Advance

Switching Schemes for Inverters

First Edition
Edited by: Ankit Yadav and M. M. Tripathi “Dedicated to
our teachers, parents and grandparents”
PREFACE
DC-AC Inverters are key components in various industrial applications that include power
DC-AC Inverters are key components in various industrial applications that include power
phase & 3-phase) are finding more attention in renewable energy systems also. These
inverters are big source of harmonics, which leads to various other problems like heating
of instruments, faulty operation of protection devices etc. The harmonics generated by
inverter can be controlled and minimise by controlling the switching of the device. Several
approaches has been made already to deal with this problem like as carrier based PWM
(pulse width modulation), selected harmonic elimination, Random PWM etc.
Although the methods named above have been succeeded in improving the quality of
power by reducing harmonics but the problem is still not solved completely, so it still
required attention. The book presents the design of new switching schemes for single
phase and three phase Voltage Source DC – AC inverters. Modified hysteresis switching
scheme as well as wavelet based modulation technique are analysed and applied for
switching of single phase and three phase voltage source inverters to produce significantly
improved fundamental components and low harmonic contents.
Ankit Yadav
M. M. Tripathi
ABOUT AUTHORS
Ankit Yadav
Ankit Yadav is Sr. Project Associate in Indian Institute of Technology (IIT) Kanpur, Uttar
Pradesh, India. He received his B. Tech. (Electrical and Electronics Engineering) from
Uttar Pradesh Technical University, Lucknow, India and M. Tech (Power system
Engineering) from Delhi Technological University (DTU), New Delhi, India. He has
worked with JK Lakshmipat University, Jaipur, India. His area of interest is power
electronics and power system.
M. M. Tripathi
Madan Mohan Tripathi is Associate Professor in Delhi Technological University, New
Delhi, India. He has done his B. E. (Electrical Engineering) from Madan Mohan Malviya
Engineering College, Gorakhpur, India and Ph. D. (Electrical Engineering) from G. B.
Technical University, Lucknow, India. He has worked with Institute for Plasma Research,
Gandhinagar, India and National Institute of Electronics & Information Technology
(NIELIT), Delhi, India. His research interests include sustainable energy, power system
restructuring, AI application to power system problems and application of IT in power
system control and monitoring. He is a member of Plasma Science Society of India
(PSSI), IEEE and IETE.
INDEX
Heading Page No.
1.0 Switching Schemes for Inverters 8
1.1 Introduction 8
1.2 Basic Concept of Converter 10
1.2.1 AC – DC Converter 10
1.2.2 DC – AC Converter (Inverter) 10
1.2.3 Firing Angle 11
1.3 Drawback of Power Electronic Devices 11
1.4 Existing Switching Schemes 16
1.4.1 Pulse Width Modulation 16
1.4.2 Selected Harmonic Elimination 18
1.4.3 Hysteresis Band Current control 19
1.4.4 Other Switching Techniques 19
1.5 Summary 19
1.6 References 20
2.0 Introduction to Wavelets 23
2.1 Introduction 23
2.2 Types of Wavelets 24
2.2.1 Haar 24
2.2.2 Daubechies 24
2.3 Wavelet Transform 25
2.4 Multi Resolution Analysis (MRA) 26
2.5 Scaling Function 27
2.5.1 Sampling 27
2.5.2 Reconstruction 27
2.6 Summary 28
2.7 References 28
3.0 Modified Hysteresis Switching Scheme for 1- phase Voltage Source Inverter 30
3.1 Introduction 30
3.2 Distinct features of modified Hysteresis Switching Scheme 31
3.2.1 Block Diagram of modified Hysteresis Switching Scheme 31 3.2.2 Reference Signal
32 3.2.3 Problems and Modification 34
3.3 Implementation 36
3.4 Testing, Results and Discussion 39 3.4.1 Linear load (R – L load) 40 3.4.2 Dynamic
load (Motor load) 42
3.5 Summary 49
3.6 References 49
4.0 Modified Hysteresis Switching Scheme for 3-phase Voltage Source Inverter 51
4.1 Modified Hysteresis Switching Scheme for three-phase Inverter 51
4.2 New Switching Strategy 51
4.2.1 Reference Signal 52
4.3 Implementation 53
4.4 Simulation Results and discussion 57 4.4.1 Linear load 57 4.4.2 Nonlinear load 64
4.5 Summary 72
4.6 References 73
5.0 Wavelet Modulation for single-phase Inverter 75
5.1 Modulation Overview 75 5.1.1 Reference Signal 75 5.1.2 Output Waveform 75 5.1.3
Sampling and Reconstruction 76 5.1.4 Flowchart 77
5.2 Wavelet Modulation 78
5.3 Modeling and Simulation 80
5.4 Results and Discussion 80 5.4.1 R – L Load 81 5.4.2 R Load 87
5.5 Summary 93
5.6 References 93

CHAPTER 1
Switching Schemes for Inverters
1.1Introduction

The world’s first power system was built at Godalming in England, in 1881 and in last 130
years, it has become the backbone of world’s civilization, with enormous number of
existing power system, where each individual system is having an ocean of generating
units, loads, and other essential equipment. In current scenario, the world cannot dare to
survive without electricity, as it was few centuries back.
In the beginning of power system network, elements like as generating units, loads,
transmission network were less in number, as well as they were less sensitive to the
disturbances. So, the quality of power was not a big issue, as it is today. In today’s era of
technology the power network has expanded drastically and is full of electronic and power
electronic devices which are highly sensitive to the disturbances in system. In the same
period need of more and more energy is leading us to look for various renewable sources
to fulfill energy requirements of society. These energy sources are operating in grid
connected or off-grid mode. Power supply from these renewable resources are not pure
sinusoidal and its conversion to sinusoid injects large amount of harmonics in system and
thus deteriorate the quality of power supply. Also now-a-days, reliability of power supply
has become need of hour. Considering the above issues, quality of power supply has
emerged as a serious concern in front of the engineers involved in power system
technology.
Power electronic devices have shown their ability to control the power flow with greater
flexibility and so, their use in power system is increasing day-by-day. With increasing
number of power electronic devices to manage the expanding power system network,
study of controlling these devices requires attention. All the power electronic devices
being used now a days are dependent on the gate pulse & firing angle for any type of
control, or we can say, only by controlling gate signal, we can control the behavior of our
power electronic devices and thus the power system network.
The operation of these devices are controlled by gate signals and thus the output,
harmonic, efficiency and other factors can be controlled by controlling gate signal. Poor
switching schemes of these power electronic devices may result in degradation of quality
of power flowing through wires and in to the systems. A proper switching strategy can
give an efficient and better control over output of power electronics device leading to
better quality of power. Several factors of operation that must be taken care while
designing switching strategy are as follows.
a. RMS values of voltage and current
b. Average and Peak value of voltage and current c. Total harmonic distortion (THD)
waveforms d. Switching losses
e. Practical implementation
Work on switching strategy of power electronic switching devices came in to existence
with birth of power electronics and has resulted in several options such as single pulse
switching, pulse width modulation (PWM), random PWM (RPWM), selected harmonic
elimination (SHE), delta modulation, hysteresis & variable band, optimal PWM. The
reason behind development of all these methods is to minimize switching losses and
higher order harmonics because the power electronics devices are major source of these
higher order harmonics. From all these methods, PWM has been widely accepted because
of its simplicity and ability to solve the problem significantly. Although PWM has proven
a better option, there is still a scope of improvement, as the problem is not eradicated
completely.
Two new switching schemes for controlling voltage source inverter namely, (i) modified
hysteresis switching scheme and (ii) wavelet modulation technique are discussed in this
book. The modified hysteresis switching scheme is modification of exiting hysteresis band
current control scheme, has been designed for single-phase as well as for three-phase
voltage source inverter. However wavelet based switching scheme has been designed for
single-phase inverter only.

1.2 Basic Concepts of Converter

1.2.1 AC – DC Converter
AC – DC converter (also known as controlled rectifier) is a thyristor based power
electronics device for rectification purpose. The figure 1.1(a) and 1.1(b) shows
singlephase and three-phase converter devices and their respective simulation outputs are
shown in figure 1.2(a) and 1.2(b). Simulation is performed with inductive load at firing
angle of 300 for single-phase converter, while it is 200 in case of three-phase converter.
Observation of output waveforms of single and three-phase converter shows that the
nature of output of the power electronic devices is dependent on the firing angle applied to
power electronics switches. Thus, the firing angle and its control are directly responsible
for operation of devices and quality of the output. Input to these converter is pure
sinusoidal AC, while the output is controlled DC [1], [2].
1.2.2 DC – AC Converter (Inverter)
In a similar fashion as AC – DC converter, DC – AC converter (inverter) can be designed.
The important point to notice is that, design of inverter using thyristor is very complex, as
thyristor requires commutation circuit, when being used for inverting purpose and at the
same instant, design of commutation circuit put more complexity as well as cost to the
device. To avoid unnecessary cost and complexity, inverters are designed using IGBT or
GTO, as these devices demand no additional circuitry for commutation.
Single-phase, as well as three-phase inverter circuits have been designed for simulation
purpose and are shown in figure 1.1(c) and 1.1(d) respectively. Simulation results of these
circuits with inductive loading are also shown in figure 1.2(c) and 1.2(d) respectively. The
input to these devices is a supply with constant DC voltage and output obtained is
controlled AC. All the figures 1.1 (a), (b), (c) and (d) have blocks named as pulse, these
block used as firing circuit of particular thyristor of IGBT.
1.2.3 Firing Angle
Firing angle can be define as the angle at which power electronics devices such as
Thyristor, IGBT, GTO etc. are being turned on by gate pulse [3], [4]. Complete operation,
output and behavior of power electronic device are controlled by controlling the firing
angle of device. Any variation in firing angle would cause direct impact over conduction
angle and commutation angle of the device.

1.3Drawback of Power Electronic Devices

Because of their wide range of controllability and applicability, power electronic devices
have made significant presence in power system network. These devices have resolve
many important issues like as filtering harmonics due to loads, maintaining voltage
profile, assuring maximum power transfer etc. and the application area of power electronic
devices is still growing. With tremendous increment in number of power electronic
devices in power system network, impact of the devices also become important concern.
Waveforms shown are voltage and current waveform respectively for each figure. The
Yaxis is either voltage or current while X- axis represents time ‘t’. The power electronic
devices are source of high frequency harmonics in power lines and causes high switching
losses. Both the problems (harmonics and switching losses) depend on switching strategy
being used for controlling the operation of devices.

1.1 (a)

1.1 (b)
1.1 (c)

1.1 (d)
Figure 1.1 Simulation model of (a) 1-phase converter, (b) 3-phase converter, (c) 1-phase
inverter, (d) 3-phase inverter
2.2 (a)

2.2 (b)
2.2 (c)

2.2 (d)
Figure 1.2 Simulation result of (a) 1-phase converter, (b) 3-phase converter, (c) 1-phase
inverter, (d) 3-phase inverter (result of phase ‘a’ is shown)

1.4Exiting Switching Schemes

As discussed in first chapter, there are several strategies have been designed and
developed over a period for generating pulses. Few important switching strategies are as
follows:-
1.4.1 Pulse Width Modulation
Pulse-width modulation (PWM) or pulse-duration modulation (PDM) is a modulation
technique that controls width of the pulse, formally the pulse duration, based on modulator
signal information [5]-[7]. Although, this modulation technique can be used to encode
information for transmission, its main use is to allow the control of power supplied to
electrical devices, especially to inertial loads such as motors. PWM switching frequency
has to be much faster than what would affect the system.

1.3 (a)

1.3 (b)
Figure 1.3 (a) Comparator for PWM, (b) Pulse generation using comparator

1.4 (a)
1.4 (b)
Figure 1.4(a) Simulation model of single-phase PWM inverter, (b) output voltage and
current waveform of 1-ph PWM inverter
The widely accepted PWM switching strategy for power electronic devices uses basic
concept of comparator, where two input ports of comparator are connected to reference
signal (sin wave) and modulating signal (triangular wave) respectively, comparator
compares both inputs and generate output according to the logic. To understand this
method one can refer to figure 1.3(a) & (b).
With the purpose of understanding the application and impact of PWM switching strategy,
we used PWM generator block in Simulink as explained in figure 1.4(a) and performed
simulation with triangular modulating signal at frequency of 2 kHz to generate output at
50 Hz. The final output is of 50 Hz is in figure 1.4(b).
1.4.2 Selected Harmonic Elimination
Selected harmonic elimination targets over specified order of harmonic (e.g.-3rd 5th 7th
etc.). For this purpose specified numbers of pulses with specified pulse–width are applied
to gate terminal of device. As mentioned in name, this method can be applied for specified
order of harmonic only. We cannot apply this method for harmonics other than specified,
which becomes major drawback of this method and limits its application.
The condition must be satisfied for selected harmonic elimination is given as in (1.1): -
k= (1.1)
Where ‘ ’ is pulse width and ‘n’ is the order of harmonic we want to eliminate.
A selected harmonic elimination method gives following output waveform for 3rd
harmonic elimination. It should be noted that, with third harmonic all other harmonics that
are multiple of three (triplents, 3, 6, 9, 12, and so on) are also eliminated.

0 k
Figure 1.5 Selected harmonic elimination for 3rd harmonic.
1.4.3 Hysteresis Band Current Control
Hysteresis band current control could be explain as shown in with the figure 1.6. In
hysteresis switching scheme, there is a reference and two bands (upper band and lower
band) [8]. The controller compares value of actual current with these bands and turns on
or off the switch pairs to generate +V or –V at output of inverter.

Figure 1.6 Hysteresis band current control

The hysteresis band concept shown here is bipolar switching scheme but it could be
unipolar as well. The switching frequency depends upon how fast current changes from
upper limit to lower limit and vice-versa. This in turns depends upon nature of load.
1.4.4 Other Switching Techniques
The importance of suitable switching techniques can be understood by fact that the list of
switching strategies doesn’t stops here, there are many more techniques have been
developed to tackle the issue. Random PWM (RPWM), optimal PWM, delta modulation,
hysteresis control, variable band technique, hybrid PWM, phase shifted PWM,
carrierbased PWM, multiple-carrier PWM, and tolerance band based PWM are few name
in the list of other switching strategies [9]-[17].

1.5Summary

Power Electronic devices provide better solution for the power control and integration of
renewable sources with existing grid system but they also brought the complexity in
operation. Operation of these devices itself are based on switching scheme and opens a
new frontier for researchers. Inverter is one of the key member of power electronic family
and several switching schemes like PWM, RPWM, SHE, HBCC etc. are developed for it
in last decades. This book present two advance switching scheme for voltage source
inverter. Power electronic devices are known for injecting harmonic in the power system
and thus research over new switching scheme to minimize this drawback become more
and more important.

References

[1]. Ned Mohan, Tore M Undeland and William P. Robbins, “Power Electronics –
Converters, Application and Design,” 3rd ed. Wiley India, Noida:2012 [2]. Dr. P. S.
Bimbhra, “Power Electronics,” 4th ed. Khanna Publishers, Delhi:2009 [3]. Jai P. Agrawal,
“Power electronic systems – Theory and Design,” 1st ed. Pearson
Education Asia, Singapore:2001
[4]. L. Umanand, “Power Electronics Essentials & Applications,” 1st ed. Wiley,
Delhi:2010 [5]. Pankaj H. Zope, Pravin G. Bhangale, Prashant Sonare and S. R. Suralkar,
“Designing
and Implementation of carrier based Sinusoidal PWM Inverter,” in International Journal of
Advanced Research in Electrical, Electronics and Instrumentation Engineering, Vol. 1,
Issue 4, October 2012, ISSN: 2278 – 8875.
[6]. Weni Zhang, Yuye Wang, Weiming Tong, Huiming Xu,and Lemin Yang, “ Study of
Voltage-Source PWM Inverter Based on State Combination Method,” in IEEE PES
Transmission and Distribution Conference & Exhibition, Asia and Pacific Dalian, China,
2005.
[7]. Wenyi Zhang and Wensheng Chan, “Research on Voltage-Source PWM Inverter
Based on State Analysis Method,” in Proc. of IEEE International Conference on
Mechantronics and Automation, China, August 9-12, 2009 .
[8]. Firuz Zare and Jafar Adabi Firouzjee, “Hysteresis Band Current Control for Single
Phase Z-Source Inverter with Symmetrical and Asymmetrical Z-Network,” in Power
conversion Conference, Nagoya, 2007, PP.143-148.
[9]. Phoivos D. Zigas, “The Delta Modulation Technique in Static PWM Inverters,” in
Phoivos D. Zigas, “The Delta Modulation Technique in Static PWM Inverters,” in 204.
[10]. M. Azizur Rahman, Jhon E. Quaicoe, and M. A. Choudhary, “Performance Analysis
of Delta Modulated PWM Inverters,” in IEEE Trans. on Power Electronics, Vol. PE-2, No.
3, July 1987, PP. 227 – 233.
[11]. B. Vasantha Reddy and B. Chitti Babu, “Hysteresis Controller and Delta Modulator –
Two Viable Scheme for Current Controlled Voltage Source Inverter,” in IEEE
International Conference on Technical Postgraduates 2009, 14-15/Dec/2009, Kuala
Lumpur, Malaysia.
[12]. Jonathan W. Kimball, Philip T. Krein and Yongxiang Chen, “Hysteresis and Delta
Modulation Control of Converters Using Sensorless Current Mode,” in IEEE Trans. on
Power Electronics, Vol. 21, No. 4, July 2006.
[13]. Yen-Shin Lai, Yong-Kai Lin and Chih-Wei Chen, “ New Hybrid Pulse Width
Modulation Technique to Reduce Current Distortion and Extend Current Reconstruction
Range for a Three-Phase Inverter Using Only DC-link Sensor,” IEEE Transaction on
Power Electronics, Vol. 28, No. 3, March 2013.
[14]. Prasopchok Hothongkham, Somkiat Kongkachat and Narongchai Thadsaporn,
“Performance Comparison of PWM and Phase-Shifted PWM Inverter Fed HighVoltage
High-Frequency Ozone Generator,” in IEEE Region 10 Conference, TENCON 2011, PP.
976-980
[15]. Wensheng Song, Xiaoyun Feng and Keyue Ma Smedley, “A Carrier-Based PWM
Strategy With the Offset Voltage Injection for Single-Phase Three-Level NeutralPoint-
Clamped Converters,” in IEEE Trans on Power Electronics, vol. 28, No. 3, March, 2013.
[16]. P.T. Krien, Xing Geng and R. Balog, “High-Frequency Link Inverter Based on
Multiple-Carrier PWM,” in Applied Power Electronics Conference and Exposition, APEC
2002, vol.2, PP. 997-1003
[17]. Istvan Nagy, “Novel Adaptive Tolerance Band Based PWM for Field-Oriented
Control of Induction Machine,” in IEEE Trans. on industrial Electronics, Vol.41, No. 4,
August 1994.

CHAPTER 2
Introduction to Wavelets
2.1Introduction

A wavelet is a wave-like oscillation with amplitude that begins at zero, increases and then
decreases back to zero [1]. It can typically be visualized as a “brief oscillation” like one
might see recorded by seismograph or heart monitor. Generally, wavelets are purposefully
crafted to have specific properties that make them useful for signal processing. Wavelets
can be combined, using a “reverse, shift, multiply and sum” technique called convolution,
with portions of unknown signal to exact information from the unknown signal.
Development of wavelets can be linked to several separate trains of thought, starting with
Haar’s work in the early 20th century. Later works by Dennis Gabor yielded Gabor atoms
(1946), which, constructed similarly to wavelets, and applied to similar purposes. Notable
contributions to wavelet theory can be attributed to Zweig’s discovery of the continuous
wavelet transform in 1975 (originally called the cochlear transform and discovered while
studying the reaction of the ear to sound). Pierre Goupillaud, Grossmann and Morlet’s
formulation of what is now known as continuous wavelet transform (CWT) (1982), Jan-
Olov Strömberg’s early work on discrete wavelets (1983), Daubechies’ orthogonal
wavelets with compact support (1988), Mallat’s multiresolution framework (1989),
Akansu’s Binomial QMF (1990), Nathalie Delprat’s time-frequency interpretation of the
CWT (1991), Newland’s harmonic wavelet transform (1993) and many others since.
Wavelets also can be applied in field of power electronics, as in this project we are
applying wavelets for modulating the gate pulse of inverter. The generation of gate pulses
for power electronic devices and using them to generate the output waveform of current
and voltage, is actually a kind of signal processing, where we desire to obtain a sinusoidal
output waveform. That means, the methods of signal processing can be applied to power
electronic devices. We have several methods in practice for signal processing like as:
Fourier transform, Laplace transform etc., but the common drawback of all existing
methods is that, at a time, they can work with either frequency or time but wavelet
transform deals with both simultaneously. In last few decades, wavelet has made its
significant place in field of signal processing and thus draws our attention toward itself.

2.2Types of wavelets

The group of wavelet includes many families named Haar, Daubenchies, Biorthogonal,
Coiflets, Symlets, Morlet, Maxican Hat, Meyer and so on. Two major families are
discussed in brief as follows.
2.2.1 Haar
Any discussion of wavelets begins with Haar wavelet, the first and simplest. Haar wavelet
is discontinuous, and resembles a step function. It represents the same wavelet as
Daubechies db1. A Haar wavelet is shown in figure 2.1.
Figure 2.1
2.2.2 Daubechies
Ingrid Daubechies, one of the brightest stars in the world of wavelet research, invented
what are called compactly supported orthonormal wavelets — thus making discrete
wavelet analysis practicable.
The names of the Daubechies family wavelets are written dbN, where N is the order, and
db the “surname” of wavelet [1]. The db1 wavelet, as mentioned above, is same as Haar
wavelet. Wavelet functions psis of the next nine members of the family are shown below:

Figure 2.2 Daubenchies wavelet family

2.3Wavelet Transform
It is traditional to introduce wavelet transform in the following way. The wavelet
transform is similar to short-time fourier transform, maps a function, S(t), in to a two –
dimensional domain (the time – scale plane) and is denoted by Ws(a,b) given by:

Ws(a,b) = S(t)h dt
= S(t)hab*(t)dt (2.1)

Where h(t) is in general called the mother wavelet, and the basis function of the transform,
called daughter wavelet, are given by:

Hab(t) = h (2.2)

Equation 2.1 is also known as the expansion formula, the forward transform or analysis;
hab(t) is a set of basis function obtained from mother wavelet h(t) by compression or
dilation using scaling parameters ‘a’ and temporal translation using shift parameter ‘b’.

C= ( ) ^ dw < infinity (2.3)

Where C is constant.
Wavelets are localized wave and they extend not from finite time duration. We can thus
think using h(t) and its scaled daughter function as the basis for a new transform [2]-[4].
The most important point of wavelet is that the basis functions are obtained by scaling and
shifting one particular function or by manipulation of single function. In that sense,
wavelet transform is similar to a fourier transform, whose basis functions are obtained by
manipulating a periodic function. Thus for periodic function, fourier analysis is ideal.
However, with wavelet transform, we are not restricted to only periodic functions but any
function, which satisfies its conditions. Implementation of a fourier transform using linear
system involve a time – variant impulse response. In contrast, wavelet transform as
defined can be implemented using time – invariant system. The components of wavelet
transform are discussed below.

2.4Multi resolution analysis (MRA)

Multi resolution analysis (MRA) can be considered as link between harmonic analysis (i.e.
time – frequency analysis) and discrete signal processing, and, in particular, discrete
multirate filter banks using or quadrature mirror filters. MRA can be defined as analysis
and synthesis of signal using wavelet basis function. Multi-resolution analysis is similar to
subband decomposition; however, for this case, time variable is continuous rather than
discrete. On multi-resolution analysis, the wavelet function is related to impulse response
of high pass filter, in a similar manner as in sub-band case, a low-pass filter is also
defined. The impulse response of this filter is the scaling function. It is defined from
mathematical standpoint as follows.
Given a sequence of embedded close subspaces i.e.

V ………… L (R) (2.4)

The goal of multi-resolution is to choose these subspaces in a way that they encompasses
whole function space without overlap or redundancy. The projection of the function gives
finer result as index j increases or function resolved with sampling period T= (1/2j)*T0
where T0 is the sampling period in V0[24].

2.5 Scaling Function

Wavelets are defined by wavelet function t) (i.e. mother wavelet) and scaling function t)
(also called father wavelet) in the time domain [5], [6].
The wavelet function is in effect a band-pass filter and scaling it for each level halves its
bandwidth. This creates problem that in order to cover the entire spectrum, an infinite
number of levels would be required. The scaling function filters the lowest level of the
transform and ensures the entire spectrum is covered. We can construct an orthogonal
basis for each space VM by dilating and scaling a single function t).
2.5.1 Sampling
Sampling is reduction of a continuous signal to a discrete signal. In process of sampling,
values of the signal are recorded at specified interval. In case of wavelet, the interval for
sampling is calculated by scaling function. The sampling process must satisfy nyquist
Sampling Rate given as fs > 2*B.
2.5.2 Reconstruction
Reconstruction is method of obtaining original signal from the sampled signal. In other
words, one may say that generating a continuous signal from discreet signal. As in
sampling, scaling function defines the interval for sampling also in case of reconstruction
same scaling function would be used for proper generation of original signal.

2.6Summary

A wavelet is a wave-like oscillation with amplitude that begins at zero, increases and then
decreases back to zero. It can typically be visualized as a “brief oscillation” like one might
see recorded by seismograph or heart monitor. The wavelets are crafted for specified
purpose. Different types of wavelets like as HAAR, Daubechies, Morlet etc. are developed
over a time by various scholars & HAAR is the first and simplest wavelet in all of them.
Wavelet Transform provide two dimensional (time – scale) analysis of signal unlike to
Fourier Transform and other methods that result in one dimensional (either time or scale)
analysis only. This two dimensional analytic property of wavelet transforms comes from
MRA (multi resolution analysis). Any wavelet is defined by wavelet function t) (i.e.
mother wavelet) and scaling function t) (also called father wavelet) in the time domain.
These wavelet function and scaling functions becomes the base for sampling and
reconstruction of the signal.

References

[1] http://en.wikipedia.org/wiki/Wavelet
[2] Agostino Abbate, Casimer M. DeCusatis and Pankaj K. Das, “Wavelet and
Subbands Fundamental and Applications,” Birkhauser, Boston:2002. [3] Stephane Mallat,
“a Wavelet Tour of Signal Processing,” 2nd ed. Academic Press,
London: 1999.
[4] S. A. Saleh, C. R. Moloney and M. A. Rahman, “Development and Testing of
Wavelet Modulation for Single-Phase Inverters,” IEEE Trans. Ind. Electronics, vol.
56, no. 7, pp. 2588 – 2599, July 2009.
[5] http://en.wikipedia.org/wiki/Multiresolution_analysis
[6] S. A. Saleh, C. R. Moloney and M. A. Rahman, “Developing a Non – Dyadaic
MRAS for Switching DC – AC inverters,” in Proc. IEEE 12th DSP Conf. , Jackson
Lake Lodge, WY, Sep. 2006, pp. 544 – 549.

CHAPTER 3
Modified Hysteresis Switching Scheme for 1 – phase Voltage
Source Inverter
3.1 Introduction

Current regulated modulations of voltage source inverter are in knowledge from very
beginning of the inverter. Tolerance band control, fixed frequency control etc. fall in of
this family. Proposed modified hysteresis switching scheme is also a current regulated
modulation technique. It is modification of existing hysteresis band current control
method. This technique uses value of output current as feedback signal and value of
desired output current as reference signal. These two signals contribute in defining the
controlling actions to be taken and accordingly the gate pulses for inverter get generated.
Thus in the proposed strategy, relation between values of the load current and reference
signal is used for generating gate pulses to IGBT switches of the Inverter.

Figure 3.1 Single-phase H-bridge VS DC – AC inverter model

The modified hysteresis switching scheme has been designed for single-phase H-bridge
VS DC–AC inverter. A circuit model of single-phase H-bridge VS DC–AC inverter is
shown in figure 3.1, which has been used throughout the process of designing and testing
of new switching scheme. The inverter model has four IGBT switches S1, S2, S3, and S4.
These IGBT switches are controlled via gate pulses provided through G1, G2, G3 and G4
respectively. Switching scheme is logic for proper sequence, which is implemented
through a programmable logic circuitry. The new switching scheme is discussed below in
fig. 3.1.

3.2 Distinct features of modified Hysteresis Switching Scheme

The designed modified hysteresis scheme is modification of exiting hysteresis current

control scheme. The new switching scheme differs from existing one in following term.
I. No Band: - The new switching scheme is free from band concept when we apply it for
three-phase system. Although, it is a hybrid of band and band less concept for single-
phase system.
II. Hybrid: - It is a hybrid of unipolar and bipolar unlike to existing schemes, as they are
either unipolar or bipolar.
III. Variable reference: - The reference in new scheme is of variable nature instead of
being fixed in earlier schemes. This reference is derived from load current itself.
3.2.1 Block Diagram of modified Hysteresis Switching Scheme
The block diagram of new switching scheme is shown in figure 3.2. In new switching
scheme readings of output current in ammeter is taken as feedback. Thus, the feedback
signal is a copy of load current. This feedback signal can be named as signal B. The
reference signal for new switching scheme is a sinusoidal signal; define as in (3.1).
A=A sin ( t) (3.1)
Where Amax is maximum value of reference signal A. The reference signal in this
switching scheme is actually a copy of desired sinusoidal load current. During the process
of generating gate pulses, switching circuit compares feedback signal B with respect to
sinusoidal reference signal A. Figure 3.3 (a) shows a conceptual diagram of new switching
scheme for positive cycle. When feedback signal B is less than reference signal A, the
switching circuit generates ON pulse to switch S1 and S2. This ON pulse brings the
switches S1 and S2 in conducting mode and as soon as switches S1 and S2 comes in
conducting mode, load current and thus the feedback signal starts increasing. As load
current and thus the feedback signal B attains higher value than reference signal A,
switching circuit generates OFF pulse for switches S1 and S2. This OFF pulse blocks the
conduction mode of switches S1 and S2. This decreases the value of load current (and
feedback signal) to level of reference or even lower than that. A similar process is being
followed for negative cycle also. Switches S1 and S2 operate during positive cycle and
switches S3 and S4 operates during negative cycle of output current. A plot showing
actual feedback, reference and gate signals are shown in figure 3.3(b).

Figure 3.2 Block diagram for new switching scheme

3.2.2 Reference Signal
The reference signal plays an important role in any switching scheme. Complete process
of switching (or pulse modulation) depends on selected reference signal. As discussed
earlier, reference signal for new switching scheme is a sinusoidal signal defined in (3.1).
Uniqueness of new strategy lies in its reference signals. Instead of having a fixed reference
signal and scaling the modulating signal to that range, a new concept of variable reference
signal has been introduced. This reduces the calculation and eliminates circuitry to scale
modulating signal. This reference signal is a copy of desired load current and load current
depends on load, so, may not have same value for all kind of load. So, the reference signal
also may not be same for all kind of application. The value of amplitude of desired load
current and thus reference signal will vary from load to load.

3.3 (a)

3.3 (b)
Figure 3.3 (a) A conceptual plot for new switching scheme, (b) Plot for new switching
scheme with practical data.
Considering the above facts, value of Amax has been kept variable and thus reference
signal has been defined in two parts.
I. Initial reference signal II. Updated reference signal
Initial reference signal is defined with a small constant value of amplitude as in (3.2). Here
amax is constant value and is less than Amax.
A=a sin ( t) (3.2)
This initial reference signal is required to initiate the switching process. As soon as
switching process begins and inverter starts delivering power to load, the reference signal
must be updated according to load demand, until its amplitude reaches to maximum value
Amax as shown in figure 3.4. The updating of reference signal follows the equation (3.3).
a =b (3.3)
Where bmaxis the maximum amplitude of the feedback signal in the ongoing cycle.

Figure 3.4 Reference signal A.

3.2.3 Problems and Modification
The defined switching scheme has a problem also. Figure 3.3 (b) shows that in second half
of positive cycle when gate pulse is OFF even though the feedback signal (load current) of
inverter is not following the reference signal. Load current is higher than reference current
during this period and just switching off switches S1 and S2 is not sufficient enough to
decrease the value of feedback signal (load current) to the level of reference signal. This
problem arises because of inductive nature of load. Inductive property of load causes the
storage of energy during first half of positive cycle when the load current is increasing and
during second half of cycle, it behaves as a current source and causes the excess value of
load current.
The solution of this problem has been driven from concept of hysteresis control (or
modulation) scheme for inverters. Similar to tolerance band of hysteresis current control,
concept of error has been introduced to tackle the above problem. By recording difference
between reference signal and feedback signal one can obtain the error signal. If we
observe error signal shown in figure 3.5, then we can analyze that the error between the
reference signal and feedback signal is almost in a fixed limit except a specific period of
second half of positive or negative cycle. The inductive effect of load is causing trouble
during this period.
As the error signal is almost in fixed limit except a specific period, we can define an error
band as a band of fixed limit for the error signal. Furthermore, two new terminologies
have been introduced.
Figure 3.5 Error signal representation.
A. Positive error
Positive error is the error between reference signal and feedback signal during positive
cycle.
B. Negative error
Similar to positive error, negative error is the error between reference signal and feedback
signal during negative cycle.
Using the above concept of error band and positive and negative error, an additional
switching scheme has been defined. Switching circuit compares error with the error band
and accordingly generates gate pulses. During the positive cycle when positive error
becomes more than the error band the switching circuit generates ON pulse for switches
S3 and S4. This establishes a negative voltage across the load with respect to existing
voltage and tries to establish a current in opposite direction and so the excess energy gets
balanced. A similar logic is applied to negative cycle also. Switches S3 and S4 get
operated during positive cycle and S1 and S2 during negative cycle.
It must be noted that this modification of error band is applied for very short duration of
time in respect of time-period of one cycle of current. During this period, switching
becomes bipolar and is unipolar for the rest of operation. Now this error band based
switching scheme is applied to inverter along with previously defined new switching
scheme.

3.3 Implementation

Designed switching scheme is implemented using S-function block of SIMULINK facility

of MATLAB R 2011b. A program has been written to realize the modified switching
scheme. Simulation and testing for modified switching scheme is performed on model of
single-phase VS DC–AC inverter shown in figure 3.1. The flowchart of complete
switching scheme is shown in figure 3.6.
The power electronic switches of inverter are controlled by gate pulses. Output of
switching circuit is applied to gate terminal of these switches. The gate terminals are
named as G1, G2, G3 and G4 for switches S1, S2, S3 and S4 respectively. Logic ‘1’ on
gate terminal of any switch represent ON pulse for respective switch and similarly logic
‘0’ represents OFF pulse. The error band is represented by Eb in description and flowchart
below. The step by step description of the modified switching scheme is as follows.
Step 1) Initialize the switching circuit.
Step 2) Define basic parameter such as frequency (fm) time period (Tm), sampling time
(Ts) and set time t=0;
Step 3) Define initial reference current as in (3.2)
A=a sin(t) (3.4) Step 4) Evaluate positive (+ve) error and negative (-ve) error.
+ve error = A B (During +ve cycle) (3.5) ve error = B A (During –ve cycle) (3.6) Step 5)
If t ; then go to step 6. Otherwise go to step 9.
Step 6) If A > ; then G1 = G2 = 1 and G3 = G4 = 0. Otherwise go to step 7. Step 7) If +ve
error > E ; then G1 = G2= 0 and G3 = G4 =1. Otherwise G1 = G2 = G3 = G4 = 0.
Step 8) Go to step 11.
Step 9) If B > ; then G1 = G2 = 0 and G3 = G4 = 1. Otherwise go to step 10. Step 10) If ve
error > E ; then G1 = G2= 1 and G3 = G4 =0. Otherwise G1 = G2 = G3 = G4 = 0
Step 11) Check the amplitude of reference signal, does it require update?
Step 12) Update the value of reference current if required, otherwise leave this step Step
13) If t T ; then reset the value of t. otherwise leave this step.
Step 14) Go to step 4.
Figure 3.6 Flowchart for modified switching scheme.
When this switching scheme is applied to motor or any other load withdrawing relatively
large load current in beginning than the load current during steady state operation, then
system faces a problem with reference signal that the large current in beginning set a
reference with relatively large amplitude than required during steady state operation. This
causes improper operation of system.
This issue is solved by employing reset logic for reference signal once the current
requirement reduced to its normal steady state condition. This reset instruction requires 10
to 20 cycles only, depending on difference between initial and updated reference signal
and it improves outcome significantly. Another way to solve this problem is to put a reset
logic, which checks the reference signal in each cycle.

3.4 Testing, Results and Discussion

Simulation and testing has been performed for designed modified switching scheme on the
model of voltage source inverter is shown in figure 3.1. The simulation has been carried
out for linear as well as dynamic loading condition. An R–L load is used for linear loading
condition while for dynamic loading condition a main and auxiliary winding type motor is
used. Simulation is also performed for PWM switching scheme under same loading
condition. The results from PWM as well as modified switching schemes are used for
comparative analysis. The voltage and current waveforms obtained under various loading
condition with both switching scheme are presented in following section. THD graphs of
voltage and current are also presented for analysis. A comparative study also can be made
by data available in table 3.1 and 3.2
The simulation is performed for load current of 50 Hz at output terminal. All the
waveforms presented are recorded under steady state condition. Various loading condition
and other parameter for simulation is as follows.
I. Source voltage (VS) = 200 V II. Source impedance (ZS III. Modulation index for PWM
= 0.8 IV. Modulation frequency for PWM = 1.2 KHz
V. R–L load = 15+ j9.425
VI. Motor loads (main and auxiliary winding motor)
a) Volt-ampere (VA) = 0.25 * 746 watt
b) Voltage rating = 220 V
c) Resistance (pu)
# main winding stator = 0.0311
# main winding rotor = 0.0635
# auxiliary winding stator = 0.110
d) Inductance (pu)
# main winding stator = 0.429 H
# main winding rotor = 0.0325 H
# auxiliary winding stator = 0.0493 H #main
winding mutual inductance= 1.0296 H
3.4.1 Linear load (R –L load)
The waveform of voltage and current for R–L load obtained using PWM scheme is shown
in figure 3.7 and the voltage and current waveform obtained from modified switching
scheme is shown in figure 3.8. THD graphs for voltage and current is in figure 3.9 and
3.10 respectively. The simulation results obtained for given R–L load for new scheme and
PWM are presented in table 3.1. By observing the waveforms and result in table 3.1, a
comparative analysis has been done.
The RMS and peak value of voltage from modified switching scheme is 127.1 V and
179.8 V respectively in respect of 112.8 V and 159.5 V respectively from PWM scheme.
Similarly in case of current the new switching scheme results in 7.177 A RMS value and
10.15 A as peak value while the PWM scheme provide result with 6.367 A as RMS value
and 9.004 A as peak value of current. The modified switching scheme has given better
result than the PWM scheme. RMS and peak value of voltage and current for the modified
switching scheme is more than that for PWM scheme. The THD (total harmonic
distortion) has also shown improvement in case modified switching scheme over PWM
scheme. THD in load current for modified switching scheme is 2.25% and same for PWM
is 3.12%. Thus the harmonic content in load current is less in new defined scheme. THD
for voltage is 38.25% for modified switching scheme in comparison of 77.06% of PWM
scheme.
TABLE 3.1
SIMULATION RESULTS FOR R – L LOAD (SINGLE PHASE) Load Type Parameters
PWM Modified Switching
Vrms 112.8 V 127.1 V Irms 6.367 A 7.177 A Vpeak 159.5 V 179.8 V R-L Load Ipeak 9.004 A
10.15 A

V THD 77.06% 38.25% I THD 3.12% 2.25%

Figure 3.7 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for R-L load using PWM switching scheme.
Similar to case of load current, harmonics content in load voltage is lesser in results
obtained from modified scheme than the results obtained from PWM scheme. The analysis
of outcome of both schemes signifies that the modified switching scheme gives better
result than the PWM scheme in terms of RMS, peak as well as THD value of voltage and
current for given load and operating condition. The observation of RMS and peak values
of current, voltage, values of THDs and their waveforms suggest that the modified scheme
is better than the PWM scheme under given operating conditions.
3.4.2 Dynamic load (Motor load)
The simulation has been performed for motor load also under the previously defined
loading conditions. The operating condition and other parameters have been kept identical
for both schemes under simulation. The motor used as load is a main and auxiliary
winding type motor which specification is described above. This motor has been operated
with constant torque of unity during simulation. Voltage and current waveforms obtained
by applying PWM scheme is shown in figure 3.11 and waveform obtained using modified
switching scheme are shown in figure 3.12. THD graphs from both schemed are presented
in 3.13 and 3.14 respectively for voltage and current. The comparative results of
simulation of both schemes are in table 3.2.

Figure 3.8 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for R-L load using modified switching scheme
Similar to case of R–L load, for motor load also, modified scheme provides higher RMS
and peak value of voltage as well as of current, than the values obtained from PWM
scheme. RMS and peak value of voltage from PWM is 109 V and 154 V respectively
while the same from modified scheme is 151.7 V and 214.5 V respectively. New
switching scheme provides result with 4.852 A RMS value and 6.862 A peak value of
current and same from PWM is 3.062 A and 4.339 A respectively.
3.9 (a)

3.9 (b)
Figure 3.9 THD in load voltage waveform of single-phase H-bridge VS DC–AC with R-L
load using (a) PWM (b) modified switching scheme.

3.10 (a)

3.10 (b)
Figure 3.10 THD in load current waveform of single-phase H-bridge VS DC–AC with R-
L load using (a) PWM (b) modified switching scheme.
The THD in load current is 5.03% for modified scheme which shows improvement over
6.26% of PWM scheme. The 46.33% THD in voltage from modified switching scheme is
also a significant improvement with respect to 77.01% THD from PWM scheme. Amount
of the harmonic distortion is reduced significantly in voltage as well as current when
inverter is operated with new scheme than the PWM scheme. The reduced value of THD
implies lesser amount of harmonics in system and so thus better quality of output power.
TABLE 3.2
SIMULATION RESULTS FOR MOTOR LOAD (SINGLE PHASE) Load Type
Parameters PWM Modified Switching
Vrms 109 V 151.7 V Irms 3.068 A 4.852 A
Vpeak 154 V 214.5 V Motor
Ipeak 4.339 A 6.826 A
V THD 77.01% 46.33%
I THD 6.26% 5.03%

Figure 3.11 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for motor load using PWM scheme
The comparative analysis of voltage waveform for both switching schemes shows that
new switching scheme requires very less amount of switching than the PWM scheme.
This can be considered as an additional advantage of modified switching scheme over
PWM. Similar to linear loading condition, observation of RMS and peak values of current,
voltage, values of THDs and their waveforms suggest that the modified scheme is better
than the PWM scheme under given operating conditions.
Figure 3.12 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for motor load using modified switching scheme.

3.13 (a)

3.13 (b)
Figure 3.13 THD in load voltage waveform of single-phase H-bridge VS DC–AC with
3.14 (a)

3.14 (b)
Figure 3.14 THD in load current waveform of single-phase H-bridge VS DC–AC with

3.5 Summary

Modified hysteresis switching scheme is one of the two advance switching scheme
developed in the book. This chapter is about the modified switching scheme for Single
phase voltage source inverter. This switching scheme lies in category of current regulated
schemes and is a modification over existing hysteresis band current control. The new
scheme has no band, it is a hybrid of unipolar and bipolar switching and also the reference
is based of load current directly and can increase or decrease on its own as soon as there is
any change in load current. Scheme has been developed and tested on SIMULINK
platform of MATLAB. R-L load and single phase motor has been selected as linear and
dynamic load. The results of modified switching are compared with result of PWM
scheme obtained under the identical loading and operating conditions for both types of
loading. The results show that the modified switching scheme has an edge over PWM
scheme in terms of RMS and Peak value of voltage and current. The THD in voltage as
well as in current are found to be lesser in results from new scheme than the results from
PWM.
References

[1]. M. Monfared, H. Rastegar, and H.M. Kojabadi, “Overview of Modulation Techniques

for the Four-Switch Converter Topology,” in IEEE International conference on Power and
Energy, PECon, Malasia, December, 2013.
[2]. A. Hassanpoor, L. Angquist, S. Norrga, and K. Ilves, “Tolerance Band Modulation
Methods for Modular Multilevel Converters,” in IEEE Trans. on Power Electronics, Vol.
30. No. 1, Pg 311-326, August-2014.
[3]. Over Bor-Ren Lin, Yu-Bin Nian and Tung-Yuan Shiau, “Resonant Converter with
Fixed Frequency Control,” in IEEE international Conference TENCON, Xi,an, October
22-25, 2013.
[4]. F. Zare, and A. Nami, “A New Random Current Control Technique for a single-Phase
Inverter with Bipolar and Unipolar Modulations,” in IEEE international conference on
Power Coversion, Nagoya, 2007.

CHAPTER 4
Modified Hysteresis Switching Scheme for 3-phase Voltage
Source Inverter
4.1 Modified Hysteresis Switching Scheme for 3-Phase Inverter

The proposed switching scheme has been designed for three-phase voltage source inverter.
The circuit diagram of three-phase inverter is shown in Figure 4.1. It contains three legs,
each with two switches. All the designing and testing has been performed using the same
circuit diagram. The IGBT has been selected to be used as power electronics switch. The
switches S1 to S6 are provided with gate signal through G1 to G6 respectively. These
pulses follow the applied switching scheme and control the inverter accordingly. The new
scheme is discussed below in detail.

Figure 4.1 Circuit diagram of three – phase voltage source inverter.

4.2 New Switching Strategy

The block diagram of new switching scheme for three-phase inverter shown in figure 4.2
is a modification of same scheme used for single-phase system discussed in chapter 3.
Value of load currents are measured through ammeter and then their values are passed to
switching circuit as feedback. Thus the feedback signals are images of original load
current waveforms. The reference signals for this scheme is sinusoidal signal defined as in
(4.1).

R =R sin(wt) (4.1(a))
Y =Y sin(wt wT 3)(4.1(b))
B =B sin(wt wT 3)(4.1(c))

Where Ra, Ya, Ba are reference signals for R, Y, B phases respectively. Ramax, Yamax,
and Bamax are the maximum amplitude of reference signals respectively.

Figure 4.2 Block diagram for modified switching scheme.

All the three reference signals are images of desired load current in respective phases. The
feedback signal also can be named as Rb, Yb and Bb respectively.
4.2.1 Reference Signals
The reference signal for three-phase system is similar to single-phase system. The only
difference is that in three-phase system we have three reference signal i.e reference for
each phase instead of single reference as in single-phase. The each reference signal
includes two parts as explained in chapter 3.
I. Initial reference signals II. Updated reference signal

4.3 Implementation

S–Function block of SIMULINK facility of MALAB R2011b has been used for designing
the switching logic. The program for switching logic has been written and then converted
in MATLAB executable files. The simulation and testing has been performed using three–
phase voltage source inverter model shown in figure 4.1. The flowchart for the scheme is
shown in figure 4.4.
Figure 4.3 Method of replacing time-interval with respective relation between references.
The reference signal for all three phase is shown in figure 4.3. The figure shows the
reference signal for time-period ‘T’. The time-period ‘T’ can be sub divided in six equal
parts T/6 as shown in figure. During the interval between t0 and t1 the reference Ra is
more than reference Ya, reference Ya is less than reference Ba and reference Ba is more
than or equal to reference Ra. A similar comparison can be observed for other time
intervals. Using this comparison’ logic of switching is defined in such a way that, it can be
applied to obtain output current of any value of frequency, without much change in
programming. Instead of defining logic of switching in terms of time interval, the
respective value of all three references has been used for it. The same has used in
designing flow chart of programming.
The IGBT switches are controlled through the gate pulses generated from switching
circuitry. The ON pulses are represented by logic ‘1’ and similarly OFF pulses by logic
‘0’. The description of the switching scheme is as follows.
Step 1) Initialize the circuit.
Step 2) Define parameters Tm (time-period), Ts (sampling time), t (set time) etc. Step 3)
Define initial reference signals as in (1).
Step 4) If(Ra > )&&(Ya < )&&( ); then to step 5. Else jump to step 9. Step 5) If(Ra > );
then G1 = 1 & G4 = 0. Else G1 = 0 & G4 = 0.
Step 6) If(Ya < ); then G3 = 0 & G6 = 1. Else G3 = 0 & G6 = 0.
Step 7) If(Ba > ); then G5 = 1 & G2 = 0. Else G5 = 0 & G2 = 1.
Step 8) Jump to step 34.
Step 9) If(Ra > )&&(Ya Ba)&&( < ); then to step 10. Else jump to step 14. Step 10) If(Ra
> ); then G1 = 1 & G4 = 0. Else G1 = 0 & G4 = 0.
Step 11) If(Ya < ); then G3 = 0 & G6 = 1. Else G3 = 1 & G6 = 0.
Step 12) If(Ba < ); then G5 = 0 & G2 = 1. Else G5 = 0 & G2 = 0.
Step 13) Jump to step 34.
Step 14) If(Ra Ya)&&(Ya > )&&( < ); then to step 15. Else jump to step 19. Step 15) If(Ra
> ); then G1 = 1 & G4 = 0. Else G1 = 0 & G4 = 1.
Step 16) If(Ya > ); then G3 = 0 & G6 = 1. Else G3 = 1 & G6 = 0.
Step 17) If(Ba < ); then G5 = 0 & G2 = 1. Else G5 = 0 & G2 = 0.
Step 18) Jump to step 34.
Step 19) If(Ra < )&&(Ya > )&&( ); then to step 20. Else jump to step 24. Step 20) If(Ra <
); then G1 = 0 & G4 = 1. Else G1 = 0 & G4 = 0.
Step 21) If(Ya > ); then G3 = 1 & G6 = 0. Else G3 = 0 & G6 = 0.
Step 22) If(Ba < ); then G5 = 0 & G2 = 1. Else G5 = 1 & G2 = 0.
Step 23) Jump to step 34.
Step 24) If(Ra < )&&(Ya Ba)&&( > ); then to step 25. Else jump to step 29. Step 25) If(Ra
< ); then G1 = 0 & G4 = 1. Else G1 = 0 & G4 = 0.
Step 26) If(Ya > ); then G3 = 1 & G6 = 0. Else G3 = 0 & G6 = 1.
Step 27) If(Ba > ); then G5 = 1 & G2 = 0. Else G5 = 0 & G2 = 0.
Step 28) Jump to step 34.
Step 29) If(Ra Ya)&&(Ya < )&&( > ); then to step 30. Else jump to step 34. Step 30) If(Ra
< ); then G1 = 0 & G4 = 1. Else G1 = 1 & G4 = 0.
Step 31) If(Ya < ); then G3 = 0 & G6 = 1. Else G3 = 0 & G6 = 0.
Step 32) If(Ba > ); then G5 = 1 & G2 = 0. Else G5 = 0 & G2 = 0.
Step 33) Jump to step 34.
Step 34) Update the reference signal
Step 35) Calculate t = t + Ts.
Step 36) If (t Tm) ; then reset ‘t’. Else go to 36.
Step 37) Jump to 4.
Figure 4.4 Flowchart for new switching scheme for three-phase voltage source inverter.

4.4 Simulation Results and Discussions

Simulation and testing have been performed on three-phase voltage source inverter as
shown in figure 4.1. Linear as well as nonlinear loading conditions have been used for
testing of designed scheme. The linear loading has been realized through R-L load while
rectifier has been used for nonlinear loading condition. Along with designed switching
scheme, PWM has also been tested and the results of both schemes have been compared
for analytical study.
The simulation has been performed under balanced loading condition for output current of
50 Hz. The results and waveforms presented are recorded under the steady state condition.
The various parameters and loading conditions are as follows:
I. Source voltage (Vdc) = 700 V II. Source impedance (Zdc III. Modulation index for PWM
= 0.8 IV. Modulation frequency for PWM = 1.2 KHz
V. R–L load = 10+ j23.5619
VI. Rectifier load
a) Resistance = 15
b) Capacitance = 100 μf
4.4.1 Linear load
The Simulation and testing has been performed and observations have been made for all
three phases of the output. The above described loading condition has been used for
simulation with both the scheme keeping all other operating conditions identical. The
waveforms of voltage and current for all three phases are presented in figure 4.9 and 4.10
for PWM and modified switching scheme respectively. Since the model is operating under
balanced condition, so result and waveforms for R-phase only are used for comparative
analysis. Waveforms of voltage and current for R-phase of three-phase voltage source
inverter using PWM scheme and modified switching scheme is shown in figure 4.5 and
figure 4.6 respectively. Figure 4.7 and figure 4.8 presents THD in voltage and current
respectively for PWM scheme as well as modified switching scheme. The result of
simulation and testing of both schemes using linear load is in table 4.1.
We can observe that the new switching scheme has performed better than its counterpart
PWM scheme under linear loading condition. The DC side voltage source is set as 700 V.
RMS value of output voltage for R-phase obtained from new scheme is 289.9 V, which is
significantly, much more than 197.9 V, that obtained using PWM scheme. The peak
voltage form PWM scheme is found to be 279.9 V which is much lesser than 410 V from
modified switching scheme. Also the value of RMS current 8.786 A, obtained from new
switching scheme is more than 6.00 A RMS current value obtained from PWM scheme.
Similarly, 12.42 A as peak value of current from modified scheme is respectively a large
value than 8.49 A as peak value from PWM scheme.
New scheme also presents a significant improvement in THD in voltage as well as in
current. The THD in voltage obtained from PWM is 91.60% while same using new
proposed switching scheme is 31.32%. Similarly, THD in output current obtained from
PWM is 2.73% and same for new proposed switching scheme is 1.00%. We can observe
from waveform of THDs that the difference in values of THD is mainly because of
difference in DC component present in output. The DC component is more when inverter
is operated with PWM scheme. In case of proposed scheme the DC components in voltage
and current are 0.08102V and 0.02939 A respectively while in case of PWM, DC
components in voltage and current are 7.388 V and 0.7458 A respectively.
The observation of RMS and peak values of current, voltage, values of THDs and their
waveforms shows, superiority of newly designed modified scheme over the PWM scheme
under given operating conditions.
TABLE 4.1
SIMULATION RESULTS FOR LINEAR LOAD (THREE-PHASE) Load Type
Parameters PWM Modified Switching Vrms 197.9 289.9
Irms 6.004 8.786 Vpeak 279.9 410 Ipeak 8.49 12.42
Linear Load V THD 91.60% 31.32% I THD 2.73% 1.00%
DC component in
Voltage 7.388 0.08102
DC component in
Current 0.7458 0.02939

Figure 4.5 Three-phase VS DC–AC inverter output (load voltage and load current)
waveform of R- Phase under linear loading using PWM scheme.

Figure 4.6 Three-phase VS DC–AC inverter output (load voltage and load current)
waveforms under linear loading using modified switching scheme.
4.7

4.7 (b)
Figure 4.7 THD in load voltage waveforms of three-phase VS DC–AC inverter with linear
load using (a) PWM (b) modified switching scheme.

4.8 Figure 4.8 THD in load current waveforms of Three-phase VS DC–AC inverter with
linear load using (a) PWM (b) modified switching scheme.
4.9 Figure 4.9 Three-phase VS DC–AC inverters output waveform under linear loading
using PWM scheme. (a) Load voltage R, Y and B-phase (b) load current R,Y and B phase.

4.10 (a)
4.10 (b)
Figure 4.10 Three-phase VS DC–AC inverter output waveforms under linear loading
using modified switching scheme. (a) Load voltage R, Y and B-phase (d) load current R,Y
and B-phase.
4.4.2 Nonlinear load
Nonlinear loading conditions for simulation and testing has been realized using rectifier
with pure resistive load with capacitor in parallel. Output waveforms of voltage and
current for all three-phase are shown in figure 4.15 and figure 4.16 for PWM and modified
switching scheme respectively. Similar to case of linear loading here also all comparative
analysis are performed over R- phase only. So, with this purpose, the waveforms of
voltage and current are presented in figure 4.11 for PWM scheme and same for modified
scheme is in figure 4.12. THD in voltage and current waveforms are shown in figure 4.13
and figure 4.14. The result of all simulation results are shown in table 4.3. All these results
are of output voltage and current at output terminal of inverter considering whole rectifier
as load. The current and voltage across resistive load is also presented in figure 4.17 and
4.18 for PWM and modified switching scheme.
Similar to linear load all testing has been performed under balanced three-phase condition.
The DC voltage source for testing and simulation is taken as 700 V. The rectifier load has
been connected to inverter through and inductive coil and values have been monitored
between output port of inverter and inductive coil. The results presented are for R-phase
only.
In this case also the RMS and peak as well as THD values of output voltage and current
obtained from newly designed switching scheme show significant improvement over
PWM scheme. RMS value of voltage is 212 V in new proposed scheme as compared to
196.7 V obtained from PWM scheme. Similar to RMS value, peak value of voltage from
modified scheme shows significant improvement over its counterpart. The peak value
from modified scheme is 299.9 V in respect of same as 278.2 V from PWM scheme. The
RMS value of current is 11.45 A for new proposed scheme which is significantly more
than 11.12 A obtained from PWM scheme. 16.2 A as the peek value of current is shows an
improvement over 15.73A from PWM scheme. THD in voltage waveform are 59.14% and
92.50% for new scheme and PWM scheme respectively. THDs in current waveform are
4.02% and 2.15% from PWM and modified scheme respectively. Similar to case of linear
loading, we can observe from waveform of THDs that the difference in values of THD is
mainly because of the difference in DC component present in output. Here also the DC
component is more when inverter is operated with PWM scheme. The new switching
strategy also shows significant improvement in terms of DC component in voltage and
current. The DC component in voltage and current are 5.832V and 0.9474A for PWM
scheme respectively and same is 0.5593V and 0.0314 in voltage and current waveforms
obtained from modified switching scheme.
Observation of waveform of current trough resistive load of rectifier shows the value of
current is more when inverter is operated with newly designed modified switching
scheme. Here again the observation of RMS and peak values of current, voltage, values of
THDs and their waveforms shows superiority of newly designed modified scheme over
PWM scheme under given operating conditions, as like to previous cases.
TABLE 4.2
SIMULATION RESULTS FOR NON - LINEAR LOAD (THREE-PHASE) Load Type
Parameters PWM Modified Switching Vrms 196.7 212 Irms 11.12 11.45 Vpeak 278.2
299.9 Ipeak 15.73 16.2
Non - linear Load
V THD 92.50% 59.14% I THD 4.02% 2.15%
DC component
in Voltage 5.832 0.5593
DC component
in Current 0.9474 0.0314

Figure 4.11 Three-phase VS DC–AC inverter output (load voltage and load current)
waveform of R- Phase under non- linear loading using PWM scheme.
Figure 4.12 Three-phase VS DC–AC inverter output (load voltage and load current)
waveform of R- Phase under non- linear loading using modified scheme.

4.13 (b)
Figure 4.13 THD in load voltage waveforms of Three-phase VS DC–AC inverter with
nonlinear load (a) PWM (b) Modified switching scheme.
Figure 4.14 THD in load current waveforms of Three-phase VS DC–AC inverter with
nonlinear load (a) PWM (b) modified switching scheme.

Figure 4.15 Three-phase VS DC–AC inverters output waveform under linear loading
using PWM scheme. (a) Voltage of R, Y and B-phase (b) Current in R,Y and B-phase.
Figure 4.16 Three-phase VS DC–AC inverters output waveform under non-linear loading
using modified scheme. (a) Voltage of R, Y and B-phase (b) current R,Y and B-phase.

Figure 4.17 Load voltage and load current waveform for load resistance of rectifier when
supplied from three-phase VS DC–AC inverter with PWM switching scheme. Figure 4.18
Load voltage and load current waveform for load resistance of rectifier when supplied
from three-phase VS DC–AC inverter with modified switching scheme.

4.5 Summary

This chapter is about the modified hysteresis switching scheme for three phase voltage
source inverter. The scheme for the three phases similar to the one for single phase system
but now it has been designed considering the three phases and their relationship with each
other. Similar to single phase, three phase scheme is also has no band, is a hybrid and has
a variable reference. While implementing the switching scheme unlike to single phase
system where time period (or frequency) has been used to define logics, in three phase
system the relation between the all three references have been utilised. This makes the
scheme free from frequency of condition of system and can operate with 50Hz, 60Hz or
any other frequency without any change in programming. The testing of switching scheme
is done over linear load (R-L load) and non-linear load (rectifier load). Here again the
modified scheme has shown and edge over the PWM scheme. THD in voltage and current
have improved in modified scheme than the PWM and also the RMS and peak value of
voltage and current is improved in modified scheme.

References
[1]. Donald Grahame Holmes, Reza Davoodnezhad and Brendan P. McGrath, “An
Improved Three-Phase Variable-Band Hysteresis Current Regulator,” in IEEE Trans. on
Power Electronics, vol. 28, No. 1, January, 2013.
[2]. Mansour Mohseni, Syed M. Islam and Mohammad A.S. Masoum “Enhanced
Hysteresis-Base Current Regulators in Vector Control of DFIG Wind Turbines,” in IEEE
Trans. on Power Electronics, vol.26, No. 1, January, 2011.
[3]. Vinciane Chereau, Francois Auger and Luc Loron, “A new three-phase varying band
hysteresis current controller for voltage-source inverters,” IEEE international conference
on Power Electronics and Drive System, PEDS, Bangkok, November, 2007

CHAPTER 5
Wavelet Modulation for single-phase Inverter
5.1 Modulation Overview

The wavelet modulation for DC-AC single-phase inverter is a process of sampling and
reconstruction of signal. Various elements of modulation are discussed below.
5.1.1. Reference Signal
The output of DC-AC single-phase inverter is sinusoidal wave. In wavelet modulation,
focus is on generating samples of a sinusoidal reference signal and reconstructing
sinusoidal output from these samples. The reference signal (SM) of frequency fM can be
written as:

S =Sin( f t) (5.1)
f = (5.2)
Samples are generated from this reference signal over a time-period of TM. Each two
samples corresponds to one group of samples (i.e. each group of sample of have two
samples) denoted by ‘s’. Thus ‘s’ is the total number of group of samples over time-period
T.
5.1.2. Output Waveform
The output current waveform of DC-AC single-phase voltage source inverter is desired to
be sinusoidal wave. It has been observed in case of PWM, as shown in figure 5.1, that the
rectangular output voltage pulses of inverter gives sinusoidal current waveform and
quality of output current waveform depends upon width of rectangular output voltage
pulses as well as their number. These output pulses are function of gate pulses resulting
from switching logic, which could be PWM, RPWM, SHE, square pulse or one used in
this paper i. e. wavelet modulation.
5.1.3. Sampling and Reconstruction
Sampling and Reconstruction is link between reference signal and output waveform. As
nature of output waveform can be controlled by varying the width of rectangular pulses
and so the sampling and reconstruction also needed to be non-uniform in nature. Thus
only a non-uniform sampling can lead to sinusoidal output. The gate pulses of inverter are
basically ON-OFF signal and ON signal can be modeled from Haar scaling function.
Duration and location of pulses can be controlled by suitable scaling function.
5.1.3.1. Sampling
For single-phase DC – AC inverter, desired output should be as close as possible to
sinusoidal reference signal SM(t) frequency fM which is defined as in (5.1) and (5.2)
This constraint on output can viewed as if each switching action is one stage in
reconstructing reference signal from its non-uniform samples. The groups of non-uniform
recurrent samples formed over one cycle of reference signal can be viewed as a set D as
follows:
D= {d0, d1, d2, … dD-1} (5.3)
Since, dyadic MRA can support only uniform sampling; we need to design a MRA that
can support non-uniform sampling theorem for switching dc – ac inverter [28].
It is to be noted that, for case of dc – ac inverter, the sampled signal are sinusoidal ones
(with a period of Tm seconds) that have a quarter – cycle symmetry property. This
property cause ‘j‘ to increase over the intervals [0,(TM/4) and [(TM/2), (3TM/4)] and to
decrease over intervals [(TM/4), (TM/2)] and [(3TM/4), TM)]. As a result, non-uniform
recurrent sample groups will follow scale j, as it changes over each cycle of SM(t).
5.1.3.2. Reconstruction For reconstruction of sinusoidal signal we need to define
interpolation function as:

j =[ ] (5.4)
j =[ ] (5.5)

Where td1 and td2 are defined as:

td1 = d + 2( ) (5.6)
td2 = d + 1 2( ) (5.7)

Where d= 0, 1, 2, …….., D-1

The interpolation function j and j have their intervals of support such that td1 t td2, where
subscript d indicates that the interpolation is over sample group of d of Xc(t). These
interpolation functions are used to activate switching element of an inverter. It is to be
noted that the properties of switching element used in inverter force their output to be
rectangular pulses. The reconstruction of reference-modulating signal SM(t) from its
nonuniform recurrent samples is carried out by using dilated and shifted version of
synthesis scaling function. Each dilated and shifted version of synthesis scaling function
has an interval of support that given by [td1 td2].
5.1.4 Flowchart
Using the above concept and a detailed analysis, a flow chart of complete process can be
drawn as presented in figure 5.15. Using this flowchart, one can define steps and strategy
to get a modulated output from a single-phase inverter as we are doing in this project.
Similar to single-phase, wavelet modulation can be implemented to three-phase inverters
also [29]-[30]. The designing of switching scheme for voltage source inverter has
performed on MATLAB using S-function block of SIMULINK. The concept of
programming and MATLAB are given in [31].
Figure 5.1 An example to view the relation of output voltage and current (in PWM
method)

5.2 Wavelet Modulation

A wavelet modulation (wavelet transformation) starts with defining wavelet function

(mother wavelet) and scaling function (father wavelet) and using these functions a new
scaling function can be derived as per requirement. Inverter gate pulses are basically
ONOFF pulses, so Haar’s wavelet and scaling function can be used for deriving a scaling
function which can be used further for switching of inverter.
The Haar wavelet function is defined as in (5.8).

1,0 t (t)= 1, t 1 (5.8) 0, otherwise

The Haar wavelet function can be represent as in figure 5.2

Figure 5.2 Haar wavelet function.

Haar Scaling function is defined as in (5.9).

(t)= 1,0 t 1 0, otherwise

Figure 5.3 Haar scaling function.
(5.9)
Haar Scaling function can be represent as in figure 5.3.
Further, wavelet modulation involves multi-resolution analysis (MRA) which is analysis
of space L2(R) consist of a sequence of nested subspaces as in (5.10).

V ………… L (R) (5.10)

As discussed above, using Haar scaling function a new scaling function has been derived
as in (5.11)

(t)= a t + a (t 1+a( ))(5.11)

Where k = 1, 2, 3….
The dual scaling function can be obtained by subtracting it from Haar scaling function.
The other factors are ts1 and ts2 which can be defined as in (5.12) and (5.13).

T =d T + 2 (5.12)
T =d T + T + 2 (5.13)
where s = 1, 2, 3, 4 ….. S
At time ts1, ON pulse start and at ts2, it stops. S is number of samples in one cycle (i.e.
time period Tm). Tp is time of one sample and mathematically can be define as TP = 1/S;

5.3 Modeling and Simulation

Simulation has been performed using SIMULINK facility of MATLAB R2011b, on 1-ph
4-pulse H-bridge voltage source DC-AC inverter. SIMULINK model of the circuit is
shown in figure 5.4. The wavelet modulation has been implemented using S-function
block of SIMULINK. Programming of S-function is done using C language. Inverter is
based on four switch topology using IGBT and switches are controlled by gate signals.
The all four IGBT switches operate according to gate pulses provided to them. When
IGBT.1 and IGBT.2 both are conducting then they generates positive cycle of voltage and
current across the load and similarly negative cycle of voltage and current gets generated
by operating IGBT.3 and IGBT.4. Thus repeated operation of both group one after another
provides an alternating voltage and current across load. The gate signal circuit block of
figure 5.4 has S-function Block inside, this S-Function block is link between programmed
file and simulation model of inverter. Programmed file contains logics required for
generating GATE pulses according to the concept of wavelet modulation. This
programmed file is initially written in C Language and then converted in MATLAB
executable file using command “mex filename”.

5.4 Results and Discussion

To analyze wavelet modulation techniques designed above, simulation has been performed
for various kinds of linear loads. Resistive load (R load) and resistive–inductive load (R–L
load) are used for simulation and testing of new wavelet modulation technique. The
simulation has also been performed for PWM and Square pulse techniques under same
loading condition. Results of all three techniques have been used for comparative analysis
of performance. All operating conditions are kept identical to all three schemes during
simulation process. The simulation tests are performed for current output of 50Hz. The
loading conditions and other parameter used for simulation are as follows-
I. Source voltage (VS) = 50 V
II. Source impedance (ZS
III. R Load = 25
IV. R – L load = 15+ j12.566
V. Modulation index for PWM = 0.8 VI. Modulation frequency for PWM = 1.2 KHz
5.4.1. R – L Load
Simulation test is performed using resistive – inductive load of given value on
SIMULINK model of single-phase H-bridge voltage source DC-AC inverter shown in
figure 5.4, using wavelet modulation by taking 12 groups of samples in each cycle.
Similarly, simulation has been performed for PWM and square pulse methods. The plots
of results obtained from simulation of single-phase H-bridge VS DC–AC inverter under
R-L loading with PWM switching scheme is presented in figure 5.5, same with square
pulse switching is shown in figure 5.6 and figure 5.7 is simulation results for wavelet
switching scheme under same condition. The THD (total harmonic distortion) in voltage
waveform and current waveform are shown in figure 5.8 and 5.9 respectively for all three
switching schemes. A comparative result of performances of all three methods is also
presented in table 5.1.
The RMS value of voltage with PWM scheme is 30.11 V, and with square pulse switching
is 44.83 V while with wavelet modulation scheme is 42.29 V. Similarly peak voltage with
PWM scheme is 42.58 V, with square pulse scheme is 63.41 V and with wavelet
modulation is 59.81 V. Also in terms of THD in voltage waveforms PWM schemes results
with 70.47% THD, square pulse provides 48.50% THD and wavelet modulation results in
47.14% THD. Thus, we can observe that in terms of RMS and peak value of voltage
wavelet modulation give much better results than that obtained from PWM scheme.
Observation of THD in voltage shows that wavelet modulation provided results with
significant improvement in level of THD with respect to PWM and square pulse scheme.
With PWM scheme, RMS value of current is 1.539 A and same with square pulse and
wavelet modulation scheme is 2.291 A and 2.161 A respectively. The peak value of
current is 2.176 A, 3.24 A and 3.057 A respectively for PWM, square pulse and wavelet
modulation schemes. In terms of THD in current PWM scheme provides results with
4.76% and square pulse and wavelet modulation schemes results in 17.73% and 16.93%
THD respectively. Thus, we can observe RMS and peak value of current obtained from
wavelet scheme shows improvement over result obtained from PWM scheme similar to
case of voltage waveform. Current waveform obtained from wavelet schemes results in
lesser THD than the same from square pulse scheme.

Figure 5.4 MATLAB model of single-phase H-bridge voltage source DC – AC inverter

using IGBTs.
In the case of PWM, THD in load current is less than the same in wavelet modulation in
present case. However it is significant point to observe that the voltage THD and output
voltage for wavelet modulation is better than PWM.
TABLE 5.1
SIMULATION RESULT FOR R-L LOAD (SINGLE-PHASE) Load Type Parameters
PWM WM Square Pulse Vrms 30.11 42.29 44.83 Irms 1.539 2.161 2.291
Vpeak 42.58 59.81 63.41 R – L load Ipeak 2.176 3.057 3.24
V THD 70.47% 47.14% 48.51% I THD 4.76% 16.93% 17.73%
Figure 5.5 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for R-L load using PWM switching scheme.

Figure 5.6 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for R-L load using square pulse switching scheme.

Figure 5.7 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for R-L load using wavelet switching scheme.
5.8 (a)

5.8 (b)

5.8 (c)
Figure 5.8 THD in load voltage waveform of single-phase H-bridge VS DC–AC with R-L
load using (a) PWM (b) square pulse (c) wavelet switching scheme.
5.9 (a)

5.9 (b)

5.9 (c)
Figure 5.9 THD in load current waveform of single phase H-bridge VS DC–AC with R-L
load using (a) PWM (b) square pulse (c) wavelet switching scheme.`
5.4.2. R Load
Similar to R–L load, Simulation is also performed with pure resistive load using
SIMULINK model of single-phase H-bridge VS DC–AC inverter shown in figure 5.4. In
this case, also simulations have been performed for all three switching techniques, wavelet
modulation, PWM and square pulse. Performance results of these simulation tests are
shown in table 5.1. The waveform of voltage and current respectively are shown in figure
5.10 for PWM scheme, in figure 5.11 for square pulse scheme and in figure 5.12 for
wavelet modulation scheme.
The RMS and peak value of voltage obtained from PWM scheme is 30.15 V and 42.62 V
respectively. In case of square pulse, RMS value of voltage is 44.83 V, peak voltage is
63.4 V, and with wavelet modulation, RMS and peak value of voltage is 42.25 and 59.75
respectively. THD obtained in voltage waveform is 69.85% for PWM scheme and same is
48.34% for square pulse scheme and 38.56% for wavelet modulation switching strategy.
Thus, the value of THD in voltage waveform shows significant improvement in case of
wavelet modulation in comparison to PWM and square pulse switching schemes. Under
resistive loading condition, PWM scheme results in 1.206 A and 1.705 A respectively as
RMS and peak value of load current. 1.793 A and 1.69 A is RMS value of current from
square pulse and wavelet modulations strategy respectively. The peak value of current for
square and wavelet modulation is 2.536 A and 2.39 A respectively. The value of THD in
current waveform from PWM is 69.85% and same from square pulse and wavelet
modulation is 48.34% and 33.74% respectively.
In comparison to PWM and Square wave, wavelet modulation has given much better
result. THD of current as well as of voltage are much lesser in case of wavelet modulation
than that in PWM and square pulse switching scheme. The RMS and peak value of load
current and voltage have higher value in case of wavelet modulation than that in case of
PWM. Since, nature of load is pure resistive, we observe identical value of THD in current
and voltage waveform for any selected switching scheme.
TABLE 5.2
SIMULATION RESULT FOR R LOAD (SINGLE-PHASE) Load Types Parameters
PWM WM 1 Pulse Vrms 30.15 42.25 44.83 Irms 1.206 1.69 1.793
Vpeak 42.62 59.75 63.4 R load Ipeak 1.705 2.39 2.563 V THD 69.85% 33.74% 48.34% I THD
69.85% 33.74% 48.34%

Figure 5.10 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for R load using PWM switching scheme.
Figure 5.11 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for R load using square pulse switching scheme.

Figure 5.12 Single-phase H-bridge VS DC–AC inverter test result (load voltage and load
current waveform) for R load using wavelet switching scheme.
5.13 (a)

5.13 (b)

5.13 (c)
Figure 5.13 THD in load voltage waveform of single-phase H-bridge VS DC–AC with R
5.14 (a)

5.14 (b)

5.14 (c)
Figure 5.14 THD in load current waveform of single-phase H-bridge VS DC–AC with R
Figure 5.15 Flowchart of wavelet modulation for single-phase inverter.

5.5 Summary
The Wavelet Modulation scheme has been developed for switching of single phase voltage
source inverter. Similar to any modulation, it also require a reference single, which then
get sampled using the sampling logic and finally the signal is reconstructed from these
samples. A sinusoidal signal has been considered as reference signal as the desired load
current of any voltage source inverter is a sinusoid. Sampling and reconstruction of the
signal requires the wavelet function, sampling function and MRA. Since the gate pulses of
any power electronic switch is either on or off. This on-off signal can be obtained from
HAAR scaling function. An advance scaling function has been derived from the HAAR
scaling function. This derived function defines the locations of on and off pulses to the
switches. This scheme has been tested on R and R-L load and results are compared with
results from PWM and square pulse scheme under the identical operating conditions. In
terms of RMS and Peak values of voltage and current the results from wavelet modulation
lies in between the PWM and square pulse switching scheme. It has shown an edge over
square pulse switching scheme. In terms of THD, the results from wavelet scheme have
shown improvement over both PWM and square pulse switching scheme. The
performance of wavelet modulation scheme depends on derived scaling function. A
different derived scaling function may give different result.
References

[1]. S. A. Saleh, C. R. Moloney and M. A. Rahman, “Development and Testing of Wavelet

Modulation for Single-Phase Inverters,” IEEE Trans. Ind. Electronics, vol. 56, no. 7, pp.
2588 – 2599, July 2009.
[2]. S. A. Saleh and M.A. Rahamn, “Performance testing of single phase voltage-source
wavelet modulated AC-DC converter”, in IEEE International conference on Electrical and
Computer Engineering, CCECE, St. John’s, NL, Canada, May 2009.
[3]. W. G. Morsi and M.E. El-Hawary, “Suitable Mother Wavelet for Harmonics and
Interharmonics Measurements Using Wavelet Packet Transform”, in IEEE International
conference on Electrical and Computer Engineering, CCECE, Vancouver, BC, Canada,
April 2007.