An Investigation of the Effects of Voltage and Current

Harmonics on an Electrical Distribution Island

Network

by

KABELO CLIFFORD MODIPANE

Dissertation

Submitted in partial fulfilment of the requirements for the degree of

Magister Ingeneriae

in

Electrical and Electronic Engineering Science

in the

Faculty of Engineering and the Built Environment

At the

University of Johannesburg

Supervisor: Prof J-H.C. Pretorius

Co-supervisor: Prof I.W. Hofsajer

July 2005
 3

Acknowledgements
• To our Creator for the privilege; without Him at my side I would not
have been able to complete this dissertation.

• To my family and friends for their support and encouragement.

• To my supervisor, Prof Jan-Harm Pretorius, and co-supervisor, Prof

Ivan Hofsajer; thank you for the guidance, support, motivation and
technical leadership.

• To Chris van Dyk from Power System Dynamics; thanks for your help
on PSCAD Software applications.

• To F Velosa; thank you for professional language editing.

 4

Abstract

With the advances in technology and the increase in industrial facilities, the
harmonic content of an electrical network has always been a major concern
among power system engineers. This is especially in an industrial
environment, where the usage of power is very high as there are many large-
scale types of equipment being used. Thus, it would be useful to know the
sources, distortion level, impact on the power system and the equipment of
harmonic currents on the harmonically rich electrical network.

The presence of harmonics on electrical networks poses many problems to

power system engineers and inconvenience or loss to the industries. Their
effects on power system apparatus include resonance, reduced operating life
of rotating machines and error in power calculations. For this study, the author
would like to find out the implications of this issues, following these harmonic
problems by analyzing, with some amount of simulations, practical
measurements and assessments. In this way, a better understanding could be
gained about these harmonic problems and harmonic contents. Steps could
be taken to protect the power system equipment that could be affected by
high harmonic currents and raise the quality of power supply.
 5

Table of contents

ACKNOWLEDGEMENTS ...............................................................................3

ABSTRACT .....................................................................................................4

TABLE OF CONTENTS ..................................................................................5

LIST OF SYMBOLS AND DEFINITIONS ........................................................9

LIST OF ABBREVIATIONS...........................................................................10

LIST OF FIGURES ........................................................................................11

LIST OF TABLES..........................................................................................12

1 INTRODUCTION ....................................................................................13
1.1 Aim of the study........................................................................................................ 14

1.2 Description of facilities being studied ................................................................... 15

1.2.1 Background of SAC and Hart RAO ......................................................................... 15

1.2.2 Power generation...................................................................................................... 16

1.2.3 Harmonics sources .................................................................................................. 17

1.3 Concerns facing SAC............................................................................................... 18

1.3.1 Harmonic distortion level of SAC network ............................................................ 18

1.3.2 Can the loads be increased?................................................................................... 18

1.3.3 A model required to simulate SAC network .......................................................... 19

1.4 Overview of the document....................................................................................... 21

2 HARMONIC PROBLEMS.......................................................................23
2.1 Introduction............................................................................................................... 23

2.1.1 Fourier series representation.................................................................................. 24

2.2 Causes of harmonics ............................................................................................... 24

2.3 Classification of power quality problems on electrical network ......................... 26

2.3.1 Transients.................................................................................................................. 26
 6

2.3.2 Long duration voltage variations............................................................................ 28

2.3.4 Voltage dip ................................................................................................................ 29

2.3.5 Voltage unbalance.................................................................................................... 30

2.3.6 Waveform distortion................................................................................................. 31

2.4 Summary ................................................................................................................... 36

3 POWER SYSTEM COMPUTER AIDED DESIGN MODEL (PSCAD) ....38

3.1 Introduction............................................................................................................... 38

3.2 Modelling of power system components............................................................... 39

3.3 Electrical diagram of SAC and Hart RAO............................................................... 40

3.4 Time domain simulation .......................................................................................... 41

3.5 Validation of the simulation results........................................................................ 42

3.6 Time domain simulation results to show the accuracy of the PSCAD model vs
real measurements................................................................................................... 43

3.7 Transmission losses ................................................................................................ 50

3.7.2 The effect on transmission efficiency if the load is three times the original load
53

3.8 Summary ................................................................................................................... 54

4 CURRENT AND VOLTAGE MEASUREMENT ......................................55

4.1 Power definitions in the presence of harmonics .................................................. 55

4.1.1 Introduction............................................................................................................... 55

4.1.2 Single-phase powers................................................................................................ 56

4.2 Limitations of the production of harmonics on an electrical network................ 58

4.3 South African Power Quality Standards ................................................................ 58

4.3.1 NRS 048 part 2: minimum standards...................................................................... 59

4.3.2 Voltage harmonics ................................................................................................... 59

4.3.3 Recommended planning levels for harmonic voltage on HV and EHV systems 60

4.4 IEEE Standard 519.................................................................................................... 60

4.5 Field measurements................................................................................................. 61

4.6 Electrical network diagram...................................................................................... 62

4.7 Measurement 1.......................................................................................................... 63

 7

4.8 Measurement 2.......................................................................................................... 66

4.9 Measurement 3.......................................................................................................... 70

4.10 Voltage and current unbalanced............................................................................. 72

4.11 New power-quality assessment criteria for supply system under unbalanced
and non-sinusoidal conditions .............................................................................. 74

4.12 Summary ................................................................................................................... 78

5 DESIGN/MANAGEMENT TOOL FOR MAINTANANCE AND FUTURE

EXPANSION...........................................................................................80
5.1 Introduction............................................................................................................... 80

5.2 Effects of harmonics on a power system .............................................................. 80

5.3 Transmission system/cable..................................................................................... 80

5.3.1 Heating and loss effects caused by harmonic currents on underground cables
81

5.3.3 Model description ..................................................................................................... 83

5.3.4 Calculation of ohmic losses .................................................................................... 83

5.3.5 De-rating due to harmonics..................................................................................... 85

5.4 Transformers............................................................................................................. 86

5.4.1 Transformer de-rating factor ................................................................................... 87

5.4.2 Estimation method for power system harmonics effects on power transformers
91

5.4.3 Methods of harmonic load content estimation...................................................... 91

5.4.4 Limitations of the crest factor and %THD methods.............................................. 92

5.4.5 Expansion of the k-factor method .......................................................................... 93

5.4.6 Effects of harmonic currents on the Hart RAO network and motor–generator set
98

5.4.6.1 Effects on the Hart RAO network............................................................................ 98

5.4.7 Effects of harmonic currents on a motor............................................................... 99

5.4.8 Unbalanced currents on a synchronous generator ............................................ 100

5.4.9 Neutral currents in a three-phase power system ................................................ 101

5.4.10 Problem of a high neutral current......................................................................... 102

5.5 Summary ................................................................................................................. 102

 8

6 CONCLUTIONS AND RECOMMENDATIONS ....................................104

6.1 Conclusions ............................................................................................................ 104

6.2 Recommendations and future work ..................................................................... 106

7 ANNEXURES .......................................................................................108
Annexure A: PSCAD software............................................................................................ 108

Annexure B: Unbalance calculations ................................................................................ 110

Annexure C: Cable parameters .......................................................................................... 111

8 LIST OF REFERENCES.......................................................................112
 9

List of symbols and definitions

EHCI Equivalent Harmonic Current Index

3Φ Three–phase
P Active Power
Q Reactive Power
IH RMS Harmonic Current
I RMS Current
I1 Fundamental Current
IEH Equivalent Harmonic Current
U1 Fundamental Voltage
THD Total Harmonic Distortion
V THD Voltage Total Harmonic Distortion
I THD Current Total Harmonic Distortion
UHIH Harmonic Apparent Power
U1IH Current Distortion Power
UHI1 Voltage Distortion Power
N Total Non-active Power
NH Total Harmonic Non-active Power
PH Total Harmonic Active Power
S Apparent Power
S1 Fundamental Apparent Power
P1 Fundamental Active Power
Q1 Fundamental Reactive Power
SN Non- fundamental Apparent Power
TWSCL Total Weighted Sum Composite Load
ao Dc Component
Vh Peak Voltage Level
F Fundamental Frequency
Θh Phase Angle
K De-rating Factor
Δ Skin Depth
σ Conductivity
Μ Permeability
IB Base Current
PU Per Unit
UB Voltage Unbalance
 10

List of abbreviations
AC Alternating Current
CSIR Council for Scientific and Industrial Research
DC Direct Current
EHV Extra High Voltage
SAC Satellite Application Centre
UPS Uninterruptible Power Supply
PF Power Factor
RMS Root Mean Square
IEEE Institute of Electrical and Electronic Engineering
SMPS Switch Mode Power Supply
IHD Individual Harmonic Distortion
PSCAD Power System Computer-aided Design
Hart RAO Hartbeeshoek Radio Astronomy Observatory
HDF Harmonic Distortion Factor
HV High Voltage
VSD Variable Speed Drives
IGY International Geophysical Year
USA United States of America
RF Radio Frequency
LV Low Voltage
MV Medium Voltage
 11

List of figures

Figure 1-1: Network diagram of SAC and Hart RAO...................................16

Figure 2-1: Distorted waveform composed of fundamental and 3rd harmonic
23
Figure 2-2: Impulsive transient ...................................................................26
Figure 2-3: Oscillatory transient waveform..................................................27
Figure 2-4: Voltage sag ..............................................................................27
Figure 2-5: Voltage swell ............................................................................28
Figure 2-6: An interruption due to a fault ....................................................28
Figure 2-7: Measured voltage dip parameters ............................................30
Figure 2-8: Balanced and unbalanced phase voltages ...............................31
Figure 2-9: Harmonic distortion...................................................................32
Figure 2-10: Representation of symmetrical components.............................33
Figure 3-1: Network diagram of SAC and Hart RAO...................................40
Figure 3-2: Simplified electrical model of SAC............................................41
Figure 3-3: Comparison of model and field measurements at the generator
43
Figure 3-4: Line currents of model and field measurements taken at the
SAC generator..........................................................................45
Figure 3-5: Normalised harmonic current of model and field measurements
45
Figure 3-6: SAC network model simulation 1..............................................46
Figure 3-7: Line currents and line-to-line voltages of simulation 1 at the SAC
generator respectively ..............................................................47
Figure 3-8: SAC network model simulation 2 after increasing the load.......48
Figure 3-9: Line-to-line voltages and line currents at the Sac generator after
increasing the load respectively ...............................................48
Figure 3-10: SAC Network model simulation 3 at the Load ..........................49
Figure 3-11: The line currents and line-to-neutral voltages of simulation 3 of at
the SAC load respectively .........................................................49
Figure 3-12: Line-to-line voltage at the generator with two times the original
load ...........................................................................................51
Figure 3-13: Line-to-line voltage at generator, three times the original load .53
Figure 4-1: SAC and Hart RAO network diagram .......................................62
Figure 4-2: SAC network model..................................................................63
Figure 4-3: The line-to-neutral voltage and the line current measured on the
ESKOM side .............................................................................64
Figure 4-4: SAC network diagram for measurement 2................................66
Figure 4-5: The line-to-neutral voltage and line current at the primary side of
the step-up transformer ............................................................66
Figure 4-6: SAC network diagram for measurement 3................................70
Figure 4-7: The line-neutral voltage and line current at the SAC load on the
secondary side of the step-down transformer...........................71
Figure 4-8: Daily trend of current non-linear index and current THD ..........78
Figure 5-2: SAC and Hart RAO network diagram .......................................98
 12

List of tables

Table 3-1: Comparison of model and real measurements in terms of power

measurement ...........................................................................46
Table 3-2: Comparison of transmission efficiency of the model and real
system ......................................................................................50
Table 3-3: Transmission efficiency using non-ideal transformer................51
Table 3-4: Transmission efficiency using the ideal transformer double the
original load ..............................................................................52
Table 3-5: Transmission efficiency using the non-ideal transformer double
the original load ........................................................................52
Table 3-6: Transmission efficiency using the ideal transformer, three times
the original load ........................................................................53
Table 3-7: Transmission efficiency using the non-ideal transformer, three
times the original load...............................................................53
Table 4-1: The maximum levels for harmonic voltages expressed as a
percentage of the declared voltage of the LV and MV power
systems ....................................................................................59
Table 4-2: Indicative values of planning levels for harmonic voltage .........60
Table 4-3: IEEE measured data red phase................................................64
Table 4-4: Comparison of NRS 048 and measurement on the ESKOM
supply to SAC and Hart RAO ...................................................65
Table 4-5: IEEE measured data blue phase ..............................................67
Table 4-6: Comparison of NRS 048 and measurements at the SAC
generator ..................................................................................68
Table 4-7: Comparison of NRS 048 measurements at the SAC generator69
Table 4-8: Comparison of NRS 048 and measurements at the SAC
generator ..................................................................................69
Table 4-9: IEEE measured data RED phase .............................................71
Table 4-10: Comparison of NRS 048 and the SAC at the LOAD side .........72
Table 4-11: Current and voltage on the three-phase at the SAC generator 73
Table 4-12: Sequential currents and voltages .............................................73
Table 4-13: Current non-linear index and current THD three-phases..........77
Table 5-1: Cable parameters .....................................................................83
Table 5-2: Ohmic losses calculated for the RED phase ............................84
Table 5-3: Ohmic losses calculated for the BLUE phase...........................84
Table 5-4: Ohmic losses calculated for the WHITE phase ........................84
Table 5-5: RMS phase and peak currents .................................................90
 13

1 INTRODUCTION

The increase in recent decades of electrical equipment that produce

harmonics has posed significant problems for electrical networks and power
quality. Harmonics are sinusoidal voltages or currents, the frequency of which
are a multiple of the fundamental frequency (50 Hz) of the power system.
Because electrical devices that act as non-linear loads draw current non-
linearly, they are responsible for injecting harmonic currents into the electricity
network. Some problems caused by harmonics include cable/conductor failure
and the overheating of transformer windings.

The effect that harmonics have on electrical equipment, especially sensitive

electronic equipment, has only become apparent in the last number of years.
Power interference in communications is only one of the undesirable side
effects that harmonics could cause when present in the electricity network.
Therefore, there is a need to investigate the amount of harmonic content or
distortion level allowable on electrical systems, and to find ways to remedy the
situation so that adverse effects and added costs are avoided, for future
expansion to be possible.

Deregulation in the electricity industry, coupled with changing customer loads

and demand, has created a growing need for electricity utilities to provide
customers with power quality [1]. Power quality is any anomalous behaviour
on a power system arising in the form of voltage and/or current, which
adversely affects the normal operation of electrical or electronic equipment
[1]. Under normal (ideal sinusoidal, symmetrical and balanced) conditions,
power quality is basically a loading issue. But with the growth in the power
electronics and control systems industries, the once mostly linear customer
loads are now being dominated by sensitive and sophisticated non-linear
customer loads, which have a catastrophic impact on a power system and
equipment.

Harmonics studies require a thorough understanding of the mechanics and

operations of power system parameters and loads. Knowledge of power
systems includes the roles of the generation, transmission and distribution
hierarchies, while an understanding of the diverse applications of a range of
 14

varying sized loads is important. The choice of measuring equipment is also

important to ascertain a solution to these harmonic problems. Harmonic
studies involve: gathering a data resource and analysing the data with power
quality standards to evaluate network compliances with appropriate
standards.

With the growing concern about the level of harmonic current and voltage
distortion present in electrical networks, a number of regional standards have
been introduced in countries throughout the world. These standards are
imposed on electrical items, electricity customers and electricity utilities. There
are many standards available, for example, the IEEE 519-1992 USA “IEEE
Recommended Practices and Requirements for Harmonic Control in Electrical
Power Systems” and The NRS 048 (South Africa).

1.1 Aim of the study

The main objectives of this study are to investigate the harmonic impact on a
power system and its equipment and the power quality of harmonic currents at
the Satellite Application Centre (SAC) power network when considering future
load growth or expansion; to determine the harmonic content or harmonic
voltage distortion levels at SAC; to see whether network voltage distortion is
exceeding the standard limits; to determine the detrimental effects of the
harmonic currents on power system equipments when considering future
expansion or load growth on the SAC network; to develop a model that could
be used to simulate SAC’s power network; and to identify general concerns
when considering expansion in the network.

The expected outcomes of this investigation are to show successfully the

harmonic content, the voltage distortion level at SAC, and the effects of
harmonic currents on power system equipment, and to develop a successful
model that could simulate the SAC power network.
 15

1.2 Description of facilities being studied

1.2.1 Background of SAC and Hart RAO

“The Satellite Applications Centre, as we know it today, had its origin at the
dawn of the space age in December 1957 when the CSIR agreed to install
and operate a 108 MHz “MINITRACK” interferometer satellite tracking system
in South Africa. This system was to be part of a world-wide tracking network of
similar systems, designed by the United States Naval Laboratories (NRL) to
track artificial earth satellites that were to be launched by the United States
with Project Vanguard, as their contribution to the International Geophysical
Year (IGY) in 1958” [2].

“The Hartebeesthoek Radio Astronomy Observatory (Hart RAO) is located

west of Johannesburg, South Africa. It operates as a National Research
Facility under the auspices of the National Research Foundation (NRF). Hart
RAO is the only major radio astronomy observatory in Africa. It is located in a
valley in the Magaliesberg hills, 50 km west of Johannesburg. The
Observatory began as Deep Space Station 51, built in 1961 by the National
Aeronautics and Space Administration (NASA) of the United States of
America. The station tracked many unmanned US space probes. These
included the Ranger, Surveyor and Lunar Orbiter spacecraft which landed on
the Moon or mapped it from orbit, the Mariner missions which explored the
planets Venus and Mars and the Pioneers which measured the Sun's winds.

The station was handed over to the Council for Scientific and Industrial
Research (CSIR) in 1975 and was converted to a radio astronomy
observatory. In 1988 the observatory became a National Facility operated by
the Foundation for Research Development (FRD). In 1999 the FRD was
restructured as the National Research Foundation (NRF)” [3].
 16

1.2.2 Power generation

The harmonic problems for the Satellite Application Centre (SAC) are
investigated in this dissertation. The Satellite Application Centre has its own
power distribution and conversion system. Eskom’s transformer supply to
SAC and Hart RAO is 11 kV 1MVA, 50 Hz. This source is then converted to
115 V (three-phase, 208 V), 60 Hz or 220 V (three-phase, 380 V), 60 Hz,
because the generation station was built by America in 1957. In case of a
power failure, SAC has two 500kVA diesel generators that are able to supply
the full load within 17 seconds of the detected supply failure. The grid is
monitored 24 hours a day. Deviation that exceeds 15% results in the diesel
generator assuming the supply of power until stability is reached in the grid for
a minimum of 25 minutes. At this point, the power supply automatically reverts
to the grid. In the event of a major grid failure, SAC has approximately 40 000
litres of diesel fuel on site. This supply is sufficient for about 10 days.

Figure 1-1 is an electrical diagram of SAC and Hart RAO.

Figure 1-1: Network diagram of SAC and Hart RAO

 17

The power to SAC and Hart RAO is 11kV, 50Hz. This is then transformed
down to 400V by 11kV to 400V delta-wye transformer. The two networks are
coupled to the same bus on the ESKOM side. But each network has its own
converter or motor-generator set. An underground cable of 16 mm2, 11kV
paper lead drain cable 3km long brings this power to the SAC main building,
where it is transformed down to 208/400V by 500kVA 11kV to 208/400V delta-
wye transformer. Hart RAO is 300m away from the generator house, with an
underground feeder. The underground cable was used in these networks to
avoid interference with sensitive equipment used at Hart RAO and SAC. SAC
is responsible for the management and maintenance of the generation station.

1.2.3 Harmonics sources

Two uninterruptible power supply systems (UPS) of 120kVA each and other
critical electronic equipment in the facility are the main source of harmonics on
the system. The presence of unbalanced loads creates unbalanced currents
components that cause voltage drops across the source impedance, hence
generating harmonic powers flowing backwards from the load to the network.

Single-phase non-linear loads, such as personal computers, electronic

ballasts and other electronic equipment, generate odd harmonics (i.e. 3rd, 5th,
7th, 9th, etc.). The troublesome harmonics for single-phase loads are the 3rd
and odd multiples of the 3rd i.e the 9th, 15th, etc. These harmonics are called
“triplens”. The A-phase triplen harmonics, B-phase triplen harmonics and C-
phase triplen harmonics are all in phase with each other [4]. They will then
add rather than cancel the neutral conductor of a three-phase four-wire
system. This can overload the neutral if it is not sized to handle this type of
load.

On the other hand, three-phase non-linear loads like three-phase adjustable

speed drives (ASD’s), three-phase direct current drives and three-phase
rectifiers generate primarily 5th and 7th current harmonics and a small amount
of 11th , 13th and higher orders. These types of loads do not generate triplen
harmonics.
 18

1.3 Concerns facing SAC

Due to current and possible future load growth on the electrical network at
SAC, there could be some major catastrophic impact on a power system
when the load keeps on increasing with a system with a high harmonic level.

The major concerns facing the power system engineers at SAC that need to
be assessed are as follows:

• Harmonic distortion level of SAC network.

• Can the loads be increased and to what power level?

• The effect of harmonics on the power system and equipment.

• A model is required to simulate the SAC power network.

1.3.1 Harmonic distortion level of SAC network

There are two reasons to undertake a system harmonic study: firstly, to study
the impact of the harmonic source and, secondly, to examine a harmonic
problem on an existing system. The harmonic contents or distortion level of
SAC need to be assessed. An approach to this is to benchmark the SAC
harmonics against utility standards for the purpose of checking the network
compliance.

1.3.2 Can the loads be increased?

SAC is a large satellite station, so the load at SAC needs to be increased (i.e.
installation of new antenna on the network). There are several things that
could have a major impact on a power system’s behaviour; thus, there is a
need to investigate the concerns before any major change can be made to the
existing loads.

Firstly, the capability of the underground cable under non-sinusoidal

conditions must be determined, i.e. could future expansion or load growth on
the network affect the performance of underground cables under time-varying
 19

non-linear load conditions? Yes, it could damage the cable because the
harmonic currents generate heat in the cable, which could shorten the life of
the cable, so the cable needs to de-rated to a certain percentage in order to
compensate for future load growth.

Secondly, the rating method for transformers and underground cables needs
to be developed. The transformer capability when supplying non-sinusoidal
loads needs to be evaluated by using a suitable rating system that could
accommodate harmonic frequencies. Non-linear load causes accelerated loss
of insulation life in transformer windings because of the heating effects of high
frequency harmonics. When significant harmonic currents are present in a
system, additional I2R losses occur in the phase or line conductors and the
neutral conductor. The primary effect of power system harmonics on
transformers is the additional heat that is generated by the losses caused by
the harmonic content of the load current. The lifespan of a transformer will be
reduced if it operates above the rated temperatures. Initially, the transformer
may be operated at the reduced loading, but in future, the loading could be
increased without considering the intended de-rating.

Thirdly, the effects on synchronous generators then keep on supplying

unbalanced currents. Unbalanced currents on three-phase machines are
undesirable, but a small amount can generally be tolerated. Excessive
unbalanced currents could shorten machine life and increase energy
consumption.

1.3.3 A model required to simulate SAC network

SAC’s component model

A power system simulation package/program tool is required for the analysis

of system behaviour and for analysing other power quality disturbances that
could occur on the network. SAC has expertise in all aspects associated with
designing, procuring, establishing, commissioning, operating and maintaining
complex satellite systems, specifically with regards to:

a) system and RF engineering activities; and

 20

b) electrical works: air-conditioning, transformers, switchboards,

distribution panels, UPS systems, power cabling, reticulation,
lightning protection, lighting at antennas and shelters (internal and
external) and power distribution.

This means that the load condition can change rapidly. So an accurate model
is required for harmonic studies, to analyse possible future expansion and to
simulate other power quality disturbances. Because the load could increase
due to the installation of new antennas and other power electronic loads,
there is a need to check power network compliance with the standards. If the
distortion level is beyond the allowable limit, steps need to be taken to remedy
the situation so that adverse effects and added costs are avoided.

Three-phase or single-phase modelling is required. The first decision to be

made in any system harmonic study is whether a three-phase model is
required or whether a single-phase model would be sufficient. The three-
phase model is required if:

• a combination of wye-wye and/or delta-wye transformers leads to

harmonic cancellation;

• ground currents are important in the study;

• significant unbalanced loading is present.

It is often recommended to implement a three-phase model on any system

study. There have, however, been numerous successful studies that have
been single phase in nature [16]. The typical instances where a single-phase
model could be sufficient are:

• A single large three-phase harmonic source is the cause of the study.

• The remaining system is well balanced.

• Ground currents are not an issue.

The single-phase model of the system could be attractive as it is one third the
size of the three phase model and the results could be more compact and
easier to interpret. Still, it should be employed only when it is clear that it
 21

would be sufficient for the study being undertaken. For this study, a three-
phase model will be used because the system is unbalanced.

1.4 Overview of the document

Chapter 1 provides a description of the facilities being studied and the
harmonic sources on the network. It also covers the concerns about the
effects of harmonic currents on power system equipment at SAC when
considering possible load growth.

Chapter 2 discusses the harmonic problems in power systems in general. It

also covers the causes of harmonics and discusses other power quality
disturbances that occur in the power system.

In chapter 3, PSCAD software is used to model the SAC electrical network,

using nameplate ratings of power system equipment. The model and field
measurements are evaluated and compared to see the accuracy of the
PSCAD software to a model real system. It covers the impact on transmission
efficiency as the load increases.

Chapter 4 provides an analysis of the current and voltage measured on the

SAC network power quality meter (power harmonic analyser) used to
measure the waveforms. Measurement data will be used to evaluate voltage
harmonic distortion at SAC using NRS 048 as the benchmark. It also covers
the IEEE power definition which will be used to analyse the measured results
to clarify some uncertainty of the measured results and to evaluate the
measured data. Furthermore, it evaluates the unbalance factor on the SAC
network using the sequential component calculations with reference to the
IEEE and NRS 048. It also covers new power quality assessment criteria for
the supply system under unbalanced and non-sinusoidal conditions.

Chapter 5 discusses the effects of harmonics on power system equipment

and methods used to estimate the equipment performance under non-
sinusoidal conditions. Furthermore, its covers a suitable rating system of
underground cables and transformers supplying non-linear loads at SAC.
 22

Finally, the document concludes with chapter 6 providing conclusions and

recommendations.
 23

2 HARMONIC PROBLEMS

2.1 Introduction
Harmonics is a more important issue for the industry, commerce and the
home consumer now than it was a few decades ago. Modern consumers and
commercial equipment are not designed to cope with the increasing levels of
power distortion or pollution. The discussion of the harmonic will starts off with
definition of harmonics, the discussion of the theory of harmonics, causes and
other power quality disturbances.

Harmonics is a sinusoidal component of a periodic wave or quantity having a

frequency that is an integral multiple of the fundamental frequency the power
frequency, fundamental frequency of 50Hz in South Africa or 60 Hz in USA.
The equation representing a harmonic frequency is given by [5]:

f h = 60 Hz × h 2-1
where h is the harmonic order. In the power system, these harmonics interact
with the fundamental frequency waveform and each other to produce a
distorted waveform as shown in Figure 2-1

Figure 2-1: Distorted waveform composed of fundamental and 3rd harmonic

 24

2.1.1 Fourier series representation

Using the Fourier series, any voltage or current waveform could be
reproduced from the fundamental frequency component and the sum of the
harmonic components.
∞
V (t ) = a 0 + ∑ Vh ⋅ sin( h ⋅ 2 ⋅ π ⋅ f ⋅ t + θ h ) 2-2
k =1

where,

ao: dc component

Vh: peak voltage level

f: fundamental frequency

θh: phase angle

The fourier series can also be used to deconstruct a waveform into the
fundamental and harmonic components. This is the principle behind
performing a harmonic analysis on a power system. A waveform is recorded
and the magnitudes of the harmonics in the wave are calculated.

2.2 Causes of harmonics

There are many causes of harmonics on a power system. Harmonics can
arise in the generation of power, in the distribution system, and from the loads
connected to the network. If a generator produces a non-ideal sinusoidal
waveform, the voltage waveform will contain a certain amount of harmonics.
Modern-day generators are however much better in this regard than the
earlier types of synchronous generators. In the distribution system,
transformers are capable of producing harmonics due to magnetic core
saturation.

This is more prevalent at a lighter loading of the transformer. By far the

greatest production of harmonics is, however, via harmonic current generation
from non-linear loads. Non-Linear loads are those in which the load does not
 25

draw a sinusoidal current. Especially with the introduction of power electronic

devices, there are now many devices that act in this manner [6].

With the advent of power electronic devices in rectifiers, motor drives and
power supplies, the increasing levels of harmonics have become quite a
concern for power system engineers. Non-linear loads cause harmonic
distortion in a power system where the current is not proportional to the
applied voltage. A common example of this is in variable speed drives, where
the diode rectifier module within the variable speed drive (VSD) is responsible
for injecting harmonic currents into the system [6].

Research has shown that non-linear loads inject harmonic currents into the
system. Therefore, the harmonic-producing loads can be treated as current
sources. However, these harmonic currents that pass through the system
cause a voltage drop for each harmonic. This results in voltage distortion at
the load bus as well as current distortion.

A particular group of harmonics that deserves special mention is the triplen

harmonics, which are odd multiples of the third harmonic. This is due to the
fact that, unlike the fundamental frequency components that cancel in the
neutral conductor, triplen harmonics coincide in phase and time, and thus
produce a third harmonic current in the neutral that can be 1.7 times the
phase current [4]. These triplen harmonics can be eliminated by the use of a
delta-wye transformer, with the source of harmonics connected to the wye
side. The triplens will flow into the neutral conductor, yet on the delta side,
they are trapped in the delta winding, and hence do not appear in the line
currents on this side.

The most notable effects that harmonics have on a power system are impact
on the quality of the AC voltage waveform (i.e., it will become distorted), thus
causing problems with other sensitive loads connected to the same supply [7].

In transformers, harmonic currents cause the RMS current to be greater than

its capacity, leading to increased conductor loss and heating. In motors,
decreased efficiency, excessive heating, and vibration are symptoms of
harmonic voltage distortion. The tripping of protective relaying, telephone
 26

interference, and false meter readings are other consequences of harmonics

in power systems.

2.3 Classification of power quality problems on electrical

network
Before conducting a power quality measurement and analysis, an extensive
knowledge of power system theory is required. This knowledge must then be
applied to specific situations to perform an accurate analysis of the data
attained during the field measurements. The following sections of this chapter
provide various power quality-related problems.

2.3.1 Transients
A transient is a disturbance in an AC wave, characterised by a discontinuity of
the waveform. These may be of either polarity and are additive or subtractive
from the nominal waveform. Transients can be classified into impulsive and
oscillatory transients. Impulsive transients are instantaneous voltage
deviations, which are often impervious to the RMS voltage due to their short
duration shown in Figure 2-2. These types of transients are often due to
lightning striking a power system. Oscillatory transients have positive and
negative polarity and are distinctive by their fast rise time and oscillatory
behaviour that decays over time are shown in Figure 2-3. Oscillatory
transients are most commonly a result of the load [8].

Figure 2-2: Impulsive transient [8]

 27

Figure 2-3: Oscillatory transient waveform [8]

Short duration voltage variations are a deviation from the nominal voltage
supply in the form of a ‘rise’, ‘dip’ or ‘supply void’. They can be classified in [9]
as sags, swells and interruptions. A sag is a decrease between 0.1pu and
0.9pu in the RMS voltage or current at the power frequency with durations
from 10msec (0.5 cycles) to one minute. A voltage sag can result from a large
load being placed on the power system, such as starting up a large motor as
shown in Figure 2-4 [9].

A swell is an increase between 1.1pu and 1.8pu in the RMS voltage or current
at the power frequency with durations from 10msec (0.5 cycles) to one minute
[10]. The causes of swells can be the consequence of removing a large load
from a power system as shown in Figure 2-5.

Figure 2-4: Voltage sag [9]

 28

Figure 2-5: Voltage swell [9]

When the supply voltage or load current decreases to less than 0.1pu for a
period not exceeding one minute, the disturbance can be classified as an
interruption [11]. An interruption can be a consequence of a power system
fault, which requires some finite time for the circuit breakers to re-close
causing a supply interruption as shown in Figure 2-6.

Figure 2-6: An interruption due to a fault [8]

2.3.2 Long duration voltage variations

Long duration variations are defined similarly to the short durations, in that
they have sags, swells and interruptions. However, long durations are merely
 29

encountered for a longer period than the short duration variations. The long
durations are classified as undervoltages, overvoltages and sustained
interruptions.

An undervoltage is a voltage sag that is longer than one minute. Similarly an

overvoltage is a voltage swell that is longer than one minute in duration. A
sustained interruption (outage) is an interruption lasting more than one minute
and often requires some human repair to the system.

2.3.3 Power frequency variations

Power frequency variations are a deviation from the nominal supply frequency
(50Hz or 60 Hz). The supply frequency is a function of the rotational speed of
the generators used to produce the electrical energy. A frequency variation
occurs if a generator becomes un-synchronous with the power system,
causing an inconsistency that is manifested in the form of a variation. The
specified frequency variation in South Africa should be within the bounds of
± 2.5% Hz at all times for grid network and ± 5.0% at all times for islanded
networks [11].

2.3.4 Voltage dip

A voltage dip has become the main power quality focus area for utilities in the
past couple of years due to the costs associated with voltage dip induced to
consumer plant and operation. A voltage dip is a reduction in the RMS voltage
for a period of 20ms to 3s, of all or any of the phase voltages of a three-phase
or single-phase system supply [11]. The duration of the voltage dip is the time
measured from the moment the RMS voltage drops below 0.9pu of the
declared voltage to when the voltage rises above 0.9pu of the declared
voltage.

Voltage dips can be represented graphically in terms of time/duration and

magnitude on a graph called a voltage dip window shown in Figure 2-7.
 30

Declared voltage

R.M.S. voltage
Dip
threshold
Dip
magnitude
(% of declared)

Duration (ms)

Figure 2-7: Measured voltage dip parameters [11]

2.3.5 Voltage unbalance

In a similar manner as for harmonics level, unbalance limits are also being lain
down by utilities. As this is done on a per consumer basis, specific guidelines
are necessary for doing this. The voltage unbalance is a potentially costly
phenomenon in the damage it can cause. The effect of voltage unbalance lies
in the damage to machines due to the unbalance currents that flow into the
machine. This could lead to overheating and eventual burn-out or tripping of
those machines. Known causes of voltage unbalance are non-linear loads.

A voltage unbalance in a three-phase system is characterised by differences

in the magnitudes and/or angles between the RMS, phase voltages. A
concept known as the voltage unbalance factor is used to determine the
extent of the unbalance, and is defined as, “The ratio of the negative or zero
phase sequence component of the voltage to the positive phase sequence
component of the voltage”. The IEEE 1159 standard states that the voltage
unbalance factor should remain less than 2% for the system to remain
balanced [5].

The positive sequence set of three vectors are all equal in magnitude and 120
degrees away from each other, as shown in Figure 2-10. While the negative
sequence set of three vectors in figure 2-10 are all equal in magnitude and
120 degrees away from each other, their phase order is reversed of the
positive sequence phase order, and the zero sequence set of three vectors
are all equal in magnitude and phase.

The voltage unbalanced factor can be expressed by [11]:

 31

Vneg
UB = × 100 2-3
V pos
where,

Vneg: negative sequence voltage

Vpos: positive sequence voltage

Figure 2-8: Balanced and unbalanced phase voltages

The compatibility level for unbalance on three phase networks is 2%, but if
there is a large number of single-phase or two-phase loads, the assessed
unbalance increases to 3% [11].

2.3.6 Waveform distortion

Waveform distortion in [12] is defined as “steady state deviation from an ideal
sinusoid”.

The distortion of waveforms can be classed into five separate categories:

• Harmonic

• Inter-harmonic

• Notching

• Noise

• Flicker
 32

The harmonic components of a waveform are sinusoidal in shape, with a

frequency equal to an integer multiple of the fundamental frequency, e.g. the
3rd harmonic has a frequency of 150Hz (3 x 50Hz).

Figure 2-9: Harmonic distortion

The amount of harmonic distortion can be measured by means of a factor

known as the total harmonic distortion (THD).

h max

∑
2
Mh
h =2
THD
M1
2-4

where M1 represents the fundamental components and h represents harmonic

number. Total harmonic distortion is the term used to describe the net
deviation of a non-linear waveform from ideal sine wave characteristics. THD
is the ratio between the RMS value of the harmonics and the RMS value of
the fundamental.

2.3.7 Fundamental of symmetrical components [13]

Symmetrical components allow unbalanced phase quantities such as currents

and voltages to be replaced by three separate balanced symmetrical
components. In the three-phase system, the phase sequence is defined as
the order in which it passes through a positive maximum.
 33

Figure 2-10: Representation of symmetrical components

The direction of the rotation of the phasors is taken to be counter-clockwise.

The three phasors are written as:

I a = I a ∠0° = I a
1 1 1

I b = I a ∠240° = a 2 I a
1 1 1
2-5
I c = I a ∠120° = aI a
1 1 1

where an operator a causes a counter-clockwise rotation of 120 ˚, such that

a = 1∠120° = −0.5 + j 0.866

a 2 = 1∠240° = −0.5 − j 0.866 2-6
a 3 = 1∠360° = 1 + j 0
From equation 2-6, it is clear that

1+ a + a2 = 0 2-7

the order of the phasors is abc. This is designated the positive phase
sequence. When the order is acb, as shown in Figure 2-10, it is designated
the negative phase sequence. The negative phase sequence quantities are
represented as:
I a = I a ∠ 0° = I a
2 2 2

I b = I a ∠240° = a I a
2 2 2
2-8
I c = I a ∠120° = a I a
2 2 2 2

The last set of balanced phasors, the zero phase sequence, are found to be in
phase with each other. Zero phase sequence currents, as shown in Figure 10-
2, would be designated as:
 34

Ia = Ib = Ic
0 0 0
2-9

The superscripts 1, 2 and 0 are used to represent positive, negative and zero-
sequence quantities respectively. However, 0, + and – can also be used
instead of 1, 2 and 0. Three-phase unbalanced phasors of a three-phase
system can be resolved into three balanced system of phasors.

Given the three-phase unbalanced currents Ia, Ib and Ic. we seek to find
symmetrical components of the currents such that:
Ia = Ia + Ia + Ia
0 1 2

Ib = Ib + Ib + Ib
0 1 2
2-10
Ic = Ic + Ic + Ic
0 1 2

According to the definition of the symmetrical components given by 2-5, 2-8,

and 2-9, equation 2-10 can be rewritten in terms of phase a components.

Ia = Ia + Ia + Ia
0 1 2

I b = I a + a 2 I a + aI a
0 1 2
2-11
I c = I a + aI a + a 2 I a
0 1 2

or

1 ⎤⎡I a ⎤
0
⎡ I a ⎤ ⎡1 1
⎢ I ⎥ = ⎢1 a 2 ⎢ 1⎥
⎢ b⎥ ⎢ a ⎥⎥ ⎢ I a ⎥ 2-12
⎢⎣ I c ⎥⎦ ⎢⎣1 a a 2 ⎥⎦ ⎢⎢ I a 2 ⎥⎥
⎣ ⎦

Equation 2-12 in matrix form:

abc 012
I = A∗ Ia 2-13

A is called the symmetrical component transformation matrix, which

abc 012
transforms phasor currents I into component currents I a and is:

⎡1 1 1⎤
A = ⎢⎢1 a 2 a ⎥⎥ 2-14
⎢⎣1 a a 2 ⎥⎦
 35

Solving equation 2-13 for symmetrical components of currents, then:

012 −1 abc
Aa =A I 2-15

The inverse of A is given by:

⎡1 1 1⎤
a 2 ⎥⎥
−1 1⎢
A = ⎢1 a 2-16
3
⎢⎣1 a 2 a ⎥⎦

From equation 2-14 and 2-16,

−1 1 *
A = A 2-17
3
−1
Substituting for A in equation 2-15 then:

⎡I a 0 ⎤ ⎡1 1 1 ⎤ ⎡I a ⎤
⎢ 1⎥ 1⎢
⎢ I a ⎥ = ⎢1 a a 2 ⎥⎥ ⎢⎢ I b ⎥⎥ 2-18
⎢ 2 ⎥ 3 ⎢1 a 2 a ⎥⎦ ⎢⎣ I c ⎥⎦
⎣⎢ I a ⎦⎥ ⎣

In component form, the symmetrical components are:

1
I a = (I a + I b + I c )
0

3
1
I a = ( I a + aI b + a 2 I c )
1
2-19
3
1
I a = ( I a + a 2 I b + aI c )
2

It can be seen from equation 2-19 that the zero–sequence current is equal to
one-third of the sum of the phase currents. When the phase currents add up
to zero, as in a three-phase system with an ungrounded neutral, the zero-
sequence current cannot exist. But if the neutral of the system is grounded,
zero-sequence current flows between the neutral and the ground. Similarly
expressions for voltages can be derived. The unbalanced phase voltages in
terms of symmetrical component voltages are:
 36

Va = Va + Va + Va
0 1 2

Vb = Va + a 2Va + aVa
0 1 2
2-20
Vc = Va + aVa + a 2Va
0 1 2

In matrix form:

= AVa
abc 012
Va 2-21

The symmetrical components in terms of the unbalanced voltages are:

1
Va = (Va + Vb + Vc )
0

3
1
Va = (Va + aVb + a 2Vc )
1
2-22
3
1
Va = (Va + a 2Vb + aVc )
2

In matrix form:

012 −1 abc
Va =A V 2-23
Unbalanced conditions can be resolved into positive, negative and zero
sequence components. The positive sequence component is similar to the
normal balanced load. The equations in this chapter will be used to assess the
concern regarding the current and voltage unbalance factor at SAC in chapter
4.

2.4 Summary
As the use of power electronic devices increases, there will be a need to
understand the effects of harmonics and the application of all the mitigation
methods. Harmonics is not a transient phenomenon, and its presence can
easily be measured and identified. In some cases, harmonics can be
tolerated, but in other cases it should be minimised or eliminated.

A power system during normal operation is in a steady-state condition even

though some voltage fluctuation is present as a result of facility switching
operations. The voltage remains within a tolerance that would normally be
expected.
 37

Disturbances are perhaps as damaging to electrical equipment, but they can

cause a variety of problems in the operation of an electrical appliance. These
disturbances can be source generated or generated within a facility due to
loads. A power system simulation tool to simulate power quality disturbances
and harmonics in the network is required. The components model method has
proved to be well accepted for conducting harmonic studies. This method will
be used in the next chapter to model the SAC network.
 38

3 POWER SYSTEM COMPUTER AIDED DESIGN

MODEL (PSCAD)

3.1 Introduction

Simulation results are obtained by using the Power System Computer-aided

Design, which is also known as the PSCAD engineering software (see
Annexure A). With the introduction of various simulation tools, simulation has
proved to be an efficient tool in identifying power quality disturbances.
Simulation results could enable engineers to identify the type of power
disturbances occurring in power system and be used to analyse a system for
future expansion. With the data collected from the simulation results,
engineers could easily pinpoint the various types of possible problems that
could occur when considering changes in the network and disturbances that
are affecting the power system, and have a clearer understanding of how
such events take place. Thus, certain appropriate measures or methods could
then be enforced/ implemented to prevent further disturbances on the system
and be used to perform some estimation when considering expansion.

The modelling of power systems involves the incorporation of device models

into a power system model. The development of accurate system models for
harmonic studies involves the selection of the devices to be included in the
model as well as the selection of device models, to achieve a balance
between complexity and accuracy for the study in question.

SAC requires a power system simulation tool to simulate the power network
before attempting expansion on the network so that possible future expansion
does not cause problems such as cable and transformer failure due to
overloading. Well–accepted component models will be used to simulate SAC
power network.
 39

3.2 Modelling of power system components

Harmonics is a high frequency phenomenon involving frequency from
fundamental 50/60 Hz to about 3000Hz. For harmonic studies, it is important
to take into account the component response characteristics at the harmonic
frequency range.

1. Source: A three-phase voltage source with a grounded neutral will be

used to model the synchronous generator for this study.

2. Underground cable: Cables can be modelled with a pi-circuit/network

using the actual information or data of the specific underground cable
to make the model more accurate. One aspect to be aware of is that
the transmission system could be a major source of harmonics for the
distribution system.

3. Transformers: A transformer is affected by a harmonic flow through its

series impedance, winding connection, and magnetizing branch. It
could be modelled by using its short circuit impedance for most
harmonic studies. It is also important to include the transformer phase-
shift effect. This phase-shifting (Δ-Y) 30° effect could lead to significant
harmonic cancellations in the system. The magnetizing branch of a
transformer is a harmonic source. Inclusion of the saturation
characteristic of the branch is important only if the harmonics
generated by a transformer are of primary concern [14].

4. Loads/power electronic devices: Power electronic devices can be a

load or a compensator. In terms of harmonic analysis, they are
harmonic sources. The most commonly acceptable way to model them
is to use the current source model. For this study, loads will be
modelled as current sources. Load modelling is difficult, as it is not
possible to identify exactly what the load is at any given point. The load
model is therefore somewhat empirical in nature.

There are different methods to analyse the harmonic propagation in an

electric network. The methods for harmonic analysis can generally be in the
frequency and time domains. Frequency domain methods include frequency
 40

the scan, harmonic penetration and harmonic load flow. On the other hand,
the time domain analysis methods are based on the calculation of the time
evolution of the system magnitudes that are generally done through simulation
of the operation. This chapter is dedicated to the time domain analysis of the
system under study.

3.3 Electrical diagram of SAC and Hart RAO

Figure 3-1: Network diagram of SAC and Hart RAO

As discussed in chapter 1, Eskom supply power to SAC and Hart RAO But
each network has its own converter or motor-generator set.
 41

Figure 3-2: Simplified electrical model of SAC

The simplified electrical configuration of the facility under study is shown in

Figure 3-2. For this study, the three-phase model was chosen to simulate
SAC because significant unbalanced loading is present and wye-wye and/or
delta-wye transformers are used. The SAC configuration will be investigated
separately in this study because of the higher level of distortion. Due to the
previous measurements taken of SAC and Hart RAO by the author, signs of
harmonics appeared on measurements taken of the SAC network and Hart
RAO which had a good power quality. Nameplate ratings of the transformers,
generator and cable data sheet (see Annexure C) from the manufacturer
(Aberdare) were used on the software package to model the SAC network.

3.4 Time domain simulation

Time domain methods for the analysis of electrical systems are generally
more accurate than other frequency domain techniques such as power flow.
On the other hand, they are more complex to implement. Some special
conditions exist in power systems that cannot be represented accurately by
the power flow technique requiring the use of time domain simulation.

These conditions can be summarised as:

 42

1. The distortion produced by the devices can vary according to the

conditions under which they are working.

2. Interactions between different harmonic orders cannot be represented

by time invariant models as the ones used for harmonic power flow.

3. When power electronic conversion is used, the presence of distortion in

the system magnitudes can influence the control of the gating signals.

Moreover, the time domain simulation allows the simultaneous existence of all
the electrical phenomena existing in the actual system operations like the
transients produced by the switching of the power electronic devices that
constitute the energy converters. In addition to a proper modelling of the
different systems components, some other consideration should be made in
order to obtain accurate results from a time domain simulation of the system
behaviour. These considerations include the proper calculation of the
initialisation process and magnitude of the different variables.

3.5 Validation of the simulation results

There are several items related to a study, which have to be set properly in
order to obtain results that match the reality. These items are generally:

1. The models of the different elements present in the system.

2. The simulations must be done by following the same operating

procedures as the real system.

Once the results are obtained from the computer simulation, it is also
important to compare the values against the measurements in the real field. In
the case of this study, the values and waveforms obtained will be matched
with field measurements. Although these studies were intended for different
purposes, the results and waveforms they provide allow for a comparison to
show the accuracy of the software.
 43

3.6 Time domain simulation results to show the accuracy

of the PSCAD model vs real measurements
In this section, the accuracy of the model to be used to model the real system
will be evaluated by using waveforms of the model and field measurements.
The results will then show whether this model built on the PSCAD can be
used to simulate the SAC network and investigate transmission efficiency
when the load increases. Once the results are obtained from the PSCAD
simulation, it is also important to compare the values against the
measurements in the real field. In this study, the waveforms obtained will be
matched with the measurements taken in the field.

It is difficult to match the exact conditions of the field measurements in the

simulation due to the unknown load powers, the unbalance currents in the
network and the variable nature of the load. The motor-generator set makes it
more difficult to simulate the network accurately, so a three-phase voltage
source was used instead of a motor-generator set. The power harmonic
analyser gives current and voltage measurements with the magnitude and
phase angle at a different harmonic frequency from the fundamental up to the
31st harmonic. This current data, taken from power the harmonic analyzer,
was used to model the current source as shown in Figure 3-3.

Figure 3-3: Comparison of model and field measurements at the generator

 44

Discussion of the model and real measurement:

• Waveforms generated by using the PSCAD model and field

measurement

• Table with power

The comparison of the results is on the following basis:

Comparison of waveforms: The waveforms from real measurement and the

PSCAD model.

The comparison is subjective:

The basis of the comparison is based on two properties as follows:

• Curve shape – comparison of the general shape of the waveforms, i.e.

similar or different.

• Power measurement – comparison of power.

The comparison of the wave shapes was difficult to do, because of the phase
rotation. The SAC network uses colour codes blue, red and yellow, and the
PSCAD software uses A, B and C. It was difficult to match the phases, so
waveforms with a similar shape/pattern from model and field measurements
were chosen. The measuring equipment measured single phase. The phase
rotation can be measured by determining the phase angle between the
measured waveforms. It can be positive or negative phase rotation, taking VA
as the reference phase. The positive phase rotation (abc) is when VA leads VB
by 120 º and VB leads VC by 120 º. The negative phase rotation (acb) is when
VA leads VC by 120 º and VC leads VB by 120 º.

The line currents of the model and field measurements are shown in Figure
3-4.
 45

Figure 3-4: Line currents of model and field measurements taken at the
SAC generator

This is excellent agreement between the measured and modeled. It was

difficult to match the model with exact conditions of field measurements due to
unknown load powers, unknown losses in the transmission system and
variable nature of the loads. The line current results of the model were smaller
than real measurements in magnitude. The waveforms in Figure 3-4 were
obtained after increasing the magnitude of the harmonic current source. The
wave-shapes, looks similar, as shown Figure 3-4.

Figure 3-5: Normalised harmonic current of model and field measurements

Figure 3-5 shows that the magnitude of the 5th harmonic current is very high
for both model and field measurements. This will be evaluated in the next
chapter with reference to the NRS 048 standard.
 46

Table 3-1: Comparison of model and real measurements in terms of power

measurement
Description Model Description Real Units
measurement
Active power (P) 146.224 Fundamental Active Power(P) 134.729 kW
Reactive power (Q) -25.977 Fundamental Reactive Power (Q) 53.248 Kvar

Apparent Power (S) 148.5 Fundamental Apparent Power(S) 144.87 kVA

The IEEE power definitions discussed in chapter 4 were used to analyse the
real measurements to ensure some uncertainty on the measured results and
the results, were similar. The PSCAD software has a power component (P
and Q) in the master library. The real/reactive power meter measures
instantaneous real power by multiplying the instantaneous voltage and current
of each phase and then adding them together. The reactive power is
calculated on the assumption that all three phases are balanced. The software
does not specify whether the calculated power is fundamental and harmonic
power.

Simulations 1: SAC generator

Figure 3-6: SAC network model simulation 1

 47

Figure 3-6 shows the point where current and voltage simulation results were
taken from, just before the step-up transformer. The cable impedance is
ignored because the length is short ( ± 6m).

The line-to-line voltages and line currents of simulation 1 at the SAC

generator are shown below in Figure 3-7.

Figure 3-7: Line currents and line-to-line voltages of simulation 1 at the SAC
generator respectively

The voltages are balanced and 120 º away from each other, but the currents
are unbalanced as shown in Figure 3-7. The line-to-line voltage is a pure
sinusoidal waveform, which means that any load connected to the same
source with the SAC that generates the harmonic current cannot be affected
by this harmonic, because the current waveform at the source depends on the
line-to-line voltage of the generator.

The Hart RAO network cannot be affected, because two networks have two
separate motor-generator set, which means that there is mechanical coupling
between the two networks and the ESKOM supply. One of the major concerns
is whether the load can be increased. This concern will be evaluated by
increasing the load on the model and calculating the transmission efficiency of
the model.
 48

Simulations 2 results after increasing the load

Figure 3-8: SAC network model simulation 2 after increasing the load

The line-to-line voltages and line currents of simulation run 2 at the SAC
generator after increasing the load are shown in Figure 3-9.

Figure 3-9: Line-to-line voltages and line currents at the Sac generator after
increasing the load respectively

It shows that when the load power increases, the distortion level of the current
waveform at the generator increases, but the voltage waveform is still purelly
sinusoidal.
 49

Simulation 3 of the SAC load

Figure 3-10: SAC Network model simulation 3 at the Load

The line-to-ground voltages and line currents of simulation 3 at the SAC load
208 V E bus supplying two UPS of 120 KVA each are shown below in Figure
3-11.

Figure 3-11: The line currents and line-to-neutral voltages of simulation 3 of

at the SAC load respectively

The line-to-neutral voltages are 120V and 120 degrees away from each other
and well balanced, as shown in Figure 3-11. Figure 3-11 shows the non-linear
current characteristics of the UPS. These power electronic loads draw power
only during portions of the applied voltage waveform.
 50

3.7 Transmission losses

The transformer model as a ideal transformer with the default values of the
PSCAD and cable parameters in Figure 3-8 were used to determine the
transmission efficiency. Power into the step-up transformer and out of the
step-down transformer was calculated by the PSCAD software to determine
the transmission efficiency, and the results are provided in Table 3-2.

Table 3-2: Comparison of transmission efficiency of the model and real

system
Description Model Description Real Units
Active power (P) in 146.224 Fundamental (P) in 134.729 kW
Reactive power (Q) in -25.977 Fundamental (Q) in 53.248 Kvar
Apparent power (S) in 148.5 Fundamental (S) in 144.87 kVA
Active power (P) out 144.61 Fundamental (P) out 131.13 kW
Reactive power (Q) out 18.277 Fundamental (Q) out 48.73 Kvar
Apparent power (S) out 145.75 Fundamental (S) out 140.062 kVA
Efficiency (η ) 98.1 96.68 %

The transmission efficiency can be calculated with the formula:

Pout
η= × 100 3-1
Pin
There is a difference between the real and model transmission efficiency. The
difference might be that there are transmission losses on the real system, and
losses on the transformers and the underground cable. The PSCAD software
uses a default value for losses in the transformer, which could differ from the
actual losses on the transformer. The other possibility for this difference could
be the way in which PSCAD calculates active and reactive power and the
IEEE definitions used to calculate power in the real system.

The transformer model as a non-ideal transformer and cable parameters in

Figure 3-10 were used to determine the transmission efficiency. Power into
the step-up transformer and out of the step-down transformer was calculated
by the PSCAD software to determine the transmission efficiency, and the
results are provided in Table 3-3.
 51

Table 3-3: Transmission efficiency using non-ideal transformer

Power Power in Power out Units

Active power (P) 146.224 142.291 KW
Reactive power (Q) -25.977 17.969 KVAR
Apparent P\power (S) 148.5 143.42 KVA
Efficiency ( η ) 96.6 %

From two simulation for cases where the transformer is treated as ideal and
non-ideal the losses was about 2%

3.7.1 The effect on transmission efficiency when the

load increases

Figure 3-12: Line-to-line voltage at the generator with two times the original
load

Figure 3-12 shows that the voltage waveforms are 120 degrees away from
each other, pure sinusoidal, and balanced, which means that doubling the
load has no effect on the line-to-line voltage. The transformer model in Figure
3-10 as an ideal transformer is used to determine the transmission efficiency
when the original load is doubled. Power into the step-up transformer and out
of the step-down transformer was calculated by the PSCAD software to
determine the transmission efficiency, and the results are provided in Table
3-4.
 52

Table 3-4: Transmission efficiency using the ideal transformer double the
original load

Power Power in Power out Units

Active power (P) 146.224 91.9943 KW
Reactive power (Q) -25.977 33.7886 KVAR
Apparent power (S) 148.5 98.003 KVA
Efficiency (η ) 65.9 %

The transmission efficiency is 65.9 % which means that doubling the original
load has an effect on the transmission efficiency of the system. It shows that if
the non-linear loads on the system increase, the losses on the cable will also
increase, which could cause overheating in the transformers and the cables in
the network. The transformer model in Figure 3-10 used as non-ideal
transformer to determine the transmission efficiency when the original load is
doubled. Power into the step-up transformer and out of the step-down
transformer was calculated by the PSCAD software to determine the
transmission efficiency, and the results are provided in Table 3-5.

Table 3-5: Transmission efficiency using the non-ideal transformer double

the original load

Power Power in Power out Units

Active power (P) 146.224 90.005 KW
Reactive power (Q) -25.977 33.2166 KVAR
Apparent power (S) 148.5 95.94 KVA
Efficiency (η ) 64.6 %

The transmission efficiency is 64.6% which means that doubling the original
load has an effect on the transmission efficiency of the system. It shows that if
the non-linear loads on the system increase, the losses on the cable will also
increase, which could cause overheating in the transformers and the cables in
the network.
 53

3.7.2 The effect on transmission efficiency if the load

is three times the original load

Figure 3-13: Line-to-line voltage at generator, three times the original load

The voltage waveforms in Figure 3-13 are still the same as the voltage on the
original load 120 degrees away from each other, pure sinusoidal and
balanced.

Table 3-6: Transmission efficiency using the ideal transformer, three times
the original load

Power Power in Power out Units

Active power (P) 146.224 50.30 KW
Reactive power (Q) -25.977 51.42 KVAR
Apparent power (S) 148.5 71.93 KVA
Efficiency (η ) 48.4 %

Table 3-7: Transmission efficiency using the non-ideal transformer, three

times the original load

Power Power in Power out Units

Active power (P) 146.224 51.91 KW
Reactive power (Q) -25.977 50.21 KVAR
Apparent power (S) 148.5 72.21 KVA
Efficiency (η ) 48.6 %
 54

Tripling the original load has a major impact on the transmission efficiency of
the model. It shows that increasing non-linear loads in the network could
cause serious problems.

3.8 Summary
It is difficult to match the exact conditions of the field measurements in the
simulation due to the unknown load powers, the unknown power losses on the
underground cable and transformers, the unbalanced currents in the network
and the variable nature of the load. The transmission efficiency of the model is
reliable, but it was shown in this chapter that as the load increases, the
transmission efficiency decreases, which means that a significant amount of
heating is generated on the transformer cores and cables on the network.
Increasing the non-linear loads in the power network has a catastrophic
impact on the current waveforms and transmission efficiency of the system.

It has been shown that as the size of the non-linear loads increases, the
current waveform distortion increases, and the transmission efficiency
decreases. Thus, additional heat is generated by the losses caused by the
harmonic content of the load current. The lifespan of a transformer and cables
the will be reduced if they operate above the rated temperatures.

If the real system expands, there could be some problems such as

transformer and cable failure due to heat. These power system components
need to be de-rated to compensate for load growth. Harmonic level studies of
the existing system needs to be conducted, which will be done the chapter 4,
before making changes to the network. Furthermore, information about the
harmonic content could be used to develop a method to de-rate the
transformers and cables in the network and be given to the manufacturer so
that the designer knows the harmonic conditions to which the new equipment
would be subjected.
 55

4 CURRENT AND VOLTAGE MEASUREMENT

4.1 Power definitions in the presence of harmonics

4.1.1 Introduction
The work of the IEEE Working Group on non-sinusoidal situations presents a
set of practical definitions for distortion power in terms of the total,
fundamental and harmonic constituents [15]. These definitions will be
analysed numerically in the measurements to ensure adequate uncertainty of
the measurements that are made. Accurate measurements of electrical power
are becoming more important, as electrical power becomes a more essential
and valuable commodity. Many factors make this measurement more difficult
and error prone. Power is becoming dirtier and more distorted, and
measurements beyond straight power and energy are becoming
commonplace.

“Because of the diversity of the approaches used in the different power

theories, different aspects of the power measurements are emphasised. Not
all the definitions have equal utility in a given application, and different
theories are employed under different circumstances for different applications.

It is true that the different definitions suit different applications better than
others and that the diversity is sometimes advantageous. It would, however,
be a great advantage if a single theory could be found that would be equally
suitable for all types of analyses under all circumstances “[16].
 56

4.1.2 Single-phase powers

The RMS value of a general complex time-dependent periodic waveform can
be expressed for the voltage as:

∑U U ∑U
∗ 2
U= n n = n 4-1
n n

Similarly, the effective value of the current is:

∑I ∑I
2
I= n In∗ = n 4-2
n n

If the fundamental components U1 and I1 are separated from the harmonic

components UH and IH then

U2 = U12 + UH2 4-3

And

I2 = I12 + IH2 4-4

Are obtained, where

∑U
2
UH = n 4-5
n ≠1

and

∑I
2
IH = n 4-6
n ≠1

From equation 4-5 and 4-6, the apparent power S can be calculated,

S = ∑|Un|2|In|2 4-7
n

and the total non-active power N as:

N = S2-P2 4-8

Alternatively, S can be defined as:

S2 = (U1I1)2 + (U1IH)2 + (UHI1)2 + (UHIH)2 4-9

Apparent power has two components:

 57

S2 = S12 + SN2 4-10

in which S1 is the fundamental apparent power and SN the non-fundamental

apparent power:

In turn:

S12 = (U1I1)2

= (U1I1cosθ1)2+ (U1I1sinθ1)2 = P12 + Q12 4-11

and

SN2 = (U1IH)2 + (UHI1)2 + (UHIH)2 4-12

The term (U1IH) is named the current distortion power, (UHI1) the voltage
distortion power and (UHIH) the harmonic apparent power. The last term could,
in turn, be broken down into the total harmonic active power PH and the total
harmonic non-active power NH.

SH2 = PH2 + NH2 4-13

in which NH is derived analogously to N in (4-8).

If (4-12) is divided by (4-11), the usefulness becomes clear:

⎛SN⎞2 ⎛IH⎞2 ⎛VH⎞2 ⎛SN⎞2

⎜ S ⎟ = ⎜ I ⎟ +⎜U ⎟ +⎜ S ⎟ 4-14
⎝ 1⎠ ⎝ 1 ⎠ ⎝ 1⎠ ⎝ 1 ⎠

to yield the normalised non-fundamental distortion power:

⎛SN⎞2 2 2 2
⎜ S ⎟ = (ITHD) + (UTHD) + (UTHD⋅ITHD) 4-15
⎝ 1⎠

The three right-hand terms of (4-15) represent the current THD, the voltage
THD and the product of the previous two respectively.
 58

4.2 Limitations of the production of harmonics on an

electrical network
The voltage and current waveforms that could exist throughout the system are
described, and the waveform distortion for the system is established. The
IEEE definition was developed to guide the quantities that should be
measured for the revenue purpose, specifically for the economic despatching
and the voltage profile determination in the system with harmonics. The
objective of the current limits is to limit the maximum individual frequency
harmonic voltage to 3% of the fundamental and the voltage THD to 5%, for
the 69kV and below.

This section discusses the NRS 048 on harmonic apportioning with reference
to the voltage and current limits that are set. A comparative evaluation of the
said standard will be done and the aim is to asses the harmonic content of the
SAC electrical network, to determine whether the harmonic level of the system
is not exceeding standards limits. Efforts are being made to regulate the levels
of harmonics injected into the power system through the introduction of
standard guidelines and recommended practices.

The purpose of the power quality standards relating to power system

harmonics is to:

1. to control the distortion of the power system current (limiting

harmonic current injection from individual customers) and voltage
waveforms to levels that the system and its associated components
can tolerate;

2. to provide customers connected to the power system with a voltage

supply waveform, that is suitable for their particular needs; and.

3. to ensure that the power system does not interfere with the
operating of other systems such as telephone networks.

4.3 South African Power Quality Standards

 59

4.3.1 NRS 048 part 2: minimum standards

Part 2 of the NRS 048 describes the voltage quality parameters that affect the
normal operation of the electricity-dependent processes of the customers. The
minimum standard in this part of the NRS 048 is intended to be applied as a
measure of power quality at the point to the end customer of electricity
utilities.

4.3.2 Voltage harmonics

The total harmonic distortion of the supply voltage, including all the harmonics
up to order 40, must not exceed 8%, for compatibility levels, for harmonic
voltages, for low (LV) and medium voltage (MV) power systems and the total
harmonic distortion for high voltage (HV) networks, is 3%.

Table 4-1: The maximum levels for harmonic voltages expressed as a

percentage of the declared voltage of the LV and MV power
systems
1 2 3 4 5 6
Odd harmonics non-multiple of 3 Odd harmonics multiple of 3 Even harmonics
Order n Harmonic voltage (%) Order h Harmonic Order h Harmonic
voltage % voltage %

5 6 3 5 2 2
7 5 9 1,5 4 1
11 3,5 15 0,3 6 0,5
13 3 21 0,2 8 0,5
17 2 >21 0,2 10 0,5
19 1,5 12 0,2
23 1,5 >12 0,2
25 1,5
>25 0,2 + 1,3 × 25/h
Total harmonic distortion (THD) ≤ 8 %
NOTE — For each harmonic, the harmonic voltage distortion compatibility level is given as a percentage of the
magnitude of the declared (fundamental frequency) voltage
 60

4.3.3 Recommended planning levels for harmonic

voltage on HV and EHV systems

Table 4-2: Indicative values of planning levels for harmonic voltage

(expressed as a percentage of the rated voltage of HV and EHV

power systems)

1 2 3 4 5 6
Odd harmonics non-multiple of 3 Odd harmonics multiple of 3 Even harmonics
Order Harmonic voltage Order Harmonic voltage Order Harmonic voltage
HV-EHV HV-EHV HV-EHV
H % H % H %
5 2 3 2 2 1,5
7 2 9 1 4 1
11 1,5 15 0,3 6 0,5
13 1,5 21 0,2 8 0,4
17 1 >21 0,2 10 0,4
19 1 12 0,2
23 0,7 >12 0,2
25 0,7
>25 0,2 +
0,5 × 25/h
Total harmonic distortion (THD) ≤ 3 % in HV networks

Part 2 of the NRS 048 describes the voltage quality parameters that affect the
normal operation of the electricity-dependent processes of customers. The
minimum standards in this part of the NRS 048 are intended to be applied as
a measure of the power quality at the point of supply to the end customers of
electricity utilities.

4.4 IEEE Standard 519

The maximum allowable voltage and current distortion is specified by IEEE
standard 519-1992. The limits on current distortion apply to the amount of
harmonic current to be supplied by the utility source to the harmonic load. For
most industrial power systems, the maximum acceptable voltage THD level is
5 %, with the maximum percentage of 3 % for individual harmonic
components [17].
 61

IEEE standard 519 imposes limits on the harmonics that a customer may
contribute to a network, and the harmonics that a utility may produce,
measured at the point of coupling to a customer. This ensures that individual
customers do not overly distort the electricity supply for other customers. The
electricity utility has also an obligation to supply clean power to all its
customers.

NRS 048 and IEEE 519-1992 determine the prescribed voltage current
distortion limits. IEEE 519 –1992 set limits for the harmonic of not only the
current caused by the operation of the load of individual customers, but also
the supply voltage. The NRS 048 sets harmonic limits for the supply voltage,
but does not provide current harmonic limits for the operating loads, as in
IEEE 519-1992. NRS 048 sets harmonic voltage limits only and leaves the
harmonic producers responsible for reducing their respective harmonic
currents so that the system voltage limits are not violated.

The aim of these measurements is to discuss NRS 048 and the IEEE
standards on harmonic apportioning with reference to the voltage limits that
are set as well as the apportioning procedure that is followed. An evaluation of
the said standards will be done using the current and voltage measured data
to check network compliance.

4.5 Field measurements

The instrument used to measure is the FLUKE 41B power harmonic
analyser), which is a special purpose instrument to measure power
harmonics. These instruments:

- display voltage, current and power waveforms;

- measure RMS, peak and total harmonic distortion of voltage and

current;

- measure harmonic power and indicate direction of flow; and

- calculate phase power, power factor and KVAR.

 62

The IEEE working Group definitions will be analysed numerically in the

measurements to ensure the adequate uncertainty measurements are made.
The calculations are done directly in the MathCAD package using the IEEE
definitions. The waveforms were also generated in the MathCAD program.
The measurements were taken during normal plant-operating conditions with
most of the equipment in the system operating.

4.6 Electrical network diagram

Figure 4-1: SAC and Hart RAO network diagram

SAC and Hart RAO receive supply from ESKOM, as shown in Figure 4-1 and
discussed and shown in chapter 3. The two networks are coupled to the same
bus on the ESKOM side, but each network has its own converter mechanism.
 63

Simplified electrical diagram model

Figure 4-2: SAC network model

The simplified model of the SAC is shown in Figure 4-2. The reason that the
SAC was the only network investigated in the dissertation was the signs of
harmonic problems had occurred on the previous measurements taken. Hart
RAO does not experience harmonic problems.

4.7 Measurement 1
Measurement 1 was taken on the secondary side of 11KV/400V 1MVA step-
down transformer, which is the ESKOM supply to SAC and Hart RAO, just
before the motor-generator set.

The line-to-neutral voltage and the line current measured in the ESKOM side
are shown in the waveforms in figure 4-3
 64

Figure 4-3: The line-to-neutral voltage and the line current measured on the
ESKOM side

There is no distortion and the voltage is purely sinusoidal, as would be

expected with such a relatively clean current waveform. Only internally
generated harmonics are considered, as shown in Figure 4-5. The ambient
harmonics on the supply from ESKOM are not a problem.

Table 4-3: IEEE measured data red phase

Description Measured Unit
RMS Voltage 230.13 V
RMS Current 229.86 A
Total Apparent Power 52.9 kVA
Total Non-active Power 0.82 kVA
Total Active Power 52.9 kW
Fundamental Apparent Power 52.8 kVA
Fundamental Active Power 52.8 kW
Fundamental Reactive Power 0.63 kVAr
Harmonic Apparent Power 36.9 kVA
Current Distortion Power 2.2 kVA
Voltage Distortion Power 868.94 VA
Normalised Harmonic App. Power 0.6 9
%
Current THD 4.2 %
Voltage THD 1.64 %
Total Power Factor 0.99 -

The most common measure of distortion is total harmonic distortion (THD).

The THD applies to both the current and the voltage, and is defined as the
RMS value of the harmonic divided by the RMS value of the fundamental, and
then multiplied by 100% THD of the current, which varies from a few per cent
to more than 100%. The THD of the voltage is usually less than 5% by IEEE
 65

standards. The THD below 5% is widely considered acceptable as for this

case in Table 4-3. This shows that the quality of the supply from ESKOM is
good.

Table 4-4: Comparison of NRS 048 and measurement on the ESKOM

supply to SAC and Hart RAO
RED PHASE

Odd harmonics non-multiple of 3 Odd harmonics multiple of 3 Even harmonics

Order (h) Harmonic voltage (%) Order (h) Harmonic voltage (%) Order (h) Harmonic voltage (%)

ESKOM NRS ESKOM NRS ESKOM NRS

5 0.23 6 3 0.23 5 2 0.05 2

7 0.31 5 9 0.02 1,5 4 0.05 1
11 0.9 3,5 15 0.01 0,3 6 0.03 0,5
13 0.55 3 21 0.03 0,2 8 0.08 0,5
17 0.24 2 10 0.07 0,5
19 0.16 1,5 12 0.03 0,2
23 0.44 1,5
25 0.2 1,5

Total harmonic distortion (THD) = 3.63 %, which is less than 8 %.

It shows that the individual harmonic does not exceed the limit of the specific
harmonic by NRS 048, but the THD voltage for all harmonics is also below the
maximum limit of NRS 048. The maximum voltage distortion prescribed by
NRS 048 part 4 is 8%. With reference to NRS 048, the voltage distortion is
4.37 % below the allowable limit. This shows that the supply from ESKOM to
SAC and Hart RAO is acceptable.
 66

4.8 Measurement 2

Measurement 2 was taken at the SAC generator, on the primary side of the
400V/11KV, 1MVA step-up transformer as shown in Figure 4-4.

Figure 4-4: SAC network diagram for measurement 2

The line-to-neutral voltage and the line current measured at the SAC
generator are shown in Figure 4-5 .

Figure 4-5: The line-to-neutral voltage and line current at the primary side of
the step-up transformer
 67

The apparent power is a quantity that is very useful in the predominantly

sinusoidal practice. However, it should be recognised that even in this
situation, none of the properties that are attributed to it are expressed
accurately. When the current and the voltage are distorted, different aspects
of the voltage and the current relationship become important, depending on
the power equipment and the problem. Figure 4-5 shows that the current
waveform is non-sinusoidal; this shows that there are non-linear loads on the
SAC network, propagating harmonics all over the network and back to the
power source.

It shows clearly that the type of load connected shows its impact on the
waveforms measured at the generator, by distorting the current waveform at
the source. It appears that SAC, which is generating harmonics, could
automatically be penalised by a higher apparent electricity consumption,
which might as well cause the supply some additional losses. It would be to
SAC’s own benefit to reduce the harmonic level to the largest possible extent
if they were coupled to the grid to avoid penalties.

Table 4-5: IEEE measured data blue phase

Description Measured Unit

RMS Voltage 276.94 V
RMS Current 132.97 A
Total Apparent Power 34.7 kVA
Total Non-active Power 11.45 kVA
Total Active Power 32.7 kW
Fundamental Apparent Power 35 kVA
Fundamental Active Power 32.8 kW
Fundamental Reactive Power 12 kVAr
Harmonic Apparent Power 684.3 VA
Current Distortion Power 11.2 kVA
Voltage Distortion Power 2.1 VA
Normalised Harmonic App. Power 1.96 %
Current THD 32.1 %
Voltage THD 9.09 %
Total Power Factor 0.8 -

The overall power factor of 0.8 in Table 4-5, compared to the fundamental
power factor of 0.93, is brought about by the relatively high level of harmonic
distortion. At a low power factor of operation for a given voltage and power
level, the current drawn by the equipment will be large, thus requiring an
increased volt-ampere rating of the utility equipment such as transformers,
 68

transmission lines and generators. Note that the voltage the THD has
increased. In measurement 1, the THD of the voltage is 1.64% and now it is
9.09%, which is unacceptable. A diverse range of problems could be expected
as a result of this, ranging from the overheating of transformers and cables, to
odd phenomena such as the overheating of lightning circuit breakers. This
could affect sensitive equipment and the load in the system.

Table 4-6: Comparison of NRS 048 and measurements at the SAC

generator
BLUE PHASE

Odd harmonics non-multiple of 3 Odd harmonics multiple of 3 Even harmonics

Order (h) Harmonic voltage Order (h) Harmonic voltage Order (h) Harmonic
(%) (%) voltage (%)
SAC NRS SAC NRS SAC NRS

5 5.04 6 3 2.51 5 2 0.15 2

7 1.15 5 9 0.87 1,5 4 0.04 1
11 0.73 3,5 15 0.03 0,3 6 0.1 0,5
13 0.36 3 21 0.05 0,2 8 0.07 0,5
17 0.11 2 10 0.05 0,5
19 0.02 1,5 12 0.02 0,2
23 0.02 1,5
25 0.01 1,5

Total harmonic distortion (THD) = 11.3 %, which is more than 8 %.

Table 4-6 shows that the individual harmonic does not exceed the limit of the
specific harmonics by NRS 048, but the THD voltage for all the harmonics is
above the allowable maximum limit of NRS 048. The maximum voltage
distortion prescribed by NRS 048 part 4 is 8% and by the IEEE it is 5% THD,
but an individual THD of 3% is prescribed for 69KV systems and below.

Table 4-6 shows that the 5th harmonic is 5.04%, which is not acceptable. With
reference to NRS 048 part 4, the voltage distortion is 3.3% above the
allowable limit, which is not acceptable.
 69

Table 4-7: Comparison of NRS 048 measurements at the SAC generator

WHITE PHASE

Odd harmonics non-multiple of 3 Odd harmonics multiple of 3 Even harmonics

Order Harmonic voltage (%) Order (h) Harmonic voltage (%) Order Harmonic voltage (%)
(h) (h)

SAC NRS SAC NRS SAC NRS

5 5.04 6 3 2.73 5 2 0.11 2

7 1.43 5 9 0.95 1,5 4 0.11 1
11 0.67 3,5 15 0.07 0,3 6 0.09 0,5
13 0.34 3 21 0.03 0,2 8 0.05 0,5
17 0.14 2 10 0.06 0,5
19 0.03 1,5 12 0.03 0,2
23 0.02 1,5
25 0.02 1,5

Total harmonic distortion (THD) = 11.62 %, which is more than 8 %.

Table 4-7 shows that the individual harmonic does not exceed the limit of the
specific harmonics by NRS 048, but the THD voltage for all harmonics is
above the allowable the maximum limit of NRS 048. The maximum voltage
distortion prescribed by NRS 048 part 4 is 8 % and by the IEEE it is 5 % THD,
but an individual THD of 3% is prescribed for 69KV systems and below. Table
4-7 shows that the 5th harmonic is 5.04%, which is not acceptable. With
reference to NRS 048 part 4, the voltage distortion is 3.62 % above the
allowable limit, which is not acceptable.

Table 4-8: Comparison of NRS 048 and measurements at the SAC generator
RED PHASE

Odd harmonics non-multiple of 3 Odd harmonics multiple of 3 Even harmonics

Order (h) Harmonic voltage (%) Order (h) Harmonic voltage (%) Order (h) Harmonic voltage
(%)
SAC NRS SAC NRS SAC NRS

5 4.64 6 3 2.9 5 2 0.03 2

7 1.39 5 9 0.77 1,5 4 0.03 1
11 0.96 3,5 15 0.04 0,3 6 0.08 0,5
13 0.34 3 21 0.04 0,2 8 0.05 0,5
17 0.1 2 10 0.02 0,5
19 0.03 1,5 12 0.06 0,2
23 0.02 1,5
25 0.01 1,5

Total harmonic distortion (THD) = 11.51%, which is more than 8 %.

 70

Table 4-8 shows that the individual harmonic does not exceed the limit of the
specific harmonic by NRS 048, but the THD voltage for all the harmonics is
above the allowable maximum limit of NRS 048. The maximum voltage
distortion prescribed by NRS 048 part 4 is 8 % and by the IEEE it is 5 % THD,
and the individual THD of 3% is prescribed for 69KV systems and below.
Table 4-8 shows that the 5th harmonic is 4.64% and the 3rd harmonic is 2.9 %,
which is not acceptable. With reference to NRS 048 part 4, the voltage
distortion is 3.51 % above the allowable limit, which is not acceptable.

4.9 Measurement 3
Measurement 3 was taken on the secondary side of the 11KV/380/208V,
500kVA step-down transformer shown in Figure 4-6.

Figure 4-6: SAC network diagram for measurement 3

The line-line voltage and line current at the primary side of the step-down
transformer are shown in Figure 4-7.
 71

Figure 4-7: The line-neutral voltage and line current at the SAC load on the
secondary side of the step-down transformer

Figure 4-7 shows the non-linear current characteristics of the UPS. These
power electronic loads draw power only during portions of the applied voltage
waveform. The voltage waveforms are not purely sinusoidal, as shown in
Figure 4-7. This is one of the observable effects that harmonics have on a
power system and impact on the quality of the AC voltage waveform
becoming distorted.

Table 4-9: IEEE measured data RED phase

Description Measured Unit

Total node voltage rms 121.58 V
Total node current Description rms 180.96 A
Total active power 17.52 kW
Total non-active power 21.591 kVA
Total apparent power 21.993 kVA
Fundamental apparent power 22 kVA
Fundamental active power 19.84 kW
Fundamental reactive power 9.03 kVAr
Harmonic apparent power 846.6331 VA
Current distortion power 8.9920 kVA
Voltage distortion power 1955.2 VA
Current THD 39.6 %
Voltage THD 9.41 %
Normalised Har. App. Power 3.726 %
Fundamental power factor 0.902 -
Total power factor 0.79 -
 72

The voltage THD is 9.41%, which is unacceptable. This high voltage distortion
can clearly be seen on the voltage waveform in Figure 4-7. The power factor
is low. This is brought about by the relatively high level of harmonic distortion.
This will affect nearby sensitive loads.

Table 4-10: Comparison of NRS 048 and the SAC at the LOAD side
UPS RED PHASE

Odd harmonics non-multiple of 3 Odd harmonics multiple of 3 Even harmonics

Order (h) Harmonic voltage (%) Order (h) Harmonic voltage Order (h) Harmonic voltage
(%) (%)

SAC NRS SAC NRS SAC NRS

5 9.32 6 3 0.73 5 2 0.12 2

7 0.38 5 9 0.17 1,5 4 0.08 1
11 0.81 3,5 15 0.06 0,3 6 0.06 0,5
13 0.1 3 21 0.03 0,2 8 0.06 0,5
17 0.21 2 10 0.07 0,5
19 0.07 1,5 12 0.04 0,2
23 0.15 1,5
25 0.06 1,5

Total harmonic distortion (THD) = 12.52%, which is higher than 8 %.

Table 4-10 shows that the individual harmonic does not exceed the limit of the
specific harmonic by NRS 048, except the 5th harmonic which is quite high,
and, the THD voltage for all the harmonics is above the allowable maximum
limit of NRS 048. The maximum voltage distortion prescribed by NRS 048 part
4 is 8 % and by the IEEE it is 5 % THD and the individual THD of 3% is
prescribed for 69KV systems and below.
Table 4-10 shows that the 5th harmonic is 9.32%, which is not acceptable.
With reference to NRS 048 part 4, the voltage total harmonic distortion is
4.25% above the allowable limit, which is not acceptable.

4.10 Voltage and current unbalanced

In this section the unbalanced amount on the motor-generator set will be
discussed. A machine operating under an unbalanced supply will draw a
current with a degree of unbalance several times that of the supply voltage.
The results of this on the three-phase currents could differ and a temperature
rise would occur in the machine. A three-phase balanced load produces a
reaction field, which is constant and rotates synchronously with the rotor field
 73

system, but any unbalanced condition could be resolved into positive,

negative and zero sequence components.

Based on the theory in [13], three-phase unbalanced phasors of a three-

phase system could be resolved into a three balanced system of phasors.
This set of equations in chapter 2 will be used in this study to evaluate the
voltage and current unbalance on the SAC network, and the results will be
compared to the IEEE assessment of the allowable percentage of unbalance
on the network and NRS 048.

Table 4-11: Current and voltage on the three-phase at the SAC generator
Fundamental phase currents (A) Fundamental Line-to-neutral Voltage (V)
Red 217.56 ∠ -20°A Red 278.07 ∠ 0°V Reference
Blue 195 ∠ 100°A Blue 277 ∠ 120°V
White 132.9 ∠ 220°A White 277.5 ∠ 240°V

The sequential components were calculated in MATHCAD using this current

and voltage in Table 4-11 and the results are presented in Table 4-12.

Table 4-12: Sequential currents and voltages

Sequential currents Sequential voltages
Zero 247,386 ∠ 72.1°A 335.243 ∠ 89.97°V
Positive 247.368 ∠ -112.1°A 335.235 ∠ -89.969 °V

Negative 57.639 ∠ 160 °A 0.353 ∠ 178.805 °V

Unbalance factor % 23.301 0.105

The voltage unbalance was calculated with the following equation:

Vneg
UB = × 100 4-16
V pos
where,

Vneg: negative sequence voltage

Vpos: positive sequence voltage

0.353
UB = = 0.105% 4-17
335.235
 74

The results show that the compatibility for unbalance on the SAC network is
allowable because it is less than 2%. According to NRS 048, if the network
has significant number of single-phase loads, the assessed unbalance factor
may increase to 3% [11]. The SAC network has a predominance of single-
phase loads, but the unbalance factor is still low. With reference to the IEEE,
the voltage unbalance is not a problem at the SAC, because the unbalance
factor is below the 2% benchmark for the system to be balanced.

Current unbalance

Unbalanced voltages will generally produce unbalanced currents that are

many times greater than the percentage of the voltage unbalance. The results
in Table 4-12 show that the current unbalance factor is 23.301% and the
voltage unbalance factor is acceptable. This is clear evidence that the voltage
unbalanced is not causing the unbalance currents. The only cause of this
problem is the unbalanced non-linear loads.

4.11 New power-quality assessment criteria for supply

system under unbalanced and non-sinusoidal
conditions [ 18]
These criteria were used to assess the supply system under unbalanced and
non-sinusoidal conditions in [18]. But for this study, the method developed in
[18] will be used for harmonic distortion evaluation of the SAC network.

Harmonic distortion evaluation

Each harmonic of the current drawn by a non-linear or time-varying load

depends on every harmonic component of the voltage. This voltage is also
determined by the network operative conditions. Then the current is affected
by network too. The existing index usually depends on the voltage and the
current, which are influenced by the conditions of the network, so they are too
variable and not suitable to characterise the nature of the load. On the basis
of this observation, a new current index to asses power quality in terms of
 75

harmonic distortion has been introduced in [18].It is less sensitive to the

network operative conditions than the current total harmonic distortion. The
meaning of this is that if the load is fixed and the network is responsible for the
voltage harmonic content variation, this new index value is almost constant.

The parameters R and L of the ideal and balanced load have been estimated
on one phase according to the following procedure in [18]. The same method
can also be used for the other two phases.

If Z 1 A is the impedance of the phase with the R-L series circuit at the
fundamental frequency, it results in:

V1 A
Z1A = 4-18
I1A
∠Z1 A = θ1 A − φ1 A = ϕ 4-19

where V1 A and I1 A are the fundamental root mean square (rms) values of the
phase A voltage and current, and θ1 A and φ1 A are the phase angles of the
same quantities, then:

R A = Z 1 A ⋅ cos ϕ1 A
4-20

X 1 A= Z 1 A ⋅ sin ϕ1 A 4-21
X 1A
LA = 4-22
2 ⋅π ⋅ f0
where X 1 A represents the ideal R-L series reactance at the fundamental
frequency, f 0 , RA and LA are the expected parameters. The values of the

reactance at the frequencies of the voltage harmonic components are:

X kA = 2 ⋅ π ⋅ k ⋅ f 0 L A k = 1,2,…,n. 4-23

If the skin effect is neglected and the value of R A is not influenced by

frequency, then:

Z kA = R 2 A + X 2 kA 4-24
⎛ X kA ⎞
ϕ kA = arctg ⎜⎜ ⎟⎟ k = 1,2,…,n 4-25
⎝ A ⎠
R
where n represents the order of the last significant voltage harmonic
component.
 76

The current that the ideal linear load on phase A should drain could expressed
as

n
2 ⋅ VkA
ilA (t ) = ∑ ⋅ sin( 2 ⋅ π ⋅ k ⋅ f 0 t + (θ kA − ϕ kA )) 4-26
k =1 Z1A
This is the calculated current [18].

The difference between the measured real current and the calculated current
can be defined as the non-linear current:

inlA (t ) = i A (t ) − ilA (t ) 4-27

The non-linear current expresses how much the real load on phase A differs
from the ideal one in terms of harmonic distortion. This current could contain
some harmonic components not present in the voltage. Therefore, a non-
linear current index that provides reliable information about the distortion
attributable to the customer can be defined as follows:

I nlA
⋅ 100% 4-28
IA
where I nlA and I A are RMS values

l
I nlA = ∑I
k =1
2
knlA 4-29

m
IA = ∑I
k =1
2
kA 4-30

and I knlA and I kA are the corresponding RMS values of the harmonic

components of inlA (t ) and i A (t ) , while l = m = n.

To investigate the effectiveness of the non-linear current index in this study

analysis carried out on a typical highly distorted current waveform measured
on the SAC network, the voltage and current on each phase have been
measured. But in this way it has been difficult to evaluate the performance of
the index of the non-linear current in this study, because the load is not fixed
and daily trends of current and voltage data are not available. The
investigation of this new index will be for one cycle or period, and the results
will be compared with the existing index in this study of current total harmonic
distortion. The comparison will be based on the behaviour of the two indexes
 77

regarding the network conditions. The aim of this study is to verify that the
current non-linear index as proved in [18] is a constant non-zero value when a
non-linear load is tested.

Table 4-13: Current non-linear index and current THD three-phases

Units
Description Red phase Blue phase White phase

%
THD I 26.29 33.23 31.9

Non-linear index %
47.57 47.98 47.87

Non-linear current 100.212 88.833 60.652 A

Calculated current 110.427 92.937 66.023 A

Voltage rms 277.509 277.737 276.532 V

Current rms 210.639 185.145 126.676 A
Fundamental P1 53.774 48.062 32.893 Kw

Fundamental Q1 22.919 18.283 12.046 Kvar

Fundamental S1 58.454 51.422 35.031 kVA

Table 4-13 shows that the non-linear current index is more invariant than the
current THD. It shows that the non-linear current index is almost the same on
all three phases measured, but the current THD varies. The current THD has
a smaller value on the phase with less distortion. In this way the evaluation of
harmonic distortion generated by the load is hardly influenced by the network
conditions. Moreover, it can be pointed out that the current non-linear index
has approximately the same value on the three phases, which means that the
same types of loads are connected. This could express the load’s ability to
affect distortion at the source.
 78

Figure 4-8: Daily trend of current non-linear index and current THD [18]

For this study it is not possible to show that the current non-linear is invariant
than the current THD, because daily trends of current measurement data are
not available. Figure 4-8 shows the daily trend of the total harmonic of the
current THD and the non-linear current index where the current THD varies
while the non-linear current index is almost invariant.

4.12 Summary
Individual harmonic does not exceed the limit of the harmonics specified by
NRS 048, except the 5th harmonic, which is high in all the measurements. But
the voltage THD in the system exceeds the maximum limit of NRS 048 by
± 3.6 %. SAC did not meet these requirements from the analyses in this
chapter.

There is excellent agreement between the measured and modelled results.

The results of the RMS value of the field measurement and the model are
similar. There is a difference in the power calculations for the model and field
measurements. The reason for this could be the way in which the PSCAD
calculates its power and the definition of fundamental power used to calculate
power in real measurements.

Voltage unbalance is not a problem at SAC, because the unbalance factor is

far less than 2% stated by the IEEE and NRS 048 for the system to be
 79

balanced. But the current unbalance is a problem caused by non-linear loads

connected and their most notable effects on a power system are the impact
on the quality of the AC voltage waveform.

The aim of the proposed index is to investigate the conformity of the

measured load to the calculated one without considering the effects of that
each harmonic could introduce into the network. It is important to know that in
the presence of distortion a single index cannot be used to carry the entire
information about the system condition.

The harmonic content study of SAC conducted in this chapter shows that the
system has a significant amount of harmonics. There is a need to establish a
mechanism or tool that could be used to manage the equipment operating
under this non-sinusoidal condition. The method of de-rating transformers and
cables operating under this condition will be discussed in the next chapter.
Harmonic data will be used to illustrate these de-rating methods. The de-
rating methods will work as the tool to control the power system equipment as
the non-liner loads increase in a harmonically rich network.
 80

5 DESIGN/MANAGEMENT TOOL FOR

MAINTANANCE AND FUTURE EXPANSION

5.1 Introduction
Methods of evaluating the performance of power system equipments
operating on harmonic load environments need to be developed. Because the
harmonic level of the network is high, and there is a possibility of expanding
these high distortion level network due to operations in the facility. There is
need to de-rate equipments such as transformers and cables. Transformers
and cables can operate at low loading, but as the network expands, the load
increases without intended de-rating. The design tool to cater for the changes
on the network needs to be developed and managed properly to avoid loss of
production due to cable or transformer failure. So before any expansion on the
existing system harmonic, studies need to be performed to analyse the
equipment capability under a harmonically loaded environment.

5.2 Effects of harmonics on a power system

There is interest in this study of harmonics because of the harmful effects they
have on power system devices and their operation. The effects of harmonics
are not known until failure occurs. In this section, the effects of harmonics on
some common power system devices are discussed.

5.3 Transmission system/cable

The flow of harmonic current in a network produces two main effects.

• One is the additional transmission loss caused by the increased RMS

value of the current waveform.

• The second effect of a harmonic current flow is the creation of

harmonic voltage drops across the various circuits’ impedance.
 81

In the case of transmission by cable, the harmonic voltage increases the

dielectric stress in proportion to the crest voltages. This effect shortens the
useful life of the cable. The effects of harmonics on corona starting and
extinction levels are functions of the peak-to-peak voltages. The peak voltage
depends on the phase relationship between the harmonic and the
fundamental. It is thus possible for the peak voltage to be above the rating
while the RMS voltage is well within this limit.

The flow of a normal 60-Hz current in a cable produce losses, and current
distortion introduces additional losses in the conductor. Also, the effective
resistance of the cable increases with the frequency due to skin effect, where
unequal flux linkages across the cross-section of the cable cause the AC
current to flow on the outer periphery of the conductor. Because of both the
fundamental and the harmonic currents that can flow in a conductor, it is
important to make sure that a cable is rated for the proper current flow.
Resistance of a conductor varies with frequency.

5.3.1 Heating and loss effects caused by harmonic

currents on underground cables
When significant harmonic currents are present in a system, additional I2R
losses occur in the phase or line conductors as well as the neutral conductor.
The presence of harmonic currents in the supply conductor affect the
ampacity of the supply because of the additional I2R losses. The
determination of ohmic losses is complicated by the fact that the resistance of
the cable depends on frequency. The resistance increases specifically with
the frequency due to the skin effect.

The effect of harmonic currents in the neutral conductor can be evaluated with
the same method as for the phase conductors. But the harmonic current
magnitudes could be different in the neutral conductors due to the fact that the
zero-sequence current is not cancelled, and the positive and negative
sequence harmonic currents are cancelled. Thus, the neutral becomes an
additional heat-generating conductor.
 82

The skin depth is dependent on the frequency and the properties of the
conductor.

This relationship is:

1
δ = 5-1
π ⋅ f ⋅ μ ⋅σ
where μ and σ are conductor permeability and conductivity. Copper as a
conductor with a frequency of 60 Hz, has a skin depth of:

1
δ =
(4 ⋅ π × 10 )⋅ 60 ⋅ π ⋅ (0.999991) ⋅ (5.8 × 10 )
−7 7

≈ 8mm

This means that if the conductor has a diameter of more than 16mm, most of
current flows in the outer rim, 8mm from the surface of the conductor. As the
frequency increases, so the skin depth decreases. For example, a seventh
harmonic that is operating at 420 Hz in a copper conductor has a skin depth of
3.2mm. For a third harmonic, which is operating at 300 Hz, the skin depth is
3.8mm.

5.3.2 Calculation of cable losses and ampacity in the

presence of harmonics
Harmonics in cables cause an increase in the ohmic losses and increased
operating temperatures. The results of the procedure for computing ohmic
losses in a cable system by extending the Neher-McGrath model for 60 Hz
losses are used to calculate the cable ampacity or the de-rating factor due to
the harmonics discussed in [19]. Simplified formulae are also given in [19] for
evaluating ohmic losses due to harmonics. The method is simple to follow and
can be performed with a calculator. The method developed in [19] will be used
in this study to analyse the cable model losses and the de-rating factor on the
underground cable at SAC.
 83

5.3.3 Model description

The ampacity of the cable is determined by the ohmic losses of the cable and
the ability of the environment to remove the heat generated by the ohmic
losses. Models to describe the heat transfer around the cable were previously
developed by Neher-McGrath, but these models neither reflect the presence
nor the absence of a harmonic current. Thus, the discussion of those models
is not necessary. The ohmic losses depend on the presence and magnitude of
the harmonic currents. In [19] the procedure and the equations for evaluating
ohmic losses in the presence of harmonics are discussed.

5.3.4 Calculation of ohmic losses

The calculation of cable ohmic losses that carry a distorted electric current is
performed by computing the loss at each harmonic and adding the results. In
order to systematise the procedure, the definition of harmonic signature is
introduced.

Harmonic current signature is defined as the sequence {IB, α1, α3, α5, α7,…},
where IB is the base RMS value (which can be selected arbitrarily), and αi is
the per unit value of the ith harmonic with respect to the base IB [19].

The method and the equations in [19] will be used in this study to compute the
cable de-rating factor and the ohmic losses on the underground cable at SAC.

Table 5-1: Cable parameters

3KM, 3 CORE, 11KV PAPER DRAIN

LEAD COPPER CABLE
1.Conductor diameter - 4.7mm
Area –16 mm2
2.Insulation thickness – 3.2 mm
Conductor resistance DC 1.15 ohms/km
AC resistance 1.37 ohms/km
Conductor inductance 0.435 ohms/km
Inductive reactance 0.137 ohms/km
Impedance/phase 1.3828 ohms/km
Capacitance 272 n F/km
Charging current 0.543 A/km
Current rating in ground 91 A
Current in air 88 A
 84

To illustrate this method, a three-phase measurement of distorted current

waveforms measured on the primary side of the step-up transformer was
taken on the SAC network, as shown in Figure 4-2. The three-phase system
carries the distorted electric currents as presented in Table 5-2, Table 5-3 and
Table 5-4

Table 5-2: Ohmic losses calculated for the RED phase

RED phase Harmonic Current A Skin effect AC Losses

signature factor resistance
ohms/km
Base -- 233.55 A -- -- --
Fundamental 1 233.55 A 0.008532 1.3761113 75.06 kW/m
3rd harmonic 0.0131 3.06 A 0.004926 1.3761248 12.89W/m
5th harmonic 0.2181 50.94 A 0.003815 1.3761588 3.57kW/m
7th harmonic 0.0573 13.38 A 0.003225 1.3761918 242.733W/m
I RMS 239.58 A
Ploss
TOTAL 78.89kW/m

Table 5-3: Ohmic losses calculated for the BLUE phase

BLUE phase Harmonic Current Skin AC Losses
signature A effect resistance
factor ohms/km
Base -- 226.56 -- -- --
Fundamental 1 226.56A 0.008532 1.3761211 70.64kW/m
3rd harmonic 0.0499 11.31 A 0.004926 1.37611248 176.03W/m
5th harmonic 0.2331 52.81 A 0.003815 1.3761588 3.84kW/m
7th harmonic 0.0596 13.5 A 0.003225 1.3761918 250.8W/m
I RMS 233.4 A
Ploss
TOTAL 74.91kW/m

Table 5-4: Ohmic losses calculated for the WHITE phase

WHITE phase Harmonic Current A Skin AC resistance Losses

signature effect ohms/km
Base -- 215.94 A -- -- --
Fundamental 1 215.94 A 0.008532 1.37611211 64.17kW/m
3rd harmonic 0.0321 6.94 A 0.004926 1.3761248 66.28W/m
5th harmonic 0.2437 52.63 A 0.003815 1.3761588 3.8kW/m
7th harmonic 0.0738 15.94 A 0.003225 1.3761918 349.7w/m
I RMS 222.95 A
Ploss
TOTALS 68.39kW/m

Steps: Determine the current flowing in each conductor. Determine the

harmonic signatures, which are the fundamental currents divided by harmonic
currents of each. Calculate the ohmic losses which are the fundamental
and/or harmonic currents squared and multiplied by the AC resistance for
 85

each harmonic frequency on each phase. Then add the ohmic losses for each
phase to obtain the total ohmic losses of the cable.

Total Ohmic Losses (Ploss) = 78.89kW/m + 74.91kW/m + 68.39kW/m 5-2

= 222.19kW/m

5.3.5 De-rating due to harmonics

De-rating factors due to the presence of harmonics can be defined only for a
specific harmonic signature. The de-rating factor is defined as follows:

Given a harmonic signature {IB, α1, α3, α5, α7…}, the de-rating factor, k is
defined as the number k such that an identical cable carrying an undistorted
current of I1 = α1IB/k has the same losses as the cable carrying a current with
the given harmonic signature [19].

The de-rating factor is:

IB
k = α1 × 5-3
I1
where α1: the fundamental harmonic signature

IB: base current

I1: fundamental current

Mathematically, this can be expressed as follows:

2
α n ⋅I B
∑
2 2
P loss IB ⋅ r ac ( n ) ⋅ α n ⋅ r ac ( 1)
2
( n⋅ ε⋅ N ) k
5-3

Thus, the de-rating factor is:

α 1 2 ⋅ I B 2 ⋅ rac (1)
k= 5-4
Ploss

Using the value of Ploss = 222.19kW/km, IB = 225.35 A, and rac (1) = 1.3761
ohms/km the de-rating factor can be calculated.

By substituting the above values in equation 5-5:

 86

α 1 2 ⋅ I B 2 ⋅ rac (1)
k= = 0.561
Ploss

k = 0.561 is the required de-rating factor of the cable using the resulting ohmic
losses calculated by using an extended Neher-McGrath Model for 60 Hz
losses. The underground cable capability with the given non-sinusoidal load
current harmonics composition is approximately 56% of its sinusoidal load
current capability. It shown in chapter 3 that as size of non-linear loads
increase the transmission efficiency decreases, which means that there are
additional I2R losses and this affect the ampacity of the underground cable.
This effect shortens the useful life of the cable.

5.4 Transformers
The primary effect of power system harmonics on transformers is the
additional heat that is generated by the losses caused by the harmonic
content of the load current. The lifespan of a transformer will be reduced if it
operates above the rated temperatures. Other problems include possible
resonance between the transformer inductances and system capacitances,
mechanical insulation stresses (winding and lamination) due to temperature
cycling, and possible small core vibrations.

The primary loss components are I2R losses, the winding eddy-current and
stay losses from electromagnetic flux in areas such as windings, the core and
tanks. The losses due to I2R components will be due to conductor heating and
the skin effect. Losses from the winding eddy-current will increase with the
square of the load current and the square of the frequency. Other stray losses
will also increase with frequency although at a power slightly less than two.

The presence of harmonic currents in an electric circuit increases the losses

in the equipment where they circulate. The losses in most of the equipment
depend in a certain way on the frequency of the circulating current. Generally ,
the higher the frequency, the larger the losses for the same current
magnitude. Therefore, the increasing frequency of the harmonic components
 87

will contribute in an appreciable way to the number of losses. This is true in

transformers that are the key component [20].

The losses in the iron core are basically due to hysteresis and eddy current
effects in the magnetic core. The eddy current losses, as in the case of the
windings, depend on the square of the frequency. The hysteresis loss
dependence with the frequency is not related in a simple direct way like the
eddy current losses. An important effect particularly relevant to power
transformers is the circulation of the triplex zero-sequence current in the delta
windings. The extra circulation currents could overrate the windings unless
they are taken into account in the design.

5.4.1 Transformer de-rating factor

Harmonics affect transformers in two ways. A voltage harmonic producing
additional losses in the transformer core and the higher frequency harmonic
voltages set up hysteresis loops, which superimpose on the fundamental loop.
Each loop represents the high magnetising power requirements large core
losses. In a more serious way the effects of harmonics are due to harmonic
frequency currents in the windings of the transformer.

They increase the RMS current flowing in the transformer windings, which
results in additional I2R losses. The winding eddy current losses also
increase. An eddy current loss can cause an excessive winding loss from the
high-frequency heating effects. This results in a higher winding temperature
rise, which could shorten the transformer life expectancy. Transformers that
supply non-linear loads must be de-rated to handle the harmonics. The de-
rating factor is based on the percentage of the harmonic currents in the load
and the rated winding eddy current losses.

The method by which transformers could be rated to handle harmonic loads is

that of the k-factor ratings. The k-factor concept is derived from the
ANSI/IEEE C57.110 standard, Recommendation Practice for Establishing
Transformer Capability when supplying Non-sinusoidal Load Currents [21].

Two methods to calculate de-rating factor of a transformer:

 88

a) Method 1

This k-factor is equal to the sum of the square of the harmonic frequency
currents multiplied by the square of the harmonic frequency numbers [21]:

2 2 2 2 2 2 2 2
k := I1 ⋅ ( 1) + I2 ⋅ ( 2) + I3 ⋅ ( 3) + ....... + In ⋅ ( n )
5-5

Where:

I1 is the ratio between the fundamental current and the total RMS current,

I2 is the ratio between the second harmonic current and the total RMS current,
and

I3 is the ratio between the third harmonic current and the total RMS current.

b) Method 2

kVA de-rated = (HDF) × (kVA Nameplate Rating) 5-6

To determine the de-rating factor for the transformer, take the peak and RMS
current measurements for the three-phase conductors. If the phases are not
balanced, average the three measurements and plug that value into the
following formula:

HDF = Harmonic Distortion Factor

(1.414) ⋅ RMSPhaseCurrent
= 5-7
RMSPeakPhaseCurrent

This formula generates a value between 0 and 1.0, typically between 0.5 and
0.9. If the phase currents are purely sinusoidal (undistorted), the
instantaneous peaks are 1.414 times the RMS value and the de-rating factor
is 1.0. If this is the case, no de-rating is required. However, with harmonics
present, the transformer rating is the product of the nameplate kVA rating
times the Harmonic De-rating Factor (HDF).

KVA de-rated = (HDF) × (KVA) 5-8

This study will be used to clarify method 1 for de-rating transformers.

Current measurements on the secondary side of the step-down transformer

as shown in Figure 4-2 will be used to clarify this method. The magnitudes of
the current are:
 89

Fundamental = 140.75 A, 3rd harmonic = 3.63 A, 5th harmonic = 70.13 A, and

7th harmonic = 19.63 A. This current magnitude will be used to determine the
k-rating of 11kV/400/208V, 500kVA step-down transformer on the SAC
network.

The total RMS current is:

(140.75)2 + (3.63)2 + (70.13)2 + (19.63)

2
I=

I = 158.516 A

140.75
I1 = = 0.888
158.516

3.63
I3 = = 0.0229
158.516

70.13
I5 = = 0.442
158.516

19.63
I7 = = 0.1238
158.516

k = (0.888) ⋅ (1) + (0.0229) ⋅ (3) + (0.442 ) ⋅ (5) + (0.1238) ⋅ (7 )

2 2 2 2 2 2 2 2

k = 6.428
k ≈6

This transformer should be capable of handling 140.75 A of the total RMS

current with a k-factor of no less than 6. Transformers are marked with k-
ratings of 4, 9, 13, 20, 30, 40 and 50, so as to address the harmonic loading
and future expansion at SAC. The transformer should be specified as being
capable of supplying a minimum of 140.75 A with a k-rating of 9. Of course, it
would be best to consider possible load growth/expansion on the network and
adjust the minimum capacity accordingly. Such a transformer would be able to
carry the full RMS load current and handle winding eddy current losses equal
to k times the normal rated eddy current losses.

This study will also be used to clarify method 2 for de-rating transformers.
 90

The 11kV/400/208V step-down transformer rated at 500 kVA will be used for
method 2. The currents measured on the secondary side of the step-down
transformer are given in Table 5-5.

Table 5-5: RMS phase and peak currents

CURRENT BLUE PHASE RED PHASE YELLOW PHASE AVERAGE

RMS phase 159 A 171.9 A 191.87 A 174.26 A
PEAK 262 A 280 A 309.56 A 283.85 A

I phase RMS avg. = 174.26 A

I peak avg. = 283.85 A

Using equation 5-9 then:

HDF =
(1.414)(174.26) × 100 = 87%
283.85

The results indicate that with the level of harmonics present the transformer
should be de-rated to 87% of its rating to prevent overheating. The
transformer capability with the given non-sinusoidal conditions composition is
approximately 87% of its sinusoidal conditions capability.

Thus,

kVAde − rated = (HDF ) × (kVA − nameplate )

kVAde − rated = 0.87 × 500kVA = 435kVA 5-9

The result shows that the 500kVA transformer needed to be de-rated to

435kVA so that it could handle the harmonics in the system.
 91

5.4.2 Estimation method for power system harmonics

effects on power transformers
Two other methods were discussed in chapter 5, section 5.4.1, but these
methods neither show nor indicate the ability of a transformer to handle
harmonic loads. The method of the estimation of power system harmonic
effect on transformers was developed in [22]. The method and the results of
this estimation method will be used in this study to determine the capability of
a transformer under non-sinusoidal conditions.

The method of estimating the harmonic current generated by the non-linear

loads connected to a single power transformer is discussed in [22]. This
method is an expansion of the existing k-factor method: The expansion
provides a useful tool for specifying power transformers for operating in a non-
sinusoidal load environment. Several methods of estimating the additional
heats created by non-linear loads are available, but only one method portrays
the effects of individual harmonic currents. The available methods will be
evaluated in this study and the method that includes multiple harmonics
generating loads to a transformer will be proposed for use in this study.

5.4.3 Methods of harmonic load content estimation

Three methods of estimating harmonic load are employed. The crest and
harmonic factor or the percentage total harmonic distortion (%THD) are the
two most used methods of harmonic estimation. The third method, the k-
factor, is the most complicated one, but also the most meaningful method.

The following is a brief introduction to these methods and an explanation of

the accuracy and usefulness thereof.

a) Crest factor

The crest factor is a measure of the RMS value. A perfect sine wave has a
crest factor of 1.414. The mathematical definition of the crest factor is the
 92

peak magnitude of the current waveform divided by the true RMS value of the
current:

Peak
Crest factor = 5-10
RMS

The logic of this equation is that any deviation from 1.414 represents a
distorted waveform. As will be shown later, this method is limited because
harmonic frequencies are not considered.

b) % THD

The %THD is the ratio of the RMS value of the harmonics to the RMS of the
fundamental. In equation form:

∑I
2
h
h≡2
%THD = 5-11
I1

%THD is a measure of the additional harmonic current’s contribution to the

total current. It is used to measure the distortion of the waveform due to the
harmonics. This equation is also limited because harmonic frequencies are
not considered.

5.4.4 Limitations of the crest factor and %THD

methods
The eddy-current loss of transformer increases in proportion to the square of
the load current and the frequency. The exponential characteristic causes
heat in the transformer and an increase in the abnormal winding temperature
[22]. In order to model the additional heating that harmonic loads generate in
the transformer, it is important for the frequency of the harmonic to be
considered. The crest factor and % THD are not adequate when specifying
equipment in a harmonic load environment. In order to model the heating
effect of the harmonic loads, the frequency must be considered. The k-factor
is the only harmonic content estimation available that takes harmonic
frequency into account [22].
 93

a) K-factor method

The k-factor calculation considers the important effect that frequency has on
transformer losses. The k-factor is defined as the sum of the squares of the
harmonic currents times the harmonic numbers squared [21]-[22]. In the form
of an equation, it is:

∞
∑ (I h) ⋅ h
2 2
K
h =1 5-12

where Ih is the harmonic current based on the magnitude of the fundamental

current and h is the harmonic number.

It considers the effect of harmonic current frequency. Thus, it is the most

accurate and most useful method of estimating the non-linear load harmonic
content for the specification of power distribution transformers.

b) Difficulties when using the k-factor method

It is difficult to implement the k-factor for both existing and new loads. To
calculate the k-factor on an existing system, a harmonic analysis must be
done to determine the harmonic behaviour of the system. It is difficult to obtain
an accurate harmonic study because the load is generally in a constant state
of transition. The analysis should be extended to show the maximum peaks
and valleys of the harmonic currents over time. If the analysis is complete, the
transformer can now be designed for the worst-case harmonic obtained. A
simple expansion of the k-factor calculation provides conservative results
when calculating the k-factor for differing multiple loads.

5.4.5 Expansion of the k-factor method

The basic principle behind this adaptation of the k-factor is using the k-factor
as a weighted sum. The current waveform will have different harmonic
characteristics due to the presence of different types of non-linear loads.
When considering each type of load separately, each non-linear component
can be considered individually and combined to create a composite harmonic.
Thus, the k-factor transformer can be specified for any particular application.
 94

1. Developing an equivalent harmonic current

The modification of the calculated k-factor is to develop an index relating the
k-factor to the harmonic current. Since the 3rd, 5th and 7th harmonics represent
the first symmetrical component’s current terms after the fundamental, the
evaluation of their usefulness in developing an index for a k-factor weighted
sum will be helpful [22].

The equation used to find the values of equivalent harmonic current, IEH, from
the three harmonic indices in [22] is:

(1 − K )
(K − h2)
I EH
5-13

where IEH is expressed as harmonic amperes, HA; K is the k-factor as

calculated by Equation (5-13); and h2 is the square of the harmonic number.

a) The 3rd harmonic index reaches an infinite value when K = 9 and is

undefined for the values of K > 9. Since standard transformer K-rating are
available as high as K-50, the 3rd harmonic is useless because harmonic
indexes for several k-factors are between K= 9 and K = 50.

b) The 5th harmonic index reaches an infinite values at K = 25 and is

undefined for the values of K > 25. Since standard K-ratings are available
as high as K-50, the 5th is useless as a harmonic index because several K-
factors are between K = 25 and K = 50.

c) The 7th harmonic index reaches an infinite value at K = 49 and is

undefined for values of K > 49. Since the index cut-off of 49 is
approximately the highest available k-rating, K-50, the 7th harmonic is a
good choice for the equivalent harmonic current Index [22].

The 7th harmonic current index has the most useful relationship with the k-
factor in the development of the weighted sum approach to supply non-linear
loads with a single transformer [22].

The following equation can be used as an index of the equivalent harmonic

current for most values of the harmonic current generating loads:
 95

1− K
I EH = × 100 5-14
K − 49

After establishing the k-factor of the load, the equivalent harmonic current
(EHC) can now be determined and combined with other loads as a weighted
sum. Then the multiple harmonic generating loads can be combined and
applied to a transformer.

2. Multiple harmonic loads

The application and usefulness of the expanded k-factor method will be
illustrated in this study. The 500kVA, 11kV/400/208V, delta-wye step-down
transformer supplying two uninterruptible power supplies (UPS) 120 kVA, k-30
each and another 260kVA of k-15 of different office equipment and power
electronic equipment. The k-rating of the equipment is not available, and
values k-30 and k-15 are assumptions to illustrate the method.

The following is the procedure to be followed when applying this method of

estimation:

1. Determine the k-factor of each type of load using equation (5-13).

2. Determine the equivalent harmonic current (EHC), IEH for each of the three
types of loads. This represents the equivalent magnitude of the 7th
harmonic to generate the same amount of heat as a specific k-rated load.

Using equation (5-15), the IEH factors are determined to be:

I EH (K − 30) =
(1 − 30) × 100 = 123.544 HA 5-15
(30 − 49)

I EH (K − 15) =
(1 − 15) × 100 = 64.168 HA 5-16
(15 − 49)
3. Calculate the weighted sum of the kVA-equivalent harmonic current factor
for the various types of loads and add them together. This is determined
by multiplying the nameplate of the load kVA by the equivalent harmonic
current determined above for the type of load. The results of these values
are:
 96

Suppose two UPS’s have the kVA rating of 120 kVA, and a k-rating of (k-
30) each and another 260kVA and k-rating of (k-15) are for different office
equipment and power electronic equipment:

KVA (k-30) = (120kVA) (123.544 HA) = 14825.28 KVA-HA

KVA (k-30) = (120kVA) (123.544 HA) = 14825.28 KVA-HA

KVA (k-15) = (260kVA) (64.168 HA) = 16683.68 KVA-HA

Total weighted sum composite load = 46334.24 KVA-HA

4. Determine the composite equivalent harmonic current, or the magnitude of

the 7th harmonic current that produces the same harmonic heat as the
combination of the harmonic load. This is done by dividing the total
weighted sum composite load found in step 3 by the total kVA load of the
transformer, 500kVA in this application. The result is:

46334 .24 kVA ⋅ HA 5-17

I EH = = 92.67 HA
500 kVA

This shows that 92.67 % of the fundamental current as the 7th harmonic
produces the same heat as the combination of non-linear loads outlined
above. The results are then substituted in equation (5-13) along with the
fundamental current, and the multiple non-linear load composite k-factor is
determined to be:

k = (1) 2 × (1) 2 + (0.9267) 2 × (7) 2

5-18
k ≈ 43

The next higher standard size k-factor transformer should be specified for the
set-up at SAC to avoid a shortened transformer life. The transformer with k-
45, 500kVA is required. K-factor ratings range between 1 and 50. The higher
the k-factor, the more heat from harmonic currents the transformer is able to
handle. A standard transformer that is designed for linear loads is said to have
a k-factor of 1, whereas a transformer with a k-factor of 50 is designed for the
harshest harmonic current environment possible. Transformers rated with k-
factors of 40 and 50 are extremely rare, very expensive and generally are not
 97

used. Making the correct selection of k-factor is important because it affects

cost and safety. Calculations of harmonic content produces a precise value of
k-factor, but power loads change constantly rendering the calculated value
questionable.

The weighted sum method for determining the resultant system load k-rating
is the most practical approach available to specify a transformer supplying
different non-linear and linear loads. This simple expansion of the IEEE
recommendation for specifying transformer supplying non-sinusoidal loads is
easy to implement. It will prevent the cost of over-designing the entire power
network. The use of de-rated standard transformers instead of k-factor
transformers carries some disadvantages because transformers can operate
at the reduced loading. But in the future, the loading could be increased
without considering the intended de-rating.
 98

5.4.6 Effects of harmonic currents on the Hart RAO

network and motor–generator set

5.4.6.1 Effects on the Hart RAO network

Figure 5-1: SAC and Hart RAO network diagram

The power system consists of nodes and branches forming an interconnected

grid. A branch consists of one element (power line, cable) or a number of
elements in series, and all the elements of a branch carry same current under
all the conditions. In Figure 5-1, SAC and Hart RAO are connected on the
same bus on the ESKOM supply, but there is a mechanical shaft or motor-
generator set separating the ESKOM side from the two networks, i.e. Hart
RAO and SAC.

There should be a sinusoidal voltage present in all the busses throughout the
power network with a single frequency of 60Hz. But non-linear loads can
cause changes to the ideal sinusoidal voltage waveform observed at the
generator or the source. Harmonic voltage drops over the system impedances
result from the harmonic currents flowing through the frequency-dependent
impedances, causing a non-sinusoidal voltage at the bus. The voltage drop
 99

superimposes on to the 60Hz voltage, resulting in a non-sinusoidal shaped

waveform, which is called a distorted wave.

Can this distorted wave feed to other loads, as for this study, Hart RAO
network, independent of whether they are linear or non-linear? The harmonic
current “propagates” through the network. Non-linear loads at SAC are
injecting harmonic currents that can travel to other locations in the power
network and eventually back into the power source. This was observed on the
wave shapes in chapter 4.

There are some concerns about harmonic currents: first, can the harmonic
currents generated by non-linear loads at SAC feed to the Hart RAO network,
which are connected to the same bus on the ESKOM supply? Secondly, with
reference to the distortion level at SAC, how long will the system operate
before major equipment failure occurs?

From the measurements taken, it is clear that Hart RAO waveforms are clean
with no distortion in the voltage and current waveforms. The Hart RAO power
network cannot be affected by harmonics generated by the SAC load because
the two networks (Hart RAO and SAC) have two separate motor-generator
sets, which means that there is a mechanical coupling between the two
networks and the ESKOM supply. Thus, harmonics generated by SAC loads
cannot affect ESKOM and Hart RAO.

The waveform shapes depend on the line-to-line voltage of the generator,

which is pure sinusoidal. The only way that Hart RAO could be affected is by
the presence of unbalanced non-linear loads that can create unbalanced
current components that cause voltage drops across the source impedance
and, hence, generate harmonic powers flowing backwards from the load to
the network.

5.4.7 Effects of harmonic currents on a motor

A non-sinusoidal voltage applied to electric machines could cause
overheating, pulsating torque or noise. Rotor overheating has been the main
problem associated with voltage distortion.
 100

Losses in electric machines are dependent on the frequency spectrum of the

applied voltage. Core and stay losses could become significant in an induction
motor with a skewed rotor supplied from an inverter producing high harmonic
frequencies. An increase in the operating temperature of a motor will cause a
reduction in the motor operating life, and single-phase motors are the most
affected [23].

The temperature rise is not uniform throughout the motor and hot spot appear
near the conductors within the iron core portions. If the harmonics are time
varying, the motor can tolerate higher peak distortion levels without increasing
the hot spot temperature. This is possible because the motor thermal time
constant is much longer than the period of the harmonic variation.

5.4.8 Unbalanced currents on a synchronous

generator
Unbalanced loading: A three-phase balanced load produces a reaction field,
which is constant and rotates synchronously with the rotor field system. Any
unbalanced condition can be resolved into positive, negative and zero-
sequence components. The positive sequence component is similar to the
normal balanced load. The zero-sequence components produce no main
armature reaction, and do not cause heating of the generator rotor, so this
component is sometimes eliminated from the measurements, by using delta-
connected current transformers [24]. This is not necessary for a generator unit
since no zero-sequence current can flow in the generator because of external
conditions [24].

The negative sequence components are similar to the positive sequence

component system, except that the resulting reaction field rotates counter to
the DC field system and, hence, produces a flux that cuts the rotor at twice the
rotational velocity, thereby inducing double frequency currents in the field
system and in the rotor body [24]. The resulting eddy-currents are very large
and can cause serve heating of the rotor.

This effect is so severe that the single-phase load equal to the normal three-
phase rated currents can quickly heat the brass rotor slot wedges to a
softening point. They can then be extruded under centrifugal force until they
 101

stand above the rotor surface, when it is possible that they could strike the
stator iron [24]. Concentration of heating occurs on portions of the coil binding
rings and here surface fusion has been known to occur. Heating depends on
the reaction field and on the load current. A machine can be assigned a
continuous negative sequence rating [24].

A source of unbalance is in the system and will affect generators in the

vicinity. These should not be disconnected unless the condition remains
uncorrected for such time that there is danger of the generator being
damaged. Unbalanced currents on three-phase motors are undesirable, but a
small amount can generally be tolerated. Excessive unbalanced currents can
shorten motor life and increase energy consumption. Unbalanced three-phase
stator currents cause double-system-frequency currents to be induced in the
rotor iron.

The currents will quickly cause rotor overheating and serious damage if the
generator is permitted to continue operating with such an unbalance.
Unbalanced currents can also cause severe vibration, but the overheating is
the most serious problem.

5.4.9 Neutral currents in a three-phase power system

A harmonic current causes abnormal high currents in the neutral conductor of
the three-phase four-wire system, which are feeding single-phase loads.
Because loads such as switch mode power supplies produce third harmonic
currents and balanced fundamental frequency, three-phase currents will result
in no neutral current. But, in three-phase circuits, third harmonic currents add
rather than cancel the neutral and could be as much as 1.7 times the phase
currents for converter loads [25].

The neutral is normally sized the same as the phase conductors. Therefore,
the neutral conductor could be overloaded. This problem occurs often in
industries where a three-phase distribution system feeds large single-phase
electronic office equipment loads.
 102

5.4.10 Problem of a high neutral current

A high neutral current in a power system can be caused by an overloaded
power source, overloaded transformers, voltage distortion, and a common
mode noise. When the neutral conductor carries harmonic currents, additional
heat is generated and the ampacity of the conductor is reduced.

One form of common mode noise in a three-phase power system is the

voltage difference between the neutral and the ground [25]. With high
harmonic neutral current, the impedance of the neutral conductor at the
harmonic frequencies can cause high neutral conductor voltage drops. The
neutral conductor voltage drop appears as a common mode noise [25].

The effect of this low frequency common mode noise voltage is somewhat
debatable, typically calling for less than 0.5 RMS, neutral-to-ground at the
source regardless of the frequency. But as we move away from the source,
this neutral-to-ground voltage increases depending on the load and the
distance between the sources. This can measure 2 to 3 V [25]. The neutral
conductor can be sized to twice the phase conductor capacity to solve the
problem.

5.5 Summary
The harmful effects of harmonic voltages and currents on transformers, cables
and motor generator set performances often go unnoticed until an actual
failure occurs. In some instances, transformers that have operated
satisfactorily for long periods have failed in a relatively short time when plant
loads were changed or a facility's electrical system was reconfigured.

The reduction in the transformer loss of life is related to the additional

increment in the temperature that the harmonics create in the equipment. The
life reduction is related to the temperature of the transformer hottest spot. The
temperature increment will lead to a faster degradation in the transformer
insulation and a shortage of the equipment life.

Non-sinusoidal loads cause an accelerated loss of insulation life in power

transformers windings and power cables because of the exponential heating
effects of high frequency harmonics. Distorted voltages applied to a machine
 103

this could cause overheating and noise. Furthermore, the significant

unbalance currents can shorten the lifespan of a generator.
 104

6 CONCLUTIONS AND RECOMMENDATIONS

6.1 Conclusions
There is agreement between the model and real measurements. It was
difficult to match the model results with exact conditions of the field
measurements for several reasons:

a) Unknown losses in the transformers and underground cable

b) Variable nature of the loads

c) Different harmonics injected at different speeds

d) Unknown load powers

A simulation result shows that the transmission efficiency is good, ± 2%

transmission losses, and losses on the real system are ± 4% due to losses on
the transformers and the cable. The simulation result shows that as the load
increases, the transmission efficiency becomes quiet bad. The author
concludes that it is not possible to increase the load at SAC without
conducting harmonic studies of the network and de-rating power system
components to compensate for that harmonically loaded environment.

NRS 048 and the IEEE follow different approaches to determine the
prescribed voltage and current distortion limits. The IEEE sets limits for the
harmonics of not only the current caused by the operation of the loads, but
also the supply voltage. NRS 048 sets harmonics limits for the supply voltage,
but does not provide current harmonic limits for the operation of loads as the
IEEE. NRS 048 sets harmonic voltage limits only and leaves the harmonic
producers responsible for reducing their respective harmonic currents so that
the system voltage limits are not violated. Signs of harmonics are occurring in
the system. The analysed field measurements in chapter 4 show that the
overall total harmonic voltage distortion is high in the SAC network, with
reference to NRS 048 and IEEE 519 standards. It shows that the individual
harmonic does not exceed the limit of the specific harmonics specified by
NRS 048 except the 5th harmonic, which is high in every point measured on
 105

the network. Therefore, if additional non-linear loads were to be connected to

this voltage distortion, they would increase the voltage distortion level above
the allowable limits.

Unbalance voltage is not a problem in the SAC network, because the voltage
unbalance factor is below 2% stated by the IEEE and NRS 048. The only
problem in the SAC network is unbalanced and distorted currents, and their
effects on the power system.

The new index discussed in chapter 4 can be used to evaluate the distortion
on the network. Because this current non-linear index is independent of the
load entity and can be used to express the capability of the non-linear load to
affect harmonic diction at the source. This index can provide information about
the harmonic distortion level and be compared to the current total harmonic
distortion to indicate its accuracy and behaviour.

It is evident that the power system engineer must be concerned about the
application of the complete range of power system devices when applied in
the harmonically loaded environment. This subject is by no means complete
and a great deal of work remains to be done. It is possible to use basic
application guidelines and limitations. The response of various types of
equipment to distorted waveforms is by no means uniform and some
components are of more concern than others. Capacitors, dry type
transformers and electronic equipment rate require the highest degree of
attention. The effects of distorted waveform on cables, motors and oil-filled
transformers cannot be ignored. Transformers supplying non-linear loads at
SAC require de-rating to avoid overheating.

Regarding the methods of calculating the heating due to harmonics, only the
expanded k-factor method was acceptable for specifying the non-sinusoidal
condition. This method takes harmonic frequency into account. The estimation
method for power system harmonics effects on power transformers was
discussed in chapter 5 for dry-type or k-rating types of transformers, but oil-
filled transformers are used at SAC. It will be costly for SAC to redesign and
replace the oil-cooling type of transformers. For an accurate estimation of the
additional heat non-sinusoidal loads inflicted on transformers, harmonic
 106

frequency must be considered. The k-factor considers the harmonic

magnitude and frequency. Because of these two facts, the k-factor is the only
accepted method for estimating harmonic content and specifying the
transformer for use under non-linear load conditions.

The set of equations for calculating total ohmic losses in the cable with
distorted current are presented in chapter 5. These equations can be used to
compute the required de-rating factor for the cable system because of the
presence of harmonics. The calculations can be performed with a handheld
calculator. Example of these calculations are included in chapter 5 of this
document.

It would be beneficial for SAC to minimise the harmonic level on the network
to avoid equipment failure and a major loss of production. As the number of
non-linear loads increases, the level of distortion will also increase. But it is
difficult to notice the effects of a harmonic current on equipment until
equipment failure occurs.

6.2 Recommendations and future work

Firstly, for future expansion at SAC, it is recommended that the designers
utilise the expanded k-factor rating system and the proposed adaptation to
indicate the ability of a transformer to handle harmonic loads, and use k-rated
transformers instead of oil-filled transformers. K-rated transformers are
designed for operation under non-sinusoidal loading. The k-factor indicates
transformer’s ability to withstand non-sinusoidal loading without exceeding the
limits of insulation temperature rise for normal life expectancy

K-rated transformers are designed for power harmonic generating loads by

manufacturers that carry additional harmonic currents, so iron losses or
copper losses can be decreased and efficiency can be increased. For this
reason, a nonlinear overload is supplied without overheating, excessive
voltage distortion and exceeding the specified capacity of the transformer.
 107

Secondly, it is recommended that for future work an analytical harmonic study

be performed when converter loads are added to a power system or when
significant changes are made to the existing system – the study should
identify the steps necessary to minimise harmonics and evaluate the effects of
the harmonics on power system components, especially synchronous
generators. When purchasing equipment for an example transformer, for use
in this harmonic loaded environment at SAC, the manufacturer should be
notified of the unusual conditions of the actual harmonic currents and voltages
to which the transformer will be subjected. The designer is in the best position
to come up with an economical design, which would still be able to meet the
duty imposed by the power system environment.

Finally, additional overcurrent protection should be considered for all the

transformers supplying non-linear loads at SAC. With three-phase
transformers, the triplen harmonics are cancelled in the delta winding and do
not appear in the input current. The output current and transformer loading are
greater than is apparent from the input current. Therefore, the transformer can
be overloaded without the primary overcurrent protection ever tripping. Adding
transformer secondary overcurrent protection helps, but it still does not protect
the transformer from the heating effects of harmonic currents. Supplemental
protection in the form of winding temperature sensors could be used to
provide an alarm and/or system shutdown in the event of overloading,
excessive harmonic currents, high ambient temperatures or inadequate
cooling.
 108

7 ANNEXURES

Annexure A: PSCAD Software

PSCAD is a powerful user friendly and accurate power simulator for all types
of power distribution systems. It is graphically interfaced that allows the user
to systemically construct any circuit, run simulation, analyse results, run-time
control and reporting. This software includes a library where most of the
components could be easily accessed for power system analysis. Own
modeling of system and components using the built-in graphical components
workshop, could also be created when the master library could not support the
simulation. Although PSCAD is improved to interface well with the other
Microsoft windows operating systems, it requires a fortran compiler to execute
the simulation. Together with the compiler, PSCAD has proved to be effective,
accurate and fast simulation software for studying transient behavior of
electrical networks.

SAC and Hart RAO diagram

Figure A-1: Network diagram of SAC and Hart RAO

 109

Figure A-2 : Simplified diagram of SAC

 110

Annexure B: Unbalance calculations

IA := 217.56( 0.93969 − 0.342i) VA := 278.07( 1)
IA = VA =
( )
arg IA = deg ( )
arg VA = deg

VB := 277⋅ ( −0.5 + 0.866i)

IB := 217.56( −0.1736 + 0.9848i)
VB =
IB =
( )
arg IB = deg ( )
arg VB = deg

IC := 195.09( −0.7660 − 0.64278i) VC := 277.5( −0.5 + −0.866i)

VC =
IC =

( )
arg IC = deg ( )
arg VC = deg

1
Va = (Va + Vb + Vc )
0
a := −0.5 + 0.866i
1
Ia = (Ia + Ib + Ic )
0
3
2 3
a = 1
Va = (Va + aVb + a 2Vc )
1
1
I a = ( I a + aI b + a 2 I c )
1
3
3 1
Va = (Va + a 2Vb + aVc )
2
1
I a = ( I a + a 2 I b + aI c )
2
3
(
Izero := IA + IB + IC ⋅
1
3
) 3
(
Vzero := VA + VB + VC ⋅
1
)
3
Izero =
Vzero =
(
arg Izero = ) (
arg Vzero = deg)

Ipos := ⎛ IA + a ⋅ IB + a⋅ IC⎞ ⋅ Vpos := ⎛ VA + a ⋅ VB + a⋅ VC⎞ ⋅

2 1 2 1
⎝ ⎠ 3 ⎝ ⎠ 3
Ipos = Vpos =

(
arg Ipos = ) ( )
arg Vpos = deg

Ineg := ⎛ IA + a IB + a ⋅ IC⎞ ⋅ Vneg := ⎛ VA + a VB + a ⋅ VC⎞ ⋅

2 1 2 1
⎝ ⎠ 3 ⎝ ⎠ 3

Ineg = Vneg =

(
arg Ineg = deg ) ( )
arg Vneg = deg

Current Unbalance Vneg

UN :=
Ineg Vpos
UNI :=
Ipos Voltage Unbalance:

UNI = % UN = %
 111

Annexure C: Cable parameters

Underground cable parameters

3km, 3 core, 11kV Paper Drain Lead Copper
Cable
1.Conductor diameter - 4.7mm
Area –16 mm2
2.Insulation thickness – 3.2 mm
Conductor resistance DC 1.15 ohms/km
AC resistance 1.37 ohms/km 4.1283 ohms
Conductor Inductance 0.435mH/km 1.305H
Inductive Reactance 0.137 ohms/km 0.411H
Impedance/phase 1.3828 ohms/km
Capacitance 272 n F/km 0.816 mF
Charging current 0.543 A/km
Current rating In Ground 91 A
Current In Air 88 A
 112

8 LIST OF REFERENCES

[1] Energex Network Investigations Department; Power Quality: A

Guide for Commerce and Industry – Effects, Advice and

Characteristics, 2nd Edition, October 1998.

[2] //http://www.satellite application centre/.htm

[3] //http://www.hart.radio.astronomy.observatory/.htm

[4] Dalia C. Navas “Power Notes” Pacific Gas Electric Company

[5] IEEE Standard 519-1992, IEEE Recommended Practices and

Requirements for Harmonic Control in Electrical Power Systems.

[6] J. Arrilaga, D.A. Bradley, P.S. Bodger, Power System Harmonics, John
Wiley & Sons, New York, Brisbane 1985.

[7] Ned Mohan, Tore M. Undeland, William P. Robbins, Power Electronics:

Converters, Applications and Design, John Wiley & Sons, New York.

[8] R.C. Dugan, M.F. McGranaghan and H.W. Beaty, Electrical Power
Systems Quality, McGraw-Hill, USA, 1996.

[9] J. Lamoree, D. Mueller, P. Vinett, W. Jones and M. Samotyj, “Voltage

Sag Analysis Case Studies”, IEEE Transactions on Industry
Applications, Vol. 30, No. 4, July/August 1994.

[10] C.J. Melhorn, T.D. Davis and G.E. Beam, “Voltage Sags: Their Impact
on the Utility and Industrial Customers”, IEEE Transactions on Industry
Applications”, Vol. 34, No. 3, May/June 1998, pp. 549-558.

[11] NRS 048-2 1996: Voltage characteristics, system acceptability levels,

compatibility levels, limits and assessment methods.

[12] Basic Measuring Instruments (BMI), Handbook of Power Signatures,

2nd Edition, Revised and Expanded, Basic Measuring Instruments,
1993.
 113

[13] Hadi, Saadat Power System Analysis, McGraw – Hill Series in

Electrical and Computer Engineering, International Editions 1999.

[14] Wilsun Xu, “Status and Future Direction of Power System Harmonic
Analysis”.

[15] IEEE Working Group on Non-sinusoidal Situations, “Practical

Definitions for Power systems with Non-sinusoidal Waveforms and
Unbalanced Loads: a Discussion”, IEEE Trans. On Power Delivery,
Vol. 11, January 1996, pp.79-87.

[16] J.H.C. Pretorius, “Modelling of distorted electrical power and its

practical compensation in industrial plant”, doctoral thesis, Rand
Afrikaans University, Dec 1998.

[17] IEEE Standard 519-1992, IEEE Recommended Practices and

Requirements for Harmonic Control in Electrical Power Systems.

[18] Antonio Dell’ Aquila, Maria Marinelli, Vito Giuseppe Monopoli, Pericle
Zanchetta, “New Power-Quality Assessment Criteria for Supply System
Under Unbalanced and Non-sinusoidal Conditions”, IEEE Transactions
on Power Deliver, vol.19, no.3, July 2004, pp 1284 – 1290.

[19] M. A. Martin, “Calculations of secondary cable losses and impacity in

the presence of harmonics”, IEEE Transaction on Power Delivery Vol.
7, No. 2, April 1992, pp 451- 459.

[20] S. P. Kennedy, C.L. Ivey, “Application, Design and Rating of

Transformers Containing Harmonic Currents”, Pulp and Paper Industry
Technical Conference, 1990, Conference Record pp: 19-31.

[21] IEEE Std C57.110-1986: National Standard Recommended Practice for

Establishing Transformer Capability when supplying Non-Sinusoidal
Loads Currents.

[22] Gregory W. Massey, “Estimation Method for Power System Harmonic

Effects on Power Distribution Transformers.” IEEE Transaction on
Industry Applications, vol. 30, No. 2, March 1994, pp 486 – 489.
 114

[23] Protective Relays Applications Guide (3rd ed), GEC Measurements,

Stafford, England, 1987

[24] C. R. Mason: The Art and Science of Protection Relaying. GE Industrial

Systems.

[25] C. Sankaran, Power Quality CRC press,2002.