ANALYSIS OF THREE-PHASE RECTIFIERS WITHAC-SIDE SWITCHES AND INTERLEAVED
THREE-PHASE VOLTAGE-SOURCE CONVERTERS
By
Stephanie Katherine Teixeira Miller
A Thesis Submitted to the Graduate
Faculty of Rensselaer Polytechnic Institute
in Partial Fulfillment of the
Requirements for the Degree of
DOCTOR OF PHILOSOPHY
Major Subject: Electrical Engineering
Approved by theExamining Committee:
Jian Sun, Thesis Adviser
Kenneth A. Connor, Member
Leila Parsa, Member
Sheppard J. Salon, Member
James Kokernak, Member
Rensselaer Polytechnic InstituteTroy, New York
November 2008(For Graduation December 2008)
UMI Number: 3357231
Copyright 2008 by Miller, Stephanie Katherine Teixeira
All rights reserved
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
______________________________________________________________
UMI Microform 3357231 Copyright 2009 by ProQuest LLC
All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.
_______________________________________________________________
ProQuest LLC 789 East Eisenhower Parkway
P.O. Box 1346 Ann Arbor, MI 48106-1346
ANALYSIS OF THREE-PHASE RECTIFIERS WITHAC-SIDE SWITCHES AND INTERLEAVED
THREE-PHASE VOLTAGE-SOURCE CONVERTERS
By
Stephanie Katherine Teixeira Miller
An Abstract of a Thesis Submitted to the Graduate
Faculty of Rensselaer Polytechnic Institute
in Partial Fulfillment of the
Requirements for the Degree of
DOCTOR OF PHILOSOPHY
Major Subject: Electrical Engineering
The original of the complete thesis is on filein the Rensselaer Polytechnic Institute Library
Examining Committee:
Jian Sun, Thesis Adviser
Kenneth A. Connor, Member
Leila Parsa, Member
Sheppard J. Salon, Member
James Kokernak, Member
Rensselaer Polytechnic InstituteTroy, New York
November 2008(For Graduation December 2008)
c Copyright 2008by
Stephanie Katherine Teixeira Miller
All Rights Reserved
ii
CONTENTS
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
ACKNOWLEDGMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Wind Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Dissertation Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Power Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Wind Turbine Topologies . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Induction Generator . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 Grid Interconnection Requirements . . . . . . . . . . . . . . . . . . . 9
1.6.1 Voltage Fault Ride-Through . . . . . . . . . . . . . . . . . . . 10
1.6.2 Power Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6.3 Power Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.7 Survey of Power Electronics Interface Topologies for Induction Gen-erators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.7.1 Diode Rectifier . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.7.2 Six Switch Voltage-Source Converter . . . . . . . . . . . . . . 17
1.7.2.1 Back-to-Back Voltage-Source Converters . . . . . . . 18
1.7.2.2 Interleaved Voltage-Source Converters . . . . . . . . 19
1.7.3 Multilevel Converters . . . . . . . . . . . . . . . . . . . . . . . 20
1.7.4 Matrix Converter . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.8 Dissertation Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2. Three-Phase Voltage-Source Converter . . . . . . . . . . . . . . . . . . . . 28
2.1 Topology and Basic Operation . . . . . . . . . . . . . . . . . . . . . . 28
2.2 Modulation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.2.1 Naturally Sampled Carrier-Based Modulation . . . . . . . . . 30
2.2.1.1 Sinusoidal Pulse-Width Modulation . . . . . . . . . . 30
2.2.1.2 Sinusoidal Pulse-Width Modulation with 3rd Har-monic Injection . . . . . . . . . . . . . . . . . . . . . 32
2.2.2 Space-Vector Modulation . . . . . . . . . . . . . . . . . . . . . 35
iii
2.2.2.1 The Plane . . . . . . . . . . . . . . . . . . . . . . 36
2.2.2.2 Voltage Space Vectors . . . . . . . . . . . . . . . . . 37
2.2.2.3 Sector Definitions . . . . . . . . . . . . . . . . . . . . 41
2.2.2.4 Space-Vector Modulation . . . . . . . . . . . . . . . 41
2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3. Three-Phase PWM Rectifiers with Ac-Side Bidirectional Switches . . . . . 47
3.1 Literature Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Converter Operation and Space Vector Description . . . . . . . . . . 51
3.3 Pulse-Width Modulation . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.4 Open-Loop Operation . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.5 Rectifier T supplied by a Squirrel-Cage Induction Generator . . . . 64
3.5.1 Induction Generator Power Requirements under Variable WindSpeed Conditions . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.5.2 Rectifier Power Handling Capability . . . . . . . . . . . . . . . 70
3.5.3 Induction Generator/Rectifier System . . . . . . . . . . . . . . 71
3.5.3.1 Variable Terminal Voltage and Frequency . . . . . . 72
3.5.3.2 Constant Terminal Voltage and Frequency . . . . . . 73
3.5.4 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . 75
3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4. Comparison of Rectifier Topologies T, TY, and TS . . . . . . . . . . . . . 80
4.1 Literature Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.2 Rectifier T Switch Loss Analysis . . . . . . . . . . . . . . . . . . . . 82
4.2.1 Conduction Losses . . . . . . . . . . . . . . . . . . . . . . . . 84
4.2.2 Switching Losses . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.3 Diode Loss Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.3.1 Rectifiers T and TY . . . . . . . . . . . . . . . . . . . . . . . 96
4.3.2 Rectifier TS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.4 Loss Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.4.1 Switch Loss Comparison . . . . . . . . . . . . . . . . . . . . . 99
4.4.2 Diode Loss Comparison . . . . . . . . . . . . . . . . . . . . . 102
4.4.3 Validity of the Lumped Stress Equations . . . . . . . . . . . . 103
4.5 Common-Mode Voltage Comparison . . . . . . . . . . . . . . . . . . . 104
4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
iv
5. Harmonic Cancellation Effects in Interleaved Three-Phase Voltage-SourceConverters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.1 Literature Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.2 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.3 Common-Mode Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.4 AC Phase Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.5 Inter-Module Circulating Current . . . . . . . . . . . . . . . . . . . . 122
5.6 Inductor Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.7 DC-Link Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
5.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
6. Low Frequency Circulating Current Characteristics of Current ControlTechniques for Parallel Voltage-Source Converter Modules . . . . . . . . . 138
6.1 Literature Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
6.2 Circulating Current Decomposition and Definitions . . . . . . . . . . 142
6.3 Linear Current Control Methods . . . . . . . . . . . . . . . . . . . . . 143
6.3.1 Abc Current Control . . . . . . . . . . . . . . . . . . . . . . . 143
6.3.2 Dq Current Control . . . . . . . . . . . . . . . . . . . . . . . . 147
6.4 Nonlinear Average Current Control . . . . . . . . . . . . . . . . . . . 148
6.4.1 Control Strategy Description for a Single VSC Module . . . . 149
6.4.2 Interleaved Voltage-Source Converters using the Nonlinear Av-erage Current Control Technique . . . . . . . . . . . . . . . . 152
6.4.2.1 Independent Control Implementation . . . . . . . . . 156
6.4.2.2 Master/Slave Control Implementation . . . . . . . . 161
6.4.2.3 Experimental Results . . . . . . . . . . . . . . . . . 173
6.4.2.4 Modular Approach with Two Current Sensors perModule . . . . . . . . . . . . . . . . . . . . . . . . . 178
6.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
APPENDICES
A. Wind Energy State-of-the-Art . . . . . . . . . . . . . . . . . . . . . . . . . 203
A.1 General Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
A.1.1 Parts of a Wind Turbine . . . . . . . . . . . . . . . . . . . . . 203
A.1.2 Upwind vs. Downwind . . . . . . . . . . . . . . . . . . . . . . 205
v
A.2 Wind Turbine Control . . . . . . . . . . . . . . . . . . . . . . . . . . 205
A.2.1 Passive Stall Control . . . . . . . . . . . . . . . . . . . . . . . 205
A.2.2 Pitch Control and Active Stall Control . . . . . . . . . . . . . 206
A.2.3 Blade Tip Control . . . . . . . . . . . . . . . . . . . . . . . . 208
A.2.4 Yaw Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
A.3 Wind Turbine Technologies . . . . . . . . . . . . . . . . . . . . . . . 209
A.3.1 Tip-Speed Ratio . . . . . . . . . . . . . . . . . . . . . . . . . 211
A.3.2 Fixed-Speed Turbines . . . . . . . . . . . . . . . . . . . . . . . 213
A.3.3 Variable-Speed Turbines . . . . . . . . . . . . . . . . . . . . . 214
A.4 Transmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
A.4.1 Gearbox Transmission . . . . . . . . . . . . . . . . . . . . . . 215
A.4.2 Direct Drive Transmission . . . . . . . . . . . . . . . . . . . . 216
A.5 Generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
A.5.1 Synchronous Generator . . . . . . . . . . . . . . . . . . . . . . 218
A.5.2 Doubly-Fed Induction Generator . . . . . . . . . . . . . . . . 219
B. Open-Loop Simulations of Rectifiers T and TY . . . . . . . . . . . . . . . 221
C. Closed-Loop Operation of Rectifier T . . . . . . . . . . . . . . . . . . . . 226
D. Derivation of the Conduction and Switching Losses of Rectifiers TY and TS 237
D.1 Rectifier TY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
D.1.1 Conduction Losses . . . . . . . . . . . . . . . . . . . . . . . . 237
D.1.2 Switching Losses . . . . . . . . . . . . . . . . . . . . . . . . . 239
D.2 Rectifier TS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
D.2.1 Conduction Losses . . . . . . . . . . . . . . . . . . . . . . . . 240
D.2.2 Switching Losses . . . . . . . . . . . . . . . . . . . . . . . . . 243
E. Mathematica Program for determining the Harmonic Components of theDc-Link Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
vi
LIST OF TABLES
1.1 Net generation of electricity by energy source in thousand megawatthours[1]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Wind turbine topologies. . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3 Maximum harmonic current distortion in percent of current (I)a [2]. . . 12
1.4 Maximum ratings of power semiconductor devices [3]. . . . . . . . . . . 20
2.1 Voltage space vectors and resulting converter voltages. . . . . . . . . . . 39
2.2 Voltage space vector lengths and angular positions. . . . . . . . . . . . 40
2.3 Time allocated to the voltage space vectors per switching period. . . . . 44
3.1 Generation of the voltage space vectors. . . . . . . . . . . . . . . . . . . 54
3.2 Switch currents of T operating in Sector I. . . . . . . . . . . . . . . . 55
3.3 Wind speed operating range of the generator/rectifier system for differ-ent slip values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.1 Zero crossing angles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.2 Voltage space vector intervals in the time domain. . . . . . . . . . . . . 88
4.3 Comparison between theoretical and simulation results for the rms cur-rents of the bidirectional switches of T. . . . . . . . . . . . . . . . . . 93
4.4 Comparison between theoretical and simulation results for the averagecurrents of one unidirectional switch of T. . . . . . . . . . . . . . . . . 94
4.5 Analytical and simulated switch and diode current of T. . . . . . . . . 104
5.1 Common-mode voltage harmonic amplitudes. . . . . . . . . . . . . . . . 117
5.2 Ac phase current harmonic amplitudes. . . . . . . . . . . . . . . . . . . 120
5.3 Ac phase current harmonic amplitudes. . . . . . . . . . . . . . . . . . . 121
5.4 Inter-module circulating current harmonic amplitudes. . . . . . . . . . . 127
5.5 Dc-link current harmonic amplitudes. . . . . . . . . . . . . . . . . . . . 136
6.1 Definitions of vp, vn, and vm per sector [4]. . . . . . . . . . . . . . . . . 149
6.2 Definitions of Sp, Dp, Sn, Dn, Sm, and associated duty cycles [4]. . . . . 150
6.3 Measured and desired dc-link current for Sector I. . . . . . . . . . . . . 155
vii
6.4 Time interval, vector combination, and the sensed dc-link current andphase being modulated of Module 1 in Sector I. . . . . . . . . . . . . . 158
6.5 Time interval, vector combination, and the sensed dc-link current andphase being modulated of Module 2 in Sector I. . . . . . . . . . . . . . 159
6.6 Time interval, vector combination, and the sensed dc-link current andphase being modulated of Modules 1 and 2 in Sector II. . . . . . . . . . 161
6.7 Free-wheeling and controllable circulating currents per sector. . . . . . . 162
6.8 Currents of the positive and negative dc-link rails of Module 1 for thevector combinations of Sector I. . . . . . . . . . . . . . . . . . . . . . . 181
6.9 Currents of the positive and negative dc-link rails of Module 1 for thevector combinations of Sector II. . . . . . . . . . . . . . . . . . . . . . . 182
6.10 Proposed current feedback signal for Module 1. . . . . . . . . . . . . . . 182
C.1 Duty cycles per sector of switches S1, S2, and S3 of TS. . . . . . . . . . 233
D.1 Comparison between theoretical and simulation results for the rms andaverage currents of the switches of TS. . . . . . . . . . . . . . . . . . . . 243
viii
LIST OF FIGURES
1.1 Global wind power growth from 1994 to 2007 [5]. . . . . . . . . . . . . . 3
1.2 Top 10 countries in installed wind generation capacity [6]. . . . . . . . . 4
1.3 Evolution of wind turbine size [7]. . . . . . . . . . . . . . . . . . . . . . 5
1.4 Wind energy system with a power electronics interface. . . . . . . . . . 6
1.5 Wind turbine block diagram. . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6 Low voltage ride-through standard [8]. . . . . . . . . . . . . . . . . . . . 11
1.7 Wound rotor induction generator with a variable rotor resistance. . . . . 14
1.8 Three-phase diode rectifier with (a) an LC output filter, (b) an inputinductive filter and capacitive output filter and (c) a capacitive outputfilter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.9 Three-phase diode rectifier with a capacitor bank. . . . . . . . . . . . . 15
1.10 High voltage dc wind farm architecture. . . . . . . . . . . . . . . . . . . 16
1.11 Six switch voltage-source converter. . . . . . . . . . . . . . . . . . . . . 17
1.12 Two-level output of the voltage-source converter. . . . . . . . . . . . . . 18
1.13 Back-to-back voltage source converters. . . . . . . . . . . . . . . . . . . 18
1.14 Interleaved voltage-source converters. . . . . . . . . . . . . . . . . . . . 19
1.15 Output voltage of (a) a three-level and (b) a five-level multilevel converter. 21
1.16 Single leg of a five-level (a) diode-clamped multilevel converter, (b)flying-capacitor multilevel converter, and (c) cascaded H-bridge mul-tilevel converter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.17 Matrix converter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.1 Three-phase voltage-source converter and realization of the bidirectionalswitch. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.2 Naturally sampled carrier-based modulation. . . . . . . . . . . . . . . . 31
2.3 Overmodulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.4 Reference voltage and its fundamental and third harmonic components. 33
2.5 Reference voltage and its fundamental and third harmonic componentsfor optimum third harmonic injection. . . . . . . . . . . . . . . . . . . . 35
ix
2.6 Reference voltage vector rotating in the plane. . . . . . . . . . . . . 37
2.7 Space-vector modulation: Eight possible switching states. . . . . . . . . 38
2.8 Voltage space vectors in the plane. . . . . . . . . . . . . . . . . . . . 40
2.9 Sector definitions in the a) time domain and b) space vector domain. . . 41
2.10 Decomposition of the reference vector for deriving the time allocated tothe active vectors (Sector I). . . . . . . . . . . . . . . . . . . . . . . . . 42
2.11 Switching sequence for switches S1, S2, and S3 in Sector I. . . . . . . . . 45
3.1 Three-phase PWM rectifiers a) T, b) TY, and c) TB. . . . . . . . . . . 48
3.2 Three-phase uncontrolled rectifier and the input current of phase a whenemploying only a capacitive filter and when employing capacitive andinductive filters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.3 Operation boundaries of rectifier topologies T, TY, and TB. . . . . . . 50
3.4 a) Three-phase rectifier of [9] and b) the Vienna rectifier. . . . . . . . . 51
3.5 Voltage space vector v1 applied to a) the six-switch PWM VSC and b)T. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.6 Voltage space vector v4 applied to a) the six-switch PWM VSC and b)T. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.7 Operation boundaries of rectifier topologies T, TY, and TB. . . . . . . 55
3.8 Contribution of the zero vector intervals to the conduction losses of Tresulting from the use of different zero vectors in Sector I. . . . . . . . . 56
3.9 Switching loss instants for topology T when using zero vector a) {1 11} and b) {1 1 0}. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.10 Switching loss instants for topology T when using zero vector a) {1 01} and b) {0 1 1}. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.11 a) Sector I drive signals for TS and T and b) the combinatorial logicfor generating the drive signals for Sab from the drive signals of TS whenusing zero vector {1 1 1}. . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.12 SABER simulation schematic of the power stage and PWM of T. . . . 62
3.13 Simulation results of the phase a current and source voltage of TY. . . . 63
3.14 Currents through switches a) Sab (using zero vector {1 1 1}), b) Sab(using a different zero vector per sector), and c) S1 of TY. . . . . . . . . 63
3.15 Per-phase model of the induction generator. . . . . . . . . . . . . . . . . 65
x
3.16 Wind turbine and induction generator model. . . . . . . . . . . . . . . . 66
3.17 Active and reactive power trajectory of the example induction generatorfor wind speeds varying from 1 m/s to 15 m/s. . . . . . . . . . . . . . . 69
3.18 Voltage-to-frequency ratio for wind speeds varying from 1 m/s to 15 m/s. 70
3.19 Active and reactive power limits of T for a -1% slip and wind speedsequal to 5, 10, and 15 m/s. . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.20 Variation of the phase of stator current as a function of the wind speedand for a constant -1% slip. . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.21 Induction generator/rectifier operation between the cut-in speed (1 m/s)and rated speed (15 m/s). . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.22 P-Q curve of the induction generator and the operation boundaries ofrectifier T for different values of slip. . . . . . . . . . . . . . . . . . . . 74
3.23 SABER simulation schematic of the generator/rectifier system. . . . . . 76
3.24 Stator phase a current during start-up and for a step transition in theslip from -2% to -5%. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.25 Detail of the stator current during the start-up transient. . . . . . . . . 78
4.1 a) Equivalent circuit for phase a and b) the corresponding phasor diagram. 83
4.2 Equivalent circuit when zero vector {1 1 1} is applied. . . . . . . . . . . 854.3 Current envelopes for switch current iSab(1t) over a line cycle. . . . . . 86
4.4 Current iSab(t) over a line cycle and over a switching cycle in Sector Iand the voltage across the switch over a switching cycle in Sector I forthe two given modulation strategies. . . . . . . . . . . . . . . . . . . . . 87
4.5 Variation of i2Sab,rms over a line cycle for operation at rated power and a30o phase lag. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.6 Variation of i2Sab,rms over a line cycle for operation at rated power and a0o phase lag. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.7 Variation of i2Sab,rms over a line cycle for operation at rated power and a24o phase lead. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.8 Current through Sab and current envelopes for l = 1 and a) a 30o phase
lag, b) unity power factor operation, and c) a 24o phase lead. . . . . . . 92
4.9 Current through Sab and current envelopes for l = 0 and a) a 30o phase
lag, b) unity power factor operation, and c) a 24o phase lead. . . . . . . 93
4.10 SABER simulation schematic for the discrete implementation of thebidirectional switches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
xi
4.11 Simplified switching instants. . . . . . . . . . . . . . . . . . . . . . . . . 95
4.12 Variation of the combined switching losses for rated power and a 24o
current phase lead. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.13 Current envelopes of iD1(t). . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.14 Variation of the square of the switch rms current: a) Operation underrated output power and b) Operation under 50% rated output powerand stator terminal voltage magnitude. A 24o leading phase angle isassumed for the current. . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.15 Practical implementation of a bidirectional switch using a) IGBTs withseries diodes, b) a series connection of IGBTs (common-emitter config-uration), and c) an anti-parallel connection of reverse-blocking IGBTs. . 101
4.16 Comparison of common-mode voltage generated by rectifier T (upperdiagram) and the six-switch VSC (lower diagram). . . . . . . . . . . . . 105
4.17 a) One leg of the diode bridge during the zero vector interval and b) mag-nified view of the reverse bias characteristic of two mismatched diodesof a leg of the diode bridge. . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.18 Laboratory test set-up. . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.1 Phase-shifted carrier signals for N interleaved modules. . . . . . . . . . 109
5.2 A three-phase voltage-source converter using N parallel modules. . . . . 111
5.3 SABER simulation schematic for two interleaved VSC modules. . . . . . 116
5.4 Spectra of the common-mode voltage of N interleaved VSCs. . . . . . . 117
5.5 a) Equivalent circuit of phase a for N interleaved modules and b) thesimplified equivalent circuit. . . . . . . . . . . . . . . . . . . . . . . . . 118
5.6 Spectra of the ac phase current of N interleaved VSCs. . . . . . . . . . 120
5.7 Definition of the inter-module circulating and module output currentsfor a) two, b) three, and c) N interleaved VSC modules. . . . . . . . . . 122
5.8 a) Voltage difference between the converter voltages of phase a, b) inter-module circulating current, c) inductor currents ia1 and ia2 and phase acurrent for a quarter of the line cycle and N = 2. . . . . . . . . . . . . . 128
5.9 Spectra of the a) inductor current, b) inter-module circulating current,and c) phase a current for the case of N = 2. . . . . . . . . . . . . . . . 130
5.10 Spectra of the dc-link current of N interleaved VSCs. . . . . . . . . . . 136
6.1 Isolated parallel voltage-source converter configurations: a) with lowfrequency transformers and b) separate dc power supplies. . . . . . . . . 139
xii
6.2 Simplified schematic of the abc current control. . . . . . . . . . . . . . . 143
6.3 a) Average model of the parallel VSC modules and b) block diagram ofthe current loop for phase a. . . . . . . . . . . . . . . . . . . . . . . . . 145
6.4 Phase a inductor current of Module 1 and reference current. . . . . . . . 147
6.5 Inductor current space vector trajectory in the 0 space. . . . . . . . . 148
6.6 a) Sector definitions and b) equivalent circuit of a single voltage-sourceconverter module. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.7 Dc-link current of a stand-alone voltage-source converter. . . . . . . . . 151
6.8 Implementation of the nonlinear average control strategy. . . . . . . . . 152
6.9 Phase a current of a voltage-source converter operating with the non-linear average current control. . . . . . . . . . . . . . . . . . . . . . . . 153
6.10 Two voltage-source converter modules connected in parallel. . . . . . . 154
6.11 a) Sector definitions and b) equivalent circuit for Sector I. . . . . . . . . 155
6.12 Modular implementation of the single module control to two interleavedVSC modules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.13 Simulation results of the inductor currents of phase a. . . . . . . . . . . 158
6.14 Vector combination in Sector I. . . . . . . . . . . . . . . . . . . . . . . . 159
6.15 Circulating currents ib and ic, control signals and integration outputof Modules 1 and 2 in Sector II. Module 1 does not commutate. . . . . 161
6.16 Circulating currents ib and ic, control signals and integration outputof Modules 1 and 2 in Sector II. Module 1 resumes normal operation. . 162
6.17 Master/slave implementation of two interleaved voltage-source converters.163
6.18 Inductor currents of phase a when operating in the master/slave config-uration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
6.19 Difference between the inductor currents of Fig. 6.18. . . . . . . . . . . 164
6.20 Spectrum of the difference between the inductor currents of phase a. . . 165
6.21 Spectra of the a) phase a module output current, b) circulating current,and c) master converter current. . . . . . . . . . . . . . . . . . . . . . . 166
6.22 Source voltages vsa and vsc and the complimentary duty cycles of themaster and slave modules. . . . . . . . . . . . . . . . . . . . . . . . . . 167
6.23 a) Equivalent circuit of the master/slave configuration and b) equivalentcircuit for the circulating current analysis. . . . . . . . . . . . . . . . . . 168
xiii
6.24 Average model of the circulating current loop. . . . . . . . . . . . . . . 168
6.25 a) Dc-link, b) ia, ib, and ib current waveforms in Sector I. . . . . . . 170
6.26 Derivative of the inherent circulating current of phase a in Sector I fordifferent values of t. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
6.27 Current flowing through the ac source of phase a. . . . . . . . . . . . . 172
6.28 Experimental set-up of two interleaved voltage-source converters. . . . . 174
6.29 Phase a source voltage and current of the single module operating withthe nonlinear average current control. . . . . . . . . . . . . . . . . . . . 174
6.30 Phase a inductor currents of the master and slave converters. . . . . . . 175
6.31 Phase b inductor currents of the master and slave converters. . . . . . . 175
6.32 Phase c inductor currents of the master and slave converters. . . . . . . 176
6.33 Currents flowing through the ac sources. . . . . . . . . . . . . . . . . . 176
6.34 Difference between the inductor currents of phase a. . . . . . . . . . . . 177
6.35 Difference between the inductor currents of phase b. . . . . . . . . . . . 177
6.36 Difference between the inductor currents of phase c. . . . . . . . . . . . 178
6.37 Spectrum of the inductor current of the master converter (phase a). . . 179
6.38 Spectrum of the circulating current of phase a. . . . . . . . . . . . . . . 179
6.39 Spectrum of the current flowing through the ac source of phase a. . . . 180
6.40 Definition of the dc-link rail currents of two parallel VSC modules. . . . 181
6.41 Implementation of the nonlinear current control using two current sen-sors per module. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
6.42 Inductor current waveforms of phase a when using two current sensorsper module. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
6.43 Current flowing through the ac source of phase a when using two currentsensors per module. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
6.44 Detail of the transition from Sector I to Sector II. . . . . . . . . . . . . 185
6.45 High frequency circulating current of phase a when using two currentsensors per module. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
6.46 Detail of the high frequency current ripple transitioning to free-wheelingcurrent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
A.1 Parts of a wind turbine. . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
xiv
A.2 a) Definition of angle of attack and rotation axis, b) Pitch control, andc) Active stall control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
A.3 Blade tip control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
A.4 Power output comparison between a variable-speed pitch-controlled tur-bine and a fixed-speed stall-regulated turbine [10]. . . . . . . . . . . . . 210
A.5 Section of power surface domain (vertical view) [11]. . . . . . . . . . . . 212
A.6 Direct drive arrangement. . . . . . . . . . . . . . . . . . . . . . . . . . . 217
A.7 Doubly-fed induction generator system. . . . . . . . . . . . . . . . . . . 220
B.1 a) Equivalent circuit for phase a and b) the corresponding phasor diagram.221
B.2 Combinatorial logic for T when using zero vectors {1 1 0} for SectorsI and IV, {1 0 1} for Sectors II and V, and {0 1 1} for Sectors III and VI.224
B.3 SABER simulation schematic of the power stage and PWM of TY. . . . 225
C.1 Schematic of the power stage of T. . . . . . . . . . . . . . . . . . . . . 227
C.2 Schematic of the 0 and dq0 transformations. . . . . . . . . . . . . . . 228
C.3 Block diagram of the plant. . . . . . . . . . . . . . . . . . . . . . . . . . 230
C.4 Schematic of the current compensator. . . . . . . . . . . . . . . . . . . . 231
C.5 a) Drive signals of switches S1, S2, and S3 of the six-switch VSC forSector I and b) derivation of the time associated with the applicationof each voltage space vector in Sector I. . . . . . . . . . . . . . . . . . . 232
C.6 Schematics of the inverse transformation and duty cycle dS1 in SectorsI and IV. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
C.7 Schematics of the duty cycle selection for switch S1 and PWM. . . . . . 235
C.8 Simulation result of the phase a source voltage and phase a input currentof T for PFC operation and closed-loop current control. . . . . . . . . 236
D.1 Current envelopes of iS1(t). . . . . . . . . . . . . . . . . . . . . . . . . . 237
D.2 Switch current iS1(t) for a switching cycle of each subsector. . . . . . . . 238
D.3 Switch currents and voltages across the switches for Subsector Ia. . . . . 239
D.4 Currents through switch S1 and diode D1 over a line cycle and over aswitching cycle for different sectors. . . . . . . . . . . . . . . . . . . . . 241
D.5 Variation of i2S1,rms for rated power operation and a 24o phase lead. . . . 242
D.6 Switching loss instants for topology TS during Subsector Ia. . . . . . . . 244
xv
ACKNOWLEDGMENT
I would like to thank my husband, Eduardo. He was my rock during these
past four years, supporting me in my decisions and understanding the long hours
and dedication required to complete my research. Thank you for your love.
I would also like to thank my parents and my brother who encouraged me to
pursue a Ph.D. and always provided me with the strength to persevere.
A special thanks to Prof. Sun for his time and dedication. His hard questions
and critiques have pushed me to constantly look for ways to improve my work.
I would like to thank my committee members, Prof. Connor, Prof. Parsa, Prof.
Salon, and Dr. Kokernak, for their time, questions, and suggestions.
I would like to thank Jerry Dziuba for always providing us with everything we
needed to successfully complete our experimentation, both in the classroom (for my
TA assignments) and in the lab.
I would like to thank John Szczesniak for his time and willingness to manu-
facture the inductor brackets for our prototype. His technical expertise, both in the
design of the brackets and in the use of the machines in his laboratory, and his good
humor made the task a breeze.
A special thanks to Ann Bruno whose friendship has meant the world to me...
Her friendly ear and words of encouragement were always there whenever I needed
them the most.
A special thanks to Priscilla Magilligan who was always there for me, as a
friend and as my go-to person for all my questions on RPI procedures.
I would also like to thank Zhonghui Bing, Troy Beechner, and Min Chen for
our time working together and for your friendship.
xvi
ABSTRACT
Of all the alternative and renewable energy sources, wind power is the fastest
growing alternative energy source with a total worldwide capacity of over 93 GW
as of the end of 2007. However, making wind energy a sustainable and reliable
source of electricity doesnt come without its set of challenges. As the wind tur-
bines increase in size and turbine technology moves towards off-shore wind farms
and direct drive transmission, the need for a reliable and efficient power electronics
interface to convert the variable-frequency variable-magnitude output of the wind
turbines generator into the fixed-frequency fixed-magnitude voltage of the utility
grid is critical.
This dissertation investigates a power electronics interface envisioned to op-
erate with an induction generator-based variable-speed wind turbine. The research
conclusions and the interface itself are applicable to a variety of applications, includ-
ing uninterruptible power supplies, industrial drives, and power quality applications,
among others. The three-phase PWM rectifiers with ac-side bidirectional switches
are proposed as the rectification stage of the power electronics interface. Modulation
strategies are proposed for the rectifiers and the operation of the rectifiers in con-
junction with an induction generator is demonstrated. The viability of using these
rectifiers in place of the standard three-phase voltage-source converter is analyzed
by comparing losses and common-mode voltage generation of the two topologies.
Parallel three-phase voltage-source converter modules operated in an inter-
leaved fashion are proposed for the inversion stage of the power electronics inter-
face. The interleaved three-phase voltage-source converters are analyzed by deriving
analytical models for the common-mode voltage, ac phase current, and dc-link cur-
rent to reveal their spectra and the harmonic cancellation effects of interleaving.
The practical problem of low frequency circulating current in parallel voltage-source
converters is also analyzed. The low frequency circulating current characteristics
of abc, dq, and nonlinear average current control are determined and experimental
results for the nonlinear average current control are presented.
xvii
CHAPTER 1
Introduction
A crisis is an opportunity riding the dangerous wind. (Chinese proverb)
It is difficult to deny the current energy crisis that has overwhelmed the U.S.
in recent years. During most of the twentieth century, the price of oil in the U.S.
was heavily regulated by production or price controls [12]. However, the price of the
barrel of oil has sky-rocketed from $15.71 per barrel in 1999 (adjusted for inflation
of 2007) [13], to $145.00 per barrel as of July 3, 2008 [14], due to the aftermath
of September 11, OPEC cuts, Hurricane Katrina, the Iraq War, the growth of the
Asian market, and the weaker U.S. dollar [12].
The pressure on the oil markets is expected to continue and increase as the
air of uncertainty enveloping the global oil supply and demand in China, the Mid-
dle East, and Latin America increase. The world consumption of liquid fuels and
petroleum by-products is expected to grow by approximately 900,000 barrels per
day in 2008 and by an additional 1.4 million barrels per day in 2009 [14].
With the governments ban on exploring new oil sites, the U.S. has had to
rely more and more on foreign oil to meet its ever increasing demand. In 1973, the
year of the oil embargo, the U.S. was importing only 24% of its oil requirements.
At the start of the Gulf War in 1990, the importation of oil had grown to 42%.
Currently, the U.S. imports approximately 70% of its need [15]. To put this in a
global perspective, the U.S. consumes almost 25% of the worlds oil for only 4% of
the worlds population [15].
At current oil prices, it is expected that the U.S. will send $700 billion to
foreign countries by the end of the year. Over a ten year period and considering a
fixed barrel price, the total cost of importing foreign oil will be $10 trillion dollars.
This will represent the largest transfer of wealth in the history of mankind [16].
The U.S. dependency on foreign oil has raised grave concerns over the economy,
the environment, and, especially, national security. Although the U.S. has begun
the process of weaning itself off foreign oil by possibly approving measures to drill in
the Gulf of Mexico and ANWR (Arctic National Wildlife Refuge in Alaska), increas-
ing the fleet of electric and ethanol-based vehicles, and researching and developing
1
alternative sources of energy, it is still at the infant stage. Although it is projected
that U.S. oil consumption will decline by approximately 400,000 barrels per day in
2008, the cost of residential energy is expected to rise by an annual average of 5.2%
in 2008 and 9.8% in 2009 [14]. It is also projected that electric energy consumption
will increase substantially when plug-in hybrid electric and electric vehicles start to
replace gasoline-based vehicles. Table 1.1 presents the net generation of electricity
in the U.S. by energy source according to the Energy Information Association [14].
Table 1.1: Net generation of electricity by energy source in thousandmegawatthours [1].
Period Coal PetroleumNatural
Nuclear Hydroelectric RenewablesGas
1994 1,690,694 105,901 460,219 640,440 260,126 76,5351995 1,709,426 74,554 496,058 673,402 310,833 73,9651996 1,795,196 81,411 455,056 674,729 347,162 75,7961997 1,845,016 92,555 479,399 628,644 356,453 77,1831998 1,873,516 128,800 531,257 673,702 323,336 77,0881999 1,881,087 118,061 556,396 728,254 319,536 79,4232000 1,966,265 111,221 601,038 753,893 275,573 80,9062001 1,903,956 124,880 639,129 768,826 216,961 70,7692002 1,933,130 94,568 691,006 780,064 264,329 79,1092003 1,973,737 119,406 649,908 763,733 275,806 79,4872004 1,978,620 120,771 708,854 788,528 268,417 82,6042005 2,013,179 122,522 757,974 781,986 270,321 87,2132006 1,990,926 64,364 813,044 787,219 289,246 96,4232007 2,020,572 65,708 893,211 806,487 248,312 102,988
The statistics are grim; however, alternative energy sources may be the source
of relief in the long run. What was once considered the brainchild of environmen-
talists is becoming a reality, fueled in part by the popularity of the global warming
theory and the public outcry over the ever increasing prices at the pump. There is a
particular interest in renewable energy sources, such as solar, wind, hydro, geother-
mal, and tidal, which are naturally replenished by nature and are domestic. The
Energy Policy Act of 2005 authorized subsidies for wind and other alternative energy
producers and established that the U.S. should obtain 7.5% of its electrical energy
from renewable energy sources by 2013 [17].
2
1.1 Wind Energy
Of all the alternative and renewable energy sources, wind power is the fastest
growing alternative energy source with a total worldwide capacity of over 93 GW as
of the end of 2007 [5]. The worldwide installed capacity, however, is only a scratch
on the surface of the total potential energy that can be captured from the wind. In a
recent study at Stanford University, it was concluded that the global locations with
sustainable Class 3 winds can produce 72 TW of energy [18]. Even if only 20% of
the 72 TW were successfully converted into electrical energy, it could satisfy many
times over the global consumption of 1.6 to 1.8 TW [18].
As can be observed in Fig. 1.1, the worldwide installed capacity has grown
continuously each year by an average of approximately 29%. The wind power indus-
try is continuously increasing the size, capacity, and efficiency of the wind turbines
due to investments in research and development. The Global Wind Energy Coun-
cil predicts that by 2012 the installed capacity will have increased by 155% of its
current size to 240 GW [6].
3.5
31G
W
4.8
21
GW
6.1
04
GW
7.6
36
GW
10.1
53
GW
13.5
94
GW
17.3
57
GW
24.3
90
GW
31.2
28
GW
39.4
31
GW
47.3
17
GW
59.0
84
GW
74.2
23
GW
10
20
30
40
50
60
70
80
90
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
Cu
mu
lati
ve
GW
Inst
all
ed
Figure 1.1: Global wind power growth from 1994 to 2007 [5].
Over 70 countries contribute to the total worldwide production of wind energy,
with Europe accounting for 65% of the total, although the largest growths over the
past years have been observed outside Europe [6]. The top ten countries in installed
3
wind generation capacity are presented in Fig. 1.2.
SpainIndia
China
Denmark
Italy
France
UKPortugal Restof
theworld Germany
USA
Country MW %
Germany
USA
China
India
Denmark
Italy
France
UK
Portugal
Restoftheworld
22,247
16,818
15,145
7,845
5,906
3,125
2,726
2,454
2,389
2,150
13,060
23.7
Spain
17.9
16.1
8.4
6.3
3.3
2.9
2.6
2.5
2.3
13.9
Figure 1.2: Top 10 countries in installed wind generation capacity [6].
The growth of installed wind power capacity can be attributed to many fac-
tors including interest in green power, increase in the price of fossil fuels, maturing
technology, involvement of more businesses in the wind industry, and government in-
centives to produce a greater percentage of national energy from renewable sources.
Significant investments have been made in the research and development of the tech-
nologies that compose the wind turbine. The results of these efforts are evident in
the increased energy capture achieved by modern turbines. This growth could not
have been possible without improvements in foundation structures, materials, blade
aerodynamics, gearbox technology, and energy conversion systems. The evolution
of the size of wind turbine diameters and energy capture capability are presented
in Fig. 1.3. The size of the turbines have increased over 1000% in the past twenty
years and have become the largest rotating machines on Earth [7].
Making wind energy a sustainable and reliable source of electricity doesnt
come without its set of challenges. First and foremost, it is still more expensive
than conventional energy sources. According to data from the Department of Energy
(DOE), wind energy currently costs $0.06 to $0.09 per kWh at good wind sites versus
$0.02 to $0.04 per kWh for energy generated by coal plants [19]. A second challenge
is the intermittent nature of wind, which requires that the system have some form
4
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2004
0.05 0.3 0.5 1.3 1.6 2 4.5 5
Year
MW
126m
12m
Figure 1.3: Evolution of wind turbine size [7].
of energy storage to be viable. However, with increasing government subsidies and
continuing research into new and more efficient energy storage components, wind
energy is projected to be competitive with conventional fossil fuel sources by 2012.
1.2 Dissertation Scope
This dissertation considers a variable-speed pitch-regulated wind turbine em-
ploying an induction generator for the electromechanical conversion of the energy
obtained from the wind. The focus of the dissertation is the power electronics inter-
face between the induction generator and the three-phase electrical grid. Although
the chosen application is wind energy conversion, the interface and research con-
clusions are applicable to a variety of applications, including uninterruptible power
supplies, industrial drives, and power quality applications, among others.
The energy conversion is considered to occur in two stages, from the wild
ac energy at the output of the induction generator to a constant dc voltage (ac/dc
conversion) and from the constant dc voltage to the fixed-frequency fixed-amplitude
three-phase ac voltage of the grid (dc/ac conversion). Each wind turbine presents its
own ac/dc converter and the farm shares a common inversion stage. The interme-
diate dc voltage is the common connection point for the wind turbines of the wind
farm, as illustrated in Fig. 1.4, forming a high voltage dc grid (HVDC). The energy
5
of the HVDC grid is then converted to a form compatible with the grid by parallel
dc/ac converters.
AC/DC
Grid
Dc-Link
AC/DC
DC/AC
Transformer
DC/ACDC/AC
Figure 1.4: Wind energy system with a power electronics interface.
The remainder of this chapter will describe the current state-of-the-art in
variable-speed pitch-reguated wind turbines and the power electronics technology
used to convert the energy of the wind into electrical energy. Further discussions on
wind energy concepts, definitions, turbine control, and fixed-speed versus variable-
speed technologies for large wind turbines can be found in Appendix A.
1.3 Power Capability
The rotor, gearbox, and generator of a wind turbine can be represented as
blocks with a given efficiency, as presented in Fig. 1.5. This block diagram will be
used in Chapter 3 for the analysis of the conversion of the energy in the wind to
electrical energy.
The power of the wind (Pw) is extracted by the rotor with a given perfor-
mance coefficient (Cp) and transformed into mechanical power in the low-speed
shaft. This mechanical energy is transformed from low-speed/high-torque energy to
high-speed/low-torque energy by the gearbox. Given the efficiency of the gearbox
(t) and the gearbox transformation ratio, the mechanical power of the high-speed
shaft (Ps) is transferred to the generator which converts the mechanical power into
6
electrical power.
Rotor( )Cp
Gearbox
( )ht
Generator
( )hg
Pw
v
Pt Pm Pe
wt wm we
Figure 1.5: Wind turbine block diagram.
The power available in the wind is determined by
Pw =1
2Av3 (1.1)
where is the air density given in kg/m3, A is the area swept by the rotor blades
given in m2, and v is the undisturbed wind speed given in m/s. The power capture
of the turbine is directly proportional to the area swept by the rotor, which explains
the effort to increase the rotor diameter of modern turbines (Fig. 1.3).
The combined average efficiency of a typical upwind stall-regulated three bladed
turbine is slightly above 20% [20]. However, the efficiency of a turbine varies sig-
nificantly with the wind speed. The wind turbine should be designed to achieve
maximum efficiency at the wind speed with the highest probability at the site. At
low wind speeds, the efficiency is not as critical. At high wind speeds, the turbine
will shed power deliberately when the power available in the wind is higher than the
rating of the generator.
As the wind passes through the rotor blades, the low-speed shaft of the turbine
will rotate with an angular velocity given by
t =v
r
where r is the radius of the rotor blades in meters and is the tip-speed ratio,
which will be explained in Section A.3.1. The transformation ratio of the gearbox,
x, multiplies rotational speed t and converts the low-speed/high-torque energy of
the low-speed shaft into high-speed/low-torque energy at the rotor of the generator
(high-speed shaft)
m = xt
These are the basic equations for any wind turbine.
7
1.4 Wind Turbine Topologies
A wind turbine is composed of the input (rotor), transmission (mechanical
energy transmission), generator (mechanical to electrical energy conversion), and
additional circuitry for connection to the utility grid. Given the variety of options
available for each component of the turbine as well as turbine control options, there
are a multitude of wind turbine topologies that can be implemented. Table 1.2
presents the current state-of-the-art in wind turbine topologies [21]. Different com-
binations of generators, gearbox/direct drive, and power electronics result in different
variable-speed wind turbine topologies. The different options for transmission and
generators are explored Appendix A. The induction generator, its power electronics
interface, and the connection of the system to the grid will be the subjects of the
follwing sections.
Table 1.2: Wind turbine topologies.
Input TransmissionMachine Type
Rotor StatorPower Electronics Grid
Fixedspeed
VariableSpeed
Gearbox
InductionGenerator Soft-starterand
Capacitorbank
3 Acgridf
SquirrelCage Wound
Gearbox
SynchronousGenerator
PermanentMagnet Wound
LargePowerElectronics
Converter
Stator-Side 3 Acor
Dcgrid
f
VariableSpeed GearboxLargePowerElectronics
Converter
Stator-Side 3 Acor
Dcgrid
fInductionGenerator
SquirrelCage Wound
VariableSpeed GearboxSmallPowerElectronics
Converter
Rotor-Side 3 Acor
Dcgrid
f
WoundWound
Doubly-FedIG
VariableSpeed
SynchronousGenerator
PermanentMagnet Wound
LargePowerElectronics
Converter
Stator-Side 3 Acor
Dcgrid
f
Direct
VariableSpeedLargePowerElectronics
Converter
Stator-Side 3 Acor
Dcgrid
f
Direct
Novelmachines
WoundWound
8
1.5 Induction Generator
The induction generator is currently the most widely used generator in com-
mercial wind turbines [22], due to its low cost and robustness. The first induction
generator-based wind turbines were built in the late 1980s and were rated for less
than 100 kW [23]. This generator has been used in both fixed-speed and variable-
speed wind turbines (refer to Table 1.2). In both cases, the cheaper and more robust
squirrel cage induction generator has been preferred over the wound rotor induction
generator. The squirrel cage is mechanically simple and is resistant to the effects of
possible dirt ingress and vibration. This limits the maintenance of the squirrel cage
induction generator to occasional bearing lubrications [22], making it a cost effective
choice for both initial system cost and long-term maintenance costs.
The main drawback of the induction generator is its reactive power consump-
tion. As the active power at its output increases, it requires more reactive power
to operate. In fixed-speed wind turbines, the need for reactive power was addressed
by connecting a capacitor bank between the stator windings and the grid. In the
case of variable-speed wind turbines, the power electronics interface is required to
provide the reactive power consumed by the generator.
The induction generator is well suited for responding to rapid wind variations
since it can increase or decrease its speed slightly if the torque varies [22] [24]. This
characteristic of the machine reduces the wear and tear on the gearbox.
1.6 Grid Interconnection Requirements
Wind turbines connected to the grid could not accommodate the voltage and
frequency transients that often occur on the electric grid [25]. It was common
practice for wind turbines to be disconnected from the grid whenever a disturbance
was detected. However, with the ever increasing penetration of wind turbines in
the grid, these turbines must now contribute to the stability of the grid and have
the capability of riding-through faults in the same fashion as conventional power
generation equipment [25].
Each country has adopted a set of standards for the interconnection of large
wind turbines and wind farms to their respective electrical grids. Some of the most
complete and developed standards have been proposed by Germany, Denmark, and
Spain due to the large penetration of wind energy in their respective electric grids
9
[26]. The United States currently uses the IEEE 1547 standard and the Federal
Energy Regulatory Commissions Order No. 661.
IEEE standard 1547 offers power quality limits regarding the harmonic content
of the currents injected into the grid by any distributed resource (wind turbines,
photovoltaic systems, fuel cell systems, etc.) with an output below 10 MW. This
standard is relevant for large wind turbines individually connected to the grid.
The Federal Energy Regulatory Commissions Order No. 661, instated in June
of 2005, determines the interconnection standards of large wind farms (over 20 MW)
regarding the power factor (reactive power capability) and low voltage ride-through
requirements.
1.6.1 Voltage Fault Ride-Through
One of the main driving factors for the use of power electronics interfaces for
wind turbines connected to the grid is the requirement that the turbine ride-through
low voltage conditions at the point of interconnection [27]. The low voltage ride-
through requirements are outlined in the Federal Energy Regulatory Commissions
Order No. 661. It states that the wind turbine system must comply with the low
voltage ride-through standard if the Transmission Providers Impact Study deems it
necessary for the safety or reliability of the overall system [8]. The voltage levels of
a disturbance and undervoltage duration that the wind turbine must accommodate
are highlighted in Fig. 1.6 [8]. Outside of the highlighted area, the turbine is not
required to remain online.
1.6.2 Power Factor
When the penetration of wind turbines was limited to small units, there was
no need for the wind turbines to provide reactive power to the grid. However, as the
size of the wind turbines increase and the penetration levels reach 10 to 15% [25],
the Transmission Provider now requires that wind turbines operate within a power
factor range to help balance the reactive power of the transmission system [8].
The Federal Energy Regulatory Commissions Order No. 661 states that the
wind turbine system must maintain a power factor within the limits of 0.95 leading
and 0.95 lagging at the high voltage side of the wind plant substation transformer
if the Transmission Providers Impact Study deems it necessary for the safety or
10
10.4
0.1
0 625 3,000 t [ms]
P.u.V
olt
age
att
he
Poin
tof
Inte
rconnec
tion
Faultoccured
Lowervalueofthevoltageband
0.2
0.3
0.5
0.6
0.7
0.8
0.9
Figure 1.6: Low voltage ride-through standard [8].
reliability of the overall system [8]. This power factor requirement should take into
consideration the voltage level at the point of interconnection and the active power
output of the wind turbine system.
The wind turbine system should also be capable of dynamic voltage regulation
in place of the power system stabilizer and automatic voltage regulation if the Impact
Study deems it necessary [8].
1.6.3 Power Quality
The IEEE 1547 standard determines the limits of harmonic current generated
by the distributed resource, in this case, the wind turbine system that can be injected
into the grid at the point of interconnection. The dc current shall not exceed 0.5%
of the rated output current at the point of interconnection [2]. The harmonics
generated by the distributed resource shall not exceed the limits established in Table
1.3 [2]. The current harmonic limits exclude those harmonics generated due to
existing voltage distortion in the local grid before the connection of the distributed
resource.
11
Table 1.3: Maximum harmonic current distortion in percent of current(I)a [2].
Oddh < 11 11 h < 17 17 h < 23 23 h < 35 35 h TDD cHarmonic
Order b
Percent 4.0 2.0 1.5 0.6 0.3 5.0
a I = the greater of the local electric power system maximum load current integrateddemand (15 or 30 minutes) without the distributed resource unit, or the distributedunit rated current capacity (transformed to the point of common connection whena transformer exists between the distributed unit and the point of common connec-tion).b Even harmonics are limited to 25% of the odd harmonics limits above.c Total Demand Distortion
1.7 Survey of Power Electronics Interface Topologies for In-
duction Generators
The wind power industry is currently moving towards direct drive transmission
and full power electronics architectures for the interface between the wind turbine
and the utility grid. The difficulties in implementing these architectures increase
substantially as the power level of the wind turbine increases. This is due to several
factors ranging from semiconductor technology, the chosen architecture, modulation
and control strategies, overall control of the system, control partitioning, efficiency,
and cost.
Wind energy systems can be divided into three main categories with regards
to the power electronics interface: systems without a power electronics interface,
systems with a partially rated power electronics interface, and systems with a fully
rated power electronics interface [28] (Table 1.2).
Wind turbines without a power electronics interface are generally stall-regulated
fixed-speed wind turbines employing an induction generator directly connected to
the grid. The reactive power consumed by the machine is dependent on the ac-
tive power that is being generated and, therefore, varies with the wind speed. The
reactive power is compensated by a capacitor bank which is composed of sets of
smaller capacitors that can be switched in an out with mechanical contactors [29].
The difference between the instantaneous voltage of the grid and the voltage across
the capacitors results in transients every time capacitors are switched in. These
transients compromise the lifetime of both the capacitors and contactors [30]. In
12
[29], a variable capacitor bank using thyristors to control the switch-in instant was
proposed to reduce these transients. The thyristors turn on, connecting the capac-
itor bank to the system, when the instantaneous voltage of the grid matches the
capacitor voltage. The variable capacitor bank can be emulated by a voltage source
converter, which provides continuous and smooth reactive power compensation [31].
This solution is similar to the use of power electronics converters for correcting power
quality problems of the grid (e.g., UPFC, DVR, STATCOM).
The problem with fixed-speed wind turbines is their limited power regulation
and controllability which does not allow them to contribute to grid stability [30].
Variable-speed wind turbines with a power electronics interface between the turbine
and the grid can be properly controlled to comply with the grid interconnection
requirements.
Wind turbines with a partially rated power electronics interface are based
on the doubly-fed induction machine or wound rotor induction generator with a
controllable rotor resistance [28]. The latter, as illustrated in Fig. 1.7, is composed
of a reactive power compensator, a soft starter, and an additional resistance added
to the rotor windings. This resistance is variable and controlled by a partially rated
power electronics converter (high current/low voltage) which allows the speed to
vary between 2 and 4% [28]. The doubly-fed induction generator, as discussed in
Section A.5.2, uses a partially rated power electronics converter connected to the
wound rotor of the induction generator through slip rings. The speed variation of
this system is directly related to the power rating of the converter.
The fully rated power electronics interface is connected between the machine
and the grid. It is the most versatile of the three interface types since the turbine
speed varies to extract maximum power at any given wind speed, machine selection
is no longer restricted to a certain technology, the gearbox can be eliminated (direct
drive transmission), and the converter that interfaces with the grid can be controlled
to comply with interconnection standards. However, it is also the interface that
sustains the highest losses since it processes all of the power extracted from the
wind. A survey of fully rated power electronics interfaces proposed in the literature
for the squirrel cage induction generators of wind turbines will be presented in the
following subsections.
13
ResistanceControl
ReactivePower
Compensator
Soft-Starter Grid
WoundRotorInductionGenerator
Figure 1.7: Wound rotor induction generator with a variable rotor resis-tance.
1.7.1 Diode Rectifier
The three-phase diode rectifier, presented in Fig. 1.8, is the simplest ac-dc
converter. The commutation of the diodes occurs naturally depending on the voltage
biasing and current flow and is, thus, said to be uncontrolled. The commutation
frequency is equal to the frequency of the stator voltage. The advantage of this
rectifier over high frequency pulse-width modulation topologies is the robustness
and reliability of the high power diodes.
a) b) c)
Figure 1.8: Three-phase diode rectifier with (a) an LC output filter, (b)an input inductive filter and capacitive output filter and (c)a capacitive output filter.
The three-phase diode rectifier, however, is an unidirectional converter, mean-
ing that power can only flow in one direction: into the converter. Therefore, the
diode rectifier on its own cannot provide the reactive power required by the induc-
14
tion machine. It must operate in conjunction with a source of reactive power. A
capacitor bank for excitation, as presented in Fig. 1.9, was proposed by [34] for a
small-scale stand-alone dc system. Induction machines with self-excitation capaci-
tors have a well-known voltage regulation problem [35], which is not necessarily a
problem in the case of wind turbines with a power electronics interface. The capac-
itor bank is designed to provide the reactive power necessary to maintain the rated
stator voltage at the rated wind speed, therefore, stator voltage and frequency will
vary significantly with changes in wind speed. However, increasing the load of the
system or low wind speeds could cause the stator voltage to collapse to zero [36].
This can be avoided by increasing the excitation capacitors; however, the efficiency
of the machine will decrease and the stator currents will increase. Therefore, the
magnetizing current of the machine should be limited and properly controlled so
that the stator currents do not exceed the ratings of the machine [36].
IG
Figure 1.9: Three-phase diode rectifier with a capacitor bank.
The current of the diode bridge is not sinusoidal and can be excessively dis-
torted if the filter inductor on the dc side is not used. The harmonic content of
the diode rectifier currents can cause serious problems to the machine, such as in-
crease in copper and iron losses, pulsating torques due to the positive and negative
sequence harmonics, increase in audible noise, overheating, and possibly damage.
The three-phase diode rectifier is usually followed by a dc-dc conversion stage.
The dc voltage at the output of the diode rectifier changes as the wind speed varies,
therefore, in order to regulate the dc-link voltage for the inversion stage, a dc-dc
converter is placed between the output of the rectifier and inversion stage [32].
The three-phase diode rectifier has been proposed as the rectification stage of a
power electronics interface connected to the grid [32], as well as the rectification stage
for a high voltage dc system (HVDC) [33] [34]. In the first case, the diode rectifier is
15
followed by an inversion stage composed of a voltage source converter or multi-level
converter. In the case of an HVDC, the rectifier can be one of many turbine/rectifier
systems supplying energy to a common dc grid [28], as illustrated in Fig. 1.10. The
HVDC system is becoming the standard for off-shore wind farms. The long distances
between the farm sites and the conversion station on land make ac transmission
impractical due to the distributed capacitance among cables, which is higher in
undersea cables than in overhead power lines and limits the power transmission
capability [32]. Dc grids have the advantages of decoupling the frequencies of the
turbine and the utility grid, isolating the wind farm disturbances on the grid and
vice-versa, reducing the power losses in the cables compared to HVAC, and increasing
the transmission capability per cable [32]. Once on-shore, the energy is converted
into a form compatible with the local utility grid by a power electronics inverter.
It should be noted that different rectifier topologies can be used in place of the
three-phase diode rectifier for HVDC systems.
IGAC
DC
IG
IG
.
.
.
Grid
HVDC TransmissionAC
DC
ACDC
DCAC
Figure 1.10: High voltage dc wind farm architecture.
16
1.7.2 Six Switch Voltage-Source Converter
The three-phase six switch voltage-source converter, presented in Fig. 1.11,
is the leading solution for medium to high power applications. This converter is
extremely flexible in its operation, capable of generating any current waveform and
operating as a rectifier or inverter (bidirectional active power flow) while provid-
ing/absorbing reactive power (bidirectional reactive power flow). Due to its ver-
satility, it has been the subject of countless publications regarding every aspect of
the its operation and use in such diverse applications as motor drives, distributed
generation, reactive power compensation, power factor correction, active filters, and
uninterruptible power supplies.
Vdc
+
_
Figure 1.11: Six switch voltage-source converter.
This topology has been proposed for both the rectification and inversion stages
of the power electronics interface of induction generator-based wind turbines [22] [32]
[38] [39] [40] [41] [42] [43], due to its bidirectional power flow.
The six switch voltage-source converter presents a two-level output, i.e., the
voltage at the mid-points of the switch legs can only present two values, as presented
in Fig. 1.12. Thus, the filtering requirements are greater in the case of the voltage-
source converter than for multi-level converters, which present N voltage levels for
N converter levels [44], and the matrix converter, which presents three voltage levels
[45].
The voltage-source converter is the subject of Chapter 2; its basic operation
and the modulation strategies that will be used in subsequent chapters will be dis-
cussed.
17
-Vdc/2
Vdc/2
t
Figure 1.12: Two-level output of the voltage-source converter.
1.7.2.1 Back-to-Back Voltage-Source Converters
The back-to-back voltage-source converter topology, as illustrated in Fig. 1.13,
is the most widely used topology for wind energy conversion [45]. It is composed of
two voltage-source converters, one connected at the terminals of the induction gen-
erator and operating as a rectifier and the other connected to the grid and operating
as an inverter. The two converters share a common dc-link which decouples of the
controls of the converters [45]. Besides its use with induction generators, the back-
to-back voltage-source converter is the standard topology for doubly-fed induction
generators [32] [40].
IG
Figure 1.13: Back-to-back voltage source converters.
The advantage of this topology is its simplicity and the fact that it can easily
meet the requirements of both the induction machine and the grid. In the case of the
grid interconnection requirements, the inverters active and reactive power flow can
be controlled quickly and independently from the generator-side converter [28] [38].
Besides meeting the grid requirements, the grid-side converter should be controlled
to maintain the dc-link voltage constant and larger than the peak line voltage of the
grid throughout the variations in wind speed [45] [39].
A disadvantage of this topology is the low efficiency of the rectifier at low wind
speeds. The inverter always operates with high efficiency, since the dc-link voltage
18
is maintained constant at a value slightly larger than the peak line voltage of the
grid [46]. The voltage amplitude generated at the stator terminals of the induction
generator is a function of the wind speed, low wind speeds result in low voltage
amplitudes. The difference in the amplitudes of the stator voltages and dc-link
voltage will force the switches of the rectifier to remain on for longer periods of time
in order to build up energy in the inductors and successfully boost the generator
voltages; this ultimately increases the losses of the rectifier [46]. As the wind speed
increases, the efficiency of the rectifier increases and becomes comparable with the
efficiency of the inverter at the rated wind speed.
1.7.2.2 Interleaved Voltage-Source Converters
The interleaved voltage-source converter topology is essentially N voltage-
source converters connected in parallel with phase-shifted drive signals (refer to
Chapter 5). The topology is presented in Fig. 1.14 for N interleaved voltage-source
converters. As in the case of a single voltage-source converter, this topology can be
used as a rectifier or an inverter.
...
Vdc
+
_
... ...
Figure 1.14: Interleaved voltage-source converters.
Each paralleled converter module processes only a fraction of the total power
extracted from the wind. This is advantageous for multi-megawatt wind turbines and
for the inversion stage of a wind farm (Fig. 1.10), where the use of a single voltage-
source converter is limited by the semiconductor technology used to implement the
switches. Table 1.4 presents the power semiconductor switches most widely used for
power electronics converters and their maximum ratings [3].
19
Table 1.4: Maximum ratings of power semiconductor devices [3].
Voltage(V)
Current(A)
Output(kVA)a
Turn-offtime( s)mb
Frequencyrange(kHz)
Driverequirement
BPT IGBT MOSFET GTO IGCT
1,200
800
480
15-25
0.5-5
Medium
6,500
3,600
4,000
1-4
2-20
Low
1,200
700
70
0.3-0.5
5-100
Low
6,000
6,000
24,000
10-25
0.2-1
High
6,000
6,000
24,000
10-15
>2
High
Symbol
a
b
Idealcapacityperswitch.
Includesdelaytimesandpartialcurrentphase.
For cost reasons, the dc-link of the power electronics interface of high power
wind turbines (> 100 kW) is generally designed to be lower than 1 kV [3]. In the
case of multi-megawatt wind turbines and wind farms, the current capability of the
switches would need to be in the kA range, which is beyond the current state-of-
the-art of power semiconductors. In order to bypass this problem, the literature
has proposed to connect converters in parallel [42] [43], or switches in parallel. The
latter can be problematic since no two switches are identical and switching times
may differ sufficiently to cause damage to the converter [43].
Paralleling voltage-source converter modules reduces the stress on the switches,
reduces losses, increases the redundancy of the system, and, consequently, the relia-
bility. By interleaving the drive signals of the modules, harmonic cancellation occurs
which ultimately reduces the filtering requirements on the ac-side of the converter.
1.7.3 Multilevel Converters
Multilevel converters are a family of power electronics converters initially
conceived to reduce the harmonic content of the synthesized voltage waveform
[47]. These converters synthesize sinusoidal voltages at their outputs by generating
stepped voltage waveforms [44]. The difference between the multilevel converters
and the two-level voltage-source converter is the number of voltage levels; while
20
the voltage-source converter produces only two voltage levels (refer to Fig. 1.12),
the multilevel converters can theoretically produce an unlimited number of voltage
levels, depending only on the number of converter levels. The minimum number
of voltage levels a multilevel converter can generate is three. Figure 1.15 presents
the voltage output of a three-level (PWM) and a five-level (non-PWM) multilevel
converter.
tt
a) b)
Figure 1.15: Output voltage of (a) a three-level and (b) a five-level mul-tilevel converter.
By increasing the number of voltage levels in the output voltage waveform,
it is possible to achieve an output voltage which more closely resembles a sinewave
and, thus, reduces the harmonic content of the voltage. The lower total harmonic
distortion reduces the size of the output filters [47], which has a significant impact
on the size and weight of the converter. The increase in the number of voltage levels
also reduces the switching dv/dt which decreases the potential for electromagnetic
interference problems [47].
The three main multilevel converter topologies are the diode-clamped multi-
level converter, the flying-capacitor multilevel converter, and the cascaded H-bridge
multilevel converter. A single leg schematic of a five-level version of each converter
is presented in Fig. 1.16. The cascaded H-bridge multilevel converter is composed of
single-phase full-bridge converters. Each H-bridge converter requires a separate and
floating dc voltage source. Due to this characteristic, this converter is not widely
used in wind energy conversion due to its high cost and poor performance; however,
it is a good candidate for photovoltaic energy conversion [46].
The diode-clamped and flying-capacitor multilevel converters have been pro-
posed to implement the power electronics interface (rectifier and/or inverter) of a
21
Vdc
+
_
a)
Vdc
+
_
b)
+
_
Vdc1
+
_
Vdc2
c)
Figure 1.16: Single leg of a five-level (a) diode-clamped multilevel con-verter, (b) flying-capacitor multilevel converter, and (c) cas-caded H-bridge multilevel converter.
wind energy system [22] [46] [48] [49]. These topologies present the same output
voltage waveform and are similar in structure. However, the component count is
larger in the case of the diode-clamped topology: one flying capacitor per phase
substitutes two diodes [22].
The common element in the multilevel converter topologies is the presence of a
large voltage achieved by small discrete dc voltages [47]. These discrete dc voltages
compose the voltage steps of the output voltage, as seen in Fig. 1.15. The dc voltages
are achieved by capacitor voltage sources. Maintaining the voltage across these
capacitors balanced is one of the difficulties and disadvantages of multilevel converter
topologies. Generally, additional components or control strategies are required to
22
maintain the voltage balance [32]. This voltage balance problem is easier to handle
in the flying-capacitor topology than in the diode-clamped topology; however, this
is only a substantial problem in the diode-clamped topology when there are more
than three levels [22].
The voltage stress on the switches is proportional to the number of dc sources.
Therefore, a high dc-link voltage level is divided among the dc sources, which limits
the voltage stress of the switches to only a fraction of the dc-link voltage and allows
the use of power devices with lower voltage ratings [47]. This is an advantage over
the voltage-source converter whose switches are subjected to the entirety of the
dc-link voltage.
However, the conduction losses in the case of the multilevel converters are
higher than in the voltage-source converter. This is due to the fact that multiple
switches are in the path of the current, whereas, only one switch of the voltage-source
converter is in the current path.
One of the difficulties regarding the multilevel converters is the redundancy
that exists in the switching states, i.e., multiple switching states achieve the same
output. Achieving an optimum switching sequence is difficult and becomes ever
more complicated as the number of converter levels increases. In general, the main
concerns in developing a modulation strategy for multilevel converters are: mini-
mization of load current harmonics, minimization of the switching frequency, main-
taining a uniform switching frequency among the switches, voltage balance among
the capacitors [47].
These topologies are especially suited for wind farms using HVDC transmission
since the high dc-link voltage can be easily distributed among the N-levels of the
converters. For practical and economical reasons, multilevel converters with four or
more levels are only recommended for power levels that exceed 300 MW. At lower
power levels, the advantages of the multilevel converters are not as pronounced and
the two-level voltage-source converter becomes more economically attractive [48].
The voltage-source converters low efficiency at low wind speeds is not a prob-
lem with the multilevel topologies. The high dc-link voltage is divided among the
capacitors, thus, clamping the voltage across each switch to one capacitor voltage.
Since the voltage across each switch is only a fraction of the dc-link voltage, the
rectifier losses at low wind speeds are significantly reduced compared to the voltage-
23
source converter [46].
1.7.4 Matrix Converter
The matrix converter, as presented in Fig. 1.17, is a semiconductor-based topol-
ogy with no passive components and incorporates the functions of rectifier and in-
verter into one compact topology. It is a three-level topology; therefore, the total
harmonic distortion of the output voltage is smaller than in the case of the back-to-
back voltage-source converter. It has been proposed for wind energy conversion in
[22], [46], [50], [51], and [52].
IG
Outputleg
Figure 1.17: Matrix converter.
The desired output current, voltage, and frequency can be achieved by properly
connecting the output terminals to the input terminals. This is accomplished by
never connecting two (or three) switches of an output leg to an input phase at any
time and by always having the three output phases connected to the input [22].
The switches of the matrix converter are bidirectional in current and voltage,
which poses a problem since pure bidirectional switches are not yet available. How-
ever, bidirectional switches can be realized by combining semiconductor devices,
which increases the component count of the topology. Depending on the realization
of the bidirectional switches, half of the commutations are soft-switching [22], which
reduces the switching losses.
One disadvantage of the matrix converter is the limited output voltage it can
achieve. The maximum output voltage is only 86% of the input voltage. In order
24
to achieve the same output power as a back-to-back voltage-source converter, the
output current would have to be 1.15 times higher. This leads to greater conduction
losses in the matrix converter than in the back-to-back voltage-source converter [45]
[22].
The absence of a dc-link capacitor is both a blessing and a curse for the matrix
converter. Without a dc-link capacitor the efficiency and lifetime of the converter
is increased compared to the back-to-back voltage-source converter. However, with
no dc-link capacitor, there is no decoupling between the input and the output.
Therefore, disturbances at the input/output will propagate through the converter
and affect the output/input [22] [45].
1.8 Dissertation Outline
The rectification stage of the proposed power electronics interface is the subject
of Chapters 2, 3, and 4. The proposed rectifier is the three-phase PWM rectifier
with ac-side bidirectional switches. It will be compared to the standard three-phase
voltage-source converter.
Chapter 2 presents a brief overview of the three-phase voltage-source con-
verter covering its basic operation and modulation schemes. The two most common
naturally-sampled carrier-based modulation strategies will be reviewed: sinusoidal
pulse-width modulation and sinusoidal pulse-width modulation with third harmonic
injection. Space-vector modulation will also be reviewed and the relationship be-
tween space-vector modulation and naturally-sampled carrier-based modulation will
be explained. This converter is used throughout the research and it is necessary to
first define the framework for the use of the converter in subsequent chapters.
A comprehensive analysis of the three-phase PWM rectifiers with ac-side bidi-
rectional switches is presented in Chapter 3. First, a literature overview of these
rectifiers is carried out focusing on the origins of the rectifiers and their applications,
modulation strategies, and performance. Modulation strategies are proposed for the
rectifiers based on understanding of space-vector modulation applied to the three-
phase voltage-source converter. The rectifiers operating with an induction generator
is presented to verify the chosen modu
Top Related