University of Wollongong

Research Online

University of Wollongong Thesis Collection University of Wollongong Thesis Collections

1954-2016

2003

Harmonic management in mv distribution systems

Duane Robinson
University of Wollongong

Follow this and additional works at: https://ro.uow.edu.au/theses

University of Wollongong
Copyright Warning
You may print or download ONE copy of this document for the purpose of your own research or study. The University
does not authorise you to copy, communicate or otherwise make available electronically to any other person any
copyright material contained on this site.
You are reminded of the following: This work is copyright. Apart from any use permitted under the Copyright Act
1968, no part of this work may be reproduced by any process, nor may any other exclusive right be exercised,
without the permission of the author. Copyright owners are entitled to take legal action against persons who infringe
their copyright. A reproduction of material that is protected by copyright may be a copyright infringement. A court
may impose penalties and award damages in relation to offences and infringements relating to copyright material.
Higher penalties may apply, and higher damages may be awarded, for offences and infringements involving the
conversion of material into digital or electronic form.

Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily
represent the views of the University of Wollongong.

Recommended Citation
Robinson, Duane, Harmonic management in mv distribution systems, PhD thesis, School of Electrical,
Computer and Telecommunications Engineering, University of Wollongong, 2003. http://ro.uow.edu.au/
theses/178

Research Online is the open access institutional repository for the University of Wollongong. For further information
contact the UOW Library: research-pubs@uow.edu.au
HARMONIC MANAGEMENT IN MV DISTRIBUTION SYSTEMS

A thesis submitted in fulfilment of the

requirements for the award of the degree

DOCTOR OF PHILOSOPHY

from

UNIVERSITY OF WOLLONGONG

by

DUANE ROBINSON, BE(ELEC)

School of Electrical, Computer and Telecommunications Engineering

2003

i
To Beverly

ii
 Abstract

With many distribution customer loads being sensitive to excessive harmonic voltage

distortion, electricity distribution network service providers should now be looking

towards preventative measures to ensure that voltage distortion levels remain within

limits set by the appropriate standards. Measures will need to be taken at the planning

stage to ensure distribution systems are able to meet harmonic limits recommended by

standards as distortion due to loads increases. This thesis describes the development of

harmonics planning and analysis tools that allow effective system modelling and

comparison with standards, particularly in the planning phase where details of loads are

usually not accessible.

In this thesis development of statistical harmonic models of residential, commercial and

industrial load types to simulate the global behaviour of distorting loads at distribution

substations is presented. Both time and phase diversities are included in the

representative load models. A method to estimate the 95th percentile cumulative

probability level of harmonic voltage distortion in an MV distribution system as

required by the present Australian harmonic standard (AS/NZS 61000.3.6) is also

developed. Results from a harmonic monitoring programme carried out on a typical MV

distribution system are used to establish parameters for the load models and also to

confirm the relative accuracy of the proposed distortion level prediction technique.

A generalised method to extend the IEC 61000-3-6 approach of allocating allowable

harmonic emissions to the case where customers are distributed along an MV

distribution system feeder having significantly different fault levels is presented. The

method involves the determination of an 'allocation constant' using the agreed loading

iii
of all customers and the system harmonic impedances. This approach typically requires

an extensive amount of data that may not always be available to distribution network

service provider engineers. An extension to the method has thus been established to

cater for the situation where only limited data is available. This is achieved by looking

at several extreme cases that classify the most common MV distribution system feeder

configurations and through the use of correction factors for the 'allocation constant'.

Several example systems have been studied to illustrate the harmonic management tools

described above. These case studies include identification of the key indicators for

harmonic performance of a distribution system.

iv
 Certification

I, Duane Robinson, declare that this thesis, submitted in fulfilment of the requirements

for the award of Doctor of Philosophy, in the School of Electrical, Computer and

Telecommunications Engineering at the University of Wollongong, is wholly my own

work unless otherwise referenced or acknowledged. The document has not been

submitted for qualification at any other academic institution.

………………….

Duane Robinson

27 November 2003

v
 Acknowledgements

I would like to express my sincere appreciation for the encouragement and guidance of

my supervisors, Prof. Vic Gosbell, and Dr. Sarath Perera.

I would also like to thank Dr. Vic Smith, Mr. Damian Mannix, and Mr. Neil Browne for

valuable technical discussions throughout the course of my candidature.

I am also grateful for the generous financial support of the Australian Research Council

Postgraduate Awards scheme and the Integral Energy Power Quality Centre.

vi
 Contents

Abstract...........................................................................................................................iii

Certification .....................................................................................................................v

Acknowledgements.........................................................................................................vi

List of symbols ...............................................................................................................xii

List of figures ................................................................................................................xiv

List of tables................................................................................................................xviii

Chapter 1
Introduction .....................................................................................................................1

1.1 Harmonics in distribution systems ................................................................1

1.2 Effects and problems caused by harmonics ..................................................3

1.3 Representation of harmonics.........................................................................4

1.3.1 Harmonic components ..........................................................................5

1.3.2 Total harmonic distortion (THD) ..........................................................6

1.4 Harmonic monitoring programmes ...............................................................7

1.5 Harmonic standards.......................................................................................9

1.6 Objective of the thesis.................................................................................11

1.7 Structure of thesis........................................................................................13

Chapter 2
Overview of harmonic monitoring in distribution systems.......................................15

2.1 Introduction .................................................................................................15

2.2 Harmonic monitoring programmes .............................................................16

2.3 Method of measurement..............................................................................17

2.4 Presentation of results .................................................................................19

2.5 Load type identification ..............................................................................21

2.6 Previous monitoring programmes ...............................................................22

vii
 2.7 Future projection of harmonics ...................................................................28

2.8 Relevant harmonic standards ......................................................................30

2.9 The summation law .....................................................................................34

2.10 Summary .....................................................................................................38

Chapter 3
Estimation of harmonic levels in MV distribution systems.......................................40

3.1 Introduction .................................................................................................40

3.2 Diversity of loads in MV distribution systems ...........................................41

3.3 Residential, commercial and industrial load characteristics .......................47

3.4 Assumptions to allow estimation of distortion levels .................................50

3.5 Key concepts for estimating distortion levels .............................................51

3.6 Method to establish distortion levels...........................................................53

3.7 Study system and harmonic monitoring results ..........................................57

3.8 Application of method to the study system.................................................58

3.9 Sensitivity results ........................................................................................60

3.10 Summary .....................................................................................................62

Chapter 4
Allocation of harmonic emissions to MV distribution system customers ................65

4.1 Introduction .................................................................................................65

4.2 Principles of IEC 61000-3-6 .......................................................................67

4.3 The allocation constant k.............................................................................69

4.4 Incomplete data approach ...........................................................................71

4.4.1 A pessimistic approximation to k........................................................72

4.4.2 Approximation to k when all feeders are similar ................................73

4.4.3 Correction factors for the pessimistic approximation of k ..................74

4.5 Case study examples ...................................................................................75

4.5.1 Homogenous example.........................................................................75

viii
 4.5.2 Extreme example.................................................................................78

4.6 Choice of allocation policy .........................................................................80

4.7 Summary .....................................................................................................81

Chapter 5
Identification of MV distribution system aggregated load models...........................83

5.1 Introduction .................................................................................................83

5.2 Selection of monitoring instrument.............................................................85

5.2.1 Relevant standards ..............................................................................85

5.2.2 Instrument accuracy ............................................................................87

5.2.3 Instrument price ..................................................................................91

5.2.4 Further instrument considerations.......................................................91

5.2.5 Instrument selection ............................................................................92

5.3 Monitoring site selection.............................................................................93

5.4 Selection of harmonics to measure .............................................................96

5.5 Results from harmonic measurements at selected sites ..............................98

5.6 Harmonic voltage and current trends ..........................................................99

5.7 Harmonic voltage and current cumulative probability..............................108

5.8 Harmonic levels compared to standards ...................................................115

5.9 Annual growth in harmonic distortion levels............................................117

5.10 Seasonal changes in harmonic distortion levels........................................119

5.11 Residential, commercial and industrial load harmonic characteristics .....120

5.12 Relationship of harmonics on individual phases.......................................121

5.13 Minimum requirements of a harmonic monitoring programme ...............124

5.14 Summary ...................................................................................................126

Chapter 6
Effects of system capacitance on harmonic levels ....................................................129

6.1 Introduction ...............................................................................................129

ix
 6.2 Modelling of the system............................................................................130

6.2.1 Line and cable impedance.................................................................131

6.2.2 Transformers .....................................................................................134

6.2.3 Load models ......................................................................................134

6.2.4 Power factor correction capacitors....................................................135

6.3 Critical lengths of feeders .........................................................................135

6.4 Detailed approach to determining system impedance...............................138

6.4.1 Location and sizing of PFC capacitors .............................................139

6.4.2 Typical resonant frequencies for different systems ..........................139

6.4.3 De-tuning of PFC capacitors.............................................................140

6.4.4 Effects of load damping ....................................................................141

6.4.5 Complexity with a number of capacitors present..............................144

6.5 Pragmatic approach in determining system impedance ............................145

6.6 Allocation of harmonic emissions with capacitors present.......................147

6.7 Summary ...................................................................................................148

Chapter 7
Application of harmonic planning methods to MV distribution systems ..............150

7.1 Introduction ...............................................................................................150

7.2 Case study 1: Homepride zone distribution system ..................................151

7.2.1 Identifying the critical impedances of the system.............................152

7.2.2 Consideration of existing loading when assessing harmonic levels .157

7.2.3 Applying suitable limits to different points within the system .........159

7.3 Case Study 2: Katoomba zone distribution system...................................163

7.3.1 Assessment of background harmonic voltage levels ........................166

7.3.2 Calculation of acceptable harmonic current emissions.....................168

7.3.3 Accounting for load interaction with system harmonic voltages......173

7.4 Case study 3: Springhill sub-transmission substation...............................176

x
 7.4.1 System layout ....................................................................................177

7.4.2 Assessment of levels and emissions..................................................177

7.4.3 Power factor correction capacitor considerations .............................179

7.4.4 Assessment of emissions...................................................................183

7.5 Summary ...................................................................................................187

Chapter 8
Conclusions ..................................................................................................................189

8.1 Conclusion and recommendations ............................................................189

8.2 Further work..............................................................................................192

Statement of original contribution ............................................................................194

Publications based on work performed on this thesis .............................................197

References ....................................................................................................................198

Appendix A Fluke 41 Power Harmonics Analyser specifications .........................205

Appendix B EDMI Mk3 Energy Meter specifications............................................206

Appendix C Results of the Homepride harmonic monitoring programme..........207

Appendix D Calculation of maximum demand for Customer A ............................210

Appendix E Calculation of acceptable emissions for Customer A .........................211

xi
 List of symbols

α exponent of the second summation law

Ah Fourier series coefficient for hth harmonic

Bh Fourier series coefficient for hth harmonic

CBD central business districts

Ch magnitude of hth harmonic voltage or current

CP95 cumulative probability 95th percentile

DC direct current

DNSP distribution network service provider

EIhi allowed harmonic current emission limit of order h for ith consumer

EUhi allowed harmonic voltage emission limit of order h for ith consumer

FFT fast Fourier transform

f frequency

GDP gross domestic product

GhMV global harmonic voltage emission of order h for all loads supplied at MV

h harmonic order

i single customer or load

Ih harmonic current of order h

k harmonic emission allocation constant

LhHV harmonic voltage planning level of order h for HV

LhMV harmonic voltage planning level of order h for MV

MV medium voltage

n number of customer PCCs along the weakest feeder only

PCC point of common coupling of the customer

PFC power factor correction

xii
r number of parallel feeders

rms root mean square

s circuit segment number

SFj total capacity of all loads along jth feeder

Si apparent maximum demand of ith customer

St total available power at saturation of the supply system capacity

T period

TDD total demand distortion

THDI total harmonic current distortion

THDV total harmonic voltage distortion

ThHM HV/MV harmonic voltage transfer coefficient for hth order harmonic

Uh harmonic voltage of order h

Uh0 background harmonic voltage of order h

Uhi harmonic voltage of order h for ith customer

u(t) periodic function of voltage or current

ψh phase of hth harmonic voltage or current

Zhi magnitude of hth order harmonic impedance of the distribution system at ith PCC

xiii
 List of figures

Figure 2.1 Symbolic illustration of a harmonic monitoring system................................17

Figure 2.2 Seven day recording of harmonic voltages and histograms [20]...................25

Figure 2.3 Probability of N harmonics of amplitude A exceeding kA [47] .....................35

Figure 3.1 Residential distribution transformer harmonic current snapshot

(THDI = 4.64%) .......................................................................................................47

Figure 3.2 Commercial distribution transformer harmonic current snapshot

(THDI = 6.19%) .......................................................................................................48

Figure 3.3 Industrial distribution transformer harmonic current snapshot

(THDI = 10.59%) .....................................................................................................48

Figure 3.4 Zone substation feeder harmonic current snapshot (THDI = 4.59%).............48

Figure 3.5 MV feeder equivalent circuit for voltage distortion calculation....................56

Figure 3.6 Simplified layout of Homepride zone MV distribution system ....................59

Figure 3.7 LV distribution system layout .......................................................................59

Figure 3.8 Sensitivity of the 5th harmonic voltage ..........................................................61

Figure 3.9 5th harmonic voltage sensitivity to alpha .......................................................62

Figure 4.1 Example radial MV distribution system ........................................................72

Figure 4.2 Homogeneous MV distribution network with six identical feeders ..............76

Figure 4.3 Harmonic current allocations using methods I-IV to calculate k ..................77

Figure 4.4 Harmonic voltages arising at each node using methods I-IV to calculate k..78

Figure 4.5 Extreme MV distribution system example ....................................................78

Figure 4.6 Acceptable emissions for equal harmonic current, power and voltage .........80

Figure 4.7 Harmonic voltages arising at each node using different allocation policies .81

Figure 5.1 Comparison of harmonic monitoring instruments at 5th harmonic [16] ........88

Figure 5.2 Comparison of harmonic monitoring instruments at 19th harmonic..............89

xiv
Figure 5.3 Waveform generator test of monitoring instruments at 5th harmonic............90

Figure 5.4 Waveform generator test of monitoring instruments at 19th harmonic..........90

Figure 5.5 EDMI 2000-04XX Energy Meter [60] ..........................................................92

Figure 5.6 Single line diagram illustrating the Homepride zone distribution system.....94

Figure 5.7 Homepride zone (Site 1) weekday harmonic voltage trend (5th, 19th, 49th).100

Figure 5.8 Homepride zone (site 1) weekly harmonic voltage trend (5th, 19th, 49th) ....100

Figure 5.9 Homepride zone (site 1) weekday harmonic current trend (5th, 19th, 49th)..100

Figure 5.10 Homepride zone (site 1) weekly harmonic current trend (5th, 19th, 49th) ..101

Figure 5.11 Homepride zone (site 1) weekly harmonic voltage trend (3rd, 5th, 7th)......102

Figure 5.12 Homepride zone (site 1) weekly harmonic current trend (3rd, 5th, 7th) ......102

Figure 5.13 Residential substation (site 5) weekly harmonic voltage trend .................104

Figure 5.14 Residential feeder (site 2) weekly harmonic current trend........................104

Figure 5.15 Residential substation (site 5) weekly harmonic current trend..................104

Figure 5.16 Commercial substation (site 6) weekly harmonic voltage trend ...............105

Figure 5.17 Commercial feeder (site 3) weekly harmonic current trend ......................105

Figure 5.18 Commercial substation (site 6) weekly harmonic current trend................105

Figure 5.19 Industrial substation (site 7) weekly harmonic voltage trend....................106

Figure 5.20 Industrial feeder (site 4) weekly harmonic current trend...........................106

Figure 5.21 Industrial substation (site 7) weekly harmonic current trend ....................107

Figure 5.22 Homepride zone (site 1) THD versus 5th harmonic...................................108

Figure 5.23 95th, 99th and 100th percentile 7th harmonic voltage at industrial substation

(site 7)....................................................................................................................109

Figure 5.24 Homepride zone substation (site 1) 3rd harmonic histogram .....................110

Figure 5.25 Residential substation (site 5) 3rd harmonic histogram..............................110

Figure 5.26 Commercial substation (site 6) 3rd harmonic histogram............................110

xv
Figure 5.27 Industrial substation (site 7) 3rd harmonic histogram ................................111

Figure 5.28 Homepride zone substation (site 1) 5th harmonic histogram .....................111

Figure 5.29 Residential substation (site 5) 5th harmonic histogram..............................111

Figure 5.30 Commercial substation (site 6) 5th harmonic histogram ............................111

Figure 5.31 Industrial substation (site 7) 5th harmonic histogram ................................112

Figure 5.32 Homepride zone substation (site 1) 7th harmonic histogram .....................112

Figure 5.33 Residential substation (site 5) 7th harmonic histogram..............................112

Figure 5.34 Commercial substation (site 6) 7th harmonic histogram ............................112

Figure 5.35 Industrial substation (site 7) 7th harmonic histogram ................................113

Figure 5.36 Spectrum of Homepride zone (site 1) instantaneous voltage waveform

capture ...................................................................................................................114

Figure 5.37 Spectrum of residential substation (site 5) instantaneous voltage waveform

capture ...................................................................................................................115

Figure 5.38 Seasonal 95th percentile voltage trend for Homepride zone (site 1)..........119

Figure 5.39 Homepride (site 1) 5th harmonic A and C phase voltage and current........122

Figure 5.40 Residential (site 5) 5th harmonic A and C phase voltage and current........122

Figure 5.41 Commercial (site 6) 5th harmonic A and C phase voltage and current......122

Figure 5.42 Industrial (site 7) 5th harmonic A and C phase voltage and current ..........123

Figure 6.1 Admittance model of transmission line [64] ...............................................132

Figure 6.2 Transformer harmonic model, where Xh = hX1 ...........................................134

Figure 6.3 CIGRE model for normal aggregate distribution loads [63] .......................135

Figure 6.4 Simple system to estimate critical lengths of feeders..................................136

Figure 6.5 Harmonic at which first resonance occurs for overhead line lengths ..........137

Figure 6.6 Harmonic at which first resonance occurs for underground cable lengths..137

Figure 6.7 Simple distribution model illustrating de-tuning of PFC capacitor.............141

xvi
Figure 6.8 Homogeneous distribution system for load damping study ........................142

Figure 6.9 Harmonic impedance with no PFC capacitors installed ..............................143

Figure 6.10 Harmonic impedance with PFC capacitor installed at zone substation.....143

Figure 6.11 Harmonic impedance with PFC capacitor installed at 2/3 of feeder length

...............................................................................................................................144

Figure 6.12 11kV worst case impedance curve from [15] ............................................146

Figure 6.13 Harmonic impedance with full load damping and two times impedance..146

Figure 7.1 5th harmonic impedance seen by load as percentage of total impedance ....153

Figure 7.2 5th harmonic voltage as percentage of level at end of distributor................154

Figure 7.3 Indicative schematic layout of a typical radial distribution system.............161

Figure 7.4 Schematic layout of Katoomba distribution system feeding Customer A ...164

Figure 7.5 Schematic layout of Customer A’s load ......................................................165

Figure 7.6 Calculated harmonic spectrum of Customer A’s installation at PCC ..........165

Figure 7.7 Calculated acceptable harmonic spectrum for Customer A’s installation ...169

Figure 7.8 Full load cycle of 5th harmonic rms voltage and current .............................174

Figure 7.9 Half load cycle 5th harmonic rms current for various loading.....................174

Figure 7.10 Springhill sub-transmission substation single line diagram ......................177

Figure 7.11 33kV harmonic impedance for 6 different capacitor combinations (no

damping) ...............................................................................................................180

Figure 7.12 Capacitor switching vs 5th harmonic current at 33kV bus.........................181

Figure 7.13 Capacitor switching vs 5th harmonic voltage at 33kV bus ........................181

Figure 7.14 33kV harmonic impedance for 6 different capacitor combinations with

damping due to 30% loading ................................................................................182

Figure 7.15 Customer B allocated and measured maximum harmonic current. ...........185

Figure 7.16 33kV network harmonic distortion as a percentage of HV planning level 186

xvii
 List of tables

Table 2.1 Voltage and current distortion levels from [17]..............................................23

Table 2.2 IEEE 519 current distortion limits for general distribution systems [12] .......31

Table 2.3 ANSI/IEEE 519 voltage distortion limits [12]................................................31

Table 2.4 Planning levels for harmonic voltages in MV, HV and EHV systems [14] ...33

Table 2.5 AS 2279.2 Harmonic voltage ratio limits at any point on the system [11].....33

Table 2.6 k and α for summation law using uniform distribution of amplitude and phase

[49] ..........................................................................................................................37

Table 3.1 Current waveforms of common domestic loads .............................................43

Table 3.2 Harmonic components of currents from common domestic loads .................44

Table 3.3 Typical supply voltage waveform for Tables 3.1 and 3.2...............................45

Table 3.4 Fundamental and 5th harmonic current 95th percentile values ........................49

Table 3.5 Example values of sampled data.....................................................................53

Table 3.6 Results of monitoring at Homepride for fundamental and 5th harmonic ........57

Table 3.7 Comparison of calculated and measured 95th percentile values .....................60

Table 4.1 Allocation constant k using different calculation methods.............................77

Table 4.2 Comparison of different allocation schemes for the extreme case .................79

Table 5.1 Harmonic monitoring instrument positions ....................................................95

Table 5.2 95th percentile values for all sites over initial monitoring period .................113

Table 5.3 Maximum values for all sites over initial monitoring period .......................116

Table 5.4 Harmonic voltages levels normalised to recommended limits .....................117

Table 5.5 Harmonic voltage 95th percentile values for August during monitoring

(1999 − 2001)........................................................................................................118

Table 5.6 Harmonic current 95th percentile values for August during monitoring

(1999 − 2001)........................................................................................................118

xviii
Table 5.7 Fundamental and 5th harmonic current 95th percentile values (A)................121

Table 5.8 Correlation coefficients for initial period of monitoring programme ...........124

Table 5.9 Weekly 95th percentile values for Homepride zone substation 5th harmonic

voltage ...................................................................................................................125

Table 5.10 Daily 95th percentile values for Homepride zone substation 5th harmonic

voltage ...................................................................................................................125

Table 7.1 Indicative 5th harmonic voltage levels for typical power system..................162

Table 7.2 5th harmonic voltage allocation for scenarios (i)-(iii) ...................................173

xix
 Chapter 1

Introduction

1.1 Harmonics in distribution systems

Harmonic distortion has been identified as existing within power systems since the

advent of alternating current electricity distribution systems [1]. Harmonic disturbances

are also possibly the most researched and published of all the power quality

disturbances, especially over the past two decades. The interest in harmonics can largely

be attributed to the advancing power electronics usage and growth in the number of new

load technologies connected to distribution systems that are sensitive to excessive

harmonic voltage levels. The increased penetration of these new technologies is for the

most part driven by the desire to have more energy efficient equipment and greater

control of equipment operation. Increased efficiency and control typically means the

equipment includes some type of semiconductor based power electronic front end. The

same new technologies, in general produce increased levels of harmonic current

emissions that in turn raise the levels of harmonic voltages existing on the system. In

conjunction with this the growing trend of electricity being sold as a product under

contractual agreements means that guaranteeing an acceptable level of power quality

disturbances is becoming of greater importance to both electricity distribution network

service providers (DNSP) and customers alike.

Indications in the literature suggest that harmonic voltage levels in transmission and

distribution systems are increasing [2-4]. Increasing levels of voltage harmonics creates

a concern for customers with sensitive equipment and also for electricity DNSPs that

face increased losses throughout their power systems. The increased losses burden

DNSPs with additional costs that have to be recovered to ensure profit margins are not
1
decreased. There are suggestions by [5] that in Canada and some major European

countries the cost of network modifications and losses due to harmonic distortion

effects may represent as much as 1% of a country’s gross domestic product (GDP).

As stated earlier it is well known that harmonic voltage distortion within power systems

is due to the interaction of harmonic current emissions from a variety of sources and the

impedance of the power system. The sources of harmonic emissions can include large

industrial loads connected at transmission or distribution levels, smaller commercial and

residential distribution system customer loads, distributed generation installations, the

power system itself, and to a lesser extent traditional generation equipment.

Harmonic current emissions originate from all types of non-linear loads. Non-linear

loads are broadly classified as loads which draw non-sinusoidal current even when the

supply voltage is perfectly sinusoidal [6]. Non-linear loads include saturated magnetic

circuits, such as those in power system transformers and rotating machines, arc

furnaces, fluorescent lighting and of course power electronic loads. Power electronic

loads by far are the most significant harmonic contributors relative to the amount of

energy they draw. Some of the more common power electronic loads include

i) Switch mode power supplies - present in computers, televisions, microprocessors

ii) Rectifiers – present in dc motor drives, regulated power supplies, battery chargers,

iii) Inverters – present in variable speed ac drives,

iv) Static VAr compensators,

v) Cyclo-converters, and

vi) High voltage DC transmission converters.

2
Mitigation techniques are often required to suppress the levels of harmonic emissions

resulting from the above mentioned power electronic equipment.

1.2 Effects and problems caused by harmonics

Harmonic distortion can have both short-term and long-term effects on distribution

system equipment and connected customer loads. Short-term effects are mainly

concerned with immediate damage, equipment malfunction, and the associated power

losses due to harmonic currents and voltages. Long-term effects include thermal losses

and reduced life span of equipment.

Although harmonic voltage and currents within most distribution systems are quite

small relative to the fundamental component, increased impedance of the network at

harmonic frequencies often gives rise to significant power losses. These power losses

include direct copper losses due to an effective increase in the level of rms current in the

presence of harmonics in combination with increases in conductor resistance due to skin

effect. Increased iron losses due to harmonics are also significant within the iron cores

of most magnetic materials. Iron losses include hysteresis losses, where losses increase

approximately in proportion to the harmonic frequency, and eddy current losses, where

losses increase in proportion to the square of the harmonic frequency. Iron losses result

in reduced efficiency of equipment and also raise core temperatures creating thermal

stresses and degrading insulation levels. Harmonics are also a major cause of metal

losses and dielectric stress in capacitor banks causing additional heating and loss of life

[7, 8]. Insulation stress due to higher peak voltages from harmonics can also result in

cable insulation breakdown and disruption to supply.

3
In some circumstances capacitors can combine with source and load inductance to form

a parallel resonant circuit. In the presence of a harmonic resonance, harmonic voltages

and currents may be amplified. The resulting voltages may highly exceed the voltage

rating of capacitors or other connected equipment and the consequence is capacitor

damage or blown fuses.

Complete or partial load disruption can also occur due to the presence of harmonics.

Harmonics interfere with protective relays, metering devices, control circuits,

communication circuits, and customer electronic equipment. Sensitive equipment can

experience malfunction or complete component failure. Shortened incandescent lamp

lifetime and failure of some types of fluorescent lights have also been recorded in [7, 9].

While some of the above mentioned problems and effects associated with harmonics

occur over short-term durations, more significant long-term problems such as degrading

insulation due to thermal stresses are usually of most concern. These problems are often

not visible to the DNSP or customer until final failure of equipment occurs. Reduced

lifespan of equipment necessitates costly repairs or replacements. Thus it is important

for both DNSP and customer to be proactive with regards to measurement and

mitigation of harmonic distortion problems.

1.3 Representation of harmonics

This section outlines some of the fundamental concepts and important mathematical

relationships for representation and analysis of harmonic distortion in power systems.

The following equations will be referred to throughout this thesis when developing

system models and describing harmonic distortion.

4
1.3.1 Harmonic components
A method to represent any non-sinusoidal periodic function u(t) using an infinite series

of sine and cosine functions and coefficients as shown in equation (1.1) was first

proposed by Baron Jean Fourier in 1822 [10]

∞
u (t ) = A0 + ∑ [ Ah cos(hω 0 t ) + Bh sin( hω 0 t )]
h =1
∞
(1.1)
= A0 + ∑ C h cos(hω 0 t + ψ h )
h =1

where u(t) is a periodic function of frequency f0, angular frequency ω0 = 2π f0, and

period T = 1/f0 = 2π /ω0. C1cos(ω0t+ψ1) represents the fundamental component, and

Chcos(hω0t+ψh) represents the hth harmonic component of amplitude Ch, frequency hω0

and phase ψh relative to the fundamental.

Generally, for power systems the fundamental frequency is either 50Hz or 60Hz.

Australian power systems are typically operated at 50Hz and thus harmonic frequencies

will appear as multiples of 50Hz (100Hz, 150Hz, 200Hz, etc.). The Fourier series

coefficients C1, C2, …Ch and relative phases ψ1, ψ2, …ψh make up the harmonic

spectrum of the waveform and are found using equations (1.2) to (1.6)

1 T 1 2π
A0 = ∫ 0 u (t )dt = ∫ 0 u (t )dx, where x = ω0t (1.2)
T 2π

2 T 1 2π
Ah = ∫ 0 u (t ) cos(hω 0 t )dt = ∫ 0
u (t ) cos(hx)dx (1.3)
T π

2 T 1 2π
Bh = ∫ 0 u (t ) sin( hω 0 t )dt = ∫ 0
u (t ) sin( hx)dx (1.4)
T π

Ch = Ah2 + Bh2 (1.5)

 Ah 
ψ h = tan −1   (1.6)
 Bh 

5
Conversely, if the harmonic spectrum of a given current or voltage waveform u(t) is

known the original waveform can be constructed using the Fourier series summation

∞
u (t ) = ∑ U h cos(hω 0 t + ψ h ) (1.7)
h =1

where Uh is the hth harmonic peak current or voltage, ψh is the hth harmonic phase, ω0 is

the fundamental angular frequency, ω0=2π f0, and f0 is the fundamental frequency,

typically 50Hz.

1.3.2 Total harmonic distortion (THD)

Although the harmonic content of a power system may be quite small relative to the

fundamental in most circumstances, for exactness the rms value of a current or voltage

waveform requires the harmonic content to be considered such that

2
 1∞

U rms = ∑  U h  (1.8)
h =1  2 

where Urms is the rms value of voltage or current. The rms voltage or current can also be

used to quantify the level of distortion of the waveform. The total harmonic distortion of

voltage or current waveform (THDU) is calculated using equation (1.9)

2
1 ∞
U 
THDU =
U1
∑ U =  rms
2
h
 − 1 (1.9)
h=2  U 1 rms 

where THDU represents voltage or current total harmonic distortion (alternatively

represented as THDV and THDI respectively) and U1rms is the rms fundamental voltage

or current. Alternatively rms voltage or current can be represented in terms of total

harmonic distortion

∞
U rms = ∑U h =1
2
h rms = U 1 rms 1 + THDU2 (1.10)

6
As distribution system fundamental voltage and current rarely remain static in

magnitude at different times throughout the day, the definition for total harmonic

distortion may at times provide a misleading value for the harmonic distortion level [2].

This is especially true for distribution system fundamental currents that fall close to zero

at certain periods of the day, resulting in large values of THDI. For this reason a

modified index for harmonic distortion may be used with the harmonic content of the

waveform expressed as a percentage of a fixed nominal value rather than the

fundamental value, giving total demand distortion (TDDU)

∞
1 1
TDDU =
U nom
∑
h=2 2
U h2 (1.11)

Total demand distortion is specified in [6, 11] instead of THD for harmonic current

distortion for the above mentioned reasons. The fixed value Unom is required to be

specified and may be a maximum rms value, maximum demand, average or selected

nominal system value.

1.4 Harmonic monitoring programmes

Until recently electricity DNSPs in Australia have been primarily reactive with regard to

harmonic distortion problems. Harmonic measurements are typically not undertaken

unless a problem is reported which cannot be attributed to conventional distribution

system problem sources, such as high or low voltages and power interruptions. Often

DNSPs are unaware of growing levels of harmonic distortion within their power

systems until a problem occurs.

7
By maintaining a proactive approach to harmonic distortion problems, DNSPs can

identify problematic loads, better plan the layout and operation of their systems to help

reduce harmonic distortion levels, and ensure an acceptable level of power quality is

being supplied to their customers. With contractual agreements between transmission

utilities, DNSPs, and customers becoming commonplace, a greater emphasis is being

placed on the level of power quality disturbances by regulators to ensure all parties

maintain compliance with the relevant power quality standards.

To be proactive, the best approach for DNSPs to take with regard to harmonic distortion

is to carry out regular power quality monitoring programmes on various sections of their

power systems. Identifying problems areas early allows planning engineers to prioritise

work to ensure harmonic levels remain below recommended levels. Identification of

excessive harmonic levels may also indicate a nearby harmonic resonance. If harmonic

resonances are recognised at an early stage, capacitors and other power system

equipment may be prevented from suffering reduced life spans through proper

mitigation, effectively creating savings on equipment replacement or repair costs.

Continuous monitoring of harmonic levels in distribution systems allows an assessment

of present levels of harmonics to be made, keeping check on the rate at which levels are

increasing or decreasing. This is important information for planning new distribution

systems or upgrades. Levels can also be benchmarked against the acceptable levels

according to the relevant power quality standards. DNSPs should have a procedure for

implementing harmonic monitoring programmes to ensure efficient processing of data

and effective reporting of results. Within Australia there have been only a few

monitoring campaigns published to allow an assessment of how harmonic levels are

8
growing in typical distribution systems. This may be partially attributed to the cost of

instrumentation, time, labour, lack of requirements from regulators, and complexities of

metering in carrying out such surveys.

1.5 Harmonic standards

To properly manage harmonic distortion levels in distribution systems there are two

approaches that may be taken: either a DNSP controls the level of harmonic emissions

by implementing emission limits on their customers, or they install mitigation

techniques as required, such as lower impedance conductors and harmonic filters, to

limit voltage levels and effects of excessive harmonic voltages and currents.

To implement restrictions on emission from customers, DNSPs require legal

documentation that can be used to regulate customers. Thus, selection of an appropriate

dedicated harmonic emission standard is required. In this way standards control

harmonic levels emitted from customers to ensure no excessive voltage distortion

appears within the distribution system. In Australia the standard that limits harmonic

emissions from customers connected to the power system has been AS 2279.2-1991

[12] in various forms, for a period of 25 years. This standard provided a simple method

of determining acceptable sizing of converter installations as well as providing a guide

for assessing emission limits for other harmonic producing installations. The application

of AS 2279.2 is reasonably straightforward and customer equipment can be relatively

easily assessed for harmonic compliance, i.e. whether the customer is emitting

acceptable levels of harmonic distortion or not.

9
With the dramatic increase in the use of high efficiency equipment and other power

electronic load technologies there now exists more significant harmonic producing

loads in distribution systems than ever before. These loads have an enormous amount of

diversity in their design and operating conditions. The AS 2279.2 harmonic standard

does not include the detail required to account for the extreme diversity of loads, and

thus Australian Standards looked to the international community to determine a new,

more applicable standard.

Two major international standards were being used by the international community for

assessment and guidance in the area of power system harmonics when a replacement for

AS2279.2 was required at the turn of the century. The first applicable standard, IEEE

519:1992 [13], has it origins in the USA. However, this standard provides assessment

and guidance based on typical data from distribution systems in the USA, which are

typically meshed systems consisting of different transformer configurations and

customer connection terminations than that of typical Australian distribution systems.

Additionally, the operating frequency of power systems in the USA is 60Hz, though this

does not have a large bearing on the application of the IEEE 519 standard.

Australian distribution systems are typically radial in nature with the exception of some

networks existing within central business districts (CBD). Feeder lengths are also

typically much longer than in the USA. Australian distribution systems are actually a

closer match to European systems with their transformer configurations and operating

voltages. For this reason and due to the World Trade Organisation agreement on

technical barriers [14], Standards Australia has adopted the internationally accepted

technical report IEC 61000-3-6:1996 [15] as the new harmonic standard, identified as

10
AS/NZS 61000.3.6:2001 [16]. The above IEC technical report was originally published

as a guide only, with recommendations on how to manage harmonic distortion in power

systems, rather than to be implemented as a clearly defined harmonic standard, as was

the case with AS 2279.2.

The increased complexity of the new harmonic standard and some of the pitfalls in

creating that complexity form the basis for the research reported in this thesis. A more

detailed assessment of AS/NZS 61000.3.6 (IEC 61000-3-6) and AS 2279.2 is provided

in Chapter 2. The research presented attempts to provide a coordinated approach to

applying IEC 61000-3-6 and expanding some of its keys concepts to use in the

management of harmonic distortion in distribution systems.

1.6 Objective of the thesis

The aim of this thesis is to develop a methodical and comprehensive approach to

harmonic analysis that may be used by a DNSP during the planning or design stage of a

distribution system to estimate the harmonic performance. As knowledge of the

harmonic nature of future loads during the planning stage will be relatively undefined,

analysis tools require pragmatic estimation techniques to broadly identify the

performance of the system. Thus before analysis tools are developed a study must be

undertaken to identify the properties of typical distribution system loads.

The first objective of this thesis is to determine the macro-characteristics of the

harmonic emission behaviour within power distribution systems. Macro-characteristics

of loads are required to enable pragmatic modelling approaches to be applied to

distribution systems during the design phase. This is required to determine the

11
harmonics capabilities of a particular distribution design. Identifying the macro-

characteristics of different load types is achieved in this thesis through the observation

of results from a benchmark harmonic monitoring programme and complementary

simulations.

The second objective is to provide harmonic planning techniques for DNSP engineers,

which can be implemented on general distribution systems to control the level of

harmonic voltage distortion. This includes a suitable method to allocate allowable

harmonic distortion contributions to customers connected to a power distribution

system. This is achieved through generalising procedures suggested by the relevant

standards and also realising the important variables that determine the harmonic

capabilities of the power distribution system.

A method to predict the level of harmonic distortion in a distribution system will also be

presented. Analysis techniques for accurate modelling of the distribution system in the

presence of harmonics are developed to allow for such a prediction. Forecasting and

quantification of the growth of harmonics at particular points in the system, taking into

account the development of industrial technology over the next decade and development

of general guidelines for identifying the most suitable connection point for future large

distorting loads may be addressed utilising such a prediction technique. General

planning strategies for maximising the harmonic capability of distribution systems will

also be investigated.

12
1.7 Structure of thesis
A brief summary of the contributions of each of the remaining Chapters of this thesis is

provided below.

Chapter 2: In this chapter a literature review on the present situation of harmonics in

distribution systems throughout the world and trends in growth from harmonic

monitoring programmes are presented. The various strategies employed by standards to

control harmonics are discussed. An introduction to and history of the summation law, a

principle concept used in this thesis is presented.

Chapter 3: A method to estimate the level of voltage harmonic distortion existing in

an MV power distribution system is proposed. The suggested method is based on

macro-modelling where the large amount of data required for cumbersome detailed

simulations is not available. The advantages and limitations of this method are

discussed.

Chapter 4: A new approach to the allocation of harmonic emissions to customers on

long MV feeders is developed. A sensitivity analysis of the method is also determined.

The proposed method allows the harmonic standard IEC 61000-3-6 to be implemented

when only a limited amount of data is available, specifically for the typical yet complex

example of loads connected along a feeder having significantly different fault levels.

Chapter 5: The global behaviour of multiple harmonic loads connected to a power

distribution system is analysed. The procedure for implementing a benchmark harmonic

monitoring programme of a distribution system is presented. An outline for presentation

13
of results and discussion of relevant outputs are included. Identification of MV

distribution system aggregated load models is determined using data from field

measurements.

Chapter 6: This chapter provides a review of the complexities involved in allocation

methods when significant system capacitance or power factor correction (PFC)

capacitors are present within the distribution system network. An assessment of some

simplifying rules is also provided.

Chapter 7: The application of harmonic allocation methods to three real case studies

is provided. The three case studies provided include an example of allocation of

harmonic emissions to a customer situated along a long MV feeder, assessment of

existing harmonic distortion levels within a distribution system, and an evaluation of a

system containing numerous PFC capacitors.

Chapter 8: The final chapter will summarise the significant conclusions from this

thesis and suggest areas of future work to complete the research.

14
 Chapter 2

Overview of harmonic monitoring in distribution systems

2.1 Introduction
There is a general concern within distribution network service providers (DNSP) about

the growth of harmonic distortion in power systems. However, as excessive harmonic

distortion levels usually only cause gradual degradation of power system equipment and

loads, effective warning on problems associated with harmonic distortion can often only

be achieved through a careful programme of monitoring. Such extensive harmonic

monitoring programmes are both costly and time-consuming for a DNSP to undertake.

Optimising the effectiveness of harmonic monitoring programmes is thus of interest to a

DNSP.

This Chapter reviews literature on harmonic monitoring programmes that have been

undertaken both nationally and internationally on transmission and distribution systems

in an attempt to identify the state-of-the-art practices in monitoring campaigns. Of

particular interest is establishing the optimal method of measurement, identification of

load types monitored, and effective presentation of results.

For a DNSP to effectively manage harmonics a projection of future harmonic levels is

required. To evaluate growth of harmonic distortion over the next decade or so a model

of distribution system load increases and changes in load technology are required in

conjunction with a benchmarking of present levels of harmonics. This Chapter includes

presentation of monitoring results from various distribution systems to help benchmark

existing harmonic levels. In the few international instances where comprehensive

15
monitoring programmes have been completed there is a general consensus of steady

harmonic distortion growth.

Finally an introduction to the harmonic summation law is presented. The summation

law is a statistical tool used to approximate the net magnitude of contributions from a

number of harmonic sources incorporating both time and phase diversity. The

summation law is a principle concept used in this thesis and provides the backbone of

the theory developed in the remaining chapters.

2.2 Harmonic monitoring programmes

The number of harmonic monitoring programmes completed internationally has

increased gradually over the past decade. To allow the results of such monitoring

programmes to be analysed effectively and measured against previous or future

programmes there are a number of factors that should be considered to ensure the

integrity of the resulting conclusions. To be included in these considerations are the

following

(i) Method of measurement,

(ii) Presentation of results, and

(iii) Load or source type identification.

The cited literature on harmonic monitoring at both national and international levels has

indicated the need to standardise this practise to ensure the ability to include monitoring

results into future references. While some of the published monitoring programmes are

carried out to establish existing harmonic distortion levels, usually monitoring

programmes result from a complaint by a customer or a specific investigation by a

16
DNSP. While satisfying the requirements of the initiator the published results of these

programmes are of limited use for future reference. The following sections provide

discussion on (i)-(iii) above to help ensure the full potential of harmonic monitoring

programmes are achieved.

2.3 Method of measurement

The method of measurement used in the cited literature, spanning the last decade, has

varied depending on the requirements of the harmonic monitoring programmes. The

first consideration to be emphasised by the literature is the type of monitoring

equipment used. It is important to laboratory test measurement devices before being

used in the field to establish device accuracy and consistency. Regular calibration is

often not the only requirement to establish instrument accuracy, the algorithms used by

instrument manufacturers to calculate the harmonic components often need to be

verified as indicated in work by [17]. Figure 2.1 illustrates a typical monitoring system

for a simple harmonic monitoring programme.

Phase Conductor

Potential Current
Transformer Transformer

Waveform Capture & Data Storage &

Data Processing Unit Presentation

Figure 2.1 Symbolic illustration of a harmonic monitoring system

It has been widely accepted through the literature that metering current transformers and

potential transformers are the most appropriate points of measurement. Metering current

transformers are nearly always available at both feeder and load connection points and

their bandwidth is usually sufficient to measure up to the 50th harmonic. However, it has

17
been found that some harmonic instruments do not measure accurately to such higher

order harmonics and thus care is to be taken [17]. In the case where potential

transformers are not available capacitive voltage dividers have been shown to give good

results for voltage harmonic measurements [18]. Capacitive voltage transformers tuned

to the fundamental frequency however are not suitable for harmonic measurements.

Harmonic distortion measurements are usually determined from a single cycle or

averaged over a small number of cycles utilising Fourier analysis. The recently

introduced IEC 61000-4-30:2003 [19] specifies 8-10 cycles to be used to establish

harmonic components but it is anticipated that it will be a while before instrument

manufacturers comply with this standard. The Fast Fourier Transform (FFT) technique

is used to determine the magnitude and phase of each of the harmonic components from

the sampled waveform. Instrument manufacturers do not usually specify the details of

the sampling or algorithms used for harmonic analysis. This information is often

important when measurements from different instruments are being compared.

Instrument sampling times must also be selected carefully. IEC 61000-3-6 [15] and

IEC 61000-4-7:2002 [20] specify that both 3 second and 10 minute samples are

required, however present harmonic monitoring instruments are often unable to provide

such data over extended periods due to memory limitations. To establish good weekly

trends of harmonics, measurements should be taken at least every half hour. Previous

surveys have sampled the system at 1-3 minute intervals to establish more accurate

trends [21, 22]. Where possible measurements should always be taken of magnitudes of

individual voltage and current harmonics rather than overall THD, as THD rarely

portrays a clear picture of problem areas. The most common harmonics from the

18
literature cited include the 3rd, 5th and 7th. Measurements should also be completed in

each of the seasons to establish seasonal variation [23].

2.4 Presentation of results

Presentation of results has been often a topic for discussion in the cited literature [18,

24, 25]. The first argument raised is whether to use the fundamental or rated quantities

as a base for harmonic distortion calculation. From the literature cited, few instances

present harmonic distortion using rated quantities as a base for voltage and current as

per equation (1.11) [18, 24]. These monitoring programmes were completed early in the

last decade with most recent programmes conforming to the IEEE 519 standard [13]

using total harmonic distortion as shown by equation (1.9).

The choice of which of the above quantities to use for the harmonic distortion

calculation is to be resolved in the newly drafted standard by IEEE [26]. This standard

suggests the use of nominal system rms voltage for voltage harmonic distortion

calculations, and maximum demand load current for current harmonic distortion

calculations. By using the maximum demand load current instead of the fundamental

current the problem of high harmonic distortion values resulting during low demand

periods is eliminated.

Data collected from harmonic monitoring programmes needs to be analysed on a

statistical basis to extract meaningful figures. The existing definitions and benchmark

data analysis techniques used for processing collected data from harmonic benchmark

monitoring programmes are outlined in [25].

19
The cumulative frequency curves and histograms are good tools to establish the extent

of harmonic problems. The 95% cumulative probability value (CP95) is often used to

provide the level of harmonic distortion for comparisons with other results. CP95 is the

statistical quantity representing the harmonic distortion value that is greater than 95% of

data obtained for each site s found using equation (2.1).

CP 95s

∑f s ( xi )
−∞
∞
= 0.95 (2.1)
∑f
−∞
s ( xi )

where xi is the steady state harmonic distortion measurement number i, and fs(xi ) is the

probability distribution function of the sampled harmonic (or THD) values for site s.

For presentation of harmonic monitoring results as an index the CP95 value requires

two additional descriptors, the type of monitoring location that represents site s, and the

duration from which the samples were obtained, e.g. one week.

To obtain a harmonic index for a system from within which a number of measurements

have been taken at different locations, [25] suggests combining each site to get the

system THD index (STHD95) by obtaining the 95% cumulative probability value using

the individual site CP95 values and equation (2.2).

STHD 95
∑ f t (CP95s ) × Ls
−∞
∞
= 0.95 (2.2)
∑ f t (CP95s ) × Ls
−∞

where Ls is the connected kVA served by site s, and ft(CP95s ) is the probability

distribution function of individual site CP95 values. An improved method for

combining sites is suggested in [27] by normalising all sites to an appropriate limit, but

this technique is still in the preliminary stages of development.

20
There are a number of other relationships given in [25] that are useful concepts for

analysing harmonic survey data. Use of statistical methods allows the results to be

evaluated effectively without the interference of harmonic transients that can distort

results. Of the literature cited only a few report results in a form that addresses the issue

effectively [21, 28].

To adequately benchmark harmonic levels, results should contain at least the maximum,

minimum, and CP95 values for the individual voltage and current harmonics, and also

the total harmonic distortion in tabulated form.

2.5 Load type identification

If harmonic monitoring results are to be used to predict future levels of harmonic

distortion in distribution systems, the load types from which results are obtained must

be recorded. Results from measurements at a point of common connection (PCC)

should also specify the nature of the loads represented at that PCC. A good example of

this is given in [24], which specifies the location, plant type (i.e. aluminium plant),

shifts worked, significant load type, and equipment ratings.

In [29] data is distinguished using three categories of load sectors; industrial (I),

commercial (C) and residential (R). Sites within the monitoring programme are

classified using the particular type of load (it may be a combination of two, i.e.

industrial and residential). Typical industrial loads are given as lighting, electric motors,

power electronic devices etc. while commercial loads are lighting, elevators, computers

etc. Residential loads are defined as lighting lamps and household appliances. This

method of identifying load types is the most common and the recommended method to

21
use. The usefulness of this method is reflected by the nature of the typical loads for each

category. A distribution feeder can be easily categorised by the industrial, commercial

and residential ratio if the customer loads are known.

A third way of identifying the load type is by the current waveform the load draws.

Loads can then be comprised of linear loads (sinusoidal current), or one of three classes

of non-linear loads. This method is utilised in [30], noting that method is usually used

for estimations in forecasting harmonic growth rather than in the harmonic monitoring

programme phase. This method is an extension of the industrial, commercial and

residential method in that it categorises the current waveform from a particular load

type, and the load type can then be assigned as industrial, commercial or residential.

2.6 Previous monitoring programmes

Throughout Europe, Asia and USA a number of distribution system harmonic

monitoring programmes have been completed. The purpose of these programmes has

been to monitor the existing levels of harmonic distortion on the distribution systems,

while also establishing the growth patterns of the levels of harmonic distortion for the

future. None of the reported programmes however offer coordinated voltage and current

measurements across a number of sites at various intervals throughout a year.

Harmonic monitoring programmes in Australia have been limited usually only to areas

of customer complaint. However, this trend is changing as indicated by DNSP

participation in a pilot study by [31]. Due to the lagging nature of Australian business in

taking up new commercial and industrial technologies the harmonic growth found

22
overseas paints a reasonable picture of the future of harmonic growth on a national

level.

Harmonic monitoring programmes performed around the world over a decade ago

showed that harmonic levels were relatively low and were of little concern due to the

tolerance of the existing electrical equipment. A programme in the USA presented in

[18] involved four years of data collection. The results of [18] shown in Table 2.1

highlights that at most of the sites monitored, harmonic distortion was well within the

recommended limits [13]. Only at one site were the limits exceeded and this was due to

the circuit resonating at the 5th harmonic because of customer capacitors. Results show

that industrial and commercial loads typically give rise to higher harmonic distortion

than residential loads.

Table 2.1 Voltage and current distortion levels from [18]

Location Average Dominant Average Dominant
voltage voltage current current
distortion harmonic distortion harmonic
level (%) and level (%) and
distortion distortion
level (%) level (%)
Residential (Res) 1.01 5th /0.54 2.58 5th /1.58
Commercial (Com) 1.92 5th /1.71 3.13 3rd /2.65
Industrial (Ind) 2.53 5th /2.26 2.79 5th /2.48
Res & Com 1.14 3rd /0.89 3.84 5th /2.72
Industrial 2.29 5th /1.06 2.52 5th /1.41
Res & Com 1.68 5th /1.49 2.36 5th /1.72
Res, Com, & Ind 4.15 5th /3.93 15.33 5th /14.97

Another harmonic monitoring programme in the USA by [24] included measurements

from 37 substations and 39 individual load sites. Of the sites monitored 48.7% were

distribution substations, 14.5% heavy industry, 30.3% light industry (less than 5MVA),

and 6.5% were commercial sites. As expected the principal harmonics were the odd

23
harmonics ranging from the 3rd to 13th. The substations primarily serving industrial and

commercial sites had rather high values of current distortion as compared to the

substations primarily serving residential customers, which had less than 5% current

distortion. As with the results from [18] the above results suggest that the industrial,

commercial, residential load classification method is useful. The monitoring programme

in [24] also found that some distribution system sites were already exceeding IEEE

recommended limits for current harmonics. Four of the 39 test sites were found to

exceed the limits while at least another ten were close to exceeding the limits. No sites

exceeded the voltage THD limit of 5% although three sites had values greater then 4%.

The monitoring programmes discussed in [18] and [24] looked at the levels of harmonic

distortion only. Other programmes were interested in the trends of the harmonic

distortion over periods of time. The usual patterns of harmonic distortion levels are

similar to that of the system loading in that each 24 hours will display peak periods at

particular times. Over a period of one week, harmonic monitoring results of [21] in the

USA found that these repeatable patterns were very similar during weekdays with

weekend levels providing their own individual patterns. An example of a weekly

harmonics trend is given in Figure 2.2 that includes the associated histogram illustrating

the frequency of the distortion magnitudes.

24
 3rd Harmonic
3rd Harmonic
Max
140
Min 160000
Weekend
120
140000

Reading Occurence
100 120000
Vrms (Volts)

100000
80 Switched
80000
60 60000 Capacitor
40000
40
20000
20 0
0 20 40 60 80 100 120 140 160 0 30 60 90 120 150 180 210
Vrms (V)
Time (hours)

Figure 2.2 Seven day recording of harmonic voltages and histograms [21]

The levels of THD monitored at distribution feeders presented in [21] were found to be

well within the US guidelines with levels lower than 2% for 99% of the time. Results

from [21] also highlight the effects of switching of PFC capacitors on the individual

harmonics. These effects of switching capacitors are made evident by the resulting

peaks in the histograms of presented data in Figure 2.2.

Further monitoring programmes have continued in the USA [28] to highlight the

behaviour of harmonic distortion at customer bus locations. The monitoring

programmes completed at small industrial and commercial sites found that they

displayed similar daily trends to that of earlier programmes. In some instances the 3rd

harmonic demonstrated higher levels than the 5th harmonic. The 3rd harmonic voltage

also indicated a growth from 0.25% to 1.0% over the period of a decade. Although this

would not be the case in Australia due to the use of delta-star distribution transformers,

it highlights the possible dramatic growth levels of harmonic distortion.

Results of a harmonic monitoring programme carried out on a distribution system in

United Kingdom are presented in [32]. This covered monitoring conducted on industrial

loads over a period of two weeks with samples of the supply voltage at 11kV and 33kV
25
taken every 20 seconds. Measurements were taken first to assess the background

voltage distortion and then the interaction between the background distortion and

particular distortion producing loads were assessed. The results of the monitoring

showed that the levels for each of the dominant harmonics i.e. 3rd, 5th (most dominant)

and 7th had their own repeatable daily pattern. Both the 5th and 7th harmonics produced

similar daily trends. Peak values of individual voltage harmonics were shown to be

around 2% of the fundamental voltage. The collected results were used by [33] to

produce mathematical models which allowed the prediction of the resulting individual

harmonic magnitude and phase for the connection of a distorting load to a distribution

system with known levels of background harmonic distortion.

Monitoring data obtained from residential loads in France [34] indicated that household

appliances are a major contributor to harmonic distortion levels. The conclusion is

drawn that harmonic distortion limits of today will soon be exceeded if emission levels

of household appliances are not reduced. It is suggested that the level of THD is relative

to the number of harmonic loads rather than the total power rating of the distorting

loads.

A major monitoring programme conducted by [23] fours years after [32] involving 46

different substations in the United Kingdom indicated that levels of harmonic distortion

on distribution systems had increased. The most significant individual harmonic, the 5th

harmonic, was found to have typical peak levels of 2.5%-3.0% at some locations. This

was mainly due to the increased use of power electronic devices connected at lower

voltages.

26
Monitoring of harmonic voltage in Northern Taiwan also indicates the same growth

patterns as in other areas around the world. Taiwan power utilities are finding that

acceptable harmonic distortion levels are already being exceeded with levels of THD

above 3% on 50% of the transformers surveyed [35]. The effects of the high harmonic

distortion levels are being noticed through damage to distribution system capacitors.

Power transformers have also had to be derated due to the effects of harmonic currents.

The need for Australian industry to analyse their harmonic problems in a joint effort

with their DNSP is proposed in [22] to ensure that future levels do not grow to the

extent that will cause sensitive equipment malfunction. The effective use of harmonic

filters is also reported in [22] where harmonics in the supplying 132kV system were

found to be higher than that of the industrial customer 33kV distribution system where

filters were installed. It is stated in [22] that harmonics are often not usually considered

when PFC capacitors are installed within industrial plants.

A monitoring programme carried out on the electrical distribution system of several

commercial buildings in the USA is reported in [36]. The method of monitoring was to

first take initial measurements using true-rms and averaging multimeters. The true-rms

meter provided measurements of the peak and rms current, while the averaging

multimeter provided measurement of the fundamental component. The crest factor and

distortion factor for each of the sites was then calculated using equations (2.3) and (2.4).

I peak
Crest Factor = (2.3)
I rms

I rms
Distortion Factor = (2.4)
I1

27
From the calculations, sites with high crest or distortion factors were selected for further

analysis using a power analyser to determine the levels of harmonics. This method can

save time when determining where high levels of harmonic voltages of currents may be

present, and may help determine where to locate permanent monitoring devices. The

results of this survey found that the sites had voltage THD values between 2.0% and

4.5%.

Throughout all harmonic monitoring programmes in the cited literature it has been

found that the 5th harmonic was the most significant. Due to this, THD usually followed

the pattern set by the level of 5th harmonic. Also noticeable in the studied literature was

the suggestion that levels of harmonic distortion were increasing at a rate so that present

day harmonic tolerances would soon be exceeded. The changing nature of loads

connected to distribution systems is also a concern for DNSPs.

2.7 Future projection of harmonics

Growth of harmonic voltages in power systems over recent years is to be expected as

harmonic distortion problems are largely derived from power electronic devices, which

are predominant in existing growth technologies. Load growth studies by utilities in the

USA in the late 1990s expected that over 50% of power system loads would be supplied

through power electronics systems by the year 2000 [37]. Some of the cited literature

suggests that acceptable harmonic distortion levels stipulated by various standards may

be exceeded in some parts of the power system in the years 2000 to 2010 [21, 28].

Measurement of growth of harmonic distortion levels requires well coordinated long-

term monitoring programmes to be established. These types of studies are becoming

possible through the introduction of permanent power quality monitoring instruments

28
installed at substations and the increase of metering devices containing power quality

monitoring functionality.

Future projections should be based upon data obtained from previous harmonic

monitoring programmes and assumptions on existing load development and load growth

models. Load growth estimations should consider technologies such as adjustable speed

drives (ASD) in air conditioners and washing machines, televisions, personal

computers, and future loads such as electric vehicle battery chargers [38, 39]. The

increase in numbers and loadings of households and industrial sites should be

incorporated into the forecast. The sensitivity of the distribution system being studied is

one of the major concerns when attempting to predict the effects of harmonic distortion

growth.

Attempts have been made to project the level of voltage THD for typical distribution

system feeders supplying typical residential, commercial and industrial customers over

a period of a decade or so [30, 40]. Of a major consideration is the harmonic

susceptibility of the distribution feeders. Some projections of distorting load growth

have suggested that short feeders are more immune to the effects of harmonic currents

[40]. Based on this and the associated effects of distribution system capacitor

connections it has been shown that for short feeders with no parallel resonances, voltage

THD growth could be as low as 0.01% per year [40], while for longer feeders with high

harmonic susceptibility, growth could be as much as 0.35% per year.

A study reported in the UK [5], based on harmonic monitoring programmes of

residential areas during 1979-1999, indicate harmonic growth of 1% per 10 years for

29
voltage THD and 1.4% per 10 years for 5th harmonic voltage. However due to

organisational constraints there was no consideration given to seasonal variations and

the measurement periods were not taken at the same time of each year. Thus future

projections based on these results could not be considered reliable.

Other areas that need to be considered when attempting to project harmonic distortion

growth are the harmonic self-compensation effect [41], generalisation of distribution

system loads, i.e. breaking down load into percentage of non-linear load, and

identification of distribution system parameters that affect harmonic susceptibility.

2.8 Relevant harmonic standards

Most countries have their own regulatory standards to control the levels of harmonic

distortion in distribution power systems. More recently a number of countries have

collectively started applying similar harmonic control methodologies and recommended

limits through the adoption of international standards such as IEC 61000-3-6 [15] and

IEEE 519 [13]. Other similar standards exist such as EN 50160 [42], however the IEC

and IEEE standards are the most commonly applied.

In an attempt to control the levels of harmonic voltage distortion within distribution

systems, most standards apply limits to harmonic current emissions in the hope that if

customers are limited appropriately the net effect of all customer emissions will result in

an acceptable level of harmonic voltage distortion. IEEE 519 restricts customer

harmonic current emissions to a value derived from the short circuit level at the point of

connection and the size of customer’s non-linear load. The recommended harmonic

current limits from IEEE 519 are given in Table 2.2. As a consequence of these limits it

30
is assumed the voltage levels will not exceed those given in Table 2.3. However it has

been found by [43] that most "utility versions" (adaptations) of the standard required

customers to also comply with the voltage limits, and thus may be disconnected if they

cause excessive voltage even when their current emissions are within specifications of

Table 2.2.

Table 2.2 IEEE 519 current distortion limits for general distribution systems [13]
Maximum Harmonic Current Distortion
in Percent of IL
Individual Harmonic Order (Odd harmonics)
ISC/IL h<11 11≤ h<17 17≤ h<23 23≤ h<35 h ≥35 TDD
< 20 4.0 2.0 1.5 0.6 0.3 5.0
20-50 7.0 3.5 2.5 1.0 0.5 8.0
50-100 10.0 4.5 4.0 1.5 0.7 12.0
100-1000 12.0 5.5 5.0 2.0 1.0 15.0
> 1000 15.0 7.0 6.0 2.5 1.4 20.0
Even harmonics are limited to 25% of the odd harmonic limits above.
Current distortions that result in a dc offset, e.g., half-wave converters, are not
allowed.
*All generation equipment is limited to these values of current distortion,
regardless of actual ISC/IL
Where
ISC = maximum short-circuit current at PCC
IL = maximum demand load current (fundamental frequency component) at PCC

Table 2.3 ANSI/IEEE 519 voltage distortion limits [13]

Bus voltage at PCC Individual Vh, % Voltage THD, %
V < 69kV 3.0 5.0
69 ≤ V < 161kV 1.5 2.5
V ≥ 161kV 1.0 1.5

The IEC approach differs slightly from the IEEE standard in that it considers future

customers in the harmonic allocation. IEC 61000-3-6 provides formulas to estimate the

allowed current emission for each customer such that all customers, including future

ones, share the harmonic absorbing capability of the system [43]. While the

recommended harmonic voltage levels from the IEC standard are more generous than
31
those of the IEEE standard, the allowable customer harmonic current contributions are

usually more restrictive, although this will depend much upon circuit configuration.

Overall, while being more complex, the IEC standard has a better philosophy in that it

incorporates future loads and also makes customers responsible for harmonic voltages

on the system.

In January 2001, Standards Australia adopted the IEC 61000-3-6 technical report as

AS/NZS 61000.3.6. This new standard supersedes the existing AS 2279.2 that has been

in use in various forms as the harmonics standard for distribution systems in Australia

for the last 25 years. The need for the new standard has come with the more prolific use

of power electronics within distribution systems. The philosophy of AS 2279.2 was

based on there being only a few large non-linear loads within each distribution system.

The allocation of customer's permissible harmonic currents is based either purely on

load size or, in more extenuating circumstances, the amount of harmonic distortion

existing on the system prior to connection of the new load, usually creating a 'first-

come-first-served' scenario as with IEEE 519. The approach of AS 2279.2 does not

adequately allow for future harmonic producing load growth.

The largest difference between IEC 61000-3-6 and AS 2279.2 is that time variation,

leading to diversity, is introduced to account for the multiple types and operating modes

of the non-linear loads with the system [44]. Also harmonic voltage planning levels for

the new standard, given in Table 2.4, are larger at lower frequencies and fall off with

increasing frequency to smaller values. This differs from the AS 2279.2 recommended

limits, which are constant for even and odd harmonics as indicated in Table 2.5.

32
 Table 2.4 Planning levels for harmonic voltages in MV, HV and EHV systems [15]
Odd harmonics Odd harmonics
Even harmonics
non-multiple of 3 multiple of 3
Harmonic Harmonic Harmonic
Order Order Order
voltage Voltage Voltage
h H h
% % %
MV HV-EHV MV HV-EHV MV HV-EHV
5 5 2 3 4 2 2 1.6 1.5
7 4 2 9 1.2 1 4 1 1
11 3 1.5 15 0.3 0.3 6 0.5 0.5
13 2.5 1.5 21 0.2 0.2 8 0.4 0.4
17 1.6 1 >21 0.2 0.2 10 0.4 0.4
19 1.2 1 12 0.2 0.2
23 1.2 0.7 >12 0.2 0.2
25 1.2 0.7
>25 0.2 + 0.2 +
25 25
0.5 h 0.5 h
NOTE – Total harmonic distortion (THD): 6.5% in MV networks; 3% in HV
networks

Table 2.5 AS 2279.2 Harmonic voltage ratio limits at any point on the system [12]
Supply system Voltage at point of Total harmonic Individual harmonic
common coupling voltage ratio voltage ratio
kV % %
Odd Even
Primary and secondary ≤ 33 5 4 2
distribution
Transmission and 22, 33 and 66 3 2 1
sub-transmission ≥ 110 1.5 1 0.5

The selection of an appropriate harmonic standard to regulate emissions from customers

remains the decision of the DNSP. However it is anticipated that acceptance of the new

AS/NZS 61000.3.6 will follow if the difficulties in its application are addressed. An

application guide for the new standard has recently been produced [45] to aid in the

interpretation of principles in AS/NZS 61000.3.6.

33
2.9 The summation law
IEC 61000-3-6 incorporates statistical quantities to assess and allocate harmonic

emissions to distribution customers. One of the key principles of the IEC approach is

the use of summation laws to simplify calculations of net harmonic current from a

number of distorting loads. The summation laws are adopted to account for time,

magnitude and phase diversity of several harmonic loads without completing a detailed

harmonic study. The first summation law is based on weighting factors that depend on

load types, and the second summation law is based on the power law. The second

summation law is a more general approach to combining the harmonic contributions

from a number of loads and is thus considered more applicable in most circumstances.

This summation law relies on the power law to incorporate the diversity of loads

allowing frequency domain studies to predict cumulative probability levels of time

varying harmonics.

Historically, loads that produced significant harmonic emissions were generally limited

to industrial applications of thyristor rectifiers. The limited types of distorting loads

allowed the summation of a number of loads to be completed using the arithmetic sum

of the loads and additional diversity factors, as suggested by [46] in 1967 and later used

in earlier versions of AS 2279.2. This approach was limited in its application and as

more types of harmonic distorting loads were connected to the power system a more

general statistical method for combining a large number of loads was required.

In 1972 [47] extended the application of the recommendations in [46] to include an rms

rule as shown by equations (2.5) and (2.6). The rms rule was used to determine the

34
resulting permissible harmonic current of multiple power electronic installations, as the

arithmetic sum rule seemed too pessimistic for most cases.

Assessed harmonic current = kArms (2.5)

Arms = A12 + A22 + ... + An2 (2.6)

where A1, A2, etc. are the rms harmonic currents of the individual loads and k is adjusted

for the probability of exceedance desired. [47] suggests that the rms rule with k=1.85 be

used for three or more sources producing harmonics with independent phase control,

provided no single load exceeds 55% of the total harmonic load at the particular

frequency. [48] also addressed the problem of the addition of a number of harmonic

sources. Sources with constant amplitude and randomly varying phase angles were

considered showing that the probability of exceedance of the various percentiles (75th,

95th, 99th etc.) could be approximated by the rms rule and various values of k.

0
10

1.41
1.82
Probability of exceedance, P(>kA)
Value of k used with rms rule, k

2.83

25.0%
1
10 15.8%

5.0%

1.0%
0.1%
Max
0 1 2
10 10 10

Number of harmonic sources, N

Figure 2.3 Probability of N harmonics of amplitude A exceeding kA [48]

Figure 2.3 illustrates the percentile exceedances of a Monte Carlo study for the addition

of N harmonic sources with constant amplitude and uniformly random phase angle [48].

35
A comparison to the rms rule using various values of k is also displayed using dashed

lines. It can be seen that the rms rule provides a good approximation to the stochastic

process for values of N greater than 10.

The work of [47] and [48] was complimented by [49] by including random amplitude in

addition to random phase. Unlike the earlier work, [49] used phasors instead of the

instantaneous sine wave components to demonstrate that through the central limit

theorem the resulting distribution of the addition of randomly varying harmonics could

be approximated as a Rayleigh distribution. [49] did not directly contribute to the

summation law but recognised that the nature of the harmonic producing loads should

be first determined to ensure that the requirements of the theoretical work were fulfilled.

Further statistical analysis of the summation of a number of harmonic sources is

presented in [50]. This includes a similar study to [49] with the inclusion of normal

distributions of amplitude and phase, however the rms rule is modified to the form of

equation (2.7). Values of k and α for the 95th percentiles are determined for various

magnitudes of uniformly random amplitude and phase and are provided in Table 2.6.

A = k α ∑ Aiα (2.7)

36
Table 2.6 k and α for summation law using uniform distribution of amplitude and phase
[50]
Range of Range of N=2 N>2
phase amplitude k α k α
angle θi Ai
0-360 0-1 1.0 2.0 1.0 2.0
0.5-1 1.3 2.0 1.3 2.0
1 1.0 1.0 1.7 2.0
0-270 0-1 0.9 1.6 0.9 1.6
0.5-1 1.0 1.4 1.0 1.4
1 1.0 1.0 1.3 1.4
0-180 0-1 0.8 1.3 0.8 1.3
0.5-1 0.9 1.2 0.9 1.2
1 1.0 1.0 1.2 1.2
0-90 0-1 0.9 1.2 0.9 1.2
0.5-1 0.9 1.1 0.9 1.1
1 1.0 1.0 1.0 1.0

Field measurements are included in [50] to establish the types of distributions most

suited for modelling actual systems. Results showed that lower order harmonics are

reasonably stable and approximately normally distributed whereas higher order

harmonics are uniformly distributed. Approximate values for k and α for ranges of

harmonic h are suggested by [50] as follows for 95th percentiles

• h = 3, 5, 7: k=1, α=1 (valid for harmonics with fixed amplitude, whose phase

angle may randomly vary between 0° and 90°)

• h = 11, 13: k=1, α=1.4 (valid for harmonics whose amplitude may vary between

half maximum and maximum and whose phase angle may randomly vary between

0° and 270°)

• h > 13: k=1, α=2 (valid for harmonics whose magnitude may randomly vary

between 0 and maximum and whose phase angle may randomly vary between 0°

and 360°).

37
The early work on the summation rule was based on the distributions of random

amplitude and phase of the individual harmonics. Also considered in [50] is the

introduction of harmonics varying randomly over periods of time. The only

disadvantage with this method is that to verify results experimentally the harmonic

nature of each of the individual loads within a distribution system need to be

determined. A final technical report [51] has been completed in relation to the

summation law as a draft version of IEC 61000-3-6. In the draft document no reasoning

is given for the values of α used but suggest α=1 for n<5, α=1.4 for 5≤n≤10, and α=2

for n>10. These values differ to those indicated above by the work in [50].

2.10 Summary
Harmonic monitoring programmes undertaken overseas and within Australia have been

reviewed. There is a large amount of inconsistency between how measurements are

taken and results reported. Conformance to guidelines specified in recently introduced

international standards are however helping to rectify these inconsistencies. If DNSPs

aim to fully utilise the results from harmonic monitoring programs undertaken

periodically or for specific investigations, measurement and reporting procedures need

to be developed to ensure these results can be used effectively for future reference.

The levels of harmonic distortion on an international basis have been investigated

through a literature review of results from harmonic monitoring campaigns. The trend is

of gradual harmonic growth with some literature estimating existing planning levels

may be exceeded within the next decade. This review highlights the need for a

commitment to monitoring of harmonic levels within distribution systems to allow

utilities and customers to plan adequately for harmonic growth. International regulatory

38
document IEC 61000-3-6 and IEEE Standard 519 are gradually being adopted in an

attempt to ensure harmonic voltages within MV distribution systems remain at

acceptable levels.

In January 2001 Australia adopted a new harmonic standard governing emission limits

of distorting loads in MV and HV power systems. The new standard AS/NZS 61000.3.6

is an adaptation of the international technical report IEC 61000-3-6. AS/NZS 61000.3.6

replaces AS 2279.2, which was first introduced in 1979 and has been the basis of setting

the standards for harmonics in Australia. Some work is still required to overcome the

complexities in the newly adopted standard.

Statistical techniques are required to combine the contributions from individual

distorting loads and the summation rule has been suggested as a useful tool to achieve

this. The limitation of the summation rule is that measured data does not exactly follow

the required pattern of the theoretical data used to obtain the required indices. However

the summation rule is designed only to give an accurate approximation of the 95th

percentile cumulative probability levels. For use in the application of harmonic

standards it is suggested the summation rule would be suitable and is adopted as one of

the key principles of the theory and methodologies developed in following chapters.

39
 Chapter 3

Estimation of harmonic levels in MV distribution systems

3.1 Introduction
Results from harmonic monitoring programmes presented in Chapter 2 have suggested a

general trend of growth in harmonic distortion levels. Electricity distribution network

service providers (DNSP) should now be looking towards preventative measures to

ensure that voltage distortion levels are within limits set by the appropriate standards

[16]. These measures will need to be taken at the planning stage to ensure distribution

systems will be able to meet harmonic standards as the distortion levels due to loads

rise.

To evaluate the harmonic performance of a distribution system design the DNSP will

require the ability to estimate distortion levels. The problem of estimating harmonics at

the design stage of a distribution system is very different to the normal investigation of

a particular harmonic load. The study must encompass many loads generating

harmonics and whose harmonic spectrum and daily variation are poorly specified. For

such modelling the aim is not to accurately model every individual load but to

determine methods for representing the statistics of large aggregations of load.

To develop a method for estimating distortion levels this Chapter investigates the

diversity of small and medium sized harmonics producing loads common to MV

distribution systems. Results from a harmonic monitoring programme of seven sites in

an MV distribution system outlined in Chapter 5 are also utilised to establish aggregated

load models. A breakdown of these loads into the residential, commercial, and industrial

load sectors is carried out. This small part of the measurements from the monitoring
40
programme is required to give the parameters of the residential, commercial and

industrial load sectors. The remainder of the monitoring results have been used to

validate a proposed method to estimate distortion levels.

The assumptions required to allow a pragmatic approach to harmonic modelling are also

addressed in this Chapter. Finally the Chapter will report on a method that can be used

to establish typical harmonic distortion levels within a distribution system and therefore

predict the effect of a change in network or load conditions. The method applies to MV

radial distribution systems and incorporates background distortion from the upstream

supply. The method to be described is able to be set up in a spreadsheet rather than

requiring specialist simulation software and allows calculation of sensitivities to various

planning parameters so that the important ones can be identified.

3.2 Diversity of loads in MV distribution systems

Within MV distribution systems there exists an endless diversity of load types that draw

distorted current from the power supply system. The usual behaviour of these loads

means that while the amount of distortion in their current may depend on the level of

harmonics existing on the system, they will usually draw non-linear current even when

the supply voltage is nearly sinusoidal. The harmonic currents drawn by these non-

linear loads in turn produce harmonic voltages on the system in proportion to the

distribution system harmonic impedance.

With the growing use of power electronic devices to achieve efficiency and flexibility,

substantial increases in harmonic distortion will become apparent if harmonics are not

considered in distribution system design. The difficulty is not on how to model each of

41
the individual loads as the DNSP engineer is not interested in this detail. Instead, a

method of modelling the combined aggregate effect of many different distorting loads at

each of the distribution transformers within the distribution system is required.

The harmonic content of currents drawn by distorting loads varies depending on

operating mode, equipment components and network properties. The phase of each of

the individual harmonics also varies with the same parameters. Thus addition of

multiple harmonic distorting loads in the frequency domain computed directly using

arithmetic addition provides a pessimistic model of harmonic distortion and hence

phasor addition has to be considered. To account for this behaviour harmonic planning

techniques must incorporate load and time diversity in their approach to control the

harmonic distortion levels. As a preliminary to producing harmonic models suitable for

planning techniques it is essential to first look at some measurements that illustrate load

type diversity.

Table 3.1 shows the current waveforms of various domestic loads within a household

supplied by Integral Energy in the Sydney area, captured using a Fluke 41 Power

Harmonic Analyser (see Appendix A for specifications). The harmonic components of

each of the load current waveforms are listed in Table 3.2 alongside a graphical

representation of the harmonic spectrum. It can be seen that the phase of the harmonic

currents vary significantly over the different load types. It is to be noted that most of

these loads are single phase. Greater diversity would be expected if more three phase

loads were also included in the study.

42
 Table 3.1 Current waveforms of common domestic loads
1000W 27L Microwave Oven 15

10

(Cooking on high) 5

Current (A)
0
0.0 3.1 6.3 9.4 12.5 15.6 18.8
-5

-10

-15
Time (ms)

100W Pentium 120MHz Processor 1.5

2

(Running a programme) 1

0.5

Current (A)
0
0.0 3.1 6.2 9.4 12.5 15.6 18.7
-0.5

-1

-1.5

-2
Time (ms)

96W Compact Audio System 0.8

1

(Playing compact disk at 80% sound

0.6
0.4

volume) Current (A)

0.2
0
-0.2 0.0 3.1 6.3 9.4 12.5 15.6 18.8

-0.4
-0.6
-0.8
-1
Time (ms)

82W 51cm Colour Television 1.5

(Normal operating mode) 0.5

Current (A)

0
0.0 3.1 6.3 9.4 12.5 15.6 18.8
-0.5

-1

-1.5
Time (ms)

1000W Vacuum Cleaner 8

(Normal operating mode) 4

2
Current (A)

0
0.0 3.1 6.3 9.4 12.5 15.6 18.8
-2

-4

-6

-8
Time (ms)

6kg Washing Machine 2.5

2

(On wash cycle with agitator)

1.5
1
Current (A)

0.5
0
-0.5 0.0 3.1 6.3 9.4 12.5 15.6 18.8

-1
-1.5
-2
-2.5
Time (ms)

3 Phase Air Conditioner – Inverter driven 15

(Normal operating mode)

10

5
Current (A)

0
0.0 3.1 6.3 9.4 12.5 15.6 18.8
-5

-10

-15
Time (ms)

43
 Table 3.2 Harmonic components of current from common domestic loads
1000W 27L Microwave Oven 7

I1 = 3.46A ∠-8°

Current (A)
4

I3 = 1.78A ∠-152°
3

I5 = 0.68A ∠-57° 1

I7 = 0.25A ∠-68° 0

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Harmonic (h)

100W Pentium 120MHz PC 0.45

0.4

0.35

0.3

I1 = 0.43A ∠15°

Current (A)
0.25

0.2

I3 = 0.36A ∠176° 0.15

I5 = 0.26A ∠-13°
0.1

0.05

I7 = 0.16A ∠160° 0

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Harmonic (h)

96W Compact Audio System 0.45

0.4

0.35

0.3

I1 = 0.42A ∠-6°
Current (A)
0.25

0.2

I3 = 0.21A ∠-169° 0.15

I5 = 0.05A ∠-1°
0.1

0.05

I7 = 0.06A ∠-9° 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Harmonic (h)

82W 51cm Colour Television 0.35

0.4

0.3

I1 = 0.35A ∠-18°
0.25
Current (A)

0.2

I3 = 0.19A ∠176° 0.15

0.1

I5 = 0.13A ∠-27° 0.05

I7 = 0.07A ∠141° 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Harmonic (h)

1000W Vacuum Cleaner 4.5

3.5

I1 = 3.93A ∠-11°
Current (A)

2.5

I3 = 0.81A ∠134°
2

1.5

I5 = 0.09A ∠-176°
1

0.5

I7 = 0.13A ∠-73° 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

Harmonic (h)

6kg Washing Machine 1.4

1.2

I1 = 1.28A ∠-1°
Current (A)

0.8

I3 = 0.32A ∠112°
0.6

0.4

I5 = 0.11A ∠138° 0.2

I7 = 0.01A ∠-94° 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

Harmonic (h)

3 Phase Air Conditioner – Inverter driven 3.5

4

I1 = 3.37A ∠-4°
2.5
Current (A)

I3 = 0.12A ∠-155° 1.5

I5 = 2.62A ∠132° 0.5

I7 = 2.20A ∠-68° 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

Harmonic (h)

44
From Table 3.1 it is clear that all the loads tested draw distorted current. The level of

voltage distortion existing on the supply will also have an effect on the level of current

distortion caused by each load. This has been documented in [52] for the case of

capacitive filtered rectifiers within switch mode power supplies (SMPS). For these

rectifiers the shape of the current waveform depends on the peak value of the voltage

supply waveform. Large groups of these rectifiers draw significant levels of 3rd

harmonic current that tend to flatten the voltage waveform, reducing the peak of the

current waveform to each rectifier and the overall amount of current distortion. A quick

comparison of the single-phase load waveforms in Table 3.1 (most of which contain a

SMPS) and the voltage waveform of Table 3.3 shows the typical single-phase load

current pulses are approximately in phase with the voltage peaks.

Table 3.3 Typical supply voltage waveform for Tables 3.1 and 3.2
400
Supply Voltage 300

200

V1 = 243.06V ∠0° 100

Voltage (V)

V3 = 1.28V (0.53%) ∠98° -100

0.0 3.1 6.3 9.4 12.5 15.7 18.8

V5 = 7.27V (2.99%) ∠-175° -200

-300

V7 = 1.86V (0.76%) ∠11° -400

Time (ms)

VTHD = 3.10%
300

250

200
Voltage (V)

150

100

50

0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

Harmonic (h)

One of the most prolific uses of power electronics involves the application of capacitor

filtered diode rectifiers. When single-phase and three-phase diode rectifiers are mixed

within distribution systems, it is possible that the total harmonic distortion may

45
decrease. This is due to the cancellation of individual harmonics, such as the 5th and 7th

harmonics, which can occur with near 180° difference in the phase angle [53].

The three-phase inverter driven air conditioner load of Table 3.1 illustrates a typical

three-phase load powered through a capacitor-filtered rectifier, which has a trough at the

peak of the voltage waveform. By combining the three-phase and single-phase loads it

can be realised that the peak of the single-phase load partially fills up the trough in the

current waveform of the three-phase load, as reported in [53]. Also with controlled

converters operating at different firing angles some compensation can occur by

harmonic currents being slightly out of phase producing net harmonic current levels less

than what would be obtained by direct addition. The diversity of loads illustrated above

creates a difficult problem for estimating the aggregate harmonic distortion emissions

from multiple loads at a distribution transformer.

The problem of modelling these loads for planning purposes is how to combine the

harmonic currents of various loads to obtain a suitable net harmonic current and

calculate the resulting harmonic voltage contribution. It is very difficult to include all

parameters and find the maximum harmonic voltage that will occur on the distribution

system. This is because each of the harmonic producing loads will have different phase

angles for each of the harmonic components as discussed above and also because it is

very unlikely that all loads will be producing their maximum harmonic current all at the

same instance in time. Harmonic planning techniques must also include the varying

nature of distribution system harmonics over time for assessment of harmonic voltage

levels.

46
3.3 Residential, commercial and industrial load characteristics
From a DNSP perspective it will be assumed that the loads on a feeder can be classified

as residential, industrial and commercial load types, each having an average aggregate

harmonic current characteristic that is roughly constant throughout the network.

To illustrate which harmonics are of most significance within a typical distribution

system the harmonic content of a number of combined loads measured on the LV side

of distribution transformers are shown in Figures 3.1 to 3.3. These harmonic spectrums

were obtained from LV monitoring locations within Integral Energy’s Homepride zone

distribution system. The harmonic content of the current measured at the sending end of

an MV feeder connected to the Homepride zone substation is also illustrated in Figure

3.4. While there is no indication of the phase of each harmonic, it can be seen that the

3rd, 5th and 7th harmonics are the dominant harmonics for the residential, commercial

and industrial load sectors at LV. These are the characteristic harmonics of most single-

phase power electronic loads, with the 5th and 7th also being the dominant harmonics for

three-phase power electronic loads.

3.5

3.0

2.5

2.0
Distortion
(%) 1.5

1.0

0.5

0.0
3 5 7 9 11 13 15 17 19 21 23 25 27 29 31
Harmonic

Figure 3.1 Residential distribution transformer harmonic current snapshot

(THDI = 4.64%)

47
 5.0

4.5

4.0

3.5

3.0
Distortion
2.5
(%)
2.0

1.5

1.0

0.5

0.0
3 5 7 9 11 13 15 17 19 21 23 25 27 29 31

Harmonic

Figure 3.2 Commercial distribution transformer harmonic current snapshot

(THDI = 6.19%)
9.0

8.0

7.0

6.0

Distortion 5.0

(%) 4.0

3.0

2.0

1.0

0.0
3 5 7 9 11 13 15 17 19 21 23 25 27 29 31
Harmonic

Figure 3.3 Industrial distribution transformer harmonic current snapshot

(THDI = 10.59%)
4.0

3.5

3.0

2.5
Distortion
2.0
(%)
1.5

1.0

0.5

0.0
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Harmonic

Figure 3.4 Zone substation feeder harmonic current snapshot (THDI = 4.59%)
48
Table 3.4 summarises the results from the harmonic distortion monitoring programme

giving the 95th cumulative probability values for the fundamental and 5th harmonic

current for the three LV locations, averaged over the entire monitoring period. The

subscript ‘1CP95’ indicates the fundamental component 95th percentile cumulative

probability level. It is required to characterise the load types so that the 95th percentile

cumulative probability levels of harmonic distortion can be predicted from known

network impedances, load magnitudes and compensations. Assuming that similar load

compositions at any power level will have the same spectrum, each load type can be

characterised conveniently by the I5CP95/I1CP95 as shown in the right hand column. This

shows that the residential load has much less distortion than the commercial and

industrial types as would be expected. Over the course of the entire monitoring

programme it was found that the ratio of 95th percentile 5th harmonic to 95th percentile

fundamental varied by less than 10% for any given week of data. Note that the

I5CP95/I1CP95 ratio for the industrial case is representative of light industry, as large

industry will usually require further consideration and in depth harmonic analysis,

especially when loads such as arc furnaces are installed.

Table 3.4 Fundamental and 5th harmonic current 95th percentile values
Harmonic Monitor Site I1CP95 (A) I5CP95 (A) I5CP95/I1CP95
Residential 133 3.36 0.025
Commercial 617 43.3 0.071
Industrial 1469 170.9 0.116

Each load type is a mixture of linear and non-linear equipment that varies throughout

the day. Since the two components vary independently to some extent, the fraction I5/I1

will also vary throughout each day. Noting that variation of the 95th percentile ratio is

less than 10% provides the basis to assume that the ratio of the statistics I5CP95/I1CP95 is

constant for all loads of a given type (residential, commercial and industrial). These
49
ratios can be found from current measurements alone at the supply point to a pure

residential, commercial or industrial load. Ideally it would be advantageous to repeat the

measurements from the monitoring programme on numerous other MV distribution

systems. However, cost and access to systems makes this task difficult to complete in

the time frame of the thesis.

The statistical harmonic models of residential, commercial and industrial load types

above have been developed to simulate the global behaviour of distorting loads at

distribution substations. The load models represent aggregates of loads and are

specifically intended for calculation of harmonic emissions for comparison with the

relevant standards. Both time and phase diversities are included in the representative

load models for frequency domain analysis.

3.4 Assumptions to allow estimation of distortion levels

Harmonic distortion surveys that have been completed over the past decade have shown

that the 5th harmonic is the most significant harmonic and usually accounts for over

80% of the total harmonic distortion (THD) within MV systems [21, 22, 54]. This is due

to the rapid fall off of harmonics with frequency, produced by present power electronic

technology (even more rapid than the fall off of voltage limits) and the removal of most

3rd harmonics by delta-star transformers stepping down to the LV system. Present trends

suggest that the 5th harmonic will continue to dominate for many years. A key

assumption in this method of estimating harmonic levels is that if the 5th harmonic

voltage is within acceptable limits, there will most likely be no harmonic problems. This

assumption will need to be revisited where there are power factor correction capacitors

50
giving amplification over some frequency range away from the 5th harmonic and this

will be discussed further in Chapter 6.

From the harmonic distortion monitoring programme results it was verified that the

amount of total harmonic voltage distortion is dominated by the 5th harmonic. This

characteristic may be used to predict the amount of total harmonic distortion existing on

the system if the 5th harmonic distortion level can be established.

3.5 Key concepts for estimating distortion levels

There are numerous commercially available software packages, such as

PSCAD/EMTDC [55], PSPICE [56], SIMULINK [57], and SUPERHARM [58], which

can be used to simulate the behaviour of harmonics within distribution systems and

estimate resulting distortion levels. These packages include the ability to perform

studies in time domain, frequency domain, or a combination of both. A common

requirement in the use of all these simulation packages for modelling purposes is a

detailed knowledge of both the distribution system and customer loads. For MV

distribution systems customer loads are most likely to include a mixture of many small

and medium sized loads incorporating a variety of power electronic products. It would

be a difficult and impractical task for a DNSP to model all of these loads individually in

a software package for planning purposes. It is desired to have a method of establishing

harmonic distortion levels within MV distribution systems using a more applicable

simplistic approach while maintaining an acceptable level of accuracy.

The method for establishing harmonic levels within a distribution system uses some of

the concepts that are given in IEC 61000-3-6 [15]. Two concepts used throughout this

51
thesis are the evaluation of emission levels by statistical means and the second

summation law.

Emission Levels - The IEC standard suggests that the emission levels of harmonic

voltage distortion should be evaluated using a statistical approach to take time

variations into account [15]. Broadly, it is proposed that a harmonic survey should be

taken over a period of at least a week and that the signal analysis should be along the

strict guidelines of [59] with the maximum and 95th percentile of the cumulative

probability distribution (CP95) determined. The resulting value must be less than the

planning level for the appropriate voltage level, of which suggested values are given.

This prevents a DNSP being penalised for a high value of harmonics that might occur

for only 5% of the day (1.2 hours).

Second Summation Law – The law is used to combine the 95th percentile cumulative

probability values of harmonic currents or voltages to give their overall combined value.

It is designed to give a value less than that which would be given by direct addition to

take account of time and phase diversity, as discussed in Chapter 2.

The second summation law is required to combine the statistical values, as direct

addition for simplified analysis does not suffice. Consider the data sets in Table 3.5,

where the first set has 90 values of “1”, followed by 5 of “2” and then 5 of “3” giving a

95th percentile cumulative value of 2. The second set shown will have a 95th percentile

cumulative value of 3. The sum of these two sets will have 100 values all of “4” giving

a 95th percentile cumulative value of 4, yet the sum of the individual 95th percentile

cumulative values is 5.

52
The second summation law is assumed to have the form

Vh = α ∑ Vhiα (3.1)

where

Vh = magnitude of resulting harmonic voltage (order h)

Vhi = magnitude of individual ith harmonic voltage (order h) to be combined.

α = exponent depending on the probability of actual value exceeding the calculated

value and the degree to which the individual harmonic voltages vary in magnitude and

phase.

Table 3.5 Example values of sampled data

% of Data Set 1 Set 2 Sum (set 1+set 2)
90% 1 3 4
5% 2 2 4
5% 3 1 4
CP95 2 3 5

For the above example, α should be 1.51. As presented in Chapter 2, for the 5th

harmonic it is suggested by [15] that α=1.4 be used, with a higher value for some of the

higher order harmonics.

3.6 Method to establish distortion levels

In order to establish typical levels of harmonic voltages a model of the distribution

system under study must be first produced. The parameters required to model an MV

distribution system for the proposed calculation method are listed below.

• The impedance, total rated capacity and total maximum demand of the HV/MV

zone substation transformer.

• Fault level of HV transmission feeder to establish harmonic impedance of HV

transmission system.

53
• Number of feeders and approximate number of distribution transformers per feeder.

• Approximate lengths and impedances of lines (either underground or overhead).

• Approximate proportion of residential, commercial and industrial loading for each

feeder.

During the design process only a qualitative estimate of voltage harmonic distortion

levels is required. To establish these typical levels of harmonic voltage distortion some

approximations must be made. The assumptions required to simplify a system such as

the distribution system on which the harmonic monitoring programme was performed

(shown in Figure 5.6, of Chapter 5) are listed below.

(i) The system is balanced with no zero sequence. This can be justified by the results

from the harmonic monitoring programme in Chapter 5, which illustrate that even

at the LV levels the loading of each phase is well balanced.

(ii) The 5th harmonic is the most significant harmonic. This is true for most balanced

MV distribution systems [21, 22, 54] and is reinforced by the results of the

harmonic monitoring programme presented in Chapter 5 and Figures 3.1 to 3.4

above.

(iii) As typical distribution system feeders are less than 10km in length the capacitance

of the feeders is assumed negligible.

(iv) All impedances (conductors and transformers) are considered to be inductive with

the resistive component assumed to be negligible at the 5th harmonic. Resistive

components may be included if available, noting that this will reduce the

simplicity of the method without adding significant accuracy.

(v) 5th harmonic currents drawn by individual loads are independent of each other and

are summated using the second summation law defined by equation (3.1).

54
(vi) Residential, commercial and industrial type loads are spread out evenly over the

length of each feeder.

The first step in predicting the harmonic voltage distortion throughout a system is to

estimate the level of harmonic currents. Using the number of customers per LV

distribution system and number of MV substations in conjunction with the proportion of

residential, commercial and industrial loads the total harmonic current drawn at the

HV/MV zone substation can be determined.

The individual load harmonics are summated together as given by equation (3.2) (α=1.4

for the 5th harmonic) to find the total harmonic current drawn from the HV transmission

system.

I5total1.4=I5load11.4+I5load21.4+I5load31.4+… (3.2)

The level of distortion at the MV bus of the zone substation (V5MV) is then the

summation of the HV background distortion (V5BG) and the voltage distortion arising

from the impedance of the HV transmission line and the HV/MV zone substation

transformer (V5HV). The background distortion is the distortion that appears on the

downstream system due to the interconnection of other loads via the upstream supply.

V5BG1.4+V5HV1.4=V5MV1.4 (3.3)

If known, the background distortion level can be directly applied or else the

approximation outlined in section 7.2.2 of [15] can be used by assuming the distortion

level at HV is the same proportion as the distortion level at MV when compared to the

harmonic limits. For example the recommended harmonic limit for the 5th harmonic at

55
HV is 2% and for MV the limit is 5% thus the background distortion can be calculated

as shown by equation (3.4).

(k2%)1.4+V5HV1.4=(k5%)1.4 (3.4)

where k is the existing proportion of recommended harmonic voltage limit.

From the MV zone substation the distortion levels downstream of each of the feeders is

calculated by the summation of the voltage distortions due to each of the individual

loads and the background distortion. This is achieved by multiplying the harmonic

current, approximated by the size of load and typical harmonic current level, with the

harmonic impedance seen by that load back to the MV zone substation bus. For the 5th

harmonic, the harmonic impedance is approximated by 5 times the fundamental

impedance. Figure 3.5 illustrates the process of the harmonic voltage calculation.

X1 X2 X3

V5BG I51 I52 I53

Figure 3.5 MV feeder equivalent circuit for voltage distortion calculation

The harmonic voltage distortion at the connection point after impedance X1, which is

due to the 5th harmonic currents produced by the individual loads downstream and the

background distortion, is given by equation (3.5) as follows.

V5X11.4=V5BG1.4+(I51X1)1.4+(I52X1)1.4+(I53X1)1.4 (3.5)

56
3.7 Study system and harmonic monitoring results
The study system on which the harmonic monitoring programme was performed is the

Homepride zone distribution system, which will be further described in Chapter 5. One

of the assumptions in the calculation of 95th percentile harmonic voltage distortion

levels is that an approximate proportion of the residential loads, commercial loads and

industrial loads are known. Further, typical values of harmonic currents can be assigned

to each of the residential, commercial and industrial loads for calculation of the

harmonic voltage distortion. To establish typical values of harmonic currents the results

from the monitoring programme are utilised. The resulting weekly average 95th

percentile cumulative probability results for each of the seven sites monitored are given

in Table 3.6.

Table 3.6 Results of monitoring at Homepride for fundamental and 5th harmonic
Harmonic Monitor Site (Voltage Level) I1CP95 I5CP95 V5CP95
(A) (A) (%)
Residential transformer (415V) 119.7 3.0 1.45
Commercial Transformer (415V) 548.9 48.9 2.89
Industrial transformer (415V) 1404.6 193.0 1.63
Residential feeder (11kV) 173.5 5.5 1.54
Commercial feeder (11kV) 114.6 3.8 1.54
Industrial feeder (11kV) 153.7 8.2 1.54
Homepride zone substation transformer (11kV) 1205.3 17.1 1.54

The magnitudes of harmonic currents were obtained from measurements taken in the

harmonic survey. The typical levels of per unit 5th harmonic current produced by

residential, commercial and industrial load types given in Table 3.4 were established

from the data collected over the entire three year monitoring programme. Similar

figures could be used during the design phase of similar distribution systems to establish

typical harmonic levels and help identify future harmonic problems.

57
3.8 Application of method to the study system
Relevant details of the Homepride zone distribution system required to establish

approximate harmonic distortion levels are listed below.

• Zone substation rating is 50MVA.

• Typical maximum demand at 33kV/11kV zone substation transformer is 0.45pu

(approximated by 95th percentile of fundamental current).

• Load is approximately 30% residential, 40% commercial, and 30% industrial.

• 10 Feeders are equally loaded.

• There are no power factor correction (PFC) capacitors installed in the system, but

these should be easy to allow for in the approach given here. This will be discussed

further in Chapter 6.

• Residential feeders (classified as 85% residential load and 15% commercial load)

typically consist of 10 substations.

• Commercial feeders (classified as 86% commercial load and 14% residential load)

typically consist of 7 substations.

• Industrial feeders (classified as 75% industrial load, 20% commercial load and 5%

residential load) typically consist of 5 substations.

• Figure 3.6 represents a simplified layout of the Homepride zone MV distribution

system. The residential, commercial and industrial feeders are shown with the

corresponding number of distribution transformers connected. The "other feeders"

consist of other residential, commercial and industrial feeders.

Except for large customers who have direct connection to the feeder (usually industrial

customers) each of the MV substations will have an LV distribution system that is also

58
owned by the DNSP. The LV distribution system typically consists of overhead or

underground cable with lengths up to 500m.

33kV/11kV 50MVA
Impedance 15%
33kV Transmission
500MVA Fault Level

Other Feeders

Industrial Feeder

Commercial Feeder

Residential Feeder

Figure 3.6 Simplified layout of Homepride zone MV distribution system

A layout of the LV distribution system is shown in Figure 3.7. The number of customers

connected to each substation will vary from site to site but the loading should remain

approximately the same due to the rating of the transformer.

11kV/415V 500kVA
Impedance 5%

11kV Substation

Customers

Figure 3.7 LV distribution system layout

59
Each of the feeders was modelled as a combination of residential, commercial and

industrial loads. Using the typical values of harmonic currents for each load type, with

the second summation law the total 5th harmonic current drawn by the MV system was

determined. With the total harmonic current satisfactorily matching the actual measured

current the progressive harmonic voltages at points further down the system could be

evaluated.

Comparisons of the calculated 95th percentile results from the simulated model and the

measured results are given in Table 3.7. As can be seen from the results the method

used to establish typical harmonic levels gives a reasonably accurate account of the 95th

percentile harmonic voltage levels occurring on the MV distribution system.

Table 3.7 Comparison of calculated and measured 95th percentile values

Site Calculated Measured
I5TOTAL 23.2A 17.1A
V511kV BUS 1.50% 1.54%
V511kVRES 1.52% -
V511kVCOM 1.51% -
V511kVIND 1.53% -

Site Calculated Measured

V5415VRES 1.64% 1.45%
V5415VCOM 2.33% 2.89%
V5415VIND 1.76% 1.63%

3.9 Sensitivity results

A sensitivity analysis on the model was completed to establish which parameters have

the greatest effect on the outputs of the model. The 5th harmonic voltage level at the MV

busbar was chosen as the representative output for the analysis. Each of the input

variables were increased individually by a factor of 1% and the percentage change in the

60
output variable was established, Figure 3.8 illustrates the resulting sensitivity for the 5th

harmonic voltage at the MV busbar.

Variable increased by 1% Percentage change in Zone Substation

5th harmonic voltage (%)
Residential Load (VA) 0.064%
Commercial Load (VA) 0.414%
Industrial Load (VA) 0.523%
Alpha (α) -3.321%
XS + XT (Fault level & zone sub transformer) 1.000%
XT (Zone substation transformer) 0.630%
Number of substations per feeder -0.284%
Residential (I5CP95 / I1CP95) 0.064%
Commercial (I5CP95 / I1CP95) 0.414%
Industrial (I5CP95 / I1CP95) 0.523%
All load types (I5CP95 / I1CP95) 1.000%
Residential Load (%) -0.792%
Commercial Load (%) 0.397%
Industrial Load (%) 0.830%

Figure 3.8 Sensitivity of the 5th harmonic voltage

The results of the sensitivity analysis illustrate that 5th harmonic voltage is less sensitive

to errors in the ‘known quantities’ of distribution system parameters such as line

impedance, fault levels, but particularly sensitive to the ‘unknown quantity’ of the

exponent α. The value used for α is based on approximated statistical distributions of

harmonic phase and amplitude obtained through experience [50]. Recommended values

for the range of alpha for each individual harmonic are given in [15].

Although the distribution system surveyed showed good correlation with the calculated

results from the model, the sensitivity analysis highlights the need for an appropriate

value for α to be used. Where possible, measurements of harmonic amplitude and phase

should be obtained to confirm an approximate uniform distribution for the lower order

harmonics, an assumption that has been used in establishing α=1.4 [50].

As the 5th harmonic voltage from the model has shown greatest sensitivity to the value

of α, a further study over the full range of α was completed. As expected the

61
relationship of the magnitude of α to error in the resulting 5th harmonic voltage was

exponential. Figure 3.9 illustrates the exponential relationship for percentage change in

the 5th harmonic voltage depending on the value of α used. If the diversity of the

individual harmonics was not considered, direct addition of harmonic components

(α=1) would largely over estimate the resulting value for the 5th harmonic voltage. By

adjusting the value of α for zero error between the calculated and measured results it

was found that α=1.42 gave the most suitable results. This necessary increase in the

value of alpha (allowing greater diversity) could partially be attributed to the normal

distribution of the 5th harmonic rather than the assumed uniform distribution.

350.00%
Error in 5th harmonic voltage (%)

300.00%

250.00%
200.00%
150.00%
α = 1.42
100.00%

50.00%
0.00%
0.5 1.0 1.5 2.0 2.5
-50.00%
-100.00%
Value of alpha (α) used

Figure 3.9 5th harmonic voltage sensitivity to alpha

3.10 Summary

Statistical harmonic models of residential, commercial and industrial load types have

been developed to simulate the global behaviour of distorting loads at distribution

substations. The load models represent aggregates of loads and are specifically intended

for calculation of harmonic emissions for comparison with the relevant standards. Both

time and phase diversities are included in the representative load models for frequency

domain analysis.

62
A method has been developed to estimate the 95th percentile cumulative probability

level of harmonic voltage distortion in an MV distribution system as required by the

present Australian harmonic standard AS/NZS 61000.3.6 (adapted IEC 61000-3-6).

This method includes techniques to overcome the difficulties in determining distortion

levels when feeders are not being loaded to their fullest extent, and allows for cases

when levels of background distortion may not be as high as limits set by the standard.

The implementation of these techniques will be illustrated in Section 7.2. The method

provides a useful tool to combine the effects of non-precise distribution system loads at

the planning stage or to evaluate distortion levels of existing systems.

By assigning typical harmonic currents to residential, commercial and industrial load

sectors a simplified harmonic analysis of voltage distortion levels within a distribution

system can be completed. The benefit of this analysis is that only a reduced order of

system data is required to produce reasonably accurate results. This technique is

applicable to most radial MV distribution systems. The harmonic prediction technique

has proven to be successful when compared to actual harmonic measurements obtained

from a harmonic monitoring programme on a suburban MV radial distribution system.

Application to additional study systems with harmonic monitoring data is required to

verify the technique. However, the time and cost constraints of this project did not allow

for supplementary harmonic monitoring programs to be completed and is an area for

further work. The ability to estimate harmonic voltage levels will enable distribution

system designers to better understand the harmonic capabilities of distribution systems.

Estimation of harmonic voltage levels within radial distribution systems using the

second summation law from [15] has shown to be successful. The value of exponent α

63
recommended by [16] has produced good results for the study system in this case, but a

sensitivity analysis has illustrated that α needs to be chosen carefully.

The work described in this Chapter contributes to the development of distribution

system planning guidelines to enable the design engineer to maximise the harmonic

capabilities of a distribution system. This can be achieved by optimising parameters

such as impedance values of lines and transformers, feeder lengths, allocation of load

types to particular connection points, location and size of power factor correction

capacitors, and connection of detuning reactors.

64
 Chapter 4

Allocation of harmonic emissions to MV distribution system

customers

4.1 Introduction
In the planning stage of an MV distribution system it is necessary to ensure only

acceptable levels of harmonic current emissions are contributed from each distribution

substation to ensure that the harmonic voltages recommended by standards are not

exceeded. With a suitable method for estimating harmonic voltage levels established in

Chapter 3, this Chapter will look at extending the method to include a general policy for

allocating harmonic emissions.

The harmonic standard IEC 61000-3-6 outlined in Chapter 2 comprises a number of

stages and tests to determine harmonic emission allowances for customers connected to

MV or HV networks. Stage 1 has three tests that base acceptance on load size as

compared to the short circuit level at the connection point. Stage 2 contains three tests

of increasing complexity depending on the amount of information known about the

system. There is also a Stage 3 where excessively distorting loads are allowed

connection on a temporary and precarious basis. It is perceived that most distorting

loads will be assessed under Stage 2 of the standard.

The application of guidelines from IEC 61000-3-6 is somewhat more difficult than the

superseded standard AS 2279.2 in that it includes time-varying situations and the use of

statistical quantities. Of particular importance is the section of the standard concerning

loads distributed along a feeder having significant variation in fault level. IEC 61000-3-

65
6 briefly covers this section in Stage 2, Test 3. The application of the principles

suggested by the standard for this section is poorly described and only a non-practical

trivial example is provided.

This Chapter presents the guiding principles of IEC 61000-3-6 that are used to allow

allocations of harmonic emission to customers distributed throughout a power system.

A general approach to harmonic emission allocation for customers spread along several

MV feeders with significant variations in fault levels is then developed. This method is

designed to encompass the poorly described section of the IEC 61000-3-6. This method

is designed such that when applied to the simplest system it aligns exactly with the

guidelines outlined in IEC 61000-3-6, while retaining the ability to be applied to far

more complex systems.

Allocating harmonic emissions to distribution system customers using the developed

method may often require an extensive amount of data. This Chapter will also

investigate the use of approximations that may simplify the allocation process and

eliminate the need for extensive data. Finally the method is applied to some general

study systems to evaluate the harmonic emission allocation technique and the associated

simplifying approximations.

In the method described to allocate harmonic emissions to individual customers, the

presence of power factor correction (PFC) capacitors is not considered other than as part

of the harmonic impedance. PFC capacitors can cause system resonances at problematic

frequencies and thus are an important consideration for harmonic analysis, an aspect

that will be further investigated in Chapter 6.

66
4.2 Principles of IEC 61000-3-6
The guidelines specified in the new standard are somewhat more difficult to apply than

in the previous harmonics standard AS 2279.2 [12]. These guidelines attempt to ensure

allocation of harmonic emission rights to customers is more equitable. A key concept is

that customers with the same maximum demand and the same point of common

coupling (PCC) are entitled to equal harmonic emission rights. The PCC is defined as

the nearest point in the power system to which another customer might be connected.

To account for time variation, customer harmonic contributions and DNSP harmonic

levels are generally assessed using the 95th percentile cumulative probability (CP95)

level. As the 95th percentile levels are statistical quantities direct summation is

inadequate for combining contributions from a number of customers. As mentioned in

Chapter 2, two summation laws are proposed by the standard:

(i) The first summation law makes use of diversity factors that require knowledge of

the load type and is suited to more individual cases.

(ii) The second summation law is a more general method that accounts for time

diversity of the individual loads on a larger scale, and is given by equation (3.1)

where the exponent α depends on the harmonic order h, and its recommended

value for the 5th harmonic is 1.4.

The second summation law provides the basis for the proposed methodology for

allocating harmonic emission rights to customers within an MV distribution system.

Although the standard encourages an equitable allocation of harmonic 'rights' to all

customers having the same maximum demand, where customers see different fault

levels the question arises as to whether these 'rights' are to equal harmonic voltage,

67
equal harmonic current, or some other right. It can be shown that allocating equal

harmonic voltage rights allows greater use of the system's harmonic absorption

capability, but customers towards the end of a weak feeder receive lower currents. The

allocation of equal current is fairer but underutilises the harmonic absorption capability.

The standard recommends a mid-way policy of equal harmonic power, which can be

shown to be equivalent to a harmonic current allocation varying with the square root of

the fault level.

IEC 61000-3-6 assumes that the harmonic voltage at the MV level is a combination of

the emissions from the MV loads and the background distortion of the HV transmission

system. Thus a fraction ThMV of the HV harmonic planning level LhHV must be included

in the MV harmonic voltage planning level LhMV. Using the second summation law the

acceptable global harmonic contribution GhMV from the MV distribution system alone

can be calculated using equation (4.1).

GhMV = a LahMV − (ThHM LhHV ) a (4.1)

An extensive amount of data is required to calculate the value of ThHM as per

IEC 61000-3-6. The most conservative approach is to assume all upstream harmonic

voltages are transferred to the downstream. Thus fraction ThHM is assumed here as unity.

For the purpose of this work only the 5th harmonic is considered as it has been shown to

be the most predominant and problematic for most MV distribution systems [21, 23].

Thus, assessment is made on the basis of the 5th harmonic alone. Three reservations

need to be considered [60].

68
(i) In the future there may be a change in power electronic technology that will

provide a spectrum of current emission different to present day,

(ii) Relatively small high frequency harmonics may provide exceedances of limits as

IEC planning levels become smaller with increasing frequency, and

(iii) Resonances caused by PFC capacitors without detuning reactors can cause

amplification around a particular harmonic order giving prominence to a harmonic

order in a higher frequency range.

4.3 The allocation constant k

When loads are spread out along a feeder and connected to points having different fault

levels, allocation of harmonic current emissions becomes difficult and the methods of

the standard need to be extended. To achieve the constant harmonic power policy

recommended in Section 4.2, the harmonic current emissions need to be allocated in

proportion to maximum demand Si and inversely proportional to the square root of the

harmonic impedance Zhi at the PCC. A suitable strategy is to allocate harmonic current

emissions EIhi using equation (4.2)

1
kS iα
E Ihi = (4.2)
Z hi

where k is termed the allocation constant [60]. The same value of k is used for all loads

supplied from a common substation. Its value is chosen such that when the substation

reaches its maximum capacity, and all loads are contributing their maximum permitted

harmonic contribution, the magnitude of the considered harmonic voltage will have a

value not exceeding the limits suggested by IEC 61000-3-6. It is easy to show that this

voltage will occur at the far end of the 'weakest' feeder.

69
The power 1/α is used in equation (4.2) for the maximum demand as a consequence of

customers connected at the same PCC typically consisting of similar or identical types

of loads. Thus for the case of all customers connected directly to the same PCC the

loads combine using the summation law to produce an undiversified aggregate load.

Exact calculation of k is possible but complex and requires an enormous amount of data.

To illustrate this process a distribution system with each non-linear load modelled as an

equivalent harmonic current source is considered. At harmonic order h, the resulting

voltages are related to the currents as shown by equation (4.3).

[Vh ] = [Z h ][I h ] (4.3)

In equation (4.3) [Vh] is the unknown harmonic voltage vector, [Zh] is the harmonic

impedance matrix, and [Ih] is the harmonic current vector. For a system with N nodes

the expanded form of equation (4.3) is as follows

 V h1   Z h11 Z h12 L Z h1 j L Z h1N   I h1 

V   Z Z h 22 L Z h2 j L Z h 2 N   I h 2 
 h 2   h 21  
 M   M M M M  M 
 =  
 V hi   Z hi1 Z hi 2 L Z hij L Z hiN   I hj 
 M   M M M M  M 
    
V hN   Z hN 1 Z hN 2 L Z hNj L Z hNN   I hN 

Using direct addition the harmonic voltage at node i is given by equation (4.4).

N
Vhi = ∑ Z hij I hj (4.4)
j

As the 95th percentile level voltages are combined using the second summation law

equation (4.4) must be rewritten to include the exponent α as shown by equation (4.5).

N
Vhia = ∑ Z hij
a
I hja (4.5)
j

70
Note that the phase of the harmonic currents and voltages are not considered in equation

(4.5) but are assumed to be taken care of by the use of the summation law exponent α.

Evaluating equations (4.2) and (4.5) and assuming the maximum harmonic voltage to be

less than the global harmonic contribution GhMV the value of k can be determined from

equation (4.6).

G hMV (4.6)
k= 1
  a
 N Z hij
a
⋅Sj 
max ∑ a

i
 j 
 Z hjj2 

Evaluation of equation (4.6) requires the projected maximum demand and system

harmonic impedance at each PCC along every feeder within the local MV distribution

system. To reduce the need for an extensive amount of data some assumptions can be

made to determine an approximate value of the harmonic allocation constant k.

4.4 Incomplete data approach

Although the 'weakest' feeder is strictly defined by equation (4.6), in most cases it will

also be the feeder with the lowest fundamental voltage when the system is loaded to the

fullest extent. Knowledge on the 'weakest' feeder allows an approximation to k to be

obtained when other data is not readily available. Three methods of approximating k

when limited data is available are provided here:

(i) A pessimistic approach assuming all loads other than the 'weakest' feeder loads

are connected to the zone substation busbar (equivalent to assuming all other

feeders to be of zero length).

71
(ii) An approach when all feeders are similar, i.e. all feeders are assumed to have the

same loading and fault level distribution as the 'weakest' feeder.

(iii) The use of (i) incorporating correction factors.

4.4.1 A pessimistic approximation to k

To illustrate how to reduce the amount of data required to calculate the value of the

allocation constant k the radial MV distribution system shown in Figure 4.1 is

considered.

Feeder 1
Zh0 Zh1 Zh2 Zhn

S1 S2 Sn
SF2 Feeder 2
SF3 Feeder 3 (r-1) feeders
SFr Feeder r

S0

Figure 4.1 Example radial MV distribution system

(Sn in feeder 1 is maximum expected demand at each takeoff point;
SFr is the maximum expected loading on feeder r)

To simplify the expression for k given by equation (4.6) it is assumed that all feeders

other than the 'weakest' feeder have zero harmonic impedance, i.e. all loads from the

other feeders are connected at the supply busbar. This assumption simplifies the amount

of data required considerably and can be justified realising that the assumption

overestimates the current on the remaining feeders and hence will be pessimistic.

Assuming that the highest harmonic voltage level will occur at the end of the 'weakest'

feeder the value of k can be estimated using equation (4.7).

72
 G hMV (4.7)
k≈ 1

 n
 a

 ∑ S i Z hi2 + S 0 Z h20 + (S F 2 + ... + S Fr )Z h20 

a a a

 i =1 

The approximation to k consists of three terms in the denominator. These terms are the

harmonic contribution from the 'weakest' feeder, the harmonic contribution from any

local load at the zone substation busbar and the harmonic contribution from the loads on

the other (r-1) feeders.

This approximation requires the projected maximum demand of each customer (Si) and

the system harmonic impedance (Zhi) at all PCC points along the 'weakest' feeder, and

also an estimate of the total maximum demand from the other feeders (SF2, SF3, .., SFr).

Further, the approximation will always ensure a slightly pessimistic result for the value

of k since it underestimates Zhi for the other feeders and therefore allocates too much

when equation (4.2) is applied.

4.4.2 Approximation to k when all feeders are similar

Various studies using the approximate value of k from equation (4.7) have shown that

this approach is most inaccurate when there are a number of weak feeders all of similar

nature. In the case where all feeders are similar in loading and impedance, a less

pessimistic approximation to k may be calculated.

In this case the harmonic contribution at the zone substation busbar due to each of the

other feeders will be equal to that of the 'weakest' feeder. To reflect this the third term in

the denominator of equation (4.7) is modified to give equation (4.8).

73
 G hMV (4.8)
k≈ 1

 n  α

 n
Z ha0 
 ∑ i hi ∑
a a

S Z 2
+ S Z 2
0 h0 + (r − 1 ) S i a 
 i =1 i =1
Z 2 
 hi 

Less data is required to evaluate k using equation (4.8) than is required for

equation (4.7) but the application is useful only when all feeders are of similar nature. It

will be shown in Section 4.5 that this method provides the optimal current allocation for

the trivial example provided in IEC 61000-3-6 standard. However, it is perceived that

this approximation to k will be rarely appropriate for real systems.

4.4.3 Correction factors for the pessimistic approximation of k

By considering the relationship between the harmonic allocation constant and ratio of

impedance at either end of a feeder the pessimistic value of k from equation (4.7) can be

corrected to be less pessimistic if the additional data is available.

By substituting the ratio of sending end to receiving end harmonic impedance into the

exact expression for k given by equation (4.6) a suitable adjustment factor can be

determined. To adjust the value of k from equation (4.7) to be less pessimistic the

harmonic emission contribution from the other feeders, the 3rd component of

denominator in equation (4.7), should be divided by the correcting factor given by

equation (4.9).

1
Z  2α
Fhr =  hnr  (4.9)
 Z h0 

where Zhnr is the harmonic impedance at the end of feeder r. While this expression is

rather complex it may be suited to certain applications where additional data is

74
available. If the system impedance (Zhnr) at the end of each of the other feeders is not

known a rule-of-thumb value of √2 for Fhr has been found to be suited to most systems.

4.5 Case study examples

To illustrate the application of the harmonic allocation constant k, two example systems

are provided. The first is a homogenous example from the IEC 61000-3-6 standard. The

homogenous example demonstrates the agreement of the harmonic allocation constant

method with the IEC 61000-3-6 approach, incorporating the various correction factors.

For the homogenous case the IEC 61000-3-6 approach matches the exact approach

outlined in Section 4.3. The second case study is a distribution system consisting of a

reduced number of feeders, but with greater variation in the loading and fault levels

along each feeder. The additional complexity in the variation of loading and feeder

strength of this example is used to demonstrate the deficiencies in the IEC 61000-3-6

approach.

4.5.1 Homogenous example

The harmonic allocation constant method is applied in conjunction with each of the

consecutive methods outlined in the IEC 61000-3-6 standard, to the example

distribution system provided in Annex E of IEC 61000-3-6. The 20kV distribution

network example is shown in Figure 4.2.

75
 20kV 132kV

5x

1x40MVA 2500MVA
XT=15%
1
2
3
4
6 feeders
5

5km 5km 5km 5km 5km 6

PCC0 PCC1 PCC2 PCC3 PCC4 PCC5

0 1 2 3 4 5 Si=500kVA

Figure 4.2 Homogeneous MV distribution network with six identical feeders

The system consists of six 20kV overhead feeders of 25km length fed by one HV/MV

40MVA transformer. The 50Hz inductive reactance of the lines is assumed to be

approximately 0.35Ω/km and the resistance is ignored. It is assumed that all loads are

directly supplied at MV and the system is at full capacity.

The example calculations are performed only for the 5th harmonic. The planning levels

for the 5th harmonic are assumed to be LhMV=5% and LhHV=2% as per IEC 61000-3-6.

Using these values and the recommended value of α=1.4 for the 5th harmonic the

resulting value for the global harmonic voltage emission GhMV from equation (4.1) is

3.97%.

All feeders in the example are identical, thus any feeder can be chosen as the 'weakest'

feeder for the calculation of the harmonic allocation constant, k. Table 4.1 shows the

results from an exact calculation of k, and the three approximation methods described

previously. As all feeders are identical in this example the approximation using the

assumption of similar feeders produces the same value as the exact value of k.

76
 Table 4.1 Allocation constant k using different calculation methods
Calculation method Allocation constant k
I. Exact value 9.20%
II. Pessimistic value 6.88%
III. Similar feeders value 9.20%
IV. Adjusted pessimistic value 9.09%

Table 4.1 suggests the adjusted pessimistic value gives good results as compared to the

exact value of k. The adjusted pessimistic approximation should be used when feeders

are not all similar and only limited data is available.

The resulting harmonic current allocations of each load along the feeder are shown in

Figure 4.3 for the different methods of calculating k.

35.0
Exact value
Harmonic current (% of fundamental)

30.0
Pessimistic value
Similar feeders value
25.0
Adjusted pessimistic value

20.0

15.0

10.0

5.0

0.0
0 1 2 3 4 5
Node

Figure 4.3 Harmonic current allocations using methods I-IV to calculate k

As can be seen in Figure 4.3 each approximation method provides a suitable value for k.

The voltages arising from the allocated harmonic currents calculated using the second

summation law are shown in Figure 4.4.

77
 8.0
Exact value
7.0 Pessimistic value
Similar feeders value
Adjusted pessimistic value

Harmonic voltage (%)

6.0

5.0

4.0

3.0

2.0

1.0

0.0
0 1 2 3 4 5
Node

Figure 4.4 Harmonic voltages arising at each node using methods I-IV to calculate k

4.5.2 Extreme example

To fully test the application of the proposed method, and highlight the deficiencies of

the IEC 61000-3-6 method, the system in Figure 4.5 containing one weak feeder, one

strong feeder and a large load at the busbar can be considered.

FL1 FL2 FL3

FL0=250MVA S0=5MVA
FL1=61MVA S1=1MVA
FL0 S1 S2 S3 FL2=35MVA S2=1MVA
FL4 FL5 FL3=24MVA S3=1MVA
FL4=222MVA S4=1MVA
FL5=200MVA S5=1MVA
S4 S5

S0

Figure 4.5 Extreme MV distribution system example

As the top feeder in Figure 4.5 is easily identifiable as the weakest feeder the calculation

for the harmonic allocation constant, k is

G hMV
k≈ 1
n α α α α
2 + (S + S + ... + S )Z 2 
 ∑ Si Z hi2 + S0Z h0
 i =1 F2 F3 Fr h0 
 

78
 3.97%
≈ 1
 1.4 1.4 1.4 1.4

1.4 1.4
 0.02 ⋅ 4.10 2 + 0.02 ⋅ 7.14 2 + 0.02 ⋅ 10.42 2 + 0.10 ⋅ 1.00 2 + (0.02 + 0.02) ⋅ 1.00 2 
 
≈ 7.97%

The acceptable harmonic emissions for each of the loads are calculated using the

resultant value of k. For example, for the load at node 3 the 5th harmonic current

allocation is

1 1
kS3α 7.97 ⋅ 0.02 1.4
E Ih3 = = = 0.15% (on a 50MVA base) = 7.56% (on own base)
Z h3 10.42

Table 4.2 provides the harmonic emission calculation results using the IEC 61000-3-6

standard approaches and the proposed allocation method.

Table 4.2 Comparison of different allocation schemes for the extreme case
PCC Load Fifth harmonic Stage 2, Test 1 Stage 2, Test 3 Proposed
node Size impedance Zh=5 EIhi EIhi Stage 2, Test 3
No. MVA (pu) (%)1) (%)1) (%)1)
0 5 1.00 24.2 5.9 15.4
1 1 4.10 9.3 14.4 12.0
2 1 7.14 5.4 10.9 9.1
3 1 10.42 3.7 9.1 7.6
4 1 1.13 34.0 27.6 23.0
5 1 1.25 30.6 26.2 21.8
Resulting harmonic voltages using the above current allocations are given below
Resulting Uhi at Node 0 4.2% 2.9% 3.7%
Resulting Uhi at Node 3 4.9% 5.0% 5.0%
1)
% of load current of each single consumer of maximum demand Si relative to own base

In this example the disadvantages of the previous IEC 61000-3-6 methods are

illustrated. In the first approximation, Stage 2, Test 1, it can be noted that the harmonic

emission allowance for the far end load on the weakest feeder is quite small and the

loads on the strong feeders receive a much larger allowance. In Stage 2, Test 3 the

allocation for the loads at the end of the weakest feeder is seen as being fairer. However,

79
the emission allocation is not determined by load size, and thus the allowance given to

the local load on the busbar is small considering the load is half the system capacity.

The proposed method is a good compromise between the other two methods in that the

local load at the busbar receives a larger proportion of the harmonic emission allowance

as it represents a large proportion of system capacity. At the same time the loads at the

end of the weakest feeder have received a considerable emission allowance.

4.6 Choice of allocation policy

To demonstrate the effect of applying the different allocation policies to customers

along a feeder the results of applying equal harmonic current, equal harmonic power

and equal harmonic voltage policies for the example distribution system of Section

4.5.1 are analysed.

The harmonic current allocations from applying the equal harmonic current, power and

voltage policies are shown in Figure 4.6.

60.0
Harmonic current (% of fundamental)

Equal harmonic current

50.0 Equal harmonic power
Equal harmonic voltage
40.0

30.0

20.0

10.0

0.0
0 1 2 3 4 5
Node

Figure 4.6 Acceptable emissions for equal harmonic current, power and voltage

80
Summing the total allocation of harmonic current from all loads in the system provides

a measure of how well the distribution system's absorption capability is being utilised.

For the case study example in Section 4.5.1 the use of equal harmonic power and equal

harmonic voltage policies increase the amount of total harmonic current allowed to be

injected into the system by 10% and 15% respectively. The equal harmonic voltage

allocation however over penalises the customer at the end of the feeder (node 5).

Taking the increase in total harmonic current into consideration and comparing the

different values in Figure 4.6 it can be found that the allocation using the equal

harmonic power policy has provided a suitable increase in the systems harmonic

capacity without unduly penalising customers at the end of the feeder. The voltages

arising from the different current allocations are shown in Figure 4.7.

8.0
Equal harmonic current
7.0 Equal harmonic power
Harmonic voltage (%)

6.0
Equal harmonic voltage

5.0

4.0

3.0

2.0

1.0

0.0
0 1 2 3 4 5
Node

Figure 4.7 Harmonic voltages arising at each node using different allocation policies

4.7 Summary
A generalised method has been developed to extend the IEC 61000-3-6 approach of

allocating allowable harmonic emissions to the case where customers are distributed

along an MV distribution system feeder having significantly different fault levels. The

method involves the determination of an 'allocation constant' using the agreed loading
81
levels of all customers and the system harmonic impedances. The allocation of

harmonic emission levels using the equal harmonic power rights policy has shown to be

the most useful for determining an ‘allocation constant’ that is ‘fairest’ to all customers.

The work presented in this Chapter is an area that IEC 61000-3-6 fails to cover

adequately making the AS/NZS 61000.3.6 standard unsuitable for application to real

systems.

The approach of allocating emissions to customers described above typically requires an

extensive amount of data. This data may not always be available to the DNSP engineer.

An extension to the generalised method has thus been completed to cater for the

complex situation where only limited data is available. This is achieved by looking at

several extreme cases that categorise the most common MV distribution system feeder

configurations, and through the use of correction factors for the 'allocation constant'.

The developed method has been applied to several study systems. This has allowed the

approximate methods to be compared against the exact form for a range of cases, where

the approximate methods have been shown to provide acceptable results when only

limited data is available. Application to real systems, including monitoring of the

subsequent performance, is an area for further work.

82
 Chapter 5

Identification of MV distribution system aggregated load

models

5.1 Introduction
Accurate harmonic modelling of distribution systems requires a detailed knowledge of

the system parameters and the associated loads connected to the system. Often such

details are unavailable or impractical to obtain. Chapter 3 has demonstrated that

pragmatic modelling of distribution systems can often provide sufficient accuracy when

only limited data is available. To obtain data on the typical behaviour of customer loads

for pragmatic modelling, investigations into a number of typical distribution systems

must be first carried out.

As part of this thesis a harmonic monitoring programme was completed on a typical

MV distribution system owned by Integral Energy with the intention of establishing the

harmonic behaviour of aggregates of load at MV distribution system substations. The

monitoring involved measurements of the harmonic current and voltage from the

residential, commercial, and industrial load sectors. It was proposed that simultaneous

measurements of current and voltage from the different load sectors would allow the

effect on the net distribution system harmonic voltages to be determined. In addition, a

further purpose of the monitoring programme was to benchmark present day harmonic

voltage distortion levels in a typical MV distribution system.

Several harmonic monitoring programmes have been undertaken both overseas [32-34]

and in Australia [22], however there seems to be no consistent approach to conduct,

83
analyse and present the monitoring results. The monitoring programme undertaken in

relation to this thesis is the first systematically conducted monitoring programme of its

kind to be completed in Australia, including measurement of both harmonic current and

voltage at each of the monitoring locations. As a preliminary to the monitoring

programme being implemented four major aspects had to be considered:

(i) Choice of harmonic monitor,

(ii) Choice of monitoring sites,

(iii) Which harmonics to measure, and

(iv) Processing and presentation of results.

This Chapter presents discussion on items (i)-(iv). Several of the monitoring instrument

types considered for implementation of the monitoring programme are tested to ensure

consistency in reporting harmonic voltage levels. While final selection of the type of

monitoring instrument and monitoring sites was not completed as part of this thesis a

short commentary on the process is presented. Preliminary results from the monitoring

program are used to determine which harmonics are of most importance and thus should

be recorded as part of the study.

An investigation into the trends of the harmonic voltages and currents from the

residential, commercial and industrial load sectors is performed to identify a suitable

measure for pragmatic customer load modelling. A small amount of the data collected

from the monitoring programme is used to both establish the load models used in

Chapter 3 and to verify the accuracy of using such models to predict harmonic

distortion levels in distribution systems. Processing and presentation of other significant

results from the monitoring programme are also included in this Chapter.

84
It is proposed that the monitoring programme would give a clear picture of present day

harmonic levels on MV distribution systems and also provide some indication as to the

growth of harmonic levels and the prevailing headroom that remains before limits

recommended by standards are exceeded.

5.2 Selection of monitoring instrument

There are a number of instruments available in the present market for making a

frequency domain study of a time-varying system. The instruments range from versions

with limited features to fully expandable models with various signal-processing options.

With such a diverse range of instruments available it is important to understand the

requirements of a harmonic monitoring programme before the instrument selection is

carried out.

5.2.1 Relevant standards

The standard outlining the measurement techniques and accuracy requirements of

harmonic monitoring instruments in Australia is AS/NZS 61000.4.7 [59], an adaptation

of IEC 61000-4-7 [20]. This standard is quite complex in that it requires a strictly

specified sampling window and calculation of harmonic orders up to the 50th. The

standard classifies the measurement requirements according to the rate of change of the

harmonic levels being recorded. As distribution system loads change continually the

appropriate classification for most harmonic monitoring programs is 'fluctuating

harmonics'. Under 'fluctuating harmonics' classification, IEC 61000-4-7 specifies that

the sampling window be strictly synchronized, must not overlap and there must be no

gaps, i.e. every cycle of the supply in the survey period must be sampled exactly once.

85
There are some practical difficulties in meeting the IEC 61000-4-7 standard with

presently available instrumentation as the standard is relatively new and some parts are

not clearly expressed [17]. Also there are no default tests that can be used to establish if

a particular instrument meets the standard at this stage. Consequently compliance with

the measurement standard of IEC 61000-4-7 could not be confirmed for the available

harmonic monitoring instruments at the time of instrument selection.

As the purpose of monitoring is to establish present day harmonic levels with reference

to present day limits it was necessary to understand the requirements of the harmonic

standards that the monitoring results would be compared against. IEC 61000-3-6 [15]

requires measurement of the following parameters for compliance from a DNSP

perspective

(i) The greatest 95% probability daily value of Uh,vs (rms value of individual

harmonic components over ‘very short’ 3 second periods) should not exceed the

planning level.

(ii) The maximum weekly value of Uh,sh (rms value of individual harmonics over

‘short’ 10 minute periods) should not exceed the planning level.

(iii) The 99.9% weekly value of Uh,vs should not exceed 1.5 to 2 times the planning

level.

The recommended planning levels from IEC 61000-3-6 where given in Table 2.4 of

Chapter 2. The above requirements (i)-(iii) are designed for strict assessment of the

harmonic levels in a DNSP’s power system. However logging of ‘very short’ 3 second

intervals over a period of one week was beyond most power quality instruments due to

the intense memory requirements. Without constantly downloading data via a

86
communications network 3 second logging filled the memory within 2 hours for the

instruments that had a ‘very short’ interval selection. Expansion of memory obviously

needs to be addressed in future designs of harmonic monitoring instruments.

Due to the restricted memory of the monitoring instruments a compromise between (a)

selection of parameters to record and (b) at what logging intervals had to be established.

For the purpose of the monitoring programme it was desired to establish weekly

harmonic current and voltage trends at each of the monitoring sites. At least two weeks

of data would be required to establish weekly patterns. For most of the available

monitoring instruments logging data at 10 or 15 minute intervals suited the available

instrument memory. Logging at such intervals can be justified knowing that damage due

to harmonics usually occurs over longer periods of time, i.e. high levels of harmonics

for a period of less than 3 seconds will not significantly affect overheating of induction

motors. It was also assumed that approximate trends of harmonic loads could be

established without adopting a ‘very short’ interval for logged data.

5.2.2 Instrument accuracy

The selection process for the harmonic monitors used in the benchmark survey involved

laboratory tests to confirm the consistency and limitations of the various makes of

monitoring instruments (for confidentiality reasons the meters will be referred to as

Meters A-E). These tests included the following

(i) Comparison tests involving simultaneous logging of a laboratory supply outlet by

the available monitors over a period of one day, and

(ii) Logging constant voltage harmonics from a programmable waveform generator

including measurement comparisons with a Voltech PM3000A (a power analyser

87
 of high accuracy but inadequate to record time-varying harmonics over an

extended period in the field).

Comparison tests revealed that some of the available instruments are seemingly

unreliable in recording fluctuating harmonics. This was possibly as a result of their

sampling process, where harmonics are recorded using a snapshot without any

averaging. Results from some of the monitoring instruments during the comparison tests

are illustrated in Figures 5.1 and 5.2. Meter C shows definite operational problems in

Figure 5.1 with recordings of near zero harmonic levels during the tests for the 5th

harmonic. In Figure 5.2 Meter F was found to give spurious results with the magnitude

of the higher order harmonics on the laboratory supply possibly being too small in

magnitude for the resolution of the instrument.

Voltage trend of 5th harmonic - Laboratory A phase

4.5

4.0

3.5
% of fundamental

3.0

2.5
Meter A
2.0
Meter B
1.5 Meter C

1.0

0.5

0.0
00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00

Time

Figure 5.1 Comparison of harmonic monitoring instruments at 5th harmonic [17]

88
 Voltage trend of 19th harmonic - Laboratory A phase
0.35

0.30

0.25

% of fundamental
0.20

Meter D
0.15 Meter E
Meter F
0.10

0.05

0.00
00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00

Time

Figure 5.2 Comparison of harmonic monitoring instruments at 19th harmonic

The waveform generator, which is essentially a programmable voltage source inverter

[61], was used to check coherence between the individual meters for quasi-stationary

harmonic voltages. Harmonics of various phase and magnitude were applied to the

monitoring instruments. Of the monitoring instruments tested three gave spurious

results as shown in Figures 5.3 and 5.4. Meter C produced consistently higher

harmonics than the PM3000A and other instruments. It was noted that all meter

manufacturers were unwilling to provide an accurate description of how harmonic

calculations were performed by the instruments. Meters A and E show fluctuating

harmonic voltage levels when the PM3000A and other instruments record levels as near

constant. The magnitude of these errors is small and it is suspected that they are caused

by higher order harmonics (near 10kHz) produced by the waveform generator causing

some difficulties with measurement of the fundamental waveform by the instruments.

89
 Harmonic generator test with 5th Harmonic at 5%
5.6

5.4

5.2

% of fundamental
Meter A
5.0
Meter B
Meter C
4.8
Meter D
Meter E
4.6
PM3000A

4.4

4.2
1 2 3 4 5 6 7 8 9 10

Sample

Figure 5.3 Waveform generator test of monitoring instruments at 5th harmonic

Harmonic generator test with 19th Harmonic at 1%

3.0

2.5
% of fundamental

2.0
Meter A
Meter B
1.5
Meter C
Meter D
1.0 Meter E
PM3000A
0.5

0.0
1 2 3 4 5 6 7 8 9 10

Sample

Figure 5.4 Waveform generator test of monitoring instruments at 19th harmonic

The different makes of instruments becoming available for different periods of the

testing phase complicated the evaluation. In one set of tests it was found that only one

of the three monitors being evaluated was able to give non-spurious results for both

tests, however most instruments provided reasonable consistency for the waveform

generator tests. It was also noted that some instruments performed well at lower

frequencies but poorly at higher frequencies such as at the 19th harmonic.

90
5.2.3 Instrument price
The price of the instruments considered for the harmonic monitoring programme varied

from $1,500 to $9,000 for different makes and models. Software capabilities and

flexibility of the meters were the greatest price driving factors. Accuracy of the

instruments was unable to be confirmed and thus no comparison can be made between

price and accuracy. From a DNSP’s perspective an instrument that is capable of

performing all required tasks at the lowest price would be a compelling reason for its

selection.

5.2.4 Further instrument considerations

Once the requirements of the standards used to evaluate results are established and a

price range is settled, there are still a number of factors to consider before the

appropriate monitoring instrument can be selected. Some further important

considerations when selecting a harmonic monitor include the following (in no

particular order):

(i) Types and number of parameters which can be recorded,

(ii) Available memory for data storage,

(iii) Data storage and transfer formats,

(iv) Sampling speed,

(v) Environmental requirements (such as ruggedness), and

(vi) Additional functions.

The Integral Energy system is spread out over an area of 24,500 square kilometres and

the University of Wollongong is situated near one extreme of the system. Thus, for

geographical reasons it was important that the harmonic monitoring instruments had

91
remote set-up, initiation and downloading capabilities, as it was possible that the chosen

monitoring site would be a considerable distance away from the University.

For numerous cases monitoring programmes will not be performed solely for the

purpose of monitoring harmonics but rather as a component of a complete power quality

survey. As such there may also be additional factors that should be considered in

instrument selection not mentioned here. In the future it is perceived that most standard

customer metering instruments will contain functionality that will allow utilities to

monitor power quality disturbances such as harmonics on a continuous basis.

5.2.5 Instrument selection

The instrument used for the harmonic monitoring programme was selected through

consultation with Integral Energy. Using the criteria discussed in the previous sections,

including results from the earlier instrument tests, the meter selected was an EDMI Mk3

Energy Meter from Electronic Design and Manufacturing International Pty. Ltd, as

shown in Figure 5.5. Further details of the EDMI Energy Meter can be found in

Appendix B.

Figure 5.5 EDMI 2000-04XX Energy Meter [62]

92
5.3 Monitoring site selection
The purpose of the harmonic monitoring programme was to produce results that are

typical to MV distribution systems as a whole rather than being specific to an individual

MV distribution system. Ideally a number of study systems should be used to produce

such results. However, due to limited access to DNSP and customer systems,

availability of monitoring instruments, time constraints, and financial limitations, only

one system could be monitored for this project. The site selection process was critical to

extract results that would be representative of typical MV distribution systems.

The criteria for selecting a suitable monitoring site were that the distribution system had

high impedance and thus anticipated harmonic problems, and a good mix of residential,

commercial and industrial load types. It was also considered that future measurements

could be made to establish the growth of harmonics with the introduction of new load

types rather than growth in load. For this reason a distribution system that was nearly

fully loaded and unlikely to change configuration over the next few years was preferred.

To isolate harmonic emission contributions from residential, commercial and industrial

load types it was hoped that a substation could be found with three feeders consisting

solely of each load type. It was initially hoped that a zone substation could be found

where measurements at all sites could be completed at 11kV.

In consultation with Integral Energy a suitable zone substation was found for the

harmonic monitoring programme to satisfy the necessary criteria. Homepride is a

typical 33/11kV zone substation in the Liverpool area of Sydney that supplies ten 11kV

radial feeders. The zone substation is supplied at 33kV from the bulk supply point of

Western Liverpool transmission network. Figure 5.6 gives the layout of the zone

93
substation and feeder system for the harmonic monitoring programme. The substation

maximum demand was approximately 22MVA (80% of capacity with N-1 transformer

redundancy) and the short circuit level at the 11kV busbar was approximately 213MVA.

The high impedance zone substation transformers, each having a leakage reactance of

approximately 17%, combined with some substantially long overhead feeders suited the

requirement of the distribution system to have relatively high impedance.

HOMEPRIDE DISTRIBUTION SYSTEM

33kV Transmission System

33/11kV
1 25MVA
11kV Zone Substation
2 3 4

Other Feeders

KEY
Harmonic monitor
Pole top transformer
Substation transformer

7
6

Predominantly Predominantly Predominantly

Residential Commercial Industrial
Feeder Feeder Feeder
Figure 5.6 Single line diagram illustrating the Homepride zone distribution system

Typically most MV distribution systems in Australia are radial in structure while HV

and EHV systems tend to be meshed. With the Homepride distribution system being
94
radial it was decided to install 7 monitors, a monitor at each of the residential,

commercial and industrial sites, a monitor at the sending end of the three individual

feeders, and a monitor at the zone substation incoming supply. The sites to be

monitored within the distribution system are outlined in Table 5.1.

Table 5.1 Harmonic monitoring instrument positions

Site ID Purpose of monitoring instrument
1 Measure total voltage and current at the 11kV zone
substation busbar
2-4 Measure the voltage and current at sending end of each of
the predominantly residential, commercial and industrial
feeders
5-7 Measure voltage and current at typical residential,
commercial and industrial sites along the feeder

Measurements at sites 1-4 were made by means of the substation voltage and current

transformers. It is suggested in [59] that there should be no problems with the voltage

transformers used as they were magnetic type. Work presented in [63] had established

that the current transformers would have adequate bandwidth for these measurements.

Further, on older Integral Energy distribution systems, metering voltage and current

transformers are not usually available for the B phase at the 11kV busbar, thus at sites

1-4 measurements were to be recorded for Vab, Vcb, Ia and Ic only.

The EDMI Mk3 Energy Meter was not weatherproof and had to be connected at an

enclosure. The MV side of distribution transformers is not accessible for most metering

applications. Instead sites 5-7 were monitored on the LV side of 11kV/415V

distribution transformers using direct connection for the voltage leads and clip-on

current transformers for the current sensing. All three line-to-neutral voltages and line

currents were recorded at the LV locations [17].

95
Sites 1-4 in Figure 5.6 are all within the substation at the sending end of the feeders

identified as being of a predominant load type. Site 5 was the last enclosed type

substation (pad mount) along the feeder route located approximately 2km from

Homepride zone substation. The remaining 2km of feeder is overhead with pole-top

mounted transformers. The site fed a recently developed residential area. Site 6 supplies

a Westfield shopping centre with a couple of large supermarkets and many small shops.

Site 7 supplies a factory manufacturing paper product such as paper towels, toilet paper

and tissues.

Based on distribution customer details obtained from Integral Energy the three feeders

in question have an estimated proportioning of load type as follows:

(i) Predominantly residential feeder – 85% residential, 15% commercial

(ii) Predominantly commercial feeder – 90% commercial, 10% residential

(iii) Predominantly industrial feeder – 75% industrial, 20% commercial, 5%

residential

The load breakdown of the three LV sites is estimated to be:

(i) Site 5: Residential transformer – 90% residential, 10% commercial

(ii) Site 6: Commercial transformer – 100% commercial

(iii) Site 7: Industrial transformer – 100% industrial

5.4 Selection of harmonics to measure

The memory capabilities of the EDMI Mk3 Energy Meter at the time of purchase

limited recordings to the fundamental current and voltage in each phase, the current and

96
voltage THD in each phase, and only three individual harmonics in each phase. For the

initial harmonic monitoring programme the harmonics chosen were as follows:

(i) 5th harmonic: had been shown to dominate in previous national and international

harmonic surveys [21, 22, 54] and perceived as the most problematic.

(ii) 19th harmonic: highest harmonic of importance for the then existing Australian

harmonic standard AS 2279.2 [12]

(iii) 49th harmonic: close to the highest harmonic of importance for the new Australian

harmonic standard AS/NZS 61000.3.6 [16]

After the initial monitoring programme had been completed it was found that the 19th

harmonics and 49th harmonics were of insignificant magnitude. Some spot checks of the

harmonic spectrum of the current and voltages at the individual monitoring sites found

that the 3rd and 7th harmonics were the next most significant harmonics. The harmonic

monitors were then reconfigured to record the 3rd and 7th harmonic currents and voltages

at each of the monitoring sites.

The harmonic monitors were set to record the 3 second average fundamental and

harmonic voltages and currents on each of the three phases. The initial monitoring

programme was planned to produce two weeks of harmonic data. The memory

restrictions of the monitoring instruments dictated that each parameter was recorded

only every 15 minutes. This was later shortened to every 10 minutes to more closely

comply with IEC 61000-3-6 but had the disadvantage of having to download twice to

obtain a full two weeks of data. The monitoring programme was to be repeated after 12

months to obtain an indication of harmonic growth. Seasonal effects of load types, such

97
as air-conditioners in summer, created the need to download more frequently to

illustrate the harmonic levels through different seasons.

5.5 Results from harmonic measurements at selected sites

The monitoring programme was completed in three phases:

(i) Four weeks of monitoring to benchmark harmonic levels and the contributions

from the residential, commercial and industrial load sectors,

(ii) Two weeks of monitoring after twelve months from the initial study and then

annually for the duration of the project with the aim to establish growth of

harmonic distortion levels, and

(iii) Ongoing monitoring to establish seasonal changes, global behaviour of the

different load types, and to confirm the validity of the monitoring programmes in

(i) and (ii) above.

The data retrieved from the harmonic monitoring programme spans from August 1999

to December 2002. The monitoring programme involved 20 downloads of two week

long blocks of data from the EDMI Mk3 Energy Meters, numerous snapshot captures of

the harmonic content of currents and voltages at each of the monitoring locations, and

instantaneous waveform captures at the Homepride zone and commercial substation

locations. In total about 200Mb of data was collected for analysis.

Considering the large amount of data obtained in a harmonic monitoring programme it

was necessary to summarise the data into a format that can be easily analysed. Careful

presentation of results is necessary to ensure that the important characteristics of the

monitoring data are not lost in summarising. For this monitoring programme the data

98
from the seven monitoring locations has been organised into three main sections

including daily and weekly trending of harmonics, the type of statistical distribution,

and the correlation between the harmonics in each of the three phases.

5.6 Harmonic voltage and current trends

The initial harmonic monitoring programme consisted of four weeks of monitoring with

two consecutive downloads of two week long blocks of data from each of the seven

EDMI Mk3 Energy Meter locations. This included monitoring data from 21 August

1999 to 17 September 1999. The aim of this monitoring programme was to examine the

harmonics to determine the contribution to the system harmonic levels from the

residential, commercial and industrial load sectors. The expected outcome in simple

terms was to determine the existing harmonic situation in a typical 11kV distribution

system.

Figure 5.7 shows typical recordings at the Homepride zone substation 11kV busbar

daily and weekly harmonic voltage trends averaged over the two line-to-line phases

measured. The 5th harmonic and the voltage THD are very close to each other verifying

that the 5th harmonic is the dominant distorting component. The 19th harmonic is low at

approximately 0.1%, which is close to the limit of resolution of the instrument. The 49th

harmonic recorded was insignificant. Figure 5.8 illustrates how the daily harmonic

voltage trends for the 5th harmonic and THD are repeated, in an approximate form, over

each of the weekdays and somewhat different over the weekend period.

99
 Homepride zone substation daily voltage trend
2.0
1.8
1.6

Voltage distortion (%)

1.4
1.2 THD

1.0 5th
19th
0.8
49th
0.6
0.4
0.2
0.0
00:00 04:00 08:00 12:00 16:00 20:00 00:00

Figure 5.7 Homepride zone (site 1) weekday harmonic voltage trend (5th, 19th, 49th)
Homepride zone substation weekly voltage trend
2.0
1.8
1.6
Voltage distortion (%)

1.4
THD
1.2
5th
1.0
19th
0.8 49th
0.6
0.4
0.2
0.0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.8 Homepride zone (site 1) weekly harmonic voltage trend (5th, 19th, 49th)
Homepride zone substation weekly current trend
25 1250

Total
harmonic
20 1000
current
Fundamental current (A)
Harmonic current (A)

5th

15 750
19th

10 500
49th

5 250 Fund

0 0
00:00 04:00 08:00 12:00 16:00 20:00 00:00

Figure 5.9 Homepride zone (site 1) weekday harmonic current trend (5th, 19th, 49th)

100
 Homepride zone substation weekly current trend
30 1200

Total
25 1000 harmonic

Fundamental current (A)

current

Harmonic current (A)

5th
20 800

19th
15 600

49th
10 400

Fund
5 200

0 0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.10 Homepride zone (site 1) weekly harmonic current trend (5th, 19th, 49th)

The harmonic current trends for Homepride are shown in Figure 5.9 and 5.10. There is a

noticeable difference in the magnitude of the 5th harmonic current and the current THD.

This suggests that the 5th harmonic may not be the most dominant harmonic component

of the current for some parts of the day. Snapshots of the harmonic content of the

current confirmed that the 7th harmonic was of similar order of magnitude as the 5th

harmonic during some periods of the day.

As illustrated in Figures 5.7 to 5.10 the initial monitoring program found that the 19th

harmonic and 49th harmonic were insignificant in magnitude. Previous monitoring

programs overseas [21, 22] had found that the more common problematic harmonics

included 3rd, 5th and 7th. Although a large component of the 3rd harmonic is blocked by

the MV/LV delta-star transformers used widely in Australia, if the system is unbalanced

or the harmonics are out of phase a significant level of 3rd harmonic can exist on the

MV system as well. After a number of harmonic snapshots of all monitoring locations it

was confirmed that the dominant harmonics were indeed the 3rd, 5th and 7th harmonics.

101
It was decided to alter the monitoring programme to replace the 19th and 49th harmonic

recordings with the 3rd and 7th harmonics.

Homepride zone substation weekly voltage trend

2.0

1.8

1.6
Voltage distortion (%)

1.4

1.2 THD
3rd
1.0
5th
0.8 7th
0.6

0.4

0.2

0.0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.11 Homepride zone (site 1) weekly harmonic voltage trend (3rd, 5th, 7th)
Homepride zone substation weekly current trend
30 1200

Total
25 1000 harmonic
current

Fundamental current (A)

Harmonic current (A)

3rd
20 800

5th
15 600

7th
10 400

Fund
5 200

0 0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.12 Homepride zone (site 1) weekly harmonic current trend (3rd, 5th, 7th)

The weekly harmonic voltage and current trends for the new harmonics recorded are

shown in Figure 5.11 and 5.12. The 3rd, 5th and 7th harmonics make up the major

components of the current THD. Although the 5th harmonic current has the highest peak

magnitude, the 7th harmonic current is the largest harmonic component for a significant

period of each day. Interestingly the peak 7th harmonic current periods correspond to

low 7th harmonic voltage recordings and vice versa. This is due to the 7th harmonic

102
current producing a voltage out of phase with the existing 7th harmonic background

voltage distortion or larger loads decreasing network impedance and thus acting as a

harmonic filter as documented in [64]. However it would be difficult to find the source

of these harmonics without further coordinated monitoring. The 3rd harmonic displays

the more expected trend of increasing harmonic current producing an increase in

harmonic voltage.

The weekly harmonic voltage and current trends for the remaining six harmonic

monitoring sites (sites 2-7) are shown in Figures 5.13 to 5.21. At all monitoring

locations it was found that the harmonic magnitudes followed a similar trend each day.

The magnitude of the 3rd harmonic current at the residential and commercial sites was

more significant than at the other locations, mostly due to the abundance of single phase

loads. The harmonic voltage at the residential substation, illustrated in Figure 5.13,

remained slightly less than the Homepride zone substation busbar harmonic voltage for

the entire monitoring period. This suggests that the contribution to the total harmonic

voltage at the zone substation from the residential loads was relatively small. This is

seemingly not consistent with suggestion that significant increase in harmonic voltage

levels in Europe are due to the use of televisions in the late afternoon periods [5].

103
 Residential substation weekly voltage trend
2.0

1.8

1.6

Voltage distortion (%)

1.4
1.2 THD
3rd
1.0
5th
0.8 7th
0.6
0.4

0.2

0.0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.13 Residential substation (site 5) weekly harmonic voltage trend

Residential feeder weekly current trend
10 200

Total
8 160 harmonic

Fundamental current (A)

current
Harmonic current (A)

3rd
6 120
5th

4 80 7th

Fund
2 40

0 0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.14 Residential feeder (site 2) weekly harmonic current trend

Residential substation weekly current trend
15.0 150

12.5 125 Total

harmonic
Fundamental current (A)

current
Harmonic current (A)

10.0 100 3rd

5th
7.5 75
7th
5.0 50
Fund
2.5 25

0.0 0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.15 Residential substation (site 5) weekly harmonic current trend

104
 Commercial substation weekly voltage trend
4.0

3.5

3.0

Voltage distortion (%)

2.5 THD
3rd
2.0 5th
7th
1.5

1.0

0.5

0.0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.16 Commercial substation (site 6) weekly harmonic voltage trend

Commercial feeder weekly current trend
7 140

6 120 Total
harmonic

Fundamental current (A)

current
Harmonic current (A)

5 100 3rd

4 80
5th

3 60
7th

2 40
Fund
1 20

0 0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.17 Commercial feeder (site 3) weekly harmonic current trend

Commercial substation weekly current trend
100 700

90 630
Total
80 560 harmonic
Fundamental current (A)
Harmonic current (A)

current
70 490 3rd
60 420
5th
50 350
40 280 7th
30 210
Fund
20 140

10 70

0 0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.18 Commercial substation (site 6) weekly harmonic current trend

105
The commercial substation had distortion levels up to twice the level of the Homepride

zone substation busbar harmonic voltage. The largest voltage distortion was produced

by the commercial substation reaching a maximum value of around 5% during the

initial monitoring period, and up to 6% a number of times during the three year

monitoring period. Figures 5.17 and 5.18 illustrate that the 3rd and 5th harmonics are the

significant harmonic components of the current on the commercial feeder. The

industrial and commercial feeders seem to contribute approximately the same level of

7th harmonic current while that of the residential feeder is significantly higher.

Industrial substation weekly voltage trend

2.0
1.8

1.6
Voltage distortion (%)

1.4

1.2 THD
1.0 3rd
5th
0.8
7th
0.6
0.4

0.2

0.0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.19 Industrial substation (site 7) weekly harmonic voltage trend

Industrial feeder weekly current trend

12 200

180
10 Total
160 harmonic
Fundamental current (A)

current
Harmonic current (A)

140 3rd
8
120
5th
6 100
80 7th
4
60

40 Fund
2
20

0 0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.20 Industrial feeder (site 4) weekly harmonic current trend

106
 Industrial substation weekly current trend
300 1800

250 1500 Total

harmonic

Fundamental current (A)

Harmonic current (A)
current
200 1200 3rd

5th
150 900

7th
100 600

Fund
50 300

0 0
Sat Sun Mon Tue Wed Thu Fri Sat

Figure 5.21 Industrial substation (site 7) weekly harmonic current trend

Figures 5.11, 5.13, 5.16 and 5.19 illustrate that the maximum harmonic voltage levels

occur during normal working hours. The blocking of the 3rd harmonic by the delta-star

MV/LV transformers is also substantiated by the high levels of 3rd harmonic current at

the LV load sites but not at the sending end of the MV feeders. Figure 5.21 shows a

sustained high level of fundamental current for the seven day monitoring program. This

is typical of an industrial load where the manufacturing operation continues 24 hours a

day seven days per week with limited downtime.

There is a small increase in the 95th percentile 5th harmonic voltage level at the

industrial substation compared to that at the corresponding feeder sending end.

However, the industrial load draws significant 3rd and 5th harmonic currents. The line

impedance between the Homepride zone substation and the industrial substation is low

and thereby the increase in 95th percentile cumulative probability voltage due to the high

harmonic current is minimal.

107
The voltage trends from all sites illustrate the dominance of the 5th harmonic voltage.

This agrees with the results from previous surveys conducted overseas and discussed in

Chapter 2. The dominance of the 5th harmonic is also illustrated using a scatter graph of

5th harmonic voltage versus THD as shown in Figure 5.22. The high correlation of the

scatter graph illustrates how the THD rises in proportion to the 5th harmonic. This result

allows the work of Chapter 3 of estimating voltage THD levels to be based primarily on

the 5th harmonic.

th
Homepride zone substation THD vs 5 harmonic
2.0

1.8

1.5

1.3
THD (%)

1.0

0.8

0.5

0.3

0.0
0.0 0.3 0.5 0.8 1.0 1.3 1.5 1.8 2.0
5th harmonic distortion (%)

Figure 5.22 Homepride zone (site 1) THD versus 5th harmonic

5.7 Harmonic voltage and current cumulative probability

From the harmonic voltage and current trends it can be seen that the changing load

causes harmonic levels to constantly fluctuate (albeit in a repeating pattern) over the

monitoring period. Due to faults and the most adverse combination of customer loads it

would be difficult for a DNSP to control the level of harmonics 100% of the time. For

these reasons it is necessary to incorporate statistical techniques to provide an

assessment method of MV distribution systems.

108
The concept of the 95th percentile cumulative probability used in [16] has already been

introduced in Chapter 2. An advantage of the cumulative probability method is it

enables comparisons without spurious or fluctuating data affecting the reported indices.

For the 95th percentile this is a direct result of the highest 5% of readings, containing

most of the outliers, being excluded from the cumulative probability indicator. Selection

of the 95th percentile rather than the 99th or 100th (maximum) can be justified using data

from the monitoring programme. Figure 5.23 illustrates weekly 95th, 99th and 100th

percentiles for the 7th harmonic voltage at the industrial site. As illustrated the variation

of the latter percentiles show a large variation for adjacent weeks, whereas the 95th

percentile, while still demonstrating weekly variations produces a more stable index.

th th th th
95 , 99 and 100 percentile 7 harmonic voltage
4.00

3.50
Voltage distortion (%)

3.00

2.50
7th-CP95
2.00
7th-CP99

1.50 7th-CP100

1.00

0.50

0.00
02-Nov

16-Nov

30-Nov
01-Jun

15-Jun

29-Jun

13-Jul

27-Jul

10-Aug

24-Aug

07-Sep

21-Sep

05-Oct

19-Oct

Date

Figure 5.23 95th, 99th and 100th percentile 7th harmonic voltage at industrial substation
(site 7)

Figures 5.24 to 5.35 illustrate the 3rd, 5th and 7th harmonic voltage and current

cumulative probability plots for all the monitoring locations throughout the Homepride

network of Figure 5.6. The 19th and 49th harmonics have not been included due to their

magnitude being relatively insignificant. The cumulative probability plots and

histograms illustrate that the magnitude of the 5th harmonic voltage and current

109
approximates a normal distribution for all monitoring locations. The 3rd harmonic also

follows this trend at all locations except at the commercial substation. The 7th harmonic

voltages and currents do not follow this normal distribution.

Histogram of Homepride 3rd harmonic voltage Histogram of Homepride 3rd harmonic current

140 100% 80 100%

90% 70 90%
120
80% 80%

Cumulative frequency
Cumulative frequency
60

Count of samples
Count of samples

100 70% 70%

60% 50 60%
80 Ia
Vab
50% 40 50%
60 Vbc Ic
40% 30 40%
40 30% 30%
20
20% 20%
20 10
10% 10%
0 0% 0 0%
0.00
0.02
0.03
0.05
0.06
0.08
0.09
0.11
0.12
0.14
0.15
0.17
0.18
0.20
0.21
0.23
0.24
0.26
0.27
0.29
0.30
0.32
0.33
0.35
0.36
0.38

0.00
0.28
0.56
0.85
1.13
1.41
1.69
1.97
2.26
2.54
2.82
3.10
3.38
3.67
3.95
4.23
4.51
4.79
5.08
5.36
5.64
5.92
6.20
6.49
6.77
7.05
Voltage distortion (%) Current distortion (A)

Figure 5.24 Homepride zone substation (site 1) 3rd harmonic histogram

Histogram of Residential 3rd harmonic voltage Histogram of Residential 3rd harmonic current

160 100% 120 100%

140 90% 90%

100
80% 80%

Cumulative frequency
Cumulative frequency

120
Count of samples
Count of samples

70% 70%
80
100 60% 60%
Va Ia
80 50% Vb 60 50% Ib
40% Vc 40% Ic
60
40
30% 30%
40
20% 20%
20
20 10% 10%
0 0% 0 0%
0.00
0.01
0.03
0.04
0.05
0.07
0.08
0.10
0.11
0.12
0.14
0.15
0.16
0.18
0.19
0.20
0.22
0.23
0.24
0.26
0.27
0.29
0.30
0.31
0.33
0.34

0.00
0.33
0.66
0.99
1.32
1.65
1.98
2.31
2.64
2.97
3.30
3.63
3.96
4.30
4.63
4.96
5.29
5.62
5.95
6.28
6.61
6.94
7.27
7.60
7.93
8.26
Voltage distortion (%) Current distortion (A)

Figure 5.25 Residential substation (site 5) 3rd harmonic histogram

Histogram of Commercial 3rd harmonic voltage Histogram of Commercial 3rd harmonic current

250 100% 250 100%

90% 90%
Cumulative frequency
Cumulative frequency

200 80% 200 80%

Count of samples
Count of samples

70% 70%
Va 150 60% Ia
150 60%
Vb 50% Ib
50%
Vc 100 40% Ic
100 40%
30% 30%
50 20% 50 20%
10%
10%
0 0%
0 0%
0.00
2.12
4.24
6.36
8.49
10.61
12.73
14.85
16.97
19.09
21.22
23.34
25.46
27.58
29.70
31.82
33.95
36.07
38.19
40.31
42.43
44.55
46.68
48.80
50.92
53.04
0.00
0.07
0.14
0.21
0.28
0.36
0.43
0.50
0.57
0.64
0.71
0.78
0.85
0.93
1.00
1.07
1.14
1.21
1.28
1.35
1.42
1.50
1.57
1.64
1.71
1.78

Voltage distortion (%) Current distortion (A)

Figure 5.26 Commercial substation (site 6) 3rd harmonic histogram

110
 Histogram of Industrial 3rd harmonic voltage Histogram of Industrial 3rd harmonic current

200 100% 300 100%

180 90% 90%
250

Cumulative frequency
160 80% 80%

Cumulative frequency

Count of samples
Count of samples
140 70% 70%
200
60% Ia
120 60% Va
150 50% Ib
100 50% Vb
40% Ic
80 40% Vc 100
30%
60 30%
20%
40 20% 50
10%
20 10%
0 0%
0 0%

0.00
5.84
11.68
17.52
23.35
29.19
35.03
40.87
46.71
52.55
58.38
64.22
70.06
75.90
81.74
87.58
93.41
99.25
105.09
110.93
116.77
122.61
128.44
134.28
140.12
145.96
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.15
0.17
0.19
0.21
0.23
0.25
0.27
0.29
0.31
0.33
0.35
0.37
0.40
0.42
0.44
0.46
0.48
0.50
0.52
Voltage distortion (%) Current distortion (A)

Figure 5.27 Industrial substation (site 7) 3rd harmonic histogram

Histogram of Homepride 5th harmonic voltage Histogram of Homepride 5th harmonic current

70 100% 60 100%
90% 90%
60 50
80%

Cumulative frequency
80%
Cumulative frequency

Count of samples
Count of samples

50 70% 70%
40
60% 60%
40 Ia
Vab 30 50%
50%
Vbc Ic
30 40% 40%
20
30% 30%
20
20% 20%
10
10 10%
10%
0 0% 0 0%
0.00
0.84
1.68
2.52
3.37
4.21
5.05
5.89
6.73
7.57
8.42
9.26
10.10
10.94
11.78
12.62
13.47
14.31
15.15
15.99
16.83
17.67
18.52
19.36
20.20
21.04
0.00
0.07
0.13
0.20
0.27
0.34
0.40
0.47
0.54
0.60
0.67
0.74
0.81
0.87
0.94
1.01
1.08
1.14
1.21
1.28
1.34
1.41
1.48
1.55
1.61
1.68

Voltage distortion (%) Current distortion (A)

Figure 5.28 Homepride zone substation (site 1) 5th harmonic histogram

Histogram of Residential 5th harmonic voltage Histogram of Residential 5th harmonic current

60 100% 60 100%
90% 90%
50 50
80%

Cumulative frequency
80%
Cumulative frequency

Count of samples
Count of samples

70% 70%
40 40
60% Va 60% Ia
30 50% Vb 30 50% Ib
40% Vc 40% Ic
20 20
30% 30%
20% 20%
10 10
10% 10%
0 0% 0 0%
0.00
0.06
0.13
0.19
0.25
0.32
0.38
0.45
0.51
0.57
0.64
0.70
0.76
0.83
0.89
0.95
1.02
1.08
1.14
1.21
1.27
1.34
1.40
1.46
1.53
1.59

0.00
0.15
0.31
0.46
0.62
0.77
0.92
1.08
1.23
1.39
1.54
1.69
1.85
2.00
2.16
2.31
2.46
2.62
2.77
2.93
3.08
3.23
3.39
3.54
3.70
3.85

Voltage distortion (%) Current distortion (A)

Figure 5.29 Residential substation (site 5) 5th harmonic histogram

Histogram of Commercial 5th harmonic voltage Histogram of Commercial 5th harmonic current

70 100% 80 100%
90% 70 90%
60
Cumulative frequency

80%
Cumulative frequency

80%
Count of samples

60
Count of samples

50 70% 70%
Va 50 60% Ia
40 60%
Vb 40 50% Ib
50%
30 Vc 40% Ic
40% 30
30% 30%
20 20
20%
20%
10 10 10%
10%
0 0%
0 0%
0.00
3.15
6.29
9.44
12.58
15.73
18.87
22.02
25.16
28.31
31.45
34.60
37.74
40.89
44.03
47.18
50.32
53.47
56.61
59.76
62.90
66.05
69.19
72.34
75.48
78.63
0.00
0.15
0.30
0.45
0.60
0.75
0.90
1.05
1.20
1.35
1.50
1.65
1.80
1.95
2.10
2.25
2.40
2.55
2.70
2.85
3.00
3.15
3.30
3.45
3.60
3.75

Voltage distortion (%) Current distortion (A)

Figure 5.30 Commercial substation (site 6) 5th harmonic histogram

111
 Histogram of Industrial 5th harmonic voltage Histogram of Industrial 5th harmonic current

60 100% 70 100%
90% 90%
60
50

Cumulative frequency
80% 80%

Cumulative frequency

Count of samples
50
Count of samples
70% 70%
40
60% 40 60% Ia
Va
30 50% 50% Ib
Vb
30 40% Ic
40% Vc
20 20 30%
30%
20%
20% 10
10 10%
10%
0 0%
0 0%

0.00
9.81
19.62
29.44
39.25
49.06
58.87
68.68
78.50
88.31
98.12
107.93
117.74
127.56
137.37
147.18
156.99
166.80
176.62
186.43
196.24
206.05
215.86
225.68
235.49
245.30
0.00
0.08
0.16
0.24
0.32
0.40
0.48
0.56
0.64
0.72
0.80
0.88
0.96
1.03
1.11
1.19
1.27
1.35
1.43
1.51
1.59
1.67
1.75
1.83
1.91
1.99
Voltage distortion (%) Current distortion (A)

Figure 5.31 Industrial substation (site 7) 5th harmonic histogram

Histogram of Homepride 7th harmonic voltage Histogram of Homepride 7th harmonic current

100 100% 80 100%

90 90% 70 90%

Cumulative frequency
80 80% 80%
Cumulative frequency
60

Count of samples
Count of samples

70 70% 70%
50 60%
60 60%
Vab 40 50% Ia
50 50%
Vbc Ic
40 40% 30 40%

30 30% 30%
20
20 20% 20%
10 10%
10 10%
0 0%
0 0%
0.00
0.69
1.38
2.07
2.76
3.46
4.15
4.84
5.53
6.22
6.91
7.60
8.29
8.99
9.68
10.37
11.06
11.75
12.44
13.13
13.82
14.52
15.21
15.90
16.59
17.28
0.00
0.02
0.05
0.07
0.09
0.12
0.14
0.17
0.19
0.21
0.24
0.26
0.28
0.31
0.33
0.35
0.38
0.40
0.42
0.45
0.47
0.50
0.52
0.54
0.57
0.59

Voltage distortion (%) Current distortion (A)

Figure 5.32 Homepride zone substation (site 1) 7th harmonic histogram

Histogram of Residential 7th harmonic voltage Histogram of Residential 7th harmonic current

90 100% 80 100%

80 90% 70 90%
80% 80%

Cumulative frequency
Cumulative frequency

70 60
Count of samples
Count of samples

70% 70%
60
50 60%
60% Va Ia
50
50% Vb 40 50% Ic
40
40% Vc 40% Ib
30
30
30% 30%
20
20 20% 20%
10 10% 10 10%
0 0% 0 0%
0.00
0.02
0.04
0.07
0.09
0.11
0.13
0.16
0.18
0.20
0.22
0.25
0.27
0.29
0.31
0.34
0.36
0.38
0.40
0.43
0.45
0.47
0.49
0.52
0.54
0.56

0.00
0.11
0.22
0.33
0.44
0.55
0.66
0.77
0.88
0.99
1.10
1.21
1.32
1.42
1.53
1.64
1.75
1.86
1.97
2.08
2.19
2.30
2.41
2.52
2.63
2.74

Voltage distortion (%) Current distortion (A)

Figure 5.33 Residential substation (site 5) 7th harmonic histogram

Histogram of Commercial 7th harmonic voltage Histogram of Commercial 7th harmonic current

90 100% 100 100%

80 90% 90 90%
Cumulative frequency

80 80%
Cumulative frequency

70 80%
Count of samples
Count of samples

70% 70 70%
60
60% Va 60 60% Ia
50 Vb 50 50% Ic
50%
40 Vc 40 40% Ib
40%
30 30% 30 30%
20 20% 20 20%
10 10% 10 10%
0 0%
0 0%
0.00
0.55
1.11
1.66
2.22
2.77
3.33
3.88
4.44
4.99
5.55
6.10
6.66
7.21
7.77
8.32
8.88
9.43
9.99
10.54
11.10
11.65
12.21
12.76
13.32
13.87
0.00
0.04
0.07
0.11
0.15
0.19
0.22
0.26
0.30
0.33
0.37
0.41
0.45
0.48
0.52
0.56
0.60
0.63
0.67
0.71
0.74
0.78
0.82
0.86
0.89
0.93

Voltage distortion (%) Current distortion (A)

Figure 5.34 Commercial substation (site 6) 7th harmonic histogram

112
 Histogram of Industrial 7th harmonic voltage Histogram of Industrial 7th harmonic current

90 100% 70 100%
90% 90%
80 60
80%

Cumulative frequency
80%

Cumulative frequency
70

Count of samples
50
Count of samples

70% 70%
60
60% 40 60% Ia
Va
50 50%
50% Vb Ic
40 30 40% Ib
40% Vc
30 20 30%
30%
20 20% 20%
10
10 10% 10%
0 0%
0 0%

0.00
1.84
3.69
5.53
7.37
9.21
11.06
12.90
14.74
16.59
18.43
20.27
22.11
23.96
25.80
27.64
29.48
31.33
33.17
35.01
36.86
38.70
40.54
42.38
44.23
46.07
0.00
0.06
0.12
0.18
0.23
0.29
0.35
0.41
0.47
0.53
0.58
0.64
0.70
0.76
0.82
0.88
0.93
0.99
1.05
1.11
1.17
1.23
1.28
1.34
1.40
1.46
Voltage distortion (%) Current distortion (A)

Figure 5.35 Industrial substation (site 7) 7th harmonic histogram

The 95th percentile cumulative probability value for the 5th harmonic voltage at the

Homepride zone substation is 1.54%, well below the recommended planning level of

5% provided by the IEC 61000-3-6 standard. Table 5.2 provides the 95th percentile

values of the Homepride zone substation, residential, commercial and industrial

monitoring locations for the 3rd, 5th, 7th and 19th harmonic voltages and currents over the

initial monitoring period.

Table 5.2 95th percentile values for all sites over initial monitoring period
Harmonic voltage (maximum of each phase)
Site/Description Fund 3rd 5th 7th *19th THD
Homepride zone
1 11097V 0.24% 1.54% 0.47% 0.14% 1.63%
Substation
5 Residential substation 248V 0.22% 1.45% 0.45% 0.15% 1.53%
6 Commercial substation 251V 1.67% 2.89% 0.74% 0.13% 3.28%
7 Industrial substation 249V 0.33% 1.63% 1.02% 0.13% 1.80%
Harmonic current (maximum of each phase)
Site/Description Fund 3rd 5th 7th *19th THD
Homepride zone
1 1205.3A 5.17A 17.06A 15.46A 1.24A 2.96%
Substation
2 Residential feeder 173.48A 0.82A 5.47A 3.32A 0.12A 5.79%
3 Commercial feeder 114.56A 0.71A 3.79A 2.06A 0.14A 6.58%
4 Industrial feeder 156.65A 1.44A 8.22A 2.21A 0.25A 5.74%
5 Residential substation 119.74A 3.96A 2.97A 1.76A 0.22A 11.76%
6 Commercial substation 548.88A 38.18A 48.89A 6.86A 0.94A 26.47%
7 Industrial substation 1404.6A 108.67A 192.98A 28.89A 3.24A 20.12%
th
* The 19 harmonic was monitored at a different period but is included to indicate
relative magnitude
113
The commercial site shows the highest level of voltage distortion at 3.28% with the 5th

harmonic contributing to over 85% of the distortion. The LV sites are expected to have

significantly higher distortion levels due to the additional impedance of the MV/LV

distribution transformer.

The residential site (1.53%) however has a slightly lower distortion level than at the

Homepride zone substation (1.63%), possibly due to the residential loads contributing a

harmonic voltage that cancels the background voltage distortion. It is also noted that the

5th harmonic voltage is lower at the remote end of the residential feeder (1.45%)

compared to that at the zone substation (1.54%). This prompted further investigation.

Using the instantaneous feature of the monitoring instruments, several waveform

captures were carried out at both the Homepride zone substation and at the residential

transformer. A representative sample of these waveforms in the form of a harmonic

spectrum is shown in Figures 5.36 and 5.37.

Homepride zone substation voltage waveform harmonic spectrum

1.4

1.2
Voltage Distortion (%)

1.0 Va
Vc
0.8

0.6

0.4

0.2

0.0
THD 5 9 13 17 21 25 29 33 37 41 45 49
THD and Harmonic Order

Figure 5.36 Spectrum of Homepride zone (site 1) instantaneous voltage waveform

capture

114
 Residential substation voltage waveform harmonic spectrum
1.4

1.2

Voltage Distortion (%)

Va
1.0
Vb
0.8 Vc

0.6

0.4

0.2

0
THD 5 9 13 17 21 25 29 33 37 41 45 49
THD and Harmonic Order

Figure 5.37 Spectrum of residential substation (site 5) instantaneous voltage waveform

capture

Each of the readings from Figures 5.36 and 5.37 confirms the expectations for an

increase in harmonics towards the remote end of the feeder. However, there remains the

possibility the harmonic voltage contributions from the residential site being out of

phase with the zone substation voltage creating a lower net voltage at the residential

site. This raises the question of discrepancies with the overall cumulative probability

and shows that most comparison techniques should not be used in isolation. Each

technique should be used as a tool to add greater understanding to what is known about

the system.

5.8 Harmonic levels compared to standards

As discussed in Chapter 2, Standards Australia has recently adopted the IEC 61000-3-6

harmonic standard as AS/NZS 61000.3.6 to replace AS 2279.2. The harmonic voltage

limits for MV distribution systems recommended by AS 2279.2 and IEC 61000-3-6

were given in Tables 2.4 and 2.5 of Chapter 2 respectively. The introduction of the new

IEC standard has incorporated an increase in the harmonic levels of the most

problematic 5th harmonic from 4% to 5% while the 7th harmonic remains at 4%. For the

115
other harmonic orders the harmonic voltage limits have decreased and fall off with

increasing harmonic order.

As AS 2279.2 does not use statistical techniques to analyse harmonic voltages the

maximum harmonic voltages at each of the monitoring locations should be first found.

These maximum voltage levels for the initial monitoring period are presented in Table

5.3. The benchmark harmonic monitoring programme has found that the ‘head room’

available in relation to the recommended harmonic voltage limits at the MV voltage

levels is quite significant at most locations. The maximum values of harmonic levels

and THD at the various sites normalised using the limits specified in AS 2279.2 and

IEC 61000-3-6 are given in Table 5.4. From this table it is evident that no exceedances

are noted except from the commercial site where the 5th harmonic and THD are close to

the limits based on AS 2279.2.

Table 5.3 Maximum values for all sites over initial monitoring period
Harmonic voltage (maximum of each phase)
Site Description 3rd 5th 7th *19th THD
Homepride zone
1 0.38% 1.96% 0.59% 0.33% 2.03%
Substation
5 Residential substation 0.36% 1.74% 0.56% 0.34% 1.82%
6 Commercial substation 1.92% 3.86% 0.98% 0.30% 4.19%
7 Industrial substation 0.52% 2.14% 1.96% 0.34% 2.36%
* see Table 5.2 footnote

116
 Table 5.4 Harmonic voltage levels normalised to recommended limits
AS 2279.2
Site Description 3rd 5th 7th *19th THD
Homepride zone
1 10% 49% 15% 9% 41%
Substation
5 Residential substation 9% 44% 14% 9% 36%
6 Commercial substation 48% 97% 25% 9% 84%
7 Industrial substation 13% 54% 49% 9% 47%
IEC 61000-3-6
Site Description 3rd 5th 7th *19th THD
Homepride zone
1 6% 31% 12% 12% 25%
Substation
5 Residential substation 6% 29% 11% 13% 24%
6 Commercial substation 42% 58% 19% 11% 50%
7 Industrial substation 8% 33% 26% 11% 28%
* see Table 5.2 footnote

5.9 Annual growth in harmonic distortion levels

Determining the level of harmonic distortion growth over a three year monitoring

period is difficult due to seasonal changes and daily fluctuations of harmonic levels.

The level of harmonics are largely dependent on weather conditions due to the use of

loads such as inverter driven air conditioners and also the length of daylight hours

which in some ways may influence the use of non-linear loads. To determine the growth

it would be preferable to have periodic data over the full three years so that a true trend

could be established. However at the beginning of the program the data from the

harmonic monitors were not periodically downloaded. A less than ideal way of

determining harmonic growth over the monitoring period is to determine the 95th

percentile probability at the same time in each year. As the initial monitoring studies

were completed at the same time each year a very crude comparison of the 95th

percentile harmonic levels can be made. The 95th percentile harmonic voltage values for

the two week period beginning in mid August in each year of the monitoring

programme are given in Table 5.5. As the most problematic harmonic is the 5th

117
harmonic, only the 5th harmonic and the THD will be included for the harmonic growth

study. Also the 3rd and 7th harmonics were not logged over the entire monitoring period.

The corresponding harmonic current growths are presented in Table 5.6.

Table 5.5 Harmonic voltage 95th percentile values for August during monitoring
(1999 − 2001)
Harmonic voltage distortion growth by 95th percentile
Site Location 1999 2000 2001
th
1 Homepride zone substation 5 1.57% 1.54% 1.29%
th
5 Residential substation 5 1.47% 1.45% 1.25%
th
6 Commercial substation 5 2.65% 2.89% 2.40%
th
7 Industrial substation 5 1.80% 1.63% 1.33%
1 Homepride zone substation THD 1.63% 1.58% 1.39%
5 Residential substation THD 1.54% 1.53% 1.34%
6 Commercial substation THD 3.07% 3.28% 2.86%
7 Industrial substation THD 2.00% 1.80% 1.93%

Table 5.6 Harmonic current 95th percentile values for August during monitoring
(1999 − 2001)
Harmonic current distortion growth by 95th percentile
Site Location 1999 2000 2001
th
1 Homepride zone substation 5 18.02A 16.93A 15.62A
th
5 Residential substation 5 2.88A 2.83A 3.56A
th
6 Commercial substation 5 43.36A 50.55A 40.38A
th
7 Industrial substation 5 134.22A 185.51A 168.53A

Tables 5.5 and 5.6 do not indicate clear harmonic growth trends. As mentioned a more

detailed study comprising of periodic downloads is required to establish true harmonic

growth trends. For the final two years of monitoring the data from the monitoring

instruments were downloaded each month to establish the seasonal harmonic trends.

The following section presents the results of that study.

118
5.10 Seasonal changes in harmonic distortion levels
For the years of 2001-2002 the harmonic monitoring instruments at the Homepride

system were frequently downloaded to establish seasonal trending. A similar study [25]

found that in the USA the lowest levels of harmonic voltages existed during the summer

months. The levels from the study [25] varied from 1.3% to about 1.8% throughout the

year. A similar variation was found on the Homepride system as shown in Figure 5.38.

A significant difference however was that the highest harmonic voltage levels occurred

during the summer months of 2001 and then during autumn of 2002.

th
2001/2002 95 Percentile seasonal voltage trend
2.50

2.25
2.00
1.75 THD-2001
Distortion (%)

3rd-2001
1.50
5th-2001
1.25 7th-2001
THD-2002
1.00
3rd-2002
0.75 5th-2002
0.50 7th-2002

0.25
0.00
J F M A M J J A S O N D
Month

Figure 5.38 Seasonal 95th percentile voltage trend for Homepride zone (site 1)

To establish harmonic growth trends it would be necessary to repeat the seasonal

harmonic voltage trend study over a period of a number of years. The study presented in

[25] found that while the seasonal trend continued each year it was found that the THD

levels increased by 10% each year. Although an initial large increase in distortion is

observed during the first half of 2002 further work needs to be completed to establish

the true growth trends at the Homepride zone distribution system. Integral Energy also

119
reported that a large industrial customer had been disconnected from Homepride zone

substation in the second half of 2002 which may have contributed to the fall in

harmonic levels illustrated in Figure 5.38.

5.11 Residential, commercial and industrial load harmonic

characteristics
From the above monitoring programme results it is seen that the amount of total

harmonic voltage distortion is dominated by the 5th harmonic. This characteristic may

be used to predict the level of total harmonic distortion existing in the system by

applying a suitable multiplying factor to the 5th harmonic, if the 5th harmonic distortion

level can be established.

Table 5.7 summarises the results from the monitoring programme using the 95th

cumulative probability values for the fundamental and 5th harmonic current for the three

LV locations, averaged over the entire monitoring period. The subscript ‘1CP95’

indicates the fundamental component 95th percentile cumulative probability level. It is

required to characterise the load types so that the 95th percentile cumulative probability

levels of harmonic distortion can be predicted from known network impedances, load

magnitudes and compositions of load types. Assuming that similar load compositions at

any power level will have the same spectrum, each load type can be characterised

conveniently by the I5CP95/I1CP95 as shown in the right hand side column. This shows that

the residential load has much less distortion than the commercial and industrial types as

would be expected. Over the course of the entire monitoring programme it was found

that the ratio of 95th percentile 5th harmonic to 95th percentile fundamental varied by less

than 10% for any given week of data.

120
 Table 5.7 Fundamental and 5th harmonic current 95th percentile values (A)
Harmonic Monitor Site I1CP95 I5CP95 I5CP95/I1CP95
Site 5 – Residential 133 3.36 0.025
Site 6 – Commercial 617 43.3 0.071
Site 7 – Industrial 1469 170.9 0.116

Assuming the data in Table 5.7 is typical to most distribution systems, the above

residential, commercial and industrial load models may be used to estimate harmonic

distortion levels for the 5th harmonic and thus the THD during the planning phase of a

distribution system. The method to estimate harmonic distortion levels is outlined in

Chapter 3. An area for future work is to verify that the data in Table 5.7 is typical to

other distribution systems.

5.12 Relationship of harmonics on individual phases

If two sets of data across all three phases correlate very well then the likely load type is

three-phase as variations in the load affect all phases similarly. This method of

comparison enables the phase-to-phase comparison of data as shown in Figures 5.39 to

5.42. At the 5th harmonic it can be seen that good correlation exists between A, B and C

phases for both voltage and current for most of the monitoring sites. This means as A

phase varies, B and C phase vary in similar proportions. The harmonic current at the

residential site is the exception (Figure 5.40), although the spread of data points is

expected as the residential substation would consist almost entirely of single-phase

loads. Scatter graphs for the 3rd, 7th and 19th harmonics can be found in Appendix C. At

the 3rd and 19th harmonics the phase to phase correlation is less obvious and the large

spread of data points indicates very little dependency between A, B and C phase at all

the monitoring sites.

121
 th th
Homepride 5 harmonic A versus C voltage Homepride 5 harmonic A versus C current
2.5 4.0

2.0
C phase distortion (%)

C phase distortion (%)

3.0

1.5

2.0
1.0

0.5 1.0

0.0
0.0
0.0 0.5 1.0 1.5 2.0 2.5
0.0 1.0 2.0 3.0 4.0
A phase distortion (%) A phase distortion (%)

Figure 5.39 Homepride (site 1) 5th harmonic A and C phase voltage and current

Residential 5th harmonic A versus C voltage Residential 5th harmonic A versus C current

2.0 7.0

6.0
C phase distortion (%)
C phase distortion (%)

1.5
5.0

4.0
1.0
3.0

2.0
0.5

1.0

0.0 0.0
0.0 0.5 1.0 1.5 2.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
A phase distortion (%) A phase distortion (%)

Figure 5.40 Residential (site 5) 5th harmonic A and C phase voltage and current

th
Commercial 5th harmonic A versus C voltage Commercial 5 harmonic A versus C current

2.5 25.0
C phase distortion (%)
C phase distortion (%)

2.0 20.0

1.5 15.0

1.0 10.0

0.5 5.0

0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 0.0 5.0 10.0 15.0 20.0 25.0 30.0

A phase distortion (%) A phase distortion (%)

Figure 5.41 Commercial (site 6) 5th harmonic A and C phase voltage and current

122
 th
th
Industrial 5 harmonic A versus C voltage Industrial 5 harmonic A versus C current
2.5 20.0

2.0

C phase distortion (%)

15.0
C phase distortion (%)

1.5
10.0

1.0

5.0
0.5

0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 0.0 5.0 10.0 15.0 20.0
A phase distortion (%) A phase distortion (%)

Figure 5.42 Industrial (site 7) 5th harmonic A and C phase voltage and current

A useful method for interpreting scatter graphs is the correlation coefficient ρ. This

coefficient varies in the range -1≤ρ≤1, where a value of 1 (or –1) implies perfect

correlation and zero indicating no correlation. This may give some additional insight as

to whether the load is predominantly single-phase or three-phase type.

The scatter graphs of Figure 5.41 of the commercial substation demonstrate good

correlation between the phases. This is a result of the load at the commercial sites being

mostly three-phase air conditioning and lighting balanced over all three phases. This

balanced load current results in symmetrical voltage correlation.

The correlation coefficients of Table 5.8 support the graphical representations of

Figures 5.39 to 5.42 with the 5th and 7th harmonics showing good correlation for voltage

and current at all sites other than for the residential substation for 3rd and 19th harmonics

where very little dependency can be found. However the 95th cumulative probabilities

for the 3rd and 19th harmonics on each of the phases were found to be similar across the

three phases (where available) for the seven monitoring locations. The figure calculated

for the correlation coefficient would suggest that it would be required to monitor all

123
three phases, however use of the cumulative probability method on one phase would be

sufficient to give the required information on harmonic levels. The 95th cumulative

probability also appears to be less sensitive to fluctuations at low levels of distortion.

Table 5.8 Correlation coefficients for initial period of monitoring programme

Harmonic 3rd 5th 7th 19th
MV Harmonic voltage correlation coefficients
Homepride zone
A vs. C 0.322 0.988 0.988 -0.005
substation
Residential feeder A vs. C 0.400 0.988 0.981 0.788
Commercial feeder A vs. C 0.378 0.983 0.980 0.049
Industrial feeder A vs. C 0.379 0.983 0.982 0.004
MV Harmonic current correlation coefficients
Homepride zone
A vs. C 0.772 0.829 0.993 0.169
substation
Residential feeder A vs. C 0.639 0.990 0.995 0.107
Commercial feeder A vs. C 0.692 0.916 0.965 0.235
Industrial feeder A vs. C 0.005 0.962 0.913 0.188
LV Harmonic voltage correlation coefficients
A vs. B -0.519 0.980 0.963 0.899
Residential
A vs. C -0.018 0.981 0.973 0.679
A vs. B 0.985 0.986 0.985 0.017
Commercial
A vs. C 0.986 0.974 0.970 0.013
A vs. B -0.277 0.975 0.950 0.144
Industrial
A vs. C -0.777 0.983 0.945 0.251
LV Harmonic current correlation coefficients
A vs. B 0.368 0.662 0.584 0.158
Residential
A vs. C -0.364 0.457 0.471 0.304
A vs. B 0.870 0.956 0.866 0.947
Commercial
A vs. C 0.975 0.991 0.841 0.969
A vs. B 0.282 0.914 0.660 0.359
Industrial
A vs. C 0.655 0.873 0.441 0.325

5.13 Minimum requirements of a harmonic monitoring programme

The information from the monitoring program was recorded over a substantial period of

time that allows the minimum monitoring requirements to be estimated. The minimum

monitoring period can be found by breaking the monitoring period into smaller separate

weekly records as shown in Table 5.9 using the Homepride zone MV readings over the

first month of recording. It can be seen that the maximum deviation from the overall
124
average reading of 1.54% is 0.14%, which is within acceptable accuracy of 0.2%. Thus

monitoring over a period as short as one week gives useful information without being

biased too much by short-term trends. Note that there is an additional long-term

seasonal trend that needs to be considered. Figure 5.38 shows a seasonal variation of

approximately 0.6% in 2001. Table 5.9 again reinforces the balanced nature of the

system with different phases sharing similar trends.

Table 5.9 Weekly 95th percentile values for Homepride zone substation 5th harmonic
voltage
5th harmonic voltage Vab (% of fundamental)
(Site 1) Week 1 Week 2 Week 3 Week 4
Vab 1.48% 1.68% 1.61% 1.45%
Vcb 1.46% 1.66% 1.56% 1.45%

The weekly readings have been further broken down into daily 95th percentile

cumulative probability values in Table 5.10 for which the following can be noted:

(i) The weekend usually has high harmonic levels although the first Sunday has one

of the lowest levels recorded,

(ii) Friday always indicates the lowest level,

(iii) No one day is typical of the week as a whole, and

(iv) The smallest subset of days giving useful results is Sunday-Tuesday, which has an

average differing from the weekly value by about 0.6%.

Table 5.10 Daily 95th percentile values for Homepride zone substation 5th harmonic
5th harmonic voltage Vab (% of fundamental)
(Site 1) Sat Sun Mon Tue Wed Thu Fri Total
Week 1 1.44 1.28 1.35 1.45 1.56 1.53 1.47 1.48
Week 2 1.73 1.90 1.55 1.53 1.32 1.26 1.24 1.68
Week 3 1.76 1.75 1.50 1.54 1.35 1.44 1.37 1.61
Week 4 1.58 1.50 1.34 1.41 1.52 1.43 1.34 1.45

125
Maximum harmonic levels occurred on weekdays (Monday-Friday) at 4-5pm and

7-9pm, different to what is reported for US levels in [18]. However, significant 95th

percentile harmonic levels on Sunday are consistent with [18] in that higher levels also

occur away from the peak loading periods. This may reflect harmonic loads such as

personal computers and lighting being on for a relatively long time of the day and the

effect of other loads absorbing harmonics at peak load periods.

5.14 Summary
Inconsistencies have been found with some types of harmonic monitoring instruments.

The indifferent performance of the selected instruments suggests that testing prior to

field application is necessary before undertaking harmonic measurements.

A harmonic monitoring programme has been designed and implemented on a typical

MV distribution system to establish harmonic voltage and current characteristics from

the residential, commercial, and industrial load sectors. The monitoring program has

been carried out using simultaneous measurement of harmonic voltages and currents at

several sites within an MV distribution system. This coordinated approach in obtaining

results has overcome some of the problems with synchronising and reporting data

outlined in [25, 28].

Statistical harmonic models of residential, commercial and industrial substation loads

have been developed to simulate the global behaviour of distorting loads at distribution

substations. The load models represent aggregates of loads and are specifically intended

for calculation of harmonic emissions for comparison with the relevant standards. Both

time and phase diversities are included in the representative load models. The

126
monitoring programme results have also been used to confirm the relative accuracy of

distortion level prediction techniques introduced in Chapter 3.

The monitoring programme has established seasonal trends of harmonics with voltage

distortion levels being significantly higher in both summer and autumn months.

Benchmarking of harmonic voltage distortion levels against the relevant standards has

also been completed with results suggesting that the amount of headroom remaining

before recommended limits are reached is sufficient to allow for some further growth in

emission levels from customer loads for the monitored system.

Results from the monitoring programme indicate a small amount of growth in harmonic

voltage levels for the study system over the three-year period. However, the amount of

growth in harmonic voltage levels is considerably less than the seasonal variations and

thus no strong conclusions can be made regarding future variations. Continual

monitoring to establish firm growth patterns is an area for further work.

The harmonic current trends reflected voltage trends as expected for each of the

substations monitored. The largest voltage distortion was produced by the commercial

substation, reaching daytime peaks of 3-4% 5th harmonic and a few peaks of 5%

through the entire three year monitoring programme.

A study of the variation of harmonics over the three phases throughout the period of a

week suggest that monitoring of a single phase in most cases will provide a sufficient

indication of harmonic levels on all three phases. Analysis of daily harmonic trends has

shown that no single day provides sufficient indication of harmonic levels for the entire

127
week. As a minimum requirement harmonic monitoring should be completed over a

whole week.

128
 Chapter 6

Effects of system capacitance on harmonic levels

6.1 Introduction
The work presented in the previous chapters has shown harmonic impedance to be a

crucial factor in the calculations required for allocation of acceptable harmonic

emissions and the prediction of harmonic voltage levels. In this chapter the effects of

system capacitance are considered in the modelling of a distribution system and

determination of harmonic impedance. Typical sources of significant system

capacitance in distribution systems include power factor correction (PFC) capacitors

and distribution line shunt capacitance on long feeders.

The introduction of significant system capacitance can give rise to harmonic resonances

at frequencies within the harmonic range covered by IEC 61000-3-6 [15]. Harmonic

resonances can lead to a significant increase in the harmonic impedance and

subsequently high harmonic voltage if resonance occurs at a problematic frequency.

Thus system capacitance has to be considered when calculating the harmonic

impedance at any given point within the network. Calculations may also have to

consider a range of network configurations and loading conditions to ensure that the

worst-case harmonic resonances are identified.

Although a documented example for the calculation of harmonic impedance at the point

of common connection (PCC) for a customer connected to a system with a PFC

capacitor has been provided in IEC 61000-3-6, the example considers only the effects of

a single capacitor. The detailed modelling difficulties due to loading variations are

discussed but not addressed in the standard. A method to estimate harmonic impedance
129
when multiple PFC capacitor installations exist is also proposed in the standard but

there are no clear guidelines on how to apply the method.

The aim of this chapter is to investigate the level of complexity required in modelling

system capacitance that will provide acceptable results in the calculation of harmonic

impedance in order to minimise computational effort. The work begins by first

investigating detailed modelling techniques for distribution systems. Initially the models

will be used to determine the lengths of distribution feeders at which significant

harmonic resonances occur due to the capacitance of cables and lines. The complexity

in incorporating damping effects of customer loads and multiple capacitor installations

will also be explored. The applicability of a pragmatic approach to the determination of

harmonic impedance introduced in the standard will be the concluding investigation in

this chapter.

6.2 Modelling of the system

For the purpose of pragmatic modelling, previous chapters have modelled transformers

and distribution lines and cables as single-phase series resistances and inductances with

sufficient accuracy to enable estimations of the harmonic voltage contributions from

customers to be established. Pragmatic techniques such as these are required to allow

distribution planning to be carried out when detailed knowledge of the system is not

available. To include system capacitance when modelling a distribution system further

detailed models of system components have to be used to ensure that an accurate

estimation of impedance and resonant frequencies are obtained.

130
6.2.1 Line and cable impedance
For pragmatic calculations at frequencies above the fundamental the series resistance of

a distribution line may be neglected due to the dramatic increase in reactance with

frequency. For such studies the fundamental impedance is assumed to be predominantly

inductive and multiplying the fundamental reactance by the harmonic order h as shown

in (6.1) approximates the harmonic impedance.

Zh ≈ h × X1 (6.1)

It is suggested by [65] that (6.1) will provide reasonable accuracy (better than 20%) for

MV systems where the reactance of the HV/MV transformer is high compared to the

high voltage supply system impedance. This type of pragmatic modelling is assumed to

be acceptable when

(i) Higher order harmonics are not of primary concern,

(ii) Distribution system feeder lengths are short and mostly overhead, and

(iii) No PFC capacitors are present.

For overhead distribution systems with short feeders resonance due to distribution line

shunt capacitance will typically appear at relatively high frequencies and thus is usually

of little concern for system harmonic planning. In special circumstances where high

frequencies are prominent in customer equipment emissions, for example with high

pulse number converters, these higher frequencies will need to be considered. For

underground feeders resonant frequencies may appear at much lower frequencies and

therefore distribution feeder shunt capacitance will need to be included in the

calculation of system impedance when allocating harmonic emissions to distribution

system customers.

131
To establish appropriate lengths of feeders beyond which shunt capacitance will cause

resonance at problematic frequencies in relation to the nominal PI model shown in

Figure 6.1 [66]. In Figure 6.1 the conductance G, susceptance B, and reactance X

combine to give the admittance GL + jBL = 1/(RL + jXL) with YC = 1/jXC, XL=ωL, and

XC=1/ωC at frequency ω (rads-1).

GL + jBL

YC YC
2 2

Figure 6.1 Admittance model of transmission line [66]

It has been shown that the nominal PI model can be improved in accuracy by dividing

the distribution line into a number of subsections of nominal PI models [67].

Additionally, in preference to subsections the model can be improved using the

evaluated distributed parameters equations (6.2) to (6.5) with series impedance, Z', and

shunt admittance, Y' [65]. This new model is usually entitled the equivalent PI model.

Z = R + jXL (6.2)

Y = jωC (6.3)

Z sinh YZ
Z' = (6.4)
YZ

YZ
tanh
Y' 2
=Y (6.5)
2 YZ

132
The increase in resistance and the reduction in inductance of conductors due to skin

effect may also be considered when modelling distribution lines and cables. There are a

number of analytical methods for estimating the magnitude of skin effect [65, 66, 68]

that typically involve complex calculations based on conductor construction and Bessel

functions. In each case the resistive component is most affected with only a small

decrease in the inductance.

For transmission lines [68] suggests the harmonic impedance incorporating skin effect

may be represented by equation (6.6) where h is the harmonic order, R1 is the line series

resistance at 50Hz and Rh is the line series resistance at harmonic order h.

Rh = √ h × R1 (6.6)

However, equation (6.6) does not show close agreement with the values suggested by

equation (6.7) from [65] or the methods from [66], suggesting that (6.6) may

overestimate the resistive component, especially for lower order harmonics.

 0.646 h2 
Rh = R1 1 + 192 + 0.518 h2 (6.7)
 

According to equation (6.7) the increase in resistance due to skin effect at the 5th

harmonic will be approximately 8%. For harmonics near the 20th the increases will be as

much as 60%. However, due to the inductive component of the series impedance

increasing linearly with frequency the increase in the overall impedance will typically

represent less than 3% due to skin effect. Skin effect has been thus neglected in the

present study as the specific application is designed for lower order harmonic

frequencies and MV/LV distribution systems rather than EHV/HV transmission systems

where skin effect is more prolific.

133
6.2.2 Transformers
For harmonic studies transformers are modelled using a combination of series and

parallel inductive and resistive components. Generally, harmonic modelling of

transformers can be completed with sufficient accuracy using the model shown in

Figure 6.2 [65].

Rp

Rs
jXh

Figure 6.2 Transformer harmonic model, where Xh = hX1

As transformers have a high reactance to resistance ratio, typically in the range of

10-20, for pragmatic modelling the resistances in the above model can be ignored and

the transformer is modelled as a single series reactance. For the detailed modelling

undertaken in this chapter full model shown in Figure 6.2 will be used.

6.2.3 Load models

For harmonic analysis load models need to incorporate both simple and complex loads.

Simple passive loads such as lighting and heating may be represented as equivalent

resistances. For pragmatic harmonic studies induction motors can also be represented

using passive components, typically series inductance and resistance. Complex loads,

such as an thyristor controlled AC loads, whose harmonic impedance will vary

according to the phase angle of the thyristor firing [69] require detailed modelling and

go beyond the scope of this thesis.

134
As planning engineers will often be more concerned with aggregates of load rather than

individual load types [65] recommends that distribution loads be modelled using the

model shown in Figure 6.3. In relation to the figure, V is the nominal voltage of the

network, P1 is the minimum active power of the load, and φ1 is the power factor angle.

Development of suitable aggregate load models is an area for further investigation.

where
XS
R = V2 / P1
XP
XS = 0.073⋅h⋅R

R h⋅R
XP =
6.7tanφ1 - 0.74

Figure 6.3 CIGRE model for normal aggregate distribution loads [65]

6.2.4 Power factor correction capacitors

PFC capacitors are installed within distribution systems to increase the utilisation of the

system at the fundamental frequency. PFC capacitors are modelled as a simple shunt

capacitor. A series resistance may be included in the model to improve accuracy around

resonant frequencies [16].

6.3 Critical lengths of feeders

As the magnitude of the distribution line shunt capacitance is relative to the length of

the line, for harmonic planning purposes it is important to know the critical lengths of

feeders at which distribution line shunt capacitance needs to be considered.

Harmonic resonances due to the cable or line impedance are due to the interaction of the

shunt capacitive reactance and the series inductive reactance. For example a system

135
model as illustrated in Figure 6.4 is considered. The system consists of a transformer

connected to a purely inductive upstream system, creating an assumed fault level of

100MVA at an 11kV busbar. The busbar is connected to a single feeder of variable

length.

11kV busbar
ZPCC
ZTX Z’

ZSYS Y' Y'

2 2

Figure 6.4 Simple system to estimate critical lengths of feeders

Using typical overhead line parameter values of R=0.3Ω/km, XL=0.3Ω/km and

XC=300kΩ.km for the equivalent PI model the harmonics at which resonance occurs

due to the series impedance and shunt admittance for a given length of feeder was

investigated with the results illustrated in Figure 6.5. The same study has also been

completed for an underground cable using R=0.12Ω/km, XL=0.06Ω/km and

XC=3kΩ.km as illustrated in Figure 6.6.

136
 50

45

40

35

Feeder length (km) 30

25

20

15

10 Equ ivalent π
Nomina l π (3 se ctio ns)
5 Nomina l π (2 se ctio ns)
Nomina l π
0
26 28 30 32 34 36 38 40
Harmo nic num ber (h)

Figure 6.5 Harmonic at which first resonance occurs for overhead line lengths

50

45

40

35
Feeder length (km)

30

25

20

15

10 Equ ivalent π
Nomina l π (3 se ctio ns)
5 Nomina l π (2 se ctio ns)
Nomina l π
0
5 10 15 20 25 30 35 40
Harmo nic num ber (h)

Figure 6.6 Harmonic at which first resonance occurs for underground cable lengths

A feeder length of up to 50km was considered although realistically lengths greater than

about 15km would be considered rare at MV distribution level. Figure 6.5 illustrates

that for overhead lines there is no significant contribution from the feeder shunt

capacitance and thus shunt capacitance may be neglect for pragmatic harmonic studies

involved with allocating harmonic emissions.

137
Figure 6.6 suggests that feeder lengths of underground cables greater than

approximately 7km need to have the shunt capacitance included in the modelling as

resonance may occur at frequencies between the 30th and 40th harmonics. Experience

has shown that harmonic planning is really only achievable at distribution level for

lower order harmonics [70]. The resonance at the higher order harmonics has not shown

to be of significant importance in most cases. Recommendations in [45] are that feeder

shunt capacitance will have a negligible effect on the harmonic impedance if the h×l

product is less than 300 for overhead line or 50 for underground cable, where line h is

the harmonic of concern and l is the length of the feeder in kilometres. Figures 6.5 and

6.6 suggest that this is conservative for typical Australian distribution systems. Thus

distribution line shunt capacitance may be neglected for most harmonic studies.

6.4 Detailed approach to determining system impedance

The installation of PFC capacitors can dramatically change the distribution system

harmonic impedance. Parallel or series resonance problems can occur at problematic

harmonics such as the 3rd, 5th and 7th harmonics if PFC capacitor installation design is

not completed carefully. The resonant frequency produced from the installation of PFC

capacitors can be approximated by equation (6.8).

1
fr = (6.8)
2π LC

where fr is the resonant frequency, L is the system inductance, and C is the capacitance

of the PFC capacitor. The resonant harmonic order hr can also be estimated using QC,

the capacitor reactive volt-amperes and SSC, the fault level at the point of capacitor

connection as given by equation (6.9).

S SC
hr = (6.9)
QC

138
6.4.1 Location and sizing of PFC capacitors
Depending on the power factor requirements and the policy of a particular DNSP, PFC

capacitors may be installed locally to the MV busbar of zone substations, at more

remote locations further downstream on distribution feeders, and within customer

installations. The fundamental frequency performance of the system is the primary

consideration when designing PFC capacitor installations. DNSP planning engineers

must also bear in mind the harmonic performance of the distribution system to reduce

the possibility of a dramatic increase in levels of harmonic voltages due to system

resonance.

6.4.2 Typical resonant frequencies for different systems

As distribution systems are largely designed to standard practises and procedures of the

DNSP, system impedances at various levels of the distribution system will often be

similar. PFC capacitors also usually come in banks of standard sizes, i.e. 2.5MVAr in

distribution substations and 300kVAr for pole mounted banks, although this will vary

depending on the DNSP. As pole mounted capacitor banks may be installed at any

location along a feeder it is highly unlikely that any typical resonant frequencies would

exist. For zone substations capacitor bank sizes will generally be around 10%-20% of

substation rating to correct power factor back to near unity.

From the data obtained from Integral Energy planning engineers, fault levels at

distribution zone substations vary typically from 150MVA to 250MVA. Given that a

typical zone substation will be rated at approximately 25MVA, capacitor banks totalling

5MVA will be usually expected to be installed. Using equation (6.9) the resulting

harmonic resonance will be between the 5th and 10th harmonic, implying that most

139
substation capacitor installations may lead to resonances around the problematic 5th and

7th harmonics. Thus, great care needs to be taken to ensure resonances do not coincide

with these harmonics, especially considering switched capacitors can often lead to

multiple resonant frequencies. Detuning reactors may be necessary in some instances to

shift resonant frequencies away from the problematic 5th and 7th harmonics.

6.4.3 De-tuning of PFC capacitors

De-tuning of PFC capacitors is achieved by connecting a reactor in series with the

capacitor. This effectively shifts the resonant frequency created by the capacitor

installation to a less problematic harmonic or sub-harmonic. Figure 6.7 illustrates the

effect of including a de-tuning reactor on the harmonic impedance for the simple model

of a distribution system. The circuit without the de-tuning reactor has a parallel

harmonic resonance at the 7th harmonic. Once the correct de-tuning inductor is installed,

both series and parallel resonance frequencies are shifted away from the 7th harmonic.

Particularly, the series resonance is now located at 4.5th harmonic and hence the

resulting impedances at the problematic 5th and 7th harmonics are both reduced.

140
 System impedance

R jXL Detuning reactor

jXD (tuned to 4.5th harmonic)

-jXC PFC capacitor

5
Impedance with no de-tuning reactor
4

3 Impedance with no PFC capacitor

2
Impedance with de-tuning reactor
1

0
0 5 10 15 20

Harmonic, h
Figure 6.7 Simple distribution model illustrating de-tuning of PFC capacitor

PFC capacitors will generally have small in-rush reactors installed for protection and

thus the addition of a slightly larger reactor will often not increase the cost by

significant levels. The increase in cost for a detuned capacitor bank will be

approximately 30% more than without detuning. This figure applies to smaller capacitor

banks (less than 5MVA) as the proportion of cost increases with larger banks, i.e. for

capacitors installed within transmission and sub-transmission systems. Further details of

planning issues regarding PFC capacitors and the application of de-tuning reactors to

reduce harmonic voltages are discussed in [71].

6.4.4 Effects of load damping

Slight variations in system harmonic impedance are expected as a network is

reconfigured and different loads are connected or disconnected. For a distribution

system with insignificant system capacitance the variations in the harmonic impedance

141
will usually not be extensive enough to cause intolerable errors in the calculations of

harmonic emissions or estimations of harmonic voltage levels. However, if significant

system capacitance is present the variations in network impedance at certain frequencies

may be considerable due to the effects of harmonic resonance.

Passive loads have a damping effect on the magnitude of harmonic resonance. To study

the extent of damping the load model illustrated in Figure 6.3 is utilised in a harmonic

impedance study of an overhead distribution system illustrated in Figure 6.8. The

system consists of six feeders fed via two 25MVA transformers, each feeder containing

six 500kVA customers with a power factor of 0.85 lagging. The distance between two

adjacent customers is 2km. The system models outlined in Section 6.2 are used to

complete the study. The overhead line parameters are R=0.3Ω/km, XL=0.3Ω/km and

XC=300kΩ.km. An X:R ratio of 20:1 is used for the system impedance on the 33kV

busbar.

33kV transmission Zone substation 11kV distribution

SSC=1000MVA 0 1 2 3 4 5

500kVA 500kVA 500kVA 500kVA 500kVA

33/11kV
Transformer
5 identical
25MVA
Xm=10% feeders

Figure 6.8 Homogeneous distribution system for load damping study

The harmonic impedances seen by customers (at nodes 0-5) in Figure 6.8 are illustrated

in Figure 6.9 for both full load damping and no load damping considered. To correct the

power factor two 5MVAr capacitor banks are installed at the zone substation 11kV

142
busbar resulting in a harmonic resonance at the 5.8th harmonic. The resulting harmonic

impedances seen by each customer are illustrated in Figure 6.10.

No PFC capacitor, no load damping No PFC capacitor, full load damping
40 40
Cust 0 Cust 0
35 Cust 1 35 Cust 1
Cust 2 Cust 2
30 Cust 3 30 Cust 3
Cust 4 Cust 4
impedance, pu

impedance, pu
25 Cust 5 25 Cust 5
20 20

15 15

10 10

5 5

0 0
0 5 10 15 20 0 5 10 15 20
harmonic, h harmonic, h

Figure 6.9 Harmonic impedance with no PFC capacitors installed

PFC capacitor at zone substation, no load damping PFC capacitor at zone substation, full load damping
60 60
Cust 0 Cust 0
Cust 1 Cust 1
50 Cust 2 50 Cust 2
Cust 3 Cust 3
40 Cust 4 40 Cust 4
impedance, pu

impedance, pu

Cust 5 Cust 5
30 30

20 20

10 10

0 0
0 5 10 15 20 0 5 10 15 20
harmonic, h harmonic, h

Figure 6.10 Harmonic impedance with PFC capacitor installed at zone substation

As an alternative to installing a capacitor at the zone substation busbar, each feeder can

be installed with a capacitor at a single location along the length of the feeder. The size

of the capacitor is chosen to be the optimum size and location as per recommendations

in [72]. The capacitors are rated at 2/3 of the total reactive volt-amperes of each feeder

and are installed at 2/3 the total length from the zone substation. The resulting harmonic

impedances are illustrated in Figure 6.11.

143
 PFC capacitor on feeder, no load damping PFC capacitor on feeder, full load damping
40 40
Cust 0 Cust 0
35 Cust 1 35 Cust 1
Cust 2 Cust 2
30 Cust 3 30 Cust 3
Cust 4 Cust 4
impedance, pu

impedance, pu
25 Cust 5 25 Cust 5
20 20

15 15

10 10

5 5

0 0
0 5 10 15 20 0 5 10 15 20
harmonic, h harmonic, h

Figure 6.11 Harmonic impedance with PFC capacitor installed at 2/3 of feeder length

From Figures 6.9 to 6.11 it can be seen that the effects of load damping on the harmonic

impedance are significant. For the cases where a PFC is present the harmonic

impedances at the resonant frequencies have reduced considerably while the resonant

frequencies themselves have changed only slightly. Damping at only full load is

considered here. Damping due to lighter loads produces less significant reductions in

resonant frequencies for the same sized PFC capacitor. It is noted that capacitor banks

would generally have a fixed and switched component, thus at light loads less

capacitance would be in service leading to a different resonant frequency (at a higher

order harmonic).

6.4.5 Complexity with a number of capacitors present

In Figure 6.11 two harmonic resonance frequencies can be identified. The two

resonances frequencies are due to the capacitor on the study feeder and a combination of

the capacitors on the remaining feeders. The inclusion of multiple capacitor installation

will often create multiple harmonic resonances. Also, disconnection of one or more

feeders or one of the zone substation transformers will dramatically shift the resonant

frequencies that exist. When allocating harmonic emissions to a customer, knowledge of

144
all possible distribution system configurations should be considered such that the worst

case scenario can be identified.

Multiple resonant frequencies and network configurations also lead to additional

difficulties in attempting to de-tune PFC capacitors. Detailed analysis should always be

completed in these cases to ensure problematic resonance do not occur.

6.5 Pragmatic approach in determining system impedance

As detailed modelling of the system may not be practical during the planning stage of a

distribution system, it is necessary to develop a pragmatic approach that may be used to

model the system effectively in the presence of significant system capacitance. While

being straightforward to implement a pragmatic approach should also provide results

within suitable levels of accuracy.

Detailed approaches to system modelling are well documented in [11, 13, 15, 65] for the

case of a single PFC capacitor. The effects of system resonance have been studied

extensively but the work described has always involved complex and detailed

calculations [73, 74]. Dedicated software packages such as PSCAD/EMTDC [55],

Superharm [58], or PSpice [56] may be used for detailed analysis of multiple capacitor

installations to calculated harmonic impedances. Also mathematical packages such as

MATLAB® may also be used. However, all of these methods are very time consuming.

IEC 61000-3-6 suggests a simplified method to determine the harmonic impedance in

the presence significant capacitance by calculating an envelope impedance curve. The

145
impedance curve is illustrated in Figure 6.12 with Zh=2hX1 up to the 8th harmonic and

Zh=hX1 beyond the 8th harmonic.

Impedance, pu

0 2 4 6 8 10 12 14 16 18
Harmonic, h

Figure 6.12 11kV worst case impedance curve from [16]

This method has been refined in [45], however still usually leads to an “over-

assessment” of the harmonic impedance. For comparison the damped harmonic

impedance of Figure 6.11 is compared with a curve equal to two times the impedance

with no PFC capacitors present (dashed lines overlayed) as shown in Figure 6.13.

PFC capacitor on feeder, full load damping

30
Cust 0
Cust 1
25 Cust 2
Cust 3
20 Cust 4
impedance, pu

Cust 5
15

10

0
0 5 10 15 20
harmonic, h
Figure 6.13 Harmonic impedance with full load damping and two times impedance

146
As shown in Figure 6.13 the calculated harmonic impedance only protrudes beyond the

two times impedance curve at the resonant frequency. This suggests that the envelope

impedance curve is a suitable tool for estimating the harmonic impedance. Trials on

variations of the example illustrated in Figure 6.8 also yielded similar results.

Further verification is required to ensure that the envelope impedance curve provides

acceptable results for all possible system configurations, i.e. multiple capacitor

installations and parameter variations. This is an area to be addressed in future work.

6.6 Allocation of harmonic emissions with capacitors present

For the allocation of harmonic emissions to a customer located electrically near a PFC

capacitor installation it will be assumed that the envelope impedance curve

recommended in IEC 61000-3-6 will provide a suitable pessimistic approach in

determining the harmonic impedance. In the event of an allocation derived from this

method being unacceptable to a customer a detailed study will be required whereby the

harmonic models for various distribution system components introduced in this Chapter

are utilised.

Detuning of PFC capacitors using series inductors can help reduce the parallel

resonance effect by shifting the resonant frequency to a non-problematic harmonic.

Thus DNSP planning engineers should ensure that PFC capacitor installation take into

account the harmonic performance of a distribution system. In this manner emission

allocation for problematic harmonics need only be completed using the assumption that

there is no significant system capacitance.

147
6.7 Summary
In this Chapter consideration has been given to the effects of PFC capacitors,

distribution line and cable shunt capacitance on the harmonic performance of MV

distribution systems. The effects on harmonic impedance are important when estimating

harmonic voltage levels and during the procedure of allocating harmonic emissions to

distribution system customers. Various models for distribution system components have

been introduced for use in detailed harmonic studies.

Using detailed analysis it has been determined that cable and line lengths at distribution

level will be usually short enough for resonance due to distribution line shunt

capacitance not to occur at the problematic lower order harmonics. Thus distribution

line shunt capacitance may be assumed to be negligible for pragmatic harmonic studies.

Installation of PFC capacitors within distribution system is often required to improve

the power factor, improve voltage stability and reduce losses. While these performance

parameters are based on the fundamental frequency, it is important for DNSP planning

engineers to also consider the harmonic performance. This is due to harmonic resonant

frequencies typically created by PFC capacitors installed within zone substations and on

feeders often coinciding with problematic lower order harmonic frequencies. By de-

tuning PFC capacitors correctly resonance at problematic harmonics can be reduced.

This has a drawback practically in that it imposes high cost in design and installation.

For detailed harmonic analysis damping at resonant frequencies due to connected loads

has shown to be significant. Thus if system capacitance is being modelled consideration

to the effects of damping from loads should also be given. Modelling of loads is very

difficult at harmonic frequencies where further research needs to be undertaken. The

148
CIGRE model used here is considered reasonable but possibly not suitable for all

systems without further research.

Provided resonance frequencies are not at the problematic harmonics, the envelope

impedance curve method described in IEC 61000-3-6 to estimate harmonic impedances

should provide suitably pessimistic results. If resonant problems are anticipated more in

depth harmonic studies should be completed to determine harmonic problem areas.

Areas for future work in determining harmonic impedance in the presence of significant

system capacitance include development of a less pessimistic yet straightforward

pragmatic method to estimate harmonic impedance when PFC capacitors are electrically

near a customer’s installation.

149
 Chapter 7

Application of harmonic planning methods to MV

distribution systems

7.1 Introduction
A general method for allocating acceptable harmonic emissions to MV customers has

been established in Chapter 4 based on the guiding principles outlined in the

IEC 61000-3-6 standard. The new method allows the same ideology of allocating

harmonic emissions to be applied to general systems with complexity beyond the simple

examples presented in the standard. Harmonic emission allocations to three real case

studies are provided in this chapter to illustrate the practical application of the new

method. The case studies presented include an assessment of existing harmonic

distortion levels within a distribution system, an example of an allocation of harmonic

emissions to a customer situated along a long MV feeder, and an evaluation of a system

containing numerous power factor correction (PFC) capacitors.

The first case study system is the Homepride zone distribution network. Details of this

distribution network can be found in Chapter 5. The system is presented here to

illustrate assessment of existing harmonic voltage levels on the network. More

specifically identifying the critical impedances of the system and applying suitable

harmonic planning levels throughout the network.

The second case study is the Katoomba Zone MV distribution network. This study

traces the application of the new harmonic emission allocation method to an MV

customer wishing to connect to the end of a relatively weak feeder. This case study was

150
the first application of the new allocation method to a real system. Discussion is

presented on the difficulties encountered including the interpretation of the pragmatic

procedures incorporated in the new method to complex load operation.

The third case study involves the Springhill sub-transmission substation. This study

addresses the practical application of pragmatic measures when dealing with the

presence of power factor correction (PFC) capacitors. The system analysed contains

sub-transmission, distribution, and a number of large industrial customers. The

difficulty of harmonic allocations within a network that experiences significant changes

in harmonic impedance due to PFC capacitor switching is specifically addressed.

In the three case studies use of the generalised methods presented in this thesis for

harmonic voltage management are illustrated. Each case study involves collection of

suitable data, application of harmonic models, and assessment of harmonic voltage

levels or customer emissions. A dependence on reliable data to complete the studies is

highlighted. Specific discussion on the above-mentioned issues is presented and the

deficiencies in the existing harmonic standard are highlighted.

7.2 Case study 1: Homepride zone distribution system

The Homepride zone distribution system is located in the western suburbs of Sydney.

Homepride is an 11kV distribution network owned and operated by Integral Energy

containing ten MV feeders supplying residential, commercial and industrial customers.

A majority of the feeders are predominantly overhead lines, however there is a

significant amount of underground cable included in some sections of the feeders. The

151
distribution system had a maximum demand of approximately 80% of capacity in the

year 2000. A schematic layout of Homepride was given in Figure 5.6 of Chapter 5.

Using data from the harmonic monitoring programme reported in Chapter 5 the

following section demonstrates the practical application of harmonic management

techniques introduced in Chapters 3 and 4. More specifically this case study presents

discussion on the following issues

(i) Identifying the critical impedances within the system,

(ii) Assessing harmonic levels given the loading of the distribution system, and

(iii) Applying suitable harmonic limits to different points within the system.

7.2.1 Identifying the critical impedances of the system

Harmonic voltages appear in distribution systems when non-sinusoidal currents from

distorting loads flow through the system. The harmonic voltages resulting from these

currents are proportional to the system harmonic impedance seen by the distorting load.

Thus, reducing the harmonic impedance of a system will lead to a ‘stronger’ system, i.e.

a system that has the ability to absorb a greater magnitude of harmonic current without

producing excessive harmonic voltage levels. For this reason it is important to identify

critical impedances that exist within the system. An assessment can then be made as to

where the impedance is most cost effective to reduce to provide increased harmonic

performance.

Using an appropriate base a diagram showing the harmonic impedance from the

transmission system down to the end of an LV distributor can be constructed. Figure 7.1

illustrates the calculated 5th harmonic impedance seen through the different sections of

152
the Homepride system and to the end of an LV distributor for customers at a residential

site (Site 4 of Figure 5.6 in Chapter 5).

HV MV feeder LV distributor

100%
90%
5 harmonic impedance

80%
70%
60%
50%
40%
30%
th

20%
10%
0%
Distance along system

Figure 7.1 5th harmonic impedance seen by load as percentage of total impedance

The HV section of the system in Figure 7.1 includes the source impedance of the

transmission system at (0.0015+j0.0271) per unit on a 100MVA base, three parallel-

connected 132kV/33kV 120MVA sub-transmission transformers, and a length of 33kV

overhead sub-transmission line. The zone substation transformer is the equivalent of

two 33/11kV 20MVA transformers connected in parallel for (N-1) redundancy. The

MV feeder consists of both overhead and underground lines and the LV distributor is a

length of overhead line. The load on the zone substation consists approximately of 50%

LV customers and 50% MV customers.

As illustrated in Figure 7.1 the impedance seen by a distorting load connected to the LV

distribution system is mainly due to the MV/LV transformer and LV distributor line.

Thus the voltage distortion contribution due to a single LV distorting customer will

predominantly occur along the LV distribution system and across the MV/LV
153
distribution transformer. However, as each customer on the LV distribution system will

typically represent less than 0.1% of the entire zone substation load the resulting

harmonic voltage may never become significant. Thus harmonic voltages from an

aggregate of all the loads within the distribution system need to be examined when

determining critical impedances. This is necessary to test whether the small

contributions of many customers will produce a significant overall voltage distortion at

different points within the distribution system.

HV MV feeder LV distributor

100%
90%
80%
5 harmonic voltage

70%
60%
50%
40%
30%
th

20%
10%
0%
Distance along system

Figure 7.2 5th harmonic voltage as percentage of level at end of distributor

A plot of the calculated 5th harmonic voltage distortion throughout the distribution

system due to a combination of smaller harmonic loads was completed using the second

summation law and is shown in Figure 7.2. For this study MV customers were

considered to give a similar harmonic contribution to that of an equivalent sized

combination of LV customers. Figure 7.2 illustrates that if harmonic contributions from

many customers are considered the largest harmonic voltage distortion increases occur

at the high impedance locations of the system transformers.

154
The small increase in harmonic voltage levels along the MV feeder shown in Figure 7.2

suggest connecting a distorting load at the end of the feeder will produce the worst case

of harmonic voltage distortion, however not significantly different from having the

same load at the beginning of the a feeder. This is due to the HV/MV transformer

providing most of the harmonic impedance across which the harmonic voltage is

developed when all distorting load currents are combined.

Figure 7.2 also suggests that if a DNSP plans to control harmonic voltage levels in a

distribution system by reducing system impedance, the best locations to do so are at the

HV/MV zone substation transformer or the MV/LV distribution transformer. Reduction

of the impedance of the MV/LV distribution transformers would be difficult to justify

for an existing system, as this would require the replacement of hundreds of

transformers, however may be viable during the planning phase of a new system.

Assuming most of the harmonic current injected in to the system by various customers

will flow through the zone substation transformer, reducing its impedance by a factor of

10% will also reduce harmonic voltage levels at the zone substation by 10%. At the end

of the LV distributor this reduction of harmonic voltage will appear as 2% overall due

to the voltage increases downstream from the zone substation transformer.

Alternatively, if an MV/LV distribution transformer impedance is reduced by 10% the

harmonic voltage level at the end of the LV distributor will reduce by 4%. A slightly

larger reduction in harmonic levels will obviously occur for MV customers, who

typically connect to the system via their own MV/LV distribution transformer with no

LV distributor lines. Use of aerial bundled conductors or underground cables along MV

155
feeders and LV distributors will also reduce the level of harmonic voltages, however not

by magnitudes as significant as that obtained by modifying transformer impedances.

Cost is a significant factor when deciding on harmonic mitigation techniques during the

planning phase of network design. A nominal figure obtained from Integral Energy for

the replacement cost of a zone substation transformer is approximately $550 000. For a

distribution transformer the cost is $25 000. Moving cable underground can cost up to

$450 per metre, offering less significant harmonic impedance reductions, but has

additional benefits of reduced voltage sags due to motor starts and arguably better

suburban streetscape environment. Thus reducing the harmonic impedance of

transformers and lines for existing systems would rarely prove to be a cost effective

method of harmonic mitigation.

Reducing system harmonic impedance during the design phase of a distribution network

is advantageous as it leads to a reduction in losses and increases the systems harmonic

absorption capability. Reduced system impedance evidently means an increase in fault

levels. This is not usually a problem if DNSP customers are aware of ratings required

for protection and switchgear. Reducing system impedance also reduces the impact of

other power quality disturbances, including voltage fluctuations, voltage sags and

voltage regulation. The cost of additional copper or aluminium in conductors will

obviously remain the most significant factor affecting decisions related to reducing

system impedances. The most cost effective method of mitigating harmonic problems

will often be the installation of harmonic filters or active compensators.

156
7.2.2 Consideration of existing loading when assessing harmonic levels
When assessing harmonic distortion levels within a distribution system it is important to

measure or have an indication of the loading on the system. The term loading here is

used to express the fraction of the total available capacity of the system that has already

been allocated to existing customers and does not include diversity. The loading is

usually a figure well known to the DNSP and is often expressed as the undiversified

maximum load.

Without the loading of the system a true indication of whether customer loads are taking

up their harmonic entitlements is not possible. For example, consider a system with a

harmonic voltage planning level of 5% for the 5th harmonic. If through a harmonic

monitoring programme the 95th percentile 5th harmonic voltage was measured to be 4%,

the DNSP may initially consider the system to have acceptable harmonic voltage levels.

However, if the system was only one third loaded it is expected that measured harmonic

levels should be less than 4%. Harmonic voltage levels should thus be measured against

a scaled harmonic voltage planning level. The scaled harmonic voltage planning level

can be determined using the second summation law from IEC 61000-3-6 and

equation (7.1).

S′T
L ′hMV = L hMV ⋅ α (7.1)
ST

where L ′UhMV is the scaled harmonic voltage planning level for order h at MV, L UhMV is

the harmonic voltage planning level for order h at MV, S′T is the system undiversified

existing loading, S T is the system total available power, and α is the exponent of the

summation law.

157
Using the scaled harmonic voltage planning level it is possible for a DNSP to assess

whether customer loads are taking up their full harmonic entitlement and whether there

is room to allow additional harmonic loads to be connected. If the scaled harmonic

voltage planning levels were used, the one third loaded system mentioned above would

be identified as having excessive harmonic voltage levels and cost effective mitigation

may be able to be achieved before harmonic levels become more problematic.

For Homepride, results from the harmonic monitoring programme at the commercial

site recorded in April 2002 are considered. The system is loaded to 24MVA with a

capacity of approximately 27.5MVA, allowing for 10% transformer overloading. The

resulting THDV and 5th harmonic voltage distortion levels at the LV commercial site

were found to have 95th percentile values of 4.17% and 3.81% respectively.

For the 5th harmonic a summation law exponent of α = 1.4 is recommended by

IEC 61000-3-6. Using equation (7.1) to evaluate the recommended harmonic planning

levels

S′T
L′5LV = L5LV ⋅ α
ST
24
= 5.00% ⋅ 1.4
27.5
= 4.54%

The recorded 95th percentile level can then be expressed as a percentage of the scaled

harmonic planning level

4.17%
U 5 COM LV =
4.54%
= 92%

158
Thus, the measured levels are quite close to the recommended harmonic planning levels

considering the loading of the system. Note that a 5th harmonic planning level of 5% has

been assumed for LV sites for this case. Such a level is considered to be quite

conservative in that the planning level recommended by IEC 61000-3-6 for the 5th

harmonic at MV is 5%, thus the planning level at LV would be expected to be slightly

higher than this in most cases.

7.2.3 Applying suitable limits to different points within the system

An important aspect of harmonic planning for a DNSP is the selection of harmonic

voltage planning levels at the different points within their distribution systems and

indeed further upstream in the sub-transmission and transmission systems. The IEC

approach to determining suitable harmonic voltage distortion planning levels in

distribution systems is to ensure that levels remain below equipment susceptibility

levels. That is, compatibility levels are determined for a broad range of equipment, and

then planning levels for the DNSP are set slightly below this to attempt to ensure the

compatibility levels are never reached. Determining compatibility levels for equipment

is an area of power quality that requires significantly more research and will change as

technology also changes. For this thesis it will be assumed that the compatibility levels

recommended by the IEC are well substantiated.

For transmission systems (HV and EHV) the harmonic voltage planning levels

suggested by IEC 61000-3-6 are 3% for voltage THD and 2% for the 5th harmonic. This

means expected harmonic voltage contributions from upstream HV and EHV networks

should be at most 2% for the 5th harmonic. Although no planning levels are specified in

the IEC standard for the lower extreme of the system, i.e. at the end of LV distributors,

159
compatibility levels for MV and LV systems are provided. Thus the standard provides

some guidance for the planning levels for the upstream and downstream extremities of

distribution systems, however insufficient guidance is provided to determine suitable

planning levels for points on the system between the extremities, e.g. the zone

substation busbar.

The Homepride system 5th harmonic voltage level at the zone substation 11kV busbar in

September 1999 was measured to be 1.61% (95th percentile). This is well below the

recommended planning level of 5.0% for the 5th harmonic. However it would be unwise

for harmonic planning levels at the MV busbar to be set at 5.0% as this does not allow

for any harmonic voltage increases along MV feeders, distribution transformers, and LV

distributors.

To determine what the levels of harmonic voltage distortion should be at the zone

substation busbar, the entire Homepride distribution system is considered as a long

feeder having distribution points with varying fault levels. The harmonic voltage

through the system and along the weakest MV feeder to the LV distributor (at which the

worst harmonic voltage is assumed to occur) is considered. To simplify calculations of

the harmonic voltages a representative system as illustrated in Figure 7.3 is utilised. The

procedure for allocating harmonic emissions outlined in Chapter 4 is then utilised to

determine appropriate voltages expected at each point on the system given that the

maximum 5.0% (for the 5th harmonic) will occur at the end of an LV distributor. Fault

levels for HV, MV and LV locations are selected to be indicative of typical sub-

transmission and distribution systems in Australia that match reasonably closely to the

Homepride system.

160
 j0.03

132kV bus, FL=3000MVA

j0.03
Other sub-transmission networks

33kV bus, FL=1500MVA

j0.93
Other zone substations

11kV bus, FL=100MVA

Other distribution feeders

j0.3 j0.3 j0.3 j0.3 j0.3

LV customers
j5.0 j5.0 j5.0 j5.0 j5.0 PCC

Aggregated models of MV and LV customers

Figure 7.3 Indicative schematic layout of a typical radial distribution system

The estimated harmonic voltages are established using the second summation law. A

background harmonic voltage is included in the calculations such that when the system

reaches the recommended harmonic limit at the LV customer’s PCC, the background

harmonic voltage combined with the harmonic voltage due to the system will equal the

recommended HV harmonic limit at the 132kV bus. Put more simply, it is assumed the

HV system will reach the respective limits the same time as the LV system. The

resulting values of the harmonic voltages at each voltage level are shown in Tables 7.1

for the 5th harmonic.

161
 Table 7.1 Indicative 5th harmonic voltage levels for typical power system
5th harmonic
Location
voltage
Background 0.9%
132kV bus 2.0%
33kV bus 2.5%
11kV bus 4.0%
11kV above PCC 4.4%
PCC LV customer 5.0%

From the results in Table 7.1 it could be suggested that indicative planning levels for the

sub-transmission 33kV bus and distribution 11kV bus should be approximately 2.5%

and 4.0% for the 5th harmonic respectively. As many systems may be weaker than the

study system in Figure 7.3, and because most of the system impedance seen by

customers appears at distribution level, it may be suitable for a DNSP to use a slightly

more conservative value for sub-transmission planning levels where possible. Use of the

HV planning levels at sub-transmission level is one suitable method to ensure fewer

harmonic problems at LV. It is assumed planning levels for harmonic orders other than

the 5th harmonic could be found by similar means or by appropriate scaling of the values

in Table 7.1 to suit the planning levels recommended by IEC 61000-3-6.

The HV harmonic voltage limits here have been fixed at the planning levels suggested

by the standard for the 132kV busbar at 2.0%. The harmonic current from the MV

distribution system itself will contribute a voltage component to this level. In Australia

the HV and EHV systems are typically meshed networks and thus it is difficult to

determine what this current contribution might be without detailed modelling, and goes

beyond the scope of this thesis.

162
7.3 Case Study 2: Katoomba zone distribution system
Pragmatic modelling of MV distribution systems and associated loads attempts to

reduce the complexity involved in determining effects of loads on the system and in

most cases the general approach will be sufficient. However, the complex behaviour of

individual loads may often provide additional difficulty when applying the procedures

to allocate harmonic emissions to customers. This case study presents the harmonic

assessment of a customer seeking approval for connection to Integral Energy’s

Katoomba zone distribution system. In particular this section presents discussion on the

following

(i) Assessment of background harmonic voltage levels,

(ii) Calculation of acceptable customer harmonic current emissions, and

(iii) Accounting for load interaction with system harmonic voltages.

The customer of interest (to be called Customer A) is located in the Blue Mountains

north of Sydney. A schematic of the Katoomba zone distribution system to which

Customer A is connected is illustrated in Figure 7.4, showing distribution transformers

and the 10 year projected maximum demand along the feeder to which Customer A is

connected.

The connection point for Customer A is located near the end of the MV feeder at

distribution transformer 23 (Tx 23). The customer is located in a mountainous area

making connection to a stronger (higher fault level) alternative feeder considerably

expensive. The fault level at the selected point of common connection (PCC) is

44MVA.

163
 66kV MV Sub-transmission

11kV MV Zone Substation Bus Bar

Tx 1 (193kVA)

Tx 2 (190kVA) 7 other feeders with equal

total load
Tx 3 (213kVA)
Key
Tx 4 (230kVA)
Pole Mounted Distribution Transformer
Tx 5 (283kVA) Pad Mounted Distribution Transformer

Tx 6 (282kVA) Tx 7 (125kVA)

Tx 8 (142kVA) Tx 9 (317kVA)

Tx 10 (160kVA)

Tx 11 (200kVA)
Tx 16 (355kVA) Tx 17 (56kVA)
Tx 12 (200kVA)
Tx 19 (20kVA) Tx 13 (482kVA) Tx 18 (170kVA)

Tx 20 (115kVA) Tx 14 (445kVA)

Tx 21 (120kVA) Tx 15 (240kVA)
Tx 22 (202kVA)

Tx 23 (210kVA + 815kVA for Customer A)

Tx 24 (130kVA)

Figure 7.4 Schematic layout of Katoomba distribution system feeding Customer A

Customer A’s load consists of a six pulse 530kW regenerative DC drive system and

50kW of auxiliary equipment with an overall approximate maximum demand of

815kVA. There is also a pre-existing load of 210kVA of the same customer at the PCC.

The load configuration for Customer A is illustrated in Figure 7.5. Passive harmonic

filters were included in the original installation by Customer A to reduce the levels of

5th, 7th, 11th and 13th harmonics injected back into the distribution system.

164
 PCC
FL=44MVA
11kV/415V Transformer
1000kVA
X = 0.05pu
Auxiliary
50kW

DC Drive
530kW
Harmonic Filters 0.7-0.8pf

=M

Figure 7.5 Schematic layout of Customer A’s load

Customer A calculated harmonic spectrum

Caculated for 100% load with reference to PCC
35
Without Filter With Filter
30

25
Current (A)

20

15

10

0
Fund 3 5 7 11 13 17 19 23 25 27 29 35
Harmonic Number

Figure 7.6 Calculated harmonic spectrum of Customer A’s installation at PCC

Customer A provided details of the load operation for the DNSP to make an appropriate

assessment of whether connection should be granted. Design calculations performed by

Customer A found that with harmonic filters connected the harmonic spectrum of the

load current produced a maximum 5th harmonic at 28% of fundamental as shown in

Figure 7.6. This 5th harmonic current is significantly above the 20% or less, a value

typically expected for a similar sized six pulse DC drive system [37].

165
Assessment of Customer A’s installation was completed after the design and installation

phase had already been completed. Ideally the customer and DNSP should consult prior

to installation to ensure that the design is acceptable regarding the level of harmonic

emissions and possible mitigation requirements before a connection agreement is

obtained. A further difficulty for this installation was that it occurred during period of

transition when Australia was transferring to a new harmonic standard from AS 2279.2

to AS/NZS 61000.3.6 (adapted IEC 61000-3-6) and there were some concerns on which

standard was most appropriate to apply. Only a discussion regarding IEC 61000-3-6 is

presented here.

7.3.1 Assessment of background harmonic voltage levels

The first stage of assessment was to determine the existing levels of voltage harmonics

on the distribution system. From voltage distortion profiles provided by the DNSP the

existing 95th percentile 5th harmonic distortion was found to be approximately 2.72%

with THD at approximately 2.84%. Although this is well within the recommended

planning level of 5% for the 5th harmonic from IEC 61000-3-6 it is important to verify

that this level is also in proportion with the existing loading. Also verification of the

growth in harmonic voltage levels is possible if measurements are taken before and after

new customer connections are completed.

Recommended planning levels for the 5th harmonic voltage at HV and MV of 2% and

5% respectively are assumed. Note that the MV planning level includes a contribution

from HV planning level. The second summation law (α=1.4 for 5th harmonic) is used to

determine the allowable contribution from MV and LV (GhMV+LV) when the system is

fully saturated as follows.

166
GhMV + LV = α LαhMV − LαhHV
= 1.4 5%1.4 − 2%1.4 (7.2)
= 3.97%

Taking into account the overload capacity and maximum utilisation factor of the

HV/MV transformer, the total power available to the distribution system was calculated

to be 27.6MVA. Considering the system was only loaded to 20MVA, the allowable

harmonic contribution from MV for the partially loaded case can be determined as

follows

1
S α
GhMV + LV ( partially loaded ) =  i  × GhMV + LV
 St 
1
 20 MVA  1.4
=  × 3.97%
 27.6MVA 
= 3.15%

Combining the allowable contribution from the MV partially loaded case, and assuming

the HV system is at the same level of loading (i.e. not utilising the full 2% of

recommended 5th harmonic limit) the recommended 5th harmonic voltage limit for the

system in existing configuration is as follows

1
 20 MVA 1.4
LhHV ( partially loaded ) =  × 2.00%
 27.6 MVA 
= 1.59%

LhMV ( partially loaded ) = 1.4 3.15%1.4 + 1.59%1.4

= 3.97%

167
As the HV contribution is of the same proportion as the MV the above recommended 5th

harmonic voltage limit could have been calculated directly from the combined harmonic

limit of 5.0% for the 5th harmonic as follows

1
S α
LhMV ( partially loaded ) =  i  × LhMV ( fullyloaded )
 St 
1
 20MVA 1.4
=  × 5.0%
 27.6MVA 
= 3.97%

Thus the existing 95th percentile level of 5th harmonic voltage distortion is within the

planning levels recommended by IEC 61000-3-6. Performing this calculation provides

an indication of whether customers are taking up their emission entitlements. The DNSP

may choose to utilise spare harmonic capacity to allow a particular customer additional

harmonic emission rights if required.

7.3.2 Calculation of acceptable harmonic current emissions

To determine acceptable harmonic emissions for Customer A’s installation according to

IEC 61000-3-6 the consecutive stages and tests of the standard need to be completed as

required. The acceptable harmonic current emissions at the given PCC as determined

according to Stage 1 and the initial tests from Stage 2 are illustrated in Figure 7.7.

Calculations related to these stages are given in Appendix E, Sections E.1 and E.2

respectively.

168
 Customer A acceptable harmonic spectrum
Calculated for 100% load with reference to PCC
2.5
Stage 1 Stage 2, Test 1
2

Current (A)
1.5

0.5

0
3 5 7 9 11 13 17 19 23 25
Harmonic Number

Figure 7.7 Calculated acceptable harmonic spectrum for Customer A’s installation

It can be seen that the harmonic currents indicated by Customer A in Figure 7.6 exceed

the acceptable emissions, thus the installation fails Stage 1 and Stage 2, Test 1 of the

standard and the more detailed harmonic allocation using Stage 2, Test 3 is required.

For the Stage 2, Test 3 to be completed the generalised method introduced in Chapter 4

is necessary due to the inadequacies of the approach outlined in the standard to be able

to deal with non-homogenous systems. Application of Stage 2, Test 3 is presented in the

remainder of this section. It is expected this test will provide a more generous

recommended allowable harmonic emission limit. The application of a Stage 3 study

may be ultimately applied at the discretion of the DNSP if connection is granted. This

stage is designed to allow connection of customers on a temporary and precarious basis.

As the amount of data required to perform the calculation of acceptable current

emissions is extensive, and not all required data was available, the following

assumptions were made

• The 5th harmonic is the only harmonic voltage of concern for this study, and

169
• All feeders supplied by the zone substation have identical total loadings to that of

the feeder supplying Customer A

Three different scenarios are considered in determining an acceptable harmonic voltage

contribution from Customer A using the method outlined in Chapter 4 for compliance

with Stage 2, Test 3 of IEC 61000-3-6. The three scenarios are as follows:

(i) The feeder on which the installation is to be installed is the "weakest feeder" and

all other feeders are significantly "stronger".

(ii) The feeder on which the installation is to be installed is the "weakest feeder" and

some other feeders are significantly "stronger".

(iii) The feeder on which the installation is to be installed is "identical" to all other

feeders supplied by the zone substation, i.e. all other feeders are equally "weak".

From a simple load flow calculation utilising MATLAB®, it was found that the PCC for

Customer A produced the lowest fundamental voltage and thus was considered to be at

the end of the weakest feeder.

The first stage in scenario (i) calculations is to determine a value for the harmonic

current 'allocation constant' for the model system. The method suggested in Chapter 4

for determining the value of the 'allocation constant' is given by equation (7.3).

G hMV
k≈ 1
(7.3)
 n α 
α

 ∑ Si Z hi2 + S 0 Z h0
α α
2
+ (S F2 + S F3 + ... + S Fr )Z h0
2

 i =1 

The harmonic voltage planning level at HV and MV are assumed to be 2% and 5%

respectively. Thus the global harmonic voltage emission, GhMV, for all loads at MV in
170
this scenario is 3.97%. Assuming that there are no local loads at the zone substation MV

busbar equation (7.3) can be simplified to equation (7.4).

G hMV
k≈ 1
(7.4)
 n α 
α

 ∑ Si Z hi2 + (S F2 + S F3 + ... + S Fr )Z h0
α
2

 i =1 

MATLAB® was used to determine the value of the 'allocation constant' for the model,

given the values of load sizes and network impedances. A spreadsheet could just as well

be used for this calculation. Using a 1MVA base the value of the 'allocation constant' k

was found to be 1.28%. Calculation given by equation (7.5) is then used to determine

the allowable harmonic current and voltage allocation. The total load at the PCC (Tx 23

of Figure 7.4) is 1025kVA (including the 815kVA for Customer A).

1
kSiα
E Ihi =
Z hi
1

1.28% × 1.025 1.4

= (7.5)
0.0888
= 4.37%

Customer A’s installation is not the only load connected at this PCC, i.e. does not make

up the total load connected. The second summation law from IEC 61000-3-6 can be

used to determine contributions from Customer A as compared to the total load.

Equation (7.6) is used to allocate the respective harmonic emission to Customer A.

1

S α
I hi =  i  I hM
 SM 
1

 0.815  1.4
=  4.37% (7.6)
 1025 
= 3.71%

171
This value for allowable harmonic current (on 1MVA base) corresponds to a 5th

harmonic voltage allocation of V5 = 0.33% (on Customer A’s load base) and a 5th

harmonic current allocation of I5 = 1.95A. This value is much lower than the calculated

harmonic current in Figure 7.6.

As it is most likely that the study feeder is not the only weak feeder on the zone

substation the first scenario result is viewed as being pessimistic. By assuming that

there are other feeders similar in nature, the rule of thumb 'divide by √2' introduced in

Chapter 4 may be applied to the third term in the denominator of equation (7.3) as per

scenario (ii).

Scenario (iii) assumes that all feeders connected to the zone substation are identical in

configuration, impedance, and load size. With this assumption a different value for the

harmonic current 'allocation constant' is calculated. As all feeders are assumed to be

identical, and that there are no local loads, equation (7.3) can be simplified to

equation (7.7).

G hMV
k≈ 1
(7.7)
 α
 α
Z α

 ∑ S i Z i2 + (n − 1)∑ S i α 
0

 i i
Z i2 


The results from scenarios (i)-(iii) are presented in Table 7.2. It can be seen that the

harmonic voltage allocation from scenarios (ii) and (iii) give similar magnitudes.

172
 Table 7.2 5th harmonic voltage allocation for scenarios (i)-(iii)
Allocation Voltage Current
Scenario
constant allocation allocation
(i) Other feeders very strong k=1.28% V5=0.33% I5=1.95A
(ii) Other feeders stronger k=1.56% V5=0.40% I5=2.37A
(iii) Other feeders equally weak k=1.70% V5=0.44% I5=2.58A

The values of 5th harmonic voltage contribution calculated by the company responsible

for Customer A’s installation suggested a value of V5 = 1.13%, which was derived from

a 5th harmonic current of I5 = 5.7A provided in Figure 7.6. For the above method of

applying IEC 61000-3-6 the calculated values are less than half the values calculated by

the company commissioned to install the load. Thus, it was perceived that the

installation would fail the IEC 61000-3-6 requirements.

If the installation was to comply with IEC 61000-3-6 the 5th harmonic current should be

approximately half that of the value calculated by the installing company. This may

require reconfiguring the load, decreasing the feeder impedance, or the addition of

larger harmonic filters that are able to absorb a greater amount of harmonic current. Of

course discretion of the applicability of IEC 61000-3-6 and the above results to

Customer A lies with Integral Energy.

7.3.3 Accounting for load interaction with system harmonic voltages

As the installation of Customer A was operational prior to assessment, field

measurements of harmonic voltage and current could be used to determine compliance

with IEC 61000-3-6. A component of the measured harmonic current would be due to

the system harmonic voltage, however it will be assumed negligible compared to the

harmonic current emissions derived purely from the DC drive. A capture of one full

173
load cycle of the DC drive for the 5th harmonic voltage and current on the LV side of

the distribution transformer was obtained as illustrated in Figure 7.8.

12 300
VA5 (V) IA5 (A)
10
250
8

6
200

RMS Current (A)

RMS Voltage (V)
4

2 150

0
100
-2

-4
50
-6

-8 0
11:42:19 11:43:19 11:44:19 11:45:19 11:46:19 11:47:19 11:48:19 11:49:19 11:50:19

Time
th
Figure 7.8 Full load cycle of 5 harmonic rms voltage and current

To verify that the loading of the drive did not alter the harmonic current significantly

loads of 0.21 Tonne (low), 0.93 Tonne (medium) and 3.53 Tonne (full) were applied

and the resulting 5th harmonic currents obtained are illustrated in Figure 7.9. As the

variation of 5th harmonic current was not significant for the purpose of compliance it

was assumed the pattern of harmonics in Figure 7.8 repeated 50-60 times over an eight

hour period, seven days a week.

200
Load 3.53T
180
Load 0.93T
160
Load 0.21T
RMS Current (A)

140

120

100

80

60

40

20

0
Time
Figure 7.9 Half load cycle 5th harmonic rms current for various loading

174
The 95th percentile values of the 5th harmonic current on the MV side of the transformer

was estimated by interpolating values from Figure 7.8 over a full week resulting in a

current of 115.2A (LV) and 4.3A (MV). This exceeds the allocated 2.58A. The 95th

percentile value for the 5th harmonic voltage on the LV side was calculated to be 4.0%,

indicating a value considerably less than this on the MV side due to the drop across the

transformer impedance. This leaves a considerable margin to the recommended

planning level of 5.0%.

It can be seen in Figure 7.8 that during approximately 40% of each cycle Customer A’s

equipment was seemingly reducing the network 5th harmonic voltage levels by injecting

an out-of-phase component of 5th harmonic current. IEC 61000-3-6 does not consider

the connection of load types that have a compensating effect on the system distortion,

indicating a problem with the inflexible application of the summation law. Perhaps the

customer should not be penalized for producing such harmonic currents. If the out-of-

phase harmonic components are excluded from calculations, the resulting 95th percentile

5th harmonic current was considerably less at 3.0A (MV), only exceeding recommended

values by a small amount.

Although Customer A’s load has failed assessment in this case study if it could be

established that the background harmonic distortion were low relative to the existing

feeder load, there may be scope to grant Customer A compliance under Stage 3 of

IEC 61000-3-6 ("Temporary and precarious basis"). To assess the extent of the

background harmonic distortion levels a further study would have to be completed

which requires data of present background levels, present loading along feeder, and

potential loading along feeder.

175
7.4 Case study 3: Springhill sub-transmission substation
This section presents a discussion on a harmonics monitoring campaign carried out at

the sub-transmission point of supply of a distribution network with the objective of

benchmarking background harmonic distortion levels prior to refurbishment of the

substation and to ensure emissions from HV customers and the downstream MV

networks are within acceptable levels.

Springhill sub-transmission substation is owned by Integral Energy and located

approximately 100km south of Sydney. Proposed refurbishment of the substation

included replacement of seven 132/33kV 60MVA transformers, with transformers of

ratings 60MVA or 120MVA, and refurbishment of one other 60MVA transformer,

effectively upgrading the total capacity of the substation by approximately 30%.

To determine existing levels of harmonics a two-week harmonics monitoring campaign

was undertaken. Assessment of harmonics levels was carried out, where possible, as per

guidelines from the relevant standard IEC 61000-3-6, although instrument limitations

prevented an exact assessment being completed. Data from the monitoring campaign

was utilised to assist in modelling the sub-transmission substation and surrounding

network to establish how the refurbishment would affect harmonic distortion levels.

Some of the issues encountered with the monitoring program and assessment relevant to

this thesis are discussed here and include

(i) Evaluating the impact of multiple combinations of switched capacitors,

(ii) Lack of clearly defined planning levels for sub-transmission systems, and

(iii) Calculation of acceptable emissions for MV customers.

176
7.4.1 System layout
A single line diagram of the Springhill sub-transmission substation is illustrated in

Figure 7.10.

Bulk Supply Point

132kV Customer C 132kV Customer D

132kV Bus Bar

Ynyn0 Ynyn0

1
3

33kV Distribution Network

2 33kV Customer B

Figure 7.10 Springhill sub-transmission substation single line diagram

Substation load is comprised of three large industrial customers and Integral Energy’s

own 33kV network. Industrial Customer B is fed at 33kV via five 60MVA transformers,

two other industrial customers (Customer C and Customer D) are fed at 132kV and

Integral Energy’s own 33kV distribution network is supplied via three 60MVA

transformers. Integral Energy’s 33kV network includes some smaller industrial

customers and a significant amount of commercial and domestic load.

7.4.2 Assessment of levels and emissions

IEC 61000-3-6 suggests that the topology of a network is an important consideration

when selecting harmonic voltage planning levels. Indicative planning levels for both

MV and HV are provided by the standards. For effective harmonic management MV

177
planning levels should be assessed at the most extreme points on the system where it is

expected harmonic levels will be highest, i.e. the end of distribution feeders for

harmonics, thus utilising MV planning levels for this sub-transmission study was not

perceived practical. There is no clear indication given by the standard on how to select

planning levels for points in the network between transmission and extremes of

distribution. The substation 33kV busbar in this study is defined as being sub-

transmission, and accordingly planning levels should, in principle, appear somewhere

between the HV and MV planning levels. However, fault levels at the 33kV busbar

were closely matched to that of the 132kV busbar, and thus for this project the HV

planning levels were used for both transmission (132kV busbar) and sub-transmission

(33kV busbar). The allocation of suitable planning levels throughout a network has

since been addressed in [45].

In order to compare DNSP harmonic levels (or consumer harmonic current emissions)

to planning levels, the minimum measurement period specified in IEC 61000-3-6 is one

week, with measurement of the following [15]:

(i) greatest 95% daily value of Uh,vs (or Ih,vs) (rms value of individual harmonic

components over very short 3 second period) should not exceed the

planning level,

(ii) maximum weekly value of Uh,sh (or Ih,vs) (rms value of individual harmonics

over short 10 minute periods) should not exceed the planning level, and

(iii) 99.9% weekly value of Vh,vs (or Ih,vs) should not exceed 1.5 to 2 times the

planning level (or emission limit).

178
However most field harmonic monitoring instruments are not able to measure and store

data at 3 second intervals for a whole week (or even a single day) due to memory

limitations. Also as damage due to harmonic voltages is rarely caused by excessive

short-term levels but more usually exposure to longer term levels, a more practical

method of assessment is using the EN 50160 standard [42] method of using only the

95th percentile 10 minute mean rms values of each of the individual harmonics (voltage

and current). For the purpose of a harmonic assessment of Springhill this method will be

used in conjunction with the appropriate IEC 61000-3-6 indicative planning levels.

7.4.3 Power factor correction capacitor considerations

A study of the effects of power factor correction (PFC) capacitors at and ‘electrically

nearby’ the Springhill sub-transmission substation was required to ensure no critical

harmonic resonance would occur that might cause excessive harmonic voltages. The

simulation study was completed for the substation before and after refurbishment, with

the altered harmonic impedance of the transformers, the most significant modification.

As there were nine capacitor banks in the vicinity of the sub-transmission substation the

number of possible combinations of capacitors connected at any one time could be up to

512 combinations, suggesting an excessive number of iterations of simulations. A

number of capacitor banks were switched based on time and load, thus it was possible

to reduce the number of possible capacitor combinations by a factor of 10 based on the

normal operating conditions of the substation before simulations were carried out.

MATLAB® was used to model the entire study system generating Tableau matrices

calculated at each increment of frequency with individual power system components

modelled as per recommendations in [65]. Simulations under various loading conditions

179
were carried out to identify possible harmonic resonances. Figure 7.11 illustrates the

resulting harmonic impedance curves at the 33kV distribution network busbar for six

different capacitor combinations with no load damping considered.

2
10

1
10
Impedance (pu)

0
10

-1
10

-2
10

-3
10
0 2 4 6 8 10 12 14 16 18 20

Harmonic (h)

Figure 7.11 33kV harmonic impedance for 6 different capacitor combinations

(no damping)

From Figure 7.11 it can be seen that one of the capacitor combinations creates a

resonance at both the 5th and 7th harmonic. It was found that this particular condition

only exists for a very short time interval as capacitors within the same bank are

gradually switched in.

180
 5th Harmonic current and capacitor switching over survey period
33kV Ia5
Ia5 (PM30) Cust
Ibhp5 B Ib5
(PM30) 33kV Ic5
Ic5 (PM30) Cap
Spr 11 Cap
Spr 2 2 Cap
Spr 33 Cap
Dap 14 Cap
Dap 25 Cap
Outer6 Cap
Bel 7
90

80

70
Distortion Level (%)

60

50

40

30

20

10

0
27/Mar 27/Mar 28/Mar 28/Mar 29/Mar 29/Mar 30/Mar 30/Mar 31/Mar 31/Mar 1/Apr 1/Apr 2/Apr 2/Apr 3/Apr 3/Apr 4/Apr 4/Apr 5/Apr 5/Apr 6/Apr 6/Apr 7/Apr 7/Apr 8/Apr 8/Apr 9/Apr 9/Apr 10/Apr
0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00

Figure 7.12 Capacitor switching vs 5th harmonic current at 33kV bus

5th Harmonic voltage and capacitor switching over survey period

33kV Vab5
Vab (PM30) 33kV Van5
Van (POLYQ) Cap
Spr 11 Cap
Spr 22 Cap
Spr 33 Cap
Dap 14 Cap
Dap 25 Cap
Outer6 Cap
Bel 7
1.8

1.6

1.4
Distortion level (%)

1.2

0.8

0.6

0.4

0.2

0
27/Mar 27/Mar 28/Mar 28/Mar 29/Mar 29/Mar 30/Mar 30/Mar 31/Mar 31/Mar 1/Apr 1/Apr 2/Apr 2/Apr 3/Apr 3/Apr 4/Apr 4/Apr 5/Apr 5/Apr 6/Apr 6/Apr 7/Apr 7/Apr 8/Apr 8/Apr 9/Apr 9/Apr 10/Apr
0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00

Figure 7.13 Capacitor switching vs 5th harmonic voltage at 33kV bus

The harmonic voltage for the very short interval for which this capacitor combination

existed was examined in conjunction with the appropriate block of survey data where an

insignificant rise in 5th and 7th harmonic voltage levels was found. Figures 7.12 and 7.13

illustrate the variation of 5th harmonic current and voltage at the 33kV busbar for

various switched capacitor combinations (capacitors 1-7). Although some small

transients in harmonic voltage and current coincide with capacitor switching instances

no problematic resonances seem to occur.

The harmonic resonance is possibly reduced greatly due to effects of load damping

which essentially trim down the peak of the harmonic resonance. A simulation of the

same capacitor combinations with a very light loading of 30% of normal load was

considered and results are illustrated in Figure 7.14. For the simulations the load was

represented by an inductor in parallel with a series combination of inductor and resistor,

as per Figure 6.3 of Chapter 6.

2
10

1
10
Impedance (pu)

0
10

-1
10

-2
10

10-3
0 2 4 6 8 10 12 14 16 18 20
Harmonic (h)

Figure 7.14 33kV harmonic impedance for 6 different capacitor combinations with
damping due to 30% loading

182
As can be seen in Figure 7.14 harmonic resonances are greatly attenuated by even a

small amount of load damping. The level of attenuation in fact reduces the magnitude of

resonances to below twice that of the case with no capacitors present (Upper line in

Figure 7.14). This seems to align with the findings in Chapter 6 of using twice the

harmonic impedance with no capacitors to provide a pessimistic approximation of

harmonic impedance with capacitors present. The simulations completed indicated that

the shift in resonant frequencies due to the changing impedance of substation

transformers did not create any new problematic harmonic resonances.

Of the problematic harmonic resonances identified in the detailed harmonic study only

one such capacitor combination occurred during the survey period. It was also noted

that the several capacitor banks were usually permanently switched in and others

switched in during rare extremely high load periods. Thus a further reduction in

capacitor combinations could have been achieved.

After the refurbishment of Springhill substation the impedance seen by the Integral

Energy 33kV network will decrease. This may give rise to different resonant

frequencies. A study of the harmonic impedance seen by the Customer B and Integral

Energy 33kV networks using the new transformer details should also be performed.

7.4.4 Assessment of emissions

Acceptable harmonic emissions from Customer B and the 33kV distribution network

were calculated as per Stage 2 Test 1 of IEC 61000-3-6. Calculations are presented here

for the 5th harmonic only. A worst-case harmonic impedance of 2 × Zh (up to 8th

harmonic) applies due to the large number of neighbouring capacitor banks possibly

183
causing resonant effects. The maximum demand of Customer B is calculated as follows

by equation (7.8)

Si = 3 × 60MVA transformers = 180MVA (7.8)

The rated capacity of the sub-transmission system at the PCC to which Customer B is

connected is

St= ∑ Sout = 300 + 180 + 60 + 60 = 600MVA (7.9)

Allocation of acceptable harmonic voltage and current contributions are then calculated

using equations (7.10) and (7.11) and assuming a planning level of 2% for the 5th

harmonic, and a fundamental source impedance of 0.0337∠86.9°pu (100MVA base).

Calculations for the other harmonics are not presented here but were completed using a

simple spreadsheet.

Si
EUhi = LhHV α (7.10)
St

180
= 0.02 1.4
600

= 0.00846 pu

E Uhi
EIhi = (7.11)
Zh

0.00846
=
2 × 5 × 0.0337sin(86.9°)

= 0.0252 pu (100MVA base)

= 11A (5th harmonic) at 132kV, or

= 44A (5th harmonic) at 33kV

184
Figure 7.15 illustrates acceptable harmonic emission limits calculated for Customer B

plotted against maximum harmonic current measured during the monitoring campaign.

Although the 5th harmonic current in Figure 7.15 is shown as exceeding the acceptable

emission limit this may be regarded as acceptable by the utility due to the following

(i) The harmonic impedance used for the allocation was approximated as twice the

impedance with no PFC present. This is conservative considering the amount of

load damping illustrated in Figure 7.14.

(ii) Harmonic emissions from other customers (Customer C and Customer D) are

lower than their allowances.

(iii) Harmonic voltage levels are not approaching recommended HV planning levels.

120.0

100.0
Current Distortion (A)

80.0

60.0 Limit
Measured
40.0

20.0

0.0
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Harmonic

Figure 7.15 Customer B allocated and measured maximum harmonic current.

Theoretically, after substation refurbishment the allocation of harmonic emissions

would require recalculation, as the total capacity of the sub-transmission substation is to

increase. This would mean a slight reduction in current allocation to Customers B, C,

and D but an increase for the 33kV distribution network. It would be very difficult and

perhaps unrealistic for utilities to impose such changes in allocations if previous

emission agreements with customers had been established.

185
If required, the harmonic voltages on the 132kV busbar can be interpolated from the

33kV distribution network results by combining the 33kV harmonic voltage, currents

and impedance using the second summation law. It is expected that harmonic voltages

on the 33kV busbar would be higher than that of the 132kV busbar, thus there is no

need to complete such a calculation in this instance, as the levels of distortion do not

approach the planning levels.

80%

Van Vbn Vcn

70%
Distortion as percentage of limit

60%

50%

40%

30%

20%

10%

0%
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 THD

Harmonic

Figure 7.16 33kV network harmonic distortion as a percentage of HV planning level

Benchmarking of results is achieved by reporting harmonic voltage as a percentage of

recommended planning levels. If emission levels from all customers are at their limits

and voltage disturbance levels are well below the planning levels, there may be scope

for allowing an increase of customer emissions if requested. Figure 7.16 illustrates the

harmonic voltage levels expressed as a percentage of HV planning levels for each phase

of the 33kV distribution network.

It is noticeable in Figure 7.16 that the 6th harmonic voltage is at an unexpectedly large

percentage of recommended planning level. The values illustrated in Figure 9 were

obtained from line-to-neutral measurements. The high value of 6th harmonic could be

186
attributed to the presence of zero sequence created by measuring the system line-to-

neutral. This was confirmed by low line-to-line measurements using another instrument.

This is consistent with the system being reasonably balanced and the 6th harmonic

appearing as a zero sequence triplen harmonic and thus cancelling in the line-to-line

voltages. It is proposed that line-to-neutral measurements may not provide an adequate

representation of three wire systems for harmonic assessment.

7.5 Summary
In three case studies investigated, the difficulties encountered highlight the complexity

and shortfalls of the IEC 61000-3-6 standard. Some of the shortfalls include

(i) No clear direction to set harmonic voltage planning levels at points within the

distribution network other than at the transmission level and end of distribution

feeders.

(ii) No generalised approach to allocating suitable emission limits to customers spread

along a distribution feeder having significantly different fault levels.

(iii) No straightforward approach to calculating harmonic impedances in the presence

of significant system capacitance

Assessment of existing harmonic voltage levels on a system has been demonstrated

using a weighted approach that depends on the level of loading on the system. The

significant harmonic impedances of a distribution network have been identified as the

zone substation and distribution transformers. A reduction in these impedances will

increase a network’s ability to absorb harmonic currents without exceeding harmonic

voltage planning levels. An attempt at determining harmonic planning levels through a

187
distribution network has been made. This method will have to be implemented on a

number of typical systems to ensure that suitable planning levels are selected.

Successful implementation of a harmonic emission allocation method for loads spread

along a feeder has been completed. Use of the different approximations outlined in

Chapter 4 to estimate the value of the harmonic allocation constant has been

demonstrated.

Pragmatic modelling of networks containing significant system capacitance have been

shown to over-estimate the harmonic impedance thus resulting in pessimistic harmonic

emission allocations. Detailed harmonic analysis has shown to be difficult due to the

number of different network configurations that may exist, including capacitor

switching.

Application to a number of different systems is required to better refine the approaches

to harmonic emission allocation and determination of harmonic impedance outlined in

Chapter 4 and Chapter 6 respectively.

188
 Chapter 8

Conclusions

8.1 Conclusion and recommendations

This thesis has described the development of guidelines and associated tools for

harmonic distortion level management in medium voltage (MV) electrical distribution

systems. The implementation of such guidelines is necessary to allow effective

management of harmonic emissions from distribution customer installations.

Statistical harmonic models of residential, commercial and industrial load types have

been developed to simulate the global behaviour of distorting loads at distribution

substations. The load models represent aggregates of loads and are specifically intended

for calculation of harmonic emissions for comparison with the relevant standards. Both

time and phase diversities are included in the representative load models.

A harmonic monitoring programme has been designed and implemented on a typical

electrical MV distribution system to establish parameters for the three types of

statistical harmonic load models mentioned above, and also to confirm the relative

accuracy of distortion level prediction techniques presented.

Suggestion of significant growth in harmonic voltage distortion levels internationally by

the literature reviewed has been verified to a limited extent by the harmonic monitoring

campaign. To determine an effective measure of harmonic growth, i.e. 0.1% per year or

similar, a minimum monitoring time of three years is recommended to ensure seasonal

variations do not distort results. Excessive levels of harmonic voltage distortion can

cause equipment malfunction or destruction, reduced lifetime of power system

189
equipment and connected customer loads, and increased power system losses. This

emphasises the need for distribution utilities to become more vigilant in their approach

to managing harmonic distortion levels on their electrical distribution systems.

For the past 25 years the harmonics standard for electrical distribution systems in

Australia has been AS 2279.2. While the procedures in this standard could be

implemented with relative ease, the increasing complexity and diversity of load

technologies existing on electrical distribution systems has made the standard obsolete.

The adoption of the international technical report IEC 61000-3-6 in 2001 as the present

harmonic standard in Australia has overcome some of the unaddressed issues in

AS2279.2. The new harmonic standard AS/NZS 61000.3.6 however is not as clearly

defined as its predecessor and there are some ambiguous sections that do not allow for a

clear approach to be taken with regard to allocation of acceptable emissions and

assessment of customer installations. Some of the tools and guidelines developed in this

standard seek to clarify the ambiguous sections of the new standard.

A method has been developed to estimate the 95th percentile cumulative probability

level of harmonic voltage distortion on an electrical MV distribution system as required

by IEC 61000-3-6. This method includes techniques to overcome the difficulties in

determining distortion levels when feeders are not loaded to their fullest extent, and

allows for cases when levels of background distortion may not be as high as limits set

by the standard. The method provides a useful tool to combine the effects of non-precise

distribution loads at the planning stage or to evaluate distortion levels of existing

systems.

190
A generalised method has been developed to extend the IEC 61000-3-6 approach of

allocating allowable harmonic emissions to the case where customers are distributed

along an MV distribution system feeder having significantly different fault levels. This

is an area that IEC 61000-3-6 fails to cover adequately to allow application to real

systems. The method involves the determination of an 'allocation constant' using the

agreed loading of all customers and the system harmonic impedances.

The generalised approach of allocating emissions to customers described above

typically requires an extensive amount of data. This data may not always be available to

the utility engineer. An extension to the generalised method has thus been completed to

cater for the complex situation where only limited data is available. This is achieved by

looking at several extreme cases that categorise the most common MV distribution

system feeder configurations, and through the use of correction factors for the

'allocation constant'.

A preliminary look at the effect of power factor correction capacitance on the harmonic

guidelines developed has shown that the presence of capacitors usually require detailed

analysis for effective harmonic management. The present standards approach of using

'twice the impedance' is illustrated to be a useful tool when resonances do not occur at

problematic harmonics.

The planning guidelines developed have been applied to example systems, both with

and without the presence of power factor correction capacitors, to illustrate the

harmonic management tools. This includes identification of the key indicators for a

distribution system's harmonic performance.

191
8.2 Further work
The harmonic monitoring program implemented on the typical distribution system has

illustrated the diversity of harmonic distortion levels. An overall growth in harmonic

levels has been found although the seasonal variation in the levels is significantly larger.

A long-term harmonic monitoring program of many distribution systems throughout

Australia would establish a better indication of the growth trends of harmonic distortion

levels. This would also allow a comparison of harmonics indices between DNSPs to

determine the performance of a particular provider.

The results presented in this thesis suggest that in most cases a detailed harmonic study

is required when multiple capacitor installation exist within a distribution system. Such

detailed studies require a significant amount of time and data. It is proposed that a rule

of thumb, similar to the ‘twice the impedance’ suggested by IEC 61000-3-6 used in this

thesis, needs to be further developed to allow a simpler approach for use in conjunction

with the present standard.

This thesis has focused on the application of harmonic management tools for MV

distribution systems. In Australia these systems are typically radial, with the exception

of some meshed networks located in inner city locations where a stronger supply is

required. Further work is required to extend the harmonic management tools introduced

in this thesis to meshed distribution networks and transmission systems, which are

typically configured as meshed networks.

Finally, compatibility levels for power system equipment remain an area yet to be fully

explored by power quality researchers. While the IEC compatibility levels have been

192
assumed to be valid in this thesis, there is no documented evidence as to why these

particular limits were selected other than they being reasonable based on existing

harmonic measurements.

193
 Statement of original contributions

As technology develops many customer loads are becoming increasingly sensitive to

excessive harmonic voltage distortion. Conversely, loads are also producing relatively

higher levels of harmonic emissions as technology drives for greater use of power

electronics to increase controllability and efficiency. Electricity distribution network

service providers (DNSPs) should now be looking towards preventative measures to

ensure that voltage distortion levels are within limits set by the appropriate standards.

Measures will need to be taken at the planning stage to ensure that the distribution

systems will be able to meet harmonic standards as load distortion rises. Tools will also

need to be developed to allow effective system modelling and comparison with

standards, especially in the planning phase where details of loads are usually not

accessible.

The original contributions in this thesis include the following to aid in the management

of harmonics in MV distribution systems:

(i) Statistical harmonic models of residential, commercial and industrial load types

have been developed to simulate the global behaviour of distorting loads at

distribution substations. The load models represent aggregates of loads and are

specifically intended for calculation of harmonic emissions for comparison with

the relevant standards. Both time and phase diversities are included in the

representative load models.

(ii) A method has been developed to estimate the 95th percentile cumulative

probability level of harmonic voltage distortion in an MV distribution system as

required by the present international harmonic standard IEC 61000-3-6. This

method includes techniques to overcome the difficulties in determining distortion

194
 levels when feeders are not loaded to their fullest extent, and allows for cases

when levels of background distortion may not be as high as limits set by the

standard. The method provides a useful tool to combine the effects of pragmatic

modelling of distribution loads at the planning stage or to evaluate distortion

levels of existing systems.

(iii) A generalised method has been developed to extend the IEC 61000-3-6 approach

of allocating allowable harmonic emissions to the case where customers are

distributed along an MV distribution system feeder having significantly different

fault levels. This is an area that IEC 61000-3-6 fails to cover adequately to allow

application to real systems. The method involves the determination of an

'allocation constant' using the agreed loading of all customers and the system

harmonic impedances.

(iv) The approach of allocating emissions to customers described in (iii) typically

requires an extensive amount of data. This data may not always be available to the

DNSP engineer. An extension to the method of (iii) has thus been completed to

cater for the complex situation where only limited data is available. This is

achieved by looking at several extreme cases which categorise the most common

MV distribution system feeder configurations, and through the use of correction

factors for the 'allocation constant'.

(v) A harmonic monitoring programme has been designed and implemented on a

typical MV distribution system to establish parameters for the three types of

statistical harmonic load models in (i), and also to confirm the relative accuracy of

distortion level prediction techniques mentioned in (ii).

(vi) A preliminary look at the effect of power factor correction capacitance on (i)-(iv)

has shown that the presence of capacitors usually requires detailed analysis for

195
 effective harmonic management. The approach of IEC 61000-3-6 of using 'twice

the impedance' is illustrated to be a useful tool when resonances do not occur at

problematic harmonics.

(vii) Example systems have been studied to illustrate the harmonic management tools

described in (i)-(iv). This includes identification of the key indicators for a

distribution system's harmonic performance.

196
 Publications based on worked performed on this thesis

I. Robinson, D., V. Gosbell and S. Perera. Harmonic allocation constant for

implementation of AS/NZS 61000.3.6. in Proc. of Australasian Universities Power
Engineering Conf. (ISBN 1 76067 068 X). 23-26 September 2001. Perth. p. 142-
147.

II. Robinson, D., V. Gosbell, S. Perera and A. Baitch. Application of Australia's new
harmonic standard to a long MV feeder. in Proc. of 6th Inter. Transmission and
Distribution Conf. 11-14 November 2001. Brisbane. Paper No.1.

III. Robinson, D., V. Gosbell, S. Perera and A. Baitch. Application of Australia's new
harmonic standard to a long MV feeder. Journal of Electrical & Electronics
Engineering Australia, 2002. Vol.21(No.3): p. 1-9.

IV. Gosbell, V., D. Robinson and S. Perera. The application of IEC 61000-3-6 to MV
systems in Australia. in Proc. of ERA Technology Inter. Conf. on Quality and
Security of Electrical Supply. February 2001. Thames. p. 7.1.1-7.1.10.

V. Robinson, D., V. Gosbell, S. Perera and D. Mannix. Establishment of typical

harmonic voltage levels in radial distribution systems. in Proc. of 9th Inter. Conf.
on Harmonics and Quality of Power. 1-4 October 2000. Orlando. Vol.3: p. 1053-
1058.

VI. Gosbell, V., D. Mannix, D. Robinson and S. Perera. Harmonic survey of an MV

distribution system. in Proc. of Australasian Universities Power Engineering
Conf. (ISBN 1 76067 068 X). 23-26 September 2001. Perth. p. 338-343.

VII. Robinson, D., V. Gosbell, S. Perera, N. Browne, Harmonics and flicker in a sub-
transmission substation: A measurement experience. in Proc. of 7th Inter. Energy
Transmission and Distribution Conf. 16-19 November 2003. Adelaide. Paper
No.198.

197
 References

1. Owen, E.L., A history of harmonics in power systems. IEEE Industry Applications

Magazine, 1998. Vol.4(Issue 1): p. 6-12.

2. Phipps, J.K., J.P. Nelson, and P.K. Sen, Power quality and harmonic distortion on
distribution systems. IEEE Trans. on Industry Applications, 1994. Vol. 30(No. 2):
p. 476-484.

3. Martinon, J., et al., A new statistical approach of harmonic propagation in

transmission systems. IEEE Trans. on Power Delivery, 1996. Vol.11(No.2): p.
1032-1040.

4. Mahmoud, A., R. Owen, and T. Ortmeyer, Power system harmonics: an overview.

IEEE Trans. on Power Apparatus and Systems, 1983. Vol. 102(No. 8): p. 2455-
2460.

5. Eurelectric, Power Quality in European electricity supply networks. 2002, NE

Standardisation Specialist Group. p. 11.

6. Lundquist, J., On harmonic distortion in power systems. 2001, Chalmers

University of Technology: Göteborg.

7. Wagner, V.E., J.C. Belda, and T.M. Barnes, Effects of harmonics on equipment.
IEEE Trans. on Power Delivery, 1993. Vol.8(No.2): p. 672-680.

8. Chen, W. and Z. Cheng, An experimental study of the damaging effects of

harmonics in power networks on the capacitor dielectrics. Proc. of Second Inter.
Conf. on Properties and Applications of Dielectric Materials, 1988. Vol.2: p. 645-
648.

9. Arseneau, R. and M. Ouelette, The effects of supply harmonics on the

performance of compact fluorescent lamps. IEEE Trans. on Power Delivery,
1993. Vol. 8(No. 2): p. 473-479.

10. Fourier, J.B., Théorie Analytique de la Chaleur. 1822.

198
11. Dugan, R.C., et al., Electrical power systems quality. 2002, New York: McGraw
Hill.

12. AS 2279.2, Disturbances in mains supply networks - Part 2: Limitation of

harmonics caused by industrial equipment. Australian Standards, 1991.

13. IEEE Std 519, IEEE Recommended practices and requirements for harmonic
control in electrical power systems. IEEE Industry Applications Society/Power
Engineering Society, 1992.

14. POLICY SG-007-1, Adoption of international standards; Standardization Guide

No.7.1. Australian Standards, 2003.

15. IEC 61000-3-6, Electromagnetic compatibility (EMC) Part 3: Limits - Section 6:

Assessment of emission limits for distorting loads in MV and HV power systems -
Basic EMC publication. International Electrotechnical Commission, 1996.

16. AS/NZS 61000.3.6, Electromagnetic compatibility (EMC) Part 3.6: Limits -

Assessment of emission limits for distorting loads in MV and HV power systems
(IEC 61000-3-6:1996, MOD). Australian Standards, 2001.

17. Gosbell, V., et al. Harmonic survey of an MV distribution system. in Australasian

Universities Power Engineering Conf. 2000. Brisbane.

18. Shuter, T.C., H.T. Vollkommer, and T.L. Kirkpatrick, Survey of harmonic levels
on the American Electric power distribution system. IEEE Trans. on Power
Delivery, 1989. Vol.4(No.4): p. 2204-2213.

19. IEC 61000-4-30, Electromagnetic compatibility (EMC) Part 4-30: Testing and
measurement techniques - Power quality measurement methods. International
Electrotechnical Commission, 2003.

20. IEC 61000-4-7, Electromagnetic compatibility (EMC) Part 4-7: Testing and
measurement techniques - Guide on harmonics and interharmonics measurements
and instrumentation, for power supply systems and equipment connected thereto.
International Electrotechnical Commission, 2002.

199
21. Emanuel, A.E., et al., A survey of harmonic voltages and currents at distribution
substations. IEEE Trans. on Power Delivery, 1991. Vol.6(No.4): p. 1883-1890.

22. Wilkie, R.G. and G.E. Dowd. Harmonic measurements in industrial applications.
in AUPEC'93. 1993. Wollongong.

23. Downs, S.P.C. and N. McGoldrick, Harmonic measurements on transmission

systems. IEE Colloquium (Digest), 1997(No.096): p. 4/1-4/8.

24. Govindarajan, S.N., M.D. Cox, and F.C. Berry, Survey of harmonic levels on the
Southwestern Electric Power Company system. IEEE Trans. on Power Delivery,
1991. Vol.6(No.4): p. 1869-1875.

25. Sabin, D.D., D.L. Brooks, and A. Sundaram, Indices for assessing harmonic
distortion from power quality measurements: Definitions and benchmark data.
IEEE Trans. on Power Delivery, 1999. Vol.14(No.2): p. 489-496.

26. IEEE Draft Std P519A/D5, Guide for applying harmonic limits on power systems.
Harmonics Working Group (IEEE PES T&D Committee), 1996.

27. Gosbell, V.J., B.S.P. Perera, and H.M.S.C. Herath. New framework for utility
power quality (PQ) data analysis. in Australasian Universities Power
Engineering Conf. 2001. Perth.

28. Emanuel, A.E., et al., A survey of harmonic voltages and currents at the
customer's bus. IEEE Trans. on Power Delivery, 1993. Vol.8(No.1): p. 411-421.

29. Wu, C.J., et al., Application of regression models to predict harmonic voltage and
current growth trend from measurement data at secondary substations. IEEE
Trans. on Power Delivery, 1998. Vol.13(No.3): p. 793-799.

30. Emanuel, A.E., et al., Voltage distortion in distribution feeders with nonlinear
loads. IEEE Trans. on Power Delivery, 1994. Vol.9(No.1).

31. Gosbell, V.J., et al., National power quality survey of Australian electricity
distributors; A pilot study. 2002, Australian power quality service centre,
University of Wollongong.

200
32. Morrison, R.E. and E. Duggan, Long term monitoring of power systems
harmonics. IEE Colloquium (Digest), 1993(Issue 120): p. 2/1-2/3.

33. Duggan, E. and R.E. Morrison, Prediction of harmonic voltage distortion when a
nonlinear load is connected to an already distorted supply. IEE Colloquium
(Digest), 1993. Issue 140(No.3).

34. Lachaume, J., T. Deflandre, and M. Meunier. Harmonics in MV and LV

distribution systems present and future levels. in IEE Disturbances and Protection
in Supply Systems. 1993. Stevenage.

35. Hu, C.H., et al., Survey of harmonic voltage and current distribution substation in
Northern Taiwan. IEEE Trans. On Power Delivery, 1997. Vol.12(No.3): p. 1275-
1284.

36. Tolbert, L.M., H.D. Hollis, and P.S. Hale. Survey of harmonics measurements in
electrical distribution systems. in IEEE Industry Applications Conf. 1996.

37. Mohan, N., T.M. Undeland, and W.P. Robbins, Power electronics: Converters,
applications, and design. 2nd ed. 1995, New York: John Wiley & Sons.

38. Staats, P.T., et al., A statistical method for predicting the net harmonic currents
generated by a concentration of electric vehicle battery chargers. IEEE Trans. on
Power Delivery, 1997. Vol.12(No.3).

39. Gorgette, F.A., J. Lachaume, and W.M. Grady. Statistical summation of the
harmonic currents produced by a large number of single phase variable speed air
conditioners: A study of three specific designs. in 8th Inter. Conf. on Harmonics
and Quality of Power. 1998. Athens.

40. Emanuel, A.E., et al., Distribution feeders with nonlinear loads in the Northeast
U.S.A.: Part 1 - Voltage Distortion Forecast. IEEE Trans. on Power Delivery,
1995. Vol.10(No.1).

41. Mansoor, A. and W.M. Grady. Analysis of compensation factors influencing the
net harmonic current produced by single-phase non-linear loads. in 8th Inter.
Conf. on Harmonics and Quality of Power. 1998. Athens.

201
42. EN 50160, Voltage characteristics of electricity supplied by public distribution
systems. CENELEC, 1994.

43. Xu, W. Comparisons and comments on harmonic standards IEC 61000-3-6 and
IEEE Std. 519. in 9th Inter. Conf. on Harmonics and Quality of Power. 2000.
Orlando.

44. Gosbell, V.J., P. Muttik, and D.K. Geddey. A review of the new Australian
harmonics standard AS/NZS 61000.3.6. in Australasian Universities Power
Engineering Conf. 1999. Darwin.

45. Gosbell, V.J., et al., Power quality - recommendations for the application of
AS/NZS 61000.3.6 and AS/NZS 61000.3.7. Standards Australia, HB 264-2003,
2003. ISBN 0 7337 5439 2.

46. Electricity Council, Supplies to converter equipment, harmonic distortion and

permissible pulse number of consumers' rectifiers and invertors. Electricity
Council recommendation G5/2, 1967.

47. Corbyn, D.B., This business of harmonics. IEE Electronics & Power, 1972.
Vol.18: p. 219-223.

48. Sherman, W.G., Summation of harmonics with random phase angles. IEE Proc.,
1972. Vol.119(No.11): p. 1643-1648.

49. Rowe, N.B., The summation of randomly-varying phasors or vectors with

particular reference to harmonic levels. IEE Conference Publication 110, 1974: p.
177-181.

50. Crucq, J.M. and A. Robert. Statistical approach for harmonics measurements and
calculations. in CIRED. 1989.

51. Robert, A., Emission limits for distorting loads in MV and HV power systems. EC
77A, 1993.

52. Mansoor, A., et al., Effect of supply voltage harmonics on the input current of
single-phase diode bridge rectifier loads. IEEE Trans. on Power Delivery, 1995.
Vol.10(No.3): p. 1416-1422.
202
53. Hansen, S. Harmonic cancellation by mixing non-linear single phase and three-
phase loads. in IEEE Industry Application Conf. 1998.

54. Nejdawi, I.M., et al., Harmonics trend in NE USA: A preliminary survey. IEEE
Trans. On Power Delivery, 1999. Vol.14(No.4): p. 1488-1494.

55. PSCAD/EMTDC Version 4.0.1. 2003, Manitoba HVDC Research Centre.

56. PSPICE Version 9.2.3. 2003, OrCAD.

57. SIMULINK Release 13. 2003, Mathworks Inc.

58. SUPERHARM Version 4.2.0.3. 2000, Electrotek Concepts Inc.

59. AS/NZS 61000.4.7, Electromagnetic compatibility (EMC) Part 4.7: Testing and
measurement techniques - General guide on harmonics and interharmonics
measurements and instrumentation, for power supply systems and equipment
connected thereto. Australian Standards, 1999.

60. Gosbell, V., D. Robinson, and S. Perera. The Application of IEC 61000-3-6 to MV
Systems in Australia. in ERA Technology Inter. Conf. on Quality and Security of
Electrical Supply. 2001. Thames.

61. Robinson, D., S. Perera, and V. Gosbell. Waveform Generator for Load
Susceptibility Testing. in Australasian Universities Power Engineering Conf.
1999. Darwin.

62. EDMI, Users Manual - EDMI 2000-04XX Energy Meter. Electronic Design and
Manufacturing International Pty. Ltd, 1997.

63. Gosbell, V.J. and G.J. Sanders. Frequency response of distribution CTs. in
Australasian Universities Power Engineering Conf. 1996. Melbourne.

64. Cavallini, A., et al. Analysis and modeling of harmonic pollution due to consumer
electronics. in 8th Inter. Conf. on Harmonics and Quality of Power. 1998. Athens.

65. Robert, A. and T. Deflandre, Guide for assessing the network harmonic
impedance. ELECTRA, 1996(No.167): p. 97-131.

203
66. Arrilaga, J., et al., Power system harmonic analysis. 1997: John Wiley & Sons.

67. Kimbark, E.W., Electrical transmission of power and signals. 1955, New York:
John Wiley & Sons.

68. Wakileh, G.J., Power systems harmonics: Fundamentals, analysis and filter
design. 2001, Berlin: Springer.

69. Oiu, M., V.K. Sood, and H. Jin. Evaluation of harmonic impedance presented by
a thyristor controlled AC load. in WESCANEX'95 Communications, Power and
Computing Conf. 1995: IEEE.

70. Robinson, D., et al. Harmonics and flicker levels at a sub-transmission substation:
A measurement experience. in Distribution 2003. 2003. Adelaide.

71. Tong, D., et al. Harmonic propagation in transmission system with multiple
capacitor installations. in PowerCon 2000. 2000.

72. Gonen, T., Electrical power distribution engineering. 1986: McGraw-Hill.

73. Bridgeman, M.J., R.E. Morrison, and T.B. Tennakoon. Measurement of harmonic
impedance on an LV system utilising power capacitor switching and consequent
predictions of capacitor induced harmonic distortion. in 8th Inter. Conf. on
Harmonics and Quality of Power. 1998. Athens.

74. Kocatepe, C. The effects of reactors and capacitors connected to power systems
on the harmonic load flow. in MELECON'96. 1996. Bari.

204
 Appendix A

Fluke 41 Power Harmonics Analyser specifications

The Fluke 41 Power Harmonics Analyser (Serial 6050090) was used to perform

harmonic measurements on various domestic loads. Specifications for the Analyser are

listed below.

Information obtained from http://au.fluke.com/auen/products/41B.htm

205
 Appendix B

EDMI Mk3 Energy Meter specifications

The Mk3 revenue meter was used as the primary instrument for the harmonic
monitoring programme on the Homepride zone distribution system. Key Features of the
meter include; four quadrant energy measurement, time of use, load profile, pulse
inputs/outputs, high accuracy, remote reading, waveform capture, sag and swell analysis
(half cycle resolution), quality of supply (on board harmonic analysis) [62].

Information obtained from http://www.edmi.com.au/products/powerquality/mk3.cfm

206
 Appendix C

Results of the Homepride harmonic monitoring programme

Residential Substation (Site 5) Annual Voltage Trend
3
VTHD 2001
V3 2001
V5 2001
2.5
V7 2001
95 Percentile Distortion (%)

VTHD 2002
2 V3 2002
V5 2002
V7 2002
1.5

1
th

0.5

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.1 Residential substation 95th percentile annual harmonic voltage trend
Commercial Substation (Site 6) Annual Voltage Trend
5
VTHD 2001
4.5 V3 2001
V5 2001
4 V7 2001
95 Percentile Distortion (%)

VTHD 2002
3.5
V3 2002
3 V5 2002
V7 2002
2.5

1.5
th

0.5

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.2 Commercial substation 95th percentile annual harmonic voltage trend
Industrial Substation (Site 7) Annual Voltage Trend
3
V THD 2001
V 3 2001
V 5 2001
2.5
V 7 2001
95 Percentile Distortion (%)

V THD 2002
2 V 3 2002
V 5 2002
V 7 2002
1.5

1
th

0.5

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.3 Industrial substation 95th percentile annual harmonic voltage trend

207
 Residential Substation (Site 5) Annual Current Trend
10
I3 2001
9 I5 2001
I7 2001
8 I3 2002

95 Percentile Current (A)

I5 2002
7
I7 2002

3
th

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.4 Residential substation 95th percentile annual harmonic voltage trend
Commercial Substation (Site 6) Annual Current Trend
100
I3 2001
90 I5 2001
I7 2001
80 I3 2002
95 Percentile Current (A)

I5 2002
70
I7 2002

60

50

40

30
th

20

10

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.5 Commercial substation 95th percentile annual harmonic voltage trend
Industrial Substation (Site 7) Annual Current Trend
300
ITHD 2001
I3 2001
I5 2001
250
I7 2001
95 Percentile Current (A)

ITHD 2002
200 I3 2002
I5 2002
I7 2002
150

100
th

50

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.6 Industrial substation 95th percentile annual harmonic voltage trend

208
 Residential Feeder (Site 2) Annual Current Trend
7
I3 2001
I5 2001
6 I7 2001
I3 2002

95 Percentile Current (A)

I5 2002
5
I7 2002
I3 2003
4 I5 2003
I7 2003
3

2
th

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.7 Residential feeder 95th percentile annual harmonic current trend
Commercial Feeder (Site 3) Annual Current Trend
7
I3 2001
I5 2001
6 I7 2001
I3 2002
95 Percentile Current (A)

I5 2002
5
I7 2002
I3 2003
4 I5 2003
I7 2003
3

2
th

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.8 Commercial feeder 95th percentile annual harmonic current trend
Industrial Feeder (Site 4) Annual Current Trend
20
ITHD 2001
18 I3 2001
I5 2001
16 I7 2001
95 Percentile Current (A)

ITHD 2002
14
I3 2002

12 I5 2002
I7 2002
10 ITHD 2003
I3 2003
8 I5 2003
I7 2003
6
th

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure C.9 Industrial feeder 95th percentile annual harmonic current trend

209
 Appendix D

Calculation of maximum demand for Customer A

The following is the calculation of the maximum demand Si for Customer A installation

introduced in Section 7.3. Refer to Figure 7.5 for assumed data.

Real power of the drive, Pdrive = 530kW

= 0.53pu

Complex power of the drive, Sdrive = Pdrive ÷ pfdrive

= 0.53pu÷0.7

= 0.757pu

Reactive power of the drive, Qdrive = 0.757 2 − 0.532

= 0.541pu

Total complex power, Sload = (0.53 + 0.05) 2 + 0.5412

= 0.793pu

Transformer reactive power, Qtrans = (0.793)2×0.05

= 0.031pu

Customer A maximum demand, Si = (0.53 + 0.05) 2 + (0.541 + 0.031) 2

= 0.815pu

210
 Appendix E

Calculation of acceptable emissions for Customer A

E.1 IEC 61000-3-6 Stage 1

Assessment of Customer A’s installation under Stage 1 of IEC 61000-3-6 requires only

a small amount of data from the customer. The data utilised in the following

calculations is provided in Figure 7.5 of Chapter 7. The maximum demand of individual

customer, Si, at the PCC for Customer A is calculated using a base of 1MVA as 0.815pu

(see appendix D). The short circuit level, SSC, is 44.0pu as indicated in Figure 7.5. As

the distorting load in Customer A’s installation contains a six pulse converter feeding a

dc drive a weighting factor of 1.0 (from Table 6 in IEC 61000-3-6) is also applied to the

proportion of Si which is non-linear.

SDwi = 1.0 × 0.757 (E.1)

Thus the total load is 0.815pu as before. A customer can obtain automatic acceptance

under Stage 1 of IEC 61000-3-6 if the load is of relatively small size, i.e. it satisfies

equation (E.2), or if the load is under 2MVA the relative harmonic currents of the total

load are less than that provided in Table 7 of IEC 61000-3-6.

S i / S SC ≤ 0.1% (E.2)

For Customer A’s installation Si and SSC are 0.815pu and 44pu respectively. Thus using

(E.2)

S i / S SC = 0.851 / 44
= 1.85%

Therefore Customer A’s installation fails the Stage 1 test.

211
E.2 IEC 61000-3-6 Stage 2 – First Approximation

At saturation of the system, taking into account overload capacity and maximum

utilisation factor of the 25MVA HV/MV zone substation transformer, the total power

available (St) is 27.5MVA. The X:R ratio of the system impedance seen by Customer A

is assumed to be 1:1. Table E.1 outlines the results from Stage 2 first approximation

calculations with the recommended allowable harmonic current emission limits given

relative to the PCC and the customer load in the right most columns.

Table E.1 Stage 2 first approximation results

Harmonic α LhMV LhHV GhMV+LV EUhi% Zh(pu) EIhi (pcc)% EIhi (load)%
3 1 4.00% 2.00% 2.00% 0.06% 0.05 1.22% 1.50%
5 1.4 5.00% 2.00% 3.97% 0.32% 0.08 3.98% 4.89%
7 1.4 4.00% 2.00% 2.85% 0.23% 0.11 2.04% 2.51%
9 1.4 1.20% 1.00% 0.41% 0.03% 0.14 0.23% 0.28%
11 2 3.00% 1.50% 2.60% 0.45% 0.18 2.52% 3.10%
13 2 2.50% 1.50% 2.00% 0.34% 0.21 1.64% 2.02%
17 2 1.60% 1.00% 1.25% 0.21% 0.27 0.79% 0.96%
19 2 1.20% 1.00% 0.66% 0.11% 0.31 0.37% 0.46%
23 2 1.20% 0.70% 0.97% 0.17% 0.37 0.45% 0.56%
25 2 1.20% 0.70% 0.97% 0.17% 0.40 0.42% 0.51%

The above calculations were completed with a system base of 1MVA. The base current

for the calculations is determined using equation (E.3).

1MVA
I base =
3 × 11kV (E.3)
= 52.5 A

The recommended allowable harmonic current emission limits in Amperes are given in

Table E.2. These limits may be increased slightly if the residential loading of HV/MV

zone substation transformer is significant (load curves of LV and MV are required for

this).

212
Table E.2 Harmonic current emission limits for Stage 2 first approximation
Harmonic EIhi (Amps)
3 0.64
5 2.09
7 1.07
9 0.12
11 1.32
13 0.86
17 0.41
19 0.20
23 0.24
25 0.22

213