DOKUZ EYLÜL UNIVERSITY

GRADUATE SCHOOL OF NATURAL AND APPLIED

SCIENCES

ANALYSIS AND MEASUREMENT OF

HARMONIC DISTURBANCES IN POWER
NETWORKS AND THEIR TRANSIENTS

by
Alpaslan KARAAĞAÇ

November, 2006
İZMİR
 ANALYSIS AND MEASUREMENT OF
HARMONIC DISTURBANCES IN POWER
NETWORKS AND THEIR TRAINSIENTS

A Thesis Submitted to the

Graduate School of Natural and Applied Sciences of
Dokuz Eylül University
In Partial Fulfillment of the Requirements for
the Degree of Master of Science in Electrical and Electronics Engineering ,

by
Alpaslan KARAAĞAÇ

November, 2006
İZMİR
 ii

M.Sc THESIS EXAMINATION RESULT FORM

We have read the thesis entitled “ANALYSIS AND MEASUREMENT OF

HARMONIC DISTURBANCES IN POWER NETWORKS AND THEIR
TRANSIENTS” completed by Alpaslan KARAAĞAÇ under supervision of Asist.
Prof. Dr. Hacer ŞEKERCİ ÖZTURA and we certify that in our opinion it is fully
adequate, in scope and in quality, as a thesis for the degree of Master of Science.

Asist. Prof. Dr. Hacer ŞEKERCİ ÖZTURA

(Supervisor)

(Jury Member) (Jury Member)

Prof.Dr. Cahit HELVACI

Director
Graduate School of Natural and Applied Sciences

ii
 iii

ACKNOWLEDGMENTS

I would like to thank my supervisor Asist. Prof. Dr. Hacer ŞEKERCİ ÖZTURA
for her valuable guidance and support during the course of this thesis. This work has
been based on field measurements and studies in networks. For helping me with data
of networks I would like to thank TEDAS İZMİR İL MÜDÜRLÜĞÜ.

Finally, I owe the nearest standing me my deepest gratitude, especially my family,

for understanding and encouragement throughout this work.

Alpaslan KARAAĞAÇ

iii
 iv

ANALYSIS AND MEASUREMENT OF HARMONIC DISTURBANCES IN

POWER NETWORKS AND THEIR TRANSIENTS

ABSTRACT

Identifying problems in an electrical network before any extensive damage occurs is

advisable for any power system. Not only does monitoring of power quality necessary
for the maintaining accurate operation of sensitive equipment, it also ensures that
unnecessary energy losses in a power system are kept at a minimum which lead to more
profits.

This thesis seeks to identify most popular power quality problem, harmonics in the
distribution System located in İzmir. For the problems identified, recommendations are
suggested to improve the power quality. The voltage harmonics caused by converter
circuits of large motors is investigated with measured data from the distribution system
and simulatedusing MATLAB/Simulink to observe the extent of the harmonic
throughout the system.

Results of the thesis include the total harmonic voltage distortion of the distribution
system is being within IEEE standards, and a correct mitigation technique available for
the identified harmonic problem if it is possible.

Keywords : Power Quality, Harmonic Distortion, Power System Monitoring,

rectifiers, Harmonic Pasive Filter.

iv
 v

GÜÇ SİSTEMLERİNDEKİ HARMONİKLERİN ÖLÇÜLMESİ ANALİZİ

VE BUNLARIN ETKİLERİ

ÖZ

Bu projede orta gerilim dağıtım sistemlerindeki güç kalitesini etkileyen

harmonikler incelenmiştir.Bu incelemeyi gerçekleştirmek için İZMİR ilindeki
sisteme bağlı bir fiderden ölçümler alınmıştır. Buradaki problemlere çözüm
getirilmeye çalışılmıştır. Dağıtım sisteminden alınan ölçüm değerleri
MATLAB/Simulink programında analiz edilmiştir. Güçlü motorlara sahip bir sistemdeki
voltaj harmoniklerine karşı bir fpasif filtre tasarlanarak sonuçlar MATLAB/Simulink
programında incelenmiştir.
Sonuç olarak bir dağıtım hattındaki harmoniklerin ölçümünü gerilimde meydana
gelen değişimleri, bunların zararlarını ve IEEE standartlarına uygunluğunu ve harmonik
problemlerin giderilmesini incelemiştir.

Anahtar sözcükler : Güç kalitesi, harmonikler, harmonik ölçümleri, harmonik

ölçümlerin analizi

v
 vi

CONTENTS

Page

THESIS EXAMINATION RESULT FORM ....................................................... ii

ACKNOWLEDGEMENTS................................................................................ iii
ABSTRACT........................................................................................................iv
ÖZ........................................................................................................................v

CHAPTER ONE- INTRODUCTION ..............................................................1

1.1 Background ...........................................................................................1
1.2. Outline of the Thesis..............................................................................2

CHAPTER TWO-TERMS AND EVENTS RELATED TO POWER

QUALITY .........................................................................................................4

2.1 Introduction...........................................................................................4

2.2 What is Power Quality? .........................................................................6

2.2.1 The Network Model...........................................................................7

2.3 Power Quality Problems ........................................................................9

2.3.1 Long Duration Voltage Variations ...................................................10
2.3.2 Short Duration Voltage Variations ...................................................11
2.3.2.1 Voltage Sags ............................................................................11
2.3.2.2 Momentary/Temporary Interruptions .......................................12
2.3.2.3 Voltage Swells .........................................................................12
2.3.2.4 Other Disturbances...................................................................14
2.3.3 Harmonics .......................................................................................16
2.3.3.1 Fourier Analysis.......................................................................16
2.3.3.2 The Consequences of Harmonic Distortion ..............................18

vi
 vii

2.3.3.3 Harmonic Reduction Techniques..............................................20

2.3.3.3.1 Active Filters .........................................................................20
2.3.3.3.2 Three Winding Transformers-Higher Pulse Bridge Circuits....20
2.3.3.3.3 Passive Filters ........................................................................21
2.3.3.3.4 Line Inductors ........................................................................23
2.3.3.3.5 Load Configuration Management – Harmonic Cancellation ...23

2.4 Cause of Power Quality Problems........................................................24

CHAPTER THREE-HARMONIC ANALYSIS of POWER SYSTEMS......27

3.1 Harmonic Analysis Methods................................................................28

3.1.1 Frequency Domain Analysis ............................................................28
3.1.2 Frequency Scan ...............................................................................31
3.1.3 Voltage Scan ...................................................................................31
3.1.4 Phase Angle of Harmonics...............................................................32
3.1.5 Newton-Raphson Method ................................................................33
3.1.6 Time Domain Analysis ....................................................................35

3.2 Harmonic Modeling Of System Components .......................................36

3.2.1 Transmission Lines..........................................................................36
3.2.2 Underground Cables ........................................................................37
3.2.3 Filter Reactors .................................................................................37
3.2.4 Transformers ...................................................................................37
3.2.5 Induction Motors .............................................................................40
3.2.6 Generators .......................................................................................42

3.3 Load Models........................................................................................42

3.4 System Impedance...............................................................................43

3.5 Power Factor and Reactive Power........................................................45

vii
 viii

CHAPTER FOUR-MONITORING AND SIMULATION OF POWER

SYSTEMS HARMONICS ..............................................................................50

4.1 Introduction.........................................................................................50

4.2 System Information Consolidation.......................................................51

4.3 Calculations Of Modeled Components In Matlab.................................54

4.3.1 Calculations Of Cable Impedance ....................................................55

4.4 Measurements and Simulation .............................................................56

4.4.1 Measurement ...................................................................................56
4.4.2 Simulation....................................................................................59

4.5 Harmonic Study of a Distorted Distribution System.............................63

4.5.1 Harmonic Calculations in the Frequency Domain…………………….66
4.5.2 MATLAB Time-Domain Modeling………………………………….71

CHAPTER FİVE-CONCLUSİONS ...............................................................80

REFERENCES................................................................................................81

APPENDIENCES ...........................................................................................84

viii
 1

CHAPTER ONE
INTRODUCTION

1.1 Background

In the past, equipment used to control industrial process was mechanical in nature,
being rather tolerant of voltage disturbances, such as voltage sags, spikes, harmonics,
etc. In order to improve the efficiency and to minimize costs, modern industrial
equipment typically uses a large amount of electronic components, such as
programmable logic controllers (PLC), adjustable speed drives (ASD), power
supplies in computers, and optical devices. Nevertheless, such pieces of equipment
are more susceptible to malfunction in the case of a power system disturbance than
traditional techniques based on electromechanical parts (Wagner, Andreshak &
Staniak, 1999). Minor power disruptions, which once would have been noticed only
as a momentary flickering of the lights, may now completely interrupt whole
automated factories because of sensitive electronic controllers or make all the
computer screens at an office go blank at once. In order to restart the whole
production, computers, etc, a considerable time might be necessary (in the range of
some hours), implying on significant financial losses to an industry (Warren, Short &
Burke, 1999).

It is thus natural that electric utilities and end-users of electrical power are
becoming increasingly concerned about the quality of electric power in distribution
systems. The term “power quality” has become one of the most common expressions
in the power industry during the current decade (Dugan, McGranaghan & Beaty,
1996, IEEE, 1995). The term power quality was first introduced in late 1960’s
(Kajihara, 1968) , (Plette, 1969) almost a century after the first power system was put
in service. It became in common use twenty years later. Before that time, it did not
attract too much attention mainly because of two reasons. One is that in the past there
were much fewer power system components that could generate power system
disturbances. The other is that in the past there were much fewer power system

1 1
 2

components that were sensitive to power system disturbances. Obviously these

two factors are mutually linked (Bollen, 1999).

The term includes a countless number of phenomena observed in power systems.

Although such disturbances have always occurred on the power systems, a great
attention has been dedicated to minimize their effects to the end-users, notably large
industrial plants (Wang, 2001).

There are two ways to deal with power system disturbances, from the viewpoints
of utilities and from the viewpoint of customers. What the utilities can do is to locate
the source of the disturbances and take necessary measures to limit the severity and
occurrence frequency of such disturbances, so as to ensure the stable and safe
operation of power systems. For the customers, it is necessary to study the impact of
power system disturbances on the equipment. Applying a particular device or control
mode may mitigate such disturbance effect. Generally speaking, system stability and
component performance are the main issues to be studied whenever there are power
system disturbances.

One difficulty in calculating harmonic voltages and currents throughout a

transmission and distribution system is the need for an adequate equivalent to
represent the distribution system and consumers loads fed radially from each busbar.

It has become evident that the use of equivalents without a comprehensive check
on the effect of all impedances actually present can lead to inaccurate estimation of
harmonic voltages and currents in the transmission system. On the other hand, it is
not practicable to obtain and represent all the system details.

1.2. Outline of the Thesis

Chapter 2 exemines the related theories behind power quality problems, that are
normally found in our electrical power supply. The description of several types of
power quality disturbances, which affect the reliability and capability of power
 3

distribution equipment will be discussed together with their common causes and
effects they have on the power system.
The following section, Chapter 3 conducts a study of the relevant literature
review. The review covers the characteristics of the MATLAB for simulating power
quality problems and incorporates this program into a simulation study of power
quality phenomena at a power distribution network in İzmir. Chapter 3 also gives a
brief introduction on MATLAB and evaluates the usefulness and effectiveness of the
program. The possible advantages and disadvantages of using this program in the
electrical transient studies will also be mentioned.

Chapter 4 describes the procedures and constructions in carrying out the system
and equipment models required to simulate and to better understand the harmonic
problems. Steps in the process; from gathering the system information to building the
components used in the MATLAB models and presentation of MATLAB simulation
components set-up parameters will be shown as well in this section. The findings of
the simulated results together with a discussion of their significance. These results
will be further analyzed and comparing them to those obtained by site monitoring.
Observations will be considered during the course of this study. Chapter 4 also
involves the power quality monitoring program conducted at a power distribution
network in İzmir. The actual power quality data obtained will be analyzed, discussed
and compared with the simulated results. Suggestion of possible solutions to these
power quality problems will also be mentioned.
 4

CHAPTER TWO
TERMS AND EVENTS RELATED TO POWER QUALITY

2.1 Introduction

The term power quality has become one of the most common expressions in the
power industry during the last ten years. The term includes a countless number of
phenomena observed in electric power systems. Although such disturbances have
always occurred in the systems, greater attention has recently been dedicated to
minimizing their effects to the end users.

In the past, it can be said that the concepts of power quality and reliability were
very similar because the loads were mostly linear and the amount of power
electronics components was negligible. The loads were typically lighting, heating,
and motors, which in general, are not very sensitive to momentary voltage variations.
Moreover, the loads were more or less isolated from each other and process
automation was almost non-existent. In resume, the loads did not properly work only
in the case of an interruption of the supplied voltage. Some aspects have contributed
to the increased interest on the subject of power quality and the clear distinction
existing nowadays between power quality and reliability:
Sensitivity of the equipment industrial processes have increasingly adopted digital
controllers at their installation in order to improve the efficiency and minimize costs.
Power electronics devices, for instance, adjustable speed drives, have also been used
in large scale. Nevertheless, such apparatuses are more susceptible to malfunction in
the case of a power system disturbance than traditional equipment based on
electromechanical devices. Controls might be affected by minor voltage
disturbances, resulting in nuisance tripping or missoperation of an important process.
Besides, these sensitive loads are interconnected in large networks and automated
processes in series. This implies that the whole system is as sensitive as the most
sensitive device.

4
 5

Market competition with the deregulation of the electrical energy

transmission/distribution market, the utilities are aware of the importance of
delivering to their customers a voltage with “good quality” in order to keep them
satisfied. A large industrial customer unsatisfied with the quality level of the voltage
delivered to its plant is a big concern to the utility that supplies this customer.
Increased knowledge due to the increased number of disturbances that affect their
production, (large) industrial customers must be more familiar with power quality
related aspects such as voltage sags/interruptions and harmonic distortion. The
interaction between utility customer is imperious and both parts must exchange their
knowledge and information about problems and possible solutions to be adopted for
a particular plant. In this context, monitoring equipment located at different points of
the system is of extreme importance in the identification of typical disturbances,
interaction with protection schemes and subsequent response of the process.

In the context of power quality, the term customer has been employed instead of
consumer as electricity is viewed as a product. Large customers are identified as
being industrial installations or commercial business complex, where the load
consists of motors, lighting, power supplies, etc. For most of the residential
customers, power quality is not so critical. Short interruption of power when they are
not at home or at night is not noticed, except for resulting reset of digital clocks with
subsequent blinking in home appliances, e.g. video cassette recorders and microwave
ovens. However, a noticeable increase of the number of home computers has
occurred. These computers can also be affected by some disturbances, increasing
possible sources of dissatisfaction of residential customers regarding the quality of
the utility service. Considering industrial customers, the level of dissatisfaction is
much more critical because some of them can strongly be affected by power system
disturbances. These disturbances may lead to the shut down of production and the
complete restart may require a considerable time interval (in the range of hours),
implying on significant economical losses. The disturbance may also result in
destruction of equipment or material being in production
 6

2.2 What is Power Quality?

A well-established definition of power quality does not exist because it depends

on one’s reference frame. For instance, whilst one customer considers a certain
voltage waveform as having a “sufficient quality” in order to maintain the production
working properly, another customer can realize that the same voltage has a “poor
quality”.

One aspect of common agreement is to consider the power quality as a customer

driven issue, i.e. the customers point of view is determinant for indicating the quality
of the power (in fact, the quality of the voltage). Based on this assumption, a power
quality problem can be defined as (Dugan, McGranaghan & Beaty, 1996): “Any
power problem manifested in voltage/current or leading to frequency deviations
that results in failure or missoperation of customer equipment”. This definition
means that the decisive measurement of power quality is taken from the performance
and productivity of end-user equipment (customer). If the electric power is
inadequate for those needs, the quality is said to be “lacking”. The IEEE Standard
1100 (“IEEE Recommended Practice for Powering and Grounding Sensitive
Electronic Equipment”, also known as “Emerald Book”) (IEEE, 1992) , published in
1992, describes power quality as “the concept of powering and grounding sensitive
equipment in a matter that is suitable to the operation of that equipment”.

It is clear that both definitions are vague; the meanings of the terms
“missoperation” and “powering” are not totally clear. Therefore, any deviation from
perfectly sinusoidal voltages or currents at fundamental frequency with rated
magnitude values is potential candidate to a power quality problem.

Another problematic issue is defining the responsible for a poor power quality at a
certain point of the power system. Apart from natural phenomena that cannot be
avoided, e.g. lightning, a usual procedure is the customers and utilities blaming each
other. Some cases can be identified as typical:
 7

• Remote faults in the system can make the voltage drop at the point where a
critical customer is connected. Although the utility might not detect any
abnormality on the feeder to this customer due to the suitable action of the
protection system, the voltage drop might be sufficient to cause an adjustable
speed drive (ASD) of a motor to trip off.
• Despite being originally supplied by a voltage with “good quality”, wrong
maneuvers, equipment malfunction or high non-linear loads at the industrial
plant can also be the source of power quality problems to other customers
supplied by an electrically close feeder.
• The owner of the equipment at an industrial plant, which is a utility customer,
usually buys equipment at lowest cost. Suppliers of equipment do not feel
encouraged to add extra features to the equipment in order to bear common
disturbances, as these features would increase the equipment cost and final
price. Moreover, many manufacturers are unable to identify the power system
disturbances that can affect their equipment.

From the above discussion, the dissemination of information on power quality by

the utilities and subsequent education of customers and manufacturers are clearly
necessary.

2.2.1 The Network Model

A basic network model commonly used when studying power quality phenomena
is shown in Figure 2.1. This model sometimes includes more elements but the main
structure is always the same. This means the model always contains a voltage
generator (in most cases an independent voltage source), a transmission line (cable,
wire etc.) and a load. The power quality phenomena are normally generated by the
load and propagate in the network influencing the supply system as well as other
loads. In general, the level of power quality is determined by the loads connected to
the network as well as by the design of the network.
 8

Generally speaking, a strong network (low source impedance) will reduce power
quality phenomena while a weak network (high source impedance) is more sensitive
to bad power quality.

Figure 2.1 Basic network model.

Consider an independent voltage generator in Figure 2.1 maintaining the voltage

UG, regardless of the current taken by the load. Also consider the source impedance
ZT to be linear and passive. When ZL is connected, a load current I will flow in the
circuit on the other hand if ZL is a nonlinear load the current I will be distorted and
the consequence will be a nonlinear voltage drop UZT across ZT. Depending on the
characteristics of the load ZL, the voltage UL will also be (more or less) distorted. If
the voltage distortion is of a certain level, it will affect the functioning of the loads
connected to the network. An alternative way to describe the influence between a
load and the rest of the network is discussed in (Dugan, McGranaghan & Beaty,
1996) and shown in Figure 2.1 As shown in Figure 2.2 (IEEE, 1992) loads are
affecting the level of power quality but they are also affected by power quality.
 9

Figure 2.2 The interaction between the power network and a load .

2.3 Power Quality Problems

Power quality problems are associated to an extensive number of electromagnetic

phenomena in power systems with a broad range of time. For instance, it includes
impulsive transients (in the range of nanoseconds) as well as frequency deviations (in
the range of some seconds). The purpose of this section is to shortly introduce the
main disturbances that will be focused along the thesis, as well as their time range.

A comprehensive description of the categories and characteristics of power

systems electromagnetic phenomena related to variations in the voltage magnitude is
available in (IEEE, 1995).

For the sake of convenience, according to the cases studied in this thesis, it is
feasible to classify them based on the duration. The classification takes the voltage
 10

into account, as the quality of the voltage is the addressed issue in most of the cases.
However, it is well known that there is always a close relationship between voltages
and currents in a power system. Specifications regarding current are applied to
dimensioning an equipment or in the case of harmonics.

2.3.1 Long Duration Voltage Variations

Deviations in the operating rms values during longer time than one minute are
usually considered long-duration variations. According to the amplitude variation,
they can be related to permanent faults, load variations, and switching operations in
the system. As an example, switching a capacitor bank or a large load can cause
noticeable changes in the voltage. If the countermeasure, in this case the voltage
regulation, acts very slowly, the voltage change can be characterized as a long-
duration variation. Depending on the magnitude of the voltage change, long-duration
voltage variations can be classified as:

• Undervoltage – decrease in the rms voltage to less than 90% of the nominal
voltage.
• Overvoltage – increase in the rms voltage to more than 110% of the nominal
voltage.
• Sustained interruption – supply voltage equal to zero for more than one
minute. These interruptions are usually permanent and require human
intervention to repair the system.

Although utilities use the term “outage” as a sustained interruption for reliability
reporting purposes, the term “outage” should be avoided in the power quality
context. The reason is that end-users associate the term “outage” to any interruption
of power that shuts down a process, even when the power supplied by the utility is
restored in a few cycles. Meanwhile, in the reliability context, the term “outage”
refers to the state of a component in a system that has failed to function as expected.
 11

The sustained interruptions are studied in the area of reliability, where the
duration and number of these interrupts are computed by different indexes. With the
help of statistics theory, these interruptions can be even predicted.

2.3.2 Short Duration Voltage Variations

This type of voltage variation is mainly caused by either fault conditions – and
associated fault currents or energization of large loads that require high starting
currents. Depending on the electrical distance - related to impedance, type of
grounding, and connection of transformers - between the fault/load location and the
analyzed node, the disturbance can cause a temporary loss of voltage (denoted
interruption) or temporary voltage reduction (denoted sag or dip) orvoltage rises
(denoted swells) at different nodes of the system. In any case, the impact on the
voltage during the disturbance is of short-duration, until protective devices start
operating.

2.3.2.1 Voltage Sags

A sag is sometimes defined as a decrease between 0.1 and 0.9 p.u. in rms voltage
at the network fundamental frequency with duration from 0.5 cycles to one minute.
According to this definition, voltage drops lasting less than half cycle cannot
effectively be characterized by a change in the rms value. In such a case, these events
are considered transients. The term sag has been used (especially in the U.S. power
quality community) to describe a short-duration (0.5 cycles – 1 min.) decrease in the
voltage, while the term dip is recommended by IEC to describe this phenomenon.

Voltage sags are usually associated with system fault currents but can also be
caused by energization of heavy loads or starting of large motors. The duration of the
sag represents the greatest difference between sags caused by a fault from those
caused by a motor start-up. Typical fault clearing times vary from 3 to 30 cycles,
depending on fault current magnitude and the type of over current protection.
Meanwhile, an induction motor can draw 6 to 10 times its full load current during
 12

start-up, which can take some seconds. If the current magnitude is relatively large
compared to the available fault current at that point (i.e. the short-circuit
capacity/power), the resulting voltage sag can be significant both in amplitude and
duration. From the analysis above, it is possible to realize the importance of limiting
over currents not only for avoiding stress on the system components, as also for
avoiding voltage sags at different points of the system.

2.3.2.2 Momentary/Temporary Interruptions

Interruptions are mainly caused by faults and equipment failures. In the first case,
its duration is determined by the operating time of the protection system. Utilities
usually adopt the instantaneous reclosing technique, i.e. a utility breaker, opened
when a fault is detected, is rapidly reclosed after the fault is cleared. If the fault is not
permanent, the interruption interval is limited and certainly less than one minute.

It is interesting to observe that some interruptions can be preceded by a voltage sag

since the protection system does not react immediately when a fault occurs. If an
upstream breaker is opened, the voltage sag may become a voltage interruption.

2.3.2.3 Voltage Swells

A swell is defined as an increase to between 1.1 and 1.8 p.u. in rms voltage at the
network fundamental frequency with duration from 0.5 cycles to one minute. The
term momentary overvoltage is also used as a synonym for swell. Switching off a
large inductive load or energizing a large capacitor bank are typical system
maneuvers that cause swells. Although not as common as voltage sags, swells are
also usually associated to system faults. The severity of a voltage swell during a fault
condition is a function of the fault location, system impedance, and grounding.
During a single phase-to-ground fault on an impedance grounded system, i.e. with
some zero sequence impedance, the non-faulted phase-to-ground voltages can
increase up to 3 times the per-unit value (in the case of a non-grounded or high
impedance grounded system). The difference in the zero- and positive-sequence
 13

impedance causes a change in the non faulted phases, not only in magnitude but also
in phase.

As an example of how voltage swells are related to faults, Figure 2.3 shows a
small part of the distribution system. Both 33 and 11 kV networks are grounded
through a high resistance. A single phase fault direct to ground is applied at the 11
kV bus after 50 ms. It is observed that the non-faulted phase-to-ground voltages are
increased by the 3 factor and phase-shifted by 30 degrees. Since the 33/11 kV
transformer is Y- connected, the voltage on the 33 kV side will scarcely be affected
during the fault. The currents along the feeder are not disturbed and consequently an
over current relay along this feeder will not identify any fault.

Figure 2.3 Single-phase fault at 11 kV bus occurring around 50 msn a high resistive grounded
system.

An interesting classification of the phenomena related to power quality is also

presented in (Bollen, 1999), dividing them according to the magnitude of the voltage
and their duration in a more straightforward manner:

• Voltage interruption – voltage magnitude equal to zero.

• Undervoltage – voltage magnitude below its nominal value.
 14

• Overvoltage – voltage magnitude above its nominal value.

Regarding the duration of the event, it can be referred not to a determined or fixed
time interval but to the type of restoration for returning to the predisturbance
situation (normal system operation):

• Very short – corresponding to transient and self-restoring events.

• Short – corresponding to an automatic recovery to the pre-event situation.
• Long - corresponding to a manual recovery to the pre-event situation.
• Very long – requiring repair or replacement of faulted components.

It can be observed that this is a qualitative approach and more practical since it
also includes the actions taken by the protection schemes. Nevertheless, it is stil
preferable to adopt the previously mentioned terms since they are internationally
used and found in the literature.

2.3.2.4 Other Disturbances

In addition to the voltage variations previously described, there are obviously

other disturbances that may affect the performance of power systems and are also
related to power quality. Nevertheless, their impact on distribution systems is located
in an electrically smaller area than the impact of voltage interruptions and sags. In
this category, unbalanced and harmonic voltages/currents, as well as voltage
fluctuation (known widely as voltage flicker), can be included.

Unbalance is a topic that will be exploited on the thesis and it is usually defined
using symmetrical components. The ratio of either the negative- or zero-sequence
component to the positive-sequence component is normally used to specify percent
unbalance. Small unbalances (less than 2%) are primarily caused by single-phase
loads operating on a three-phase circuit. Blown fuses in one-phase of a three-phase
capacitor bank can also result in voltage unbalance.
 15

A considerable amount of loads on distribution systems in Turkey are single

phase, connected to neutral. As the system neutral is usually multiply grounded, the
currents are very often unbalanced. In such a case, a zero-sequence current will flow
through the neutral. Therefore, ground fault relays cannot be sensitive because the
ground current can be substantial even at normal operation. Multiple grounded
systems have the advantage that disturbance propagation is attenuated by the
grounds.

On the other hand, the great majority of European countries employ impedance
grounded neutral for their distribution systems on medium-voltage level. In the event
of unbalanced currents, the neutral voltage is increased, which is a problem in terms
of safety. These systems are also more susceptible to overvoltages originated by
lightning and require higher insulation levels for end-user equipment. Nevertheless,
the ground fault protection can be tuned very sensitive, as the neutral current is very
small at normal operation.

Other topics related to power quality that have received a lot of attention from the
utilities are voltage fluctuations and harmonics:

• Loads exhibiting continuous, rapid variations in the load current magnitude

can cause voltage variations that are often referred to as flicker. The term
“flicker” is derived from the impact of the voltage fluctuation on lamps such
that they are perceived to flicker by the human eye. Technically speaking,
flicker is not a power system disturbance, but only the undesirable result of
the voltage fluctuation in some loads, especially electric arc furnaces (Hager,
1997) and wind generators (Brauner, 1997).
• Other non-linear loads, such as diode/thyristor rectifiers and cycloconverters,
widely used at industrial processes and railway networks, draw nonsinusoidal
currents from utilities that can result in voltage harmonics/interharmonics.
The term “interharmonics” refers usually to components not integer multiple
of the fundamental frequency.
 16

It should also be observed that harmonics and flicker problems are caused by
loads, namely converters (especially rectifiers) and electric arc furnaces. Meanwhile,
disturbances such as voltage sags and interruptions are mainly originated in the
transmission and distribution networks and they affect the loads. Power electronics
based compensators have been proved to be very useful in the mitigation of both
voltage fluctuations and harmonics on the load site, avoiding in this way their
spreading through the power system (Palesjö, 1992), (Aredes & Watanabe 1995).
However, flickers will not be studied in this thesis.

2.3.3 Harmonics

Harmonics can be referred to as a type of pollution that occurs in electrical

systems. The consequence of this type of pollution includes degradation of telephone
communications systems due to induced harmonic noise, incorrect operation of
control and protection equipment and the overloading of power systems and
equipment. The negative effect that harmonics have on a power system is to decrease
the life expectancy of connected equipment and increase losses. Arrillaga et al.
(1985) described a future concern in the area of harmonics and power supply as
being the cumulative effect of multiple small power converters which have not
generally included harmonic controls (Arrillaga, Bradley & Bodger, 1985) importent
that the effect of increased harmonic loading upon a power system be investigated
and addressed.

2.3.3.1 Fourier Analysis

A distorted waveform can be represented by a number of perfect sine waves at

frequency multiples of the fundamental combined together. The fundamental and
harmonic components of a periodic non sinusoidal waveform can be calculated by
means of Fourier analysis (Soliman & Srinath, 1998)

A non-sinusoidal waveform f(t) repeating with an angular frequency can be

expressed as equation (2.1):
 17

∞
F (t ) = F0 + ∑ f n (t )
n =1
(2.1)
1 ∞
= a0 + ∑ [an cos(nwt ) + bn sin (nwt ) + ]
2 n=2

where,

1
F0 = a0 (2.2)
2

is the average value

2π
1
an =
π ∫ f (t )cos(nwt )∂(wt )
0
n=0,1,2,……..∞ (2.3)

2π
1
bn =
π ∫ f (t )sin (nwt )∂(wt )
0
n=0,1,2,……..∞ (2.4)

Distortion of voltage and current is generally quantified as total harmonic

distortion (THD). THD is the ratio of the root mean square of all the harmonics to the
root mean square of the fundamental component can be expressed as equation (2.5)

∑V
n =2
n
2

VTHD =
V1

∑I
2
n
n =2
I THD = (2.5)
I1

and phase sequences of harmonics can be expressed as equation (2.6)

vah (t ) = 2Vh sin(hω0t + θ h )

vbh (t ) = 2Vh sin(hω0t − 2hπ / 3 + θ h )

 18

vch (t ) = 2Vh sin(hω0t + 2hπ / 3 + θ h )

(2.6)
where,

ITHD : Total Harmonic Distortion Current

VTHD: Total Harmonic Distortion Voltage
I1 : Fundamental Current RMS value
In : Harmonics Current RMS value
n : integer (n > 2)
h : Harmonic number

An alternative method for quantifying harmonic distortion is to calculate

harmonic distortion normalized to a constant base. This quantity is termed Distortion
Factor (DF) and the current distortion factor can be expressed as equation (2.7)

∑I
n= 2
n
2

I DB = (2.7)
Base Current

Note that for Odd functions, only the bn coefficients and odd harmonics will exist.
Hence for distorted sine waves, the harmonics are of order 3,5,7,9……etc.

2.3.3.2 The Consequences of Harmonic Distortion

A lower quality of electric power is, as mentioned in the previous section, of

concern to both the consumer, and the supply and distribution companies. From a
consumer point of view, a distorted voltage/current causes additional losses in
rotating machines as well as a reduced lifetime of winding insulation due to
overheating. The supply and distribution companies also have losses associated with
a distorted electrical supply. The transmission losses are increased and cause
harmonic voltage drops across circuit impedances. Furthermore harmonic distortion
 19

can lead to derating of equipment such as transformers, transmission lines and

cables.

The concern and problems caused by harmonic distortion are well founded
because of present and expected increases in non-linear loads connected to the
electrical network. An extensive list of problems experienced by distorted voltages
and currents are outlined below.

• Lowered efficiency in electromagnetic equipment, such as generators,

because of increased heating due to iron and copper losses at harmonic
frequencies.
• Oscillating torque generated in rotating machinery due to the varying
sequence (positive, negative and zero) components of different harmonics.
• Audible noise in motors and transformers.
• Derating of power cables due to increased resistance from the skin effect and
the proximity effect and associated additional i2R losses.
• Increased heat and dielectric stress in capacitors.
• Shifted zero voltage crossings causing malfunction in control systems whose
operation depends on the zero voltage crossing.
• Metering errors due to phase shifting.
• Derating of equipment such as switchgear and fuses due to increased heating
and losses.
• Disturbance of communication systems by power conversion apparatus
producing electric and magnetic fields.
• Lower power factor.

The many problems associated with harmonic distortion can lead to costly
equipment failures. Capital expenditure increases due to derating or over sizing of
equipment. Other financial losses can be associated with the risk of malfunctioning
control systems and metering errors. Hence the potential costs involved with
harmonic distortion can be quite significant. However, a cost-benefit analysis is
always recommended, when dealing with capital expenditures that are not due to
 20

compliance problems, but are solely due to minimising the adverse affects of
distortion. The subsequent section deals with harmonic reduction techniques.

2.3.3.3 Harmonic Reduction Techniques

The reduction of harmonics within a system can be performed in many different

ways and the selection of a method is often determined by the nature of the system
itself. Methods commonly used to reduce harmonics will normally fall into one of
the following categories: active filters, three winding transformers – higher pulse
bridge circuits, passive filters, line inductors, current injection and load configuration
management-harmonic cancellation.

2.3.3.3.1 Active Filters. Active filters operate by injecting a compensating

current into the system. Often this is performed by comparing the current waveform
to that of a perfect sine wave. A current that is equal to the difference between the
actual waveform and the sine wave is injected to counter the distortion in the current
waveform. This method of harmonic reduction is relatively expensive as the
requirement is for each filter to be custom built. There is also the possibility that in
producing a compensating current, the filter is itself producing harmonics.

2.3.3.3.2 Three Winding Transformers-Higher Pulse Bridge Circuits. Three

phase bridge circuits are in widespread use in industrial applications. These bridge
circuits take the form of six -, twelve -, and higher pulse circuits. Consider a three
phase thyristor fired bridge rectifier. If the thyristors are fired at the earliest possible
moment (α = 0°) then the mean value of the d.c. voltage will be the maximum and
called ideal value. As described by Schlabbach et al. (Schlabbach, Blume &
Stephanblome, 2001), Fourier analysis of the current on the three phase side
produces harmonics in accordance with:

h = np ± 1 (2.7)
 21

where: n = 1, 2, 3, 4……..and p = the number of commutations or pulses during a

period. Accordingly, harmonics of the orders h = 5, 7, 11, 13, 17, 19……are
generated.

The approach of the three winding transformer is to converter a six pulse

converter to a twelve pulse converter. This can only be done on bridge rectifier
circuits. The converter will produce harmonic currents of magnitude inversely
proportional to the harmonic order. The relationship is described by the following:

Ih 1
= (2.8)
I1 h

Hence the magnitude of the 5th harmonic is approximately equivalent to 1/5th of

the fundamental component in an ideal case. The circuit shown in Figure 2.4 shows
how a l-l-D transformer can be used to change a three phase six pulse bridge into a
three phase twelve pulse bridge.

The twelve pulse bridge provides cancellation of harmonics of order 5, 7, 17,

19…etc in the transformer and hence the relationship: 1 ± = np h holds true. The
harmonic currents generated by the twelve pulse bridge circuit are of the orders 11,
13, 23, 25…..etc. The advantage of converting a six pulse bridge to a twelve pulse
bridge is a clear reduction in the THD as the harmonics are converted to higher order
harmonics with a lower magnitude. The downfall of this approach is that it only
works on bridge circuits and is relatively expensive.

2.3.3.3.3 Passive Filters. Passive filters are the most common method of
decreasing harmonic currents. Passive filters are sets of LC filters that are tuned to
specific harmonic frequencies. The filter is based on the series resonance condition
and is used to trap the harmonic current. A filter is tuned to a particular frequency so
as to provide low impedance to the specific harmonic current. Common, basic filter
configurations are shown in Figure 2.5.
 22

Whilst this is a very popular method of controlling harmonic currents, care must
be taken with regard to parallel resonance frequencies. Passive filters also make use
of large numbers of low tolerance components that can deteriorate with age and
diminish the effectiveness of the filter.

Figure 2.4 Twelve Pulse Three Phase Bridge Circuit (l-l-D)

Figure 2.5 Common Passive Filters

 23

2.3.3.3.4 Line Inductors. Very basically, this approach involves placing

inductors in series with the line currents. The disadvantage of this approach is that
line inductors lead to increased losses.

2.3.3.3.5 Load Configuration Management – Harmonic Cancellation. Load

configuration management - harmonic cancellation utilises the characteristics of
individual load harmonic profiles to enable selective combination of loads on
individual phases of a supply. The aim is to reduce current distortion as measured by
THDi.

Hence, harmonic cancellation can be achieved by combining loads that have 180°
shifted harmonic order currents. An example of this approach is presented by
W.Lawrance et al. (Lawrance, Hovingh & Lo, bt.) whereby harmonic current load
profiles are compared for two fluorescent lights and a personal computer. Lawrance
et al. (Lawrance, Hovingh & Lo, bt.) found that the third harmonic of a personal
computer was 180° shifted to the third harmonic of two fluorescent tubes. The
combination of these two loads provides a decrease in the combined third order
current harmonic. Other studies in this area have looked at the combination of single
phase nonlinear and three phase loads (Mansoor & other, 2000) and distributed
single phase computer loads (MJ, 1995).

To visualize the concept of load configuration management - harmonic

cancellation, Figure 2.6 shows four vectors representing the third harmonic of four
different loads. By combining the loads that are 180° phase shifted and having them
on the same supply phase, 100% cancellation of the third harmonics would be
achieved where the magnitude of the 180° phase shifted vectors are identical. In
Figure 2.6, it can be seen that the combination of loads A and C will result in
cancellation and some cancellation will also result from the combination of loads B
and D. The degree of cancellation is dependant on both the magnitude and phase
angle of the harmonics with respect to the fundamental frequency.
 24

This approach to dealing with harmonics is the most cost effective of all methods
considered to reduce harmonic current distortion as there are no hardware
components required and no capital costs. As harmonic current magnitudes are
generally inversely proportional to the harmonic order, low order harmonics are
those for which cancellation should be prioritised (i.e the third and fifth harmonic).

Figure 2.6 Graphical example of third harmonics for four different loads

2.4 Cause of Power Quality Problems

Some typical disturbances to power systems, which may cause power quality
problems, are listed below:

• Lightning and natural phenomena.

• Energization of capacitor banks and transformers.
• Switching or start-up of large loads e.g. motors.
• Operation of non-linear and unbalanced loads
 25

• Failure of equipment, e.g. transformers and cables.

• Wrong maneuvers in distribution substations and plants.

Although all disturbances mentioned above are of concern in the power quality
context, there is no doubt that the most problematic issue is the occurrence of faults,
which is the most exploited topic along this work. System faults can produce voltage
variations at different points of the system with different magnitudes and time scales,
depending on how far the analyzed point is from the fault location, the fault clearing
procedure, and system impedances.

The large majority of faults on a utility system are single phase-to-ground

temporary faults. Nevertheless, most of the three-phase breakers and reclosers on
utility distribution system work on all three phases in order to prevent single phasing
of three-phase loads such as large three-phase motors. It can thus be said that the
single-phase fault will have the same effect downstream to the fault as a three-phase
fault after the actuation of the protection scheme. Operating the circuit breakers and
reclosers only on the faulted phase is a usual practice if the feeder serves only single-
phase loads, which is common in Turkey.

Faults in transmission systems usually do not cause sustained interruptions, as the

transmission systems are mostly meshed. In the case of a fault, the electric power
flow is transferred to another path through the action of the protection system. On the
other hand, faults in distribution systems are prone to cause sustained interruptions
because distribution systems are radially operated or with very slow redundancy
capability (in the range of hours). Nevertheless, faults in both transmission and
distribution systems can produce short duration voltage variations that can be very
critical to certain customers/processes.

In systems where overhead lines are predominant, natural phenomena are

responsible for the majority of faults in transmission and distribution systems,
especially lightning (Dugan, McGranaghan & Beaty, 1996), (Bollen, 1999). In
principle, a lightning stroke is a transient increase in the voltage along the line.
 26

However, an arc is created between the phase hit by the stroke and ground and
consequently the voltage is depressed to zero. The voltage in the fault proximity has
a characteristic squared waveform until the fault is cleared by the protection system
(Brauner, 1997)

Heavy snowfalls, winds, and thunderstorms are also severe disturbances in

medium- and high voltage distribution networks. Heavy loading of snow/ice directly
over lines or on adjacent trees can lead to line or tower failures, giving a considerable
contribution to the number of voltage interruptions and sags.

The main concern regarding underground cables, mainly employed in urban areas,
is the significant degradation of the insulation against voltage surges with aging.
Surge arresters (varistors) are commonly used in order to absorb surges caused by
lightning and capacitor switching.
 27

3 CHAPTER THREE
HARMONIC ANALYSIS OF POWER SYSTEMS

The purpose of harmonic analysis is to ascertain the distribution of harmonic cur-

rents, voltages, and harmonic distortion indices in a power system. This analysis is
then applied to the study of resonant conditions and harmonic filter designs and also
to study other effects of harmonics on the power system, i.e., notching and ringing,
neutral currents, saturation of transformers, and overloading of system components.
On a simplistic basis, harmonic simulation is much like a load flow simulation. The
impedance data from a short-circuit study can be used and modified for the higher
frequency effects. In addition to models of loads, transformers, generators, motors,
etc., the models for harmonic injection sources, arc furnaces, converters, SVCs, etc.
are included. These are not limited to characteristics harmonics and a full spectrum
of load harmonics can be modeled. These harmonic current injections will be at
different locations in a power system. As a first step, a frequency scan is obtained
which plots the variation of impedance modulus and phase angle at a selected bus
with variation of frequency or generates R-X plots of the impedance. This enables
the resonant frequencies to be ascertained. The harmonic current flows in the lines
are calculated, and the network, assumed to be linear at each step of the calculations
with added constraints, is solved to obtain the harmonic voltages. The calculations
may include the following:

. Calculation of harmonic distortion indices.

. Calculation of TIF, KVT, and IT.
. Induced voltages on communication lines.
. Sensitivity analysis, i.e., the effect of variation of a system component.

This is rather a simplistic approach. The rigorous harmonic analysis gets involved
because of interaction between harmonic producing equipment and the power
system, the practical limitations of modeling each component in a large power
system, the extent to which the system should be modeled for accuracy and the types
of component and nonlinear source models. A furnace arc impedance varies

27
 28

erratically and is asymmetrical. Large power high-voltage dc (HVDC) converters and

FACTS devices have large nonlinear loads and superimposition is not valid.
Depending on the nature of the study simplistic methods may give erroneous results.

3.1 Harmonic Analysis Methods

There are a number of methodologies for calculation of harmonics and effects of

nonlinear loads. Direct measurements can be carried out, using suitable
instrumentation. The analytical analysis can be carried out in the frequency and time
domains. Another method is to model the system to use a state-space approach. The
differential equations relating current and voltages to system parameters are found
through basic circuit analysis. We will discuss frequency and time-domain methods.

3.1.1 Frequency Domain Analysis

For calculations in the frequency domain, the harmonic spectrum of the load is
ascertained and the current injection is represented by a Norton’s equivalent circuit.
Harmonic current flow is calculated throughout the system for each of the harmonics.
The system impedance data are modified to account for higher frequency, and are
reduced to their Thevenin equivalent. The principal of superposition is applied. If all
nonlinear loads can be represented by current injections, the following matrix
equations can be expressed as equation (3.1), (3.2)

Vh = ZhIh (3.1)

I h = Y hV h (3.2)

where,

h: Harmonic number
 29

The formation of bus impedance and admittance matrices has already been dis-
cussed. The distribution of harmonic voltages and currents are no different for net-
works containing one or more sources of harmonic currents. During the steady state
the harmonic currents entering the network are considered as being produced by
ideal sources that operate without repercussion. The entire system can then be mod-
eled as an assemblage of passive elements. Corrections will be applied to the impe-
dance elements for dynamic loads, e.g., generators and motors’ frequency dependent
characteristics at each incremental frequency chosen during the study can be mod-
eled. The system harmonic voltages are calculated by direct solution of the linear
matrix equations (3.1) and (3.2).

In a power system, the harmonic injection will occur only on a few buses. These
buses can be ordered last in the Y matrix and a reduced matrix can be formed. For n-
nodes and n-j+ 1 injections, the reduced Y matrix is

Ij Y jj . Y jn V j
. = . . . . (3.3)
I n Ynj . Ynn Vn

where diagonal elements are the self-admittances and the off-diagonal elements
are transfer impedances as in the case of load flow calculations.

The admittance matrix is formed from a primitive admittance matrix by

transformation:

Y abc = A' Y prim A (3.4)

and the symmetrical component transformation is given by

−1
Y 012 = TS Y abc TS (3.5)

The vector of nodal voltages is given by

 30

−1
Vh =Yh Ih = ZhIh (3.6)

For the injection of a unit current at bus k:

Z kk = V k (3.7)

where Zkk is the impedance of the network seen from bus k. The current flowing in
branch jk is given by

I jk = V jk (V j −V k ) (3.8)

Where Yjk is the nodal admittance matrix of the branch connected between j and k.
Variation of the bus admittance matrix, which is produced by a set of modifications
in the change of impedance of a component, can be accommodated by modifications
to the Y-bus matrix as discussed before. For harmonic analysis the admittance matrix
must be built at each frequency of interest, for component level RLC parameters for
circuit models of lines, transformers, cables, and other equipment. Thus, the
harmonic voltages can be calculated. A new estimate of the harmonic injection
currents is then obtained from the computed harmonic voltages, and the process is
iterative until the convergence on each bus is obtained. Under resonant conditions,
large distortions may occur and the validity of assumption of linear system
components is questionable.

Harmonic impedance is important in the response of the system to harmonics.

There can be interaction between harmonic sources throughout the system, and if
these are ignored, the single-source model and the superposition can be used to
calculate the harmonic distortion factors and filter designs. The assumption of
constant system impedance is not valid, as the system impedance always changes,
say due to switching conditions, operation, or future additions. These impedance
changes in the system may have a more profound effect on the ideal current source
modeling than the interaction between harmonic sources. A weak ac/dc
interconnection defined with a short-circuit ratio (short-circuit capacity of the ac
system divided by the dc power injected by the converter into converter bus) of <3
 31

may have voltage and power instabilities, transient and dynamic overvoltages, and
harmonic overvoltages.

3.1.2 Frequency Scan

A frequency scan is merely a repeated application in certain incremental steps of

some initial value of frequency to the final value, these two values spanning the
range of harmonics to be considered. The procedure is equally valid whether there
are single or multiple harmonic sources in the system, so long as the principal of
super-imposition is held valid. Then, for unit current injection the calculated voltages
give the driving point and transfer impedance, both modulus and phase angle. The Y
bus contains only linear elements for each frequency. Varying the frequency gives a
series of impedances which can be plotted to provide an indication of the resonant
conditions. Figure 3.1 shows a frequency scan of impedance modulus versus
frequency. The parallel resonances occur at peaks, which give the maximum
impedances and the series resonances at the lowest points of the impedance plots.
Figure 3.1 shows parallel resonance at two frequencies fp1 and fp2 and series
resonance at fs1 and fs2. Such a frequency scan can be obtained with two single-tuned
shunt filters. Multiphase frequency scans can identify the harmonic resonance caused
by single-phase capacitor banks.

3.1.3 Voltage Scan

A voltage scan may similarly be carried out by applying unit voltage to a node
and calculating the voltages versus frequency in the rest of the system. The resulting
voltages represent the voltage transfer function to all other nodes in the system. This
analysis is commonly called a voltage transfer function study. The peaks in the scan
identify the frequencies at which the voltages will be magnified and the lowest points
indicate frequencies where these will be attenuated.
 32

3.1.4 Phase Angle of Harmonics

For simplicity, all the harmonics may be considered cophasial. This does not
always give the most conservative results, unless the system has one predominant
harmonic, in which case only harmonic magnitude can be represented. The phase
angles of the current sources are functions of the supply voltage phase angle and are
expressed as

θ h = θ h , spectrun + h(θ 1 − θ1,spectrun ) (3.9)

where θ1 is the phase angle obtained from fundamental frequency load flow
solution, and θ1,spectrum is the typical phase angle of harmonic current source
spectrum. The phase angles of a three-phase harmonic source are rarely 1200 apart, as
even a slight unbalance in the fundamental frequency can be reflected in a
considerable unbalance in the harmonic phase angle.

Figure 3.1 Frequency scan showing parallel and series resonance frequencies.
 33

3.1.5 Newton-Raphson Method

The Newton-Raphson method can be applied to harmonic current flow (Taylor,

Hill, Burden & Mattern, 1998). This is based on the balance of active power and
reactive volt-amperes, whether at fundamental frequency or at harmonics. The active
and reactive power balance is forced to zero by the bus voltage iterations.

Consider a system with n+1 buses, bus 1 is a slack bus, buses 2 through m - 1 are
conventional load buses, and buses m to n have nonsinusoidal loads. It is assumed
that the active power and the reactive volt-ampere balance are known at each bus and
that the nonlinearity is known. The power balance equations are constructed so that
∆P and ∆Q at all nonslack buses is zero for all harmonics. The form of ∆P and ∆Q as
a function of bus voltage and phase angle is the same as in conventional load flow,
except that Ybus is modified for harmonics. The current balance for fundamental
frequency is written as

I r ,m (
g r , m Vm , Vm , Κ , α m
5
βm )
I i ,m
=
g i ,m (V m
5
, Vm , Κ , α m βm ) (3.10)
− −
I i ,n (
g i , n Vn , Vn , Κ , α n
5
βn )

where Ir,m and Ii,m are real and imaginary bus injection currents at bus m at the
fundamental, a is the firing angle, and ft is the commutation parameter. This equation
is modified for buses with harmonic injections as

I rk,1 0
I k
i ,1
0
.. ..
I k
i , m −1
0
=− (3.11)
I k
r ,m
(
g r , m Vm , Vm , Κ , α m
5
βm )
I k
i ,m
g i ,m (V m
5
, Vm , Κ , α m βm )
.. ..
I k
i ,n
(
g i , n Vn , Vn , Κ , α n
5
βn )
 34

where I rk,1 is the real and I ik,1 is the imaginary part of the current at the kth harmonic,

g ik and g rk are the imaginary and real parts of the current equation at the kth har-
monic, and Vm with superscript is the voltage at that harmonic.

The third equation is the apparent volt-ampere balance at each bus:

S L2 = ∑ PL2 + ∑ QL2 + ∑ D L2 (3.12)

s s s

where the third term of the equation denotes distortion power at bus L, which is
not considered as an independent variable, as it can be calculated from real and
imaginary components of currents.

The final equations for the harmonic power flow become:

∆W J1 J5 J7 .. 0 ∆V 1
∆I 1 YG1,1 YG1,5 YG 1, 7 .. H 1 ∆V 1
∆I = YG 5,1 YG 5,5 YG 5, 7 .. H 5 .. (3.13)
∆I 7 YG 7 ,1 YG 7 ,5 YG 7, 7 .. H 7 ..
Κ ∆α
where all elements in Eq. (3.13) are sub-vectors and sub-matrices partitioned from
∆M (apparent mismatches), J, and ∆U, i.e., ∆M = J∆U.

∆W= mismatch active and reactive volt-amperes

∆I 1 = mismatch, fundamental current
∆I k = mismatch harmonic currents at kth harmonic
J 1 = conventional power flow Jacobian
J k = Jacobian at harmonic k
(YG )k , j = Y k ,k + G k ,k (k = j ) = G k , j (k≠j)

where Yk,k is an array of partial derivatives of injection currents at the kth

harmonic with respect to the kth harmonic voltage, and Gk,j are the partials of the kth
 35

harmonic load current with respect to the jth harmonic supply voltage; Hk are the
partial derivatives of nonsinusoidal loads for real and imaginary currents with respect
to α and β. Ref. [6] discusses the impedance matrix method of harmonic analysis.

3.1.6 Time Domain Analysis

The simplest harmonic model is a rigid harmonic source and linear system impe-
dance. A rigid harmonic source produces harmonics of a certain order and constant
magnitude and phase, and the linear impedances do not change with frequencies.
Multiple harmonic sources are assumed to act in isolation and the principal of
superimposition applies. These models can be solved by iterative techniques, and the
accuracy obtained will be identical to that of time-domain methods. For arc furnaces
and even electronic converters under resonant conditions an ideal current injection
may cause significant errors. The nonlinear and time-varying elements in the power
system can significantly change the interaction of the harmonics with the power
system. Consider the following:

• Most harmonic devices that produce uncharacteristic harmonics as

terminal conditions are in practice not ideal, e.g., converters operating with
unbalanced voltages.
• There is interaction between ac and dc quantities and there are interactions
between harmonics of different order, given by switching functions
(defined later).
• Gate control of converters can interact with harmonics through synchro-
nizing loops.

Time-domain analyses have been used for transient stability studies, transmission
lines, and switching transients. It is possible to solve a wide range of differential
equations for the power system using computer simulation and to build up a model
for harmonic calculations, which could avoid many approximations inherent in the
frequency domain approach. Harmonic distortions can be directly calculated and
making use of Fast Fourier Transform (FFT) these can be converted into frequency
 36

domain. The graphical results are waveforms of zero crossing, ringing, high dv=dt,
and commutation notches. The transient effects can be calculated, e.g., the part-
winding resonance of a transformer can be simulated. The synchronous machines can
be simulated with accurate models to represent saliency, and the effects of frequency
can be dynamically simulated. EMTP and MATLAB is one very widely used
program for simulation in the time domain.

For analysis in the time domain, a part of the system of interest may be modeled
in detail. This detailed model consists of three-phase models of system components,
transformers, harmonic sources, and transmission lines and it may be coupled with a
network model of lumped RLC branches at interconnection buses to represent the
driving point and transfer impedances of the selected buses. The overall system to be
studied is considerably reduced in size and time-domain simulation is simplified.

3.2 Harmonic Modeling Of System Components

3.2.1 Transmission Lines

The transmission line model to be used is determined by the wavelength of the

highest frequency of interest. Long-line effects should be represented for lines of
length 150/h miles, where h is the harmonic number. The effect of higher frequencies
is to increase the skin effect and proximity effects. A frequency-dependent model of
the resistive component becomes important, though the effect on the reactance is
ignored. The resistance can be multiplied by a factor g(h) (Knoller &Lonnstam,
1969):

R(h ) = Rdc g (h ) (3.14)

g (h ) = 0.035 X 2 + 0.938 yX > 2.4 (3.15)

g (h ) = 0.35 X + 0.3 yX ≤ 2.4 (3.16)

 37

where
fh h
X = 0.3884 (3.17)
f Rdc

where fh = harmonic frequency and f = system frequency.

3.2.2 Underground Cables

Cables have more significant capacitance than overhead lines. An estimate of

length where long-line effects are modeled is 90/h miles.

3.2.3 Filter Reactors

The frequency-dependent Q of the filter reactors is especially important, as it

affects the sharpness of tuning. Resistance at high frequencies can be calculated by
the following expressions:

 0.115h 2 + 1
Rh =  R f for aluminum reactors (3.18)
 1.15 

 0.55h 2 + 1
Rh =  R f for copper reactors (3.19)
 1.055 

3.2.4 Transformers

A linear conventional T-circuit model of the transformer fundamental frequency

values of resistance and reactance can be found by a no-load and short-circuit test on
the transformer. The resistance of the transformer can be modified with increase in
frequency according to Fig. 3.2. While the resistance increases with frequency, the
leakage inductance reduces. The magnetizing branch in the transformer model is
often omitted if the transformer is not considered a source of the harmonics. This
 38

simplified model may not be accurate, as it does not model the nonlinearity in the
transformer on account of:

• Core and copper losses due to eddy current and hysteresis effects. The core
loss is a summation of eddy current and hysteresis loss; both are frequency
dependent:

Pc = K e B 2 f 2 + K h B S f 2 (3.20)

where B is the peak flux density, s is the Steinmetz constant (typically 1.5-2.5,
depending on the core material), f is the frequency, and Ke and Kh are constants.
• Leakage fuxes about the windings, cores, and surrounding medium.

Figure 3.2 Increase in transformer resistance with frequency

 39

• Core magnetization characteristics, which are the primary source of

transformer non-linearity.

A number of approaches can be taken to model the nonlinearities. EMTP

transformer models are Satura and Hysdat. These models consider only core
magnetization characteristics and neglect nonlinearities or frequency dependence of
core losses and winding effects.

Capacitances of the transformer windings are generally omitted for harmonic

analysis of distribution systems; however, for transmission systems capacitances are
included. A simplified model with capacitances is shown in Figure 3.3; Cbh is the
high voltage bushing capacitance, Cb is the secondary bushing capacitance, Chg and
Clg are distributed winding capacitances, and Chl is the distributed winding
capacitance between the windings. Typical capacitance values for core-type
transformers are shown in Table 3.1.

Figure 3.3 Simplified capacitance model of a two-winding transformer

Converter loads may draw dc and low-frequency currents through the trans-
formers, i.e., a cycloconverter load. Geomagnetically induced currents flow on the
 40

earth’s surface due to geomagnetic disturbance are typically at low frequencies

(0.001-0.1 Hz), reaching a peak value of 200 A. These can enter the transformer
windings through grounded wye neutrals and bias the cores to cause half-cycle
saturation.

Table 3.1 Capacitance of Core-Type Transformers in nF

MVA rating of Chg Cgl Clg

transformer (µF) (µF) (µF)

1 1.2–14 1.2–17 3.1–16

2 1.4–16 1–18 3–16
5 1.2–14 1.1–20 5.5–17
10 4–7 4–11 8–18
25 2.8–4.2 2.5–18 5.2–20
50 4–6.8 3.4–11 3–24
75 3.5–7 5.5–13 2.8–30
100 3.3–7 5–13 4–40

3.2.5 Induction Motors

Figure 3.4 shows the equivalent circuits of an induction motor. The shunt ele-
ments gc and bm are relatively large compared to R1, r2, X1, and x2. Generally, the
locked rotor current of the motor is known and is given by Eq. (3.21).

V1
I 1r = (3.21)
(R1 + r2 ) + j ( X 1 + x2 )

At fundamental frequency, neglecting magnetizing and loss components, the

motor reactance is

X f = X 1 + x2 (3.22)
 41

Figure 3.4 Equivalent circuit of an induction motor for balanced positive sequence voltages.

and the resistance is

r2
R f = R1 + (3.23)
s

This resistance is not the same as used in short-circuit calculations. At harmonic

frequencies, the reactance can be directly related to the frequency:

X h = hX f (3.24)

though this relation is only approximately correct. The reactance at higher

frequencies is reduced due to saturation. The stator resistance can be assumed to vary
as the square root of the frequency:

R1h = h (R1 ) (3.25)

The harmonic slip is given by related :

for positive sequence harmonics are given in eq. (3.26)

h −1
sh = (3.26)
h
 42

for negative sequence harmonics are given in eq. (3.27)

h +1
sh = (3.27)
h
The rotor resistance at harmonic frequencies is

(1 ± h )
r2 h = (3.28)
sh

3.2.6 Generators

The average inductance experienced by harmonic currents, which involve both the
direct axis and quadrature axis reactances, is approximated by

L"d + L"q
Average Inductance = (3.29)
2

At harmonic frequencies the fundamental frequency reactance can be directly pro-

portioned. The resistance at harmonic frequencies is given by

[
Rh = Rdc 1 + 0.1(h f f )
1.5
] (3.30)

This expression can also be used for calculation of harmonic resistance of

transformers and cables having copper conductors.

3.3 Load Models

Figure 3.5 (a) shows a parallel RL load model and figure 3.5 (b) CIGRE type-C
model It represents bulk power load as an RL circuit connected to ground. The
resistance and reactance components are calculated from fundamental frequency
voltage, reactive volt-ampere and power factor.
 43

V2 V2
R= L= (3.31)
S cos φ 2πfS sin φ
The reactance is frequency dependent and resistance may be constant or it can
also be frequency dependent. Alternatively, the resistance and reactance may remain
constant at all frequencies.

Figure 3.5(b) shows a CIGRE (Conference Internationale des Grands Reseaux

Electriques a Haute Tension) type-C load model [10], which represents bulk power,
valid between 5th and 30th harmonics. Here, the following relations are applicable:

V2 hRs
Rs = X s = 0.073hRs Xp = (3.32)
P Q
6.7 − 0.74
P

Figure 3.5 Load model (a) Parallel RL load model; (b) CIGRE type-C model.

3.4 System Impedance

The system impedance to harmonics is not a constant number. Figure (3.6) shows
the R-X plot of a system impedance. The fundamental frequency impedance is induc-
tive, its value representing the stiffness of the system. The resonances in the system
make the R–-X plots a spiral shape, and the impedance may become capacitive. Such
spiral shaped impedances have been measured for high-voltage systems, and reso-
nances at many frequencies are common. These frequencies at resonance points and
 44

also at a number of other points on the spiral shaped curves are indicated as shown in
Figure 3.6. At the resonance the impedance reduces to a resistance. The system
impedance can be ascertained by the following means:

• A computer solution can be used to calculate the harmonic impedances.

• In noninvasive measurements the harmonic impedance can be calculated
directly from the ratio of harmonic voltage and current reading.
• In another measurement method, a shunt impedance is switched in the circuit
and the harmonic impedance is calculated by comparing the harmonic
voltages and currents before and after switching.

The spiral shaped impedance plots can be bounded in the Z plane by a circle on
the right side of the Y axis and tangents to it at the origin at an angle of 75°. This
configuration can also be translated in the Y plane that is shown (Figure 3.7) (Wang,
2001).

Figure 3.6 R-X plot of a supply system source impedance.

 45

Figure 3.7 Generalized impedance plot (a) in R-X plane and (b) in Y plane.

3.5 Power Factor and Reactive Power

The power factor of a converter is given by the expression:

q π 
Total PF = sin   (3.33)
π q

where q is the number of converter pulses and nq is the angle in radians. This
ignores commutation overlap, no-phase overlap, and neglects transformer
magnetizing current. For a six-pulse converter the power factor is 3/π = 0.955. A 12-
pulse converter has a theoretical power factor of 0.988. With commutation overlap
and phase retard the power factor is given by (Palesjö, 1992):

Ed' I d 3 1  Ex 
PF = =  cos α −  (3.34)
3 EL I L π 1 − 3 f ( µ ,α )  Edo 

where
Ed’ = Ed + Er + Ef
Ed = average direct voltage under load
Er = resistance drop
Ef = total forward drop per circuit element
Id = dc load current in average amperes
 46

EL = primary line-to-line ac voltage

IL =ac primary line current in rms amperes
α = phase retard angle
µ = angle of overlap or commutation angle
Edo = theoretical dc voltage
Ex = direct voltage drop due to commutation reactance

and

sin µ  2 + cos ( µ + 2α )  − µ 1 + 2 cos α cos ( µ + α ) 

f ( µ ,α ) = 2
(3.35)
 2π cos α − cos ( µ + α ) 

The displacement power factor is

sin 2 µ
cos φ1 = (3.36)
µ 2 sin 2 µ − 2 µ sin µ cos µ

This relationship neglects transformer magnetizing current. The correction for

magnetizing current is approximately given by

cos φ1 = cos  arccos φ1 '+ arctan ( I mag I1 )  (3.37)

where cosφ1 is the is displacement power factor without transformer magnetizing

current. The power factor of converters will vary with the type of converter and dc
filter. In a PWM inverter, driven from a dc link voltage with a reactor and capacitor,
the drive motor power factor is not truly reflected on the ac side, and is compensated
by the filter capacitor and reactor.
Line commutated inverters require reactive power from the supply system. The
closer is the operation to zero voltage dc the more the reactive power required.
Figure 3.8 shows the reactive power requirement of a fully controlled bridge circuit
 47

versus half-controlled bridge circuit. The maximum reactive power input for a half-
controlled circuit is seen to be half of the fully controlled circuit.

In the case of sinusoidal voltage and current the following relationship holds:

S 2 = P2 + Q2 (3.38)

where P is the active power, Q is the reactive volt ampere and S is the volt
ampere. This relationship has been amply explored in load flow section: S = VfIf, Q =
VfIf sin(θf-δf), and PF = P/S

Figure 3.8 Reactive power requirements of fully controlled and half-controlled bridge circuits.
 48

In the case of nonlinear load or when the source has nonsinusoidal waveform, the
reactive power Q can be defined as

h =∞
Q = ∑ Vh I h sin (θ h − δ h ) (3.39)
h =1

and the apparent power can be defined as

S 2 = P2 + Q2 + D2 (3.40)

where D is the distortion power. Consider D2 up to the third harmonic:

D 2 = (V0 2 + V12 + V2 2 + V32 )( I 0 2 + I12 + I 2 2 + I 32 )

2
− (V0 I 0 + V1I1 cos θ1 + V2 I 2 cos θ 2 + V3 I 3 cos θ3 )
2
− (V1I1 sin θ1 + V2 I 2 sin θ 2 + V3 I 3 sin θ 3 )

An expression for distortion power factor can be arrived from current and voltage
harmonic distortion factors. From the definition of these factors, rms harmonic
voltages and currents can be written as

2
Vrms ( h) = V f 1 + (THDV /100 ) (3.41)

2
I rms ( h) = I f 1 + (THDI /100 ) (3.42)

Therefore, the total power factor is

P
PFtot = (3.42)
2 2
V f I f 1 + (THDV 100 ) 1 + (THDI 100 )
 49

Neglecting the power contributed by harmonics and also voltage distortion, as it is

generally small

1
PFtot = cos (θ f − δ f ) (3.43)
2
1 + (THDI 100 )

The total power factor is the product of displacement power factor (which is the
same as the fundamental power factor) and is multiplied by the distortion power
factor as defined above.
 50

4 CHAPTER FOUR
MONITORING AND SIMULATION OF POWER SYSTEMS HARMONICS

4.1 Introduction

The power quality monitoring can have a number of benefits for electrical
systems. One of these benefits is the advantage of quantifying the magnitude of the
problem in a system. This stems to even recognising the extent of the problems. As
mentioned earlier, one of the reasons for the interest in power quality is due to the
growing need for standardisation and performance criteria and the fact that
equipment has become more sensitive to disturbances.

Monitoring of Power quality has become common place in industry. With

standards describing power quality phenomena, measurement techniques have been
developed, and specifically designed to capture the power quality problems. The
MTE PTS 400.3 ( Modular three-phase Portable Test System) is one available power
quality measuring instrument capable of sampling power quality phenomena
accurately and with ease. Guidelines for instrumentation requirements associated
with different types of power quality phenomena are being developed by the IEEE
1159.1 Working Group.

Simulation is an important tool that helps to produce better results faster. It can
used to predict the cause and results of certain actions, understand why observed
events occur, identify problem areas before implementation and evaluate ideas and
identify inefficiencies. It is important to study these issues offline before any changes
in the operations could be made. Simulation methods are usually more cost-effective
and flexible than the physical testing methods

A simulation study of power quality problems was conducted on a distribution

system of the University of EGE campus in response to the power quality problems
reported over the past few years. The site examined was at the medium voltage (MV)
(34.5 kV) feeder of EGE TM, where the load distribution system mainly includes

50
 51

workshop motors, PCs, fluorescent lighting and air-conditioners. The loads often
lead to tripping or missoperation of the equipment during normal operation and have
caused inconvenience to the operating staff.

The focuses on various causes of power quality problem to be simulated are

namely harmonics.

A program known as MATLAB Power System Toolbox will be used to perform

the simulation of harmonics as mentioned above.

4.2 System Information Consolidation

Before any simulation on several processes of various power quality problems can
be performed, there is a need to create the models of distribution system components.
Precise information about the system involved must first be obtained so as to
simulate power quality problems accurately. The procedures on information
consolidation are important and will be illustrated.

Field data provided by the University of EGE are used in this study such as
schematic diagrams on substations, information on transformers, cables and types of
loads.

The following figure is a simplified one line diagram illustrating the information
on the local distribution substation.

The distribution system is supplied by the local utility through a 34.5 kV

substation, which is located on the University of EGE campus.

Harmonic values were measured from Fakulteler feeder of University TM and

figure 4.1. One line diagram of EGE TM substation, figure 4.2
 52

Figure 4.1. One line diagram of Universite TM substation

 53

Figure 4.2 One line diagram of Fakulteler feeder substation

The following table summarizes a detail breakdown of all system information

required to model the University distribution system.
 54

Table 4.1 Power information of the substation system

Cabin Number Power (kVA)

M-86 250
M-86/A 1250
M-87 160
M-88 Bypass
M-88/A 250
M-89 500+400
M-90 1000
M-91 250
M-93 630+800
M-94 500
M-94/A 1600
M-95 500
M-97 500
M-98 1000
M-99 400
M-290 Bypass
M-295 630
M-309 400
Total 8690

4.3 Calculations Of Modeled Components In MATLAB

The very first approach taken towards simulation is the calculations of parameters
for the modeled components. Care must be taken in calculating the parameters in
order to achieve accuracy.

This section will describe the calculations carried out and methods used to build
the models to simulate harmonic distortion.
 55

Table 4.2 Cable information of the substation system

Cabin Number Cable

Feeder-M-290 240m, 4 conductors per phase Area of core =240mm2
M-290--M-99 315m, 4 conductors per phase Area of core =150mm2
M-90--M-99 415m, 4 conductors per phase Area of core =150mm2
M-90--M-98 193m, 4 conductors per phase Area of core =150mm2
M-90--M-89 456m, 4 conductors per phase Area of core =150mm2
M-88--M-89 289m, 4 conductors per phase Area of core =150mm2
M-88--M-87 240m, 4 conductors per phase Area of core =150mm2
M-86/A--M-87 367m, 4 conductors per phase Area of core =150mm2
M-86/A--M-86 430m, 4 conductors per phase Area of core =150mm2
M-86--M-93 272m, 4 conductors per phase Area of core =150mm2
M-309--M-93 425m, 4 conductors per phase Area of core =150mm2
M-94--M-93 524m, 4 conductors per phase Area of core =150mm2
M-94--M-94/A 414m, 4 conductors per phase Area of core =150mm2
M-98--M-94/A 380m, 4 conductors per phase Area of core =150mm2
M-240--M-295 300m, 4 conductors per phase Area of core =150mm2
M-95--M-295 150m, 4 conductors per phase Area of core =150mm2

4.3.1 Calculations Of Cable Impedance

The cable model used in MATLAB for this simulation is coupled PI cables
component since they are fairly short. The inductance and capacitance of the cables
will be neglected and consider only the resistance. The resistance of various lengths
can be calculated and are shown below using the formulae:

ρl
R= (4.1)
A

where

R - resistance
ρ - resistivity of the cable is 1.724 x 10-8 Ωm
l -length of cable
 56

A - area of cable core

4.4 Measurements and Simulation

Harmonics can be simulated MATLAB by either building models of nonlineer

devices or by injecting the appropriated harmonic currents into a bus. Values of
injected current should be determined by measurement. The approach used in this
project is in injecting currents into a bus.

4.4.1 Measurement
The harmonic analyzer has been connected one week along everday beetwen
hours 15.30, to the distribution system. An instant harmonic voltage and current
values are record. The simulations are done according to these measurements. The
connection of harmonic analyzer to the system is shown in Figure 4.3 and MTE PTS
400.3 is the name of the instrument be used that is shown in Figure 4.4

Figure 4.3 MTE PTS 400.3 Modular three phase portable analyzer
 57

Figure 4.4 Connection of harmonic analyzer to substation system

The measured voltage and current harmonics are shown in Table 4.3 and 4.4
respectively. According to these measurements and the calculated values of system
parameters the MATLAB model of the substation system can be construct.

Figure 4.5 Measurement system value of fakülteler feeder in EGE TM

 58

Table 4.3. Measurement system electrical value of Fakülteler feeder

Phase 1 Phase 2 Phase 3 Total

Voltage Y L-N 56,940 V 54,652 V 57,598 V
Voltage V L-L 97,967 V 97,715 V 97,838 V
Current 1,832 A 1,891 A 1,921 A
Active Power 99,844 W 11,084 W 107,477 W 309,127 W

Reactive Power 26,124 var 23,584 var 29,678 var 79,385 var

Apparent Power 104,699 VA 103,604 VA 111,254 VA 319,16 VA

Power Factor 0,958 0,979 0,967 0,969
Reactive Factor 0,286 0,206 0,254 0,249
Phase Shift 16,604 ° 11,864 ° 14,694 °

Frequency 50,000 Hz
Voltage Transformer Current Transformer
Up 34500 V Ip 150 A
Us 100 V Is 5 A
Up/Us 345 Ip/Is 30

Table 4.4 Harmonic values of Fakülteler Feeder

L1
Harmonic Phase Real Harmonic Phase Real
% %
Current angle Value Current angle Value
I1 1,8231 A 16,604 100 54,693 I11 0,0037 A -136,53 0,205 0,111
I2 0,0054 A 27 0,295 0,162 I12 0,0026 A 90 0,141 0,078
I3 0,0035 A -9 0,191 0,105 I13 0,0027 A -115,21 0,148 0,081
I4 0,0028 A -0,09 0,153 0,084 I14 0,0016 A -15,2 0,087 0,048
I5 0,0569 A -0,92 3,123 1,707 I15 0,0023 A 32,6 0,128 0,069
I6 0,0038 A -26,5 0,208 0,114 I16 0,0011 A -16 0,06 0,033
I7 0,0334 A 175 1,831 1,002 I17 0,0039 A -12,9 0,216 0,117
I8 0,0031 A 270 0,17 0,093 I18 0,0007 A 136 0,041 0,021
I9 0,0031 A -9 0,169 0,093 I19 0,0036 A 120 0,197 0,108
I10 0,0021 A 90 0,117 0,063 I20 0,001 A -63 0,057 0,03
 59

2
1,8
1,6
1,4
1,2

current
1 Seri 1
0,8
0,6
0,4
0,2
0
I1 I2 I3 I4 I5 I6 I7 I8 I9 I10 I11 I12 I13 I14 I15 I16 I17 I18 I19 I20
harm onic num ber

Figure 4.6 Current harmonics

4.4.2 Simulation

Figure 4.7 shows the MATLAB model of the substation system. The
measurements can be taken only from the Fakülteler feeder of the substation system.
So it can not be defined to show the distribution harmonic currents through the
power system. The analysis were made using MATLAB FFT toolbox
 60

Figure 4.7 MATLAB model

 61

Figure 4.8 Harmonic currents subsystem of the MATLAB model

Figure 4.8 shows the subsystem harmonic current model. The simulations are
performed in time domain. In this type of simulation rigid harmonic source produces
harmonics of a certain order and constant magnitude and phase, and the linear
impedances do not change with frequencies. Multiple harmonic sources are assumed
to act in isolation and the principal of superimposition applies.
 62

Figure 4.9 Simulation results voltage

Figure 4.10 Simulation results current

 63

In Figure 4.9 and 4.10 the simulation results of the system is shown. Figure 4.9
and 4.10 below gives a summary of the harmonic spectra of the voltage and current
waveform on a phase. These can be compared with the results given by the
MATLAB. When comparing these figures, it is important to keep in mind that the
simulations were made on the basis of educated guesses about the type and sizes of
the loads in the building.

Examining, for example, Figure 4.9, and comparing it with Figure 4.10, shows
that the simulation results are actually quite acceptable – the lower order harmonics
are present in similar levels, and the 5 th is dominant, giving similar THDs for both.

The measurements and simulation results show that the THD values of voltage
and current is 0.83% and %3.67. This is compliance with IEEE 519 and IEC 61000-
2-4 standarts. IEEE 519 recommends Vh ≤ 5% for V< 69 kV. The harmonic values
that we measured were not high when we compared them with the values that we
expected. So it is not necessary to add a harmonic filter to this distribution system.

4.5 Harmonic Study of a Distorted Distribution System

Consider the industrial system of Figure 4.11. The system consists of two buses
IND1 and IND2 connected through a short 3-phase, 4-wire line. The system is
supplied by the utility through a 69kV/13.8 kV transformer. A line commutated
power converter is connected on bus IND2. the harmonic voltage and THD on Buses
IND1 and IND2 are calculated and simulated with MATLAB.
 64

UTIL
69kV
∆
20,000 kVA
Y-g IND1
13.8 kV
PFC

Line
10,000 kW, 0.85 lag

IND2
PFC

3,000 kW, 0.9 lag

5,000 kW, 0.85 lag

Figure 4.11 Single line diagram of the distribution system.

The system data are as follows:

Transformer : 69kV-∆/13.8kV-Y-g, 20 MVA, Z=0.5+j0.8.

Line :Short distribution line Z=0.02+j0.06 Ω.

Load on IND1 :10,000 kW, 0.85 lag pf. Of this load, 60% is motive.

Load on IND2 :3,000 kW, 0.9 lag pf. Largely residential and commercial.

Converter on IND2 :3-phase line commutated rectifier . 5,000 kW, 0.85 lag pf. The
rectifier produces the full spectrum of its characteristic orders at their normal
amplitude and phase. Non characteristic harmonic orders are not produced.

The calculations will be done using the p.u. system equivalent. The base
quantities are:
 65

Power base=10,000 kVA

S base3Q = 3.S base1Q 2
2
Vbase ( L− L)
V base( L − L) Z base =
Z base = 3.Sbase1Q
S base3Q
2
( 3.13.8kV ) (23.9) 2
Impedance base (on the 13.8 kV side) = = = 19 Ω
3.10 MVA 30
In this system, only positive and negative sequence harmonics will be considered.
Thus, we will assume a symmetric and balanced system. Two methods of solution
are presented and compared: Frequency domain and time domain using MatLab. The
associated files are given in Appendices.

The system impedance is expressed in pu on the given basis. The system load
flow data are shown on table 4.5. We assume that the utility voltage is 1 pu. The
calculations to obtain the load flow data including the pu system impedance are
shown in the associated M file. The load flow solution is shown on table 4.6.
ρ
ρ ρ Actual Value
Per Unit Quality = ρ
Base Value

Table 4.5 Branch data.

From Charging
To bus R (pu) X (pu) Element Type
bus (pu)
UTIL IND1 0.0025 0.04 0 Transformer
IND1 IND2 0.0011 0.0032 0 Line

Table 4.6 Bus data.

Bus P-gen Q-gen P-load Q-load Shunt V phase

UTIL 1.860 0.152 0 0 0 1.0 0.0

IND1 0 0 1.0 0.62 0.62 0.992 -4.16

IND2 0 0 0.8 0.455 0.455 0.991 -4.3

 66

4.5.1 Harmonic Calculations in the Frequency Domain

The positive/negative sequence equivalent of the system harmonic impedance is

shown in Figure 4.12 (h is the harmonic index). The system is balanced, therefore,
the positive and negative sequence equivalents are identical. Thus, only one circuit is
needed for the harmonic calculations.

The approach used for the system modeling is described in Chapter 3 of the
tutorial. The detail calculations are shown here. With reference to Figure 4.12, each
device is modeled as follows:

UTIL

0.0025

j0.04h
3.2

Transformer

IND1

0.021
2.117

j0.62h

j0.167h
j0.9h

PCF

Motive
Resistive
Part
Part
0.0011
j0.0032h

IND2
2.7

j0.455h
Ich
j1.307h

PCF Resistive
Load Converter

Figure 4.12 Harmonic impedance diagram of the example system.

Utility system: It is a sinusoidal voltage source at fundamental frequency with

zero impedance.
Transformer: The transformer is represented by a combination of series and
parallel impedance between Buses UTIL and IND1. With reference to Figure 4.12,
the values of these elements, at the hth harmonic order, are calculated directly from
the fundamental frequency series impedance of the transformer:
 67

Transformer series impedance at fundamental: ZT=RT+jXT=0.0025+j0.04 pu.

Series resistance in the harmonic equivalent: Rs=RT=0.0025 pu.
Parallel reactance: Xp=hXT=0.04h.
Parallel resistance: Rp=80XT=3.2 pu.

Thus, the equivalent harmonic impedance of the transformer between UTIL and
IND1 is given by (4.5).
h 2 X T2 R p hX T R p2
Z Th = R s + +j h >1 (4.4)
R p2 + h 2 X T2 R p2 + h 2 X T2

Transmission line: The line is of short length, the charging capacitance is

neglected and only the series impedance is used (4.5).

Z Lh = R L + jhX L = 0.0011 + j 0.0032h h ≥ 1 (4.5)

Where, RL and XL are, respectively, the pu series resistance and reactance of the
line at fundamental frequency.

Load at IND1: The load of IND1 consists of 60% induction motors and of 40%
resistive load (including coupling transformer reactance).

The motive part is represented by a series R-X circuit associated with locked rotor
impedance of the motor. Assume 0.8 lag as aggregate power factor for the motors.

Motor real power : PM=0.6PIND1=0.6 pu.

Motor installed power (apparent power): SM=PM/(0.8)=0.75 pu.

Assume a severity factor KI=8. The severity factor is the ratio between the starting
and the rated current of the motor. From that, the combined pu locked rotor reactance
of the motive part at fundamental frequency is calculated (4.6):
 68

V2 1
XM = = = 0.167 pu (4.6)
K I S M 8S M

We assume 1 pu operating voltage for the motor. The pu resistive part of the
locked rotor impedance is found by assuming a quality factor of 8 for the rotor
circuit:

RM=XM/8=0.021 pu.

Thus, the motor harmonic impedance in Figure 4.12 is given by (4.7).

Z Mh = 0.021 + j 0.167h (4.7)

The resistive part at IND1 is represented by a series R-X impedance.

Power of resistive load :PR=0.4PIND1=0.4 pu.

Reactive power :QR=QIND1-QM=0.62-0.6SM=0.17 pu.
Installed power of resistive part :SR=0.435 pu at 0.92 lag (φR=23°.03).

The pu value of the fundamental impedance that produces the same power under 1
pu voltage is:
ZR=1/SR=2.3 pu. XR/RR=tan(φR)=0.425. Therefore: RR=2.117 pu, XR=0.9 pu.

The load harmonic impedance at Bus IND1 is:

Z Rh = 2.117 + j 0.9h (4.8)

Resistive load at IND2: The load harmonic impedance is, therefore:

Z R 2 h = 2.7 + j1.307 h (4.10)
 69

The pu harmonic current of the converter is given in the following Table.

Table 4.7 Currents of the line commutated converter.

h 5 7 11 13 17 19 23 25 29
%Ic1 0.2 0.143 0.091 0.077 0.059 0.053 0.043 0.04 0.034
Ich 0.119 0.085 0.054 0.046 0.035 0.031 0.026 0.024 0.020
θh -π 0 -π 0 -π 0 -π 0 -π

(continued)
h 31 35 37 41 43 47 49
%Ic1 0.032 0.029 0.027 0.024 0.023 0.021 0.02
Ich 0.019 0.017 0.016 0.014 0.014 0.012 0.012
θh 0 -π 0 -π 0 -π 0

The Nodal admittance matrix is used for the calculation of the system harmonic
voltage. With reference to the circuit topology in Figure 4.12, at each frequency of
the converter harmonic orders, this matrix is computed according to (4.11).

 1 1 1 1 1 
 Z + Z + Z + Z + j 0.62h −
Z Lh 
Yh =  Th Lh Mh Rh
 (4.11)
 1 1 1
− + + j 0.455h 
 Z Lh Z Lh Z R 2 h 

Inverting the nodal admittance matrix we obtain the nodal impedance matrix at
each harmonic order.

Z h = Yh−1 (4.12)

Subsequently, the harmonic voltage at Bus IND1 (corresponding to row and

column 1 of the matrices) is found using the transfer impedance, Zh(1,2), between
Buses 2 and 1 (IND2 and IND1). The harmonic voltage at Bus IND2 (corresponding
to row and column 2 of the matrices) is found using the driving point impedance,
Zh(2,2), at Bus 2.
 70

V IND1, h = Z h (1,2) I ch
(4.13)
V IND 2, h = Z h (2,2) I ch

Note that the converter current represents an injection at node 2 of the network.

Figure 4.13 shows the driving point and transfer impedance. As it can be seen
from the figure, the system forms two resonance frequency, one around the 5th
harmonic and one around the 35th harmonic.

Harmonic impedances Z11,Z12

2

1.8

1.6

1.4

1.2
p.u. ohm s

0.8

0.6

0.4

0.2

0
0 5 10 15 20 25 30 35 40 45 50
Harmonic index

Figure 4.13 System harmonic impedance.

Figure 4.14 shows the harmonic voltage at the two system buses. It can be seen
from this figure that the harmonic voltage of both buses at the 5th and the 35th orders
is amplified due to the impedance resonance.
 71

Harmonic Voltage at IND1 and IND2

0.12

0.1

0.08

p.u. k V
0.06

0.04

0.02

0
5 10 15 20 25 30 35 40 45 50
Harmonic index

Figure 4.16 System harmonic voltage.

4.5.2 MATLAB Time Domain Modeling

The purpose of the following calculations is to show how our harmonic problem
can be solved using MATLAB modeling and simulation.

The positive/negative sequence equivalent-circuit used for the MATLAB

simulation is shown in Figure 4.15 With reference to this figure, the circuit has
identical topology as the circuit used for the frequency domain calculations.

The circuit data are expressed in terms of inductance and capacitance. A harmonic
current source is connected at the converter bus. Additional sources of fundamental
frequency are included for the derivation and plotting of the system waveforms. It
should be noted that the inclusion of these fundamental sources is not necessary for
this example, since the circuit is linear. That is, the same response for the harmonic
voltage would be obtained, if these sources were absent. Fundamental frequency
sources are necessary, if accurate system waveforms are desired for demonstration
purpose or if the circuit is nonlinear. The calculation of the system data is as follows:
 72

cos(2πt)

UTIL

0.0025

j0.04
3.2
Transformer

IND1

2.117
j0.62

0.021
j0.9
PCF

j0.167
5.46cos(2πt+87.32)

Resistive Motive
Part Part

0.0011
j0.0032
2.7
IND2

j0.455 ich(t)
0.594cos(2πt-36.09)
j1.307

PCF
Converter
Resistive
Load

Figure 4.15 MATLAB model of the example system (One phase simulation).

Calculation of Inductance and Capacitance. This would require converting the

known pu values of the fundamental impedance to corresponding H of F. However,
we have the alternative option of entering inductance and capacitance in pu. PU
calculation of inductance and capacitance uses pu time. Here for convenience, the
time base is set to 1/50 Hz. That is 1 pu time is equivalent to 1 fundamental period.
The inductance and capacitance pu calculations are as follows:

di di di X L' di X L' di '

v(t ) = L = L = fL = XB ⇒ v' (t ' ) = (4.14)
dt dt ' t B dt ' 2π dt ' 2π dt '

dv dv BC' dv BC' dv'

i (t ) = C = fC = BB ⇒ i ' (t ' ) = (4.15)
dt dt ' 2π dt ' 2π dt '

Where: tB=1/f is the time base.

t' is the time in pu.
XB, BB are respectively the base of reactance and susceptance.
 73

XL', BC' are respectively the pu fundamental reactance and susceptance of the
inductor and capacitor respectively.

Thus, the pu inductance and capacitance are found from the pu fundamental
reactance and susceptance respectively dividing by 2π. MATLAB data entry permits
further simplification of these conversions. Figure 4.15 shows the pu fundamental
reactance and susceptance of the system as they were entered in the MATLAB data
file. These values are retrieved directly from the previous frequency domain data.

Source Modeling. To obtain the accurate waveform of the voltage, the

fundamental as well as the harmonic sources must be described with their correct
angle. With reference to Figure 4.15, the external utility is modeled as a voltage
source at UTIL with unity amplitude and zero angle. Note that the cos function is
used for describing the sources. This is consistent with the MATLAB source
modeling and also makes the calculation of harmonic phase angles easier.
The induction motor at Bus IND1 must be properly augmented by a fundamental
current source, because the series R-X representation alone does not account for the
motor's full fundamental current. Equation (4.16) is the current of the motor drawn at
fundamental frequency as it is calculated from the load flow. Equation (4.17) is the
current of the motor drawn at fundamental frequency as it is calculated by the circuit
representation of the motor in Figure 4.15

I 1 = S M* / V M* ≈ 0.75 ∠ −36.87 / 1∠ 4.16 = 0.75 ∠ − 41.03 puA (4.16)

I 1' = V M /( R + jX ) + I M = 1∠ − 4.16 /(0.021 + j 0.167) + I M = 5.95 ∠ −87 + I M (4.17)

Where: IM is the augmented current source. The phase angles are obtained from
the load flow solution and the power factor. Equating (4.16) and (4.17) we obtain the
value of the augmented current source in Figure 4.15.

I M = I 1 − I 1' = 5.46 ∠87.32

Table 4.8 shows the description of the equivalent harmonic current sources as
they are included in the MATLAB file. (Note: time is in pu.)
 74

Table 4.8 Equivalent current sources at IND2

Harmonic Order Description
5 0.8os(5t-0.45)
7 0.085cos(7t+107.37)
11 0.054cos(11t+143)
13 0.046cos(13t-109.17)
17 0.035cos(17t-73.53)
19 0.031cos(19t+34.29)
23 0.026cos(23t+69.93)
25 0.024cos(25t-112.32)
29 0.020cos(29t-146.61)
31 0.019cos(31t-38.70)
35 0.017cos(35t-3.15)
37 0.016cos(37t+104.67)
41 0.014cos(41t+140.31)
43 0.014cos(43t-111.87)
47 0.012cos(47t-76.23)
49 0.012cos(49t+31.59)

The MATLAB model of the system is shown in Figure 4.15.

 75

AC Voltage Source

+
v
-
Voltage Measurement2 Scope3
0.0025

3,2 0,04/(2.PI.50)

+
v
-
Voltage Measurement1 Scope2

1/(2.PI.50.0,62) AC current
0,021
Source of
2,117
M Motive

0,0011

0,9/(2.PI.50) 0,167/(2.PI.50) Z
0,0032/(2.PI.50)
Impedance Measurement

+
v
-
Voltage Measurement Scope

1/(2.PI.50.0,455)
AC Current
+

CM Source
i
-

Conn1 Scope4

Conn2 5 th harmonic
2,7 filter
converter
1,307/(2.PI.50)

Discrete,
Ts = 5.208e-005 s.

powergui

Figure 4.16 MATLAB model of the system

The current harmonics of the converter are modelled as current source in

Converter Block according to the values in Table 4.8.

The voltage waveforms of UTIL, IND1 and IND2 node is shown in Fig 4.16. It
can be seen that the total harmonic distortions of the nodes are very high anda re not
complience with harmonic limits.
 76

(a)

(b)
 77

(c)
Figure 4.16 The Total harmonic distortions and of nodes without filter (a) UTIL (b) IND1 (c)IND2

The highest harmonic distortion is at 5th harmonic. So it is necessary to add a

harmonic filter for this frequency component. When the S switch is closed and the
filter is on, the voltages of the nodes are shown in Figure 4.17.
 78

(a)

(b)
 79

(c)
Figure 4.17 The Total harmonic distortions and of nodes with filter (a) UTIL (b) IND1 (c)IND2

It can be seen that the distortions of the voltages are extremely low when the filter
is on. And they are compliance with IEEE 519 standard.
 80

5 CHAPTER FİVE
6 CONCLUSİONS

The main emphasis of this thesis is to analyse and measure harmonic disturbances
in power networks and their trainsients
To simulate harmonics using an analytical tool, MATLAB Simulink Power
System Toolbox, conducted at the medium voltage (MV) (34.5 kV) feeder of
ÜNİVERSİTE TM distribution system and regular measurements were taken. This
feeder is chosen as there were some power quality disturbances.

Furthermore passive filter was added to a similar system and this system was
simulated with MATLAB Simulink Power System Toolbox.

Simulations are important to study the harmonic penetration in power systems

and the effects on loads, but also here there is often difficult to obtain corresponding
values to the monitored ones. This is due to the lack of accurate harmonic models of
aggregated loads, linear and nonlinear, and the power system. Monitoring the
distortion is important to get real data of the harmonics amplitude and phase angle
and of the variation over the time, i.e. the day, week or year. But there are several
possible sources of error when analyzing the monitored data, not only the technical
measurement, due to harmonic interaction in the system.

80
 81

REFERENCES

Aredes M. and Watanabe E.H.(1995). New Control Algorithms for Series and Shunt
Three-Phase Four-Wire Active Power Filters, IEEE Trans. on Power Delivery,
Vol. 10, No. 3, pp. 1649-1656.

Arrillaga, J., Bradley, D.A & Bodger, PS. (1985). Power System Harmonics, John
Wiley & Sons.

Bollen M. H. J.(1999). Understanding power quality problems:voltage, sags and

interruptions, New York: IEEE Press.

Bollen M.H.J.(1999). Understanding Power Quality Problems: Voltage Sags and

Interruptions, IEEE Press, USA.

Brauner G. (1997). Voltage Fluctuations Caused by Windparks, Proceedings of the

Power Quality Applications´97 (Europe) Conference, Section 2B: Connection of
Particular Loads, Stockholm, Sweden.

Davis T., Beam G.E., and Melhorn C.J., (1996). Voltage Sags: Their Impact on the
Utility and Industrial Customers, Proceedings of the Industrial and Commercial
Power Systems Technical Conference, pp. 65-73, New Orleans-Louisiana, USA,.

Dugan R.C., McGranaghan M.F., and Beaty H.W.(1996). Electric Power Systems
Quality, McGraw-Hill.

Fujita H. and Akagi H. (1998). The Unified Power Quality Conditioner: the
Integration of Series- and Shunt-Active Filters, IEEE Trans. on Power
Electronics, Vol. 13, No. 2, pp. 315-322.
 82

Häger M., Wallin L., and Swärd C.(1997). Flicker Influence from a 100 MVA Arc
Furnace Installation in a Weak Network, Proceedings of the Power Quality
Applications´97 (Europe) Conference, Section 2B: Connection of Particular
Loads, Stockholm, Sweden.

IEEE, (1995), IEEE Recommended Practice for Monitoring Electric Power

Quality,” IEEE Std. 1159-1995.

IEEE (1992).IEEE Recommended Practice for Powering and Grounding Sensitive

Electronic Equipment, IEEE Std. 1100-1992.

Kajihara H. H.(1968). Quality power for electronics, Electro Technology, Vol. 82,
No. 5, p. 46

Knoller P. M. and Lonnstam L.,(1969). Voltage quality and voltage tendency

recorders, Siemens Review, Vol. 36, No. 8, pp. 302 – 303

Larsson T. and Poumarède C.(1999). STATCOM, an Efficient Means for Flicker

Mitigation, IEEE Power Engineering Society 1999 Winter Meeting, Vol. 2, pp.
1208-1213, New York-NY, USA.

Lawrance, W., Hovingh, B., Lo, TY., and Mistry,B. A Scheme for Harmonic
Attenuation in Commercial Buildings, School of Electrical and Computer
Engineering, Curtin University of Technology, Perth Western Australia.

Mansoor, A., Grady, W.M., Staats, P.T., Thallum, R.S., Doyle, M.T. and Samotyj,
Hansen,S., Nielsen, P. and Blaabjerg, F. (2000). Harmonic Cancellation by
Mixing Nonlinear Single-Phase and Three-Phase Loads, IEEE Transactions on
IndustryApplications, Vol. 36, No. 1.
 83

M.J. (1995). Predicting the Net Harmonic Currents Produced by Large Numbers of
Distributed Single-Phase Computer Loads, IEEE Transactions on Power
Delivery, Vol. 10, No. 4.

Palesjö A. (1992). An Analysis of Compensators Used for Flicker Reduction,

Technical Licentiate Thesis, ISSN 1100-1615, TRITA-EHE 9203, The Royal
Institute of Technology, Stockholm, Sweden,.

Plette D. L.(1969). The effects of improved power quality on utilization equipment,

Proceedings of the IEEE National Aerospace Electronics Conference, , Dayton,
OH, pp. 243 – 250

Schlabbach, J., Blume, D. & Stephanblome, T. (2001), “Voltage Quality in

Electrical Power Systems, IEE, London United Kingdom.

Soliman, S.S. & Srinath, M.D. (1998), Continuous and Discrete Signals and Systems,
2nd edition, Prentice Hall, New Jersey

Taylor G.A., Hill J.E., Burden A., Mattern K. (1998), Responding to the Changing
Demands of Lower Voltage Networks – the Utilization of Custom Power Systems,
Paper No. 14-303 presented at CIGRÉ 1998 Session.

Wagner V.E., Andreshak A.A., and J.P. Staniak, (1990). Power Quality and Factory
Automation, IEEE Trans. on Power Delivery, Vol. 26, No. 4, pp. 620-626.

Wang F., (2001). On power quality and protection, Licentiate Thesis, Chalmers
University of Technology

Warren C.A., Short T.A., Burke J.J., Morosini H., Burns C.W., and Storms J. (1999).
Power Quality at Champion Paper – The Myth and the Reality, IEEE Trans. on
Power Delivery, Vol. 14, No. 2, , pp. 636-641.
 84

8 APPENDIENCES

MATLAB program to make the frequency domain analysis of the distribution

system in Figure 4.14

clear all
clc
close all
% Calculate harmonic impedance
ZTh='0.0025+(h^2*0.04^2*3.2+i*h*0.04*3.2^2)/(3.2^2+(h*0.04)^2)';
%Transformer impedance
ZLh='0.0011+0.0032i*h'; %Line Impedance
ZMh='0.021+0.167i*h'; %Motor impdance
ZR1h='2.117+0.9i*h'; %Load Impedance at IND1
ZR2h='2.7+1.307i*h'; %Load impedance at IND2
B1h='0.62i*h'; %Compensation PFC1
B2h='0.455i*h'; %Compensation PFC2

H=[5;7;11;13;17;19;23;25;29;31;35;37;41;43;47;49]; % Define the rectifier

current harmonic orders
Ic=[0.119;0.085;0.054;0.046;0.035;0.031;0.026;0.024;...
0.020;0.019;0.017;0.016;0.014;0.014;0.013;0.012]; %Define the rectifier
harmonic current

%Start calculating harmonic voltage at bus IND1 (#1) and IND2 (#2)
for n=1:16,1;
h= H(n);
% Reset harmonic bus admittance matrix, Yh
Yh=[0,0;0,0];

% Form Yh
Yh(1,1)=1/eval(ZTh)+1/eval(ZMh)+1/eval(ZR1h)+1/eval(ZLh)+eval(B1h);
 85

Yh(1,2)=-1/eval(ZLh);
Yh(2,1)=Yh(1,2);
Yh(2,2)=1/eval(ZLh)+1/eval(ZR2h)+eval(B2h);

% Calculate harmonic bus impedance matrix, Zh

Zh= Yh^(-1);

V1h(n)= abs(Zh(1,2)*Ic(n)); %calculate harmomonic voltage at IND1

V2h(n)= abs(Zh(1,1)*Ic(n)); %calculate " " at IND2
end;

stem(H,V1h),xlabel('Harmonic index'),ylabel('p.u. kV');

title('Harmonic Voltage at IND1 and IND2');
hold on;
stem(H,V2h,'r');
hold off;

% Computation and Plot of harmonic impedance

Dh=0.5;
for n=1:100,1;
h= n*Dh;
hw(n)= h;
% Form Yh
Yh(1,1)=1/eval(ZTh)+1/eval(ZMh)+1/eval(ZR1h)+1/eval(ZLh)+eval(B1h);
Yh(1,2)=-1/eval(ZLh);
Yh(2,1)=Yh(1,2);
Yh(2,2)=1/eval(ZLh)+1/eval(ZR2h)+eval(B2h);
% Calculate Zbush
Zh= Yh^(-1);
Z11(n)=abs(Zh(1,1));%Thevenin impedance at converter bus
Z12(n)=abs(Zh(1,2));%Transfer impedance between IND1 and IND2.
 86

end;
figure;
plot(hw,Z11,'b',hw,Z12,'r');
xlabel('Harmonic index');
ylabel('p.u. ohms');
title('Harmonic impedances Z11,Z12')