MATLAB/SIMULINK IMPLEMENTATION AND ANALYSIS

OF THREE PULSE-WIDTH-MODULATION (PWM)

TECHNIQUES

by
Phuong Hue Tran

A thesis
submitted in partial fulfillment
of the requirements for the degree of
Master of Science in Electrical Engineering
Boise State University

May 2012

c 2012
Phuong Hue Tran
ALL RIGHTS RESERVED
 BOISE STATE UNIVERSITY GRADUATE COLLEGE

DEFENSE COMMITTEE AND FINAL READING APPROVALS

of the thesis submitted by

Phuong Hue Tran

Thesis Title: MATLAB/Simulink Implementation and Analysis of Three Pulse-

Width-Modulation (PWM) Techniques

Date of Final Oral Examination: 11 May 2012

The following individuals read and discussed the thesis submitted by student Phuong Hue
Tran, and they evaluated her presentation and response to questions during the final oral
examination. They found that the student passed the final oral examination.

Said Ahmed-Zaid, Ph.D. Chair, Supervisory Committee

Elisa Barney Smith, Ph.D. Member, Supervisory Committee

John Chiasson, Ph.D. Member, Supervisory Committee

The final reading approval of the thesis was granted by Said Ahmed-Zaid, Ph.D., Chair of
the Supervisory Committee. The thesis was approved for the Graduate College by John
R. Pelton, Ph.D., Dean of the Graduate College.
 ACKNOWLEDGMENTS

I would like to take this opportunity to express my sincere appreciation to all the people
who were helpful in making this thesis successful. First of all, I would like to express
my gratitude to my advisor, Dr. Said Ahmed-Zaid for his valuable comments, guidance,
and discussions that guided me well in the process of my thesis. I am thankful to my
committee members ( Dr. John Chiasson and Dr. Elisa Barney Smith) for their comments
and suggestions and their willingness to be my committee members.
Most importantly, I would like to express my deepest thanks to my family for giving
me a life of opportunities and backing me up in every corner where my life has been
difficult. Thanks for all their care, support, encouragement, and inspiration all along in my
academic endeavors.

iv
 ABSTRACT

With advances in solid-state power electronic devices and microprocessors, various

pulse-width-modulation (PWM) techniques have been developed for industrial applica-
tions. For example, PWM-based three-phase voltage source inverters (VSI) convert DC
power to AC power with variable voltage magnitude and variable frequency.
This thesis discusses the advantages and drawbacks of three different PWM tech-
niques: the sinusoidal PWM (SPWM) technique, the third-harmonic-injection PWM (THIPWM)
technique, and the space-vector PWM (SVPWM) technique. These three methods are
compared by discussing their ease of implementation and by analyzing the output har-
monic spectra of various output voltages (poles voltages, line-to-neutral voltages, and
line-to-line voltages) and their total harmonic distortion (THD).
The simulation results show that both the THIPWM and SVPWM techniques have
lower total harmonic distortion than the SPWM technique. The THIPWM and SVPWM
techniques in the under-modulation region can both increase the fundamental output volt-
age by 15.5% over the SPWM technique. Moreover, the SVPWM technique can increase
the fundamental output by about 5% in each of the overmodulation regions 1 and 2,
respectively.

v
 TABLE OF CONTENTS

ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Research Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Sinusoidal PWM and Third-Harmonic-Injection PWM . . . . . . . . . . . . . . . . . 6

2.1 Sinusoidal PWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.1 Sinusoidal PWM Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.2 Modulation Index of Sinusoidal PWM . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Third-Harmonic-Injection PWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.1 Concept and Calculation of Optimum Distortion . . . . . . . . . . . . . . . 10

3 Space Vector PWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2 Principle of Space Vector PWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3 Implementation Procedure of a Two-Level Space Vector PWM . . . . . . . . . . 24
3.3.1 Angle and Reference Voltage Vector . . . . . . . . . . . . . . . . . . . . . . . . 26

vi
 3.3.2 Modulation Index of Linear Modulation . . . . . . . . . . . . . . . . . . . . . 27
3.3.3 Sector Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.4 Time Durations Ta , Tb , T0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.5 Determination of the Switching Times for Each Transistor Switch
(S1 - S6 ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3.6 Types of Different Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4 Space Vector PWM in Over-modulation Region . . . . . . . . . . . . . . . . . . . . . . . 44

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.2 Over-Modulation Mode 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.3 Over-Modulation Mode 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

5 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.2 Sinusoidal PWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.3 THIPWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.4 Under-Modulation of Space Vector PWM . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.5 Mode 1 in Over-Modulation of SVPWM . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6 Comparison Between SPWM, THIPWM, and SVPWM Techniques . . . . . . . 75

6.1 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.2 Total Harmonic Distortion (THD) Comparison . . . . . . . . . . . . . . . . . . . . . . 79

7 Conclusion and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

7.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
7.2 Recommendations for Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

vii
 LIST OF TABLES

3.1 Space Vectors, Switching States, and On State Switches . . . . . . . . . . . . . . . 21

3.2 Vectors, Switching Vectors, Phase Voltages and Line to Line Voltages as a
Function of the DC Bus Voltage Vdc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.3 Voltage Vectors, Switching Vector, α and β . . . . . . . . . . . . . . . . . . . . . . . . 25
3.4 Sector Definition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.5 Time Intervals Ta and Tb for Each Sector . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.6 Seven-Segment Switching Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.7 Switching Pulse Pattern for the Three Phase for Each Sector . . . . . . . . . . . . 41
3.8 Switching Sequence for Three-Phase PWM Technique . . . . . . . . . . . . . . . . 42

4.1 The Time Intervals Ta0 and Tb0 for Each Sector . . . . . . . . . . . . . . . . . . . . . . . 48

6.1 Pole Voltage Vao in SPWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

6.2 Voltages Vab , Van , and Vao in THIPWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.3 Harmonics of Vao in Under-Modulation range of SVPWM . . . . . . . . . . . . . 82
6.4 Vao in Over-Modulation Region 1 of SVPWM . . . . . . . . . . . . . . . . . . . . . . 83
6.5 Vab in the SPWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.6 Van in the SPWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.7 Vab in Under-Modulation of SVPWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.8 Van in Under-Modulation of SVPWM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.9 Vab in Over-Modulation Region 1 of SVPWM . . . . . . . . . . . . . . . . . . . . . . 85
6.10 Van in Over-Modulation Region 1 of SVPWM . . . . . . . . . . . . . . . . . . . . . . 85

viii
 LIST OF FIGURES

2.1 Control Signal Generator for SPWM [27]. . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 Three-Phase Sinusoidal PWM Inverter [27]. . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Three-Phase Sinusoidal PWM: a). Reference Voltages (a,b,c) and Trian-
gular Wave b). Vao , c) Vbo , d) Vco e) Line-to-Line Voltages [27]. . . . . . . . . . 8
2.4 One-Phase Third-Harmonic Injection PWM [17]. . . . . . . . . . . . . . . . . . . . . 14
2.5 Reference Voltages (a,b,c), Triangular Waveforms (VT ), and Output Volt-
age (Vao ,Vbo ,Vco ). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3.1 Under-modulation and Over-modulation Regions in Space Vector Repre-

sentation [12]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Three-Phase Bridge Inverter [11]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3 Eight Switching Configuration of a Three-Phase Inverter [7]. . . . . . . . . . . . 20
3.4 Space Vectors of Three-Phase Bridge Inverter [12]. . . . . . . . . . . . . . . . . . . . 22
3.5 Flow Diagram for SVPWM Implementation [11]. . . . . . . . . . . . . . . . . . . . . 25
3.6 Fundamental of Voltage Waveform [10]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.7 Construction of Symmetrical Pulse Pattern for Three-Phase [20]. . . . . . . . . 35
3.8 Vre f Falls into Sector 1 [21]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.9 Switching Patterns in the Six Sectors [20]. . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.10 Switching Patterns of Six Sectors in Circle [20]. . . . . . . . . . . . . . . . . . . . . . 40
3.11 Switching Sequence of all Six Sectors [4]. . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.1 The Two Over-Modulation Regions in Space Vector Representation [12]. . . 45

ix
4.2 Crossover Angle vs. Modulation Index in the Mode 1. . . . . . . . . . . . . . . . . 46
4.3 Over-Modulation Mode Region 1 [12]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.4 Trajectory of Reference Voltage Vector and Phase Voltage Waveform in
Mode 1 [14]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.5 Holding Angle vs. Modulation Index in Mode 2 [11]. . . . . . . . . . . . . . . . . . 51
4.6 Angular Displacement of Reference and Actual Voltage Vectors in Mode
2 [12]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.7 Trajectory of Reference Voltage Vector and Phase Voltage Waveform in
Mode 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

5.1 SPWM System Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

5.2 Vao , Vbo , and Vco of SPWM Waveforms. . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3 Vao , Vbo , and Vco of SPWM after Filtering. . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.4 Neutral Voltages Van , Vbn , and Vcn of SPWM after Filtering. . . . . . . . . . . . . 59
5.5 Neutral Voltages Van , Vbn , and Vcn of THIPWM after Filtering. . . . . . . . . . . 60
5.6 Neutral Voltages Vab , Vbc , and Vca of THIPWM after Filtering. . . . . . . . . . . 60
5.7 Linear Modulation of SVPWM Simulation System. Diagrams for Blocks
1 to 7 are Shown in Figures 5.8 to 5.13. . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.8 Three-Phase Input Sinusoidal Voltages and Modulation Index are Detail
for Block 1 in Figure 5.7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.9 α, β Voltages and Modulation Index are Detail for Block 2 in Figure 5.7. . . 64
5.10 Detail for Block 3 in Figure 5.7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.11 Switching Time Calculation is Detail for Block 4 in Figure 5.7. . . . . . . . . . 65
5.12 Timing Signals and Triangular Waveform Details for Block 6 in Figure 5.7. 65
5.13 Inverter Output Signals to Neutral Voltages is Detail for Block 7 in Figure
5.7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.14 Filtered Timing Signals of Three-Phase and Triangular Waveforms in SVPWM. 67

x
5.15 Vao , Vbo , and Vco for Linear Modulation SVPWM. . . . . . . . . . . . . . . . . . . . . 68
5.16 Vao , Vbo and Vco for Linear Modulation SVPWM after Filtering. . . . . . . . . . 69
5.17 Line to Line Voltages of Linear-Modulation Region. . . . . . . . . . . . . . . . . . . 70
5.18 Neutral Voltages Van , Vbn , and Vcn of Linear Modulation SVPWM after
Filtering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
5.19 Detail of the Lookup Table in the First Block of SPVWM Over-Modulation
Region 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.20 Detail of the Fourth Block of Space Vector PWM in Region 1. . . . . . . . . . . 72
5.21 Vao , Vbo , and Vco of Over-Modulation Region 1 before and after Filtering. . . 73
5.22 Line-to-Neutral of Over-Modulation Region 1 after Filtering. . . . . . . . . . . . 74
5.23 Line-to-Line Voltages of Over-Modulation Region 1 after Filtering. . . . . . . 74

6.1 Locus Comparison of Maximum Peak Voltage in SPWM and SVPWM [28]. 76
6.2 Loci of SPWM, SVPWM and Region 1 of SVPWM. . . . . . . . . . . . . . . . . . 77
6.3 Locus of SPWM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.4 Locus of SVPWM in Linear Modulation Range. . . . . . . . . . . . . . . . . . . . . . 78
6.5 Loci of THIPWM and SVPWM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.6 Locus of Over-Modulation Region 1 at a Cross Angle of 15 Degrees. . . . . . 80
6.7 Locus of Over-Modulation Region 1 at a Cross Angle of 0 Degree. . . . . . . . 80
6.8 Spectrum of Vao for SPWM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.9 Spectrum of Vao for THIPWM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.10 Spectrum of Vao in The Under-Modulation Range of SVPWM. . . . . . . . . . . 83
6.11 Spectrum of Vao for SVPWM (Over-Modulation Region 1). . . . . . . . . . . . . 84

xi
 1

CHAPTER 1

INTRODUCTION

1.1 Introduction

Pulse-width modulation (PWM) is a technique where the duty ratio of a pulsating wave-
form is controlled by another input waveform. The intersections between the reference
voltage waveform and the carrier waveform give the opening and closing times of the
switches.

PWM is commonly used in applications like motor speed control, converters, audio
amplifiers, etc. For example, it is used to reduce the total power delivered to a load without
losses, which normally occurs when a power source is limited by a resistive element.
PWM is used to adjust the voltage applied to the motor. Changing the duty ratio of the
switches changes the speed of the motor. The longer the pulse is closed compared to the
opened periods, the higher the power supplied to the load is. The change of state between
closing (ON) and opening (OFF) is rapid, so that the average power dissipation is very low
compared to the power being delivered. PWM amplifiers are more efficient and less bulky
than linear power amplifiers. In addition, linear amplifiers that deliver energy continuously
rather than through pulses have lower maximum power ratings than PWM amplifiers.

There is no single PWM method that is the best suited for all applications and with ad-
vances in solid-state power electronic devices and microprocessors, various pulse-width-
modulation (PWM) techniques have been developed for industrial applications. For these
reasons, the PWM techniques have been the subject of intensive research since 1970s.
 2

1.2 Literature Review

With advances in solid-state power electronic devices and microprocessors, various in-
verter control techniques employing pulse-width-modulation (PWM) techniques are be-
coming increasingly popular in AC motor drive applications. These PWM-based drives are
used to control both the frequency and the magnitude of the voltages applied to motors [1].
Various PWM strategies, control schemes, and realization techniques have been developed
in the past two decades [2]. PWM strategy plays an important role in the minimization of
harmonics and switching losses in converters, especially in three-phase applications. The
first modulation techniques were developed in the mid-1960s by Kirnnich, Heinrick, and
Bowes as reported in [3]. The research in PWM schemes has intensified in the last few
decades. The main aim of any modulation technique is to obtain a variable output with a
maximum fundamental component and minimum harmonics [4].

The carrier-based PWM methods were developed first and were widely used in most
applications. One of the earliest modulation signals for carrier-based PWM is sinusoidal
PWM (SPWM). The SPWM technique is based on the comparison of a carrier signal and a
pure sinusoidal modulation signal. It was introduced by Schonung and Stemmler in 1964
as reported in [5]. The utilization rate of the DC voltage for traditional sinusoidal PWM is
only 78.5% of the DC bus voltage, which is far less than that of the six-step wave (100%).
Improving the utilization rate of the DC bus voltage has been a research focus in power
electronics [6].

This problem of the under-utilization of the DC bus voltage led to the development
of the third-harmonic-injection pulse-width modulation (THIPWM). In 1975, Buja de-
veloped this improved sinusoidal PWM technique, which added a third-order harmonic
content into the sinusoidal reference signal leading to a 15.5% increase in the utilization
rate of the DC bus voltage [6].

Another method of increasing the output voltage is the space-vector PWM (SVPWM)
 3

technique. SVPWM was first introduced in the mid-1980s and was greatly advanced by
Van Der Broeck in 1988 [9]. Compared to THIPWM, the two techniques have similar re-
sults but their methods of implementation are completely different. With the development
of microprocessors, SVPWM has become one of the most important PWM methods for
three-phase inverters [7]. Many SVPWM schemes have been developed and extensively
investigated in the literature. The goal in each modulation strategy is to lower the switching
losses, maximize bus utilization, reduce harmonic content, and still achieve precise control
[8].

The SVPWM technique utilizes the DC bus voltage more efficiently and generates less
harmonic distortion when compared with the SPWM technique [4]. The maximum peak
fundamental magnitude of the SVPWM technique is about 90.6% of the inverter capacity.
This represents a 15.5% increase in the maximum voltage compared with conventional
sinusoidal modulation [10].

In 1991, Holtz proposed a classical over-modulation technique based on SVPWM,

which divided the over-modulation range into two modes of operation and increased the
utilization rate of the DC voltage to that of a six-step wave [6,11,12,13]. Holtz pro-
posed this technique using switching time calculations in the over-modulation region of
SVPWM.

In 1998, Lee analyzed Holtz’s over-modulation technique graphically, gave some ap-
proximate linear expressions between the modulation index and its own auxiliary param-
eter, and discussed the harmonic problem [14]. He showed that this technique generated
less harmonic distortion in the output voltages (or) the currents applied to the phases of
an AC motor and provided more efficient use of the DC input voltage. Because of its
superior performance characteristics, it has found widespread application in recent years
[4]. Accordingly, many other researchers have explored various aspects of this technique
in the literature [6,15,16,17].
 4

1.3 Research Motivation

The SPWM technique is the easiest modulation scheme to understand and to implement
in software or hardware but this technique is unable to fully utilize the DC bus supply
voltage available to the voltage source inverter. This drawback led to the development of
THIPWM and SVPWM. THIPWM is a technique that adds a third-order harmonic content
to a sinusoidal reference signal thereby increasing the utilization rate of the DC bus voltage
by 15.5%.

The implementation of the conventional SVPWM is especially difficult because it

requires complicated mathematical operations. In the under-modulation region, this al-
gorithm provides 15.5% higher output voltages compared to the SPWM technique. More-
over, the utilization of the DC bus voltage can be further increased when extending into
the over-modulation region of SVPWM.

Three-phase voltage source pulse-width modulation inverters have been widely used
for DC to AC power conversion since they can produce outputs with variable voltage
magnitude and variable frequency. For example, modern power electronics controllers
have been rapidly moving toward digital implementation. Typical solutions employ mi-
crocontrollers or DSPs [18].

This thesis discusses the principles, theories, mathematical equations, and procedures
involved for the software (MATLAB/Simulink package) implementation of these tech-
niques. This thesis synthesizes and compares the main theories behind three-phase genera-
tion of SPWM, THIPWM, and SVPWM. These three techniques are used to generate their
respective output PWM signals, which are then compared based on harmonic content and
distortion using the total harmonic distortion (THD) measure of various output voltages.
 5

1.4 Thesis Organization

This thesis is comprised of seven chapters. Chapter 1 is an introduction including a

literature review, research motivation, and thesis outline. Chapters 2, 3, and 4 provide an
in-depth look into the concepts, mathematical equations, implementation, and waveforms
generated by the SPWM, THIPWM, and SVPWM techniques. Chapter 5 presents the
Simulink models and output plots obtained from the simulation of SPWM, THIPWM,
and SVPWM in the under-modulation region and over-modulation region 1. Chapter 6
compares the results of these three techniques from Chapter 5 along with a discussion
of the output harmonic spectra of various output voltages and a THD measure for each
output. Chapter 7 summarizes the thesis with conclusions and recommendations for future
research.
 6

CHAPTER 2

SINUSOIDAL PWM AND THIRD-HARMONIC-INJECTION PWM

2.1 Sinusoidal PWM

2.1.1 Sinusoidal PWM Concept

The sinusoidal pulse-width modulation (SPWM) technique produces a sinusoidal wave-

form by filtering an output pulse waveform with varying width. A high switching fre-
quency leads to a better filtered sinusoidal output waveform. The desired output voltage
is achieved by varying the frequency and amplitude of a reference or modulating volt-
age. The variations in the amplitude and frequency of the reference voltage change the
pulse-width patterns of the output voltage but keep the sinusoidal modulation.

As shown in Figure 2.1, a low-frequency sinusoidal modulating waveform is compared

with a high-frequency triangular waveform, which is called the carrier waveform. The
switching state is changed when the sine waveform intersects the triangular waveform.
The crossing positions determine the variable switching times between states.

In three-phase SPWM, a triangular voltage waveform (VT ) is compared with three

sinusoidal control voltages (Va , Vb , and Vc ), which are 120◦ out of phase with each other
and the relative levels of the waveforms are used to control the switching of the devices in
each phase leg of the inverter.

A six-step inverter is composed of six switches S1 through S6 with each phase output
connected to the middle of each inverter leg as shown in Figure 2.2. The output of the
comparators in Figure 2.1 form the control signals for the three legs of the inverter. Two
 7

Figure 2.1: Control Signal Generator for SPWM [27].

Figure 2.2: Three-Phase Sinusoidal PWM Inverter [27].

switches in each phase make up one leg and open and close in a complementary fashion.
That is, when one switch is open, the other is closed and vice-versa. The output pole
voltages Vao , Vbo , and Vco of the inverter switch between -Vdc /2 and +Vdc /2 voltage levels
where Vdc is the total DC voltage.

The peak of the sine modulating waveform is always less than the peak of the triangle
 8

carrier voltage waveform. When the sinusoidal waveform is greater than the triangular
waveform, the upper switch is turned on and the lower switch is turned off. Similarly,
when the sinusoidal waveform is less than the triangular waveform, the upper switch is
off and the lower switch is on. Depending on the switching states, either the positive or
negative half DC bus voltage is applied to each phase. The switches are controlled in pairs
((S1 , S4 ), (S3 , S6 ), and (S5 , S2 )) and the logic for the switch control signals is:

◦ S1 is ON when Va >VT S4 is ON when Va <VT

◦ S3 is ON when Vb >VT S6 is ON when Vb <VT

◦ S5 is ON when Vc >VT S2 is ON when Vc <VT .

Figure 2.3: Three-Phase Sinusoidal PWM: a). Reference Voltages (a,b,c) and Triangular
Wave b). Vao , c) Vbo , d) Vco e) Line-to-Line Voltages [27].
 9

As seen in Figure 2.3, the pulse widths depend on the intersection of the triangular and
sinusoidal waveforms. The inverter output voltages are determined as follows:

if Va >VT then Vao = 0.5Vdc

Vb >VT then Vbo = 0.5Vdc

Vc >VT then Vco = 0.5Vdc

and if Va <VT then Vao = −0.5Vdc

Vb <VT then Vbo = −0.5Vdc

Vc <VT then Vco = −0.5Vdc .

The inverter line-to-line voltages are obtained from the pole voltages as:

Vab = Vao −Vbo

Vbc = Vbo −Vco

Vca = Vco −Vao .

2.1.2 Modulation Index of Sinusoidal PWM

The Fourier series expansion of a symmetrical square wave voltage with a peak magnitude
of Vdc /2 has a fundamental of magnitude 2Vdc /π. The maximum of the output voltage
generated by the SPWM method is Vdc /2. The modulation index is defined as the ratio of
the magnitude of output voltage generated by SPWM to the fundamental peak value of the
maximum square wave. Thus, the maximum modulation index of the SPWM technique is

Vdc
VPW M 2 π
MI = = 2Vdc
= ≈ 0.7855 = 78.55%,
Vmax−sixstep 4
π
 10

where VPW M is the maximum output voltage generated by a SPWM and Vmax−sixstep is the
fundamental peak value of a square wave.

2.2 Third-Harmonic-Injection PWM

2.2.1 Concept and Calculation of Optimum Distortion

The sinusoidal PWM is the simplest modulation scheme to understand but it is unable to
fully utilize the available DC bus supply voltage. Due to this problem, the third-harmonic-
injection pulse-width modulation (THIPWM) technique was developed to improve the
inverter performance.

Following Reference [17], consider a waveform consisting of a fundamental compo-

nent with the addition of a triple-frequency term:

y = sin θ + A sin 3θ , (2.1)

where θ = ωt and A is a parameter to be optimized while keeping the maximum amplitude

of y(t) under unity. The maximum value of y(t) is found by setting its derivative with
respect to θ equal to zero. Thus,

dy
= cos θ + 3A cos 3θ = cos θ (12A cos2 θ − (9A − 1)) = 0. (2.2)
dθ

The maximum and minimum of the waveform therefore occur at

cos θ = 0 (2.3)

and
 11

 1
9A − 1 2
cos θ = (2.4)
12A

which yield, respectively,

sin θ = 1 (2.5)

and

 1
1 + 3A 2
sin θ = . (2.6)
12A

The peak value of y can be found by substituting the values obtained for sin θ in (2.5) and
(2.6) into (2.1). Using the following trigonometric identity,

sin 3θ = 3 sin θ − 4 sin3 θ , (2.7)

Equation (2.1) becomes

y = (1 + 3A) sin θ − 4A sin3 θ . (2.8)

Substituting the values in (2.5) and (2.6), for sin θ we have

yb = 1 − A (2.9)

and

 3
1 + 3A 2
yb = 8A , (2.10)
12A

where yb is the peak value of y.

 12

The optimum value for A is that value which minimizes yb and can be found by dif-
ferentiating (2.10) for yb with respect to A and equating the result to zero. Then, Equation
(2.10) becomes,

 1  
db
y 1 + 3A 2 1
= 2− = 0. (2.11)
dA 12A 3A

Thus, the two possible values of A are

1 1
A=− and A= . (2.12)
3 6

From Equation (2.9), we can see that the negative value of A makes yb greater than unity.
Therefore, the only valid solution for A is 1/6 and the required waveform is

1
y = sin θ + sin 3θ . (2.13)
6

From Equation (2.3), cos θ = 0 yields θ = π/2. Substituting the value of 1/6 for A in (2.4)
gives cos θ = 1/2, i.e., θ = π/3, 2π/3, etc. All triple harmonics pass through zero at these
values of θ . If we substitute the values of θ = nπ/3 in (2.13), then we have a maximum
√
amplitude of yb = ± 3/2 at these angles.

In Figure 2.4, it is shown that the addition of a third harmonic with a peak magnitude
of one sixth to the modulation waveform has the effect of reducing the peak value of the
√
output waveform by a factor of 3/2 without changing the amplitude of the fundamental.
It is possible to increase the amplitude of the modulating waveform by a factor K, so
that the full output voltage range of the inverter is again utilized [17]. If the modulating
waveform is expressed as

1
y = K(sin θ + sin 3θ ), (2.14)
6
 13

the required factor K for a peak value of unity should satisfy the constraint

√
1 = K 3/2 (2.15)

and, therefore,

2
K=√ . (2.16)
3

On the same figure, we see that the addition of this third harmonic produces a 15.5%
increase in the amplitude of the fundamental of the phase voltages. Figure 2.4(a) does
not have a third harmonic, only a peak value and amplitude of fundamental equal 1. The
√
peak of Figure 2.4(b) is 3/2 with one-sixth of the third harmonic added. The amplitude
of the fundamental equals 1. The peak amplitude in Figure 2.4(c) equals 1 while the
√
peak amplitude of the fundamental equals 2/ 3 with one-sixth of third harmonic added.
Injecting a third harmonic component to the fundamental component gives the following
modulating waveforms for the three-phase [17]:

 
2 1
Van = √ sin(ωt) + sin(3ωt) (2.17)
3 6
 
2 1
Vbn = √ sin(ωt − 2π/3) + sin(3ωt) (2.18)
3 6
 
2 1
Vcn = √ sin(ωt + 2π/3) + sin(3ωt) . (2.19)
3 6

The THIPWM is implemented in the same manner as the SPWM, that is, the reference
waveforms are compared with a triangular waveform. As a result, the amplitude of the
reference waveforms do not exceed the DC supply voltage Vdc /2, but the fundamental
component is higher than the supply voltage Vdc . As mentioned above, this is approx-
imately 15.5% higher in amplitude than the normal sinusoidal PWM. Consequently, it
 14

Figure 2.4: One-Phase Third-Harmonic Injection PWM [17].

provides a better utilization of the DC supply voltage.

The three reference voltages and triangular waveform of a three-phase THPWM pro-
duce the following output pole voltages Vao ,Vbo ,Vco shown in Figure 2.5.
 15

Figure 2.5: Reference Voltages (a,b,c), Triangular Waveforms (VT ), and Output Voltage
(Vao ,Vbo ,Vco ).
 16

CHAPTER 3

SPACE VECTOR PWM

3.1 Introduction

Another method for increasing the output voltage about that of the SPWM technique is
the space vector PWM (SVPWM) technique. Compared to THIPWM, the two methods
have similar results but their methods of implementation are completely different. In the
SVPWM technique, the duty cycles are computed rather than derived through comparison
as in SPWM. The SVPWM technique can increase the fundamental component by up to
27.39% that of SPWM. The fundamental voltage can be increased up to a square wave
mode where a modulation index of unity is reached.

SVPWM is accomplished by rotating a reference vector around the state diagram,

which is composed of six basic non-zero vectors forming a hexagon. A circle can be
inscribed inside the state map and corresponds to sinusoidal operation. The area inside
the inscribed circle is called the linear modulation region or under-modulation region. As
seen in Figure 3.1, the area between the inside circle and outside circle of the hexagon
is called the nonlinear modulation region or over-modulation region. The concepts in the
operation of linear and nonlinear modulation regions depend on the modulation index,
which indirectly reflects on the inverter utilization capability.
 17

Figure 3.1: Under-modulation and Over-modulation Regions in Space Vector Representa-

tion [12].

3.2 Principle of Space Vector PWM

A three-phase mathematical system can be represented by a space vector. For example,

given a set of three-phase voltages, a space vector can be defined by

→
− 2 2π 4π
V (t) = [Va (t)e j0 +Vb (t)e j 3 +Vc (t)e j 3 ], (3.1)
3

where Va (t), Vb (t), and Vc (t) are three sinusoidal voltages of the same amplitude and
frequency but with ±120o phase shifts. The space vector at any given time maintains its
magnitude. As time increases, the angle of the space vector increases, causing the vector
to rotate with a frequency equal to that of the sinusoidal waveforms. When the output
 18

voltages of a three-phase six-step inverter are converted to a space vector and plotted on
the complex plane, the corresponding space vector takes only on one of six discrete angles
as time increases. The central idea of SVWPM is to generate appropriate PWM signals so
that a vector with any desired angle can be generated.

SVPWM is a form of PWM proposed in the mid-1980s that is more efficient compared
to natural and regularly-sampled PWM. In the space-vector modulation, a three-phase
two-level inverter can be driven to eight switching states where the inverter has six active
states (1-6) and two zero states (0 and 7).

A typical two-level inverter has 6 power switches (labeled S1 to S6 ) that generate three-
phase voltage outputs. A detailed drawing of a three-phase bridge inverter is shown in
Figure 3.2. The circuit has a full-bridge topology with three inverter legs, each consisting
of two power switches. The circuit allows only positive power flow from the supply system
to the load via a full-bridge diode rectifier. Negative power flow is not possible through
the rectifier diode bridge.

The six switching power devices can be constructed using power BJTs, GTOs, IGBTs,
etc. The choice of switching devices is based on the desired operating power level,
required switching frequency, and acceptable inverter power losses. When an upper tran-
sistor is switched on, the corresponding lower transistor is switched off. Therefore, the
ON and OFF states of the upper transistors S1 , S3 , S5 can be used to determine the current
output voltage. The ON and OFF states of the lower power devices are complementary
to the upper ones. Two switches on the same leg cannot be closed or opened at the same
time.

The basic principle of SVPWM is based on the eight switch combinations of a three-
phase inverter. The switch combinations can be represented as binary codes that corre-
spond to the top switches S1 , S3 , and S5 of the inverter as shown in Figure 3.2. Each
switching circuit generates three independent pole voltages Vao , Vbo , and Vco , which are
 19

Figure 3.2: Three-Phase Bridge Inverter [11].

the inverter output voltages with respect to the mid-terminal of the DC source marked as
‘O’ on the same figure. These voltages are also called pole voltages.

The pole voltages that can be produced are either Vdc /2 or −Vdc /2. For example,
when switches S1 , S6 , and S2 are closed, the corresponding pole voltages are Vao = Vdc /2,
Vbo = −Vdc /2, and Vco = −Vdc /2. This state is denoted as (1,0,0) and, according to
→
−
Equation (3.1), may be depicted as the space vector V (t) = 23 [Vdc e j0 ]. Repeating the
same procedure, we can find the remaining active and non-active states shown in Figure
3.3.

The three-phase inverter is therefore controlled by six switches and eight inverter
configurations. The eight inverter states can be transformed into eight corresponding space
vectors. In each configuration, the vector identification uses a ‘0’ to represent the negative
phase voltage level and a ‘1’ to represent the positive phase voltage level. The relationship
between the space vector and the corresponding switching states is given in Table 3.1
and Figure 3.2. In addition, the switches in one inverter branch are in controlled in a
complementary fashion (1 if the switch is on and 0 if it is off). Therefore,
 20

Figure 3.3: Eight Switching Configuration of a Three-Phase Inverter [7].

S1 + S4 = 1;

S2 + S6 = 1;

S5 + S2 = 1.

We use orthogonal coordinates to represent the three-phase two-level inverter in the

phase diagram. There are eight possible inverter states that can generate eight space
vectors. These are given by the complex vector expressions:

 2 Vdc e j(k−1) π3

if k = 1, 2, 3, 4, 5, 6
→
− 3
Vk= (3.2)

 0 if k = 0, 7.

The entire space is divided into six equal-size sectors of 60o . Each sector is bounded
→
− →
−
by two active vectors. V 0 and V 7 are two voltage vectors with zero amplitude located at
the origin of the hexagon. The eight active and non-active state vectors are geometrically
drawn in Figure 3.4.
 21

Table 3.1: Space Vectors, Switching States, and On State Switches

Space Vetor Swtiching State On-state Switch Vector Definition

→
− →
−
V0 [000] S4,S6,S2 V0=0
→
− →
−
V1 [100] S1,S6,S2 V 1 = 32 Vdc e j0
→
− →
− π
V2 [110] S1,S3,S2 V 2 = 23 Vdc e j 3
→
− →
− 2π
V3 [010] S4,S3,S2 V 3 = 32 Vdc e j 3
→
− →
− 3π
V4 [011] S4,S3,S5 V 4 = 32 Vdc e j 3
→
− →
− 4π
V5 [001] S4,S6,S5 V 5 = 32 Vdc e j 3
→
− →
− 5π
V6 [101] S1,S6,S5 V 6 = 32 Vdc e j 3
→
− →
−
V7 [111] S1,S3,S5 V 7 = 32 Vdc e j0

→
−
The reference voltage vector V re f rotates in space at an angular velocity ω = 2π f ,
where f is the fundamental frequency of the inverter output voltage. When the reference
voltage vector passes through each sector, different sets of switches in Table 3.1 will
be turned on or off. As a result, when the reference voltage vector rotates through one
revolution in space, the inverter output varies one electrical cycle over time. The inverter
output frequency coincides with the rotating speed of the reference voltage vector. The
→
− →
− →
− →
−
zero vectors ( V 0 and V 7 ) and active vectors ( V 1 to V 6 ) do not move in space. They are
→
−
referred to as stationary vectors. Figure 3.4 shows the reference vector V re f in the first
sector.

The six active voltage space vectors are shown on the same graph with an equal
magnitude of 2Vdc /3 and a phase displacement of 60o . The inverter cannot produce a
desired reference voltage vector directly. It is possible to decompose the reference vector
into vectors that lie on two adjacent active vectors and two zero vectors, which are located
at the center of the hexagon.

The relationship between the switching variable vector [S1 , S3 , S5 ] and the line-to-line
 22

Figure 3.4: Space Vectors of Three-Phase Bridge Inverter [12].

voltage vector [Vab ,Vbc ,Vca ] is shown in Equation (3.1). When the upper or lower transistor
of a phase is ON, the switching signal of that phase is ‘1’ or ‘-1’, and when an upper or
lower transistor is OFF, then the switching signal is ‘0’. The eight combinations and the
derived output line-to-line and phase voltages in terms of the DC supply voltage are:
    
 Vab   1 −1 0   S1 
    
V  = Vdc  0 1 −1  S . (3.3)
 bc    3 
    
Vca −1 0 1 S5

Choosing a neutral load point n, we have:





 Va0 = Van +Vn0


Vb0 = Vbn +Vn0


(3.4)



 Vc0 = Vcn +Vn0



 V +V +V
an bn cn = 0.
 23

From Equations (3.3) and (3.4), the output phase voltages of the inverter depend on the
relationship between the switching variables [S1 , S3 , S5 ] and the DC voltage as follows:

    
 Van   2 −1 −1   S1 
  Vdc   
V =  −1 2 −1   S  . (3.5)
 bn 3 


 3  



Vcn −1 −1 2 S5

According to the Equations (3.3) to (3.5), the eight switching vectors, output line-to-
neutral voltage (phase voltages), and output line-to-line voltages in terms of DC link Vdc
→
− →
−
are given in Table 3.2 along with the eight inverter voltage vectors ( V 0 to V 7 ).

Table 3.2: Vectors, Switching Vectors, Phase Voltages and Line to Line Voltages as a
Function of the DC Bus Voltage Vdc .

Voltage Switching Line-to-Neutral Line-to-Line

Vectors Vectors Voltages Voltages

a b c Van Vbn Vcn Vab Vbc Vca

V0 0 0 0 0 0 0 0 0 0
2Vdc
V1 1 0 0 3 − V3dc − V3dc Vdc 0 -Vdc

V2 1 1 0 Vdc
3
Vdc
3 − 2V3dc 0 Vdc -Vdc
2Vdc
V3 0 1 0 − V3dc 3 − V3dc -Vdc Vdc 0

V4 0 1 1 − 2V3dc Vdc
3
Vdc
3 -Vdc 0 Vdc
2Vdc
V5 0 0 1 − V3dc − V3dc 3 0 -Vdc Vdc

V6 1 0 1 Vdc
3 − 2V3dc Vdc
3 Vdc -Vdc 0

V7 1 1 1 0 0 0 0 0 0

The space vector can be also represented in another reference frame with two orthog-
onal axes (α and β ). We assume that the α axis is aligned in the horizontal direction
and that the β axis is vertical. Then the abc three-phase voltage vector given in Equation
(3.1) can be transformed into a vector with αβ coordinates. The αβ vector is used to find
 24

the sector of the αβ plane in which the reference voltage vector lies. The phase voltages
corresponding to the eight combinations of switching patterns can be mapped into αβ
coordinates:
→
− 2 2π 2π
V re f = Vα + jVβ = (Va +Vb e j 3 +Vc e− j 3 ) (3.6)
3
2 2π 2π 2 2π 2π
Vα + jVβ = (Va +Vb cos( ) +Vc cos( )) + j (Vb sin( ) −Vc cos( )).
3 3 3 3 3 3

Equating real and imaginary parts, we get

2 2π 2π
Vα = (Va +Vb cos( ) +Vc cos( )) (3.7)
3 3 3

2 2π 2π
Vβ = (Vb sin( ) −Vc cos( )). (3.8)
3 3 3

In matrix form, these equations become:

 
 Va 
   
→
− Vα  2 1 −1/2 −1/2  
V re f = =   V . (3.9)

3 0 √3/2 −√3/2  b 

Vβ  
Vc

The values of Vα and Vβ in Table 3.3 are called the α and β components of the space
vector, and the last column in Table 3.3 shows the reference space vector. The voltages Vα
and Vβ become the inputs for dwelling time calculations in the space vector PWM and are
→
−
used to compute the scalar magnitude of the reference voltage | V Re f |.

3.3 Implementation Procedure of a Two-Level Space Vector PWM

The SVPWM scheme is more complicated than that of the conventional SPWM. It re-
quires the determination of a sector, calculation of vector segments, and it involves region
identification based on the modulation index and calculation of switching time durations.
 25

Table 3.3: Voltage Vectors, Switching Vector, α and β

Voltage Vector a b c Vα Vβ Vector

V0 0 0 0 0 0 0
2Vdc
V1 1 0 0 3 0 V0o
Vdc V
√dc
V2 1 1 0 3 3
V60o

V3 0 1 0 − V3dc V
√dc
3
V120o
2Vdc
V4 0 1 1 3 0 V180o

V5 0 0 1 − V3dc − V√dc3 V240o

Vdc
V6 1 0 1 3 − V√dc3 V300o

V7 1 1 1 0 0 0

Figure 3.5: Flow Diagram for SVPWM Implementation [11].

A simplified flow diagram for the implementation of the SVPWM algorithm is shown
in Figure 3.5. The procedure for implementing a two-level space vector PWM can be
summarized as follows:

→
−
1. Calculate the angle θ and reference voltage vector V re f based on the input voltage
components.

2. Calculate the modulation index and determine if it is in the over-modulation region.

→
− →
− →
−
3. Find the sector in which V re f lies, and the adjacent space vectors of V k and V k+1
 26

based on the sector angle θ .

4. Find the time intervals Ta and Tb and T0 based on Ts , and the angle θ . (For over-
modulation, find Ta0 , Tb0 and T00 is zero.)

5. Determine the modulation times for the different switching states.

3.3.1 Angle and Reference Voltage Vector

In the Space Vector PWM, the three-phase output voltage vector is represented by a
reference vector that rotates at an angular speed of ω = 2π f . The Space Vector PWM
→
−
uses the combinations of switching states to approximate the reference vector V re f . A
→
−
reference voltage vector V re f that rotates with angular speed ω in the αβ plane represents
three sinusoidal waveforms with angular frequency ω in the abc coordinate system. Each
output voltage combination in Table 3.3 corresponds to a different voltage space vector.
Three sinusoidal and balanced voltages are given by the relations:

Va (t) = Vre f cos(ωt) (3.10)

Vb (t) = Vre f cos(ωt − 2π/3) (3.11)

Vc (t) = Vre f cos(ωt + 2π/3). (3.12)

For any three-phase system with three wires and equal load impedances, we have

Va (t) +Vb (t) +Vc (t) = 0. (3.13)

The space vector with magnitude Vre f rotates in a circular direction at an angular velocity
of ω where the direction of rotation depends on the phase sequence of the voltages. If it
has a positive phase sequence, then it rotates in the counterclockwise direction. Otherwise,
it rotates in the clockwise direction with a negative phase sequence.
 27

The three-phase voltages could be described with only two components, α and β , in a
two-dimensional plane. The magnitude of each active vector is 2Vdc /3. The active vectors
are 60o apart and describe a hexagon boundary. The locus of the circle projected by the
→
− →
− → − → − → − → − → − → − → −
space reference vector V re f depends on V 0 , V 1 , V 2 , V 3 , V 4 , V 5 , V 6 , V 7 ,

→
− 2
V re f = [Va + aVb + a2Vc ] (3.14)
3

where a = e j2π/3 . The magnitude of the reference vector is:

→
− q
| V re f | = Vα2 +Vβ2 . (3.15)

The phase angle is evaluated from

Vβ
θ = tan−1 ( ), (3.16)
Vα

where θ ∈ [0,2π].

3.3.2 Modulation Index of Linear Modulation

In the linear region, the rotating reference vector always remains within the hexagon. The
largest output voltage magnitude is the radius of the largest circle that can be inscribed
within the hexagon. This means that the linear region ends when the reference voltage is
equal to the radius of the circle inscribed within the hexagon.

The fundamental component of the voltage waveform is shown in Figure 3.6. From a
Fourier analysis, the fundamental voltage magnitude is given by
 28

Z π/3 
4 Vdc 2Vdc
Z π/2
Vmax−sixstep = sin θ dθ + sin θ dθ
π 0 3 π/3 3
4Vdc h π π πi
= (− cos + 1) + (−2 cos + 2 cos
3π 3 2 3 (3.17)
4Vdc h πi
= 1 + cos
3π 3
2Vdc
= .
π

→
−
The ratio between the reference vector V re f and the fundamental peak value of the square

Figure 3.6: Fundamental of Voltage Waveform [10].

phase voltage wave (2Vdc /π) is called the modulation index. The mode of operation is
determined by the modulation index (MI). In this linear region, the MI can be expressed
as:

→
−
V re f
MI = . (3.18)
Vmax−sixstep

→
−
From the geometry of Figure 3.1, the maximum modulation index is obtained when V re f
equals the radius of the inscribed circle.

→
− 2
V re f (max) = Vdc cos(π/6). (3.19)
3
 29

Therefore,

2 π
3 Vdc cos( 6 )
MImax = 2Vdc
= 0.907. (3.20)
π

3.3.3 Sector Determination

It is necessary to know in which sector the reference output lies in order to determine the
switching time and sequence. The identification of the sector where the reference vector
is located is straightforward. The phase voltages correspond to eight switching states: six
non-zero vectors and two zero vectors at the origin. Depending on the reference voltages
Vα and Vβ , the angle of the reference vector can be used to determine the sector as per
Table 3.4.
Table 3.4: Sector Definition.

Sector Degrees

1 0 < θ ≤ 60 o

2 60 < θ ≤ 120o

3 120 < θ ≤ 180o

4 180 < θ ≤ 240o

5 240 < θ ≤ 300o

6 300 < θ ≤ 360o

3.3.4 Time Durations Ta , Tb , T0

The duty cycle computation is done for each triangular sector formed by two state vectors.
The magnitude of each switching state vector is 2Vdc /3 and the magnitude of a vector
√
to the midpoint of the hexagon line from one vertex to another is Vdc / 3. In the under-
modulation, the maximum possible modulation index is 0.907 as derived previously.
 30

The reference space vector rotates and moves through different sectors of the complex
→
−
plane as time increases. In each PWM cycle, the reference vector V re f is sampled at
a fixed input sampling frequency fs . During this time, the sector is determined and the
→
−
modulation vector V re f is mapped onto two adjacent vectors. The non-zero vectors can
be represented by

→
− 2 π
V k = Vdc e j(k−1) 3 (3.21)
3

for k=1, 2, 3, 4, 5, 6.
→
− →
−
Therefore, the non-zero vectors for V k and V k+1 become

→
− 2 π π
V k = Vdc [cos(k − 1) + j sin(k − 1) ]
3 3 3
→
− 2 π 2 kπ kπ
V k+1 = Vdc e jk 3 = Vdc [cos + j sin ].
3 3 3 3

Due to symmetry in the patterns in the six sectors, the integration

Ts T0 T0 T0 Ts
→
− →
− 4 +Ta →
− 4 +Ta +Tb →
− →
−
Z Z Z Z Z
2 4 2
V re f dt = V 0 dt + T0
V k dt + T0
V k+1 dt + T0
V 7 dt (3.22)
0 0 4 4 +Ta 4 +Ta +Tb

can be carried out for only half of the pulse-width modulation period (Ts /2). Zero voltages
are applied during the null state times:

T0 Ts
→
− →
−
Z Z
4 2
V 0 dt = T0
V 7 dt = 0. (3.23)
0 4 +Ta +Tb

Then Equation (3.22) becomes:

Ts T0 T0
→
− 4 +Ta →
− 4 +Ta +Tb →
−
Z Z Z
2
V re f dt = T0
V k dt + T0
V k+1 dt. (3.24)
0 4 4 +Ta

→
−
Thus, the product of the reference voltage vector V re f and Ts /2 equals the sum of the
 31

voltage multiplied by the time interval of the chosen space vectors. The reference voltage
→
−
vector V re f can be expressed as function of Vk and Vk+1 as

→
− Ts → − →
−
V re f = V k Ta + V k+1 Tb
2 (3.25)
→
−
V re f = Vα + jVβ ,

→
− →
−
where Ta and Tb denote the required on-time of the active-state vectors V k and V k+1
during each sample period, and k is the sector number denoting the reference location.
The calculated times Ta and Tb are applied to the switches to produce space vector PWM
switching patterns based on each sector. The switching time is arranged according to the
first half of the switching period while the other half is a reflection forming a symmetrical
pattern (see Figure 3.7). T0 and T7 are the times of the null state vectors in Figure 3.7.
→
− →
−
If V re f lies exactly in the middle between two vectors, for example between V 1 and
→
− →
− →
− →
−
V 2 with an angle of π/6, Ta for V 1 will be equal to Tb for V 2 . If V re f is closer to
→
− →
− →
− →
−
V 2 than V 1 , it means that Tb will be greater than Ta . If V re f coincides with V 2 , then
Ta will be equal to zero. If the reference keeps making a circular trajectory inside the
hexagon, then T0 is greater than zero, the output voltage will be a sinusoidal waveform in
the under-modulation region.
→
− →
−
Assuming that the reference voltage and the voltage vectors V k and V k+1 are constant
→
−
during each pulse-width modulation period Ts and splitting the reference voltage V re f into
its real and imaginary components (Vα and Vβ ) gives the following result:

       
 Vα  Ts 2Vdc   cos (k−1)π
3 cos kπ
3
 = Ta   + Tb 
   
2 3
  
(k−1)π kπ
Vβ sin 3 sin 3

or
 32

  
2 cos (k−1)π
3 cos kπ
  Ta 
3
= Vdc  . (3.26)

3

sin (k−1)π
3 sin kπ
3 Tb

These equations require computations involving trigonometric functions. From the Equa-
tion (3.26), the inverse matrix is used to calculate Ta and Tb as:
    
√ kπ kπ
 Ta  3Ts  sin 3 − cos 3   Vα 
= . (3.27)
2Vdc
  
Tb − sin (k−1)π
3 cos (k−1)π
3 Vβ

The Space Vector PWM produces the following balanced set of three-phase voltages with
→
−
magnitude V re f and angular frequency ω, given by:

Va = Vre f cos(ωt) (3.28)

2π
Vb = Vre f cos(ωt − ) (3.29)
3
4π
Vc = Vre f cos(ωt − ). (3.30)
3

The corresponding reference voltage space vector can be expressed as

→
− →
− →
−
V re f = | V re f |e jωt = | V re f |(cos ωt + j sin ωt). (3.31)

Equation (3.31) then becomes:

    
√ →
− kπ kπ
 Ta  3Ts | V re f |  sin 3 − cos 3   cos nωTs 
= . (3.32)
2Vdc
  
Tb − sin (k−1)π
3 cos (k−1)π
3 sin nωTs

The modulation index is defined in Equation (3.18) as the ratio of the desired peak funda-
mental magnitude to the maximum fundamental output in a six-step mode:
 33

→
−
| V re f | πVre f
MI = = (3.33)
Vmaxs ixstep 2Vdc

or
→
− 2
| V re f | = MI Vdc . (3.34)
π

Substituting this equation into the above equation for Ta and Tb leads to the following
duration times:
    
√ kπ kπ
 Ta  MI 3Ts  sin 3 − cos 3   cos nωTs 
 =   . (3.35)
π (k−1)π (k−1)π
Tb − sin 3 cos 3 sin nωTs

Since the sum of 2Ta and 2Tb should be less than or equal to Ts , the inverter has to stay in a
zero state for the rest of the period. The duration of the null vectors is the remaining time
in the switching period. Since

Ts = T0 + 2(Ta + Tb ) (3.36)

then the time interval for the zero voltage vectors is

T0 = Ts − 2(Ta + Tb ). (3.37)

The switching times are arranged symmetrical around the center of the switching
→
−
period as shown in Figure 3.7. The zero vector V 7 (1,1,1) is placed at the center of
→
−
the switching period, and the zero vector V 0 (0,0,0) at the start and the end, and the total
period for a zero vector is divided equally among the two zero vectors.

In the under-modulation region, as the modulation index increases, the reference volt-
age vector grows outward in magnitude. It reaches the inscribed circle of the hexagon and
T0 will reduce to zero whenever the tip of the reference voltage vector is on the hexagon. If
 34

the modulation index increases further, then T0 becomes negative and meaningless. There-
fore, the modulation index will reach a maximum of 0.907 in the linear under-modulation
region.
The calculated values of Ta and Tb in term of Ts /Vdc for all six sectors are listed in Table
3.5. The time durations of two adjacent nonzero vectors in each sector are calculated

Table 3.5: Time Intervals Ta and Tb for Each Sector

Sector θ Ta Tb
√ √
3Vα 3Vβ 3Vβ
1 0 < θ ≤ 60o 4 − 4 0Vα + 2
√ √
3Vα 3V −3Vα 3Vβ
2 60 < θ ≤ 120o 4 + 4
β
4 + 4
√ √
3V −3Vα 3Vβ
3 180 < θ ≤ 240o 0Vα + 2 β 4 − 4
√ √
−3Vα 3Vβ 3Vβ
4 120 < θ ≤ 180o 4 + 4 0Vα − 2
√ √
−3Vα 3Vβ 3Vα 3Vβ
5 240 < θ ≤ 300o 4 − 4 4 − 4
√ √
3V 3Vα 3V
6 180 < θ ≤ 240o 0Vα − 2 β 4 + 4
β

based on the magnitude and phase of the reference voltage. From Figure 3.7, a zero state
vector is applied followed with two adjacent active vectors in half of the switching period.
The next half of the switching period is symmetrical to the first half.
To generate the signals that produce the rotating vector, an equation is required to
determine the time intervals for each sector. Figure 3.8 shows the pulse patterns generated
by space vector PWM in sector 1.
→
− →
−
For example [19], when V re f is in sector 1 as shown in Figures 3.7 and 3.8, the V 1
→
−
vector is applied to the inverter during Ta interval, and consequently the vector V 2 is
→
− →
−
applied during T2 interval. The PWM period is shared between V 1 and V 2 for durations
→
− →
−
Ta and Tb , respectively, and the zero vectors V 0 and V 7 for a duration T0 . The switching
→
− → − → − → − → − → − → − → −
sequence is given by V 0 - V 1 - V 2 - V 7 - V 7 - V 2 - V 1 - V 0 during two half-sampling periods.
The generated space vector PWM waveforms are symmetrical with respect to the middle
 35

Figure 3.7: Construction of Symmetrical Pulse Pattern for Three-Phase [20].

of each PWM period. The switching frequency is the same as the sampling frequency of
the inverter. An example of a symmetric space vector PWM waveform is shown in Figure
→
−
3.7 where it is assumed that the reference voltage is in the sector formed by vector V 1 and
→
−
V 2 with angle 0 < θ ≤ 60o . In Figure 3.7, switching states are required to change from
one state to the next. The progress of switching states from the left to the right of that
figure with following steps:

→
−
1. When the circuit configuration is in the V 0 state (time interval is T0 /2), all top
switches (S1 , S3 , and S5 ) of Figure 3.2 are opened.
→
−
2. When it is in the V 1 state (with a time interval Ta ), switch S1 is closed.
→
−
3. When it is in the V 2 state (with a time interval Tb ), switch S3 is closed (S1 is still
closed).
→
−
4. When it is in the V 7 state (with a time interval T0 /2), switch S5 is closed. (S1 , and
S3 , are still closed.)
 36

After the first half of the switching period is done, the switching combination is re-
versed. All switches are closed for T0 /2 seconds before the circuit configuration is back to
→
− →
− →
−
V 2 , then V 1 , and V 0 with corresponding time intervals of Tb , Ta , and T0 /2. Following a
similar process, the switching cycles are determined for the five remaining vectors.
From Tables 3.1 and 3.2 for this example the magnitude of all the space vectors is
2Vdc /3 and the phase voltages are Van = 2Vdc /3,Vbn = −Vdc /3, Vcn = −Vdc /3 and the
line-to-line voltages are Vab = Vdc , Vbc = 0, Vca = −Vdc .

Figure 3.8: Vre f Falls into Sector 1 [21].

3.3.5 Determination of the Switching Times for Each Transistor Switch (S1 - S6 )

It is necessary to arrange the switching sequence so that the switching frequency of each
inverter leg is minimized. There are many switching patterns that can be used to implement
SVPWM. To minimize the switching losses, only two adjacent active vectors and two zero
vectors are used in a sector [15,20,21]. To meet this optimal condition, each switching
period starts with one zero vector and end with another zero vector during the sampling
time Ts . This rule applies normally to three-phase inverters as a switching sequence.
 37

Therefore, the switching cycle of the output voltage is double the sampling time, and
the two output voltage waveforms become symmetrical during Ts . Table 3.6 presents a
symmetric switching sequence.
Referring to this table, the binary representations of two adjacent basic vectors differ in
only one bit, so that only one of the upper transistors switches is closed when the switching
pattern moves from one vector to an adjacent one. The two vectors are time-weighted in a
sample period Ts to produce the desired output voltage.

Table 3.6: Seven-Segment Switching Sequence

Sector Switching Segment

1 2 3 4 5 6 7

1 ~V0 ,[000] ~V1 ,[100] ~V2 ,[110] ~V7 ,[111] ~V2 ,[110] ~V1 ,[100] ~V0 , [000]

2 ~V0 ,[000] ~V3 ,[010] ~V2 ,[110] ~V7 ,[111] ~V2 ,[110] ~V3 ,[010] ~V0 , [000]

3 ~V0 ,[000] ~V3 ,[010] ~V4 ,[011] ~V7 ,[111] ~V4 ,[011] ~V3 ,[010] ~V0 , [000]

4 ~V0 ,[000] ~V5 ,[001] ~V4 ,[011] ~V7 ,[111] ~V4 ,[011] ~V5 ,[001] ~V0 , [000]

5 ~V0 ,[000] ~V5 ,[001] ~V6 ,[101] ~V7 ,[111] ~V6 ,[101] ~V5 ,[001] ~V0 , [000]

6 ~V0 ,[000] ~V1 ,[100] ~V6 ,[101] ~V7 ,[111] ~V6 ,[101] ~V1 ,[100] ~V0 , [000]

3.3.6 Types of Different Schemes

There are two modes of operation available for the PWM waveform: symmetric and
asymmetric PWM. The pulse of an asymmetric edge aligned signal always has the same
side aligned with one end of each PWM period. On the other hand, the pulse of symmetric
signals is always symmetric with respect to the center of each PWM period. The sym-
metrical PWM signal is often preferred because it has been shown to have the lowest total
harmonic distortion (THD), and has been implemented in [22,23,24]. Output patterns for
each sector are based on a symmetrical sequence. There are different schemes in space
 38

vector PWM and they are based on their repeating duty distribution. The seven-segment
technique is studied in this thesis and will be referred to as the symmetric technique.
Based on the equations for Ta , Tb , T0 , T7 , and according to the principle of symmetrical
PWM, the switching sequence in Table 3.7 is shown for the upper and lower switches.
Figure 3.10 shows the switching patterns of all six sectors in the circle. As shown in
the same figure, the space vector for a three-phase voltage source inverter is divided into
six sectors based on six fundamental vectors. Any voltage vector in this vector space can
be synthesized using two adjacent vectors. One switching period is depicted in the same
figure. In sector 1, for example, switching is achieved by applying a zero state vector
followed by two adjacent active state vectors in a half switching period. The next half of
the switching period is the mirror image of the first half.
In order to reduce the switching loss of the power components of the inverter, it
is required that at each time only one bridge arm is switched. After re-organizing the
switching sequences, a scheme with center-aligned pulses is obtained as shown in Figure
3.9.
The switching pulse patterns of six different sectors in Figure 3.11 are shown for the
upper and lower switches of a three-phase inverter. It is obvious that in the odd sector
the active state sequence is in ascending-descending order; whereas, it is in a descending-
ascending order in an even sector. For example:

→
− → − → −
1. In an odd sector 1, the state sequence of space vectors is in the order V 0 , V 1 , V 2 ,
→
− → − → − → − → −
V 7, V 7, V 2, V 1, V 0.
→
− → − → − → − → − → −
2. In an even sector 2, the state sequence of space vectors is: V 0 , V 3 , V 2 , V 7 , V 7 , V 2 ,
→
− → −
V 3, V 0.

Following the same procedure, we have the switching sequence summarized in Table 3.8
for all six sectors.
Figure 3.9: Switching Patterns in the Six Sectors [20].
39
 40

Figure 3.10: Switching Patterns of Six Sectors in Circle [20].

 41

Table 3.7: Switching Pulse Pattern for the Three Phase for Each Sector

Sector Upper Switches: S1, S3, S5 Lower Switches: S4, S6, S2

1 S1 = Ta + Tb + T0 /2 S4 = T0 /2

S3 = Tb + T0 /2 S6 = Ta + T0 /2

S5 = T0 /2 S2 = Ta + Tb + T0 /2

2 S1 = Ta + T0 /2 S4 = Tb + T0 /2

S3 = Ta + Tb + T0 /2 S6 = T0 /2

S5 = T0 /2 S2 = Ta + Tb + T0 /2

3 S1 = T0 /2 S4 = Ta + Tb + T0 /2

S3 = Ta + Tb + T0 /2 S6 = T0 /2

S5 = Tb + T0 /2 S2 = Ta + T0 /2

4 S1 = T0 /2 S4 = Ta + Tb + T0 /2

S3 = Ta + T0 /2 S6 = Tb + T0 /2

S5 = Ta + Tb + T0 /2 S2 = T0 /2

5 S1 = Tb + T0 /2 S4 = Ta + T0 /2

S3 = T0 /2 S6 = Ta + Tb + T0 /2

S5 = Ta + Tb + T0 /2 S2 = T0 /2

6 S1 = Ta + Tb + T0 /2 S4 = T0 /2

S3 = T0 /2 S6 = Ta + Tb + T0 /2

S5 = Ta + T0 /2 S2 = Tb + T0 /2
 42

Table 3.8: Switching Sequence for Three-Phase PWM Technique

Sector Switching Sequence of the Three Phase Modulation

1 ~V0 -~V1 -~V2 -~V7 -~V2 -~V1 -~V0

2 ~V0 -~V3 -~V2 -~V7 -~V2 -~V3 -~V0

3 ~V0 -~V3 -~V4 -~V7 -~V4 -~V3 -~V0

4 ~V0 -~V5 -~V4 -~V7 -~V4 -~V5 -~V0

5 ~V0 -~V5 -~V6 -~V7 -~V6 -~V5 -~V0

6 ~V0 -~V1 -~V6 -~V7 -~V6 -~V1 -~V0

 43

Figure 3.11: Switching Sequence of all Six Sectors [4].

 44

CHAPTER 4

SPACE VECTOR PWM IN OVER-MODULATION REGION

4.1 Introduction

Full utilization of the DC bus voltage is important for achieving, for example, maximum
output torque under all operating conditions in AC machine drive applications. Over-
modulation aims to extend the linear operation region of the SVPWM, which leads to a
better utilization of the DC bus voltage and enhances the power utilization of the voltage
source inverter. Several approaches to over-modulation have been proposed to extend the
linear region of SVPWM over the years. It is an advanced method and possibly the best
technique for variable frequency drive applications.

For SPWM, the highest possible peak phase fundamental voltage is 0.5Vdc . With
SVPWM, the peak phase fundamental voltage can be as high as 0.577Vdc during under-
modulation when the reference vector makes a circular trajectory and it can be higher in
the nonlinear modulation region when the desired trajectory partly passes outside of the
hexagon.

The modulation index of the over-modulation region ranges from 0.907 up to 1 [11].
The over-modulation range can be considered as one region or it can be divided into two
regions (see Figure 4.1). In this thesis, the over-modulation region is considered as two
regions with two modes of operation depending on the modulation index values.

• In mode 1, the actual voltage vector keeps the angular speed of a modified reference
vector constant, but its amplitude changes over time.
 45

Figure 4.1: The Two Over-Modulation Regions in Space Vector Representation [12].

• In mode 2, both the amplitude and angle of the modified reference vector are varied.

4.2 Over-Modulation Mode 1

The over-modulation region starts when the reference voltage exceeds the hexagon bound-
ary, and the MI is larger than 0.907. The boundary between the under-modulation zone
and over-modulation region 1 starts when MI=0.907 and the boundary between the over-
modulation region 1 and the over-modulation region 2 starts when MI=0.952. In region
1, the crossover angle (αr ) is measured from a hexagon vertex to the intersection of the
compensated voltage vector trajectory with a hexagon side (see Figure 4.3). The crossover
angle decreases as the modulation index is increased until at the limit of region 1. It
depends on the value of the modulation index and varies between π/6 at the beginning
of region 1 when MI=0.907 and 0o at the end of region 1 when MI =0.952 as shown in
Figure 4.2. When the crossover angle equals zero degrees, the reference vector is fully on
 46

the hexagon. The fundamental peak value generated in this way voltage is about 95% of
the peak voltage of the square wave. This gives a modulation index of 0.952 [23].
Cross Angle vs. Modulation Index
30

Cross Angle (degree) 25

20

15

10

0
0.91 0.915 0.92 0.925 0.93 0.935 0.94 0.945 0.95 0.955
Modulation Index

Figure 4.2: Crossover Angle vs. Modulation Index in the Mode 1.

In mode 1, the amplitude of the desired voltage is modified to fit inside a hexagon
→
−
while the desired phase angle is not modified. It starts when the reference voltage V re f
crosses the hexagon at two points in each sector. Figure 4.3 shows two crossing points at
B and C. There is a loss of fundamental voltage in the region where the reference vector
exceeds the hexagon boundary of B-C. To compensate for that loss, the output voltage
→
−
must track the modified reference voltage V re f , where it must change to a larger radius.
→
−
This leads to a new reference voltage ( V 0re f ), which crosses the hexagon at an angle αr .
The angle of the new reference vector is transmitted without any change.

The characteristic of the over-modulation mode 1 is that the voltage space vector moves
partially on the hexagon (along B-C) and partially on the circle providing a continuous op-
eration. The output voltage waveform is approximated by linear segments on the hexagon
trajectory and by sinusoidal segments on the circular trajectory.

When the reference voltage vector is within the hexagon, near the outer corners of a
sector on curve length A-B in the same figure, it is on a circular trajectory with switching
times of Ta , Tb , and T0 as in the under-modulation mode, but with a modified reference
voltage.
During the nonlinear modulation, only the active vectors are taken into account while
 47

Figure 4.3: Over-Modulation Mode Region 1 [12].

→
−
the zero states are neglected. So, when V re f is on the hexagon, T00 =0, the switching times
must be calculated using the modified equations [10]:

 
Ts Ta
Ta0 = (4.1)
2 T + Tb
  a
Ts Tb
Tb0 = . (4.2)
2 Ta + Tb

When the original reference trajectory passes outside of the hexagon, the tip of the
reference vector is located in the area between the inside circle and the outside circle of
the hexagon. The Ta and Tb equations in linear space vector PWM yield a negative and
meaningless duration for the zero vectors. Whether the trajectory is inside the hexagon or
not can be easily recognized using the following explanation.
 48

The magnitude of the new reference voltage is a function of the phase angle θ . If
the modulation index is above 0.907, the desired trajectory passes outside of the hexagon
and the interval time T0 is negative on the zero state vectors. In this case, the converter
exceeds its linear region of operation and enters the over-modulation region. Therefore,
the synthesized waveform becomes distorted.

If T00 is equal to zero, Ta0 and Tb0 should be reduced with an equal ratio so that their
sum should equal Ts . We need to use Equations (4.1) and (4.2) for Ta0 and Tb0 in the region
where the modified reference trajectory moves along the hexagon. The durations of the
active vectors during one switching period Ts are calculated using

√
3 cos(α − (k − 1) π3 ) − sin(α − (k − 1) π3 )

Ts
Ta0 = √ (4.3)
2 3 cos(α − (k − 1) π3 ) + sin(α − (k − 1) π3 )
Tb0 = Ts − Ta0 (4.4)

T00 = 0, (4.5)

where k is the sector number. Equation (4.3) to (4.5) are used to calculate the durations of
the active vectors for each sector. Table 5.1 shows those durations for each sector. Using

Table 4.1: The Time Intervals Ta0 and Tb0 for Each Sector

Sector θ Ta0 Tb0

√
3Vα −Vβ 2Vβ
1 0 < θ ≤ 600 √
3Vα +Vβ
√
3Vα +Vβ
√ √
3Vα +Vβ − 3Vα +Vβ
2 60 < θ ≤ 1200 2Vβ 2Vβ
√
2V − 3Vα −Vβ
3 120 < θ ≤ 1800 √ β
− 3Vα +Vβ
√
− 3Vα +Vβ
√
− 3Vα +Vβ −2Vβ
4 180 < θ ≤ 2400 √
− 3Vα −Vβ
√
− 3Vα −Vβ
√ √
− 3Vα −Vβ 3Vα −Vβ
5 240 < θ ≤ 3000 −2Vβ −2Vβ
√
−2Vβ 3Vα +Vβ
6 300 < θ ≤ 3600 √
3Vα −Vβ
√
3Vα −Vβ
 49

Equation (3.22) in the under-modulation chapter, we can find the new reference voltage
space vector in the over-modulation region of mode 1:

→
− Ts → − →
−
V re f = V k .Ta + V k+1 .Tb . (4.6)
2

Then,

→
− 0 Ts →
− 0 → −
V re f = V k Ta + V k+1 Tb0
2  
→
−0 2 →− →
−  Ts

V re f = V k Ta + V k+1 Tb (4.7)
Ts 2(Ta + Tb )
 
→
−0 Ts →
−
V re f = V re f .
2(Ta + Tb )

Referring to Figure 4.4, consider the trajectory of three voltage vectors rotating in a
complex plane (left-hand side) and the phase voltage waveform of an actual voltage refer-
ence vector transformed in a time domain (right-hand side), which is actually modulated
by the inverter [14]. The phase voltage waveform is divided into four segments and the

Figure 4.4: Trajectory of Reference Voltage Vector and Phase Voltage Waveform in Mode
1 [14].
 50

voltage equations in each segment are expressed as follows [14]:

Vdc π
f1 = √ tan θ , for ( − αr ) > θ ≥ 0
3 6
Vdc π π
f2 = √ sin θ , for ( + αr ) > θ ≥ ( − αr )
3 cos( π6 − αr ) 6 6
(4.8)
Vdc π π
f3 = √ sin θ , for ( − αr ) > θ ≥ ( + αr )
π
3 cos( 3 − αr ) 2 6
Vdc π π
f4 = √ sin θ , for ( ) > θ ≥ ( − αr ),
π
3 cos( 6 − αr ) 2 2

where θ = ωt and ω is the angular velocity of the fundamental voltage reference vector.

The modification of the reference voltage vector is the same in all of the six sectors of
vector diagram, the relationship between the crossover angle (αr ) and the new reference
voltage is given by:

√
π (Vdc / 3)
cos( − αr ) = →− . (4.9)
6 ( V 0re f )

By expanding those four segment equations with a Fourier series, the fundamental com-
ponent of the phase voltage can be expressed as:

π π π π
6 −θ 6 +θ 2 −θ
Z 
4
Z Z Z
2
F(αr ) = f1 sin θ dθ + f2 sin θ dθ + f3 sin θ dθ .
f4 sin θ dθ (4.10)
π 0 π
6 −θ
π
6 +θ
π
2 −θ

This integral ranges from an angle of 0 to π/2 as shown in the complex plane (right part)
of the same figure. We can obtain the value of Fourier series F(αr ) with respect to the
angle αr . Since F(αr ) is the peak of the reference voltage,

2
F(αr ) = Vdc × MI (4.11)
π

and according to the definition of the modulation index,

 51

F(αr )
MI = 2
. (4.12)
π Vdc

The Fourier series F(αr ) can be generated from a MATLAB script. The relationship
between MI and the angle αr gives us the plot shown in Figure 4.2, which shows the
correlation between MI and the crossover angle αr .

4.3 Over-Modulation Mode 2

At the end of mode 1, the component of the reference voltage changes to a piecewise
linear waveform. When the modulation index is higher than 0.952, the second region of
the over-modulation region algorithm is entered. Both the reference magnitude and the
phase angle have to be changed compared to the linear region. The modified reference
vector is held at a vertex of the hexagon in every sector for the rest of the switching period.
The angle for which the active switching state vector is held constant is called the holding
angle αh . Figure 4.5 shows the holding angle in region 2 as a function of the modulation
index.

Figure 4.5: Holding Angle vs. Modulation Index in Mode 2 [11].

At the end of Mode 2, the linear segments vanish giving a six-step or square-wave
operation where the modified vector is held at the hexagon corner for 60o , that is, αh =30o
 52

[11]. An expression for the modified angle αh in Mode 2 can be given as


0 if : 0 ≤ α ≤ αh





αm = α−αh
∗ π6 if : αh < α < (π/3 − αh ) (4.13)
 π/6−αh


π
if: (π/3 − αh ) ≤ α ≤ π/3,


3

αm is the modified angle of αh .

The holding angle increases from 0 to π/6 as modulation index goes from 0.955 to
1. The relation between the holding angle and the modulation index is nonlinear [16].
Therefore, the actual trajectory is modified so that the output fundamental voltage matches
that of the reference voltage. The operation in this region is characterized by partly holding
the modified vector at a hexagon corner for a holding angle αh , and partly by tracking the
hexagon sides in every sector. During the holding angle, the magnitude of Van remains
constant, whereas during hexagon tracking, the voltage changes almost linearly. In Figure
4.6, the trajectory shows five steps following by four angle ranges.

Figure 4.6: Angular Displacement of Reference and Actual Voltage Vectors in Mode 2
[12].

1. If the angle falls between zero and the holding angle (αh ), the modified reference
voltage vector is held at a vertex, while the fundamental voltage vector is continu-
 53

ously rotating.

2. If the angle range is from the holding angle to (π/6-αh ), the modified reference
vector moves along the hexagon, but lags the fundamental voltage vector.

3. When α equals π/6 exactly, the modified reference vector catches up with the
reference vector.

4. When α < ( π3 −αh ), the modified reference vector speeds past it until α = π/3−αh .

5. In the last angle range (π/3 > α ≥ π/3 − αh ), the modified reference voltage vector
has just arrived at the other vertex. It remains there until the reference voltage vector
has caught up with the modified reference voltage vector.

The modulation will gradually change to a six-step mode for αh equal zero and generate a
square wave when the modulation index gets close to 1. The six-step mode is characterized
by a selection of the switching vector for one-sixth of the fundamental period. In this case,
the maximum possible inverter output voltage is generated. Mode 2 ends when the holding
angle is π/6.
The voltage equations in the four segments (see Figure 4.7) are expressed as:

Vdc π
f1 = √ . tan α p , for 0 ≤ θ < ( − αh )
3 6
Vdc π π
f2 = , for ( − αh ) ≤ θ < ( + αh )
3 6 6
Vdc sin(α p0 ) (4.14)
π π
f3 = √ , for ( + αh ) ≤ θ < ( − αh )
π 0
3. cos( 3 − α p ) 6 2
2Vdc π π
f4 = , for ( − αh ) ≤ θ < )
3 2 2

where

6 θ π θ − π6 + αh θ 0 + αh
α p0 = αp − = − = = (4.15)
π 1 − 6απ h 6 1 − 6απ h 1 − 6απ h

Equations (4.10) to (4.12) are then used to find the values of modulation index and Fourier
 54

series F(αh ) with respect to the holding angle. A relationship between MI and holding
angle gives us the plot shown in Figure 4.5 [14]. The modulation index increases from 0
to π/6 as the holding angle increases from 0.955 to 1.

Figure 4.7: Trajectory of Reference Voltage Vector and Phase Voltage Waveform in Mode
2.
 55

CHAPTER 5

SIMULATION RESULTS

5.1 Introduction

This chapter discusses the MATLAB/Simulink software implementation of SPWM, THIPWM,

and SVPWM in the under-modulation region and the over-modulation region 1. We
describe Simulink models built on the corresponding equations in previous chapters. In
addition to the simulation results, this chapter includes detailed subsystems of the Simulink
models as well as an explanation of the role of every subsystem. Low-pass filters are
required at the outputs in order to filter out the PWM waveforms and visualize the funda-
mental results. Our simulation analysis does not include the programming of dead time
for the switching of complementary switches in an inverter leg.

Simulation results are presented in three groups based on three simulation models.
The first group (Figures 5.1-5.6) shows simulation results for SPWM and THIPWM. The
second group (Figures 5.7-5.18) presents the under-modulation results for SVPWM. The
third group shows the simulation results of region 1 in the over-modulation SVPWM
(Figures 5.19-5.23).

The simulation of the over-modulation region 2 of SVPWM has been excluded from
this thesis. A different method needs to be implemented that ensures that the output voltage
varies continuously in this mode.
 56

5.2 Sinusoidal PWM

SPWM is very popular and easy to implement using comparators. The SPWM simulink
system model is built in Figure 5.1. It has the following blocks: (1) Sinusoidal wave
generators, (2) High-frequency triangular wave generator, and (3) comparators.

Figure 5.1: SPWM System Model.

The SPWM technique treats each modulating voltage as a separate entity that is com-
pared to the common carrier triangular waveform. A three-phase voltage set (Va , Vb ,
and Vc ) of variable amplitude is compared in three separate comparators with a common
triangular carrier waveform of fixed amplitude as shown in the same figure. The output
(Vao , Vbo , and Vco ) of the comparators form the control signals for the three legs of the
inverter composed of the switch pairs (S1 , S4 ), (S3 , S6 ), and (S5 , S2 ), respectively. From
these switching signals and the DC bus voltage, PWM phase-to-neutral voltages (Van ,
Vbn , and Vcn ) are obtained. In the simulink model, the simulation is performed under the
following conditions:
 57

Vdc = 300 V

Switching frequency = 1800 Hz

Inverter frequency = 60 Hz

Vre f = 150 V.

Figure 5.2: Vao , Vbo , and Vco of SPWM Waveforms.

Figure 5.3 shows the simulated pole voltages Vao , Vbo , and Vco . The relationship
between the line-to-neutral voltages and the switching states (S1 , S3 , and S5 ) in a balanced
three-phase load are:

Vdc
Van = (2S1 − S3 − S5 )
3
Vdc
Vbn = (−S1 + 2S3 − S5 )
3
Vdc
Vcn = (−S1 − S3 + 2S5 ).
3
 58

Figure 5.3: Vao , Vbo , and Vco of SPWM after Filtering.

To visualize the actual results, filtering of the PWM waveforms is required. The line-
to-neutral voltages Van , Vbn , and Vcn are shown after they have been passed through a
low-pass filter. As expected, magnitudes of the phase voltages (Van , Vbn , and Vcn ) are about
0.5Vdc . The output voltage waveforms show that the higher the switching frequencies, the
smoother the output voltage waveforms, as expected.

5.3 THIPWM

The THIPWM simulink system model is the same as that of the SPWM system except for
the modulating waveform voltages, which are generated by injected the third-harmonic
components as follows:
 59

Figure 5.4: Neutral Voltages Van , Vbn , and Vcn of SPWM after Filtering.

 
2 1
Va = √ sin(ωt) + sin(3ωt)
3 6
 
2 1
Vb = √ sin(ωt − 2π/3) + sin(3ωt)
3 6
 
2 1
Vc = √ sin(ωt + 2π/3) + sin(3ωt) .
3 6

Figure 5.5 shows the line-to-neutral voltages and Figure 5.6 shows the line-to-line
p
voltage of THIPWM. As expected, the magnitude of the phase voltage is about 300/ (3)
(Volts) and 300 (Volts) for line-to-line voltage, and the results show that THIPWM im-
proves the fundamental voltages compared to SPWM.

5.4 Under-Modulation of Space Vector PWM

A two-level three-phase inverter has eight possible inverter switching states that can gen-
→
− →
− →
−
erate eight space vectors: six non-zero vectors ( V 1 to V 6 ) and two zero vectors ( V 0 and
→
−
V 7 ). These vectors are applied during the switching times Ta , Tb , and T0 . According to
the time duration equations in the chapter on under-modulation SVPWM, the average pole
voltage vector over one PWM period can be averaged using two adjacent vectors and the
null vectors.
 60

200

150

100 Van
Vbn
Vcn
50

−50

−100

−150

−200
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018

Figure 5.5: Neutral Voltages Van , Vbn , and Vcn of THIPWM after Filtering.

300

200

100

−100
Vab
Vbc
Vca
−200

−300
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018

Figure 5.6: Neutral Voltages Vab , Vbc , and Vca of THIPWM after Filtering.
 61

To generate a rotating space vector with constant amplitude, the reference voltage
vector must be limited to the inscribed circle inside the hexagon. The simulation model
used to verify the under-modulation SVPWM scheme is shown in Figure 5.7 and it has
seven main subsystems or blocks, which are shown in detail in Figures 5.8-5.13.

1. The first block Figure 5.8 is used to generate three-phase sinusoidal input voltages
with variable frequency, amplitude, direction, and DC bus voltage. The three signals
are delayed by 120o from each other.

2. The three-phase abc voltages are then converted to two-phase αβ voltages given in
the second block as:

2 1 1
Vα = Va − Vb − Vc
3 3 3
1 1
Vβ = √ Vb − √ Vc .
3 3

It is necessary to know in which sector the reference output is in order to determine

the switching time. The reference voltages Vα and Vβ are utilized to determine the
sector of the vectors from 1 to 6 as shown in Figure 5.9. These values are the inputs
to the third block.

3. Equations (5.1) and (5.2) in the third block calculates the phase angle,

Vβ
θ = tan−1 ( ) (5.1)
Vα
θ ∈ [0,2π] (5.2)

and using Table 3.5, it can be used to identify the sector of the reference voltage as
shown in Figure 5.10.
The modulation index is entered in the first block. It is the ratio of the amplitude of
the output sinusoidal voltage to the maximum fundamental voltage.
Figure 5.7: Linear Modulation of SVPWM Simulation System. Diagrams for Blocks 1 to 7 are Shown in Figures 5.8 to 5.13.
62
 63

Figure 5.8: Three-Phase Input Sinusoidal Voltages and Modulation Index are Detail for
Block 1 in Figure 5.7.

4. In the fourth block, the switching time calculator is used to calculate the timing of the
reference voltage vector. The inputs are the sector in which the voltage vector lies,
the modulation index, the sampling time period of switching frequency, and cos ωt,
and sin ωt. The duration time of the active and zero vectors are then calculated using
    
√ kπ kπ
 Ta  MI 3Ts  sin 3 − cos 3   cos nωTs 
 =   .
Tb π − sin (k−1)π
3 cos (k−1)π
3 sin nωTs
 64

Figure 5.9: α, β Voltages and Modulation Index are Detail for Block 2 in Figure 5.7.

Figure 5.10: Detail for Block 3 in Figure 5.7.

In Figure 5.11, the time for the active and zeros vectors are arranged in the switching
pattern sequence shown in Table 3.7. In the same block (Figure 5.11), we also have
Sample & Hold blocks after sector Ta and Tb . The purpose of these blocks is to hold
the values of Ta and Tb fixed during each TPW M period [25].

5. The fifth block is a triangular generator used to produce a unit triangular waveform
 65

Figure 5.11: Switching Time Calculation is Detail for Block 4 in Figure 5.7.

at the PWM switching frequency.

6. The gate timing signals from the fourth block are compared with the triangular
generator of fifth block, producing the outputs for the six switches of the inverter
(Figure 5.12).

Figure 5.12: Timing Signals and Triangular Waveform Details for Block 6 in Figure 5.7.
 66

7. The seventh block is built to simulate a voltage source inverter (Figure 5.13).
As seen in Figure 5.14, the inputs for the sixth block are the output time signals from

Figure 5.13: Inverter Output Signals to Neutral Voltages is Detail for Block 7 in Figure
5.7.

the fourth block and the triangular waveform is output from the fifth block. Figure
5.15 shows the pole voltages Vao , Vbo , and Vco for Linear Modulation SVPWM. Due
to the relationship between the DC bus voltage and the switching states of the output
from the sixth block, we obtain the following PWM phase-to-neutral voltages:

Vdc
Van = (2S1 − S3 − S5 )
3
Vdc
Vbn = (−S1 + 2S3 − S5 )
3
Vdc
Vcn = (−S1 − S3 + 2S5 )
3

The simulation of the under-modulation of SVPWM is performed under the follow-

ing conditions:
 67

−5
x 10

Triangular Wave
Time Signal of phase A
6
Time Signal of phase B
Time Signal of phase C

−1
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018

Figure 5.14: Filtered Timing Signals of Three-Phase and Triangular Waveforms in

SVPWM.

Vdc = 300 V

Sampling time period = 1/1800 Sec

Inverter frequency = 60 Hz
300
Vre f = √ V.
3

Figure 5.16 shows the output voltage waveforms obtained from the SVPWM strat-
egy. The modulation waveform of the seven-segment SVPWM has a saddle shape
when passed through a low-pass filter (Figure 5.16) and its line-to-neutral voltages
are sine waveforms because of the under-modulation PWM region (Figure 5.17).
√
The maximum neutral voltages Van , Vbn , and Vcn in Figure 5.18 is Vdc / 3 (173.2
Volts).
 68

Figure 5.15: Vao , Vbo , and Vco for Linear Modulation SVPWM.

5.5 Mode 1 in Over-Modulation of SVPWM

In the over-modulation region 1, additional calculations are required to compute the ref-
erence space vector. As mentioned before, there is a loss of fundamental voltage in the
region where the reference vector exceeds the hexagon boundary. To compensate for this
loss, the reference vector amplitude is increased in the region where the reference vector
is inside the hexagon boundary. The magnitude of the reference vector is changed from
the reference voltage vector to a modified reference voltage vector. A modified reference
voltage trajectory proceeds partly on the hexagon and partly on the circle. When it remains
on the circular part, the switching time of Ta , Tb , and T0 are similar to the equations used
 69

Figure 5.16: Vao , Vbo and Vco for Linear Modulation SVPWM after Filtering.

for under-modulation. When the reference voltage vector passes outside of the hexagon,
the value of T0 is negative and meaningless. This problem is overcome by rescaling the
duration times. Thus, T00 =0, and Ta0 and Tb0 are the new time intervals shown in Table 4.1.

The simulation system model of the over-modulation region 1 of SVPWM is the same
as that of the under-modulation SVPWM system except for the lookup table in first block
and an added fourth block. The lookup table in the first block is the relationship between
MI and the crossover angle, which is shown in Figure 4.2. For an added fourth block, the
switches will be changed based on the duration T0 of the zero state vectors.
 70

300
Vab
Vbc
Vca
200

100

−100

−200

−300
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018

Figure 5.17: Line to Line Voltages of Linear-Modulation Region.

Van, Vbn and Vcn of Linear Modulation of Space Vector PWM

200
Van
Vbn
150 Vcn

100

50

−50

−100

−150

−200
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018

Figure 5.18: Neutral Voltages Van , Vbn , and Vcn of Linear Modulation SVPWM after
Filtering.

• When T0 is greater than zero, the duration times T0 , Ta , and Tb shown in Table 3.5
are used.

• When T0 is less than zero, the active duration times become Ta0 and Tb0 , as shown in
 71

Table 4.1.

Figure 5.19 shows the lookup table in the first block. Figure 5.20 shows the additional

Figure 5.19: Detail of the Lookup Table in the First Block of SPVWM Over-Modulation
Region 1.

block, which is a judgment condition for the over-modulation based on the sign of T0 . In
red circle, if T0 is negative, then Ta0 and Tb0 are used in the blue block.

As mentioned before, in the over-modulation region 1, the magnitude of the reference

voltage vector is changed while the angle remains unchanged. This region ends when the
reference voltage is traveling along the sides of the hexagon. The simulation results are
 72

Figure 5.20: Detail of the Fourth Block of Space Vector PWM in Region 1.

presented to verify the effectiveness of this analysis. Figure 5.21 shows the output voltages
for region 1 of the over-modulation.
Compared to Figure 5.16, the phase voltage waveforms of Figure 5.21 do not have the
saddle waveform shape seen in the under-modulation SVPWM, but they stay at the limit
of the hexagon.
Figure 5.22 shows the neutral voltages Van , Vbn , and Vcn in region 1 of the over-
modulation SVPWM after filtering. The voltage waveforms are no longer pure sine wave-
forms. Since the desired trajectory passes outside of the hexagon, the converter exceeds
its linear region of operation and enters the over-modulation region. Therefore, the syn-
thesized waveforms become distorted.
The low pass filter removes the high-frequency switching components, leaving the
triple order harmonics in the phase voltage waveforms. These harmonics do not affect
the line-to-line voltages [1]. As shown in Figure 5.23, the waveforms of the line-to-line
voltages Vab , Vbc , and Vca are less distorted compared to the neutral voltages in Figure 5.22.
 73

Figure 5.21: Vao , Vbo , and Vco of Over-Modulation Region 1 before and after Filtering.
 74

Figure 5.22: Line-to-Neutral of Over-Modulation Region 1 after Filtering.

400 Vab
Vca
Vbc
300

200

100

−100

−200

−300

−400
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018

Figure 5.23: Line-to-Line Voltages of Over-Modulation Region 1 after Filtering.

 75

CHAPTER 6

COMPARISON BETWEEN SPWM, THIPWM, AND SVPWM

TECHNIQUES

6.1 Simulation Results

The simulation studies of Chapter 5 confirmed that the THIPWM and SVPWM techniques
have a better DC bus voltage utilization than SPWM. As seen in Figure 6.1, the smaller
circle represents the operating region of the SPWM technique and the larger inscribed
circle represents the operating region of the SVPWM technique in the under-modulation
→
− →
−
region. In SVPWM, the length of each discrete space vector V 1 through V 6 is 2Vdc /3.
Each side of the hexagon midpoint is tangential to the inscribed circle. The largest possible
magnitude of the reference voltage can be calculated as

√
π 2Vdc 3 Vdc
MO = OL × cos( ) = × =√ . (6.1)
6 3 2 3

Since NM is perpendicular to OL, then

√
π Vdc 3 Vdc
ON = OM × cos( ) = √ × = . (6.2)
6 3 2 2

Thus, for SPWM, the smaller circle has a maximum magnitude of Vdc /2. It was al-
ready mentioned in Chapter 3 that the linear under-modulation region has a modulation
∗ = (2/3)V ×
index that approaches 90.7% for a maximum output fundamental of Vmax dc
 76

√
cos(π/6) = Vdc / 3. The line-to-line voltage magnitude in linear SVPWM is thus equal
to Vdc .

In SPWM, the maximum modulation index is 78.55%, the maximum output funda-
√
mental is 0.5 Vdc , and the maximum amplitude of the line-to-line voltage is 3Vdc /2. The
line-to-line voltage of SVPWM is then increased by about:

√
3Vdc
Vdc − 2
√ × 100 ≈ 15.5%. (6.3)
3Vdc
2

Figure 6.1: Locus Comparison of Maximum Peak Voltage in SPWM and SVPWM [28].

From the simulation studies, the diagram in Figure 6.2 shows the evolution of the
→
−
voltage reference vector V Re f in the complex plane. The space voltage vector, which
rotates with constant length and constant circular frequency, has the same features as
that of a line-to-neutral phase voltage. The rotating space voltage vector is inside the
 77

hexagon for under-modulation and between the inscribed circle of the hexagon and the
circumscribed circle of the hexagon for over-modulation.

Figure 6.2: Loci of SPWM, SVPWM and Region 1 of SVPWM.

→
−
Figure 6.3 shows the locus of SPWM from the simulation results. When V re f is
desired to produce a balanced set of three-phase sinusoidal voltages, then the locus of
→
−
V re f is a circle inscribed inside the hexagon (Figure 6.4) when these space vectors are
plotted on real and imaginary axes.

Figure 6.5 shows the maximum circle loci of THIPWM and under-modulation of
SVPWM. They have the same radius.

In Figures 6.6 and 6.7, the locus of the space vector applied is partly circular and partly
hexagonal. In the hexagon portion, only two active states are applied. The simulated
 78

Figure 6.3: Locus of SPWM.

200

150

100

50

−50

−100

−150

−200
−200 −150 −100 −50 0 50 100 150 200

Figure 6.4: Locus of SVPWM in Linear Modulation Range.

results show that the over-modulation region 1 reported a modulation index that can reach
to 0.952. This is an extension of about 5%, which is a significant improvement.
 79

Figure 6.5: Loci of THIPWM and SVPWM.

6.2 Total Harmonic Distortion (THD) Comparison

In this section, the four different PWM techniques (SPWM, THIPWM, SVPWM in linear
modulation range, and SVPWM in over-modulation range 1) are compared in terms of
total harmonic distortion (THD). A Fast Fourier Transform (FFT) analysis in MATLAB is
used to conduct the harmonic analysis. The harmonic spectrum of the inverter voltage
waveforms of these techniques are presented with different modulation indices. The
frequency modulation ratio in this analysis is 30 using a modulation frequency of 60
Hz and a carrier frequency of 1800 Hz. The results shown in tables and figures below
include the first 50 harmonics. The sidebands frequency are positioned on both sides of
the carrier frequencies with a frequency separation of ±µ fo , where fo is the frequency of
 80

200

150

100

50

−50

−100

−150

−200
−200 −150 −100 −50 0 50 100 150 200

Figure 6.6: Locus of Over-Modulation Region 1 at a Cross Angle of 15 Degrees.

200

150

100

50

−50

−100

−150

−200
−200 −150 −100 −50 0 50 100 150 200

Figure 6.7: Locus of Over-Modulation Region 1 at a Cross Angle of 0 Degree.

the reference sinusoidal and µ is an integer that depends on the carrier harmonic frequency
[26]. The harmonic number of individual sidebands can be found using the following
 81

formula: fc ± µ fo , where fc is carrier frequency.

Figure 6.8: Spectrum of Vao for SPWM.

Table 6.1: Pole Voltage Vao in SPWM

MI THD(%) Fund h26 h28 h30 h32 h34

0.733 113.3 140 2.06 42.73 101.34 42.7 2.06
0.7854 99.51 150 2.68 47.69 90.19 47.67 2.66

Figure 6.7 and Table 6.1 show the harmonic results for the pole voltage Vao in the
SPWM technique. As expected, we do not see a third harmonic in SPWM. Comparing
the data of two different modulation indices in Table 6.1, we conclude that the higher
modulation index has a lower THD (%).
Figure 6.8 and Table 6.2 show the harmonic results of the THIPWM technique. This
technique has clearly increased the output voltage of the inverter compared to the SPWM
technique. Thus, the third harmonics is presented and has increased phase and line-to-line
voltages as well. As expected, the value of third harmonic in the Vao is about one-sixth
that of the fundamental voltage.
 82

Figure 6.9: Spectrum of Vao for THIPWM.

Table 6.2: Voltages Vab , Van , and Vao in THIPWM

THD(%) Fund h3 h26 h28 h30 h32 h34 Voltages

51.76 300.08 ≤1 37.56 71.42 ≤1 71.42 37.49 Vab
51.74 173.3 ≤1 21.68 41.22 ≤1 41.18 21.66 Van
70.12 173.2 28.74 21.68 41.21 56.21 41.18 21.67 Vao

Compared to SPWM, the values in 26th and 34th harmonic of THIPWM are very high
and THD of THIPWM is significantly lower than for SPWM. The values of the 24th , and
36th harmonics are less than 5 Volts.

Table 6.3: Harmonics of Vao in Under-Modulation range of SVPWM

MI THD (%) Fund h3 h26 h28 h30 h32 h34

0.733 111.34 139.72 28.56 18.25 25.9 95.73 25.43 17.63
0.7854 97.38 149.7 30.61 20.48 29.03 83.53 28.57 19.85
0.806 92.03 153.71 31.41 21.40 30.30 78.53 29.84 20.76
0.9069 67.61 172.9 35.37 25.74 36.38 54.06 35.97 25.11

Figure 6.9 and Table 6.3 show the harmonic results of Vao in the under-modulation
 83

Figure 6.10: Spectrum of Vao in The Under-Modulation Range of SVPWM.

region of SVPWM. In Table 6.3, there are four different modulation indices arranged in
increasing order of magnitude. As in the SPWM technique, a lower modulation index has
a higher THD (%). As can be seen from the figure, there are also a few values in the 22th ,
24th , 36th , and 40th harmonic values but they are lower than 5 Volts.

Compared to SPWM with the same modulation index, the THD of SVPWM is slightly
lower. The fundamental voltages (Fund) of both techniques are the same. Compared to
THIPWM, the THD(%) of SVPWM (linear-modulation) is slightly lower than THIPWM
and the value in the 3rd harmonic of SVPWM (under-modulation) is slightly higher than
that in THIPWM. The fundamental voltages of both techniques are comparable.

Table 6.4: Vao in Over-Modulation Region 1 of SVPWM

MI THD(%) Fund h3 h26 h28 30 h32 h34

0.932 66.69 177.9 37.52 24.82 35.92 47.02 37.63 27.53
0.952 58.60 181.8 39.35 24.83 36.06 41.08 37.74 27.51
 84

Figure 6.11: Spectrum of Vao for SVPWM (Over-Modulation Region 1).

Figure 6.10 and Table 6.4 show the harmonic results of Vao in the over-modulation
region 1 of SVPWM. Overall, the spectrum of this technique is similar to the under-
modulation of SVPWM. However, it has a higher fundamental voltage and a higher value
of the 3rd harmonic. We also see additional harmonic values around the carrier frequency
(30th harmonic) and a lower THD (%) with a higher modulation index.
Tables 6.5-6.10 show the harmonic results for Van and Vab . Tables 6.5-6.6 show the
results of the SPWM technique. Tables 6.7-6.8 show the results of the SVPWM under-
modulation technique. Table 6.8 shows the result of the SVPWM over-modulation region
1 technique. The line-to-line voltages are undistorted since the third harmonic components
in the phase waveforms cancel out. The line-to-neutral voltages do not contain the third
harmonic because this is a three-phase three-wire system. This means that there is no
neutral connection to the wye-connected balanced load, and the triplen harmonic of the
load currents must be zero.
 85

Table 6.5: Vab in the SPWM

MI THD(%) Fnd h26 h28 h32 h34

0.733 75.32 242.50 3.58 73.97 73.99 3.57
0.7854 68.09 259.77 4.64 82.57 82.58 4.6

Table 6.6: Van in the SPWM

MI THD(%) Fnd h26 h28 h32 h34

0.733 75.32 140.01 2.07 42.71 42.71 2.06
0.7854 68.09 149.98 2.68 47.67 47.68 2.66

Table 6.7: Vab in Under-Modulation of SVPWM

MI THD (%) Fnd h26 h28 h32 h34

0.733 73.3 242 31.49 45.06 43.93 30.60
0.7854 66.01 259.3 35.24 50.50 49.36 34.44
0.806 63.14 266.24 36.89 52.70 51.58 36.01
0.9069 49.54 300 44.38 63.28 62.21 43.56

Table 6.8: Van in Under-Modulation of SVPWM

MI THD (%) Fnd h26 h28 h32 h34

0.733 73.29 139.7 18.52 25.88 25.44 17.62
0.7854 66 149.72 20.49 29.01 28.58 19.84
0.806 63.13 153.72 21.39 30.29 29.86 20.74
0.9069 49.51 172.9 25.74 36.37 35.97 25.10

Table 6.9: Vab in Over-Modulation Region 1 of SVPWM

MI THD(%) Fnd h26 h28 h32 h34

0.932 47.86 308 42.97 62.45 65.18 47.56
0.952 44.78 315 42.95 62.75 65.45 47.54

Table 6.10: Van in Over-Modulation Region 1 of SVPWM

MI THD(%) Fnd h26 h28 h32 h34

0.932 47.72 178 24.97 35.86 37.47 27.60
0.952 44.64 182.1 24.95 36.01 37.61 27.56
 86

CHAPTER 7

CONCLUSION AND FUTURE WORK

7.1 Conclusion

This thesis has evaluated three different PWM techniques, namely SPWM, THIPWM, and
SVPWM (in the linear modulation region and over-modulation mode 1). The contributions
of the thesis are as follows:

• The thesis has provided a thorough review of the each techniques with a special
focus on the operation of SVPWM in the under-modulation and over-modulation
modes.

• In this thesis, Simulink models for all three techniques have been developed and
tested in the MATLAB/Simulink environment. The SVPWM model is able to gen-
erate both the operation of the under-modulation region of SVPWM as well as the
over-modulation region 1.

• The thesis discusses the advantages and drawbacks of each technique. Their simu-
lation results are compared and analyzed by plotting the output harmonic spectra of
various output voltages, and computing their total harmonic distortion (THD).

As seen from the simulation results, SVPWM and THIPWM have a superior performance
compared to SPWM, especially in the over-modulation region of SVPWM. The SPWM
technique is very popular for industrial converters. It is the easiest modulation scheme to
understand and implement. This technique can be used in single-phase and three-phase
 87

inverters.
The THIPWM technique operates by adding a third harmonic component to the sinu-
soidal modulating wave. It is possible to increase the fundamental by about 15.5% and,
hence, allow a better utilization of the DC power supply. From the shape of the line-to-line
voltages, the resulting flat-topped waveforms allow over-modulation with respect to the
original SPWM technique.
The SVPWM technique can only be applied to a three-phase inverter and it increases
the overall system efficiency. The SVPWM is used for controlling the switching of the
machine side converter. Advantages of this method include a higher modulation index,
lower switching losses, and less harmonic distortion compared to SPWM [27]. SVPWM
research has been widespread in recent years making it one of the most popular methods
for three-phase inverters because it has a higher fundamental voltage output than SPWM
for the same DC bus voltage. The SVPWM is significantly better than SPWM by approxi-
mately 15.5%. However, the SVPWM technique is complex in implementation, especially
in the over-modulation region.
The SVPWM technique has been deeply studied in the over-modulation region due
to its performance benefits when compared to other modulation techniques. Numerous
over-modulation algorithms have been proposed in the literature for the control of voltage
source inverters [14,12,24]. The simulated results confirm that the over-modulation region
1 leads to a modulation index up to 0.952. This is an extension of around 5%, which is a
significant improvement.

7.2 Recommendations for Future Work

There are couple of interesting topics suggested for future research:

• Further simulation studies should be performed in region 2 of over-modulation SVPWM

 88

with neural networks and The Neural Network toolbox of MATLAB. Neural net-
work implementation is very fast and can increase the switching frequency of power
switches in the inverter.

• Many papers in the literature have reported that the dead time of space vector PWM
has an influence on drive systems. To circumvent this problem, it is important
to research how to compensate for the dead time effect in order to increase the
performance of drive systems.
 89

REFERENCES

[1] H. Salehfar, “DSP-Based Implementation of Vector Control of Induction Motor

Drives,” Taylor & Francis Group, LLC, 2005.

[2] R.K. Pongiannan, and N. Yadaiah, “FPGA Based Three Phase Sinusoidal PWM
VVVF Controller,” IEEE ICEES (International Conference on Electrical Energy
Systems), pp. 34-39, 2011.

[3] A. Aktaibi, A. Rahman, and A. Razali, “A Critical Review of Modulation Tech-

niques.” Available from: http://necec.engr.mun.ca/ocs2010/viewpaper.php?id=13&
print = 1.

[4] K.V. Kumar, P.A. Michael, J.P. John and S.S. Kumar, “Simulation and Comparison of
SPWM and SVPWM control for Three Phase Inverter,” Asian Research Publishing
Network, Vol. 5, No. 7, pp. 61-74, July 2010.

[5] J.Y. Lee, and Y.Y. Sun, “A New SPWM Inverter with Minimum Filter Requirement,”
International Journal of Electronics, Vol. 64, No. 5, pp. 815-826, 1988.

[6] H. Quan, Z.Gang, C. Jie, Z. Wu, and Z. Liu, “Study of A Novel Over-modulation
Technique Based on Space-Vector PWM,” IEEE Computer Distributed Control and
Intelligent Environmental Monitoring (CDCIEM), pp. 295-298, 2011.

[7] K. Zhou and D. Wang, “Relationship Between Space-Vector Modulation and Three-
Phase Carrier-Based PWM: A Comprehensive Analysis,” IEEE Transactions on
Industrial Electronics, Vol. 49, No. 1, pp. 186-196, February 2002.

[8] W.F. Zhang and Y.H. Yu, “Comparison of Three SVPWM Strategies,” Journal of
Electronic Science and Technology of China, Vol. 5, No. 3, pp. 283-287, September
2007.

[9] A.W. Leedy, and R.M. Nelms, “Harmonic Analysis of a Space Vector PWM Inverter
using the Method of Multiple Pulses,” IEEE Transactions on Industrial Electronics,
Vol. 4, pp. 1182-1187, July 2006.

[10] “Implementing Space Vector Modulation with the ADMCF32X,” Analog Devices
Inc., January 2000.

[11] B.K. Bose. Modern Power Electrics and AC Drives. Prentice-Hall, Inc., 2002.
 90

[12] J. Holtz, W. Lotzkat, and A.M. Khambadkone, “On Continuous Control of PWM
Inverters in the Over-modulation Range Including the Six-Step Mode,” IEEE Trans-
actions on Power Electronics, Vol. 8, No. 4, pp. 546-553, October 1993.
[13] J. Holtz, “On Continuous Control of PWM Inverters in Over-modulation Range
Including the Six-Step, ” IEEE Transactions on Power Electronics, Vol. 8, No. 4,
pp. 546-553, 1993.
[14] D.C. Lee, “A Novel Over-modulation Technique for Space Vector PWM inverters,”
IEEE Transactions on Power Electronics, Vol. 13, No. 6, pp. 1144-1151, Nov. 1998.
[15] D.G. Holmes and T.A. Lipo. Pulse Width Modulation for Power Converters. John
Wiley & Sons, Inc. 2003.
[16] A.M. Khambadkone, and J. Holtz, “Current Control in Over-modulation Range for
Space Vector Modulation based Vector Controlled Induction Motor Drives,” IEEE
Industrial Electronics Society, Vol.2, pp. 1134-1339, 2000.
[17] J.A. Houldsworth, and D.A. Grant, “The Use of Harmonic Distortion to Increase the
Output Voltage of a Three-Phase PWM Inverter,” IEEE Transactions on Industry
Applications, Vol. IA-20, No. 5, pp. 1124-1228, September-October 1984.
[18] M. Giesselmann, H. Salehfar, H.A. Toliyat, and T.U. Rahman, “Modulation Strate-
gies,” CRC Press LLC, 2002.
[19] A. Savulescu, “The Analysis and The Simulation of The SVM Generator Used
for The Control of The Electric Drives with Asynchronous Motors,” International
Conference on Electromechanical and Power Systems, pp. 47-50, October 4-6, 2007.
[20] E. Hendawi, F. Khater, and A. Shaltout, “Analysis, Simulation and Implementation
of Space Vector Pulse Width Modulation Inverter,” International Conference on
Application of Electrical Engineering, pp. 124-131, 2010.
[21] B. Wu. High-Power Converters and AC Drives. IEEE Press, John Wiley and Sons,
Inc., 2006.
[22] MathWorks, 2010, SimPowerSystems, User’s Guide.
[23] R. Cordero, and J.O.P. Pinto, “Relationship Between SPWM
and SVPWM for Undermodulation and Overmodulation
Modes Based on Modified Carrier Signal.” Available from:
http://www.labplan.ufsc.br/congressos/Induscon%202010/fscommand/
web/docs/I0275.pdf
[24] V.H. Prasad D. Boroyevich and R. Zhang, “Analysis and Comparison of Space Vector
Modulation Schemes for Three-Leg and Four Leg Voltage Source Inverters,” IEEE
Applied Power Electronics Conference and Exposition, Vol. 2, pp. 864-871, February
1997.
 91

[25] A. Oteafy, “How to Use a Sample & Hold Block after sector, duration time switch-
ing,” Personal Communication.

[26] H. Patangia and D. Gregory, “A Harmonic Reduction Scheme in SPWM,” IEEE

Asia Pacific Conference on Circuits and Systems, pp. 1737-1740, 2006.

[27] C.U. Ogbuka and M.U. Agu, “A Generalized Rectified Sinusoidal PWM Technique
for Harmonic Elimination,” The Pacific Journal of Science and Technology, Vol. 10,
No. 2, pp. 21-26, November 2009.

[28] Texas Instruments Europe, “Field Orientated Control of 3-Phase AC-Motors,” Texas
Instruments Incorporated, Literature Number: BPRA073, February 1998.