Studies on Power Quality

and
An LSE & PSOPC Based Harmonic Estimation to Real
Time Data


THESIS

Submitted in Partial Fulfillment of the Requirements for the Degree of


BACHELOR OF TECHNOLOGY
in
ELECTRICAL ENGINEERING



by
AMIT KUMAR SHUKLA
10502025


under the guidance of
PROF. SHARDINDU GHOSH
and
MR. SANDIP GHOSH




Department of Electrical Engineering
NATIONAL INSTITUTE OF TECHNOLOGY
ROURKELA

2009

ii



Certificate
This is to certify that the thesis entitled, Studies on Power Quality and; An LSE &
PSOPC based Harmonic Estimation to the Real Time Data submitted by Shri Amit
Kumar Shukla in partial fulfillment of the requirements for the award of Bachelor of
Technology degree in Electrical Engineering at the National Institute of Technology
Rourkela (Deemed University), is an authentic work carried out by him under my supervision
and guidance.
To the best of my knowledge the matter embodied in the thesis has not been submitted to any
other University/Institute for the award of any degree or diploma.


Date:
Place: Rourkela
Mr. Sandip Ghosh
Department of Electrical Engineering
National Institute of Technology Rourkela



iii

Abstract
Electric power as a commodity is very different from other products it is generated far from the
point of use, is fed to the grid together with the output of many other generators and arrives at the
point of use via several transformers and many kilometers of overhead and possibly underground
cabling. Quality check of power is very important. But it has constantly been polluted by both
consumers (non-linear loads) and utility companies (due to increased use of semi conductor based
power processors). Assuring the quality of delivered power to be good enough for the application at
hand calls for measurement and study of power quality at the point of common coupling.
There are various categories of Power Quality issues. Issues related to voltage amplitude variations
will include voltage sags/dips and voltage swells. Issue related to frequency is harmonics in the power
system. Short term problems include transients and momentary interruptions. We have carried out a
brief study of all these PQ issues, and problems caused by them.
Developments in the field of power electronics has brought tremendous power processing capability
to engineers but it has also polluted the power system with harmonics. Then there are non linear loads
that inject harmonics in the power system. Measurement and estimation of harmonics in power system
has become one of the most important tasks in the hands of power system engineers for increasing
reliability and quality of supplied power. For harmonic analysis FFT based schemes are unable to give
enough information for PQ mitigation. So researchers have been attempting other methods to estimate
harmonics in sampled voltage and current signals. Among these; least square estimation, for
estimating amplitudes and particle swarm optimization with passive congregation, for estimating
phase and frequency used together has most attractive features of compactness and fastness. We have
developed a harmonic estimator based on LSE and PSOPC to estimate amplitude, phase and
frequency of the sampled signal. We have used the estimator on real time data acquired using a digital
storage oscilloscope in Electrical Machines Laboratory of Department of Electrical Engineering at
NIT Rourkela. The results for estimating integral harmonics are attractive and encourage researchers
to modify the estimator code to make it able to estimate inter-harmonics and sub-harmonics also.

iv

Acknowledgements
I have been very fortunate to start my thesis work under the supervision and guidance of Prof.
Shardindu Ghosh. He introduced me to the field of Power Quality, educated me with the methods and
principles of research, and guided me patently until he was here in NIT Rourkela. He is the whole
Philosopher and Guide behind this thesis. I am extremely thankful to him for giving me the special
attention and support in between his very busy hours as the Head of the Department. Working with
him, a person of values has been a rewarding experience.
At this stage, I would like to express my sincere gratitude to Mr. Sandip Ghosh for helping me in
carrying on with the thesis work. I am highly indebted and express my deep sense of gratitude for his
invaluable guidance, constant inspiration and motivation with enormous moral support during
difficult phase to complete the work. I acknowledge his contributions and appreciate the efforts put by
him for helping me complete the thesis.
I would like to take this opportunity to thank Prof. B. D. Subudhi, the Head of the Department for
letting me use the laboratory facilities for my project work. I am thankful to him for always extending
every kind of support to me.
I am thankful to Prof. P. C. Panda, Prof. A. K. Panda, Mr. S. Mohanty, Mr. B. Chitti Babu, and Dr.
Sharmili Das for their constructive criticism and for extending their support in my project.
At this moment I would also like to express my gratitude for the technical staff of our laboratories.
They have always helped me in every-way they can during my experimental phase of the work.



1

Contents

Cover Page i
Certificate ii
Abstract iii
Acknowledgements iv

List of Figures ............................................................................................................................ 4
List of Tables ............................................................................................................................. 4
Chapter 1 Introduction ............................................................................................................... 5
1.1 Motivation of the Thesis .................................................................................................. 6
1.2 Organization of the Thesis ............................................................................................... 7
Chapter 2 Introduction to Power Quality ................................................................................... 9
2.1 Definitions........................................................................................................................ 9
2.2 Power Quality: The Buzzword ....................................................................................... 10
2.3 Power Quality Defects ................................................................................................... 11
2.3.1 Pictorial representation of power quality events ..................................................... 14
2.3.2 Causes of power quality defects ............................................................................. 17
2.4 Typical Problems Due to Poor Power Quality ............................................................... 18
2.5 The Cost of Poor Power Quality .................................................................................... 24
2.5.1 Harmonic distortion ................................................................................................ 25
2.5.2 Blackouts ................................................................................................................. 26
2.5.3 Dips or sags ............................................................................................................. 27
2.5.4 Transients ................................................................................................................ 30
2.6 Solutions Power Quality Issues ..................................................................................... 31
2.7 Standards ........................................................................................................................ 33
2.8 Conclusion ..................................................................................................................... 35
2.9 References ...................................................................................................................... 36
Chapter 3 Harmonics ............................................................................................................... 37
3.1 Causes of Harmonic Distortion ...................................................................................... 38
3.2 Problems Caused by Harmonics .................................................................................... 39


2

3.3 Conclusion ..................................................................................................................... 42
3.4 References ...................................................................................................................... 42
Chapter 4 Fourier analysis for Harmonics ............................................................................... 44
4.1 Theoretical Background ................................................................................................. 44
4.2 Fourier series .................................................................................................................. 46
4.3 Fourier Transform .......................................................................................................... 49
4.4 Discrete & Fast Fourier Transform ................................................................................ 50
4.5 Power Quality Terms Related to Harmonics ................................................................. 53
4.5.1 Total harmonic distortion (THD) ............................................................................ 53
4.5.2 Other harmonic PQ Terms ...................................................................................... 54
4.6 References ...................................................................................................................... 56
Chapter 5 Survey on Harmonic Estimation ............................................................................. 57
5.1 Harmonic Measurements ............................................................................................... 58
5.1.1 Basic equipments used for the analysis of non-sinusoidal voltages and currents ... 59
5.1.2 Requirements for instrument response .................................................................... 60
5.2 Harmonic Analysis of Sampled Data ............................................................................. 61
5.2.1 Application of FFT for harmonic analysis .............................................................. 62
5.2.2 Application of Kalman filtering algorithm for optimal harmonic estimation ......... 63
5.2.3 Application of wavelet transform for harmonic estimation .................................... 64
5.2.4 Application of artificial intelligence for harmonic estimation ................................ 64
5.3 Conclusion ..................................................................................................................... 65
5.4 References ...................................................................................................................... 66
Chapter 6 A Harmonic Estimation Technique based on LSE & PSOPC ................................ 68
6.1 Introduction to the method of Least Square Estimation ................................................ 68
6.1.1 Principle of least squares ........................................................................................ 68
6.1.2 Method of least squares: ......................................................................................... 68
6.2 Introduction to the Particle Swarm Optimization based Techniques ............................. 72
6.2.1 Standard Particle Swarm Optimizer (SPSO) .......................................................... 73
6.2.2 Particle Swarm Optimizer with Passive Congregation (PSOPC) ........................... 76
6.3 LSE and PSOPC based Harmonic Estimation Technique ............................................. 80
6.4 References ...................................................................................................................... 84
Chapter 7 Real Time Data Collection for the Purpose of Harmonic Estimation ..................... 87
7.1. Aim of the Experiment .................................................................................................. 87


3

7.1.1 Equipments used ..................................................................................................... 87
7.1.2 Circuit diagram of the experimental setup .............................................................. 89
7.1.3 Procedure followed ................................................................................................. 90
7.1.4 Observations ........................................................................................................... 90
7.1.5 Conclusions from the images .................................................................................. 92
Chapter 8 Application of the LSE & PSOPC Based Harmonic Estimation Technique to the
Real Time Data ........................................................................................................................ 94
8.1 Output of the script estimating odd harmonics .............................................................. 94
8.2 Output of the script estimating all (even +odd ) harmonics ........................................ 101
8.3 Conclusion ................................................................................................................... 106
8.4 Scope of Future Work .................................................................................................. 107
Appendix ................................................................................................................................ 108
Matlab Code for LSE+PSOPC Based Harmonic Estimation Technique .......................... 108
Snapshots of the Experimental Setup ................................................................................ 122
Bibliography .......................................................................................................................... 123




4

List of Figures
Figure 2-1: Most prevalent PQ problems, measured at 1,400 sites in 8 European countries by
European Copper Institute ....................................................................................................... 19
Figure 2-2: ITIC curve ............................................................................................................. 29
Figure 2-3: Losses due to voltage dips .................................................................................... 29
Figure 2-4: Most prevalent PQ solutions, in terms of % adoption rate at 1,400 sites in 8
European countries ................................................................................................................... 33
Figure 4-1: Fundamental with 3rd & 5th Harmonics ............................................................... 44
Figure 4-2: Distorted current waveform .................................................................................. 45
Figure 6-1: Interaction between particles and the best particle gbest ...................................... 73
Figure 6-2: Interactions of particles with passive congregation .............................................. 78
Figure 6-3: Flow chart for LSE & PSOPC based Harmonic Estimation Technique ............... 83
Figure 7-1: Circuit diagram of the experimental setup ............................................................ 89
Figure 8-1: Spectrum of data sample no. 1 .............................................................................. 96
Figure 8-2: Estimation Error for data sample no. 1 ................................................................. 96
Figure 8-3: Spectrum of data sample no. 2 .............................................................................. 97
Figure 8-4: Estimation error for data sample no. 2 .................................................................. 97
Figure 8-5: Spectrum of data sample no. 3 .............................................................................. 98
Figure 8-6: Estimation error in data sample no. 3 ................................................................... 98
Figure 8-7: Spectrum for data sample no. 4 ............................................................................. 99
Figure 8-8: Estimation error for data sample no. 4 ................................................................ 100
Figure 8-9: Spectrum for data sample no. 5 ........................................................................... 103
Figure 8-10: Estimation error for data sample no. 5 .............................................................. 103
Figure 8-11: Spectrum for data sample no. 6 ......................................................................... 105
Figure 8-12: Estimation error for data sample no. 6 .............................................................. 105

List of Tables
Table 5-1: Minimum Required Attenuation (dB) .................................................................... 61
Table 6-1: Pseudo-code for the PSOPC algorithm .................................................................. 79
Table 7-1: Acquired Wave form and Spectrum images ........................................................... 91
Table 0-1: Snapshots of Experimental Setup ......................................................................... 122




5


Chapter 1 Introduction

Electrical power has become the 4
th
essential element for survival of our modern civilized
world after bread, clothing and shelter. Our civilization has commoditized all these elements.
What makes electric power different from other commodities is that it requires a continuous
flow - it cannot be conveniently stored in quantity - and it cannot be subject to quality
assurance checks before it is used. For commodity of this type, the reliability of the supply
must be known and the pliability of the supply dependent processes must be understood.
Electric power is actually very different from any other product it is generated far from the
point of use, is fed to the grid together with the output of many other generators and arrives at
the point of use via several transformers and many kilometers of overhead and possibly
underground cabling. Where the industry has been deregulated, these network assets will be
owned, managed and maintained by a number of different organizations. Assuring the quality
of delivered power at the point of use is no easy task and there is no way that sub-standard
electricity can be withdrawn from the supply chain or rejected by the customer.
From the consumers point of view the problem is even more difficult. There are very limited
statistics available on the quality of delivered power, but the acceptable quality level as
perceived by the supplier (and the industry regulator) may be very different from that
required, or perhaps desired, by the consumer. The most obvious power defects are complete
interruption (which may last from a few seconds to several hours) and voltage dips or sags
(where the voltage drops to a lower value for a short duration). Naturally, long power


6

interruptions are a problem for all users, but many operations are very sensitive to even very
short interruptions. Examples of sensitive operation are:
Continuous process operations, where short interruptions can disrupt the
synchronization of the machinery and result in large volumes of semi-processed
product. A typical example is the paper making industry where the clean-up operation
is long and expensive.
Multi-stage batch operations, where an interruption during one process can destroy
the value of previous operations. An example of this type is the semiconductor
industry, where the production of a wafer requires a few dozen processes over several
days and the failure of a single process is catastrophic.
Data processing, where the value of the transaction is high but the cost of processing
is low, such as share and foreign exchange dealing. The inability to trade can result in
large losses that far exceed the cost of the operation.
These are examples of the most sensitive industries, but it is surprising how many apparently
mundane operations have quite critical power supply requirements. Examples include large
retail units with computerized point of sale and stock control equipment and manufacturing
plant with distributed control.
1.1 Motivation of the Thesis
The importance of clean power has become very vital these days. But which type of power
can be said as clean power may depend upon ones reference frame. The same power can be
clean for one user and polluted for the other. So it is very necessary to decide on the criteria
of power quality. In this regard one need to get familiar with various Power Quality issues in
general. So study of power quality is very important.


7

The rapid developments in the field of high capacity semi-conductor devices has on one hand
brought increased power processing capability to our service (which we wished for) and on
the other hand it has polluted the power system a lot by injecting current harmonics to it
(which we never wished for). Apart from power processors, there are many other non linear
loads that inject harmonic currents. The irony of situation is that the device that injects
harmonics in the power system is most affected by the polluted power. These power
processors are very sensitive to harmonics and they mal-operate in presence of harmonics,
sometimes creating havoc. If control scheme of these devices is not operating perfectly, they
inject non-characteristic harmonics, and sometimes inter-harmonics and sub-harmonics also.
So cognition of these harmonics is very essential for the purpose of successful mitigation.
Harmonic estimation thus, is very important.
1.2 Organization of the Thesis
The thesis has two major goals 1) study of power quality, and 2) Harmonic Estimation in
sampled data. We have divided the thesis in two sections, first section (chapter 2) deals with
introduction to power quality, and second section (chapter 3 onwards) is related to
measurement and estimation of harmonics in power system.
Chapter 2 gives introduction to power quality. Various power quality issues, problems caused
by them and solution to these problems have been discussed in brief in this chapter. Criteria
to estimate cost associated with poor power quality have also been given.
Chapter 3 focuses on the harmonics in power system in detail and establishes the importance
of knowing the degree of harmonic pollution in the system
Chapter 4 gives classical methods of harmonic estimation. This chapter also includes some
popular measures of harmonic content.


8

Chapter 5 is a survey on harmonic estimation problem. It has been found that particle swarm
optimization with passive congregation has quite attractive features to be considered for
estimating non-linear parameters of the harmonic estimation problem.
Chapter 6 has the documentation of least squares and particle swarm optimization (PSO)
techniques. It also documents the LSE and PSOPC based harmonic estimation technique
towards the end of the chapter.
Chapter 7 is devoted to the method adopted for collecting real time data for harmonic
analysis to be carried out using the technique based on LSE and PSOPC.
Chapter 8 is devoted to the results obtained by applying the harmonic estimator to the real
data obtained during the experiment. In the last we have given the conclusions.



9


Chapter 2 Introduction to Power Quality

What do we mean by power quality? : A perfect power supply would be one that is always
available, always within voltage and frequency tolerances, and has a pure noise free
sinusoidal wave shape. Just how much deviation from perfection can be tolerated depends on
the users application, the type of equipment installed and his view of his requirements.
2.1 Definitions
Institute of Electrical and Electronic Engineers (IEEE) Standard 1100 defines power quality
as the concept of powering and grounding sensitive electronic equipment in a manner
suitable for the equipment.
Mr. C. Sankaran in his book Power Quality published by CRC Press [2.1] says that Power
quality is a set of electrical boundaries that allows a piece of equipment to function in its
intended manner without significant loss of performance or life expectancy. This definition
embraces two things that we demand from an electrical device: performance and life
expectancy.
Dugan et. al. in their book Electrical Power Systems Quality (Second edition) [2.2] use the
following definition of PQ, Any power problem manifested in voltage, current, or frequency
deviations that results in failure or mal-operation of customer equipment.


10

There can be completely different definitions for power quality, depending on ones frame of
reference. For example, a utility may define power quality as reliability and show statistics
demonstrating that its system is 99.98 % reliable. Criteria established by regulatory agencies
are usually in this vein. A manufacturer of load equipment may define power quality as those
characteristics of the power supply that enable the equipment to work properly. These
characteristics can be very different for different criteria. Power quality is ultimately a
consumer-driven issue, and the end users point of reference takes precedence.
2.2 Power Quality: The Buzzword
The term power quality has become one of the most prolific buzzwords in the power industry
since the late 1980s. It is an umbrella concept for a multitude of individual types of power
system disturbances. The issues that fall under this umbrella are not necessarily new. What is
new is that engineers are now attempting to deal with these issues using a system approach
rather than handling them as individual problems [2.2].
The phenomenon of disturbances caused by power system equipments resulting in a non-
ideal waveform of the supply voltage is as old as the power network itself. First
considerations made at the end of the 19
th
century identified transformers and rotating
machinery as the main source of the waveform distortion. The development of technology
over the years, especially the growth of the use of switched power semiconductor devices has
brought about many technical conveniences and economical profits, but on the other hand it
has exerted a continual growing influence on the power systems operating conditions. This
ever increasing process is stimulated mostly by the way the power system has changed under
new challenges imposed by the prescriptions on the energy market introduced by many
countries. The resulted thereby growing share of renewable energy sources in power systems,


11

which rely heavily on sophisticated power electronic equipments to operate properly. We see
the following four main reasons of increased concern:
i. Newer-generation load equipment, with microprocessor-based controls and power
electronic devices, is more sensitive to power quality variations than was equipment
used in the past.
ii. The increasing emphasis on overall power system efficiency has resulted in continued
growth in the application of devices such as high-efficiency, adjustable-speed motor
drives and shunt capacitors for power factor correction to reduce losses. This is
resulting in increasing harmonic levels on power systems and has many people
concerned about the future impact on system capabilities.
iii. End users have an increased awareness of power quality issues. Utility customers are
becoming better informed about such issues as interruptions, sags, and switching
transients and are challenging the utilities to improve the quality of power delivered.
iv. Many things are now interconnected in a network. Integrated processes mean that the
failure of any component has much more important consequences.
The ultimate reason that we are interested in power quality is economic value. There are
economic impacts on utilities, their customers, and suppliers of load equipment.
2.3 Power Quality Defects
Power quality defects the deviations from perfection in the wave shape fall into five
categories:
Harmonic distortion
Interruptions/Blackouts


12

Under or over voltage
Dips (or sags) and surges
Transients
The issue of short interruptions and voltage dips highlights the difference in perspective
between supplier and customer. They are by definition short term events so that unless there
is a permanent monitor installed the very existence of the event is difficult to prove. It is even
more difficult to attribute a business loss to a particular event. The electricity supply industry
tends to value an interruption in terms of the cost of the electricity that was not supplied as a
result, while the consumer values it in terms of the revenue lost as a consequence of the break
in production. Electricity is relatively cheap and the supply interruption relatively short, while
lost production can be very valuable (as in the case of semiconductor industry) and the
downtime very long to allow for clean up (as in the paper making industry). The two parties
therefore have completely different views of the importance of voltage dips and on the level
of investment in reduction equipment that is justified.
Longer interruptions power cuts are usually thought of as being caused by the supplier but
can also be caused by the failure of on-site equipment, conductors and connections. Careful
design using high resilience techniques can minimize the effects. The objective is to identify
single points of failure and eliminate them by providing redundant equipment or alternative
supply paths so that operation can continue despite a single failure. Systems designed in this
way are easier to maintain and are better maintained as a result. It is important that
maintenance procedures are developed at an early stage as part of the resilient design
concept. Standby generation and UPS systems, required to cover short and longer term power
cuts are essential elements of a resilient system.


13

While the majority of voltage dips and interruptions originate in the transmission and
distribution system and are the responsibility of the supplier, harmonic problems are almost
always the responsibility of the consumer. It is harmonic currents that cause problems in
installations and when these currents flow back into the supply impedance at the point of
common coupling, a harmonic voltage is developed. This voltage distortion, or at least some
components of it, are distributed around the system and are combined with the background
harmonic voltage distortion present in any transmission system (due to the non-linearity of
transformers for example). By limiting the harmonic current consumers are permitted to
draw, the level of voltage distortion on the supply is kept within acceptable limits.
Determining the source of harmonic distortion can be difficult and this often leads to
consumers blaming the supplier for the problem. In fact, it is unusual for harmonic problems
within an installation to arise from external causes the cause is almost always due to the
equipment on site and the installation practice used.
Transient disturbances are high frequency events with durations much less than one cycle of
the supply. Causes include switching or lightning strikes on the network and switching of
reactive loads on the consumers site or on sites on the same circuit. Transients can have
magnitudes of several thousand volts and so can cause serious damage to both the installation
and the equipment connected to it. Electricity suppliers and telecommunications companies
go to some effort to ensure that their incoming connections do not allow damaging transients
to propagate into the customers premises. Nevertheless, non-damaging transients can still
cause severe disruption due to data corruption. The generation and influence of transients is
greatly reduced and the efficacy of suppression techniques greatly enhanced where a good
high integrity earthing system has been provided. Such an earthing system will have multiple
ground connections and multiple paths to earth from any point, so ensuring high integrity and
low impedance over a wide frequency band.


14

Power quality problems present designers with many questions, perhaps the greatest of which
is, How good is good enough? This question is impossible to answer. While it is relatively
simple to quantify the behavior of a particular piece of equipment to voltage dips,
determining the likely incidence of voltage dips at a particular location on the supply system
is rather more difficult; it will change over time as new consumers are added and assets
replaced. It is extremely difficult to collect any meaningful data on the sensitivity of
equipment to harmonic voltage distortion, and even on the harmonic current distortion caused
by equipment. The real question is one of compatibility between the equipment and the
supply.
There are some international standards available that set limits of voltage variation and
harmonic voltage distortion below which equipment should function without error. Similarly,
there are standard limits for voltage deviation and harmonic voltage distortion of the supply.
Ideally, there should be a guard band a safety margin between the two limits but because
supply quality is difficult to measure on a continuous basis, the supply limits are set in
statistical terms and not as hard limits.
2.3.1 Pictorial representation of power quality events
This section gives some waveforms and identifies power quality defects in them. This gives a
very quick knowledge about the power quality (PQ) issues/events.


Dist
Volt
torted Wave
15
tage Sag

eforms (Har

rmonics)





OOsciIatory trransient wav
Voltag
veform
O
16
ge Notching

Im
Outage

mpulsive Trransient Wa


aveform



2.3.2 Ca
Each of t
shared in
some cus
problem
same sub
installatio
Harmoni
proven re
Nonlinea
the comin
nonlinear
load will
severe th
energy po
auses of pow
these power
nfrastructure
stomers and
on one cust
bsystem. Ot
on and may
c problems
eduction equ
ar equipment
ng years it c
r character.
l decrease th
han it is toda
ollution are
C
wer quality
quality prob
e. For examp
d the higher
tomers site
ther problem
y or may not
can be dea
uipment.
ts have beco
can be expec
Considering
he problem o
ay. This is w
causing som
Categories o
y defects
blems has a d
ple, a fault o
the level of
may cause a
ms, such as
t propagate
lt with by a
ome a part o
cted that mo
g the fact th
of maintainin
why harmonic
me alarm. Alw
17
of PQ Varia
different cau
on the netw
f the fault, t
a transient th
s harmonics
onto the net
a combinatio
of the world
st of the loa
hat at the dis
ng the prope
cs, inter-harm
ways presen
ations
use. Some pr
work may ca
the greater t
hat affects a
s, arise with
twork and s
on of good
s directives
ds in distrib
stribution le
r quality of
monics and
nt in electrica
roblems are
ause a dip th
the number
all other cust
hin the cus
so affect oth
design prac
for energy
ution system
evel the shar
supply will
other condu
al networks,

a result of th
hat will affe
affected, or
tomers on th
stomers ow
her customer
ctice and we
saving and i
ms will have
re of resistiv
become mor
ucted forms o
they are no

he
ct
a
he
wn
rs.
ell
in
a
ve
re
of
w


18

proliferating to epidemic proportions. As a result, new challenges for power system operation
studies have been created giving a decisive impulse to the growing interest in the wide power
quality area, which is spread from fields relating to the problems of EMC, like emission or
compatibility limits and voltage or current characteristics to the traditional power system
areas of voltage continuity studies in terms of reliability and availability considerations.
2.4 Typical Problems Due to Poor Power Quality
The following are snippets from a study carried out by European Copper Institute in 2001,
covering 1,400 sites in 8 countries, any given site in Europe has a 5-20 % probability that it
will suffer from one or more of the PQ problems listed below. Typically, half of sites in
energy-intensive industries or mission-critical office buildings will suffer from two or more
problems. Very few sites are trouble free. Figure 2-1 gives the frequency distribution of the
most common PQ problems.
Of course, poor power quality is not the cause of every occurrence of the problems listed. For
example, computer lockups can be software related. In addition, attribution of the origin of
the problem to causes either before the meter (i.e. on the suppliers side of the point of
common coupling (PCC)) or after the meter (i.e. on the customers installation side of the
PCC) is often difficult without detailed measurement and analysis.
COMPUTER LOCKUP
Earth current originating in the equipment, results in a voltage drop between the equipment
and true earth. Although small, this noise voltage may be significant as compared to the
signal voltages (of a few volts) on which IT equipment operates. PC hardware is designed to
minimize sensitivity to this kind of disturbance but it cannot be eliminated entirely, especially
as the noise frequency rises. Modern communications protocols have error detection and

correctio
conseque
a frequen
In a TN-
voltage d
longer at
connecte
Figure 2-1
Copper In
F
Triple-n
and prote
building.
n algorithm
ently reducin
nt phenomen
-C network,
drops. The
t the same po
d to earth at
1: Most preva
nstitute
FLICKERIN
harmonic cu
ective condu
As a result
ms built in, r
ng the data th
non in today
the combin
earth refere
otential. Cur
t both ends fo
alent PQ prob
NG SCREEN
urrents sum
uctor are com
t, neutral retu
requiring ret
hroughput. A
s office env
ned neutral-e
ence plane o
rrents will fl
for EMC com
blems, measur
NS
in the neutra
mbined and
urn currents
19
transmission
As a result, P
vironments.
earth conduc
of different
low, for exam
mpliance.
red at 1,400 si
al conductor
connected i
s can flow an
n of erroneo
PCs will oft
ctor actively
computers
mple along t
ites in 8 Euro
r. In a TN-C
n many plac
nywhere in t
ously receive
en slow dow
y carries cur
on different
the shields o
opean countrie
configuratio
ces to the str
the metal str
ed data - an
wn or lock-up
rrent, creatin
t floors is n
of data cable
es by Europea
on the neutr
ructure of th
ructure of th

nd
p,
ng
no
es,

an
ral
he
he


20

building and create uncontrolled and uncontrollable magnetic fields. In extreme cases, these
fields can result in flicker of computer screens. Neutral current always needs to be returned to
the point of common coupling using a separate conductor as in the TN-S and TN-C-S
systems. In fact, the discipline of having one and only one neutral-earth connection point in
the installation improves safety and EMC.
FLICKERING LIGHTS
Short duration voltage changes; resulting from switching, short-circuits and load changing
can result in light flicker. The permissible magnitude of light flicker is regulated by
International Standards, based on perception criteria. Excessive flicker can cause migraine
and is responsible for some instances of the so called sick building syndrome.
OVERHEATING OF TRANSFORMERS AT MODERATE LOAD
Harmonics cause additional losses in the transformer. When the transformer is close to
maximum load, these losses can lead to early failure due to overheating and hot spots in the
winding. With the current trend to push equipment harder to its limits, and the increasing
harmonic pollution in low-voltage networks, this problem is occurring ever more frequently.
Losses in transformers are due to stray magnetic losses in the core, and eddy current and
resistive losses in the windings. Of these, eddy current losses are of most concern when
harmonics are present, because they increase approximately with the square of the frequency.
In a typical mixed load building the transformer eddy current losses will be about 9 times
higher than would be expected, approximately doubling the total load losses. Before the
excess losses can be determined, the harmonic spectrum of the load current must be known.



21

INDUCTION MOTORS
Voltage harmonics cause extra losses in direct line-connected induction motors. The 5th
harmonic creates a counter-rotating field, whereas the 7th harmonic creates a rotating field
beyond the motors synchronous speed. The resulting torque pulsing causes wear and tear on
couplings and bearings. Since the speed is fixed, the energy contained in these harmonics is
dissipated as extra heat, resulting in premature ageing. Harmonic currents are also induced
into the rotor causing further excess heating. The additional heat reduces the rotor/stator air
gap, reducing efficiency even further.
Variable speed devices cause their own range of problems. They tend to be sensitive to dips,
causing disruption of synchronized manufacturing lines. They are often installed some
distance from the motor, and cause voltage spikes due to the sharp voltage rise times.
Special care has to be taken at start-up of motors after a voltage dip when the motor is
normally operating at close to full load. The extra heat from the inrush current at start-up may
cause the motor to fail. Optimum sizing of motors should take into account:
that the motor has been designed to run at maximum efficiency at about 70 % load
frequency of voltage dips, and time one can afford to wait to resume motor operation.

OVERHEATING OF CONDUCTORS DUE TO SKIN EFFECT
All harmonics cause additional losses in the phase conductors. The skin effect, which is
negligible at 50 Hz, starts to play a role from 350 Hz (7th harmonic) and upwards. For
example, a conductor with 20 mm diameter has 60 % more apparent resistance at 350 Hz
than its dc-resistance. The increased resistance, and even more, the increased reactance (due


22

to higher frequency), will result in an increased voltage drop and an increased voltage
distortion.
CORRECT FUNCTIONING OF PROCESS CONTROL EQUIPMENT
Severe harmonic distortion can create additional zero-crossings within a cycle of the sine
wave, affecting sensitive measurement equipment. Synchronization of process control
equipment in continuous manufacturing may be disturbed and PLC devices may lock up.
DATA NETWORK CONGESTION
Earth leakage currents cause small voltage drops along the earthing conductor. In a TN-C
network, the combined earth-neutral conductor will constantly carry significant current,
dominated by triple-n harmonics. Due to the increasing use of low-voltage signals in IT
equipment, bit error rate increases, up to the point that the entire network locks up. How
many large and small, privately owned networks enjoy this phenomenon almost on a weekly
basis- For an unexplained reason, the network locks up, e-mail services fail; it is no longer
possible to print.
PROBLEMS WITH POWER FACTOR CORRECTION EQUIPMENT
Harmonic frequencies may coincide with resonant frequencies of the combined stray
inductance and power factor correction (PFC) equipment, creating excessive voltage or
current and leading to premature failure. Moreover, as a general problem, measurement
devices may not correctly measure the loading of the PFC, as they incorrectly measure the
harmonic content in the current.



23

PROBLEMS WITH SPECIFIC (LONG) LINES OR WHEN SWITCHING
HEAVY LOADS
Long lines mean higher impedance, resulting in higher voltage disturbances from inrush
currents, for example when a heavy motor starts up, or when switching on computers.
Harmonic currents generated by variable speed drives, or switch-mode power supplies,
located at the end of long lines, result in higher harmonic voltage distortion. Therefore,
upsize long power lines for low voltage drop is the solution. As a side benefit, upsized power
lines will have lower losses. When loaded more than 3,000 hours, the economic payback will
be very short.
OVERLOADED NEUTRALS
In a 3-phase circuit, there are 3 active conductors, and a return conductor, which carries the
unbalance between the 3 phases. However, with the triple-n harmonics adding up, significant
currents flow in the neutral conductor. As many neutral conductors have been, in the past,
half-sized, this situation can become critical, even when the phase conductors are operating
well below full load.
NUISANCE TRIPPING OF PROTECTIVE DEVICES
Inrush currents may trip circuit breakers. Circuit breakers may not correctly sum the current
contained in the fundamental and various harmonics and so trip erroneously, or not at all,
when they should. Leakage currents may reach thresholds that cause residual current devices
to trip.
The remedies against erroneous tripping must not compromise safety of personnel on-site.
The general solution is to reduce the inrush currents and earth leakage currents by splitting


24

equipment over more circuits, each feeding fewer loads. Specially designed circuit breakers;
that can cope with harmonics, should be used. Over sizing is never the correct solution.
UTILITY CLAIMS RESULTING FROM HARMONICS AFFECTING
SUPPLY
Not many utilities charge (yet) for harmonic pollution, as they currently do for reactive
power. However, they may start to do so in the future, as harmonics also lead to sub-optimal
exploitation of the electricity distribution system.
2.5 The Cost of Poor Power Quality
Electrical power is an important raw material for all commercial operations and, like any
other raw material, the quality of supply is very important. The nature and causes of power
quality defects have been discussed in earlier sections; this section concentrates on the effects
of defects on production and the costs that can be expected to occur. The five basic defects,
each with different causes and effects have different cost implications.
It is estimated that power quality problems cost industry and commerce in the EU about 10
billion per annum while expenditure on preventative measures is less than 5 % of this [CDA].
The question is the obvious one: How much money should be invested in prevention to
balance the risk of failure? and the answer depends on the nature of the business. The first
step is to understand the nature of the problems and assess how each of them relates to the
business and what losses might result. The following sections discuss power quality problems
from the point of view of their business interruption potential.



25

2.5.1 Harmonic distortion
Harmonic distortion caused by non-linear loads on the electricity supply system, result in
currents in the system that are of higher magnitude than expected and contain harmonic
frequency components. These currents cannot be adequately measured by some of the lower
cost portable test meters commonly used by installation and maintenance technicians, leading
to current levels being seriously under-estimated sometimes by as much as 40 %. This error
in magnitude alone can result in circuits being installed with conductors that are too small.
Even if the current is within the capacity of the overcurrent protection device, conductors run
at higher temperatures and waste energy typically 2-3 % of the load. Frequently the
overcurrent protection device rating is too close to the real load current (because it was
under-estimated) and the circuit is prone to so-called nuisance tripping.
The harmonic frequency components cause greatly increased eddy current losses in
transformers because such losses are proportional to the square of the frequency. Because the
losses are higher, the operating temperature of the transformer is higher and the lifetime is
considerably shortened. Even moderately loaded transformers supplying IT loads will have
much lower lifetimes than expected unless proper precautions are taken.
The economic effects of harmonics are shorter equipment lifetime, reduced energy efficiency
and a susceptibility to nuisance tripping. The cost of nuisance tripping, like any other
unplanned shortage, can be very significant and is discussed further in the section on voltage
dips. Shorter equipment lifetime can be very expensive. Equipment such as transformers is
usually expected to last for 30 or 40 years and having to replace it in 7 to 10 years can have
serious financial consequences. The cost of avoidance is relatively small, requiring only good
installation practice and proper equipment selection. Installing cables that are one to two sizes


26

greater than the calculated minimum reduces losses and operating costs at very little increase
in initial cost.
2.5.2 Blackouts
Blackouts are the most basic of power quality problems, lasting from several seconds up to,
in one famous extreme case, months. In the UK, the average blackout lasts for about 100
minutes and occurs every 15 months, but individual events may be very short and much more
frequent. Of course, the public supply is not the only source of failures. Within the
installation in a building or plant there will be many areas where the failure of a single
component, cable or connection will cause complete shutdown.
Protection against complete power failure requires two types of action. The installation must
be designed to eliminate single points of failure, or at least those identified in a risk
assessment as posing the greatest risk, and steps must be taken to identify the need for a
backup power supply. The techniques required are neither difficult nor particularly expensive
but can, alone, provide considerable benefit. As always, these techniques are far cheaper to
apply at the initial design stage than during commissioning. Alternative sources of power can
be very expensive both to purchase and to maintain there is little point in having a standby
generator, for example, if it is not ready for an instant start-up and the need for and type of
supply required must be carefully considered. In judging the economic viability of investing
in on-site generation plant it should be remembered that, once installed, it will protect against
failures over many years.
Large, power-hungry industries such as steel or paper-making will require a second supply
taken from a different section of the grid so that a single failure is very unlikely to affect both
supplies. Alternatively, on-site full demand generation may be viable, if suitable fuel supplies


27

are available. In either case the initial cost of this is likely to be very high, but so is the
potential cost of power failure. Paper, for example, is manufactured in a continuous process
requiring precisely controlled speeds of hundreds of rollers in a machine that may be over
500 meters in length. Any failure of the power supply, even a dip, will cause loss of
synchronization, and halt the process. All the part processed paper and pulp must be cleared
from the machine and the surrounding area before restarting; this can take many hours. Apart
from the loss of output, the waste of raw materials and manpower, the inability to supply the
customer is very important. Newsprint, for example, is used in such vast quantities that it is
impossible for either supplier or customer to keep a buffer stock. It is required just in time,
and it is manufactured, used and discarded in just a few days. Failure of the papermaker to
deliver means that the publisher cannot print and, since yesterdays news has no value (but
considerable costs), there are severe financial consequences. This may lead to a change of
supplier or a change in the terms of the supply contract, such as penalty clauses.
For smaller industries with lower power requirements it may be viable to have on-site
generation to handle essential equipment during blackouts and to reduce demand peaks. This
is much cheaper, but the cost must still be judged against the risks of failure an assessment
that can only be made by the organization itself. It must be remembered that a standby
generator takes time to start, so some other backup supply, such as a UPS, must be provided
for sensitive loads. A UPS has limited capacity and it is important that it is used to support
only the essential loads, such as the computer network servers and critical workstations, and
nothing else. As always, good maintenance procedures are essential.
2.5.3 Dips or sags
Dips are short-term reductions in the RMS magnitude of the supply voltage lasting from a
fraction of a second up to several seconds. Dips are described in terms of duration and


28

retained voltage, i.e. the percentage of the nominal supply voltage (RMS) remaining during
the event. Note that; a very short but complete loss of supply is called an interruption but is
frequently referred to as a dip. The Information Technology Industry Council (ITIC) curve,
formerly known as the Computer and Business Equipment Manufacturers Association
(CBEMA) curve, shown in Figure 2-2, describes the tolerance of equipment to voltage
disturbances of all types.
The solid lines represent the maximum and minimum voltages that can be tolerated without
malfunction plotted against time. For example, data processing equipment should tolerate an
over-voltage of five times nominal supply for a duration of 100 s but only 20 % over-
voltage for 10 ms. On the under voltage side, a complete loss of supply should be tolerated
for up to 20 ms (one mains cycle) but for 100 ms the minimum retained voltage must be 70 %
of nominal. The curve was originally produced to help users of IT equipment in resolving
power quality problems with electricity suppliers. By standardizing the requirements of
equipment it became much easier to determine by site measurement whether the supply was
adequate or not. The ITIC curve presents a rather optimistic view of the performance of
supply networks! Many dips are caused by faults on the supply network with the severity of
the dip depending on the relative positions of the generator, fault and measurement point.
There are no official statistics on the severity and distribution of voltage dips but some
medium scale measurements are now in progress and can be expected to yield valuable
information in due course. One study, carried out by a major generator, measured voltage
disturbances at 12 sites with demand between 5 and 30 MVA. In a ten-month period 858
disturbances were logged, 42 of which resulted in disruption and financial loss. Although all
12 sites were low technology manufacturing operations making low value added products the
financial loss totaled 600 000 (average 14 300 per event or 50 000 per site), with the
highest individual loss of 165 000. Clearly plants making high value added products and

those req
much hig
Figure 2-2
Figure 2-3
quiring mul
gher losses. T
2: ITIC curve
3: Losses due t
tistage man
The table in
to voltage dips
nufacturing p
figure 2-3 g
s
29
processes, s
gives some ty
such as sem
ypical values
miconductors
s.

s, would fac

ce



30

These are huge costs for what might seem to be trivial events lasting less than a second. The
problem is that, since the response of individual items such as data processing equipment, or
variable speed motor drives to dips is undefined, the behavior of a system is impossible to
predict or control. For continuous processes, such as papermaking, the effect of a dip is just
as serious as a complete blackout, with the same cleanup costs, raw material losses and lost
production. For computer based operations the time taken to re-boot a large number of
workstations and recover pending transactions and unsaved documents can take several
hours. The semiconductor industry is particularly vulnerable because wafers require two
dozen or so manufacturing stages to be completed over several days. If a wafer is spoiled
towards the end of the process all the value of the work done is lost. The rate of development
in semiconductors is now so fast, competition so intense and product life cycles so short that
the loss of product is a major concern not only to the suppliers but also to their customers
who cannot build and ship their own products.
On-line uninterruptible power supplies, in which the load power is continuously generated
from a storage battery constantly charged from the supply, inherently provide immunity to
dips. Off-line units are less secure because the absence of supply has to be detected before the
load is switched to the internal generator. If the detection threshold is too high, the UPS is
switched in and out frequently and unnecessarily while if the limit is too low, damaging dips
will be fed to the load. Detailed specifications should be consulted before selecting a
particular model.
2.5.4 Transients
Transients are voltage disturbances of very short duration (up to a few milliseconds) but high
magnitude (up to several thousand volts) with a very fast rise time. Most transients arise from
the effects of lightning strikes or switching of heavy or reactive loads. Because of the high


31

frequencies involved they are considerably attenuated as they propagate through the network
so that those occurring close to the point of interest will be much larger than those originating
further away. Protective devices in the network ensure that transients are generally kept to a
safe level and most problems arise because the source of the transient is close to or within the
installation.
The damage that results may be instantaneous, such as the catastrophic failure of electrical
plant or appliances, or the corruption of data within computers or on network cabling or it
may be progressive with each event doing a little more damage to insulation materials until
catastrophic failure occurs. The cost of replacing the failed equipment and the cost of the
downtime involved must be considered. Protection is relatively cheap. The basic requirement
is that the installation earthing system should be designed to have low impedance over a wide
frequency band, with a good low impedance connection to the earth electrode system. The
lightning protection system should be designed appropriately, taking into account local
factors, such as the number of lightning days per year. Transient protection should be fitted at
the entry of all incoming conductors, including telephone and other communications lines.
The manufacturer should have provided suppression of transients from switching equipment
and good maintenance procedures should be introduced to ensure that it continues to be
effective.
2.6 Solutions Power Quality Issues
Electricity suppliers argue that critical users must bear the costs of ensuring supply quality
themselves rather than expect the supply industry to provide a very high reliability supply to
every customer everywhere on the network. Such a guaranteed quality supply would require a
very substantial investment in additional network assets for the benefit of relatively few
customers (in numerical, not consumption, terms) and would be uneconomic. It is also


32

doubtful whether it would be technically feasible within the current social and legal
framework in which any customer is normally entitled to be connected to the supply and
utility providers have the right to excavate roadways with the risk of cable damage. Weather
conditions, such as high winds and freezing rain, frequently cause damage to overhead lines,
which, under the same conditions, are difficult and time consuming to repair. It is therefore
the consumers responsibility to take steps to ensure that the quality of power delivered to his
process is good enough, with the clear implication that this quality level may well be higher
than that delivered to the plant by the supplier.
There are a variety of engineering solutions available to eliminate or reduce the effects of
supply quality problems and it is a very active area of innovation and development. As such,
customers need to be aware of the range of solutions available and the relative merits and
costs. Users are faced with the need to make design investment decisions about the type and
quantity of additional plant required to achieve the quality of supply required. Unfortunately,
some vital information is missing the extent and severity of power quality problems likely
to be experienced in any particular location is largely unknown. Because there are so few
published statistics it is very difficult for consumers to quantify the cost of failure and justify
the cost of preventative measures.
The list of possible solutions for power quality problems is again long and non-complete.
Figure 2-4 lists the power quality solutions adopted according to the survey conducted by
Copper Institute at 1,400 sites in 8 European countries.

Figure 2-4
countries

2.7 Stan
Followin
some of t
Institute
IEEE 644
IEEE C6
IEEE 518
4: Most prev
ndards
ng is a list of
the standard
of Electrical
4 S
M
63.12 R
8 G
N
valent PQ solu
f power qua
ds listed are i
l and Electro
Standard Pro
Magnetic Fie
Recommend
Guide for th
Noise Inputs
utions, in ter
ality and rela
in existence
onic Enginee
ocedure for
elds from AC
ded Practice
he Installatio
s to Controll
33
rms of % ado
ated standard
at this time,
ers (IEEE);
Measureme
C Power Lin
for Electrom
n of Electric
ers from Ext
option rate a
ds from two
while others
Piscataway,
ent of Power
nes
magnetic Com
cal Equipme
ternal Sourc
t 1,400 sites
renowned o
s are still in
NJ; http://w
r Frequency
mpatibility L
ent to Minim
ces
in 8 Europea
organization
process:
www.ieee.org
y Electric an
Limits
mize Electric


an
ns;
g
nd
al


34

IEEE 519 Recommended Practices and Requirements for Harmonic Control in
Electrical Power Systems
IEEE 1100 Recommended Practice for Powering and Grounding Sensitive
Electronic Equipment
IEEE 1159 Recommended Practice for Monitoring Electric Power Quality
IEEE 141 Recommended Practice for Electric Power Distribution for Industrial
Plants
IEEE 142 Recommended Practice for Grounding of Industrial and Commercial
Power Systems
IEEE 241 Recommended Practice for Electric Power Systems in Commercial
Buildings
IEEE 602 Recommended Practice for Electric Systems in Health Care Facilities
IEEE 902 Guide for Maintenance, Operation and Safety of Industrial and
Commercial Power Systems
IEEE C57.110 Recommended Practice for Establishing Transformer Capability when
Supplying Non-sinusoidal Load
IEEE P1433 Power Quality Definitions
IEEE P1453 Voltage Flicker
IEEE P1564 Voltage Sag Indices



35

International Electrotechnical Commission (IEC); Geneva, Switzerland; http://www.iec.ch
IEC/TR3 61000-2-1 Electromagnetic Compatibility Environment
IEC/TR3 61000-3-6 Electromagnetic Compatibility Limits
IEC 61000-4-7 Electromagnetic Compatibility Testing and Measurement
Techniques General Guides on Harmonics and Interharmonics
Measurements and Instrumentation
IEC 61642 Industrial a.c. Networks Affected by Harmonics Application of
Filters and Shunt Capacitors
IEC SC77A Low Frequency EMC Phenomena
IEC TC77/WG1 Terminology
IEC SC77A/WG1 Harmonics and Other Low Frequency Disturbances
IEC SC77A/WG6 Low Frequency Immunity Tests
IEC SC77A/WG2 Voltage Fluctuations and Other Low Frequency Disturbances
IEC SC77A/WG8 Electromagnetic Interference Related to the Network Frequency
IEC SC77A/WG9 Power Quality Measurement Methods
2.8 Conclusion
The literature survey in the field of power quality reveals that PQ has become one of the most
talked about concern of the utility company as well as consumer. We have carried out an
introductory level study of power quality events. We have also come to know about the PQ
issues and how they affect industry and our normal day to day life. We have also come across


36

some of the chosen ones PQ mitigation techniques, and come to know how these can be
applied to solve the problem of poor power quality.

2.9 References
[2.1] Sankaran C., Power Quality, CRC Press, New York, 2001
[2.2] Dugan et al., Electrical power Systems Quality, Second Edition, McGraw Hill
Publishers. New York
[2.3] Baggini Angelo, Handbook of Power Quality, John Wiley and Sons, London, 2008




37


Chapter 3 Harmonics

As discussed earlier, harmonic distortion has been present within the power systems since the
advent of alternating current electricity distribution systems [3.1]. Over the past two decades
a lot of research has been done in the field of harmonic disturbances. The interest in
harmonics can largely be attributed to the advancing power electronics usage and growth in
the number of new load technologies connected to distribution systems that are sensitive to
excessive harmonic voltage levels. The increased penetration of these new technologies is for
the most part driven by the desire to have more energy efficient equipment and greater
control of equipment operation. Increased efficiency and control typically means the
equipment includes some type of semiconductor based power electronic front end. The same
new technologies, in general produce increased levels of harmonic current emissions that in
turn raise the levels of harmonic voltages existing on the system. In conjunction with this the
growing trend of electricity being sold as a product under contractual agreements in some
parts of the world means that guaranteeing an acceptable level of power quality disturbances
is becoming of greater importance to both electricity utility companies and customers alike.
Indications in the literature suggest that harmonic voltage levels in transmission and
distribution systems are increasing [3.2, 3.3, and 3.4]. Increasing levels of voltage harmonics
creates a concern for customers with sensitive equipment and also for electricity utility
companies that face increased losses throughout their power systems. The increased losses
burden utility companies with additional costs that have to be recovered to ensure profit


38

margins are not decreased. There are suggestions by [3.5] that in Canada and some major
European countries the cost of network modifications and losses due to harmonic distortion
effects may represent as much as 1% of a countrys gross domestic product (GDP).
3.1 Causes of Harmonic Distortion
As stated earlier it is well known that harmonic voltage distortion within power systems is
due to the interaction of harmonic current emissions from a variety of sources and the
impedance of the power system. The sources of harmonic emissions can include large
industrial loads connected at transmission or distribution levels, smaller commercial and
residential distribution system customer loads, distributed generation installations, the power
system itself, and to a lesser extent traditional generation equipment. Harmonic current
emissions originate from all types of non-linear loads. Non-linear loads are broadly classified
as loads which draw non-sinusoidal current even when the supply voltage is perfectly
sinusoidal [3.6]. Non-linear loads include saturated magnetic circuits, such as those in power
system transformers and rotating machines, arc furnaces, fluorescent lighting and of course
power electronic loads. Power electronic loads by far are the most significant harmonic
contributors relative to the amount of energy they draw. Some of the more common power
electronic loads include
i. Switch mode power supplies(SMPS) - present in computers, televisions,
microprocessors
ii. Rectifiers present in dc motor drives, regulated power supplies, battery chargers,
iii. Inverters present in variable speed ac drives and now a days in Indian homes also,
iv. Static VAr compensators,
v. Cyclo-converters, and
vi. High voltage DC transmission converters.


39

Mitigation techniques are often required to suppress the levels of harmonic emissions
resulting from the above mentioned power electronic equipments.
3.2 Problems Caused by Harmonics
How is an equipment affected by harmonics?-was discussed in last chapter, now well see the
general problems caused by them. Harmonic distortion can have both short-term and long-
term effects on distribution system equipment and connected customer loads. Short-term
effects are mainly concerned with immediate damage, equipment malfunction, and the
associated power losses due to harmonic currents and voltages. Long-term effects include
thermal losses and reduced life span of equipment.
HEATING: It is common to refer to heating as I
2
R losses. By using superposition,
the total losses can be expressed as a sum of the individual harmonic losses.
I
2
R = I
50 Hz
2
R
50 Hz
+I
150 Hz
2
R
150 Hz
+I
250 Hz
2
R
250 Hz
+
Although harmonic voltage and currents within most distribution systems are quite
small relative to the fundamental component, increased impedance of the network at
harmonic frequencies often gives rise to significant power losses. These power losses
include direct copper losses due to an effective increase in the level of rms current in
the presence of harmonics in combination with increases in conductor resistance due
to skin effect. Since most ac equipment ratings are based on 50-Hz losses, the addition
of the harmonic loss components requires derating the equipment [3.10]. The total
losses (50 Hz losses plus harmonic losses) are to be within the specified equipment
ratings. Regardless of actual quantities, it can be seen that they add to the total amount
of heating. The amount of derating will depend upon which harmonics are present and
the magnitudes of the individual harmonic current and resistance. Other results of
heating are


40

o excessive losses and heating in motors, capacitors, and transformers, and
o blown capacitor fuses.
INCREASED IRON LOSSES: Increased iron losses due to harmonics are also
significant within the iron cores of most magnetic materials. Iron losses include
hysteresis losses, where losses increase approximately in proportion to the harmonic
frequency, and eddy current losses, where losses increase in proportion to the square
of the harmonic frequency. Iron losses result in reduced efficiency of equipment and
also raise core temperatures creating thermal stresses and degrading insulation levels.
PROBLEMS IN CAPACITORS: Harmonics are also a major cause of metal losses
and dielectric stress in capacitor banks causing additional heating and loss of life [3.7,
3.8]. Insulation stress due to higher peak voltages from harmonics can also result in
cable insulation breakdown and disruption to supply.
HARMONIC RESONANCE: In some circumstances capacitors can combine with
source and load inductance to form a parallel resonant circuit. In the presence of a
harmonic resonance, harmonic voltages and currents may be amplified. The resulting
voltages may highly exceed the voltage rating of capacitors or other connected
equipment and the consequence is capacitor damage or blown fuses.
RELAY/CIRCUIT BREAKER MALFUNCTION: Complete or partial load
disruption can also occur due to the presence of harmonics. Harmonics interfere with
protective relays, metering devices, control circuits, communication circuits, and
customer electronic equipment. Sensitive equipment can experience malfunction or
complete component failure.
SHORTENED LAMP LIFE: Shortened incandescent lamp lifetime and failure of
some types of fluorescent lights have also been reported in [3.7, 3.9].


41

COMMUNICATION INTERFERENCE: Magnetic (or electrostatic) coupling
[3.10, 3.11] between electrical power circuits and communication circuits can cause
what is known as communication interference. Current flowing in the power circuit
produces a magnetic (or electrostatic) field that will induce a current (or voltage) in
the nearby conductors of the communication circuit. The amount of interference will
depend upon the magnitude of the induced current (or voltage), frequency [3.11], and
the efficiency of the magnetic (electrostatic) coupling. Other types of communication
interference are:
o induced line noise,
o interference with power line carrier systems, and
o relay malfunctions can be there due to communication failure.
SOLID-STATE DEVICE MALFUNCTIONS: Harmonics can cause solid-state
devices to malfunction if the equipment is sensitive to zero crossings [3-13]. A
resonance condition can cause a current waveform to have zero crossings occur more
than once every half-cycle. Problems with solid-state devices also develop with
equipment that operates in response to peak values of voltage. The typical example of
this mode of operation is the diode. Because the operational settings for such a device
are often given in terms of rms values for the user's convenience, the device may not
operate properly in the presence of harmonics because it is sensing a peak value that
does not directly correspond to the rms value of the waveshape. Other solid-state
device malfunctions are:
o errors in measurement equipment,
o nuisance tripping of relays and breakers,
o unstable operation of zero voltage crossing firing circuits, and
o interference with motor controllers.


42

While some of the above mentioned problems and effects associated with harmonics occur
over short-term durations, more significant long-term problems such as degrading insulation
due to thermal stresses are usually of most concern. These problems are often not visible to
the utility company or customer until final failure of equipment occurs. Reduced lifespan of
equipment necessitates costly repairs or replacements. Thus it is important for both utility
company and customer to be proactive with regards to measurement and mitigation of
harmonic distortion problems.
3.3 Conclusion
Now a day, it is very important to know which harmonics are present in power system. We
also need to estimate their amplitude and possible phase also, for designing the solution to the
harmonic problem. Well discuss about the methods of harmonic estimation in chapter 5.

3.4 References
[3-1] Owen, E.L., A history of harmonics in power systems. IEEE Industry Applications
Magazine, 1998. Vol.4(Issue 1): p. 6-12.
[3-2] Philipps, J .K., Nelson, J .P., and Sen, P.K., Power quality and harmonic distortion on
distribution systems. IEEE Trans. on Industry Applications, 1994. Vol. 30(No. 2): p.
476-484.
[3-3] Martinon, J ., et al., A new statistical approach of harmonic propagation in
transmission systems. IEEE Trans. on Power Delivery, 1996. Vol.11(No.2): p. 1032-
1040
[3-4] Mahmoud, A., Owen, R., and Ortmeyer, T., Power system harmonics: an overview.
IEEE Trans. on Power Apparatus and Systems, 1983. Vol. 102(No. 8): p. 2455-2460.


43

[3-5] Eurelectric, Power Quality in European electricity supply networks. 2002, NE
Standardisation Specialist Group. p. 11.
[3-6] Lundquist, J ., On harmonic distortion in power systems. 2001, Chalmers University
of Technology: Gteborg.
[3-7] Wagner, V.E., Belda, J .C., and Barnes, T.M., Effects of harmonics on equipment.
IEEE Trans. on Power Delivery, 1993. Vol.8(No.2): p. 672-680.
[3-8] Chen, W. and Cheng, Z., An experimental study of the damaging effects of harmonics
in power networks on the capacitor dielectrics. Proc. of Second Inter. Conf. on
Properties and Applications of Dielectric Materials, 1988. Vol.2: p. 645- 648.
[3-9] Arseneau, R. and Ouelette, M., The effects of supply harmonics on the performance
of compact fluorescent lamps. IEEE Trans. on Power Delivery, 1993. Vol. 8(No. 2):
p. 473-479.
[3.10] Miller, T. J . E., Reactive Power Control in Electric Systems. New York: Wiley,
1982, ch. 10.
[3.11] IEEE Guide for Harmonic Control and Reactive Compensation of Static Power
Converters, IEEE Standard 519, 1981.
[3.12] Shipp, D. D., Harmonic analysis and suppression for electrical systems supplying
static power converters and other nonlinear loads, IEEE Trans. Ind. Appl., vol. IA-
15. no. 5 , pp. 453-458, Sept./Oct. 1979.
[3.13] Subjak J r, J oseph S., and Mcquilkin, J ohn S., Harmonics - Causes, Effects,
Measurements, and Analysis: An Update, IEEE Transactions on Industry
Applications. Vol. 26. No. 6. November-December 1990
[3.12] Tagare, D. M., Reactive Power Management, New Delhi, Tata McGraw Hill, 2004



Chap

Harmoni
fundamen
250 Hz.
amplitud
4.1 The
Any peri
sum of s
wave wit
Figure 4-1
pter 4
c frequencie
ntal of 50 H
Plot that sh
de spectrum a
eoretical B
iodic signal
sinusoids wh
th third and f
1: Fundament
Fouri
es are integr
Hz, the third
hows amplit
and that show
Backgroun
which is not
hich have ha
fifth harmon
al with 3rd &
ier an
ral multiples
harmonic w
tude versus
ws phase ve
nd
t a perfect s
armonic freq
nics.
5th Harmoni
44
nalysi
s of the fund
would be 150
frequency c
rsus frequen
sinusoid can
quencies. F

ics
s for H
damental sup
0 Hz and the
characteristic
ncy is called
be expresse
igure 4-1 sh
Harm
pply frequen
e fifth harmo
cs of the sig
phase spectr
ed in a Fouri
hows a fund
monics
ncy, i.e. for
onic would b
gnal is calle
rum.
ier series as
damental sine

s
a
be
ed
a
e-

Figure 4-
Note tha
example,
Figure 4-2
This wa
equipmen
readings.
equipmen
non-fund
harmonic
the harm
useful co
but it is
detail lat
to branch
is very im
are explic
-2 shows a f
at in practice
, containing
2: Distorted cu
aveform is
nt, such as
. Note also
nt that uses
damental fre
cs in power
monics origin
onclusions ca
still commo
ter) figures q
h circuits not
mportant tha
citly specifie
fundamental
e most disto
many more
urrent wavefo
clearly not
averaging
that there ar
zero crossin
equencies a
installations
nate as curre
an be drawn
on to find o
quoted. Whe
t concerned
at both volta
ed as voltage
l with 70 %
orted curren
harmonics w
orm
a sine-wav
reading rm
re six zero c
ng as a refe
and has to
s it is the cur
ents and mos
without kno
only the tota
en harmonic
with carryin
age and curr
e and curren
45
third harmo
nt waveform
with a more
ve and tha
ms-calibrated
crossing poi
erence will m
be treated
rrent harmon
st of the ill
owledge of t
al harmonic
s propagate
ng the harmo
rent values
nt values.
onic and 50
ms are much
complex pha
at means th
d multimete
ints per cycl
malfunction.
d accordingl
nics that are
effects are d
the spectrum
distortion (W
around a di
onic current,
are measure
% fifth har
more comp
ase relations
hat normal
ers, will giv
le instead o
. The wavef
ly. When t
of most con
due to these
m of the curre
Well discu
istribution sy
they do so a
ed and that q
rmonic adde
plex than th
ship.
measuremen
ve inaccura
f two, so an
form contain
talking abou
ncern becaus
e currents. N
ent harmonic
ss about it i
ystem, that i
as voltages.
quoted value

d.
his
nt
te
ny
ns
ut
se
No
cs
in
is,
It
es


46

Conventionally, current distortion measurements are suffixed with I, e.g. 35 % THDI, and
voltage distortion figures with V, e.g. 4 % THDV. Harmonic currents have been present in
the electricity supply system for many years. Initially they were produced by the mercury arc
rectifiers used to convert AC to DC current for railway electrification and for DC variable
speed drives in industry. More recently the range of types and the number of units of
equipment causing harmonics have risen sharply, and will continue to rise, so designers and
specifiers must now consider harmonics and their side effects very carefully.
4.2 Fourier series
If a function g(t) is periodic with period I
p
,[4.1] i.e.,
g(t _ I
p
) = g(t) (4-1)
and in any finite interval g(t) has at most a finite number of discontinuities and a finite
number of maxima and minima (Dirichlets conditions), and in addition,
_ g(t)Jt <
1
p
0

(4-2)
i.e. g(t) can be absolutely integrated over a period, then g(t) can be expressed as a series of
sinusoids. That is,
g(t) =
o
0
2
+ o
n
cos n
0
t +b
n
sin n
0
t

n=1
(4-3)





47

where

0
=
2n
I
p

(4-4)
and the Fourier coefficients o
n
and b
n
are determined by the following equations.
o
n
=
2
I
p
_ g(t) cos n
0
t Jt n = u, 1, 2, S,
t
0
+1
p
t
0
(4-4)
b
n
=
2
I
p
_ g(t) sinn
0
t Jt n = u, 1, 2, S,
t
0
+1
p
t
0

(4-5)
Equation (4-3) is called the trigonometric Fourier series. The term
u
0
2
in Equation (4-3) is the
dc component of the series and is the average value of g(t) over a period. The term
o
n
cos n
0
t + b
n
sin n
0
t is called the n
th
harmonic. The first harmonic is obtained when
= 1 . The latter is also called the fundamental with the fundamental frequency of
0
. When
n = 2 , we have the second harmonic and so on.
Equation (4-3) can be rewritten as,
g(t) =
o
0
2
+ A
n
cos(n
0
t +
n
)

n=1
(4-5)
where
A
n
= o
n
2
+ b
n
2

(4-6)
and

n
= -tan
-1
_
b
n
o
n
] (4-7)


48

The total power in g(t) is given by the Parsevals equation:
P =
1
I
p
_ g
2
(t)Jt = [
o
0
2

2
+
A
n
2
2

n=1
t
0
+1
p
t
0
(4-8)
By using the Eulers identity, the cosine and sine functions of Equation (4-3) can be replaced
by exponential equivalents, yielding the expression
g(t) = c
n
c
]no
0
t

n=-
(4-9)
where
c
n
=
1
I
p
_ g(t)c
-]nw
0
t
Jt
1
p
2
-1
p
2
(4-10)
and

0
=
2n
I
p
(4-11)
Equation (4-9) is termed the exponential Fourier series expansion. The coefficient c
n
is
related to the coefficients o
n
and b
n
of Equations (4-4) and (4-5) by the expression
c
n
=
1
2
o
n
2
+b
n
2
z-tan
-1
_
b
n
o
n
] (4-12)
In addition, c
n
relates to A
n
and
n
of Equations (4-6) and (4-7) by the relation
c
n
=
A
n
2
z
n

(4-13)
The plot of |c
n
| versus frequency is termed the discrete amplitude spectrum or the line
spectrum. It provides information on the amplitude spectral components of (t) . A similar


49

plot of zc
n
versus frequency is called the discrete phase spectrum and the latter gives
information on the phase components with respect to the frequency of g(t).
4.3 Fourier Transform
If g(t) is a non-periodic deterministic signal expressed as a function of time t , then the
Fourier transform of g(t) is given by the integral expression [4.2]:
0() = _ g(t)c
-]2n]t
Jt

-
(4-14)
where denotes frequency g(t) can be obtained from the Fourier transform 0() by the
Inverse Fourier Transform formula:
g(t) = _ 0()c
]2n]t
J

-
(4-15)
For a signal g(t) to be Fourier transformable, it should satisfy the Dirichlets conditions that
were discussed in earlier section. If g(t) is continuous and non-periodic, then 0() will be
continuous and periodic. However, if g(t) is continuous and periodic, then G(f) will discrete
and nonperiodic; that is
g(t) = g(t _I
p
) (4-1)
where
I
p
= pcrioJ
then the Fourier transform of g(t) is


50

0() =
1
I
p
c
n
o _ -
1
I
p
_

n=-
(4-16)
where
c
n
=
1
I
p
_ g(t)exp (-]2n
0
t)Jt
1
p
2
-1
p
2
(4-17)
4.4 Discrete & Fast Fourier Transform
Fourier series links a continuous time signal into the discrete-frequency domain. The
periodicity of the time-domain signal forces the spectrum to be discrete. Aperiodic finite
energy signals have continuous spectra. The discrete time Fourier transform of a discrete-
time signal g|n] is given as [4.3]
0() = g|n]exp (-]n)

n=-
(4-18)
Since the 0() is periodic with period equal to 2n, so we can take N sample of this spectra
in one frequency period. Thus we get the N point Discrete Fourier Transform (DFT) as given
by following equation:
0|k] = g|n]exp (-]2nkn N )
N-1
n=0
; k = u, 1, 2, , (N - 1) (4-19)
The inverse discrete Fourier transform, g|n] is
g|n] =
1
N
0|k]exp (]2nkn N )
N-1
k=0
; n = u, 1, 2, , (N - 1) (4-20)
where


51

N is the number of time sequence values of g[n]. It is also the total number frequency
sequence values in G[k].
I is the time interval between two consecutive samples of the input sequence g|n].
F is the frequency interval between two consecutive samples of the output sequence
0|k].
N , and F are related by the expression
NI =
1
F

(4-21)
NI is also equal to the record length. The time interval, I, between samples should be chosen
such that the Shannons Sampling theorem is satisfied. This means that I should be less than
the reciprocal of 2
H
, where
H
is the highest significant frequency component in the
continuous time signal g(t) from which the sequence g|n] was obtained. Several fast DFT
algorithms require N to be an integer power of 2.
A discrete-time function will have a periodic spectrum. In DFT, both the time function and
frequency functions are periodic. Because of the periodicity of DFT, it is common to regard
points from n = 1 through n = N2 as positive, and points from n = N2 through
n = (N 1) as negative frequencies. In addition, since both the time and frequency
sequences are periodic, DFT values at points n = N2 through n = (N - 1) are equal to
the DFT values at points n = N2 through n = 1.
In general, if the time-sequence is real-valued, then the DFT will have real components which
are even and imaginary components that are odd. Similarly, for an imaginary valued time


52

sequence, the DFT values will have an odd real component and an even imaginary
component.
If we define the weighting function w
N
as
w
N
= c
-
]2n
N
= c
-]2nP1

(4-22)
Equations (4-19) and (4-20) can be re-expressed as
0|k] =
1
N
g|n]w
N
kn
N-1
n=0
(4-23)
and
g|n] = 0|k]w
N
-kn
N-1
k=0
(4-24)
The Fast Fourier Transform, FFT, is an efficient method for computing the discrete Fourier
transform. FFT reduces the number of computations needed for computing DFT. For
example, if a sequence has N points, and N is an integral power of 2, then DFT requires N
2

operations, whereas FFT requires
N
2
log
2
N complex multiplication,
N
2
log
2
N complex
additions and
N
2
log
2
N subtractions. For N = 1u24, the computational reduction from DFT
to FFT is more than 2uu to 1.
The FFT can be used to (a) obtain the power spectrum of a signal, (b) do digital filtering, and
(c) obtain the correlation between two signals.




53

4.5 Power Quality Terms Related to Harmonics
We have seen the methods to get the spectrum of a signal in the above sections. Now if we
have the harmonic spectrum of a current or voltage waveform u(t), the original waveform
can be constructed using the Fourier series summation:
u(t) = u
h
cos(b
0
t +
h
)

h=1
(4-25)
where u
h
is the h
th
harmonic peak current or voltage,
h
is the h
th
harmonic phase,
0
is the
fundamental angular frequency,
0
= 2n
0
, and
0
is the fundamental frequency, typically
50Hz.
4.5.1 Total harmonic distortion (THD)
Although the harmonic content of a power system may be quite small relative to the
fundamental in most circumstances, for exactness the rms value of a current or voltage
waveform requires the harmonic content to be considered such that
u
ms
= __
1
2
u
h
]
2

h=1

(4-26)
where u
ms
is the rms value of voltage or current. The rms voltage or current can also be used
to quantify the level of distortion of the waveform. The total harmonic distortion of voltage or
current waveform (THDU) is calculated using equation (4-27)
IE
0
=
1
u
1
_u
h
2

h=2
=
_
_
u
ms
u
1ms
]
2
-1
(4-27)


54

where IE
0
represents voltage or current total harmonic distortion (alternatively represented
as IE
v
and IE
I
respectively) and u
1ms
is the rms fundamental voltage or current.
Alternatively rms voltage or current can be represented in terms of total harmonic distortion
u
ms
= _u
hms
2

h=1
= u
1ms
_1 +IE
0
2

(4-28)
As distribution system fundamental voltage and current rarely remain static in magnitude at
different times throughout the day, the definition for total harmonic distortion may at times
provide a misleading value for the harmonic distortion level [4.4]. This is especially true for
distribution system fundamental currents that fall close to zero at certain periods of the day,
resulting in large values of IE
I
For this reason a modified index for harmonic distortion
may be used with the harmonic content of the waveform expressed as a percentage of a fixed
nominal value rather than the fundamental value, giving total demand distortion I
0
:
I
0
=
1
u
nom
__
1
2
u
h
]
2

h=2

(4-29)
Total demand distortion is specified in [4.5] instead of THD for harmonic current distortion
for the above mentioned reasons. The fixed value u
nom
is required to be specified and may
be a maximum rms value, maximum demand, average or selected nominal system value.
4.5.2 Other harmonic PQ Terms
The following definitions have been adapted from IEEE Standard 519-1992.
deviation from a sine wave. A single number measure of the distortion of a sinusoid due to
harmonic components. It is equal to the ratio of the absolute value of the maximum difference
between the distorted wave and the crest value of the fundamental.


55

deviation from a sine wave, maximum theoretical. For a nonsinusoidal wave, the ratio of
the arithmetic sum of the amplitudes (rms) of all harmonics in the wave to the amplitude
(rms) of the fundamental.
distortion factor (harmonic factor). The ratio of the root-mean-square of the harmonic
content to the root-mean-square value of the fundamental quantity, expressed as a percent of
the fundamental.
F = _
sum o squorcs o omplituJcs o oll bormonics
squorc o omplituJc o unJomcntol
1uu%
harmonic. A sinusoidal component of a periodic wave or quantity having a frequency that is
an integral multiple of the fundamental frequency.
harmonic, characteristic. Those harmonics produced by semiconductor converter
equipment in the course of normal operation. In a six-pulse converter, the characteristic
harmonics are the nontriple odd harmonics, for example, the 5th, 7th, llth, 13th, etc.
b = kq _ 1
k = ony intcgcr
q = pulsc numbcr o con:crtcr
harmonic, noncharacteristic. Harmonics that are not produced by semiconductor converter
equipment in the course of normal operation. These may be a result of beat frequencies; a
demodulation of characteristic harmonics and the fundamental; or an imbalance in the ac
power system, asymmetrical delay angle, or cycloconverter operation.


56

power factor, displacement. The displacement component of power factor; the ratio of the
active power of the fundamental wave, in watts, to the apparent power of the fundamental
wave, in voltamperes (including the exciting current of the thyristor converter transformer).
power factor, total. The ratio of the total power input, in watts, to the total voltampere input
to the converter.
short-circuit ratio. For a semiconductor converter, the ratio of the short-circuit capacity of
the bus, in MVA, at the point of converter connection to the rating of the converter, in MW.

4.6 References
[4.1] Kreszig E., Advanced Engineering Mathematics, 8
th
Edition, 1999, New Delhi,
Wiley Publishers (India)
[4.2] Phillips, Parr, and Riskin, Signals, Sytems and Transforms, 3
rd
Edition, 2003, New
Delhi, Pearson Education
[4.3] Proakis & Manolakis, Digital Signal Processing, Principles, Algorithm and
Application, 3
rd
Edition, 1996, New Jersy, Prentice Hall Internatinal
[4.4] Philipps, J.K., Nelson, J.P., and Sen, P.K., Power quality and harmonic distortion on
distribution systems, IEEE Trans. on Industry Applications, 1994. Vol. 30(No. 2): p.
476-484.
[4.5] Dugan, R.C., et al., Electrical power systems quality. 2002, New York: McGraw
Hill.



57


Chapter 5 Survey on Harmonic Estimation

The severity of problems; caused by the presence of harmonics depend upon the sensitivity of
existing equipment and the magnitude of the waveform distortion. These problems range
from occasional nuisance trips to persistent problems, e.g., errors in kilowatt-hour metering,
and may even involve major damage including catastrophic failure.
Not only is excessive harmonic distortion unacceptable, it need not be tolerated. It is well-
known that many harmonic related problems can be corrected through relatively simple
system reconfiguration or the installation of harmonic filters. However, harmonic distortions
and their sources must be accurately identified before a verifiable solution can be
implemented. Knowing the location where problems have occurred within the system and the
location of harmonic sources and capacitor banks is prerequisite to any harmonic
measurement. Knowing these facts helps to determine where measurements should be made,
what types of measurements to take, and what type of result to expect. In chapter 3 we have
discussed the importance of harmonic analysis. Now that we know mathematics of harmonic
analysis from chapter 4 well go for harmonic analysis. There are two challenges in handling
the measurement of harmonics:
i. The first one is to get the sampled data. In this regard various standards are available.
ii. Next one is to analyze the data. We have carried out a literature survey to know about
the method available.


58

5.1 Harmonic Measurements
IEEE std.519-1992 [5.1] suggests that measurements of current and voltage harmonics are
essential for the reliable distribution of electric energy. The following are a few reasons that
highlight the importance of measurements:
Monitoring existing values of harmonics and checking against recommended or
admissible levels.
Testing equipment that generates harmonics.
Diagnosing and troubleshooting situations in which the equipment performance is
unacceptable to the utility or to the user.
Observing existing background levels and tracking the trends in time of voltage and
current harmonics (daily, monthly, seasonal patterns). Measuring for verification of
simulation studies that include harmonic load flow.
Measuring harmonic currents and harmonic voltages with their respective phase
angle. Such measurements can be made with and without a part of the nonlinear loads
connected, and can help determine the harmonic driving point impedance at a given
location.
The techniques used for harmonics measurements differ from those used for ordinary power
system measurement. The frequency bandwidth of the ordinary measurements of voltage,
current, and power can be accomplished with attention to a narrow band of frequencies near
the distribution frequency. Substantially wider bandwidths (up to 3 kHz) are required in the
study of power system harmonics.



59

5.1.1 Basic equipments used for the analysis of non-sinusoidal voltages and
currents
i. Oscilloscope: The display of the waveform on the oscilloscope gives immediate
qualitative information on the degree and type of distortion. Sometimes cases of
resonances are identifiable through the visible distortion that is present in the current
and voltage waveforms.
ii. Spectrum Analyzers: These instruments display the power distribution of a signal as a
function of frequency. A certain range of frequencies is scanned, and all the
components, harmonics, and interharmonics of the analyzed signal are displayed. The
display format may be a CRT or a chart recorder.
iii. Harmonic Analyzers or Wave Analyzers: These instruments measure the amplitude
(and in more complex units, the phase angle) of a periodic function. These
instruments provide the line spectrum of an observed signal. The output can be
recorded, or it can be monitored with analog or digital meters.
iv. Distortion Analyzers: These instruments indicate total harmonic distortion (THD)
directly.
v. Digital Harmonics Measuring Equipment: Digital analysis can be performed with
two basic techniques:
By means of digital filter: This method is similar to analog filtering. Dual-
channel digital signal analyzers include digital filtering. In the setup for a
particular measurement, the frequency range to be measured sets up the
digital filters for that range. Also, the bandwidth is varied to optimize the
capture of smaller harmonics in the presence of a very large fundamental.


60

The Fast Fourier Transform technique. These are real-time, very fast
methods of performing a spectrum analysis that permit the evaluation of a
large number of functions. Multichannel analog-digital conversion and micro
or mini computers are used for real time data acquisition.
When the waveform is recorded with suitable bandwidth using either analog or digital
techniques on-line, the Fast Fourier Transform (FFT) calculation of harmonic components,
the conversion to engineering units, the calculation of statistics, and the plotting and printing
of results can be performed off-line in the laboratory using suitable facilities.
5.1.2 Requirements for instrument response
For accurate harmonics measurements, the following important requirements must be met
[IEEE std 519-1992].
i. Accuracy: The instrument must perform the measurement of a constant (steady state)
harmonic component with an error compatible with the permissible limits. It is
reasonable to use an instrument with an uncertainty no larger than 5% of the
permissible limit. For example, assume a 480 V, three-phase system in which the 11
th

harmonic should be less than 0.70%. The line-neutral 11th harmonic, V
11
, is less than
1.94 V. This indicates that the instrument should have an uncertainty of less than
_(. 5)(1. 94) = _. 97 F .
ii. Selectivity: The selectivity of the instrument is an indication of its ability to separate
harmonic components of different frequencies. One practical way to ensure good
selectivity is to define requirements for minimum attenuation of an injected
frequency, while the instrument is set (tuned) at a frequency f
h
=60 Hz. Table 5.1
gives minimum required attenuation.


61

Table 5-1: Minimum Required Attenuation (dB)
Injected Frequency
(Hz)
Frequency-Domain
Instrument
Time-Domain
Instrument
60 0 0
30 50 60
120 to 720 30 50
720 to 1200 20 40
1200 to 2400 15 35

In many applications, the fundamental current may be very large in comparison to the
harmonic currents. The harmonic currents may be significant enough to cause serious
disturbance, as in the case of telephone interference. In such situations, the dynamic
range required for overall harmonic surveillance in a power system is important.
Almost all harmonic measuring devices can meet 60 dB (0.1% of fundamental)
minimum. Extra cost instruments reach down to 90 dB (0.00316%).
iii. Averaging or Snapshot. If the measured harmonics vary in time, it is necessary to
smooth out the rapidly fluctuating components over a period of time. Two factors
become important in this case: dynamic response and bandwidth.
iv. Dynamic Response. If, for example, an average over a period of 3 s is desirable, then
the response of the output meter should be identical to a first order low-pass filter
with a time constant of 1.S _ u.1S s.
5.2 Harmonic Analysis of Sampled Data
Analysis of acquired data is necessary to get useful information from it. The following works
have been found in literature survey:


62

5.2.1 Application of FFT for harmonic analysis
The DFT and FFT algorithms have been applied to many useful applications in power system
phasor measurements and harmonic analysis [5.2, 5.3]. However, misapplication of the FFT
algorithm would lead to incorrect results. There are basic assumptions embodied in the
application of DFT and FFT. These assumptions are:
i. The signal is stationary (constant magnitude);
ii. The sampling frequency is equal to the number of samples multiplied by the
fundamental frequency assumed by the algorithm;
iii. The sampling frequency is greater than twice the highest frequency in the signal to be
analyzed; and
iv. Each frequency in the signal is an integer multiple of the fundamental frequency.
Basically, the fundamental frequency is the reciprocal of the window length of data (T).
When these assumptions are satisfied, the results of the DFT or FFT are accurate.
5.2.1.1 Pit falls in FFT Algorithm
There are three major pitfalls in the application of FFT:
i. Aliasing,
ii. Leakage, and
iii. Picket-fence effect [5.4].
Aliasing can be alleviated by increasing the sampling frequency (
s
). However, pseudo-
aliasing may still occur even if the highest frequency component is not higher than
s
2 . This
may be caused by the presence of a fraction of a cycle of data or the presence of a white noise
in the measurements.
The term "leakage" refers to the apparent spreading of energy from one frequency into
adjacent ones. It arises due to the truncation of the sequence such that a fraction of a cycle
exists in the analyzed waveform. Data acquisition systems sampling rates are usually set at


63

fixed values, multiples of kHz. If the sampled waveform does not contain an integer number
of samples per integer number of cycles, the results of the DFT algorithm will include errors.
The resulting error is known as spectral "leakage" [5.4, 5.5]. The DFT and FFT of such a
sampled waveform will incorrectly indicate non-zero values for all of the harmonic
frequencies.
The picket-fence effect [5.4] occurs if the analyzed waveform includes a frequency which is
not one of the discrete frequencies (an integer times the fundamental). This influences the
accuracy of the magnitudes of each harmonic. Since the FFT is discrete, only those
frequencies that fall exactly on the discrete points in the frequency domain are indeed their
true value. The resulting frequency spectrum can be considered as the true spectrum viewed
through a picket fence with only those frequencies corresponding to gaps in the fence visible.
The peak values may be hidden from view behind the fence. Once established, the same
fundamental frequency, e.g., 50 Hz, must always be sampled. A varying frequency, as found
in a diesel generator that is not synchronized to line, may have fundamental or harmonic
frequencies that would be one of the in-between frequencies rather than 50 Hz and its
harmonics.
5.2.2 Application of Kalman filtering algorithm for optimal harmonic estimation
Application of Kalman filtering for the purpose of optimal harmonic estimation was reported
in the early 90s [5.6, 5.7, and 5.8]. The Kalman filtering approach provides a means for
optimally estimating phasors and the ability to track time-varying parameters. Kalman
Filtering technique represents the signal in state space form and then applies state space
theory to estimate the harmonics. This approach is simple, linear, and robust; however, it
requires a priori knowledge of the statistics of the electrical signal and the state matrix needs
to be defined accurately as well.


64

To overcome the short comings of Kalman filter in this regard, we require higher order FIR
filters, which are limited by ability of computer to perform calculations in real time.
Developments in this direction have origins in the theory of non-linear parameter estimation.
5.2.3 Application of wavelet transform for harmonic estimation
A new algorithm proposed in 1999 by V.L.Pham and KRWong [5.9] is among the first
attempts to identify all harmonics including integer, non-integer and sub-harmonics. In the
first step of the approach, the frequency spectrum of the waveform is decomposed into sub-
bands using discrete wavelet packet transform filter banks. In the second step, continuous
wavelet transform is applied to nonzero sub-bands to evaluate the harmonic contents. In the
first step, there is the problem of images due to down samplings and up samplings of the
waveform. A method for alleviating this image problem is developed in their paper. Methods
are developed to accurately quantify harmonics frequency amplitude and phase.
5.2.4 Application of artificial intelligence for harmonic estimation
In the very recent years artificial neural networks [5.10], evolutionary algorithms [5.11, 5.12],
optimization techniques [5.12, 5.13] have been proposed by various researchers.
Noting that the actual models of voltage/current signals are nonlinear in phase and linear in
amplitude, a hybrid algorithm iterating between genetic-algorithm (GA)-based phase
estimation and LS-based amplitude estimation is proposed in [5.11]. As a population-based
algorithm, unlike the conventional gradient based methods, GA does not rely on the gradient
information of an objective function; thus, it is unlikely to be trapped in local minima.
However, the efficiency of GA is significantly degraded when it is applied to a function
where the parameters that are optimized are highly correlated [5.13] (e.g., in harmonic
estimation, the change of the phase of the fundamental harmonic may alter the phases of


65

other harmonics). Thus, the optimization process of the GA will converge prematurely,
leading to the deterioration of the accuracy of the optimization results.
Lu et al proposed a very novel approach [5.13] based on particle swarm optimization with
passive congregation (PSOPC) to estimate harmonics. They proposed to apply PSOPC to
optimize the phase of each individual harmonic, because the particle swarm optimizer (PSO)
has shown a faster convergence rate than other evolutionary algorithms on some problems
and has fewer parameters to adjust [5.14]. By introducing passive congregation, the
information sharing mechanism is improved and the optimization result is more accurate.
5.3 Conclusion
PSOPC seem to be most attractive among all the algorithms discussed above, because of its
ease of implementation. It is also faster than most of the other algorithms. This algorithm first
requires estimation of amplitude by some linear estimation procedure. Lu et al. proposed to
use Least Square Estimation method. Well also stick to the same method for our harmonic
analysis. Before we go onto discussing particle swarm optimization, well discuss about least
square estimation method first.









66

5.4 References
[5.1] IEEE std. 519:1992: IEEE Recommended Practices and Requirements for Harmonic
Control in Electrical Power Systems
[5.2] Thorp, J . S., Phadke, A. G., and Karimi, K. I., Real-Time Voltage Phasor
Measurements for Static-State Estimation, IEEE Transactions on Power Auuaratus
and Svstems, Vol. PAS-104, No. 11, November 1985, pp.3099-3 106.
[5.3] Heydt, G. T. , A New Method for the Calculation of Subtransmission and
Distribution System Transients Based on the FFT, IEEE Transactions on Power
Delivery, Vol. 4, No. 3, J uly 1989, pp.1869-1875.
[5.4] Girgis, A. A., and Ham, F., A Qualitative Study of Pitfalls in FFT, IEEE Transactions
on Aerosvace and Electronic Svstems. Vol. AES 16, No. 4, pp.434-439, J uly 1980.
[5.5] Ham, F. M., and Girgis , A. A., Measurement of Power Frequency Fluctuations
Using FFT, IEEE Transactions on Industrial Electronics, Vol. IE-32, No. 8, August
1985, pp.199-204.
[5.6] Girgis et. al., Measurement and characterization of Harmonic and high frequency
distortion for large industrial load, IEEE Transactions 1989
[5.7] Girgis et al., Digital recursive measurement scheme for online tracking of power
system harmonics, IEEE Transactions on Power Delivery ,Vol. 6, No. 3, J uly 1991
[5.8] Ma, H. and Girgis, A. A., Identification and tracking of harmonic sources in a power
system using kalman filter, IEEE Trans. Power Del., vol. 11, no. 4, pp. 16591665,
Oct. 1996.
[5.9] Pham, V.L. and Wong, K. R., Wavelet-transform-based algorithm for harmonic
analysis of power system waveforms, IEEE Proc.-Gener. Tronsm. Disirib.. Vol. 146.
No. 3, May I999
[5.10] Kaewarsa et al., Recognition of power quality events by using multiwavelet-based


67

neural networks , Elsevier J ournal of Electrical Power and Energy Systems 30 (2008)
254260
[5.11] Mishra, S., A Hybrid Least Square-Fuzzy Bacterial Foraging Strategy for Harmonic
Estimation, IEEE Transactions on Evolutionary Computation, Vol. 9, No. 1,
February 2005
[5.12] Bettayeb, M. and Qidwai, U., A hybrid least squares-GA-based algorithm for
harmonic estimation, IEEE Trans. Power Del., vol. 18, no. 2, pp. 377382, Apr.
2003.
[5.13] Lu, Z., J i , T. Y., Tang, W. H., Optimal Harmonic Estimation Using A Particle
Swarm Optimizer, IEEE Transactions on Power Delivery, Vol. 23, No. 2, April 2008
[5.14] Kennedy, J . and Eberhart, R. C., Swarm Intelligence. San Mateo, CA: Morgan
Kaufmann, 2001.




68


Chapter 6 A Harmonic Estimation
Technique based on LSE & PSOPC

The algorithm discussed toward the end of this chapter is based on Linear Least Square
Estimator (LSE) and Particle Swarm Optimizer (PSO). Well discuss LSE and PSO in the
next two sections, and then well propose the harmonic estimation technique.
6.1 Introduction to the method of Least Square Estimation
6.1.1 Principle of least squares
According to Gauss the principle of least squares is that the unknown parameters of a model
should be chosen in such a way that
The sum of the squares of the differences between the actually observed and computed
values multiplied by numbers that measure the degree of precision is a minimum.
To be able to give an analytic solution, the computed values must be linear functions of the
unknown parameters.
6.1.2 Method of least squares:
Least squares estimation becomes particularly simple if the mathematical model can be
written in the form:


69

y(t) = p
1
(t)0
1
+p
2
(0
2
) ++ p
n
(t)0
n
=
1
(t). 0 (6-1)
Where y = observed variable
0
1
, , 0
n
= Unknown parameters
p
1
, , p
n
= Known functions that may depend on other known
variables.
The variables p
1
, p
2
, p
N
are called regression variables or regressors and the model of
equation (6-1) is called a regression model.
Define

1
(t) = |p
1
(t) p
2
(t) p
n
(t)]
0
1
= |0
1
0
2
0
n
]
Then we can write
y(t) =
1
(t)
1n
. 0
n1
(6-2)
If N measurements are taken then equation can be written as
y(1) = |p
1
(1) p
2
(1) p
n
(1)] _
0
1
0
2
.
0
n
_
y(2) = |p
1
(2) p
2
(2) p
n
(2)] _
0
1
0
2
.
0
n
_


70

y(N) = |p
1
(N) p
2
(N) p
n
(N)] _
0
1
0
2
.
0
n
_
or, =
where = N 1 vector
= N n Matrix
and
= _
p
1
(1) p
2
(1) p
n
(1)
p
1
(2) p
2
(2) p
n
(2)
. . . .
p
1
(N) p
2
(N) p
n
(N)
_
1 =
l
l
l
l

1
(1)

1
(2)
.

1
(N)1
1
1
1

The problem is to determine the parameters in such a way that the output computed from the
modes of equation (6-1) agrees as closely as possible with the measured variables y(i) in the
sense of least squares.
To obtain the least squares estimate we minimize
[ =
1
N
c
2
(t)
N
t=1
=
1
N
(y(t) -
1
(t). 0)
2
N
t=1

In vector form, we want to minimize
[ =
1
N
( -0
`
)
1
( - 0
`
) =
1
N
|
1
-
1
0
`
- 0
`
1

1
+0
`
1

1
0
`
]
All terms in this expression are scalar. We can therefore transpose a scalar without changing
it. Now


71

(
1
0
`
)
1
= 0
`
1

1
=
1
0
`


Therefore,
[ =
1
N
|
1
- 0
`
1

1
- 0
`

1
+ 0
`
1

1
0
`
]

Now setting to zero the derivative of [ with respect to 0
`
.
o[
o0
`
= -2
1
+ 2
1
0
`
= u
or,

1
=
1
0
`

or,
0

= (4
T
4)
-1
4
T
Y (6-3)
This solve for a unique minimum if the second derivative matrix
6
2
]
60
2
= 2(4
T
4)
is positive definite.







72

6.2 Introduction to the Particle Swarm Optimization based Techniques
The particle swarm optimizer (PSO) is a population-based algorithm that was invented by
Kennedy and Eberhart [6.1] in 1995. It is inspired by the social behavior of animals such as
fish schooling and bird flocking. Similar to other population-based algorithms, such as
evolutionary algorithms, PSO can solve a variety of difficult non-linear optimization
problems and has shown a faster convergence rate than other evolutionary algorithms on
some problems [6-2]. Another advantage of PSO is that it has very few parameters to adjust,
which makes it particularly easy to implement.
The foundation of PSO is based on the hypothesis that social sharing of information among
con-specifics offers an evolutionary advantage [6.1]. The SPSO model is based on the
following two factors [6.1]:
i. The autobiographical memory, which remembers the best previous position of each
individual (P

) in the swarm;
ii. The publicized knowledge, which is the best solution (P
g
) found currently by the
population.
Therefore, the sharing of information among con-specifics is achieved by employing the
publicly available information P
g
, shown in Figure 6-1. There is no information sharing
among individuals except that P
g
broadcasts the information to the other individuals.
Therefore, the population may lose diversity and is more likely to confine the search around
local minima if committed too early in the search to the global best found so far.


73


Figure 6-1: Interaction between particles and the best particle gbest
Angeline [6.3] pointed out that although PSO may outperform other evolutionary algorithms
in the early iterations, its performance may not be competitive as the number of generations is
increased. Recently, several investigations have been undertaken to improve the performance
of standard PSO (SPSO). Lbjerg et al. [6.4] presented a hybrid PSO model with breeding
and subpopulations. Kennedy and Mendes [6.5] investigated the impacts of population
structures to the search performance of SPSO. Other investigations on improving PSOs
performance were undertaken using cluster analysis [6.6] and fuzzy adaptive inertia weight
[6.7].
Biologists have proposed four types of biological mechanisms that allow animals to
aggregate into groups: passive aggregation, active aggregation, passive congregation, and
social congregation [6.8]. There are different information sharing mechanisms inside these
forces. It has been found that the passive congregation model is suitable to be incorporated in
the SPSO model.
6.2.1 Standard Particle Swarm Optimizer (SPSO)
PSO is a population-based optimization algorithm. The population of PSO is called a swarm
and each individual in the population of PSO is called a particle. The i
th
particle at iteration k
has the following two attributes:


74

i. A current position in an N-dimensional search space X

k
= (x
1
k
, x
2
k
, , x
N
k
), where
x
n
k
e |l
n
, u
n
], 1 n N, l
n
onJ u
n
are the lower and upper bound for the n
th

dimension, respectively.
ii. A current velocity , I

k
= (:
1
k
, :
2
k
, , :
N
k
), which is bounded by a maximum velocity
I
mux
k
= (:
1mux
k
, :
2mux
k
, , :
Nmux
k
) and a minimum velocity
I
mn
k
= (:
1mn
k
, :
2mn
k
, , :
Nmn
k
).
In each iteration of PSO, the swarm is updated by the following equations [6.1]:
I

k+1
= I

k
+ c
1
r
1
(P

k
- X

k
) +c
2
r
2
(P
g
k
-X

k
) (6-4)
X

k+1
= X

k
+I

k+1
(6-5)
where P

is the best previous position of the ith particle (also known as pbest). According to
the different definitions of P
g
, there are two different versions of PSO.
If P
g
is the best position among all the particles in the swarm (also known as gbest),
such a version is called the global version.
If P
g
is taken from some smaller number of adjacent particles of the population (also
known as lbest), such a version is called the local version. P
i
and P
g
are given by the
following equations, respectively:
P

= _
P

; i [(X

) P

; i [(X

) < P

_ (6-6)
P
g
e {P
0
, P
1
, , P
m
]|[(P
g
) = min ([(P
0
), [(P
1
), , [(P
m
)) (6-7)


75

where [ is the objective function, m Hand H is the total number of particles, r
1
and r
2
are
elements from two uniform random sequences in the range (0,1): r
1
U(0,1); r
2
U(0,1), and
is an inertia weight [6.9] which is initialized typically in the range of [0,1]. A larger inertia
weight facilitates global exploration and a smaller inertia weight tends to facilitate local
exploration to fine-tune the current search area [6.10]. The variables c
1
and c
2
are acceleration
constants[6.7], which control how far a particle will move in a single iteration.
Another important variant of standard PSO is the constriction factor approach PSO (CPSO),
which was proposed by Clerc and Kennedy [6.11]. The velocity of CPSO is updated by the
following equation:
I

k+1
= _(I

k
+c
1
r
1
(P

k
-X

k
) +c
2
r
2
(P
g
k
-X

k
)) (6-8)
where is called a constriction factor, given by:
_ =
2
|2 -
2
-4|

where, = c
1
+c
2
, > 4
(6-9)
The CPSO ensures the convergence of the search procedures and can generate higher-quality
solutions than the standard PSO with inertia weight on some studied problems [6.12].






76

6.2.2 Particle Swarm Optimizer with Passive Congregation (PSOPC)
The PSO algorithm is inspired by social behaviors such as spatial order, more specially,
aggregation such as bird flocking, fish schooling, or swarming of insects. Each of these cases
has stable spatio-temporal integrities of the group of organisms: the group moves persistently
as a whole without losing the shape and density.
For each of these groups, different biological forces are essential for preserving the groups
integrity. Parrish and Hamner [6.13] proposed mathematical models of the spatial structure of
animal groups to show how animals organize themselves. In these models, aggregation
sometimes refers to a grouping of the organisms by non-social, external, physical forces.
There are two types of aggregation: passive aggregation and active aggregation. Passive
aggregation is a passive grouping by physical processes. One example of passive aggregation
is the dense aggregation of plankton in open water, in which the plankton are not attracted
actively to the aggregation but are transported passively there via physical forces such as
water currents. Active aggregation is a grouping by attractive resource, such as food or space,
with each member of the group recruited to a specific location actively.
Congregation, which is different from aggregation, is a grouping by social forces that is the
source of attraction is the group itself. Congregation can be classified into passive
congregation and social congregation. Passive congregation is an attraction of an individual
to other group members but where there is no display of social behavior. Social
congregations usually happen in a group where the members are related (sometimes highly
related). A variety of inter-individual behaviors are displayed in social congregations,
necessitating active information transfer [6.13]. For example, ants use antennal contacts to
transfer information about individual identity or location of resources [6.14].


77

From the definitions above, the third part of Equation (6-1): c
2
r
2
(P
g
k
-X

k
)can be classified
as either active aggregation or passive congregation. But since P
g
is the best solution the
swarm has found so far, which can be regarded as the place with most food, researchers argue
that it is better to classify c
2
r
2
(P
g
k
- X

k
)as active aggregation.
It has been discovered that in spatially well-defined congregations, such as fish schools,
individuals may have low fidelity to the group because the congregations may be composed
of individuals with little to no genetic relation to each other [6.15]. Schooling fish are
generally considered a selfish herd[6.16], in that each individual attempts to take the
sweeping generalization advantage from group living, independent of the fates of neighbors
[6.17]. In these congregations, information may be transferred passively rather than actively
[6.18]. Such asocial types of congregations can be referred to as passive congregation.
Because PSO is inspired by fish schooling, it is therefore natural to ask if a passive
congregation model can be employed to increase the performance of SPSO. Here, other
models have not been considered such as passive aggregation, because PSO is not aggregated
passively via physical processes. And social congregation usually happens when group
fidelity is high, such that the chance of each individual meeting any of the others is high
[6.19]. Social congregations frequently display a division of labor. In a social insect colony,
such as an ant colony, large tasks are accomplished collectively by groups of specialized
individuals, which is more efficient than performing sequentially by unspecialized
individuals [6.20]. The concept of labor division can be employed by data clustering, sorting
[6.21] and data analysis [6.22].
Group members in an aggregation can react without direct detection of incoming signals from
the environment, because they can get necessary information from their neighbors [6.8].
Individuals need to monitor both environment and their immediate surroundings, such as the


78

bearing and speed of their neighbors [6.8]. Therefore, each individual in an aggregation has a
multitude of potential information from other group members that may minimize the chance
of missed detection and incorrect interpretations [6.8]. Such information transfer can be
employed in the model of passive congregation. Inspired by this result, and to keep the model
simple and uniform with SPSO, a hybrid PSO with passive congregation is proposed as
follows:
I

k+1
= I

k
+ c
1
r
1
(P

k
-X

k
) + c
2
r
2
(P
g
k
- X

k
) +c
3
r
3
(R

k
-X

k
) (6-10)
X

k+1
= X

k
+ I

k+1
(6-11)
where R
i
is a particle selected randomly from the swarm, c
3
the passive congregation
coefficient, and r
3
is a uniform random sequence in the range (0,1): r
3
U(0,1). The
interactions between individuals of PSOPC are shown in Figure. 6-2, the pseudo-code for
PSOPC is listed in Table 6-1.

Figure 6-2: Interactions of particles with passive congregation





79

Table 6-1: Pseudo-code for the PSOPC algorithm
Set k:=0;
Randomly initialize positions and velocities of all particles;
WHILE (the termination conditions are not met)
FOR (each particle i in the swarm)
Calculate fitness: Calculate the fitness value of current particle: [(X

)
Update phext : Compare the fitness value of pbest with [(X

). If [(X

)is better than the fitness

value of pbest, then set pbest to the current position X
i
;
Update ghext : Find the global best position of the swarm. If [(X

)is better than the fitness

value of gbest, then gbest is set to the position of the current particle X
i
;
Update R
|
: Randomly select a particle from the swarm as R

;
Update velocities: Calculate velocities V
i
using Equation (6-7). If I

> I
mux
then I

= I
mux
. If
I

< I
mn
then I

= I
mn

Update positions: Calculate positions X
i
using Equation. (6-8);
END FOR
Set k:=k+1;
END WHILE




80

6.3 LSE and PSOPC based Harmonic Estimation Technique
This section presents an estimaton procedure for harmonic estimation using PSOPC (a
nonlinear optimization algorithm) and the linear LS method. Assume an electrical signal
Z(t) is described as
Z(t) = A
n
sin(
n
t +
n
)
N
n=1
+ :(t) (6-12)
where n = 1, 2, S, , N represents the order of the harmonic; A
n
,
n
onJ
n
are the
amplitude, phase angle, and angular frequency of the n
th
harmonic, respectively,
n
= 2n
n
;
and :(t) is the additive noise.
To estimate A
n
,
n
onJ
n
of each harmonic, the following function is constructed:
Z(t)

= A
n

sin(
n
t +
n

)
N
n=1
(6-2)
where A
n

,
n
, onJ
n

are the estimatio of A

n
,
n
onJ
n
respetively. Thereby, the original
signal can be represented as
Z(t) = Z(t)

+r(t) (6-3)
where r(t) is the estimation error or estimation residual that indicates the difference between
Z(t) and Z(t)

. the estimation goal is to search for the best A

n
,
n
onJ
n
; sufficing that
r(t) is the smallest in condition of N being determined.
Since the phases of the harmonics in the model are nonlinear and are restricted within |u, 2n],
the recursive optimization algorithm is utilized to estimate the values of
n
onJ
n
. Once the
phases and the frequencies are estimated in each iteration, the amplitude A
n
is obtained by the
standard LS method.


81

Sampling the signal Z(t) withK samples, the discrete linear model is given as
Z(k) = E(k) A + :(k), k = 1, 2, , K (6-4)
E(k) =
l
l
l
l
l
l
sin(
1
t
1
+
1
) sin(
N
t
1
+
N
)
sin(
1
t
2
+
1
) sin(
N
t
2
+
N
)
. . .
sin(
1
t
k
+
1
) sin(
N
t
k
+
N
)
. .
sin(
1
t
K
+
1
) sin(
N
t
K
+
N
)1
1
1
1
1
1
(6-5)
where Z(k) is the k th sample of the measured signal with additive noise :(k) ; A =
|A
1
A
2
A
N
]
1
is the amplitude vector. We aim to estimate it using LS algorithm. E(k)
is the system structure matrix. As stated earlier, the goal is to find the best A
n

,
n
, onJ
n

which minimizes the difference between (k), onJ Z(k)

= E(k)

A
`
.
After the values of
n
, onJ
n

are evaluated using a certain optimization algorithm (PSOPC

is our case), E(k)

is calculated consequently. Assuming E(k)

is a full rank matrix, the

estimation of A
`
can be obtained via the standard LS algorithm as
A
`
= |E(k)

1
E(k)

]
-1
E(k)

1
Z(k)
(6-6)
which ensures that the estimation of the signal
Z(k)

= E(k)

A
`
(6-7)
is best in the condition of
n
, onJ
n

. However, the values of

n
, onJ
n

are not the best

solution and needed to be optimized. Therefore, in the next iteration,
n
, onJ
n

are updated
using a PSOPC, according to the cost function [, which is defined as


82

[ =
(Z(k) - Z(k)

)
2 K
k=1
Z
k
2 K
k=1

The process repeats until the convergence condition is met.
Letting F = |
1

2

N
] and = |
1

2

N
] we represent i th particle as
X

= |
F
]. The flow chart for the applied algorithm is given in figure 6-3.
Based on this flow chart and knowledge acquired in previous sections we have written a
Matlab script that can estimate a given number of odd harmonics in the sampled signal whose
sampling interval is known. The script is given in appendix. A variant of the above script is
used to get all the harmonics in the signal. This script is also given in the appendix. We have
also found the total harmonic distortion and deviation in sine wave also for each case.



83


Figure 6-3: Flow chart for LSE & PSOPC based Harmonic Estimation Technique



84

6.4 References
[6.1] Kennedy, J . and Eberhart, R., Particle swarm optimization, IEEE International
Conference on Neural Networks vol. 4, IEEE Press (1995).
[6.2] Kennedy, J . and Eberhart, R.C., Swarm Intelligence, Morgan Kaufmann Publishers
(2001).
[6.3] Angeline, P., Evolutionary optimization versus particle swarm
[6.4] Lbjerg, M., Rasmussen, T.K., Krink, K., 2001. Hybrid particle swarm optimizer
with breeding and subpopulations. In: Proceedings of the Third Genetic and
Evolutionary Computation Conference (GECCO-2001), vol. 1. pp. 469476.
[6.5] Kennedy, J . and Mendes, R., Population structure and particle swarm performance,
Proceedings of the 2002 Congress on Evolutionary Computation CEC2002, IEEE
Press (2002), pp. 16711676.
[6.6] Kennedy, J ., Stereotyping: improving particle swarm performance with cluster
analysis, Proceedings of the IEEE International Conference on Evolutionary
Computation (2000), pp. 15071512.
[6.7] Eberhart, R.C. and Shi, Y., Particle swarm optimization: developments, applications
and resources, Proceedings of IEEE International Conference on Evolutionary
Computation (2001), pp. 8186.
[6.8] Parrish, J .K. and Hamner, W.M., Animal Groups in Three Dimensions, Cambridge
University Press, Cambridge, UK (1997).
[6.9] Shi, Y. and Eberhart, R.C., A modified particle swarm optimiser, Proceedings of the
IEEE International Conference on Evolutionary Computation (1997), pp. 303308.
[6.10] Shi, Y., Eberhart, R.C., 1998. Parameter selection in particle swarm optimization. In:
Evolutionary Programming VII (1998). Lecture Notes in Computer Science, vol.
1447. Springer, pp. 591600.


85

[6.11] Clerc, M. and Kennedy, J ., The particle swarm: explosion, stability, and convergence
in a multi-dimensional complex space, IEEE Trans. Evolut. Comput. 6 (2002) (1),
pp. 5873.
[6.12] Eberhart, R. and Shi, Y., Comparing inertia weights and constriction factors in
particle swarm optimization, Proceedings of the Congress on Evolutionary
Computation (CEC2000) (2000), pp. 8488.
[6.13] Parrish, J .K. and Hamner, W.M., Animal Groups in Three Dimensions, Cambridge
University Press, Cambridge, UK (1997).
[6.14] Gordon, D.M., Paul, R.E. and Thorpe, K., What is the function of encounter pattern
in ant colonies?, Anim. Behav. 45 (1993), pp. 10831100
[6.15] Hilborn, R., Modelling the stability of fish schools: exchange of individual fish
between schools of skipjack tuna (Katsuwonus pelamis), Can. J. Fish. Aquat. Sci. 48
(1991), pp. 10801091.
[6.16] Hamilton, W.D., Geometry for the selfish herd, J. Theor. Biol. 31 (1971), pp. 295
311.
[6.17] Pitcher, T.J . and Parrish, J .K., Functions of shoaling behaviour in teleosts In: T.J .
Pitcher, Editors, Behaviour of Teleost Fishes, Chapman and Hall, London (1993), pp.
363439.
[6.18] Magurran, A.E. and Higham, A., Information transfer across fish shoals under
predator threat, Ethology 78 (1988), pp. 153158.
[6.19] Alexander, R.D., The evolution of social behaviour, Annu. Rev. Ecol. Syst. 5 (1974),
pp. 325383.
[6.20] Bonabeau, E., Dorigo, M. and Theraulaz G., Swarm Intelligence: From Natural to
Artificial Systems, Oxford University Press (1999).
[6.21] Deneubourg, J .L., Goss, S., Franks, N., Sendova-Franks, A., Detrain C. and Chretien


86

L., The dynamics of collective sorting: robot-like ant and ant-like robot, Proceedings
of the First Conference on Simulation of Adaptive Behavior: from Animals to
Animals (1991), pp. 356365.
[6.22] Lumer, E. and Faieta B., Diversity and adaptation in population of clustering ants,
Proceedings of the Third International Conference on Simulation of Adaptive
Behavior: From Animals to Animals (1994), pp. 499508.
[6.23] He, S. et al., A particle swarm optimizer with passive congregation, Elsevier J ournal
of BioSystems 78 (2004) 135147




87


Chapter 7 Real Time Data Collection for
the Purpose of Harmonic Estimation

Our study concentrates on the power quality problems related to harmonics, so we decided to
have an estimate of harmonic content in voltage supply in our institute premises only. We
conducted an experiment to collect the wave shape data in the Electrical Machines
Laboratory of Electrical Engineering Department. We also wanted to see the effect of loading
a good capacity industrial rectifier on the supply of a Diesel Engine Driven Alternator of
comparable capacity. So we collected data in various operating environments of the
laboratory. Well describe the methodology in the form of an experiment report.
7.1. Aim of the Experiment
To collect the wave shape/spectrum data for power quality analysis of voltage waves
generated by a 3 phase Diesel Engine-Alternator set and supply mains.
7.1.1 Equipments used
1. Diesel Engine-Alternator Set
a. Alternator
Manufacturer Kirlosker Ltd.
Frame CM200A
kVA 40.0


88

RPM 1500
VOLT 415
AMP 55.8
Rotation Reversible viewing at the driven end
Rating S1, BS 2613-1970
Insulation Class ST-F, RT-F, EX-F
Connection Y
Phase 3
Hz 50
Excitation 285 V, 4.3 A
PF 0.8
Phase Sequence BS 822-UVW
P Code/ Machine No. 292421805/ 8602AD-27
b. Diesel Engine
Manufacturer Kirlosker Ltd.
Type BB44
RPM 1500
Bore x Stroke 110 x 116
kW 37.2
BHP 50.5 B55514
Engine No. 481208185175
2. Rheostats: 100 ohm, 5 A (3 in no.)
3. Ammeters: 0-3 A (3 in no.), MI Type
4. Lamp Load: 3 Phase type (had 2; 200 W/215 V bulbs in series in each phase)



89

5. Digital Storage Oscilloscope
Manufacturer Tektronix Ltd.
Model No. TDS 2022 B
Channels 2
Sample Rate 2.0 GS/s
Band width 200MHz
Display Color
Record Length 2500 data points
PC Connectivity Through USB & Open Choice PC Communication software
Probe
*
P2220
*
We used attenuation setting at 10X to utilize the whole available band width. At 1X
band width is only 6 MHz for the instrument.
6. Digital Multi-meter available in the Laboratory
7. PC: HP Pavilion dv2025TU Notebook PC
7.1.2 Circuit diagram of the experimental setup

Figure 7-1: Circuit diagram of the experimental setup


90

7.1.3 Procedure followed
1. Connected the equipments and instruments as shown in the circuit diagram.
2. Connected the oscilloscope to the PC through USB port.
3. We assumed that all the circuit elements we linear and did inject no harmonics into the
circuit. To get the voltage waveform we used rheostats as potential dividers and
saw/acquired the waveform/spectrum across almost 10 ohm resistance (measured using
multi-meter). We ensured that the potential difference across the two terminals of the
probe was within the safe limits. For this also we used multi-meter.
4. We used ammeters to ensure that the load was balanced.
7.1.4 Observations
We observed/acquired the following things:
1. Voltage waveform
2. Spectrum
Both of the above; under various operating condition inside the Laboratory. We acquired the
data into the PC through the USB interface provided with the oscilloscope. Sampling time of
the storage oscilloscope was 0.00004 sec. For the purpose of review the acquired images of
wave form and spectrum are given here.






Table 7-1:: Acquired Wa
Wa
Wa
3.
Wa
Wa
ave form and
aveform
2. Di
aveform
. Service M
aveform
aveform
Spectrum ima
1. S
esel Engine-
Mains Loaded
91
ages
Service Main
-Alternator s
d with Recti
ns
Sp
set on No Lo
Sp
fier and DC
Sp
Sp
pectrum
oad
pectrum
Machines
pectrum
pectrum


7.1.5 Con
1. M
o
2. U
w
3. O
o
4.
Wa
nclusions fr
Mains supply
f the laborat
Under no loa
within the lim
On a normal
f loads:
a. Rectif
b. DC R
c. AC R
d. Trans
Diesel Eng
aveform
rom the ima
y being draw
tory.
ad condition
mits and are d
day of the la
fier
Rotating Mac
Rotating Mac
formers
gine-Alterna

ages
wn from the
s the alterna
due to the to
aboratory th
chines
chines
92
ator set on no
national gri
ator produce
oothed struct
he alternator
ormal Labor
Sp
d id not affe
es insignifica
ture of the st
is burdened
ratory load
pectrum
ected by the
ant harmoni
tator and roto
d with the fo
rectifier loa
ics. These ar
or.
ollowing type

ad
re
es


93

Out of these loads rectifier is the biggest culprit for injecting harmonics into the
electrical circuit of the laboratory. This is observed in the wave shape and spectrum
also. So the data is suggested to be analyzed for harmonics.



94


Chapter 8 Application of the LSE & PSOPC
Based Harmonic Estimation Technique to
the Real Time Data

We have found the following results by applying the proposed LSE+PSOPC based Harmonic
Estimation technique to the real time data we have collected. We are giving here, the output
of the estimator scripts (Matlab code of the Estimator is given in the appendix).
8.1 Output of the script estimating odd harmonics
The script PSOPC.m was run for 15 number of particle and 100 generations in the following
data samples. The results were published in word format using the GUI interface of Matlab,
and the same are pasted here.



95

i. Output of the the script for data
sample no. 1 (data1.m).
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
DateofExp:28Jan2009
Source:DGSetphaseA
WorkEnvironmentoftheLaboratory:Withoutanyload

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

ComputedOutputintheformat:
Amplitude(V)Freq.(Hz)Phase(rad)
1.0e+003*

0.01310.04640.0012
0.00000.13930.0002
0.00020.23220.0006
0.00020.32510.0016
0.00000.41790.0001
0.00000.51080.0031
0.00000.60370.0055
0.00000.69650.0007
0.00010.78940.0041
0.00010.88230.0035
0.00000.97520.0033
0.00001.06800.0018
0.00001.16090.0018
0.00001.25380.0041
0.00001.34660.0020

NormalizedOutputintheformat:
Amplitude(%)Freq.(Hz)Phase(deg)
1.0e+003*
0.10000.04640.1116
0.00030.13930.1889
0.00140.23220.1472
0.00130.32510.0895
0.00010.41790.0061
0.00020.51080.0038
0.00020.60370.3168
0.00010.69650.0406
0.00040.78940.2341
0.00040.88230.0195
0.00010.97520.1863
0.00001.06800.0747
0.00011.16090.0785
0.00011.25380.0530
0.00001.34660.0666

DfSW:Deviationfromsinewave(in%)
DfSW=
5.2525

THD:TotalHarmonicDistortion(in%)
THD=
2.1097


96


Figure 8-1: Spectrum of data sample no. 1

Figure 8-2: Estimation Error for data sample no. 1

i. Output of the script running
for data sample no. 2
(data2.m).

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
DateofExp:28Jan2009
Source:MainsphaseA
WorkEnvironmentoftheLaboratory:Noloadinsidethelab

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
ComputedOutputintheformat:
Amplitude(V)Freq.(Hz)Phase(rad)
1.0e+003*

0.01350.04940.0031
0.00020.14840.0029
0.00040.24730.0001
0.00010.34620.0011
0.00000.44510.0016
0.00010.54400.0005
0.00000.64290.0012
0.00000.74180.0015
0.00000.84070.0025
0.00000.93960.0025
0.00001.03850.0020
0.00001.13740.0000
0.00001.23640.0009
0.00001.33530.0001
0.00001.43420.0025

NormalizedOutputintheformat:
Amplitude(%)Freq.(Hz)Phase(deg)
1.0e+003*

0.09990.04940.0005
0.00120.14840.0130
0.00300.24730.0048
0.00040.34620.0652
0.00000.44510.0921
0.00070.54400.1505
0.00030.64290.1127
0.00000.74180.0840


97

0.00000.84070.1424
0.00030.93960.1411
0.00011.03850.1161
0.00001.13740.1782
0.00011.23640.1271
0.00001.33530.0064
0.00001.43420.0374

DfSW:Deviationfromsinewave(in%)
DfSW=
6.3402

THD:TotalHarmonicDistortion(in%)
THD=
3.3011

Figure 8-3: Spectrum of data sample no. 2

Figure 8-4: Estimation error for data sample no. 2

ii. Output of the script running
for data sample no. 3
(data4.m).

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
DateofExp:29Jan2009
Source:DGSetphaseA
WorkEnvironmentoftheLaboratory:NormalWorkingDay

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
ComputedOutputintheformat:
Amplitude(V)Freq.(Hz)Phase(rad)
1.0e+003*

0.01920.04600.0032
0.00000.13820.0004
0.00050.23030.0010
0.00050.32240.0005
0.00010.41460.0004
0.00000.50670.0016
0.00010.59880.0004
0.00010.69100.0011


98

0.00020.78310.0002
0.00010.87530.0016
0.00010.96740.0015
0.00011.05950.0014
0.00001.15170.0016
0.00001.24380.0026
0.00001.33590.0013

NormalizedOutputintheformat:
Amplitude(%)Freq.(Hz)Phase(deg)
1.0e+003*

0.09990.04600.0026
0.00000.13820.1586
0.00260.23030.1218
0.00240.32240.0310
0.00040.41460.0228
0.00020.50670.0857
0.00060.59880.1572
0.00030.69100.1187
0.00120.78310.1706
0.00030.87530.0897
0.00050.96740.0862
0.00031.05950.0822
0.00001.15170.0900
0.00021.24380.0317
0.00011.33590.0747
DfSW:Deviationfromsinewave(in%)
DfSW=
1.1463

THD:TotalHarmonicDistortion(in%)
THD=
3.8806

Figure 8-5: Spectrum of data sample no. 3

Figure 8-6: Estimation error in data sample no. 3
iii. Output of the script for data
sample no. 4 (data7.m).

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
DateofExp:30Jan2009
Source:DGSetphaseB
WorkEnvironmentoftheLaboratory:NormalWorkingDay

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
ComputedOutputintheformat:


99

Amplitude(V)Freq.(Hz)Phase(rad)
1.0e+003*

0.01880.04630.0000
0.00020.13910.0032
0.00080.23190.0031
0.00070.32460.0020
0.00000.41740.0008
0.00030.51020.0013
0.00010.60300.0031
0.00000.69570.0012
0.00000.78850.0023
0.00000.88130.0025
0.00000.97410.0008
0.00001.06690.0018
0.00001.15960.0011
0.00001.25240.0019
0.00001.34520.0011

NormalizedOutputintheformat:
Amplitude(%)Freq.(Hz)Phase(deg)
1.0e+003*

0.09980.04630.0012
0.00110.13910.0015
0.00450.23190.1770
0.00350.32460.0629
0.00010.41740.0470
0.00150.51020.0765
0.00040.60300.1751
0.00020.69570.1094
0.00010.78850.1297
0.00000.88130.0381
0.00010.97410.0437
0.00021.06690.1043
0.00011.15960.0611
0.00011.25240.0697
0.00011.34520.0654

DfSW:Deviationfromsinewave(in%)

DfSW=
2.0354

THD:TotalHarmonicDistortion(in%)
THD=

6.0416

Figure 8-7: Spectrum for data sample no. 4


100








Figure 8-8: Estimation error for data sample no. 4




101

8.2 Output of the script estimating all (even +odd ) harmonics
In the following outputs, the script PSOPCh.m was run for 60 generations with 10 number of
particles.

i. For data sample no. 5 (data3.m)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
DateofExp:29Jan2009
Source:DGSetphaseA
WorkEnvironmentoftheLaboratory:Withoutconverterload

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
ComputedOutputintheformat:
Amplitude(V)Freq.(Hz)Phase(rad)
1.0e+003*

0.01930.04700.0031
0.00000.09400.0014
0.00000.14110.0024
0.00010.18810.0023
0.00040.23510.0028
0.00000.28220.0006
0.00020.32920.0018
0.00000.37620.0020
0.00000.42330.0015
0.00000.47030.0015
0.00010.51740.0024
0.00000.56440.0009
0.00000.61140.0024
0.00000.65850.0024
0.00000.70550.0016
0.00000.75250.0016
0.00000.79960.0032
0.00000.84660.0019
0.00010.89360.0008
0.00000.94070.0025
0.00000.98770.0016
0.00001.03480.0006
0.00001.08180.0014
0.00001.12880.0028
0.00001.17590.0035
0.00001.22290.0029
0.00001.26990.0026
0.00001.31700.0027
0.00001.36400.0014
0.00001.41100.0028
0.00001.45810.0034
0.00001.50510.0025
0.00001.55220.0019
0.00001.59920.0014
0.00001.64620.0037
0.00001.69330.0016
102

102

0.00001.74030.0019
0.00001.78730.0035
0.00001.83440.0019
0.00001.88140.0016
NormalizedOutputintheformat:
Amplitude(%)Freq.(Hz)Phase(deg)
1.0e+003*

0.10000.04700.0048
0.00020.09400.0788
0.00010.14110.0445
0.00030.18810.0502
0.00200.23510.1584
0.00010.28220.0322
0.00110.32920.0762
0.00020.37620.0675
0.00020.42330.0962
0.00000.47030.0945
0.00050.51740.1402
0.00010.56440.0496
0.00020.61140.0434
0.00010.65850.0439
0.00020.70550.0888
0.00020.75250.0876
0.00020.79960.1820
0.00000.84660.1086
0.00030.89360.0461
0.00010.94070.1431
0.00010.98770.0912
0.00001.03480.1450
0.00011.08180.0981
0.00011.12880.0216
0.00011.17590.0184
0.00011.22290.0143
0.00001.26990.1489
0.00001.31700.0254
0.00011.36400.0972
0.00001.41100.0197
0.00011.45810.0158
0.00001.50510.0371
0.00011.55220.0697
0.00001.59920.0818
0.00001.64620.0313
0.00001.69330.0914
0.00001.74030.0695
0.00001.78730.0215
0.00001.83440.0736
0.00001.88140.0877

DfSW:Deviationfromsinewave(in%)
DfSW=
0.5016

THD:TotalHarmonicDistortion(in%)
THD=
2.4740
103

103


Figure 8-9: Spectrum for data sample no. 5

Figure 8-10: Estimation error for data sample no. 5

ii. For data sample no.6 (data7.m)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
DateofExp:30Jan2009
Source:DGSetphaseB
WorkEnvironmentoftheLaboratory:NormalWorkingDay

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
ComputedOutputintheformat:
Amplitude(V)Freq.(Hz)Phase(rad)
1.0e+003*

0.01640.03880.0036
0.00480.07770.0041
0.00190.11660.0015
0.00120.15540.0009
0.00070.19430.0021
0.00030.23310.0041
0.00000.27200.0043
0.00070.31080.0031
0.00020.34970.0008
0.00000.38860.0007
0.00010.42740.0037
0.00010.46630.0034
0.00000.50510.0003
0.00000.54400.0002
0.00010.58290.0020
0.00010.62170.0024
0.00000.66060.0010
0.00000.69940.0041
0.00000.73830.0031
0.00000.77720.0007
0.00000.81600.0001
0.00010.85490.0025
0.00010.89370.0003
0.00000.93260.0004
0.00000.97150.0012
0.00001.01030.0018
0.00001.04920.0030
0.00001.08800.0009
0.00001.12690.0006
104

104

0.00001.16570.0001
0.00001.20460.0029
0.00001.24350.0029
0.00001.28230.0023
0.00001.32120.0016
0.00001.36000.0010
0.00001.39890.0010
0.00001.43780.0006
0.00001.47660.0035
0.00001.51550.0001
0.00001.55430.0014

NormalizedOutputintheformat:
Amplitude(%)Freq.(Hz)Phase(deg)
1.0e+003*

0.09500.03880.0268
0.02770.07770.2330
0.01130.11660.0942
0.00710.15540.0513
0.00400.19430.0624
0.00170.23310.2352
0.00020.27200.2478
0.00400.31080.0030
0.00090.34970.1349
0.00020.38860.0412
0.00040.42740.2104
0.00060.46630.1922
0.00010.50510.1605
0.00030.54400.1701
0.00070.58290.0629
0.00060.62170.0423
0.00030.66060.0597
0.00000.69940.2351
0.00010.73830.1784
0.00000.77720.1414
0.00020.81600.0056
0.00030.85490.0345
0.00040.89370.0198
0.00010.93260.0203
0.00020.97150.0681
0.00001.01030.0755
0.00001.04920.1745
0.00021.08800.0543
0.00021.12690.0355
0.00011.16570.0038
0.00021.20460.1655
0.00001.24350.0116
0.00011.28230.0460
0.00021.32120.0855
0.00021.36000.0572
0.00001.39890.2395
0.00011.43780.0325
0.00011.47660.1994
0.00011.51550.1737
0.00021.55430.0985

DfSW:Deviationfromsinewave(in%)

105

105

DfSW=
12.4179

THD:TotalHarmonicDistortion(in%)
THD=
32.9514

Figure 8-11: Spectrum for data sample no. 6

Figure 8-12: Estimation error for data sample no. 6




106

8.3 Conclusion
This thesis presents application of Particle Swarm Optimization with Passive Congregation for
estimation of phase and frequency; and application of Least Square Algorithm for estimating the
amplitudes in a power signal. From the experiments we conclude that the proposed algorithm can
accurately estimate the amplitudes and phases of the harmonics contained in a voltage or current
waveform for PQ monitoring. The estimation process is iterative and; in each iteration, the
algorithm first applies PSOPC to estimate the phases or the frequencies, and then calculates the
amplitudes using the well-known LS method.
As far as integral harmonics estimation are concerned, the proposed algorithm achieves
improved performance and better results as compared to conventional methods of GA and DFT
schemes in the presence of noise. Moreover, unlike DFT, the proposed algorithm can be made
capable of estimating inter-harmonics and sub-harmonics. The experimental results demonstrate
that the proposed algorithm can be applied to the estimation of harmonics even in the case of the
deviation of the fundamental frequency.
Remark on the chosen values of PSOPC parameters
The values of PSOPC parameters (, c
1
, c
2
, c
3
) are very critical to the problem at hand, and the
same set of parameters may not give satisfactory results for two different problems. The usual
ranges of the PSOPC parameters are given in chapters 6. We have also given some criteria for
choosing appropriate vales, yet one needs to do a little bit of trial and error before he can get
satisfactory values specific to his problem. For harmonic estimation problem = u.8, c
1
= u.2 ,
c
2
= u.2, onJ c
3
= u.6 seem to work quite well. We have chosen these values in our estimator
and the results are good.



107

8.4 Scope of Future Work
At present the scripts, PSOPC.m and PSOPCh.m, presented in appendix are not able to estimate
inter-harmonics. But the authors (Lu et al.) of this Algorithm claim that the algorithm can be
used to estimate inter-harmonics also. For that we believe that Matlab script will need
modifications. So; one can go for developing a script to estimate inter-harmonics and sub-
harmonics.
One can also compare the results obtained with PSOPC and other those obtained with neural
networks, bacteria foraging techniques etc. also.



123

Bibliography

Akagi, H. (1996). New Trends in Active Filters for Power Conditioning. IEEE Transactions on
Industry Applications , 32 (6).
Aplonio, R., de Vasconcellos, A. B., & de Oliveira, J . C. (2004). Three-phase Electromagnetic
Filter for Zero Sequence Harmonics. 11th International Conference on Harmonics and Quality
of Power (pp. 613-618). IEEE.
Arseneau, R., Heydt, G. T., & Kempler, M. J . (1997). Application of IEEE Standard 519-1992
Harmonic Limits for Revenue Billing Meters. IEEE Transactions on Power Delivery , 12 (1).
Barros, E., & Perez, J . (2008). An extended Kalman filtering approach for detection and analysis
of voltage dips in power systems. Elsevier Journal of Electric Power Systems Research , 618
625.
Basu, M., Das, S. P., & Dubey, G. K. (2007). Comparative evaluation of two models of UPQC
for suitable interface to enhance power quality. Elsevier Journal of Electric Power Systems
Research (77), 821-830.
Bettayeb, M., & Qidwai, U. (2003). A hybrid least squares-GA-based algorithm for harmonic
estimation. IEEE Transactions on Power Delivery , 377382.
Bhat, A. H., & Agarwal, P. (2008). Three-phase, power quality improvement ac/dc converters.
Elsevier Journal of Electric Power Systems Research (78), 276-289.


124

Bhende, C. N., Mishra, S., & Panigrahi, B. K. (2008). Detection and classification of power
quality disturbances using S-transform and modular neural network. Elsevier Journal of Electric
Power Systems Research (78), 122-128.
Biswal, B., Dash, P. K., & Panigrahi, B. K. (2008). Non-stationary power signal processing for
pattern recognition using HS-transform. Elsevier Journal of Applied Soft Computing .
Cerqueira, A. S., Ferreir, D. D., Ribeiro, M. V., & Duque, C. A. (2008). Power quality events
recognition using a SVM-based method. Elsevier Journal of Electric Power Systems Research .
Clayton, R. P. (2006). Introduction to Electromagnetic Compatibility. Wiley Interscience.
Clerc, M., & Kennedy, J . (2002). The particle swarm: explosion, stability, and convergence in a
multi-dimensional complex space. IEEE transactions on Evolutinary Computing , 5873.
Dugan, R. C., McGranaghan, M. F., Santoso, S., & Beaty, H. W. (2004). Electrical Power
Systems Qauality. McGraw Hill.
Eberhart, R. C., & Shi, Y. (2001). Particle swarm optimization: developments, applications and
resources. Proceedings of IEEE International Conference on Evolutionary Computation, (pp.
8186).
Eberhart, R., & Shi, Y. (2000). Comparing inertia weights and constriction factors in particle
swarm optimization. Proceedings of the Congress on Evolutionary Computation (CEC2000),
(pp. 8488).
Esfandiari, A., Parniani, M., & Mokhtari, H. (2004). Power Quality Improvements in Three
Phase Four-Wire System using Shunt Active Filters. CIEP .


125

Ferrero, A. (2008). Measuring electric power quality: Problems and perspectives. Elsevier
Journal of Measurement (41), 121-129.
Gallo, D., Landi, C., & Pasquino, N. (2008). An instrument for objective measurement of light
flicker. Elsevier Journal of Measurement , 334340.
Gerek, O. N., & Ece, D. G. (2008). Compression of power quality event data using 2D
representation. Elsevier Journal of Electric Power Systems Research (78), 1047-1052.
Girgis. (1991). Digital recursive measurement scheme for online tracking of power system
harmonics. IEEE Transactions on Power Delivery , 6 (3).
Girgis. (n.d.). Measurement and characterization of Harmonic and high frequency distortion for
large industrial load. IEEE Transactions .
Girgis, A. A., & Ham, F. (1980). A Qualitative Study of Pitfalls in FFT. IEEE Transactions on
Aerosvace and Electronic Svstems , AES16 (4), 434-439.
Girgis, A. A., & Ham, F. M. (1985). Measurement of Power Frequency Fluctuations Using FFT.
IEEE Transactions on Industrial Electronics , IE32 (8), 199-204.
Girgis, A. A., & Ma, H. (1996). Identification and tracking of harmonic sources in a power
system using kalman filter. IEEE Trans. Power Delivery , 11 (4), 16591665.
Halpin, S. M. (n.d.). Harmonic Modeling and Simulation Requirements for the Revised IEEE
Standard 519-1992.
He, S. (2004). A particle swarm optimizer with passive congregation. Elsevier Journal of
BioSystems , 78, 135147.


126

Heydt, G. T. (1989). A New Method for the Calculation of Subtransmission and Distribution
System Transients Based on the FFT. IEEE Transactions on Power Delivery , 4 (3), 1869-1875.
Kaewarsa. (2008). Recognition of power quality events by using multiwavelet-based neural
networks . Elsevier Journal of Electrical Power and Energy Systems , 254260.
Kaewarsa, S., Attakitmongcol, K., & Kulworawanichpong, T. (2008). Recognition of power
quality events by using multiwavelet-based neural networks. Elsevier Journal of Electrical
Power and Energy Systems (30), 254-260.
Kamal, A.-H., & Mandache, L. (2005). New High Precision Harmonic Analysis Method for
Power Quality Assessment. IEEE , 1958-1961.
Kennedy, J . (2000). Stereotyping: improving particle swarm performance with cluster analysis.
Proceedings of the IEEE International Conference on Evolutionary Computation, (pp. 1507
1512).
Kennedy, J ., & Eberhart, R. C. (2001). Swarm Intelligence. San Mateo, CA: Morgan Kaufmann.
Kennedy, J ., & Eberhart, R. (1995). Particle swarm optimization. IEEE International Conference
on Neural Networks .
Kennedy, J ., & Mendes, R. (2002). Population structure and particle swarm performance.
Proceedings of the 2002 Congress on Evolutionary Computation CEC2002 (pp. 16711676).
IEEE Press.
Kong, W., Donga, X., & Chen, Z. (2008). Voltage sag source location based on instantaneous
energy detection. Elsevier Journal of Electric Power Systems Research .


127

Liu, Z. Z., Tao, J ., & Wang, X. H. (2003). The study and design of power quality control in HT-
7U power supply system. Elsevier Journal of Fusion Engineering and Design , 1109-1113.
Lbjerg, M., Rasmussen, T. K., & Krink, K. (2001). Hybrid particle swarm optimizer with
breeding and subpopulations. Proceedings of the Third Genetic and Evolutionary Computation
Conference (GECCO-2001), (pp. 469476).
Lu, Z., J i, Y., & Tang, H. (2008). Optimal Harmonic Estimation Using A Particle Swarm
Optimizer. IEEE Transactions on Power Delivery , 23 (2).
Mishra, S. (2005). A Hybrid Least Square-Fuzzy Bacterial Foraging Strategy for Harmonic
Estimation. IEEE Transactions on Evolutionary Computation .
Parrish, J . K., & Hamner, W. M. (1997). Animal Groups in Three Dimensions. Cambridge:
Cambridge University Press.
Pedra, J ., Sainz, L., Corcoles, F., Bergas, J ., & de Blas, A. (2008). Effects of balanced and
unbalanced voltage sags on DC adjustable-speed drives. Elsevier Journal of Electric Power
Systems Research (78), 957-966.
Ramos, P. M., & Serra, A. C. (n.d.). Comparison of Frequency Estimation Algorithms for Power
Quality Assessment. Elsevier Journal of Measurement .
Rash, A. (n.d.). Power Quality and harmonics in the supply network : a look at common
practices and standards. IEEE , 1219-1223.
Rivera, I. J ., Ramirez, A. B., & Rodriguez, D. (2005). A Time-Frequency Signal Analysis
System for Power Quality Assessment. IEEE , 1277-1280.


128

Sankaran, C. (2001). Power Quality. New York: CRC Press.
Shi, Y., & Eberhart, R. C. (1997). A modified particle swarm optimiser. Proceedings of the IEEE
International Conference on Evolutionary Computation, (pp. 303308).
Shi, Y., & Eberhart, R. C. (1998). Parameter selection in particle swarm optimization. In
Evolutionary Programming VII (1998). Lecture Notes in Computer Science (pp. 591600).
Springer.
Shireen, W., & Tao, L. (2008). A DSP-based active power filter for low voltage distribution
systems. Elsevier Journal of Electric Power Systems Research .
Thorp, J . S., Phadke, A. G., & Karimi, K. I. (1985). Real-Time Voltage Phasor Measurements
for Static-State Estimation. IEEE Transactions on Power Auuaratus and Svstems , 3099-3106.
Uyara, M., Yildirima, S., & Gencoglu, M. T. (2008). An effective wavelet-based feature
extraction method for classification of power quality disturbance signals. Elsevier Journal of
Electric Power Systems Research .
Wong, K. R., & Pham, V. L. (1999). Wavelet-transform-based algorithm for harmonic analysis
of power system waveforms. IEE Proceedings on Generation, Tronsmission, and Disiribution ,
146 (3).
Yilmaz, A. S., Alkan, A., & Asyali, M. H. (2008). Applications of parametric spectral estimation
methods on detection of power system harmonics. Elsevier Journal of Electric Power Systems
Research , 683693.
Zadeh, H. K., & Li, Z. (2008). A novel power swing blocking scheme using adaptive neuro-
fuzzy inference system. Elsevier Journal of Electric Power Systems Research (78), 1138-1146.


129

Zobaa, A. F. (2006). Voltage Harmonic Reduction for Randomly Time-Varying Source
Characteristics and Voltage Harmonics. IEEE Transactions on Power Delivery, VOL. 21, NO. 2 ,
816-822.