary parts of the positive- and negative-sequence voltage vectors, respectively.

For perfect pre-sag ﬁeld orientation, i.e., θ

1

= θ

g

, (7.5) can be transformed to the synchro-

nous reference frame by substituting E

s

g0

= E

g0

e

jθ

g

and solving the resulting equation for

E

g0

:

E

g0

= j

E

p0

+E

n0

e

−j2θ

g

. (7.7)

As expected, the negative sequence becomes in the synchronous reference frame a compo-

nent with a frequency of twice the fundamental frequency, i.e. −2ω

g

. From (7.6), it is seen

that minimal modulus of the positive-sequence voltage vector is E

p

= 2E

g,nom

/3 and that

the maximum negative-sequence voltage vector is E

n

= E

g,nom

/3. This occurs when V = 0,

i.e., a total loss of voltage in the faulted phase.

The initial error angle of the positive-sequence voltage vector due to a SLGF is

˜

θ

0

= arg(E

p0

) = arctan

V sin φ

2E

g,nom

+ V cos φ

. (7.8)

Eventually, the PLL will track the position of the positive-sequence voltage vector, such

that, ideally, E

p

becomes real valued and, hence,

˜

θ ≈ 0. Consider the following values for a

numerical example: E

g,nom

= 1 p.u., V = 0.5 p.u. and φ = −45

◦

. This gives an initial error

angle of

˜

θ = arctan[−0.5 0.71/(2+0.5 0.71)] ≈ −0.15 rad ≈ −9

◦

. The initial error angle

becomes even smaller if V is smaller than 0.5 p.u.;

˜

θ

0

= 0 for V = 0, for instance.

Two-Lines-to-Ground Fault

After a two-lines-to-ground fault (TLGF) between the ﬁrst and second phase, the grid phase

voltages can be expressed as

E

1

(t = 0

+

) =

√

2V cos(θ

g

+ π/2 + φ) (7.9)

E

2

(t = 0

+

) =

√

2V cos(θ

g

+ π/2 −2π/3 + φ) (7.10)

E

3

(t = 0

+

) =

√

2E

g,nom

cos(θ

g

+ π/2 + 2π/3) (7.11)

which correspond to the following space vector in the synchronous reference frame:

E

g0

= j

E

p0

+E

n0

e

−j2θ

g

(7.12)

where

E

p0

=

E

g,nom

3

+

2

3

V e

jφ

, E

n0

=

E

∗

p0

−V e

−jφ

e

−jπ/3

. (7.13)

91

From (7.12) and (7.13), it can be seen that the minimal modulus for the positive-sequence

voltage vector is E

p

= E

g,nom

/3 and that the maximal modulus for the negative-sequence

voltage vector is E

n

= E

g,nom

/3 for V = 0.

The initial error angle directly after a TLGF is

˜

θ

0

= arg(E

p0

) = arctan

2V sin φ

E

g,nom

+ 2V cos φ

. (7.14)

An initial error angle of −23

◦

is obtained as a numerical example using the same values as

in the previous section.

Line-to-Line Fault

Directly after a worst-case (no feeder impedance) line-to-line fault (LLF) between phases 2

and 3, the grid phase voltages are found as

E

1

=

√

2E

g,nom

cos(θ

g

+ π/2) (7.15)

E

2

=

E

g,nom

√

2

cos(θ

g

+ π/2 −π) (7.16)

E

3

=

E

g,nom

√

2

cos(θ

g

+ π/2 −π) (7.17)

which correspond to the following space vector in the synchronous reference frame:

E

g0

= j

1 −e

−j2θ

g

E

g,nom

2

. (7.18)

Obviously, the modulus of the positive- and negative-sequence voltage vectors both equal

E

g,nom

/2, and the initial error angle moments after the LLF equals zero.

7.2 Full-Power Converter

In this section, the voltage sag response of PWM rectiﬁers, designed for the rated WT power,

is analyzed. The system conﬁguration consists of a generator and two converters connected

“back-to-back” as depicted in Fig. 7.2. The main focus of this section is put on the PWM

rectiﬁer and the achieved results and conclusions are independent of the type of converter

at the generator side. As a result of the analysis, accurate estimates of the transient and

steady-state response of the grid current and dc-link voltage during voltage sags are provided.

These results can be useful when designing a PWM rectiﬁer for various grid codes and

requirements.

7.2.1 Analysis

First, the dc-link voltage dynamics are analyzed for various disturbances and voltage sags.

The dc-link voltage controller presented in Section 4.5.2 will be considered, with the ex-

ception of P

r

= −P

t

. Note that for this case, the grid-ﬁlter current equals the grid current.

This exception indicates that the rotor power, P

r

, used for the DFIG is changed to the total

turbine power, P

t

, for the full-power converter analyzed in this section. It is assumed that

92

SG Grid

Grid ﬁlter

PWM “rectiﬁer”

L

g

C

dc

Fig. 7.2. Wind turbine with a full power rectiﬁer.

the dc-link capacitance is accurately modeled, i.e.,

ˆ

C

dc

= C

dc

. The transfer function from

the turbine power P

t

to the error signal, e

w

= W

ref

dc

− W

dc

, will be considered which, with

(4.120), becomes

G

Pe

(p) = −G

PW

(p) =

−2p

C

dc

p

2

+ 2α

w

ξp + α

2

w

ξ

. (7.19)

Since the dc-link dynamics are considered to be much slower than the switching and sam-

pling frequency, f

sw

, of the PWM rectiﬁer, the grid-ﬁlter current dynamics and the switching

transients at the dc link are, thus, neglected. For instance, if the bandwidth of the dc-link volt-

age control loop is α

w

= 0.2 p.u. and the switching and sampling frequency is f

sw

= 4.9

kHz. Then, for a 50-Hz grid, α

w

is 4900/(0.2 50) = 490 times smaller than f

sw

. Moreover,

steady-state condition, symmetrical and nominal grid voltage, and perfect ﬁeld orientation

are assumed to precede the different disturbances.

Minimal DC-link Capacitance

In PWM rectiﬁers, the current in the dc-link capacitors is heavily distorted which gives rise

to a small (compared to diode rectiﬁers) ripple in the dc-link voltage. To ensure that this

voltage ripple remains below a tolerable value, the dc-link capacitance should be selected no

smaller than [59]

C

dc,min

=

√

3i

nom

fq

8f

sw

˜ v

p-p

dc

(7.20)

where i

nom

fq

= 1 p.u. is the nominal q-axis current and ˜ v

p-p

dc

is the tolerable peak-to-peak ripple

for the dc-link voltage. The value ˜ v

p-p

dc

= 0.028 pu, which corresponds to 1 % peak-to-peak

ripple at v

ref

dc

= 2.8 and f

sw

= 4.9 kHz, are considered for a numerical example. For a base

frequency of ω

b

= 314 rad/s, these values yield C

dc,min

=

√

3 1/(8 4900/314 0.028) ≈

0.5 p.u. However, very high demands [50] are placed on the dc-link voltage control loop

when using such small a dc-link capacitance, so (7.20) is mainly a benchmark that can be

used for comparison to more realistic operating conditions. Henceforth, a dc-link capacitance

of C

dc

= 3.5 p.u. is considered, which equals the capacitance of the experimental setup in

Section 7.2.3.

93

Assessment of Turbine Power Reduction

The grid-voltage modulus is normally close to its nominal value. Therefore, it is natural to

let E

gq

= E

g,nom

when analyzing the capability of the dc-link voltage control loop to reject

disturbances in P

t

. Therefore, (7.19) is reduced to

G

nom

Pe

(p) = G

Pe

(p)

E

gq

=E

g,nom

=

−2p

C

dc

(p + α

w

)

2

. (7.21)

For a step in the turbine power, from P

t

(0

−

) = 0 to P

t

(0

+

) = ΔP, the error, e

w

(t), becomes

e

w

(t) = L

−1

G

nom

Pe

(p)

ΔP

p

= −

2ΔP

C

dc

e

−α

w

t

t. (7.22)

Depending on whether ΔP is positive or negative, (7.22) has a local minimum or maximum

for t = 1/α

w

(determined by solving ˙ e

w

(t) = 0). Then, the maximum/minimum value of

e

w

(t) is

e

max/min

w

(t = 1/α

w

) = −

2ΔP

α

w

C

dc

e

−1

≈ −

0.74ΔP

α

w

C

dc

. (7.23)

The values ΔP = −1.5 p.u. (50 % of nominal power), α

w

= 0.2 p.u. and C

dc

= 3.5 p.u.

are considered for a numerical example which yield a local maximum for e

w

(t) at e

max

w

=

0.74 1.5/(0.2 3.5) ≈ 1.6 p.u. With v

ref

dc

= 2.8 p.u., this corresponds to a minimum dc-link

voltage of v

min

dc

= ((v

ref

dc

)

2

−e

max

w

)

0.5

= (2.8

2

−1.6)

0.5

= 2.5 p.u.

Response to Symmetrical Voltage Sags

As mentioned earlier, it is assumed that symmetrical voltage sags are preceded by symmetri-

cal and nominal grid voltage, perfect ﬁeld orientation and steady-state condition, i.e.,

˙

W = 0.

This implies that:

t < 0 : 0 = −3E

g,nom

i

fq

+ P

t0

(7.24)

where P

t0

is the pre-sag turbine power. Moments after a symmetrical voltage sag occurs,

it can be assumed that E

gq

= V while P

t

remains at its pre-sag value. These assumptions

imply the following dynamics for W:

t ≥ 0 :

1

2

C

dW

dt

= −3V i

fq

+ P

t0

(7.25)

after a sag at t = 0. Since the power to the grid ﬁlter is P

f

= 3E

gq

(t)i

fq

(v

dc

), where

E

gq

(t) changes stepwise at t = 0 and i

fq

(v

dc

) is a function of the dc-link voltage (via the

v

dc

control loop), the dynamics in (7.25) appear to be time-varying. However, this is not the

case, though, which can be deduced by multiplying (7.24) by V/E

g,nom

:

t < 0 : 0 = −3V i

fq

+ P

t0

V

E

g,nom

. (7.26)

Then by introducing the “new” turbine power, P

t

(t) = P

t0

V/E

g,nom

, it follows from (7.25)

and (7.26) that a symmetrical voltage sag is equivalent to a positive step in P

t

, which changes

from P

t

(0

−

) = P

g0

V/E

g,nom

to P

t

(0

+

) = P

t0

. This means that the net power step is ΔP

=

(1 − V/E

g,nom

)P

t0

. Meanwhile, the q-axis grid voltage can considered to be constant at

94

E

gq

= V provided that accurate ﬁeld orientation is maintained. The single exception to

this power step equivalence is when V = 0, which corresponds to that power cannot be

transferred to the utility grid.

Once the equivalence to turbine power steps has been revealed, the dynamics of e

w

during

symmetrical voltage sags are, hence, given by (7.19). By substituting E

gq

= V , the poles of

this transfer function are

p

1,2

= −

α

w

E

g,nom

V ±j

E

g,nom

V −V

2

. (7.27)

For normal operation, i.e., V = E

g,nom

, the poles are located at −α

w

, as seen in (7.21).

Moreover, the poles of (7.27) are well damped for V ≥ E

g,nom

/2. More troublesome how-

ever, is that symmetrical voltage sags may require very large i

fq

in order to counteract the

reduction in the grid rms voltage such that P

f

= P

g

in the steady state. Consider P

t

= 3 p.u.

(nominal power) and V = 0.1 p.u., for instance, which demands for i

fq

= 3/(3 0.1) = 10

p.u. in order to regain steady-state conditions during a sag. Remedies for avoiding severe

overcurrents during symmetrical voltage sags are discussed in Section 7.2.2.

Provided that overcurrent is avoided, e

w

(t), after a symmetrical voltage sag, is obtained

from the inverse Laplace transform of (7.19) multiplied by the step ΔP

/p:

e

w

(t) = L

−1

G

Pe

(p)

ΔP

p

= −

2ΔP

ω

w

C

e

−α

w

ξt

sin (ω

w

t) (7.28)

where ω

w

= α

w

(1 −ξ)ξ. Depending on the sign of ΔP

**, (7.28) has a local minimum or
**

maximum for t = arcsin(

√

1 −ξ)/ω

w

. By substituting this instant in (7.28), the extreme

value for e

w

(t) is obtained as

e

w

= −

2ΔP

ω

w

C

exp

−

ξ

1 −ξ

arcsin

1 −ξ

1 −ξ. (7.29)

The values C = 3.5 p.u., α

w

= 0.2 p.u., P

t0

= −1.5 p.u., V = 0.6 p.u. are considered

for a numerical example. This means that ΔP

= −1.5 0.4 = −0.6 p.u., ξ = 0.6 and

ω

w

= 0.2

√

0.4 0.6 ≈ 0.1 p.u. This yields a local maximum of e

max

w

≈ 0.94 p.u. With

v

ref

dc

= 2.8 p.u., this corresponds to a minimal dc-link voltage of v

min

dc

=

√

2.8

2

−0.94 ≈ 2.6

p.u., i.e., the dc-link voltage decreases by 0.2/2.8 100 ≈ 7 %.

Response to “Phase-Angle Jumps”

For reasons of simplicity and clarity, it is assumed that the modulus of the grid voltage vector

remains constant at E

g,nom

, i.e., no voltage sag accompanies the “phase-angle jump.” The

response of PWM rectiﬁers to “phase-angle jumps” is, to a large extent, determined by the

dynamics of the error angle. For a PLL tuned assuming a bandwidth of ρ, the time function

of the error angle after a “phase-angle jump” can be modeled as

˜

θ(t) =

˜

θ

0

e

−ρt

. (7.30)

In the time interval when

˜

θ(t) converges exponentially to zero with the rise time 1/ρ, the

q-axis grid voltage varies as E

gq

(t) = E

g,nom

cos[

˜

θ(t)] which yields the instantaneous grid-

ﬁlter power as P

f

= 3E

g,nom

i

fq

(v

dc

) cos[

˜

θ(t)]. Since P

f

is a function of time and v

dc

, the

95

dynamics of W, are time varying during “phase-angle jumps” in contrast to symmetrical

voltage sags. As a remedy for this, the seemingly daring assumption of nearly constant

dc-link voltage during “phase-angle jumps” is adopted. This assumption is validated by

simulations and experiments in Section 7.2.3 which show that the approximation is, indeed,

reasonable. With

˙

W ≈ 0, the “dynamics” after a “phase-angle jump” at t = 0 simplify to

t ≥ 0 : 0 ≈ −3E

g,nom

i

fq

cos

˜

θ + P

t0

= −3E

g,nom

i

fq

+

P

t0

cos

˜

θ

. (7.31)

The approximated “dynamics” in (7.31) are time-invariant, since P

f

= 3E

g,nom

i

fq

(v

dc

)

is a function of v

dc

only. Therefore, from (7.24) and (7.31), a “phase-angle jump” is in

close correspondence to a time varying P

t

which changes from P

t

(0

−

) = P

t0

to P

t

(t) =

P

t0

/ cos[

˜

θ(t)] at constant E

gq

= E

g,nom

. For small

˜

θ, such that 1/ cos

˜

θ ≈ 1 +

˜

θ

2

/2, the net

change in P

t

is

ΔP(t) =

1 +

˜

θ

2

(t)

2

P

t0

−P

t0

=

ˆ

θ

2

0

2

e

−2ρt

P

t0

(7.32)

where the latter expression results from (7.30). The time function of the error signal can be

derived by taking the inverse Laplace transform of the product of (7.21) and ΔP(p):

e

w

(t) = L

−1

¸

G

nom

Pe

(p)L¦ΔP(t)¦

¸

=

2

˜

θ

2

0

P

t0

ρC

dc

e

−ρt

+

ρ

2

t −1

e

−ρt

(7.33)

where α

w

= ρ is assumed since proper rejection of the negative-sequence voltage requires a

PLL bandwidth of ρ ≈ 0.2 p.u. [76] (this happens to coincide with the selection of α

w

= 0.2

p.u. in the beginning of this section). Within a short time interval after a “phase-angle jump,”

e

w

(t) can be approximated by

e

w

(t) ≈

2

˜

θ

2

0

P

t0

ρC

dc

e

−ρt

−1

e

−ρt

= e

w

(t) (7.34)

since e

−ρt

initially decays faster than ρt/2 increases. The error signal e

w

(t) has a local

minimum/maximum for t = ln 2/ρ. By substituting this instant in (7.33), the extreme value

for e

w

(t) is found as

e

w

≈

2

˜

θ

2

0

P

t0

ρC

dc

1

2

+

ln 2

2

−1

1

2

≈ −

0.15

˜

θ

2

0

P

t0

ρC

dc

. (7.35)

The values P

t0

= −1.5 p.u.,

˜

θ

0

= −π/4 rad, α

w

= ρ = 0.2 p.u. and C

dc

= 3.5 p.u. are

considered for a numerical example, which gives a local maximum of e

max

w

= 0.15 1.5

π

2

/(4

2

0.2 3.5) ≈ 0.2 p.u. With v

ref

dc

= 2.8 p.u., this implies that the dc-link voltage

decreases to v

min

dc

= (2.8

2

− 0.2)

0.5

≈ 2.76 p.u., i.e., a decrement by 0.04/2.8 100 ≈ 1 %.

From this analytic ﬁnding, which is supported by simulations and experiments in Section

7.2.3, it can be concluded that “phase-angle jumps” are believed not to be critical for PWM

rectiﬁers.

Response to Unsymmetrical Voltage Sags

The response of PWM rectiﬁers to unsymmetrical voltage sags is partly similar to the re-

sponse to symmetrical voltage sags, although less critical, since the remaining positive-

sequence voltage of unsymmetrical sags is never as small as that of the worst-case sym-

metrical sag, as discussed previously. A unique property of unsymmetrical sags is, on the

96

other hand, that the negative-sequence voltage vector introduces a ripple in the instanta-

neous grid power. This power ripple in turn gives rise to ripple in the dc-link voltage and

in i

fq

, which can be expressed as i

fq

= i

avg

fq

+ ˜ı

fq

, where i

avg

fq

is the average value of i

fq

and ˜ı

fq

is the current ripple. Two simpliﬁcations are introduced in order to analyze these

ripples. First shortly after an unsymmetrical sag it is assumed that the PLL recovers the po-

sition of the positive-sequence voltage vector, i.e., the q-axis grid voltage eventually varies

as E

gq

(t) = E

p

+ E

n

cos(−2θ

g

+ ϕ) where ϕ is the angle of the negative-sequence voltage

vector for

˜

θ = t = 0. Provided with this expression for E

gq

and, secondly, small current

ripple, such that [i

avg

fq

[ [˜ı

fq

[, the dc-link voltage dynamics simplify to

1

2

C

dc

dW

dt

≈ −3E

p

i

fq

−3E

n

i

avg

fq

cos(2θ

g

−ϕ) + P

t

. (7.36)

From (7.36), the power ripple can be treated as a turbine power disturbance, denoted by

˜

P

t

= 3E

n

i

avg

q

cos(2θ

g

− ϕ) =

˜

P

pk

t

cos(2θ

g

− ϕ), with constant E

gq

= E

p

. The dc voltage

ripple that results from

˜

P

t

are obtained from the static gain of G

Pe

(p) in (7.19) at the relevant

frequency 2ω

g

and E

gq

= E

p

:

[G

Pe

(j2ω

g

)[ =

4ω

g

C

dc

(α

2

w

ξ

p

−4ω

2

g

)

2

+ (4ω

g

α

w

ξ

p

)

2

(7.37)

where ξ

p

= E

p

/E

g,nom

. If α

w

is selected at least three times smaller than ω

g

, i.e., smaller

than 0.3 p.u., such that ω

2

g

α

2

w

, then (7.37) can be approximated as

[G

Pe

(j2ω

g

)[ ≈

4ω

g

C

dc

(−4ω

2

g

)

2

=

1

ω

g

C

dc

. (7.38)

Hence, the ripple in e

w

, due to an unsymmetrical voltage sag, is determined, to a large

extent, by the dc-link capacitance. An LLF is considered for a numerical example. The

values E

n

= 0.5 p.u., i

avg

fq

= 1 p.u., v

ref

dc

= 2.8 p.u. and C

dc

= 3.5 p.u. yield a peak ripple

of ˜ e

pk

w

≈

˜

P

pk

t

/(ω

g

C

dc

) = 3 0.5 1/(1 3.5) = 0.43 p.u. at a frequency of 2ω

g

. The

corresponding peak value of the v

dc

ripple is ˜ v

pk

dc

= 2.8 − (2.8

2

− 0.43)

0.5

≈ 0.08 p.u., or a

ripple of 0.08/2.8 100 ≈ 3 %. This is a fairly small ripple which is not critical for the proper

operation of a PWM rectiﬁer. As for symmetrical sags, the modulus of the positive-sequence

voltage vector is the most critical consequence. This indicates that if E

p

= 0.5 p.u., a q-axis

current of i

avg

fq

= 1 p.u. yields P

f

= 3E

p

i

avg

fq

= 1.5 p.u., whereas nominal power, i.e., P

f

= 3

p.u., requires i

avg

fq

= 1/0.5 = 2 p.u. This may be too large a current to be tolerated in a WT

application; remedies for avoiding large q-axis currents are to be discussed in Section 7.2.2.

As previously discussed, the power ripple during unsymmetrical voltage sags also trans-

fers to the q-axis current, via the dc voltage control system. In order to analyze the resulting

q current ripple, which adds to the grid current distortion during faults, the transfer function

from P

t

to i

fq

can be derived from Fig. 4.15, which results in

G

Pi

(p) =

−2[G

a

+ F(p)]/(pC

dc

)

1 −6E

q

[G

a

+ F(p)]/(pC

dc

)

=

2α

w

(p + α

w

/2)

3E

g,nom

(p

2

+ 2α

w

E

gq

/E

g,nom

p + α

2

w

E

gq

/E

g,nom

)

.

(7.39)

97

By substituting E

gq

= E

p

in this expression, the static gain of G

Pi

(p) at the relevant fre-

quency 2ω

g

is obtained as

[G

Pi

(j2ω

g

)[ =

α

w

3E

g,nom

α

2

w

+ 16ω

2

g

(α

2

w

ξ

p

−4ω

2

g

)

2

+ (4ω

g

α

w

ξ

p

)

2

≈

α

w

3ω

g

E

g,nom

(7.40)

where the latter approximation holds when α

w

is selected at least three times smaller than

ω

g

. The relation in (7.40) implies that the resulting ripple in i

fq

is mainly determined by the

bandwidth of the dc voltage control system. Therefore, a less distorted grid current during

unsymmetrical voltage sags can be obtained by selecting α

w

smaller. This yields a peak

ripple of ˜ı

pk

fq

≈ α

w

˜

P

pk

t

/(3ω

g

E

g,nom

) = 0.2 1.5/3 = 0.1 p.u. during an LLF, with identical

values as previously and α

w

= 0.2 p.u. This is a fairly large ripple although the previous

assumption on [i

avg

fq

[ [˜ı

fq

[ is still reasonable.

7.2.2 Discussion

In general, WTs using PWM rectiﬁers are robust towards voltage sags but large reductions

in modulus of the positive-sequence voltage vector appear to be critical. For a voltage sag

where the modulus reduces to V , no more than P

max

f

= 3V i

max

fq

can be transferred to the

utility grid. Depending on the wind situation when the voltage sag occurs, it may happen

that the turbine power is larger compared to P

max

f

. For such operating conditions, the dc-link

voltage begins to increase, unless the excess energy is somehow stored or dissipated. The

design of such energy storages depends on several factors of which some are:

• Cost.

• Grid codes.

• The remaining modulus of the positive-sequence grid voltage vector and the duration

of the voltage sag.

Depending on these factors, one, or possibly a combination, of the following four solutions

may be applicable:

Rotor Energy Storage

In this solution, the turbine power is controlled to P

t

= P

f

by changing the torque reference

for the turbine. If the pre-sag grid power must be restored moments after the voltage sag

is cleared, the blades should preferably remain in their pre-sag position, unless the WT ap-

proaches overspeed. If there is no need for instantaneous power restoration, the blades can

be pitched out of the wind directly.

“Braking” Chopper

A “braking” chopper, acting as a load dump, can be installed at the dc link. The limiting fac-

tor of this solution is the heat generated by the “braking” resistor which may be troublesome

to remove for long-duration voltage sags or interruptions.

98

DC-Link Energy Storage

Alarge dc-link capacitor bank can possibly be used, such that energy fromthe WT is buffered

at the dc link during the sag. The required size of the capacitor bank can be calculated

by substituting

˙

W = W

Δ

/t

Δ

in (7.25), assuming i

fq

= i

max

fq

, and solving the resulting

expression for C

dc

:

C

dc

=

2Δt

ΔW

−3V i

max

fq

+ P

t

. (7.41)

If the dc-link voltage is allowed to increase by no more than 10 %, then W

Δ

= (1.1v

dc

)

2

−

(v

ref

dc

)

2

= 0.21(v

ref

dc

)

2

. The values ΔW = 0.21 2.7

2

≈ 1.5 p.u., Δt = 0.25 s, V = 0,

i

max

fq

= 1 p.u. and P

t

= 3 p.u. are considered for a numerical example, which gives C

dc

=

2 0.25 314 3/1.5 ≈ 310 p.u. This is a very large value, so a dc-link energy storage appears

to be suitable mainly for small voltage sags that appear for a short period of time.

Overcurrent

The PWM rectiﬁer can be designed for overcurrent, i.e., i

max

fq

> 1 pu. However, the thermal

limit of the utility grid may not be designed for such overcurrent, especially if several WTs

are connected to a common point.

7.2.3 Evaluation

This section presents simulated and experimental results of a PWM rectiﬁer which is sub-

jected to various disturbances and voltage sags. The base values are 85 A, 105 V, 50 Hz,

and 1.2 Ω. The dc-link capacitance is C

dc

= 9.2 mF, which corresponds to C

dc

= 3.5 p.u.

The PWM rectiﬁer uses 4.9 kHz sampling and switching frequency and the reference for the

dc-link voltage is normally 2.8 p.u. The PWM rectiﬁer is loaded by a dc-link resistor which

corresponds to P

t

= −1.5 p.u. and i

fq

= −0.5 pu at v

dc

= 2.8 p.u.

The closed-loop grid current and dc-link voltage control loops are tuned for the band-

widths 2.3 p.u. and α

w

= 0.2 p.u. respectively, which corresponds to a current rise time

of 3 ms and a dc-link voltage rise time of 35 ms. The d current reference equals zero, the

maximum current modulus allowed is 1 pu, and the PLL bandwidth is ρ = 0.2 p.u.

Fig. 7.3 shows the results from the ﬁrst simulation and experiment. A pure “phase-angle

jump” of φ = −45

◦

≈ −0.8 rad occurs at t = 0.05 s, which yields an initial error angle of

˜

θ

0.05

= −0.8 rad. As seen from

˜

θ in Fig. 7.3b), the PLL recovers accurate ﬁeld orientation

at approximately 40 ms after the “phase-angle jump” so the PWM rectiﬁer is hardly affected

by the “jump,” as already concluded. A symmetrical voltage sag occurs in the time interval

t =0.1–0.3 s, giving, eventually, E

g

= 0.5 p.u. and requiring i

q

to be close to 1 p.u. in the

steady state. Moments after t = 0.1 s, the grid voltage modulus is E

g

≈ 0.6 p.u. which

causes the dc-link voltage to drop to v

dc

= 2.65 p.u. at t = 0.12 s. This is close to the

predicted value v

dc

= 2.6 p.u., resulting from the numerical example in Section 7.2.1. The

grid voltage is recovered at t = 0.32 s, causing the dc-link voltage to increase to v

dc

= 3 p.u.

Fig. 7.4 shows the results of an unsymmetrical voltage sag, characterized by E

p

= 0.6

p.u. and E

n

= 0.4 p.u. The sag occurs in the time interval t =0.1–0.3 s. In all other aspects,

the simulation and corresponding experiment are carried out under similar conditions as in

Fig. 7.3. The simulated and experimental waveforms are similar to those in Fig. 7.3 except

99

0 0.1 0.2 0.3 0.4

2.5

3

3.5

0 0.1 0.2 0.3 0.4

2.5

3

3.5

0 0.1 0.2 0.3 0.4

−1.5

−1

−0.5

0

0.5

0 0.1 0.2 0.3 0.4

−1.5

−1

−0.5

0

0.5

0 0.1 0.2 0.3 0.4

0

1

2

0 0.1 0.2 0.3 0.4

0

1

2

0 0.1 0.2 0.3 0.4

−1

0

1

0 0.1 0.2 0.3 0.4

−1

0

1

a)

v

d

c

[

p

.

u

.

]

v

d

c

[

p

.

u

.

]

b)

c)

i

f

q

[

p

.

u

.

]

i

f

q

[

p

.

u

.

]

d)

e)

E

q

[

p

.

u

.

]

E

q

[

p

.

u

.

]

f)

g)

˜ θ

1

[

r

a

d

]

˜ θ

1

[

r

a

d

]

h)

Time [s] Time [s]

Fig. 7.3. Response to “phase-angle jump” and symmetrical voltage sag. a) DC-link voltage, v

dc

(simulation). b) DC-link voltage, v

dc

(experiment). c) Grid-ﬁlter current, i

fq

(simulation).

d) Grid-ﬁlter current, i

fq

(experiment). e) Grid voltage, E

q

(simulation). f) Grid voltage,

E

q

(experiment). g) PLL error angle,

˜

θ

1

(simulation). h) PLL error angle,

˜

θ

1

(experiment).

that a ripple of approximately 0.1 p.u. is superimposed on the dc-link voltage, and the ripple

in the q-axis current is close to 0.12 p.u. These ripples are in close correspondence to the

values predicted by the numerical example in the analysis section.

In the last experiment, the load power is stepped fromP

t

= 0 to P

t

= −1.5 p.u. at t = 0.1

s, and 0.2 s later, the reference for the dc-link voltage changes stepwise from v

ref

dc

= 2.8 p.u.

to v

ref

dc

= 3 p.u. Fig. 7.5 shows the results. The dc-link voltage reduces to 2.5 p.u., as

predicted in Section 7.2.1, 15 ms after the load power step. The step response for t > 0.3 s

is well damped and the dc voltage rise time (10–90 % of the ﬁnal value) appears to equal the

intended 35 ms.

7.2.4 Conclusion

The voltage sag response of PWM rectiﬁers has been investigated for a candidate dc-link

voltage control system. A method of analysis was derived, which showed good agreement

between analytical predictions and experimental results. For several types and magnitudes

of voltage sags, the candidate dc-link voltage control system can successfully reduce dis-

turbances from both symmetrical and unsymmetrical voltage sags such that nominal power

100

0 0.1 0.2 0.3 0.4

2.5

3

3.5

0 0.1 0.2 0.3 0.4

2.5

3

3.5

0 0.1 0.2 0.3 0.4

−1.5

−1

−0.5

0

0.5

0 0.1 0.2 0.3 0.4

−1.5

−1

−0.5

0

0.5

0 0.1 0.2 0.3 0.4

−0.5

0

0.5

1

1.5

0 0.1 0.2 0.3 0.4

−0.5

0

0.5

1

1.5

0 0.1 0.2 0.3 0.4

−1

0

1

0 0.1 0.2 0.3 0.4

−1

0

1

a)

v

d

c

[

p

.

u

.

]

v

d

c

[

p

.

u

.

]

b)

c)

i

f

q

[

p

.

u

.

]

i

f

q

[

p

.

u

.

]

d)

e)

E

q

[

p

.

u

.

]

E

q

[

p

.

u

.

]

f)

g)

˜ θ

1

[

r

a

d

]

˜ θ

1

[

r

a

d

]

h)

Time [s] Time [s]

Fig. 7.4. Response to “phase-angle jump” and LLF unsymmetrical voltage sag. a) DC-link voltage,

v

dc

(simulation). b) DC-link voltage, v

dc

(experiment). c) Grid-ﬁlter current, i

fq

(simu-

lation). d) Grid-ﬁlter current, i

fq

(experiment). e) Grid voltage, E

q

(simulation). f) Grid

voltage, E

q

(experiment). g) PLL error angle,

˜

θ

1

(simulation). h) PLL error angle,

˜

θ

1

(experiment).

0 0.1 0.2 0.3 0.4

2.4

2.6

2.8

3

3.2

0 0.1 0.2 0.3 0.4

2.4

2.6

2.8

3

3.2

0 0.1 0.2 0.3 0.4

−1.5

−1

−0.5

0

0.5

0 0.1 0.2 0.3 0.4

−1.5

−1

−0.5

0

0.5

a)

v

d

c

[

p

.

u

.

]

v

d

c

[

p

.

u

.

]

b)

c)

i

f

q

[

p

.

u

.

]

i

f

q

[

p

.

u

.

]

d)

Time [s] Time [s]

Fig. 7.5. Steps in P

s

and v

ref

dc

. a) DC-link voltage, v

dc

(simulation). b) DC-link voltage, v

dc

(experi-

ment). c) Grid-ﬁlter current, i

fq

(simulation). d) Grid-ﬁlter current, i

fq

(experiment).

101

production can be restored once the grid voltage recovers. However, large reductions in the

positive-sequence voltage were found to be critical. Unless suitable actions are taken, such a

voltage reduction sag may result in a dc-link overvoltage since the transferable active power

reduces with reducing grid voltage. Remedies for avoiding overvoltage at the dc link have

also been discussed.

7.3 Doubly-Fed Induction Generator with Shunt Converter

Fig. 7.6 shows a principle sketch of the DFIG. In the ﬁgure a crowbar is also depicted,

which short-circuits the rotor circuit in case of too large a grid disturbance causing high

rotor current, and thereby protects the rotor converter. After such an action, rotor current

control has been lost and the turbine must be disconnected from the grid. The crowbar in

Crowbar

DFIG Grid

=

= ≈

≈

Fig. 7.6. Doubly-fed induction generator system with a crowbar.

Fig. 7.6 consists of a diode rectiﬁer and a thyristor that is triggered when the rotor circuit

should be short circuited. One disadvantage with this system is that once the crowbar has

been triggered, the turbine must be disconnected from the grid, since the current through the

thyristor is a continuous dc current and can only be interrupted if the turbine is disconnected

from the grid [72]. However, one possibility is to still have a rotor converter, but one that can

handle a higher current for a short period of time of some 100s of ms. Assuming such a short

over-current time, this means that only the IGBT modules need to be designed for a higher

current while the rotor winding and the converter (cooling etc) still can be designed according

to the slip power only. This means that the converter shortly can handle a higher current and

thereby stay connected to the grid longer without any crowbar action. Still, this system will

have high fault currents from the stator during the voltage sag. Since the relatively low

power losses in the power electronic equipment were a major reason for selecting a DFIG, it

is accordingly important to study how the ride-through system inﬂuences the power losses,

since additional hardware or modiﬁcations may reduce the efﬁciency.

Before explaining the candidate DFIG ride-through system, we will look further into the

dynamics of the DFIG during a voltage sag.

102

7.3.1 Response to Small Voltage Sags

In order to explain what happens to the DFIG during and after a voltage sag, we will start

by looking at the ﬂux dynamics. First, we will assume that the converter is ideal an can

supply the desired rotor voltage and current, and steady-state conditions are assumed to

precede the voltage sag. As discussed several times before, the ﬂux dynamics of the DFIG

are poorly damped. As also previously shown a voltage sag will cause the stator ﬂux to

enter a poorly damped oscillation with an oscillating frequency close to the line frequency.

The amplitude of the oscillation will be proportional to the size of the voltage sag, which

can be realized from the fact that ψ

s

≈ E

g

/ω

g

. If the DFIG system survives the voltage

sag, i.e., no crowbar action, the amplitude of the ﬂux oscillations can, after the voltage sag,

i.e., when the voltage returns, vary between zero and close to twice the ﬂux oscillations in

the beginning of the sag. The reason that the amplitude of the stator ﬂux oscillation can

almost vary between zero and twice the initial amplitude is that steady-state condition hardly

precedes the returning of the voltage. This can be realized from Fig. 7.7, where a phase

portrait and corresponding time series of the ﬂux can be seen. Note, that in order to make the

ﬁgure more lucid, the duration of the voltage sags in the time series is two periods longer.

In the ﬁgure, two different voltage sags are shown, with the duration of the sag as the only

difference. The cross marks the equilibrium point during normal operation and the circle

marks the equilibrium point during the voltage sag. It can be seen in the ﬁgure that directly

after the voltage sag has occurred, the ﬂux will circularly approach the “new” equilibrium

point (the circle) very slowly. This is indicated by the dashed lines in the ﬁgure. Then, when

the voltage returns, the ﬂux will again approach circularly the equilibrium point (the cross)

indicated with the solid line. However, as indicated by the difference between Fig. 7.7a) and

Fig. 7.7b), the duration of the sag is important. In Figs. 7.7a) and c), the voltage returns

when the ﬂux is close to the original equilibrium point (the cross), which leads to that the

ﬂux oscillations after the voltage sag are relatively small. However, as shown in Fig. 7.7b)

and d), if the voltage returns at an unfortunate moment, when the ﬂux is far away from the

original equilibrium point (the cross), the oscillations become even worse when the voltage

returns after the sag.

Symmetrical Voltage Sags

In this section, the dc-link dynamics of the DFIG system will be analyzed in a similar way as

for the full-power converter system in Section 7.2. In the analysis below, it will be assumed

that the “disturbance” is applied at t = 0 and that steady-state conditions precede the fault

causing the voltage sag. In order to analyze the response to voltage sags we will assume that

the magnitude of the stator ﬂux can be expressed in a similar way as in (5.47) and (5.48) as

t < 0 : Ψ

s

(t) = ψ

s0

≈

E

g,nom

ω

g

(7.42)

t ≥ 0 : Ψ

s

(t) ≈ ψ

s0

V

E

g,nom

+

1 −

V

E

g,nom

ψ

s0

e

−R

s

t/L

M

e

−jω

g

t

(7.43)

103

0.5 0.75 1 1.25 1.5

−0.5

−0.25

0

0.25

0.5

0.5 0.75 1 1.25 1.5

−0.5

−0.25

0

0.25

0.5

0 0.02 0.04 0.06 0.08 0.1

−0.5

0

0.5

1

1.5

0 0.02 0.04 0.06 0.08 0.1

−0.5

0

0.5

1

1.5

2

a)

ψ

sd

ψ

sd

ψ

s

q

ψ

s

q

b)

c)

d

d

q

q

Time [s] Time [s]

ψ

s

d

,

ψ

s

q

ψ

s

d

,

ψ

s

q

d)

Fig. 7.7. Phase portrait and time series of the ﬂux dynamics during a symmetrical voltage sag. For

clarity of the ﬁgure, the voltage sag duration for the time series is two periods longer. a) 18.5

ms long voltage sag. b) 10 ms long voltage sag. c) Time series of the ﬂux in a). d) Time

series of the ﬂux in b).

or as

t < 0 : Ψ

s

(t) = ψ

s0

≈

E

g,nom

ω

g

(7.44)

t ≥ 0 : Ψ

s

(t) ≈ ψ

s0

V

E

g,nom

+ ψ

pk

s

e

−R

s

t/L

M

e

−jω

g

t

(7.45)

where ψ

pk

s

= (1 −V/E

g,nom

)ψ

s0

is the peak value of the stator ﬂux oscillation. The expres-

sion for the stator ﬂux in (7.43) and (7.45) can be found by solving the differential equation

in (6.1). The dynamics of the dc-link are described by (4.48), and are governed both by the

rotor power P

r

and the grid-ﬁlter power P

f

, which can be approximated as

P

r

≈ 3E

gq

i

Rq

−3ω

r

Re[jΨ

s

i

∗

R

] ≈ 3E

gq

i

Rq

−3ψ

s

i

Rq

ω

r

(7.46)

P

f

≈ 3E

gq

i

fq

. (7.47)

104

The expression for P

r

is derived from the fact that P

r

= Re[3v

R

i

∗

R

] and by using the approx-

imation of the rotor voltage given by (5.38), i.e., v

R

≈ E

gq

−jω

r

Ψ

s

, where the stator voltage

has been changed to the grid voltage. This means that just before the voltage sag, P

r

and P

f

equal

P

r

(t = 0

−

) ≈ 3E

gq

i

Rq

−3ψ

s0

i

Rq

ω

r

(7.48)

P

f

(t = 0

−

) ≈ 3E

gq

i

fq

. (7.49)

Moreover, since steady-state conditions are assumed to precede the sag, we have that P

r

(t =

0

−

) = −P

f

(t = 0

−

), giving i

fq

(t = 0

−

) = −(E

gq

i

Rq

−ψ

s,0

i

Rq

ω

r

)/E

gq

≈ −(1−ω

r

/ω

g

)i

Rq

.

Under the assumption that the rotor current controller and grid-ﬁlter controller manage to

keep the current at (or at least close to) its reference value, moments after the sag has oc-

curred, it is possible to express P

r

and P

f

as

t ≥ 0 : P

r

(t) ≈ 3V i

Rq

−3

V

E

g,nom

i

Rq

ω

r

ψ

s0

+ ψ

pk

s

i

R

ω

r

e

−R

s

t/L

M

sin(ω

g

t + φ

r

)

(7.50)

t ≥ 0 : P

f

(t) ≈ 3V i

fq

≈ −3V

1 −

ω

r

ω

g

i

Rq

(7.51)

where i

R

= [i

R

[ and φ

r

= ∠i

R

. Then, as the stator ﬂux prior to the voltage sag can be

approximated as ψ

s0

≈ E

g,nom

/ω

g

, the above expression can be further reduced as

t ≥ 0 : P

r

(t) ≈ 3V

1 −

ω

r

ω

g

i

Rq

−3ψ

pk

s

i

R

ω

r

e

−R

s

t/L

M

sin(ω

g

t + φ

r

) (7.52)

t ≥ 0 : P

f

(t) ≈ −3V

1 −

ω

r

ω

g

i

Rq

. (7.53)

The dc-link dynamics in (4.48) are governed by the term −P

r

− P

f

. This means that the

power drop in the ﬁrst term of (7.52) is compensated for by the same drop in P

f

as can be

seen in (7.53). However, the second term in (7.52) will act as a disturbance to the dc-link, as

˜

P

r

= 3ψ

pk

s

i

R

ω

r

e

−R

s

t/L

M

sin(ω

g

t + φ

r

) =

˜

P

pk

r

e

−R

s

t/L

M

sin(ω

g

t + φ

r

). (7.54)

This disturbance will cause a ripple in the dc-link voltage with the frequency ω

g

. In order to

determine the amplitude of the ripple the static gain of (4.120), with G

Pe

(p) = −G

PW

(p),

at the relevant frequency can be used. This yields

[G

Pe

(jω

g

)[ =

2ω

g

C

dc

(α

2

w

ξ −ω

2

g

)

2

+ (2ω

g

α

w

ξ)

2

(7.55)

where ξ = V/E

g,nom

. If ω

g

α

w

, the above expression can be further approximated as

[G

Pe

(jω

g

)[ ≈

2

C

dc

ω

g

. (7.56)

For example, consider a voltage sag with V = 0.75 p.u., and C

dc

= 3.5 p.u. This means

that ψ

pk

s

= (1 − V/E

g,nom

)ψ

s0

= (1 − 0.75) 1 = 0.25 p.u., yielding

˜

P

pk

r

= 3ψ

pk

s

i

R

ω

r

=

105

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

1

1.25

0 0.1 0.2 0.3 0.4 0.5

−0.5

0

0.5

1

1.5

0 0.1 0.2 0.3 0.4 0.5

−100

0

100

0 0.1 0.2 0.3 0.4 0.5

2.5

2.6

2.7

2.8

2.9

3

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

1

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

a)

E

g

[

p

.

u

.

]

b)

Ψ

s

[

p

.

u

.

]

d

q

c)

P

r

+

P

f

[

%

]

d)

v

d

c

[

p

.

u

.

]

e)

i

R

q

[

p

.

u

.

]

f)

i

f

q

[

p

.

u

.

]

Time [s] Time [s]

Fig. 7.8. Simulation of the response of the DFIG system to a small symmetrical voltage sag. a) Grid

voltage. b) Stator ﬂux. c) Sum of rotor and grid-ﬁlter power. d) DC-link voltage. e) q-

component rotor current. f) q-component grid-ﬁlter current.

3 0.25 1 1 1.3 = 0.98 p.u. Then, according to (7.56), the peak ripple ˜ e

pk

w

in the error

signal e

w

= W

ref

dc

− W

dc

will be ˜ e

pk

w

= 2

˜

P

pk

r

/(C

dc

ω

g

) = 2 0.98/(3.5 1) = 0.56 p.u. The

corresponding peak value of the ripple in v

dc

is ˜ v

pk

dc

= 2.8−(2.8

2

−0.56)

0.5

= 0.1 p.u. when

v

ref

dc

= 2.8 p.u. In Fig. 7.8, a corresponding simulation is shown. The simulation veriﬁes the

ﬁnding that the sum of the rotor and grid-ﬁlter powers consists of a corresponding oscillating

power at a frequency of ω

g

. Moreover, the ripple in the dc-link voltage is 0.1 p.u., which is

according to the analytical result. However, as previously discussed, at the time when the

voltage returns, the amplitude of the stator-ﬂux oscillations can be close to twice the value

at the beginning of the sag. This means, of course, that the amplitude in the dc-link voltage

ripple will be increased accordingly. Since there is ripple in the dc-link voltage, this will also

be transferred to i

fq

, since it is used for controlling the dc-link voltage. With the transfer

function in (7.39) the static gain of the ripple in i

fq

can be calculated (note that −G

Pi

is

actually used since here P

t

= −P

r

). This yields

[G

Pi

(jω

g

)[ =

α

w

3E

g,nom

α

2

w

+ 4ω

2

g

(α

2

w

ξ

p

−ω

2

g

)

2

+ (2ω

g

α

w

ξ

p

)

2

≈

2α

w

3ω

g

E

g,nom

(7.57)

where the approximation holds if ω

g

α

w

. For the example given above we have, with

α

w

= 0.2, that ˜ı

pk

fq

= 2α

w

˜

P

pk

r

/(3ω

g

E

g,nom

) = 2 0.2 0.98/(3 1 1) = 0.13 p.u. This value

106

is also conﬁrmed by the simulation shown in Fig. 7.8. Moreover, the stator-ﬂux oscillations

will also cause a ripple in the stator current. The stator current can be found from (4.40) as

i

s

=

Ψ

s

L

M

−i

R

. (7.58)

Then if the rotor current is controlled accurately, i.e., i

R

= i

ref

R

, the ripple in the stator current

will be ˜ı

pk

s

= ψ

pk

s

/L

M

, which, with L

M

= 4.6 p.u., yields ˜ı

pk

s

= 0.25/4.6 = 0.05 p.u.

“Phase-Angle Jumps”

In this section we will study how the DFIG system responds to small “phase-angle jumps.”

Moreover, in the analysis it will be assumed that the magnitude of the grid voltage remains

at its nominal value after the “phase-angle jump.” This has been done in order to study the

effect of the actual “phase-angle jump” and not the inﬂuence of a voltage sag. After a pure

“phase-angle jump,” i.e, without any voltage sag, the grid voltage vector can be expressed as

E

g

= jE

g,nom

e

j

˜

θ(t)

≈ jE

g,nom

1 + j

˜

θ(t)

≈ jE

g,nom

1 + j

˜

θ

0

e

−ρt

(7.59)

where

˜

θ(t) is the error angle and the approximation holds if

˜

θ(t) is small. In (7.59) the error

angle

˜

θ(t) is modeled as in (7.30). Substituting (7.59) in (6.1) and solving the differential

equation, the following solution is obtained

t < 0 : Ψ

s

(t) = ψ

s0

≈

E

g,nom

ω

g

(7.60)

t ≥ 0 : Ψ

s

(t) ≈

E

g,nom

ω

g

ω

2

g

+ ρ

2

+ (1 + j)ω

g

ρ

˜

θ

0

e

−ρt

−

ω

g

ρ + jω

2

g

˜

θ

0

e

−(R

s

/L

M

+jω

g

)t

ω

2

g

+ ρ

2

(7.61)

if the stator resistance in the solution is assumed to be zero—except in e

−R

s

t/L

M

—and ψ

s0

≈

E

g,nom

/ω

g

. If ω

g

ρ, it is possible to further approximate the above equation as

t < 0 : Ψ

s

(t) = ψ

s0

≈

E

g,nom

ω

g

(7.62)

t ≥ 0 : Ψ

s

(t) ≈

E

g,nom

ω

g

+ j

E

g,nom

ω

g

˜

θ

0

e

−ρt

−e

−R

s

t/L

M

e

−jω

g

t

. (7.63)

Using (7.46) and (7.47), the rotor and grid ﬁlter powers can be determined in a similar way

as for the symmetrical voltage sag as

t ≥ 0 :

P

r

(t) ≈3E

g,nom

1 −

ω

r

ω

g

i

Rq

+ 3E

g,nom

ω

r

ω

g

˜

θ

0

i

Rd

e

−ρt

−i

R

e

−R

s

t/L

M

cos(ω

g

t + φ

r

)

(7.64)

t ≥ 0 : P

f

(t) ≈−3E

g,nom

1 −

ω

r

ω

g

i

Rq

. (7.65)

As for the case with symmetrical voltage sags, the dc-link dynamics in (4.48) are governed by

the term−P

r

−P

f

. This means that the power drop in the ﬁrst term of (7.64) is compensated

107

for by the same drop in P

f

; see (7.65). However, the second term in (7.64) will act as a

disturbance to the dc link, as

˜

P

r

= −3E

g,nom

˜

θ

0

ω

r

ω

g

i

Rd

e

−ρt

−i

R

e

−R

s

t/L

M

cos(ω

g

t + φ

r

)

. (7.66)

In (7.66), the disturbance consists of two terms: one that depends on the bandwidth, ρ, of

the PLL-type estimator and one that depends on the stator ﬂux dynamics. It is difﬁcult to

use the disturbance in (7.66) in order to ﬁnd the extreme value in the error signal e

w

since

it consists of two terms of which one is sinusoidal. One way of estimating the “worst case”

impact of a speciﬁc “phase-angle jump” is to treat the two terms independently and then add

them together. Of course, the result should be used with care since adding the results will

not, generally, give mathematically correct results. However, the analysis will still give some

valuable information of the system. The ﬁrst term’s impact on the dc-link dynamics can be

found from the extreme value of

e

w

(t) = L

−1

G

Pe

(p)L

−3E

g,nom

˜

θ

0

ω

r

ω

g

i

Rd

e

−ρt

= 3E

g,nom

˜

θ

0

ω

r

ω

g

i

Rd

(α

w

t −2)te

α

w

t

C

dc

(7.67)

where G

Pe

(p) is given by (7.21). The extreme value of (7.67) occurs for t = (2 −

√

2)/α

w

if ρ = α

w

. This means that the extreme value of (7.67) becomes

e

max/min

w

= 3E

g,nom

˜

θ

0

ω

r

ω

g

i

Rd

2(−1 +

√

2)e

−2+

√

2

C

dc

α

w

≈ 3E

g,nom

˜

θ

0

ω

r

ω

g

i

Rd

0.46

C

dc

α

w

(7.68)

if ρ = α

w

. The second term in (7.66) will cause ripple in the dc-link voltage with the static

gain according to (7.56).

The values

˜

θ

0

= −15

◦

≈ −0.26 rad, α

w

= ρ = 0.2 p.u., C

dc

= 3.5 p.u., w

r

= 1.3 p.u.,

i

R

= 1 p.u., and i

Rd

= 0.34 (corresponding to unity power factor) are used for a numerical

example. From (7.68) we have that e

max/min

w

= 3(−0.26)1.30.340.46/(3.50.2) = −0.23

p.u., which corresponds to a dc-link voltage of v

dc

= (2.8

2

− (−0.23))

0.5

= 2.84 p.u. The

amplitude of the second termin (7.66) becomes −3E

g,nom

˜

θ

0

ω

r

/ω

g

i

R

= −3(−0.26)1.31 =

1.01 p.u., giving according to (7.56) a ripple with the amplitude 1.01 2/3.5 = 0.58 p.u.,

which will cause a ripple in the dc-link voltage of ˜ v

pk

dc

= 2.8 − (2.8

2

− 0.58)

0.5

= 0.11 p.u.

This means that the “worst case” dc-link voltage could be v

wc

dc

= 2.84 + 0.11 = 2.95 p.u.

In Fig. 7.9 shows a simulation of the “phase-angle jump” used in the example. It can be

seen that the amplitude of the oscillation in the dc-link voltage and the maximum value of

the dc-link voltage is close to the predicted values. Eq. (7.57) can be used to determine the

amplitude of the oscillation in i

fq

. Since the amplitude of the oscillation in

˜

P

r

is 1.01 p.u.

the oscillation in i

fq

becomes ˜ı

pk

fq

= 2 0.2 1.01/(3 1 1) = 0.13 p.u.

For comparison to larger “phase-angle jumps” a corresponding simulation of a −45

◦

“phase-angle jump” can be seen in Fig. 7.10. With the same analysis as above will give a

dc-link voltage ripple of 0.33 p.u. and a “worst case” dc-link voltage of v

wc

dc

= 3.25 p.u.

Unsymmetrical Voltage Sags

Similar analysis, as for symmetrical voltage sags and “phase-angle jumps,” for unsymmetri-

cal sags is more difﬁcult to derive, since the system also will be excited with the negative-

sequence voltage. Therefore, the analysis here will be limited to simulations. In Fig. 7.11

108

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

1

1.25

0 0.1 0.2 0.3 0.4 0.5

−1

0

1

2

0 0.1 0.2 0.3 0.4 0.5

−50

0

50

0 0.1 0.2 0.3 0.4 0.5

2.6

2.8

3

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

1

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

a)

E

g

[

p

.

u

.

]

b)

Ψ

s

[

p

.

u

.

]

d

d

q

q

c)

P

r

+

P

f

[

%

]

d)

v

d

c

[

p

.

u

.

]

e)

i

R

q

[

p

.

u

.

]

f)

i

f

q

[

p

.

u

.

]

Time [s] Time [s]

Fig. 7.9. Simulation of the response of the DFIG system to a small “phase-angle jump” of −15

◦

.

a) Grid voltage. b) Stator ﬂux. c) Sum of rotor and grid-ﬁlter powers. d) DC-link voltage.

e) q-component rotor current. f) q-component grid-ﬁlter current.

the response to an SLGF occurring at t = 0.1 ms with V = 0.75 p.u. is presented. In the

ﬁgure it can be seen that in, for instance, the ﬂux, the dc-link voltage, and in the grid-ﬁlter

current oscillations with both frequencies of ω

g

and 2ω

g

. The oscillation with the frequency

ω

g

arises from the ﬂux dynamics while the oscillation with the frequency 2ω

g

arises from the

negative-sequence voltage. However, depending on the phase angle at the time instance of

the sag the oscillation at ω

g

can in principle be removed for an SLGF. This is indicated in

Fig. 7.12 where the sag occurs at t = 0.105 ms, all other conditions are as in Fig. 7.11.

Summary

The response of the DFIG system due to different grid disturbances has been investigated.

It has been shown that the amplitude of the ﬂux oscillation when the voltage returns after a

voltage sag can vary between zero and twice the initial amplitude of the ﬂux oscillations due

the sag. Moreover, the DFIG system has been analyzed for symmetrical voltage sags with

good agreement. However, the response to “phase-angle jumps” and unsymmetrical voltage

sags are analytically harder to derive. Moreover, the DFIG system is roughly as sensitive to

“phase-angle jumps” as to symmetrical voltage sags.

109

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

1

1.25

0 0.1 0.2 0.3 0.4 0.5

−2

0

2

0 0.1 0.2 0.3 0.4 0.5

−200

−100

0

100

0 0.1 0.2 0.3 0.4 0.5

2

2.5

3

3.5

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

1

0 0.1 0.2 0.3 0.4 0.5

−0.5

0

0.5

1

a)

E

g

[

p

.

u

.

]

b)

Ψ

s

[

p

.

u

.

]

d

d

q

q

c)

P

r

+

P

f

[

%

]

d)

v

d

c

[

p

.

u

.

]

e)

i

R

q

[

p

.

u

.

]

f)

i

f

q

[

p

.

u

.

]

Time [s] Time [s]

Fig. 7.10. Simulation of the response of the DFIG system to a “phase-angle jump” of −45

◦

. a) Grid

voltage. b) Stator ﬂux. c) Sum of rotor and grid-ﬁlter powers. d) DC-link voltage. e) q-

component rotor current. f) q-component grid-ﬁlter current.

7.3.2 Response to Large Voltage Sags

In previous section the voltage sags were assumed to be small enough, not causing the rotor

converter to fail in controlling the rotor current. However, this cannot be assumed for larger

voltage sags. As shown in (5.40), the rotor voltage will change in proportion to the depth

of the voltage sag. So, for larger voltage sags, the rotor voltage will hit its maximum value

and lose control of the rotor current. In this section, the DFIG system will be analyzed and it

is assumed that the converter is large enough to handle excess currents. This has been done

in order to study the behavior of the DFIG and not the inﬂuence of the converter and the

crowbar. Still, the rotor voltage in the simulations is, anyhow, limited to ±0.4 p.u. (referred

to the stator circuit). This limitation of the rotor voltage is a major difference compared to

the analysis in the previous section, since the converter will lose control of the rotor current.

In Fig. 7.13 a simulation of a symmetrical voltage sag (at 0.05 s) down to 0.25 p.u. is

presented. Before the voltage sag the DFIG is running at rated power and a rotor speed

of 1.3 p.u. The duration of the voltage sag in the simulation is 102 ms and 92 ms. In

the ﬁgure it can be seen that the rotor voltage will hit its maximum value directly after the

voltage sag. This means that the converter loses control of the rotor current, leading to an

uncontrolled rotor current. As shown in Section 7.3.1 the situation might be even worse when

the voltage returns. This is also indicated in Fig. 7.13, i.e., with two identical simulations

110

0 0.1 0.2 0.3 0.4 0.5

−0.25

0

0.25

0.5

0.75

1

1.25

0 0.1 0.2 0.3 0.4 0.5

−1

0

1

2

0 0.1 0.2 0.3 0.4 0.5

−50

0

50

0 0.1 0.2 0.3 0.4 0.5

2.7

2.8

2.9

3

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

1

1.25

0 0.1 0.2 0.3 0.4 0.5

0

0.2

0.4

a)

E

g

[

p

.

u

.

]

b)

Ψ

s

[

p

.

u

.

]

d

d

q

q

c)

P

r

+

P

f

[

%

]

d)

v

d

c

[

p

.

u

.

]

e)

i

R

q

[

p

.

u

.

]

f)

i

f

q

[

p

.

u

.

]

Time [s] Time [s]

Fig. 7.11. Simulation of the response of the DFIG system to an unsymmetrical (SLGF) voltage sag.

a) Grid voltage. b) Stator ﬂux. c) Sum of rotor and grid-ﬁlter powers. d) DC-link voltage.

e) q-component rotor current. f) q-component grid-ﬁlter current.

except the duration of the voltage sag. It can be seen that the maximum rotor current can be

much higher when the voltage returns than when the voltage drops at the beginning of the

disturbance, if the machine ﬂux is in the wrong “direction.” In Fig. 7.14, the maximum rotor

current and the maximum rotor power due to a symmetrical voltage sag for three different

operating conditions can be seen. In the ﬁgure, the effect of returning voltage has not been

taken into consideration. From the ﬁgure it can be seen that the maximum rotor current due

to a voltage sag will increase with the magnitude of the sag. Moreover, the maximum rotor

power that is fed into the dc link can be up to almost 250% of nominal power. It should be

kept in mind that for the ordinary DFIG system, the converter and dc link are only rated for

30–40% of the nominal power. This means that there is a huge rotor power that needs to be

dealt with. Based on these ﬁndings a candidate ride-through system will be presented in the

next section.

7.3.3 Candidate Ride-Through System

The aim of this section is to present a candidate ride-through DFIG system, based on the

result in the previous section. The main idea is to overdimension the valves of the power

electronic converter so that they can handle the rotor current occurring at deep voltage sags.

However, as indicated in Fig. 7.13, the maximum rotor current actually might be much

111

0 0.1 0.2 0.3 0.4 0.5

−0.25

0

0.25

0.5

0.75

1

1.25

0 0.1 0.2 0.3 0.4 0.5

−1

0

1

2

0 0.1 0.2 0.3 0.4 0.5

−50

0

50

0 0.1 0.2 0.3 0.4 0.5

2.75

2.8

2.85

2.9

0 0.1 0.2 0.3 0.4 0.5

0

0.25

0.5

0.75

1

1.25

0 0.1 0.2 0.3 0.4 0.5

0.25

0.3

0.35

a)

E

g

[

p

.

u

.

]

b)

Ψ

s

[

p

.

u

.

]

d

d

q

q

c)

P

r

+

P

f

[

%

]

d)

v

d

c

[

p

.

u

.

]

e)

i

R

q

[

p

.

u

.

]

f)

i

f

q

[

p

.

u

.

]

Time [s] Time [s]

Fig. 7.12. Simulation of the response of the DFIG system to an unsymmetrical (SLGF) voltage sag.

a) Grid voltage. b) Stator ﬂux. c) Sum of rotor and grid-ﬁlter powers. d) DC-link voltage.

e) q-component rotor current. f) q-component grid-ﬁlter current.

higher, i.e., up to twice as high, when the voltage returns than at the voltage drop. Of course,

this means that the valves have to be even more overdimensioned. In order to avoid these

high current when the voltage returns, anti-parallel thyristors can be connected in series with

the stator in order to achieve a quick disconnection of the stator circuit [20]. By interrupt-

ing the stator circuit, the ﬂux oscillation will also be interrupted. As soon as the ﬂux is

interrupted it is possible to remagnetize the DFIG quickly through the rotor converter and

connect the stator circuit to the grid again. The converter needs not to be disconnected from

the rotor circuit since the valves of the converter are overdimensioned. In order remove the

excess power that is fed into the dc link, a “dc-link breaking chopper” is used to dissipate

the excess power. Moreover, if required, the grid-side converter may provide the grid with

reactive power during the sag. The system with anti-parallel thyristors is illustrated in Fig.

7.15. The anti-parallel thyristor switch can disconnect the stator within a half cycle, i.e., in

10 ms, [9, 20]. The anti-parallel thyristor switch needs to be equipped with a forced commu-

tation unit in case of a dc component in the stator fault current [20]. Another option would

be to have gate-turn-off thyristors; then, the disconnection time can be lowered. However,

a complex driving circuit is needed, since typically a large negative gate current is required

to turn off that device [67]. A third option would be to have anti-parallel IGBTs. Since a

normal IGBT, a so-called punch-through IGBT, cannot handle reverse voltages as high as the

forward blocking voltages, a diode has to be put in series with the IGBT. Instead of using the

112

0 0.1 0.2 0.3

0

0.5

1

1.5

0 0.1 0.2 0.3

0

0.5

1

0 0.1 0.2 0.3

0

5

10

0 0.1 0.2 0.3

0

5

10

0 0.1 0.2 0.3

0

1

2

3

0 0.1 0.2 0.3

−4

−2

0

2

a)

E

g

[

p

.

u

.

]

v

R

[

p

.

u

.

]

b)

i

s

[

p

.

u

.

]

c)

i

R

[

p

.

u

.

]

d)

ψ

s

[

p

.

u

.

]

e)

P

r

[

p

.

u

.

]

f)

Time [s] Time [s]

Fig. 7.13. Simulation of the response of a DFIG system to a voltage sag down to 0.25 p.u. Solid line

correspond to a voltage sag of 102 ms and dashed line correspond to a voltage sag of 92

ms. a) Stator voltage. b) Rotor voltage magnitude. c) Stator current magnitude. d) Rotor

current magnitude. e) Stator ﬂux magnitude. f) Rotor power.

punch-through IGBT, the so-called non-punch through IGBT can be used. The non-punch

through IGBT can handle a reverse voltage as high as the forward blocking voltage. How-

ever, this device has higher on-state losses [67]. One last option would be to have a contactor

as a circuit breaker. However, the disconnection time for the contactor will be longer, while

on the other hand, in principle, without losses.

Thus, the IGBT modules of the machine-side converter are designed for a higher current

rating, in order to withstand voltage sags. Moreover, in order to avoid the possible higher

rotor currents when the voltage returns, the stator circuit is disconnected from the grid. Then,

after a disconnection, the rotor current controller re-magnetizes the DFIG, and then the DFIG

can be synchronized to the grid as soon as the voltage has returned to an acceptable, prede-

ﬁned, level. For the investigated system, the maximum rotor current and rotor power due to

symmetrical voltage sag is shown in Fig. 7.16, if the stator circuit is disconnected within 10

ms.

Of course, for IGBT modules that can handle higher currents temporarily, the stator

circuit can be disconnected less often due to voltage sags. Fig. 7.17 shows the maximum

rotor current and rotor power when the stator is not disconnected from the grid and the grid

voltage returns at the worst instance around 50 ms. This means that the duration of the

voltage sag is approximately 50 ms. It can be seen in the ﬁgure that for voltage sags down to

113

0 0.5 1

0

1

2

3

4

5

0 0.5 1

−250

−200

−150

−100

−50

0

50

M

a

x

i

R

[

p

.

u

.

]

a)

Voltage sag V [p.u.] Voltage sag V [p.u.]

M

a

x

P

r

[

%

]

b)

Case 1

Case 1

Case 2

Case 2

Case 3

Case 3

Fig. 7.14. Maximum rotor current and rotor power for three different operating conditions. Case 1

corresponds to rated power and a rotor speed of 1.3 p.u. Case 2 corresponds to 23% of

rated power and a rotor speed of 1.0 p.u. Case 3 corresponds to 11% of rated power and

a rotor speed of 0.7 p.u. a) Maximum rotor current. b) Maximum rotor power (Note that

negative rotor power means that the power is fed into the dc link).

Machine-side Grid-side

converter converter

DFIG

Switch

Y

Δ

“Breaking chopper”

=

= ≈

≈

Fig. 7.15. DFIG with anti-parallel thyristors in the stator circuit.

0.5 p.u. the maximum rotor current is approximately as large as the maximum rotor current

when the stator is disconnected. However, the maximum rotor power fed to the dc link

becomes higher compared to when the stator circuit is disconnected.

7.3.4 Evaluation of the Ride-Through System

The aim of this section is to make a theoretical case study on the candidate voltage sag

ride-through system presented in the previous section. This study will focus on the energy

production and energy production cost of such a system for a 2-MW DFIG WT.

114

0 0.5 1

0

1

2

3

4

5

0 0.5 1

−250

−200

−150

−100

−50

0

50

M

a

x

i

R

[

p

.

u

.

]

a)

Voltage sag V [p.u.] Voltage sag V [p.u.]

M

a

x

P

r

[

%

]

b)

Case 1

Case 1

Case 2

Case 2

Case 3

Case 3

Fig. 7.16. Maximum rotor current and rotor power for three different operating conditions if the

stator circuit is disconnected within 10 ms. Case 1 corresponds to rated power and a rotor

speed of 1.3 p.u. Case 2 corresponds to 23% of rated power and a rotor speed of 1.0 p.u.

Case 3 corresponds to 11% of rated power and a rotor speed of 0.7 p.u. a) Maximum rotor

current. b) Maximum rotor power (note that negative rotor power means that the power is

fed into the dc link).

0.4 0.6 0.8 1

0

2

4

6

8

0.4 0.6 0.8 1

−400

−300

−200

−100

0

M

a

x

i

R

[

p

.

u

.

]

a)

Voltage sag V [p.u.] Voltage sag V [p.u.]

M

a

x

P

r

[

%

]

b)

Case 1

Case 1

Case 2

Case 2

Case 3

Case 3

Fig. 7.17. Maximum rotor current and rotor power fed to the dc link for three different operating

conditions with returning voltage at the worst instance around 50 ms. Case 1 corresponds

to rated power and a rotor speed of 1.3 p.u. Case 2 corresponds to 23% of rated power and

a rotor speed of 1.0 p.u. Case 3 corresponds to 11% of rated power and a rotor speed of 0.7

p.u. a) Maximum rotor current. b) Maximum rotor power (note that negative rotor power

means that the power is fed into the dc link).

Calculation of Power Losses

The losses taken into consideration are the losses in the aerodynamic conversion, gearbox,

generator and in the semiconductor devices, i.e., the same losses as in Chapter 3. The calcu-

lation of the will also follow the loss models used in Chapter 3.

The losses in the anti-parallel thyristor switch used in the stator circuit, see Fig. 7.15, can

115

be described by (3.4). Since the switching occurs at zero current and at low frequency (the

grid frequency), the switching losses in the switch will be neglected. The parameters for the

thyristors used here are r

Thy

= 0.164 mΩ and V

Thy

= 0.88 V [93]. The dc-link “chopper” is

not used during normal operation; hence, it will not inﬂuence the energy production.

In Fig. 7.18 the losses in the semiconductor devices are shown. It can be noticed in Fig.

5 10 15 20 25

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

L

o

s

s

e

s

[

%

]

Wind speed [m/s]

Ordinary DFIG

Candidate DFIG

Thyristor switch

Grid-side converter

Fig. 7.18. The losses in the semiconductor devices used in this work. The solid lines show the losses

in the MSC for ordinary and the candidate DFIG system. The dashed line shows the losses

for the GSC and the dotted line shows the losses in the anti-parallel thyristor switch.

7.18, that the losses of the MSC can be reduced by approximately 0.05 percentage units at

rated operation by increasing the current rating of the valves. This reduction is only due

to that the resistance in the valves decreases with an increasing current rating. However, the

losses of the thyristor switch are much larger than the reduction of losses due to the increased

current rating of the valves. The steps in the curves at 7.8 m/s are due to that the stator of the

generator is switch from a Y connection to a Δ connection, as discussed in Section 2.3.4.

In Fig. 7.19 the expected efﬁciency of the candidate DFIG system as a function of the

average wind speed is presented. In the ﬁgure expected efﬁciencies of the ordinary DFIG

system as well as a system that utilizes a full power electronic converter can also be seen.

In the ﬁgure it can be seen that the ordinary DFIG system has the highest efﬁciency, even

though the difference towards the candidate DFIG system is relatively small. See Section

2.1.1 for details of calculation of the expected efﬁciency.

Energy Cost

For the calculation of the energy production cost, it has been assumed that the standard

2-MW DFIG WT costs e1600000 [65] and that one IGBT converter and the anti-parallel

thyristor switch costs e6000/p.u. current. The cost of the IGBTs is an estimate based on

cost information obtain from some IGBT manufactures. In Fig. 7.20, the relative energy

production cost of the candidate system normalized with the energy production cost of the

ordinary DFIG system can be seen. As could be expected, the energy production cost of the

candidate system is higher than for the ordinary DFIG system. The energy production cost

116

5 6 7 8 9 10

91

92

93

94

95

96

Full power converter system

Candidate DFIG system

Ordinary DFIG system

E

f

ﬁ

c

i

e

n

c

y

[

%

]

Average wind speed [m/s]

Fig. 7.19. Expected efﬁciency as a function of the average wind speed for the candidate DFIGsystem,

the ordinary DFIG system, and a system with a full-power converter.

5 6 7 8 9 10

1

1.01

1.02

1.03

1.04

1.05

Candidate DFIG system without thyristor switch

Full power converter system

Candidate DFIG system

R

e

l

a

t

i

v

e

e

n

e

r

g

y

c

o

s

t

[

%

]

Average wind speed [m/s]

Fig. 7.20. Energy cost of the candidate DFIG, candidate DFIG without thyristor switch, and the

system that utilizes a full-power converter. The energy cost is related to the ordinary DFIG

system.

of the candidate system is approximately two percentage units higher then for the ordinary

DFIG system. A full power electronic conversion, which handles the total power, has ap-

proximately three percentage units higher energy cost than the ordinary DFIG system. The

increase in energy production cost is due to the lower energy production of the candidate

system, but it is mainly due to the increased cost of the valves. In Fig. 7.20 the normalized

energy production cost of a modiﬁed candidate system, i.e., without anti-parallel thyristor

switch in the stator circuit, is also shown.

117

Conclusion

The inﬂuence on the energy production of a DFIG ride-through system has been investigated.

This system is based on increased current rating of the converter and anti-parallel thyristors

in the stator circuit. It has been found that the increased cost for a ride-through system for a

DFIG turbine might be reasonable, in comparison to the cost of full-power converter system

connected to a cage-bar induction generator.

7.4 Doubly-Fed Induction Generator with Series Converter

After a voltage sag, the stator ﬂux of the DFIG will start to oscillate. This oscillation often

causes very high rotor currents, which necessitates a disconnection of the WT. Today, the

grid-side converter is connected to the grid in a shunt conﬁguration, see Fig. 7.21. This

means that the converter injects a current into the grid. However, if the converter is instead

connected in series with the grid, a voltage is introduced in series with the stator voltage,

i.e., the stator voltage of the DFIG is the sum of the grid and converter voltages. Then, the

series voltage can be used in order to control the stator ﬂux of the machine and prevent, for

instance, high rotor current with resulting disconnection of the turbine. Kelber has shown

that such a system can effectively damp the ﬂux oscillations caused by voltage sags [55].

Conv. Conv.

DFIG DFIG

Grid Grid

Shunt connected Series connected

Fig. 7.21. Schematic ﬁgure showing shunt- and series-connected converters for doubly-fed induction

generator systems.

The contribution and purpose of this section is to analyze and present the advantages and

drawbacks of a DFIG system for a wind-turbine application with a series converter with the

focus on handling voltage sags. In addition, a goal is also to study the energy efﬁciency,

and, in particular, compare it to a system that utilizes a full-power converter. The reason for

comparing these two systems is that they are both capable of voltage sag ride-through.

7.4.1 Possible System Conﬁgurations

As mentioned in the Introduction, the idea is to have a converter connected in series with

the stator circuit and the grid. Fig. 7.21 shows both the ordinary DFIG system where the

converter is connected in shunt to the grid, and the system where it is connected in series.

The purpose of the series-connected converter is to control the stator ﬂux of the DFIG, and in

this way be able to control the DFIG during voltage sags. By having the converter connected

118

in series, the stator voltage v

s

of the DFIG is, ideally, the sum of grid voltage E

g

and the

voltage v

c

from the series converter:

v

s

= E

g

+v

c

. (7.69)

Some of the demands on the series converter for a DFIG system may be:

• A sufﬁciently fast stator-ﬂux control in order to damp the oscillations and control the

stator ﬂux.

• Accurate control of the dc-link voltage.

There are at least two methods of accomplishing this series voltage, which are presented

below.

Series-Injection Transformer

In this conﬁguration, the voltage source converter is connected to the grid via a series-

injection transformer, as depicted in Fig. 7.22. This conﬁguration of a series-injection trans-

former and a voltage source converter is also used in dynamic voltage restorers (DVRs)

[27, 49]. The protection system of such a system is complicated since a simple discon-

nection does not work [70]. Normally the system is equipped with an LC ﬁlter with the

objective of reducing voltage and current harmonics generated from the voltage source con-

verter. Note that the LC ﬁlter can be placed on either side of the transformer [49]. The

series-injection transformer is necessary for galvanic insulation. Moreover, in order to avoid

magnetic saturation, the series-injection transformer must be rated to handle twice the nom-

inal ﬂux [27]. Another option, in order to avoid the series-injection transformer, is to have

a converter for each phase with separate dc links [63]. For DVRs, there are, at least, three

Converter

DFIG

Grid

Transformer

Fig. 7.22. Doubly-fed induction generator (DFIG) with the grid-side converter connected in series

via a series-injection transformer.

methods of controlling the series voltage: 1) in an open-loop manner [45], 2) directly control

the series voltage [71], and 3) by two control loops, i.e., an inner fast current control loop

that controls the current through the inductance and an outer cascade loop controlling the

capacitor (series) voltage [10]. One advantage of controlling both the inductor current and

capacitor voltage is that it is easy to avoid the resonant frequency of the LC ﬁlter. How-

ever, a drawback for the DFIG is that bandwidth is lost for the stator-ﬂux controller, since

the stator ﬂux is then controlled in cascade with both the capacitor (series) voltage and the

inductor current. For example, if it is desired to separate the control loops by one decade,

119

the bandwidth of the ﬂux control loop is a factor of hundred lower than the current control

loop. This means that a very high bandwidth of the current control loop is necessary and,

accordingly, a very high switching frequency is needed.

Converter in the Y Point of the DFIG

The second method of accomplishing a series voltage for the DFIG is to connect a voltage

source converter where the Y point of the stator circuit usually is [54, 55]. Hereafter, this

will be referred to as the Y point. In [54, 55], this is accomplished using an additional (third)

converter, which is only used to damp the occurring stator ﬂux oscillations. During normal

operation, the extra converter voltage is zero. In [54, 55], the converter in the Y point of the

DFIG system is only used to damp disturbances, while here it can also be used to control the

magnitude of the stator ﬂux and the dc-link voltage.

Fig. 7.23 shows a principle sketch of the system when a voltage source converter is

connected to the Y point of the stator circuit of the DFIG. For this system, the converter

Converter

Stator circuit

Rotor circuit

Grid

Fig. 7.23. Doubly-fed induction generator (DFIG) with the grid-side converter connected to the Y

point of the stator circuit.

voltage is directly used to control the stator ﬂux in the machine, while the rotor current is

controlled by the machine-side converter. One disadvantage of this method is that all of

the stator current is passed through the Y-point converter, which may cause additional high

losses in the power electronic equipment.

7.4.2 System Modeling

As mentioned earlier the stator voltage is ideally the sum of the grid and series voltage. This

means that (4.38) and (4.39) become

v

s

= E

g

+v

c

= R

s

i

s

+

dΨ

s

dt

+ jω

1

Ψ

s

(7.70)

v

R

= R

R

i

R

+

dΨ

R

dt

+ jω

2

Ψ

R

(7.71)

where v

c

is the series voltage. The dc-link dynamics are described by

C

dc

2

dv

2

dc

dt

= v

dc

C

dc

dv

dc

dt

= −P

r

−P

c

(7.72)

120

where P

r

is the rotor power and P

c

is the power from the grid-side converter, which are given

by

P

r

= 3Re [v

R

i

∗

R

] (7.73)

P

c

= 3Re [v

c

i

∗

s

] (7.74)

or as

P

r

= 3

R

R

i

2

Rq

+ R

R

ψ

2

sd

L

2

M

+ ψ

sd

i

Rq

ω

2

(7.75)

P

c

= 3

R

s

i

2

Rq

+ E

g

i

Rq

−ψ

sd

i

Rq

ω

1

(7.76)

in the steady state if ψ

sq

= 0, i.e., controlled to be zero. Moreover, in the above equations

the d component of the rotor current is controlled so that the system operates at unity power

factor, i.e., i

Rd

= ψ

sd

/L

M

.

Steady-State Operation

For a constant dc-link voltage it is required that P

r

= −P

c

. This means that ψ

sd

must be

used to control the dc-link voltage, and in the steady state ψ

sd

approximately becomes

ψ

sd

≈

E

g

ω

r

(7.77)

where the approximation is that the stator resistance has been neglected. Since the magne-

tizing losses depend on the ﬂux in the machine, (7.77) indicates that the system will have an

undesirable feature. At low wind speeds (low power), the rotor speed is also low, causing

the ﬂux to be high and thereby also the magnetizing losses to increase. At high wind speeds

(high power), the rotor speed is low, which means that the stator ﬂux is low. Since the stator

ﬂux is low, a higher torque-producing current is needed.

In the steady state, the rotor voltage can be expressed as

v

R

= (R

R

+ jω

2

L

σ

) i

ref

R

+ jω

2

Ψ

s

= (R

R

+ jω

2

L

σ

) i

ref

R

+ jω

2

ψ

sd

. (7.78)

If ψ

sd

= E

g

/ω

r

, (7.78) becomes

v

R

= (R

R

+ jω

2

L

σ

) i

ref

R

+ jω

2

E

g

ω

r

(7.79)

= (R

R

+ jω

2

L

σ

) i

ref

R

+ j

ω

1

ω

r

−1

E

g

(7.80)

≈ j

ω

1

ω

r

−1

E

g

. (7.81)

For example, we have, with ω

1

= E

g

= 1 p.u.

v

R

≈

0.43 p.u. for ω

r

= 0.7 p.u.

−0.23 p.u. for ω

r

= 1.3 p.u.

(7.82)

121

showing that the rotor voltage is not symmetrically distributed around the synchronous speed,

as for the case with constant stator ﬂux.

The series voltage v

c

of the grid-side converter is given by

v

c

= R

s

i

s

+ jω

1

Ψ

s

−E

g

≈ jω

1

Ψ

s

−E

g

(7.83)

in steady state. Then, as ψ

sq

= 0 and E

g

= jE

g

, we have

v

c

≈ jω

1

ψ

sd

−jE

g

= jω

1

E

g

ω

r

−jE

g

= j

ω

1

ω

r

−1

E

g

. (7.84)

As seen in (7.81) and (7.84), the rotor voltage will approximately equal the converter voltage,

i.e., v

R

≈ v

c

.

Close to No-Load Operation

It is required that (7.72) equals zero at steady-state operation in order to have a constant dc-

link voltage. If the rotor current and stator ﬂux are controlled with high-gain feedback, we

have that

R

R

i

2

Rq

+ R

R

ψ

2

sd

L

2

M

+ ψ

sd

i

Rq

ω

2

+ R

s

i

2

Rq

+ E

g

i

Rq

−ψ

sd

i

Rq

ω

1

= 0 (7.85)

in order to have a constant dc-link voltage. Note that in (7.85), ψ

sq

= 0. Eq. (7.85) can be

rewritten as

ψ

2

sd

−ω

r

i

Rq

L

2

M

R

R

ψ

sd

+

L

2

M

R

R

(R

s

+ R

R

)i

2

Rq

+ E

g

i

Rq

= 0 (7.86)

which has the following solution:

ψ

sd

= ω

r

i

Rq

L

2

M

2R

R

±

ω

2

r

i

2

Rq

L

4

M

4R

2

R

−

L

2

M

R

R

(R

s

+ R

R

)i

2

Rq

+ E

g

i

Rq

. (7.87)

The expression under the square root cannot be negative, implying that

L

2

M

R

R

i

Rq

¸

ω

2

r

L

2

M

4R

R

−(R

s

+ R

R

)

i

Rq

−E

g

≥ 0 (7.88)

or

[i

Rq

[ ≥

E

g

ω

2

r

L

2

M

4R

R

−(R

s

+ R

R

)

≈

4E

g

R

R

ω

2

r

L

2

M

. (7.89)

If [i

Rq

[ < 4E

g

R

R

/(ω

2

r

L

2

M

), ψ

sd

cannot keep the dc-link voltage constant. For the values

given in the Appendix, the constraint becomes [i

Rq

[ < 4 1 0.009/(1

2

4.6

2

) = 0.0017 p.u.,

which is a small value. In order to handle this problem an extra converter that controls the

dc-link voltage is added. It might be possible to use either a diode rectiﬁer (depending on the

power ﬂow) or an IGBT converter as the extra converter. However, later on when the losses

122

and efﬁciency are calculated, an IGBT converter has been used. Here two different sizes of

this extra converter will be investigated:

Option 1. In this case, the extra converter is designed to be as small as possible. However,

since this converter would be small and only used at very low powers another way could be

to increase energy storage on the dc-link, and thereby make the third converter unnecessary.

Option 2. For this option the extra converter is designed so that it is used in the whole

operating region. This means that for this option the stator ﬂux is not used for controlling

the dc-link voltage. Kelber et al. used this option [54, 55]. However, in contrast to [54, 55],

the stator ﬂux is here controlled to reduce the magnetizing losses. This means as the extra

converter controls the dc-link voltage, the stator-ﬂux reference value is set to minimize the

losses of the generator. Note, that for this option the ﬂux does not follow (7.77).

7.4.3 Control

The basic idea of the control system is to have an inner fast rotor current controller. The rotor

current is controlled with the machine-side converter. With the rotor current it is possible to

control the active and reactive powers. Then, the stator ﬂux is controlled using the grid-side

converter. The stator-ﬂux control loop is about a decade slower than the current control loop.

Then, ﬁnally, the dc-link voltage is controlled in cascade with the stator ﬂux in order to keep

the dc-link voltage constant.

As mentioned in the previous section, to be able to control the dc-link voltage with the

stator ﬂux there is a minimum rotor current. Therefore, when designing the control laws it

will be assumed that an additional power electronic device keeps the dc-link voltage constant

when i

Rq

is small. This means that when i

Rq

is below a certain value, ψ

sd

is not used to con-

trol the dc-link voltage anymore. For this case, the stator ﬂux can be controlled “arbitrarily,”

meaning that the stator ﬂux can be controlled so that the losses are reduced.

Rotor Current Control

The d component of the rotor current, i

Rd

, is used to control the reactive power while the q

component of the rotor current, i

Rq

, controls the active power or the torque, for details see

Chapter 4.

Stator-Flux Control

The stator voltage equation (7.70) can be rewritten as

E

g

+v

c

= −R

s

i

ref

R

+

dΨ

s

dt

+

R

s

L

M

+ jω

1

Ψ

s

(7.90)

where the rotor current has been put to its reference value. Then, v

c

= v

c

−E

g

+(jω

1

−Ω

a

)Ψ

s

is chosen where Ω

a

is the “active damping.” The above equation then reduces to

v

c

= −R

s

i

ref

R

+

dΨ

s

dt

+

R

s

L

M

+ Ω

a

Ψ

s

. (7.91)

123

The term R

s

i

ref

R

is treated as a disturbance and the transfer function from v

c

to Ψ

s

is found

as

G(p) =

Ψ

s

(p)

v

c

(p)

=

1

p + R

s

/L

M

+ Ω

a

. (7.92)

IMC yields the following PI controller tuned for a closed-loop bandwidth α

f

F(p) =

α

f

p

G

−1

(p) = α

f

+ α

f

R

s

/L

M

+ Ω

a

p

. (7.93)

If the “active damping” is set to Ω

a

= α

f

− R

s

/L

M

a disturbance is damped with the

same bandwidth as the closed-loop stator-ﬂux control loop, i.e., the transfer function from a

disturbance, D(p), to Ψ

s

is

G

DΨ

s

(p) =

Ψ

s

(p)

D(p)

=

p

(p + α

f

)

2

. (7.94)

DC-Link Voltage Control

For a dc-link voltage controller with a shunt converter, see Section 4.5.2. If the resistive

losses are treated as a disturbance, the dc-link dynamics in (7.72) can be written as

C

dc

2

dv

2

dc

dt

= 3i

Rq

(ψ

sd

ω

r

−E

g

) + D (7.95)

where D is the disturbance. Moreover, if the variable substitution W = v

2

dc

is made, the

following system is obtained

C

dc

2

dW

dt

= 3i

ref

Rq

ψ

ref

sd

ω

r

−E

g

+ D. (7.96)

where i

Rq

and ψ

sd

are put to their reference values. By choosing the reference value of the

ﬂux as

ψ

ref

sd

=

ψ

ref

sd

−G

a

W

i

ref

Rq

ω

r

+

E

g

ω

r

(7.97)

the dc-link dynamics are reduced to

C

dc

2

dW

dt

= 3ψ

ref

sd

−3G

a

W + D. (7.98)

In (7.97), G

a

is the “active damping.” Then, the transfer function, treating D as a disturbance,

becomes

G(p) =

W(p)

ψ

ref

sd

=

3

C

dc

/2p + 3G

a

. (7.99)

By using IMC, we obtain the following PI controller

F(p) =

α

dc

p

G

−1

(p) =

C

dc

α

dc

6

+

G

a

α

dc

p

. (7.100)

Then, if G

a

= C

dc

α

dc

/6, the transfer function from D to W becomes

W(p)

D(p)

=

G

1 + FG

=

p

C

dc

2

(p + α

dc

)

2

. (7.101)

This means that a disturbance is damped with the same bandwidth as the dc-link voltage

control loop.

124

Simulation of Electromechanical Torque Steps

In the simulation shown here, it is assumed that the rotor speed can by the pitch mechanism

of the wind turbine, if desired, be controlled with a bandwidth of α

s

.

Fig. 7.24 shows a simulation of the investigated system during current (or torque) control

mode. After 50 ms, i

Rq

is stepped from −0.9 p.u. to −0.1 p.u., and between 0.5 s and 1.0 s

the rotor speed is ramped from 1.2 p.u. down to 0.8 p.u. using pitch control. Finally, at 1.25 s,

i

Rq

is stepped back to −0.9 p.u. It can be seen in the ﬁgure that the control system manages

0 0.5 1 1.5

−1

−0.5

0

0.5

0 0.5 1 1.5

−1

0

1

2

0 0.5 1 1.5

2.6

2.8

3

0 0.5 1 1.5

0.8

1

1.2

0 0.5 1 1.5

−0.5

0

0.5

0 0.5 1 1.5

−0.5

0

0.5

i

R

[

p

.

u

.

]

a)

Ψ

s

[

p

.

u

.

]

b)

v

d

c

[

p

.

u

.

]

c)

ω

r

[

p

.

u

.

]

d)

v

R

[

p

.

u

.

]

e)

Time [s] Time [s]

v

c

[

p

.

u

.

]

f)

d d

d

d

q q

q

q

Fig. 7.24. Simulation of the system when the DFIG is in current control (or torque) mode. a) Rotor

current. b) Stator ﬂux. c) DC-link voltage. d) Rotor speed. e) Rotor voltage. f) Series

voltage.

to control i

Rd

, i

Rq

, and v

dc

well. Moreover, the simulation veriﬁes the result previously

presented in Section 7.4.2, which indicated that the q component of the rotor voltage and

converter voltage are close to identical. There is a small difference in the d component due

to the fact that the rotor converter also supplies the magnetizing current.

7.4.4 Speed Control Operation

At low wind speeds, the pitch angle of the turbine is ﬁxed and the DFIG is operated in speed

control operation. In this section, a rotor speed control law will be derived using IMC. The

125

mechanical dynamics are given by

J

n

p

dω

r

dt

= T

e

−T

s

(7.102)

where T

e

is the electromechanical torque, T

s

is the shaft torque, J is the inertia and, n

p

is the

number of pole pairs. Assuming ψ

sq

= 0, the electromechanical torque can be expressed as

T

e

= −3n

p

ψ

sd

i

Rq

. Then, since ψ

sd

≈ E

g

/ω

r

, (7.102) can be expressed as

J

n

p

dω

r

dt

= −3n

p

E

g

ω

r

i

ref

Rq

−T

s

(7.103)

where i

Rq

has been changed to its reference value. Now, we choose

i

ref

Rq

=

ω

r

3n

p

E

g

i

ref

Rq

+ B

a

ω

r

(7.104)

where B

a

is the “active damping.” This means that the mechanical dynamics can be rewritten

as

J

n

p

dω

r

dt

= −i

ref

Rq

−B

a

ω

r

−T

s

. (7.105)

Then, with IMC, the following controller is obtained

F(p) = k

p

+

k

i

p

= −α

s

J

n

p

−α

s

B

a

p

(7.106)

and if B

a

= α

s

J/n

p

, a change in T

s

is damped with the same bandwidth, α

s

, as the speed

control loop:

G

Tω

(p) =

P

J

n

p

(p + α

s

)

2

(7.107)

Fig. 7.25 shows an example of the proposed DFIG series system with the DFIG operated

in speed-control mode. In order to validate the performance, the machine is exposed to shaft

torque steps of 30% to 100% of rated torque. These torque steps are much faster than what

would be the case in reality and they are performed for veriﬁcation purposes only. The rotor

speed is controlled by the DFIG to be 0.8 p.u. during the whole simulation. It can be seen in

the ﬁgure that the control system manages to control i

Rd

, ω

r

, and v

dc

well.

7.4.5 Response to Voltage Sags

Fig. 7.26 shows the response to a 0% voltage sag, i.e., the remaining voltage is 0 p.u. The

voltage drops after 0.1 s and the sag has a duration of 250 ms. This is an extreme voltage sag

and if the system manages this sag, it manages the Swedish transmission system operator’s

demands for large production facilities [96]. During the simulation, the DFIG is operated

at i

Rq

= −1 p.u. which corresponds to generator operation at rated current (full power). In

this case, the rotor speed is controlled by the pitch mechanism to 1.2 p.u. During the sag,

the stator ﬂux is controlled by the series-connected converter to be close to zero. The rotor

126

0 0.5 1 1.5 2

0

0.25

0.5

0 0.5 1 1.5 2

−0.5

0

0.5

1

1.5

0 0.5 1 1.5 2

2.79

2.8

2.81

2.82

0 0.5 1 1.5 2

0.5

0.75

1

i

R

[

p

.

u

.

]

a)

Ψ

s

[

p

.

u

.

]

b)

d

d

q

q

v

d

c

[

p

.

u

.

]

c)

ω

r

[

p

.

u

.

]

d)

Time [s] Time [s]

Fig. 7.25. Simulation of the system when the DFIG is in speed-control mode. a) Rotor current.

b) Stator ﬂux. c) DC-link voltage. d) Rotor speed.

current is practically constant during the sag. Some minor current transients can be observed

at the instant of the sag and at the instant where the voltage returns. The main reason for

this is that both the rotor and converter voltages have been limited to their maximum values.

Otherwise, the dynamic performance of the system is promising.

Although the sag in the simulation is only 250 ms, the system can stay connected to the

grid for indeﬁnitely long voltage sags. This can be realized from the fact that the stator ﬂux is

controlled down to an appropriate level. The system then returns to a steady-state operating

condition, in this case at a voltage level of 0 p.u. However, one issue that must be kept

in mind is that, since the stator ﬂux is reduced according to the voltage sag, the maximum

torque that can be handled by the generator is reduced in proportion to the voltage sag. This

means that the pitch mechanism must reduce the incoming torque accordingly, as otherwise

overspeed occurs and the overspeed protection trips the turbine.

7.4.6 Steady-State Performance

As mentioned in the Introduction, the main reasons for choosing a DFIG system are cost and

efﬁciency. Therefore, when modifying the DFIG system it is necessary to evaluate how the

modiﬁcations affects both cost and efﬁciency. In Chapter 3, the average efﬁciency of the or-

dinary DFIG WT system has been calculated and compared to other electrical conﬁgurations

used in wind turbine systems. This study serves as a basis for the efﬁciency calculations and

comparisons in this section. Details of the calculations methods used here are described in

Chapter 3. Moreover, the systems are compared to an ordinary DFIG system and a system

that utilizes a full-power converter.

In this section, the efﬁciency will be calculated for the two options presented in the last

part of Section 7.4.2.

127

0 0.1 0.2 0.3 0.4 0.5

−2

−1

0

1

0 0.1 0.2 0.3 0.4 0.5

−1

0

1

2

0 0.1 0.2 0.3 0.4 0.5

2.5

3

3.5

4

0 0.1 0.2 0.3 0.4 0.5

1

1.1

1.2

1.3

0 0.1 0.2 0.3 0.4 0.5

−0.5

0

0.5

0 0.1 0.2 0.3 0.4 0.5

−0.5

0

0.5

i

R

[

p

.

u

.

]

a)

Ψ

s

[

p

.

u

.

]

b)

v

d

c

[

p

.

u

.

]

c)

ω

r

[

p

.

u

.

]

d)

v

R

[

p

.

u

.

]

e)

Time [s] Time [s]

v

c

[

p

.

u

.

]

f)

d d

d

d

q q

q

q

Fig. 7.26. Response to a 0% symmetrical voltage sag. a) Rotor current. b) Stator ﬂux. c) DC-link

voltage. d) Rotor speed. e) Rotor voltage. f) Series voltage.

Series-Injection Transformer

Fig. 7.27 shows the converter losses for the ordinary DFIG system, and in addition it presents

the losses of the series-injection transformer for the two options of the DFIG series system

presented earlier.

In Fig. 7.28, the average efﬁciencies of the ordinary DFIG system, the series system

with the two options, and a system with a full-power converter are shown as functions of

the average wind speed. As seen in Fig. 7.28, the efﬁciency of the standard DFIG system

is highest. The efﬁciency of the DFIG series system with Option 1 is roughly same as the

system that utilizes a full-power converter, although the efﬁciency is slightly lower at low

average wind speeds and slightly higher at higher average wind speeds. The efﬁciency of

the DFIG series system with Option 2 is between the ordinary DFIG system and the full-

power converter system. Accordingly, this is the most energy efﬁcient system with voltage

sag ride-through facility.

Converter in the Y Point of the DFIG

Fig. 7.29 shows the magnetizing and resistive losses of the generator and the converter losses

for the ordinary DFIG system and for the two options of the DFIG series system.

In Fig. 7.30, the efﬁciencies of the ordinary DFIG system, full-power converter system,

128

5 10 15 20 25

0

0.5

1

1.5

5 10 15 20 25

0

0.5

1

1.5

2

5 10 15 20 25

0

0.5

1

1.5

5 10 15 20 25

0

0.5

1

M

a

g

n

e

t

i

z

i

n

g

l

o

s

s

e

s

[

%

]

a)

R

e

s

i

s

t

i

v

e

l

o

s

s

e

s

[

%

] b)

C

o

n

v

e

r

t

e

r

l

o

s

s

e

s

[

%

]

c)

T

r

a

n

s

f

o

r

m

e

r

l

o

s

s

e

s

[

%

]

Wind speed [m/s] Wind speed [m/s]

d)

Fig. 7.27. Losses of the system with a series-injection transformer with the same turns ratio as the

stator-to-rotor turns ratio. Dashed lines correspond to the series DFIG system with Option

1, solid line to the series DFIG with Option 2, and dotted line to the ordinary DFIG system.

a) Magnetizing loses of the generator. b) Resistive losses of the generator. c) Converter

losses. d) Transformer losses.

5 6 7 8 9 10

90

91

92

93

94

95

96

E

f

ﬁ

c

i

e

n

c

y

[

%

]

Average wind speed [m/s]

Option 1

Option 2

Fig. 7.28. Expected efﬁciency as a function of the average wind speed for the system with a series-

injection transformer. Dashed line corresponds to the ordinary DFIG system, solid line to

the DFIG series system, and dotted line to a system with a full-power converter.

and the two different options for the series DFIG system are shown as functions of the av-

erage wind speed. If Fig. 7.30 is compared to Fig. 7.28, it can be seen that the results are

almost identical when connecting the converter to the Y point of the stator circuit. One rea-

son for this is that the increased losses in the converter are almost the same as the losses of

129

5 10 15 20 25

0

0.5

1

1.5

2

C

o

n

v

e

r

t

e

r

l

o

s

s

e

s

[

%

]

Wind speed [m/s]

Fig. 7.29. Converter losses when the converter is connected to the Y point of the stator circuit of

the DFIG. Dashed line is the series DFIG system with Option 1, solid line is the series

DFIG with Option 2, and dotted line is the ordinary DFIG system. The generator losses

are identical to that of Fig. 7.27.

5 6 7 8 9 10

90

91

92

93

94

95

96

E

f

ﬁ

c

i

e

n

c

y

[

%

]

Average wind speed [m/s]

Option 1

Option 2

Fig. 7.30. Expected efﬁciency as a function of the average wind speed. The converter is connected to

the Y point of the stator circuit of the DFIG. Dashed line corresponds to the ordinary DFIG

system, solid line to the DFIG series system, and dotted line to a system with a full-power

converter.

the series-injection transformer.

Energy Production Cost

Fig. 7.31 shows the relative energy cost of the DFIG series system in comparison to the

ordinary DFIG system. From an initial cost perspective, an extra converter for the DFIG

series system seems to be disadvantageous. However, as indicated in Fig. 7.31, from the

130

5 6 7 8 9 10

1

1.01

1.02

1.03

1.04

1.05

R

e

l

a

t

i

v

e

e

n

e

r

g

y

c

o

s

t

[

p

.

u

.

]

Average wind speed [m/s]

Option 1

Option 2

Fig. 7.31. Energy cost of the DFIG series system. The energy cost is related to the ordinary DFIG

system. Solid lines are with the series-injection transformer, dashed lines are with the

converter connected to the Y point, and dashed-dotted line is the system that utilize a

full-power converter.

energy cost point of view it is beneﬁcial to use the extra (third) converter to control the dc-

link voltage. In the ﬁgure it is shown that the increased energy cost for this series system

using Option 2 is approximately 1.5 percentage units. Moreover, as seen in the ﬁgure, for

the system with a full-power converter the corresponding energy cost is approximately 1.5

percentage unit higher than for the series system with Option 2.

7.4.7 Discussion and Conclusion

A control law for the doubly-fed induction generator with the grid side converter connected

in series with the stator circuit has been derived. The rotor current (torque and power factor),

stator ﬂux, and dc-link voltage are controlled. Simulations showed that the dynamic perfor-

mance of the system is promising both during normal operation and during conditions when

voltage sags are present in the grid. The derived control law is not capable of controlling

the dc-link voltage at very low loads. As a remedy for this, two different options using an

additional converter to solve this problem have been proposed and investigated. It was found

that the best option was to use an additional converter for controlling the dc-link voltage in

the whole operating area. Then, the series-connected converter can be used to control the

ﬂux to an optimal value from an overall efﬁciency point of view.

Two different methods of connecting the series converter resulted in almost the same

efﬁciency. The efﬁciency of the DFIG series system with the best performance was found

to be between the ordinary DFIG system and a system that utilizes a full-power converter

system, i.e., a cage-bar induction generator equipped with a back-to-back converter.

131

7.5 Conclusion

In this chapter, voltage sag ride-through of variable-speed wind turbines has been investi-

gated. It has been shown that a variable-speed wind turbine with a full-power converter

system, e.g., a cage-bar induction generator with a back-to-back converter, can successfully

reduce disturbances from both symmetrical and unsymmetrical voltage sags. Two candidate

methods, with one shunt-connected and one series-connected grid-side converter respec-

tively, of improving the voltage sag ride-through of DFIG variable-speed wind turbines have

also been investigated. The shunt connected DFIG system with ride-through capabilities still

suffers, at least initially, from high fault currents, while the series-connected DFIG system

seems to have similar dynamic performance as the full-power converter system. However,

the control of the DFIG series system is much more complicated than that of the full-power

converter system. Another drawback of the series-connected DFIG system in comparison

to the full-power converter system is that the maximum torque that can be handled by the

generator is reduced in proportion to the voltage sag.

The energy production cost of the full-power converter system was found to be three per-

centage units higher than that of the ordinary DFIG system. The shunt DFIG system and the

series system have approximately the same energy production cost, which is approximately

1.5 percentage unit higher compared to the ordinary DFIG system.

132

Chapter 8

Flicker Reduction of Stalled-Controlled

Wind Turbines using Variable Rotor

Resistances

Although there will be very large wind power installations, the installations of small-scale

wind turbines (WTs) will most likely proceed. Small WTs, 1 MW and below, have been de-

veloped successfully using the ﬁxed-speed stall-regulated concept, and will probably domi-

nate the small-turbine market also in the near future. Worth pointing out is that ﬁxed-speed

WTs have the same energy production given a certain rotor diameter as variable-speed WTs

(see Chapter 3).

The power quality impact, for instance the ﬂicker (or voltage ﬂuctuations) contribution,

of WTs is an important concern for grid owners. For individual installations of these types

of WTs, the ﬂicker contribution can be the limiting factor from a power quality point of

view, especially in weak grids [64]. One possibility to reduce ﬂicker from a stall-controlled

WT with an induction generator (IG) directly connected to the grid could be to introduce a

variable rotor resistance. In other words, the rotor resistance could be used to control the

rotor speed in a limited range and, in this way, absorb torque ﬂuctuations and thereby reduce

the ﬂicker emission. The purpose of this chapter is to derive a rotor resistance control law,

with the objective of minimizing torque ﬂuctuations and ﬂicker, for a stall-controlled WT.

8.1 Modeling

In Fig. 8.1, the system with turbine, gearbox, generator, and external rotor resistances, is

presented. It is possible to control the slip of the IG with the external rotor resistances. The

value of the external rotor resistances is adjusted with the power electronic equipment. How-

ever, in this chapter the power electronic equipment is not included in the model, i.e., it is

assumed to be ideal. Therefore, the external rotor resistances can be treated as a continuous

variable.

One way of representing the IG dynamically is to the use the so called Γ model as de-

133

IG

Gear-

box

External

rotor resistances

Grid

Fig. 8.1. Wind turbine with variable-rotor-resistance induction generator.

scribed in Section 4.2.1. The mechanical dynamics are described by

J

n

p

dω

g

dt

= T

g

−

T

t

g

r

(8.1)

where T

g

is the electromechanical torque produced by the generator, T

t

is the torque pro-

duced by the turbine, on the low-speed side of the gearbox, and g

r

is the gear ratio of the

gearbox. The drive train (soft axis) is not included in the model, since the objective is to

investigate the relative performance of the derived control law and, for instance, absolute

ﬂicker values are of minor importance.

For the 1-MW IG considered in this chapter, operated at 690 V and 50 Hz the following

parameters are used: R

s

=0.007 p.u., R

max

R

=0.05 p.u., R

min

R

=0.01 p.u., R

avg

R

=0.03 p.u., L

M

=5

p.u., L

σ

=0.2 p.u., n

p

=2, g

r

=61, and, J=32000 p.u. (without turbine J=3000 p.u.).

8.1.1 Reduced-Order Model

A common way to reduce the order of the induction machine model in (4.38) and (4.39) is

to neglect the stator-ﬂux dynamics. Then, the electrical dynamics of the induction machine

dynamics are described by (4.39). Eliminating ψ

R

from (4.39) yields

0 = (R

R

+ jω

2

L

σ

)i

R

+ L

σ

di

R

dt

+ jω

2

Ψ

s

.

(8.2)

Note that in the above equation, the stator-ﬂux dynamics have also been neglected. Further,

if i

Rd

can be assumed constant or at least small, and Ψ

s

≈ ψ

sd

≈

v

s

ω

1

, the dynamic system

reduces to

L

σ

di

Rq

dt

= −R

R

i

Rq

−(ω

1

−ω

g

)L

σ

i

Rd

−(ω

1

−ω

g

)

v

s

ω

1

(8.3)

J

n

p

dω

g

dt

= −k

T

i

Rq

−

T

t

g

r

(8.4)

where k

T

= 3v

s

n

p

/ω

1

. This means that the model has been reduced to the second order, i.e.,

one electrical and one mechanical equation.

In Fig. 8.2, simulations of the induction machine are presented, both with the ﬁfth-order

and the second-order model of the system. In the simulations, the rotor resistance is increased

by 40% after 50 ms and, after 250 ms, the shaft torque is increased from half of the rated

torque to rated. Note that in this simulation, only the inertia of the generator has been taken

into account and not the inertia of the turbine. This has been done in order to get a quicker

134

0 0.1 0.2 0.3 0.4

−0.5

0

0.5

1

0 0.1 0.2 0.3 0.4

−1.2

−1

−0.8

−0.6

−0.4

0 0.1 0.2 0.3 0.4

1.01

1.02

1.03

1.04

1.05

1.06

0 0.1 0.2 0.3 0.4

0.95

1

1.05

R

o

t

o

r

c

u

r

r

e

n

t

[

p

.

u

.

]

d

q

a)

T

o

r

q

u

e

[

T

n

o

m

]

b)

R

o

t

o

r

s

p

e

e

d

[

p

.

u

.

]

c)

S

t

a

t

o

r

ﬂ

u

x

[

p

.

u

.

]

Time [s] Time [s]

Time [s] Time [s]

d)

Fig. 8.2. Example of the response of the induction machine due to a step in the rotor resistance.

The rotor resistance is increased 40% after 50 ms and after 250 ms the shaft torque is

increased to the rated torque. Solid lines correspond to the ﬁfth-order model while dashed

lines correspond to the second-order model. a) Rotor current, b) Torque, c) Rotor speed and

d) Stator ﬂux.

response of the rotor speed and thereby a more lucid ﬁgure. The ﬁgure shows that both

models produce approximately the same results. However, there is a small deviation in the d

component of the rotor current. This reduced-order model will be used to derive the control

law.

8.2 Current Control

In order to remove the multiplication between the R

R

and i

Rq

, i.e., the term R

R

i

Rq

, in (8.3),

we will introduce the following non-linear control law

R

ref

R

=

R

R

+ R

Ra

i

Rq

i

Rq

i

Rq

= 0

(8.5)

where R

Ra

is an “active damping,” which can be used damp disturbances as described earlier.

How to chose R

Ra

will be described in the next section. Substitution of the above control

135

law in (8.3) yields

L

σ

di

Rq

dt

= −R

R

−R

Ra

i

Rq

+ D (8.6)

D = −(ω

1

−ω

t

)i

Rd

L

σ

−(ω

1

−ω

t

)

v

s

ω

1

= −ω

2

i

Rd

L

σ

+

v

s

ω

1

(8.7)

where a term D has been introduced. By treating the term D as a disturbance the following

open-loop transfer function can be found

G

ol

(p) =

i

Rq

(p)

R

R

(p)

=

−1

L

σ

p + R

Ra

. (8.8)

Then, by using IMC, the following current controller is obtained

F

c

(p) = k

pc

+

k

ic

p

= −L

σ

α

c

−

R

Ra

α

c

p

(8.9)

where α

c

is the closed-loop bandwidth of the current control loop. A block diagram of the

current control loop is shown in Fig. 8.3.

¸ ¸

R

Ra

i

ref

Rq

i

Rq

F

c

(p) 1/i

Rq

G

ol

(p)

R

ref

R

+

+

+

−

Fig. 8.3. Current Control Block Diagram.

Determination of the Active Damping

The transfer function, from a disturbance D to the current i

Rq

, is found as

G

D,i

Rq

(p) =

−p

L

σ

p

2

+ (R

Ra

+ L

σ

α

c

)p + R

Ra

α

c

. (8.10)

If R

Ra

= L

σ

α

c

, the above transfer function is reduced to

G

D,i

Rq

(p) =

−p

L

σ

(p + α

c

)

2

. (8.11)

This choice of R

Ra

causes a disturbance to be damped with the same time constant as the

current control loop. A Bode diagram of (8.11) can be seen in Fig. 8.4 for three different

values of the current control loop bandwidth α

c

.

136

0 50 100 150 200 250

10

−3

10

−2

10

−1

10

0

10

1

10

2

G

a

i

n

Frequency [Hz]

Fig. 8.4. Bode diagram. Solid α

c

=22 rad/s, dashed α

c

=220 rad/s and dotted α

c

=2200 rad/s.

0 0.05 0.1 0.15 0.2 0.25 0.3

0.4

0.5

0.6

0 0.05 0.1 0.15 0.2 0.25 0.3

−0.65

−0.6

−0.55

−0.5

−0.45

0 0.05 0.1 0.15 0.2 0.25 0.3

0

0.02

0.04

0.06

i

R

q

[

p

.

u

.

]

a)

T

g

,

T

s

[

T

n

o

m

]

b)

R

R

[

p

.

u

.

]

Time [s]

c)

Fig. 8.5. Example of current control of an IG with external rotor resistances. a) q component of

the rotor current (dashed line is the reference value), b) Torque (dashed line is the shaft

torque), c) Rotor resistance (dashed line is the minimum, average and maximum value of

the available rotor resistance).

137

8.2.1 Evaluation

Fig. 8.5 shows a simulation with the above derived rotor current control law. In the simulation

the bandwidth of the current control loop is set to 220 rad/s which corresponds to a (10–90%)

rise time of 10 ms. In the ﬁgure, it is seen that the current controller manages to control the

rotor current with the desired bandwidth. Moreover, the controller manages to keep the

generator torque at the shaft torque step (at 150 ms) until the current reference is adjusted

according to the new shaft torque (at 250 ms). However, the rotor resistance varies over its

entire range even for small current variations and shaft torque steps, as seen in the ﬁgure. It

is also seen that when the shaft and generator torques differ (between 150–250 ms), the rotor

resistance is constantly increased (or decreased for opposite sign of the torque difference).

If the rotor resistance has to be limited, the current controller will not manage to keep the

rotor current and thereby the generator torque. Because of the limited range in which the

rotor resistance can vary, the setting of the current reference will be of great importance for

the over-all performance of the system. How to set the rotor current reference will be further

addressed in the next section. First, however, in this section, a brief analytical investigation

of how the rotor resistance varies due to a shaft torque step is made.

By controlling the rotor current with a high-gain feedback, the rotor-current dynamics in

(8.3) can be expressed as

L

σ

di

Rq

dt

= −R

ref

R

i

Rq

−ω

2

L

σ

i

Rd

+

v

s

ω

1

= 0. (8.12)

This implies that the rotor-resistance reference value varies as

R

ref

R

= −

ω

2

i

ref

Rq

L

σ

i

Rd

+

v

s

ω

1

≈ −

ω

2

i

ref

Rq

v

s

ω

1

i

ref

Rq

= 0.

(8.13)

From (8.13), it is seen that the rotor resistance is depending on the slip, ω

2

, and the operating

condition, i.e. i

ref

Rq

. Moreover, if the generator is exposed to a shaft-torque step, ΔT

t

, the

generator speed becomes according to (8.1)

J

n

p

dω

g

dt

= −

ΔT

t

g

r

(8.14)

if the system initially was in the steady state and the electromechanical torque, T

g

, is kept

constant. This means that (8.13) can be rewritten as

R

ref

R

≈ −

v

s

ω

1

i

ref

Rq

n

p

Jg

r

ΔT

t

dt

i

ref

Rq

= 0

(8.15)

since ω

g

= ω

1

− ω

2

and dω

1

/dt = 0. The integral can be evaluated easily since ΔT

t

is

constant. This means that rotor resistance has changed ΔR

R

over the time

Δt = −

Jg

r

ω

1

n

p

v

s

i

ref

Rq

ΔT

t

ΔR

R

. (8.16)

138

For the shaft torque step at 150 ms in Fig. 8.5, the increase in rotor resistance (ΔR

R

= 0.024)

would, according to the above formula, take 0.2 s, which also can be seen in the ﬁgure.

Moreover, if ±ΔR

max

R

is the maximum available rotor resistance, the time, Δt

lim

, to reach

maximum or minimum value of the rotor resistance becomes

Δt

lim

=

Jg

r

ω

1

n

p

v

s

i

ref

Rq

ΔT

t

ΔR

max

R

. (8.17)

This means that for a given step in the shaft torque, the time until the rotor resistance must

be limited depends on i

ref

Rq

. That is, smaller values of i

ref

Rq

imply a shorter time until the rotor

resistance must be limited. This nonlinearity makes the setting of the rotor current reference

i

ref

Rq

more difﬁcult.

8.3 Reference Value Selection

In the steady state, the rotor resistance should be (or at least close to) its desired value, R

R0

.

One idea is to set i

ref

Rq

as

i

ref

Rq

= k

R

(R

R0

−R

R

)dt −B

a

ω

2

(8.18)

where only an integration term of the error in the rotor resistance is used in order to avoid an

algebraic loop. If the current control loop is fast, i.e., i

Rq

= i

ref

Rq

= k

R

I − B

a

ω

2

, where I is

the integration of the error in the rotor resistance, the system becomes

J

n

p

dω

2

dt

= k

T

i

ref

Rq

+

T

t

g

r

= k

T

(k

R

I −B

a

ω

2

) +

T

t

g

r

(8.19)

dI

dt

= R

R0

−R

R

. (8.20)

Note that the slip dynamics are found from (8.1), ω

g

= ω

1

−ω

2

and dω

1

/dt = 0. Moreover,

since the bandwidth of the current control loop is fast, it can be assumed that R

R

= R

ref

R

.

Therefore, according to (8.13), R

R

equals to

R

R

= R

ref

R

≈ −

ω

2

i

ref

Rq

v

s

ω

1

= −

ω

2

k

R

I −B

a

ω

2

v

s

ω

1

. (8.21)

This means, ﬁnally, that the following system must be analyzed

J

n

p

dω

2

dt

= k

T

(k

R

I −B

a

ω

2

) +

T

t

g

r

(8.22)

dI

dt

= R

R0

+

ω

2

k

R

I −B

a

ω

2

v

s

ω

1

. (8.23)

The above system has an equilibrium point at

ω

2,0

=

R

R0

T

t0

ω

1

g

r

k

T

v

s

(8.24)

I

0

=

T

t0

(ω

1

R

R0

B

a

−v

s

)

g

r

k

R

k

T

v

s

. (8.25)

139

Linearization and insertion around the equilibrium point yields

Δ˙ x =

⎡

⎣

−B

a

k

T

n

p

/J k

R

k

T

n

p

/J

g

r

k

T

(ω

1

R

R0

B

a

−v

s

)

T

t0

ω

1

−g

r

k

R

k

T

R

R0

T

t0

⎤

⎦

Δx +

¸

n

p

g

r

J

0

¸

Δu (8.26)

where

Δx =

¸

Δω

2

ΔI

Δu = ΔT

t

. (8.27)

Now, it is interesting to see how a change in the incoming torque inﬂuences the rotor resis-

tance. Therefore, one option is to study the error in the rotor resistance, e = R

R0

− R

R

.

However, e cannot be found directly from the state variables but since I is the integration of

e, it is possible to use the derivative of I. This means that

G

T

t

e

= p G

T

t

I

(p) (8.28)

where G

T

t

I

(p) is the transfer function from T

t

to I which can be found from the system in

(8.26). If k

R

and B

a

are chosen as

k

R

=

α

2

R

JT

t0

ω

1

g

r

k

2

T

n

p

v

s

(8.29)

B

a

= −

α

2

R

J

2

R

R0

ω

1

−2a

R

Jk

T

n

p

v

s

k

2

T

n

2

p

v

s

(8.30)

where α

R

is a parameter that can be set “freely,” the above transfer function G

T

t

e

(p) becomes

G

T

t

e

(p) = −

(k

T

n

p

v

s

−α

R

JR

R0

ω

1

)

2

Jk

T

n

p

v

s

ω

1

T

t0

p

(p + α

R

)

2

(8.31)

which is a band-pass ﬁlter centered at α

R

. Moreover, the damping of the above transfer

function and the parameter k

R

is dependent on the operating condition, i.e., T

t0

.

8.3.1 Evaluation

For a given operating condition it possible to express (8.31) as

G

T

t

e

(p) = K

p

(p + α

R

)

2

(8.32)

where K is a constant that depends on the operating condition. If the system is exposed to a

step, we will get

e(t) = L

−1

¸

1

p

G

T

t

e

(p)

= Kte

−α

R

t

(8.33)

where L is the Laplace transformation symbol. From the above equation, it is seen that

after a torque step, the rotor resistance returns to its desired value R

R0

, i.e., e(t →∞) = 0.

140

0 5 10 15 20

0.2

0.4

0.6

0.8

1

0 5 10 15 20

−1

−0.8

−0.6

−0.4

−0.2

0 5 10 15 20

0

0.02

0.04

0.06

i

R

q

[

p

.

u

.

]

a)

T

g

,

T

s

[

T

n

o

m

]

b)

R

R

[

p

.

u

.

]

Time [s]

c)

Fig. 8.6. Example of outer reference selection control loop. a) q component of the rotor current

(dashed line is the reference value), b) Torque (dashed line is the shaft torque), c) Rotor

resistance (dashed line is the minimum, average and maximum value of the available rotor

resistance).

Moreover, by looking at the derivative of the above function it is possible to determine that

the function has a maximum at

t

max(R

R

)

=

1

α

R

. (8.34)

Fig. 8.6 shows a simulation of the system with the reference selection control loop. The

bandwidth of the current control loop is set to a high value (2200 rad/s) and the parameter α

R

is set to 1 rad/s. It is seen in the ﬁgure that after the torque step (at t = 1 s) the rotor resistance

has its maximum value after 1 s (at t = 2 s), which is also veriﬁed by the expression (8.34).

Moreover, after the torque step the rotor resistance is returning to its desired value.

8.4 Evaluation

In order to evaluate the derived control law, the ﬂicker emission is compared to a similar

system with uncontrolled rotor resistances, i.e., R

R

is ﬁxed. Flicker emission or rapid voltage

ﬂuctuations can be described with the dimensionless quantity P

st

: the short-term severity

index. In the standard IEC 61000-21, it is described how this value is determined [51]. The

141

system is simulated for 10 minutes, since the 10 minute P

st

-value is used. The applied shaft

torque has been precalculated using blade element momentum theory with different average

wind speeds and turbulence intensities. Then the P

st

value has been calculated on a ﬁctive

grid with a short-circuit power of 50 times the nominal power of the WT and with an X/R

ratio of 0.5. The average torque, corresponding to the average wind speed, for each 10 minute

period is used to set the parameters that are dependent on the operating condition, i.e., k

R

.

Naturally, in a real system, they can be adjusted according to a changing operating condition.

However, since this should be done on a much slower time scale than the bandwidths of the

control loops, it has been ignored in the simulation presented here.

Fig. 8.7 shows an example of how the derived rotor resistance control law operates for a

short piece of one of the above mentioned 10-minute simulation. The average wind speed in

the 10 minute simulation was 14 m/s and the turbulence intensity was 25%. The bandwidth

0 5 10 15 20

−1.5

−1

−0.5

0

0 5 10 15 20

−4

−3

−2

−1

0

0 5 10 15 20

0

0.02

0.04

0.06

T

o

r

q

u

e

[

r

a

t

e

d

]

a)

S

l

i

p

[

%

]

b)

R

o

t

o

r

r

e

s

i

s

t

a

n

c

e

[

p

.

u

.

]

c)

Time [s]

Fig. 8.7. Example of the behaviour of the derived control law. a) Torque (generator torque is solid

and turbine torque is dashed), b) Slip and c) Rotor resistance.

of the current control loop, α

c

, is 2200 rad/s and the parameter α

R

, of the reference value

selection control loop is 0.5 rad/s. The set point value for the rotor resistance, R

R0

, has been

set to the average value of the available rotor resistance R

avg

R

.

142

8.4.1 Flicker Contribution

In Fig. 8.8, the P

st

value is seen for a system with ﬁxed rotor resistance and with the derived

control law as a function of the turbulence intensity and for different average wind speeds.

The control parameter is as in Fig. 8.7, except for the parameter α

R

that is 1 rad/s. In the

5 10 15 20 25

0

0.05

0.1

0.15

5 10 15 20 25

0

0.05

0.1

0.15

5 10 15 20 25

0

0.05

0.1

0.15

a)

b)

F

l

i

c

k

e

r

P

s

t

F

l

i

c

k

e

r

P

s

t

F

l

i

c

k

e

r

P

s

t

c)

Turbulence intenisity [%]

Fig. 8.8. Flicker as a function of the turbulence intensity. Solid line is WT with controlled rotor

resistance, dashed line is with ﬁxed rotor resistance, R

R

= R

avg

R

and dotted line is with

ﬁxed rotor resistance, R

R

= R

max

R

. The average wind speed is a) 6 m/s, b) 14 m/s and c) 20

m/s.

ﬁgure, the system with ﬁxed rotor resistance has been simulated with two different values

of the rotor resistance, i.e., the average value, R

avg

R

, and the maximum value (in continuous

operation), R

max

R

, of the available rotor resistance. It can be seen that the derived control

law produces lower P

st

values than the system with ﬁxed rotor resistance. Even though

the P

st

value for the ﬁxed rotor resistance system with R

R

= R

max

R

is close to the system

with controlled rotor resistances, it suffers from a drawback, namely, that the higher the

rotor resistance is, the higher the losses in the rotor resistance will be. These higher losses

imply that it will be necessary to increase the cooling of the generator. Finally, during the

simulation, the average value of the rotor resistance R

R

is very close to R

avg

R

.

143

8.4.2 Flicker Reduction

In Fig. 8.9 the relative ﬂicker contribution for the proposed controller for ﬁve different values

of a

R

is shown. The ﬂicker in the comparison is related to a system with a ﬁxed rotor

resistance. The rotor resistance of this system is set to the average value of the available

rotor resistance, i.e., R

R

= R

avg

R

. A relative ﬂicker of 1 corresponds to a ﬂicker contribution

equal to that of the ﬁxed rotor resistance system. Lower values of the relative ﬂicker imply

a lower ﬂicker contribution and vice versa. The relative ﬂicker is given as a function of

turbulence intensities for an average wind speed of 6 m/s. In general, it can be seen that the

5 10 15 20 25

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0.25 rad/s

0.5 rad/s

1 rad/s

1.5 rad/s

2 rad/s

R

e

l

a

t

i

v

e

ﬂ

i

c

k

e

r

Turbulence intensity [%]

Fig. 8.9. Reduction in ﬂicker for different bandwidths of α

R

. The average wind speed is 6 m/s.

lower the parameter α

R

is, the more reduction in the P

st

value is achieved. However, if the

frequency is too low or the turbulence intensity is too high, the rotor resistance will hit its

maximum or minimum value to a high extent which will make the result worse. For example,

for the case with α

R

put to 0.5 rad/s in the ﬁgure, the number of times the rotor resistance

has to be limited is rapidly increased from a turbulence intensity of 7% and upwards. For

the case with α

R

equal 0.25 rad/s the rotor resistance has been limited to its maximum or

minimum value between 20–70% of the total simulation time depending on the turbulence

intensity. Due to this fact, the P

st

value is actually worse for this case than for the case with

ﬁxed rotor resistance.

In Figs. 8.10 and 8.11 the corresponding diagrams for an average wind speed of 14 and

20 m/s are shown. It is seen that when the turbulence intensity becomes higher, for low

values of α

R

, the rotor resistance can not follow its reference value and has to be limited to a

higher and higher degree (i.e., the same phenomena as in Fig. 8.9). This will have a negative

impact on the performance.

As mentioned earlier, the damping of the ﬂicker (or the torque ﬂuctuation) is dependent

on the operating condition. This is also veriﬁed by the simulation since it is possible to

reduce more of the ﬂicker at higher average wind speeds (i.e., higher average torques). On

the other hand, the ﬂicker contribution is lower at lower average wind speeds. Moreover,

from the ﬁgures it can be seen that in order to have an “optimal” reduction in ﬂicker, over the

whole operating area, with the derived control law, the parameter α

R

should be a function of

144

5 10 15 20 25

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0.25 rad/s

0.5 rad/s

1 rad/s

1.5 rad/s

2 rad/s

R

e

l

a

t

i

v

e

ﬂ

i

c

k

e

r

Turbulence intensity [%]

Fig. 8.10. Reduction in ﬂicker for different bandwidths of α

R

. The average wind speed is 14 m/s.

5 10 15 20 25

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0.25 rad/s

0.5 rad/s

1 rad/s

1.5 rad/s

2 rad/s

R

e

l

a

t

i

v

e

ﬂ

i

c

k

e

r

Turbulence intensity [%]

Fig. 8.11. Reduction in ﬂicker for different bandwidths of α

R

. The average wind speed is 20 m/s.

both the average torque and turbulence intensity.

8.5 Conclusion

A non-linear rotor resistance control law has been derived with the objective of minimizing

the ﬂicker contribution of a stall-controlled ﬁxed-speed wind turbine to the grid.

It was shown that it is possible to reduce the ﬂicker contribution by utilizing the derived

rotor resistance control law with 40–80% depending on the operating condition. However,

since the rotor resistance can be varied only within a limited range, the reduction in the ﬂicker

contribution will be dependent on the operating condition. Moreover, the non-linearity of the

system will make an “optimal” reduction in ﬂicker, over the whole operating area, difﬁcult.

145

146

Chapter 9

Conclusion

The electrical energy efﬁciency of wind turbine systems equipped with doubly-fed induction

generators in comparison to other wind turbine generator systems has been investigated. It

was found that the energy efﬁciency of a doubly-fed induction generator system is a few per-

centage units higher compared to a system using a cage-bar induction generator, controlled

by a full-power converter. In comparison to a direct-driven permanent-magnet synchronous

generator, controlled by a converter or a two-speed generator system the difference in energy

efﬁciency was found to be small. Moreover, the converter losses of the doubly-fed induction

generator can be reduced if the available rotor-speed range is made smaller. However, the

aerodynamic capture of the wind turbine is reduced with a smaller rotor-speed range. This

means that the increased aerodynamic capture that can be achieved by a larger converter has,

thus, a greater impact than the increased converter losses. Finally, two methods to reduce the

magnetizing losses of the doubly-fed induction generator system, have been investigated. It

was found that the method, utilizing a Y-Δ switch in the stator circuit had the largest gain in

energy, of the two investigated methods.

In order to evaluate different methods of reducing the inﬂuence of the back EMF on the

rotor current control loop, a general rotor current control law has been derived with the op-

tion of having feed-forward compensation of the back EMF and “active resistance.” It was

found that the method that combines both the feed-forward compensation of the back EMF

and the “active resistance” manages to suppress the inﬂuence of the back EMF on the rotor

current best and was found to be the least sensitive to erroneous parameters. The choice of

current control method is of greater importance if the bandwidth of the current control loop

is low. Moreover, it has been shown that by using grid-ﬂux orientation, the stability and

the damping of the system is independent of the rotor current, in contrast to the stator-ﬂux

oriented system.

Dynamic models of the DFIG wind turbines have been experimentally veriﬁed, with a

850-kW wind turbine. Simulations and experimental results of the dynamic response to

symmetrical as well as unsymmetrical voltage sags of a DFIG wind turbine were presented.

Simulations were carried out both with a full-order model, and also with a reduced-order

(second-order) model. Both models produced acceptable results.

Voltage sag ride-through capabilities of some different variable-speed wind turbines have

been investigated and compared. A variable-speed wind turbine with a full-power converter

147

system can handle voltage sags very well. Two candidate methods for improving the voltage

sag ride-through capability of DFIG variable-speed wind turbines have been investigated.

One of the methods still suffers, at least initially, from high fault currents, while the other

method seems to have similar dynamical performance as the full-power converter system.

However, the control of the latter method is much more complicated than that of the full-

power converter system. In addition, the maximum torque that can be handled by the gener-

ator is reduced in proportion to the voltage sag. The energy production cost of the full-power

converter system was found to be three percentage units higher than that of the ordinary

DFIG system without ride through capability. The two DFIG candidate methods have ap-

proximately the same energy production cost, which is approximately 1.5 percentage units

higher in comparison to the ordinary DFIG system.

Finally, a non-linear rotor resistance control law has been derived with the objective of

minimizing the ﬂicker contribution of a stall-controlled ﬁxed-speed wind turbine to the grid.

It has been found that the ﬂicker contribution can be reduced with 40–80%, depending on

the operating condition, with the derived control law.

9.1 Future Research

The following candidate topics are proposed for future research:

• Development of a uniﬁed estimator for both stator-ﬂux and grid-ﬂux ﬁeld orientation.

Since the ﬂux dynamics are poorly damped, a desired property would be a relatively

good damping of the ﬂux dynamics.

• More thorough dynamic, steady-state, and experimental analysis of the voltage sag

ride-through systems for the DFIG wind turbine. In addition, it is essential to study

the hardware conﬁguration of the voltage sag ride-through systems.

• Development of mathematical models of wind turbines with voltage sag ride-through

properties. Experimental evaluation of the developed models with commercial wind

turbines with voltage sag ride-through properties.

• Derivation of analytical expressions for the response of the DFIG to unsymmetrical

voltage sags.

148

References

[1] SIMPOW Power System Simulation Software, Brochure, ABB Power Systems Analy-

sis, V¨ aster˚ as, Sweden.

[2] F. Abrahamsen, “Energy optimal control of induction motor drives,” Ph.D. disserta-

tion, Aalborg Univ., Aalborg, Denmark, Feb. 2000.

[3] T. Ackermann and L. S¨ oder, “An overview of wind energy-status 2002,” Renew. Sus-

tain. Energy Rev., vol. 6, no. 1–2, pp. 67–128, Feb./Apr. 2002.

[4] H. Akagi and H. Sato, “Control and performance of a doubly-fed induction machine

intended for a ﬂywheel energy storage system,” IEEE Trans. Power Electron., vol. 17,

no. 1, pp. 109–116, Jan. 2002.

[5] H. Akagi, Y. Kanazawa, and A. Nabae, “Instantaneous reactive power compensators

comprising switching devices without energy storage components,” IEEE Trans. Ind.

Applicat., vol. 20, no. 3, pp. 625–630, May/June 1984.

[6] J. Bendl, M. Chombt, and L. Schreier, “Adjustable-speed operation of doubly fed

machines in pumped storage power plants,” in Proc. Ninth International Conference

on Electrical Machines and Drives, Sep., 1–3, 1999, pp. 223–227.

[7] I. Boldea and S. A. Nasar, Electric Drives. CRC Press LCC, 1999.

[8] S. Bolik, “Grid requirements challenges for wind turbines,” in Proc. Int. Work. Large-

Scale Integration Wind Power Transmission Networks Offshore Wind Farms, Billund,

Denmark, Oct., 20–21, 2003.

[9] M. H. Bollen, Understanding Power Quality Problems: Voltags Sags and Interup-

tions. Piscataway, NJ, USA: IEEE Press, 2002.

[10] M. Bongiorno, “Control of voltage source converters for voltage dip mitigation in

shunt and series conﬁguration,” Chalmers University of Technology, G¨ oteborg, Swe-

den, Licentiate Thesis 515L, Nov. 2004.

[11] T. Burton, D. Sharpe, N. Jenkins, and E. Bossanyi, Wind Energy Handbook. John

Wiley & Sons, Ltd, 2001.

[12] O. Carlson, J. Hylander, and K. Thorborg, “Survey of variable speed operation of wind

turbines,” in Proc. of European Union Wind Energy Conference, G¨ oteborg, Sweden,

May, 20–24, 1996, pp. 406–409.

149

[13] L. Congwei, W. Haiqing, S. Xudong, and L. Fahai, “Research of stability of double fed

induction motor vector control system,” in Proc. of the Fifth International Conference

on Electrical Machines and Systems, vol. 2, Shenyang, China, Aug., 18–20, 2001, pp.

1203–1206.

[14] R. L. Cosgriff, Nonlinear Control Systems. McGraw-Hill, 1958.

[15] R. Datta and V. T. Ranganathan, “Asimple position-sensorless algorithmfor rotor-side

ﬁeld-oriented control of wound-rotor induction machine,” IEEE Trans. Ind. Electron.,

vol. 48, no. 4, pp. 786–793, Aug. 2001.

[16] ——, “Variable-speed wind power generation using doubly fed wound rotor induction

machine-a comparison with alternative schemes,” IEEE Trans. Energy Conversion,

vol. 17, no. 3, pp. 414–421, Sept. 2002.

[17] ——, “Decoupled control of active and reactive power for a grid-connected doubly-

fed wound rotor induction machine without position sensors,” in Proc. Conference

Record of the 1999 IEEE Industry Applications Conference, vol. 4, Phoenix, AZ,

USA, Oct. 1999, pp. 2623–2628.

[18] F. B. del Blanco, M. W. Degner, and R. D. Lorenz, “Dynamic analysis of current

regulators for ac motors using complex vectors,” IEEE Trans. Ind. Applicat., vol. 35,

no. 6, pp. 1424–1432, Nov./Dec. 1999.

[19] DEWIND. (2005, Jan.) The D8 series. Brochure. [Online]. Available: http:

//www.dewind.de/en/downloads/D8-2000-100-eng.pdf

[20] A. Dittrich and A. Stoev, “Grid voltage fault proof doubly-fed induction generator

system,” in Proc. Power Electronics and Applications (EPE), Toulouse, France, Sep.

2003.

[21] J. L. Duarte, A. V. Zwam, C. Wijnands, and A. Vandenput, “Reference frames ﬁt for

controlling pwm rectiﬁers,” IEEE Trans. Power Electron., vol. 46, no. 3, pp. 628–630,

June 1999.

[22] J. B. Ekanayake, L. Holdsworth, and N. Jenkins, “Comparison of 5th order and 3rd or-

der machine models for doubly fed induction generator (DFIG) wind turbines,” Elec-

tric Power Systems Research, vol. 67, pp. 207–215, Dec. 2003.

[23] Elforsk. (2005) Driftuppf¨ oljning av vindkraftverk december 2004. [Online].

Available: http://www.elforsk.se/varme/underlag/vstat0412.pdf

[24] Elsam. (2003, Oct.) Horns rev offshore wind farm. Brochure. [Online]. Available:

http://www.hornsrev.dk/nyheder/brochurer/Horns Rev GB.pdf

[25] (2004) Enercon website. [Online]. Available: http://www.enercon.de/

[26] A. Feij ´ oo, J. Cidr´ as, and C. Carrillo, “A third order model for the doubly-fed induction

machine,” Electric Power Systems Research, vol. 56, pp. 121–127, Nov. 2000.

150

[27] C. Fitzer, A. Arulampalam, M. Barnes, and R. Zurowski, “Mitigation of saturation

in dynamic voltage restorer connection transformers,” IEEE Trans. Power Electron.,

vol. 17, no. 6, pp. 1058–1066, Nov. 2002.

[28] J. Fortmann, “Validation of DFIG model using 1.5 MW turbine for the analysis of its

behavior during voltage drops in the 110 kV grid,” in Proc. Int. Work. Large-Scale

Integration Wind Power Transmission Networks Offshore Wind Farms, Billund, Den-

mark, Oct. 2003.

[29] GE Wind Energy. (2005, Jan.) 3.6s offshore wind turbine. Brochure. [On-

line]. Available: http://www.gepower.com/prod serv/products/wind turbines/en/

downloads/ge 36 brochure.pdf

[30] ——. (2005, Jan.) Low voltage ride-thru technology. Brochure. [Online].

Available: http://www.gepower.com/businesses/ge wind energy/en/downloads/ge

lvrt brochure.pdf

[31] T. Glad and L. Ljung, Reglerteori: ﬂervariabla och olinj¨ ara metoder. Lund, Sweden:

Studentlitteratur, 1997, (in Swedish).

[32] A. Grauers and S. Landstr¨ om, “The rectiﬁers inﬂuence on the size of direct-

driven generators,” in Proc. of European Wind Energy Conference and Exhibition

(EWEC´99), Nice, France, Mar., 1–5, 1999.

[33] A. Grauers, “Synchronous generator and frequency converter in wind turbine

applications: system design and efﬁciency,” Chalmers University of Technology,

G¨ oteborg, Sweden, Licentiate Thesis 175L, May 1994. [Online]. Available: http://

www.elteknik.chalmers.se/Publikationer/EMKE.publ/Abstracts/old/Grauers Lic.pdf

[34] ——, “Design of direct-driven permanent-magnet generators for wind turbines,”

Ph.D. dissertation, Chalmers University of Technology, G¨ oteborg, Sweden, Nov.

1996. [Online]. Available: http://www.elteknik.chalmers.se/Publikationer/EMKE.

publ/Abstracts/old/Grauers PhD Thesis.pdf

[35] ——, “Efﬁciency of three wind energy generator systems,” IEEE Trans. Energy Con-

version, vol. 11, no. 3, pp. 650–657, Sept. 1996.

[36] L. H. Hansen, L. Helle, F. Blaabjerg, E. Ritchie, S. Munk-Nielsen, H. Bindner,

P. Sørensen, and B. Bak-Jensen, “Conceptual survey of generators and power elec-

tronics for wind turbines,” Risø National Laboratory, Roskilde, Denmark, Tech. Rep.

Risø-R-1205(EN), ISBN 87-550-2743-8, Dec. 2001.

[37] L. Harnefors and H.-P. Nee, “A general algorithm for speed and position estimation

of ac motors,” IEEE Trans. Ind. Electron., vol. 47, no. 1, pp. 77–83, Feb. 2000.

[38] L. Harnefors, “On analysis, control and estimation of variable-speed drives,” Ph.D.

dissertation, Royal Institute of Technology, Stockholm, Sweden, 1997.

[39] ——, Control of Variable-Speed Drives. V¨ aster˚ as, Sweden: Department of Elec-

tronics, M¨ alardalen University, 2002.

151

[40] L. Harnefors and H.-P. Nee, “Model-based current control of ac machines using the

internal model control method,” IEEE Trans. Ind. Applicat., vol. 34, no. 1, pp. 133–

141, Jan./Feb. 1998.

[41] L. Harnefors, K. Pietil¨ ainen, and L. Gertmar, “Torque-maximizing ﬁeld-weakening

control: design, analysis, and parameter selection,” IEEE Trans. Ind. Electron.,

vol. 48, no. 1, pp. 161–168, Feb. 2001.

[42] S. Hartge and V. Diedrichs, “Ride-through capability of ENERCON-wind turbines,”

in Proc. Int. Work. Large-Scale Integration Wind Power Transmission Networks Off-

shore Wind Farms, Billund, Denmark, Oct. 2003.

[43] M. Heller and W. Schumacher, “Stability analysis of doubly-fed induction machines

in stator ﬂux reference frame,” in Proc. of 7th European Conference on Power Elec-

tronics and Applications, vol. 2, Brussels, Belgium, Sept., 8–10, 1997, pp. 707–710.

[44] K. Hentabli, M. E. H. Benbouzid, and D. Pinchon, “CGPC with internal model struc-

ture: Application to induction motor control,” in Proc. of the 1997 IEEE International

Conference on Control Applications, Hartford, CT, USA, Oct., 5–7, 1997, pp. 235–

237.

[45] N. G. Hingorani and L. Gyugyi, Understanding FACTS: Concepts and Technology of

Flexible AC Transmission Systems. Piscataway, NJ, USA: IEEE Press, 2000.

[46] B. Hopfensperger, D. J. Atkinson, and R. A. Lakin, “Stator-ﬂux-oriented control of a

doubly-fed induction machine with and without position encoder,” IEE Proc. Electr.

Power Appl., vol. 147, pp. 241–250, July 2000.

[47] B. Hopfensperger, D. Atkinson, and R. A. Lakin, “Stator ﬂux oriented control of a

cascaded doubly-fed induction machine,” IEE Proc. Electr. Power Appl., vol. 146,

no. 6, pp. 597–605, Nov. 1999.

[48] B. Hopfensperger and D. Atkinson, “Doubly-fed a.c. machines: classiﬁcation and

comparsion,” in Proc. Power Electronics and Applications (EPE), Graz, Austria,

Aug., 27–29, 2001.

[49] C.-J. Huang, S.-J. Huang, and F.-S. Pai, “Design of dynamic voltage restorer with

disturbance-ﬁltering enhancement,” IEEE Trans. Power Electron., vol. 18, no. 5, pp.

1202–1210, Sep. 2003.

[50] N. Hur, J. Jung, and K. Nam, “A fast dynamic dc-link power-balancing scheme for

a PWM converter–inverter system,” IEEE Trans. Ind. Electron., vol. 48, no. 4, pp.

794–803, Aug. 2001.

[51] Measurement and assessment of power quality characteristics of grid connected wind

turbines (11/2000), International Electrotechnical Commission Std. IEC 61 000-21,

2000.

[52] M. G. Ioannides and J. A. Tegopoulos, “Optimal efﬁciency slip-power recovery drive,”

IEEE Trans. Energy Conversion, vol. 3, no. 2, pp. 342–348, June 1988.

152

[53] G. L. Johnsson, Wind Energy Systems. Englewood Cliffs, N.J., USA.: Prentice-Hall,

1985.

[54] C. R. Kelber and W. Schumacher, “Active damping of ﬂux oscillations in doubly-

fed ac machines using dynamic variation of the system’s structure,” in Proc. Power

Electronics and Applications (EPE), Graz, Austria, Aug., 27–29 2001.

[55] C. R. Kelber, “Aktive d¨ ampfung der doppelt-gespeisten Drehstrommaschine,” Ph.D.

dissertation, Technishen Universit¨ at Carolo-Wilhelmina, 2000, (in German).

[56] E. H. Kim, S. B. Oh, Y. H. Kim, and C. H. Kim, “Power control of a doubly fed induc-

tion machine without rotational transducers,” in Proc. of 3rd International Conference

on Power Electronics and Motion Control, vol. 2, Beijing, China, Aug., 15–18, 2000,

pp. 951–955.

[57] P. Kokotovi´ c, H. K. Khalil, and J. O’Reilly, Singular Perturbation Methods in Control:

Analysis and Design. London: Academic Press, 1986.

[58]

˚

A. Larsson, P. Sørensen, and F. Santjer, “Grid impact of variable speed wind tur-

bines,” in Proc. of European Wind Energy Conference and Exhibition (EWEC´99),

Nice, France, Mar., 1–5, 1999.

[59] A. Lasson and P. Lundberg, “Trefas sj¨ alvkommuterade str¨ omriktare anslutna till

eln¨ atet—En inledande studie,” Chalmers University of Technology, G¨ oteborg, Swe-

den, Tech. Rep., 1990.

[60] P. Ledesma and J. Usaola, “Effect of neglecting stator transients in doubly fed induc-

tion generator models,” IEEE Trans. Energy Conversion, vol. 19, pp. 459–461, June

2004.

[61] W. Leonhard, Control of Electrical Drives, 2nd ed. Berlin, Germany: Springer-

Verlag, 1996.

[62] M. Lindgren, “Modeling and control of voltage source converters connected to the

grid,” Ph.D. dissertation, Chalmers University of Technology, G¨ oteborg, Sweden,

Nov. 1998. [Online]. Available: http://www.elteknik.chalmers.se/Publikationer/

EMKE.publ/Abstracts/old/Lindgren PhD Thesis.pdf

[63] B. Li, S. Choi, and D. Vilathgamuwa, “Transformerless dynamic voltage restorer,”

IEE Proceedings Generation, Transmission and Distribution, vol. 149, no. 3, pp.

1202–1210, May 2002.

[64] S. Lundberg, T. Petru, and T. Thiringer, “Electrical limiting factors for wind energy

installations in weak grids,” International Journal of Renewable Energy Engineering,

vol. 3, no. 2, pp. 305–310, Aug. 2001.

[65] S. Lundberg, “Performance comparison of wind park conﬁgurations,” Department

of Electric Power Enginering, Chalmers University of Technology, Department of

Electric Power Enginering, G¨ oteborg, Sweden, Tech. Rep. 30R, Aug. 2003.

153

[66] EMTDC Users’s Guide, On-line help, Manitoba HVDC Research Centre Inc., Win-

nipeg, Manitoba, Canada, 1998.

[67] N. Mohan, T. Undeland, and W. Robbins, Power Electronics Converter, Applications

and Design. New York, USA: John Wiley & Sons, 1995.

[68] L. Morel, H. Godfroid, A. Mirzaian, and J. Kauffmann, “Double-fed induction ma-

chine: converter optimisation and ﬁeld oriented control without position sensor,” IEE

Proc. Electr. Power Appl., vol. 145, no. 4, pp. 360–368, July 1998.

[69] P. Mutschler and R. Hoffmann, “Comparison of wind turbines regarding their energy

generation,” in Proc. 2002 IEEE 33rd Annual IEEE Power Electronics Specialists

Conference, vol. 1, Cairns, Qld., Australia, June, 23–27, 2002, pp. 6–11.

[70] M. J. Newman and D. G. Holmes, “An integrated approach for the protection of series

injection inverters,” IEEE Trans. Ind. Applicat., vol. 38, no. 3, pp. 679–687, May/June

2002.

[71] J. G. Nielsen, M. Newman, H. Nielsen, and F. Blaabjerg, “Control and testing of a dy-

namic voltage restorer (DVR) at medium voltage level,” IEEE Trans. Power Electron.,

vol. 19, no. 3, pp. 806–813, May 2004.

[72] J. Niiranen, “Voltage dip ride through of doubly-fed generator equipped with active

crowbar,” in Proc. Nordic Wind Power Conference (NWPC), G¨ oteborg, Sweden, Mar.,

1–2 2004.

[73] Nordex. (2005, Jan.) N80/2500 kW N90/2300 kW. Brochure. [Online]. Available:

http://www.nordex-online.com/ e/online service/download/ dateien/PB N80 GB.pdf

[74] R. Ottersten, A. Petersson, and K. Pietil¨ ainen, “Voltage sag response of PWM recti-

ﬁers for variable-speed wind turbines,” in Proc. IEEE Nordic Workshop on Power and

Industrial Electronics (NORpie´2004), Trondheim, Norway, June, 14–16 2004.

[75] R. Ottersten, “Vector control of a double-sided PWM converter and induction

machine drive,” Chalmers University of Technology, G¨ oteborg, Sweden, Licentiate

Thesis 368L, Dec. 2000. [Online]. Available: http://www.elteknik.chalmers.se/

Publikationer/EMKE.publ/Abstracts/old/otterstenLic/ottersten lic thesis.pdf

[76] ——, “On control of back-to-back converters and sensorless induction machine

drives,” Ph.D. dissertation, Chalmers University of Technology, G¨ oteborg, Sweden,

June 2003. [Online]. Available: http://www.elteknik.chalmers.se/Publikationer/

EMKE.publ/Abstracts/2003/RolfPhD.pdf

[77] M. P. Papadopoulos, S. A. Papathanassiou, N. G. Boulaxis, and S. T. Tentzerakis,

“Voltage quality change by grid-connected wind turbines,” in European Wind Energy

Conference, Nice, France, 1999, pp. 783–785.

[78] R. Pena, R. Cardenas, G. Asher, J. Clare, J. Rodriguez, and P. Cortes, “Vector control

of a diesel-driven doubly fed induction machine for a stand-alone variable speed en-

ergy system,” in IEEE Annual Conference of the Industrial Electronics Society, vol. 2,

Nov., 5–8, 2002, pp. 985–990.

154

[79] R. Pena, R. Cardenas, R. Blasco, G. Asher, and J. Clare, “A cage induction generator

using back to back PWM converters for variable speed grid connected wind energy

system,” in Proc. IEEE IECON’01, vol. 2, Denver, CO, Nov. 2001, pp. 1376–1381.

[80] R. Pena, J. C. Clare, and G. M. Asher, “Doubly fed induction generator using back-to-

back PWM converters and its application to variable-speed wind-energy generation,”

IEE Proc. Electr. Power Appl., vol. 143, pp. 231–241, May 1996.

[81] A. Petersson, “Analysis, modeling and control of doubly-fed induction generators for

wind turbines,” Chalmers University of Technology, G¨ oteborg, Sweden, Licentiate

Thesis 464L, Feb. 2003. [Online]. Available: http://www.elteknik.chalmers.se/

Publikationer/EMKE.publ/Abstracts/2003/AndreasLic.pdf

[82] T. Petru and T. Thiringer, “Active ﬂicker reduction from a sea-based 2.5 MW wind

park connected to a weak grid,” in Proc. Nordic Workshop on Power and Industrial

Electronics, Aalborg, Denmark, June, 13–16, 2002.

[83] ——, “Modeling of wind turbines for power system studies,” IEEE Trans. Power

Syst., vol. 17, pp. 1132–1139, Nov. 2002.

[84] M. A. P¨ oller, “Doubly-fed induction machine models for stability assessment of wind

farms,” in Proc. IEEE Bologna Power Tech, Bologna, Italy, June, 23–26 2003.

[85] PSS/E-Introduction to Dynamic Simulation, Manual, Power Technologies, Inc., Sch-

enectady, NY, U.S.A., 2001.

[86] B. Rabelo and W. Hofmann, “Optimal active and reactive power control with the

doubly-fed induction generator in the MW-class wind-turbines,” in Proc. Interna-

tional Conference on Power Electronics and Drives Systems, vol. 1, Denpasar, In-

donesia, Oct., 22–25, 2001, pp. 53–58.

[87] R. Richter, Electrische Machinen, 2nd ed. Basel/Stuttgart: Verlag Birkh¨ auser, 1954,

(in German).

[88] B. Schmidtbauer, Analog och digital reglerteknik, 2nd ed. Lund, Sweden: Studentlit-

teratur, 1995, (in Swedish).

[89] Semikron. (2004, Dec.) SKiiP 1203GB172-2DW. Data sheet. [Online]. Available:

http://www.semikron.com/internet/ds.jsp?ﬁle=1266.html

[90] ——. (2004, Dec.) SKiiP 1803GB172-3DW. Data sheet. [Online]. Available:

http://www.semikron.com/internet/ds.jsp?ﬁle=1264.html

[91] ——. (2004, Dec.) SKiiP 2403GB172-4DW. Data sheet. [Online]. Available:

http://www.semikron.com/internet/ds.jsp?ﬁle=1261.html

[92] ——. (2004, Dec.) SKiiP 513GD172-3DUL. Data sheet. [Online]. Available:

http://www.semikron.com/internet/ds.jsp?ﬁle=1268.html

[93] ——. (2004, Dec.) Thyristor SKT 2400. Data sheet. [Online]. Available:

http://www.semikron.com/internet/ds.jsp?ﬁle=516.html

155

[94] G. R. Slemon, “Modelling of induction machines for electric drives,” IEEE Trans. Ind.

Applicat., vol. 25, no. 6, pp. 1126–1131, Nov./Dec. 1989.

[95] J.-J. E. Slotline and W. Li, Applied Nonlinear Control. Upper Saddle River, New

Jersey, USA: Prentice Hall, 1991.

[96] Svenska Kraftn¨ at, “Aff¨ arsverket svenska kraftn¨ ats f¨ oreskrifter om

drifts¨ akerhetsteknisk utformning av produktionsanl¨ aggningar,” Svenska Kraftn¨ at,

V¨ allingby, Sweden, Tech. Rep., 2004, Draft version. [Online]. Available:

http://www.svk.se/upload/3645/Foreskriftprod remiss.pdf

[97] Swedish Energy Agency, “Climate report 2001,” Swedish Energy Agency, Tech.

Rep. ER 6:2002, 2002. [Online]. Available: http://www.stem.se/web/biblshop.nsf/

FilAtkomst/ER62002.pdf/$FILE/ER62002.pdf?OpenElement

[98] ——, “The energy market 2004,” Swedish Energy Agency, Tech. Rep.

ET 29:2004, 2004. [Online]. Available: http://www.stem.se/web/biblshop.nsf/

FilAtkomst/ER62002.pdf/$FILE/ER62002.pdf?OpenElement

[99] Y. Tang and L. Xu, “Flexible active and reactive power control strategy for a variable

speed constant frequency generating system,” IEEE Trans. Power Electron., vol. 10,

no. 4, pp. 472–478, July 1995.

[100] T. Thiringer and J. Linders, “Control by variable rotor speed of a ﬁxed-pitch wind

turbine operating in a wide speed range,” IEEE Trans. Energy Conversion, vol. 8,

no. 3, pp. 520–526, Sept. 1993.

[101] T. Thiringer and J. Luomi, “Comparison of reduced-order dynamic models of induc-

tion machines,” IEEE Trans. Power Syst., vol. 16, no. 1, pp. 119–126, Feb. 2001.

[102] J. L. Thomas and M. Boidin, “An internal model control structure in ﬁeld oriented

controlled vsi induction motors,” in Proc. of the 4th European Conference on Power

Electronics and Applications, vol. 2, 1991, pp. 202–207.

[103] K. Thorborg, Power Electronics – in Theory and Practice. Lund, Sweden: Studentlit-

teratur, 1997.

[104] Vestas. (2004, May) V52-850 kW. Brochure. [Online]. Available: http://www.vestas.

com/produkter/pdf/updates 020304/V52 UK.pdf

[105] ——. (2005, Jan.) V120-4.5 MW Offshore leadership. Brochure. [Online]. Available:

http://www.vestas.com/pdf/produkter/AktuelleBrochurer/v120/V120%20UK.pdf

[106] A. K. Wallace, R. Spee, and G. C. Alexander, “The brushless doubly-fed machine: its

advantages, applications and design methods,” in Sixth International Conference on

Electrical Machines and Drives), Oxford, UK, Sept., 8–10, 1993, pp. 511–517.

[107] S. Wang and Y. Ding, “Stability analysis of ﬁeld oriented doubly-fed induction ma-

chine drive based on computer simulation,” Electric Machines and Power Systems,

vol. 21, no. 1, pp. 11–24, 1993.

156

[108] S. Williamson, A. C. Ferreira, and A. K. Wallace, “Generalised theory of the brushless

doubly-fed machine. part 1: Analysis,” IEE Proc. Electr. Power Appl., vol. 144, no. 2,

pp. 111–122, Mar. 1997.

[109] L. Xu, F. Liang, and T. A. Lipo, “Transient model of a doubly excited reluctance

motor,” IEEE Trans. Energy Conversion, vol. 6, no. 1, pp. 126–133, Mar. 1991.

[110] L. Xu and C. Wei, “Torque and reactive power control of a doubly fed induction

machine by position sensorless scheme,” IEEE Trans. Ind. Applicat., vol. 31, no. 3,

pp. 636–642, May/June 1995.

[111] D. Zhou and R. Spee, “Field oriented control development for brushless doubly-fed

machines,” in Proc. IEEE Industry Applications Conference, vol. 1, Oct., 6–10, 1996,

pp. 304–310.

[112] D. S. Zinger and E. Muljadi, “Annualized wind energy improvement using variable

speeds,” IEEE Trans. Ind. Applicat., vol. 33, no. 6, pp. 1444–1447, Nov./Dec. 1997.

157

158

Appendix A

Nomenclature

Symbols

A

r

swept area

C capacitor

C

p

power coefﬁcient

E back EMF

E

g

, E

g

grid-voltage modulus and space vector

F controller

f(w) probability density function

G transfer function

g

r

gearbox ratio

I steady-state complex-valued current

i, i current modulus and space vector

J inertia

j

√

−1

k

E

, k

R

coefﬁcients in the rotor current control law

k

p

, k

i

proportional and integral gain

L inductance

L Laplace transform

L

−1

inverse Laplace transform

n

p

number of pole pairs

n

s

/n

r

stator-to-rotor turns ratio

P active power

p d/dt

Q reactive power

R resistance

S apparent power

s slip

T

e

, T

s

electromechanical and shaft torque

T

sample

sample time

V steady-state complex-valued voltage

V remaining voltage

v, v voltage modulus and space vector

159

α closed loop bandwidth

β pitch angle

λ tip-speed ratio

ρ density of air or bandwidth of PLL

Ψ ﬂux space vector or steady-state complex-valued ﬂux

ψ ﬂux modulus

ω

1

, θ

1

synchronous frequency and angle

ω

2

slip frequency

ω

g

, θ

g

grid frequency and angle

ω

r

(electrical) rotor speed of generator

˜ error

ˆ estimated

Superscripts

avg average

max maximum

min minimum

s stator-oriented reference frame

pk peak

ref reference

Subscripts

cl closed loop

co cut off

d real part of synchronous-frame space vector

f (grid-) ﬁlter or ﬂux

g grid

GB gearbox

m mutual

M mutual (Γ representation)

mech mechanical

n negative sequence

nom nominal

R rotor (Γ representation)

r rotor

s stator

sw switch

t turbine

q imaginary part of synchronous-frame space vector

p positive sequence

λ leakage

σ leakage (Γ-representation)

160

Abbreviations

DFIG doubly-fed induction generator

EMF electromotive force

FSIG ﬁxed-speed wind turbine with an induction generator

G generator

GSC grid-side converter

IG induction generator

IGBT insulated gate bipolar transistor

IMC internal model control

LLF line-to-line fault

MSC machine-side converter

PLL phase-locked loop

PMSG permanent-magnet synchronous generator

p.u. per unit

PWM pulse width modulation

RMS root mean square

SG synchronous generator

SLGF single-line-to-ground fault

TLGF two-lines-to-ground fault

VSIG variable-speed wind turbine with an induction generator and

a full-power converter

WT wind turbine

161

162

Appendix B

Data and Experimental Setup

B.1 Data of the DFIG

These data and parameters of the DFIG are used throughout the thesis if not otherwise stated.

In Table B.1, Table B.2, and in Table B.3 the nominal values, base values, and the parameters

of the DFIG are shown respectively.

TABLE B.1. NOMINAL VALUES OF THE DFIG.

Rated voltage (Y) V

n,p−p

690 V

Rated current I

n

1900 A

Rated frequency f

n

50 Hz

Rated power P

n

2 MW

Number of pole pairs n

p

2

TABLE B.2. BASE VALUES.

Base voltage (phase-neutral) V

b

400 V

Base current I

b

1900 A

Base frequency ω

b

2π 50 Hz

Base impedance Z

b

= V

b

/I

b

0.21 Ω

TABLE B.3. PARAMETERS OF THE INDUCTION MACHINE.

Stator resistance R

s

0.0022 Ω ⇔ 0.01 p.u.

Rotor resistance R

r

0.0018 Ω ⇔ 0.009 p.u.

Rotor resistance (Γ equivalent) R

R

0.0019 Ω ⇔ 0.0093 p.u.

Stator leakage inductance L

sλ

0.12 mH ⇔ 0.18 p.u.

Rotor leakage inductance L

rλ

0.05 mH ⇔ 0.07 p.u.

Leakage inductance (Γ equivalent) L

σ

0.18 mH ⇔ 0.27 p.u.

Magnetizing resistance R

m

42 Ω ⇔ 198 p.u.

Magnetizing inductance L

m

2.9 mH ⇔ 4.4 p.u.

Magnetizing inductance (Γ equivalent) L

M

3.1 mH ⇔ 4.6 p.u.

A dc-link capacitor of C

dc

= 53 mH = 3.5 p.u. is used.

163

B.2 Laboratory Setup

The laboratory setup consists of one slip-ringed wound rotor induction machine, one voltage

source converter, two measurement boxes, one digital signal processing (DSP) system and

one measurement computer. Data of the induction machine is given in Section B.2.1. Fig.

B.1 shows a principle sketch of the laboratory setup. In the measurement boxes voltages and

IM dc mach.

DSP Converter

Meas.

computer

dc supply

ac supply

θ

r

v, i

v, i

Fig. B.1. Laboratory setup. Thick lines indicates cables with power while dashed lines implies mea-

surements signals.

currents are measured. One measurement box is attached to the stator circuit while the other

measure the rotor circuit. There is also a resolver that measure the rotor position, θ

r

, of the

induction machine. When running the machine as doubly-fed the stator circuit is directly

connected to the grid (during the experiments in this thesis the stator circuit was connected

to a 230-V, 50-Hz source, note that the nominal voltage of the induction machine is 380 V).

Normally, the converter operates as a back-to-back converter, but during the experiments the

converter was directly fed by a dc source of 450 V dc. Although the converter here is fed

directly from a dc source, it is possible to run it as a back-to-back converter. The loading dc

machine is fed through a thyristor inverter and could be both speed or torque controlled.

The control laws were all written in the C-language and downloaded to the DSP-unit

(Texas TMS320c30). The DSP-unit has 16 analog input channels, for measurement signals,

and 8 analog output channels, for signals that is desired to be fed to the measurement com-

puter. The voltage references to the converter are modulated digitally and via optic ﬁbers

sent to the converter.

The measurement system consists of one ﬁlter box and one computer equipped with the

LabView software. With this system it is possible to measure up to 16 channels, i.e., from

the measurements boxes or from the DSP unit.

A more thorough description of the laboratory set up can be found in [75].

B.2.1 Data of the Induction Generator

In Table B.4, Table B.5, and in Table B.6 the nominal values, base values, and the parameters

of the laboratory DFIG are shown respectively.

164

TABLE B.4. NOMINAL VALUES OF THE INDUCTION GENERATOR.

Rated voltage (Y) V

n,p−p

380 V

Rated current I

n

44 A

Rated frequency f

n

50 Hz

Rated rotor speed n

n

1440 rpm

Rated power P

n

22 kW

Rated torque T

n

145 Nm

Power factor 0.89

TABLE B.5. BASE VALUES.

Base voltage (phase-neutral) V

b

220 V

Base current I

b

44 A

Base frequency ω

b

2π 50 Hz

Base impedance Z

b

= V

b

/I

b

5 Ω

TABLE B.6. PARAMETERS OF THE INDUCTION MACHINE.

Stator resistance R

s

0.115 Ω ⇔ 0.0230 p.u.

Rotor resistance R

r

0.184 Ω ⇔ 0.0369 p.u.

Stator leakage inductance L

sλ

1.65 mH ⇔ 0.104 p.u.

Rotor leakage inductance L

rλ

1.68 mH ⇔ 0.106 p.u.

Magnetizing resistance R

m

224 Ω ⇔ 44.9 p.u.

Magnetizing inductance L

m

46.6 mH ⇔ 2.93 p.u.

Inertia J 0.334 kgm

2

⇔ 178 p.u.

B.3 Jung Data Acquisition Setup

The experiments were made on a VESTAS V-52 850 kWWT, located at the inland (≈100 km

from the west coast) in the southern part of Sweden. The wind turbine is located in a ﬂat

surroundings and is connected to the 10-kV distribution grid via a transformer, which trans-

forms the voltage to the wind-turbine voltage of 690 V. See Fig. B.2 for a picture of the

turbine and the data acquisition computer. In Table B.7 some data of VESTAS V-52 850 kW

WT is given. The currents and voltages are measured using transformers, which transform

TABLE B.7. DATA OF VESTAS V-52 850 KW WT [104].

Rated voltage (Y) 690 V

Rated power 850 kW

Rotor diameter 52 m

Rotor speed 14.0–31.0 rpm (26 rpm)

Cut-in wind speed 4 m/s

Nominal wind speed 16 m/s

Maximum wind speed 25 m/s

the current to 5 A and the voltage to 110 V. In addition, the stator currents are also measured

directly using LEM modules.

165

Fig. B.2. Jung wind turbine and the data acquisition computer.

166

Analysis, Modeling and Control of Doubly-Fed Induction Generators for Wind Turbines ANDREAS PETERSSON ISBN 91-7291-600-1

c ANDREAS PETERSSON, 2005.

Doktorsavhandlingar vid Chalmers tekniska h¨ gskola o Ny serie nr. 2282 ISSN 0346-718x

Division of Electric Power Engineering Department of Energy and Environment Chalmers University of Technology SE-412 96 G¨ teborg o Sweden Telephone + 46 (0)31-772 1000

Chalmers Bibliotek, Reproservice G¨ teborg, Sweden 2005 o

Analysis, Modeling and Control of Doubly-Fed Induction Generators for Wind Turbines ANDREAS PETERSSON Division of Electric Power Engineering Department of Energy and Environment Chalmers University of Technology

Abstract

This thesis deals with the analysis, modeling, and control of the doubly-fed induction generator (DFIG) for wind turbines. Different rotor current control methods are investigated with the objective of eliminating the inﬂuence of the back electromotive force (EMF), which is that of, in control terminology, a load disturbance, on the rotor current. It is found that the method that utilizes both feed forward of the back EMF and so-called “active resistance” manages best to suppress the inﬂuence of the back EMF on the rotor current, particularly when voltage sags occur, of the investigated methods. This method also has the best stability properties. In addition it is found that this method also has the best robustness to parameter deviations. The response of the DFIG wind turbine system to grid disturbances is simulated and veriﬁed experimentally. A voltage sag to 80% (80% remaining voltage) is handled very well. Moreover, a second-order model for prediction of the response of small voltage sags of the DFIG wind turbines is derived, and its simulated performance is successfully veriﬁed experimentally. The energy production of the DFIG wind turbine is investigated and compared to that of other wind turbine systems. The result found is that the energy capture of the DFIG wind turbine is almost the same as for an active stall-controlled ﬁxed-speed (using two ﬁxed speeds) wind turbine. Compared to a full-power-converter wind turbine the DFIG wind turbine can deliver a couple of percentage units more energy to the grid. Voltage sag ride-through capabilities of some different variable-speed wind turbines has been investigated. It has been found that the energy production cost of the investigated wind turbines with voltage sag ride-through capabilities is between 1–3 percentage units higher than that of the ordinary DFIG wind turbine without the ride-through capability. Finally, a ﬂicker reduction control law for stall-controlled wind turbines with induction generators, using variable rotor resistance, is derived. The ﬁnding is that it is possible to reduce the ﬂicker contribution by utilizing the derived rotor resistance control law with 40– 80% depending on the operating condition. Keywords: Doubly-fed induction generator, wind turbine, wind energy, current control, voltage sag, power quality. iii

iv .

Thanks goes to my fellow Ph. and Oskar Wallmark for a good companionship and valuable discussions. especially with the analysis of the full-power converter. who have assisted me during the work of this Ph. v .D. students who have assisted me: Stefan Lundberg for a pleasant collaboration with the efﬁciency calculations. The ﬁnancial support provided by the Swedish National Energy Agency is gratefully acknowledged. Many thanks go to the colleagues at the Division of Electric Power Engineering and the former Department of Electric Power Engineering. Finally. Dr. thesis.Acknowledgements This research project has been carried out at the Department of Energy and Environment (and the former Department of Electric Power Engineering) at Chalmers University of Technology. inspiration. Tore Undeland for valuable comments and encouragement. Torbj¨ rn Thiringer and Prof. Rolf Ottersten for many interesting discussions and a nice cooperation. Lennart Harnefors o for help. Tom´ s Petr˚ for valuable and time aˇ u saving collaboration with practical ﬁeld measurement set-ups. Dr. I would also like to thank my examiner Prof.D. I would like to thank my supervisors Dr. and encouragement. I would like to thank my family for their love and support.

vi .

. . . . . . . . 3. . .5 Total Losses . . .3. . . . . . . . . .1 Equivalent Circuit of the Doubly-Fed Induction Generator . . . . . . 3. .1 Wind Energy Conversion . 2. . . . . . . . . . . . . . . . . . 2. . . . . . . . . . . . .2 Variable-Speed Wind Turbine . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . .3 Induction Generator Losses . . . . . . . . . . . . . . . . . . . 1. . . .4 Converter Losses . . . . . . . . . . . . . . . . . . . .3 Aerodynamic Conversion . . . . . . . . 3.3 Variable-Speed Wind Turbine with Doubly-Fed Induction Generator 2. . . . .1. . . . . . . . . . . .2 Wind Turbine Systems . . . . . . . . . . . . 3 Energy Efﬁciency of Wind Turbines 3. . . 2. . . . . 2. . . . . .1 Aerodynamic Losses . . . . . . . . . . . . 3. .2 Purpose and Contributions . . . . . . . . . . 3. . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . . . . . . . 2. . . . . . .2. . . . . . . . . . . . . .1. . . . .5 Other Types of Doubly-Fed Machines . . . . . . 2. . . . . . . . . . . . . .1 Wind Distribution . . . . . . . .Table of Contents Abstract Acknowledgements Table of Contents 1 Introduction 1. . . . . . . . . .3. . .1 Determination of Power Losses . . . . . . . .2 Aerodynamic Power Control . . . . . . . . . . . . .2. .1 Review of Related Research . . . . . . . . . . . . . . . . . . . . . . . .3 List of Publications . . . . 2. . . . . . . . . . . . . . . . . . 2. . . 2. . . . . . . . . . . . . . . . . . . .2 Gearbox Losses . . . . . . . . . . .3 Doubly-Fed Induction Generator Systems for Wind Turbines . . . . . . . . . . . . 2. . . . . . . . .1 Investigation of the Inﬂuence of the Converter’s Size on the Energy Production . . . 2 Wind Energy Systems 2.4 Lowering Magnetizing Losses . . .2. . .1 Fixed-Speed Wind Turbine .1. 3. . . . . vii iii v vii 1 2 4 5 7 7 7 8 8 9 11 11 12 13 14 16 17 18 19 23 23 23 24 24 26 28 29 29 . . . . . . . . . . . . . . .1.1. . . . . . . . . . . . . . . . . . .3. . . 2. . . . . 1. . . . . . 2. . . . . . . . . . . . . . . . . . . . .2 Power Flow . . . . .3. . . .2 Energy Production of the DFIG System . . . . . . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. . . .1. . . .3 Stator-to-Rotor Turns Ratio . . .1. .

. . . . . . . . . . . . . . . . . . 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . .1 Current Control of Grid Filter . . . . . . . . . . . . . . . . . .4 Summary . . . 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . . . . 4. 5. . . . . . . . . . 4. . . . . . 4. . .2 Mathematical Models of the DFIG System . . . . . . . . . . .4. . . . . viii . . 4. . . . . . .4. . . . . . .2. . . 4. 4. . . . . . . . . .3. . . . . . . . . . . . Rs and RR . .5. . . 5. . . . . . . . . .5 Sensorless Operation . 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . .1. . . . . . . . .1 Comparison Between Stator-Flux and Grid-Flux-Oriented System 5. 5. . . . 5. . . . . .2 Torque Control . . .6 Saturation and Integration Anti-Windup . . . . . . . . . .3. . . . . . . . . . .4 Impact of Stator Voltage Sags on the Current Control Loop . . .5. . .1 Stator-Flux-Oriented System . . .2 Grid-Flux-Oriented System . . . . . . . . 5. . . . . . . 4. . . .1. . Conclusion . . . . . . . . . . . . . . . .1. . 5. . . . . .1 Machine Model . 5. .7 Discretization . . . . . . . . . . . . . . . . . .3 Conclusion . .2. . . . . . . . . . . .4. . 4. . . 5. .2 Inﬂuence of Erroneous Parameters on Stability . . . . . . . . . . . . . . . . .2. . . . . . . . . . .2 Grid-Filter Model . . . . . . . . . . . . . . . . . .5 4 3. . . . . . . . .4. . . .2 Grid-Flux Orientation . . . . . .2 DC-Link Voltage Control . . . . . . . . . . . . . 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Control of Machine-Side Converter . . .3. . . . . . 4. . . . . . . . 4. . . . . . . . . . . 5. . .2 Generation Capability During Voltage Sags . . . . . . . . . . . . . .1. . . . . .3 Experimental Evaluation .1 Stator-Flux Orientation . . . . . . . . .4 Reactive Power Control . . . . . . . . . . . . . . .1 Space Vectors .3 Phase-Locked Loop (PLL)-Type Estimator .1 Leakage Inductance. . . .3 DC-Link Model . . . . . 4. . . . . . . . . . . . . . .2. . . . . . . . . . . . . .1. . . . . . . Comparison to Other Wind Turbine Systems Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . .5 Control of Grid-Side Converter . . 5 Evaluation of the Current Control of Doubly-Fed Induction Generators 5. . . . . . . . . .4. . . . . . 31 31 33 34 35 35 35 36 36 37 38 40 40 41 41 43 44 45 45 45 46 47 47 50 50 52 54 55 56 56 59 59 59 64 66 66 67 67 70 71 71 73 74 74 76 Control of Doubly-Fed Induction Generator System 4. .1. . . . Lσ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Current Control .1 Stability Analysis . . . . . . . . . . . .1 Inﬂuence of Erroneous Parameters . . . . . . 4. . . . . . . . . . . . . 5. . . . . . . . . . . . .1. . .2 Power and Reactive Power in Terms of Space Vectors 4. . . . 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 Flux Damping . . . . . . . . .3 Speed Control . . . . . . .1. . . . . . . . . . . . . . . . . .3 3. . . . . . . .2 Reduction of Magnetizing Losses . . . . . . . . .1 Introduction . . . . . . . . . . . . . . . .1. . . . . . . . . . .2. . . . . . . . . . .4 3. . . .2. . 5. . . . . . . . . . . 4. . . . . . . . 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Stator-Flux Orientation . . . .4. . . . . . . . . . . . . . . . . . . . .2 Stator and Rotor Resistances. . . . . . .4 Internal Model Control (IMC) . . . . . . . . . . . . .1.3 Field Orientation . 4. . . . . . . 4.5 “Active Damping” . .4. . . . . . . . . . . . . . . . . . . . . . 4.

. . . . . . . 134 . . . . . .1. . . . . . . . .4. . . . . .3. . . . . . . . . . . .1 Stator-Flux Orientation . . . . . . . . . . .5. . . . . . . .2 Current Control . . 7. . . . . . . . . . . . . . . . . . . . . . 135 .5. . . . . . . . . . . . . . . . . . . . . . 6. . . . . . .1 Voltage Sags .5. . . ix . . . . . . . . .1 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . .2. . . . . . . . .3. .4 Conclusion . . . . . . . 7. . . . . . . . . . . . . . . . . . . .1 Reduced-Order Model . . . . . . . . . 7. . . . . . . . . . .4 Evaluation . . . . . . . . 7. . . .3 Reference Value Selection . . . 7. . . . . . . .2 System Modeling .5 Conclusion . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . .4. . 8. . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . .1 Response to Small Voltage Sags .1. . . . . . . . . . . . . .1 Reduced-Order Model . . . . . . . . . 138 . . . .4 Implementation in Grid Simulation Programs . . . . . 7. . . .7 Discussion and Conclusion . . . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Candidate Ride-Through System . . . .4. . . . . 76 76 77 77 79 81 81 81 82 82 83 87 88 89 90 90 90 92 92 98 99 100 102 103 110 111 114 118 118 120 123 125 126 127 131 132 6 Evaluation of Doubly-Fed Induction Generator Systems 6. 8 Flicker Reduction of Stalled-Controlled Wind Resistances 8. . 133 . . . . . . . . . . . . . . . . . . . 7. . . .2. . . . . . . . . 8. . . . . . . . . .3 Doubly-Fed Induction Generator with Shunt Converter . . . . .2 Full-Power Converter . . . . . . . . . . . . . . . . . . . . . . . 8. . . .1 Evaluation . . . . 7. 5. . . . . . . . . 7. . . . .3 Parameter Selection . .2. . . . . . . . . . . . . . . . . .2. . . . . . . . . . . . . . . . . . . . .4.3. . . . . . . .1 Modeling . . . . 7 Voltage Sag Ride-Through of Variable-Speed Wind Turbines 7. 6.2. . . 6. .1. . . . . . . .4 Speed Control Operation . 7. . . .2 Discretization of the Doubly-Fed Induction Generator 6. . . . . . . . . . . 7. . . . . 7. . . . . .6 5. . . . . . . . . . 133 . . . . 7. . . . . . . . . 139 . . . . . . . . . . . 5. . . . . . .2 Discussion . . . . . . . . . .2.2 Response to Large Voltage Sags . . . . . . . . . . .4 Evaluation of the Ride-Through System .6 Steady-State Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Possible System Conﬁgurations . 7. . . . . . . . . . . . . . .5 Response to Symmetrical Voltage Sags Conclusion . . . . . .5 Response to Voltage Sags . . .4. . . . . . . . . .4. . . . . . . . . . . . . . . . . . . .3 Control . . . . . . . . . . . . . . . . .2 Unsymmetrical Voltage Sags . . 8. . . . . . . . . . . . .3 Response to Grid Disturbances . . . . . . . . . . . . . . 7. . . . . . 7. . . . . . . . .3. . . . .1 Analysis . . . .2 Grid-Flux Orientation . . . . . . . . . 7. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. . . . . . . . . . . . . . . . . . . . . . Turbines using Variable Rotor . .1 Symmetrical Voltage Sags . . . . 5. . . . . . . 7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Grid-Flux Orientation . . . . . . . .5. . . . . . . . . . .2. . . . . . 140 . . . . . . . . . . . . . . 7. . . . . . . . . 7.4 Doubly-Fed Induction Generator with Series Converter . . . . . .5 Summary . . . . . . . . . . .3 Evaluation . . . . . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . 6. .

. . . . . . . . . . . . . . . . . . . . .8. . . . . . . .3 Jung Data Acquisition Setup .1 Data of the Induction Generator B. . . . . . . B. . . . . . . . . .1 Data of the DFIG . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . . . Conclusion . .2. . . . . . . . . . . .5 9 Evaluation . . . . . . . . . . . . . 141 143 144 145 147 148 149 159 163 163 164 164 165 Conclusion 9. . . . . . . .1 Flicker Contribution 8. . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x . . . . . .1 Future Research . . . . . .2 Flicker Reduction . . . . . .4 8. . . . . .2 Laboratory Setup . . . . . . . . . . . . . . . . . . . . . . . . . References A Nomenclature B Data and Experimental Setup B. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . B. . . . . . . . . .

The major advantage of the doubly-fed induction generator. 98]. Nuclear power production shall be replaced by improving the efﬁciency of electricity use. These wind turbines produced 0. conversion to renewable forms of energy and other environmentally acceptable electricity production technologies [97]. i. 110]. 29. Fixed-speed induction generators with stall control are regarded as unfeasible [3] for these large wind turbines. by the end of 2004. The ﬁrst nuclear reactor of Barseb¨ ck a was shut down 30th of November 1999. Today. that the nuclear power production is to be phased out at a slow rate so that the need for electrical energy can be met without risking employment and welfare.Chapter 1 Introduction The Swedish Parliament adopted new energy guidelines in 1997 following the trend of moving towards an ecologically sustainable society. According to [97] wind power can contribute to fulﬁlling several of the national environmental quality objectives decided by Parliament in 1991. This means that the losses in the power electronic equipment can 1 . corresponding to approximately 0. among those are possibilities to reduce stresses of the mechanical structure.8 TWh of electrical energy in 2004. 73. The Swedish National Energy Agency suggest that the planning objectives for the expansion of wind power should be 10 TWh/year within the next 10–15 years [97]. There are several reasons for using variable-speed operation of wind turbines. For a variablespeed wind turbine the generator is controlled by power electronic equipment. In Sweden. These large wind turbines are all based on variable-speed operation with pitch control using a directdriven synchronous generator (without gearbox) or a doubly-fed induction generator (DFIG).5% of the total generated and imported electrical energy [23. The decision also conﬁrmed that the 1980 and 1991 guidelines still apply. corresponding to 1% of the total installed electric power in the Swedish grid [23. Continued expansion of wind power is therefore of strategic importance.e. Wind turbines (WTs) can either operate at ﬁxed speed or variable speed. 68. For a ﬁxedspeed wind turbine the generator is directly connected to the electrical grid. Most of the major wind turbine manufactures are developing new larger wind turbines in the 3-to-5-MW range [3]. is that the power electronic equipment only has to handle a fraction (20–30%) of the total system power [36. doubly-fed induction generators are commonly used by the wind turbine industry (year 2005) for larger wind turbines [19. The energy policy decision states that the objective is to facilitate a change to an ecologically sustainable energy production system. which has made it popular.. 98]. acoustic noise reduction and the possibility to control active and reactive power [11]. 105]. there was 442 MW of installed wind power.

Control of the DFIG is more complicated than the control of a standard induction machine. also.be reduced in comparison to power electronic equipment that has to handle the total system power as for a direct-driven synchronous generator. blade design. rotor speed. One method is to reduce the bandwidth of the current controllers [43]. [43] and Congwei et al. at least for a stator-ﬂux-oriented system. Wang et al. In order to control the DFIG the rotor current is controlled by a power electronic converter. apart from the cost saving of using a smaller converter. According to [16] the energy capture can be signiﬁcantly increased by using a DFIG.1 Review of Related Research According to [12] the energy production can be increased by 2–6% for a variable-speed wind turbine in comparison to a ﬁxed-speed wind turbine. selected blade proﬁle. harder to ﬁnd. or with air-gap-ﬂux orientation [107. One exception is in [16]. 99]. These poorly damped poles inﬂuence the rotor current dynamics through the back electromotive force (EMF). A comparison to other electrical systems for wind turbines are. while in [112] it is stated that the increase in energy can be 39%. The author has. 86. what kind of power that should be used as a common basis for comparison. and stator ﬂux. which is dependent on the stator voltage. If the stator resistance can be considered small. and. [107] have by simulations found that the ﬂux is inﬂuenced both by load changes and stator power supply variations. 61. [54] have used 2 . where Datta et al. Heller et al. that the system is unstable for certain operating conditions. Several vector control schemes for the DFIG have been proposed. [13] have investigated the stability of the DFIG analytically. affect the outcome of the investigations. The ﬂux oscillations can be damped in some different ways. Kelber et al. selection of maximum speed of the WT. not found in the literature any evaluation of the performance of different rotor current control laws with respect to eliminating the inﬂuence of the back EMF. Wang et al. 80. There is thus a need to clarify what kind of energy capture gain there could be when using a DFIG WT. 68]. Efﬁciency calculations of the DFIG system have been presented in several papers. One common way is to control the rotor current with stator-ﬂux orientation [46. in the rotor current. 99]. In [69] it is shown that the gain in energy generation of the variable-speed wind turbine compared to the most simple ﬁxed-speed wind turbine can vary between 3–28% depending on the site conditions and design parameters. They state an increased energy capture of a DFIG by over 20% with respect to a variable-speed system using a cage-bar induction machine and by over 60% in comparison to a ﬁxed-speed system. stator-ﬂux orientation gives in principle orientation also with the stator voltage (grid-ﬂux orientation) [17. both compared to another variable-speed WT and towards a traditional ﬁxed-speed WT. showing that the dynamics of the DFIG have poorly damped eigenvalues (poles) with a corresponding natural frequency near the line frequency. however. however. Factors such as speed control of variable-speed WTs. have made a comparison of the energy capture for various WT systems. 61. missing facts regarding the base assumptions etc. 110]. 1. One of the reasons for the various results is that the assumptions used vary from investigation to investigation. One common way of controlling the rotor current is by means of ﬁeld-oriented (vector) control. The ﬂux response to a disturbance is a damped oscillation. for instance [52. [107] have introduced a ﬂux differentiation compensation that improves the damping of the ﬂux.

This is a common way to reduce the DFIG model in classical control theory stability analysis [13. New grid codes will require WTs and wind farms to ride through voltage sags. meaning that normal power production should be re-initiated once the nominal grid voltage has been recovered. As described earlier a dominating feature of the DFIG system is the natural frequency of the ﬂux dynamics. Such codes are in progress both in Sweden [96] and in several other countries [8]. The possibility to use it as simulation model remains to be shown.another possibility. Another option proposed in [72] is to use an “active” crowbar. which at large grid disturbances has to short circuit the rotor circuit in order to protect the converter. In the literature there are some different methods to modify the DFIG system in order to accomplish voltage sag ride-through proposed. 72] in order to comply. It was found that the methods with a ﬂux differentiation compensation and the method with an extra converter manage to damp the oscillations best. the efﬁciency and cost of the different voltage sag ride-through system might also inﬂuence the choice of system. a slightly different model approach must be made. The response of wind turbines to grid disturbances is an important issue. 20. a third-order model has been proposed that neglects the stator-ﬂux dynamics of the DFIG. it is obvious that a second-order model is the simplest that can be used. 60. which has initiated industrial research efforts [8. 55]. Kelber has shown that such a system can effectively damp the ﬂux oscillations caused by voltage sags. This leads to that the turbine must be disconnected from the grid. it is important for utilities to be able to study the effects of various voltage sags and. 26. After the DFIG WT has been disconnected. after a large voltage sag. The DFIG system. In [22. This means that new WTs have to ride through these voltage sags. 43]. 42. 30. has a crowbar in the rotor circuit. when modifying the DFIG system for voltage sag ride3 . Kelber has in [55] made a comparison of different methods of damping the ﬂux oscillations. it is of importance to have as simple models as possible that still manage to model the dynamics of interest. Therefore. and thereby be able to remagnetize the generator and reconnect the stator to the grid as fast as possible.e. A third method. For calculations made using grid simulation programs.. the DFIG WT will be disconnected from the grid when large voltage sags appear in the grid. which can break the short circuit current in the crowbar. In order to preserve the behavior of an oscillatory response. In order to preserve the dynamic behavior of the DFIG system. is to use an additional converter to substitute the Y point of the stator circuit [54. especially since the rated power of wind-turbine installations steadily increases. that has been mentioned earlier. which is close to the line frequency. 84]. the corresponding wind turbine response. All of these systems have different dynamical performance. Since the dynamics of the DFIG are inﬂuenced by two poorly damped eigenvalues (poles) it would be natural to reduce the model of the DFIG to the ﬂux dynamics described by a second-order model. for instance. Today. i. of today. to use an extra (third) converter that substitutes the Y point of the stator winding. In [55]. Therefore. it takes some time before the turbine is reconnected to the grid. Moreover. This model gives a correct mean value [22] but a drawback is that some of the main dynamics of the DFIG system are also neglected. 28. These grid codes will inﬂuence the choice of electrical system in future WTs. an extra degree of freedom is introduced that can be used to actively damp the ﬂux oscillations. In [20] anti-parallel thyristors is used in the stator circuit in order to achieve a quick (within 10 ms) disconnection of the stator circuit.

The investigated systems are two ﬁxed-speed induction generator systems and three variable-speed systems. a comparison of the energy efﬁciency of DFIG system to other electrical systems is presented. 4 . in both the stator-ﬂux-oriented and the grid-ﬂux-oriented reference frames. Simulations are veriﬁed with experimental results. for both correctly and erroneously known parameters. • In Chapter 6. which inﬂuences the aerodynamical efﬁciency. another objective is also to study how a reduced-order (second-order) model manages to predict the response of the DFIG system. The variable-speed systems are: a doubly-fed induction generator. the control and the modeling are also important parts of the thesis. the grid-fault response of a DFIG wind turbine system is studied. 41]. in Chapter 3. an induction generator (with a full-power converter) and a direct-driven permanent-magnet synchronous generator system. one of the main advantage with the DFIG system was that losses of the power electronic equipment is reduced in comparison to a system where the power electronic equipment has to handle the total power.2 Purpose and Contributions The main purpose of this thesis is the analysis of the DFIG for a WT application both during steady-state operation and transient operation. On the other hand it also implies a smaller variable-speed range. In order to make the comparison as fair as possible the base assumption used in this work is that the maximum (average) shaft torque of the wind turbine systems used should be the same. since such a system can be considered to have excellent voltage sag ride-through performance (as also will be shown in Chapter 7) [74]. • In Chapter 3 an investigation of the inﬂuence of the converter’s size on the energy production for a DFIG system is analyzed. Finally. Stability analysis of the system is performed for different combinations of the terms introduced in the current control law.” “Active resistance” has been used for the squirrel-cage induction machines to damp disturbances. such as varying back EMF [18. 1. as mentioned earlier.through it is necessary to evaluate consequences for cost and efﬁciency. Consequences for the efﬁciency is an important issue since. Any evaluation of different voltage sag ride-through methods for DFIG wind turbines and how they affect the efﬁciency is hard to ﬁnd in the literature. Important electrical and mechanical losses of the systems are included in the study. it is necessary to compare the ride-through system with a system that utilizes a full-power converter. Moreover. with details being as follows. Further. Hence. Terms are introduced in order to allow the possibility to include feed-forward compensation of the back EMF and/or “active resistance. In order to analyze the DFIG during transient operation both the control and the modeling of the system is of importance. The main contribution of this thesis is dynamic and steady-state analysis of the DFIG. Moreover. A smaller converter implies that the converter losses will be lower. • In Chapter 4 a general rotor current control law is derived for the DFIG system. The main contribution of Chapter 5 is an evaluation of different rotor current control laws with respect to eliminating the inﬂuence of the back EMF. two different methods of reducing the magnetizing losses of the DFIG system are compared.

3. Lundberg. a rotor resistance control law for a stall-controlled wind turbine is derived and analyzed. Petersson. July 13– 17. “A DFIG Wind-Turbine Ride-Through System Inﬂuence on the Energy Production. vol.3 List of Publications Some of the results presented in this thesis have been published in the following publications. “Flicker Reduction of Stall-Controlled Wind Turbines using Variable Rotor Resistances. Moreover. S. Thiringer. dynamically and in the steady state. in Chapter 8. Toronto. Petr˚ . T. 2003. G¨ teborg. 1–2. The organizing committee of the conference recommended submission of this paper to Wind Energy. 3. 2. IEEE Nordic Workshop on Power and Industrial Electronics (NORpie/2002). A. o In this paper a rotor resistance control law is derived for a stall-controlled wind turbine. Petersson and S. Canada. The objective of the control law is to minimize torque ﬂuctuations and ﬂicker.• The contribution of Chapter 7 is to analyze. The reason for comparing these two systems with a system that utilizes a full-power converter is that the latter system is capable of voltage sag ride-through. 12–14.” in Proc. 2004. 1. Stockholm. Aug.” in Proc. The grid disturbance response to ﬁxed-speed wind turbines and wind turbines with DFIG were presented. This paper is an early version of the material presented in Chapter 3. This study is presented in Chapter 8. Sweden. Mar. these two methods are also compared to a system that utilizes a full-power converter. Nordic Wind Power Conference. and L. A. “Grid Disturbance Response of Wind Turbines u Equipped with Induction Generator and Doubly-Fed Induction Generator. 1–2. • Finally. Petersson. Sweden. Sweden. The objective of the control law is to minimize the ﬂicker (or voltage ﬂuctuations) contribution. Petersson.” in Proc. Nordic Wind Power Conference. and T. G¨ teborg. IEEE Power Engineering Society General Meeting. 5 . o In this paper a voltage sag ride-through system for a DFIG WT based on increased rating of the valves of the power electronic converter was investigated. Harnefors. A. 2002. 1. Mar. “Energy efﬁciency comparison of electrical systems for wind turbines. The paper has been accepted for publication. Lundberg. Thiringer. and T. A. 4. pp. This paper presents one of the voltage sag ride-through system for a DFIG wind turbine that is compared in Chapter 7. 2004. 1542–1547.” in Proc. Thiringer. The efﬁciency of some different electrical systems for wind turbines are compared. two different voltage sag ride-through systems for the DFIG. T.

Germany. 470–475. Norway. A. Harnefors. 2004. In this paper the analysis of the stator-ﬂux oriented current control of the DFIG presented in Chapter 5 was studied. IEEE Nordic Workshop on Power and Industrial Electronics (NORpie/2004). 8. Petersson. IEEE Power Electronics Specialists Conference (PESC’04). A. Harnefors. Thiringer and A. L. “Modeling and Experimental u Veriﬁcation of Grid Interaction of a DFIG Wind Turbine. 6. and T. Petr˚ . 6 . R. their power quality impact.” IEEE Trans. L. pp. Petersson. A. 7. 2004. “Voltage Sag Response of PWM Reca tiﬁers for Variable-Speed Wind Turbines. “Evaluation of Current Control Methods for Wind Turbines Using Doubly-Fed Induction Machines. 20. no. A. and its the response to grid disturbances. This paper gives an overview of the three most common wind turbine systems. June 14–16. The paper has been accepted for publication. and T. June 20–25.” IEEE Trans. These results are also presented in Chapter 6. The organizing committee of the conference recommended submission of this paper to EPE Journal. no. vol. The models are experimentally veriﬁed with an 850 kW DFIG wind turbine. 1. Aachen. 1. Energy Conversion (accepted for publication) Here a full-order model and a reduced-order model of the DFIG is compared during grid disturbances. vol. Thiringer. and K. Ottersten. 482–486. Jan.5. 2004. The comparison between grid-ﬂux and stator-ﬂux-oriented current control of the DFIG presented in Chapter 5 were studied in this paper. T. Power Electron.” in Proc. The voltage sag response of a PWM rectiﬁer for wind turbines that utilizes a full-power converter were studied. pp. Petersson. Thiringer. Trondheim. “Comparison Between Stator-Flux and Grid-Flux Oriented Rotor-Current Control of Doubly-Fed Induction Generators.. 5. pp. This paper serves as a basis for the comparison of ride-through systems of wind turbines in Chapter 7. T. Pietil¨ inen. 2005. and T. 9. Harnefors. Thiringer.” in Proc. Petersson. 227–235. Petersson. “Grid Integration of Wind Turbines. L.” Przeglad Elektrotechniczny.

can be calculated from ∞ c 1 w= wf (w)dw = Γ (2. will be presented. will be described. Thus.Chapter 2 Wind Energy Systems 2..e. for example. i. will be presented.g.4) c = √ w.2) k k 0 where Γ is Euler’s gamma function. Γ(z) = 0 ∞ tz−1 e−t dt. the average wind speed (or the expected wind speed). c is a scale parameter and w is the wind speed.. Finally. i. The average wind speeds in the ﬁgure are 5.2 m/s. For the Rayleigh distribution the scale factor. 2. Then different methods to control the aerodynamic power will be described. π In Fig. i. produced power. given the average wind speed can be √ found from (k=2. 6. while 8–9 m/s are wind speeds available at sites located outside the Danish west coast [24]. the aerodynamic conversion. the wind speed probability density function of the Rayleigh distribution is plotted.8 m/s.1. First the wind distribution. which are of interest in this thesis. and 8. β)-curve. the probability of a certain average wind speed.3) If the shape parameter equals 2. e.. c. w.4 m/s. The Weibull distribution is described by the following probability density function k w k−1 −(w/c)k e (2. properties of the wind. and Γ( 1 ) = π) 2 2 (2. 7 . [11. the so-called Cp (λ. the Weibull distribution is known as the Rayleigh distribution. (2. 2. A wind speed of 5. The wind distribution can be used to determine the expected value of certain quantities.1 Wind Energy Conversion In this section.4 m/s correspond to a medium wind speed site in Sweden [100].1 Wind Distribution The most commonly used probability density function to describe the wind speed is the Weibull functions [53]. 53].1) f (w) = c c where k is a shape parameter.e. The interested reader can ﬁnd more information in.e.1.

2.2 m/s (dotted).. Below rated wind speed the turbine should produce as much power as possible. It is also possible to increase the angle of attack towards stall in order to limit the aerodynamic power. 2.e.5) (2.. For steady-state calculations of the mechanical power from a wind turbine.2 Aerodynamic Power Control At high wind speeds it is necessary to limit the input power to the wind turbine. The mechanical power. i.. Stall control implies that the blades are designed to stall in high wind speeds and no pitch mechanism is thus required [11].e.e.05 0 0 5 10 15 20 25 Wind speed [m/s] Fig. Probability density of the Rayleigh distribution.1.8 m/s (dashed) and 8.e. β)w3 2 Ωr rr λ= w 8 (2. for newer larger wind turbines.3 Aerodynamic Conversion Some of the available power in the wind is converted by the rotor blades to mechanical power acting on the rotor shaft of the WT. i. In order to limit the aerodynamic power. the so called Cp (λ. can be determined by [53] 1 Pmech = ρAr Cp (λ. pitch. by stall. 6. Almost all variable-speed wind turbines use pitch control. There are three major ways of performing the aerodynamic power control. The average wind speeds are 5. Pitch control is the most common method of controlling the aerodynamic power generated by a turbine rotor.0. at high wind speeds. using a pitch angle that maximizes the energy capture. the angle between the chord line of the blade and the relative wind direction [53]. or active stall control. i.1.4 m/s (solid).6) . β)-curve can be used.1. Pmech . This control method is known as active stall or combi stall [11]. the pitch angle is controlled to decrease the angle of attack. i.15 Probability density 0.1 0. 2.. This method can be used to ﬁne-tune the power level at high wind speeds for ﬁxed-speed wind turbines. aerodynamic power control. Above rated wind speed the pitch angle is controlled in such a way that the aerodynamic power is at its rated [11].

5). as a function of the tip speed ratio. In Fig.. λ. according to (2. Cp . This choice is to a high extent dependent on the average wind speed of the site.e.u. Note that there is a possibility to optimize the radius of the wind turbines rotor to suit sites with different average wind speeds. However.2 Wind Turbine Systems Wind turbines can operate with either ﬁxed speed (actually within a speed range about 1 %) or variable speed. derived from the Cp (λ.3 0. rr .4 b) 120 100 Cp (λ) 0. 2. the pitch angle is at higher wind speeds controlled in order to limit the input power to the wind turbine. This implies that the nominal power will be reached for a lower wind speed.2. As mentioned before. i.e. at a wind speed of approximately 13 m/s. β) curve. increasing the rotor radius implies that for higher wind speed the output power must be even more limited. β) curve and the shaft power as a function of the wind speed for rated rotor speed. In Fig. Ωr is the rotor speed (on the low-speed side of the gearbox). 2. can be seen. 1 p. b) Mechanical power as a function of wind speed at rated rotor speed (solid line is ﬁxed pitch angle..2. Cp is kept at maximum as long as the power or rotor speed is below its rated values.5 0. by pitch control. stall control and dashed line is active stall). As seen in Fig. 2. λ is the tip speed ratio.3b). i.. e. w is the wind speed. when the turbine has reached the rated power. Fig. a) The power coefﬁcient. ρ is the air density and Ar is the area swept by the rotor.where Cp is the power coefﬁcient.1 0 0 5 10 15 20 Power [%] 80 60 40 20 0 5 10 15 20 25 Tip speed ratio Wind speed [m/s] Fig.. the output power of the turbine is also increased. 2. is increased. while dashed line corresponds to a varying β (active stall). β. there is a trade-off between the rotor radius and the nominal power of the generator.e. referred to Fig. an example of a Cp (λ. 2. pitch angle. β is the pitch angle. so that the nominal power of the generator is not exceeded. 2. For example.2b) the solid line corresponds to a ﬁxed a) 0.2 0. rr is the rotor-plane radius. The rotor speed in the variable-speed area is controlled in order to keep the optimal tip speed ratio. Therefore.3 shows an example of how the mechanical power.. λ.g. For ﬁxed-speed wind turbines.3b) the turbine in this example reaches the rated power. the generator (induction generator) is di9 . 2. if the rotor radius. a ﬁxed-speed wind turbine. i. and the rotor speed vary with the wind speed for a variable-speed wind turbine.

2. it is not possible to store the turbulence of the wind in form of rotational energy. the power quality impact caused by the wind turbine can be improved compared to a ﬁxed-speed turbine [58]. Hence. a description of some of these systems can be found in [36]. 10 . Variable-speed wind turbine equipped with a cage-bar induction generator or synchronous generator. Therefore. In this way the power ﬂuctuations caused by wind variations can be more or less absorbed by changing the rotor speed [82] and thus power variations originating from the wind conversion and the drive train can be reduced. b) Mechanical power as a function of wind speed. 3. Variable-speed wind turbine equipped with a doubly-fed induction generator. 4. The rotational speed of a wind turbine is fairly low and must therefore be adjusted to the electrical frequency. and most certainly not controllable. Fixed-speed wind turbine with an induction generator. Since the speed is almost ﬁxed to the grid frequency. The number of pole pairs sets the mechanical speed of the generator with respect to the electrical frequency and the gearbox adjusts the rotor speed of the turbine to the mechanical speed of the generator. 2. a) Rotor speed as a function of wind speed. For a variable-speed wind turbine the generator is controlled by power electronic equipment. and thus affect the power quality of the grid [77]. Variable-speed wind turbine equipped with multiple-pole synchronous generator or multiple-pole permanent-magnet synchronous generator.3. which makes it possible to control the rotor speed. This can be done in two ways: with a gearbox or with the number of pole pairs of the generator. rectly connected to the grid. There are also other existing wind turbine concepts. Typical characteristic for a variable-speed wind turbine. In this section the following wind turbine systems will be presented: 1. for a ﬁxed-speed system the turbulence of the wind will result in power variations.a) 25 b) 100 80 20 Rotor speed [rpm] Power [%] 5 10 15 20 25 60 40 20 15 10 0 5 10 15 20 25 Wind speed [m/s] Wind speed [m/s] Fig.

The gearbox is designed so that maximum rotor speed corresponds to rated speed of the generator. The ﬁxed-speed wind turbine system has often two ﬁxed speeds.6. or it can be a generator with two windings having different ratings and pole pairs.4. see Fig. 2. 11 . 2. This is accomplished by using two generators with different ratings and pole pairs. A synchronous generator with multiple poles as a wind turbine generator is successfully manufactured by Enercon [25]. The rotor speed of the ﬁxed-speed wind turbine is in principle Gearbox IG Soft starter Transformer Capacitor bank Fig. Fixed-speed wind turbine with an induction generator.2. generator or a synchronous generator. 61]. This leads to increased aerodynamic capture as well as reduced magnetizing losses at low wind speeds. determined by a gearbox and the pole-pair number of the generator. 2.2. 39. 2.5 consists of a wind turbine equipped with a converter connected to the stator of the generator. The generator could either be a cage-bar induction Gearbox G ≈ = = ≈ Transformer Power electronic converter Fig. Since this “full-power” converter/generator system is commonly used for other applications.4.5. 2.2.2 Variable-Speed Wind Turbine The system presented in Fig. Synchronous generators or permanent-magnet synchronous generators can be designed with multiple poles which implies that there is no need for a gearbox. one advantage with this system is its well-developed and robust control [7. This system (one or two-speed) was the “conventional” concept used by many Danish manufacturers in the 1980s and 1990s [36].1 Fixed-Speed Wind Turbine For the ﬁxed-speed wind turbine the induction generator is directly connected to the electrical grid according to Fig. Variable-speed wind turbine with a synchronous/induction generator. 2.

the losses in the power electronic converter can be reduced. This is mainly due to the fact that the power electronic converter only has to handle a fraction (20–30%) of the total power [36. 2. 2. Therefore. Variable-speed direct-driven (gear-less) wind turbine with a synchronous generator (SG). 105]. see Fig.6. Variable-speed wind turbine with a doubly-fed induction generator (DFIG). In addition. see Chapter 3. erators for variable-speed wind turbines [36]. DeWind. consists of a wind turbine with doubly-fed induction generator. for example. that produce wind turbines with the doubly-fed induction machine as generator are.7. 29. 2. This means that the stator is directly connected to the grid while the rotor winding is connected via slip rings to a converter. compared to a system where the converter has to handle the total power. There exists a variant of the DFIG method that uses controllable external rotor resistances (compare to slip power recovery). 110].7. This system have recently become very popular as gen- Transformer Gearbox DFIG ≈ = = ≈ Power electronic converter Fig. Nordex.SG ≈ = = ≈ Transformer Power electronic converter Fig. and Vestas [19. GE Wind Energy. Manufacturers.3 Variable-Speed Wind Turbine with Doubly-Fed Induction Generator This system.2. the cost of the converter becomes lower. 73. 2. Some of the drawbacks of this method are that energy is unnecessary dissipated in the external rotor resistances and that it is not possible to control the reactive power. 12 .

The stator circuit of the DFIG is connected to the grid while the rotor circuit is connected to a converter via slip rings. as energy storage. 110].8.g. 2. that are connected “back-to-back. 2. the DFIG and also the power factor at the stator terminals. of the DFIG system with a back-to-back converter can be seen in Fig. machine-side converter and grid-side converter. This means that the losses in the power electronic converter can be reduced compared to a system where the converter has to handle the total power.2. As mentioned earlier the reason for this is that power electronic converter only has to handle a fraction (20–30%) of the total power [36.9. As also seen in the ﬁgure. while the main objective for the grid-side converter is to keep the dc-link voltage constant.9.10 [61]. see Fig.8.” Between the two converters a dc-link capacitor is placed. In addition. the DFIG can max operate both in motor and generator operation with a rotor-speed range of ±Δωr around the synchronous speed. DFIG system with a back-to-back converter. With the machine-side converter it is possible to control the torque or the speed of DFIG dc link ≈ = Machine-side converter = ≈ Grid-side converter Grid Fig.3 Doubly-Fed Induction Generator Systems for Wind Turbines For variable-speed systems with limited variable-speed range. Principle of the doubly-fed induction generator. e.. 2. the cost of the converter becomes lower. 13 . in order to keep the voltage variations (or ripple) in the dc-link voltage small. the DFIG can be an interesting solution [61]. 2. ω1 .e. The speed–torque characteristics of the DFIG system can be seen in Fig. i. 2. A more detailed picture Converter Fig. The back-to-back converter consists of two converters. ±30% of synchronous speed.

Equivalent circuit of the DFIG. in Fig. besides wind turbines. since they operate in a limited speed range of approximately ±30%.T Motor ωr ω1 Generator max 2Δωr Fig. pumped storage power plants [6.3.8) (2.11. A typical application. 43]. as mentioned earlier. 2.7) (2. In the case that the DFIG is Δ-connected the machine can still be represented by this equivalent Y representation. Vr . In this section the jω-method is adopted for calculations. that if the rotor voltage.11 yields [87] Vs = Rs Is + jω1 Lsλ Is + jω1 Lm (Is + Ir + IRm ) Vr Rr = Ir + jω1 Lrλ Ir + jω1 Lm (Is + Ir + IRm ) s s 0 = Rm IRm + jω1 Lm (Is + Ir + IRm ) 14 (2. Note. Applying Kirchhoff’s voltage law to the circuit in Fig. for example. 2. Speed–torque characteristics of a DFIG. Other applications.9) .1 Equivalent Circuit of the Doubly-Fed Induction Generator The equivalent circuit of the doubly-fed induction generator. for DFIG is wind turbines. 2.10. 2. Rs Is jω1 Lsλ jω1 Lrλ Ir Rr /s + Vr s − jω1 Lm Rm IRm the equivalent circuit for the DFIG becomes the ordinary equivalent circuit for a cage-bar induction machine. can be seen in Fig. ﬂywheel energy storage system [4].11. with inclusion of the magnetizing losses. stand-alone diesel systems [78]. or rotating converters feeding a railway grid from a constant frequency public grid [61]. 2. This equivalent circuit is valid for one equivalent Y phase and for steady-state calculations. 2. is short circuited + Vs − Fig. for the DFIG systems are.11.

magnetizing inductance.u. magnetizing resistance.04 0. ω1 − ωr ω2 = ω1 ω1 (2. stator current..01 0.12) (2. The slip.9). rotor resistance.0 15 .01 0.2 2.14) (2.0 Medium Machine 100 kW 0.5 Large Machine 800 kW 0.10) where ωr is the rotor speed and ω2 is the slip frequency. stator leakage inductance. ω1 s slip.15) (2. as Te = 3np Im Ψm I∗ = 3np Im Ψr I∗ r r (2. stator frequency. Te . T YPICAL PARAMETERS OF THE I NDUCTION M ACHINE IN P.where the following notation is used.e. s. stator ﬂux and rotor ﬂux are deﬁned as Ψm = Lm (Is + Ir + IRm ) Ψs = Lsλ Is + Lm (Is + Ir + IRm ) = Lsλ Is + Ψm Ψr = Lrλ Ir + Lm (Is + Ir + IRm ) = Lrλ Ir + Ψm the equations describing the equivalent circuit.7)–(2. can be rewritten as Vs = Rs Is + jω1 Ψs Rr Vr = Ir + jω1 Ψr s s 0 = Rm IRm + jω1 Ψm .11) (2. rotor leakage inductance. (2. Stator and rotor resistance Rs and Rr Leakage inductance Lsλ + Lrλ ≈ Lσ Magnetizing inductance Lm ≈ LM Small Machine 4 kW 0. The resistive losses of the induction generator are Ploss = 3 Rs |Is |2 + Rr |Ir |2 + Rm |IRm |2 and it is possible to express the electro-mechanical torque.3 3. [101]. Ir IRm magnetizing resistance current. Is rotor current.1. equals s= Rs Rr Rm Lsλ Lrλ Lm stator resistance. if the air-gap ﬂux. Moreover.17) (2. U .). Vs stator voltage.1 shows some typical parameters of the induction machine in per unit (p.3 4. i.13) where np is the number of pole pairs.16) (2.. Table 2. TABLE 2.18) (2. Vr rotor voltage.

Moreover. using the expressions in the previous section. In the ﬁgure it can be seen how the mechanical power divides Pmech DFIG Converter sPmech /(1 − s) Fig. the rotor power is approximately minus the stator power times the slip: Pr ≈ −sPs .3.21) (2. given the mechanical power.19) (2.2 Power Flow In order to investigate the power ﬂow of the DFIG system the apparent power that is fed to the DFIG via the stator and rotor circuit has to be determined. r 2 2 (2.12.2.24) 3ω1 sIm [Ψm I∗ ] r ≈ −3ω1 sIm [Ψm I∗ ] r where the approximations are because the resistive losses and the magnetizing losses have been neglected. Pmech /(1 − s) Pmech Grid between the stator and rotor circuits and that it is dependent on the slip. From the above equations the mechanical power produced by the DFIG can be determined as the sum of the stator and rotor power as Pmech = 3ω1 Im [Ψm I∗ ] − 3ω1 sIm [Ψm I∗ ] = 3ωr Im [Ψm I∗ ] . The stator apparent power Ss and rotor apparent power Sr can be found as Ss = 3Vs I∗ = 3Rs |Is |2 + j3ω1 Lsλ |Is |2 + j3ω1 Ψm I∗ s s Sr = 3Vr I∗ = 3Rr |Ir |2 + j3ω1 sLrλ |Ir |2 + j3ω1 sΨm I∗ r r which can be rewritten. Therefore.23) (2. as |Ψm |2 Ss = 3Rs |Is | + j3ω1 Lsλ |Is | + j3ω1 + 3Rm |IRm |2 − j3ω1 Ψm I∗ r Lm Sr = 3Rr |Ir |2 + j3ω1 sLrλ |Ir |2 + j3ω1 sΨm I∗ . 2.25) Then. by dividing Pmech with mechanical rotor speed. In Fig.20) (2. the produced electromechanical torque. r r r (2.18). as mentioned earlier. 2. as given in (2. can be found. 16 . Moreover. ωm = ωr /np .22) Now the stator and rotor power can be determined as Ps = Re [Ss ] = 3Rs |Is |2 + 3Rm |IRm |2 + 3ω1 Im [Ψm I∗ ] ≈ 3ω1 Im [Ψm I∗ ] r r Pr = Re [Sr ] = 3Rr |Ir | − 2 (2. the rotor converter can be rated as a fraction of the rated power of the DFIG if the maximum slip is low. An example of how the stator and rotor powers depend on the slip is shown in Table 2.12 the power ﬂow of a “lossless” DFIG system can be seen. Power ﬂow of a “lossless” DFIG system.2. It can be seen in the table that the power through the converter. this means that Ps ≈ Pmech /(1 − s) and Pr ≈ −sPmech /(1 − s).

the rotor voltage will approximately be rotor = s/(ns /nr )Vs = 0.TABLE 2. 75% of the stator voltage.] 0.13 the stator power is at maximum only 0. ns /nr .7 0 1. as seen in Fig. 2.3/0. slip. i. but the stator and rotor power of the DFIG system is also shown and instead of the rotor speed the slip is shown. In Fig. if the slip s of the DFIG is 30%. [%] rotor speed.13. 2. 2. Moreover.75Vs . rotor power (solid) and stator power (dashed) as a function of wind speed. 0 0.3 rotor power. 2. Typical characteristic for a variable speed DFIG wind turbine.43 · Pmech. Ps 1.4Vs = 0. s. for the wind turbine in Fig. For a wind turbine. stator power. E XAMPLE OF THE P OWER F LOW FOR D IFFERENT S LIPS OF THE DFIG SYSTEM . the average value of the rotor-speed range corresponds to synchronous speed of the DFIG.14 a transformer is placed between the rotor circuit and the converter. Note that VR 17 .u.43 · Pmech. However. ωr .3. the case is not as shown in Table 2. Pmech. if the stator-to-rotor turns ratio. For example. is 0.3 Stator-to-Rotor Turns Ratio Since the losses in the power electronic converter depend on the current through the valves.13.2. b) Mechanical power (dotted). [p.23 · Pmech.e. at low mechanical power the slip is positive and for high mechanical power the slip is negative.2. This is due to the factor 1/(1 − s) in the expressions for the rotor power. for a wind turbine.3 0. The ﬁgure is actually the same as Fig. is higher for positive slips (ωr < ω1 ). The transformer is to highlight and indicate the stator-to-rotor turns ratio.7 times the rated power.77 · Pmech.4. a) Slip as a function of wind speed. the rotor current is approximately 0. In the ﬁgure it is assumed that the gearbox ratio is set in such a way that a) 50 25 b) 100 80 Power [%] 5 10 15 20 25 Slip [%] 60 40 20 0 0 −25 −50 5 10 15 20 25 Wind speed [m/s] Wind speed [m/s] Fig. in general.. if the magnetizing current is neglected.3. which leaves VR rotor = (nr /ns )VR is the actual (physical) room for a dynamic control reserve.4 times smaller than the stator current.0 −0. Pr −0. Moreover. 2. 2. it is important to have a stator-to-rotor turns ratio of the DFIG that minimizes the rotor current without exceeding the maximum available rotor voltage. but it does not exist in reality.3 1. 0.

2.” In the ﬁgure two switches can be seen. This can be done by: 1.4 Lowering Magnetizing Losses In an ordinary induction machine drive the stator is fed by a converter. Still. 2. the DFIG can be controlled in a similar way as an ordinary cage-bar induction generator.DFIG Converter ns /nr Fig. referred to as the Y-Δ-connected DFIG.16 presents a set-up of the Y-Δ-connected DFIG. and accordingly the ﬂux level is closely linked to the stator voltage.14. However. Y-Δ-connected DFIG Fig. As shown in the ﬁgure. while VR is rotor voltage referred to the stator circuit. all rotor variables and parameters are referred to the stator circuit if not otherwise stated. and transmitting all the turbine power through the converter. at low wind speeds the ﬂux level in the generator can be lowered. two methods to lower the magnetizing losses of the DFIG. Before a 18 . short-circuiting the stator of the induction generator at low wind speeds. leading to a better efﬁciency. which means that the magnetizing losses are lowered. that in this operating condition. For instance. Switch S2 is used to disconnect the turbine from the grid and switch S1 is then used to short-circuit the stator of the DFIG. Stator-to-rotor turns ratio indicated with a “virtual” transformer. for the DFIG system there are. At low loads it is possible to reduce the ﬂux level. The inﬂuence that these two methods have on the overall efﬁciency of a DFIG system will be further analyzed in Chapter 3. 2. A brief description of these two systems follows: “Short-Circuited DFIG” Fig. having the stator Δ-connected at high wind speeds and Y-connected at low wind speeds. This means.3. in the DFIG system the stator is connected to the grid. 2.15 shows a diagram of the “short-circuited DFIG. This set-up is referred to as the short-circuited DFIG. at least. a device for changing between Y and Δ connection has been inserted in the stator circuit. Now the turbine is operated as a cage-bar induction machine. which means that it is possible to reduce the losses in the machine by using an appropriate ﬂux level. In this thesis. 2. rotor voltage. except that the converter is connected to the rotor circuit instead of the stator circuit.

16. 2. See Fig.5 Other Types of Doubly-Fed Machines In this section a short presentation of other kinds of doubly-fed machines is made: a cascaded doubly-fed induction machine. Principle of the “short-circuited DFIG.3.15. a single-frame cascaded doubly-fed induction machine. Cascaded Doubly-Fed Induction Machine The cascaded doubly-fed induction machine consists of two doubly-fed induction machines with wound rotors that are connected mechanically through the rotor and electrically through the rotor circuits. 2. 2.S1 Transformer DFIG S2 ≈ = Power electronic converter Fig.” = ≈ Transformer DFIG Y/Δ S1 ≈ = ≈ = Power electronic converter Fig. Principle of the Y-Δ-connected DFIG.17 for a principle diagram. change from Y to Δ connection (or vice versa) the power of the turbine is reduced to zero and the switch S1 disconnects the stator circuit from the grid. Then the stator circuit is connected in Δ (or vice versa) and the turbine is synchronized to the grid. a brushless doubly-fed induction machine. and a doubly-fed reluctance machine. The stator circuit of one of the machines is directly connected to the grid while the other machine’s stator is connected via a 19 . 2.

2. It is doubtful whether it is practical to combine two individual machines to form a cascaded doubly-fed induction machine. Principle of cascaded doubly-fed induction machine. see [106.Converter Fig. while the rotor is based on the principle of reluctance. To avoid a direct transformer coupling between the two-stator windings. it suffers from comparatively low efﬁciency [48]. 111]. it is possible to control the induction machine that is directly connected to the grid with the other induction machine. That is. they can not have the same number of pole pairs. to avoid unbalanced magnetic pull on the rotor the difference between the pole pairs must be greater than one [106]. one winding for the power and one winding for the control. in spite the 20 . Since the rotor voltages of both machines are equal. Due to a large amount of windings. The number of poles in the rotor must equal the sum of the number of poles in the two stator windings [106]. the losses are expected to be higher than for a standard doubly-fed induction machine of a comparable rating [48]. converter to the grid. For further information and more details. Although this machine is mechanically more robust than the cascaded doubly-fed induction machine. 2.18 for a principle sketch. 108. Furthermore. but with the two induction machines in one common frame. Brushless Doubly-Fed Induction Machine This is an induction machine with two stator windings in the same slot. Doubly-Fed Reluctance Machine The stator of the doubly-fed reluctance machine is identical to the brushless doubly-fed induction machine. An equivalent circuit with constant parameters can be obtained for the doubly-fed reluctance machine. See Fig. even though it is the basic conﬁguration of doubly-fed induction machine arrangement. Single-Frame Cascaded Doubly-Fed Induction Machine The single-frame cascaded doubly-fed induction machine is a cascaded doubly-fed induction machine.17. It is possible to achieve decoupled control of active and reactive power of the cascaded doubly-fed induction machine in a manner similar to the doubly-fed induction machine [47].

fact that the machine is characterized by a pulsating air-gap ﬂux.Converter Fig. 2. 21 . It has almost the same equivalent circuit as the standard doubly-fed induction generator [109].18. Principle of the brushless doubly-fed induction machine.

22 .

2) inﬂuence of the converter’s size on the energy production (i. In the ﬁgure it is seen that the ﬁxed-speed system with only one generator has a lower input power at low wind speeds. as described in Section 2. Moreover.2. The following systems are included in this study: • FSIG 1 system — Fixed-speed system. the rated WT power used in this chapter is 2 MW. 3. generator losses and converter losses.1. smaller converter implies a smaller variable-speed range for the DFIG system) and ﬁnally 3) comparison of the DFIG system to other electrical systems. In order to make the comparison as fair as possible the base assumption used in this work is that the maximum (average) shaft torque of the wind turbine systems used should be the same. 3.1.3. as described in Section 2. with one generator.2. in order to compare the performance of the DFIG system. The other systems produce almost 23 . gearbox losses. • VSIG system — Variable-speed system with an induction generator and a full-power converter. the power losses of other systems with induction generators will also be presented.1.2.1 Determination of Power Losses Steady-state calculations are carried out in this section in order to determine the power losses of the DFIG system. 3. The following losses are taken into account: aerodynamic losses.e. Moreover. as described in Section 2. • DFIG system — Variable-speed system with a DFIG.1 Aerodynamic Losses Fig.1 shows the turbine power as a function of wind speed both for the ﬁxed-speed and variable-speed systems.Chapter 3 Energy Efﬁciency of Wind Turbines The purpose of this chapter is to investigate the energy efﬁciency of the DFIG system and to relate this study to other types of WTs with various electrical systems.. This study focuses on 1) reducing the magnetizing losses of the DFIG system. as described in Section 2. • FSIG 2 system — Fixed-speed system.2. with two generators or a pole-pair changing mechanism.2.

The solid line corresponds to the variable-speed systems (VSIG and DFIG) and the two-speed system (FSIG 2).1) where η is the gear-mesh losses constant and ξ is a friction constant. has been used. for instance.3. for a 2-MW gearbox. According to [34]. the stator-to-rotor turns ratio for the DFIG is adjusted so that maximum rotor voltage is 75% of the rated grid voltage.02 and ξ = 0. The dotted line corresponds to a ﬁxed-speed system (FSIG 1). i. This is done in order to have safety margin. 3.nom (3.1. [33].e. a wind gust. In order to avoid making the results dependent on the torque. is. In order to calculate the power in Fig.1.1.3 Induction Generator Losses In order to calculate the losses of the generator. 3. 3.2 Gearbox Losses One way to estimate the gearbox losses.e. For the DFIG system. the equivalent circuit of the induction generator. Ploss. identical results. 3.1.005 are reasonable. a so-called Cp (λ. a dynamic reserve to handle. the inﬂuence of the turbulence is ignored. derived using blade-element theory has been used used.1. speed and pitch control strategy.1. and anyway the settings used by the manufacturers are not in detail known by the authors. In Fig. as described in Section 2. Moreover. 24 ..GB = ηPlowspeed + ξPnom Ωr Ωr. Turbine power. 3.2 the gearbox losses are shown for the investigated systems. the same effect can also be obtained by using different rated voltages on the rotor and stator [81]. the constants η = 0. Observe that instead of using a varying turns ratio. that vary from turbine to turbine. β)-curve. with inclusion of magnetizing losses.GB . only the average wind speed is used in the calculations.. i. the voltage drop across the slip rings has been neglected.3. The power is given in percent of maximum shaft power. see Section 2. Ploss. The interested reader can ﬁnd information of the inﬂuence of the turbulence on the energy production in [69].100 Turbine power [%] 80 60 40 20 0 5 10 15 20 25 Wind speed [m/s] Fig.

DFIG is solid. The losses are given in percent of maximum shaft power.5 1 0. given the same input power. 3.5 0 5 10 15 20 25 Wind speed [m/s] Fig. FSIG 1 and 2 (both one-speed and two-speed generators). Induction generator losses.5 1 FSIG 1 FSIG 2 0.3 2. The solid line corresponds to the variable-speed systems (VSIG and DFIG). It can also be noted that the losses of the DFIG are higher than those of the 3 2.3 the induction generator losses of the DFIG system are shown. dashed is the variable-speed system (VSIG) and dotted are the ﬁxed-speed systems (FSIG 1 and 2). Gearbox losses. 3.5 VSIG 0 4 6 8 10 12 14 16 Wind speed [m/s] Fig. The dotted lines correspond to ﬁxed-speed systems.e. i.3. This implies that the shaft torque of the generators will be different for the two systems.5 FSIG 1 & 2 VSIG DFIG Generator losses [%] 2 1.. 3. The reason for this is that the ﬂux level of the VSIG system has been optimized from an efﬁciency point of view while for the DFIG system the ﬂux level is almost ﬁxed to the stator voltage. VSIG for low wind speeds. The losses are given in percent of maximum shaft power. This means that for the VSIG system a lower ﬂux level is 25 .2. In Fig. The reason that the generator losses are larger for high wind speeds for the VSIG system compared to the DFIG system is that the gearbox ratio is different between the two systems.5 Gearbox losses [%] 2 1.

i. and np = 3. 0. The transistor and the diode can be modeled as constant voltage drops. Rr = 0.4 Converter Losses In order to be able to feed the IG with a variable voltage and frequency source... see Fig.1.u.2) 2 2 rCE Irms mi 4rCE Irms mi cos(φ) √ +√ − 3 cos(φ)6π 45π 3 √ 2 VT 0 Irms 2 Irms VT 0 mi cos(φ) rT Irms √ − Pc.3) 2 2 rT Irms mi 4rT Irms mi cos(φ) √ −√ + 45π 3 3 cos(φ)6π 26 . 3. The conducting losses arise from the current through the transistors and diodes. respectively.4.u. are (with a third harmonic voltage injection) [2] √ 2 VCE0 Irms 2 Irms VCE0 mi cos(φ) rCE Irms √ + + Pc. T1 T2 T3 ⇔ VCE0 rCE VT 0 rT T4 T5 T6 ⇔ Fig.e. 3.. Converter scheme.04 p. The losses of the converter can be divided into switching losses and conducting losses. Lm = 3. For the diode the switching losses mainly consist of turn-off losses [103].u. for a transistor leg.7 p. reverse-recovery energy. Lrλ = 0. For the IGs used in this chapter operated at 690 V 50 Hz and with a rated current of 1900 A and 390 A.e. VCE0 and VT 0 .used for low wind speeds. the following parameters are used: 2-MW power: See Appendix B.. where each transistor.01 p.4-MW power: Rs = 0.T = π 2 6 (3.12 p. Simpliﬁed expressions of the transistor’s and diode’s conducting losses.g. The switching losses of the transistors are the turn-on and turn-off losses. i. the magnetizing losses are reduced. Lsλ = 0. T1 and T4). and a resistance in series. 3..04 p.4.. rCE and rT .. an equivalent circuit of the converter is drawn. A PWM circuit switches the transistors to on and off states.u. The turn-on and turn-off losses for the transistor and the reverse-recovery energy loss for a diode can be found from data sheets.4.D = + π 2 6 (3.. Rm = 192 p. 3. T1 to T6. The duty cycle of the transistor and the diode determines whether the transistor or a diode is conducting in a transistor leg (e.u. is equipped with a reverse diode.u. the IG can be connected to a pulse-width modulated (PWM) converter. In Fig.1.

0 mΩ 128 mJ 2400 A 1200 V 1. 90. In the equations above. Since.8 mΩ 1150 mJ 1. Nominal current IC. The switching frequency used in this thesis is 5 kHz.1 for actual values). C ONVERTER C HARACTERISTIC DATA (IGBT AND I NVERSE D IODE ). be written as √ 2 2 2 Irms + rIGBT Irms . for a given dc-link voltage. it is possible to determine a resistance. IC. have been introduced. rIGBT. This means that for a given dc-link voltage and switching frequency (which both are assumed in this thesis). mi is the modulation index. two voltage drops.6) Irms π IC.nom π where Eon and Eoﬀ are the turn-on and turn-off energy losses.8 mΩ 171 mJ losses can. with the above-mentioned approximation.nom Operating dc-link voltage VCC VCEO Lead resistance (IGBT) rCE Turn-on and turn-off Eon + Eoﬀ energy (IGBT) VT O Lead resistance (diode) rT Reverse recovery Err energy (diode) 500 A 1200 V 1. which are based on [89. as is assumed here [2]. and φ is the phase shift between the voltage and the current.0 V 0.T (3.1 V 1. Pc = Pc.0 V 1 mΩ 863 mJ 1.1 A /IC. Vsw.nom are practically constant for all the valves in Table 3. Hence. that is valid for a nominal current IC.nom and rT IC.nom also are practically constant and equal to each other. rIGBT = rCE ≈ rT and VIGBT = VCEO ≈ VT O .1 V 2.0 V 1. 91. By performing the above simpliﬁcation the model of the IGBT and valve can be scaled to an arbitrary rating. the switching losses of the IGBT and diode can be modeled as a constant voltage drop that is independent of the current rating of the valves.T = (Eon + Eoﬀ ) fsw ≈ Vsw. Using 27 . This is possible since the ratios (Eon + Eoﬀ )/IC. for the transistor.nom = 2Irms max where Irms is the maximum RMS value of the current in the valve.0 V 3 mΩ 288 mJ 1.1 V 0.5 mΩ 86 mJ 1800 A 1200 V 1. for the values in this chapter. This implies that the switching losses from the transistor and the inverse diode can be expressed as √ √ 2 2 Irms 2 2 Ps. it is possible to reduce the loss model of the transistor and the diode to the same model. The conduction TABLE 3.4) The switching losses of the transistor can be considered to be proportional to the current.1 V 1.nom = 1 A.nom and Err /IC.D (3.nom .T + Pc.nom is the nominal current of the valve. the resistance of a speciﬁc valve can be determined from rIGBT = rIGBT.nom is chosen as IC.D = VIGBT π (3. where max IC.nom is the nominal current through the transistor. Moreover.D = Err fsw ≈ Vsw.D .6 mΩ 43 mJ 1200 A 1200 V 1.nom π √ √ 2 2 Irms 2 2 Ps. In this thesis.5) Irms π IC.T and Vsw.1. since the products rCE IC.1. Err is the reverse recovery energy for the diode and IC.1 A . Then. respectively. 92] (see Table 3.5 mΩ 575 mJ 1.where Irms is the root mean square (RMS) value of the (sinusoidal) current to the grid or the generator.

From the ﬁgure it can be noted that the DFIG system and the two-speed system (FSIG 2) has roughly the same total losses while the full-power converter system has higher total losses.GSC + Ploss. Irms .MSC .5 Total Losses The total losses (aerodynamic. assuming a switching frequency of 5 kHz and the resistance rIGBT.38 V.5.8) The total converter losses are now presented as a function of wind speed in Fig. gearbox) are presented in Fig. The losses are given in percent of maximum shaft power. 28 . Now. From the ﬁgure it can. be noted that the converter losses in the DFIG system are 3 2. 3. much lower compared to the full-power converter system. When determining Vsw. converter. the total losses from the three transistor legs of the converter become Ploss = 3(Pc + Ps.T √ 2 2 2 Irms + rIGBT Irms .converter = Ploss. For the MSC.the values given in Table 3.5 Converter losses [%] VSIG 2 1. and rIGBT.T .5 0 5 10 15 20 25 Wind speed [m/s] Fig.1 it is possible to determine the voltage drops Vsw. as expected. Thus it is now possible to calculate the losses of the back-to-back converter as Ploss. The reactive current can be freely chosen. the current through the valves.T + Ps.5 V and Vsw.D . 3. is the stator current for the VSIG system or the rotor current for the DFIG system. 3.T = 2. generator. + Vsw. adjusted with the ratio between machine-side voltage and the grid voltage. Converter losses.1.5. DFIG is solid and VSIG is dashed.5 1 DFIG 0. 3. (3.76 Ω.1 A the average values of all of the valves in Table 3. Vsw.1 A = 1.D = 0.D ) π (3.D ) = 3 (VIGBT + Vsw.6.1 has been used.7) The back-to-back converter can be seen as two converters which are connected together: the machine-side converter (MSC) and the grid-side converter (GSC). One way of calculating Irms for the GSC is by using the active current that is produced by the machine.

3.6. the power loss as a function of transmitted power (or wind speed) was determined. the most important quantity is the energy delivered to the grid (electric energy capture). FSIG 1 and 2 is dotted and VSIG is dashed. 3. for low wind speeds.7 illustrates the impact of having a smaller converter and thus a smaller rotor-speed range. Given a probability density functions. 3. 3. In order to do this. in this section the results in the previous section have been used to determine the energy capture (or energy efﬁciency) of the various systems. f (w). However.. the aerodynamic losses become higher.8 the converter losses are presented for different designs of the rotor-speed range. This means that the smaller the converter is. As mentioned earlier one commonly used probability density functions to describe the wind speed is the Rayleigh function [53]. 3. the distribution of wind speeds must be known. i. can be found as Pavg = 0 ∞ P (w)f (w)dw (3. The losses are given in percent of maximum shaft power. In Fig. Total losses. it is not possible to obtain a full speed range with the DFIG system if the converter is smaller than the rated power of the turbine. It can be seen in the ﬁgure that the converter losses are lower for smaller rotor-speed ranges (or smaller converter ratings).e. the more the WT will operate at a non-ideal tip-speed ratio.9) where w is the wind speed.. Note. the average (or expected) value of the power. a smaller rotor-speed range implies smaller ratings of the converter. as mentioned earlier.2 Energy Production of the DFIG System In the previous section. λ.2.e. P (w).8 7 VSIG Total losses [%] 6 5 4 3 2 1 0 5 10 15 20 25 FSIG 1 & 2 DFIG Wind speed [m/s] Fig. i. that the stator-to-rotor turns ratio has to be 29 . DFIG is solid. for the wind-turbine application. Fig. Accordingly.1 Investigation of the Inﬂuence of the Converter’s Size on the Energy Production As was mentioned earlier.

4 0. that the rotor-speed range is of greater importance for a low average wind-speed compared to a high average wind speed. as expected. 3. it is possible to run at optimal tip-speed ratio in the whole variable-speed area.26 20 22 Turbine power [%] 4 6 8 10 Rotor speed [rpm] 15 18 10 14 5 10 0 4 6 8 10 Wind speed [m/s] Wind speed [m/s] Fig. It can be seen in the ﬁgure that the gain in energy increases with the rotor-speed range (converter size). designed according to desired variable-speed range in order to minimize the converter losses. Rotor speed and the corresponding turbine power. even though the converter losses of the DFIG system increase with the rotor-speed range (converter size). However. 3.8.1 0 5 10 15 20 15-25 rpm 18-25 rpm 21-25 rpm 25 Average wind speed [m/s] Fig. 0. It can be seen in the ﬁgure.8 12-25 rpm 0. The increased aerodynamic capture has thus a larger impact than the increased converter losses. In Fig.7. 3.6 0. 3.2 0. the most interesting information is the total energy efﬁciency.3 0.5 0.8. as shown in Fig. 30 .7 Converter losses [%] 0. Converter losses for some different rotor-speed ranges as a function of the wind speed.9 the energy efﬁciency of the DFIG for different rotor-speed ranges (or converter sizes) can be seen. If the rotor-speed range is set to 12–25.

4 there are at least two ways of lowering the magnetizing losses. 2. this can be done by: 1. referred to as the Y-Δ-connected DFIG. The systems are the DFIG system.2. the full variable-speed 31 .4 m/s (solid).11 the produced grid power together with the various loss components for an average wind speed of 6 m/s are presented for the various systems. and the other variants will not be subjected to any further studies. It can be seen in the ﬁgure that the Y-Δ-connected DFIG system produces approximately 0. 3. i. Note that the aerodynamic efﬁciency is also taken into account. 6. Efﬁciency.3 Comparison to Other Wind Turbine Systems The base assumption made here is that all wind turbine systems have the same average maximum shaft torque as well as the same mean upper rotor speed.8 m/s (dashed) and 8.9.2 Reduction of Magnetizing Losses As presented in Section 2. 3. The break-even point of the total losses or the rated values of the equipment determines the switch-over point for the doubly-fed generators. short-circuiting the stator of the induction generator at low wind speeds. 3. of the DFIG system as a function of the rotor-speed range.e.96 94 Efﬁciency [%] 92 90 88 86 24−25 20−25 16−25 12−25 Rotor-speed range [rpm] Fig. This set-up is referred to as the short-circuited DFIG.. i. 3. for average wind speeds of 5.3. In Fig. 3. having the stator Δ-connected at high wind speeds and Y-connected at low wind speeds.. Since the Y-Δ-connected DFIG system performs better than the short-circuited DFIG system the Y-Δ-connected DFIG system will henceforth be referred to as the DFIG system. the Y-Δ coupling or the synchronization of the stator voltage to the grid. at least for low average wind speeds.2 m/s (dotted). and transmitting all the turbine power through the converter.10 the energy gain using the two methods are presented.e. In Fig.2 percentage units more energy than the short-circuited DFIG system.

The produced average grid power and generator.10. a variablespeed system equipped with a permanent magnet synchronous generator (PMSG). aerodynamic efﬁciency.2 Gain in energy [%] 1 0. with respect to the rotor speed.1.6 0. a transistor rectiﬁer has the potential to utilize the generator best [32]. which may give a possibility to reduce the converter losses [32]. The average efﬁciency for the PMSG is taken from [34]. 32 . It would also be possible to have the PMSG connected to a diode rectiﬁer with series or shunt compensating capacitors. The converter losses of the PMSG system are assumed equal to that of the VSIG system. Gain in energy production by lowering the magnetizing losses for a DFIG system as a function of the average wind speed. 100% correspond to the input turbine power at optimal.11. system (VSIG). However. 3.2 0 5 6 7 8 9 10 Average wind speed [m/s] Fig.8 0.4 0. In the ﬁgure it can be seen that the one-speed system 100 Average power [%] 95 90 85 Gearbox losses Generator losses Converter losses Grid power FSIG 1 FSIG 2 VSIG PMSG DFIG 80 Fig. one-speed system (FSIG 1).4 1. and. 3. Solid line is the Y-Δ-connected DFIG and dashed line is short-circuited DFIG. two-speed system (FSIG 2). converter and gearbox losses for an average wind speed of 6 m/s.

that the base assumptions differ. it was found in this investigation that there is a possibility to gain a few percentage units (approximately 2%) in energy using the DFIG system compared to the full variable-speed system. 3. However. PMSG and the DFIG system as a function of the average wind speed. VSIG. and the PMSG system have almost the same efﬁciency. However. 96 94 Produced energy [%] 92 90 88 86 84 5 6 7 8 9 10 PMSG FSIG 2 FSIG 1 VSIG DFIG Average wind speed [m/s] Fig.4 Discussion In Fig. 3. 3. the boost converter on the dc-link. for various average wind speeds. Reference [16] sets the rating of the stator windings equal while we choose the shaft power and maximum speed instead. transistor converter towards the grid and the magnetizing system of the generator are in the same range as the PMSG system. Further.12. the DFIG is operated with a rotor-speed range set to 12–25 rpm. However. PMSG and DFIG). This can be compared to a gain of 20% for the DFIG system compared to the variable-speed system reported in [16].(FSIG 1) has the disadvantage of poor aerodynamic efﬁciency. in this study the produced energy of the systems was found to be similar. the DFIG system. The reason for the difference is again. with the two-speed system (FSIG 2) the aerodynamic efﬁciency is improved and close to the variable speed systems (VSIG. 3. In [16] it was found that the DFIG system produced 60% more energy compared to a ﬁxed-speed system. 33 . it is reasonable to assume that the losses in the diode rectiﬁer connected to the stator. In the ﬁgure.12 it can be seen that the two-generator system (FSIG 2). a lower ﬁxed-speed IG WT efﬁciency was reported than in this study. The difference between the result here and in [16] is due to the different base assumptions used.12 the produced energy of the different systems. The reason for this is that the IG in this study has two generators and lower iron losses. FSIG 2. In Fig. In [35]. Detailed information of the gearless electrically magnetized generator system was not available. are presented. Energy efﬁciency of the FSIG 1.

two methods to reduce the magnetizing losses (and thereby increase the gain in energy) of the DFIG system. it has been found that there is a possibility to gain a few percentage units in energy efﬁciency for a doubly-fed induction generator system compared to a cage-bar induction generator. So. only the average wind speed has been used and the inﬂuence of turbulence has not been treated. among other factors. the turbulence intensity. aerodynamics must be accounted for when ﬁxed-speed and variable-speed turbines are compared. Also. controlled by a full-power converter. the aerodynamic capture of the wind turbine is reduced with a smaller rotor-speed range. However. Reference [69] showed that a two-speed active stall turbine and a variable-speed pitch turbine is fairly unaffected by the turbulence intensity for turbulence intensities up to 15%. Worth stressing is that the main reason for using a variable-speed turbine instead of a ﬁxed-speed turbine is not the energy efﬁciency. thus. In comparison to a direct-driven permanent-magnet synchronous generator. For the more unusual turbulence intensities of 20–25% the variable-speed turbines gained a couple of percentage units in energy production compared to the two-speed active stall-regulated system. the time delay between generator switchings for the ﬁxed-speed systems. Finally. instead it is the possibility of lowering the mechanical stresses [53] and also improving the power quality [58]. Moreover. the ambition has instead been to make the comparison as clean as possible. This means that the increased aerodynamic capture that can be achieved by a larger converter has. it was found that the converter losses of the DFIG can be reduced if the available rotor-speed range is made smaller. using only the facts that can be presented clearly and with best certainty. Another problem when incorporating the effect of the turbulence intensity is that the selection of torque. start and stop. in order to perform a detailed energy capture calculation. regardless of value plays an unimportant role since the torque and speed control of the turbines are in principle the same.The focus in this chapter is on the electrical energy efﬁciency of the DFIG-system in relation to other systems.5 Conclusion In this chapter. However. a greater impact than the increased converter losses. 3. 34 . have been investigated. In order to reduce the number of uncertainties. Of great importance to point out is that when comparing the DFIG system to the full variable system. speed and pitch control inﬂuences the result. (The rotor-speed range of the DFIG system is assumed to be almost the same as for the full-variable speed system). in order not to include uncertainties that might not be the best chosen. controlled by a converter or a two-speed generator system the difference in energy efﬁciency was found to be small. It was found that the method utilizing a Y-Δ switch in the stator circuit had the largest gain in energy of the two investigated methods. Δ-Y-reconnections for the DFIG-systems must be known.

1 Space Vectors The idea behind space vectors is to describe the induction machine with two phases instead of three. forms a rotating ﬂux in the air gap. which is supplied with three-phase currents.1 Introduction In this section.1: Principle of space vectors.1. 4. space vector. A three-phase stator winding. different aspects of designing and implementing control systems for doublyfed induction generators (DFIGs) are treated. 4. 4.1) . The same rotating ﬂux could also be formed with only two phases. as seen in Fig. of the three-phase quantities. sb . sa . In order to determine the Im ψ Im ψ Re ⇐⇒ Re Fig.1. the following transformation is applied [39] ss = sα + jsβ = 2K sa + asb + a2 sc 3 35 (4. This is the principle of space vectors. Space vectors were originally invented to describe the spatial ﬂux of an ac machine [39]. ss .Chapter 4 Control of Doubly-Fed Induction Generator System 4. and sc .

1.e.. 4. i. A general space vector can be expressed as ss = sej(θ1 +φ) (4. as the corresponding imaginary part of the above equation [5]: Q = 3Im [vi∗ ] . ω1 .. The constant K can be chosen arbitrary.where K is the space-vector scaling constant and a = ej2π/3 .1.2) 2 the space vector will be scaled according to the RMS value of the three-phase quantities. (4.4) The synchronous coordinate system is not indicated by a superscript.8) (4. e. the grid voltage. This choice of K will be used throughout this thesis.g.e. for example. Superscript “s” indicates that the space vectors are referred to the reference frame of the stator of the induction machine. though if it is chosen as 1 K=√ (4.7) 4. However.2 Power and Reactive Power in Terms of Space Vectors The instantaneous power. and its corresponding angle.9) . 2 2 2K 2K √ The above-mentioned scaling. normally the stator or rotor ﬂux of an induction machine. the same as active power in terms of phasors.3) where φ is a phase shift and θ1 is the synchronous angle corresponding to the synchronous frequency. The PLL-type estimator used in this thesis is described by [37] dˆ 1 ω = γ1 ε dt ˆ dθ1 = ω1 + γ2 ε ˆ dt 36 (4. Q. θ1 . i. yields P = 3Re [vi∗ ] . ω1 . K = 1/ 2. the synchronous frequency. (4.5) (4. It is also possible to deﬁne a quantity the instantaneous reactive power.3 Phase-Locked Loop (PLL)-Type Estimator A PLL-type estimator can be used for estimation of the angle and frequency of a signal.6) In (4. Space vectors in synchronous coordinates will be dc quantities in the steady state. (4. in a three-phase system is given by 3 3 P = v a ia + v b ib + v c ic = Re [vs (is )∗ ] = Re [vi∗ ] .6) the instantaneous power is the real part of voltage times the complex conjugate of the current. as dω1 /dt = θ1 . P . It is possible to transform the vector to synchronous coordinates (dq coordinates) as s = sd + jsq = e−jθ1 ss = sejφ . The synchronous coordinate system has to be aligned with a quantity. it is also possible to align the synchronous coordinate system with.

The notation “ˆ” indicates an estimated variable or parameter.1. asymptotically. 44.9). ˆ G(p) and a process model.11) (4. This implies that ω1 and θ1 will converge to ω1 and θ1 ˆ1 is small.8)–(4. Controller design is then just a matter of choosing the “right” transfer function C(p).15) . If the true frequency and position are given by dω1 /dt = 0 and dθ1 /dt = ω1 .8). 4. it is possible to approximate respectively. If the parameters are chosen as γ1 = ρ 2 γ2 = 2ρ (4. such as the grid frequency. If the difference θ1 − θ ˆ ˆ sin(θ1 − θ1 ) ≈ θ1 − θ1 .2. the order of the denominator is equal to or greater than that of the numerator.4 Internal Model Control (IMC) Due to the simplicity of IMC for designing controllers. see Fig.14) (4. i. For small bandwidths the rejection is twice as good. IMC can.12) In [76] it is shown that the modiﬁed PLL-type estimator rejects better voltage harmonics than the PLL-type estimator in (4.9).ˆ ˆ where ε = sin(θ1 − θ1 ) and θ1 − θ1 is the error in the estimated angle.8)–(4. and the following characteristic polynomial of the system described by (4. Modiﬁed PLL-Type Estimator If the PLL-type estimator should synchronize to a constant (or at least close to constant) frequency. The idea behind IMC is to augment the error between the process. be used for designing current or speed control laws of any ac machine [40. then the modiﬁed PLL-type estimator becomes [76] ˆ dθ1 = ω1 = ωg + ρε. The closed-loop system will be ˆ Gcl (p) = G(p) 1 + C(p)[G(p) − G(p)] which simpliﬁes to Gcl (p) = G(p)C(p) = 37 α p+α n −1 C(p) (4. G(p).e.10) then ρ can be adjusted to the desired bandwidth of the PLL-type estimator. by a transfer function C(p). 102].13) where n is chosen so that C(p) become proper. it is possible to simplify the PLL-type estimator in (4. 4. γ2 } > 0. In the above equations γ1 and γ2 are gain parameters. dt (4.. This is done by neglecting (4.8) and (4. this method will be used throughout this thesis.9) can be found: p 2 + γ2 p + γ1 where p = d/dt. for instance. then it is ˜ ˆ ˆ shown in [37] that the estimation error equations for ω1 = ω1 − ω1 and θ1 = θ1 − θ1 are ˜ ˆ ˆ asymptotically stable if {γ1 . One common way is [31] C(p) = α p+α n ˆ G−1 (p) (4.

1. The relationship between the bandwidth and the rise time (10%–90%) is α = ln 9/trise . the transfer function from the load disturbance E to output signal i is given by GEi (p) = G(p) p = G(p) 1 + G(p)F (p) p+α (4. as the dynamics of the process. 4. can improve the disturbance rejection.18) if all parameters are assumed to be known. (inside the dashed area in Fig. addition of an inner feed-back loop.20) GEi (p) = K (p + α)2 38 .17) 4. the transfer function in (4. −1 (p) + R p + α 1 + G(p)R p+αG (4. Then. (4. 4.2. ˆ when G(p) = G(p). Principle of IMC. However. the disturbance rejection is to a large extent determined by the process [76].18) is changed to GEi (p) = p G(p) p 1 = .16) For a ﬁrst-order system. The controller then typically becomes an ordinary PI controller: F (p) = α ˆ −1 ki G (p) = kp + . see Fig. The controller.3.2) becomes ˆ F (p) = 1 − C(p)G(p) −1 C(p). G(p). is set to the desired bandwidth of the closed-loop system. p p (4. the load disturbance rejection should be sufﬁcient.19) For a ﬁrst-order system it is possible to choose R so that the above transfer function can be reduced to p (4. Therefore. F (p). The parameter α is a design parameter. which for n = 1. are normally much slower than the dynamics of the closed-loop system. n = 1 is sufﬁcient. 4.iref + − C(p) v G(p) ˆ G(p) − + i F (p) Fig. If the dynamics of G(p) are fast.5 “Active Damping” For a ﬁrst-order system designed with IMC.

the following PI controller can be found: F (p) = R + Ra α −1 G (p) = αL + α . Principle of “active damping.24) Then. L (p + α)2 (4.iref + − F (p) v + − v + E(p) + G(p) i R Fig. −1 (p) + R p+αG a p .” “Active damping” has been used for the cage-bar induction machine to damp disturbances.21) where i is the state (current).3.22) which has the following transfer function G(p) = (4.26) This means that the disturbance is damped with the same time constant as the dynamics of the control loop. 39 . = v (p) Lp + R + Ra (4. p p (4. where the term Ra i is the “active damping” term. and e is a load disturbance (back emf). consider the following ﬁrst-order system (e. v is the input signal (applied voltage).23) By using IMC. This will be refereed to as “active damping” or “active resistance. the transfer function from the load disturbance e to the output i can be determined as Gei (p) = p 1 .19). For example. 41]. 4. This means that a load disturbance E is damped with the same time constant as the control loop.” where K is a constant. the transfer function is reduced to Gei (p) = (4. such as varying back EMF [18. a dc machine) L di = v − Ri + e dt (4. from (4. Then the “active damping” can be introduced by letting v = v − Ra i. the system can be rewritten as L di = v − (R + Ra )i + e dt i(p) 1 ..25) By choosing Ra = αL − R.g. Then.

28) (4. so called integrator wind-up. when implementing control laws in computers. For a linear time-invariant discrete system the corresponding stability region is inside the unit circle [88]. This might cause overshoots in the controlled variable since the integration part of control law will keep the ideal control signal high even when the controlled variable is getting closer to the reference value [39]. does not exceed the maximum value.1.4.. The algorithm can be described as [39] u = kp e + ki I usat = sat(u) dI usat − u =e+ dt kp where e is the control error and I is the integral of the control error.7 Discretization Throughout the thesis. i.. Therefore.30) where n indicates the sample number.29) 4. This causes the integral part of the PI controller to accumulate the control error during the saturation. Mapping the unit circle 40 . The idea behind the back-calculation method is to modify the reference value in case of saturation. Stability of a linear time-invariant continuous system requires that the poles are in the left half plane. they have to be discretized.27) (4. the control signal must be limited (saturated). (4. However. The forward Euler discretization can also be written as p −→ q−1 Tsample (4. the control signal cannot be arbitrary large due to design limitations of the converter or the machine.e.6 Saturation and Integration Anti-Windup When designing control laws.e. For a continuous system given as x(t) = Ax(t) + Bu(t) ˙ y(t) = Cx(t) the discrete equivalent using the forward Euler method becomes x(n + 1) = (I + ATsample )x(n) + Tsample Bu(n) y(n) = Cx(n). One method to avoid integration wind-up is to use the “back-calculation” method [39]. differential equations and control laws will be described in continuous time.34) (4. u. so that the ideal control signal. a derivative is approximated as x(n + 1) − x(n) dx ≈ dt Tsample (4.35) (4.31) (4. at time t = nTsample .32) where q is the forward shift operator. The forward Euler method will be used.33) (4.1. |u| = umax . i.

the system consists of a DFIG and a back-to-back voltage source converter with a dc link.1 Machine Model Due to its simplicity for deriving control laws for the DFIG. As mentioned earlier. in this thesis the L-grid ﬁlter will be used. The name is due to the formation of a “Γ” of the inductances.6.2. a grid ﬁlter is placed in between the GSC and the grid. Fig. 4. the rotor and the stator leakage inductance have the same effect. that from a dynamic point of view. the Γ representation of the IG model will be used.5 an equivalent circuit of the DFIG system can be seen. 0) in order for the forward Euler discretization to be stable. as shown in Fig. see Fig. in (4. since both the grid and the voltage source converter are voltage stiff and to reduce the harmonics caused by the converter.5. 4. In addition. becomes stable. As seen in the ﬁgure the poles must be inside a circle with the radius of 1/Tsample with the center point located at (−1/Tsample . The back-to-back converter consists of a grid-side converter (GSC) and a machine-side converter (MSC). it is possible to use a different representation of the Park model in which the leakage inductance is placed in the rotor circuit. the so-called Γ representation of the induction machine [94]. More detailed description of the models of the components of the DFIG system will be performed in the following sections. 4.4: Region of stability. the variables and the parameters in Fig.4 shows where the poles of continuous system must be located so that the forward Euler discretization. 4. 4.35) gives the region in the p plane where the poles of the continuous system must be located in order to get a stable discretization [38]. 4. This model is described by the following space-vector 41 . However. onto the continuous p plane using (4. Note.5 will also be explained. Therefore.35). For voltage source converters the grid ﬁlter used is mainly an L-ﬁlter or an LCL-ﬁlter [62]. 4.2 Mathematical Models of the DFIG System In Fig. 4. Moreover.Im p plane 1/Tsample Re Fig.

39) . equations in stator coordinates [94]: dΨs s = + dt dΨs R s s vR = RR iR + − jωr Ψs R dt s vs Rs is s (4. Equivalent circuit of the DFIG system.5. The model can also be described in synchronous coordinates as dΨs + jω1 Ψs dt dΨR vR = RR iR + + jω2 ΨR dt vs = Rs is + where the following notation is used: 42 (4.36) (4.37) where superscript s indicates stator coordinates. Γ representation of the IG in stator coordinates.DFIG is g + Es g − is f is s + s vs Rs LM Lσ RR jωr Ψs R is R + s vR MSC ≈ = − Rf Lf + s vf − + Cdc = GSC vdc − dc-link ≈ − Grid ﬁlter Fig. 4. Superscript “s” indicates that the space vectors are referred to the reference frame of the stator of the DFIG.6. 4.38) (4. + s vs is s Rs Lσ RR jωr Ψs R + − iR s + s vR LM − − Fig.

ω2 slip frequency. Lσ is the leakage inductance.2. and vf is the grid-ﬁlter voltage supplied from the grid-side converter. stator current. Finally. and electromechanical torque are given by Ψs = LM (is + iR ) ΨR = (LM + Lσ )iR + LM is = Ψs + Lσ iR Te = 3np Im Ψs i∗ R (4. Ψs stator ﬂux. The stator ﬂux. ω1 synchronous frequency. 4. rotor ﬂux.44) where Eg is the grid voltage. if is the grid-ﬁlter current.42) where LM is the magnetizing inductance.vs stator voltage. vR rotor voltage.43) where J is the inertia and Ts is the shaft torque. Applying Kirchhoffs voltage law to the Rf Lf is f + Es g − + s vf − Fig. Rs stator resistance.2 Grid-Filter Model In Fig. circuit in the ﬁgure the following model in synchronous coordinates can be found: Eg = − (Rf + jω1 Lf ) if − Lf dif + vf dt (4. The quantities and parameters of the Γ model relate to the Park model (or the T representation) as follows: vR = γvR RR = γ 2 Rr iR = ir ΨR = γΨr γ Lσ = γLsλ + γ 2 Lrλ γ= LM Lsλ + Lm Lm = γLm .40) (4. ΨR rotor ﬂux. the mechanical dynamics of the induction machine are described by J dωr = Te − Ts np dt (4. 4. 4. The ﬁlter consists of an inductance Lf and its resistance Rf .7 the equivalent circuit of the grid ﬁlter in stator coordinates can be seen. 43 . and np is the number of pole pairs. iR rotor current. Grid-ﬁlter model in stator coordinates.7.41) (4. is RR rotor resistance.

Moreover. 4.u. dt 2 dt This means that the dc-link voltage will vary as Cdc vdc dvdc = −Pf − Pr dt (4. (4.8.8 an equivalent circuit of the dc-link model.3 DC-Link Model The energy. if the switching frequency of the converter is ω = 100 p. 4. GSC + vdc − Pf ≈ = considered small and thereby be neglected.u. and Lf = 0.2 p. is given by 1 2 Wdc = Cdc vdc 2 (4.. DC-link model. Cdc . can be seen. then the gain of the grid ﬁlter is |Gf (j100)| ≈ 0. Wdc . (corresponding to a damping of 26 dB) if Rf can be neglected. Gf (p) = s = vf (p) Lf p + Rf This means that the damping of the grid ﬁlter is given by |Gf (jω)| = 1 2 L2 ω 2 + Rf f (4. In Fig. where the deﬁnition of the power ﬂow through the grid-side converter (GSC) and the machineside converter (MSC. Pr .2.05 p. and the power delivered to the rotor circuit of the DFIG.Harmonics The transfer function. stored in the dc-link capacitor. 44 .49) (4. the energy in the dc-link capacitor is dependent on the power delivered to the grid ﬁlter. the gain can be approximated as |Gf (jω)| ≈ 1/(Lf ω).47) where vdc is the dc-link voltage. as [76] dWdc d 2 1 = Cdc vdc = −Pf − Pr .48) which means that Pf = −Pr for a constant dc-link voltage.46) If Lf ω Rf . 4. Pf . of the grid ﬁlter can be expressed as is (p) 1 f .u.45) . Gf (p). if the losses in the actual converter can be MSC ≈ Cdc = Pr Fig. For example.

i.55) (4. 99]. pure stator-voltage orientation can be done without any signiﬁcant error.50) that the stator voltage. (4. According to [17].3 Field Orientation In order to control the rotor current of a DFIG by means of vector control. If the stator resistance is considered to be small. 68].53) (4. the reference frame has to be aligned with a ﬂux linkage.2.57) (4. 61. presented in Fig. 61.e.50) (4. As illustrated by the ﬁgure there is only a small angular difference between the grid-voltage and stator-ﬂux space vectors in the stator-ﬂux reference frame compared to the grid-ﬂux reference frame.52) (4. the machine is aligned with a virtual grid ﬂux. from now on.9 shows an example of the space vectors of the grid voltage and the stator ﬂux.1 Stator-Flux Orientation For a stator-ﬂux-oriented system the synchronous angle θ1 is deﬁned as θ1 = ∠Ψs s 45 (4.51) (4. Note that in this thesis stator-voltage orientation will be.56) (4. 4. 110].60) .54) Note that in (4.3. 4. vs . referred to as grid-ﬂux orientation [21]..4. 80.59) = Eg − Rs is − jω1 Ψs = vR − RR iR − jω2 ΨR = vf − (Rf + jω1 Lf ) if − Eg = −Pf − Pr = Te − Ts (4. or with air-gap-ﬂux orientation [107. 4.5 can now be summarized in synchronous coordinate.4 Summary The total model of the DFIG system. Fig. One common way is to control the rotor currents with stator-ﬂux orientation [46.58) (4. as dΨs dt dΨR dt dif Lf dt dvdc Cdc vdc dt J dωr np dt where Ψs = LM (is + iR ) ΨR = Lσ iR + LM (is + iR ) Te = 3np Im Ψs i∗ R Pr = 3Re [vR i∗ ] R Pf = 3Re vf i∗ f . Eg . stator-ﬂux orientation gives orientation also with the stator voltage [17. 4. has been changed to the grid voltage.

i.. is θ1 = ∠Ψs = ∠ − jEs = θg − g g π . ˆ i. 2 (4. and the grid voltage angle. the space vector of the ﬂux is real valued.64) where Eg is the grid voltage magnitude. that Eg = jEg . rotor current and the rotor position are measured. 4.e. ψs is the stator ﬂux magnitude. since then vs = Eg = Rs is + dΨs + jω1 Ψs ≈ jω1 Ψs dt 46 (4. θ1 . This means that for perfect ﬁeld orientation.63) Then.e. Note that the grid-ﬂux orientation is equal to the stator-ﬂux orientation in the steady state. ˆ θ1 = θ1 . that Ψs = ψs .2 Grid-Flux Orientation The basic idea behind grid-ﬂux orientation is to deﬁne a virtual grid ﬂux. Since ωg is close to constant. i. Ψs . b) Grid-ﬂux orientation. if the stator current.3. i. the virtual grid ﬂux is linked to the grid voltage. and θ1 = θ1 − θ1 is the error between the synchronous angle and its estimate. the grid voltage can be transformed to synchronous coordinates as Eg = Es e−j θ1 = jEg ej θ1 g ˆ ˜ (4. for a grid-ﬂux oriented (or stator-voltage oriented) system.9.q Eg q Eg Ψs d b) Ψg d Ψs a) Fig.62) where ωg is the frequency of the grid voltage and θg is the corresponding angle.65) . Then the stator ﬂux can be transformed to s synchronous coordinates as Ψs = Ψs e−j θ1 = ψs ej θ1 s ˆ ˜ (4.. Space-vector diagram of grid voltage and stator ﬂux. the stator ﬂux can be calculated and thus the transformation angle can be found. the space vector of the grid voltage is imaginary. 4.. This means that for perfect ﬁeld orientation.e. θ1 = θ1 . where Ψs is the stator ﬂux in stator coordinates. if the stator resistance can be neglected. Moreover.61) ˆ ˜ ˆ where θ1 is the estimate of θ1 . as [21. a) Stator-ﬂux orientation. This means that the relationship between the synchronous angle. 76] g Ψs = g Es jEg ejθg g =− jωg ωg (4.e. θg .

which yields dΨs Rs + jω1 Ψs + dt LM diR dΨs vR = (RR + jω2 Lσ )iR + Lσ + + jω2 Ψs dt dt diR = (RR + Rs + jω2 Lσ )iR + Lσ +E dt Rs + jωr Ψs E = vs − LM vs = −Rs iR + (4.4 Control of Machine-Side Converter The main task of the machine-side converter is. 80]. ˆ ˆ (4. ˆ an estimate of the whole back EMF. (4.3. it is common to control the rotor current with either stator-ﬂux orientation or grid-ﬂux orientation. it is possible to include a feed-forward compensating term in the control law that will compensate for the tracking error caused by variations in the back EMF.4. Further. Here.” Ra . In [46. for instance. for example. it is advantageous to eliminate is and ΨR from (4. can be used to track the grid frequency and its corresponding angle.66) (4. 80] this is done by feed forward of the term jω2 Ψs and neglecting the derivative of the ﬂux in (4. has been introduced.1. However.67) (4. the reactive power at the stator terminals. 47 .39). in (4. 4.68) where E is the back EMF. an “active resistance. E.1 Current Control As mentioned earlier.67)—in the control law [17. as described in Section 4.38) and (4.67).and ω1 = ωg . of course.69) where “ˆ” indicates an estimated quantity. will be used: ˆ ˆ ˆ vR = vR + (j ω2 Lσ − Ra )iR + kE E = kp e + ki ˆ e dt + (j ω2 Lσ − Ra )iR + kE E. A coefﬁcient kE is introduced in order to make the control law more general and to simplify the analysis in Chapter 5: kE = 0 for control without feed forward of E 1 for control with feed forward of E. In order to derive the rotor-current control law. 4. It is possible to decouple the cross coupling between the d and q components of the rotor current—jω2 Lσ iR in (4. The ﬁeld orientation could. This is done by having an inner fast ﬁeld-oriented current control loop that controls the rotor current. 61. The “active resistance” is used to increase the damping of disturbances and variations in the back EMF. either be aligned with the stator ﬂux of the DFIG or the grid ﬂux.69). For both reference frames the q component of the rotor current largely determines the produced torque while the d component can be used to control.70) Furthermore. The transformation angle for a grid-ﬂux oriented system can be found directly from measurements of the stator voltage. to control the machine. in order to have some ﬁltering effect a PLL-estimator.

the d and the q components of the rotor current will not be decoupled. kE = 0 in (4.. is ˆ put to zero in (4. p + αc (4. since it is an order of magnitude smaller than the term jω2 Ψs . 41].Similar approaches have been used for the squirrel-cage IG [18. the rotor current dynamics formed by the inner loop in Fig.10.69).67). it is treated as a disturbance to the rotor current dynamics.10 are now given by Lσ diR = vR − (RR + Rs + Ra )iR dt (4. 4.72) where αc is closed-loop bandwidth of the current dynamics.73) DFIG ˆ E E vR + +vR + − − iref+ R − kp + ki p G(p) iR ˆ ˆ Ra − j ω2 Lσ Fig. If the estimate of the slip frequency. ω2 .69).69) in (4. The transfer function from vR to iR is G(p) = 1 pLσ + RR + Rs + Ra which via (4. Block diagram of the current control system. since for a DSP-based digital controller it is easy to implement. 4. i. 48 . How to choose the “active resistance” will be shown in next section.e. here the d and q components of the rotor current will be decoupled.71) where the estimated parameters in the control law are assumed to have the correct values.17) yields the following controller parameters ˆ kp = αc Lσ ˆ ˆ ki = αc (RR + Rs + Ra ) (4. giving Gcl (p) = p . Nevertheless. If the back EMF is not compensated for. In [46] it is stated that the inﬂuence of the decoupling term jω2 Lσ iR is of minor importance. Substituting (4. The dashed box is the model for the doublyfed induction generator.

typiαc Lσ .u.75) −p kR = 0 (p + αc )(pLσ + RR + Rs ) GEi (p) = −p ⎪ ⎩ kR = 1 Lσ (p + αc )2 (4. Since Ra should be greater than zero. which can be considered as a low value since for modern drive.. the damping of low-frequency disturbances is signiﬁcantly improved.74). It can be seen that when using “active resistance. is given as GEi (p) = −p/(p + αc ) pLσ + Lσ αc kR + (1 − kR )(Rr + Rs ) (4. E.10.70): kR = This yields 0 for control without “active resistance” 1 for control with “active resistance. A parameter kR is introduced in a fashion similar to (4. 49 . cf. with and without “active resistance”: Gr (ω) = |GEi (jω)|kR =1 = |GEi (jω)|kR =0 (RR + Rs )2 + ω 2 L2 σ .78) The following two extreme values of the above ratio are worth noting: ⎧ ⎨ RR + Rs when ω −→ 0 Gr (ω) = αc Lσ ⎩ 1 when ω −→ ∞ (4. a current control loop bandwidth of 7 p.min equals 0.11. Fig. Gr (ω). For the investigated system. In Fig. the transfer function from the back EMF. corresponding to a rise time of 1 ms at a base frequency of 50 Hz. 4. the minimum bandwidth of the current control loop when using “active resistance” becomes (4. RR +Rs of 7 p. iR .” the damping of low-frequency disturbances has been signiﬁcantly improved.74) if all model parameters are assumed to be accurate. In order to investigate the performance of the “active resistance” with regards to damping of disturbances.u. is reasonable [40]. 4. we study the ratio between the moduli of the frequency function corresponding to (4. 2 (αc + ω 2 )L2 σ (4. since.u. to the current.76) This means that the above choice of Ra will force a change in the back EMF to be damped with the same bandwidth as the closed-loop current dynamics.77) αc.08 p.min = (RR + Rs )/Lσ . αc. is depicted for a current control loop bandwidth cally.” ⎧ ⎪ ⎨ (4.79) which shows that while the “active resistance” has little impact on high-frequency disturbances.Selection of the “Active Resistance” ˆ ˆ ˆ If the “active resistance” is set to Ra = kR (αc Lσ − RR − Rs ).

81) ˆ Instead of using the actual ﬂux in (4. 4.4. i. as iref Rq Teref =− . 4.. Rq can be determined from the reference torque.42) as Te = 3np Im Ψs i∗ ≈ −3np ψs iRq . Since it is difﬁcult to measure the torque.u.81). iRq .80) For a stator-ﬂux-oriented system the above approximation is actually an equality. The bandwidth of the current control loop is set to 7 p.12 shows a block diagram of the open-loop torque control scheme. the approximation ψs ≈ Eg. should be set much faster than the speed dynamics. with the bandwidth αc . Since the stator ﬂux. is almost ﬁxed to the stator voltage. ψs . of the back EMF.] Fig.4.e. 4.12: Block diagram of the open-loop torque control. the q component reference current.82) . Teref . it is most often controlled in an open-loop manner. Therefore. Teref − 1 ˆ 3np ψs iref Rq Fig.u. 4.2 Torque Control The electromechanical torque can be found from (4. R (4.0 Gr (ω) [dB] −50 −100 0 5 10 15 20 25 ω [p. ˆ 3np ψs (4. iref .3 Speed Control Since the current dynamics. 4. Ratio of the damping improvement when using “active resistance” as a function of the frequency. ω. the torque can be controlled by the q component of the rotor current. The mechanical dynamics are described by J dωr = Te − Ts np dt 50 (4.nom /ω1 can be used.11. the speed can be controlled in cascade with the current. Fig.

Fig. p ωr (p) = ˆ Ts J 2 p + (Ba + kp )p + ki np 51 (4.83) where an “active damping” term. Choosing the “Active Damping” The transfer function from the shaft torque. that can be used to improve the damping of disturbances. as described in Section 4. to the rotational speed can be described by. 4.13. 4.where Te is the electromechanical torque and Ts is the shaft torque. treating the shaft torque. ref (p) J Te p + Ba np (4. Ts . as mentioned earlier.87) . then becomes G(p) = 1 ωr (p) = . is introduced.84) (4.13 shows a block diagram of the speed control system. How to determine the “active damping” will be shown in the next section. Speed control loop. as Te = Teref where the reference torque is set to Teref = Teref − Ba ωr (4. Ts ref ωr + T ref F (p) e + − − Teref + − np Jp ωr Ba Fig. the following PI controller can be found: F (p) = ˆ ki Jαs Ba αs αs −1 + G (p) = kp + = p p np p (4. under the assumption that the current dynamics are much faster than the speed dynamics. an inner feedback loop [41].1. Ts . This is. Ba . The electromechanical torque can be expressed.85) Using IMC.13. The transfer function from Teref to ωr . 4.4.86) where αs is the desired closed-loop bandwidth of the speed-control loop and the notation “ˆ” indicates an estimated quantity. as a disturbance. see Fig.

. b) Rotor current (q component).87).u.14 shows a simulation of the speed control loop with rated driving torque.ˆ if the active damping term is chosen as Ba = αs J/np .25 p.u..] a) 1. is set to 0. After 4 s it is changed back to 1. This causes the rise time of the rotor speed to be longer than the ideal.] b) Time [s] Fig.4 p.u.u. 4.014 p. after 1 s it is changed to 0.4 p.u. since the maximum rotor current has been reached. it can be seen in the simulation that the speed reference step at 1 s forces limitation of the rotor current. Moreover. becomes ωr (p) p = J Ts (p + αs )2 np (4. Rotor Speed [p.14.89) .5 0 0 5 10 15 Rotor Current [p. after 13 s the driving torque is changed to 50% of its rated value.88) all parameters are assumed to be ideal.5 1 0. and the bandwidth of the speed control loop.4 Reactive Power Control The instantaneous apparent power at the stator terminals.u. (4.75 p.u.u. Simulation of the speed-control loop. 4.. a) Rotor speed.u. Initially the speed reference is set to 1. Finally. and after 7 s the reference is ramped down during 3 s to 1 p. Evaluation Fig.5 1 0.5 0 5 10 15 1. corresponds to 0.25 p. The bandwidth of the current control loop is set to 1. αs .4. can now be found as Ss = 3vs i∗ = 3 Rs is + s 52 dΨs + jω1 Ψs i∗ s dt (4. A bandwidth of 1.5 s. corresponds to a rise time of 5 ms and 0. Ss = Ps + jQs .u. In (4. The simulations shows that the speed-control loop behaves as expected. is damped with the same time constant as the bandwidth of the speed-control loop.e. change in the shaft torque.88) i. 4.014 p. Ts .

nom p (4. Closed-Loop Reactive Power Control Since the ﬂux for a DFIG system can be considered as constant. the integration of the controller will compensate for the magnetizing current. since.96) Of course.93) it can be seen that if iref = Rd the DFIG is operated at unity power factor.93) (4. the expression in (4. However. ψs ≈ Eg. i. This means that IMC yields in an I controller.e. the controller reduces to FQ = − or as iref = − Rd αQ 3Eg. For a stator-ﬂux-oriented system.91) For a grid-ﬂux-oriented system. LM ψs ψ2 iRd + 2s LM LM − 3ω1 ψs iRq (4. in the above control law. i. From (4.Then the active and reactive power. GQiRd . ψsd = ψs and ψsq = 0.94) where αQ is the bandwidth of the reactive power control loop. the above is reduced to Ps = 3Rs |is |2 + 3ω1 ψs isq = 3Rs |iR |2 − 2 Qs = 3ω1 ψs isd = 3ω1 ψs ψs − iRd . as FQ = αQ −1 αQ 1 GQiRd = − ˆ p 3ω1 ψs p (4. neglecting the derivative of the stator ﬂux.97) αQ 1 3Eg. Therefore. since ψs /LM is close to constant. can thus be written as Ps = 3Rs |is |2 + 3ω1 (ψsd isq − ψsq isd ) Qs = 3ω1 (ψsd isd + ψsq isq ) . still holds approximately since Rs can be considered as small.. in the steady ˆ state.nom /ω1 .nom Qref − Qs dt. it would be possible to add a feed-forward term in order to compensate for the magnetizing current. where the voltage is aligned with the q axis. 53 .92) (4. Moreover. s (4. ψs /LM .93). there will be a static relationship between the reactive power and the d component of the rotor current.e.95) ˆ ψs ˆ LM (4..92) and (4.90) (4. feed-forward compensation has not been considered for the reactive power control loop.

where the voltage drop across the stator resistance can not be neglected.98) ˆ ˆs and the estimate of the transformation angle. Taking the real part of the above equation and neglecting the ﬂux dynamics yield ˜ vsd = Rs isd − ω1 ψs sin(θ1 ).100) ˜ ˆ where θ1 = θ1 − θ1 is the angular estimation error.1. 56] ˆs Ψs = s ˆ s (vs − Rs is )dt (4. it is possible to form an error signal suitable for the PLL-type estimator. s dt dt (4. and their corresponding angles. or the voltage drop across the stator resisˆ tance is negligible.3. the above equation can be rewritten in synchronous coordis nates as vs = Rs is + dψs j θ1 ˜ ˜ e + jω1 ψs ej θ1 dt (4. For a stator-ﬂux-oriented control of the doubly-fed induction generator. θ1 .98) is an open-loop integration. must be estimated. so ω1 ψs ≈ vs . as ˆ ˆ vsd − Rs isd vsd − Rs isd ˆ ˜ ε = sin(θ1 − θ1 ) = sin(θ1 ) = − ≈− ω1 ψs vs (4. yields s s vs = Rs is + s ˆ ˆ dΨs dψs jθ1 s = Rs is + e + jω1 ψs ejθ1 . can then be found from θ1 = ∠Ψs .3.1. Since the estimator in (4. The purpose of this section is to give an overview of some different estimation techniques that are available in literature. described in Section 4.4.99) s If vs = vs ej θ1 and is = is ej θ1 . 54 . it has to be modiﬁed in order to gain stability. It is also possible to estimate the transformation angle in synchronous coordinates. This means that the stator frequency. and the slip frequency. Starting with the stator voltage equation in stator coordinates and taking into account that for a stator-ﬂux-oriented system.5 Sensorless Operation “Sensorless” operation implies in this thesis that neither the rotor position nor the rotor speed is measured. it might be unnecessary to estimate ω1 .101) Now. This could be done by replacing the open-loop integration with a low-pass ﬁlter [39]. it is marginally stable. Estimation of Synchronous Frequency Angle For a system which is oriented with the grid ﬂux.36) as [46. the angle θ1 can easily be found by using a PLL. ω2 . the stator ﬂux can be estimated in stator coordinates using (4. θ1 and θ2 .4. Ψs = ψs ejθ1 . on the measured grid (stator) voltage. ω1 .102) where the approximation is due to the fact that the stator is directly connected to the grid. Thus. The notation “ ˆ ” is used for estimated variables and parameters. (4. see Section 4. Note that if no stator variables exist in the control law.

e. In [56] a stator-ﬂux-oriented sensorless control using the rotor voltage circuit equation is proposed. and an outer slower control loop that controls the dc-link voltage. i. This means that the q component of the grid-ﬁlter current will control the active power delivered from the converter and the d component of the ﬁlter current will. The reference frame of the inner current control loop will be aligned with the grid ﬂux. in synchronous coordinates. The control of the grid-side converter consists of a fast inner current control loop. which. The rotor voltage equation is given by dΨR vR = RR iR + (4.103) and since the stator ﬂux is known. θ2 . the estimate of the slip angle. dt Neglecting the derivative of the ﬂux. 68] it has been done in synchronous coordinates. This implies that the outer dc-link voltage control loop has to act on the q component of the grid-ﬁlter current. 55 . is given by Ψs = ψs = LM (is + iR ) (4. can be estimated from the imaginary part of the above equation as ˆ ˆ vRq − RR iRq vRq − RR iRq = . Starting with the stator ﬂux. In the ﬁrst method.Estimation of Slip-Frequency Angle In the literature there are at least two methods to perform sensorless operation. it is to a great extent determined by the stator voltage. as ˆ ˆ θ2 = ω2 dt. can be found from integration of the estimate of the slip frequency. ω2 .108) 4. (4. ˆ (4. R The second method is based on determining the slip frequency by the rotor circuit equation.5 Control of Grid-Side Converter The main objective of the grid-side converter is to control the dc-link voltage.107) ω2 = ˆ ˆ ψRd ψs + Lσ isd ˆ Then. a set of variables is estimated or measured in one reference frame and then the variables are used in another reference frame to estimate the slip angle θ2 . which controls the current through the grid ﬁlter. The magnitude of the stator ﬂux can ˆ be estimated as ψs = vs /ω1 [46].105) θ2 = ∠ir − ∠ˆR . accordingly. the slip frequency. ω2 . if the rotor current is measured in rotor coordinates the estimate of the slip angle can be found as ˆ i (4. it is possible to use the above-mentioned equation to estimate the rotor current as follows: ˆ ˆR = ψs − is (4. control the reactive power.. see previous section for determination of this angle.104) i ˆ LM where the stator current has been measured and transformed with the transformation angle θ1 . In [15] the estimation of the rotor currents has been carried out in stator coordinates. The method will here be described in synchronous coordinates. Estimating the rotor currents from the ﬂux and the stator currents can do this. while in [46. Then.106) + jω2 ΨR .

to the grid-ﬁlter current with ideal parameters then becomes p . (4.. Moreover. is chosen as ˆ vf = vf − (Raf − jω1 Lf )if . by using IMC a PI controller can be determined with the bandwidth αf . i.110) This means that the inner closed-loop transfer function. the applied grid-ﬁlter voltage. by assuming the current control loop to be fast. dt (4.1. reduced to the following linear system dW 1 Cdc = −Pf − Pr 2 dt (4. assuming ideal parameters.5. By ˆ ˆ choosing the active damping according to Section 4.e. and adding an “active damping” term as fq iref = ifref + Ga W fq q 56 (4.113) ˆ ˆ Raf = αf Lf − Rf . (4. the grid-ﬁlter current control law can now be written as vf = where ˆ kpf = αf Lf 2ˆ ˆ kif = αf (Rf + Raf ) = αf Lf kpf + kif p ˆ (iref − if ) − (Raf − jω1 Lf )if f (4. (4.4. The dc-link dynamics (4.48) are.109) In order to introduce “active damping” and decouple the d and the q components of the grid-ﬁlter current.e. the power delivered to the grid ﬁlter can be approximated as Pf ≈ 3Egq if q . the nonlinear dynamics of the dc link are transformed into an equivalent linear system where linear control 2 techniques can be applied [95].. if q = iref . If the power losses of the grid ﬁlter are small and the current control of the grid ﬁlter is aligned with the grid ﬂux. One way of simplifying the control of the dc-link voltage is by utilizing feedback linearization.5. Pf is the power delivered to the grid ﬁlter and Pr is the power delivered to the rotor circuit of the DFIG.114) 4.44) the dynamics of the grid ﬁlter are described: Lf dif = vf − (Rf + jω1 Lf ) if − Eg .5. vf . i.111) and. This can be done by letting W = vdc [50. hence. the transfer function from grid voltage (“back emf”). as mentioned earlier.2 DC-Link Voltage Control The dc-link voltage control in this thesis is essentially following [76]. i. becomes G(p) = if (p) 1 = vf (p) Lf p + Rf + Raf (4. Eg . Raf = αf Lf − Rf .116) . 76..112) GEg if (p) = Lf (p + αf )2 Finally.115) where. thus.1 Current Control of Grid Filter In (4.e. 79].

ξ = 1.e. i.” it is possible to write the dc-link dynamics as 1 dW Cdc = −3Eg ifref − 3Eg Ga W − Pr . 57 . a disturbance. will be damped as GP W (p) = Cdc (p2 −2p 2 + 2αw ξp + αw ξ) (4. Eg .e.15. the following PI controller is obtained F (p) = ˆ αw −1 αw Cdc αw Ga − G (p) = − p 6Eg.120) ˆ where ξ = Egq /Eg.118) (4.. by utilizing IMC.nom . q 2 dt The inner closed-loop transfer function becomes G (p) = −6Eg W (p) = . If the active damping is chosen as ˆ Ga = αw Cdc /(6Eg. Pr . and αw is the bandwidth of the dc-link voltage control loop.. A block diagram of the dc-link voltage controller is depicted in Fig. ref pCdc + 6Eg Ga if q (p) (4.117) Then.119) where the magnitude of the grid voltage. i. Pr W ref + − i F (p) f q + + Ga if q 3Egq − − G (p) 2 pCdc 2 W = vdc Fig.15. 4. DC-link voltage control loop. is put to its nominal value. With Egq = Eg.nom . Eg.nom ).nom p (4.nom and Cdc = Cdc .where Ga is the gain of the “active damping. 4. GP W (p) is reduced to GP W (p) = −2αw p Cdc (p + αw )2 (4.121) which means that a disturbance is damped with the same bandwidth as the dc-link voltage control loop.

58

**Chapter 5 Evaluation of the Current Control of Doubly-Fed Induction Generators
**

In this chapter the current control law derived for the DFIG in the previous chapter is analyzed with respect to eliminating the inﬂuence of the back EMF, which is dependent on the stator voltage, rotor speed, and stator ﬂux, in the rotor current. Further, stability analysis of the system is performed for different combinations of these terms in both a stator-ﬂux and grid-ﬂux-oriented reference frame, for both correctly known and erroneously parameters.

5.1 Stability Analysis

In order to investigate the inﬂuence of the feed-forward compensation of the back EMF and the inﬂuence of the “active resistance” on the stability of the system, an analysis is performed in this section. The analysis will be performed both for a stator-ﬂux-oriented system and for a grid-ﬂux-oriented system. In this section a full-order analysis of the system is performed, since one of the objectives is to study the impact of the current control law derived in the previous chapter.

**5.1.1 Stator-Flux-Oriented System
**

Consider the system described by (4.66)–(4.68). Splitting (4.66) into real and imaginary parts, assuming stator-ﬂux orientation, i.e., Ψs = ψs ejθ1 , the stator voltage equals the grid s voltage, i.e., vs = jEg ej(θg −θ1 ) . Making the variable substitution Δθ = θg − θ1 , the system model can be rearranged as dI dt dψs dt dΔθ dt diR dt =e =− Rs ψs − Eg sin(Δθ) + Rs iRd LM dθg dθ1 Eg cos(Δθ) + Rs iRq = − = ωg − dt dt ψs ˆ kp e + ki I + (j ω2 Lσ − Ra )iR + kE E (RR + Rs + jω2 Lσ )iR + E ˆ ˆ − = Lσ Lσ 59 (5.1) (5.2) (5.3) (5.4)

where E = Eg − Rs Rs + jωr Ψs = jEg ejΔθ − ( + jωr )ψs . LM LM (5.5)

In (5.1)–(5.4), the term I is the integration variable of the control error and e = iref −iR is the R control error. Note that (5.1) and (5.4) are complex-valued equations while (5.2) and (5.3) are real-valued equations. In the following analysis, the rotational speed ωr will be assumed to be varying slowly, and is, therefore, treated as a parameter. Throughout this section, the machine model parameters will be assumed to be ideal and known. If, for example, the rotor current is controlled by a high-gain feedback, it is possible to force the system to have both slow and fast time scales, i.e., the system behaves like a singularly perturbed system [57]. This means, that if the bandwidth of the current control loop is high enough, it is sufﬁcient to study the system described by (5.2) and (5.3) in order to analyze the dynamic behavior of the DFIG. A stability analysis, assuming fast current dynamics, can be found in [13, 43]. Later on, analysis not neglecting the current dynamics will be compared to analysis neglecting the current dynamics; therefore a short summary will be presented. By linearization of the nonlinear system described by (5.2) and (5.3), the characteristic polynomial can be found. A ﬁrst-order Taylor series expansion of the characteristic polynomial around Rs = 0 (as Rs is small, typically less than 0.1 p.u.) yields p2 + Rs iref 2 ωg LM iref Rs Rq Rd 2− p+ 1− ωg LM Eg Eg (5.6)

where iref is the active current reference and iref is the magnetization current supplied from Rq Rd the rotor converter. Since Rs is small (< 0.1 p.u.), iref will only have a minor inﬂuence on Rq the dynamics. However, iref will inﬂuence the dynamic performance. It is required that Rd iref < Rd 2Eg ωg LM (5.7)

in order to maintain stability. A similar constraint can be found in [13, 43]. In order to operate the DFIG with unity power factor, one should select [99] iref = Rd ψs Eg ≈ LM ωg LM (5.8)

which value is half of the value in the condition in (5.7). For the case when it is not possible to separate the time scales by a high-gain feedback in the current control loop, a full-order analysis should be performed. By linearizing the nonlinear system described by (5.1)–(5.4) in a similar manner as previously, the characteristic polynomial for the complete system can be found. A ﬁrst-order Taylor series expansion of the characteristic polynomial around Rs = 0 yields (p + αc ) p + kR αc + (1 − kR ) RR (p4 + a3 p3 + a2 p2 + a1 p + a0 ). Lσ 60 (5.9)

The IM is running as a generator at half of the rated torque.1. it is shown how one of the complex-conjugated poles given by the fourth-degree factor move with increasing bandwidth of the current control loop.9). I NVESTIGATED C URRENT C ONTROL M ETHODS . This means that.1.where expressions for the coefﬁcients a3 to a0 become 2Rs Rs ωg iref RR + 2Rs Rd − + (1 − kR ) LM Eg Lσ ref 2 iRq Rs ωg 2 iref ωg 2 2 a2 = αc kR + ωg − + (1 + kR )αc Rs − Rd Eg LM Eg ref 1 − kR i ωg 1 − kR 2 + αc RR + 2Rs + RR Rs − Rd Lσ Lσ LM Eg 2 2 ref 2αc kR Rs 2αc RR Rs αc kR Rs iRd ωg αc iref RR Rs ωg + (1 − kR ) − − (1 − kR ) Rd a1 = LM LM Lσ Eg Lσ Eg ref iRq Rs Rs ωr ωg 2 − (1 − kE ) + (1 + kR )αc ωg 1 − Lσ Eg ref iRq Rs Rs 2 RR 2 1− + (1 − 2kR + kE )ωg + (1 − kR )ωg Lσ Eg Lσ 2 2 αc iref Rs ωg kR iref RR Rs Rq Rq 2 2 2 1 − kR RR + 2Rs − . the characteristic polynomial (5. Method II is unstable for bandwidths of the current control loop between 1.10) (5.1. −RR /LM for Method I and two at −αc for Method II) and four poles given by the fourth-degree factor. In Fig. according to Table 5.11) (5.13) The parameters kE and kR affect the roots of (5. It can be seen in the ﬁgure that the poorly damped poles of Method II move with increasing bandwidth of the current control loop from stable to unstable and back to be stable again. while for Method I the poles are stable. 5.9) is investigated for the four different options. TABLE 5. directly and via a0 –a3 . The other complex-conjugated poles given by the fourth-degree factor are well damped and are therefore not shown in the ﬁgure. Method I Method II Method III Method IV kE 0 0 1 1 kR 0 1 0 1 Methods I and II Both methods give two real-valued poles (at −αc .0–5. and is magnetized from the rotor circuit. + αc ωg a0 = αc ωg kR − Eg Lσ Eg a3 = αc (1 + kR ) + (5. 5.u. for the above mentioned operating point. The four different combinations of kE and kR available are.12) (5. when approximating fast current dynamics (marked with “x” in the ﬁgure). Below. synchronous speed. Moreover.6 p. termed Methods I–IV.1. αc . the real part of the poorly damped pole is very small. even a small error (due to the approximation of a fast current dynamics) may play a signiﬁcant 61 . as shown in Fig.

1 0.995 0.99

Im

0.985 0.98 0.975 0.97 −0.05

−0.025

0

0.025

0.05

Re

Fig. 5.1. Root loci of one of the poorly damped poles of the doubly-fed induction generator using current control methods without feed forward of the back EMF without “active resistance” (Method I, solid) and with “active resistance” (Method II, dashed). The arrow shows how the poles move with increasing bandwidth (0.5–15 p.u.) of the current control loop. The symbol “x” indicates the pole location when the current dynamics are neglected.

role for the result of the stability analysis. Hence, it is necessary to make a careful stability analysis, at least when using Methods I or II. A similar approach, as will be performed in the next section, with Routh’s table produces very large expressions of which it is difﬁcult to determine any constrains for stability. Therefore, the approach with Routh’s table is not carried out for Methods I and II. Method III When using Method III, i.e., feed forward of the back EMF, the characteristic polynomial in (5.9) is reduced to (p + αc )2 p + RR (p3 + b2 p2 + b1 p + b0 ). Lσ (5.14)

The system has at least three real-valued poles, two located at −αc and one at −RR /Lσ . The coefﬁcients in the third-degree factor become b2 = 2Rs RR + 2Rs iref Rs ωg + − Rd LM Lσ Eg 2 (Eg − iref Rs )ωg RR Rs (2Eg − iref LM ωg ) Rq Rd b1 = + Eg LM Lσ Eg b0 =

2 − iref RR Rs + (RR + 2Rs )Eg ωg Rq

(5.15) (5.16)

Lσ Eg

.

(5.17)

As can be seen, the coefﬁcients are not dependent on αc for Method III. 62

TABLE 5.2. ROUTH ’ S TABLE .

p3 p2 p1 p0

1 b2 b2 b1 − b0 B= b2 Bb0 − 0 = b0 B

b1 b0 0 0

To investigate the stability of the system, Routh’s table can be used [14], see Table 5.2. In order for the system to be stable, the coefﬁcients in the ﬁrst column must not change sign. Since the ﬁrst coefﬁcient in Routh’s table is 1, all other coefﬁcients must be positive in order to maintain stability. The expression for the coefﬁcient B becomes quite complex; an approximation is 2 2 Rs 2Eg − ωg LM iref (RR + ωg L2 ) σ Rd (5.18) B≈ LM Lσ RR Eg where a ﬁrst-order Taylor series expansion of the coefﬁcient B with respect to the stator resistance, Rs (around Rs = 0), has been carried out. The following constraint can be set on iref in order to keep the coefﬁcient b2 positive: Rd iref < Rd Eg 2Rs RR + 2Rs . + Rs ωg LM Lσ 2Eg . ωg LM (5.19)

For keeping the coefﬁcient B positive, the following constraint has to be set iref < Rd (5.20)

Since the term −iref RR Rs in b0 is at least one order of magnitude lower than the term (RR + Rq 2Rs )Eg , b0 can be considered to be positive. The constraint in (5.20) is “harder” than the constraint in (5.19). The constraint in (5.20) is identical to the constraint in (5.7) where the stability analysis was performed assuming fast current dynamics. The system has two poorly damped poles, caused by the ﬂux dynamics, and the constraint on iref relates to the Rd ﬂux dynamics. Therefore, the constraint on iref , which relates to the ﬂux dynamics, can be Rd found more easily assuming fast current dynamics. Generally, a full-order analysis is still valuable, if the current dynamics are not fast, since other parameters also may inﬂuence the stability (for stability analysis assuming fast current dynamics). Method IV For Method IV, i.e., with feed forward of the back EMF and “active resistance,” the characteristic polynomial in (5.9) is reduced to (p + αc )4 p2 + Rs iref 2 ωg LM iref Rs Rq Rd 2− p+ 1− ωg . LM Eg Eg (5.21)

The characteristic polynomial has four real roots located at −αc . The second-degree factor is identical to (5.6), where the current dynamics were neglected. Therefore, for Method IV, the same analysis as for the case with the assumption of fast current dynamics can be used. 63

**5.1.2 Grid-Flux-Oriented System
**

The corresponding dynamics for the grid-ﬂux-oriented system become dI =e dt dΨs = Eg − dt (5.22) (5.23) (5.24)

Rs + jωg Ψs + Rs iR LM ˆ diR kp e + ki I + (j ω2 Lσ − Ra )iR + kE E (RR + Rs + jω2 Lσ )iR + E ˆ ˆ = − . dt Lσ Lσ

Note that (5.22)–(5.24) are complex-valued equations. As for the case with stator-ﬂux oriented analysis, the rotational speed ωr will be assumed to be varying slowly and is therefore treated as a parameter. Throughout this section, parameters will exactly as in the previous section be assumed to be ideal and known. If, as for the stator-ﬂux orientation, the rotor current is controlled by a high-gain feedback, it is sufﬁcient to study the dynamics described by (5.23), which have the following equilibrium points: ψsd0 ψqd0

2 LM iref Rs + LM iref Rs + Eg ωg Eg + Rs iref Rq Rq Rd = ≈ 2 + L2 ω 2 Rs ωg M g

(5.25) (5.26)

LM Rs iref Rs + Eg − iref LM ωg Rs (Eg − iref LM ωg ) Rq Rd Rd = ≈ 2 + L2 ω 2 2 Rs LM ωg M g

where the approximation is due to a ﬁrst-order Taylor series expansion of Rs around Rs = 0. Then, the following characteristic polynomial can be determined: p2 + 2 Rs R2 2 p + ωg + 2s . LM LM (5.27)

In (5.27) it can be seen that the DFIG is poorly damped, and that the damping is only dependent of Rs and LM . Moreover if the PLL-type estimator, described in Section 4.1.3 is used to track the grid voltage, the dynamics of the PLL will be separated from the ﬂux dynamics in (5.27). If the rotor currents cannot be neglected, a full-order analysis has to be performed. As in the previous section, the dynamic systems described by (5.22)–(5.24) consists of two parameters kE and kR that could be either set to zero or unity. This yields, in the same way as for the stator-ﬂux-oriented system, four different options, Method I to Method IV, for the current control law, see Table 5.1. Methods I and II Linearizing of the non-linear system described by (5.22)–(5.24), its characteristic polynomial can be found. A ﬁrst-order Taylor series expansion of the characteristic polynomial with respect to the stator resistance, Rs (around Rs = 0) yields (p + αc ) p + kR αc + (1 − kR ) RR (p4 + a3 p3 + a2 p2 + a1 p + a0 ). Lσ 64 (5.28)

97 −0.1. The second-complex conjugated poles are well damped and are therefore not shown in the ﬁgure.99 Im 0. 5.32) In Fig. 5. for Method II.05 Re Fig. might not be high enough in order to be able to make the assumption of a fast current dynamics (marked with “x” in the ﬁgure). 65 . as given by the fourthorder characteristic polynomial.2.985 0.29) (5.05 −0.31) (5. for Method I the poorly damped pole is stable and for Method II the poorly damped pole moves with increasing bandwidth of the current control loop from stable to unstable and back to be stable again. 5. the system is unstable for bandwidths between 1.u.2 it is shown how one of the complex-conjugated poles.u. since the root loci are plotted with numerical values the result are only valid for the given operation conditions and for the investigated machine. αc .u.995 0.5–15 p. The arrow shows how the poles move with increasing bandwidth (0.u.30) (5.where the coefﬁcients a3 to a0 become a3 = 2 Rs RR + 2Rs + αc (1 + kR ) − (kR − 1) LM Lσ 2(1 + kR )αc Rs (kR − 1) (2RR Rs + αc LM (RR + 2Rs )) 2 2 a2 = αc kR + − + ωg Lσ LM Lσ 2 αc kR Rs 2 a1 = 2 + αc (1 + kR )ωg LM 2 2αc RR Rs + (RR + 2Rs )LM ωg Rs ωg ωr − (kR − 1) −2 LM Lσ Lσ 2 (kR − 1)(RR + 2Rs )αc ωg 2 2 a0 = αc kR ωg − .) of the current control loop. For the in the ﬁgure investigated case. Root loci of one of the poorly damped poles of the doubly-fed induction generator using current control methods without feed forward of the back EMF without “active resistance” (Method I. It can be seen in the ﬁgure that 1 0. As could also be seen for the stator-ﬂux oriented case.025 0 0. The symbol “x” indicates the pole location when the current dynamics are neglected. Lσ (5.8 p. a current control loop bandwidth of approximately 15 p.025 0. move with increasing bandwidth of the current control loop.1 p. dashed). The operating condition is as in Fig. solid) and with “active resistance” (Method II. Of course.975 0.98 0. and 3.

at least for Method I.1. Further. two located at −αc and two at −(RR + Rs )/Lσ .2. feed forward of the back EMF.3 Conclusion It has been shown that by using grid-ﬂux orientation the stability and the damping of the system is independent of the rotor current. LM LM (5.e..4. LM LM (5. the same analysis as for the case with the assumption of fast current dynamics can be used. it is possible to magnetize the DFIG entirely from the rotor circuit without reducing the damping of the system. This implies that for a grid-ﬂux-oriented system. for higher bandwidths of the current control loop.e.. Method IV with both feed-forward compensation and “active resistance” can be assumed to be the best one of the investigated methods.24) as (p + αc ) 2 RR + R s p+ Lσ 2 p2 + 2 Rs R2 2 p + 2s + ωg .22)–(5. The system has at least four real-valued poles. as shown in Section 4. given by (4. By utilizing the feed-forward compensation. given by (4.33) Note that the above characteristic polynomial has not been expanded by a Taylor series. stability of the derived current control law is independent of the bandwidth of the current control loop and the order of the system to analyze is reduced. 5. The second-degree factor is identical to the characteristic polynomial in (5.” the characteristic polynomial becomes (p + αc )4 p2 + 2 Rs R2 2 p + 2s + ωg . Therefore. Therefore. Method III When using Method III. for Method IV.. it is possible to produce as much reactive power as possible and still have a stable system with the same damping from a stability point of view. 5. i. assumed to be much faster than the ﬂux dynamics.73) and the transfer function from a disturbance to the rotor current. the inclusion of the “active resistance” improves signiﬁcantly the damping of low-frequency disturbances. for the grid-ﬂux-oriented system.e. Gcl (p).34) Note that the above characteristic polynomial has not been expanded by a Taylor series. see Fig.1. 5. The characteristic polynomial has four real roots located at −αc .27) where the current dynamics were neglected. i. with feed forward of the back EMF and “active resistance. in contrast to stator-ﬂux orientation. Method IV For Method IV. i. GEi (p).2 Inﬂuence of Erroneous Parameters on Stability We now study how the closed-loop current-control transfer function. the characteristic polynomial can be found from the system (5.74) are 66 . since the error is in the same order of magnitude as the real part of the pole. Moreover.

However.2. In order to study the behavior for larger Lσ . the back EMF will not be totally compensated for due to non-ideal parameters. 5. the error in a parameter is denoted with the symbol ˜. 5.1. From (5. see (4. Since LM is only included in the feed-forward compensation.37) . This means that when using “active resistance. For Methods I and III. 5. For ideal parameters the rotor current is determined by iR = Gcl (p)iref + GEi (p)E. Hence.35) The methods where the back EMF is compensated for using feed forward (Methods III and IV). Rs and RR Since errors in Rs and RR inﬂuence the performance in the same way. the rotor speed is set to 1.35) if 2αc Lσ dependent on errors in Rs and RR . we will study the sum ˜ ˜ ˜ of the errors in the resistances: R = Rs + RR . ˆ it is preferable to overestimate Lσ . ˜ In the analysis below. and RR . The parameters to be studied in the following are Lσ . Rs . 5. The operating condition corresponds to that for Method I with three different values of L of Fig.1 Leakage Inductance.3 p. This is shown in Fig.3 that the inﬂuence ˆ of errors in Lσ is small for the investigated 2-MW DFIG. it has no impact in the following analysis. It can be seen in Fig. 5.u.72). 5. the difference is larger. so that the effect of the cross coupling between the d and the q components is included. the rotor current.4. This means that the conditions for impact of parameter variations also hold for the methods with feed forward of the back EMF. e.” the system is not (5. even though the effect might be less severe.35) is independent of the ﬁeld orientation.inﬂuenced by non-ideal parameters. A similar analytical expression for larger errors in Lσ is ˜ difﬁcult to derive. the bandwidth of the current control loop is ˆ increased if Lσ is overestimated. Note that (5. it is not included. is given by iR ≈ αc αc ˜ ˜ 1 + j Lσ ω2 GEi (p) iref + 1 − j Lσ ω2 R p + αc p + αc GEi (p)E (5. and. root loci are shown in Fig. the rotor current is found to be iR ≈ ˜ αc (Lσ p + Rs + RR − R) iref 2 + α L p + (R + R − R)α R ˜ c Lσ p c σ s R p − E ˜ Lσ p2 + αc Lσ p + (Rs + RR − R)αc 67 (5. Clearly.35) for ideal parameters. given by (5. for smaller DFIGs such as the 22-kW laboratory DFIG. One reason for this is that the proportional part of the controller will be increased. For Methods II and IV where “active resistance” is used. Lσ = ˆ Lσ − Lσ .2. Lσ ˆ For errors in Lσ . hence.36) it can be seen that small values of Lσ do not signiﬁcantly ˜ L and Lσ ˆ inﬂuence the dynamic performance.3 ˆ σ .g. the rotor current is given by ˜ R.2 Stator and Rotor Resistances. R (5. however.36) ˜ ˜ where the approximation is due to a ﬁrst-order Taylor series expansion of Lσ around Lσ = 0 ˜ σ .

5.1 −0.5Lσ . dashed Lσ = −0.96 0.u. 5.9 Im 0.” Fig. b) Grid-ﬂux orientation. However.9 0. 68 .025 0 0.37) it can be seen that if the resistances ˜ are overestimated. as for the case with errors in Lσ . ˜ σ = −0.99 0.5–15 p. i.95 −0.e.5–15 p.5Lσ .98 Im 0.97 0.05 −0.05 0. a) 1 b)1 0.95 −0.15 −0.5 shows the root loci for Method I of the investigated 2-MW DFIG. the difference is larger for smaller DFIGs.) of the current control loop. Solid is Lσ = 0.97 0.85 −0. the approximation assuming αc Lσ Rs + RR .98 b)1 0. a) Statorﬂux orientation. the damping of the current dynamics are actually improved.5Lσ . The arrow shows how the poles move with ˜ dashed L increasing bandwidth (0. This is shown in Fig.85 −0.u. b) Grid-ﬂux orientation.05 Im 0. a) Stator-ﬂux orientation.4. R < 0.99 0. 5.025 Re Re Fig.) of the current control loop. the same phenomenon as using “active resistance. and dotted Lσ = 0.96 0.05 Re Re Fig. such as the 22-kW laboratory DFIG.15 0 0.95 Im 0. Root loci of one of the poorly damped poles of the laboratory 22 kW doubly-fed induction ˜ generator using Method I for three different errors in the leakage inductance parameter Lσ .1 −0. ˜ σ = 0. i.95 0. The arrow shows how the ˜ ˜ Solid is L poles move with increasing bandwidth (0.025 0 0.4. In (5.025 −0.05 −0. 5. Root loci of one of the poorly damped poles of the doubly-fed induction generator using ˜ ˜ Method I for three different errors in the leakage inductance parameter Lσ .a) 1 0.e. In the ﬁgure it can be seen that the inﬂuence of ˆ errors in the resistance is small...3.05 0 0.5Lσ . and dotted Lσ = 0.

97 −0.995 0.95 Im 0.u. the system can become a) 1 b)1 0.025 0 0. 5.) of the current control loop.5–15 p.) of the current control loop. Root loci of one of the poorly damped poles of the doubly-fed induction generator using ˜ Method I for three different errors in the stator and rotor resistances parameters R.1 −0.995 0. Solid ˜ = 0.85 −0. 5. The arrow shows how the poles ˜ ˜ is R move with increasing bandwidth (0. a) Stator-ﬂux orientation.975 0.5R. the poorly damped poles (corresponding to the ﬂux dynamics) could be unstable for certain operating conditions.9 Im 0.99 b)1 0.985 0. as shown previously when only using “active resistance” (Method II).05 Re Re Fig.5R. 5. and dotted R = 0. Root loci of one of the poorly damped poles of the laboratory 22 kW doubly-fed induction generator using Method I for three different errors in the stator and rotor resistances para˜ ˜ ˜ ˜ meters R.95 0. dashed R = −0. as illustrated in Fig. dashed R = −0.05 0. The arrow shows how the poles move with increasing bandwidth (0.u. It can also be seen in the ﬁgures that the grid-ﬂux-oriented system seems.1 −0. especially for Method I and smaller machines.025 Im 0.5R.6.985 0. a) Stator-ﬂux orientation.5–15 p.025 Re Re Fig. b) Grid-ﬂux orientation. even though the difference is small.15 −0. Solid is R = 0.98 0.05 −0.a) 1 0.97 −0.05 −0. b) Grid-ﬂux orientation.05 −0.15 0 0.99 Im 0.05 0 0. unstable if the resistances are overestimated.9 0. to 69 .025 0 0. Moreover.98 0.5R. and dotted R = 0.975 0.6.5.85 −0. Therefore.

u.25 d 0. reaches 0. The DFIG is magnetized entirely from the stator. Data have been sampled with 10 kHz and low-pass ﬁltered with a cut-off frequency set to 5 kHz.25 p. Experiment of the stator-ﬂux oriented current control step responses of the q component of the rotor current. to 0.25 0. and vice versa when the rotor speed reaches 0.u.75 1 Method III 0 0. 5.be less sensitive to overestimated R = Rs + RR in comparison to the stator-ﬂux-oriented system. the system might be degraded..2 for data and parameters of the laboratory setup.25 d 0. depending on the bandwidth of the current control loop. Rd Further.5 q 0.75 1 0 0. A scrutiny investigation of Fig.7 shows that Method II gives a 50-Hz ripple.5 0.5 0 −0. 5.25 p. This has been done by letting iref change from −0.25 d 0.75 1 Method IV 0 0.5 0.u.. iref = 0. 5. a 100-Hz frequency component in the stator voltage. unstable.u.4 p. see Fig. as shown earlier. See Appendix B.5 0 −0.5 Method II 0. which inﬂuenced the performances of the current control Methods III and VI. when the rotor speed.5 q d Method I 0 −0. i. However.7.3 Experimental Evaluation The performance of the various current control methods are evaluated by reference step responses. the stator voltage of the DFIG was 230 V.5 q 0.16 p.32 p.5 q 0. 5. since the stator voltage is included in the control law. a notch ﬁlter limited the inﬂuence of the 100-Hz frequency component. i.5 0 −0.75 1 Time [s] Fig.5 0.. The reason for this is that by using only “active resistance” to damp the back EMF. and is operated under no-load conditions.u. Rq ωr .e.5 0 0. In the measurements the bandwidth of the current control was set to 1. Offsets in the stator voltage measurements caused 0. Even 70 .7.e.

] 1 0.8 shows an experimental case of a stator-ﬂux-oriented and a grid-ﬂux-oriented rotor current control. after 0. that by using grid-ﬂux orientation the stability and the damping of the system is independent of the rotor current. d) Stator-ﬂux magnitude (grid-ﬂux orientation). Experimental comparison between stator-ﬂux-oriented and grid-ﬂux-oriented systems.4 Impact of Stator Voltage Sags on the Current Control Loop Due to the poorly damped poles.2 0.5 p.u.5 q d d) 1.3 p.4 0.1 0. In the ﬁgures the d component of the rotor current is increased from 0 p.1 0.2 0.5 c) 2 Current [p.u.7 that Method IV managed best to follow its reference values in this comparison.1.u. it can be seen that the stator-ﬂux-oriented system becomes unstable with an increasing amplitude of the ﬂux oscillations.u.2 0. in contrast to stator-ﬂux orientation In Fig.1 0.32 s the rotor current is put to zero in order to put back the system into a stable operating condition. During this evaluation.u. c) Rotor current (grid-ﬂux orientation). 5. When iRd is a) 2 Current [p.3 0.3 0. After 0. 5. to 1 p.3 0. and the rotor speed.4 0.5 q d b) 1.1 Comparison Between Stator-Flux and Grid-Flux-Oriented System The aim of this section is to experimentally verify the analytical result obtained in Section 5.3. in case of a voltage sag.5 0 0 0.though the difference is fairly small. increased to 1 p.1 s.] 1 0.u.5 Flux [p. the grid-ﬂux-oriented system remains stable throughout the whole experiment.u.4 0. 5.3 0.2 0. 5.u.] 1 0 −1 0 0.1 0.] 1 0 −1 0 0. 5. was controlled by a d.8. The q component of the rotor current is set to 0. machine to be 1 p. b) Stator-ﬂux magnitude (stator-ﬂux orientation). a) Rotor current (stator-ﬂux orientation).5 Flux [p. the bandwidth of the current control loop was set to 2. As expected from the analytical results. it can be seen in Fig.c.4 0.5 Time [s] Time [s] Fig.5 0 0 0. the ﬂux will enter a damped oscillation. ωr .u.u. It is essential that the magnitude of the rotor current is below the rated value of 71 .

This implies that the worst case occurs for ωr > ωg according to (5. It can also be noted that. according to (5. be instantaneously decreased. since it is actually unstable for certain operating conditions as indicated by Fig. For a wind turbine.40) it can be seen that for ωr > ωg . Further. The DFIG is running as a generator at rated torque and is fully magnetized from the rotor circuit. the magnitude of the rotor voltage will be instantaneously increased with ΔEg . Neglecting the current dynamics. implying vR ≈ −j0.u.3 − j0.3 before the voltage sag. at the time instant tsag . 5. is close to the maximum value needed in order to achieve the desired variable-speed range for a wind turbine.1 and Method III due that the results are relatively similar to those of Method IV. the rotor voltage as given in (4.40). This implies that the rotor voltage is approximately 0. the maximum rotor voltage needed due to a symmetrical voltage sag for current control Methods I and IV can be seen. The rotor speed is 1.u. especially for low bandwidths. For higher bandwidths.39) and (5. then the value of the rotor voltage magnitude will. (5.nom to Eg at tsag . In Fig.3 p.the converter in order not to force the crowbar to go into action.3 p.u. Further. since it manages to keep the rotor current unaffected during the voltage sag. then. In Fig. Assuming that the grid voltage (or stator voltage) drops from Eg. Method I requires slightly more rotor voltage than Method IV. 5.39) If the rotor voltage is vR. especially for low bandwidths of the current control loop. Method IV is not shown in the ﬁgure. the rotor voltage will be vR (t = tsag ) = vR. the corresponding maximum rotor current needed due to the voltage sag for Method I can be seen.40).9. the rotor voltage will drop ΔvR (t = tsag ) = jEg. if ωr = 1.0 + jΔEg = −j0.40) since the stator ﬂux and the rotor speed will not change instantaneously. For example. From (5.0 − vR . generally. 5. it can be seen that 72 . and thereby lose control of the rotor currents and thus the power production. for higher bandwidths of the current control loop.7 for a grid voltage drop ΔEg =0.4 p. accordingly. It can be seen in the ﬁgure that the maximum rotor current increases with the size of the voltage sag. this operating condition is disadvantageous since a rotor voltage of 0. the maximum rotor voltage is relatively independent of the bandwidth of the current control loop for Method IV.38) From this equation it can be noted that (since vs ≈ jωg ψs ) Im[vR ] > 0 if ωr < ωg Im[vR ] < 0 if ωr > ωg Im[vR ] ≈ 0 if ωr = ωg .10. it can be seen that the increase in rotor voltage due to a voltage sag follows (5. then the change in the rotor voltage will be ΔvR = vR. immediately before the voltage sag occurs.39) and (5.u.40). then. It can be seen that the maximum rotor voltage will increase with the size of the voltage sag. Method II is not considered. If ωr < ωg . with known parameters. (5.3 p.0 before the voltage sag.3.nom ejΔθ0 − jEg ejΔθ ≈ j(Eg.67) can be expressed as vR = (Rr + Rs + jω2 Lσ )iR + vs − Rs + jωr ψs LM ≈ vs − jωr ψs = jEg ej(θg −θ1 ) − jωr ψs .nom − Eg ) = jΔEg (5.4 = −j0.

If the current control loop bandwidth is below 2 p. for Method I. thus. due to a symmetrical voltage sag as a function of the sag size. vR .] b) 1.3 ΔEg [p. The reason is that when the bandwidth is increased. independent of the voltage sag magnitude.1 Inﬂuence of Erroneous Parameters As mentioned earlier. not only necessary to design the converter according to the desired variablespeed range.u. By using Method IV. ΔEg .] 0 0 0. see (4.3 0. Maximum rotor voltage.74).1 0.u.u.u.u. 5. the methods are mostly sensitive to an underestimated Lσ . in principle. especially for low bandwidths of the current control loop. ΔEg . see Fig.2 0.4 ΔEg [p. the difference in the maximum rotor current 73 .. 5. the inﬂuence of an erroneous value of Lσ is.] 1 0. a) Method I (stator-ﬂux-oriented system).] max Fig. 5.] αc [p.] 0 0 0. the maximum rotor current is practically constant. and the current control bandwidth.1 0. due to a symmetrical voltage sag as a function R of the sag size.2 0. Simulations ˜ with Lσ = 0.5 0 10 5 1 0.4.u. 4 imax [p.] R 3 2 1 0. Maximum rotor current.5 0 10 5 αc [p. αc (stator-ﬂux-oriented system).2 0. removed. imax .a) 1. mainly since the bandwidth of the current control loop then becomes lower than the desired. the “need” for compensating the back EMF vanishes.9.4 0 10 5 αc [p.u. αc .10.u.11.u. but also according to a certain voltage sag to withstand.] Fig. and the current control bandwidth.] 0 0 0. 5. It is.1 0. b) Method IV (stator-ﬂux-oriented system).5 max vR [p.u.5Lσ shows that Method I is very sensitive to an underestimated Lσ during voltage sags.5 max vR [p.3 0.4 ΔEg [p.

As mentioned before.5 0 −0. 5.] 0 0 0. when the R ˜ leakage inductance is underestimated. Method IV manages to survive deeper sags than Method I. is below 0.2 Generation Capability During Voltage Sags As an example of this.5 Flux Damping As previously mentioned there are different methods of damping the ﬂux oscillations.2 0. ΔEg .u.u. for higher values of αc ...1 0. RR . as indicated by the ﬁgure.08 Δimax [p. Lσ = 0. while for Method I.06 0. a feedback of the derivative of ﬂux was introduced in order to improve the damping of the ﬂux.u.u. for higher bandwidths. 5.4 αc [p. for Method IV produces very similar results as a bandwidth of 1 p.] R 0. αc (stator-ﬂux-oriented system). and the current control bandwidth. A bandwidth of 7 p. This means that when the current controller needs to put out a higher rotor voltage in order to compensate for the sag. Δimax . i.u. an extra degree of freedom is introduced that can be used to actively damp out the ﬂux oscillations.11.2 0.u. as can be seen in Fig.e.u. The maximum rotor voltage is limited to 0. it will lose control of the rotor current.u.02 p.02 0 10 5 0.] 0 0 0. However.] Fig. 5. one method is to reduce the bandwidth of the current control loop [43]. For Method IV and variations in Rs . and is therefore not shown in the ﬁgure.25 p. Fig.5Lσ . the difference between the methods vanishes. From the ﬁgure it can be seen that for low bandwidths of the current control loop (allowing a lower switching frequency). 74 .04 0.u.u.3 ΔEg [p.1 0. Rs and RR have small impacts for smaller αc . the difference in maximum rotor current is insigniﬁcant.4 0. 5. and LM with ±50%.12 shows the minimum remaining grid voltage that can be handled without triggering the crowbar as a function of the power. due to a symmetrical voltage sag as a function of the sag size. and the crowbar may be triggered if the rotor current becomes too high.4.u. and the crowbar short circuits the rotor circuit when the rotor current is above 1.11.a) 1 b) 0.] αc [p. this impact is also insigniﬁcant. Increased maximum rotor current.5 10 5 0.] R Δimax [p.4 p. In [107]. However.3 ΔEg [p. Another possibility is to use a converter to substitute the Y point of the stator winding.. [54]. 5. for Method I (a) and Method IV (b).

the ﬂux oscillations will be damped by feedback of the derivative of the ﬂux. and 4) the method with a converter substituting the star point in the stator winding.u.5Lσ ˜ αc = 1.3 0. The reason that this method is chosen is that it has low cost (i..12. the system with the converter in Y point becomes very interesting and is accordingly further investigated. is easy to implement. Minimum remaining voltage without triggering the crowbar of a voltage sag. implying an increase of the losses. and the increased cost for an extra converter this method. The methods are 1) reducing the bandwidth of the current control loop. for Methods I (solid) and IV (dashed) as a function of the power. later on in Chapter 7 where different methods for voltage sag ride-through are discussed and compared. and a rotor current above 1. the cost is also increased. 2) compensation of the transformation angle (to synchronous coordinates). The maximum rotor voltage is set to 0. vanishes..4 p. 5.6 0. is not considered in this section since some of the beneﬁts and reasons for the doubly-fed induction generator. Compensation of the transformation angle method improves the damping only slightly. If we add a 75 . In this section. Lσ = 0 20 40 60 80 100 Power before voltage sag [%] Fig. the disadvantage of this method is that the method cause relatively high rotor currents. Lσ = 0 ˜ αc = 1.25 p. and can damp the ﬂux oscillations well.4 0. triggers the crowbar (stator-ﬂux-oriented system).g.7 0.u. Lσ = 0. Lσ = 0 ˜ αc = 7.9 0.2 0 ˜ αc = 1. 3) feedback of the derivative of the ﬂux.5 0. but the disadvantage of this method is the required addition in hardware and software.8 0. Due to the fact that the method with an extra converter connected to the Y point of the stator winding has to handle the stator current.1 Remaining voltage [p. no extra hardware).u. Kelber made a comparison of different methods of damping the ﬂux oscillations in [55]. It is concluded in [55] that the method of reducing the bandwidth works quite well. although it has the disadvantage of slowly damping of a grid disturbances. e. Feedback of the ﬂux derivative method performs well and has a low cost. The method with a converter in the star point of the stator winding performs very well. smaller (cheaper) converter and lower losses. Since there is a need for another converter.] 0.e. However. but the d component of the current can be used to damp the oscillations and improve stability. The q component of the rotor current is used for controlling the torque.

it is seen that the constraint on the d component rotor current has increased 1 + αd LM /(2Rs ) times.1 for typical parameters. since Rs is small and LM is large. Moreover.41). 76 .42) 5.46) 5. if a ﬂux estimator is used to determine the ﬂux.0 p + αd + 2− LM Eg 2 Eg − iref LM ωg 2 Rs iref Rq Rd. a ﬂux differentiation compensation term has been introduced. the bandwidth of the damper. see Table 2.6) can be rewritten as (with correctly known parameters) ωg LM iref Rs Rd. a high-pass ﬁlter is used since a pure differentiation is not implementable.0 < 2 + αd Eg LM Rs ωg LM (5.44) in order to guarantee stability. the characteristic polynomial in (5.7). of the low-pass ﬁlter must be set lower than the oscillating frequency in order to be able to damp the oscillation at all.27) is changed to (with correctly known parameters) p2 + αd + 2 Rs LM p+ αd Rs R2 2 + 2s + ωg LM LM (5. αd and αco .5. which we control as Rd Δiref (p) = − Rd p αco αd ψs = − 1 − ˆ p + αco Rs p + αco αd ψ ˆ s Rs (5.43) With the inclusion of a ﬂux damping. so that the ﬂux damper becomes slower than the current dynamics. must be chosen smaller than the bandwidth of the current control loop. the constraint on the d component becomes iRd. αd . αd . + αd Rs 2 Eg Eg LM ωg (5.0 p+ 1− ωg . 5. Of course.2 Grid-Flux Orientation For a grid-ﬂux-oriented system the characteristic polynomial in (5. This means that iref is set to Rd iref = iref + Δiref Rd Rd. must be smaller than the bandwidth of the ﬂux estimator.1 Stator-Flux Orientation Under the assumption that the current dynamics are set much faster than the ﬂux dynamics and αco is small. αco . Rd. that have to be determined. The damping term.component Δiref to the d component of the rotor current reference.5. that will improve the damping of the system. the cut-off frequency.0 (5. the ﬂux damping uses two parameters. (5. αc .41) then. it is possible to approximate the above characteristic polynomial as 2 p2 + αd p + ωg . Obviously. Comparing to (5.0 Rd where iref is used to control the reactive power as discussed in a previous chapter. In the above equation.45) if αco is small.3 Parameter Selection As can be seen in (5.5.

95 0 0. Since it is difﬁcult to see the effect of the ﬂux damper in a measured time series.8 1 Time [s] Fig. The reference frame is aligned with the stator ﬂux. 5. a factor of ten.3 p. Simulation of current control using a stator-ﬂux oriented reference frame with (solid) and without (dashed) damping of the ﬂux oscillations.u.6 0.6 0.u.u. with and without ﬂux damping. and αco is set to 0.5 p. αco . It is assumed that before and directly after the voltage sag.u. and.5. by the ﬂux damper. due to noise. was set to 2.u. was set to 0.3.2 0.1 s changed to Rq 0.2 0.7 p. In the simulation it is assumed that the ﬂux can be determined from measurements of the stator and the rotor currents.u.5.0 changed to 0.. 5.5 Response to Symmetrical Voltage Sags In this section the ﬂux damper’s response will be analyzed with respect to symmetrical voltage sags. The ﬁgure shows that the oscillations in the ﬂux has been damped with the ﬂux damper.4. The reference value of iref is initially zero and is at 0. αd . 5. is set to 4..u.4 Evaluation Fig.5 1 0 −1 0 0.1 (kE = 1 and kR = 1).2 0. b) iRq .2.u.6 0.7 p..7 p.] a) 1.4 s Rd.14. a frequency spectra of the ﬂux magnitude has been plotted instead in Fig. a) iRd . In the ﬁgure the current control method with feed forward of the back EMF and with “active resistance” has been used. αd is set to 0.05 p. i.8 1 0 0. the damping term.u. while Current [p.4 0.13 shows a simulation of a vector-controlled doubly-fed induction generator..] Current [p.8 1 c) Flux [p. The bandwidth of the current control loop.. was set to 0. c) ψs . αc . with and without ﬂux damping. The reference value iref is initially zero and is at 0. The DFIG is operated as in Section 5.] b) 1.5 p. αc .7 s to −0. according to Section 4. It can be seen in the ﬁgure that the 50-Hz component has been to a large extent damped.4 0. 5. The bandwidth of the system.05 1 0. and at 0.5 0 −0. the cut-off frequency term. The frequency spectra is based on a 6 s long measurement on the laboratory DFIG setup described in Appendix B.05 p.u.u. the magnitude of the 77 .4 0.5 1 0.13.5 p.e.5.u.

This means that the term 1 − V /Eg. 5.] 0 1 2 10 −2 10 −3 10 −3 10 −4 10 −4 10 −5 10 −5 10 10 10 10 0 10 1 10 2 Frequency [Hz] Frequency [Hz] Fig.51) t≥0: Δiref (t) ≈ − Δψs (t) = − (ψs (t) − ψs0 ) Rd Rs Rs which can be written as t≥0: Δiref (t) ≈ Rd αd Rs 1− V Eg.u. stator ﬂux can be expressed as t<0: t≥0: ψs (t) = ψs0 ψs (t) ≈ ψs0 V Eg.e.nom + 1− V Eg.u. from (5. The reference frame is aligned with the stator ﬂux. (5.53) . b) With ﬂux damping.nom ψs0 e−αd t/2 cos(ωg t) (5.52) 2 2 The above expression has a local maximum for t = arccos −2ωg / αd + 4ωg /ωg .nom ψs0 1 − e−αd t/2 cos(ωg t) . the low-pass ﬁlter αco /(p + αco ) has low bandwidth.47) (5. This means ever. i. (5. Frequency spectra of the ﬂux (data from measurements).] 10 −2 b) 10−1 Flux amplitude [p.nom . Then. if αd that the extreme value of Δiref (t) due to a symmetrical voltage sag can be expressed as Rd Δiref (t = π/ωg ) ≈ Rd αd ψs0 1 + e−αd π/(2ωg ) Rs 78 1− V Eg.nom corresponds to the magnitude of the sag..14. a) Without ﬂux damping.48) where ψs0 is the steady-state stator ﬂux prior the voltage sag and V is the remaining voltage after the voltage sag.49) If αco is considered small. it is possible to describe Δiref (t) after the voltage sag as Rd αd αd (5.a) 10−1 Flux amplitude [p.50) αco p + αco αd L {ψs (t)} Rs (5.41) the response in ΔiRd is estimated as Δiref (t) = L−1 − 1 − Rd or as Δiref (t) = L−1 − Rd αd Rs L {ψs (t)} − αco L {ψs (t)} p + αco . How2 2 4ωg it is possible to approximate t as t ≈ arccos(−1)/ωg = π/ωg . (5.

This value is.u. Δiref [p.u. It was found that the method that combines both the feed-forward compensation of the back EMF and the “active resistance” manages best to suppress the inﬂuence of the back EMF on the rotor current. 79 ..3 p.u.85 V = 0.” in order to eliminate the inﬂuence of the back EMF on the rotor current.7/0.8 V = 0.2 0.01 p. 5.u.9 p. Rd an unrealistically high value. Maximum of Δiref due to a voltage sag as a function of αd . Solid lines correspond to Rd simulation and dashed lines correspond to results from an analytical expression. 20 Max. then some of the advantage of the ﬂux damper is lost.01 · 1 1 + e−0.u.u.7π/(2·1) (1 − 0..53) is shown. 5.15.9/1) = 9. the stability and the damping of the system is independent of the rotor current. In Fig.4 0. The results in the ﬁgure shows that the ﬂux damper is very sensitive to voltage sags. The choice of current control method is of greater importance if the bandwidth of the current control loop is low.95 10 5 0 0 0. However. for a numerical example. in contrast to the stator-ﬂux-oriented system. has been analyzed. The results are presented for four different symmetrical voltage sags between V = 0.Consider the following values: V = 0.u. This means that the maximum value of Δiref due to the voltage Rd sag is Δiref = 0.1 0.7 αd [p.6 0. the bandwidth.8 to V = 0.15 the maximum value of Δiref due to a voltage sag as a function Rd of the bandwidth αd of the ﬂux damper can be seen.95 p.6 Conclusion In this chapter.] Rd 15 V = 0. In the ﬁgure both simulated results (using stator-ﬂux orientation) as well as analytically results from (5.. even with erroneous parameters. of course.u. (note that V is the remaining voltage). 5. during a voltage sag. It has been shown that by using grid-ﬂux orientation.7 p. αd = 0. it indicates that the ﬂux damper is very sensitive to voltage sags. αd . this method was found to be the least sensitive one to erroneous parameters and it manages to keep the rotor current close to unaffected.9 V = 0. of the ﬂux damper should be small.3 0. ψs0 = 1 p. although the analytical results are generally slightly higher. the general rotor current control law derived in Chapter 4. Moreover. However. Both methods produce similar results. with the option of including feed-forward compensation of the back EMF and “active resistance. This means that if the ﬂux damper should work during (small) voltage sags.] Fig. and Rs = 0.5 0.

Finally. By utilizing feed-forward compensation. it is shown that the design of the converter for a doubly-fed induction generator should also take into account a certain voltage sag to withstand and not only the desired variable-speed range. stability of the system resulting from the proposed current controller was found independent of the bandwidth of the current control loop. and the order of the system to analyze could be reduced. 80 . The introduction of an “active resistance” in the current control law improves the damping of low-frequency disturbances signiﬁcantly.This implies that for a grid-ﬂux-oriented system it is possible to magnetize the DFIG entirely from the rotor circuit without reducing the damping of the system.

as described in Section 4. if the frequency of the grid is constant (or at least close to constant).e.. While for a grid-ﬂux oriented system the above equation can be used directly. Either a PLL-estimator. 0) in order for the forward Euler discretization to be stable. the module must be discretizised. it is possible to force the system to have both slow and fast time scales. the rotor current dynamics and the grid-ﬁlter current dynamics are controlled by a high-gain feedback.35). care must be taken not to use a too long time step or sampling period. If the current control loop is much faster than the ﬂux dynamics.e. or. i.1. i. the poles must be inside a circle with a radius of 1/Tsample and the center point located at (−1/Tsample .1.. 81 . the system behaves like a singularly perturbed system [57]. However. ω1 = ωg .u. dΨs = Eg − dt Rs + jω1 Ψs + Rs iref R LM (6. The forward Euler discretization is given by (4.. For instance. can be used to track the frequency of the grid voltage. i.1 Reduced-Order Model If.7. where the equation is in polar form. For a stator-ﬂux-oriented system the above equation can be reduced to (5. ω1 . the synchronous frequency. As pointed out in the Introduction. that the rotor and grid-ﬁlter current can be assumed to follow their reference values accurately.Chapter 6 Evaluation of Doubly-Fed Induction Generator Systems 6. with an oscillating frequency close to 1 p. As mentioned in Section 4. the ﬂux dynamics of the DFIG are strongly inﬂuenced by a pair of poorly damped poles.7. the synchronous frequency can be put equal to the grid frequency.e.2) and (5.e. in PSCAD/EMTDC [66]. see Section 4. 6.3). for example. and this often due to its simplicity results in using the forward Euler method.1. Tsample . close to the line frequency. i.2 Discretization of the Doubly-Fed Induction Generator If the simple-to-use forward Euler method. when writing user-deﬁned modules. is used to simulate the system.3.1) where the stator voltage has been put equal to the grid voltage. must be determined. This means. it is sufﬁcient to study only the ﬂux dynamics and put the rotor current to its reference value.

. 6. the above expression is reduced to Rd Tsample < (6.2. ±ωg LM < 1 Tsample . the above-mentioned poles should be located inside the circle. − which yields ωg LM iref Rs Rd 2− ωg LM iref Rs LM Eg Rd 2− .6) is found as p1.2) ωg LM iref Rs Rd 2− ± jωg .6) In order for the discretization to be stable.It should be pointed out that in some other programs.7) 82 .2 Grid-Flux Orientation The solution to the characteristic polynomial in (5. 6. ±ωg 2LM Eg < 1 Tsample (6.0 − − Rs .1 Stator-Flux Orientation The solution to the characteristic polynomial for a stator-ﬂux oriented system in (5. user-deﬁned modules return expressions for the derivatives and advanced integration algorithms are used.e. < 2 < 2 ωg LM Eg ωg LM iref 2 Rs Rd 2− L2 Eg M 2 + ωg 4 Rs . is found as p1.2 = Rs ± jωg LM (6. corresponding to grid-ﬂux-oriented system.. iref = ψs /LM ≈ Eg /(ω1 LM ).6 µs.e. for instance Simpow [1] and PSS/E [85]. i.0 − − ωg LM iref Rs Rd 2− .3) Tsample (6. 2LM Eg In order for the discretization to be stable.e. In this case. (6. − 1 Tsample .27). 2 ωg LM 1 Tsample . i. the allowed time step can be made longer. i.2 = − ± ≈− ωg LM iref Rs Rd 2− 2LM Eg 2 ωg LM iref Rs Rd 2− 2 4LM Eg 2 Rs iref Rq 2 ωg − 1− Eg (6.4) For unity power factor. the above-mentioned poles should be located inside the circle.2. the sampling period should be Ts < 4.5) For the system investigated later on in this chapter and using the forward Euler method.

. Fig. 6. 6. simulations and experimental results of the response of a DFIG wind turbine to voltage sags are presented. and Li = 0. at t=0. 6.The solution to the above equation becomes Tsample < 1 2Rs 2Rs ≈ 2 2 LM Rs LM ωg 2 + ωg L2 M (6. Comparing the two ﬁgures it is seen that the agreement between the experiment and simulation is quite satisfactory. RR = 0. is presented.5 shows a simulation of the response to the same voltage sag as in Fig. similar results from simulations can also be found from a 83 . One reason for this is that the bandwidth of the current control loops (of the machine and grid-side converter) are set to 7 p.. The following parameters were used in the simulations: Rs = 0. 6.u. The WT now produces approximately 10% of its nominal power.. the simulations presented are carried out on a ﬁctitious 850-kW DFIG WT. Before the disturbance the WT is producing approximately half of its rated power. The dc-link capacitance is set to Cdc = 2. caused by the poorly damped poles due to the voltage sag. and Lσ = 0. LM = 4.u.01 p.u. As mentioned earlier.1 s the fault causing the voltage sag on the grid is cleared.1. In this case the disturbance is so large that the over voltage protection short-circuits the rotor and.7.1 s. In Fig.01 p.e.. the simulations shown in this section are carried out for a statorﬂux-oriented system. since real machine parameters were unknown. for the full-order model and Fig. In Fig. However.6 shows the corresponding simulation for the reduced-order model. 6. it is seen that the agreement is quite satisfactory.u. Fig. Moreover. The grid ﬁlter for the grid-side converter Ri = 0.u.3 shows the corresponding simulation with the reduced-order model of the system.1.4.8 p. of the d component of the rotor current.21 p. and after 0. The voltage drops down approximately 25%. 6.4.). and the voltage starts to recover. The wind turbine produces about 20% of the nominal power. as shown in Fig.5). Moreover. is clearly seen.2. which is sufﬁciently higher than the eigenfrequency of the ﬂux dynamics (close to 1 p. The experiments were made on a VESTAS V-52 850 kW WT and in Appendix B.u. a severe voltage disturbance is presented.u. a 75% sag. experimental results of the response due to an unsymmetrical voltage sag are presented. 6. It can be seen in the ﬁgure that the full-order model and the reducedorder model produce almost the same results. shown in Section 5. Fig. An exact agreement is not to be expected.8) which is twice the value obtained by (6. the minimum sample time for the grid-ﬂux-oriented system is independent. in contrast to a stator-ﬂux-oriented system. the breaker disconnects the stator from the grid.u.3 Response to Grid Disturbances In this section. for the full-order model. Again.07 p.0071 p. after 40 ms.u. i..9 p.1 shows experimental results of the response of a DFIG wind turbine to a voltage sag. 6. The simulations have been carried out both with a “full-order” model and a second-order model. In Fig. 6. a simulation of the response to the same voltage sag.3 a short description of the used data acquisition setup is presented. 6. The oscillation close to 50 Hz.

c) Active power.2.1 0.2 0.u.3 0.3 0.2 0. 6. stator-voltage oriented (or grid-ﬂux oriented) system.1 0.7 0 0. b) Grid-current magnitude.3 0.2 0.4 Reactive power [%] c) 30 Power [%] Time [s] d) 10 Time [s] 0 −10 0 0.2 0.] 0.3 0.1 0.u.8 0.4 Current [p.u.8 0.2 0.1 Voltage [p.2 0.1.1 0.3 0. d) Reactive power.3 0.u.2 0.] 0.4 b) 0.] 1 0. b) Grid-current magnitude.2 0.3 0.1 0 0 0.4 20 10 0 0 0.4 Current [p. 84 .4 20 10 0 0 0.4 b) 0.1 0 0 0. a) 1.a) 1.4 Time [s] Time [s] Fig.4 Time [s] Time [s] Fig. d) Reactive power.1 Voltage [p.1 0. c) Active power. 6.3 0.1 0.9 0.1 0.2 0.9 0. Experiment of the response to a voltage sag. Simulation of the response to a voltage sag with the full-order model. a) Grid-voltage magnitude.1 0.4 Reactive power [%] c) 30 Power [%] Time [s] d) 10 Time [s] 0 −10 0 0.7 0 0.2 0. a) Grid-voltage magnitude.] 1 0.3 0.3 0.

6.2 0.1 Voltage [p.4 Time [s] Time [s] Fig.7 0 0. c) Active power.3 0. Simulation of the response to a voltage sag with the reduced-order model.4 b) 0.3 0.3 0. Experiment of the response to a unsymmetrical voltage sag.15 0.25 Current [p. d) Reactive power.2 0. b) Grid-current magnitude.7 0 0.2 0.2 0.] 0.u.1 0.u. a) Grid-voltage magnitude.4 b) 0.3 0.4.4 Reactive power [%] c) 30 Time [s] d) 10 5 0 −5 −10 0 0.1 0.4 Reactive power [%] c) 30 Power [%] Time [s] d) 10 Time [s] 0 −10 0 0.3 0.4 Current [p.4 20 10 0 0 0. d) Reactive power.3 0.1 0.1 0. a) 1.1 0 0 0.1 0. a) Grid-voltage magnitude.4 Power [%] 20 10 0 −10 0 0.3 0.2 0.1 0.2 0.8 0.05 0 0 0.1 0.2 0.a) 1. b) Grid-current magnitude.] 0.2 0.3 0.u.1 Time [s] 0.3.9 0.1 Voltage [p.u. 85 .] 1 0.2 0.1 0. c) Active power.4 Time [s] Time [s] Fig.8 0.2 0.3 0.9 0. 6.] 1 0.

6.3 0.3 0. a) 1.4 b) 0.25 Current [p.3 0.1 Time [s] 0.2 0.1 0.3 0.3 0.2 0.u.2 0.05 0 0 0.a) 1.] 1 0.4 Power [%] 20 10 0 −10 0 0.2 0.2 0. c) Active power. d) Reactive power.1 0.2 0.u.u. Simulation of the response to an unsymmetrical voltage sag.1 0.1 0. 6.1 Voltage [p.15 0.2 0.2 0.8 0.15 0.05 0 0 0.3 0.2 0.3 0.5.4 Power [%] 20 10 0 −10 0 0. The simulation has been performed with the full-order model.25 Current [p.9 0.2 0.u.3 0.4 Reactive power [%] c) 30 Time [s] d) 10 5 0 −5 −10 0 0.1 0.4 Time [s] Time [s] Fig. The simulation has been performed with the reduced-order model.1 Voltage [p. b) Grid-current magnitude.7 0 0.7 0 0.4 Reactive power [%] c) 30 Time [s] d) 10 5 0 −5 −10 0 0. a) Grid-voltage magnitude.1 Time [s] 0. Simulation of the response to an unsymmetrical voltage sag.] 1 0. 6.9 0.4 b) 0.1 0.] 0.8 0.] 0. d) Reactive power.1 0.1 0. 86 .4 Time [s] Time [s] Fig. c) Active power. a) Grid-voltage magnitude. b) Grid-current magnitude.

02 0.1 Current (q) [p. an example is PSS/E. Examples are EMTDC and Simpow. This means that if stator ﬂux transients are negligible a steady-state model of the DFIG dynamics are sufﬁcient as long as the rotor circuit is not short-circuited due to a too large grid disturbance. and for these programs. Severe voltage disturbance.06 0.02 0.4 Implementation in Grid Simulation Programs Some grid simulation programs can handle three-phase instantaneous quantities.] b) 0 4 2 0 0 0.] c) 0. it may not be possible to use such a short time step (about 5 µs) as is required in order to simulate the control of the DFIG system. a) Grid voltage.5 1 0. since the time step is often too large for these oscillations. However.] 1.04 0. the machine will act as a standard squirrel-cage induction machine which can be adequately modeled with a ﬁfth-order model of the induction machine [83]. the stator ﬂux transients may be negligible from the power system stability analysis point of view.a) Voltage [p. The suggested approach is to simply ignore the 50-Hz oscillations when the DFIG system is implemented in simulations with long time steps. Other programs are designed to handle the voltages as phasors. b) d component of the grid current.06 0. 50-Hz oscillations in the output quantities cannot be captured. a steady-state model of the DFIG is sufﬁcient.02 0.04 0.06 0. 87 . c) q component of the grid current. as long as the disturbances are small enough not to cause the rotor to be short-circuited.u.7. However. 60]. 6.08 0.04 0. For this case.08 0.1 4 2 0 0 0. when handling simulations of large systems. As pointed out in [84.u.1 Time [s] Fig.5 0 Current (d) [p. if a disturbance is large enough to cause the rotor to be short-circuited.08 0.u. 6.

5 Summary In this chapter. Both models produced acceptable results. 88 . simulations and experimental veriﬁcation of the dynamic response to voltage sags of a DFIG wind turbine were presented. The response to symmetrical as well as unsymmetrical voltage sags was veriﬁed.6. Simulations were carried out using a fullorder model and a reduced-order model. Perfect correspondence with experiments were not expected since the simulations were carried out on a ﬁctitious DFIG wind turbine.

7.e. WTs have to ride through these voltage sags. Proposed regulations from the Swedish national grid company.Chapter 7 Voltage Sag Ride-Through of Variable-Speed Wind Turbines As mentioned in the Introduction.1. simple space vector models will be presented for some common voltage sags that will be used in this chapter. Then.6 0.8 0.6 0.4 0. This investigation will the serve as a basis for the comparison of DFIG ride-through systems. Svenska Kraftn¨ t [96]. First.2 0. This means that new wind turbine installations have to stay connected to the grid for voltage sags above a certain reference sag. the proposed Swedish requirements for voltage sags is depicted. 89 . new grid codes are in progress both in Sweden and other countries.2 0 −0. the voltage sag response of a WT that utilizes a full-power converter is investigated.] 0.u.3–100 MW.2 Time [s] Fig.1. these systems will be compared dynamically as well as for steady-state operation.2 0 0. Solid a line is the requirement for wind parks with a rated power larger than 100 MW. and systems for voltage sag ride-through will be investigated. In Fig.2 1 Voltage [p.4 0.. In the next sections the voltage sag response of the DFIG will be further analyzed.8 1 1. 7. Dashed line is the requirement for wind turbines and wind parks with a rated power between 0. Finally. i. 1.

0.4) .5–30 cycles.1 Voltage Sags With the expression “voltage sag. and the remaining rms voltage can be as low as V = 0 for a direct-to-ground fault. In this section. the impedance of each symmetrical component can be hard to derive.u. 7. it is assumed that the source and feeder impedance are much larger compared to the line-toground impedance. Single-Line-to-Ground Fault After a single-line-to-ground fault (SLGF) in the ﬁrst phase the grid phase voltages can be expressed as E1 (t = 0+ ) = E2 (t = 0+ ) = E3 (t = 0+ ) = √ √ √ 2V cos(θg + π/2 + φ) 2Eg.” it is normally implied that the grid rms voltage drops from 1 p. e. There are other protection devices. Zero sequences are not critical for a PWM rectiﬁer since such sequences ideally disappear from the phase currents due to the absence of a neutral conductor. negative-.g.” The majority of all “phase-angle jumps” are smaller than 45◦ [9]..2) (7. Directly after a symmetrical voltage sag. and zero-sequence impedance are assumed to be equal. and φ is the “phase-angle jump.1. The fault clearing time for protective relays varies from 50 ms up to 2000 ms [9]. These models are developed in [74] and the aim of the models are to estimate the moduli of the positive. for instance.1 Symmetrical Voltage Sags Symmetrical (or balanced) voltage sags implies an equally reduction of the rms voltage and. a “phase-angle jump” in all three phases [9]. the positive-. that might have a shorter fault clearing time (less than one cycle)..3) (7.u. to 0.9 p.1–0.e.2 Unsymmetrical Voltage Sags Unsymmetrical (or unbalanced) voltage sags are more difﬁcult to model since.1.1) ˜ where V is the remaining rms voltage in the faulted phases. For ground faults.nom cos(θg + π/2 + 2π/3) 90 (7. The origin and classiﬁcation of voltage sags are well explained in [9]. possibly. θ0 is the initial error angle. current-limiting fuses. in order to simplify the derivation of models suitable for unsymmetrical voltage sags. The impedance between the two faulted lines for a line-to-line fault is neglected. i.7.nom cos(θg + π/2 − 2π/3) 2Eg. simple space vector models will be presented for some common voltage sags. However.and negative-sequence voltage vectors for different types of sags. the grid voltage vector can be expressed in the synchronous reference frame as Eg (t = 0+ ) = jV ej θ0 = jV ejφ ˜ (7. 7. The duration of voltage sags is mainly determined by the clearing time of the protection used in the grid [9]. for a short period of time. Voltage sags caused by these fuses are short and deep if the fault is in the local distribution network but if the fault is in a remote distribution network the sag is short and shallow [9].

i. E2 .5 p. (7. i.. respectively.5) are the stationary parts of the positive. and E3 are the grid phase voltages directly after the sag. Ep becomes real valued and. ideally. such ˜ that.5 · 0. V = 0.5 · 0.71)] ≈ −0.nom 2 jφ + V e .nom cos(θg + π/2 + 2π/3) which correspond to the following space vector in the synchronous reference frame: Eg0 = j Ep0 + En0 e−j2θg where Ep0 = Eg. it is seen that minimal modulus of the positive-sequence voltage vector is Ep = 2Eg.e.nom /3 and that the maximum negative-sequence voltage vector is En = Eg.6).15 rad ≈ −9◦ .. θ ≈ 0.u. the PLL will track the position of the positive-sequence voltage vector. Two-Lines-to-Ground Fault After a two-lines-to-ground fault (TLGF) between the ﬁrst and second phase. a total loss of voltage in the faulted phase. En0 = E∗ − V e−jφ e−jπ/3 . p0 3 3 91 (7. θ1 = θg .7) Eg0 = j Ep0 + En0 e−j2θg .4) is then found as Es = j Ep0 ejθg + En0 e−jθg g0 where Ep0 = 1 1 2Enom + V ejφ . As expected. For perfect pre-sag ﬁeld orientation. From (7. The space vector in a stationary reference frame that corresponds to (7. the grid phase voltages can be expressed as √ (7.u.71/(2 + 0. The initial error angle of the positive-sequence voltage vector due to a SLGF is ˜ θ0 = arg(Ep0 ) = arctan V sin φ 2Eg.u. respectively. and φ = −45◦ .e.9) E1 (t = 0+ ) = 2V cos(θg + π/2 + φ) √ (7..e.11) E3 (t = 0+ ) = 2Eg.. hence.12) (7.13) . Consider the following values for a numerical example: Eg. the negative sequence becomes in the synchronous reference frame a component with a frequency of twice the fundamental frequency.5 p. (7.10) E2 (t = 0+ ) = 2V cos(θg + π/2 − 2π/3 + φ) √ (7. and V and φ are the remaining rms voltage and “phase-angle jump” in the ﬁrst phase. This gives an initial error ˜ angle of θ = arctan[−0.nom /3.where E1 .nom + V cos φ . En0 = Enom − V e−jφ 3 3 (7. The initial error angle ˜ becomes even smaller if V is smaller than 0.and negative-sequence voltage vectors. θ0 = 0 for V = 0.nom = 1 p. i.6) (7.5) can be transformed to the synchronous reference frame by substituting Es = Eg0 ejθg and solving the resulting equation for g0 Eg0 : (7. −2ωg .2)–(7. for instance.8) Eventually. This occurs when V = 0.

5. 7.nom /3 for V = 0. 7. the dc-link voltage dynamics are analyzed for various disturbances and voltage sags. Line-to-Line Fault Directly after a worst-case (no feeder impedance) line-to-line fault (LLF) between phases 2 and 3. The system conﬁguration consists of a generator and two converters connected “back-to-back” as depicted in Fig.2. it can be seen that the minimal modulus for the positive-sequence voltage vector is Ep = Eg. with the exception of Pr = −Pt . designed for the rated WT power. The dc-link voltage controller presented in Section 4. the voltage sag response of PWM rectiﬁers.nom (7.From (7. the grid phase voltages are found as √ (7. (7. the modulus of the positive. This exception indicates that the rotor power.nom .17) E3 = √ cos(θg + π/2 − π) 2 which correspond to the following space vector in the synchronous reference frame: Eg0 = j 1 − e−j2θg Eg.14) An initial error angle of −23◦ is obtained as a numerical example using the same values as in the previous section. As a result of the analysis. The main focus of this section is put on the PWM rectiﬁer and the achieved results and conclusions are independent of the type of converter at the generator side. accurate estimates of the transient and steady-state response of the grid current and dc-link voltage during voltage sags are provided. the grid-ﬁlter current equals the grid current. Pr .2 will be considered. Pt . Note that for this case.nom cos(θg + π/2) Eg.2 Full-Power Converter In this section.15) E1 = 2Eg.and negative-sequence voltage vectors both equal Eg.1 Analysis First.nom /3 and that the maximal modulus for the negative-sequence voltage vector is En = Eg.nom + 2V cos φ .18) Obviously.2.nom (7. 7.13). for the full-power converter analyzed in this section. and the initial error angle moments after the LLF equals zero.nom /2. is analyzed. used for the DFIG is changed to the total turbine power. The initial error angle directly after a TLGF is ˜ θ0 = arg(Ep0 ) = arctan 2V sin φ Eg.16) E2 = √ cos(θg + π/2 − π) 2 Eg. 2 (7.12) and (7. These results can be useful when designing a PWM rectiﬁer for various grid codes and requirements. It is assumed that 92 .

2 · 50) = 490 times smaller than fsw . Then.028) ≈ 0. Wind turbine with a full power rectiﬁer.9 kHz.9 kHz. so (7. will be considered which. The transfer function from ref the turbine power Pt to the error signal. To ensure that this voltage ripple remains below a tolerable value. the grid-ﬁlter current dynamics and the switching transients at the dc link are.3. which equals the capacitance of the experimental setup in Section 7. 7. 93 . symmetrical and nominal grid voltage. αw is 4900/(0. these values yield Cdc. ew = Wdc − Wdc . neglected. very high demands [50] are placed on the dc-link voltage control loop when using such small a dc-link capacitance.19) Cdc p2 Since the dc-link dynamics are considered to be much slower than the switching and sampling frequency. steady-state condition. is the nominal q-axis current and vdc is the tolerable peak-to-peak ripple fq p-p for the dc-link voltage. The value vdc = 0.u. is considered. with (4.5 p.u. Henceforth.2 p.20) is mainly a benchmark that can be used for comparison to more realistic operating conditions. of the PWM rectiﬁer. which corresponds to 1 % peak-to-peak ˜ ref a ripple at vdc = 2. and perfect ﬁeld orientation are assumed to precede the different disturbances.u. Minimal DC-link Capacitance In PWM rectiﬁers. for a 50-Hz grid. becomes GP e (p) = −GP W (p) = −2p . the dc-link capacitance should be selected no smaller than [59] √ nom 3if q (7.PWM “rectiﬁer” Grid ﬁlter Lg Cdc SG Grid Fig. are considered for √numerical example.8 and fsw = 4.u.min = 8fsw vdc ˜p-p ˜p-p where inom = 1 p. Moreover.min = 3 · 1/(8 · 4900/314 · 0. fsw . i. For a base frequency of ωb = 314 rad/s.e.028 pu. if the bandwidth of the dc-link voltage control loop is αw = 0.2. Cdc = Cdc . 2 + 2αw ξp + αw ξ (7.120).5 p. ˆ the dc-link capacitance is accurately modeled. a dc-link capacitance of Cdc = 3.. the current in the dc-link capacitors is heavily distorted which gives rise to a small (compared to diode rectiﬁers) ripple in the dc-link voltage. thus.20) Cdc. For instance. and the switching and sampling frequency is fsw = 4.2. However.

are considered for a numerical example which yield a local maximum for ew (t) at emax = w ref 0. where Egq (t) changes stepwise at t = 0 and if q (vdc ) is a function of the dc-link voltage (via the vdc control loop). w Response to Symmetrical Voltage Sags As mentioned earlier. Therefore. though.24) by V /Eg. (50 % of nominal power).24) where Pt0 is the pre-sag turbine power. the dynamics in (7. perfect ﬁeld orientation and steady-state condition. becomes ew (t) = L−1 Gnom (p) Pe ΔP p =− 2ΔP −αw t e t. it is assumed that symmetrical voltage sags are preceded by symmetri˙ cal and nominal grid voltage.nom if q + Pt0 (7. Meanwhile. (7. from Pt (0− ) = 0 to Pt (0+ ) = ΔP . Then. i.5 = (2. it follows from (7. With vdc = 2.e. Cdc (7. ew (t). This means that the net power step is ΔP = (1 − V /Eg. the maximum/minimum value of ˙ ew (t) is 2ΔP −1 0.nom when analyzing the capability of the dc-link voltage control loop to reject disturbances in Pt .u. Pt (t) = Pt0 V /Eg.22) Depending on whether ΔP is positive or negative. However.25) appear to be time-varying. the q-axis grid voltage can considered to be constant at 94 .2 p. αw = 0. (7.nom ..Assessment of Turbine Power Reduction The grid-voltage modulus is normally close to its nominal value.8 p.19) is reduced to Gnom (p) = GP e (p) Pe = Egq =Eg.25) and (7. and Cdc = 3.82 − 1. the error.u.u. this is not the case. (7.26) Then by introducing the “new” turbine power.26) that a symmetrical voltage sag is equivalent to a positive step in Pt .22) has a local minimum or maximum for t = 1/αw (determined by solving ew (t) = 0).u. this corresponds to a minimum dc-link min ref voltage of vdc = ((vdc )2 − emax )0. which changes from Pt (0− ) = Pg0 V /Eg.25) after a sag at t = 0. These assumptions imply the following dynamics for W : t≥0: 1 dW C = −3V if q + Pt0 2 dt (7.5 p.nom )Pt0 . Cdc (p + αw )2 (7.2 · 3.5 p.5) ≈ 1. (7.6)0. Since the power to the grid ﬁlter is Pf = 3Egq (t)if q (vdc )..74ΔP e ≈− . Moments after a symmetrical voltage sag occurs. W = 0.5 p.u. Therefore.nom .74 · 1. it is natural to let Egq = Eg.nom : t<0: 0 = −3V if q + Pt0 V Eg.5 = 2. it can be assumed that Egq = V while Pt remains at its pre-sag value.21) For a step in the turbine power.nom to Pt (0+ ) = Pt0 . This implies that: t < 0 : 0 = −3Eg.23) emax/min (t = 1/αw ) = − w αw Cdc αw Cdc The values ΔP = −1.u.nom −2p . which can be deduced by multiplying (7.5/(0.6 p.

1) = 10 p.1 p.1 p. Remedies for avoiding severe overcurrents during symmetrical voltage sags are discussed in Section 7.2 0.6 p.u. The single exception to this power step equivalence is when V = 0. which corresponds to that power cannot be transferred to the utility grid.nom ..4 = −0. Once the equivalence to turbine power steps has been revealed... i. is obtained from the inverse Laplace transform of (7.nom cos[θ(t)] which yields the instantaneous grid˜ ﬁlter power as Pf = 3Eg. in order to regain steady-state conditions during a sag. the ˜ q-axis grid voltage varies as Egq (t) = Eg.94 p.28).28) has a local minimum or √ maximum for t = arcsin( 1 − ξ)/ωw . (7. the poles of this transfer function are αw V ± j Eg.u.82 − 0.2/2.u.nom V − V 2 . determined by the dynamics of the error angle.u.29) The values C = 3.2. Depending on the sign of ΔP .6 ≈ 0. ξ = 0. Consider Pt = 3 p. for instance.19) multiplied by the step ΔP /p: ew (t) = L−1 GP e (p) ΔP p =− 2ΔP −αw ξt sin (ωw t) e ωw C (7.u. the dynamics of ew during symmetrical voltage sags are. With w ref min vdc = 2.u. given by (7.19). the extreme value for ew (t) is obtained as ew = − 2ΔP exp − ωw C ξ arcsin 1−ξ 1−ξ 1 − ξ. αw = 0.5 · 0.2. Since Pf is a function of time and vdc .6 and √ ωw = 0.u.” The response of PWM rectiﬁers to “phase-angle jumps” is.u. This yields a local maximum of emax√≈ 0.27) are well damped for V ≥ Eg..5 p.. is that symmetrical voltage sags may require very large if q in order to counteract the reduction in the grid rms voltage such that Pf = Pg in the steady state.5 p.e.21)..8 · 100 ≈ 7 %.2 = − Eg. V = Eg. By substituting Egq = V .. More troublesome however. i.30) ˜ In the time interval when θ(t) converges exponentially to zero with the rise time 1/ρ. (nominal power) and V = 0.6 p.u. it is assumed that the modulus of the grid voltage vector remains constant at Eg. no voltage sag accompanies the “phase-angle jump.8 p.Egq = V provided that accurate ﬁeld orientation is maintained.u. the time function of the error angle after a “phase-angle jump” can be modeled as ˜ ˜ θ(t) = θ0 e−ρt . which demands for if q = 3/(3 · 0.nom /2.nom . to a large extent.27) p1.nom if q (vdc ) cos[θ(t)]. Moreover. are considered for a numerical example. the poles of (7. i. By substituting this instant in (7.6 p. (7. the 95 .94 ≈ 2.. Provided that overcurrent is avoided.28) where ωw = αw (1 − ξ)ξ.. after a symmetrical voltage sag.e..u. (7. Response to “Phase-Angle Jumps” For reasons of simplicity and clarity. V = 0. Pt0 = −1.4 · 0. For a PLL tuned assuming a bandwidth of ρ. (7. the dc-link voltage decreases by 0.e.nom For normal operation. the poles are located at −αw . this corresponds to a minimal dc-link voltage of vdc = 2.2 p. This means that ΔP = −1. as seen in (7. ew (t). hence.u.

nom . With vdc = 2. this implies that the dc-link voltage min decreases to vdc = (2. the seemingly daring assumption of nearly constant dc-link voltage during “phase-angle jumps” is adopted. and Cdc = 3.2)0.e. ˙ reasonable.nom if q (vdc ) is a function of vdc only.15 · 1.31) are time-invariant. The time function of the error signal can be derived by taking the inverse Laplace transform of the product of (7.2.5 p. The error signal ew (t) has a local minimum/maximum for t = ln 2/ρ.8 p.2 p. such that 1/ cos θ ≈ 1 + θ2 /2. Therefore. θ0 = −π/4 rad.u. ≈− 2 ρCdc (7. (7..5) ≈ 0. from (7. on the 96 .35) ˜ The values Pt0 = −1.76 p. which is supported by simulations and experiments in Section 7.82 − 0. From this analytic ﬁnding.u. For small θ. since the remaining positivesequence voltage of unsymmetrical sags is never as small as that of the worst-case symmetrical sag.3.21) and ΔP (p): ew (t) = L−1 Gnom (p)L{ΔP (t)} = Pe ˜2 2θ0 Pt0 −ρt ρ e + t − 1 e−ρt ρCdc 2 (7.u.dynamics of W . it can be concluded that “phase-angle jumps” are believed not to be critical for PWM rectiﬁers. since Pf = 3Eg. in the beginning of this section).5 ≈ 2.33).15θ0 Pt0 .2 · 3.31) t ≥ 0 : 0 ≈ −3Eg.2. are considered for a numerical example.30).34) since e−ρt initially decays faster than ρt/2 increases. i.04/2.” ew (t) can be approximated by ew (t) ≈ ˜2 2θ0 Pt0 −ρt e − 1 e−ρt = ew (t) ρCdc (7. are time varying during “phase-angle jumps” in contrast to symmetrical voltage sags.nom if q cos θ + Pt0 = −3Eg.u..32) 2 2 where the latter expression results from (7.31).8 · 100 ≈ 1 %. a decrement by 0. as discussed previously.2 p. the “dynamics” after a “phase-angle jump” at t = 0 simplify to Pt0 ˜ .u. As a remedy for this.nom if q + ˜ cos θ The approximated “dynamics” in (7. Response to Unsymmetrical Voltage Sags The response of PWM rectiﬁers to unsymmetrical voltage sags is partly similar to the response to symmetrical voltage sags. [76] (this happens to coincide with the selection of αw = 0.u. By substituting this instant in (7. This assumption is validated by simulations and experiments in Section 7. With W ≈ 0.5 · w ref π 2 /(42 · 0.24) and (7. A unique property of unsymmetrical sags is. which gives a local maximum of emax = 0. a “phase-angle jump” is in close correspondence to a time varying Pt which changes from Pt (0− ) = Pt0 to Pt (t) = ˜ ˜ ˜ ˜ Pt0 / cos[θ(t)] at constant Egq = Eg.. the extreme value for ew (t) is found as ew ≈ ˜2 2θ0 Pt0 ρCdc 1 ln 2 + −1 2 2 ˜2 1 0.. although less critical. the net change in Pt is ˜ ˆ θ2 (t) θ2 ΔP (t) = 1 + Pt0 − Pt0 = 0 e−2ρt Pt0 (7.u.u.2 p.3 which show that the approximation is. Within a short time interval after a “phase-angle jump.2 p.5 p. αw = ρ = 0.33) where αw = ρ is assumed since proper rejection of the negative-sequence voltage requires a PLL bandwidth of ρ ≈ 0. indeed.

As for symmetrical sags.8 − (2.e. a q-axis current of iavg = 1 p.nom ) 97 (7. Provided with this expression for Egq and. which can be expressed as if q = iavg + ˜f q .08 p.u. The ref values En = 0.. remedies for avoiding large q-axis currents are to be discussed in Section 7. Pf = 3 fq fq p. where iavg is the average value of if q fq fq and ˜f q is the current ripple. small current |˜f q |...u..36) From (7.u..u. then (7.nom .nom (p2 + 2αw Egq /Eg. i..2.5 p. yields Pf = 3Ep iavg = 1.5 ≈ 0.37) where ξp = Ep /Eg. is determined.nom p + αw Egq /Eg. or a ripple of 0. The ˜w ˜pk corresponding peak value of the vdc ripple is vdc = 2.36).. and Cdc = 3. the dc-link voltage dynamics simplify to ı ripple. ωg Cdc (7.43)0. such that |iavg | fq dW 1 Cdc ≈ −3Ep if q − 3En iavg cos(2θg − ϕ) + Pt . This may be too large a current to be tolerated in a WT fq application. 4. to a large extent. the ripple in ew . such that ωg αw . fq 2 dt (7.e.u. vdc = 2.5 = 2 p. Two simpliﬁcations are introduced in order to analyze these ı ripples.2. due to an unsymmetrical voltage sag.82 − 0. smaller 2 2 than 0. whereas nominal power.u.8 p.other hand. the modulus of the positive-sequence voltage vector is the most critical consequence. by the dc-link capacitance. the power ripple during unsymmetrical voltage sags also transfers to the q-axis current.43 p.8·100 ≈ 3 %.. An LLF is considered for a numerical example. the q-axis grid voltage eventually varies as Egq (t) = Ep + En cos(−2θg + ϕ) where ϕ is the angle of the negative-sequence voltage ˜ vector for θ = t = 0. that the negative-sequence voltage vector introduces a ripple in the instantaneous grid power. which adds to the grid current distortion during faults. This indicates that if Ep = 0.5 p.. First shortly after an unsymmetrical sag it is assumed that the PLL recovers the position of the positive-sequence voltage vector.u. 2 3Eg.u.37) can be approximated as |GP e (j2ωg )| ≈ Cdc 4ωg 2 (−4ωg )2 = 1 .e. with constant Egq = Ep .5) = 0. via the dc voltage control system. The dc voltage q ˜ ripple that results from Pt are obtained from the static gain of GP e (p) in (7.15.08/2.3 p. In order to analyze the resulting q current ripple.19) at the relevant frequency 2ωg and Egq = Ep : |GP e (j2ωg )| = Cdc 4ωg 2 2 (αw ξp − 4ωg )2 + (4ωg αw ξp )2 (7. yield a peak ripple fq ˜ of epk ≈ Ptpk /(ωg Cdc ) = 3 · 0. If αw is selected at least three times smaller than ωg . at a frequency of 2ωg . which results in GP i (p) = −2[Ga + F (p)]/(pCdc ) 1 − 6Eq [Ga + F (p)]/(pCdc ) 2αw (p + αw /2) = . denoted by ˜ ˜ Pt = 3En iavg cos(2θg − ϕ) = Ptpk cos(2θg − ϕ).u. As previously discussed.u.5 p. iavg = 1 p.5 · 1/(1 · 3.u. This is a fairly small ripple which is not critical for the proper operation of a PWM rectiﬁer. requires iavg = 1/0. i.u.39) . the transfer function from Pt to if q can be derived from Fig. the power ripple can be treated as a turbine power disturbance.5 p.. i.38) Hence. This power ripple in turn gives rise to ripple in the dc-link voltage and ı in if q . secondly.

40) implies that the resulting ripple in if q is mainly determined by the bandwidth of the dc voltage control system. This yields a peak ˜ ripple of ˜pk ≈ αw Ptpk /(3ωg Eg. Therefore. The relation in (7.40) where the latter approximation holds when αw is selected at least three times smaller than ωg . • Grid codes. it may happen max that the turbine power is larger compared to Pf . Depending on these factors.u. The design of such energy storages depends on several factors of which some are: • Cost. “Braking” Chopper A “braking” chopper.nom ) = 0.2 · 1. 98 . the turbine power is controlled to Pt = Pf by changing the torque reference for the turbine. a less distorted grid current during unsymmetrical voltage sags can be obtained by selecting αw smaller.nom (7. the blades should preferably remain in their pre-sag position.1 p. For such operating conditions.5/3 = 0. acting as a load dump. the dc-link voltage begins to increase. ı assumption on |iavg | fq 7. or possibly a combination. If there is no need for instantaneous power restoration. the static gain of GP i (p) at the relevant frequency 2ωg is obtained as |GP i (j2ωg )| = αw 3Eg. • The remaining modulus of the positive-sequence grid voltage vector and the duration of the voltage sag. during an LLF.u. with identical ıf q values as previously and αw = 0. no more than Pf = 3V imax can be transferred to the fq utility grid.2 p. can be installed at the dc link. The limiting factor of this solution is the heat generated by the “braking” resistor which may be troublesome to remove for long-duration voltage sags or interruptions.2. This is a fairly large ripple although the previous |˜f q | is still reasonable.By substituting Egq = Ep in this expression.nom 2 2 αw + 16ωg αw ≈ 2 2 (αw ξp − 4ωg )2 + (4ωg αw ξp )2 3ωg Eg. the blades can be pitched out of the wind directly. one. Depending on the wind situation when the voltage sag occurs.2 Discussion In general. If the pre-sag grid power must be restored moments after the voltage sag is cleared. unless the WT approaches overspeed. unless the excess energy is somehow stored or dissipated. For a voltage sag max where the modulus reduces to V . of the following four solutions may be applicable: Rotor Energy Storage In this solution. WTs using PWM rectiﬁers are robust towards voltage sags but large reductions in modulus of the positive-sequence voltage vector appear to be critical.

The grid voltage is recovered at t = 0.u.u. Fig.21 · 2.u.8 rad occurs at t = 0. max if q = 1 p.2 p.3 s. in the steady state. The PWM rectiﬁer uses 4.3.3 s.DC-Link Energy Storage A large dc-link capacitor bank can possibly be used.25 s. Fig. In all other aspects.8 p. characterized by Ep = 0.5 p.1vdc )2 − ref ref (vdc )2 = 0.3 p. This is close to the predicted value vdc = 2. This is a very large value. which gives Cdc = 2 · 0. A pure “phase-angle jump” of φ = −45◦ ≈ −0. i.4 shows the results of an unsymmetrical voltage sag. 7. Eg = 0.e. are considered for a numerical example.21(vdc )2 . The closed-loop grid current and dc-link voltage control loops are tuned for the bandwidths 2. and if q = −0.32 s.2.u. such that energy from the WT is buffered at the dc link during the sag.3 Evaluation This section presents simulated and experimental results of a PWM rectiﬁer which is subjected to various disturbances and voltage sags. which corresponds to Cdc = 3. the simulation and corresponding experiment are carried out under similar conditions as in Fig.41) Cdc = −3V imax + Pt .1.1–0.6 p. then WΔ = (1. at t = 0.6 p. fq ΔW If the dc-link voltage is allowed to increase by no more than 10 %. 7. The required size of the capacitor bank can be calculated ˙ by substituting W = WΔ /tΔ in (7.5 p.9 kHz sampling and switching frequency and the reference for the dc-link voltage is normally 2. so a dc-link energy storage appears to be suitable mainly for small voltage sags that appear for a short period of time.65 p. respectively. which corresponds to a current rise time of 3 ms and a dc-link voltage rise time of 35 ms. The d current reference equals zero.u. and 1.u. 7.2 p. assuming if q = imax . imax > 1 pu. 7.5 p.u.2.” as already concluded.12 s. 50 Hz. which yields an initial error angle of ˜ ˜ θ0.1 s.. 105 V.6 p.u.05 s. Δt = 0. 7.3 shows the results from the ﬁrst simulation and experiment. which causes the dc-link voltage to drop to vdc = 2.25 · 314 · 3/1. However. and requiring iq to be close to 1 p. eventually.u. and Pt = 3 p. Overcurrent The PWM rectiﬁer can be designed for overcurrent.72 ≈ 1.8 rad. 7. causing the dc-link voltage to increase to vdc = 3 p.8 p. resulting from the numerical example in Section 7. and αw = 0. The sag occurs in the time interval t = 0.2 mF. the grid voltage modulus is Eg ≈ 0.u. Moments after t = 0.3 except 99 . As seen from θ in Fig. V = 0.25).u.5 ≈ 310 p. the thermal fq limit of the utility grid may not be designed for such overcurrent..u.4 p.05 = −0. and solving the resulting fq expression for Cdc : 2Δt (7.. giving.u.u. and En = 0. the PLL recovers accurate ﬁeld orientation at approximately 40 ms after the “phase-angle jump” so the PWM rectiﬁer is hardly affected by the “jump.3b). especially if several WTs are connected to a common point. the maximum current modulus allowed is 1 pu.u.u.5 pu at vdc = 2.u. The dc-link capacitance is Cdc = 9. The base values are 85 A.2 Ω. The simulated and experimental waveforms are similar to those in Fig. and the PLL bandwidth is ρ = 0.u. The values ΔW = 0. The PWM rectiﬁer is loaded by a dc-link resistor which corresponds to Pt = −1.1–0. A symmetrical voltage sag occurs in the time interval t = 0.5 p.u.

4 Conclusion The voltage sag response of PWM rectiﬁers has been investigated for a candidate dc-link voltage control system. 15 ms after the load power step.2 0. e) Grid voltage. is superimposed on the dc-link voltage. which showed good agreement between analytical predictions and experimental results.1 0.] 0. the reference for the dc-link voltage changes stepwise from vdc = 2.5 p.u.2 0.u.5 vdc [p.3 0.4 e) Eq [p.u.5 2 0 0.2 0.4 1 0 1 0 g) ˜ θ1 [rad] 1 0 −1 0 0.u. The step response for t > 0.5 0 if q [p. a) DC-link voltage.1 0.1 ref s.1 0. These ripples are in close correspondence to the values predicted by the numerical example in the analysis section.3 0.] 0 0.u. b) DC-link voltage. ˜ ˜ Eq (experiment).8 p.u..4 3 2.2.4 Time [s] Time [s] Fig. and the ripple in the q-axis current is close to 0. The dc-link voltage reduces to 2.1 0.1.a) 3.5 p.u. the candidate dc-link voltage control system can successfully reduce disturbances from both symmetrical and unsymmetrical voltage sags such that nominal power 100 . as predicted in Section 7. d) Grid-ﬁlter current.2 0. 7. and 0. if q (experiment).1 0.2 0.] 0 −0.4 0 0.3 0. A method of analysis was derived.5 vdc [p. if q (simulation). θ1 (experiment).12 p. the load power is stepped from Pt = 0 to Pt = −1. Fig.3 0.3 s is well damped and the dc voltage rise time (10–90 % of the ﬁnal value) appears to equal the intended 35 ms.1 0.3 0.2 0.5 c) 0. vdc (experiment). that a ripple of approximately 0. 7.1 p.2 0.] 0 0. f) Grid voltage. c) Grid-ﬁlter current.] 3 2.3.2 s later.3 0.u.] 0 −0.2 0. Eq (simulation). h) PLL error angle.2.u.1 0.u.5 2 d) 0.5 if q [p.5 −1 −1.5 −1 −1.3 0.3 0.4 0 0.5 b) 3.4 f) Eq [p. vdc (simulation). In the last experiment. g) PLL error angle.u.u. θ1 (simulation). Response to “phase-angle jump” and symmetrical voltage sag.5 shows the results. For several types and magnitudes of voltage sags. at t = 0. ref to vdc = 3 p.1 0. 7.4 h) ˜ θ1 [rad] 1 0 −1 0 0.

5 if q [p.] 0 0. e) Grid voltage.3 0. if q (experiment). vdc (experiment).2 0. if q (simulation). g) PLL error angle.5 d) 0.2 vdc [p.5 if q [p.2 vdc [p.4 0.3 0.u.2 0. c) Grid-ﬁlter current.2 0.4 b) 3.5 b) 3.4 h) ˜ θ1 [rad] 0.3 0.4 0 0. vdc (experiment).1 0.3 0.] 0 −0.3 0.4 Time [s] Time [s] Fig.2 0.5 0 0.u.3 0.4 c) 0.5 0 if q [p.4 e) 1. a) DC-link voltage. c) Grid-ﬁlter current. 101 .8 2.5 if q [p.u. h) PLL error angle.6 2. a) DC-link voltage. d) Grid-ﬁlter current.1 0.1 0.5 −1 −1.5 −1 −1.5 d) 0.u. Eq (simulation).5 0 Eq [p.4 0 0.3 0.] 0 −0.u.1 0.4 f) 1.3 0.2 0.5 0 Eq [p.1 0. if q (simulation).5 c) 0.3 0. if q (experiment).1 0.2 0.1 0.2 0.u.4 0 0.6 2. 7.] 3 2.a) 3.4 Time [s] Time [s] ref Fig.3 0.] 0 −0. vdc (simulation).5 1 0 −1 0 0 0.2 0.2 0.1 0.u.5 1 0 −1 0 0.4 0 0.u.] 0 −0.4 g) ˜ θ1 [rad] 0. d) Grid-ﬁlter current. θ1 (experiment). Response to “phase-angle jump” and LLF unsymmetrical voltage sag.5 0. f) Grid ˜ ˜ voltage.5 −1 −1. θ1 (simulation). 7.5 vdc [p.4 3 2.2 0.] 3 2.] 1 0. Eq (experiment).5 0 −0.] 3 2.1 0.1 0.8 2. Steps in Ps and vdc .2 0.u.2 0.5 0 −0.4.1 0.5 −1 −1.1 0.u. a) 3. b) DC-link voltage. vdc (simulation).] 1 0.3 0.3 0.5 vdc [p. b) DC-link voltage.5.

large reductions in the positive-sequence voltage were found to be critical. However.6 shows a principle sketch of the DFIG. rotor current control has been lost and the turbine must be disconnected from the grid. Unless suitable actions are taken. Remedies for avoiding overvoltage at the dc link have also been discussed. this means that only the IGBT modules need to be designed for a higher current while the rotor winding and the converter (cooling etc) still can be designed according to the slip power only. The crowbar in DFIG ≈ = = ≈ Grid Crowbar Fig. since the current through the thyristor is a continuous dc current and can only be interrupted if the turbine is disconnected from the grid [72].6. 102 . Assuming such a short over-current time. However. After such an action. Since the relatively low power losses in the power electronic equipment were a major reason for selecting a DFIG.production can be restored once the grid voltage recovers. Fig. 7. 7. one possibility is to still have a rotor converter.3 Doubly-Fed Induction Generator with Shunt Converter Fig. which short-circuits the rotor circuit in case of too large a grid disturbance causing high rotor current. Doubly-fed induction generator system with a crowbar. One disadvantage with this system is that once the crowbar has been triggered. it is accordingly important to study how the ride-through system inﬂuences the power losses. the turbine must be disconnected from the grid. This means that the converter shortly can handle a higher current and thereby stay connected to the grid longer without any crowbar action. and thereby protects the rotor converter. Still. we will look further into the dynamics of the DFIG during a voltage sag. Before explaining the candidate DFIG ride-through system. 7. since additional hardware or modiﬁcations may reduce the efﬁciency. such a voltage reduction sag may result in a dc-link overvoltage since the transferable active power reduces with reducing grid voltage. In the ﬁgure a crowbar is also depicted. but one that can handle a higher current for a short period of time of some 100s of ms.6 consists of a diode rectiﬁer and a thyristor that is triggered when the rotor circuit should be short circuited. 7. this system will have high fault currents from the stator during the voltage sag.

nom ψs0 e−Rs t/LM e−jωg t (7. 7. no crowbar action. However. the ﬂux will circularly approach the “new” equilibrium point (the circle) very slowly. the ﬂux will again approach circularly the equilibrium point (the cross) indicated with the solid line. the ﬂux dynamics of the DFIG are poorly damped. It can be seen in the ﬁgure that directly after the voltage sag has occurred. the voltage returns when the ﬂux is close to the original equilibrium point (the cross). The amplitude of the oscillation will be proportional to the size of the voltage sag. as shown in Fig.7. when the voltage returns. which can be realized from the fact that ψs ≈ Eg /ωg .. i.e. i. the oscillations become even worse when the voltage returns after the sag. if the voltage returns at an unfortunate moment. 7.42) V Eg.nom . with the duration of the sag as the only difference.7a) and c). However.7b) and d). The reason that the amplitude of the stator ﬂux oscillation can almost vary between zero and twice the initial amplitude is that steady-state condition hardly precedes the returning of the voltage. we will assume that the converter is ideal an can supply the desired rotor voltage and current.7b).1 Response to Small Voltage Sags In order to explain what happens to the DFIG during and after a voltage sag. and steady-state conditions are assumed to precede the voltage sag. This is indicated by the dashed lines in the ﬁgure.47) and (5. As also previously shown a voltage sag will cause the stator ﬂux to enter a poorly damped oscillation with an oscillating frequency close to the line frequency.43) Eg.7a) and Fig.3. In the ﬁgure. two different voltage sags are shown. 7. the duration of the voltage sags in the time series is two periods longer. which leads to that the ﬂux oscillations after the voltage sag are relatively small. If the DFIG system survives the voltage sag. the duration of the sag is important. Note. as indicated by the difference between Fig. when the voltage returns. In order to analyze the response to voltage sags we will assume that the magnitude of the stator ﬂux can be expressed in a similar way as in (5. Symmetrical Voltage Sags In this section. 7. Then.e.2. it will be assumed that the “disturbance” is applied at t = 0 and that steady-state conditions precede the fault causing the voltage sag. In the analysis below.7. when the ﬂux is far away from the original equilibrium point (the cross). we will start by looking at the ﬂux dynamics.nom ωg V + 1− 103 t<0: t≥0: Ψs (t) = ψs0 ≈ Ψs (t) ≈ ψs0 (7. the amplitude of the ﬂux oscillations can. where a phase portrait and corresponding time series of the ﬂux can be seen. 7. As discussed several times before.. First. the dc-link dynamics of the DFIG system will be analyzed in a similar way as for the full-power converter system in Section 7. This can be realized from Fig. vary between zero and close to twice the ﬂux oscillations in the beginning of the sag. In Figs. after the voltage sag. The cross marks the equilibrium point during normal operation and the circle marks the equilibrium point during the voltage sag. that in order to make the ﬁgure more lucid.48) as Eg.

7.25 −0.04 0.02 0.5 ms long voltage sag.nom ωg V pk + ψs e−Rs t/LM e−jωg t (7.25 1.nom pk where ψs = (1 − V /Eg.a) 0.5 0.5 0. For clarity of the ﬁgure.46) (7. d) Time series of the ﬂux in b). a) 18.75 1 1. The expression for the stator ﬂux in (7. and are governed both by the rotor power Pr and the grid-ﬁlter power Pf . which can be approximated as Pr ≈ 3Egq iRq − 3ωr Re[jΨs i∗ ] ≈ 3Egq iRq − 3ψs iRq ωr R Pf ≈ 3Egq if q . the voltage sag duration for the time series is two periods longer.5 0.5 −0.5 1 ψsd d) 2 1.02 0.25 −0.1).06 0.5 0 −0.1 Time [s] Time [s] Fig. ψsq 1 0.08 0.75 1 1.25 b) 0.25 ψsq 0 ψsq 0. ψsq ψsd .47) .nom )ψs0 is the peak value of the stator ﬂux oscillation.5 0 q q −0.5 0 0.48).43) and (7. b) 10 ms long voltage sag.5 d ψsd . c) Time series of the ﬂux in a).45) can be found by solving the differential equation in (6.5 0 −0. Phase portrait and time series of the ﬂux dynamics during a symmetrical voltage sag.06 0. The dynamics of the dc-link are described by (4. or as t<0: t≥0: Ψs (t) = ψs0 ≈ Ψs (t) ≈ ψs0 Eg.7.45) Eg. 104 (7.08 0.5 d 0.5 ψsd c) 1.5 0.25 1.44) (7.1 0 0.5 0.04 0.

with GP e (p) = −GP W (p). yielding Prpk = 3ψs iR ωr = 105 . 1− ωg (7.52) will act as a disturbance to the dc-link. This means pk pk ˜ that ψs = (1 − V /Eg. the second term in (7. This means that the power drop in the ﬁrst term of (7. (7.53) The dc-link dynamics in (4.120). where the stator voltage has been changed to the grid voltage.nom .25 p.52) (7..53). since steady-state conditions are assumed to precede the sag.u. Then. the above expression can be further reduced as t≥0: t≥0: Pr (t) ≈ 3V Pf (t) ≈ −3V 1− ωr pk iRq − 3ψs iR ωr e−Rs t/LM sin(ωg t + φr ) ωg ωr iRq .75) · 1 = 0.55) ξ)2 αw .49) Moreover.nom )ψs0 = (1 − 0. If ωg 2ωg 2 (αw ξ − 2 ωg )2 + (2ωg αw (7.50) t≥0: Pf (t) ≈ 3V if q ≈ −3V 1− ωr ωg iRq (7. (7.e.nom pk iRq ωr ψs0 + ψs iR ωr e−Rs t/LM sin(ωg t + φr ) (7.75 p. However. vR ≈ Egq − jωr Ψs . at the relevant frequency can be used.48) are governed by the term −Pr − Pf . Pr and Pf equal Pr (t = 0− ) ≈ 3Egq iRq − 3ψs0 iRq ωr Pf (t = 0− ) ≈ 3Egq if q .0 iRq ωr )/Egq ≈ −(1 − ωr /ωg )iRq . giving if q (t = 0− ) = −(Egq iRq − ψs.u. In order to determine the amplitude of the ripple the static gain of (4.The expression for Pr is derived from the fact that Pr = Re[3vR i∗ ] and by using the approxR imation of the rotor voltage given by (5. Cdc ωg (7.56) For example..u.5 p. This means that just before the voltage sag. it is possible to express Pr and Pf as t≥0: Pr (t) ≈ 3V iRq − 3 V Eg. as the stator ﬂux prior to the voltage sag can be approximated as ψs0 ≈ Eg.52) is compensated for by the same drop in Pf as can be seen in (7.54) This disturbance will cause a ripple in the dc-link voltage with the frequency ωg . as pk ˜ ˜ Pr = 3ψs iR ωr e−Rs t/LM sin(ωg t + φr ) = Prpk e−Rs t/LM sin(ωg t + φr ). and Cdc = 3. moments after the sag has occurred.51) where iR = |iR | and φr = ∠iR .. the above expression can be further approximated as |GP e (jωg )| ≈ 2 . This yields |GP e (jωg )| = Cdc where ξ = V /Eg. i. we have that Pr (t = 0− ) = −Pf (t = 0− ). Under the assumption that the rotor current controller and grid-ﬁlter controller manage to keep the current at (or at least close to) its reference value.48) (7.nom /ωg . consider a voltage sag with V = 0.38).

98/(3 · 1 · 1) = 0. when ˜pk ref vdc = 2.8 p.u. The corresponding peak value of the ripple in vdc is vdc = 2.u.4 0.3 0.2. Since there is ripple in the dc-link voltage.2 0. d) DC-link voltage.7 2.8. c) Sum of rotor and grid-ﬁlter power.5 Time [s] Time [s] Fig.5 b) 1.9 2.8 2. a corresponding simulation is shown. 7. a) Grid voltage.82 − 0.3 = 0.4 0.] 1 0. the amplitude of the stator-ﬂux oscillations can be close to twice the value at the beginning of the sag.25 0 0 0.75 0.] 1 0.] 0.] 1 0.2 0.1 0.5 3 2.u.8 − (2.4 0.2 · 0.u.25 · 1 · 1 · 1. this will also be transferred to if q .1 0.5 · 1) = 0.3 0..5 0.39) the static gain of the ripple in if q can be calculated (note that −GP i is actually used since here Pt = −Pr ). In Fig. b) Stator ﬂux.3 0.5 0 0. that the amplitude in the dc-link voltage ripple will be increased accordingly.98 p. 7.u.u.u.25 0 0 0.a) 1.1 0.75 0. Moreover.nom 2 2 αw + 4ωg 2αw ≈ 2 ξ − ω 2 )2 + (2ω α ξ )2 (αw p 3ωg Eg.5 = 0.nom ) = 2 · 0.u. according to (7. which is according to the analytical result.5 0 −0. of course.75 if q [p. This yields |GP i (jωg )| = αw 3Eg.1 0.25 Eg [p.2 0.4 0.2 0. with pk ˜rpk /(3ωg Eg.2 0. that ˜f q = 2αw P ı 106 .5 q d c) 100 Pr + Pf [%] d) vdc [p. The simulation veriﬁes the ﬁnding that the sum of the rotor and grid-ﬁlter powers consists of a corresponding oscillating power at a frequency of ωg .] 0 0.u.56). This value αw = 0.5 0. Simulation of the response of the DFIG system to a small symmetrical voltage sag. With the transfer function in (7.nom g w p g (7.1 p. This means. However. 3 · 0.1 0. For the example given above we have. since it is used for controlling the dc-link voltage.13 p.5 Ψs [p. the ripple in the dc-link voltage is 0.25 0 0 0. the peak ripple epk in the error ˜w ref ˜ ˜w signal ew = Wdc − Wdc will be epk = 2Prpk /(Cdc ωg ) = 2 · 0. e) qcomponent rotor current.u.56 p.98/(3.4 0.8.5 0. f) q-component grid-ﬁlter current.5 0 −100 e) iRq [p.5 0 0.6 2. as previously discussed.3 0.5 f) 0. at the time when the voltage returns.2 0.u.4 0.1 0.3 0. Then.3 0.57) where the approximation holds if ωg αw .1 p.56)0.

46) and (7.u. LM (7. This means that the power drop in the ﬁrst term of (7.nom ˜ −ρt Ψs (t) ≈ θ0 e − e−Rs t/LM e−jωg t . with LM = 4.nom 1 + j θ(t) ≈ jEg. ıs “Phase-Angle Jumps” In this section we will study how the DFIG system responds to small “phase-angle jumps.25/4. the rotor and grid ﬁlter powers can be determined in a similar way as for the symmetrical voltage sag as Pr (t) ≈ 3Eg. If ωg ρ.65) As for the case with symmetrical voltage sags.64) t≥0: (7.nom ωg (7. iR = iref .05 p.” i.30).nom 1 + j θ0 e−ρt (7. the ripple in the stator current R pk ıs will be ˜pk = ψs /LM .nom 1 − ωr iRq ωg ωr ˜ + 3Eg.62) (7.e.6 p.” Moreover. the dc-link dynamics in (4.8. the stator-ﬂux oscillations will also cause a ripple in the stator current.59) ˜ ˜ where θ(t) is the error angle and the approximation holds if θ(t) is small.64) is compensated 107 .60) 2 2 ˜ ˜ Eg.1) and solving the differential equation. yields ˜pk = 0.58) Then if the rotor current is controlled accurately. it is possible to further approximate the above equation as t<0: t≥0: Ψs (t) = ψs0 ≈ Eg. the grid voltage vector can be expressed as ˜ ˜ ˜ Eg = jEg.59) the error ˜ angle θ(t) is modeled as in (7. +j ωg ωg (7. i. which. Substituting (7. After a pure “phase-angle jump..47). Moreover. without any voltage sag.nom /ωg . In (7.. in the analysis it will be assumed that the magnitude of the grid voltage remains at its nominal value after the “phase-angle jump.6 = 0. the following solution is obtained t<0: t≥0: Ψs (t) = ψs0 ≈ Ψs (t) ≈ Eg.nom Eg.63) Using (7.59) in (6. 7.nom θ0 iRd e−ρt − iR e−Rs t/LM cos(ωg t + φr ) ωg ωr iRq .nom ej θ(t) ≈ jEg.nom 1 − ωg t≥0: (7.” This has been done in order to study the effect of the actual “phase-angle jump” and not the inﬂuence of a voltage sag.61) if the stator resistance in the solution is assumed to be zero—except in e−Rs t/LM —and ψs0 ≈ Eg.e. The stator current can be found from (4. Pf (t) ≈ − 3Eg.nom ωg Eg.u.48) are governed by the term −Pr − Pf .nom ωg + ρ2 + (1 + j)ωg ρθ0 e−ρt − ωg ρ + jωg θ0 e−(Rs /LM +jωg )t 2 ωg ωg + ρ2 (7.is also conﬁrmed by the simulation shown in Fig.40) as is = Ψs − iR .

58 p. 7.26 rad.u..u.u.11 = 2. (7.u. In Fig.u.5 p.68) ωg Cdc αw ωg Cdc αw if ρ = αw .67) becomes √ √ ωr 2(−1 + 2)e−2+ 2 0.66) will cause ripple in the dc-link voltage with the static gain according to (7.u.67) αw t ˜0 ωr iRd (αw t − 2)te = 3Eg. and iRd = 0.5 = 0. the second term in (7. 7.5 = 2. the result should be used with care since adding the results will not. pk ı the oscillation in if q becomes ˜f q = 2 · 0.for by the same drop in Pf .u.66) Pr = −3Eg. which will cause a ripple in the dc-link voltage of vdc = 2.nom θ0 iRd ≈ 3Eg.8 − (2.2) = −0.9 shows a simulation of the “phase-angle jump” used in the example.34·0.u.3·0. The extreme value of (7.57) can be used to determine the ˜ amplitude of the oscillation in if q . Eq. It can be seen that the amplitude of the oscillation in the dc-link voltage and the maximum value of the dc-link voltage is close to the predicted values. The ˜ amplitude of the second term in (7.25 p.66). wr = 1.26)·1. as for symmetrical voltage sags and “phase-angle jumps.26)·1. The ﬁrst term’s impact on the dc-link dynamics can be found from the extreme value of ˜ ωr ew (t) = L−1 GP e (p)L −3Eg.01 · 2/3.2 · 1.65).. One way of estimating the “worst case” impact of a speciﬁc “phase-angle jump” is to treat the two terms independently and then add them together.95 p. Therefore. and a “worst case” dc-link voltage of vdc = 3.” for unsymmetrical sags is more difﬁcult to derive.nom θ0 ωg In (7.nom θ0 iRd (7.23 p.01 p.67) occurs for t = (2 − 2)/αw if ρ = αw .nom θ ωg Cdc √ where GP e (p) is given by (7. ˜ (7.3·1 = 1. 7.5 = 0. iR = 1 p.10.nom θ0 ωr /ωg iR = −3·(−0.01 p.66) becomes −3Eg.01/(3 · 1 · 1) = 0.13 p. ˜pk wc This means that the “worst case” dc-link voltage could be vdc = 2...58)0.11 p.56).u. the analysis here will be limited to simulations. However.u..2 p.68) we have that ew = 3·(−0. since the system also will be excited with the negativesequence voltage. However.5·0.64) will act as a disturbance to the dc link. For comparison to larger “phase-angle jumps” a corresponding simulation of a −45◦ “phase-angle jump” can be seen in Fig. see (7..11 108 . generally.46/(3. giving according to (7.34 (corresponding to unity power factor) are used for a numerical max/min example.21). ˜ The values θ0 = −15◦ ≈ −0. Cdc = 3. In Fig. αw = ρ = 0.u.33 p. Of course. the analysis will still give some valuable information of the system.3 p. the disturbance consists of two terms: one that depends on the bandwidth. It is difﬁcult to use the disturbance in (7.46 max/min ˜ ˜ ωr ew = 3Eg. ρ.82 − 0.66) in order to ﬁnd the extreme value in the error signal ew since it consists of two terms of which one is sinusoidal.nom θ0 iRd e−ρt ωg (7.82 − (−0. The second term in (7. With the same analysis as above will give a wc dc-link voltage ripple of 0. This means that the extreme value of (7. Unsymmetrical Voltage Sags Similar analysis. of the PLL-type estimator and one that depends on the stator ﬂux dynamics.84 + 0.56) a ripple with the amplitude 1. From (7.u. Since the amplitude of the oscillation in Pr is 1.u.u.23))0.84 p.. which corresponds to a dc-link voltage of vdc = (2. as ˜ ωr iRd e−ρt − iR e−Rs t/LM cos(ωg t + φr ) . give mathematically correct results.

The oscillation with the frequency ωg arises from the ﬂux dynamics while the oscillation with the frequency 2ωg arises from the negative-sequence voltage.25 0 0 q b) Ψs [p.2 0. 7.4 0.4 0.1 0.3 0.75 0. b) Stator ﬂux.1 0.1 0.5 if q [p.12 where the sag occurs at t = 0. e) q-component rotor current.1 0.8 −50 0 0. In the ﬁgure it can be seen that in.5 0.25 0 0 0. is presented.] 1 0. Summary The response of the DFIG system due to different grid disturbances has been investigated. Simulation of the response of the DFIG system to a small “phase-angle jump” of −15◦ . However.3 0.3 0. for instance.2 0.2 0.3 0. and in the grid-ﬁlter current oscillations with both frequencies of ωg and 2ωg .5 f) 0.] 0. all other conditions are as in Fig.2 0. the response to an SLGF occurring at t = 0.4 0. This is indicated in Fig.4 0.1 ms with V = 0. the dc-link voltage. It has been shown that the amplitude of the ﬂux oscillation when the voltage returns after a voltage sag can vary between zero and twice the initial amplitude of the ﬂux oscillations due the sag.u. the ﬂux.u.5 0.] 2 1 0 −1 q d d 0.1 0.5 2. 7. the DFIG system is roughly as sensitive to “phase-angle jumps” as to symmetrical voltage sags.u.4 0.5 c) 50 Pr + Pf [%] d) vdc [p.a) 1.] 1 0.5 Time [s] Time [s] Fig. Moreover. 7.] 0 3 0.6 0 0. depending on the phase angle at the time instance of the sag the oscillation at ωg can in principle be removed for an SLGF.4 0. c) Sum of rotor and grid-ﬁlter powers. However.5 e) iRq [p. Moreover.5 0 2. a) Grid voltage. d) DC-link voltage.25 Eg [p.u.3 0.9.3 0.105 ms.2 0.u.75 0. f) q-component grid-ﬁlter current. 109 .25 0 0 0.1 0. the response to “phase-angle jumps” and unsymmetrical voltage sags are analytically harder to derive.2 0. the DFIG system has been analyzed for symmetrical voltage sags with good agreement.11.u.75 p.

u.3 0.4 0. this cannot be assumed for larger voltage sags. a) Grid voltage.5 0 0. This has been done in order to study the behavior of the DFIG and not the inﬂuence of the converter and the crowbar.5 vdc [p. Still.3 p. anyhow. not causing the rotor converter to fail in controlling the rotor current. the rotor voltage will hit its maximum value and lose control of the rotor current.4 0. i. e) qcomponent rotor current.13.2 0. b) Stator ﬂux. However.5 c) 100 Pr + Pf [%] 0 −100 −200 0 0. limited to ±0. (referred to the stator circuit).1 0.1 0.u.5 e) iRq [p. In Fig.] 1 0.u. with two identical simulations 110 .4 0.2 0.1 0.1 0.10. 7.e.75 0.4 p.25 p. As shown in (5. for larger voltage sags.4 0.3 0.2 0. So.1 0.3 0. 7.2 0. In the ﬁgure it can be seen that the rotor voltage will hit its maximum value directly after the voltage sag.5 0.40). The duration of the voltage sag in the simulation is 102 ms and 92 ms.5 0. c) Sum of rotor and grid-ﬁlter powers.u.4 0.4 0.3 0.5 0 −0.] 1 0. d) DC-link voltage.75 0.1 the situation might be even worse when the voltage returns.u.. f) q-component grid-ﬁlter current.5 −2 0 0. the DFIG system will be analyzed and it is assumed that the converter is large enough to handle excess currents. This is also indicated in Fig. This means that the converter loses control of the rotor current.25 0 0 q b) Ψs [p.1 0.5 Time [s] Time [s] Fig. As shown in Section 7. This limitation of the rotor voltage is a major difference compared to the analysis in the previous section.u.2 0.3 0. Simulation of the response of the DFIG system to a “phase-angle jump” of −45◦ .a) 1.5 d) 3.13 a simulation of a symmetrical voltage sag (at 0.25 0 0 0.] 2 q 0 d d 0.5 f) if q [p. the rotor voltage will change in proportion to the depth of the voltage sag. is presented.5 2 0 0. the rotor voltage in the simulations is. leading to an uncontrolled rotor current.25 Eg [p.] 1 0. since the converter will lose control of the rotor current.u.2 0.3.3.2 Response to Large Voltage Sags In previous section the voltage sags were assumed to be small enough. In this section.u.] 3 2. 7. Before the voltage sag the DFIG is running at rated power and a rotor speed of 1.3 0. 7.05 s) down to 0.

2 0.25 0 0 0. In the ﬁgure.] 0 0.1 0.] q d d c) 50 Pr + Pf [%] d) vdc [p. Based on these ﬁndings a candidate ride-through system will be presented in the next section. 7.u.3 0.u.3 0.4 0.5 0. the maximum rotor current and the maximum rotor power due to a symmetrical voltage sag for three different operating conditions can be seen. d) DC-link voltage. e) q-component rotor current.4 0. except the duration of the voltage sag.2 0 0 0. It should be kept in mind that for the ordinary DFIG system.] 0.8 2. c) Sum of rotor and grid-ﬁlter powers.] 0.3 0.3 0.5 2 1 0 −1 0 0.75 0.2 0.25 0 −0.5 0.25 iRq [p.25 0 q b) Ψs [p. 7. b) Stator ﬂux.5 3 2.4 if q [p. if the machine ﬂux is in the wrong “direction. From the ﬁgure it can be seen that the maximum rotor current due to a voltage sag will increase with the magnitude of the sag.5 f) 0.3 0. 7. It can be seen that the maximum rotor current can be much higher when the voltage returns than when the voltage drops at the beginning of the disturbance. However.75 0.” In Fig.3 Candidate Ride-Through System The aim of this section is to present a candidate ride-through DFIG system.5 Eg [p. 7. as indicated in Fig. the effect of returning voltage has not been taken into consideration.13.4 0.1 0.2 0.1 0. This means that there is a huge rotor power that needs to be dealt with.4 0. Moreover. a) Grid voltage. f) q-component grid-ﬁlter current.14.4 0. the maximum rotor current actually might be much 111 .5 0 −50 e) 1.3.4 0.11. based on the result in the previous section.1 0. The main idea is to overdimension the valves of the power electronic converter so that they can handle the rotor current occurring at deep voltage sags.u.u.9 2.] 1 0.25 1 0.2 0.2 0. Simulation of the response of the DFIG system to an unsymmetrical (SLGF) voltage sag.2 0. the maximum rotor power that is fed into the dc link can be up to almost 250% of nominal power.a) 1.5 Time [s] Time [s] Fig.3 0.1 0. the converter and dc link are only rated for 30–40% of the nominal power.7 0 0.u.1 0.

Instead of using the 112 .25 0 0 0. In order to avoid these high current when the voltage returns..3 0.1 0. the ﬂux oscillation will also be interrupted.2 0. Moreover.] 0.2 0. higher. when the voltage returns than at the voltage drop. if required.3 0.5 0.25 0 −0.85 2.75 0 0. i. in 10 ms. c) Sum of rotor and grid-ﬁlter powers.1 0.1 0. a) Grid voltage.u. anti-parallel thyristors can be connected in series with the stator in order to achieve a quick disconnection of the stator circuit [20]. A third option would be to have anti-parallel IGBTs. As soon as the ﬂux is interrupted it is possible to remagnetize the DFIG quickly through the rotor converter and connect the stator circuit to the grid again. a so-called punch-through IGBT.u.4 0. then.3 0.u. a diode has to be put in series with the IGBT.4 0. Another option would be to have gate-turn-off thyristors. cannot handle reverse voltages as high as the forward blocking voltages.] 0 0. In order remove the excess power that is fed into the dc link.2 0.5 Time [s] Time [s] Fig.] 0.25 0 q b) Ψs [p. a “dc-link breaking chopper” is used to dissipate the excess power.] q d d c) 50 Pr + Pf [%] d) 2.4 0.12. f) q-component grid-ﬁlter current. since typically a large negative gate current is required to turn off that device [67].25 iRq [p. up to twice as high. the grid-side converter may provide the grid with reactive power during the sag. However. i.5 0. this means that the valves have to be even more overdimensioned.a) 1.e.4 0.2 0.u.3 0. The anti-parallel thyristor switch can disconnect the stator within a half cycle.35 if q [p. Of course.5 2 1 0 −1 0 0. Since a normal IGBT. By interrupting the stator circuit. Simulation of the response of the DFIG system to an unsymmetrical (SLGF) voltage sag.9 vdc [p.5 2. 20].1 0. a complex driving circuit is needed.5 Eg [p.2 0.75 0. b) Stator ﬂux. The anti-parallel thyristor switch needs to be equipped with a forced commutation unit in case of a dc component in the stator fault current [20].5 0 −50 e) 1..] 1 0. The system with anti-parallel thyristors is illustrated in Fig.3 0. 7.8 2.3 0. The converter needs not to be disconnected from the rotor circuit since the valves of the converter are overdimensioned.2 0.u.75 0. 7.4 0.4 0.15. [9.e.1 0.25 0 0.3 0.5 f) 0. e) q-component rotor current.25 1 0. d) DC-link voltage.1 0. the disconnection time can be lowered.

5 0 0 0.u.] 1 0.1 0. a) Stator voltage.1 0.17 shows the maximum rotor current and rotor power when the stator is not disconnected from the grid and the grid voltage returns at the worst instance around 50 ms.3 f) Pr [p. 7. in order to avoid the possible higher rotor currents when the voltage returns.u.u. 7.3 b) vR [p.2 0. The non-punch through IGBT can handle a reverse voltage as high as the forward blocking voltage.a) 1. the stator circuit can be disconnected less often due to voltage sags. d) Rotor current magnitude. the disconnection time for the contactor will be longer. f) Rotor power.] 2 0 −2 −4 0 0.5 Eg [p. Simulation of the response of a DFIG system to a voltage sag down to 0. However. after a disconnection. the so-called non-punch through IGBT can be used.] 0 0.1 0.1 0. Then.1 0. this device has higher on-state losses [67]. Solid line correspond to a voltage sag of 102 ms and dashed line correspond to a voltage sag of 92 ms. 7.13.2 0. the stator circuit is disconnected from the grid.3 c) 10 is [p.] d) 10 iR [p.2 0.2 0. However.3 Time [s] Time [s] Fig. This means that the duration of the voltage sag is approximately 50 ms. for IGBT modules that can handle higher currents temporarily. and then the DFIG can be synchronized to the grid as soon as the voltage has returned to an acceptable. punch-through IGBT. It can be seen in the ﬁgure that for voltage sags down to 113 . in principle. c) Stator current magnitude.3 e) ψs [p. if the stator circuit is disconnected within 10 ms. the IGBT modules of the machine-side converter are designed for a higher current rating. e) Stator ﬂux magnitude. while on the other hand. Fig. in order to withstand voltage sags. without losses. Moreover.5 0 0 0.u.u. For the investigated system.25 p. the maximum rotor current and rotor power due to symmetrical voltage sag is shown in Fig. b) Rotor voltage magnitude. Of course.1 0.] 1 0. Thus.2 0. level. the rotor current controller re-magnetizes the DFIG. predeﬁned.u. One last option would be to have a contactor as a circuit breaker.3 5 5 0 0 0 0.] 3 2 1 0 0 0.2 0.16.u.

] Voltage sag V [p. Switch Δ Y ≈ = = ≈ DFIG Grid-side Machine-side converter converter “Breaking chopper” Fig.4 Evaluation of the Ride-Through System The aim of this section is to make a theoretical case study on the candidate voltage sag ride-through system presented in the previous section.5 1 Voltage sag V [p.14. 7. the maximum rotor power fed to the dc link becomes higher compared to when the stator circuit is disconnected. 0.3.u.15.u.7 p.0 p. b) Maximum rotor power (Note that negative rotor power means that the power is fed into the dc link).] Fig.u. However. 114 . DFIG with anti-parallel thyristors in the stator circuit. This study will focus on the energy production and energy production cost of such a system for a 2-MW DFIG WT.u. the maximum rotor current is approximately as large as the maximum rotor current when the stator is disconnected.u. Case 3 corresponds to 11% of rated power and a rotor speed of 0.] 3 Case 3 Max Pr [%] Case 2 −50 −100 −150 −200 −250 Case 3 Case 2 Case 1 2 1 0 0 0. Maximum rotor current and rotor power for three different operating conditions. Case 2 corresponds to 23% of rated power and a rotor speed of 1.a) 5 Case 1 b) 50 0 4 Max iR [p.u. 7.5 1 0 0.5 p.3 p. 7. Case 1 corresponds to rated power and a rotor speed of 1. a) Maximum rotor current.u.

6 0.u.a) 5 Case 1 b) 50 0 4 Max iR [p. generator and in the semiconductor devices.7 p.16.4 0.e.u.u.u.] 3 Case 3 Max Pr [%] Case 2 −50 −100 −150 −200 −250 Case 3 Case 2 Case 1 2 1 0 0 0. gearbox.] Fig.8 1 0. a) 8 Case 1 b) 0 Max iR [p.15.5 1 0 0. Case 3 corresponds to 11% of rated power and a rotor speed of 0. Case 1 corresponds to rated power and a rotor speed of 1. i.u.5 1 Voltage sag V [p. Case 1 corresponds to rated power and a rotor speed of 1.] Voltage sag V [p. Case 3 corresponds to 11% of rated power and a rotor speed of 0.8 1 Voltage sag V [p. Case 2 corresponds to 23% of rated power and a rotor speed of 1.u.] Fig.0 p. Calculation of Power Losses The losses taken into consideration are the losses in the aerodynamic conversion.. The losses in the anti-parallel thyristor switch used in the stator circuit. 7.17.3 p.4 Case 2 Case 1 0 0.u.] Voltage sag V [p. b) Maximum rotor power (note that negative rotor power means that the power is fed into the dc link). Maximum rotor current and rotor power fed to the dc link for three different operating conditions with returning voltage at the worst instance around 50 ms. The calculation of the will also follow the loss models used in Chapter 3. Maximum rotor current and rotor power for three different operating conditions if the stator circuit is disconnected within 10 ms.u.6 0.u.3 p. 7.7 p. Case 2 corresponds to 23% of rated power and a rotor speed of 1. the same losses as in Chapter 3. a) Maximum rotor current. see Fig. a) Maximum rotor current.u.u. b) Maximum rotor power (note that negative rotor power means that the power is fed into the dc link).0 p.] 6 −100 4 Case 3 Max Pr [%] Case 2 −200 Case 3 2 −300 −400 0. 7. can 115 .u.

8 m/s are due to that the stator of the generator is switch from a Y connection to a Δ connection. Since the switching occurs at zero current and at low frequency (the grid frequency). In Fig.35 Ordinary DFIG 0. hence.18.3. In Fig. the switching losses in the switch will be neglected.18 the losses in the semiconductor devices are shown. as discussed in Section 2. The dashed line shows the losses for the GSC and the dotted line shows the losses in the anti-parallel thyristor switch. it will not inﬂuence the energy production.164 mΩ and VThy = 0.25 Losses [%] 0. the energy production cost of the candidate system is higher than for the ordinary DFIG system. The solid lines show the losses in the MSC for ordinary and the candidate DFIG system.05 Grid-side converter 0 5 10 15 20 25 Wind speed [m/s] Fig.18. Energy Cost For the calculation of the energy production cost. The energy production cost 116 .1 for details of calculation of the expected efﬁciency.u. As could be expected.be described by (3. In the ﬁgure it can be seen that the ordinary DFIG system has the highest efﬁciency. even though the difference towards the candidate DFIG system is relatively small. 0.2 0. The dc-link “chopper” is not used during normal operation. This reduction is only due to that the resistance in the valves decreases with an increasing current rating. The losses in the semiconductor devices used in this work.4. It can be noticed in Fig.05 percentage units at rated operation by increasing the current rating of the valves. The cost of the IGBTs is an estimate based on cost information obtain from some IGBT manufactures. that the losses of the MSC can be reduced by approximately 0.3 0. 7. 7. current. The parameters for the thyristors used here are rThy = 0. it has been assumed that the standard 2-MW DFIG WT costs e1600000 [65] and that one IGBT converter and the anti-parallel thyristor switch costs e6000/p.88 V [93].1. In the ﬁgure expected efﬁciencies of the ordinary DFIG system as well as a system that utilizes a full power electronic converter can also be seen. See Section 2.20. the losses of the thyristor switch are much larger than the reduction of losses due to the increased current rating of the valves. the relative energy production cost of the candidate system normalized with the energy production cost of the ordinary DFIG system can be seen.4). In Fig. The steps in the curves at 7.19 the expected efﬁciency of the candidate DFIG system as a function of the average wind speed is presented. 7. However.1 0. 7. 7.15 Candidate DFIG Thyristor switch 0.

The increase in energy production cost is due to the lower energy production of the candidate system.e. i. is also shown.01 Candidate DFIG system without thyristor switch 1 5 6 7 8 9 10 Average wind speed [m/s] Fig. 7.02 1.03 Candidate DFIG system 1. and a system with a full-power converter. which handles the total power. the ordinary DFIG system. 117 .96 Ordinary DFIG system 95 Efﬁciency [%] 94 Candidate DFIG system 93 92 Full power converter system 91 5 6 7 8 9 10 Average wind speed [m/s] Fig. In Fig.19.04 1.05 Full power converter system Relative energy cost [%] 1. has approximately three percentage units higher energy cost than the ordinary DFIG system. but it is mainly due to the increased cost of the valves. 7. of the candidate system is approximately two percentage units higher then for the ordinary DFIG system. A full power electronic conversion.. and the system that utilizes a full-power converter. 1. without anti-parallel thyristor switch in the stator circuit. The energy cost is related to the ordinary DFIG system.20 the normalized energy production cost of a modiﬁed candidate system. 7. Energy cost of the candidate DFIG. candidate DFIG without thyristor switch.20. Expected efﬁciency as a function of the average wind speed for the candidate DFIG system.

Conclusion The inﬂuence on the energy production of a DFIG ride-through system has been investigated. This system is based on increased current rating of the converter and anti-parallel thyristors in the stator circuit. It has been found that the increased cost for a ride-through system for a DFIG turbine might be reasonable, in comparison to the cost of full-power converter system connected to a cage-bar induction generator.

7.4

Doubly-Fed Induction Generator with Series Converter

After a voltage sag, the stator ﬂux of the DFIG will start to oscillate. This oscillation often causes very high rotor currents, which necessitates a disconnection of the WT. Today, the grid-side converter is connected to the grid in a shunt conﬁguration, see Fig. 7.21. This means that the converter injects a current into the grid. However, if the converter is instead connected in series with the grid, a voltage is introduced in series with the stator voltage, i.e., the stator voltage of the DFIG is the sum of the grid and converter voltages. Then, the series voltage can be used in order to control the stator ﬂux of the machine and prevent, for instance, high rotor current with resulting disconnection of the turbine. Kelber has shown that such a system can effectively damp the ﬂux oscillations caused by voltage sags [55]. Grid DFIG DFIG Grid

Conv. Shunt connected

Conv. Series connected

Fig. 7.21. Schematic ﬁgure showing shunt- and series-connected converters for doubly-fed induction generator systems.

The contribution and purpose of this section is to analyze and present the advantages and drawbacks of a DFIG system for a wind-turbine application with a series converter with the focus on handling voltage sags. In addition, a goal is also to study the energy efﬁciency, and, in particular, compare it to a system that utilizes a full-power converter. The reason for comparing these two systems is that they are both capable of voltage sag ride-through.

**7.4.1 Possible System Conﬁgurations
**

As mentioned in the Introduction, the idea is to have a converter connected in series with the stator circuit and the grid. Fig. 7.21 shows both the ordinary DFIG system where the converter is connected in shunt to the grid, and the system where it is connected in series. The purpose of the series-connected converter is to control the stator ﬂux of the DFIG, and in this way be able to control the DFIG during voltage sags. By having the converter connected 118

in series, the stator voltage vs of the DFIG is, ideally, the sum of grid voltage Eg and the voltage vc from the series converter: v s = E g + vc . Some of the demands on the series converter for a DFIG system may be: • A sufﬁciently fast stator-ﬂux control in order to damp the oscillations and control the stator ﬂux. • Accurate control of the dc-link voltage. There are at least two methods of accomplishing this series voltage, which are presented below. Series-Injection Transformer In this conﬁguration, the voltage source converter is connected to the grid via a seriesinjection transformer, as depicted in Fig. 7.22. This conﬁguration of a series-injection transformer and a voltage source converter is also used in dynamic voltage restorers (DVRs) [27, 49]. The protection system of such a system is complicated since a simple disconnection does not work [70]. Normally the system is equipped with an LC ﬁlter with the objective of reducing voltage and current harmonics generated from the voltage source converter. Note that the LC ﬁlter can be placed on either side of the transformer [49]. The series-injection transformer is necessary for galvanic insulation. Moreover, in order to avoid magnetic saturation, the series-injection transformer must be rated to handle twice the nominal ﬂux [27]. Another option, in order to avoid the series-injection transformer, is to have a converter for each phase with separate dc links [63]. For DVRs, there are, at least, three Transformer Grid DFIG (7.69)

Converter

Fig. 7.22. Doubly-fed induction generator (DFIG) with the grid-side converter connected in series via a series-injection transformer.

methods of controlling the series voltage: 1) in an open-loop manner [45], 2) directly control the series voltage [71], and 3) by two control loops, i.e., an inner fast current control loop that controls the current through the inductance and an outer cascade loop controlling the capacitor (series) voltage [10]. One advantage of controlling both the inductor current and capacitor voltage is that it is easy to avoid the resonant frequency of the LC ﬁlter. However, a drawback for the DFIG is that bandwidth is lost for the stator-ﬂux controller, since the stator ﬂux is then controlled in cascade with both the capacitor (series) voltage and the inductor current. For example, if it is desired to separate the control loops by one decade, 119

the bandwidth of the ﬂux control loop is a factor of hundred lower than the current control loop. This means that a very high bandwidth of the current control loop is necessary and, accordingly, a very high switching frequency is needed. Converter in the Y Point of the DFIG The second method of accomplishing a series voltage for the DFIG is to connect a voltage source converter where the Y point of the stator circuit usually is [54, 55]. Hereafter, this will be referred to as the Y point. In [54, 55], this is accomplished using an additional (third) converter, which is only used to damp the occurring stator ﬂux oscillations. During normal operation, the extra converter voltage is zero. In [54, 55], the converter in the Y point of the DFIG system is only used to damp disturbances, while here it can also be used to control the magnitude of the stator ﬂux and the dc-link voltage. Fig. 7.23 shows a principle sketch of the system when a voltage source converter is connected to the Y point of the stator circuit of the DFIG. For this system, the converter Stator circuit Grid

Rotor circuit

Converter

Fig. 7.23. Doubly-fed induction generator (DFIG) with the grid-side converter connected to the Y point of the stator circuit.

voltage is directly used to control the stator ﬂux in the machine, while the rotor current is controlled by the machine-side converter. One disadvantage of this method is that all of the stator current is passed through the Y-point converter, which may cause additional high losses in the power electronic equipment.

7.4.2 System Modeling

As mentioned earlier the stator voltage is ideally the sum of the grid and series voltage. This means that (4.38) and (4.39) become vs = Eg + vc = Rs is + vR = RR iR + dΨs + jω1 Ψs dt (7.70) (7.71)

dΨR + jω2 ΨR dt

**where vc is the series voltage. The dc-link dynamics are described by
**

2 dvdc Cdc dvdc = vdc Cdc = −Pr − Pc 2 dt dt

(7.72)

120

Since the stator ﬂux is low. ωr Eg ωr ω1 − 1 Eg ωr (7.43 p.77) where the approximation is that the stator resistance has been neglected. In the steady state.e.u. causing the ﬂux to be high and thereby also the magnetizing losses to increase.u. i. −0.u. for ωr = 0.75) (7.3 p.81) (7.74) (7.. (7.where Pr is the rotor power and Pc is the power from the grid-side converter.u. Since the magnetizing losses depend on the ﬂux in the machine. vR ≈ 0.78) For example.80) (7.73) (7. Moreover. At low wind speeds (low power).78) becomes vR = (RR + jω2 Lσ ) iref + jω2 R = (RR + jω2 Lσ ) iref + j R ≈j ω1 − 1 Eg .7 p. iRd = ψsd /LM . R R If ψsd = Eg /ωr .e. At high wind speeds (high power). the rotor voltage can be expressed as vR = (RR + jω2 Lσ ) iref + jω2 Ψs = (RR + jω2 Lσ ) iref + jω2 ψsd . controlled to be zero. (7. i.82) .79) (7. the rotor speed is also low.. for ωr = 1.77) indicates that the system will have an undesirable feature. which are given by Pr = 3Re [vR i∗ ] R ∗ Pc = 3Re [vc is ] or as Pr = 3 RR i2 + RR Rq 2 ψsd + ψsd iRq ω2 L2 M Pc = 3 Rs i2 + Eg iRq − ψsd iRq ω1 Rq (7.23 p. with ω1 = Eg = 1 p. 121 (7.u.76) in the steady state if ψsq = 0. we have. This means that ψsd must be used to control the dc-link voltage. Steady-State Operation For a constant dc-link voltage it is required that Pr = −Pc . and in the steady state ψsd approximately becomes ψsd ≈ Eg ωr (7. which means that the stator ﬂux is low. in the above equations the d component of the rotor current is controlled so that the system operates at unity power factor. a higher torque-producing current is needed. the rotor speed is low.

81) and (7. Note that in (7. ψsq = 0.86) which has the following solution: ψsd = ωr iRq L2 M ± 2RR 2 ωr i2 Rq L2 L4 M − M (Rs + RR )i2 + Eg iRq . vR ≈ vc . Then.e. However. (7. For the values M given in the Appendix.85) in order to have a constant dc-link voltage.showing that the rotor voltage is not symmetrically distributed around the synchronous speed.84) (7.84)..0017 p.009/(12 · 4.85) can be rewritten as 2 ψsd − ωr iRq L2 L2 M ψsd + M (Rs + RR )i2 + Eg iRq = 0 Rq RR RR (7.72) equals zero at steady-state operation in order to have a constant dclink voltage.u.. Eq. 2 ωr L2 M (7.85). ωr (7. Close to No-Load Operation It is required that (7. later on when the losses 122 . as for the case with constant stator ﬂux. we have that R R i2 + RR Rq 2 ψsd + ψsd iRq ω2 + Rs i2 + Eg iRq − ψsd iRq ω1 = 0 Rq L2 M (7. which is a small value.62 ) = 0. as ψsq = 0 and Eg = jEg .83) As seen in (7.88) ≈ (7. we have vc ≈ jω1 ψsd − jEg = jω1 Eg − jEg = j ωr ω1 − 1 Eg . implying that L2 M iRq RR or |iRq | ≥ Eg 2 ωr L2 M 2 ωr L2 M − (Rs + RR ) iRq − Eg ≥ 0 4RR 4Eg RR . It might be possible to use either a diode rectiﬁer (depending on the power ﬂow) or an IGBT converter as the extra converter.87) The expression under the square root cannot be negative. the rotor voltage will approximately equal the converter voltage. the constraint becomes |iRq | < 4 · 1 · 0. If the rotor current and stator ﬂux are controlled with high-gain feedback. In order to handle this problem an extra converter that controls the dc-link voltage is added. The series voltage vc of the grid-side converter is given by vc = Rs is + jω1 Ψs − Eg ≈ jω1 Ψs − Eg in steady state. Rq 2 4RR RR (7. ψsd cannot keep the dc-link voltage constant.89) 4RR − (Rs + RR ) 2 If |iRq | < 4Eg RR /(ωr L2 ). i.

an IGBT converter has been used. 7. when designing the control laws it will be assumed that an additional power electronic device keeps the dc-link voltage constant when iRq is small. 55].70) can be rewritten as Eg + vc = −Rs iref + R dΨs + dt Rs + jω1 Ψs LM (7. 55]. is used to control the reactive power while the q component of the rotor current. that for this option the ﬂux does not follow (7. This means as the extra converter controls the dc-link voltage. Kelber et al. the stator ﬂux is controlled using the grid-side converter. to be able to control the dc-link voltage with the stator ﬂux there is a minimum rotor current.” The above equation then reduces to vc = −Rs iref + R dΨs + dt 123 Rs + Ωa Ψs . used this option [54. ψsd is not used to control the dc-link voltage anymore. For this case. Then.90) where the rotor current has been put to its reference value. Therefore. However. the stator ﬂux can be controlled “arbitrarily. For this option the extra converter is designed so that it is used in the whole operating region. Then. Option 2. In this case.3 Control The basic idea of the control system is to have an inner fast rotor current controller. Note. Stator-Flux Control The stator voltage equation (7. ﬁnally. However.4. iRd . controls the active power or the torque. the stator-ﬂux reference value is set to minimize the losses of the generator. This means that for this option the stator ﬂux is not used for controlling the dc-link voltage. the dc-link voltage is controlled in cascade with the stator ﬂux in order to keep the dc-link voltage constant. the stator ﬂux is here controlled to reduce the magnetizing losses. LM (7.77). vc = vc −Eg +(jω1 −Ωa )Ψs is chosen where Ωa is the “active damping.91) .and efﬁciency are calculated. This means that when iRq is below a certain value. As mentioned in the previous section. Here two different sizes of this extra converter will be investigated: Option 1. iRq . in contrast to [54. Rotor Current Control The d component of the rotor current. The rotor current is controlled with the machine-side converter. for details see Chapter 4. the extra converter is designed to be as small as possible. Then. The stator-ﬂux control loop is about a decade slower than the current control loop. and thereby make the third converter unnecessary.” meaning that the stator ﬂux can be controlled so that the losses are reduced. since this converter would be small and only used at very low powers another way could be to increase energy storage on the dc-link. With the rotor current it is possible to control the active and reactive powers.

e. becomes W (p) 3 G(p) = = .93) p p If the “active damping” is set to Ωa = αf − Rs /LM a disturbance is damped with the same bandwidth as the closed-loop stator-ﬂux control loop. D(p). If the resistive losses are treated as a disturbance. 124 .96) = 3iref ψsd ωr − Eg + D.98) = 3ψsd − 3Ga W + D. (7. the dc-link dynamics in (7.5. 2 dt In (7.72) can be written as 2 Cdc dvdc (7.101) Cdc D(p) 1 + FG 2 (p + αdc ) 2 This means that a disturbance is damped with the same bandwidth as the dc-link voltage control loop. Rq 2 dt where iRq and ψsd are put to their reference values.95) = 3iRq (ψsd ωr − Eg ) + D 2 dt 2 where D is the disturbance. the transfer function. treating D as a disturbance.97) ψsd = sd ref ωr iRq ωr Ψs (p) p . = D(p) (p + αf )2 (7. Ga is the “active damping. (7.92) = G(p) = vc (p) p + Rs /LM + Ωa IMC yields the following PI controller tuned for a closed-loop bandwidth αf αf −1 Rs /LM + Ωa G (p) = αf + αf . to Ψs is F (p) = GDΨs (p) = DC-Link Voltage Control For a dc-link voltage controller with a shunt converter. if the variable substitution W = vdc is made. the transfer function from D to W becomes G p W (p) = = . see Section 4. the transfer function from a disturbance.. (7. (7. if Ga = Cdc αdc /6. (7.99) ref Cdc /2p + 3Ga ψsd By using IMC.97).The term Rs iref is treated as a disturbance and the transfer function from vc to Ψs is found R as 1 Ψs (p) . Moreover.” Then.2. the following system is obtained Cdc dW ref (7.94) the dc-link dynamics are reduced to Cdc dW ref (7. we obtain the following PI controller αdc −1 Cdc αdc Ga αdc F (p) = G (p) = + .100) p 6 p Then. i. By choosing the reference value of the ﬂux as ψ ref − Ga W Eg ref + (7.

1 p.u. the pitch angle of the turbine is ﬁxed and the DFIG is operated in speed control operation. if desired.9 p.5 1 1.5 1 1.4.5 Time [s] Time [s] Fig.8 2. it is assumed that the rotor speed can by the pitch mechanism of the wind turbine.5 d q vc [p. After 50 ms.5 0 0.5 d) 1.2 p. 7. b) Stator ﬂux.u. Moreover.5 0 0. a) Rotor current.5 vR [p.] 0 f) 0.] 0 0.5 −0. to −0.5 1 1. using pitch control. be controlled with a bandwidth of αs . the simulation veriﬁes the result previously presented in Section 7.Simulation of Electromechanical Torque Steps In the simulation shown here. There is a small difference in the d component due to the fact that the rotor converter also supplies the magnetizing current. at 1. down to 0. to control iRd .24. The 125 .. In this section.5 1 1.u. a rotor speed control law will be derived using IMC. f) Series voltage. e) Rotor voltage.u.u.5 1 1.8 p.u.0 s the rotor speed is ramped from 1. It can be seen in the ﬁgure that the control system manages a) 0.] 2 1 0 −1 iR [p.u.] d q 0 0.2.24 shows a simulation of the investigated system during current (or torque) control mode.5 d 0 b) Ψs [p.25 s. Fig. iRq is stepped back to −0.5 1 0.4. 7. and between 0. iRq is stepped from −0.5 q −0.2 2. Finally.5 e) 0.4 Speed Control Operation At low wind speeds.5 −1 c) vdc [p. which indicated that the q component of the rotor voltage and converter voltage are close to identical.6 0 0.5 1 1. 7. d) Rotor speed. and vdc well. iRq .u.9 p. c) DC-link voltage.u.] 0 3 0.u.5 s and 1. Simulation of the system when the DFIG is in current control (or torque) mode.u.8 ωr [p.] 0 d q −0.

we choose iref = Rq ωr i ref + Ba ωr 3np Eg Rq (7.” This means that the mechanical dynamics can be rewritten as J dωr ref = −iRq − Ba ωr − Ts . the DFIG is operated at iRq = −1 p.mechanical dynamics are given by J dωr = Te − Ts np dt (7.103) where Ba is the “active damping. It can be seen in the ﬁgure that the control system manages to control iRd . These torque steps are much faster than what would be the case in reality and they are performed for veriﬁcation purposes only. The rotor 126 . The voltage drops after 0. J is the inertia and. the stator ﬂux is controlled by the series-connected converter to be close to zero.104) (7. during the whole simulation. αs . it manages the Swedish transmission system operator’s demands for large production facilities [96].5 Response to Voltage Sags Fig..106) (7. This is an extreme voltage sag and if the system manages this sag. The rotor speed is controlled by the DFIG to be 0.u.102) where Te is the electromechanical torque.102) can be expressed as J dωr Eg = −3np iref − Ts np dt ωr Rq where iRq has been changed to its reference value. the rotor speed is controlled by the pitch mechanism to 1.1 s and the sag has a duration of 250 ms. with IMC.e. ωr .105) and if Ba = αs J/np . the electromechanical torque can be expressed as Te = −3np ψsd iRq . In this case.u. since ψsd ≈ Eg /ωr .107) Fig. which corresponds to generator operation at rated current (full power).25 shows an example of the proposed DFIG series system with the DFIG operated in speed-control mode.u.8 p. 7. the remaining voltage is 0 p. 7. Now. In order to validate the performance. np is the number of pole pairs.4. Assuming ψsq = 0.26 shows the response to a 0% voltage sag. (7. a change in Ts is damped with the same bandwidth.2 p. np dt Then. as the speed control loop: GT ω (p) = P J (p + αs )2 np (7. Ts is the shaft torque. During the sag. the following controller is obtained F (p) = kp + J Ba ki = −αs − αs p np p (7.u. and vdc well. 7. During the simulation. the machine is exposed to shaft torque steps of 30% to 100% of rated torque. Then. i.

a) 0. Therefore. the main reasons for choosing a DFIG system are cost and efﬁciency.5 1 1. 7. Otherwise.] 2. The main reason for this is that both the rotor and converter voltages have been limited to their maximum values. in this case at a voltage level of 0 p. In this section. However. Simulation of the system when the DFIG is in speed-control mode. Some minor current transients can be observed at the instant of the sag and at the instant where the voltage returns.u. Moreover.] 1 0.u. This means that the pitch mechanism must reduce the incoming torque accordingly. In Chapter 3. a) Rotor current. 7.82 vdc [p.5 0 −0. The system then returns to a steady-state operating condition.5 2 d) ωr [p. d) Rotor speed. This can be realized from the fact that the stator ﬂux is controlled down to an appropriate level.25. Details of the calculations methods used here are described in Chapter 3.5 Ψs [p.5 1 1.81 2. when modifying the DFIG system it is necessary to evaluate how the modiﬁcations affects both cost and efﬁciency.5 2 0 0 0.25 q 0.] 1 0. the system can stay connected to the grid for indeﬁnitely long voltage sags.5 2 Time [s] Time [s] Fig.4.5 iR [p. the systems are compared to an ordinary DFIG system and a system that utilizes a full-power converter. the dynamic performance of the system is promising. the average efﬁciency of the ordinary DFIG WT system has been calculated and compared to other electrical conﬁgurations used in wind turbine systems. current is practically constant during the sag.u.] q d b) 1.5 2 c) 2. 127 .5 0 0. as otherwise overspeed occurs and the overspeed protection trips the turbine. Although the sag in the simulation is only 250 ms.u.5 1 1. one issue that must be kept in mind is that. b) Stator ﬂux.4.5 1 1.5 0 d 0.6 Steady-State Performance As mentioned in the Introduction.8 2.2. the efﬁciency will be calculated for the two options presented in the last part of Section 7. the maximum torque that can be handled by the generator is reduced in proportion to the voltage sag. This study serves as a basis for the efﬁciency calculations and comparisons in this section.79 0 0.u. c) DC-link voltage.75 0. since the stator ﬂux is reduced according to the voltage sag.

c) DC-link voltage.a) iR [p.3 0. As seen in Fig.u. Accordingly.u.27 shows the converter losses for the ordinary DFIG system.28.] d 2 1 0 −1 d q 0 0. d) Rotor speed.3 0. The efﬁciency of the DFIG series system with Option 2 is between the ordinary DFIG system and the fullpower converter system. 7. 7.5 0 0.2 1.2 0. e) Rotor voltage.5 d) 1. the average efﬁciencies of the ordinary DFIG system. 128 .3 0.5 q c) vdc [p.4 0. Series-Injection Transformer Fig. this is the most energy efﬁcient system with voltage sag ride-through facility. 7.] 1.4 0.5 e) 0. f) Series voltage. In Fig. b) Stator ﬂux.3 0. and a system with a full-power converter are shown as functions of the average wind speed. the efﬁciency of the standard DFIG system is highest.26.1 0. 7. and in addition it presents the losses of the series-injection transformer for the two options of the DFIG series system presented earlier.29 shows the magnetizing and resistive losses of the generator and the converter losses for the ordinary DFIG system and for the two options of the DFIG series system. a) Rotor current.u.5 vR [p.u. the efﬁciencies of the ordinary DFIG system.1 0.u.5 0 0. 7.28.5 0 0. Converter in the Y Point of the DFIG Fig.u.5 3 2.1 0.3 ωr [p.5 d q vc [p.] 1 0 −1 −2 0 0.] 0 d q −0.2 0.30.1 0. 7.4 0. although the efﬁciency is slightly lower at low average wind speeds and slightly higher at higher average wind speeds.2 0. Response to a 0% symmetrical voltage sag. full-power converter system. In Fig. The efﬁciency of the DFIG series system with Option 1 is roughly same as the system that utilizes a full-power converter.5 Time [s] Time [s] Fig.4 0.1 0.2 0.1 1 0 0.] 4 3.5 −0.2 0.2 0.3 0. the series system with the two options.4 0.3 0.4 0.5 b) Ψs [p.1 0.] 0 f) 0.

d) Transformer losses. 7. it can be seen that the results are almost identical when connecting the converter to the Y point of the stator circuit.5 0 5 10 15 20 10 15 20 Transformer losses [%] Resistive losses [%] a) b) 2 1.27.5 1 0. and dotted line to the ordinary DFIG system.28. and the two different options for the series DFIG system are shown as functions of the average wind speed. 7.5 0 5 10 15 20 25 25 Wind speed [m/s] Wind speed [m/s] Fig.5 1 0.Converter losses [%] Magnetizing losses [%] 1. solid line to the DFIG series system. Dashed lines correspond to the series DFIG system with Option 1.30 is compared to Fig. b) Resistive losses of the generator.28. If Fig. solid line to the series DFIG with Option 2. c) Converter losses.5 1 0. 7. 7. Dashed line corresponds to the ordinary DFIG system. and dotted line to a system with a full-power converter. Expected efﬁciency as a function of the average wind speed for the system with a seriesinjection transformer. a) Magnetizing loses of the generator. 96 Option 2 95 Efﬁciency [%] 94 93 92 91 Option 1 90 5 6 7 8 9 10 Average wind speed [m/s] Fig.5 0 5 1. Losses of the system with a series-injection transformer with the same turns ratio as the stator-to-rotor turns ratio.5 0 5 1 10 15 20 25 25 c) d) 0. One reason for this is that the increased losses in the converter are almost the same as the losses of 129 .

Expected efﬁciency as a function of the average wind speed.31 shows the relative energy cost of the DFIG series system in comparison to the ordinary DFIG system. and dotted line is the ordinary DFIG system. 7. solid line is the series DFIG with Option 2. 7.27. Energy Production Cost Fig. 7. solid line to the DFIG series system. However. The generator losses are identical to that of Fig. from the 130 . Dashed line is the series DFIG system with Option 1.2 Converter losses [%] 1.29. Dashed line corresponds to the ordinary DFIG system. 96 Option 2 95 Efﬁciency [%] 94 93 92 91 Option 1 90 5 6 7 8 9 10 Average wind speed [m/s] Fig.30. an extra converter for the DFIG series system seems to be disadvantageous. 7. Converter losses when the converter is connected to the Y point of the stator circuit of the DFIG. 7. From an initial cost perspective. The converter is connected to the Y point of the stator circuit of the DFIG.5 1 0. the series-injection transformer. as indicated in Fig.5 0 5 10 15 20 25 Wind speed [m/s] Fig. and dotted line to a system with a full-power converter.31.

131 . The rotor current (torque and power factor). two different options using an additional converter to solve this problem have been proposed and investigated. Two different methods of connecting the series converter resulted in almost the same efﬁciency.02 1. a cage-bar induction generator equipped with a back-to-back converter. It was found that the best option was to use an additional converter for controlling the dc-link voltage in the whole operating area. as seen in the ﬁgure.u. energy cost point of view it is beneﬁcial to use the extra (third) converter to control the dclink voltage.4. and dc-link voltage are controlled.01 Option 2 1 5 6 7 8 9 10 Average wind speed [m/s] Fig. the series-connected converter can be used to control the ﬂux to an optimal value from an overall efﬁciency point of view. Simulations showed that the dynamic performance of the system is promising both during normal operation and during conditions when voltage sags are present in the grid.5 percentage units. In the ﬁgure it is shown that the increased energy cost for this series system using Option 2 is approximately 1. 7.05 Relative energy cost [p.1. The derived control law is not capable of controlling the dc-link voltage at very low loads.7 Discussion and Conclusion A control law for the doubly-fed induction generator with the grid side converter connected in series with the stator circuit has been derived. for the system with a full-power converter the corresponding energy cost is approximately 1. dashed lines are with the converter connected to the Y point.31. and dashed-dotted line is the system that utilize a full-power converter. The efﬁciency of the DFIG series system with the best performance was found to be between the ordinary DFIG system and a system that utilizes a full-power converter system.e.] Option 1 1. The energy cost is related to the ordinary DFIG system. Then. i. 7. stator ﬂux. As a remedy for this.04 1.03 1.. Moreover. Energy cost of the DFIG series system. Solid lines are with the series-injection transformer.5 percentage unit higher than for the series system with Option 2.

5 Conclusion In this chapter. Two candidate methods. However.. while the series-connected DFIG system seems to have similar dynamic performance as the full-power converter system. The shunt connected DFIG system with ride-through capabilities still suffers. can successfully reduce disturbances from both symmetrical and unsymmetrical voltage sags. of improving the voltage sag ride-through of DFIG variable-speed wind turbines have also been investigated. e. Another drawback of the series-connected DFIG system in comparison to the full-power converter system is that the maximum torque that can be handled by the generator is reduced in proportion to the voltage sag.5 percentage unit higher compared to the ordinary DFIG system. The shunt DFIG system and the series system have approximately the same energy production cost. 132 . at least initially. voltage sag ride-through of variable-speed wind turbines has been investigated. from high fault currents.7. the control of the DFIG series system is much more complicated than that of the full-power converter system. with one shunt-connected and one series-connected grid-side converter respectively.g. a cage-bar induction generator with a back-to-back converter. The energy production cost of the full-power converter system was found to be three percentage units higher than that of the ordinary DFIG system. It has been shown that a variable-speed wind turbine with a full-power converter system. which is approximately 1.

8. One possibility to reduce ﬂicker from a stall-controlled WT with an induction generator (IG) directly connected to the grid could be to introduce a variable rotor resistance. i. it is assumed to be ideal. of WTs is an important concern for grid owners. It is possible to control the slip of the IG with the external rotor resistances. Worth pointing out is that ﬁxed-speed WTs have the same energy production given a certain rotor diameter as variable-speed WTs (see Chapter 3). have been developed successfully using the ﬁxed-speed stall-regulated concept. and external rotor resistances. However. For individual installations of these types of WTs. Small WTs.. One way of representing the IG dynamically is to the use the so called Γ model as de133 . gearbox. is presented. 1 MW and below. for a stall-controlled WT. The purpose of this chapter is to derive a rotor resistance control law. generator. 8.e. Therefore. In other words. in this way. and will probably dominate the small-turbine market also in the near future. the rotor resistance could be used to control the rotor speed in a limited range and. with the objective of minimizing torque ﬂuctuations and ﬂicker. for instance the ﬂicker (or voltage ﬂuctuations) contribution. in this chapter the power electronic equipment is not included in the model. The value of the external rotor resistances is adjusted with the power electronic equipment. absorb torque ﬂuctuations and thereby reduce the ﬂicker emission. the system with turbine. the installations of small-scale wind turbines (WTs) will most likely proceed. the ﬂicker contribution can be the limiting factor from a power quality point of view.Chapter 8 Flicker Reduction of Stalled-Controlled Wind Turbines using Variable Rotor Resistances Although there will be very large wind power installations. the external rotor resistances can be treated as a continuous variable. The power quality impact.1. especially in weak grids [64].1 Modeling In Fig.

simulations of the induction machine are presented.05 p. and Ψs ≈ ψsd ≈ ω1 reduces to Lσ vs diRq = −RR iRq − (ω1 − ωg )Lσ iRd − (ω1 − ωg ) dt ω1 J dωg Tt = −kT iRq − np dt gr (8. RR =0. J=32000 p.u.2.u.. the shaft torque is increased from half of the rated torque to rated.). The drive train (soft axis) is not included in the model. RR =0. vs .4) where kT = 3vs np /ω1 . Note that in this simulation. In Fig. LM =5 p. and.u. RR =0. This has been done in order to get a quicker 134 .u. 8. the rotor resistance is increased by 40% after 50 ms and. Tt is the torque produced by the turbine. the dynamic system if iRd can be assumed constant or at least small..1) where Tg is the electromechanical torque produced by the generator. one electrical and one mechanical equation. the stator-ﬂux dynamics have also been neglected.01 p. and gr is the gear ratio of the gearbox. after 250 ms.. np =2. i.2 p... both with the ﬁfth-order and the second-order model of the system. Wind turbine with variable-rotor-resistance induction generator. dt (8. For the 1-MW IG considered in this chapter.39). 8.2) Note that in the above equation.3) (8. Eliminating ψR from (4. gr =61.e.007 p.39) yields 0 = (RR + jω2 Lσ )iR + Lσ diR + jω2 Ψs .39) is to neglect the stator-ﬂux dynamics.u..1. scribed in Section 4. only the inertia of the generator has been taken into account and not the inertia of the turbine.u.1.03 p. The mechanical dynamics are described by Tt J dωg = Tg − np dt gr (8. the electrical dynamics of the induction machine dynamics are described by (4. Lσ =0.1. In the simulations. 8. absolute ﬂicker values are of minor importance.u. operated at 690 V and 50 Hz the following avg max min parameters are used: Rs =0.2. on the low-speed side of the gearbox.38) and (4. Further. This means that the model has been reduced to the second order.Gearbox IG Grid External rotor resistances Fig. (without turbine J=3000 p. since the objective is to investigate the relative performance of the derived control law and. for instance.1 Reduced-Order Model A common way to reduce the order of the induction machine model in (4. Then..u.

Solid lines correspond to the ﬁfth-order model while dashed lines correspond to the second-order model.3 0.4 Time [s] d) Stator ﬂux [p.a) Rotor current [p.] 1 q b) Torque [× Tnom ] 0.6 −0.4 Time [s] Time [s] Fig.1 0.4 −0.2 0. 8.2 0.4 Time [s] c) Rotor speed [p.5) where RRa is an “active damping.1 0.05 1 0.2. in (8.05 1.u.2 0.06 1. we will introduce the following non-linear control law ref RR = iRq RR + RRa iRq iRq =0 (8.2 0. b) Torque. i.95 0 0.” which can be used damp disturbances as described earlier. Example of the response of the induction machine due to a step in the rotor resistance. response of the rotor speed and thereby a more lucid ﬁgure.01 0 0.u. the term RR iRq .4 −0. Substitution of the above control 135 . a) Rotor current.] 1.04 1. The rotor resistance is increased 40% after 50 ms and after 250 ms the shaft torque is increased to the rated torque.3). 8. The ﬁgure shows that both models produce approximately the same results. However.3 0.03 1.2 Current Control In order to remove the multiplication between the RR and iRq ..1 0.02 1. c) Rotor speed and d) Stator ﬂux.e. there is a small deviation in the d component of the rotor current.3 0.1 0. This reduced-order model will be used to derive the control law.] 1. How to chose RRa will be described in the next section.8 −1 −1.2 0.u.3 0.5 0 d −0.5 0 0 0.

By treating the term D as a disturbance the following open-loop transfer function can be found Gol (p) = iRq (p) −1 . by using IMC. 8. + (RRa + Lσ αc )p + RRa αc (8.iRq (p) = −p . the following current controller is obtained Fc (p) = kpc + kic RRa αc = −Lσ αc − p p (8.6) (8. A block diagram of the current control loop is shown in Fig.9) where αc is the closed-loop bandwidth of the current control loop.3.3.11) This choice of RRa causes a disturbance to be damped with the same time constant as the current control loop.4 for three different values of the current control loop bandwidth αc . = RR (p) Lσ p + RRa (8. Determination of the Active Damping The transfer function. Current Control Block Diagram.7) where a term D has been introduced. the above transfer function is reduced to GD. iref + Rq − + + RRa ref RR Fc (p) 1/iRq iRq Gol (p) Fig. is found as GD. Lσ (p + αc )2 (8.10) Lσ If RRa = Lσ αc . 8. 136 .11) can be seen in Fig.3) yields Lσ diRq = −RR − RRa iRq + D dt vs D = −(ω1 − ωt )iRd Lσ − (ω1 − ωt ) ω1 vs = −ω2 iRd Lσ + ω1 (8.law in (8.iRq (p) = p2 −p . 8.8) Then. A Bode diagram of (8. from a disturbance D to the current iRq .

Bode diagram. a) iRq [p.3 0.65 0 0.1 0.45 −0. c) Rotor resistance (dashed line is the minimum.4.55 −0.15 0.6 0.04 0.u.15 0.05 0. average and maximum value of the available rotor resistance). a) q component of the rotor current (dashed line is the reference value). dashed αc =220 rad/s and dotted αc =2200 rad/s.1 0.] 0.05 0.2 0.15 0.u.] 0.05 0. Solid αc =22 rad/s. 8.3 Time [s] Fig.2 0.1 0.06 0.25 0. 137 . 8.25 0.5.5 −0.5 0.6 −0. Ts [× Tnom ] 0 −0.4 b) Tg . Example of current control of an IG with external rotor resistances.02 0 0 0.10 10 10 2 1 0 Gain 10 10 −1 −2 10 −3 0 50 100 150 200 250 Frequency [Hz] Fig.2 0. b) Torque (dashed line is the shaft torque).25 0.3 c) RR [p.

the current controller will not manage to keep the rotor current and thereby the generator torque. as seen in the ﬁgure.13) can be rewritten as ref RR ≈ − vs ω1 iref Rq np ΔTt dt Jgr (8. iref . the Rq generator speed becomes according to (8.e. the setting of the current reference will be of great importance for the over-all performance of the system.13).3) can be expressed as Lσ vs diRq ref = −RR iRq − ω2 Lσ iRd + dt ω1 = 0.14) if the system initially was in the steady state and the electromechanical torque. The integral can be evaluated easily since ΔTt is constant. it is seen that the current controller manages to control the rotor current with the desired bandwidth.16) . np vs ΔTt 138 (8.5 shows a simulation with the above derived rotor current control law. This means that (8.12) This implies that the rotor-resistance reference value varies as ref RR = − ω2 iref Rq Lσ iRd + vs ω1 ≈− ω2 vs iref ω1 Rq (8. 8. ω2 . How to set the rotor current reference will be further addressed in the next section. is kept constant.1) ΔTt J dωg =− np dt gr (8. First. the rotor resistance varies over its entire range even for small current variations and shaft torque steps. In the ﬁgure. however. In the simulation the bandwidth of the current control loop is set to 220 rad/s which corresponds to a (10–90%) rise time of 10 ms. and the operating condition. (8. This means that rotor resistance has changed ΔRR over the time Δt = − Jgr ω1 iref Rq ΔRR . However.8.2. the rotor resistance is constantly increased (or decreased for opposite sign of the torque difference). the rotor-current dynamics in (8. i. if the generator is exposed to a shaft-torque step. From (8.1 Evaluation Fig. If the rotor resistance has to be limited. a brief analytical investigation of how the rotor resistance varies due to a shaft torque step is made. It is also seen that when the shaft and generator torques differ (between 150–250 ms). in this section. Because of the limited range in which the rotor resistance can vary.15) iref Rq =0 since ωg = ω1 − ω2 and dω1 /dt = 0. By controlling the rotor current with a high-gain feedback. it is seen that the rotor resistance is depending on the slip. ΔTt .13) iref Rq = 0. the controller manages to keep the generator torque at the shaft torque step (at 150 ms) until the current reference is adjusted according to the new shaft torque (at 250 ms). Tg . Moreover. Moreover.

024) would. dt (8. it can be assumed that RR = RR . ref ω kR I − Ba ω2 ω1 iRq 1 (8. One idea is to set iref as Rq iref = kR Rq (RR0 − RR )dt − Ba ω2 (8. Moreover.19) (8. Therefore.13). the system becomes Tt Tt J dω2 = kT (kR I − Ba ω2 ) + = kT iref + Rq np dt gr gr dI = RR0 − RR . smaller values of iref imply a shorter time until the rotor Rq Rq resistance must be limited. i. according to the above formula. I0 = gr kR kT vs 139 (8. ωg = ω1 − ω2 and dω1 /dt = 0. according to (8. iRq = iref = kR I − Ba ω2 .5.0 = RR0 Tt0 ω1 gr kT vs Tt0 (ω1 RR0 Ba − vs ) . Δtlim . the increase in rotor resistance (ΔRR = 0. the time.22) (8.17) This means that for a given step in the shaft torque. If the current control loop is fast.23) . = RR0 + dt kR I − Ba ω2 ω1 The above system has an equilibrium point at ω2.2 s.3 Reference Value Selection In the steady state. That is. if ±ΔRR is the maximum available rotor resistance. ref since the bandwidth of the current control loop is fast. 8. ﬁnally.18) where only an integration term of the error in the rotor resistance is used in order to avoid an algebraic loop.25) (8. RR0 .. This nonlinearity makes the setting of the rotor current reference iref more difﬁcult. to reach maximum or minimum value of the rotor resistance becomes Δtlim = Jgr ω1 iref Rq max ΔRR .21) This means. where I is Rq the integration of the error in the rotor resistance.24) (8.For the shaft torque step at 150 ms in Fig. the time until the rotor resistance must be limited depends on iref . the rotor resistance should be (or at least close to) its desired value.1).e. that the following system must be analyzed J dω2 Tt = kT (kR I − Ba ω2 ) + np dt gr ω2 dI vs . which also can be seen in the ﬁgure. np vs ΔTt (8. Rq 8. take 0.20) Note that the slip dynamics are found from (8. max Moreover. RR equals to ref RR = RR ≈ − vs ω2 vs ω2 =− .

26) (8. This means that GTt e = p GTt I (p) (8.e.Linearization and insertion around the equilibrium point yields ⎡ ⎤ np kR kT np /J −Ba kT np /J Δx = ⎣ gr kT (ω1 RR0 Ba − vs ) −gr kR kT RR0 ⎦ Δx + gr J Δu ˙ 0 Tt0 ω1 Tt0 where Δx = Δω2 ΔI Δu = ΔTt . the rotor resistance returns to its desired value RR0 .30) where αR is a parameter that can be set “freely.” the above transfer function GTt e (p) becomes GTt e (p) = − (kT np vs − αR JRR0 ω1 )2 p JkT np vs ω1 Tt0 (p + αR )2 (8.27) Now.28) where GTt I (p) is the transfer function from Tt to I which can be found from the system in (8. the damping of the above transfer function and the parameter kR is dependent on the operating condition.32) where K is a constant that depends on the operating condition. it is possible to use the derivative of I.. e cannot be found directly from the state variables but since I is the integration of e. If kR and Ba are chosen as 2 αR JTt0 ω1 2 gr kT np vs α2 J 2 RR0 ω1 − 2aR JkT np vs Ba = − R 2 kT n2 vs p kR = (8. e = RR0 − RR .33) where L is the Laplace transformation symbol. Tt0 . one option is to study the error in the rotor resistance. it is interesting to see how a change in the incoming torque inﬂuences the rotor resistance. However. Moreover. we will get e(t) = L−1 1 GT e (p) = Kte−αR t p t (8. i. If the system is exposed to a step.29) (8.26). From the above equation. 8. it is seen that after a torque step. i.1 Evaluation For a given operating condition it possible to express (8.31) as GTt e (p) = K p (p + αR )2 (8. 140 . Therefore. e(t → ∞) = 0. (8.3.e.31) which is a band-pass ﬁlter centered at αR ..

by looking at the derivative of the above function it is possible to determine that the function has a maximum at t max(RR ) = 1 . it is described how this value is determined [51].02 0 0 5 10 15 20 Time [s] Fig. Example of outer reference selection control loop.6 shows a simulation of the system with the reference selection control loop. 8.34).a) iRq [p.2 −0.2 0 5 10 15 20 b) Tg .u.6. b) Torque (dashed line is the shaft torque). a) q component of the rotor current (dashed line is the reference value). the ﬂicker emission is compared to a similar system with uncontrolled rotor resistances.] 1 0. Flicker emission or rapid voltage ﬂuctuations can be described with the dimensionless quantity Pst : the short-term severity index. Ts [× Tnom ] −0. In the standard IEC 61000-21. It is seen in the ﬁgure that after the torque step (at t = 1 s) the rotor resistance has its maximum value after 1 s (at t = 2 s). average and maximum value of the available rotor resistance).. Moreover.4 Evaluation In order to evaluate the derived control law.8 −1 0 0. Moreover. RR is ﬁxed. which is also veriﬁed by the expression (8.4 −0.] 0.6 0.8 0.e.4 0. The bandwidth of the current control loop is set to a high value (2200 rad/s) and the parameter αR is set to 1 rad/s. 8. after the torque step the rotor resistance is returning to its desired value.6 −0. αR (8.04 0. The 141 .06 5 10 15 20 c) RR [p.u. i.34) Fig. c) Rotor resistance (dashed line is the minimum. 8.

has been avg set to the average value of the available rotor resistance RR .5 −1 −1. b) Slip and c) Rotor resistance. a) Torque (generator torque is solid and turbine torque is dashed). The bandwidth Torque [× rated] a) 0 −0. Fig. The set point value for the rotor resistance.7. 8. The average wind speed in the 10 minute simulation was 14 m/s and the turbulence intensity was 25%. corresponding to the average wind speed.5 rad/s. 8.04 0. Naturally. However. since this should be done on a much slower time scale than the bandwidths of the control loops. they can be adjusted according to a changing operating condition. since the 10 minute Pst -value is used.] Time [s] Fig. The average torque.5. kR .5 0 0 −1 −2 −3 −4 0 0. The applied shaft torque has been precalculated using blade element momentum theory with different average wind speeds and turbulence intensities. for each 10 minute period is used to set the parameters that are dependent on the operating condition. it has been ignored in the simulation presented here. of the current control loop.system is simulated for 10 minutes.. i.e. 142 . RR0 . αc .7 shows an example of how the derived rotor resistance control law operates for a short piece of one of the above mentioned 10-minute simulation. Then the Pst value has been calculated on a ﬁctive grid with a short-circuit power of 50 times the nominal power of the WT and with an X/R ratio of 0. of the reference value selection control loop is 0. is 2200 rad/s and the parameter αR .u.02 0 0 5 10 15 20 5 10 15 20 5 10 15 20 b) Slip [%] c) Rotor resistance [p.06 0. Example of the behaviour of the derived control law. in a real system.

1 0. that the higher the rotor resistance is. The control parameter is as in Fig. 8. the average value of the rotor resistance RR is very close to RR .8.1 0.05 0 b) Flicker Pst 5 0. the average value. it suffers from a drawback. These higher losses imply that it will be necessary to increase the cooling of the generator. during the avg simulation.1 Flicker Contribution In Fig. Flicker as a function of the turbulence intensity. 8. RR . RR . It can be seen that the derived control law produces lower Pst values than the system with ﬁxed rotor resistance.15 0. In the a) Flicker Pst 0. Solid line is WT with controlled rotor avg resistance. ﬁgure.15 0. RR = RR .7.4. dashed line is with ﬁxed rotor resistance. RR = RR and dotted line is with max ﬁxed rotor resistance. 143 .05 0 5 10 15 20 25 10 15 20 25 Turbulence intenisity [%] Fig.05 0 10 15 20 25 c) Flicker Pst 5 0. and the maximum value (in continuous max operation). The average wind speed is a) 6 m/s. b) 14 m/s and c) 20 m/s.e.8. i.. namely. except for the parameter αR that is 1 rad/s. 8.15 0. the system with ﬁxed rotor resistance has been simulated with two different values avg of the rotor resistance.1 0.8. the Pst value is seen for a system with ﬁxed rotor resistance and with the derived control law as a function of the turbulence intensity and for different average wind speeds. Finally. of the available rotor resistance. Even though max the Pst value for the ﬁxed rotor resistance system with RR = RR is close to the system with controlled rotor resistances. the higher the losses in the rotor resistance will be.

higher average torques). It is seen that when the turbulence intensity becomes higher. the number of times the rotor resistance has to be limited is rapidly increased from a turbulence intensity of 7% and upwards. for low values of αR . This is also veriﬁed by the simulation since it is possible to reduce more of the ﬂicker at higher average wind speeds (i. Reduction in ﬂicker for different bandwidths of αR .11 the corresponding diagrams for an average wind speed of 14 and 20 m/s are shown. with the derived control law. In general. for the case with αR put to 0. 8.2 0 5 10 15 20 25 Turbulence intensity [%] Fig. from the ﬁgures it can be seen that in order to have an “optimal” reduction in ﬂicker. The relative ﬂicker is given as a function of turbulence intensities for an average wind speed of 6 m/s. the more reduction in the Pst value is achieved. As mentioned earlier. lower the parameter αR is.6 0.2 0. 8. The average wind speed is 6 m/s. A relative ﬂicker of 1 corresponds to a ﬂicker contribution equal to that of the ﬁxed rotor resistance system.. The rotor resistance of this system is set to the average value of the available avg rotor resistance. For example. the Pst value is actually worse for this case than for the case with ﬁxed rotor resistance. This will have a negative impact on the performance.e.25 rad/s the rotor resistance has been limited to its maximum or minimum value between 20–70% of the total simulation time depending on the turbulence intensity.9. if the frequency is too low or the turbulence intensity is too high..8 0. Due to this fact.25 rad/s 0.5 rad/s 2 rad/s Relative ﬂicker 1 0. the same phenomena as in Fig.e. 8. In Figs. the damping of the ﬂicker (or the torque ﬂuctuation) is dependent on the operating condition.8. the rotor resistance will hit its maximum or minimum value to a high extent which will make the result worse.4. i. it can be seen that the 1. the rotor resistance can not follow its reference value and has to be limited to a higher and higher degree (i. The ﬂicker in the comparison is related to a system with a ﬁxed rotor resistance. 8.9 the relative ﬂicker contribution for the proposed controller for ﬁve different values of aR is shown.4 0.. On the other hand. For the case with αR equal 0. RR = RR . the parameter αR should be a function of 144 .e.10 and 8.6 1. Lower values of the relative ﬂicker imply a lower ﬂicker contribution and vice versa. However. Moreover.4 1.2 Flicker Reduction In Fig.5 rad/s in the ﬁgure. over the whole operating area. the ﬂicker contribution is lower at lower average wind speeds.5 rad/s 1 rad/s 1.9).

the reduction in the ﬂicker contribution will be dependent on the operating condition. However.6 0.4 1.4 1.2 0.2 0.6 1.5 rad/s 1 rad/s 1. 145 . The average wind speed is 14 m/s. difﬁcult.11.5 rad/s 2 rad/s Relative ﬂicker 1 0. Reduction in ﬂicker for different bandwidths of αR .2 0 5 10 15 20 25 Turbulence intensity [%] Fig. It was shown that it is possible to reduce the ﬂicker contribution by utilizing the derived rotor resistance control law with 40–80% depending on the operating condition.6 1.5 rad/s 1 rad/s 1. The average wind speed is 20 m/s. Reduction in ﬂicker for different bandwidths of αR .8 0. the non-linearity of the system will make an “optimal” reduction in ﬂicker. 1.10.5 rad/s 2 rad/s Relative ﬂicker 1 0.4 0.4 0.8 0.5 Conclusion A non-linear rotor resistance control law has been derived with the objective of minimizing the ﬂicker contribution of a stall-controlled ﬁxed-speed wind turbine to the grid. Moreover.25 rad/s 0. 8. both the average torque and turbulence intensity. 8.2 0 5 10 15 20 25 Turbulence intensity [%] Fig. 8. over the whole operating area.6 0.1. since the rotor resistance can be varied only within a limited range.25 rad/s 0.

146

Chapter 9 Conclusion

The electrical energy efﬁciency of wind turbine systems equipped with doubly-fed induction generators in comparison to other wind turbine generator systems has been investigated. It was found that the energy efﬁciency of a doubly-fed induction generator system is a few percentage units higher compared to a system using a cage-bar induction generator, controlled by a full-power converter. In comparison to a direct-driven permanent-magnet synchronous generator, controlled by a converter or a two-speed generator system the difference in energy efﬁciency was found to be small. Moreover, the converter losses of the doubly-fed induction generator can be reduced if the available rotor-speed range is made smaller. However, the aerodynamic capture of the wind turbine is reduced with a smaller rotor-speed range. This means that the increased aerodynamic capture that can be achieved by a larger converter has, thus, a greater impact than the increased converter losses. Finally, two methods to reduce the magnetizing losses of the doubly-fed induction generator system, have been investigated. It was found that the method, utilizing a Y-Δ switch in the stator circuit had the largest gain in energy, of the two investigated methods. In order to evaluate different methods of reducing the inﬂuence of the back EMF on the rotor current control loop, a general rotor current control law has been derived with the option of having feed-forward compensation of the back EMF and “active resistance.” It was found that the method that combines both the feed-forward compensation of the back EMF and the “active resistance” manages to suppress the inﬂuence of the back EMF on the rotor current best and was found to be the least sensitive to erroneous parameters. The choice of current control method is of greater importance if the bandwidth of the current control loop is low. Moreover, it has been shown that by using grid-ﬂux orientation, the stability and the damping of the system is independent of the rotor current, in contrast to the stator-ﬂux oriented system. Dynamic models of the DFIG wind turbines have been experimentally veriﬁed, with a 850-kW wind turbine. Simulations and experimental results of the dynamic response to symmetrical as well as unsymmetrical voltage sags of a DFIG wind turbine were presented. Simulations were carried out both with a full-order model, and also with a reduced-order (second-order) model. Both models produced acceptable results. Voltage sag ride-through capabilities of some different variable-speed wind turbines have been investigated and compared. A variable-speed wind turbine with a full-power converter 147

system can handle voltage sags very well. Two candidate methods for improving the voltage sag ride-through capability of DFIG variable-speed wind turbines have been investigated. One of the methods still suffers, at least initially, from high fault currents, while the other method seems to have similar dynamical performance as the full-power converter system. However, the control of the latter method is much more complicated than that of the fullpower converter system. In addition, the maximum torque that can be handled by the generator is reduced in proportion to the voltage sag. The energy production cost of the full-power converter system was found to be three percentage units higher than that of the ordinary DFIG system without ride through capability. The two DFIG candidate methods have approximately the same energy production cost, which is approximately 1.5 percentage units higher in comparison to the ordinary DFIG system. Finally, a non-linear rotor resistance control law has been derived with the objective of minimizing the ﬂicker contribution of a stall-controlled ﬁxed-speed wind turbine to the grid. It has been found that the ﬂicker contribution can be reduced with 40–80%, depending on the operating condition, with the derived control law.

9.1

Future Research

The following candidate topics are proposed for future research: • Development of a uniﬁed estimator for both stator-ﬂux and grid-ﬂux ﬁeld orientation. Since the ﬂux dynamics are poorly damped, a desired property would be a relatively good damping of the ﬂux dynamics. • More thorough dynamic, steady-state, and experimental analysis of the voltage sag ride-through systems for the DFIG wind turbine. In addition, it is essential to study the hardware conﬁguration of the voltage sag ride-through systems. • Development of mathematical models of wind turbines with voltage sag ride-through properties. Experimental evaluation of the developed models with commercial wind turbines with voltage sag ride-through properties. • Derivation of analytical expressions for the response of the DFIG to unsymmetrical voltage sags.

148

2002. Bollen. and K. Chombt. pp. Aalborg.. Aalborg Univ. [3] T. [11] T. Sharpe. “Energy optimal control of induction motor drives. of European Union Wind Energy Conference. Wiley & Sons. Nov. Licentiate Thesis 515L. Bossanyi.. “Grid requirements challenges for wind turbines. John [12] O. Denmark. o May.” IEEE Trans. “An overview of wind energy-status 2002. [6] J. Oct. Piscataway. 20–24. Ninth International Conference on Electrical Machines and Drives. 6. 20–21. “Instantaneous reactive power compensators comprising switching devices without energy storage components. 1–3. Kanazawa. Burton. S¨ der. Sweden. 2002.” in Proc. Thorborg. Suso tain. [5] H. Boldea and S. [7] I. Bongiorno. “Control of voltage source converters for voltage dip mitigation in shunt and series conﬁguration. H. 2001.” in Proc. NJ. no. Nasar. and E.” IEEE Trans.” Ph. Y. and A. no. 1999. [4] H.References [1] SIMPOW Power System Simulation Software. “Survey of variable speed operation of wind turbines. Sep. Denmark. Abrahamsen. Hylander. a a [2] F. pp. Power Electron. [10] M. 223–227. pp. N. CRC Press LCC.. Understanding Power Quality Problems: Voltags Sags and Interuptions. “Adjustable-speed operation of doubly fed machines in pumped storage power plants. “Control and performance of a doubly-fed induction machine intended for a ﬂywheel energy storage system. 406–409. A. 3. Carlson../Apr. Ind. and L. 1999.. Bolik.. Jan. 1. 20. Feb. Ackermann and L. G¨ teborg. Feb. Sweo den. dissertation. M. J. vol. pp. 2002. Jenkins. Wind Energy Handbook. vol. Ltd. vol. 2003. USA: IEEE Press. [9] M. Schreier. Brochure.” Chalmers University of Technology. Sweden. G¨ teborg. 2000. Bendl.” in Proc. 1996. 625–630. Energy Rev. no. LargeScale Integration Wind Power Transmission Networks Offshore Wind Farms. Work. 149 . V¨ ster˚ s. Sato. May/June 1984. 67–128. Nabae. [8] S. 2004. Billund. 17. Int. pp.D. D. 1–2. Akagi.” Renew. Electric Drives. Akagi and H. ABB Power Systems Analysis. Applicat. 109–116.

Available: http://www. 2. 1203–1206. no. 207–215. (2005) Driftuppf¨ ljning av vindkraftverk december 2004. C. Ind. “Dynamic analysis of current regulators for ac motors using complex vectors. 3. of the Fifth International Conference on Electrical Machines and Systems. Oct. Fahai. Carrillo. Power Electronics and Applications (EPE). 17.dk/nyheder/brochurer/Horns Rev GB.. Ekanayake. D. (2005. 1999. and A. Vandenput.” in Proc. 121–127.. pp. pp. Aug. 2623–2628. [21] J.. Cosgriff. pp. [17] ——. Stoev. Power Electron. M. [15] R. and L.se/varme/underlag/vstat0412. [19] DEWIND.) Horns rev offshore wind farm.dewind. Congwei. no. vol. Duarte.hornsrev. Nonlinear Control Systems. Available: //www.pdf http: [20] A. “Reference frames ﬁt for controlling pwm rectiﬁers. A. B. Degner.pdf [25] (2004) Enercon website. pp. Xudong. pp. 46. W. Datta and V. Haiqing.[13] L. 2003. Ind. 1424–1432. Sep. 2003. 56. 2001. Conference Record of the 1999 IEEE Industry Applications Conference. 35.pdf [24] Elsam. “Variable-speed wind power generation using doubly fed wound rotor induction machine-a comparison with alternative schemes. June 1999. W. 1999. vol. Phoenix.” IEEE Trans. vol. [22] J. 150 . McGraw-Hill. “Decoupled control of active and reactive power for a grid-connected doublyfed wound rotor induction machine without position sensors. L. Feij´ o. no. 3. vol. Shenyang. pp. Lorenz. AZ. [23] Elforsk. “Grid voltage fault proof doubly-fed induction generator system. Applicat. and R. Nov. France. Dec. Jenkins. USA. Toulouse. vol. S. Brochure.” IEEE Trans. vol. Brochure. 4. Wijnands. 67. 1958. and C. (2003.. L. T. pp.de/en/downloads/D8-2000-100-eng./Dec. 786–793. Cidr´ s. Holdsworth.) The D8 series. [14] R.” in Proc. pp. V. “Comparison of 5th order and 3rd order machine models for doubly fed induction generator (DFIG) wind turbines.” in Proc.enercon. vol. 18–20. 6. 414–421. Energy Conversion. [Online]. Dittrich and A. vol. 4. China. 48. and N. “A simple position-sensorless algorithm for rotor-side ﬁeld-oriented control of wound-rotor induction machine. [Online]. [18] F.” Electric Power Systems Research. B. o Available: http://www. 2001. L.” IEEE Trans. Aug. Electron.” Electric Power Systems Research. “Research of stability of double fed induction motor vector control system. Zwam. Ranganathan. 628–630. [Online]. Available: http://www. J. no. Sept. Jan. del Blanco. Oct.elforsk. 2000. Nov. 2002. “A third order model for the doubly-fed induction o a machine. [Online]. [16] ——.de/ [26] A.” IEEE Trans.

Available: http://www.” Risø National Laboratory. Rep. “Design of direct-driven permanent-magnet generators for wind turbines. “The rectiﬁers inﬂuence on the size of directo driven generators. Jan. [28] J. G¨ teborg.) Low voltage ride-thru technology. 2002. no. Dec.” in Proc. Denmark. E. no. o 1996. [39] ——. Roskilde. H.pdf [31] T. 77–83. Blaabjerg. [33] A. 1999. G¨ teborg. 1058–1066.pdf [34] ——.D. Nee. [38] L. of European Wind Energy Conference and Exhibition (EWEC´99). a 151 V¨ ster˚ s. Arulampalam. 2002. Sweden. Jan. Billund. “Synchronous generator and frequency converter in wind turbine applications: system design and efﬁciency. 1. Reglerteori: ﬂervariabla och olinj¨ ra metoder. L.) 3. Denmark. “A general algorithm for speed and position estimation of ac motors.[27] C. dissertation. Sørensen. and B. Harnefors and H. “Efﬁciency of three wind energy generator systems. Lund. 1996. Munk-Nielsen. Nov.gepower. 3. Electron. Grauers.publ/Abstracts/old/Grauers Lic. 1997. Harnefors. 2003. M. Feb.D. Ljung. “Validation of DFIG model using 1. May 1994.” Chalmers University of Technology. 2000. Ind. Sweden. ISBN 87-550-2743-8. pp. “Conceptual survey of generators and power electronics for wind turbines. publ/Abstracts/old/Grauers PhD Thesis. [Online]. dissertation. Fortmann. Grauers and S. Helle. Tech. Bindner.com/businesses/ge wind energy/en/downloads/ge lvrt brochure.” IEEE Trans. (2005. Stockholm.” Ph.chalmers. [32] A. vol. Glad and L. H. Fitzer. P. Chalmers University of Technology.com/prod serv/products/wind turbines/en/ downloads/ge 36 brochure. 1997. “On analysis.5 MW turbine for the analysis of its behavior during voltage drops in the 110 kV grid. Nice. Int. Mar. control and estimation of variable-speed drives.se/Publikationer/EMKE. [29] GE Wind Energy. Ritchie. 17. 650–657. 1–5. Available: http://www. Zurowski. 6. [36] L. (2005.elteknik.” Ph. M¨ lardalen University.. Risø-R-1205(EN).gepower. S.” IEEE Trans. pp.se/Publikationer/EMKE. 11. [37] L. A. Landstr¨ m. “Mitigation of saturation in dynamic voltage restorer connection transformers. [Online].6s offshore wind turbine. Energy Conversion. no. Sept. 47. [Online]. Barnes. Power Electron. vol. Brochure. [Online]. Nov. Large-Scale Integration Wind Power Transmission Networks Offshore Wind Farms. Royal Institute of Technology. Sweden: Department of Eleca a . Oct. Available: http:// o www.pdf [35] ——. Brochure..” in Proc.-P. France. Control of Variable-Speed Drives.” IEEE Trans. tronics. and R. vol.pdf [30] ——. 2001. Hansen.chalmers. pp.elteknik. (in Swedish). Sweden. F. Sweden: a Studentlitteratur. Work. Licentiate Thesis 175L. Available: http://www. Bak-Jensen..

8–10. Harnefors. USA: IEEE Press.” IEEE Trans. 152 . Pietil¨ inen. [52] M. Electron. and L. Understanding FACTS: Concepts and Technology of Flexible AC Transmission Systems.. 2. no. Oct. 133– 141. 48. no. A. Power Appl. [49] C. Huang. Hartford. 34. [46] B. 241–250. [41] L. Graz. A. Jan. Hopfensperger. July 2000. vol. Sept. M.-P. NJ. [48] B. and R. Hur. Schumacher. vol. IEC 61 000-21. Gyugyi. Sep. 2003.. 3. 5–7.” IEEE Trans. vol. Hartge and V. pp. Heller and W. 597–605. International Electrotechnical Commission Std. G. Nee.c. Ind. Energy Conversion. Piscataway. 2001. and K. Electron. vol. pp. [47] B. Applicat. Pai. Tegopoulos. Nov. Belgium. vol. no. Atkinson. 707–710. no. Oct. Benbouzid. Hingorani and L. vol. “Doubly-fed a. 794–803. H. pp.. “Stator-ﬂux-oriented control of a doubly-fed induction machine with and without position encoder. pp.-S. and F. pp. “Optimal efﬁciency slip-power recovery drive. of the 1997 IEEE International Conference on Control Applications.” IEE Proc. Atkinson.. Lakin. “Model-based current control of ac machines using the internal model control method. Power Appl.. “Torque-maximizing ﬁeld-weakening a control: design. Harnefors and H. Power Electronics and Applications (EPE). 18. CT. machines: classiﬁcation and comparsion. and D. [42] S. vol. Electr. 2001. 1997. [50] N. J. Huang.” in Proc. 1997. Hopfensperger. Ind. Aug. pp. Large-Scale Integration Wind Power Transmission Networks Offshore Wind Farms. Electr. E. D. “Stator ﬂux oriented control of a cascaded doubly-fed induction machine. 1. 342–348.” IEEE Trans. 1202–1210. Austria.” in Proc. 6. no. USA. S. [44] K. and R. “A fast dynamic dc-link power-balancing scheme for a PWM converter–inverter system. June 1988. 5. 1998.” IEE Proc. no. Work. [51] Measurement and assessment of power quality characteristics of grid connected wind turbines (11/2000). analysis. 2000. Jung. Hentabli. pp. 2. 161–168. “Ride-through capability of ENERCON-wind turbines.” in Proc.-J. 48. Ioannides and J. 4. D. G. 1. Brussels.. “Design of dynamic voltage restorer with disturbance-ﬁltering enhancement. Feb. pp./Feb.” IEEE Trans. Atkinson. Power Electron... Int. 235– 237. Hopfensperger and D.. Ind. Pinchon. J. Denmark. Gertmar. vol. A. pp. Aug.” IEEE Trans. 1999.” in Proc.[40] L. 27–29. “CGPC with internal model structure: Application to induction motor control. Lakin. Nam. 146. 2003. 2000. K. [43] M. “Stability analysis of doubly-fed induction machines in stator ﬂux reference frame. 147. Diedrichs. of 7th European Conference on Power Electronics and Applications.-J. 2001. Billund. and parameter selection. [45] N.

no. France. T. a dissertation.publ/Abstracts/old/Lindgren PhD Thesis. Austria. Control of Electrical Drives. “Power control of a doubly fed induction machine without rotational transducers. 2nd ed. Vilathgamuwa. G¨ teborg. Department of Electric Power Enginering. o 153 . H.: Prentice-Hall. pp. Oh. Lundberg. Kelber. “Performance comparison of wind park conﬁgurations. 1996. P. Santjer.” Chalmers University of Technology. and T. 3. “Aktive d¨ mpfung der doppelt-gespeisten Drehstrommaschine. Tech. 2000.” in Proc. R. Swea o den. 2.pdf [63] B. 3. Larsson. Sweden. Nice. 1990. and F. Rep. Lasson and P. Berlin. London: Academic Press. vol. [60] P. Johnsson. 305–310. May 2002. Choi. Aug. “Effect of neglecting stator transients in doubly fed induction generator models. vol. Wind Energy Systems..J. Tech. China..se/Publikationer/ EMKE. S. Mar. 1986. 1202–1210. Lindgren. Kim. Petru. 15–18. “Transformerless dynamic voltage restorer. 1998. pp. Germany: Springer- [62] M. Beijing. Kim.. [54] C. 951–955. L. 149. Technishen Universit¨ t Carolo-Wilhelmina. USA. o Nov. no.” Department of Electric Power Enginering. Y.” in Proc. Usaola. 27–29 2001. Ledesma and J. Graz. 1–5. 2001. Verlag. of 3rd International Conference on Power Electronics and Motion Control. [61] W. and D. Rep. Lundberg. and J. Khalil. 30R. 19. G¨ teborg.” Ph. S. pp. “Active damping of ﬂux oscillations in doublyfed ac machines using dynamic variation of the system’s structure. Aug. of European Wind Energy Conference and Exhibition (EWEC´99).elteknik. vol.” IEEE Trans. [65] S. Transmission and Distribution. H. 2. dissertation. G¨ teborg. 1999. “Grid impact of variable speed wind turbines. [57] P. pp. Thiringer.chalmers. Energy Conversion.” in Proc. Kelber and W.[53] G. 1985. Power Electronics and Applications (EPE). Chalmers University of Technology. Sweden.” IEE Proceedings Generation. Sørensen. “Trefas sj¨ lvkommuterade str¨ mriktare anslutna till a o eln¨ tet—En inledande studie. 459–461. Singular Perturbation Methods in Control: c Analysis and Design. H. Li. and C. Lundberg. 2003. Kim. Aug. Chalmers University of Technology.” International Journal of Renewable Energy Engineering. R. O’Reilly. Kokotovi´ .D. N. [55] C. 2000. B. Englewood Cliffs.. [64] S. K. June 2004. a [56] E. vol.. [Online]. “Modeling and control of voltage source converters connected to the grid. ˚ [58] A.D. Schumacher.” Ph. Leonhard. Aug. “Electrical limiting factors for wind energy installations in weak grids. H. Available: http://www. (in German). [59] A.

D. G¨ teborg. Dec. no. N. Tentzerakis. [70] M.se/ Publikationer/EMKE. Winnipeg. A. June.. 2002. IEEE Nordic Workshop on Power and Industrial Electronics (NORpie´2004). dissertation. G¨ teborg. 38. “On control of back-to-back converters and sensorless induction machine drives. [75] R..” in Proc. Nice. T. G. 154 . Sweden. (2005. 1998. Pietil¨ inen. May 2004. Rodriguez. pp. Cortes. Godfroid. pp. Ottersten.publ/Abstracts/2003/RolfPhD.” in Proc. Cairns. [Online]. 4. 985–990.. Robbins.) N80/2500 kW N90/2300 kW. “Control and testing of a dynamic voltage restorer (DVR) at medium voltage level. Boulaxis..” in Proc.elteknik. o 1–2 2004. Newman and D. Morel. and K. 360–368. [73] Nordex. Power Appl. June. J.nordex-online.pdf [74] R. July 1998.publ/Abstracts/old/otterstenLic/ottersten lic thesis.” IEEE Trans. May/June 2002. 23–27.” Ph. vol.chalmers. 5–8. vol. [67] N. 145. pp. 19. [Online]. “Vector control of a diesel-driven doubly fed induction machine for a stand-alone variable speed energy system. Applications and Design. Norway. [71] J. J. Ind. Pena. G¨ teborg. Ottersten. Jan.” Chalmers University of Technology. On-line help. “Voltage sag response of PWM rectia ﬁers for variable-speed wind turbines. and J.” IEE Proc. J. Hoffmann. [72] J. Nordic Wind Power Conference (NWPC). Nielsen. “Voltage dip ride through of doubly-fed generator equipped with active crowbar. Applicat. Manitoba. R. 2.. 783–785. 806–813.elteknik. 1999. Available: http://www. 1. Holmes.. [78] R.” in IEEE Annual Conference of the Industrial Electronics Society.se/Publikationer/ EMKE. France. pp. Blaabjerg. 3. Mirzaian. and F. Licentiate o Thesis 368L. Available: http://www. Mutschler and R. [Online]. Mohan. Available: http://www. Australia. Cardenas. and S. 14–16 2004. M. Papathanassiou. 3. vol. Nov. Sweden..pdf [76] ——. Petersson.[66] EMTDC Users’s Guide. Power Electron. Niiranen. 2000. no. A. [68] L. Nielsen.chalmers. G. H. Manitoba HVDC Research Centre Inc. USA: John Wiley & Sons. vol. G. “Double-fed induction machine: converter optimisation and ﬁeld oriented control without position sensor. H. 2002. Power Electronics Converter. New York. no. “Vector control of a double-sided PWM converter and induction machine drive. Asher. Undeland. 6–11.” IEEE Trans. “Voltage quality change by grid-connected wind turbines. Electr. vol. Mar.” in European Wind Energy Conference. Kauffmann. Sweden. 679–687.pdf [77] M. and W. pp. and P. Papadopoulos. Qld.com/ e/online service/download/ dateien/PB N80 GB. “Comparison of wind turbines regarding their energy generation. Newman. Clare. pp. P. Brochure. S. [69] P. 1995. T. o June 2003. 2002 IEEE 33rd Annual IEEE Power Electronics Specialists Conference. G. Chalmers University of Technology. A. Canada. “An integrated approach for the protection of series injection inverters. Trondheim.

semikron. [Online]. Denver.) SKiiP 2403GB172-4DW. [86] B. Nov. 2. R. Analog och digital reglerteknik. Dec. Bologna. [87] R. Data sheet.A.) SKiiP 1803GB172-3DW. R. (in Swedish). P¨ ller. Nov.html [91] ——. NY.” IEEE Trans.jsp?ﬁle=1261. Feb. G¨ teborg. 2001.S. Pena. Indonesia. 22–25.se/ Publikationer/EMKE. (2004. vol. 1132–1139. A. 231–241. “Doubly fed induction generator using back-toback PWM converters and its application to variable-speed wind-energy generation. “Active ﬂicker reduction from a sea-based 2. [84] M. 17.semikron.. Denpasar. Available: http://www. Data sheet. and J. Data sheet. Sweden. C. (2004. “Optimal active and reactive power control with the doubly-fed induction generator in the MW-class wind-turbines. May 1996. (2004. 2nd ed. [85] PSS/E-Introduction to Dynamic Simulation. Dec. [89] Semikron. pp. Italy. Asher. a (in German). “Analysis.) SKiiP 1203GB172-2DW.” IEE Proc.) SKiiP 513GD172-3DUL. Nordic Workshop on Power and Industrial Electronics. 1995. Data sheet. Blasco. CO. modeling and control of doubly-fed induction generators for wind turbines. (2004... Rabelo and W. G. IEEE Bologna Power Tech. Dec. Dec. Dec.elteknik. (2004. “Modeling of wind turbines for power system studies. 2002. International Conference on Power Electronics and Drives Systems. Electrische Machinen. Licentiate o Thesis 464L. Thiringer. and G.html 155 .com/internet/ds.” in Proc. 1376–1381. Denmark. Data sheet.[79] R.” in Proc. Available: http://www. U.. 2nd ed..) Thyristor SKT 2400. [Online]. Pena. “A cage induction generator using back to back PWM converters for variable speed grid connected wind energy system. pp. 2002. June.jsp?ﬁle=516. [Online]. Electr. pp.pdf [82] T.semikron. “Doubly-fed induction machine models for stability assessment of wind o farms. Basel/Stuttgart: Verlag Birkh¨ user. Clare. J.semikron. Power Appl. [Online]. Power Syst. 13–16. Cardenas.5 MW wind park connected to a weak grid. Available: http://www.” in Proc.jsp?ﬁle=1268. M. IEEE IECON’01.chalmers.jsp?ﬁle=1266. Available: http://www.html [93] ——. Clare. 1954. Available: http://www. Sweden: Studentlitteratur. Asher. [81] A.com/internet/ds. Manual. [83] ——.com/internet/ds. pp.html [92] ——. Hofmann. 143. [Online].html [90] ——. 1.” in Proc. Schmidtbauer.” Chalmers University of Technology. Available: http://www.semikron. vol.jsp?ﬁle=1264. Lund. Petru and T. 23–26 2003.publ/Abstracts/2003/AndreasLic. 2003. Petersson. [Online]. Oct. Inc. Aalborg. [80] R.com/internet/ds. vol. 53–58. Richter. Power Technologies. June. [88] B. 2001.com/internet/ds. Schenectady. vol. 2001.

com/pdf/produkter/AktuelleBrochurer/v120/V120%20UK. Available: a http://www. 10. Rep. 3. Applicat. [Online]. Nov. Sept. vol. pp. Linders. Spee.svk. 2004. 1989.se/web/biblshop. Lund. Energy Conversion. [104] Vestas. Thiringer and J. Wallace. pp. vol. pp. pp. “Flexible active and reactive power control strategy for a variable speed constant frequency generating system.” IEEE Trans. 25.pdf?OpenElement [99] Y.” Swedish Energy Agency. Available: http://www.” Swedish Energy Agency.pdf?OpenElement [98] ——.nsf/ FilAtkomst/ER62002. 472–478. Thomas and M.” IEEE Trans. July 1995. Power Electronics – in Theory and Practice. 21. [101] T.pdf/$FILE/ER62002. 4. 1991. “Control by variable rotor speed of a ﬁxed-pitch wind turbine operating in a wide speed range.vestas. [Online].pdf/$FILE/ER62002. Available: http://www. New [96] Svenska Kraftn¨ t. 2.” Electric Machines and Power Systems. Jersey. no. pp. Sweden: Studentlitteratur. Ding. 11–24. 1991. no.pdf [97] Swedish Energy Agency. R. UK. “Comparison of reduced-order dynamic models of induction machines. May) V52-850 kW. [103] K. Jan. (2004. “An internal model control structure in ﬁeld oriented controlled vsi induction motors. 1997. Wang and Y. 1. pp. [Online].5 MW Offshore leadership. no. 156 . Applied Nonlinear Control. a a a V¨ llingby. Li. of the 4th European Conference on Power Electronics and Applications. 8..nsf/ FilAtkomst/ER62002. “Modelling of induction machines for electric drives. Power Syst. Boidin. vol.pdf [106] A. Luomi. “Stability analysis of ﬁeld oriented doubly-fed induction machine drive based on computer simulation. [102] J. Tech. [100] T. 6. Rep. 2002. K. Rep. a “Aff¨ rsverket a svenska kraftn¨ ts a f¨ reskrifter o om drifts¨ kerhetsteknisk utformning av produktionsanl¨ ggningar. Slemon. 1993.) V120-4. Slotline and W. vol. 1.” IEEE Trans. and G. Brochure. USA: Prentice Hall.. [107] S. com/produkter/pdf/updates 020304/V52 UK.se/web/biblshop. Power Electron. ET 29:2004. vol. C. Oxford.[94] G.stem. ER 6:2002. 1993. 119–126. (2005. 1993.se/upload/3645/Foreskriftprod remiss. 8–10. Sept. applications and design methods. “The energy market 2004.” in Sixth International Conference on Electrical Machines and Drives). Feb. E. Xu.pdf [105] ——.-J. L.” in Proc. Tang and L. Ind.vestas.” IEEE Trans. Brochure.stem. 2001. Thiringer and J. Available: http://www./Dec. R. [Online]. “The brushless doubly-fed machine: its advantages. Upper Saddle River. Tech. 511–517. [Online]. 520–526. Thorborg. 16. pp. “Climate report 2001.” Svenska Kraftn¨ t. Sweden. vol. Tech. Alexander. 2004. [95] J.. 202–207. Available: http://www.. Draft version. no. 1126–1131. no..

K. Energy Conversion. 111–122. 1997. Williamson. Wei. Xu and C. vol.. vol.” IEEE Trans. 33. 157 . no. Ferreira. Wallace. “Transient model of a doubly excited reluctance motor. 304–310. pp. “Generalised theory of the brushless doubly-fed machine. 636–642. pp. [109] L. [111] D. Zhou and R. 6–10. Muljadi. 126–133. S. no. Ind. Spee. A. [110] L. “Torque and reactive power control of a doubly fed induction machine by position sensorless scheme. Liang. 1997. 31. no. 1. part 1: Analysis.” in Proc. C. Lipo. pp. Mar.. 1991. A. Ind. 2. IEEE Industry Applications Conference.. pp. and A. Zinger and E. 1996. Electr. Nov. vol. 144. 6. “Field oriented control development for brushless doubly-fed machines. 6.” IEEE Trans. Applicat.” IEEE Trans.. no. May/June 1995. “Annualized wind energy improvement using variable speeds. F. Mar. Applicat.” IEE Proc. Oct. [112] D. Power Appl.[108] S. Xu. 3. 1./Dec. pp. and T. vol. vol. 1444–1447.

158 .

Ts Tsample V V v.Appendix A Nomenclature Symbols Ar C Cp E Eg . v swept area capacitor power coefﬁcient back EMF grid-voltage modulus and space vector controller probability density function transfer function gearbox ratio steady-state complex-valued current current modulus and space vector inertia √ −1 coefﬁcients in the rotor current control law proportional and integral gain inductance Laplace transform inverse Laplace transform number of pole pairs stator-to-rotor turns ratio active power d/dt reactive power resistance apparent power slip electromechanical and shaft torque sample time steady-state complex-valued voltage remaining voltage voltage modulus and space vector 159 . E g F f (w) G gr I i. ki L L L−1 np ns /nr P p Q R S s Te . i J j kE . kR kp .

α β λ ρ Ψ ψ ω1 . θ1 ω2 ωg . θg ωr ˜ ˆ closed loop bandwidth pitch angle tip-speed ratio density of air or bandwidth of PLL ﬂux space vector or steady-state complex-valued ﬂux ﬂux modulus synchronous frequency and angle slip frequency grid frequency and angle (electrical) rotor speed of generator error estimated Superscripts avg max min s pk ref average maximum minimum stator-oriented reference frame peak reference Subscripts cl co d f g GB m M mech n nom R r s sw t q p λ σ closed loop cut off real part of synchronous-frame space vector (grid-) ﬁlter or ﬂux grid gearbox mutual mutual (Γ representation) mechanical negative sequence nominal rotor (Γ representation) rotor stator switch turbine imaginary part of synchronous-frame space vector positive sequence leakage leakage (Γ-representation) 160 .

u.Abbreviations DFIG EMF FSIG G GSC IG IGBT IMC LLF MSC PLL PMSG p. PWM RMS SG SLGF TLGF VSIG WT doubly-fed induction generator electromotive force ﬁxed-speed wind turbine with an induction generator generator grid-side converter induction generator insulated gate bipolar transistor internal model control line-to-line fault machine-side converter phase-locked loop permanent-magnet synchronous generator per unit pulse width modulation root mean square synchronous generator single-line-to-ground fault two-lines-to-ground fault variable-speed wind turbine with an induction generator and a full-power converter wind turbine 161 .

162 .

is used.3. TABLE B.21 Ω TABLE B. Base voltage (phase-neutral) Base current Base frequency Base impedance Vb Ib ωb Zb = Vb /Ib 400 V 1900 A 2π · 50 Hz 0. 198 p.009 p.0022 Ω 0.u.1 mH ⇔ ⇔ ⇔ ⇔ ⇔ ⇔ ⇔ ⇔ ⇔ 0.p−p In fn Pn np 690 V 1900 A 50 Hz 2 MW 2 TABLE B.u. Stator resistance Rotor resistance Rotor resistance (Γ equivalent) Stator leakage inductance Rotor leakage inductance Leakage inductance (Γ equivalent) Magnetizing resistance Magnetizing inductance Magnetizing inductance (Γ equivalent) Rs Rr RR Lsλ Lrλ Lσ Rm Lm LM 0.05 mH 0.4 p.1.07 p.27 p. 4. base values.u.9 mH 3.1.0018 Ω 0.18 mH 42 Ω 2.5 p.3 the nominal values. PARAMETERS OF THE INDUCTION MACHINE .18 p.u.0093 p.01 p.6 p.2.u. 0. 0.12 mH 0. 0. and the parameters of the DFIG are shown respectively.u.u. 163 . 0.2. A dc-link capacitor of Cdc = 53 mH = 3.u. Table B. 0. N OMINAL VALUES OF THE DFIG. and in Table B. BASE VALUES . Rated voltage (Y) Rated current Rated frequency Rated power Number of pole pairs Vn. In Table B.Appendix B Data and Experimental Setup B.u.u.0019 Ω 0.1 Data of the DFIG These data and parameters of the DFIG are used throughout the thesis if not otherwise stated. 4.

two measurement boxes. B. the converter operates as a back-to-back converter..5.1. Laboratory setup. Fig. base values. The measurement system consists of one ﬁlter box and one computer equipped with the LabView software. The control laws were all written in the C-language and downloaded to the DSP-unit (Texas TMS320c30). i.1 Data of the Induction Generator In Table B. B. In the measurement boxes voltages and dc mach. of the induction machine.2. IM v. Normally.2.6 the nominal values. for signals that is desired to be fed to the measurement computer. for measurement signals. 164 . and 8 analog output channels. There is also a resolver that measure the rotor position.4. The loading dc machine is fed through a thyristor inverter and could be both speed or torque controlled. B. The DSP-unit has 16 analog input channels.1.B. Table B. The voltage references to the converter are modulated digitally and via optic ﬁbers sent to the converter.1 shows a principle sketch of the laboratory setup. One measurement box is attached to the stator circuit while the other measure the rotor circuit. one voltage source converter.e. and in Table B.2 Laboratory Setup The laboratory setup consists of one slip-ringed wound rotor induction machine. computer Fig. A more thorough description of the laboratory set up can be found in [75]. but during the experiments the converter was directly fed by a dc source of 450 V dc. and the parameters of the laboratory DFIG are shown respectively. Thick lines indicates cables with power while dashed lines implies measurements signals. Data of the induction machine is given in Section B. θr . When running the machine as doubly-fed the stator circuit is directly connected to the grid (during the experiments in this thesis the stator circuit was connected to a 230-V. currents are measured. Although the converter here is fed directly from a dc source. With this system it is possible to measure up to 16 channels. i θr ac supply dc supply Converter DSP Meas. from the measurements boxes or from the DSP unit. note that the nominal voltage of the induction machine is 380 V). 50-Hz source. it is possible to run it as a back-to-back converter. i v. one digital signal processing (DSP) system and one measurement computer.

65 mH 1. 165 . 0. located at the inland (≈ 100 km from the west coast) in the southern part of Sweden.9 p. 2.6.0–31.u. The wind turbine is located in a ﬂat surroundings and is connected to the 10-kV distribution grid via a transformer. The currents and voltages are measured using transformers. BASE VALUES .68 mH 224 Ω 46. 178 p.2 for a picture of the turbine and the data acquisition computer.89 TABLE B. DATA OF VESTAS V-52 850 K W WT [104].u.115 Ω 0.6 mH 0.93 p. 0.184 Ω 1.0369 p. Rated voltage (Y) Rated current Rated frequency Rated rotor speed Rated power Rated torque Power factor Vn.0230 p. which transforms the voltage to the wind-turbine voltage of 690 V.u.u. In Table B. Stator resistance Rotor resistance Stator leakage inductance Rotor leakage inductance Magnetizing resistance Magnetizing inductance Inertia Rs Rr Lsλ Lrλ Rm Lm J 0.p−p In fn nn Pn Tn 380 V 44 A 50 Hz 1440 rpm 22 kW 145 Nm 0. See Fig.7.TABLE B.104 p.u.3 Jung Data Acquisition Setup The experiments were made on a VESTAS V-52 850 kW WT.5. the stator currents are also measured directly using LEM modules. Base voltage (phase-neutral) Base current Base frequency Base impedance Vb Ib ωb Zb = Vb /Ib 220 V 44 A 2π · 50 Hz 5Ω TABLE B.u. B. PARAMETERS OF THE INDUCTION MACHINE . B. 44. which transform TABLE B. 0.106 p. N OMINAL VALUES OF THE INDUCTION GENERATOR .4.7 some data of VESTAS V-52 850 kW WT is given.u.334 kgm2 ⇔ ⇔ ⇔ ⇔ ⇔ ⇔ ⇔ 0. Rated voltage (Y) Rated power Rotor diameter Rotor speed Cut-in wind speed Nominal wind speed Maximum wind speed 690 V 850 kW 52 m 14.0 rpm (26 rpm) 4 m/s 16 m/s 25 m/s the current to 5 A and the voltage to 110 V. In addition.

2. Jung wind turbine and the data acquisition computer.Fig. B. 166 .