Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Modeling, Control and Analysis of Doubly Fed Induction Generator Based On Wind Turbine System

Download as pdf or txt
Download as pdf or txt
You are on page 1of 116

MODELING, CONTROL AND ANALYSIS OF

DOUBLY FED INDUCTION GENERATOR BASED


ON WIND TURBINE SYSTEM

2022
MASTER THESIS
ELECTRICAL-ELECTRONICS ENGINEERING

Samatar ABDI YONIS

Thesis Advisor
Prof. Dr. Ziyodulla YUSUPOV
MODELING, CONTROL AND ANALYSIS OF DOUBLY FED INDUCTION
GENERATOR BASED ON WIND TURBINE SYSTEM

Samatar ABDI YONIS

T.C.
Karabuk University
Institute of Graduate Programs
Department of Electrical-Electronics Engineering
Prepared as
Master Thesis

Thesis Advisor
Prof. Dr. Ziyodulla YUSUPOV

KARABUK
June 2022
I certify that in my opinion the thesis submitted by Samatar ABDI YONIS titled
“MODELING, CONTROL AND ANALYSIS OF DOUBLY FED INDUCTION
GENERATOR BASED ON WIND TURBINE SYSTEM” is fully adequate in scope
and in quality as a thesis for the degree of Master of Science.

Prof. Dr. Ziyodulla YUSUPOV ..........................


Thesis Advisor, Department of Electrical-Electronics Engineering

This thesis is accepted by the examining committee with a unanimous vote in the
Department of Electrical-Electronics Engineering as a Master of Science thesis. June
16, 2022

Examining Committee Members (Institutions) Signature

Chairman : Assoc. Prof. Dr. Adem DALCALI (BANU) ..........................

Member : Prof. Dr. Ziyodulla YUSUPOV (KBU) ..........................

Member : Assist. Prof. Dr. Hüseyin ALTINKAYA (KBU) ..........................

The degree of Master of Science by the thesis submitted is approved by the


Administrative Board of the Institute of Graduate Programs, Karabuk University.

Prof. Dr. Hasan SOLMAZ ..........................


Director of the Institute of Graduate Programs

ii
“I declare that all the information within this thesis has been gathered and presented
in accordance with academic regulations and ethical principles and I have according
to the requirements of these regulations and principles cited all those which do not
originate in this work as well.”

Samatar ABDI YONIS

iii
ABSTRACT

M. Sc. Thesis

MODELING, CONTROL AND ANALYSIS OF DOUBLY FED INDUCTION


GENERATOR BASED ON WIND TURBINE SYSTEM

Samatar ABDI YONIS

Karabuk University
Institute of Graduate Programs
The Department of Electrical-Electronics Engineering

Thesis Advisor:
Prof. Dr. Ziyodulla YUSUPOV
June 2022, 96 pages

In the beginning of the 21st century, there was a growing interest in renewable energy
sources particularly the wind energy for power generation. The researchers aimed to
enhance the development of the design parameters for wind energy generations.
However, doubly fed induction generator (DFIG) has been frequently used in wind
turbines because of variable speed operations. DFIG has some merits such as, cost
effectiveness, higher efficiency and reliability and the fact that it operates in both
standalone and when it’s connected to the grid. Complete modeling, control and
analysis of DFIG based on wind turbine system has been developed in this thesis.

iv
This thesis includes detailed developments of the wind turbine model such as, wind
speed, aerodynamic system, drive train and generator model. Moreover, the dynamic
model of DFIG in a rotating 𝑑𝑞 reference frame has been derived. Vector control
approach of electrical machines is employed in the whole system. The complete
system which consists of models and sub models are integrated and simulated using
MATLAB/Simulink environment. Furthermore, to get the maximum output power
from the wind turbine model, it’s necessary to design control strategy. This includes
rotor side converter (RSC) control, proportional-integral (PI) control, indirect speed
control and grid side converter (GSC) control. PI controller has been designed in order
the current components follow their references and the current loops has been tuned
until the desired value is obtained. The RSC controls torque, flux and active and
reactive power. The indirect speed control which is used to drive the torque to follow
the maximum power curve is designed at the RSC control. The function of GSC
control is to maintain the DC bus voltage constant irrespective of the region of
operation and it operates at a unit power factor.

Moreover, the steady-state analysis for the most representative magnitudes of DFIG
based wind turbine has been investigated in MATLAB. Correspondingly, the steady-
state reactive power capability of DFIG is examined. Two case study is considered
during steady-state analysis. In case 1, reactive stator power is set at zero which results
the value of the maximum stator current to be lowered. In case 2, direct rotor current
is set at zero which provides lower stator current, but the reactive power became high.

Another objective of the study is to compare the performances of most representative


magnitudes of DFIG based wind turbine using PI controller and the proposed method
in terms of settling time. The simulated result demonstrates that the performances of
the proposed method improves and it has less perturbations compared to the PI
controller. Moreover, the proposed method gives strong robustness for controlling the
system with variable parameters.

The dynamic analysis of current loops behavior in a wind turbine based DFIG has been
studied. It is summarized that, to follow properly the dynamic performances of DFIG,
it is not required to reach the saturations for the voltage references. Besides, the

v
behavior of the system under different grid fault conditions such as, symmetrical
(balanced) and asymmetrical (unbalanced) voltage dips using crowbar protection are
analyzed. The simulation result confirms the effectiveness of the proposed system.

Key Words : Wind turbine, doubly fed induction generator, indirect speed control,
dynamic analysis, symmetrical, asymmetrical, crowbar protection.
Science Code : 90526

vi
ÖZET

Yüksek Lisans Tezi

ÇİFT BESLEMELİ ASENKRON GENERTÖRLER TABANLI RÜZGAR


TÜRBİNİN MODELLENMESİ, KONTROLÜ VE ANALİZİ

Samatar ABDI YONIS

Karabük Üniversitesi
Lisansüstü Eğitim Enstitüsü
Elektrik-Elektronik Mühendisliği Anabilim Dalı

Tez Danışmanı:
Prof. Dr. Ziyodulla YUSUPOV
Haziran 2022, 96 sayfa

Yirmi birinci yüzyılın başında, yenilenebilir enerji kaynaklarına, özellikle de elektrik


üretimi için rüzgar enerjisine artan bir ilgi vardı. Araştırmacılar, rüzgar enerjisi
nesilleri için tasarım parametrelerinin gelişimini geliştirmeyi amaçlıyor. Ancak çift
beslemeli asenkron generatörler (ÇBAG), değişken hızlı işlemler nedeniyle rüzgar
türbinlerinde sıklıkla kullanılmaktadır. ÇBAG'nin maliyet etkinliği, daha yüksek
verimlilik ve güvenilirlik ve hem bağımsız hem de şebekeye bağlıyken çalışması gibi
bazı avantajları vardır. Bu tezde, ÇBAG'nin rüzgar türbini sistemine dayalı komple
modellenmesi, kontrolü ve analizi geliştirilmiştir.

Bu tez, rüzgar hızı, aerodinamik sistem, aktarma organları ve jeneratör modeli gibi
rüzgar türbini modelinin detaylı geliştirmelerini içermektedir. Ayrıca, dönen bir dq
referans çerçevesinde ÇBAG'nin dinamik modeli türetilmiştir. Tüm sistemde elektrik
makinelerinin vektör kontrol yaklaşımı kullanılmıştır. Modeller ve alt modellerden

vii
oluşan sistemin tamamı MATLAB/Simulink ortamı kullanılarak entegre edilmiş ve
simüle edilmiştir. Ayrıca rüzgar türbini modelinden maksimum çıkış gücü elde etmek
için kontrol stratejisinin tasarlanması gerekmektedir. Buna rotor tarafı dönüştürücü
(RTD) kontrolü, oransal-integral (OI) kontrolü, dolaylı hız kontrolü ve şebeke tarafı
dönüştürücü (ŞTD) kontrolü dahildir. OI denetleyici, akım bileşenlerinin referanslarını
takip etmesi için tasarlanmış ve istenen değer elde edilene kadar akım döngüleri
ayarlanmıştır. RTD, torku, akıyı ve aktif ve reaktif gücü kontrol eder. Maksimum güç
eğrisini takip etmek için torku sürmek için kullanılan dolaylı hız kontrolü, RTD
kontrolünde tasarlanmıştır. ŞTD kontrolünün işlevi, çalışma bölgesinden bağımsız
olarak DA bara voltajını sabit tutmaktır ve bir birim güç faktöründe çalışır.

Ayrıca, MATLAB'de ÇBAG tabanlı rüzgar türbininin en temsili büyüklükleri için


kararlı durum analizi incelenmiştir. Buna uygun olarak, ÇBAG'nin kararlı durum
reaktif güç kapasitesi incelenir. Kararlı durum analizi sırasında iki vaka çalışması ele
alınır. Durum 1'de, reaktif stator gücü sıfıra ayarlanır ve bu, maksimum stator akımının
değerinin düşürülmesiyle sonuçlanır. 2. durumda, doğrudan rotor akımı, daha düşük
stator akımı sağlayan sıfıra ayarlanır, ancak reaktif güç yükselir.

Çalışmanın bir diğer amacı, ÇBAG tabanlı rüzgar türbininin en temsili


büyüklüklerinin performanslarını OI kontrolör ve önerilen yöntemin oturma süresi
açısından karşılaştırmaktır. Simüle edilen sonuç, önerilen yöntemin performanslarının
arttığını ve OI denetleyiciye kıyasla daha az bozulmaya sahip olduğunu
göstermektedir. Ayrıca önerilen yöntem, sistemi değişken parametrelerle kontrol
etmek için güçlü bir sağlamlık sağlar.

Bir rüzgar türbini tabanlı ÇBAG'de akım döngülerinin davranışının dinamik analizi
incelenmiştir. ÇBAG'nin dinamik performanslarını doğru bir şekilde takip etmek için
voltaj referansları için doygunluklara ulaşılmasına gerek olmadığı özetlenmiştir.
Ayrıca, levye koruması kullanılarak simetrik (dengeli) ve asimetrik (dengesiz) gerilim
düşüşleri gibi farklı şebeke arıza koşulları altında sistemin davranışı analiz edilir.
Simülasyon sonucu, önerilen sistemin etkinliğini doğrulamaktadır.

viii
Anahtar Kelimeler : Rüzgar türbini, çift beslemeli endüksiyon jeneratör, dolaylı hız
kontrolü, dinamik analiz, simetrik, asimetrik, levye koruması.
Bilim Kodu : 90526

ix
ACKNOWLEDGMENT

First and foremost, I would like to express my special thanks to the greatest of all time,
ALLAH, the most Gracious, the most Merciful, who has given me the health, strength
and patience to complete my Master study.

I would like to give thanks to my supervisor, Prof. Dr. Ziyodulla YUSUPOV, from the
bottom of my heart for his support and guidance, outstanding advice and patience
during my Master journey.

I would also like to thank my committee members, Assoc. Prof. Dr. Adem DALCALI
and Assist. Prof. Dr. Hüseyin ALTINKAYA for their time dedicated to examine and
review my work.

Finally, my special gratitude goes to my parents and my brothers, Mr. Saad and Mr.
Mahamed and of course my sister, Mrs. Hibo, for their continuous support.

x
CONTENTS
Page
APPROVAL ................................................................................................................ ii
ABSTRACT ................................................................................................................ iv
ÖZET ......................................................................................................................... vii
ACKNOWLEDGMENT .............................................................................................. x
CONTENTS ................................................................................................................ xi
LIST OF FIGURES .................................................................................................. xiv
LIST OF TABLES ................................................................................................... xvii
SYMBOLS AND ABBREVIATIONS INDEX ..................................................... xviii

PART 1 ........................................................................................................................ 1
INTRODUCTION ....................................................................................................... 1
1.1. MOTIVATION AND BACKGROUND ......................................................... 2
1.2. PURPOSE AND CONTRIBUTION ................................................................ 3
1.3. METHODOLOGY OF THE THESIS ............................................................. 3
1.4. LITERATURE REVIEW ................................................................................. 4
1.5. THESIS STRUCTURE .................................................................................... 7

PART 2 ........................................................................................................................ 9
WIND ENERGY SYSTEM ......................................................................................... 9
2.1. WIND ENERGY CONVERSION SYSTEM .................................................. 9
2.2. WIND TURBINE SYSTEM .......................................................................... 10
2.2.1. Fixed Speed Wind Turbines ................................................................... 11
2.2.2. Variable Speed Wind Turbines .............................................................. 11
2.3. WIND TURBINE MODEL ........................................................................... 12
2.3.1. Wind Speed Model ................................................................................. 14
2.3.2. Aerodynamic Model ............................................................................... 16
2.3.3. Drive Train Model .................................................................................. 17
2.3.4. Generator Model ..................................................................................... 18

xi
Page
PART 3 ...................................................................................................................... 20
DYNAMIC MODELING OF DOUBLY FED INDUCTION GENERATOR ......... 20
3.1. DFIG MODEL IN ABC REPRESENTATION ............................................ 20
3.2. THE TRANSFORMATION TECHNIQUE .................................................. 22
3.2.1. The Clarke Transformation .................................................................... 23
3.2.2. The Park Transformation ........................................................................ 25
3.3. THE DYNAMIC MODEL OF DFIG IN DQ REPRESENTATION ............ 26
3.4. CONVERTER AND DC LINK MODEL ...................................................... 29

PART 4 ...................................................................................................................... 32
CONTROL DESIGN OF DFIG BASED WIND TURBINE .................................... 32
4.1. RSC CONTROL ............................................................................................ 33
4.1.1. Transfer Function for the RSC ............................................................... 35
4.2. GSC CONTROL ............................................................................................ 37
4.2.1. Transfer Function for the GSC ............................................................... 39
4.3. PHASE LOCKED LOOP............................................................................... 42
4.4. CONTROL STRATEGY ............................................................................... 42
4.4.1. Proportional and Integral Control ........................................................... 42
4.4.2. Indirect Speed Control ............................................................................ 46
4.5. STEADY-STATE ANALYSIS OF THE DFIG ............................................ 47
4.5.1. Case 1: When Reactive Stator Power is Zero ......................................... 47
4.5.2. Case 2: When Direct Rotor Current is Zero ........................................... 49

PART 5 ...................................................................................................................... 50
ANALYSIS OF DFIG BASED WIND TURBINE DURING VOLTAGE DIPS ..... 50
5.1. SYMMETRICAL VOLTAGE DIPS ............................................................. 50
5.1.1. Crowbar Protection ................................................................................. 51
5.2. ASYMMETRICAL VOLTAGE DIPS .......................................................... 53

PART 6 ...................................................................................................................... 55
SIMULATION RESULTS ........................................................................................ 55
6.1. SIMULATION OF THE WIND TURBINE MODEL ................................... 55

xii
Page
6.1.1. Simulation of the Wind Speed Model .................................................... 55
6.1.2. Simulation of the Aerodynamic Model .................................................. 56
6.2. SIMULATION OF THE DFIG SYSTEM ..................................................... 57
6.3. STEADY-STATE ANALYSIS OF DFIG ..................................................... 58
6.4. SIMULATION OF DFIG BASED WIND TURBINE SYSTEM ................. 60
6.4.1. Simulation Results using PI Controller .................................................. 60
6.4.2. Simulation Results of the Proposed Method .......................................... 63
6.4.3. Comparison between Typical PI Controller and Proposed Method ....... 65
6.5. IMPLEMENTATION OF DFIG BASED WIND TURBINE SYSTEM....... 66
6.5.1. Implementation of the RSC Control ....................................................... 66
6.5.2. Implementation of the GSC Control....................................................... 68
6.6. DYNAMIC ANALYSIS OF CURRENT LOOPS BEHAVIOR ................... 69
6.7. PERFORMANCE ANALYSIS UNDER VOLTAGE DIPS ......................... 71
6.7.1. Symmetrical Voltage Dips...................................................................... 71
6.7.2. Asymmetrical Voltage Dips ................................................................... 74

PART 7 ...................................................................................................................... 76
CONCLUSION .......................................................................................................... 76
7.1. CONTRIBUTIONS OF THESIS ................................................................... 77
7.2. FUTURE WORK ........................................................................................... 77
REFERENCES .......................................................................................................... 79

APPENDIX A ............................................................................................................ 84
APPENDIX B ............................................................................................................ 86
APPENDIX C ............................................................................................................ 91
APPENDIX D ............................................................................................................ 94

RESUME ................................................................................................................... 96

xiii
LIST OF FIGURES
Page
Figure 2.1. Shows components of grid-connected WECS ......................................... 9
Figure 2.2. Upwind and downwind type wind turbines ........................................... 10
Figure 2.3. Configuration of fixed speed wind turbine. ........................................... 11
Figure 2.4. Scheme of variable speed synchronous generator. ................................ 12
Figure 2.5. Schematic diagram of Mitsubishi MWT 92 wind turbine ..................... 13
Figure 2.6. Shows a typical block design of the wind turbine model ...................... 13
Figure 2.7. Aerodynamic system ............................................................................. 17
Figure 2.8. Drive train model. .................................................................................. 18
Figure 3.1. Shows stationary abc reference frame. .................................................. 21
Figure 3.2. DFIG equivalent circuits. ...................................................................... 21
Figure 3.3. The Clarke transform representation ..................................................... 23
Figure 3.4. Transform representation between stator and rotor parameters. ........... 24
Figure 3.5. The representation of Park transforms................................................... 25
Figure 3.6. General diagram of the transformation techniques. .............................. 26
Figure 3.7. Electrical circuit for DFIG model in dq representation. ........................ 27
Figure 3.8. Power circuit of DFIG in wind turbine.................................................. 29
Figure 3.9. DFIG power converter circuit................................................................ 30
Figure 3.10. DC link model ....................................................................................... 31
Figure 4.1. Phasor diagram of a synchronous rotating dq reference frame. ............ 33
Figure 4.2. Control design of the RSC ..................................................................... 35
Figure 4.3. Equivalent second order system of a closed loop current control with
regulator ................................................................................................ 35
Figure 4.4. Bode plot diagram of the closed loop for RSC...................................... 36
Figure 4.5. Power flow design. ................................................................................ 37
Figure 4.6. Grid voltage alignment with the d-axis. ................................................ 37
Figure 4.7. Control block design of the GSC ........................................................... 38
Figure 4.8. Current loops for the L filter.................................................................. 39
Figure 4.9. Bode plot diagram of both RSC and GSC. ............................................ 41

xiv
Page
Figure 4.10. Step response of the GSC. ..................................................................... 41
Figure 4.11. Classical PLL block design ................................................................... 42
Figure 4.12. PI with rotor speed ................................................................................. 43
Figure 4.13. PI with direct rotor current and voltage. ................................................ 44
Figure 4.14. PI in the GSC. ........................................................................................ 44
Figure 4.15. Torque coefficient versus lambda and power curve. ............................. 44
Figure 4.16. MPPT stability curve. ............................................................................ 46
Figure 4.17. Indirect speed control strategy. .............................................................. 47
Figure 5.1. (a) Equivalent circuit for voltage dips analysis of DFIG and (b) Space
vector phasor diagram as sub synchronous mode ................................. 51
Figure 5.2. (a) DFIG system supplied with three phase DC crowbar protection and
(b) one phase equivalent circuit of DFIG system when the crowbar is
activated ................................................................................................ 52
Figure 5.3. Dual vector control strategy of a DFIG ................................................. 53
Figure 5.4. Sequence decomposition for the current loops ...................................... 54
Figure 6.1. Wind speed model at 7.5 m/s. ............................................................... 55
Figure 6.2. Wind speed profile. ............................................................................... 56
Figure 6.3. Simulink design of the aerodynamic model .......................................... 56
Figure 6.4. DFIG Simulink design. .......................................................................... 57
Figure 6.5. Magnitudes of the DFIG with the rotor speed and electromagnetic torque
............................................................................................................... 58
Figure 6.6. Steady-state analysis of DFIG. .............................................................. 59
Figure 6.7. Performance of DFIG with PI control at different wind speed. ............ 60
Figure 6.8. Typical PI controller design of the DFIG system. ................................. 61
Figure 6.9. Steady-state response of DFIG with typical PI controller at 7.5 m/s. ... 62
Figure 6.10. Proposed Simulink design of DFIG system. ......................................... 63
Figure 6.11. Performance of the proposed controller at different wind speed........... 63
Figure 6.12. Steady-state performance of the proposed controller at 7.5 m/s. .......... 64
Figure 6.13. Steady-state performance with PI and proposed method at 7.5 m/s...... 65
Figure 6.14. Overall implementation of DFIG based wind turbine. .......................... 66
Figure 6.15. Indirect speed control. ........................................................................... 67

xv
Page
Figure 6.16. Steady-state response for the magnitudes of RSC at different wind speeds
............................................................................................................... 67
Figure 6.17. Steady-state response for the magnitudes of GSC at different wind speed
............................................................................................................... 68
Figure 6.18. Most representative magnitudes of a wind turbine based DFIG in the RSC
............................................................................................................... 69
Figure 6.19. Most representative magnitudes of a wind turbine based DFIG in the GSC.
............................................................................................................... 70
Figure 6.20. Stator flux and crowbar current under symmetrical voltage dip. .......... 71
Figure 6.21. Performance of RSC magnitudes under symmetrical voltage dips. ...... 72
Figure 6.22. Performance of GSC magnitudes under symmetrical voltage dips. ...... 73
Figure 6.23. Stator flux and crowbar current under asymmetrical voltage dips. ....... 74
Figure 6.24. Performances of RSC magnitudes under asymmetrical voltage dips. ... 74
Figure 6.25. Performances of GSC magnitudes under asymmetrical voltage dips ... 75
Figure Appendix B.1. RSC control block.………………………………………....87
Figure Appendix B.2. GSC control block. ............................................................... 87
Figure Appendix B.3. Cancellations of coupling terms. .......................................... 88
Figure Appendix B.4. d/q to D/Q transformation. ................................................... 88
Figure Appendix B.5. D/Q to abc transformation. ................................................... 88
Figure Appendix B.6. abc to D/Q transformation.................................................... 89
Figure Appendix B.7. D/Q to d/q transformation. ................................................... 89
Figure Appendix B.8. PLL angle calculator. ........................................................... 89
Figure Appendix B.9. Third harmonic injection. ....................................................... 90

xvi
LIST OF TABLES
Page
Table 2.1. Indicates different value of z_o for various topology type [43]. .............. 15
Table 4.1. The values of wind speed and the generated power………………………45
Table 4.2. Steady-state magnitudes of the DFIG for Qs=0 [55]. ............................... 48
Table 4.3. Steady-state magnitudes of the DFIG for idr=0 [55]. ............................... 49
Table 6.1. Simulated results of the DFIG with PI controller at different wind speed...
………………………………………………………………………….61
Table 6.2. Settling time of DFIG components with typical PI at different wind speed
................................................................................................................ 62
Table 6.3. Results of the DFIG with the proposed method at different wind speeds...
................................................................................................................ 64
Table 6.4. Settling time of the proposed method at different wind speed. ................ 65
Table Appendix A. 1. Initial parameters of DFIG and wind turbine………………..85

xvii
SYMBOLS AND ABBREVIATIONS INDEX

SYMBOLS

F : frequency
Ps : stator active power
Qs : reactive stator power
V : stator voltage
I : stator current
smax : maximum slip
Rs : stator resistance
Ls : stator inductance
n : rotational speed
Ts : sample time
Vbus : dc-bus voltage
Rr : rotor resistance
Lr : rotor inductance
N : gearbox ratio
U : stator rotor turns ratio
Lm : mutual inductance
Fs : stator flux
D : damping ratio
Vw : wind speed
J : moment of inertia
fsw : switching frequency
idr : direct rotor current
idg : direct grid current
iqr : quadrature rotor current
iqg : quadrature grid current

xviii
ABBREVIATIONS

WECS : Wind Energy Conversion System


IEA : International Energy Agency
SFOVC : Stator Flux Oriented Vector Control
GVOVC : Grid Voltage Oriented Vector Control
MPPT : Maximum Power Point Tracking
RSC : Rotor Side Converter
EEWS : Effective Estimations of Wind Speed
GSC : Grid Side Converter
PI : Proportional-Integral
SMC : Sliding Mode Control
DTC : Direct Torque Control
DC : Direct Current
DPC : Direct Power Control
LVRT : Low Voltage Ride Through
DVR : Dynamic Voltage Restorer
SMES : Superconducting Magnetic Energy Storage
FRT : Fault Ride Through
KVL : Kirchhoff’s Voltage Law
IGBT : Insulated Gate Bipolar Transistor
FOC : Field Oriented Control
KCL : Kirchhoff’s Current Law
PLL : Phase Locked Loop
PWM : Pulse Width Modulation
DFIG : Doubly Fed Induction Generator

xix
PART 1

INTRODUCTION

Normally in order to use the wind energy, a revolving machine called wind turbine is
required to exist [1]. For instance, the process where the kinetic energy from the
revolving wind turbine is converted in to mechanical energy and this energy is then
converted in to electrical energy by generator is known as wind energy conversion
system (WECS) [2, 3]. It is commonly used in standalone systems, distributed
generation and microgrids, smart grids and grid-connected systems. However, in all
these scenarios, proper energy management procedures must be implemented to
optimize the power generation of the wind turbines and distribute the generated power
to the system with highly efficient and acceptable power quality. In order to do that,
the structure of the power converter and effective control strategy must be chosen.

For the last few years, DFIG has been the most preferable options in wind turbines.
Due to its numerous advantages compared to other types of generators that are now
obtainable in the market. However, DFIG based wind turbines are sensitive to the grid
disturbances which is a major problem in the present power situations. This research
aims to model and study the operation of a wind turbine based DFIG system.
Moreover, it investigates the dynamic behavior of current loops when the system is
connected to the grid. It further analyses the behavior of a grid-connected DFIG based
wind turbine system during symmetrical and asymmetrical voltage dips. This chapter
will provide an introduction to the WECS based on DFIG by first discussing the
motivation behind the thesis with some related background, followed by purpose and
contribution, methodology of the thesis, literature review and finally, the thesis
structure.

1
1.1. MOTIVATION AND BACKGROUND

Wind energy has been utilized for at least three thousand years. It was employed to
produce mechanical energy to pump the water or grind grain until early 20 th century
[4-8]. In the beginning of the modern industrialization, variable wind energy source
was replaced by electrical grids or fossil fuel engines, which supplied a more sustained
source of power [6]. After the first oil price changes in the early 1970s, interest in wind
power resurfaced. However, the main focus this time was on wind power, as a source
of electrical energy rather than mechanical energy. As a result, alternative energy
systems such as the electrical grid was used as a backup to offer a stable and consistent
power supply [6]. Accordingly, wind energy is one of the most reliable and growing
renewable energy sources in the recent years. The most important contributing reasons
for the growth of the wind energy system includes: greater public awareness toward
the renewable energy, government assistance, and fast improvement in the power
electronics sector. According to international energy agency (IEA), in 2020 alone, a
total of 1592 TWh of electricity were generated from the wind turbine [9]. Which, for
instance, is 12% increase over 2019. Thus, in order to achieve the needed 8000 TWh
which is forecasted in 2030, according to IEA report [9], the generation should rise by
an average of 18% each year between 2021 and 2030. As previously mentioned, DFIG
are commonly used in wind turbines because they employ variable speed turbines.
Moreover, DFIG are highly efficient and reliable system and they have the ability to
control active and reactive power independently by utilizing partial converters.
Normally, DFIG based wind turbines are located far away from the power grid which
may result difficulties of controlling the voltages. Correspondingly, they have
drawbacks such as being sensitive to the grid disturbance and limited reactive power
capability when the system is connected to a weak grid. Therefore, this type of
technology requires more research and development in order to be implemented into
the grid. Thus, in order to implement the DFIG system into the present energy system,
it is essential to analyze the steady-state performance, the dynamic behavior of the
system, and the effect on the interconnected power system in terms of voltage control,
and active and reactive power capability.

2
1.2. PURPOSE AND CONTRIBUTION

The main purpose behind this thesis is to implement and model the operation of DFIG
based wind turbine system. The major focus of the study presented here is to design
and control the wind turbine based DFIG according to the company’s manual. It first
generates a code created in MATLAB environment to analyze the different steady-
state dynamic performance of the system, in order to compare and verify with the
designed system. Field oriented or vector control strategy has been employed in order
to implement decouple control of active and reactive powers. Moreover, maximum
power point tracking (MPPT) strategy by indirect speed control has been carried out
to extract the maximum power available from the wind. The main contribution of the
thesis is presented as follows:

▪ Simulation of the DFIG system for the RSC control using stator flux-oriented
vector control (SFOVC) strategy.
▪ Steady-state analysis of the various magnitudes of the DFIG system.
▪ Comparison between the typical PI controller and the proposed method by
investigating the settling time for different wind speed.
▪ Implementation of DFIG based wind turbine system for the RSC.
▪ Implementation of the system in GSC using grid voltage-oriented vector
control (GVOVC) strategy.
▪ Dynamic analysis of current loops behavior in a wind turbine based DFIG.

Besides, this thesis investigates the performance analysis of a grid-connected DFIG


based wind turbine under symmetrical and asymmetrical voltage dips.

1.3. METHODOLOGY OF THE THESIS

The methodology of this thesis has been performed according to the following steps:

▪ Complete literature review on relevant publications from journals, conferences


and the different books has been investigated.

3
▪ Complete modeling of the wind turbine such as: aerodynamic, electrical, and
mechanical part has been carried out.
▪ Dynamic modeling of the DFIG in the 𝑑𝑞 reference frame is derived to study
the operation of the system.
▪ Developing the steady-state analysis in MATLAB for the various magnitudes
of the DFIG.
▪ Design and control of the DFIG for the RSC.
▪ Implement and control the most representative magnitudes of a grid-connected
DFIG based wind turbine system.

The overall system has been implemented and confirmed in MATLAB/Simulink


environment. This software is applied for modeling and controlling the DFIG based
wind turbine system because it includes library tools and protection models.

1.4. LITERATURE REVIEW

In DFIG design, two back-to back linked converters supply the desired magnetization
current at the rotor side [10]. For instance, the grid power of DFIG based wind turbine
is linked to a three-phase converter (GSC), which passes to the DC bus voltage.
Moreover, the DC bus voltage can be applied as a source for the other side of the
converter, which is directly connected to the rotor windings (RSC).

Ref [11] describes an overview of grid-connected WECS in details. Hence, it is noted


that detecting the grid voltage angle correctly is an essential task in most of the control
strategies in the GSC. It has been suggested that for grid synchronization, all advanced
wind turbines must include a faster and efficient tracking equipment to determine the
grid voltage and frequency.

The authors in Ref [12] provides a review on the issue of effective estimations of wind
speed (EEWS). Most of the main EEWS strategies and their implementations are
extensively addressed. The paper realizes that, accurate estimations of the wind speed
to be effective in terms of energy capture.

4
Improving the effectiveness of vector control strategy has attracted many researchers.
In Ref [13-16], the authors presented that the DFIG based wind turbine provides
decouple control of active and reactive powers of generator, resulting good dynamic
performance, higher efficient energy output, and good power quality.

It is very important to bear in mind that the DFIG is more difficult to control compared
to the typical induction generator. Hence, the rotor current is controlled by a power
electronics converter in the rotor circuit, this results the control of the DFIG [17].
Vector control is a popular strategy for controlling the DFIG based wind turbine [18,
19]. Different vector control strategy has been presented in the literature. Stator flux-
oriented vector control is increasingly employed in the present technology.

Ref [20] introduced a new integrated frequency control technique that dynamically
integrates inertial and pitch angle control to improve the performance of DFIG-based
wind turbines to the power system frequency control. In Ref [21], a novel controller
has been introduced that allows the DFIG based wind turbine to positively contribute
to the grid operation.

In Ref [22], comparative analysis between robust sliding mode control (SMC) and
typical PI controller in DFIG based WECS has been presented. Their simulated results
demonstrates that the proposed SMC provides faster dynamic response with very little
steady-state error compared with the PI.

Moreover, Ref [23] compared PI and fuzzy control based DFIG wind turbine, and it is
obvious that the fuzzy control is more robust against machine parametric perturbations
and delivers faster convergence. From Ref [24] we can understand that using fuzzy-PI
controller, settling time and the value of peak overshoot are lowered, and variations
are damped down faster compared to the typical PI controller. The transient response
given by fuzzy-PI controller has also proved to be superior to the typical PI controller.

Ref [25] proposes a dual-loop control strategy to enhance the dynamic performance of
DFIG based wind turbine that may be subjected to the grid disturbance. This strategy

5
can quickly deteriorate the fluctuations of the stator flux and this can successfully
mitigate the impact of stator transient flux on the performances of the DFIG.

Ref [26] studies comparative investigations of control strategies in DFIG based wind
turbine. It states that the direct torque control (DTC) is quicker than vector control in
terms of transient response, that can be very beneficial and provide a lot of capacity
when specific control manipulations are required.

Ref [27] introduced a coordinated control strategy for DFIG with DC connection. The
experimental results shows that the control strategy is available for the proposed power
grid configuration. It concludes that the DC/DC converter to be the most important
component in WECS into the integration of DFIG system. From Ref [28] it can be
realized that the study places a PI controller before the conventional DTC or direct
power control (DPC) blocks to improve the dynamic performances. Consequently, the
study shows that DTC has higher operating performance than DPC, since DTC directly
controls the torque.

However, to our knowledge, none of the aforementioned literature investigates the


implementation and analysis of the magnitudes of grid-connected DFIG system by
using indirect speed control strategy. Moreover, this study describes the comparison
of PI controller and the proposed method in terms of settling time.

The authors in Ref [29] presents a review on control approach for DFIG based wind
turbine under asymmetrical faults in the GSC. It addresses the difficulties encountered
during the dip. Ref [30] presented a crowbar less low voltage ride through (LVRT)
strategy to enhance the fault ride through (FRT) capabilities of DFIG based wind
turbine under transient state operation. Ref [31] introduced an effective FRT system
based DFIG wind turbine. consequently, the design combines the transient stability
characteristics of SMC with the active and reactive power capabilities of dynamic
voltage restorer (DVR), and the instantaneous power availability of superconducting
magnetic energy storage (SMES).

6
Ref [32] describes the applications of a DVR in DFIG based wind turbine to enable
uninterruptible FRT of voltage dips that meet the grid code specifications. The DVR
can recompense for the faulty line voltage, allowing the system to keep functioning
normally as required by the grid codes.

Ref [33] developed a new organized performance modification strategy to enhance the
operations of DFIG during symmetrical and asymmetrical voltage dips. The new
strategy efficiently protects the converter by lowering the overcurrent in the RSC and
maintaining the dc bus voltage within an appropriate limit.

The authors in Ref [34] investigates the improvement of crowbar hardware design for
FRT capabilities in DFIG based wind turbines. The simulated results shows that the
oscillation that occurred after the transient event, were immediately damped with the
linkage of the DFIG to the crowbar hardware element, which was improved by the
stator and rotor dynamic model.

1.5. THESIS STRUCTURE

The structure of this thesis is classified into seven chapters. As previously stated,
chapter 1 provides an introduction to the WECS based on DFIG by first discussing the
motivation behind the thesis with some related backgrounds, followed by purpose and
contribution, methodology of the thesis, literature review and finally, the structure of
the thesis.

Chapter 2 studies the different categories of wind turbine system. It further presents
the detailed model of the wind turbine system by first modeling the wind speed,
followed by the aerodynamic model, drive train model, and generator model.

Chapter 3 describes the dynamic modeling of the DFIG in the 𝑑𝑞 reference frame.
Moreover, it presents DFIG model in the 𝐴𝐵𝐶 representation and the transformation
techniques.

7
Chapter 4 presents the control design of DFIG based wind turbine for the RSC and
when the system is connected to the grid. PI control and indirect speed control strategy
has been studied. Moreover, it investigates the steady-state analysis of DFIG and the
derivations of the transfer function for the RSC and GSC.

Chapter 5 investigates the analysis of grid-connected DFIG based wind turbine during
symmetrical and asymmetrical voltage dips.

Chapter 6 demonstrates the design and implementation of the DFIG based wind turbine
using PI controller and indirect speed control strategy. Detailed simulation results are
presented and discussed.

Chapter 7 gives the conclusion and recommendations for future work on the operation
and control of DFIG based wind turbine system.

8
PART 2

WIND ENERGY SYSTEM

Wind energy is among the most well-known renewable energy sources, and its use is
increasing steadily around the world. In 2019, alone, more than 60 GW of wind energy
capacity was installed globally, which is a 19% rise over installation in 2018, showing
that wind energy is gradually dominating the energy industry. DFIG has been an
extensively known machine compared to other forms of the wind energy system. It
contains a wind turbine attached to the shaft of an induction generator. A DFIG is built
such that: the stator part connects to the power grid directly, whereas the rotor part
links to the grid via a pulse width modulation (PWM) converter that controls the slip
energy. This section includes the different types of wind turbines both according to the
position of the axis of rotation and by the operation of their system. Moreover, it
presents the mathematical modeling of the wind turbine system.

2.1. WIND ENERGY CONVERSION SYSTEM

WECS is the entire system that uses mechanical power to convert wind energy into
useable electrical power. It is mainly divided into three parts: aerodynamic,
mechanical, and electrical [35]. Figure 2.1 depicts the primary parameters of a typical
wind power conversion system.

Figure 2.1. Shows components of grid-connected WECS [35].

9
Wind turbines can be categorized into two types of rotational axis; horizontal axis wind
turbines (HAWTs) and vertical axis wind turbines (VAWTs) [36]. HAWTs currently
account for most installed solutions around the world. As a result, HAWTs account for
about 75% of wind turbines, which is owing to their great efficiency when compared
to installation and maintenance expenses [37]. The name HAWT is given based upon
the fact that its blades spin on an axis parallel to the earth's surface. HAWTs can be
classified into two parts: downwind HAWTs and upwind HAWTs, as shown in Figure
2.2.

It can be seen that the rotational blade of downwind is on the backside of the turbine.
Unlike downwind, the rotor blade of the upwind turbine is directly toward the wind.
Furthermore, they don’t self-align toward the wind, necessitating the use of a tail vane
or a yaw system to keep the rotor facing the wind. For this purpose, most of the wind
turbines that are used nowadays are upwind HAWTs. Thereby, upwind HAWTs are
mostly presented in this work.

Figure 2.2. Upwind and downwind type wind turbines [38].

2.2. WIND TURBINE SYSTEM

Wind turbine systems may be divided into two categories based on how their system
works: Fixed and variable speed wind turbines.

10
2.2.1. Fixed Speed Wind Turbines

The rotor speed of a fixed-speed wind turbine (FSWT) can be specified and governed
by the frequencies of the grid, the gear ratio, and generator design, regardless of wind
speed [39]. A typical configuration of FSWT can be seen in Figure [2.3]. FSWTs
include a squirrel cage induction generator (SCIG) that links to a Soft-Starter that
connects to a capacitor which is linked to the grid through a transformer to adjust the
reactive power [39]. The main disadvantage is that these wind turbine generators are
unable to track whenever the value of the speed of the wind is changed. Furthermore,
the efficiency of power harvesting may not be good like in variable speed turbines. For
this reason, in this thesis, we consider variable speed turbines.

Transformer
+
Grid

Gear Soft-
box
SCIG Starter

Squirrel-Cage
Induction Capacitor
Generator bank

Figure 2.3. Configuration of fixed speed wind turbine.

2.2.2. Variable Speed Wind Turbines

Variable speed generators, differs from the fixed speed turbines in that, they can adjust
the rotational speed, as the speed of wind changes to maximize energy capturing from
the wind. The rotational speed of the turbine can be continuously modified in response
to wind speed. Furthermore, lambda (𝜆) can be kept at an optimal range to ensure
maximum power conversion efficiency at different wind speed [40]. The configuration
of variable speed turbines can be seen in Figure 2.4. A PWM converter links the
generator to the power grid then the converter allows the wind turbines to operate at
various speeds. Variable speed generator has several benefits, including enhanced

11
wind power output, greater power quality, and less mechanical impacts. Contrarily,
higher manufacturing costs and power losses are associated with the usage of power
converters [40].

Electronic Grid
converter

Gear
box

Synchronous
generator

Figure 2.4. Scheme of variable speed synchronous generator.

2.3. WIND TURBINE MODEL

It is very important to note that the data provided by the company in their manual is
relatively limited. The purpose of this section is to present some basic data and
concepts for a wind turbine model. A practical wind turbine has a nominal output
power of 1.5 to 3 MW. For this purpose, in this thesis, a 2.4 MW of wind turbine has
been chosen for the system simulation. A schematic diagram of the Mitsubishi MWT
92 wind turbine is presented in Figure 2.5. From the information gathered from the
company’s manuals and other technical references, several factors that characterize
wind turbine’s dynamic performance and its basic electrical and mechanical behavior
can be presented. To model, we need to consider the parameters of some features
including, the ratio of the blade, the nominal power for a certain wind speed, the
gearbox ratio, rotational speed, and the maximum speed of the blade’s tip which are
all presented in Appendix A.

Wind turbines produce electrical power by converting the mechanical energy provided
by kinetic energy in the wind [41]. They can be described as a dimensionless power
coefficient curve function of lambda and inclination angle. In this section, we assume
that the wind turbine model comprises several parts: a wind speed, an aerodynamic
model, a drive train model, and a generator. The corresponding wind speed supplied
12
to the aerodynamic system is estimated using a wind speed model [42]. The amount
of rotational speed divided by the real wind speed (𝑣𝑤 ) is called lambda. Wind speed,
blade’s pitch angle and speed of the rotor are all together fed into the aerodynamic
system as shown in Figure 2.6. The output torque is fed to the drive train model along
with the speed of the generator. Drive train model, thereby generates electromagnetic
torque on high-speed shaft (ℎ𝑠𝑠), which is then transmitted to the generator system.

Figure 2.5. Schematic diagram of Mitsubishi MWT 92 wind turbine [42].


Pref
Blade Angle
Control Pmeas

θblade

veq Thss Generator Transformer


Wind Aerodynamic Ta Drive-train
Speed System System System
Grid

ωrot ωgen
θrot

Figure 2.6. Shows a typical block design of the wind turbine model.
13
2.3.1. Wind Speed Model

Wind speed varies depending on the environment and it can also fluctuate randomly
over time. This explains the fact that it has a significant impact on the electromagnetic
torque and hence it seems to have a major impact on the power generated by the three
blades [43]. For this purpose, in order to simulate the dynamic behavior of the wind
turbine model, we need to take into consideration the wind speed. In this work, wind
speed can be modeled by adding the following four constituents:

𝑣𝑤 (𝑡) = 𝑣𝑤𝑎 (𝑡) + 𝑣𝑤𝑡 (𝑡) + 𝑣𝑤𝑟 (𝑡) + 𝑣𝑤𝑔 (𝑡) (2.1)

Where 𝑣𝑤𝑎 (𝑡), 𝑣𝑤𝑟 (𝑡), 𝑣𝑤𝑔 (𝑡) and 𝑣𝑤𝑡 (𝑡) are the constant, turbulence, ramp, and gust
constituents, respectively. The gust constituent can be modeled to represent unusual
transient increases in wind speed and can be expressed as follows:

0, 𝑓𝑜𝑟 𝑡 < 𝑇𝑠𝑔


𝑡−𝑇𝑠𝑔
𝑣𝑤𝑔 (𝑡) = 𝐴𝑔 (1 − 𝑐𝑜𝑠 [2𝜋 (𝑇 )]) , 𝑓𝑜𝑟 𝑇𝑠𝑔 ≤ 𝑡 ≤ 𝑇𝑒𝑔 (2.2)
𝑒𝑔 −𝑇𝑠𝑔

{0, 𝑓𝑜𝑟 𝑇𝑒𝑔 < 𝑡

Where 𝑇𝑠𝑔 and 𝑇𝑒𝑔 are start and end time of gust constituents and 𝐴𝑔 is gust amplitude.
Finally, the turbulence constituent can be represented by a signal with the following
power density [43].

ℎ −2
𝑙𝑣𝑤 [ln( )]
𝑧𝑜
𝑃𝐷𝑡 (𝑓) = 𝑓𝑙 5/3
(2.3)
[1+1.5 ]
𝑣𝑤

Where 𝑙 is turbulence scale, ℎ is the height and 𝑙 = 20ℎ, that has maximum height of
300 meter; 𝑧0 is the raggedness length value that varies due to topography type as it
can be seen in Table 2.1 [43].

14
Table 2.1. Indicates different value of z_o for various topology type [43].
Types of topologies Scale of 𝒛𝒐 in meter
High seas 0.001-0.01
Snow surface 0.01-0.05
Mown grass 0.01-0.03
Rocky ground 0.02-0.1
Mountainous 2-6

Since we know all the other values, the next stage is to generate a signal with power
spectrum. Considering that the 𝑃𝐷𝑡 is close to the responses of first-order filter, then
the suggested transfer function is as follows:

𝐾
𝐻(𝑠) = 𝑠+𝑝 (2.4)

Where 𝑝 and 𝐾 are formulated as follows:

3/5
2𝜋((𝐾12 ) −1)
𝑝= (2.5)
𝐾2 √𝐾12 −1

𝐾 = 𝐾1 𝑝 (2.6)

And 𝐾1 and 𝐾2 are expressed as:

ℎ −2
𝐾1 = 𝑙𝑣𝑤 [ln (𝑧 )] (2.7)
𝑜

𝑙
𝐾2 = 1.5 𝑣 (2.8)
𝑤

The calculated signal of the power spectrum is defined as:

𝑘2
𝑝2
𝑃𝑓𝑖𝑙𝑡𝑒𝑟 = 4𝜋2
(2.9)
1+ 2 2
𝑝 𝑓

15
2.3.2. Aerodynamic Model

Aerodynamic model calculates electromagnetic torque as a function of rotational speed


on the blades. A simple aerodynamic system is presented in Figure 2.7. The torque
coefficient can be obtained by using look up table from Simulink/MATLAB. In order
to calculate the average electromagnetic torque in the low-speed shaft, wind speed can
be defined as the average incidence of wind speed on the brushed area by the blade.
The torque generated by the three-blade wind turbine can be formulated as [42].

𝑇𝑡 = 0.5𝜌𝜋𝑅 3 𝑉𝑤2 𝐶𝑡 (2.10)

Where, 𝜌, 𝑅, 𝑉𝑤 , and 𝐶𝑡 are air density, ratio (meter), wind speed and torque coefficient
of wind turbine, respectively.

The most basic approach to characterize the power and torque coefficient is by a
mathematical calculation as a function of lambda and pitch angle, as we presented in
section (2.3). The power coefficient can be represented by the following expression:

151
𝐶𝑝 = 0.773 ( 𝜆 − 0.58𝛽 − 0.002𝛽 2.14 − 13.2) (𝑒 −18.4/𝜆𝑖 ) (2.11)
𝑖

With
1 0.003
𝜆𝑖 = 𝜆+0.02𝛽 − 𝛽3+1 (2.12)

The tip step ratio 𝜆 is obtained as:

𝑅𝛺𝑡
𝜆= (2.13)
𝑉𝑤

16
. −

Figure 2.7. Aerodynamic system.

2.3.3. Drive Train Model

The drive train system of wind turbine can be used to characterize either by 1-mass, 2-
mass, 3-mass or even 6-mass model [44]. The drive train system consists of a turbine,
generator, low-speed shaft, gearbox, and a high-speed shaft [45, 46]. In order to model
drive train system, we need to include both dynamic and typical parameters. At first,
we need to consider power train’s resonance frequency. The power transmission train
is formed by the blades connected to the hub, attached to the slow shaft, which is
coupled to the gearbox, which multiplies the rotational speed of the fast shaft attached
to the generator [42]. In this thesis, we consider the 2-mass model as illustrated in
Figure 2.8. The second resonant frequency is very higher but it’s lower in magnitude.
All the magnitudes are supposed to be in the fast shaft. Inertia (𝐽𝑡 ) affects the turbine
side masses, whereas (𝐽𝑚 ) affects the electrical generator. The stiffness coefficient (𝑘)
and damping coefficient (𝑐) represent the dynamic linkage between the two inertias.
The dynamics of the system is characterized using Newton's laws as follows:

−𝑣 2 𝑐 𝑣𝑐 𝑣2𝑘 𝑣𝑘 1
𝜔̇ 𝑚 − 𝜔𝑚 0
𝐽𝑚 𝐽𝑚 𝐽𝑚 𝐽𝑚 𝐽𝑚
𝜔̇ 𝑣𝑐 𝑐 𝑣𝑘 𝑘 𝜔𝑡 1 𝜏𝑚
{ 𝑡} = − − {𝜃 } + 0
𝐽𝑡 {𝜏 } (2.14)
𝜔𝑚 𝐽𝑡 𝐽𝑡 𝐽𝑡 𝐽𝑡 𝑚 𝑡
𝜔𝑡 1 0 0 0 𝜃𝑡 0 0
( 0 1 0 0 ) (0 0)

Where θt , θm , 𝜔t and 𝜔𝑚 are turbine angle, generator angle, turbine angular speed
and generator angular speed, respectively. 𝜏t is torque supplied to rotor shaft and 𝜏m
is the generator torque.

17
θm
Tm
ωm θt
Jm

Jt
ωt
k c
v Tt

Figure 2.8. Drive train model.

2.3.4. Generator Model

A wound rotor asynchronous machine is the generator of doubly fed wind turbine
system. The impact of magnetic saturations and the capacitance of all windings are
ignored, and the neutrality of the stator and rotor windings is assumed to be moving
and the currents flowing toward the machine are supposed to be positive and assuming
that the stator and rotor windings to be placed symmetric and sinusoidal. It is possible
to implement the relationship between voltage and current on the generator, as well as
their first derivatives using synchronous 𝑑𝑞 representation and can be formulated as
follows [43].

𝑣𝑠𝑞 𝐿𝑠 0 𝑀 0 𝑖𝑠𝑞 𝑟𝑠 𝐿𝑠 𝜔𝑠 0 𝑀𝜔𝑠 𝑖𝑠𝑞


𝑣𝑠𝑑 0 𝐿𝑠 0 𝑀 𝑑 𝑖𝑠𝑑 −𝐿𝑠 𝜔𝑠 𝑟𝑠 −𝑀𝜔𝑠 0 𝑖𝑠𝑑
{𝑣 } = [ ] { }+[ ] { }(2.15)
𝑟𝑞 𝑀 0 𝐿𝑟 0 𝑑𝑡 𝑖𝑟𝑞 0 𝑠𝑀𝜔𝑠 𝑟𝑟 𝑠𝐿𝑟 𝜔𝑠 𝑖𝑟𝑞
𝑣𝑟𝑑 0 𝑀 0 𝐿𝑟 𝑖𝑟𝑑 −𝑠𝑀𝜔𝑠 0 −𝑠𝐿𝑟 𝜔𝑠 𝑟𝑟 𝑖𝑟𝑑

Where 𝐿𝑠 and 𝐿𝑟 are the coefficients of stator and rotor inductance, 𝑀 is magnetizing
inductance and s is the slip.

Moreover, the torque and stator reactive power (𝑄𝑠 ), which is used to control the (rotor
side) can be represented as follows:

3
𝑇𝑚 = 2 𝑃𝑀(𝑖𝑠𝑞 𝑖𝑟𝑑 − 𝑖𝑠𝑑 𝑖𝑟𝑞 ) (2.16)

18
And

3
𝑄𝑠 = 2 (𝑣𝑠𝑞 𝑖𝑠𝑑 − 𝑣𝑠𝑑 𝑖𝑠𝑞 ) (2.17)

Where, 𝑃 is the number of pole pair.

19
PART 3

DYNAMIC MODELING OF DOUBLY FED INDUCTION GENERATOR

Recently, the doubly fed induction machine is commonly utilized in wind generation
system. These thesis presents the dynamic model and the performance analysis of
steady-state in DFIG. The dynamic model of DFIG allows one to know the continual
performance (not only under steady-state) of machine parameters, but also during
various voltage providing situations. This section first develops the dynamic model of
DFIG in 𝑎𝑏𝑐 stationary natural frame. The dynamics of the system is then converted
to a rotating 𝑑𝑞 reference frame using Clarke and Park transformation method for
cancellation of coupling between the two stator and rotor windings. The dynamic
model of DFIG in a rotating 𝑑𝑞 axis is very important for designing the properties of
synchronous and hyper-synchronous operating system as well as implementing vector
control strategy for controlling active and reactive output power. In this section, a brief
model of DFIG, a dc-link model, power electronic converter model and power and
torque calculation has been presented.

3.1. DFIG MODEL IN ABC REPRESENTATION

The generalized form of the DFIG model can be represented as three phase stator and
rotor windings, as shown in Figure 3.2. These three phase windings are the idealized
form of the real system. Figure 3.1 shows the relation between stationary 𝑎𝑏𝑐 natural
frame and rotating space vector. Whereby the space vector spins with a random speed
ω with respect to 𝑎𝑏𝑐 natural frame. The space vector and 𝑎𝑏𝑐 references are separated
1200 each other [42]. By using Kirchhoff’s voltage law (KVL) it’s possible to
characterize the instantaneous voltage, current and flux in the stator side of the
machine as follows:

20
𝑑𝜓𝑎𝑠 (𝑡)
𝑣𝑎𝑠 (𝑡) = 𝑅𝑠 𝑖𝑎𝑠 + 𝑑𝑡
𝑑𝜓𝑏𝑠 (𝑡)
𝑣𝑏𝑠 (𝑡) = 𝑅𝑠 𝑖𝑏𝑠 + (3.1)
𝑑𝑡
𝑑𝜓𝑐𝑠 (𝑡)
{ 𝑣𝑐𝑠 (𝑡) = 𝑅𝑠 𝑖𝑐𝑠 + 𝑑𝑡

Figure 3.1. Shows stationary abc reference frame [42].

These idealistic 𝑎𝑏𝑐 representations are the actual machine, which facilitates the
development of an equivalent electrical circuits as illustrated in Figure 3.2.

Figure 3.2. DFIG equivalent circuits [42].

21
Where 𝑖𝑎𝑠 (𝑡), 𝑖𝑏𝑠 (𝑡), and 𝑖𝑐𝑠 (𝑡) are the currents of the stator in the 𝑎, 𝑏, 𝑐 phases
respectively; 𝑣𝑎𝑠 (𝑡), 𝑣𝑏𝑠 (𝑡), and 𝑣𝑐𝑠 (𝑡) are the voltages supplied in the stator; 𝜓𝑎𝑠 (𝑡),
𝜓𝑏𝑠 (𝑡), and 𝜓𝑎𝑠 (𝑡) are stator fluxes and 𝑅𝑠 is the resistance of the stator.

For instance, the magnitudes in the rotor side can also be expressed as follows:

𝑑𝜓𝑎𝑟 (𝑡)
𝑣𝑎𝑟 (𝑡) = 𝑅𝑟 𝑖𝑎𝑟 + 𝑑𝑡
𝑑𝜓𝑏𝑟 (𝑡)
𝑣𝑏𝑟 (𝑡) = 𝑅𝑟 𝑖𝑏𝑟 + (3.2)
𝑑𝑡
𝑑𝜓𝑐𝑟 (𝑡)
{ 𝑣𝑐𝑟 (𝑡) = 𝑅𝑟 𝑖𝑐𝑟 + 𝑑𝑡

Where 𝑖𝑎𝑟 (𝑡), 𝑖𝑏𝑟 (𝑡), and 𝑖𝑐𝑟 (𝑡) are the currents of the rotor in the 𝑎, 𝑏, 𝑐 phases
respectively; 𝑣𝑎𝑟 (𝑡), 𝑣𝑏𝑟 (𝑡), and 𝑣𝑐𝑟 (𝑡) are the voltages supplied in the stator refer to
rotor; 𝜓𝑎𝑟 (𝑡), 𝜓𝑏𝑟 (𝑡), and 𝜓𝑎𝑟 (𝑡) are fluxes of the rotor and 𝑅𝑟 is the resistance of the
rotor. Moreover, the magnitudes of the rotor have constant rotational frequencies 𝜔𝑟
during steady-state condition. The relationship between stator and rotor angular
frequency is described as follows:

𝜔𝑠 = 𝜔𝑟 + 𝜔𝑚 (3.3)

with

𝜔𝑚 = 𝑃𝛺𝑚 (3.4)

Where 𝜔𝑚 is the electric rotational frequency, 𝛺𝑚 is the mechanical rotor speed and
𝑃 is the number of poles of the machine.

3.2. THE TRANSFORMATION TECHNIQUE

In order to use the vector control approach, we need to model the dynamic equation of
doubly fed induction machine (DFIM), which then provides the instantaneous voltages
and currents in order to compute and control the parameters. Even though the rotor
currents cannot be determined with cage motors, thus the current is substituted by an

22
appropriate amount expressed in a rotating 𝑑𝑞 representation by using a well-known
technique called the Clarke and Park transformation.

3.2.1. The Clarke Transformation

The Clarke transform is a mathematical technique that converts a 3-phase 𝑎𝑏𝑐 natural
frame into an orthogonally structure (𝛼𝛽) [47]. Let’s consider a symmetric 3-phase
induction generator having stationary 𝑎𝑏𝑐 axis with 1200 apart as illustrated in Figure
3.3. Then it’s possible to convert the 𝑎𝑏𝑐 constituents into 𝛼𝛽 through the Clarke
method and it can be expressed as the following matrix equation:

𝑥𝛼 𝑥𝑎
[𝑥𝛽 ] = 𝑇𝐶 [𝑥𝑏 ] (3.5)
𝑥0 𝑥𝑐

Where, 𝑇𝐶 is the Clarke transform matrix that can be represented as,

1 1
1 − −
2 2
2 √3 √3
𝑇𝐶 = 3 0 − 2 (3.6)
2
1 1 1
[2 2 2 ]

β
b
xβs x

ɑ, α
xαs

Figure 3.3. The Clarke transform representation.

23
For a symmetric 3-phase, 𝑥0 is equal to zero and the transform representation for stator
and rotor current shown in Figure 3.4 is defined as:

𝑖𝛼𝑠 𝑖𝑎𝑠
[𝑖𝛽𝑠 ] = 𝑇𝐶 [𝑖𝑏𝑠 ] (3.7)
𝑖0𝑠 𝑖𝑐𝑠

And the rotor side can be defined as:

𝑖𝛼𝑟 𝑖𝑎𝑟
[𝑖𝛽𝑟 ] = 𝑇𝐶 [𝑖𝑏𝑟 ] (3.8)
𝑖0𝑟 𝑖𝑐𝑟

Where, 𝑖0𝑠 and 𝑖0𝑟 are the stator and rotor currents at zero constituents respectively.

For instance, it’s also possible to transform the rotor representation refer to the stator
and it can be expressed by the following matrix equation:

𝑠
𝑖𝛼𝑟 𝑖𝛼𝑟
[ 𝑠 ] = 𝑇𝜃𝑟 [𝑖 ] (3.9)
𝑖𝛽𝑟 𝛽𝑟

with

cos 𝜃𝑟 −sin 𝜃𝑟
𝑇𝜃𝑟 = [ ] (3.10)
sin 𝜃𝑟 cos 𝜃𝑟

βr βs

αr
θr αs
Figure 3.4. Transform representation between stator and rotor parameters.

24
3.2.2. The Park Transformation

The Park transform technique was first introduced by an American engineer known as
R.H. Park in the late 1920s to simplify a new strategy for the analysis of a 3-phase
electrical machine. This method was introduced to transform a 2-phase 𝛼𝛽 reference
frame into a rotating 𝑑𝑞 reference frame (see Figure 3.5) through a rotating transform
matrix 𝑇𝑃 (𝜃) that can be expressed as:

𝑖𝑑𝑠 𝑖𝛼𝑠
[𝑖 ] = 𝑇𝑃 (𝜃) [𝑖 ] (3.11)
𝑞𝑠 𝛽𝑠

Where, 𝑇𝑃 (𝜃) is the transform matrix and is given as:

cos 𝜃 −sin 𝜃
𝑇𝑃 (𝜃) = [ ] (3.12)
sin 𝜃 cos 𝜃

Where, 𝜃 is the angle between d and q axis.

Figure 3.6 represents the general transformation design which is transforming from 3-
phase 𝑎𝑏𝑐 to 2-phase 𝛼𝛽 stationary reference frame and to rotating 𝑑𝑞 representation.

b
q ω1 rad/sec d

θ a,α

c
Figure 3.5. The representation of Park transforms.

25
Natural Stationary Rotating
Reference Frame Reference Frame Reference Frame

3ϕ α αβ d
β q
αβ dq

θ
Figure 3.6. General diagram of the transformation techniques.

3.3. THE DYNAMIC MODEL OF DFIG IN DQ REPRESENTATION

For a proper understanding of the behaviors of a DFIG, it is important to use a basic


and special model. In order to do that it’s necessary to model a rotating 2-phase 𝑑𝑞
that is provided by Park transform technique [48]. The equivalent electrical circuit can
be seen in Figure 3.7. The dynamic equation of DFIG is described as a fourth order
state space based synchronous 𝑑𝑞 representation, the stator and rotor voltage equations
can be expressed as follows [49]:

𝑑𝜓𝑑𝑠
𝑣𝑑𝑠 = 𝑅𝑠 𝑖𝑑𝑠 + − 𝜔𝑠 𝜓𝑞𝑠
𝑑𝑡
𝑑𝜓𝑞𝑠
𝑣𝑞𝑠 = 𝑅𝑠 𝑖𝑞𝑠 + + 𝜔𝑠 𝜓𝑑𝑠
𝑑𝑡
𝑑𝜓𝑑𝑟
(3.13)
𝑣𝑑𝑟 = 𝑅𝑟 𝑖𝑑𝑟 + − 𝜔𝑟 𝜓𝑞𝑟
𝑑𝑡
𝑑𝜓𝑞𝑟
{𝑣𝑞𝑟 = 𝑅𝑟 𝑖𝑞𝑟 + 𝑑𝑡
+ 𝜔𝑟 𝜓𝑑𝑟

the stator and rotor fluxes can be determined as:


𝜓𝑑𝑠 = 𝐿𝑠 𝑖𝑑𝑠 + 𝐿𝑚 𝑖𝑑𝑟
𝜓𝑞𝑠 = 𝐿𝑠 𝑖𝑞𝑠 + 𝐿𝑚 𝑖𝑞𝑟
(3.14)
𝜓𝑑𝑟 = 𝐿𝑟 𝑖𝑑𝑟 + 𝐿𝑚 𝑖𝑑𝑠
{ 𝜓𝑞𝑟 = 𝐿𝑟 𝑖𝑞𝑟 + 𝐿𝑚 𝑖𝑞𝑠

We can solve Equation (3.14) in terms of current as follows:

26
1 𝐿
𝐼𝑑𝑠 = 𝜎𝐿 𝜓𝑑𝑠 − 𝜎𝐿𝑚𝐿 𝜓𝑑𝑟
𝑠 𝑠 𝑟
1 𝐿𝑚
𝐼𝑞𝑠 = 𝜎𝐿 𝜓𝑞𝑠 − 𝜎𝐿 𝜓𝑞𝑟
𝑠 𝑠 𝐿𝑟
1 𝐿 (3.15)
𝐼𝑑𝑟 = 𝜎𝐿 𝜓𝑑𝑟 − 𝜎𝐿𝑚𝐿 𝜓𝑑𝑠
𝑟 𝑠 𝑟
1 𝐿𝑚
{ 𝐼𝑞𝑟 = 𝜎𝐿𝑟 𝜓𝑞𝑟 − 𝜎𝐿𝑠 𝐿𝑟 𝜓𝑞𝑠

Where 𝑣𝑑𝑠 , 𝑣𝑞𝑠 , 𝜓𝑑𝑠 , 𝜓𝑞𝑠 , 𝑖𝑑𝑠 , 𝑖𝑞𝑠 and 𝜔𝑠 are the stator side voltages, fluxes, currents
and angular speed of the 𝑑𝑞 representation respectively; 𝑣𝑑𝑟 , 𝑣𝑞𝑟 , 𝜓𝑑𝑟 , 𝜓𝑞𝑟 , 𝑖𝑑𝑟 , 𝑖𝑞𝑟
and 𝜔𝑟 are the rotor side voltages, fluxes, currents and angular speed of the 𝑑𝑞
representation respectively; 𝑅𝑠 and 𝐿𝑠 are the stator side resistance and inductance; 𝑅𝑟
and 𝐿𝑟 are the rotor side resistance and inductance; and 𝐿𝑚 is the magnetizing
𝐿𝑠 𝐿𝑟 −𝐿2𝑚
inductance and 𝜎 is the leakage coefficients formulated as; 𝜎 = 𝐿𝑠 𝐿𝑟

_ ωsψqs + Lσs Lσr +


ωrψqr _
ids Rs Rr idr

+ + + +
DC DC
Vds Lm Vdr
_ _ _ _

ωsψds _ _ ωrψdr
+ Lσs Lσr +
iqs Rs Rr iqr

+ + +
_ +
Vqs DC DC
Lm Vqr
_ _ _ _

Figure 3.7. Electrical circuit for DFIG model in dq representation.

Once we get the parameters such as voltages, currents, fluxes and inductance of the
stator and rotor side of the 𝑑𝑞 representation and magnetizing inductance, in order to
finalize the system, it’s important to adjust the active and reactive power in stator and
rotor side of the machine (let’s assume we neglect power loss) and it’s formulated as
follows:

27
3
𝑃𝑠 = 2 (𝑣𝑑𝑠 𝑖𝑑𝑠 + 𝑣𝑞𝑠 𝑖𝑞𝑠 )
3
𝑃𝑟 = 2 (𝑣𝑑𝑟 𝑖𝑑𝑟 + 𝑣𝑞𝑟 𝑖𝑞𝑟 )
3 (3.16)
𝑄𝑠 = 2 (𝑣𝑞𝑠 𝑖𝑑𝑠 − 𝑣𝑑𝑠 𝑖𝑞𝑠 )
3
{𝑄𝑟 = 2 (𝑣𝑞𝑟 𝑖𝑑𝑟 − 𝑣𝑑𝑟 𝑖𝑞𝑟 )

Electromagnetic torque can be formulated as follows:

3 𝐿𝑚
𝑇𝑒𝑚 = 2 𝑃 (𝜓𝑞𝑠 𝑖𝑑𝑟 − 𝜓𝑑𝑠 𝑖𝑞𝑟 ) (3.17)
𝐿𝑠

It’s important to represent the dynamics of DFIG in state space equation to perform
the simulation. The state space is commonly used during transient conditions. The
basic equation used for time dependent state space is expressed as:

𝑋̇(𝑡) = 𝐴𝑋(𝑡) + 𝐵𝑈(𝑡)


{ (3.18)
𝑌(𝑡) = 𝐶𝑋(𝑡) + 𝐷𝑈(𝑡)

Where 𝐴, 𝐵, 𝐶 and 𝐷 are state, input, output and forward matrixes respectively; 𝑋 and
𝑌 are the state and output vector.

For instance, substitute Equation (3.15) into Equation (3.13) it gives (considering the
flux as a state variable) as follows:

𝑑𝜓𝑑𝑠 𝑅 𝑅 𝐿𝑚
= − 𝜎𝐿𝑠 𝜓𝑑𝑠 + 𝜔𝑠 𝜓𝑞𝑠 + 𝜎𝐿𝑠 𝜓𝑑𝑟 + 𝑣𝑑𝑠
𝑑𝑡 𝑠 𝑠 𝐿𝑟
𝑑𝜓𝑞𝑠 𝑅𝑠 𝑅 𝐿
= −𝜔𝑠 𝜓𝑑𝑠 − 𝜓 + 𝑠 𝑚 𝜓𝑞𝑟 + 𝑣𝑞𝑠
𝑑𝑡 𝜎𝐿𝑠 𝑞𝑠 𝜎𝐿𝑠 𝐿𝑟
𝑑𝜓𝑑𝑟 𝑅𝑟 𝐿𝑚 𝑅𝑟
(3.19)
= 𝜓𝑑𝑠 − 𝜓 + 𝜔𝑟 𝜓𝑞𝑟 + 𝑣𝑑𝑟
𝑑𝑡 𝜎𝐿𝑠 𝐿𝑟 𝜎𝐿𝑟 𝑑𝑟
𝑑𝜓𝑞𝑟 𝑅𝑟 𝐿𝑚 𝑅
{ = 𝜓𝑞𝑠 − 𝜔𝑟 𝜓𝑑𝑟 − 𝜎𝐿𝑟 𝜓𝑞𝑟 + 𝑣𝑞𝑟
𝑑𝑡 𝜎𝐿 𝐿
𝑠 𝑟 𝑟

From Equation (3.18) the state space equation can be written as:

28
𝑅 𝑅𝑠 𝐿𝑚
− 𝜎𝐿𝑠 𝜔𝑠 0
𝜓̇𝑑𝑠 𝑠 𝜎𝐿𝑠 𝐿𝑟
𝜓𝑑𝑠 1 0 0 0 𝑣𝑑𝑠
𝑅 𝑅𝑠 𝐿𝑚
𝜓̇𝑞𝑠 −𝜔𝑠 − 𝜎𝐿𝑠 0
𝑠 𝜎𝐿𝑠 𝐿𝑟 𝜓𝑞𝑠 0 1 0 0 𝑣𝑞𝑠
= +[ ][ ] (3.20)
𝜓̇𝑑𝑟 𝑅𝑟 𝐿𝑚
0
𝑅𝑟
− 𝜎𝐿 𝜔𝑟 𝜓𝑑𝑟 0 0 1 0 𝑣𝑑𝑟
𝜎𝐿𝑠 𝐿𝑟 𝑟
[ 𝜓𝑞𝑟 ] 0 0 0 1 𝑣𝑞𝑟
[𝜓̇𝑞𝑟 ] 𝑅𝑟 𝐿𝑚 𝑅
[ 0 −𝜔𝑟 − 𝜎𝐿𝑟 ]
𝜎𝐿𝑠 𝐿𝑟 𝑟

3.4. CONVERTER AND DC LINK MODEL

The power circuit of DFIG in wind turbine (shown in Figure 3.8) consists of a DFIG,
a dc bus (commonly employs dc capacitor), a two back-to-back power electronic
converter, each converter is connected to the rotor and grid circuit, in which both the
rotor and grid side converter are connected to a controller namely RSC control and
GSC control respectively. The converters are mainly voltage source converters (VSC)
that allows power to flow in two directions [50].

Figure 3.8. Power circuit of DFIG in wind turbine [42].

The inverter gate uses an IGBTs equipped with flyback diodes. An LC filters is
supplied on each side of converters output in order to minimize the harmonics and
insulate the components [43]. In this work, RL grid side filter is used. Modeling the
converters is not an easy operation, and various level of details can be obtained by
taking some considerations. In this thesis, it’s considered a high value of switching
frequency (𝑓𝑠𝑤) of 4000 Hz in order to completely filter the voltage signals produced

29
by the converters and switching power losses are considered to be neglected. By taking
into this consideration, the dynamics equation of the machine between the voltage of
the grid side and the rotor side of the converter (assume current to be positive while it
flows to the machine) can be formulated as:

Figure 3.9. DFIG power converter circuit [43].

1
𝑑
𝑣𝑧𝑎𝑏𝑐 − 𝑣𝑙𝑎𝑏𝑐 − (𝑣𝑐𝑛 − 𝑣𝑧𝑛 ) {1} = 𝑟𝑙 𝑖𝑙𝑎𝑏𝑐 + 𝐿𝑙 𝑖𝑙𝑎𝑏𝑐 (3.21)
𝑑𝑡
1

and

1
(𝑣𝑐𝑛 − 𝑣𝑧𝑛 ) = (𝑣𝑧𝑎 + 𝑣𝑧𝑏 + 𝑣𝑧𝑐 − 𝑣𝑙𝑎 − 𝑣𝑧𝑏 − 𝑣𝑧𝑐 ) (3.22)
3

Where 𝑣𝑙𝑎𝑏𝑐 and 𝑣𝑧𝑎𝑏𝑐 are the rotor and grid side 𝑎𝑏𝑐 vector voltages of the converters;
𝑟𝑙 and 𝐿𝑙 are the resistance and inductance filters.

The DC component of the power electronics converter is commonly known as the DC


link [51]. The dc link stores the energy in the capacitor and the purpose of the capacitor
is to keep the voltage terminals constant. The dc link is connected between the rotor
side and grid side converters [52]. The energy storage through the capacitor can be
formulated as:

30
1 2
𝑊𝑐 = 2 𝐶𝑑𝑐 𝑣𝑑𝑐 (3.23)

Considering the converter losses to be small and negligible, the active power in the dc
link can be given as [53]:

RSC GSC
DC-Link
AC +
DC
Cdc vdc
DC - AC

Pr Pg
Figure 3.10. DC link model.

𝑑𝑊𝑐 𝑑𝑣𝑑𝑐
𝑃𝑑𝑐 = = 𝐶𝑑𝑐 = 𝑃𝑟 − 𝑃𝑔 (3.24)
𝑑𝑡 𝑑𝑡

Where 𝐶𝑑𝑐 and 𝑣𝑑𝑐 are the capacitor and the voltage in the dc link, 𝑃𝑟 and 𝑃𝑔 are the
rotor side and grid side active power. The currents in the dc link can be obtained by
using Kirchhoff’s current law (KCL).

𝑖𝑐 = 𝑖𝐷𝐶𝑙 − 𝑖𝐷𝐶𝑟 (3.25)

For instance, the dc bus voltage can be expressed as:

1 𝑡
𝑣 = 𝑣0 + 𝐶 ∫0 (𝑖𝐷𝐶𝑙 − 𝑖𝐷𝐶𝑟 ) 𝑑𝑡 (3.26)

Where 𝑖𝐷𝐶𝑟 and 𝑖𝐷𝐶𝑙 are the DC current in the rotor side and grid side converter
respectively.

31
PART 4

CONTROL DESIGN OF DFIG BASED WIND TURBINE

There are different strategies for controlling the DFIG. In this thesis vector control
(VC) which is also called field oriented control (FOC) of electrical machines has been
implemented. VC employs in both the rotor side and grid side converter. Using this
control strategy it is possible to decouple and construct the control of 𝑑 component
and 𝑞 component separately. Which minimizes the control design complication and
accomplishes various control goals. The main purpose of the RSC is to control the
active and reactive power. That allows the turbine to capture the maximum power from
the wind, while also providing reactive power supports to the grid. Correspondingly,
GSC aims to keep the DC bus voltage constant. In this section, the control of RSC is
presented. In RSC control, we have the current loops of the DFIG and the mechanical
speed (rotor speed controller). Direct current (𝑖𝑑𝑟 ) is proportional and controls the
reactive power, whereas, the quadrature currents are proportional to the toque or stator
active power. This section also presents the GSC control, this control includes the bus
voltage references and controls the grid reactive power reference (𝑄𝑔 _𝑟𝑒𝑓). It also
includes the derivation of the transfer function of a closed-loop second order system
for the rotor side and grid side of a DFIG. Moreover, an angle calculator also called
phase-locked loop (PLL) is designed in both RSC and GSC control, in order to achieve
a three-phase stator voltage grid synchronization, that offers the reliability to estimate
and reject smaller harmonics. Furthermore, PI controller and the MPPT approach using
indirect speed controller has been applied in order for the mechanical torque (𝑇𝑡 ) to
follow the maximum power curve in response to the wind speed variations.

32
4.1. RSC CONTROL

One of the main objectives of RSC control is the mechanical speed to obtain the
maximum power available from the wind and vary according to the MPPT of the wind
[54]. Figure 4.2 shows the control design of RSC (of current control loops). The
control of RSC contains two control loops: internal and external current control loops.
The internal current loop controls direct current (𝑖𝑑𝑟 ) and quadrature current (𝑖𝑞𝑟 ) and
the external current loop controls the speed of the machine (𝜔𝑚 ) and reactive stator
power (𝑄𝑠 ). The 𝑞 rotor current is in proportion to the electromagnetic torque (𝑇𝑒𝑚 ),
which makes 𝑖𝑞𝑟 to control the electromagnetic torque as well as the mechanical speed
(𝜔𝑚 ) if requires. Correspondingly, direct rotor current (𝑖𝑑𝑟 ) is responsible of stator
reactive power (𝑄𝑠 ).

As it can be seen from Figure 4.2 that the difference of the stator reactive power and
its reference as well as the difference of the machine speed and its references are both

fed to the regulators and gives an output of direct (𝑖𝑑𝑟 & 𝑖𝑑𝑟 ) and quadrature (𝑖𝑞𝑟 &
∗ ∗
𝑖𝑞𝑟 ) rotor currents, respectively. The values of 𝑖𝑞𝑟 and 𝑖𝑞𝑟 is subtracted and fed to the

regulator and, similarly, the values of 𝑖𝑑𝑟 and 𝑖𝑑𝑟 is subtracted and fed to the regulator.
The vector control of a DFIG can be carried out in a synchronous rotating 𝑑𝑞 reference
frame [56], where for instance, the 𝑑-axis is lined up with the stator flux (𝜓𝑠 ) as shown
in Figure 4.1.

q β
=
=0
ωs
d

θs α
Figure 4.1. Phasor diagram of a synchronous rotating dq reference frame.

33
For instance, from section (3.3), DFIG model in synchronous 𝑑𝑞 representation, if we
substitute Equation (3.14) into the rotor voltages of Equation (3.13), we get the rotor
voltages (consider 𝜓𝑞𝑠 = 0):

𝑑𝑖𝑑𝑟 𝐿𝑚 𝑑
𝑣𝑑𝑟 = 𝑅𝑟 𝑖𝑑𝑟 + 𝜎𝐿𝑟 − 𝜔𝑟 𝜎𝐿𝑟 𝑖𝑞𝑟 + ⃗⃗⃗⃗
|𝜓𝑠 |
𝑑𝑡 𝐿𝑠 𝑑𝑡
{ 𝑑𝑖𝑞𝑟 𝐿
(4.1)
𝑣𝑞𝑟 = 𝑅𝑟 𝑖𝑞𝑟 + 𝜎𝐿𝑟 + 𝜔𝑟 𝜎𝐿𝑟 𝑖𝑑𝑟 + 𝜔𝑟 𝐿𝑚 |𝜓⃗⃗⃗⃗𝑠 |
𝑑𝑡 𝑠

In DFIG the stator side is directly linked to the grid at a constant AC voltage, so the
𝑑
𝜓𝑠 is constant; therefore, the derivatives of ⃗⃗⃗⃗𝑠 | = 0. Equation (4.1) demonstrates
|𝜓
𝑑𝑡

that the direct and quadrature current controls can be achieved by using a PI controller
for each component, as shown in Figure 4.2. At the output of the current control, cross
terms of functions can be added in order to support the PI controllers.

After implementing the current control loops and flux angle calculations, the overall
control design can be presented. As it can be seen in Figure 4.1, the 𝑑-axis reference
⃗⃗⃗⃗𝑠 ), the torque equation in 𝑑𝑞 representation is expressed
is lined up with a stator flux (𝜓
as follows:

3 𝐿𝑚 3 𝐿𝑚
𝑇𝑒𝑚 = 2 𝑝 (𝜓𝑞𝑠 𝑖𝑑𝑟 − 𝜓𝑑𝑠 𝑖𝑞𝑟 ) ⇒ 𝑇𝑒𝑚 = − 2 𝑝 ⃗⃗⃗⃗𝑠 |𝑖𝑞𝑟 ⇒ 𝑇𝑒𝑚 = 𝐾𝑇 𝑖𝑞𝑟
|𝜓 (4.2)
𝐿𝑠 𝐿𝑠

Equation (4.2) shows that the quadrature rotor current 𝑖𝑞𝑟 is directly proportional to
the electromagnetic torque (𝑇𝑒𝑚 ), that is, 𝑖𝑞𝑟 controls 𝑇𝑒𝑚 . Similarly, it is possible to
derive stator reactive power (𝑄𝑠 ) equation in the 𝑑𝑞 frame, which indicates that the
direct rotor current (𝑖𝑑𝑟 ) is responsible for 𝑄𝑠 .

⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗
3
𝑄𝑠 = 2 (𝑣𝑞𝑠 𝑖𝑑𝑠 − 𝑣𝑑𝑠 𝑖𝑞𝑠 ) = − 2 𝜔𝑠
3 𝐿𝑚
⃗⃗⃗⃗𝑠 | (𝑖𝑑𝑟 − |𝜓𝑠 |) ⇒ 𝑄𝑠 = 𝐾𝑄 (𝑖𝑑𝑟 − |𝜓𝑠 |) (4.3)
|𝜓
𝐿𝑠 𝐿 𝑚 𝐿 𝑚


Hence, because of the orientation we chose, we can see that both current (𝑖𝑞𝑟 , 𝑖𝑞𝑟 ) and

(𝑖𝑑𝑟 , 𝑖𝑑𝑟 ) allow us to control the torque and active/reactive power independently. Based
on the above expressions, Figure 4.2 shows the control design of RSC.

34
Figure 4.2. Control design of the RSC [55].

4.1.1. Transfer Function for the RSC

Figure 4.3 shows the equivalent block diagram of a second order closed loop system
for 𝑖𝑑𝑟 , 𝑖𝑞𝑟 and their references, where both current loops are equals, having two poles
and a zero with a PI controller.

Figure 4.3. Equivalent second order system of a closed loop current control with
regulator [55].

35
The transfer function for both current control loops can be expressed as follows:

𝑖 (𝑠) 𝑖𝑞𝑟 (𝑠) 𝑠𝑘𝑝 +𝑘𝑖 𝑠𝑘𝑝 +𝑘𝑖


𝑇𝐹 = 𝑖𝑑𝑟
∗ (𝑠) = 𝑖 ∗ = 𝜎𝐿 2 =𝐿 2 (4.4)
𝑑𝑟 𝑞𝑟 (𝑠) 𝑟 𝑠 +(𝑘𝑝 +𝑅𝑟 )𝑠+𝑘𝑖 𝑓 𝑠 +(𝑘𝑝 +𝑅𝑓 )𝑠+𝑘𝑖

This can be equivalent to:

𝑖 (𝑠) 𝑖𝑞𝑟 (𝑠) 𝑖(𝑠) (𝑠𝑘𝑝 +𝑘𝑖 )/𝐿 (𝑠𝑘𝑝 +𝑘𝑖 )/𝐿 (𝑠𝑘𝑝 +𝑘𝑖 )/𝐿
𝑇𝐹 = 𝑖𝑑𝑟
∗ (𝑠) = 𝑖 ∗ = 𝑖 ∗(𝑠) = (𝑘𝑝 +𝑅) 𝑘 ⇒ 𝑠2 +2𝜉𝜔 2 = (𝑠+𝜔𝑛 )2
(4.5)
𝑑𝑟 𝑞𝑟 (𝑠) 𝑠2 + 𝑠+ 𝑖 𝑛 𝑠+𝜔𝑛
𝐿 𝐿

Where:

𝑘𝑝 = 2𝜔𝑛 𝐿 − 𝑅 𝑎𝑛𝑑 𝑘𝑖 = 𝜔𝑛2 𝐿 (4.6)

So here in Equation (4.5), we can put as a general inductance and general resistance,
and it can be applied to many machines. Furthermore, the proportional and integral
gains can be obtained by Equation (4.6) and we can choose 𝜔𝑛 , which are the poles of
the Closed loop transfer functions according to the dynamic performance we want to
achieve. Figure 4.4 represents the bode plot diagram of 𝑑𝑞 closed loop transfer
function for the RSC.

Figure 4.4. Bode plot diagram of the closed loop for RSC.

36
4.2. GSC CONTROL

It is very important to control some components in the GSC. Because it is impossible


to make it operate effectively without controlling certain magnitudes in the grid side
system. one of the purposes of the GSC control is to maintain the DC bus voltage
constant, irrespective of the region of operation. It also operates at a unit power factor.
In this section, vector control strategy in the GSC is presented. Figure 4.5 describes
the power flow design from the GSC to the grid.

pf pg
Qf Qg
Lf
Rf

Grid Side
Grid
Converter

Figure 4.5. Power flow design.

The GSC is responsible for controlling parts of the power flow of the DFIG. The power
produced by the three-blade wind turbine is partially supplied through the rotor of the
DFIG. For instance, the power flow that passes through the rotor also passes through
the dc-link and then sent to the grid by the GSC. Figure 4.6 represents grid voltage
alignment with the 𝑑-axis. The GSC is connected to the grid through an 𝑅𝐿 filter.
Modifying the grid voltages to be lined up with the 𝑑 component (𝑣𝑞𝑔 ), the filter
voltages of the 𝑑𝑞 references can be described as:

𝑑𝑖
𝑣𝑑𝑓 = 𝑅𝑓 𝑖𝑑𝑔 + 𝐿𝑓 𝑑𝑔 + 𝑣𝑑𝑔 − 𝜔𝑠 𝐿𝑓 𝑖𝑞𝑔
𝑑𝑡
{ 𝑑𝑖𝑞𝑔
(4.7)
𝑣𝑞𝑓 = 𝑅𝑓 𝑖𝑞𝑔 + 𝐿𝑓 𝑑𝑡 + 𝜔𝑠 𝐿𝑓 𝑖𝑑𝑔

q β

ωa =ω
d
δ
ϕ θs
α

Figure 4.6. Grid voltage alignment with the d-axis.


37
Voltage oriented vector filters for the LCL filter can be given as:

𝑑𝑖 𝑑𝑖
𝑣𝑑𝑓 = 𝑅𝑓 𝑖𝑑𝑓 + 𝐿𝑓 𝑑𝑡𝑑𝑓 + 𝑅𝑔 𝑖𝑑𝑔 + 𝐿𝑔 𝑑𝑡𝑑𝑔 +𝑣𝑑𝑔 − 𝜔𝐿𝑓 𝑖𝑞𝑓 − 𝜔𝐿𝑔 𝑖𝑞𝑔
{ 𝑑𝑖 𝑑𝑖𝑞𝑔
(4.8)
𝑣𝑞𝑓 = 𝑅𝑓 𝑖𝑞𝑓 + 𝐿𝑓 𝑑𝑡𝑞𝑓 + 𝑅𝑔 𝑖𝑑𝑔 + 𝐿𝑔 𝑑𝑡 + 𝜔𝐿𝑓 𝑖𝑑𝑓 + 𝜔𝐿𝑔 𝑖𝑑𝑔

So, the filters active and reactive power can be given as follows:

3
𝑃𝑔 = 2 (𝑣𝑑𝑔 𝑖𝑑𝑔 + 𝑣𝑞𝑔 𝑖𝑞𝑔 )
{ 3 (4.9)
𝑄𝑔 = (𝑣𝑞𝑔 𝑖𝑑𝑔 − 𝑣𝑑𝑔 𝑖𝑞𝑔 )
2

Moreover, the topology used for controlling the GSC is a two-level voltage source
converter (2L-VSC). The pulses for the controlled switches (𝑆𝑎_𝑔 , 𝑆𝑏_𝑔 , 𝑆𝑐_𝑔 ) of the
2L-VSC are, for instance, the output voltages of the converter, are generated in order
to control the DC bus voltage (𝑉𝑏𝑢𝑠 ) and the grid reactive power (𝑄𝑔 ). In this section
grid voltage-oriented vector control (GVOVC) approach has been presented.

Figure 4.7. Control block design of the GSC [42].

For instance, the control block design of the GSC can be seen in Figure 4.7. From the
references (𝑉𝑏𝑢𝑠 , 𝑄𝑔∗ ), it creates pulses for the controlled switches 𝑆𝑎_𝑔 , 𝑆𝑏_𝑔 , 𝑆𝑐_𝑔 .
Therefore, the modulator produces the pulses from the 𝑎𝑏𝑐 voltage references for the
∗ ∗ ∗
GSC: 𝑣𝑎𝑓 , 𝑣𝑏𝑓 , 𝑣𝑐𝑓 . Thus, these 𝑎𝑏𝑐 voltage references are first produced in 𝑑𝑞 axis
∗ ∗ ∗ ∗
(𝑣𝑑𝑓 , 𝑣𝑞𝑓 ), then converted to 𝛼𝛽 axis (𝑣𝛼𝑓 , 𝑣𝛽𝑓 ), and eventually generates 𝑎𝑏𝑐 voltage

38
∗ ∗
references. Then the 𝑑𝑞 voltage references (𝑣𝑑𝑓 , 𝑣𝑞𝑓 ) are separately produced by the
∗ ∗
𝑑𝑞 current (𝑖𝑑𝑔 , 𝑖𝑞𝑔 ) references. This indicates that by changing 𝑣𝑑𝑓 ; 𝑖𝑑𝑔 is mainly
changed. Moreover, there is also cancellation of coupling terms at the output of the
current PI regulators, for a good dynamic performance and it can be given as:

𝑒𝑑𝑓 = −𝜔𝑠 𝐿𝑓 𝑖𝑞𝑔


{ (4.10)
𝑒𝑞𝑓 = 𝜔𝑠 𝐿𝑓 𝑖𝑑𝑔


As shown in Figure 4.7, the direct grid current references (𝑖𝑑𝑔 ) is proportional and
controls the grid active power (𝑃𝑔∗ ) and the quadrature grid current (𝑖𝑞𝑔

) implies and
controls grid reactive power (𝑄𝑔∗ ). Note that, under ideal conditions, the term 𝑣𝑑𝑔 is
equal to the amplitude of the grid voltage and is constant. Thus, the constant terms
required can be deduced from Equation (4.8) and is given by:
1
𝑘𝑃𝑔 = 3
𝑣
2 𝑑𝑔
{ 1 (4.11)
𝑘𝑃𝑔 = 3
− 𝑣𝑑𝑔
2

4.2.1. Transfer Function for the GSC

Figure 4.8 demonstrates the current loops of the L filter. From the voltage Equation
(4.7), the L filters at an open loop can be represented as:
𝑑𝑖
𝑣𝑑𝑓 = 𝑅𝑓 𝑖𝑑𝑔 + 𝐿𝑓 𝑑𝑡𝑑𝑔
{ 𝑑𝑖𝑞𝑔
(4.12)
𝑣𝑞𝑓 = 𝑅𝑓 𝑖𝑞𝑔 + 𝐿𝑓 𝑑𝑡

Figure 4.8. Current loops for the L filter [42].

39
By using Laplace transform the following transfer function is obtained:

𝑖𝑑𝑔 (𝑠) 1
=𝐿 (4.13)
𝑣𝑑𝑓 (𝑠) 𝑓 𝑠+𝑅𝑓

and
𝑖𝑞𝑔 (𝑠) 1
=𝐿 (4.14)
𝑣𝑞𝑓 (𝑠) 𝑓 𝑠+𝑅𝑓

As a result, the control block design can be used to express this fact, as shown in Figure
4.8 using this method, we take into consideration an ideal PI regulator. As a result,
transfer function of the current closed loop can be given as:

𝑖𝑑𝑠 (𝑠) 𝑠𝑘𝑝 +𝑘𝑖


∗ (𝑠) = 𝑠2 (𝐿 (4.15)
𝑖𝑑𝑠 𝑓 )+𝑠(𝑅𝑓 +𝑘𝑝 )+𝑘𝑖

and
𝑖𝑞𝑠 (𝑠) 𝑠𝑘𝑝 +𝑘𝑖
∗ (𝑠) = 𝑠2 (𝐿 (4.16)
𝑖𝑞𝑠 𝑓 )+𝑠(𝑅𝑓 +𝑘𝑝 )+𝑘𝑖

Now, if we equate the denominator of Equation (4.15) or (4.16) to the denominator of


standard second order transfer function, we get:

𝑠 2 (𝐿𝑓 ) + 𝑠(𝑅𝑓 + 𝑘𝑝 ) + 𝑘𝑖 = 𝑠 2 + 2𝜉𝜔𝑛 𝑠 + 𝜔𝑛2 (4.17)

Thus, the integral (𝑘𝑖 ) and proportional (𝑘𝑝 ) constant values are:
𝑘𝑖 = 𝐿𝑓 𝜔𝑛2
{ (4.18)
𝑘𝑝 = 𝐿𝑓 2𝜉𝜔𝑛 − 𝑅𝑓

As we explained in previous section, we can choose 𝜔𝑛 and 𝜉 according to the dynamic


performance we want to achieve (Mitsubishi MWT 92 manuals).

40
Figure 4.9. Bode plot diagram of both RSC and GSC.

In Figure 4.9, the blue one indicates the GSC current loops and the red one shows RSC
current loops, we have current loops of both GSC and RSC with a band width very
similar. Figure 4.10 illustrates the step response for the grid side and in 15 milliseconds
it reaches the steady state.

Figure 4.10. Step response of the GSC.

41
4.3. PHASE LOCKED LOOP

An angle calculator called PLL is designed in both the RSC control and GSC control,
for voltage and current coordinate transformations. Beacause of it’s closed loop nature,
it assures the stability and rejection of perturbations to the angle estimation. Figure
4.11 shows the PLL block design at the grid, that takes an input of grid 𝑎𝑏𝑐 voltages
from the measured DFIG and then transforms to 𝑑𝑞 reference frame. Then the PLL
aligns the quadrature grid voltage (𝑣𝑞𝑔 ) by comparing direct grid voltage (𝑣𝑞𝑔 ) with
zero reference voltage. The error signal passes through the PI regulator that provides
the angular frequency (𝜔𝑠 ) of the grid voltages. The conversion method and the PLL
MATLAB/Simulink design is shown in Appendix B.
vag vdg = error
vbg abc ωs θg
vqg PI dt
dq
vcg

θg

Figure 4.11. Classical PLL block design.

Finally, the control strategy (GVOVC) that has been presented is able to control the
DC bus voltage (𝑉𝑏𝑢𝑠 ) and grid reactive power (𝑄𝑔 ) as specified, it also provides good
performance in the dynamic response because of the vector control structure. The
Simulations results in detail is presented in Part 7.

4.4. CONTROL STRATEGY

This section presents the control strategy we applied in this work. This includes typical
PI control and MPPT by indirect speed control strategy.

4.4.1. Proportional and Integral Control

The PI controllers are some kinds of regulators that combines proportional and integral
operation in one unit [57]. It is important to create a PI regulator, in order to assure

42
that the rotor speed, generated torque, the direct (𝑖𝑑𝑟 ) and quadrature (𝑖𝑞𝑟 ) rotor
currents follow their references. Figure 4.12 illustrates the PI controller of rotor speed
and its reference.

ωref + Te,ref
_ PI

ωm
Figure 4.12. PI with rotor speed.

𝑝
𝐺(𝑠) = 𝐽𝑠 (4.19)

and
𝑅(𝑠) = 𝑘𝑝 + 𝑘𝑖 /𝑠 (4.20)

Transfer function for closed loop can be expressed as:


𝑅(𝑠)𝐺(𝑠)
𝐻(𝑠) = 1+𝑅(𝑠)𝐺(𝑠) (4.21)

Which gives:
𝑘𝑝 𝑠+𝑘𝑖
𝐻(𝑠) = 𝑗/𝑝𝑠2 +𝑘 (4.22)
𝑝 𝑠+𝑘𝑖

Second order transfer function is represented as:


𝜔𝑐2
(𝑠) = (4.23)
𝑠2 +2𝜉𝜔𝑐 𝑠+𝜔𝑐2

If we equate the denominator of Equation (4.22) with (4.23) it gives:


2𝐽
𝑘𝑝 = 𝑝 𝜉𝜔𝑛
{ 𝜔2 𝐽
(4.24)
𝑘𝑖 = 𝑝𝑛

Here we can have the PI controller for direct current of the RSC.

43
ird,ref + vrd + vrd
PI
_
+

ird vrd,comp
Figure 4.13. PI with direct rotor current and voltage.

Since the GSC controls the dc bus voltage, we can have the PI controller for 𝑉𝑏𝑢𝑠 and
𝑖𝑑𝑔 component.

+ igd,ref + v gd + vgd
_ PI _ PI
+

igd vgd,comp
vdc
Figure 4.14. PI in the GSC.

Typical wind turbine has a nominal power between 1.5 and 3 MW. One of the most
significant models presented in this thesis is Mitsubishi MWT 92. Mitsubishi started
with variable speed wind turbine (VSWT) DFIG based technology with models: MWT
92 rated at 2300 kW, MWT 95, 100 and 102 rated at 2400 kW and MWT 92 rated at
2.4 MW.

Figure 4.15. Torque coefficient versus lambda and power curve.

44
The proposed output power curve shown in Figure 4.15 has been confirmed by
comparing with the output of the generated power with different wind speed. Table
shows the values of different wind speed and the generated output power.

Table 4.1. The values of wind speed and the generated power.

Wind speed Power

7.50 m/s 0.6342 MW

9.50 m/s 1.2840 MW

11.8 m/s 2.3635 MW

For instance, from Equation (4.6), we can obtain the values of proportional (𝑘𝑝 ) and
integral (𝑘𝑖 ) gains for the rotor side and can be given as (note: 𝑘𝑖 has the same value
in both 𝑑 and 𝑞 current components and 𝑘𝑝 also has the same values for both 𝑑 and 𝑞
components):

𝑘𝑝 = 0.5771
{
𝑘𝑖 = 491.5995

Moreover, when we consider PI controller as a control strategy, from Equation (4.24)


the value of 𝑘𝑝 and 𝑘𝑖 will be:
𝑘𝑝 = 5080
{
𝑘𝑖 = 203200

The same is applied to the grid side. The values of the PI controller coefficients for the
grid side are:
𝑘𝑝 = 0.3016
{
𝑘𝑖 = 56.8489

The value of gains, considering PI controller as a control strategy (having an input of


bus voltage and its reference) has been chosen until the desired value is obtained:

𝑘𝑝 = −1000
{
𝑘𝑖 = −300000

45
4.4.2. Indirect Speed Control

Different strategies have been presented in the literature to control the wind turbine
under partial load while tracking the maximum power extraction. At this point, there
are two types of speed control: Direct and Indirect speed control. In this thesis, we
consider MPPT by indirect speed control. This approach is important because it
ensures a variable speed operating system, to optimizes the power output across a
given wind speed. The power generated by the three blades becomes maximum if we
consider maximizing the power coefficient (𝐶𝑝 ).

Figure 4.16. MPPT stability curve [42].

As a result, we need to maximize lambda (𝜆) such that the blades capture maximum
amount of wind energy. Assume the variable speed turbine is functioning at ‘a’ level
of the curve that can be seen in Figure 4.16 with speed of the wind and rotor torque
constant. When the rotor speed is lowered to 𝛺𝑡−𝑏 , the operational level shifts to level
‘b’ and torque becomes 𝑇𝑡−𝑏 . Mechanical torque is held at the previous level relating
to 𝑇𝑡−𝑎 , thus 𝑇𝑡−𝑏 is greater than electromagnetic torque 𝑇𝑒𝑚 and rotor speed rises till
it stabilizes near 𝛺𝑡−𝑎 level once again. Thus, the reference of rotor speed at optimum
lambda can be given as:

𝜆𝑜𝑝𝑡 𝑣
𝜔𝑟𝑒𝑓 (𝑢) = (4.25)
𝑅

46
Figure 4.17 shows the scheme of MPPT by indirect speed control strategy.

Figure 4.17. Indirect speed control strategy [42].

The electromagnetic torque (𝑇𝑒𝑚 ) generated can be expressed as:

1 𝑅 2 𝛺2 (𝐶𝑝 )𝑚𝑎𝑥
𝑇 = 2 𝜌𝜋𝑅 3 2 (4.26)
(𝜆𝑜𝑝𝑡 ) 𝜆𝑜𝑝𝑡

Equation (4.26) can be simplified as

𝑇 = −𝑘𝑜𝑝𝑡 𝛺 2 (4.27)
Equation (4.27) leads to the control strategy shown in Figure 4.17.

4.5. STEADY-STATE ANALYSIS OF THE DFIG

It is very important to study the operation of the steady-state analysis of DFIG, in order
to know the stable operating point of the system. This study can be important for
determining the appropriate magnetizing method in terms of efficiency and estimating
the necessary rotor currents at a specific operating point for dimensioning the power
electronic converter. Thus, to analyze the steady-state, the following equations in
Table 4.2 and Table 4.3 are applied.

4.5.1. Case 1: When Reactive Stator Power is Zero

Reactive stator power (𝑄𝑠 ) is set to zero and all the magnitudes of the DFIG at steady-
state is analyzed by using steady-state equations shown in Table 4.2.
47
Table 4.2. Steady-state magnitudes of the DFIG for Qs=0 [55].

Stator
fluxes

Rotor
fluxes

Stator
currents

Rotor
currents

Stator
voltages

Rotor
voltages

Slip

Active
powers

Reactive
powers

Efficiency

48
4.5.2. Case 2: When Direct Rotor Current is Zero

Here, the quadrature rotor current is set to zero and all the magnitudes of the DFIG at
steady-state is analyzed by using steady-state equations shown in Table 4.3.

Table 4.3. Steady-state magnitudes of the DFIG for idr=0 [55].

Stator
fluxes

Rotor
fluxes

Stator
currents

Rotor
currents

Stator
voltages

Rotor
voltages

Slip

Active
powers

Reactive
powers

Efficiency

49
PART 5

ANALYSIS OF DFIG BASED WIND TURBINE DURING VOLTAGE DIPS

Voltage dips are short term drops in rms voltages that can be caused by an overload,
short circuit or the beginning of electrical generators. It mostly occurs when the rms
voltage reduces between 10 to 90 percent of its normal value during 0.5 cycles to 1
minute. Voltage dips affect all electrical drives attached to a grid varying degree. On
the other hand, one of the principal disadvantages of the DFIG is that they are very
sensitive to this grid disturbances. In this section, basic space vector magnitudes for
certain typical voltage dips are presented. It also presents the analysis of DFIG system
during symmetrical and asymmetrical voltage dips. The aim of this vector magnitude
is to investigate the positive and negative sequence voltage vectors for various forms
of dips.

5.1. SYMMETRICAL VOLTAGE DIPS

In this section, the analysis of symmetrical voltage dips is only presented. For instance,
when the voltage dip is observed immediately by the stator of the DFIG, it is very
important to analyze the behavior of the stator flux in order to comprehend the
difficulties that results from the disturbance of the dip. In order to do that, it must be
taken into consideration with park transformation model from chapter 3 (Figure 3.5),
the stator side voltage and flux references can be given as:

𝑠
𝑠 𝑠 ⃗⃗⃗⃗⃗𝑠
𝑑𝜓
𝑣𝑠 = 𝑅𝑠 ⃗⃗𝑖𝑠 +
⃗⃗⃗ (5.1)
𝑑𝑡

and
𝑠 𝑠 𝑠
⃗⃗⃗⃗
𝜓𝑠 = 𝐿𝑠 ⃗⃗𝑖𝑠 + 𝐿𝑚 ⃗⃗𝑖𝑟 (5.2)

𝑠
Thus, by comparing stator current ⃗⃗𝑖𝑠 of Equation (5.1) and (5.2) we get:

50
𝑠
⃗⃗⃗⃗⃗𝑠
𝑑𝜓 𝑠 𝑅 𝑠 𝐿𝑚 𝑠
𝑣𝑠 − 𝐿 𝑠 ⃗⃗⃗⃗
= ⃗⃗⃗ 𝜓𝑠 + 𝑅𝑠 ⃗⃗𝑖 (5.3)
𝑑𝑡 𝑠 𝐿𝑠 𝑟

Alternatively, it is important to change the equivalent electrical circuit of the DFIG,


that we explained in section (3.3), to a smaller equivalent circuit. Therefore, from
Equation (5.1) and (5.2), the following expressions are obtained. Which leads to the
circuit diagram shown in Figure 5.1 (a).

𝑟 𝐿𝑚 𝑟 𝑟 𝐿 𝑟 2 𝑑 𝑟
𝑣𝑟 =
⃗⃗⃗ ⃗⃗⃗𝑠 − 𝑗𝜔𝑚 ⃗⃗⃗⃗
(𝑣 𝜓𝑠 ) + [𝑅𝑟 + ( 𝑚 ) 𝑅𝑠 ] ⃗⃗𝑖𝑟 + 𝜎𝐿𝑟 ⃗⃗𝑖 (5.4)
𝐿𝑠 𝐿𝑠 𝑑𝑡 𝑟

(a) (b)
Figure 5.1. (a) Equivalent circuit for voltage dips analysis of DFIG and (b) Space
vector phasor diagram as sub synchronous mode [42].

From Equation (5.4), it can be seen that the rotor voltage is described as a function of
rotor current, stator voltage and flux and equivalent resistance and inductance. Figure
5.1 (b) shows the space vector phasor diagram.

5.1.1. Crowbar Protection

In [42] it is very recommendable to work with the equivalent circuit of the DFIG
represented in Figure 5.2 (a) and at steady state and normal operation conditions shown
in Figure 5.2 (b). What happens is that when there is a smaller voltage dip, the stator
𝑟
⃗⃗⃗𝑠 ) that was in normal operation is suddenly reduced to a much smaller value
voltage (𝑣
𝑟
⃗⃗⃗⃗⃗⃗
(𝑢 𝑅𝐿 ). For this purpose, the stator flux cannot evolve rapidly to a very small value.
𝑟 𝑟
⃗⃗⃗⃗⃗⃗
Therefore, in order to keep the voltage (𝑢 ⃗⃗⃗𝑟 ).
𝑅𝐿 ) small, we need a big rotor voltage (𝑣

51
Normally, the back-to-back converters are not dimensions with such big voltages
𝑟
⃗⃗⃗𝑟 ), for that purpose it loses control. Thus, in order to avoid this problematic issue,
(𝑣
it’s very important to put a hardware element called crowbar protection.

There are different hardware crowbar configurations but most of them are equivalent.
In this thesis, the crowbar protection has been chosen as one diode bridge, one switch
and one resistance. Figure 5.2 (a) and (b) illustrates DFIG system supplied with three
phase DC crowbar protection and one phase equivalent circuit of DFIG system when
the crowbar is activated, respectively.

(a)

(b)
Figure 5.2. (a) DFIG system supplied with three phase DC crowbar protection and
(b) one phase equivalent circuit of DFIG system when the crowbar is activated [42].

52
As shown in Figure 5.2 (b) the crowbar resistance is placed in each terminal of the
𝑟
⃗⃗⃗⃗𝑠 ) evolution more faster during the
rotor. Thus, it is possible to make the stator flux (𝜓
𝑟
⃗⃗⃗𝑠 ) at the
dip and reach the new steady state much faster with the reduced voltages (𝑣
grid and therefore at the stator.

5.2. ASYMMETRICAL VOLTAGE DIPS

This section presents analysis of asymmetrical voltage dips. Practically, asymmetrical


(unbalanced) faults are mostly experienced than symmetrical faults. DFIG can be
required to work under unbalanced grid voltages in a certain circumstances and
applications, such as in weak grids where the nonlinear loads cause unbalance of grid
voltages. For instance, the unbalanced voltage has direct impact on the operation of
the DFIG. Which affects the performance of DFIG, if no additional action is performed
[42]. The unbalanced occurrence can be investigated using a technique called sequence
decomposition. Accordingly, this unbalanced voltage indicates the existence of both a
positive and negative sequence. Therefore, in order to overcome these kinds of
problematic issues, it’s better to use a control strategy called dual vector control
strategy of a DFIG.

Figure 5.3. Dual vector control strategy of a DFIG [55].

53
Since Figure 5.3 deals with positive and negative sequence of voltages and currents
with the DFIG, it should include negative sequence current loops in order to be able to
properly control the generator. Furthermore, the filter (in Figure 5.4) is set in order to
eliminate the coupling oscillations between the sequences. Thus, the filter could be
tuned up to 100 Hz (which is double of the frequency of the stator) and in order to get
an acceptable performance of the loops, this double frequency should be removed.

Figure 5.4. Sequence decomposition for the current loops [55].

54
PART 6

SIMULATION RESULTS

In this section, a detailed simulation result of a DFIG based wind turbine system is
presented. The overall system has been carried out in MATLAB/Simulink software.
This system consists of different models and sub-models, this includes the DFIG, wind
speed, wind turbine model, RSC control, universal bridges (converters), PWM, GSC
control, PI controller, and MPPT by indirect speed control strategy.

6.1. SIMULATION OF THE WIND TURBINE MODEL

As previously discussed in section 2, a practical wind turbine has a rated power of


between 1.5 and 3 MW. Thus, this thesis chooses a 2.4 MW three-blade wind turbine
for system simulations. Here, we assume that the wind turbine model is made up of
several parts: a wind speed, an aerodynamic model, a drive train model, and an
electrical generator.

6.1.1. Simulation of the Wind Speed Model

In this work, the wind speed model is designed by adding four constituents such as
constant, turbulence, ramp, and gust constituents. The gust constituents can be used to
describe unusual sudden changes in wind speed.

Figure 6.1. Wind speed model at 7.5 m/s.

55
Equation (2.4), explained in chapter 2, leads to the wind speed model represented in
Figure 6.1. Thus, the simulation result of this model at 7.5 m/s of wind speed is shown
in Figure 6.2.

Figure 6.2. Wind speed profile.

6.1.2. Simulation of the Aerodynamic Model

It is possible to design the aerodynamic model by determining the turbine torque


generated by the three-blade wind turbine. To do that, some parameters should be
considered, such as the radius of the blade, turbine angular speed, wind speed, and
turbine coefficient. Moreover, the gearbox ratio should also be considered in both the
low-speed shaft and high-speed shaft.

Figure 6.3. Simulink design of the aerodynamic model.

56
Equation (2.10) discussed in chapter 2 leads to the Simulink design of the aerodynamic
model shown in Figure 6.3. The speed of the wind is set to 7.5 m/s and the value of
the tip step ratio (lambda) which is 7.21 is maintaining the specified requirement where
the optimum lambda has been considered to be 7.2 and the value of turbine coefficient
(𝐶𝑡 ) is 0.06. Thus, the torque generated by the three-blade wind turbine is -4907 N.m.
The negative shows that the machine is operating as a generator convention.

6.2. SIMULATION OF THE DFIG SYSTEM

Figure 6.4 represents the Simulink design of the DFIG system. The system consists of
a wound asynchronous induction machine (as a DFIG), two back-to-back converters,
and RSC control.

Figure 6.4. DFIG Simulink design.

The simulation result for the above model is presented in Figure 6.5. The speed
reference is set at 1500 and 1800 rev/min. The electromagnetic torque is modified as
the rotor speed changes. Accordingly, the electromagnetic torque depends on the rotor
current. Thus, electromagnetic torque is regulated by the quadrature rotor current.
Furthermore, the direct rotor current is always set to zero because the machine is
magnetizing from the stator. Moreover, the stator voltage is constant and operates
properly. Similarly, the rotor voltage references and stator and rotor currents are also
working properly.

57
Figure 6.5. Magnitudes of the DFIG with the rotor speed and electromagnetic torque.

6.3. STEADY-STATE ANALYSIS OF DFIG

The MATLAB plot shown in Figure 6.6 is the most representative magnitudes of the
model equations discussed in chapter 4 (see Table 4.2 and 4.3). In this case, as input
is, a torque as a function of the rotational speed (rev/min). It can be realized that the
torques are negative; therefore, the machine (DFIG) is operating as a generator. The
machine represents something similar to the operation at the steady-state of a wind
turbine based DFIG system.

58
Figure 6.6. Steady-state analysis of DFIG.

The rotational speed is set between 900 and 1800 rpm and the value of the total
mechanical power is almost the same as the product of torque and rotational speed.
Furthermore, it has been calculated and realized that the stator active power (2 MW)
is much greater than the rotor active power (400 kV). The rotor active power can be
negative or positive depending on the rotational speed. Obviously, in case 1, when the
stator reactive power (𝑄𝑠 ) is set at zero, the value of the maximum stator current will
be lower. On the other hand, when direct rotor current (𝑖𝑑𝑟 ) is set to zero, it provides a
lower stator current, but the reactive power will be high. Moreover, the stator voltage
amplitude is always constant (690 V), while the rotor voltage depends on the rotor
speed. For instance, the designed system (Figure 6.6) is tested for different rotational
speeds. For example, at 1500 rpm the torque is -7341 Nm and at 1800 rpm the torque
is -10890 Nm. Similarly, the results of the DFIG system (in Figure 6.5) exactly show

59
the same result. This clearly shows that the DFIG system is operating according to the
Mitsubishi design. The objective of this steady-state analysis is to know the stable
operating point of the system. Finally, it can be realized that the efficiency of the DFIG
is good, reliable, sufficient, and constant between 900 and 1800 rpm.

6.4. SIMULATION OF DFIG BASED WIND TURBINE SYSTEM

In this section, detailed results and comparisons of a DFIG based wind turbine using
MPPT by indirect speed control with typical PI and proposed controller is presented.

6.4.1. Simulation Results using PI Controller

Figure 6.7 shown below represents the simulated result of the system using PI control
for different wind speed values (7.5 and 9.5 m/s). The rotational speed is set at 1800
rpm (188.5 rad/sec) through the operation.

Figure 6.7. Performance of DFIG with PI control at different wind speed.

60
Figure 6.8. Typical PI controller design of the DFIG system.

Table 6. represents the value of the DFIG components at steady state-response with PI
control at different wind speeds.

Table 6.1. Simulated results of the DFIG with PI controller at different wind speeds.
Power
Wind speed Rotor speed Torque
(generated)

7.5 m/s 128.1 rad/sec -4860 Nm -0.622 MW

9.5 m/s 162.5 rad/sec -7825 Nm -1.271 MW

11.8 m/s 201.8 rad/sec -12100 Nm -2.44 MW

From Table 6.1 it can be realized that, as the wind speed increases, the rotor speed,
torque, and generated power also increases. However, in this work, the wind speed
doesn’t go beyond 11.8 m/s. For instance, in case it exceeds that value, the pitch control
will take place to limit the maximum extracted power. Even though, pitch controller
is not implemented, instead a 2.4 MW three blade wind turbine has been carried out in
MATLAB. Thus, the simulation result shown in Figure 6.7 has been verified with the
proposed system (see Figure 4.15). Figure 6.9 represents steady-state response of the
system with typical PI controller.

61
Figure 6.9. Steady-state response of DFIG with typical PI controller at 7.5 m/s.

Table 6.2 analyses the settling time of the DFIG components at different wind speed
value for typical PI control.

Table 6.2. Settling time of DFIG components with typical PI at different wind speed.

Settling Time (5%) (seconds)

Wind speed
Rotor speed Torque 𝒊𝒒𝒓 𝒊𝒅𝒓
(m/s)

7.5 4.665 4.896 5.168 1.117

9.5 3.71 3.914 4.476 1.03

11.8 3.218 3.64 4.1 0.91

62
6.4.2. Simulation Results of the Proposed Method

As shown in Figure 6.9 the performance of the system when PI controller is applied,
it could not satisfy the specification of the DFIG, as there is higher settling time of the
parameters and high overshoot/undershoot for all parameters. Therefore, an efficient
controller is proposed by tunning the gains of 𝑖𝑑 and 𝑖𝑞 until the desired value is
obtained. Figure 6.10 illustrates the proposed Simulink design.

Figure 6.10. Proposed Simulink design of DFIG system.

Figure 6.11. Performance of the proposed controller at different wind speed.

63
Table 6.3. represents the value of the DFIG components at steady state-response with
the proposed method. It further shows that the generated powers at different wind
speeds corresponds to Table 6.1. Therefore, at speed less than 11.8 m/s, the proposed
system performs as desired. Figure 6.12 illustrates the steady-state performance of the
proposed controller at 7.5 m/s.

Table 6.3. Results of the DFIG with the proposed method at different wind speeds.
Power
Wind speed Rotor speed Torque
(generated)

7.5 m/s 128.24 rad/sec -4900 Nm -0.63 MW

9.5 m/s 162.79 rad/sec -7850 Nm -1.28 MW

11.8 m/s 201.86 rad/sec -12200 Nm -2.46 MW

Figure 6.12. Steady-state performance of the proposed controller at 7.5 m/s.

64
Table 6.4. Settling time of the proposed method at different wind speed.

Settling Time (5%) (seconds)

Wind speed
Rotor speed Torque 𝒊𝒒𝒓 𝒊𝒅𝒓
(m/s)

7.5 1.483 2.115 1.894 0.922

9.5 1.35 1.94 1.76 0.81

11.8 1.33 1.7 1.613 0.705

6.4.3. Comparison between Typical PI Controller and Proposed Method

When compared the relevant values in Table 6.2 and Table 6.4, it can be seen that the
performance of DFIG based wind turbine with the proposed method improves in terms
of settling time.

Figure 6.13. Steady-state performance with PI and proposed method at 7.5 m/s.
65
Figure 6.13. shows the steady-state performance for the PI controller and proposed
method at 7.5 m/s. Moreover, it can be observed that the proposed method has less
perturbations compared to the PI controller. Besides, the proposed method provides
strong robustness for controlling the system with variable parameters. From the results,
it is clearly obvious that the proposed method has better performance than the PI.

6.5. IMPLEMENTATION OF DFIG BASED WIND TURBINE SYSTEM

In this section, the implementation and control of a wind turbine based DFIG in RSC
and GSC control is presented. Figure 6.14 represents the overall implementation of a
wind turbine based DFIG.

Figure 6.14. Overall implementation of DFIG based wind turbine.

6.5.1. Implementation of the RSC Control

The main purpose of RSC is to control the stator active power that enables the turbine
to capture the maximum power, while also providing reactive power support to the
grid. The RSC control is also responsible for the MPPT, Figure 6.15 shows indirect
speed control Simulink design.

66
Figure 6.15. Indirect speed control.

In the RSC, we have rotor currents and rotor speed references which is used to generate
the rotor voltage references.

Figure 6.16. Steady-state response for the magnitudes of RSC at different wind
speeds.

Figure 6.16 illustrates the most representative magnitudes of DFIG in the RSC control
at different wind speed. From the result, it can be observed that the performance of the
stator active power (Ps) is almost proportional to the torque. Furthermore, MPPT by
indirect speed control can be realized. When the wind speed changes (8.5 to 11.5 m/s),
67
in this way, MPPT control algorithm operates. Accordingly, the rotor speed changes
to a new (higher) operating point and the quadrature rotor current component (Iq )
becomes higher, which maximizes the stator active power (Ps). For instance, as the
direct rotor current (𝑖𝑑𝑟 ) component changes from 0 to 850 ampere, corresponding
stator reactive power (Qs) also modifies. Thus, this provides decouple control of active
and reactive power.

6.5.2. Implementation of the GSC Control

The main objectives of GSC are to maintain the DC bus voltage constant irrespective
of the operating zone and to control grid reactive power reference. It operates at a unit
power factor. Moreover, the power generated by wind turbine is partly delivered
through the rotor of the DFIG. For example, the power flow that pass via the rotor also
flows through the DC-link and then sent to the grid by the GSC.

Figure 6.17. Steady-state response for the magnitudes of GSC at different wind speed.
68
Figure 6.17 indicates the most representative magnitudes of a vector controlled DFIG
for the GSC at various wind speed. It can be realized that at the steady-state the DC
bus voltage is kept constant. Moreover, the grid voltage references (𝑣𝑑𝑔 and 𝑣𝑞𝑔) are
very small voltages because, we are using a feedback voltage of the grid. At 6 seconds,
it has been performed a variation from 0 to -400 kVAR reactive power at the GSC and
therefore, the quadrature grid current (𝑖𝑞𝑔) is also being modified.

6.6. DYNAMIC ANALYSIS OF CURRENT LOOPS BEHAVIOR

This section investigates the dynamic analysis of current loops behavior in a wind
turbine based DFIG. Figure 6.18 demonstrates the most representative magnitudes in
the RSC.

Figure 6.18. Most representative magnitudes of a wind turbine based DFIG in the
RSC.

69
The implementation of this thesis represents the most representative magnitudes of a
2 MW grid-connected two-level (2L) converter. Figure 6.19 represents the magnitudes
of DFIG based wind turbine in the GSC. The 𝑎𝑏𝑐 grid voltage is connected to the
converter and this converter also controls the DC bus voltage.

Figure 6.19. Most representative magnitudes of a wind turbine based DFIG in the
GSC.

In Figure 6.18, it can be realized that the 𝑖𝑑 component has been a little bit degraded
but thanks to the compensation that the theoretical model (𝑖𝑑𝑟_𝑚) fits quite well with
the actual one (𝑖𝑑𝑟). Correspondingly, at the grid side (Figure 6.19) the theoretical
model (𝑖𝑞𝑔_𝑚) is also similar to the actual model (𝑖𝑞𝑔). Obviously, the actual one has
a ripple due to the switching frequency that the simplified model (𝑖𝑑𝑟_𝑚 & 𝑖𝑞𝑟_𝑚)
doesn’t take into account. In general hardware implementations, there is one delay
between control and PWM or modulation implementations. In case the compensation
of this delay is not used, the theoretical and actual model will be further way for both

70
RSC and GSC. Moreover, it can be observed how 𝑖𝑑𝑔 and 𝑖𝑞𝑔 components are very
accurate. It should be noted that to follow properly the dynamic performance of DFIG
it is not required to reach the saturation of the voltage references. Finally, at the time
instant 2.5 seconds, there is a reactive power variation from 0 to 150 kVAR and it can
be observed, as the grid reactive power references (𝑄_𝑟𝑒𝑓) increases, the magnitudes
of the grid current also increase.

6.7. PERFORMANCE ANALYSIS UNDER VOLTAGE DIPS

This section graphically presents the performance analysis of a grid-connected DFIG


based wind turbine during symmetrical and asymmetrical voltage dips. The main
findings are drawn from the parameters of a 2 MW, 690V DFIG based wind turbine.

6.7.1. Symmetrical Voltage Dips

This section demonstrates the simulation analysis of symmetrical voltage dips. When
grid fault occurs, the converter is damaged. Hence, in order to overcome this kind of
problematic issues, we have included a hardware element called crowbar protection.
This protection is suddenly activated when the voltage dip occurs and return it under
normal operation condition. Moreover, during the crowbar activation, the RSC should
be disabled in order to protect the converter. Figure 6.20 shows the stator flux and the
crowbar current fault analysis.

Figure 6.20. Stator flux and crowbar current under symmetrical voltage dip.

71
Figure 6.21 and 6.22 demonstrates the performances of the RSC and GSC during the
dip.

Figure 6.21. Performance of RSC magnitudes under symmetrical voltage dips.

From the above simulation, the voltage dip has been performed at 3 seconds and the
remaining voltage at the grid is only 10 % of the normal voltage. Once the voltage dip
occurs, the stator flux cannot progress to the steady-state as fast as the stator voltage.
Then the crowbar protection is suddenly activated, this helps the stator flux to progress
much faster and recover. Thus, during this time all the rotor currents goes through the
crowbar protection, which results the RSC to be protected. Furthermore, during the
voltage dip as demanded by the grid codes, it has been performed a high value of 𝑖𝑑
rotor current, in order to provide reactive power through the stator.

72
Figure 6.22. Performance of GSC magnitudes under symmetrical voltage dips.

Once the voltage at the grid is totally recovered, the control of the quadrature rotor
current (𝑖𝑞) component is regained and this provides the torque and active power
demanded by the MPPT algorithm. This means that the rotational speed is controlling
properly. At the GSC, it can be realized that the voltage dip has provoked some
variations in the DC bus voltage. However, this is not that much problematic because,
the GSC is working properly during the dip. It has to be noted that, as demanded by
the grid codes, a huge amount of reactive currents should be provided by a means of
GSC and this can be done by providing grid reactive power reference (𝑄𝑔 _𝑟𝑒𝑓) during
the dip.

73
6.7.2. Asymmetrical Voltage Dips

This section exhibits the simulation analysis of asymmetrical voltage dips. Hence, the
performance of the proposed control system is analyzed for 10 % asymmetrical voltage
dip. Figure 6.23 represents the stator flux and the crowbar current performances.

Figure 6.23. Stator flux and crowbar current under asymmetrical voltage dips.

Figure 6.24. Performances of RSC magnitudes under asymmetrical voltage dips.


74
Figure 6.25. Performances of GSC magnitudes under asymmetrical voltage dips.

From the simulation above at the RSC, it can be observed that there is variation in the
rotor speed and this is mainly because there is non-total control of voltage at the grid.
Furthermore, the stator flux has moved down during the voltage dip and therefore there
is no total control on the torque. Besides, there is oscillations at the torque and those
oscillations are mainly because of the imbalance situation. Additionally, the zero
sequence of the rotor currents that has been performed is maintaining (zero) properly.
At the time instant 4.2 seconds, the grid voltage that recover again is a perturbation for
the DFIG magnitudes but the system is still able to operate properly. However, the
asymmetry has not been performed at the GSC for the sake of simplicity, but the GSC
has been capable to cope with the DC bus voltage perturbation and the system is
controlling the quadrature grid current (𝑖𝑞𝑔) component to zero during the voltage dip.

75
PART 7

CONCLUSION

In this thesis, a detailed simulation result of DFIG based wind turbine connected to the
power grid is extensively investigated. The main finding has been drawn from the
specifications of a 2 MW, 690 V DFIG system. At first, all the key elements have been
modeled including the wind turbine such as: wind speed model, aerodynamic model,
drive train model, electrical model and DFIG. Furthermore, a comprehensive dynamic
model of DFIG has been developed in a rotating 𝑑𝑞 reference frame by using Clarke
and Park transformation technique, by first deriving and modeling the 𝑎𝑏𝑐 natural
reference frame.

The most representative magnitudes, performance, and control of a wind turbine based
DFIG has been illustrated. It further contains the findings of the steady-state analysis
and simulation results that confirm the performance and control of the DFIG system.
Different wind speed with the system has been evaluated and the finding shows that
the system is capable of controlling the torque and power at variable speed. This means
that the system is also capable of extracting the maximum power available by using
MPPT. An indirect speed control strategy has been applied to drive the torque to track
the maximum power curve.

Moreover, vector control approach of electrical machines has been implemented in the
study. Vector control provides dynamic control of DFIG and it employs RSC and GSC.
The objective of the RSC is to control the torque and flux of the machine and it’s also
responsible for the MPPT. Besides, the RSC assures decouple control of active and
reactive stator power. The main purpose of GSC is to maintain the dc bus voltage
constant irrespective of the region of operation and it also operates at a unit power
factor. In addition, the comparisons of the performances of a wind turbine based DFIG
with PI control and proposed method for different wind speed has been investigated.

76
Finally, the study provides extensive investigations of the transient behavior of DFIG
based wind turbine under symmetrical and asymmetrical voltage dips. Furthermore,
the asymmetrical voltage shows the existence of positive and negative sequence. Thus,
dual vector control strategy has been implemented. It has been determined that this
strategy is suitable for controlling the reactive power properly during the dip. Crowbar
protection has been applied in order to protect the converter during the voltage dip and
the simulation result of the most representative magnitudes of a wind turbine based
DFIG has been analyzed.

7.1. CONTRIBUTIONS OF THESIS

The main contributions are presented as follows:

▪ Implementation and simulations of the DFIG system in the RSC using SFOVC.
▪ The study of steady-state analysis of the different magnitudes of DFIG system.
▪ Comparison between the typical PI controller and the proposed method in
terms of settling time for different wind speeds.
▪ Implementation of grid-connected DFIG based wind turbine in the GSC using
GVOVC.
▪ The study for dynamic analysis of current loops behavior in a wind turbine
based DFIG.
▪ Investigation and performance analysis of DFIG based wind turbine connected
to the grid under symmetrical and asymmetrical voltage dips.

7.2. FUTURE WORK

Even though various developments have been done in DFIG wind turbine, some future
work that require to develop are listed below.

▪ Since vector control strategy which, for instance, require current control loops
has been performed in this thesis. Thus, direct torque control and direct power
control strategy should be examined because, these strategies do not require
current control loops.

77
▪ Pitch controller should be designed in order to achieve good performances for
much higher wind speed.
▪ Advanced control strategies of DFIG based wind turbine should be considered.
Robust control strategy, hybrid control and optimal control strategies should
be investigated for better transient and robustness.
▪ The DFIG system during asymmetrical voltage dips that has been developed
in this thesis still need to improve by just integrating dc choppers instead of
crowbar to further modify the oscillations in the dc bus voltage.

78
REFERENCES

1. Thakur, A., Panigrahi, S., and Behera, R. R., “A review on wind energy conversion
system and enabling technology”, IEEE International Conference on Electrical
Power and Energy Systems, 527-532 (2016).

2. Annoukoubi, M., Essadki, A., Laghridat, H., and Nasser, T., “Comparative study
between the performances of a three-level and two-level converter for a Wind
Energy Conversion System”, IEEE International Conference on Wireless
Technologies, Embedded and Intelligent Systems, 1-6 (2019).

3. Morsi, A., Abbas, H. S., and Mohamed, A. M., “Wind turbine control based on a
modified model predictive control scheme for linear parameter-varying
systems”, IET Control Theory and Applications, 11(17): 3056-3068 (2017).

4. El-Sharkawi, M. A., “Wind energy: an introduction”, Taylor and Francis book on


History of the Wind Energy Development, 1-5 (2015).

5. Bianchi, F. D., De Battista, H., and Mantz, R. J., “Wind turbine control systems:
principles, modelling and gain scheduling design”, Springer Book Advances in
Industrial Control, (19): (2007).

6. Ackermann, T., “Wind power in power systems” 200(5): (2005).

7. Sathyajith, M., and Philip, G. S. (Eds.)., “Advances in wind energy conversion


technology”, Springer Science and Business Media, 3-8 (2011).

8. Pardalos, P. M., Rebennack, S., Pereira, M. V., Iliadis, N. A., and Pappu, V. (Eds.).,
“Handbook of wind power systems”, Springer Berlin Heidelberg, 21-30 (2013).

9. IEA., “International Energy Agency Ministerial Meeting on Wind Power”,


https://www.iea.org/reports/wind-power (2022).

10. Tanvir, A. A., Merabat, A., and Beguenane, R., “Real-time control of active and
reactive powerfor doubly fed induction generator (DFIG)-based wind energy
convertion system”, Energies, 8(9): 10389-10408 (2015).

11. Jain, B., Jain, S., and Nema, R. K., “Control strategies of grid interfaced wind
energy conversion system”, An overview. Renewable and Sustainable Energy
Reviews, (47): 983-996 (2015).

79
12. Jena, D., and Rajendran, S., “A review of estimation of effective wind speed based
control of wind turbines”, Renewable and Sustainable Energy Reviews, (43):
1046-1062 (2015).

13. Pati, S., and Samantray, S., “Decoupled control of active and reactive power in a
DFIG based wind energy conversion system with conventional PI controllers”,
IEEE International Conference on Circuits, Power and Computing
Technologies, 898-903 (2014).

14. Ouyang, J., Tang, T., Yao, J., and Li, M., “Active voltage control for DFIG-based
wind farm integrated power system by coordinating active and reactive powers
under wind speed variations”, IEEE Transactions on Energy Conversion, 34(3):
1504-1511 (2019).

15. Kerrouche, K., Mezouar, A., and Belgacem, K., “Decoupled control of doubly fed
induction generator by vector control for wind energy conversion system”, Energy
procedia, (42): 239-248 (2013).

16. Ghoudelbourk, S., Dib, D., and Omeiri, A., “Decoupled control of active and
reactive power of a wind turbine based on DFIG and matrix converter”, Energy
Systems, 7(3): 483-497 (2016).

17. Aydin, E., Polat, A., and Ergene, L. T., “Vector control of DFIG in wind power
applications”, IEEE International Conference on Renewable Energy Research
and Applications, Birmingham, UK, 11, 20 (2016).

18. Jabr, H. M., Lu, D., and Kar, N. C., “Design and implementation of neuro-fuzzy
vector control for wind-driven doubly-fed induction generator” IEEE
Transactions on Sustainable Energy, 2(4): 404-413 (2011).

19. Li, S., Haskew, T. A., Williams, K. A., and Swatloski, R. P, “Control of DFIG
wind turbine with direct-current vector control configuration”, IEEE transactions
on Sustainable Energy, 3(1): 1-11 (2011).

20. Wang, T., Ding, L., Yin, S., Jiang, J., Cheng, F., and Si, J., “A new control strategy
of DFIG-based wind farms for power system frequency regulation”, IEEE PES
Asia-Pacific Power and Energy Engineering Conference, 1-5 (2015).

21. Hughes, F. M., Anaya-Lara, O., Jenkins, N., and Strbac, G., “Control of DFIG-
based wind generation for power network support”, IEEE Transactions on Power
Systems, 20(4): 1958-1966 (2005).

22. Fdaili, M., Essadki, A., and Nasser, T., “Comparative analysis between robust
SMC & conventional PI controllers used in WECS based on DFIG”, International
Journal of Renewable Energy Research, 7(4): 2151-2161 (2017).

23. Azzaoui, M. E., and Mahmoudi, H., “Fuzzy-PI control of a doubly fed induction
generator-based wind power system”, International Journal of Automation and
Control, 11(1): 54-66 (2017).

80
24. Hamane, B., Benghanemm, M., Bouzid, A. M., Belabbes, A., Bouhamida, M., and
Draou, A., “Control for variable speed wind turbine driving a doubly fed induction
generator using Fuzzy-PI control”, Energy Procedia, (18): 476-485 (2012).

25. Zhu, R., Chen, Z., Tang, Y., Deng, F., and Wu, X., “Dual-loop control strategy for
DFIG-based wind turbines under grid voltage disturbances”, IEEE Transactions
on Power Electronics, 31(3): 2239-2253 (2015).

26. Tremblay, E., Atayde, S., and Chandra, A., “Comparative study of control
strategies for the doubly fed induction generator in wind energy conversion
systems: A DSP-based implementation approach”, IEEE Transactions on
sustainable energy, 2(3): 288-299 (2011).

27. Nian, H., and Yi, X., “Coordinated control strategy for doubly-fed induction
generator with DC connection topology”, IET Renewable power generation, 9(7):
747-756 (2015).

28. Wu, Y. K., and Yang, W. H., “Different control strategies on the rotor side
converter in DFIG-based wind turbines”, Energy Procedia, (100): 551-555
(2016).

29. Tourou, P., and Sourkounis, C., “Review of control strategies for DFIG-based
wind turbines under unsymmetrical grid faults”, IEEE Ninth international
conference on ecological vehicles and renewable energies, 1-9 (2014).

30. Justo, J. J., Mwasilu, F., and Jung, J. W., “Enhanced crowbarless FRT strategy for
DFIG based wind turbines under three-phase voltage dip”, Electric Power Systems
Research, (142): 215-226 (2017).

31. Musarrat, M. N., Fekih, A., and Islam, M. R., “An Improved Fault Ride Through
Scheme and Control Strategy for DFIG-Based Wind Energy Systems”, IEEE
Transactions on Applied Superconductivity, 31(8): 1-6 (2021).

32. Wessels, C., Gebhardt, F., and Fuchs, F. W., “Fault ride-through of a DFIG wind
turbine using a dynamic voltage restorer during symmetrical and asymmetrical
grid faults”, IEEE Transactions on Power Electronics, 26(3): 807-815 (2010).

33. Das, T. K., and Zhang, J., “Counteracting the effects of symmetrical and
asymmetrical voltage sags on DFIG-based wind power systems”, IEEE
International Conference on Advanced Mechatronic Systems, Melbourne,
Australia, 11, 30 (2016).

34. Döşoğlu, M. K., “Crowbar hardware design enhancement for fault ride through
capability in doubly fed induction generator-based wind turbines”, ISA
transactions, (104): 321-328 (2020).

35. Altimania, M. R., “Modeling of doubly-fed induction generators connected to


distribution system based on eMEGASim® real-time digital simulator” (2014).

81
36. Gasch, R., and Twele, J. (Eds.)., “Wind power plants: fundamentals, design,
construction and operation”, Springer Science and Business Media, 47-59 (2011).

37. Aissaoui, A. G., and Tahour, A. (Eds.)., “Wind turbines: design, control and
applications”, BoD–Books on Demand. 83-92 (2016).

38. Hermann, J. W., and Mathur, J., “Introduction to wind energy systems: basics,
technology and operation”, Springer, 50-51 (2009).

39. Bustos, G., Vargas, L. S., Milla, F., Sáez, D., Zareipour, H., and Nuñez, A.,
“Comparison of fixed speed wind turbines models: a case study”, 38th Annual
Conference on IEEE Industrial Electronics Society, 961-966 (2012).

40. Wu, B., Lang, Y., Zargari, N., and Kouro, S., “Power conversion and control of
wind energy systems”, John Wiley & Sons, 12-19 (2011).

41. Noussi, K., Abouloifa, A., Katir, H., and Lachkar, I., “Modeling and control of a
wind turbine based on a doubly fed induction generator”, IEEE 4th World
Conference on Complex Systems, 1-5 (2019).

42. Abad, G., Lopez, J., Rodriguez, M., Marroyo, L., and Iwanski, G., “Doubly fed
induction machine: modeling and control for wind energy generation”, John Wiley
& Sons, (2011).

43. Junyent-Ferré, A., Gomis-Bellmunt, O., Sumper, A., Sala, M., and Mata, M.,
“Modeling and control of the doubly fed induction generator wind
turbine”, Simulation Modelling Practice and Theory, 18(9): 1365-1381 (2010).

44. Jing, X., “Modeling and control of a doubly-fed induction generator for wind
turbine-generator systems” (2012).

45. Michas, M., “Control of turbine-based energy conversion systems” (2018).

46. Yang, B., Jiang, L., Wang, L., Yao, W., and Wu, Q. H., “Nonlinear maximum
power point tracking control and modal analysis of DFIG based wind
turbine”, International Journal of Electrical Power and Energy Systems, (74):
429-436 (2016).

47. Instruments, T., “Clarke & park transforms on the tms320c2xx”, Application
Report BPRA, (48): (1997).

48. Lamnadi, M., Trihi, M., Bossoufi, B., and Boulezhar, A., “Modeling and control
of a doubly-fed induction generator for wind turbine-generator systems”,
International Journal of Power Electronivs and Drive System, 7(3): 973-985
(2016).

49. Arifujjaman, M., Iqbal, M. T., & Quaicoe, J. E., “Vector control of a DFIG based
wind turbine”, IU-Journal of Electrical & Electronics Engineering, 9(2): 1057-
1066 (2009).

82
50. Meera, G. S., and Divya, N. A., “Rotor side converter control of DFIG based wind
energy conversion system”, International Journal of Engineering Research and
Technology, 4(8): 607-612 (2015).

51. Tuka, M. B., “DC Link Voltage and Power Flow Control of Doubly Fed Induction
Generator in Wind Power System”, IEEE PES/IAS PowerAfrica, 1-5 (2020).

52. Hallak, M., Hasni, M., and Menaa, M., “Modeling and control of a doubly fed
induction generator base wind turbine system”, IEEE International Conference
on Electrical Sciences and Technologies in Maghreb, 1-5 (2018).

53. Bhushan, R., and Chatterjee, K., “Mathematical modeling and control of DFIG‐
based wind energy system by using optimized linear quadratic regulator weight
matrices”, International Transactions on Electrical Energy Systems, 27(11): 1-
23 (2017).

54. Puchalapalli, S., and Singh, B., “A novel control scheme for wind turbine driven
DFIG interfaced to utility grid”, IEEE Transactions on Industry
Applications, 56(3): 2925-2937 (2020).

55. Abu-Rub, H., Malinowski, M., and Al-Haddad, K., “Power electronics for
renewable energy systems, transportation and industrial applications”, John Wiley
& Sons, 270-317 (2014).

56. Elbashir, O. E., Zezhong, W., and Qihui, L., “Modeling and analysis of DFIG in
wind energy conversion system”, International journal of Energy and
environment, (5): 239-250 (2014).

57. Nadour, M., Essadki, A., and Nasser, T., “Comparative analysis between PI &
backstepping control strategies of DFIG driven by wind turbine”, International
Journal of Renewable Energy Research, 7(3): 1307-1316 (2017).

58. Al Zabin, O., & Ismael, A., “Rotor current control design for DFIG-based wind
turbine using PI, FLC and fuzzy PI controllers”, IEEE International Conference
on Electrical and Computing Technologies and Applications, UAE, 9 (2019).

59. Yonis, S. A., and Yusupov, Z., “Dynamic Analysis of Current Loops Behavior in
a Wind Turbine Based Doubly-fed Induction Generator”, Avrupa Bilim ve
Teknoloji Dergisi, Konya, Turkey, 3, 11 (2022).

60. Kumhar, A. R., “Vector Control Strategy to Control Active and Reactive Power of
Doubly Fed Induction Generator Based Wind Energy Conversion System”, IEEE
2nd International Conference on Trends in Electronics and Informatics, 1-9
(2018).

61. Haseeb, I., Basit, A., Khan, R., and Asif, M., “Designing variable speed small
hydro turbine with doubly fed induction generator (DFIG)”, International
Journal of Renewable Energy Sources, (4): 1-10 (2019).

83
APPENDIX A

DFIG AND WIND TURBINE PARAMETERS

84
Table Appendix A.1. Initial parameters of DFIG and wind turbine.
PARAMETERS VALUES (UNITS)
STATOR POWER 2 MW
ROTOR SPEED 1800 rev/min
NOMINAL VOLTAGE 690 V
NOMINAL CURRENT 1.76 kA
ELECTROMAGNETIC TORQUE 12.7 kN.m
STATOR FREQUENCY 50 HZ
ROTOR VOLTAGE 2.07 kV
ROTOR RESISTANCE 0.0029 Ω
ROTOR INDUCTANCE 0.0026 H
STATOR RESISTANCE 0.0026 Ω
STATOR INDUCTANCE 0.0026 H
STATOR/ROTOR TURNS RATIO 0.34
RATED WIND SPEED 11.8 m/s
GEARBOX RATIO 100
RADIUS OF THE BLADE 42 m
AIR DENSITY 1.225 kg/m2
POLE PAIR 2

85
APPENDIX B

MATLAB/SIMULINK MODELS AND SUB-MODELS

86
Figure Appendix B.1. RSC control block.

Figure Appendix B.2. GSC control block.

87
Figure Appendix B.3. Cancellations of coupling terms.

Figure Appendix B.4. d/q to D/Q transformation.

Figure Appendix B.5. D/Q to abc transformation.

88
Figure Appendix B.6. abc to D/Q transformation.

Figure Appendix B.7. D/Q to d/q transformation.

Figure Appendix B.8. PLL angle calculator.

89
Figure Appendix B.9. Third harmonic injection.

90
APPENDIX C

INITIALIZATION PROGRAM

91
% Parameters of the DFIG

P= 2e6;
V= 0.69e2;
I = 1.76e3;
T = 12.7e3;
F= 50;
P= 2; % number of poles
vr = 2.07e3;
smax = 0.34;
n= 1.5e3;
vr-stator = (Vr*Smax) *u;
U= 0.34;
Rr = 2.9e-3;
Ls =0.087e-3;
Rs = 2.6e-3;
Lm = 2.5e-3;
Vbus = 1150;
σ = 1- Lm^2/(Ls*Lr); % sigma
Ls = Lm + Ls; % stator inductance
Lr = Lm + Ls; % rotor inductance
Fs = V*Sqrt(2/3)/(2*pi*F);%Stator flux (approx.)

Ts = 1/fsw/50; % sample time


J = 63.5; % inertia
fsw = 4000; % switching frequency
D = 0.001; % damping ratio

% The PI regulator

Tau_i = (σ *Lr)/Rr;
Tau_n = 0.05;
Wni = 100*(1/tau_i);
Wnn = 1/tau_n;

% The 𝑑𝑞 current regulators

Kp_id = (2*Wni* σ *Lr)-Rr;


Kp_iq = Kp_id;
Ki_id = (Wni^2)*Lr* σ;
Ki_iq = Ki_id;

92
% The regulators tuned as controller

Kp_n = (2*Wnn *J)/P;


ki_n = ((wnn^2) *J)/p;

% Parameter of wind turbines

Vw=11.8; %Wind speed


N = 100;
Ratio = 42;
ro = 1.225; %Air density

93
APPENDIX D

PUBLISHED PAPERS

94
TR Dizin-indexed Journal

1. Yonis, S. A., and Yusupov, Z., “Dynamic Analysis of Current Loops Behavior
in a Wind Turbine Based Doubly-fed Induction Generator”, Avrupa Bilim ve
Teknoloji Dergisi, (34): 415-420 (2022).

CONFERENCES

1. Yonis. S. A., and Yusupov, Z., “Voltage Dips Analysis in a Doubly-fed


Induction Generator Based Wind Turbine Using Active Crowbar Protection”,
International Graduate Research Symposium, Istanbul, Turkey, 3, 6 (2022).

95
RESUME

Samatar ABDI YONIS was born in Djibouti city in 1994 and he graduated first and
elementry education in Dire-dawa, Ethiopia. He completed high school education in
Al-falah secondary and preparatory islamic school, after that, he started undergraduate
program in Haramaya University Department of Electrical and Computer Engineering
in 2013. Then in 2020, he started a master degree in Karabuk University.

96

You might also like