Fulltext

Application of Optical Current
Transformers in Digital Substations
Ingvill Urdal Rian
Master of Energy and Environmental Engineering

Submission date: June 2018
Supervisor: Hans Kristian Høidalen, IEL
Co-supervisor: Magnus Kolgrov, Statnett SF
Rannveig Løken, Statnett SF
Norwegian University of Science and Technology

Department of Electric Power Engineering
Problem Description
Digitalization of substations will substantially reduce the required size of buildings and
the amount of copper wiring. The required space is reduced because digital relays can
contain more functions and the copper wiring is replaced by a process bus where a fiber
optical ring contains all signals. The engineering time can also be reduced, and faster
upgrade and renewal could be possible with more plug & play capabilities of the relays or
intelligent electronic devices (IEDs). Digital substations also open up for the application of
non-conventional instrument transformers (NCITs), such as optical current transformers
(OCTs), were current and voltages are measured with other principles than traditional
magnetic coupling. In order to integrate conventional instrument transformers into digital
substations, stand-alone merging units (SAMUs) are required, which digitize the currents
and voltages according to the IEC 61850 standard.
The following tasks are given:
• Describe a digital substation environment and the involved components.

• Study the technology used in optical current transformers and discuss error mech-
anisms, accuracy and specifications.
• Discuss benefits and drawbacks of applying optical current transformers in digital
substations compared with conventional technologies.
• Evaluate whether optical current transformers are suitable for application in digital
transmission substations.
• Study the response of the stand-alone merging unit SAM600 in the lab in terms
of steady state behavior and transient capabilities. Study possible implications on
relay response.
i
Preface
This thesis finalizes my master’s degree in Energy and Environmental Engineering at

the Norwegian University of Science and Technology (NTNU), Department of Electric
Power Engineering. The master’s thesis is a continuation of a specialization project on
digital substations, which was carried out in the fall semester of 2017. The master’s
thesis has been realized in cooperation with Statnett and with the guidance of Professor
Hans Kristian Høidalen at NTNU. The rapport is the result of work which was conducted
during the spring semester of 2018.
Trondheim, 22 Juni 2017
Ingvill Urdal Rian
iii
Acknowledgement
I would like to thank the following persons:
• My supervisor at NTNU, Hans Kristian Høidalen, for good guidance during the
work on this master’s thesis.
• My co-supervisors at Statnett, Magnus Kolgrov and Rannveig Løken, for their co-
operation and for answering questions.
• PhD students, Maciej Grebla and Jaya Raghavendra Arun Yellajosula, for helping
with the practical work and for giving good advice during the work process.
v
Abstract
In this master’s thesis optical current transformers are studied in order to evaluate whether
such sensors are suitable for application in digital transmission substations. The analysis
includes a discussion of benefits and drawbacks and an evaluation of performance, matu-
rity and fulfillment of substation application requirements as well as the digital substation
concept. Experimental testing of a stand-alone merging unit is also performed with the
objective of investigating its performance and comparing it to the performance of a con-
ventional analog acquisition system. Potential impacts on the performance of transformer
differential protection is also investigated.
The laboratory tests show that the stand-alone merging unit has a steady state perfor-
mance which is comparable to that of a conventional analog acquisition system and a
significantly better transient performance. As a result of the differing transient perfor-
mances, fundamental frequency differential currents arise when these two technologies are
combined in a hybrid mode configuration of a transformer differential protection. The
resulting fundamental frequency differential currents are not big enough to affect the
performance of the transformer differential protection in the tested cases.
Optical current transformers offer many benefits compared to conventional current trans-
formers. Some of these are improved safety, reduced environmental impact, reduced need
for maintenance and improvements in steady state and transient performance. The low
weight facilitates integration with other primary equipment, which allows space to be
saved in the substation yard. The combination of optical current transformers and a
process bus makes the design of digital substations flexible and facilitates cost reductions.
The standards which apply to non-conventional instrument transformers and digital sub-
stations often leave some room for interpretation. This may lead to lack of interoperability
between devices from different vendors in a digital substation. Competence building will
also be required, and digital substations may be expected to become more vulnerable
than the present day conventional substations due to outdoor placement of electronics
and stringent time synchronization requirements.
It is concluded that optical current transformers seems to be suitable for application in

digital transmission substations. However, there are several signs that the products and
solutions are not completely mature yet. There are several investigations to be made,
questions to be answered and issues to be solved before a migration from the present day
conventional substations with conventional current transformers to digital substations
with optical current transformers can take place.
vii
Sammendrag
I denne masteroppgaven studeres optiske transformatorer for å vurdere hvorvidt slike

sensorer egner seg til bruk i digitale transmisjonsstasjoner. Analysen omfatter en disku-
sjon av fordeler og ulemper og en evaluering av oppførsel, modenhet og oppfyllelse av
applikasjonskrav samt digital stasjon-konseptet. Testing av en stand-alone merging unit
utføres også for å undersøke enhetens oppførsel og sammenligne den med oppførselen til
et konvensjonelt analogt system. Potensiell innvirkning på responsen til et transformer-
differensialvern blir også undersøkt.
Laboratorietestene viser at stand-alone merging unit-en har en stasjonær oppførsel som

er sammenlignbar med den stasjonære oppførselen til et konvensjonelt analogt system og
en betydelig bedre transient oppførsel. Som et resultat av ulikhetene i de transiente opp-
førselene, oppstår differensialstrømmer når disse to teknologiene kombineres i en hybrid
konfigurasjon av et transformatordifferensialvern. De resulterende differensialstrømmene
er ikke store nok til å påvirke responsen til transformatordifferensialvernet i de testede
tilfellene.
Optiske strømtransformatorer gir mange fordeler sammenlignet med konvensjonelle strøm-

transformatorer. Noen av disse er forbedret sikkerhet, redusert miljøpåvirkning, redusert
behov for vedlikehold og forbedringer i stasjonær og transient oppførsel. Den lave vekten
legger til rette for integrasjon med annet primært utstyr, noe som fører til at plass kan
spares i anlegget. Kombinasjonen av optiske strømtransformatorer og prosessbuss gjør
utformingen av digitale stasjoner fleksibel og legger til rette for kostnadsreduksjoner.
Standardene som gjelder for ikke-konvensjonelle måletransformatorer og digitale stasjoner

gir ofte noe rom for tolkning. Dette kan føre til mangel på interoperabilitet mellom enheter
fra forskjellige leverandører i en digital stasjon. Kompetansebygging vil også kreves, og
digitale stasjoner kan forventes å bli mer sårbare enn dagens konvensjonelle stasjoner på
grunn av utendørs plassering av elektronikk og strenge tidssynkroniseringskrav.
Det konkluderes med at optiske strømtransformatorer synes å være egnet til bruk i digitale
transmisjonsstasjoner. Det er imidlertid flere tegn på at produktene og løsningene ikke er
helt modne ennå. Det er flere undersøkelser som skal gjøres, spørsmål som skal besvares
og problemer som skal løses før en overgang fra dagens konvensjonelle stasjoner med kon-
vensjonelle strømtransformatorer til digitale stasjoner med optiske strømtransformatorer
kan finne sted.
ix
Contents
1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.5 Structure of the Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Background 5
2.1 Conventional Substations . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Conventional Current Transformers . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Digital Substations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.1 Non-Conventional Instrument Transformers . . . . . . . . . . . . . 12
2.3.2 Merging Units and Stand-Alone Merging Units . . . . . . . . . . . 12
2.3.3 Switchgear Control Units . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.4 Process Bus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.5 Motivations Behind Digital Substations . . . . . . . . . . . . . . . 14
2.4 Standards and Implementation Guidelines . . . . . . . . . . . . . . . . . . 14
2.4.1 IEC 61850-9-2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.2 IEC 61850-9-2 LE . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.3 IEC 61869 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.4 IEC 60044-8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Regulations and Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3 Optical Current Transformers 19

3.1 Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1.1 Polarized Light . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1.2 Birefringence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.1.3 The Faraday Effect . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2 Detection Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2.1 Polarimetric Detection . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2.2 Interferometric Detection . . . . . . . . . . . . . . . . . . . . . . . 23
3.3 Types of Optical Current Transformers . . . . . . . . . . . . . . . . . . . . 24
3.3.1 Optical Fiber Current Sensors . . . . . . . . . . . . . . . . . . . . 24
3.3.2 Bulk Glass Optical Current Sensors . . . . . . . . . . . . . . . . . 29
3.4 Merging Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.5 Sources of Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
xi
3.5.1 Linear Birefringence . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.5.2 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.5.3 Vibration and Stray Magnetic Fields . . . . . . . . . . . . . . . . 33
3.5.4 Non-Ideal and Aging Optical Parts . . . . . . . . . . . . . . . . . 34
3.6 Performance Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.6.1 Linearity and Saturation . . . . . . . . . . . . . . . . . . . . . . . 35
3.6.2 Bandwidth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.6.3 Steady State Performance . . . . . . . . . . . . . . . . . . . . . . 37
3.6.4 Transient Performance . . . . . . . . . . . . . . . . . . . . . . . . 38
3.7 Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4 Commercial Products 41
4.1 Types of Optical Current Transformers . . . . . . . . . . . . . . . . . . . . 41
4.2 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.3 Interfaces and Time Synchronization . . . . . . . . . . . . . . . . . . . . . 42
4.4 Sensitivity to Environmental Disturbances . . . . . . . . . . . . . . . . . . 43
4.5 Expected Lifetime and Maintenance . . . . . . . . . . . . . . . . . . . . . 43
4.6 Weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.7 Price . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.8 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5 Stand-Alone Merging Units 45

5.1 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.2 Working Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.3 Steady State and Transient Performance . . . . . . . . . . . . . . . . . . . 46
6 Considerations in Digital Substations 49

6.1 Measurement Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
6.2 Requirements in Digital Substations . . . . . . . . . . . . . . . . . . . . . 50
6.2.1 Requirements of the Protection Systems . . . . . . . . . . . . . . . 50
6.2.2 Requirements of the Optical Current Transformers . . . . . . . . . 55
6.2.3 EMP and EMI Requirements . . . . . . . . . . . . . . . . . . . . . 55
6.3 Application of Differential Protection . . . . . . . . . . . . . . . . . . . . . 56
6.4 Application of Protection IEDs . . . . . . . . . . . . . . . . . . . . . . . . 58
6.5 Application of Energy Meters . . . . . . . . . . . . . . . . . . . . . . . . . 58
6.6 Application of Fiber Optic Cables . . . . . . . . . . . . . . . . . . . . . . 58
7 Laboratory Test Setup and Method 59

7.1 Test Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
7.1.1 OMICRON CMC 356 Test Set . . . . . . . . . . . . . . . . . . . . 60
xii
7.1.2 SAM600 Stand-Alone Merging Unit . . . . . . . . . . . . . . . . . 61
7.1.3 SEL-2488 Satellite-Synchronized Network Clock . . . . . . . . . . 62
7.1.4 RET670 Transformer Protection IED . . . . . . . . . . . . . . . . 62
7.1.5 Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
7.2 ATPDraw Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
7.2.1 Basic Simulation Model . . . . . . . . . . . . . . . . . . . . . . . . 65
7.2.2 Simulation Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
7.3 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
7.3.1 Steady State Performance . . . . . . . . . . . . . . . . . . . . . . 72
7.3.2 Transient Performance . . . . . . . . . . . . . . . . . . . . . . . . 73
8 Laboratory Results 75
8.1 Steady State Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
8.1.1 Amplitude . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
8.1.2 Phase Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.2 Transient Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.2.1 Faults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.2.2 Current Transformer Saturation . . . . . . . . . . . . . . . . . . . 84
8.2.3 Transformer Inrush . . . . . . . . . . . . . . . . . . . . . . . . . . 85
8.3 Response of Transformer Differential Protection . . . . . . . . . . . . . . . 87
9 Discussion of Laboratory Tests 89

9.1 Steady State Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
9.2 Transient Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
9.2.1 Faults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
9.2.2 Current Transformer Saturation . . . . . . . . . . . . . . . . . . . 97
9.2.3 Transformer Inrush . . . . . . . . . . . . . . . . . . . . . . . . . . 97
9.3 Performance of Transformer Differential Protection . . . . . . . . . . . . . 99
10 Conclusion of Laboratory Tests and Future Work 101

10.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
10.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
11 Discussion 103
11.1 Health, Safety and Environment . . . . . . . . . . . . . . . . . . . . . . . 103
11.2 Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
11.3 Substation Layout and Footprint . . . . . . . . . . . . . . . . . . . . . . . 104
11.4 Installation and Replacement . . . . . . . . . . . . . . . . . . . . . . . . . 105
11.5 Maintenance and Expected Lifetime . . . . . . . . . . . . . . . . . . . . . 106
11.6 Interoperability and Compatibility . . . . . . . . . . . . . . . . . . . . . . 106
xiii
11.7 Maturity and Availability of Products . . . . . . . . . . . . . . . . . . . . 107
11.8 Vulnerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
11.9 Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
11.10 Hybrid Mode Differential Protection . . . . . . . . . . . . . . . . . . . . . 110
11.11 Fulfillment of Application Requirements . . . . . . . . . . . . . . . . . . . 111
11.12 Supported Functionality in Digital Substations . . . . . . . . . . . . . . . 111
11.13 Fulfillment of the Digital Substation Concept . . . . . . . . . . . . . . . . 112
12 Conclusion and Future Work 113

12.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
12.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
A E-Mail 127
A.1 Statnett . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
A.1.1 First E-Mail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
A.1.2 Second E-Mail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
A.1.3 Third E-Mail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
A.2 ABB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
A.3 Arteche . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
A.3.1 First E-Mail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
A.3.2 Second E-Mail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
A.4 GE Grid Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
B Formulas 139
B.1 Supply Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
B.2 Current Transformers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
B.3 Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
C MATLAB Scripts 143

C.1 Fundamental Frequency Phase Current . . . . . . . . . . . . . . . . . . . . 143
C.2 Fundamental Frequency Differential Current . . . . . . . . . . . . . . . . . 145
C.3 Second Harmonic Component . . . . . . . . . . . . . . . . . . . . . . . . . 146
xiv
List of Figures
2.1 Layout of a conventional substation. . . . . . . . . . . . . . . . . . . . . . 6

2.2 Simplified equivalent circuit and phasor diagram of a CCT. . . . . . . . . . 8
2.3 Simplified magnetization curve for a CCT. . . . . . . . . . . . . . . . . . . 8
2.4 Current transformer saturation. . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5 Connection of a CCT to a protection IED. . . . . . . . . . . . . . . . . . . 10
2.6 Layout of a digital substation. . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1 Polarized light. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.2 The Faraday effect. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.3 Basic polarimetric optical fiber current sensor. . . . . . . . . . . . . . . . . 25
3.4 Dual polarimetric optical fiber current sensor. . . . . . . . . . . . . . . . . 25
3.5 Sagnac loop interferometer optical fiber current sensor. . . . . . . . . . . . 27
3.6 Reflective interferometer optical fiber current sensor. . . . . . . . . . . . . 28
3.7 Bulk glass optical current sensor. . . . . . . . . . . . . . . . . . . . . . . . 29
3.8 Sensitivity of a Sagnac loop and a reflective interferometer optical fiber
current sensor to vibrations. . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.9 Linearity of an optical fiber current sensor compared to the linearity of a
CCT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.10 Frequency response of an optical fiber current sensor. . . . . . . . . . . . . 37
3.11 Measured accuracy of a reflective interferometer optical fiber current sensor. 38
3.12 Transient response of an optical fiber current sensor. . . . . . . . . . . . . 39
4.1 A commercially available OCT. . . . . . . . . . . . . . . . . . . . . . . . . 42
5.1 Connection of CITs to secondary IEDs. . . . . . . . . . . . . . . . . . . . . 45
6.1 Total fault clearance time in a digital substation. . . . . . . . . . . . . . . 51

6.2 Operate-restrain characteristic of transformer differential protection. . . . . 57
7.1 Transformer differential protection in hybrid mode configuration. . . . . . . 59

7.2 Laboratory test setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
7.3 Operate-restrain characteristic of RET670 transformer differential protection. 63
7.4 Possible time synchronization solutions. . . . . . . . . . . . . . . . . . . . . 64
7.5 ATPDraw basic simulation model and simulation model of simulation case
1: normal load conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
7.6 ATPDraw simulation model of simulation case 2: internal high voltage side
fault. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
xv
7.7 ATPDraw simulation model of simulation case 3: internal low voltage side
fault. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
7.8 ATPDraw simulation model of simulation case 4: external fault. . . . . . . 70
7.9 ATPDraw model of three saturable CTs. . . . . . . . . . . . . . . . . . . . 70
7.10 Magnetization curves of the modelled CTs. . . . . . . . . . . . . . . . . . . 71
7.11 ATPDraw simulation model of simulation case 6: transformer inrush. . . . 71
8.1 Steady state response of the SAMU and the analog acquisition system. . . 76
8.2 Average amplitude error of phase A. . . . . . . . . . . . . . . . . . . . . . . 77
8.3 Average amplitude error of phase B. . . . . . . . . . . . . . . . . . . . . . . 78
8.4 Average amplitude error of phase C. . . . . . . . . . . . . . . . . . . . . . . 79
8.5 Average peak values of Irec_samu . . . . . . . . . . . . . . . . . . . . . . . . . 80
8.6 Average phase difference between the phase angles of Irec_samu and Irec_analog . 81
8.7 Transient response of the SAMU and the analog acquisition system to fault
currents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
8.8 Transient response of the SAMU and the analog acquisition system to fault
currents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
8.9 Transient response of the SAMU and the analog acquisition system to
saturated fault currents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
transformer inrush currents. . . . . . . . . . . . . . . . . . . . . . . . . . . 85
transformer inrush currents. . . . . . . . . . . . . . . . . . . . . . . . . . . 86
9.1 Fundamental frequency phase currents of the recorded currents from the
SAMU and the analog acquisition system in Figure 8.7. . . . . . . . . . . . 93
9.2 Fundamental frequency phase currents of the recorded currents from the
SAMU and the analog acquisition system in Figure 8.8. . . . . . . . . . . . 94
9.3 Fundamental frequency differential phase currents from the external fault
simulation case. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
9.4 Second harmonic phase currents of the recorded currents from the SAMU
and the analog acquisition system in Figure 8.10. . . . . . . . . . . . . . . 98
xvi
List of Tables
4.1 Overview of the specifications of some commercially available OCTs. . . . . 44
7.1 Settings of RET670 transformer differential protection. . . . . . . . . . . . 63

7.2 Ratings and connections of the power transformer. . . . . . . . . . . . . . . 66
7.3 Short-circuit test results for the power transformer. . . . . . . . . . . . . . 67
7.4 No-load test results for the power transformer. . . . . . . . . . . . . . . . . 67
8.1 RMS values of phase A of Irec_samu and Irec_analog . . . . . . . . . . . . . . . 80

8.2 Response of the transformer differential protection. . . . . . . . . . . . . . 87
B.1 Supply grid data at different short-circuit capacities. . . . . . . . . . . . . 140

B.2 Resistance of the load at different short-circuit capacities. . . . . . . . . . . 141
xvii
List of Abbreviations
ALF Accuracy Limit Factor
CCT Conventional Current Transformer
CIT Conventional Instrument Transformer
CT Current Transformer
CVT Conventional Voltage Transformer
DFT Discrete Fourier Transform
EMI Electromagnetic Interference
EMP Electromagnetic Pulse
GNSS Global Navigation Satellite System
GOOSE Generic Object Oriented Substation Event
HMI Human Machine Interface
HR High Remanence
HSR High-Availability Seamless Redundancy
HV High Voltage
I/O Input/Output
IED Intelligent Electronic Device
LR Low Remanence
LV Low Voltage
MMS Manufacturing Message Specification
MU Merging Unit
NCC Network Control Center
xix
NCCT Non-Conventional Current Transformer
NCIT Non-Conventional Instrument Transformer
NCVT Non-Conventional Voltage Transformer
NR Non Remanence
NTNU Norwegian University of Science and Technology
NTP Network Time Protocol
OCT Optical Current Transformer
PPS Pulse Per Second
PRP Parallel Redundancy Protocol
PTP Precision Time Protocol
RMS Root-Mean-Square
SAMU Stand-Alone Merging Unit
SAS Substation Automation System
SCU Switchgear Control Unit
SLED Superluminescent Diode
SNR Signal-to-Noise Ratio
SV Sampled Values
TSO Transmission System Operator
xx
Chapter 1
Introduction
1.1 Motivation
There is a growing interest in digitalization within almost every industry, including the
electric power industry. Several Transmission System Operators (TSOs) in different coun-
tries have initiated the work to digitize their substations and are presently working on
or have already implemented pilot digital substations [1], [2], [3], [4]. This includes the
Norwegian TSO, Statnett, which has an ongoing research and development project where
a small scale digital substation has been implemented as a part of an existing substation
for testing purposes [4]. The main goal is to achieve a substation solution with improved
safety, reduced costs and a reduction in the time spent on the building and upgrading
process [5].
The migration from the present day substations to digital substations opens up the possi-
bility to integrate Non-Conventional Instrument Transformers (NCITs), which are based
on other measuring principles than the conventional iron core transformer principle [6],
[7]. One type of NCIT is the Optical Current Transformer (OCT), which is promised to
bring significant improvements compared to Conventional Current Transformers (CCTs).
This applies to, among other things, the accuracy, dynamic range, transient performance,
safety and size of such sensors [8].
Statnett has installed an OCT in their pilot digital substation in addition to a Stand-Alone
Merging Unit (SAMU) [4], which allow existing Conventional Instrument Transformers
(CITs) to be integrated into digital substations [6]. It is thus of interest to gain a deeper
understanding of OCTs in order to evaluate whether such sensors are suitable for appli-
cation in digital substations in the transmission network. Since SAMUs are likely to be
applied in future digital substations, it is also relevant to perform experimental testing
of such a device in order to investigate its performance and potential issues related to its
application in digital substations.
1
CHAPTER 1. INTRODUCTION
1.2 Objectives
The main objective of this thesis is to evaluate whether OCTs are suitable for application
in digital substations in the transmission network. This includes answering the following
main research questions.
• What are the benefits and drawbacks of applying OCTs in digital transmission
substations compared to the use of CCTs in conventional transmission substations?
• Are there any limitations in terms of performance which may prevent the application
of OCTs in digital transmission substations?
• Are the products and solutions mature enough?
• Do OCTs meet the requirements of transmission substation applications?
• Will the overall solution fulfill the digital substation concept?
In addition to the investigations on OCTs, a SAMU shall be tested in the laboratory

at the Norwegian University of Science and Technology (NTNU). The objectives of the
laboratory tests are as follows.
• Investigate the steady state and transient performance of a SAMU.

• Compare the steady state and transient performance of a SAMU to those of a
conventional analog acquisition system.
• Investigate whether a hybrid mode configuration could have an impact on the per-
formance of transformer differential protection.
1.3 Limitations
This thesis has been conducted in cooperation with Statnett. The focus is therefore put
on the application of OCTs in digital substations in the transmission network. Other
types of Non-Conventional Current Transformer (NCCT) have not been considered, nor
Non-Conventional Voltage Transformers (NCVTs). The author has chosen to go deeper
into OCTs instead of treating different types of NCITs briefly. There was no OCT present
to be tested. Thus, the study on OCTs was limited to a literature review. Experimental
tests of an actual OCT would indeed have enriched this thesis.
There were also limitations in time and in the laboratory test setup, which limited the
experimental testing of the SAMU. The phase accuracy of the SAMU could not be inves-
tigated due to lack of time synchronization between the output from the test set and the
output from the SAMU. In addition, the highest current which could be applied to the
2
SAMU was limited by the maximum current output of the test set. This was 6.4 times
the rated input current of the SAMU.
Only the current module of the SAMU was tested in the laboratory. The SAMU also
incorporates a voltage module, which could have been tested. In the laboratory test
setup, the SAMU was directly connected to the protection IED by a single fiber optic link,
representing an ideal process bus without any significant time delays or other impacts on
the digital output from the SAMU. The effects of an actual process bus incorporating
switches and other connected devices was thus not investigated.
1.4 Approach
This thesis may be divided into two parts: a literature review and product survey on
OCTs, and experimental testing of a SAMU.
The first part of this thesis is based on a literature review on the technology and applica-
tion of OCTs in digital substations in addition to a product survey of four commercially
available OCTs. The product survey was conducted by studying the available technical
documentation of the different products as well as contacting the vendors in order to get
answers to questions which were not covered by the technical documentation. During
the fall semester of 2017, the author visited one one Statnett’s transmission substations
and Statnett’s pilot digital substation. Experiences from these two visits and information
obtained from contact with Statnett have also provided a basis for the content of this
thesis.
The second part of this thesis is based on laboratory work which was performed in the
ProSmart lab at NTNU during the spring semester of 2018. The steady state and tran-
sient performance of a SAMU was experimentally tested. In order to test the transient
performance of the SAMU, simulations were created in ATPDraw to generate different
transient current waveforms.
1.5 Structure of the Report

The rest of this thesis is structured as follows.
Chapter 2 introduces background information on conventional substations, CCTs, digital

substations and relevant standards and regulations. This chapter is partially based on the
specialization project report, "Digital Substations in Transmission Networks" [9], which
was written by the same author during the fall semester of 2017.
3
Chapter 3 treats OCTs. This chapter starts with an introduction of some basic concepts
and continues with descriptions of common detection methods, different types of OCTs
and the associated Merging Unit (MU). The main sources of error and the characteristics
of OCTs in terms of performance and other attributes are also described.
Chapter 4 presents the results of the product survey on OCTs. In this chapter the at-
tributes of some of the OCTs which are commercially available at the present time are
presented and compared.
In Chapter 5 the working principle and performance of SAMUs are described.
In Chapter 6 some considerations regarding the use of OCTs and other NCITs as well as
SAMUs in digital substations are presented.
Chapter 7 describes the test setup and method which was used during the laboratory
tests of the SAMU.
Chapter 8 presents the most important results from the laboratory tests.
In Chapter 9 the results of the laboratory tests which are given in Chapter 8 are discussed.
In Chapter 10 some preliminary conclusions are drawn from the discussion of the labora-
tory tests in Chapter 9. This chapter also includes some suggestions for future laboratory
testing of SAMUs.
Chapter 11 provides a discussion of the benefits and drawbacks of applying OCTs in

digital substations compared to the use of CCTs in conventional substations. It is also
discussed whether OCTs are suitable for application in digital transmission substations.
In Chapter 12 the main research questions, which were presented in Section 1.2, are
answered, and final conclusions are drawn. This chapter also includes some suggestions
for future work.
4
Chapter 2
Background
In this chapter relevant background information is presented. Section 2.1 describes present
day transmission substations, hereafter referred to as conventional substations. Section
2.2 treats CCTs and Section 2.3 describes digital substations. In Section 2.4 and Section
2.5 some relevant standards and regulations are presented, respectively.
The text in Section 2.1, Section 2.3 and Subsection 2.4.2 is taken from the specialization
project report, “Digital Substations in Transmission Networks” [9], written by the same
author during the fall semester of 2017. Some adaptions have been made. Section 2.2 is
also partially based on the same specialization project report. However, additional parts
have been added.
2.1 Conventional Substations

Transmission substations play an important role in transmission networks by connecting
transmission lines and providing important functions, such as transformation of voltage
levels, voltage control, reactive power control and power flow control [10]. In “Energiloven”
§ 1-5, the transmission network is defined as follows.
“The transmission network comprises of facilities for the transmission of elec-

trical energy at voltage levels of at least 200 kV, and facilities at 132 kV which
are of major importance for the operation of these facilities. The transmission
network also includes facilities for electrical energy conversion, when the con-
version facility is directly connected to facilities for transmission as mentioned
in the first paragraph and transforms to a voltage level of at least 33 kV. ...”
[11] (Translated from Norwegian)
Transmission substations generally comprise of a lot of different equipment, which may be

divided into two categories: primary equipment and secondary equipment. The primary
equipment is high voltage equipment, such as switching equipment, instrument transform-
5
CHAPTER 2. BACKGROUND
ers, power transformers, reactors, capacitor banks, busbars, power cables and power lines.
The primary equipment is located in the substation yard and is typically organized in high
voltage bays. The secondary equipment comprises of low voltage devices for protection,
control and monitoring [12], [13], [14]. These devices are typically located in the control
room of a control building. Together with the communication systems, these secondary
devices constitute the Substation Automation System (SAS) [15]. SASs have three basic
functions: to monitor, protect and control the primary equipment in the substation [16],
[12].
The architecture of a SAS can be divided into three levels: the process level, the bay
level and the station level [17]. The process level represents the interface to the primary
process [15]. At this level, data is acquired and operations are executed in order to control
the primary process [16]. The process level includes equipment such as CITs and different
types of switching equipment [15].
The bay level comprises of Intelligent Electronic Devices (IEDs), such as protection IEDs
and bay controllers, which perform protection and control functions, respectively [17].
Other bay level devices are disturbance recorders, electrical energy meters and devices
used for quality metering. The station level incorporates the substation Human Machine
Interface (HMI) and gateways which enable the communication between the substation
and the Network Control Center (NCC). The station level provides the operators with
status data from the substation equipment for the purpose of monitoring and control [17].
Figure 2.1: Layout of a conventional substation [18]. © 2017 IEEE
6
The interface between the station level and the bay level is a communication network
referred to as the station bus, as shown i Figure 2.1. The station bus connects the bay
level devices to the station level devices, thereby enabling the communication between
the two levels as well as the communication between different bay level devices [12].
The connections between the process level equipment and the bay level devices, on the
other hand, are still hardwired, as shown in Figure 2.1. There is thus a large number
of copper cables running from the control building to the individual high voltage bays
in conventional substations [4]. These are typically laid in large concrete cable trenches.
The hardwired circuits comprise of the following [14].
• Analog circuits for currents and voltages acquired by CITs.

• Binary circuits for status information, such as position of switching equipment.
• Binary circuits for protection and control signals.
• AC and DC power supply circuits.
The primary process data is not digitized before it reaches the bay level. The secondary
devices perform the conversion from analog or binary signals to digital and vice versa.
2.2 Conventional Current Transformers

CCTs scale the primary power circuit currents down to a level which is safe and practical
to use as an input for protection, control and measuring devices and create a galvanic
separation between the primary power circuit and the secondary devices [19], [20]. CCTs
have a standard rated secondary current of 1 A or 5 A, as defined in IEC 61869-2 [21].
The standardization facilitates the use and interchange of CCTs and secondary devices
from different vendors in a substation [20].
High voltage CCTs are inductive transformers with a conductive bar as the primary wind-
ing and one or more secondary windings, which are wound around iron cores encircling
the primary bar [22]. A single Current Transformer (CT) core cannot satisfy the require-
ments of all of the applications which are needed in a substation. Due to the conflicting
requirements of various applications, such as measuring and protection, different types of
transformer cores have been designed specifically to meet the accuracy, dynamic range
and bandwidth requirements of these applications [23]. The performance of a CCT is
determined by the area and diameter of the core cross section as well as the magnetic
properties [24]. The iron core causes measurement inaccuracies, which arise as a con-
sequence of magnetizing current, flux leakage, magnetic saturation, eddy current losses
and hysteresis losses. By designing different types of cores for specific applications, the
measurement inaccuracies can be reduced to a minimum [25], [26].
7
Figure 2.2: Simplified equivalent circuit and phasor diagram of a CCT. The figure is
adapted from [27] and [28].
A simplified equivalent circuit of a CCT and the associated phasor diagram is shown in
Figure 2.2. As shown in the figure, the primary current reffered to the secondary side (Ip ’)
comprises of two components: the magnetizing current (Im ) and the secondary current
(Is ). Since the magnetizing current is not transformed to the secondary side of the CT,
the primary current (Ip ) will be transformed with some errors in the amplitude and phase.
These are referred to as the ratio error (ε) and the phase displacement (δ), respectively
[26].
Figure 2.3: Simplified magnetization curve for a CCT. The figure is adapted from [27].
A simplified magnetization curve for a CCT is shown in Figure 2.3. The value of the
magnetizing impedance (Zm ) in the equivalent circuit shown in Figure 2.2 is proportional
to the slope of the magnetization curve [27]. Under normal load conditions CCTs typ-
ically operate in the unsaturated region of the magnetization curve. In this region the
8
magnetizing impedance has a much higher value than the secondary circuit impedance
(Rct + Rb ). Hence, most of the secondary current flows through the secondary circuit.
Under faulted conditions, on the other hand, CCTs may go into saturation due to high
primary currents. In the saturated region of the magnetization curve, the value of the
magnetizing impedance is much smaller than in the unsaturated region. Most of the sec-
ondary current will flow through the magnetizing impedance for a period of time each
cycle. The result is a saturated secondary current, as shown by the blue current waveform
in Figure 2.4. The saturation causes a high transformation error [27], [29]. CT satura-
tion may for instance result in incorrect operation of differential protection functions or
differential protection functions not operating when required to [30].
Figure 2.4: Current transformer saturation [31]. © 2002 IEEE
Measuring applications require high accuracy for currents up to the nominal primary
current [32], [26]. In order to make the measurements as accurate as possible, measuring
CT cores are designed with smaller cross sections. The magnetizing current is thereby
minimized. At high current levels the measuring cores may saturate, which is beneficial
in order to protect the measuring devices from high current levels [24] [26]. Protective
applications, on the other hand, require a wide dynamic range and accurate measurements
of fault currents [32], [26]. In order to avoid saturation at high current levels, protection
CT cores are designed with larger cross sections [24].
CCTs for protection can roughly be divided into three types based on how the iron core
is designed, if there are air gaps in the core and if so, how big the air gaps are. The
three types are High Remanence (HR), Low Remanence (LR) and Non Remanence (NR).
These have no air gaps in the core, small air gaps in the core and large air gaps in the
core, respectively. For HR type CTs there is no limit specified for the remanent flux.
The remanent flux of LR type CTs can be up to 10 % of the saturation flux, and the
remanent flux limit for NR type CTs is set to zero [27]. CCTs with air gaps in the core
are less prone to saturation since the air gaps reduce the magnetizing inductance and the
remanence. However, the phase displacement and the ratio error will increase because a
larger fraction of the secondary current will flow through the magnetizing impedance. In
9
addition, such CTs cannot reproduce the DC component in transient primary currents
properly. The bigger the air gaps, the faster the DC component will be damped out [27],
[33].
CCTs may be equipped with up to eight CT cores. The CCTs in present day transmission
substations typically have protection cores which are assigned to separate protective appli-
cations in addition to separate measuring cores for energy metering and measuring. The
secondary circuits of CCTs are connected in series with the input elements of secondary
devices, as shown in Figure 2.5 [32].
In conventional transmission substations the secondary CT circuits are connected to the

back of the protection and control panels in the control room. With one multiple cored
CCT per phase of each feeder, there is a great number of copper cables running from
the high voltage bays to the control building [34]. There is also a risk that one of these
secondary CT circuits can be inadvertently opened. If this happens, the CT will saturate
due to the infinite secondary impedance. A very high voltage is induced across the primary
winding, which in turn causes a very high voltage across the secondary winding. As a
result, arcing and possibly explosion of the CT can occur [35], [26]
Figure 2.5: Connection of a CCT to a protection IED [36]. © 2011 IEEE
High voltage CCT typically have oil-paper or SF6 insulation. In the Norwegian transmis-
sion grid, oil-paper insulated CTs are often used. Since these CTs contain oil, inspections
have to be performed regularly were the oil level is checked. Instrument readings may
also be performed. At Statnett this inspection is carried out each month. It is estimated
that such an inspection takes about half an hour per CT [Appendix A].
CCT are quite heavy devices. As an example, for a system voltage of 420 kV, one vendor
offers a top-core oil-paper insulated CT with a weight of 920 kg and a top-core SF6 gas
insulated CT with a weight of 1700 kg [37]. CCTs have been in wide use for a long
time and have proven to be reliable and show a good stability over time [38], [39]. The
expected lifetime is typically 25-40 years [7]. When it comes to the price, a CCT rated for
10
a maximum system voltage of 420 kV, which has three protection cores and two measuring
cores, is estimated to cost up to 8 500 € [Appendix A].
2.3 Digital Substations

The term digital substation has no standard definition. However, in general, the term
refers to substations where most of the primary process data is converted to digital form
already in the substation yard, close to where it is measured. The exchange of those
measured values as well as commands and signals between the secondary and the primary
devices takes place via a communication network in accordance with the IEC 61850 stan-
dard. The communication network is referred to as the process bus [40], [4]. The process
bus replaces the copper cables used for measuring and control circuits, which run from the
control building to the separate high voltage bays in conventional substations. In digital
substations the only copper cables left between the control building and the high voltage
bays will be those used for AC and DC power supply circuits [41]. The layout of a digital
substation is shown in Figure 2.6, with the station bus and process bus marked in dark
and light blue, respectively.
Figure 2.6: Layout of a digital substation [18]. © 2017 IEEE
11
2.3.1 Non-Conventional Instrument Transformers
The use of NCITs in substations has so far been limited. Typically, NCITs do not provide
the standard analog 1 A/5 A or 110 V output required by conventional protection, control
and monitoring devices. The introduction of a standardized process bus simplifies the
integration of NCITs into substations [6]. Since NCITs typically require digital signal
processing, a digital signal is a natural choice of output from such sensors [7].
Unlike CITs, NCITs are not based on the iron core transformer principle. The limitations
of CITs due to the iron core can thus be overcome [7]. NCITs are based on electrical or non-
electrical measuring principles. Examples of NCITs which employ electrical measuring
principles are Rogowski coil current sensors and capacitive voltage dividers. OCTs use a
non-electrical measuring principle based on the Faraday effect [42]. In Chapter 3 OCTs
will be treated in further detail.
2.3.2 Merging Units and Stand-Alone Merging Units
Merging units provide an interface for the connection of NCITs or CITs to the process
bus [43]. Merging units sample the current and voltage measurements from instrument
transformers and distribute the digital output as Sampled Values (SV) on the process
bus in accordance with IEC 61850-9-2 LE. The secondary devices which are connected to
the process bus can subsequently subscribe to these SV [44]. In order for the secondary
devices to be able to align SV from different merging units, the SV have to be time
stamped by the merging units and synchronized in time [45], [33]. In general, there are
two types of merging units. These are referred to as Merging Unit (MU) and Stand-Alone
Merging Unit (SAMU), respectively. MUs are applied in conjunction with NCITs and are
typically integrated with the associated sensor electronics. SAMUs, on the other hand,
are used to interface CITs [4]. In Section 3.4 and Chapter 5, the associated MU of OCTs
and SAMUs, will be treated in further detail, respectively.
There are several reasons why SAMUs are likely to be used in future digital substations.
First of all, as described in Section 2.1, transmission networks also include voltage levels
which are lower than 200 kV. In the lower voltage parts of transmission substations, a
continuous use of CITs may be desirable due to lack of commercially available OCTs
rated for these voltage levels. This could for instance be on the Low Voltage (LV) side
of a power transformer. In this thesis OCTs from four different vendors are considered.
At the present time, the OCTs offered by these vendors do not cover the lowest voltage
levels which are included in the transmission network [46], [47], [48], [49]. In addition,
the migration from the present day conventional substations to digital substations will be
12
a gradual process, and CITs are likely to be replaced on a step-wise basis [50]. SAMUs
allow existing CITs to be integrated into digital substations [6], thereby enabling a gradual
transition to digital substations.
2.3.3 Switchgear Control Units
Conventional switching equipment is integrated into digital substations by the use of

Switchgear Control Units (SCUs). SCUs provide a binary interface for the connection of
conventional switching equipment to the process bus in such a way that the switchgear
can be monitored and controlled [43]. SCUs acquire information regarding the position
and status of the switching equipment and distribute this information on the process
bus according to IEC 61850-8-1. In addition, SCUs actuate the switchgear based on
commands and signals which are received from the protection and control IEDs over the
process bus [1].
2.3.4 Process Bus
The process bus is the communication network which replaces the traditional hardwired
point-to-point connections between the primary equipment and the secondary devices in
conventional substations. The communication network allows process level devices to
exchange time-critical information with bay level devices and vice versa [17]. The process
bus is realized by the use of fiber optic cables, in contrast to the copper cables which are
used for the measuring and control circuits in conventional substations [4]. The primary
process data is distributed on the process bus and may be accessed by all of the secondary
devices which are connected to this communication network [40], [4].
Process bus communication is primarily of three types: SV streams, Generic Object

Oriented Substation Event (GOOSE) messages and client-server communication based
on the Manufacturing Message Specification (MMS) protocol [51]. Sampled current and
voltage measurements from SAMUs and the associated MUs of NCITs are published to
the process bus and sent to subscribing secondary devices as SV in accordance with
IEC 61850-9-2 [52]. By the use of multicast messages, the same SV can be sent to
several secondary devices simultaneously [41]. Status data and alarm information from
the primary process are also transmitted to the secondary devices, typically as GOOSE
messages in accordance with IEC 61850-8-1. Trip signals and open/close commands are
sent from the protection and control IEDs to the SCUs as GOOSE messages, also in
accordance with IEC 61850-8-1 [52].
13
2.3.5 Motivations Behind Digital Substations
There are several motivations behind digital substations. Some of them are as follows.
First of all, the process bus can be realized by a much smaller number of fiber optic
cables than the large number of copper cables which are required for the measuring and
control circuits in conventional substations. The costs associated with cabling can thus
be reduced [4]. Digital substations also allow the space which is required in the control
building to be reduced. The secondary devices can be made smaller than today since the
analog and binary Input/Output (I/O) boards are no longer needed in these devices [53].
In addition, the exchange of primary process data, commands and signals over the process
bus facilitates the integration of several functions into the same IED, which in turn can
lead to a reduced number of IEDs in the control room [4]. The reduction in cabling
and number of devices in the control building also allows the time spent on engineering,
installing, testing and replacing secondary systems to be significantly reduced [18], [53].
2.4 Standards and Implementation Guidelines

In this section, some of the standards and implementation guidelines which are relevant to
the application of OCTs in digital substations are presented. This includes IEC 61850-9-2,
IEC 61850-9-2 LE, IEC 61869 and IEC 60044-8.
2.4.1 IEC 61850-9-2
IEC 61850-9-2 is one of several parts of the IEC 61850 standard, “Communication net-
works and systems for power utility automation”, which is the international standard
for substation communication based on Ethernet [7]. IEC 61850-9-2 is called “Specific
communication service mapping (SCSM) - Sampled values over ISO/IEC 8802-3” [54].
This part of the IEC 61850 standard concerns the transmission of SV over Ethernet.
However, the standard leaves too much room for interpretation for actual systems to be
implemented. For instance, the standard does not specify what the dataset of SV should
include nor what the sample rate should be [55]. The solution came with the implemen-
tation guideline IEC 61850-9-2 LE, which will be described in the next subsection.
2.4.2 IEC 61850-9-2 LE
IEC 61850-9-2 LE, short for “IEC 61850-9-2 Light Edition”, is an implementation guide-
line made by the UCA International Users Group [55]. The implementation guideline
was created to ease the implementation of IEC 61850-9-2 and to facilitate interoperability
14
between different devices [55], [56]. IEC 61850-9-2 LE specifies two sample rates, one for
protection and metering applications and one for power quality applications. For protec-
tion and metering applications a sample rate of 80 samples per cycle is defined and for
power quality applications a sample rate of 256 samples per cycle. The resulting sample
rates in a 50 Hz network are thus 4 000 Hz and 12 800 Hz, respectively. IEC 61850-9-2
LE defines the content of the SV dataset to be one sample of each of the three phase
currents and three phase voltages from one network node in addition to neutral current
and neutral voltage [55]. The neutral values can either be measured by CITs or calculated
by the merging unit [56].
2.4.3 IEC 61869
The IEC 61869 standard, “Instrument transformers”, applies to both CITs and NCITs.
This standard comprises of several parts. Part 1, Part 6 and Part 9 are product family
standards which define general requirements applying to all types of instrument trans-
formers, additional general requirements applying to NCITs and additional requirements
applying to instrument transformers with a digital output, respectively [57], [29], [33]. In
addition to the product family standards, IEC 61869 comprises of several product stan-
dards. The product standards define additional requirements applying to specific types of
instrument transformers [57]. IEC 61869 gradually replaces the old standard for instru-
ment transformers, IEC 60044 [18]. At the present time, only Part 7 and Part 8 of IEC
60044, which apply to electronic voltage and current transformers, respectively, are still
active [58].
IEC 61869-2
IEC 61869-2, “Additional requirements for current transformers”, is a product standard

which applies to inductive current transformers. In this part of IEC 61869, standard
accuracy classes and corresponding error limits are defined for measuring current trans-
formers and protective current transformers, respectively. IEC 61869-2 has defined the
following standard measuring accuracy classes: 0.1, 0.2, 0,2 S, 0.5, 0,5 S, 1, 3 and 5.
Measuring accuracy class 0,2 S is specified by limits of ratio error and phase displacement
at 1 % to 120 % of the rated current. The standard accuracy classes for protective cur-
rent transformers are divided into different groups based on how the classes are specified.
For class P protective current transformers, the following standard accuracy classes are
defined: 5P and 10P. Protective accuracy class P is specified by limits of ratio error and
phase displacement at the rated primary current in addition to limits of composite error
at the rated accuracy limit primary current. The rated accuracy limit primary current
15
is a multiple of the rated primary current. The multiplier is called the Accuracy Limit
Factor (ALF), and its standard value is 5, 10, 15, 20 or 30 [21].
IEC 61869-6
IEC 61869-6, “Additional general requirements for low-power instrument transformers”,

specifies additional requirements concerning NCITs. In this part of IEC 61869, the ana-
log and digital output errors of NCITs are defined. IEC 61869-6 also covers additional
bandwidth requirements for NCITs [29].
IEC 61869-9
Part 9 of IEC 61869 is called “Digital interface for instrument transformers”. IEC 61869-
9 concerns instrument transformers which have a digital output and gives requirements
for the digital communication of measurements from such instrument transformers. IEC
61850-9-2 LE partially formed the basis of this standard. However, additions and im-
provements have been made [33]. For instance, instead of defining the sample rate of the
digital output in terms of samples per cycle, IEC 61869-9 defines sample rates which are
not dependent of the rated frequency of the power system. The preferred sample rates
are 4800 Hz for measuring and protective applications and 14 400 Hz for quality metering
applications. The same publishing rate of 2400 frames per second is defined for both
sample rates and is attained by packing 2 and 6 samples in each frame, respectively [55],
[33].
According to IEC 61869-9, the specifications of the accuracy classes for NCITs with a dig-
ital output apply to the digital output from the MU. In other words, all errors which occur
in the current measurement between the primary side of the NCIT and the digital output
are included. This includes errors caused by inaccuracies in the time synchronization [33].
IEC 61869-13
Part 13 of IEC 61869, “Stand-Alone Merging Unit”, has not been released yet. According
to IEC’s website, the forecasted publishing date is September 2019 [59]. This part of IEC
61869 will apply to SAMUs, as the name suggests. Among other things, this standard
will define requirements for the transient performance of SAMUs. IEC 61869-13 will
recommend making use of input elements which cannot saturate [60].
16
2.4.4 IEC 60044-8
IEC 60044-8 has partially been replaced by Part 6, 9 and 10 of IEC 61869. However, IEC
61869-8 is missing. This part has been planned for the future, but has not been developed
yet [59]. Thus, IEC 60044-8 is still applicable to OCTs [18]. The same standard measuring
accuracy classes are defined in IEC 60044-8 as for CITs in IEC 61869-2. The limits of error
are also equivalent. For protective current transformers three standard accuracy classes
are defined: 5 P, 10 P and 5TPE. The error limits for class 5 P and 10 P are equivalent
to those specified for CITs in IEC 61869-2. In addition to the error limits specified for
class 5 P, 5TPE also has a limit for the maximum peak instantaneous error at accuracy
limit conditions [61].
2.5 Regulations and Guidelines

Transmission substations in Norway, both conventional substations with CITs and dig-
ital substations with NCITs, must conform to the requirements of certain regulations,
including the ones listed below.
• “Forskrift om elektriske forsyningsanlegg” [62]
• “Forskrift om leveringskvalitet i kraftsystemet” [63]
• “Forskrift om forebyggende sikkerhet og beredskap i energiforsyningen (Beredskaps-
forskriften)” [64]
• “Forskrift om systemansvaret i kraftsystemet” [65]
In addition to the requirements specified in regulations, the owner may have their own
requirements for the design of the substation.
§ 14 of “Forskrift om systemansvaret i kraftsystemet” states the following.
“§ 14. Planning and commissioning of technical facilities in the

power system
The concessionaire shall inform the system operator about plans for new facil-
ities or alteration of its own facilities in or connected to the regional or central
grid, when other concessionaires are affected by this. The system operator shall
issue a decision on approval of new facilities or alterations of existing facilities
before these can be commissioned. ...” [65] (Translated from Norwegian)
The decision of the system operator is normally based on conditions which are indicated by
the guideline “Funksjonskrav i kraftsystemet” [66]. This guideline has been developed by
17
the system operator. Digital substations with OCTs must therefore meet the specifications
in this guideline.
Given below are some excerpts from “Beredskapsforskriften”, which the author finds par-
ticularly relevant when considering digital substations with OCTs.
§ 7-13 of “Beredskapsforskriften” states the following.
“§ 7-13. Protection against electromagnetic pulse and interference

The company must assess the vulnerability of the operational control system to
electromagnetic pulse (EMP) or electromagnetic interference (EMI). If vulner-
abilities are identified, security or emergency measures shall be implemented
according to the operational control system’s significance for safe operation
and restoration of its function in the power supply.” [64] (Translated from
Norwegian)
In “Beredskapsforskriften” facilities are classified based on the voltage level and the power
rating. There are three classes: Class 1, Class 2 and Class 3. The requirements for Class
3 facilities are the strictest. These are facilities which have the greatest importance for
the power supply [64].
“Class 3 includes:
...
b. Transformer station with overall main transformer rating of more than 100
MVA and built for a maximum voltage level of at least 200 kV and transfor-
mation to a secondary voltage level in networks of at least 30 kV.
c. Independent switching station in the power system built for a voltage level
of at least 200 kV.
...” [64] (Translated from Norwegian)
Substations in the transmission network are thus typically classified in class 3, based on
the definition of the transmission network which was given in Section 2.1. There are some
special requirements for the protection of facilities which are classified in Class 3. One of
the requirements applying to Class 3 transformer stations and switching stations is the
following [64].
“3.2.7 The facility shall have duplicated and physically independent routing
of cables for control and communication, emergency and station power and
high voltage, respectively, so that a single error or event cannot knock out vital
functions.” [64] (Translated from Norwegian)
18
Chapter 3
Optical Current Transformers
This chapter treats OCTs whose measuring principle is based on the Faraday effect. In
Section 3.1 basic concepts which are important for the understanding of the working
principle of OCTs are described. Section 3.2 presents two common detection methods,
and Section 3.3 describes some types of OCTs and their working principle. In Section 3.4
the associated MU of OCTs is shortly described. Section 3.5 treats the main sources of
error in OCTs and some solutions to how these can be mitigated. In Section 3.6 important
performance characteristics of OCTs are described, and in Section 3.7 other characteristics
are presented.
3.1 Basic Concepts

In order to understand the operating principle and the error sources of OCTs, familiarity
with the concepts of polarized light, birefringence and the Faraday effect is essential. The
purpose of this section is to provide the necessary knowledge.
3.1.1 Polarized Light
Polarized light has an electrical field which oscillates in an orderly manner [67]. Depending
on how the orientation of the electrical field vector evolves as the light wave propagates,
polarized light can be characterized as being linearly, circularly or elliptically polarized.
As shown in Figure 3.1a, the electric field vector of a linearly polarized light wave draws
a line in a plane which is perpendicular to the propagation direction of the light. The
electric field vector of a circularly polarized light wave, on the other hand, draws a circle,
as seen in Figure 3.1b, while the electric field vector of elliptically polarized light draws
an ellipse [68], [69]. Depending on the rotational direction of the electric field vector,
circularly and elliptically polarized light may further be characterized as being right-hand
or left-hand polarized [70].
19
CHAPTER 3. OPTICAL CURRENT TRANSFORMERS
(a) Linearly polarized light [71]. (b) Circularly polarized light [72].
Figure 3.1: Polarized light.
Any type of polarized light can be decomposed into two orthogonal linearly polarized
components [73]. Assuming that a polarized light wave is propagating along the z-axis,
as in Figure 3.1a and Figure 3.1b, the light wave can be decomposed into a component
polarized along the x-axis and a component polarized along the y-axis. In Figure 3.1
these two components are drawn in blue and green, respectively. The resulting light
waves are drawn in red. Linearly, circularly and elliptically polarized light result from
different magnitudes and relative phase shifts between the two orthogonal components.
If the x and y components are in phase, as seen in Figure 3.1a, the resulting light will
be linearly polarized. Circularly polarized light, on the other hand, results if the x and y
components have the same amplitude and are 90° out of phase, as shown in Figure 3.1b.
Any other mix of amplitudes and relative phase shifts gives elliptically polarized light [70].
Linearly polarized light can also be decomposed into two orthogonal circularly polarized
components. Left-hand and right-hand circularly polarized light waves are orthogonal.
Hence, a linearly polarized light wave can be seen as a superposition of a right-hand and
a left-hand circularly polarized light wave [68].
3.1.2 Birefringence
A medium is said to be birefringent if the value of its refractive index is dependent on the
propagation direction and the polarization of light [73]. The refractive index of a light
wave in a medium determines the speed of the light through the medium, as shown by
Equation (3.1),
c
v= (3.1)
n
20
where c is the speed of light in vacuum and n is the refractive index of light in a medium
[69]. Birefringence may be divided into linear and circular birefringence. Depending on
the propagation direction and the polarization of an incident light wave, a birefringent
medium will either maintain or change the polarization of the light [68].
In a linearly birefringent medium, there are two orthogonal birefringence axes with differ-
ent refractive indexes, commonly referred to as the fast axis and the slow axis. In Figure
3.1a, the fast and the slow axes could be the x-axis and the y-axis, respectively. The po-
larization of incident light which is linearly polarized along the fast or the slow axis of a
linearly birefringent medium will remain unchanged as the light wave propagates through
the medium. In Figure 3.1a this applies to the blue and the green linearly polarized light
waves. However, these two light waves will experience different refractive indexes and
thus travel at different speeds through the medium. If a linearly polarized light wave
enters a linearly birefringent medium at an angle to the fast and the slow axes, the light
wave will be split into two components due two the different refractive indexes of the fast
and the slow axes. In Figure 3.1a this applies to the red wave, which will be split into
the blue and the green linearly polarized components. Since the two components travel
at different speeds through the medium, a phase shift will develop between them. As
a result, the incident linearly polarized light is transformed into elliptically or circularly
polarized light, depending on the value of the differential phase shift which develops [74],
[75].
From the above-given explanation, it can be concluded that a linearly birefringent medium
will maintain the polarization state of linearly polarized light which is polarized along the
medium’s fast axis or slow axis, and change the polarization state of linearly polarized
light which is polarized at an angle to the fast and the slow axes [74], [75]. A circularly
birefringent medium maintains the polarization states of right-hand and left-hand circu-
larly polarized light [75]. However, these two polarization states will experience different
refractive indexes in the medium and thus travel at different speeds [74]. The polariza-
tion state of linearly polarized light, on the other hand, will be rotated. This phenomena
can be explained by considering the linearly polarized light wave as a superposition of a
right-hand and a left-hand circularly polarized light wave. Due to the relative phase shift
which develops between these two components in a circularly birefringent medium, the
plane of polarization of the linearly polarized light wave will be rotated [75].
21
3.1.3 The Faraday Effect
The plane of polarization of a linearly polarized light wave will be rotated when the wave
propagates through a magneto-optical medium which is placed in a magnetic field. This
effect is known as the Faraday effect and is illustrated in Figure 3.2. The angle of rotation
of the polarization plane is given by Equation (3.2),
θ = V HL (3.2)
where V is the Verdet constant of the medium, H is the magnetic field strength parallel to
the propagation direction of the light and L is the distance which is travelled by the light
in the medium [76]. The Verdet constant is dependent on the properties of the medium
and varies with temperature and the wavelength of the light [69].
Figure 3.2: The Faraday effect [77]. © 2017 IEEE
The Faraday effect is actually circular birefringence induced by a magnetic field [74]. As
already explained, a linearly polarized light wave can be considered as a superposition of
a right-hand and a left-hand circularly polarized light wave. These two components will
experience different refractive indexes in a magneto-optical medium which is subjected to
a magnetic field. The result is a difference in the speeds of the two components through
the medium. The polarization plane of the resultant linearly polarized light will thus be
rotated [76]. If right-hand circularly polarized light propagates in the direction of the
magnetic field, the speed of the light will be decreased. The speed of right-hand circularly
polarized light which is propagating in the opposite direction of the magnetic field, on the
other hand, will be increased. The opposite applies to left-hand circularly polarized light
[68].
22
3.2 Detection Methods

OCTs make use of the Faraday effect to measure the current in the primary conductor
[78]. Polarized light is sent through the sensing element, which is typically placed around
the primary conductor. The light will thus propagate along the magnetic field which is
produced by the primary current [74]. Two principal detection methods can be applied to
measure the Faraday rotation of the light due to the magnetic field. These are polarimetric
detection and interferometric detection [79].
3.2.1 Polarimetric Detection
Polarimetric detection is based on measuring the Faraday rotation angle of linearly po-
larized light which is propagating through the sensing element of the OCT. If the sensing
element encloses the primary conductor, Equation (3.2) can be combined with Ampere’s
law by expressing the equation as a line integral. The Faraday rotation angle can then
be expressed as in Equation (3.3),
I
θ=V ~ ~l = V N I
Hd (3.3)
C
where V is the Verdet constant of the sensing element, H is the magnetic field strength
parallel to the propagation direction of the light, N is the number of turns of the sensing
element around the primary conductor and I is the current in the primary conductor [76].
If the sensing element encloses the primary conductor, the Faraday rotation angle will
thus be proportional to the current to be measured. The Faraday rotation angle cannot
be detected directly. Instead it is transformed into a light intensity, which can be detected
[76].
3.2.2 Interferometric Detection
An alternative to measuring the Faraday rotation angle of linearly polarized light is to

measure the differential phase shift which accumulates between right-hand and left-hand
circularly polarized light propagating through the sensing element, due to the Faraday
effect. This is known as interferometric detection. As with polarimetric detection, the
phase shift must be converted into a light intensity in order to be detectable [76].
23
3.3 Types of Optical Current Transformers

OCTs comprise of the following parts: the sensor heads housing the sensing elements, the
sensor electronics and the fiber optic cables connecting the sensor electronics to the sensor
heads [80]. The MUs is typically integrated with the sensor electronics [36]. OCTs may
be divided into two groups based on the material which is used for the sensing element:
optical fiber current sensors and bulk glass optical current sensors [81]. These two types
of sensors will be described in further detail in the following subsections. The main focus
will be on optical fiber current sensors since most of the commercially available OCTs
which are considered in this thesis use this technology. This technology is also considered
as a better solution for new OCTs [80].
3.3.1 Optical Fiber Current Sensors
The sensing element of optical fiber current sensors comprises of an optical fiber which
is wound around the primary conductor one or several times [38]. One of the benefits of
this type of OCT is that the sensitivity of the sensor can easily be adapted by choosing
a different number of sensing fiber coil turns [79]. Optical fiber current sensors may
use either polarimetric or interferometric detection methods [68]. In the rest of this
subsection, some types of polarimetric and interferometric optical fiber current sensors
will be presented and their working principles explained.
Polarimetric Optical Fiber Current Sensors
Figure 3.3 shows the schematic of an optical fiber current sensor with a basic polarimetric
detection scheme. The working principle of the sensor is as follows. A light beam is
emitted from a light source and transformed into linearly polarized light by a polarizer
before it enters the sensing fiber coil. As the linearly polarized light beam propagates
through the sensing fiber coil, the polarization plane of the linearly polarized light is
rotated due to the Faraday effect [76]. At the end of the sensing fiber coil, the polarization
plane of the light has been rotated by an angle which is given by Equation (3.3) [68].
24
Figure 3.3: Basic polarimetric optical fiber current sensor [68]. © 2009 IEEE
When exiting the sensing fiber coil, the light beam passes through a second polarizer with
a transmission axis which is oriented at 45° with respect to the transmission axis of the
first polarizer. The intensity of the output light is detected by a photodetector and is
given by Equation (3.4),
I0
Id = (1 + sin 2θ) (3.4)
2
where I0 is the intensity of the input light and θ is the angle of rotation of the plane of
polarization due to the Faraday effect [68]. For sufficiently small values of θ, the transfer
function of the sensor is approximately linear. However, since the output light intensity
is dependent on the intensity of the input light, the sensor will be sensitive to fluctuations
in the input light. A solution to this issue is to modify the configuration of the sensor to
the schematic shown in Figure 3.4 [68], [79].
Figure 3.4: Dual polarimetric optical fiber current sensor [68]. © 2009 IEEE
25
In this configuration the linearly polarized light passes through a polarizing beam splitter
after propagating through the sensing fiber coil. The beam splitter splits the linearly
polarized light into two orthogonal linearly polarized light beams polarized at -45° and
+45° with respect to the polarization of the light from the first polarizer, respectively [79],
[76], [68]. The intensity of the two light beams is detected by two separate photodetectors.
A signal processor calculates an output signal which is given by Equation (3.5),
I1 − I2
S= = sin 2θ ≈ 2 V N I (3.5)
I1 + I2
where I1 and I2 are the output signals from the two photodetectors [76], [68]. Equation
(3.5) is only valid as long as the Faraday rotation angle is sufficiently small, that is when
sin 2θ ≈ 2θ. The linearity of this type of sensor is therefore limited [68].
Sagnac Loop Interferometer Optical Fiber Current Sensor
Figure 3.5 displays the schematic of a Sagnac loop interferometer optical fiber current
sensor, which measures the Faraday rotation interferometrically. A light beam is emitted
from the light source and converted into linearly polarized light by the polarizer. The
second coupler splits the light into two linearly polarized light beams, which subsequently
propagate from the sensor electronics to the sensor head through separate polarization-
maintaining fiber optic cables. Each light beam is transformed into circularly polarized
light by a quarter-wave retarder before entering the sensing fiber coil [82], [76]. Quarter-
wave retarders are made by the use of birefringent materials and have the following
working principle. By orienting the polarization plane of the incident linearly polarized
light at 45° with respect to the fast and slow axes of the retarder, the light beam is split
into two orthogonal linearly polarized components with an accumulating differential phase
shift. The length of the retarder is chosen in such a way that the phase shift between
the two linearly polarized components equals 90° at the end of the retarder. The incident
linearly polarized light is thus transformed into circularly polarized light [82].
26
Figure 3.5: Sagnac loop interferometer optical fiber current censor. The figure is adapted
from [82] © 2007 IEEE and [83] © 2002 IEEE.
As shown in Figure 3.5, the two circularly polarized light waves have the same sense of
rotation, but propagate in opposite directions through the sensing fiber coil and thus with
or against the direction of the magnetic field which surrounds the primary conductor [83].
Due to the Faraday effect, the speed of one of the light waves will increase, and the speed
of the other decrease. When exiting the sensing fiber coil, the circularly polarized light
waves are transformed back to linearly polarized light by the quarter-wave retarders. Due
to the speed difference between the two light waves in the sensing fiber coil, a differential
phase shift has accumulated, which is given by Equation (3.6) [68].
∆φS = 2 V N I (3.6)
The returning linearly polarized light waves interfere at the polarizer, and the intensity
of the output light, which is given by Equation (3.7), is detected by a photo diode [68].
I0
Id = (1 + cos ∆φS ) (3.7)
2
As can be seen from Equation (3.7), the sensitivity of the sensor is severely limited when
the differential phase shift is close to zero. However, an output signal which is linearly
related to the differential phase shift for ∆φS 90° can be obtained by the use of a
phase modulator. The linearity of the sensor can be significantly improved by employing a
closed-loop control scheme, as shown in Figure 3.5. In the closed-loop control scheme, the
27
phase modulator is used to compensate the phase shift which has accumulated between the
two returning linearly polarized light waves. The differential phase shift is compensated
in such a way that the two light waves are in phase when they are brought to interference
at the polarizer. The size of the differential phase shift and thus the value of the primary
current to be measured can be deduced from the control signal which is fed back to the
phase modulator. With this control scheme the linearity of the measurement is maintained
over a much wider range than for an open-loop sensor [68], [83], [84].
Reflective Interferometer Optical Fiber Current Sensor
The reflective interferometer optical fiber current sensor is another type of interferometric
current sensor. The schematic of this sensor is displayed in Figure 3.6, As in the Sagnac
loop configuration, the light source emits a light beam, which is subsequently transformed
into linearly polarized light by the polarizer. However, in the reflective interferometer con-
figuration, the linearly polarized light beam is split into two orthogonal linearly polarized
light beams by the use of a 45° splice, and the two light beams travel towards the sensor
head in the same polarization-maintaining fiber optic cable [76].
Figure 3.6: Reflective interferometer optical fiber current sensor. The figure is adapted
from [85] © 2018 IEEE and [83] © 2002 IEEE.
Before entering the sensing fiber coil, the two linearly polarized light waves are converted
into a right-hand circularly polarized light wave and a left-hand circularly polarized light
wave by a quarter-wave retarder. At the end of the sensing fiber coil, the light waves
are reflected in a mirror, and the polarization states of the waves are swapped. Thus,
the forward-propagating right-hand circularly polarized light wave propagates back to the
28
quarter-wave retarder as a left-hand circularly polarized light wave and vice versa. Due
to the Faraday effect, a phase shift accumulates between the two light waves as they
propagate through the sensing fiber coil. When the light waves exit the sensing fiber coil,
the differential phase shift is given by Equation (3.8). As can be seen, the differential
phase shift is twice as big as for the Sagnac loop interferometer configuration due to the
polarization swapping. Thus, the sensitivity of the sensor is also twice as high [68], [76],
[83].
∆φR = 4 V N I (3.8)
The circularly polarized light waves are converted back to two orthogonal linearly po-
larized light waves by the quarter-wave retarder and are brought to interference at the
polarizer. In order to obtain the differential phase shift which is caused by the Fara-
day effect and thereby the value of the primary current, similar open-loop and closed-loop
schemes to those described for the Sagnac loop interferometer configuration can be applied
[68], [76].
3.3.2 Bulk Glass Optical Current Sensors
The other type of OCTs is the bulk glass optical current sensors. Instead of using optical
fiber as the sensing element, these current sensors make use of magneto-optical glass or
crystal. The sensing element can be shaped as a rectangular block, as shown in Figure 3.7,
a triangular block or a glass ring enclosing the primary conductor. The sensor head can
also be constructed as an open core which do not enclose the primary conductor entirely
or simply a rod which is placed near the primary conductor [86], [87].
Figure 3.7: Bulk glass optical current sensor [74]. © 2015 IEEE
29
Bulk glass optical current sensors make use of polarimetric detection schemes in order to
measure the Faraday rotation of the polarized light. The basic principle is the same as
explained for the polarimetric optical fiber current sensor [74] in Subsection 3.3.1 and will
thus not be repeated here.
3.4 Merging Unit

The associated MUs of OCTs is typically integrated with the sensor electronics. The
signal processor, as shown in Figure 3.5 and Figure 3.6, outputs a digital signal, which
is re-sampled and time stamped by the MU [35]. Standard sample rates are defined
in IEC 61850-9-2 LE [54] and IEC 61869-9 [33], as mentioned in Subsection 2.4.2 and
Subsection 2.4.3. The MU formats the digital samples as SV according to IEC 61850-9-2
LE and subsequently publishes the SV messages to the process bus [36]. The MU may
be equipped with analog inputs, which can be used to interface Conventional Voltage
Transformers (CVTs) and CCTs. These current and voltage signals are synchronized and
merged with the currents which are measured by the OCT [48].
3.5 Sources of Error

Errors in OCT measurements are mainly caused by the effects of birefringence, temper-
ature, vibration and non-ideal optical parts [74]. These effects will be described in more
detail in the following subsections. For the same reasons stated in Subsection 3.3, the
focus will be put on sources of error in optical fiber current sensors.
3.5.1 Linear Birefringence
Linear birefringence in the sensing fiber is the main source of error in optical fiber current
sensors [74]. The linear birefringence may be intrinsic or induced by external effects. The
intrinsic birefringence is caused by the fact that the shape of the sensing fiber core is not
completely perfect. When it comes to induced birefringence, on the other hand, external
effects, such as temperature variations and vibration, can change the linear birefringence
of the sensing fiber. Linear birefringence can also be induced when the sensing fiber is
bent around the primary conductor [88], [89].
Linear birefringence in the sensing fiber adversely affects the performance of optical fiber
current sensors in several ways. First of all, linear birefringence may reduce the sensitivity
of polarimetric sensors. As explained in Section 3.1.2, a linearly birefringent medium can
30
transform linearly polarized light into elliptically or circularly polarized light. If the
output light from the sensing fiber coil turns into elliptically polarized light due to the
linear birefringence in the sensing fiber, the sensor will be less sensitive to measuring
the Faraday rotation which is induced by the magnetic field surrounding the primary
conductor. If the output light is transformed into circularly polarized light, the Faraday
rotation can no longer be measured [76].
The accuracy and stability of optical fiber current sensors are also disturbed by linear
birefringence in the sensing fiber [76]. As already mentioned, environmental disturbances,
such as vibration and temperature changes, can cause linear birefringence in the sensing
fiber. As long as these disturbances are felt by the whole sensing fiber and do not vary in
time, the phase shifts of the polarized light waves due to linear birefringence will generally
be cancelled out in both the Sagnac loop and the reflective interferometer configurations.
In the Sagnac interferometer configuration, the two counter-propagating light waves will
gain phase shifts due to linear birefringence in the sensing fiber, which are equal. These
phase shifts will thus be cancelled out when the two waves interfere at the polarizer. In
the reflective interferometer configuration, the phase shifts due to linear birefringence are
cancelled out as a consequence of the polarization swapping [90], [83].
Vibration and variations in temperature can also induce time-varying linear birefringence
in the sensing fiber [88]. The effects of time-varying linear birefringence will not be
cancelled out. In the Sagnac loop and the reflective interferometer configurations, the
two polarized light waves are not at exactly the same point in the sensing fiber at the
same time. Thus, the phase shifts which the light waves gain due to time-varying lin-
ear birefringence in the sensing fiber, will be different [78], [83]. However, the reflective
interferometer configuration is about 1000 times less sensitive to vibration and temper-
ature variations which cause time-varying linear birefringence in the sensing fiber, than
the Sagnac configuration. This is due to the fact that the relative delay in time between
the two co-propagating light waves in a reflective interferometer configuration is much
smaller than the relative delay in time between the two counter-propagating waves in a
Sagnac loop configuration [83].
There are several solutions which can be employed in order to reduce the effects of linear
birefringence in optical fiber current sensors. First of all, the choice of optical fiber material
has a great impact on the amount of linear birefringence in the sensing fiber. The proper
choice of sensing material can reduce the amount of intrinsic linear birefringence in the
sensing fiber as well as the linear birefringence caused by bending. Further reduction of the
intrinsic linear birefringence of the sensing fiber can be achieved by annealing. The effects
of linear birefringence can also be suppressed by inducing circular birefringence in the
fiber. This can be done by twisting the sensing fiber [74], [88]. Descriptions of additional
31
solutions where the configuration of the sensing fiber is optimized, the configuration of
the optical fiber current sensor is altered or additional optical devices are added to the
sensor, can be found in [74].
3.5.2 Temperature
The accuracy of OCTs should be maintained over the whole operating temperature range
[38]. In IEC 60044-8 three temperature categories are defined for NCCTs under normal
service conditions. The widest ambient air temperature range is -40 °C to +40 °C [61].
All of the commercial OCTs which are considered in this thesis have specified outdoor
operating temperature ranges which are equal to or wider than the one defined in IEC
60044-8 [49], [48], [46], [47].
The accuracy of OCTs may be affected by variations in the ambient temperature. This
applies to both the sensor heads and the sensor electronics [74]. Variations in the temper-
ature of the sensor heads change the Verdet constant of the optical sensing fiber. Since
the amount of Faraday rotation depends upon the value of the Verdet constant, as shown
in Equation (3.2), the Faraday rotation angle will change if the sensor head is subjected
to temperature variations. The temperature dependency of the Verdet constant for fused
silica fiber is given by Equation (3.9) [83].
1 dV
= 0.7 × 10−4 °C −1 (3.9)
V dT
As a result of the temperature dependency, the measured signal will have an error of up
0.56 % over the operating temperature range from -40 °C to +40 °C. This is not sufficient
to satisfy the requirements of measuring class 0,2 S over the entire current range which
is specified in IEC 60044-8. In order to meet the accuracy requirements, the temperature
drift of the Verdet constant must be compensated for [74]. One solution is to equip the
sensing fiber with a temperature sensor and implement temperature compensation as a
part of the signal processing. IEEE recommends periodical verification of the accuracy
of the temperature sensor if such a solution is employed [38]. Another solution is to
make use of a quarter-wave retarder with a temperature dependency which counteracts
the temperature dependency of the Verdet constant [68]. Bohnert, Gabus, Nehring and
Brändle [83] demonstrated that an OCT with such a quarter-wave retarder maintained
an accuracy of 0.2 % over an operating temperature range from -35 °C to + 85 °C [83].
A part of the sensor electronics which is especially sensitive to temperature variations

is the light source. Superluminescent Diodes (SLEDs) are commonly used as the light
32
sources in OCTs [91]. The central wavelength of the light which is emitted from such
a broadband semiconductor light source varies with the temperature. Since the value
of the Verdet constant is dependent upon the wavelength of the light, variations in the
temperature of the light source will cause variations in the Verdet constant and thus
variations in the amount of Faraday rotation. In order to keep the wavelength of the light
source stable, the temperature of the light source has to be controlled or compensated for
[74].
3.5.3 Vibration and Stray Magnetic Fields
In a substation environment, OCTs can be subjected to vibrations, for instance due to

operation of circuit breakers. Such vibrations should not affect the accuracy of OCTs
[38]. However, the accuracy of Sagnac loop optical fiber current sensors may be affected by
angular vibrations due to the Sagnac effect. In this configuration, two counter-propagating
light waves travel around a closed circuit, as explained in Subsection 3.3.1. If the sensor
experiences angular vibrations, the Sagnac effect will cause a phase shift between the two
light waves, which creates an error in the measured current. However, if the OCT is
constructed with a reflective interferometer configuration, the optical path does not form
a closed circuit. Hence, there will be no Sagnac effect, eliminating the errors which are
caused by it [76], [86], [78].
Bohnert, Gabus, Nehring and Brändle [83] experimentally compared the sensitivity of a
reflective interferometer and a Sagnac loop interferometer fiber optic current sensor to
linear vibrations. The sensing fiber coils of both sensors were subjected to vibrations at
a frequency of 50 Hz. The acceleration of the vibrations was increased from 0 g to 10 g.
Figure 3.8 shows the results of the tests. As seen in the figure, the vibrations did not affect
the performance of the current sensor with the reflective interferometer configuration, in
contrast to the one with the Sagnac loop interferometer configuration [83]. The results
are in compliance with what was explained in Subsection 3.5.1 regarding the sensitivity of
the two sensor configurations to vibrations which lead to time-varying linear birefringence
in the sensing fiber.
33
Figure 3.8: Sensitivity of a Sagnac loop and a reflective interferometer optical fiber cur-
rent sensor to vibrations. The signals which were induced by the vibrations are expressed
as equivalent RMS current [83]. © 2002 IEEE
The accuracy of optical fiber current sensors is not affected by external stray magnetic
fields as long as the sensing fiber encloses the primary conductor entirely. If this is the
case, the only current contributing to the Faraday rotation, will be the primary current
which is enclosed by the sensing fiber coil. Optical fiber current sensors are therefore
not affected by the magnetic fields from for instance other phases in a substation. In
addition, the Faraday rotation is independent of the shape and diameter of the sensing
fiber loop and of the position of the primary conductor. Thus, if the position of the
primary conductor changes due to vibrations, this will not affect the accuracy of the fiber
optic current sensor [82], [74], [68].
3.5.4 Non-Ideal and Aging Optical Parts
Another source of error in OCTs is optical parts which are not ideal. For instance,
measurement errors may arise if the output light from the polarizers is not entirely linear
or if the transmission axes of the two polarizers in a polarimetric detection scheme are
not oriented at a relative angle of exactly 45°. In interferometric detection schemes, the
retarders can cause errors in the measurement. The polarization plane of the input light
has to be oriented at an angle of exactly 45° with respect to the fast and slow axes of
the retarder. In addition, the relative angle which accumulates between the two linearly
polarized components has to be exactly 90° when the light exits the retarder. If not,
measurement errors will be introduced [74].
Aging of optical parts may also affect the accuracy of OCTs. For instance, as the light
34
source ages, the wavelength of the light changes, which will change the amount of Faraday
rotation, as explained in Subsection 3.5.2. Unless the sensor electronics compensate for the
changing wavelength, the accuracy of the OCT will change over time, which is undesirable
since the stability of OCTs should be at least as good as that of CCTs [38].
3.6 Performance Characteristics

This section presents some of the most important performance characteristics of OCTs.
In order to substantiate the descriptions, some experimental results obtained by other
researchers will also be presented.
3.6.1 Linearity and Saturation
As explained in Subsection 3.3.1, OCTs have a transfer function which is linear over a
limited dynamic range. However, by the use of digital signal processing, a measurement
which is linear over a wide dynamic range is achieved [68]. Since OCTs do not have an
iron core and employ a saturation-free measurement principle, OCTs cannot saturate [31].
Nevertheless, the measured current may be distorted in such a way that the peaks are
clipped if the primary current exceeds the maximum detectable setting current of the
OCT. Proper dimensioning of OCTs is sufficient to prevent distortion of the measured
current. The maximum detectable setting current must simply exceed the highest short-
circuit current which can possibly occur at the measurement point. This is done by
designing the optical fiber current sensors with the proper number of sensing fiber coil
turns. There is an inversely proportional relation between the number of sensing fiber
coil turns and the maximum detectable current [50].
In 2010, Kucuksari and Karady [35] reported on an experimental comparison of a CCT

and an optical fiber current sensor. Among other things, a comparison of the linearity
of the two CTs was performed. The results are shown in Figure 3.9. Primary currents
were applied to the optical fiber current sensor in the range from 5 to 20 times its rated
current, and to the CCT in the range of 0.25 to 1.25 times its rated current. As shown
in the figure, the linearity of the measurements from the optical fiber current sensor is
maintained over the whole current range. The measurements from the CCT, on the other
hand, become nonlinear already at an applied current of 0.8 times the rated current [35].
35
Figure 3.9: Linearity of an optical fiber current sensor compared to the linearity of a
CCT. 1 pu equals 200 A and 400 A for the optical fiber current sensor and the CCT
respectively [35]. © 2010 IEEE
3.6.2 Bandwidth
The bandwidth of the sensing element of OCTs is only limited by the propagation time of
light between the light source and the detector. Hence, signals in a very wide frequency
range (DC to GHz AC) can be measured. However, the bandwidth of the final output
signal of OCTs is significantly lower. The output bandwidth is limited by the signal
processing electronics, by the sample rate and by the digital communication [92], [38].
As mentioned in Section 2.4.2, the SV sample rates which are specified in IEC 61850-9-2
LE are 80 samples per cycle for measuring and protective applications and 256 samples
per cycle for quality metering applications. According to the Nyquist sampling theorem,
the sampling frequency should be twice as high as the highest frequency in the signal in
order avoid aliasing of the input signal [7]. If the digital output of the OCT complies
with the specifications in IEC 61850-9-2 LE, the output bandwidth will thus be limited
to 2 kHz for measuring and protective applications and to 6 400 Hz for quality metering
applications in a 50 Hz system [Appendix A].
Kucuksari and Karady [35] experimentally tested the bandwidth of the digital output
of an optical fiber current sensor. A primary current with a constant amplitude and a
frequency which varied from 60 Hz to 4.8 kHZ was applied to the optical fiber current
sensor. The results of the test are shown in Figure 3.10. As shown in the figure, the
optical fiber current sensor proved to have a bandwidth of about 2.4 kHz, which is half
of the sample rate of 80 samples per cycle in a 60 Hz system [35].
36
Figure 3.10: Frequency response of an optical fiber current sensor [35]. © 2010 IEEE
3.6.3 Steady State Performance
OCTs can provide a high accuracy [8] and are typically specified to conform to measuring
accuracy class 0,2 S, as defined in IEC 60044-8 [38]. In 2016, Thomas et al. [93] reported
on measuring the accuracy of a reflective interferometer optical fiber current sensor with
a rated current of 2 kA. The results are shown in Figure 3.11. Figure 3.11a and Figure
3.11b display the magnitude and phase error of the OCT, respectively, at different current
levels. The black lines indicate the requirements of measuring accuracy class 0,2 S, as
defined in IEC 60044-8, and the blue dots represent the measured accuracy. The standard
deviation of the measurements are represented by the red bars and the expected current
error is indicated by the green lines. As shown in the figure, the magnitude and phase
errors are within the standardized limits over the entire current range. However, a greater
deviation in the measurements can be observed at lower current levels [93].
The limiting factor of the accuracy of OCTs is electronic noise, especially at low current
levels [Appendix A]. At current levels which are significantly below the rated current of
an OCT, the level of noise in the measured current may be significant [38]. In order to
increase the Signal-to-Noise Ratio (SNR) of an optical fiber current sensor, the sensor can
be designed with an increased number of fiber coil turns [31].
37
(a) Magnitude error. (b) Phase error.
Figure 3.11: Measured accuracy of a reflective interferometer optical fiber current sensor.
The black lines indicate the requirements of measuring accuracy class 0,2 S [93]. © 2016
IEEE
Rahmamatian and Blake [8] demonstrated the ability of a reflective interferometer optical
fiber current sensor to maintain a high accuracy over a wide dynamic range. The optical
fiber current sensor had a rated current of 2 kA and maintained a measuring accuracy of
0.1 % over a range from 0.1 % to 180 % of the rated current (2 kA - 180 kA) [8].
3.6.4 Transient Performance
As mentioned in Subsection 2.4.4, the protective accuracy classes for OCTs are 5 P, 10 P
and 5TPE, as defined in IEC 60044-8. The transient performance of OCTs is dependent
on the associated sensor electronics [29]. OCTs are linear over a wide dynamic range and
cannot saturate, as already stated. The primary current will thus be transformed without
any distortion as long as the maximum detectable setting current of the OCT exceeds the
highest short-circuit current which can possibly occur at the measurement point.
The excellent transient performance of OCTs has been confirmed by a number of re-
searchers. Kucuksari and Karady [35] reported on performing an experimental test of the
transient performance of an optical fiber current sensor. The current which was applied
to the primary conductor had a maximum peak value of 20.65 times the rated primary
current of the optical fiber current sensor and a DC offset which decayed within three
power cycles. The resulting transient response of the OCT is shown in Figure 3.12. As
seen in the figure, no signs of saturation or distortion can be observed [35].
38
Figure 3.12: Transient response of an optical fiber current sensor [35]. © 2010 IEEE
Another transient performance test of OCTs was performed by Normandeau and Mah-
seredjian [94] who tested the transient performance of three OCTs. Both symmetrical and
asymmetrical primary currents with a peak value in the range from 7 kA to 57 kA were
applied to the OCTs. The composite and instantaneous errors of two of the OCTs were
well within the limits of 5% and 10 % which are specified for protection class 5TPE in
IEC 60044-8. The highest composite and instantaneous errors calculated were 1.39 % and
4.15 %, respectively. The composite and instantaneous errors of the third OCT exceeded
the specified limits at an applied primary current of 57 kApeak . However, it was discovered
that the OCT was not correctly calibrated. The incorrect calibration was regarded as the
cause of the composite and instantaneous errors exceeding the specified limits [94].
3.7 Attributes
OCT technology has been known for several decades. However, OCTs have not been
widely adopted in substations yet [80], [38]. The different parts of an OCT have different
expected lifetimes. The expected lifetime of the sensor heads is 30+ years, while the
sensor electronics have an expected lifetime of around 15 years. The sensor electronics
should thus be replaced once during the life of the sensor heads [Appendix A].
The insulation of OCTs can be made simpler than for CCTs and without use of SF6 gas or
oil. This is due to the fact that fiber optic cables are non-conductive and provide galvanic
isolation [38]. OCTs have a small size and low weight, which facilitate integration into
other primary equipment. This could for instance be a circuit breaker or a bushing [95].
For specific weights of commercially available OCTs, the reader is referred to Chapter 4.
39
Chapter 4
Commercial Products
In this chapter the attributes of some of the OCTs which are commercially available at
the present time are presented and compared. Section 4.8 presents an overview of the
specifications of these OCTs. OCTs from four different vendors have been considered.
These will be referred to as OCT 1, 2, 3 and 4, respectively. The whole sensor electronics
unit will hereafter be referred to as the MU. The information in this chapter is based on
the available technical documentation of the different OCTs in addition to information
obtained from contact with the vendors. In Appendix A the relevant e-mails are given.
4.1 Types of Optical Current Transformers

All of the considered OCTs except one are interferometric optical fiber current sensors
with different sensor configurations [49], [48], [46], [Appendix A]. OCT 2 has a Sagnac
loop interferometer configuration, while OCT 1 and OCT 3 have reflective interferometer
configurations [84], [Appendix A]. The last vendor uses glass rings instead of optical fiber
as the sensing element [47].
4.2 Design
All four vendors offer free-standing OCTs were each sensor head is mounted on top of
an insulator. An example is shown in Figure 4.1. The insulators are dry-type insulators,
either solid insulators or insulators filled with nitrogen or air, and are thus free of SF6
gas and oil [49], [48], [46], [47]. There are also possibilities for integration of the sensor
heads into other primary equipment, such as disconnecting switches and circuit breakers,
or installation of the sensor heads in conjunction with other primary components, such as
cable terminals and bushings [48], [47]. Other products, such as a disconnecting circuit
breaker with an integrated optical current sensor and a combined optical current and
voltage sensor are also on the market. The disconnecting circuit breaker combines the
functionality of a disconnector, a circuit breaker and an optical current sensor into one
41
CHAPTER 4. COMMERCIAL PRODUCTS
device [93] [Appendix A].
All of the OCTs are said to be equally suited for both measuring and protective appli-
cations [Appendix A]. The specified accuracy classes are given in Table 4.1. OCT 1 is
designed with two sensing fiber coils in each sensor head and can be equipped with two
MUs for redundancy. The sensor heads of OCT 2 can also be made with two sensing
fiber coils in each. The only vendor which has intended to place the MU outdoor in the
substation yard is the vendor of OCT 1. This MU shall be installed in a heated outdoor
cubicle, which is mounted to one of the insulator support frames. The fiber optic cables
connecting the MU to the sensor heads can be up to 300 m long [49], [84]. The other
three vendors offer solutions where the MU is installed indoor in a protection and control
panel [48], [46], [47]. The maximum length of the fiber optic cables connecting the MU
to the sensor heads is typically 1000 m [47], [Appendix A].
Figure 4.1: A commercially available OCT. Courtesy of ARTECHE [48].
4.3 Interfaces and Time Synchronization

The digital output of all MUs conform to the IEC 61850-9-2 LE implementation guideline.
OCT 2 is also specified to comply with the newer IEC 61869-9 standard [49], [48], [46],
[47]. For time synchronization different solutions are used. At the present time only the
MU of OCT 2 supports both Precision Time Protocol (PTP) and 1 Pulse Per Second
(PPS) time synchronization [48]. Other vendors make use of 1PPS and possibly IRIG-B.
42
Only one of the vendors answered to questions regarding support of network redundancy
protocols. This OCT supports both Parallel Redundancy Protocol (PRP) and High-
Availability Seamless Redundancy (HSR) [Appendix A]. The MUs of OCT 2 and OCT 3
are equipped with three-phase analog voltage inputs to interface with conventional voltage
transformers [48], [Appendix A]. The MU of OCT 2 also has analog current inputs, which
could for instance be used to acquire the neutral current [48].
4.4 Sensitivity to Environmental Disturbances

Only one of the vendors answered to questions regarding the sensitivity of the OCT to
temperature, vibration and stray magnetic fields. This vendor guarantees that the accu-
racy of the OCT is maintained over the whole operating temperature range. Temperature
compensation is employed by the use a temperature sensor which senses the temperature
of the sensing fiber [Appendix A]. Two other vendors also specify using temperature sen-
sors [96], [46]. The operating temperature range of the different OCTs are given in Table
4.1. The contacted vendor also stated that the OCT is not sensitive to vibration or stray
magnetic fields [Appendix A].
4.5 Expected Lifetime and Maintenance

The expected lifetime of the OCTs is considered to be approximately 15 years for the
sensor electronics and 30 years or more for the sensor heads [Appendix A]. The vendors
specify that their OCTs require no maintenance or a low level of maintenance [46], [47],
[Appendix A]. According to the vendor of OCT 2, the entire OCT does not have to be
replaced if a single sensor head or the MU fails. Any sensor head or MU can be replaced
[Appendix A].
4.6 Weight
The total weight of one of the OCTs is estimated to be approximately 160 kg for a
maximum system voltage rating of 420 kV. The weight of the sensor head is 15 kg.
Another vendor specifies a total weight of 65 kg for an OCT with a maximum system
voltage rating of 245 kV [Appendix A]. Information on the approximate weights of the
other OCTs could not be obtained.
43
4.7 Price
One of the vendors estimates the price of a three-phase OCT with a maximum system
voltage rating of 420 kV to be approximately 45 000 €. This includes a set of three sensor
heads and one MU. Once industrialized, this OCT is expected to be more cost-effective
than CCTs for voltage levels higher than 245 kV [Appendix A]. Information on the prices
of the other OCTs could not be obtained.
4.8 Overview
An overview of the specifications of the different OCTs which were considered in this
chapter is given in Table 4.1.
Table 4.1: Overview of the specifications of some commercially available OCTs [49], [48],
[46], [47], [Appendix A].
OCT 1 OCT 2 OCT 3 OCT 4

Maximum
245 kV-800 kV 145 kV-550 kV 72.5 kV-800 kV 123 kV-800 kV
System Voltage
Rated Primary
2 kA-4 kA Up to 2.5 kA Up to 4 kA Up to 5 kA
Current
Rated Dynamic
164 kA 187.5 kA 210 kA 200 kA
Current
Measuring
0,2 S 0,2 S 0,2 S Up to 0,2
Accuracy Class
Protective
5 P, 5TPE 5P 5P Up to 5P
Accuracy Class
Time
1PPS 1PPS, PTP 1PPS, IRIG-B 1PPS
Synchronization
Network
Not given PRP, HSR Not given Not given
Redundancy
Temperature
Range of -40°C-+45°C -40°C-+85°C -40°C-+55°C -40°C-+40°C
Sensor Head
Temperature
Range of
Not given -5°C-+55°C -5°C-+40°C Not given
Sensor
Electronics
44
Chapter 5
Stand-Alone Merging Units
This chapter treats SAMUs. Section 5.1 describes the purpose of applying SAMUs in
digital substations. In Section 5.2, the working principle of SAMUs is described and in
Section 5.3 the steady state and transient performance of SAMUs are treated.
5.1 Purpose
The application of SAMUs allows CITs to be integrated into digital substations, as shown
to the right in Figure 5.1. The figure compares the hardwired connections of CITs in
conventional substations to the connection of CITs to the process bus by means of SAMUs
in digital substations [36]. SAMUs convert the analog signals from CCTs and CVTs to
digital form and publish the measurements as SV on the process bus in accordance with
IEC 61850-9-2 LE [6]. This makes the measurements from CITs accessible to all of the
secondary devices which are connected to the process bus. SAMUs are equipped with
analog inputs where the secondary circuits of CITs can be connected [36]. Different
vendors offer slightly different solutions in terms of, among other things, the number of
analog inputs, support of time synchronization and the placement of the SAMU indoor
in the control room or in an outdoor cubicle [48], [97], [98].
Figure 5.1: Connection of CITs to secondary IEDs [36]. © 2011 IEEE

a. Hardwired connections in a conventional substation.
b. Connection to the process bus via a SAMU in a digital substation.
45
CHAPTER 5. STAND-ALONE MERGING UNITS
5.2 Working Principle

The working principle of a SAMU is as follows. First, the analog input signals have to
be scaled down to levels which are appropriate for the electronics [44]. Different manu-
facturers may employ different types of analog input elements for this purpose. For the
current input elements, shunts, Hall elements, Rogowski coils or various types of CCTs
may be used [27]. The analog signals are then filtered by an anti-aliasing filter in order
to avoid aliasing of the input signal. Following the filtering, the analog signals are peri-
odically sampled and converted from analog to digital form. The above-mentioned steps
may introduce nonlinearities in the magnitude and phase angles of the signals as well
as time delays. In order to remove these, digital filtering and calibration is performed.
The digital data may subsequently be re-sampled in order to provide an output with the
correct sample rate [44]. Typically the sample rate of the output is 80 or 256 samples per
cycle, as defined in IEC 61850-9-2 LE [56].
In order to synchronize the sampling, SAMUs are equipped with an internal clock, which
shall be synchronized to an external time source [33]. For this purpose, SAMUs typically
receive 1PPS signals via dedicated wiring or is synchronized from the process bus by the
use of PTP [7]. SV messages have a sample counter attribute called SmpCnt with a value
which is set by the merging unit and used to time stamp SV. Take a sample rate of 80
samples per cycle, that is 4000 Hz, as an example. The SmpCnt of the first SV in each
second will be given a value of 0000. For each successive SV within the same second, the
value of the SmpCnt attribute will be increased by one until a value of 3999 is reached.
The SmpCnt of the next SV is set to 0000 and the counting continues [44]. The SAMU
encodes the digital data as SV and publish the messages to the process bus in accordance
with IEC 61850-9-2 LE [6]. The subscribing IEDs make use of the SmpCnt attribute in
combination with their own time synchronization to align the received SV in time [99].
5.3 Steady State and Transient Performance

Dutra et al. [100] compared the performance of a SAMU to the performance of a conven-
tional analog acquisition system by the use of a digital fault recorder. The analog outputs
of a relay test set were connected directly to the analog inputs of the disturbance recorder
and via the SAMU. There were only small differences in the amplitude and phase angles
of the recorded currents and voltages from the analog acquisition system and the SAMU
when nominal currents and voltages were applied. The response of the SAMU and the
analog acquisition system to transitory signals were also indistinguishable. It was further
found that the SAMU had a linear output [100].
46
CHAPTER 5. STAND-ALONE MERGING UNITS
The transient performance of SAMUs will among other things depend on the chosen
technology for the analog input elements. The current input elements could be shunts,
Hall elements, Rogowski coils or various types of CCTs, as already mentioned. In present
day protection IEDs, the current input elements are typically CCTs. These input current
transformers could for instance be of the LR or NR types, which are designed with small
and large air gaps in the CT cores, respectively. As mentioned in Section 2.2, air gaps
in the CT core reduce the chance of saturation, but also result in a CT which cannot
properly reproduce the DC component in transient primary currents. How quickly the
DC component is damped out depends on the size of the air gaps [27]. As mentioned in
Subsection 2.4.3, Part 13 of IEC 61869, which will apply to SAMUs once it is released,
will recommend using non-saturable input elements in future SAMUs, that is other types
of input elements than CCTs [60].
The typical waveform of a saturated secondary current from a CCT was shown in Figure
2.4 in Section 2.2. In the figure it can be seen that the saturated current has a very
distinct waveform, which is easy to recognize. Recognizing the saturated current of input
current transformers, on the other hand, may not be as simple. When a CCT saturates,
most of the secondary current will flow through the magnetizing impedance for a period of
time each cycle, as explained in Section 2.2. This is due to the fact that the magnetizing
impedance will have a value which is considerably lower that the value of the secondary
circuit impedance.
However, when it comes to the input current transformers in SAMUs or protection IEDs,
the secondary circuit impedance is typically very small. Thus, a much larger current may
flow through the secondary winding during saturation, and the measured current will not
have the typical saturated waveform. It can be difficult to see the difference between the
waveform of a saturated current from an input current transformer and the waveform of
a secondary current from a CT which cannot reproduce the DC component in transient
currents properly [27]. Holst and Zakonjšek [27] have presented examples of secondary
current waveforms from HR, LR and NR types of CTs and a comparison of a saturated
current waveform from an input current transformer and a current waveform from a CT
which cannot reproduce DC components properly.
47
Chapter 6
Considerations in Digital Substations
In this chapter some considerations regarding the use of OCTs and other NCITs as well as
SAMUs in digital substations are treated. Section 6.1 points out the differences between
the measurement chains in conventional and digital substations, and Section 6.2 describes
some of the requirements which will apply in digital substations. In Section 6.3, Section
6.4, Section 6.5 and Section 6.6 some considerations regarding the use of differential
protection, protection IEDs, energy meters and fiber optic cables in digital substations
are described, respectively.
6.1 Measurement Chain

There are some important differences between the measurement chains in digital and
conventional substations. In conventional substations, the hardwired cables from the high
voltage bays, which are used for measurement and control circuits, are directly connected
to the analog and binary inputs of the protection, control and monitoring devices in the
control building. In other words, the secondary devices are hardwired to specific pieces of
primary equipment [100]. The conversion from analog to digital form of the current and
voltage measurements takes place in the respective protection, control and monitoring
devices.
In digital substations, on the other hand, the analog to digital conversion of current and
voltage measurements has been moved from the protection, control and monitoring devices
to MUs associated with NCITs and SAMUs. The current and voltage measurements from
instrument transformers are converted to digital form by MUs or SAMUs in the case
of NCITs and CITs, respectively. The measurements are subsequently published as SV
streams to the process bus, which the secondary devices can subscribe to [100].
49
CHAPTER 6. CONSIDERATIONS IN DIGITAL SUBSTATIONS
6.2 Requirements in Digital Substations
6.2.1 Requirements of the Protection Systems
In conventional substations, the protection systems have some requirements for the analog
measuring chain in order to ensure satisfactory performance of the protection functions.
Examples are requirements for the protective CT cores and the power which is provided
by the CCTs in the secondary circuits. This subsection aims to point out some of the
requirements which protection systems will have for the digital measuring chains in digital
substations.
Transient Performance of NCITs and SAMUs
The transient performance of SAMUs and NCITs with associated MUs may affect the
performance of the protection functions in digital substations. The transient performance
of NCITs is specified in IEC 60044 and IEC 61869. Together with IEC 61850-9-2 and IEC
61850-9-2 LE, theses standards define the digital measuring chain of NCITs, which is of
great importance in order to facilitate the realization of multivendor digital substations
[18].
At the present time, manufacturers of protection IEDs specify requirements for the dimen-
sioning of CCTs in accordance with the guidelines given in IEC 60255, "Measuring relays
and protection equipment". For protection IEDs subscribing to SV over the process bus,
on the other hand, guidelines on how to specify requirements for NCITs have not been
defined. A possible way for the manufacturers of protection IEDs to define requirements
for NCITs is to specify requirements for the frequency response of the associated MUs,
similar to those which are defined in IEC 61869-6 [101]. IEC 61869-6 uses a frequency
mask concept where prohibited regions of MUs’ frequency response have been specified.
MUs are required to operate within the defined frequency mask limits. However, these
limits are are not particularly strict. For instance, IEC 61869-6 permits a low cut-off fre-
quency of up to 1 Hz. This may cause some challenges for the manufacturers of protection
IEDs [60].
In 2016 IEC Technical Committee 95, "Measuring Relays and Protection Equipment",
(IEC TC 95) decided to create a new Ad hoc group (AHG 3) called "Use case of digital
sampled values instead of analog input" [101]. This group is evaluating the requirements
for measuring relays and protection equipment in relation to the requirements for instru-
ment transformers [Appendix A]. One of the issues which is under consideration is if and
how manufacturers of protection IEDs should define requirements for NCITs in order to
50
ensure satisfactory performance of the protection functions [101].
Part 13 of IEC 61869, which will apply to SAMUs, is a work in progress and has not
yet been released, as mentioned in Subsection 2.4.3. At the present time, the transient
performance of SAMUs is thus not standardized. It is therefore important to verify
that SAMUs and protection IEDs from different vendors are interoperable in terms of
transient performance [18]. As mentioned in Section 5.3, different manufacturers may
employ different types of analog input elements in SAMUs, such as shunts, Hall elements,
Rogowski coils or various types of input current transformers with different transient
performances. Combining SAMUs which have input elements with different capabilities
of reproducing DC components and low frequency components in the primary current
could potentially lead to incorrect operation of protection functions, especially differential
protection functions [27].
Time Delays
In digital substations, the total fault detection time is directly affected by the processing
delay time of merging units, the transfer time of SV over the process bus and the com-
munication stack processing time in the subscribing protection IEDs. These times must
be added to the fault detection time of the protection IEDs in order to estimate the total
fault detection time, which in turn will affect the total fault clearance time [44], [52],
[18]. The different components of the total fault clearance time in a digital substation are
illustrated in Figure 6.1.
Figure 6.1: Total fault clearance time in a digital substation [18]. © 2016 IEEE
Since protection functions are time-critical functions, requirements are put on the maxi-
mum processing delay time of merging units and the maximum transfer time of SV. Ap-
plication specific requirements for the maximum processing delay time of merging units
51
are defined in IEC 61869-9. For protective and measuring applications the maximum
processing delay time is 2 ms [33].
Requirements for the transfer time of SV have also been specified. These are found
in IEC 61850-5, where different transfer time classes and different message types with
associated performance classes are defined. For SV (message type 4), two performance
classes are defined, P7 and P8. SV which are intended for protection applications belong
to performance class P7 and must satisfy transfer time class TT6 with a maximum transfer
time of 3 ms [102].
As shown in Figure 6.1, the transfer time of a SV message ((tTT1 ) comprises of the
communication stack processing time of the merging unit (ta1 ), the network transfer time
over the process bus (tb1 ) and the communication stack processing time of the receiving
protection IED (tc1 ) [52], [18], [102]. The communication stack processing times of the
merging unit and the protection IED shall amount to a maximum of 40 % of the total
transfer time each. Hence, 20 % or more of the total transfer time is left for the network
transfer time. For protection applications this results in a maximum communication stack
processing time in merging units of 1.2 ms and a network transfer time of 0.6 ms [52],
[18], [103].
Time Synchronization
In digital substations, application functions which are dependent on current and/or volt-
age measurements from several sources require highly accurate time synchronization. Ex-
amples of such applications are differential protection and distance protection [102]. SV
from different merging units cannot be aligned unless they are synchronized in time [45].
In addition, the accuracy of the sampling synchronization directly affects the accuracy
of the signal processing in protection IEDs and thus the performance of the protection
functions [104], [52]. Inaccuracies in the sampling synchronization give phase errors in
the measured data, which could for instance result in incorrect operation of differential
protection functions [45], [105].
IEC 61850-5 defines time synchronization classes for the synchronization of IEDs. The
time synchronization classes T4 and T5 have the highest accuracy requirements, 4 µs and
1 µs, respectively. In a 50 Hz system, a time synchronization inaccuracy of 4 µs will result
in a phase angle error of 0.1°. The resulting phase angle error of a time synchronization
inaccuracy of 4 µs is 0.02° [102].
According to IEC 61850-9-2 LE, the sampling synchronization accuracy in a merging

unit should satisfy the requirements of time synchronization class T4 [56]. Sampling
52
synchronization with an accuracy of ±4 µs will be sufficient for most protection and control
applications [52]. The accuracy requirement also applies for a specified amount of time
in a potential situation where the merging unit looses the external time synchronization
[52]. This time is specified to be minimum 5 s in IEC 61869-9 [33], [52]. SV messages have
a SmpSynch attribute which contains information regarding the quality of the sampling
synchronization [33]. How measuring relays and protection equipment should behave in
case they receive SV which are not time synchronized, is one of the issues IEC TC 95
AHG 3 is presently working on [101].
In order to achieve highly accurate time synchronization in a digital substations, different

methods can be applied. In IEC 61869-9-2 LE it is stated that merging units shall be
synchronized with a 1PPS input [56]. IEC 61869-9, on the other hand, recommends using
PTP for time synchronization [33]. The distribution of 1 PPS signals requires dedicated
wiring, while PTP is a network-based protocol, which allows time synchronization signals
to be distributed over the process bus [106], [107]. In any case, the time synchronization
source shall provide a time signal with an accuracy which is better than ±1 µs, as defined
in both IEC 61850-9-2 LE [56] and IEC 61869-9 [33].
Sampled Values Message Integrity
SV messages that are malformed, missing or arrive at the protection IEDs out of sequence
may affect the performance of the protection functions in digital substations [44]. Up to a
certain limit, protection IEDs are typically still able to operate correctly if a some of the SV
are lost [108]. Chen, Guo and Crossley [109] experimentally tested the impact of SV loss
on the performance of two protection IEDs from different vendors. Six tests were carried
out in total where the protection IEDs subscribed to SV generated by merging units
from three different vendors, respectively. The results showed that the protection IEDs
were affected differently by the loss of SV and that the type of merging unit generating
the SV also had an impact. For one of the protection IEDs, the loss of SV resulted
in unstable readings of the voltage. At certain loss rates, the protection IED stopped
issuing trip signals. For the other protection IED, the loss of SV had different impacts on
the protection IEDs within specific loss rate intervals. Based on these results it can be
expected that different protection IEDs will have different limits for the acceptable loss
rate of SV. [109]. How measuring relays and protection equipment should behave if SV
are lost is another issue IEC TC 95 AHG 3 is presently working on [101].
53
Process Bus Network
A well functioning process bus network is critical for the performance of the protection
functions in digital substations since the process bus enables the time-critical transmission
of the required SV and protection trip GOOSE messages. The process bus network should
therefore be designed in such a way that it does not limit the performance of the protection
functions in digital substations. This puts requirements on, among other things, the
network bandwidth, speed of transmission and availability of the network [110], [45]. The
availability of the protection functions in a digital substation is directly affected by the
availability of the process bus network. Thus, the process bus network has to satisfy
requirements of high availability [104], [111]. In order to satisfy these requirements, the
process bus network should provide bumpless redundancy. This means that there has
to be an alternative path for the data traffic in the network and that the switchover
from one path to the other has to take place without any time passing. Zero switchover
time is achieved by having both paths in operation during normal operating conditions.
IEC 61850 recommends the use of either PRP or HSR protocols to achieve bumpless
redundancy [104], [7].
A poorly designed process bus network can for instance result in unacceptably long trans-
fer times of SV or loss of SV [108], [44]. Multiple SV streams generate quite a lot of
continuous data traffic in the process bus. In addition, there is other data traffic, such
as GOOSE messages and time synchronization data. Designing the process bus with a
sufficient bandwidth and employing network traffic management are some of the keys to
achieve a well functioning process bus network [112], [104]. Statnett has designed the
process bus network in their pilot digital substation with a bandwidth of 1 Gbps, which is
also the bandwidth planned for future digital substations. A bandwidth of 5 Gbps would
also be possible. The connection of the protection IEDs to the network is typically limited
to 100 Mbps [Appendix A].
Interoperability
An important requirement in digital substations is interoperability between secondary

IEDs, SAMUs and NCITs with associated MUs. Interoperability between devices from
the same vendor is typically tested by the manufacturer in order to make sure that the
different devices will be interoperable in a system. Interoperability testing between devices
from different vendors, on the other hand, is typically more limited. Conformity with the
IEC 61850 standard shall ensure interoperability between devices from different vendors.
However, standards may be interpreted differently [108].
An example of standards being interpreted differently is the handling of the SV quality
54
attribute in merging units. The quality attribute contains information regarding the
quality of the SV. Aune [113] performed research on how the quality attribute of SV
are handled in digital substations [113]. IEC 61850-7-3 leaves some degrees of freedom in
terms of setting the validity of the SV to questionable or invalid based on how some detail
qualities, such as outOfRange and inaccurate, are set. Additional specifications for how
this should be handled are given in IEC 61869-9. However, Aune’s work, which was based
on participation in a factory acceptance test and contact with another vendor, showed
that different vendors have interpreted the handling of quality attributes differently. In
addition, it has not yet been defined in any of the standards how the subscribing IEDs
should respond to the quality declaration of the received SV [113], [33], [114]. This is
another issue IEC TC 95 AHG 3 is presently working on [101].
6.2.2 Requirements of the Optical Current Transformers
According to one of the vendors which was contacted during the product survey on OCTs,
their OCT does not pose any specific requirements for the surrounding systems in order
to perform satisfactory, except for the following. There are certain requirements for the
connection of the MU to the process bus in terms of physical interface, type of optical
connectors and type of fiber optic cables [Appendix A]. IEC 61869-9 recommends using
duplex LC type of optical connectors [33]. However, different IEDs, MUs and SAMUs
may be equipped with different types of optical contacts. This was experienced during the
laboratory tests which were performed as a part of this thesis. Chen, Guo and Crossley
[108] also reported that the IEDs used in an interoperability performance assessment of
a multivendor process bus had optical contacts which were not compatible with those
of other vendors [108]. Hence, different types of fiber optic patch cords may have to be
used in order to connect the devices in a digital substation. Another requirement is that
the protection, control and monitoring devices have to comply with IEC 61850-9-2 LE or
possibly IEC 61869-9 in order to receive the SV streams from the MUs [49], [Appendix
A].
6.2.3 EMP and EMI Requirements
As mentioned in Section 4.2, different suppliers of OCTs have different solutions for the
location of the associated MUs, either indoor installation in a protection and control panel
or in an outdoor cubicle. Statnett is planning to employ a flexible solution with regards
to the placement of the MU indoor or outdoor [Appendix A]. As mentioned in Section
2.5, there are regulatory requirements for the protection against Electromagnetic Pulse
(EMP) and Electromagnetic Interference (EMI) in transmission substations. Electronic
55
components which are to be placed outdoor in the substation yard, have to be installed
in cabinets which are secured againts EMP and EMI [4]. This applies to both the cabinet
itself and the cable feedthroughs.
6.3 Application of Differential Protection

When applying differential protection in digital substations, there are some new factors
which have to be considered compared to conventional solutions with analog inputs to
the protection IEDs. First of all, as already mentioned, inaccuracies in the sampling
synchronization of merging units could potentially cause incorrect operation of differential
protection functions. It is thus important that the sampling synchronization is sufficiently
accurate. Ingram, Schaub, Taylor and Campbell investigated how errors in the sampling
synchronization of merging units influence the performance of transformer differential
protection. Errors from 4 µs to 1000 µs were introduced in the time synchronization
signal to the merging units under test. The test results showed that the transformer
differential protection operated at lower differential currents than specified by the operate-
restrain characteristic when synchronizing errors of 500 µs and 1000 µs were introduced. A
synchronizing error of 100 µs did not affect the performance of the transformer differential
protection [112].
An example of an operate-restrain characteristic of a conventional transformer differential

protection is shown in Figure 6.2. If the differential current, which is the sum of the
current phasors from the CTs on each side of the transformer, is above the operate-
restrain characteristic and the operation of the differential protection function has not
been blocked, the differential protection will trip [115]. The operate-restrain characteristic
comprises of three sections with different purposes. The purpose of section 2 and 3 in the
figure is mainly to prevent the differential protection from operating due to inaccuracies
caused by CCTs. The performance of the CCTs on each side of the transformer may not be
exactly the same, especially when the current level is high. The third section has a steeper
slope in order to prevent the differential protection from operating due to CT saturation
during external faults. Since OCTs are linear over a wide dynamic range and cannot
saturate, the slopes of the operate-restrain characteristic can be significantly reduced in
digital substation. The result is a differential protection with a higher sensitivity [18].
56
Figure 6.2: Operate-restrain characteristic of a transformer differential protection [112].

© 2013 IEEE
However, the maximum detectable setting current of OCTs has to be taken into consid-
eration. When applying OCTs for line differential protection, it is important to ensure
that the maximum detectable currents of the OCTs are equal. Unwanted operations of
the differential protection may occur if the short-circuit current exceeds the maximum
detectable current of one or both of the OCTs since this will cause unequal distortion of
the currents which are measured. For line differential protection, the maximum detectable
currents of the OCTs do not have to exceed the maximum possible short-circuit current
as long as the maximum detectable currents of the OCTs are equal. If the measured cur-
rents are distorted, they will be equally distorted. When applying OCTs for busbar and
transformer differential protection, on the other hand, the maximum detectable currents
of the OCTs should exceed the maximum possible short-circuit current since different CT
ratios are involved [50].
In future substations, differential protection may be employed in schemes which involves

combinations of NCITs and SAMUs. As mentioned in Subsection 2.3.2, the migration
from the present day conventional substations to digital substations will be a gradual
process, and CITs are likely to be replaced on a step-wise basis. In addition, a continuous
use of CITs may be desirable in some parts of substations due to lack of commercially
available OCTs specified for the lowest voltage levels. An example of such a hybrid
solution could be a transformer differential protection with an OCT on the High Voltage
(HV) side of the power transformer and a SAMU on the LV side. In such a scheme, it
is of great importance to make sure that any differences in the transient response of the
different NCITs and SAMUs will not affect the performance of the differential protection
functions [50], [99]. Preferably the transient responses should be similar and the currents
from different MUs and SAMUs have to be accurately synchronized [50].
57
6.4 Application of Protection IEDs

At the present time, there are many vendors which supply protection IEDs supporting IEC
61850-9-2 LE process bus [Appendix A]. Statnett uses frame agreements with suppliers of
protection IEDs and requires two separate protection IEDs supplied by different vendors
for distance protection as a redundancy measure [116]. Currently, Statnett has a frame
agreement with two suppliers of protection IEDs. Both of these offer protection IEDs
which support IEC 61850-9-2 LE process bus. However, these protection IEDs are newer
versions of protection IEDs which have been approved by Statnett. These versions will
thus have to be approved before they can be applied in digital substations. In addition,
it is uncertain whether good solutions for busbar protection in bigger substations are
presently available [Appendix A].
6.5 Application of Energy Meters

There are regulatory requirements which apply to the electrical energy metering in trans-
mission substations. An accredited meter inspection has to be performed by an approved
accreditation company after the electrical energy meter has been put into operation in or-
der to control the accuracy. There is currently an electrical energy meter available which
can subscribe to SV measurements from an IEC 61850-9-2 LE process bus. However, so
far a solution to how accredited meter inspections can be performed in digital substations
has not been worked out. The issue is being discussed both internationally and within
Statnett [4], [Appendix A].
6.6 Application of Fiber Optic Cables

The application of fiber optic cables reduces the need for large cable trenches. In digital
substations, the large number of copper cables for measurement and control circuits, which
run from the control building to the high voltage bays will be replaced by a process bus,
which can be realized by a smaller amount of fiber optic cables [4]. Fiber optic cables
can be laid in pipes or smaller cable trenches. The installation can be performed by
blowing or pulling the cables into the pipes. It is expected that the cost of cabling and
cable trenches in addition to the cost of installing the cables will be reduced in digital
substations. However, special competence will be needed in order to connect the fiber
optic cables [Appendix A]. Fiber optic cables are not conducting and will thus provide
galvanic isolation between the connected devices [19]. In addition, fiber optic cables are
not sensitive to EMI, in contrast to copper cables [117].
58
Chapter 7
Laboratory Test Setup and Method
In this chapter the test setup and method of the laboratory tests which were conducted
in the ProSmart lab at NTNU during the spring semester of 2018 is presented. In Section
7.1 the equipment and test setup is described. Section 7.2 presents the simulations which
were performed in ATPDraw and Section 7.3 describes the test method.
7.1 Test Setup

A SAM600 SAMU was tested in a laboratory test setup emulating the hybrid mode con-
figuration of a transformer differential protection which is shown in Figure 7.1. The reason
behind the choice of this particular configuration is that it allows the steady state and
transient performance of the SAM600 SAMU to be investigated and compared to those
of a conventional analog acquisition system. In addition, the test setup allows the perfor-
mance of a transformer differential protection (RET670) in hybrid mode configuration to
be investigated.
Figure 7.1: Transformer differential protection in hybrid mode configuration.
59
CHAPTER 7. LABORATORY TEST SETUP AND METHOD
The laboratory test setup is illustrated in Figure 7.2.
Figure 7.2: Laboratory test setup.
The test setup comprises of a OMICRON CMC 356 test set, three SAM600 SAMU mod-
ules, a SEL-2488 network clock and a RET670 transformer protection IED. In the follow-
ing subsections the equipment and connections will be described in further detail.
7.1.1 OMICRON CMC 356 Test Set
The OMICRON CMC 356 test set has has six current sources, each with a maximum
output of 32 ARMS (45 Apeak ). For connected loads with a resistance Rload ≤ 0.5 Ω, the
accuracy of the current outputs is given by Equation (7.1) [118].
Error < 0.05 % of reading value + 0.02 % of upper range value (7.1)
60
In this test setup the six current sources were configured as two three-phase current
sources. The test set has an associated software, OMICRON Test Universe, which com-
prises of various test modules. By the use of the Quick CMC test module, the test set can
generate symmetric steady state AC currents. The Advanced TransPlay test module can
be used to reproduce current waveforms from an actual fault recording or a simulation
[119].
7.1.2 SAM600 Stand-Alone Merging Unit
SAM600 is a modular SAMU, which encompasses three types of modules, SAM600-TS,

SAM600-CT and SAM600-VT, each with its own functionality [97]. This test setup
included one of each of the three types of SAM600 modules. The modules can be connected
in a chain or a ring to form a SAM600 system. Such a system may comprise of up to 10
modules, depending on the requirements for analog interfaces, communication ports and
time synchronization [97].
The SAM600-CT and SAM600-VT modules are analog input modules used to interface
CCTs and CVTs, respectively. Each analog input module has four analog input channels
and four Ethernet communication ports, two of which are optical, the other two electrical.
The SAM600-CT module can be ordered with an input current rating of 1 ARMS or 5 ARMS
[97]. The SAM600-CT module which was used in this test setup has an input current
rating of 5 ARMS . The analog input elements of the SAM600-CT module are shunt
resistors, and the measurement range is up to 80 times the rated current. The rated
input of the SAM600-VT module can be set in the range from 100 V to 125 V. Neutral
current and voltage can be configured to be measured or calculated by the analog input
modules [97]. In this test setup the neutral values were set to be calculated by the analog
input modules. The SAM600 system was configured by the use of ABB’s Protection and
Control IED manager software, PCM600.
The SAM600-TS module does not have any analog input channels and is an optional part
of a SAM600 system. The purpose of the SAM600-TS module is to provide 1PPS time
synchronization functionality and additional access points for connecting the SAM600
system to the process bus. The modules in a SAM600 system are time synchronized
internally by the use of PTP. Any of the modules may be connected to a PTP synchronized
network and used to synchronize the SAM600 system to an external time source. However,
if 1PPS time synchronization is desired, a SAM600-TS module is required. The SAM600-
TS module has one 1PPS input and five 1PPS outputs. The 1PPS input can be used to
synchronize the SAM600-TS module and thereby the rest of the SAM600 modules in a
SAM600 system to an external time source. The 1PPS outputs can be used to synchronize
61
other devices, such as secondary IEDs, to the SAM600 system [97].
7.1.3 SEL-2488 Satellite-Synchronized Network Clock
The SEL-2488 satellite-synchronized network clock has an associated antenna, which re-
ceives Global Navigation Satellite System (GNSS) time signals. Based on these time
signals, the network clock can supply precise time, which can be used to synchronize dif-
ferent devices. The SEL-2488 network clock provides several options for the distribution
of time. SEL-2488 is equipped with eight configurable BNC ports, which can provide
IRIG-B, PPS or kPPS time signals. The network clock also supports network-based
time distribution. The clock has four Ethernet ports supporting Network Time Protocol
(NTP) and PTP. In this test setup, the network clock was configured to use PTP for time
synchronization, providing a time stamp accuracy of ±100 ns [120].
7.1.4 RET670 Transformer Protection IED
The RET670 transformer protection IED which was used in this test setup has 6 input
current transformers with a rated current of 1 ARMS . The nominal range of the input
current transformers is 0.2 to 40 times the rated current [121]. The RET670 protection
IED was configured as a transformer differential protection for a two winding transformer
by the use of PCM600. The transformer differential protection obtains the fundamental
frequency differential currents in a phase-wise manner. For each phase the fundamen-
tal frequency differential current is calculated as the sum of the fundamental frequency
current phasors from the LV and the HV side of the protected power transformer. The
highest fundamental frequency current after taking the phase shift and ratio of the power
transformer into account, is chosen as the restrain current. All fundamental frequency
phase currents are referred to one side of the protected power transformer, typically the
HV side [115].
The operate-restrain characteristic of the transformer differential protection is given in

Figure 7.3 and the associated settings in Table 7.1. These settings are the default set-
tings. Idmin is the sensitivity of Section 1, IdUnrestrained is the unrestrained protection limit
and I2/I1 is the maximum ratio of the second harmonic component to the fundamental
frequency differential current [115]. The base current, IBase , was set equal to 412 A, which
is the rated HV side current of a 420 kV/132 kV power transformer with 300 MVA rat-
ing. The RET670 protection IED has a disturbance recorder function, which was used to
record the currents from the SAMU and the analog acquisition system. The associated
software, Wavewin, was used to analyze the recorded currents.
62
11
Operate unconditionally
10
9
Section 1 Section 2 Section 3
8
Operate Current [ IBase]
6
Operate conditionally
5
3
Restrain
2
0
0 1 2 3 4 5 6 7 8 9 10 11
Restrain Current [ IBase]
Figure 7.3: Operate-restrain characteristic of RET670 transformer differential protec-

tion. The figure is adapted from [115].
Table 7.1: Settings of RET670 transformer differential protection.
Data Value
End of Section 1 [×I Base ] 1.25
End of Section 2 [×I Base ] 3
IdMin [×I Base ] 0.30
Slope of Section 2 [%] 40
Slope of Section 3 [%] 80
IdUnrestrained [×I Base ] 10
I2/I1 [%] 15
7.1.5 Connections
As shown in Figure 7.2, the three SAM600 modules (SAM600-CT, SAM600-VT and
SAM600-CT) were connected by Ethernet cables in a chain to form a SAM600 system.
Each of the two sets of three-phase current outputs of the OMICRON CMC 356 test set
were connected in a star configuration to the current inputs of the SAM600-CT module
and the RET670 protection IED, respectively. The process bus represented an ideal pro-
cess bus and was realized as a direct fiber optic link between the SAM600-TS module and
the RET670 protection IED. For this particular test setup, several time synchronization
methods and configurations were possible. The following solutions for time synchroniza-
tion were considered. These are also illustrated in Figure 7.4.
63
1. Network-based time synchronization by the use of PTP.

2. Time synchronization by the use of 1PPS signals from SEL-2488.
3. Network-based time synchronization of SAM600-TS by the use of PTP. 1PPS signals
from SAM600-TS is used to synchronize RET670.
4. Time synchronization of RET670 by the use of a GPS antenna. 1PPS signals from
RET670 is used to synchronize SAM600-TS.
Figure 7.4: Possible time synchronization solutions.
The simplest and preferred choice for time synchronization of the devices in the test setup
is the first solution. However, when using PTP for process bus time synchronization, every
device which is connected to the network must support PTP [104]. Since the RET670
protection IED which was used in this test setup does not support PTP, PTP could only
be used to synchronize the SAM600-TS module to the SEL-2488 network clock and to
synchronize the three SAM600 modules to each other.
The second option required additional time synchronization modules to be installed in the
RET670 protection IED since these were not initially present. In addition, it turned out
that the 1PPS output contacts of the SEL-2488 network clock were not directly compatible
with the 1PPS input contacts of the RET670 protection IED and the SAM600-TS module.
The SEL-2488 network clock is equipped with electrical output contacts for for 1PPS time
synchronization signals, whereas the RET670 protection IED and the SAM600-TS module
64
both have optical input contacts. Hence, an electrical to optical converter would have been
needed, possibly introducing additional errors.
By the use of the third option, the issues of non-compatible 1PPS contacts were solved.
The SAM600-TS module was connected to the SEL-2488 network clock by an Ethernet
cable and synchronized by the use of PTP. The RET670 protection IED was connected
to the SAM600-TS module with an additional fiber optic cable for time synchronization
and synchronized by the use of 1PPS signals.
Another time synchronization method which was tested out initially is the fourth solution.
This solution is not preferred since it differs significantly from the typical configuration
in a digital substation. However, this solution was the only feasible solution before the
time synchronization modules were installed in the RET670 protection IED. The method
was tested out successfully, but was later replaced in favor of the third method following
the installation of the tims synchronization modules.
7.2 ATPDraw Simulations

In order to test the SAMU in the transformer differential protection configuration shown
in Figure 7.1 under different operating conditions, models for five simulation cases were
created in ATPDraw. The simulation cases are listed below.
Simulation Case 1: Normal Load Conditions

Simulation Case 2: Internal High Voltage Side Fault
Simulation Case 3: Internal Low Voltage Side Fault
Simulation Case 4: External Fault
Simulation Case 5: Current Transformer Saturation
Simulation Case 6: Transformer Inrush
The basic simulation model and the different simulation cases will be described in the
following subsections.
7.2.1 Basic Simulation Model
Figure 7.5 shows the basic simulation model, which represents normal load conditions.
This model provides the basis for the development of the models used in the other simu-
lation cases. The basic simulation model comprises of a supply grid, a power transformer,
CTs located on each side of the power transformer and a load.
65
Figure 7.5: ATPDraw basic simulation model and simulation model of simulation case
1: normal load conditions.
Supply Grid
The supply grid is modelled as a three-phase AC voltage source in series with a symmetric
three-phase RL coupled line. The rated line-to-line voltage of the source is 420 kVRMS
and the rated frequency 50 Hz. The short-circuit capacity of the supply grid was initially
set to 10 000 MVA. However, in some of the simulation cases, the short-circuit capacity
had to be reduced due to the limited maximum current output of the OMICRON CMC
356 test set. The positive and zero sequence impedances of the RL coupled line were
calculated based on the short-circuit capacity of the supply grid in each simulation case.
The reactance of the line was assumed to be ten times the resistance of the line and the
zero sequence impedance three times the positive sequence impedance. The formulas used
in these calculations and the resulting impedances are given in Appendix B.1.
Power Transformer
The power transformer is modelled as a three-phase three-winding hybrid transformer

with a three-leg stacked core. The ratings and connections of the power transformer are
given in Table 7.2. The transformer is modelled based on data of a real transformer.
Table 7.2: Ratings and connections of the power transformer.
Data Primary (P) Secondary (S) Tertiary (T)

Line-to-Line Voltage [kV] 420 132 12
Apparent Power [MVA] 300 300 100
Connection Y Y D
Phase shift [Degrees] 0 0 330
66
The leakage inductance and winding resistance of the power transformer is determined
based on the short-circuit test results given in Table 7.3.
Table 7.3: Short-circuit test results for the power transformer.
Data P-S P-T S-T

Impedance [%] 15 8 16
Apparent Power [MVA] 300 100 100
Loss [kW] 1000 200 250
The magnetization and losses of the legs and yokes are determined based on the results
from a no-load test performed at the secondary side of the power transformer. The no-load
test results are given in Table 7.4.
Table 7.4: No-load test results for the power transformer.
Voltage [%] Loss [kW] Current [%]

90 75 0.03
100 100 0.04
110 125 0.1
Current Transformers
Each CT is modelled as an ideal CT by the use of a current probe whose output is

divided by the rated CT ratio. For the hybrid mode configuration of the transformer
differential protection shown in Figure 7.1, the current probe on the HV side of the power
transformer represents the CT which is hardwired to the RET670 protection IED, while
the current probe on the LV side of the power transformer represents the CT which is
connected to the SAM600 SAMU. The rated secondary currents of the two CTs are set
equal to the rated input currents of the RET600 protection IED and the SAM600 SAMU,
respectively. The rated primary currents of the CTs are determined based on the rated
power and voltages of the power transformer, which are given in Table 7.2. The formulas
used in these calculations are given in Appendix B.2. The current transformer on the HV
side of the power transformer has a ratio of 500 A/1 A and the current transformer on
the LV side has a ratio of 1500 A/5 A.
67
Load
The load is modelled as a three-phase pure resistive linear branch element. The load is kept
constant in all simulation cases and equal to the rated power of the power transformer
(300 MW). An initial value for the resistance of the load was calculated based on the
rated power and LV side voltage of the power transformer. The formulas used in these
calculations and the resulting impedances are given in Appendix B.3.
7.2.2 Simulation Cases
In simulation cases 2, 3, 4 and 5, a three-phase-to-ground fault is introduced at differ-

ent locations in the basic simulation model. The fault resistance is set to zero and the
resistance to ground to 1 Ω.
Simulation Case 1: Normal Load Conditions
The model for the first simulation case is the basic simulation model displayed in Figure
7.5. This simulation case represents normal load conditions. The short-circuit capacity
of the grid is set to 10 000 MVA.
Simulation Case 2: Internal High Voltage Side Fault
The model for the second simulation case is displayed in Figure 7.6. This simulation case
represents a three-phase-to ground fault occurring after 380.7 ms on the HV side of the
power transformer within the protected zone. The short-circuit capacity of the grid is set
to 6 800 MVA.
68
Figure 7.6: ATPDraw simulation model of simulation case 2: internal high voltage side
fault.
Simulation Case 3: Internal Low Voltage Side Fault
The model for the third simulation case is displayed in Figure 7.7. This simulation case
represents a three-phase-to ground fault occurring after 380.6 ms on the LV side of the
power transformer within the protected zone. The short-circuit capacity of the grid is set
to 10 000 MVA.
Figure 7.7: ATPDraw simulation model of simulation case 3: internal low voltage side
fault.
69
Simulation Case 4: External Fault
The model for the forth simulation case is displayed in Figure 7.8. This simulation case
represents a three-phase-to ground fault occurring after 380.9 ms on the low voltage side
of the power transformer outside of the protected zone. The short-circuit capacity of the
grid is set to 3 300 MVA.
Figure 7.8: ATPDraw simulation model of simulation case 4: external fault.
Simulation Case 5: Current Transformer Saturation
In the fifth simulation case, the ideal CTs in the previous fault simulation cases are
replaced by saturable CTs. The CTs are modelled as shown in Figure 7.9.
Figure 7.9: ATPDraw model of three saturable CTs.
70
The magnetization curves of the two CTs are shown in Figure 7.10.
(a) HV side current transformer. (b) LV side current transformer.
Figure 7.10: Magnetization curves of the modelled CTs.
Simulation Case 6: Transformer Inrush
The model for the sixth simulation case is shown in Figure 7.11. This simulation case
represents a transformer inrush. The load is replaced by a resistive three-phase branch
with a per-phase resistance of 10 000 000 Ω in order to simulate no-load conditions. The
power transformer is energized after 5 ms. The simulated current is multiplied by a factor
of 10 in order to generate high enough currents for the SAMU and the protection IED to
detect in the laboratory tests.
Figure 7.11: ATPDraw simulation model of simulation case 6: transformer inrush.
71
7.3 Method
The laboratory test setup presented in Section 7.1 and shown Figure 7.2 was used through-
out all of the tests. Two sets of three-phase currents were generated by the OMICRON
CMC 356 test set and applied to the current inputs of the RET670 IED and the SAM600-
CT module, respectively. The applied currents will be referred to as Iapp , as indicated in
Figure 7.2. The disturbance recorder of the RET670 protection IED was used to record
the currents from the SAMU and the analog acquisition system. The recorded currents
will be referred to as Irec_samu and Irec_analog , respectively. The recordings were exported
from the RET670 protection IED, loaded into Wavewin for analysis and exported as
CSV-files in order to further processes the results.
7.3.1 Steady State Performance
For the steady state tests, the current transformer ratio settings for the SAMU and the
protection IED were set to 2500 A/5 A and 500 A/1 A, respectively, in order to simplify
the comparison of Irec_samu and Irec_analog . The power transformer ratio was set to 1:1. The
steady state tests were performed using the QuickCMC test module of the OMICRON
Test Universe software. Two equal sets of 50 Hz three-phase currents were generated
by the test set and applied to the current inputs of the SAMU and the protection IED,
respectively. Eight tests were performed in total at increasing current levels. The RMS
values of Iapp were: 1 ARMS , 5 ARMS , 10 ARMS , 15 ARMS , 20 ARMS , 25 ARMS , 30 ARMS , 32
ARMS .
The amplitudes of Irec_samu and Irec_analog were compared with the amplitude of Iapp . The
amplitude errors were calculated separately for the negative peaks and the positive peaks
according to Equation (7.2) and Equation (7.3) and averaged over a time period of 100
cycles.

ˆ
Irec_samu − Iâpp

Amplitude Error SAM U [%] =
ˆ
× 100% (7.2)
Iapp

ˆ
Irec_analog − Iâpp

Amplitude Error Analog [%] =
ˆ
× 100% (7.3)
Iapp
where Iˆrec_samu and Iˆrec_analog are the peak values of the recorded currents from the SAMU
and the analog acquisition system, respectively, and Iâpp is the peak value of the applied
72
currents from the test set. The Root-Mean-Square (RMS) values of Irec_samu and Irec_analog
were also calculated and compared with the RMS value of Iapp . The RMS values were
calculated over a time period of one cycle (80 samples) according to Equation (7.4).
s
1 2
iRM S = (i + i22 + ... + i2n ) and n = 80 (7.4)
n 1
where n is the number of samples and in is the instantaneous value of current sample n.
The phase difference between the phase angles of Irec_samu and Irec_analog was calculated
according to Equation (7.5) for each current sample and averaged over a time period of
100 cycles.
P hase Dif f erence [°] = φsamu − φanalog (7.5)
where φsamu and φanalog is the phase angles of Irec_samu and Irec_analog , respectively. Due
to lack of time synchronization between Iapp and Irec_samu /Irec_analog , the phase angles of
Irec_samu and Irec_analog could only be compared to each other, not to the phase angle of
Iapp .
7.3.2 Transient Performance
The transient performance tests were performed using the Advanced TransPlay test mod-
ule of the OMICRON Test Universe software. These tests were based on the ATPDraw
simulations which were described in Section 7.2. The currents from the two CTs in the dif-
ferent simulation case models were exported in PL4-files, imported to Advanced TransPlay
and applied from the test set to the current inputs of the SAMU and the protection IED,
respectively. The transient performance tests were performed in two parts.
First part: Performance of the SAMU
In order to assess the transient performance of the SAMU and to compare it to the
transient performance of the analog acquisition system, the same three-phase currents
were applied to the current inputs of the SAMU and the protection IED. That is to
say, the currents from either the LV side or the HV side of the power transformer in
the simulations were applied to both current inputs. Similarly to the steady state tests,
the current transformer ratio settings for the SAMU and the protection IED were set to
2500 A/5 A and 500 A/1 A, respectively. The power transformer ratio was set to 1:1.
73
A Discrete Fourier Transform (DFT) was performed in MATLAB in order to filter out
the fundamental frequency and second harmonic components from Irec_samu and Irec_analog
when fault currents and transformer inrush currents were applied, respectively.
Second part: Performance of the Transformer Differential Protection
In order to investigate the response of the transformer differential protection, the actual
three-phase currents from the simulations cases were applied to the inputs of the SAMU
and the protection IED. That is to say, the currents from the HV side of the power trans-
former were applied to the protection IED and the currents from the LV side of the power
transformer were applied to the SAMU. The current transformer ratio settings for the
SAMU and the protection IED were set to 1500 A/5 A and 500 A/1 A, respectively, in
accordance with the CT ratios in the simulation model. The power transformer ratio
was also set in accordance with the power transformer ratio in the simulation models. A
DFT was performed in MATLAB in order to filter out the fundamental frequency com-
ponents from Irec_samu and Irec_analog when fault currents were applied. The fundamental
frequency differential currents were subsequently calculated. The response times of the
transformer differential protection to the fault currents from the different simulation cases
were recorded by the disturbance recorder in the protection IED.
74
Chapter 8
Laboratory Results
In this chapter the most important results from the laboratory tests which were described
in Chapter 7 are presented. The findings will be discussed in Chapter 9. The results of
the steady state and transient performance tests are given in Section 8.1 and Section 8.2,
respectively. A summary of the response of the transformer differential protection to the
currents from the different simulation cases is presented in Section 8.3.
8.1 Steady State Performance

Figure 8.1 shows the steady state response of the SAMU and the analog acquisition system
at applied currents (Iapp ) of 1 ARMS and 32 ARMS , respectively. These two current levels
represent the lowest and the highest currents which were applied to the current inputs of
the SAMU and the analog acquisition system during the steady state tests. Irec_samu_x and
Irec_analog_x are the recorded currents from the SAMU and protection IED, respectively,
where x = A, B, C indicates the three phases A, B and C.
75
CHAPTER 8. LABORATORY RESULTS
(a) Steady state response at Iapp = 1 ARMS .
(b) Steady state response at Iapp = 32 ARMS .
Figure 8.1: Steady state response of the SAMU and the analog acquisition system.
76
8.1.1 Amplitude
Figure 8.2, Figure 8.3 and Figure 8.4 display the average amplitude errors of phase A, B
and C, respectively, of Irec_samu and Irec_analog at increasing levels of Iapp . The amplitude
errors were calculated by the use of Equation (7.2) and Equation (7.3) given in Section
7.3.1.
(a) Average amplitude error of Irec_samu .
(b) Average amplitude error of Irec_analog .
Figure 8.2: Average amplitude error of phase A. The dots represent the average ampli-
tude errors and the error bars the errors in the current output from the test set.
77
Figure 8.3: Average amplitude error of phase B. The dots represent the average ampli-
78
Figure 8.4: Average amplitude error of phase C. The dots represent the average ampli-
79
In Figure 8.5 the average peak values of Irec_samu at increasing levels of Iapp are plotted.
The plot of the peak values of Irec_analog was practically identical and is therefore not
shown.
Figure 8.5: Average peak values of Irec_samu .
The RMS values of phase A of Irec_samu and Irec_analog are given in Table 8.1. The RMS
values of the other two phases were almost identical and are therefore not given. The
RMS values were calculated by the use of Equation (7.4) given in Section 7.3.1.
Table 8.1: RMS values of phase A of Irec_samu and Irec_analog .
Iapp [ARMS ] Irec_samu [ARMS ] Irec_analog [ARMS ]

1A 1.00 1.00
5A 5.00 5.00
10 A 10.0 9.94
15 A 15.00 14.88
20 A 20.00 19.84
25 A 25.00 24.85
30 A 30.01 29.89
32 A 32.01 31.91
80
8.1.2 Phase Angle
Figure 8.6 displays the average phase difference between the phase angles of Irec_samu and
Irec_analog at increasing levels of Iapp . The phase differences were calculated by the use of
Equation (7.5) given in Section 7.3.1.
Figure 8.6: Average phase difference between the phase angles of Irec_samu and Irec_analog .
The dots represent the average phase difference and the error bars the minimum and
maximum phase differences.
8.2 Transient Performance
8.2.1 Faults
The transient response of the SAMU and the analog acquisition system to fault currents
of different magnitudes are shown in Figure 8.7 and Figure 8.8. Iapp represents the cur-
rents on the HV side and the LV side of the power transformer during an external fault,
respectively. The transient responses to the internal faults on the HV side and the LV of
the power transformer were very similar to the ones shown in Figure 8.7 and Figure 8.8
and are therefore not given.
81
(a) Transient response of phase A.
(b) Transient response of phase B.
(c) Transient response of phase C.
Figure 8.7: Transient response of the SAMU and the analog acquisition system to fault
currents. Iapp represents the fault currents on the HV side of the power transformer during
an external fault. Iapp cannot be seen in the plots since it is hidden by Irec_samu . Ierror is
the difference between Irec_samu and Irec_analog .
82
Figure 8.8: Transient response of the SAMU and the analog acquisition system to fault
currents. Iapp represents the fault currents on the LV voltage side of the power transformer
during an external fault. Iapp cannot be seen in the plots since it is hidden by Irec_samu .
Ierror is the difference between Irec_samu and Irec_analog .
83
8.2.2 Current Transformer Saturation
Figure 8.9 shows the transient response of the SAMU and the analog acquisition system
to saturated fault currents. No saturation occurred in phase A, and the transient response
of this phase is therefore not given. Iapp represents the saturated fault currents on the
LV side of the power transformer during an external fault. The transient responses to
the saturated currents from the internal faults on the HV side and the LV of the power
transformer were very similar to the ones shown in Figure 8.9 and are therefore not given.
(a) Transient response of phase B.
(b) Transient response of phase C.
Figure 8.9: Transient response of the SAMU and the analog acquisition system to
saturated fault currents. Iapp represents the saturated fault currents on the LV voltage
side of the power transformer during an external fault.
84
8.2.3 Transformer Inrush
The transient response of the SAMU and the analog acquisition system to transformer
inrush currents is shown in Figure 8.10 and zoomed in in Figure 8.11.
transformer inrush currents. Iapp cannot be seen in the plots since it is hidden by Irec_samu .
85
(c) Transient response of phase C

.
transformer inrush currents. Iapp cannot be seen in the plots since it is hidden by Irec_samu .
86
8.3 Response of Transformer Differential Protection

Table 8.2 summarizes the response of the transformer differential protection to the cur-
rents from the simulation cases presented in Subsection 7.2.2. The response time of the
differential protection was recorded by the disturbance recorder.
Table 8.2: Response of the transformer differential protection.
Simulation Case Response Response Time

1) Normal load conditions No trip -
2) Internal HV side fault Trip 11 ms
3) Internal LV side fault Trip 11 ms
4) External fault No trip -
5a) CT saturation during an internal HV side fault Trip 11 ms
5b) CT saturation during an internal LV side fault Trip 10 ms
5c) CT saturation during an external fault No trip -
87
Chapter 9
Discussion of Laboratory Tests
In this chapter the laboratory results which were presented in Chapter 8 are discussed.
The objective of the laboratory tests was threefold: to investigate the steady state and
transient performance of a SAMU, to compare the steady state and transient performance
of a SAMU to those of a conventional analog acquisition system and to investigate whether
a hybrid mode configuration could have an impact on the performance of transformer
differential protection. The results of the laboratory tests are discussed in light of these
objectives.
9.1 Steady State Performance

In Figure 8.2a, Figure 8.3a and Figure 8.4a, it can be observed that there are some small
differences between the amplitude accuracies of the recorded currents of phase A, B and
C from the SAMU, particularly at the lowest levels of applied current. In all three phases
the average amplitude error in percent of the applied current is highest at an applied
current of 1 ARMS , which is 20 % of the nominal input current of the SAMU. The highest
average amplitude error occurs in phase A and is approximately 3 %, which is equivalent
to 0.04 A. At the lowest levels of applied current, there seems to be a small DC offset in
the recorded currents from the SAMU since the average amplitude errors of the negative
peaks are approximately equal to the average amplitude errors of the positive peaks and
of opposite sign. This DC offset is also observable in the steady state response of the
SAMU which is shown in Figure 8.1a. However, the DC offset is not consistent and does
not occur in the recorded currents from the SAMU at all levels of applied current, as
shown in Figure 8.1b. Either the DC offset only occurs at low levels of applied current or
there are other sources of error. Anyhow, the average amplitude errors are so small that
they do not affect the linearity of the SAMU, as can be seen in Figure 8.5. Dutra et al.
[100], whose work was mentioned in Section 5.3, also found that the SAMU which they
tested had a linear output.
89
CHAPTER 9. DISCUSSION OF LABORATORY TESTS
There are some small differences between the amplitude accuracy of the recorded currents
from the SAMU and the analog acquisition system, as shown in Figure 8.2, Figure 8.3
and Figure 8.4. Judging from the results, the analog acquisition system seems to have a
slightly better amplitude accuracy at 1 ARMS and 5 ARMS and the SAMU in the range
from 10 ARMS to 25 ARMS . The results in Table 8.1 show that the RMS values of the
recorded currents from the SAMU and the analog acquisition system are almost equal and
coincide well with the RMS values of the applied current. The observed amplitude errors
are so small that they should not affect the performance of the transformer differential
protection in hybrid mode configuration.
There is a small phase difference between the phase angles of the recorded currents from
the SAMU and the analog acquisition system, as can be seen in Figure 8.6. At all levels of
applied current above 1 ARMS , the recorded currents from the SAMU leads the recorded
currents from the analog acquisition system. In all three phases the phase difference
increases with the applied current. The phase difference could partially be caused by
inaccuracies in the time synchronization of the SAMU and the protection IED. However,
since the phase difference increases with the applied current, it cannot be caused solely
by inaccuracies in the time synchronization. A possible explanation could be that the
phase difference is caused by the different types of input elements which are used in the
SAMU and in the protection IED, that is shunt elements in the SAMU and input current
transformers in the protection IED. The phase difference could possibly be a result of the
inductance of the input current transformers in the protection IED. The highest phase
difference observed was 0.31°, which corresponds to a magnitude error of only 0.54 %.
The insignificance of the observed phase difference can also be proved by considering the
operate-restrain characteristic of the transformer differential protection which is shown
in Figure 7.3 in Subsection 7.1.4. The most critical point on the characteristic is at the
end of Section 1. The applied current level closest to this point is 1 ARMS , corresponding
to a primary current of 500 ARMS or 1.21 times the base current of 412 ARMS . Under
these conditions, the phase difference between the phase angles of the recorded currents
from the SAMU and the analog acquisition system would have had to be 14.3° in order to
cause an incorrect operation of the differential protection. By comparing the value of the
highest phase difference which was observed during the laboratory tests to this number,
it can be seen that the observed phase difference is insignificant.
Figure 8.6 shows that the phase difference is not equal in all three phases. The phase
differences of phase A and C are almost identical at all levels of applied current. The phase
difference of phase C, on the other hand, differs from the others. However, the change
of the phase differences with increasing levels of applied current is similar in all three
phases. Dutra et al. [100] also investigated the phase difference between the recorded
90
currents from a SAMU and an analog acquisition system, but only at an applied current
of 5A. Similar results were obtained at this current level, though with smaller differences
between two of the phases and the third than what was observed during these laboratory
tests.
9.2 Transient Performance
9.2.1 Faults
The transient response of the SAMU to various faults is almost identical to the applied
fault currents. Two examples are shown in Figure 8.7 and Figure 8.8. The recorded
currents from the SAMU are not distorted, and the DC component in the fault currents
is reproduced well. When comparing the transient response of the SAMU to that of the
analog acquisition system, on the other hand, differences can be observed. As mentioned in
Section 5.3, some types of input current transformers cannot reproduce the DC component
in transient primary currents properly. Judging from the transient response shown in
Figure 8.7, this seems to be the case for the input current transformers of the protection
IED.
It was also mentioned in Section 5.3 that the waveform of saturated currents from input
current transformers may not have the typical appearance of saturated CT currents due to
the very low impedance of the secondary circuit. The transient response which is shown in
Figure 8.8 is the result of applying fault currents with a maximum value of approximately
32 times the rated input current the protection IED. Comparing with previous work of
Holst and Zakonjšek [27], which was mentioned in Section 5.3, it may look like the input
current transformers of the protection IED are saturating after approximately 410 ms. The
error between the recorded currents from the SAMU and the analog acquisition system is
much larger in Figure 8.8 than in Figure 8.7, where fault currents with a maximum value
of only 6 times the rated input current of the protection IED were applied. The error
currents in Figure 8.8 also vary much more over time than the error currents in Figure
8.7. Based on a comparison with the waveforms of currents from different types of CTs
which was presented in the work of Holst and Zakonjšek [27], it may look like the input
current transformers of the protection IED are of the LR CT type, that is the type with
small air gaps in the CT core.
The maximum value of the fault currents in Figure 8.8 is within the measurement range
of both the protection IED and the SAMU. However, this current is only about 6 times
higher than the rated input current of the SAMU, compared to 32 times the rated input
current of the protection IED. In order to properly compare the transient response of the
91
SAMU and the analog acquisition system, the SAMU and the protection IED should have
had the same rated input current. On the other hand, it is not expected that the transient
response of the SAMU will show any signs of saturation even at an applied current of
32 times its rated input current since the SAMU has shunts as input elements instead of
input current transformers.
In order to evaluate whether the observed differences in the transient response of the
SAMU and the analog acquisition systems may affect the performance of the differential
protection in a hybrid mode configuration, a digital filter was implemented in MATLAB
by the use of DFT. The MATLAB script is given in Appendix C.1. The filter was used
to extract the fundamental frequency phase currents from the recorded phase currents of
the SAMU and the analog acquisition system. The fundamental frequency phase currents
of the recorded currents in Figure 8.7 and Figure 8.8 are shown in Figure 9.1 and Figure
9.2, respectively.
In Figure 9.1 it can be observed that there is a small difference between the amplitudes
of the fundamental frequency phase currents from the SAMU and the analog acquisition
system in phases B and C. The difference is likely to be caused by the differing abilities
of the SAMU and the analog acquisition system to reproduce the DC component in the
fault current. A larger difference between the fundamental frequency phase currents from
the SAMU and the analog acquisition system can be observed in Figure 9.2. In phases
B and C, the fundamental frequency phase currents from the analog acquisition system
have a higher amplitude than the fundamental frequency phase currents from the SAMU
and oscillates around a lower value after fault occurrence.
92
(a) Fundamental frequency currents of phase A.
(b) Fundamental frequency currents of phase B.
(c) Fundamental frequency currents of phase C.
Figure 9.1: Fundamental frequency phase currents of the recorded currents from the
SAMU and the analog acquisition system in Figure 8.7. Figure 8.7 shows the transient
response of the SAMU and the analog acquisition system to fault currents on the HV side
of the power transformer during an external fault.
93
(a) Fundamental frequency currents of phase A.
(b) Fundamental frequency currents of phase B.
(c) Fundamental frequency currents of phase C.
Figure 9.2: Fundamental frequency phase currents of the recorded currents from the
SAMU and the analog acquisition system in Figure 8.8. Figure 8.8 shows the transient
response of the SAMU and the analog acquisition system to fault currents on the LV side
of the power transformer during an external fault.
94
It should be noted that these two cases do not represent true fault scenarios since the same
phase currents were applied to both the SAMU and the analog acquisition system. The
observed differences between the fundamental frequency phase currents from the SAMU
and the analog acquisition system were further analyzed by considering the external fault
simulation case. In this test currents from the HV side of the power transformer were fed
to the analog acquisition system and currents from the LV side of the power transformer
to the SAMU. A DFT was performed in MATLAB and the fundamental frequency dif-
ferential current of each phase calculated. The MATLAB script is given in Appendix C.2
and the results in Figure 9.3.
Fundamental frequency differential currents can be observed around t = 0 ms in Figure

9.3. These are most likely caused by some irregularities in the recorded currents from the
SAMU and the analog acquisition system at the instant when the currents from the test
set were applied. These may be ignored. In phases B and C, where the recorded currents
from the analog acquisition system were affected by the limited ability of the input current
transformers to reproduce the DC component in the fault currents, increased fundamental
frequency differential currents can be observed around t = 400 ms. The maximum values
are 20.0 ARMS and 23.5 ARMS in phase B and C, respectively, compared to 8.8 ARMS
in phase A where the recorded currents from the analog acquisition system were almost
equal to the recorded currents from the SAMU.
As described in Subsection 7.1.4, the base current in the transformer protection IED
was set to 412 ARMS and the minimum differential current (IdMin ) to 30 % of the base
current. The differential current would thus have to be at least 123.6 A in order for the
transformer differential protection to trip. In addition, this is the limit of Section 1 in
the operate-restrain characteristic of the transformer differential protection. Based on the
fundamental frequency phase currents shown in Figure 8.7, it can be estimated that the
highest fundamental frequency current and thus the restrain current was at least three
times the base current. This number marks the end of Section 2 in the operate-restrain
characteristic. Thus, the fundamental frequency differential current would have to be
even greater than 123.6 A in order for the transformer differential protection to trip.
Exactly how this particular transformer differential protection filters out the fundamental
frequency phase currents and calculates the fundamental frequency differential currents
is not known. The fundamental frequency phase currents shown in Figure 9.1 and Figure
9.2 and the fundamental frequency differential currents displayed in Figure 9.3 may not
exactly represent the currents calculated by the protection IED. However, the approach
applied here should represent a fairly good approximation.
95
(a) Fundamental frequency differential current of phase A.
(b) Fundamental frequency differential current of phase B.
(c) Fundamental frequency differential current of phase C.
Figure 9.3: Fundamental frequency differential phase currents from the external fault
simulation case.
96
9.2.2 Current Transformer Saturation
The transient response of the SAMU to saturated fault currents is almost identical with
that of the analog acquisition system, as shown in Figure 8.9. Neither the SAMU nor
the analog acquisition system manage to reproduce the steepest peaks of the simula-
tion. Apart from that, both the SAMU and the analog acquisition system reproduce the
waveforms of the saturated fault currents well.
9.2.3 Transformer Inrush
There are distinct differences between the transient response of the SAMU and the analog
acquisition system to transformer inrush currents, as shown in Figure 8.10 and Figure 8.11.
In all three phase currents, there is a DC component. As with the fault cases, the analog
acquisition system cannot reproduce the DC component in the primary currents properly.
The SAMU on the other hand, reproduces the waveforms of the applied inrush currents
well.
Since transformer inrush currents are characterized by the presence of second harmonics, a
DFT was performed in MATLAB in order to compare the content of second harmonics in
the recorded currents from the SAMU and the analog acquisition system. The MATLAB
script is given in Appendix C.3. The resulting second harmonic phase currents are given
in Figure 9.4. In the figure it can be observed that the content of second harmonics in
the recorded currents from the SAMU and the analog acquisition system is similar. The
biggest difference occurs around 40 ms in all phases. Judging from the results, it does
not seem like the differing transient responses of the SAMU and the analog acquisition
system have a significant impact of the content of second harmonics.
97
(a) Second harmonic currents of phase A.
(b) Second harmonic currents of phase B.
(c) Second harmonic currents of phase C.
Figure 9.4: Second harmonic phase currents of the recorded currents from the SAMU
and the analog acquisition system in Figure 8.10. Figure 8.10 shows the transient response
of the SAMU and the analog acquisition system to transformer inrush currents.
98
9.3 Performance of Transformer Differential

Protection
The transformer differential protection operated correctly for all simulation cases, as
shown in Table 8.2. The response time of the differential protection was approximately
the same in all cases. Judging from these results, the hybrid mode configuration does
not seem to affect the performance of the transformer differential protection. However,
these simulation cases only represent a sparse selection of possible operating scenarios.
In addition, the response time of the transformer differential protection was only roughly
estimated. In order to properly assess the influence of a hybrid mode configuration on the
transformer differential protection, the response time should have been accurately mea-
sured. Another aspect to consider is that the process bus in an actual digital substation
will not be ideal as it was in the tests performed here. Among other things, possible loss
of SV and network delays may affect the stream of SV from the SAMU to the subscribing
protection IED.
99
Chapter 10
Conclusion of Laboratory Tests and

Future Work
In this chapter some preliminary conclusions are drawn from the discussion of the labo-
ratory tests in Chapter 9. Suggestions for future laboratory work are also described. The
final conclusion to this thesis will be given in Chapter 12.
10.1 Conclusion
The laboratory tests showed that the SAMU under test has a steady state and transient
performance which makes it adequate for use in digital substations. The SAMU provided
SV with a small, but acceptable amplitude error. The amplitude error was highest at
applied currents below the rated current of the SAMU. In order to achieve the best
performance possible, the rated secondary current of the associated CIT and thus the
rated current of the SAMU should be chosen in such a way that the secondary current is
higher than the rated current under most operating conditions.
The tested SAMU has a steady state performance which is comparable to the steady
state performance of a conventional analog acquisition system. There is a small phase
difference, which increases with the applied current, between the phase angles of the
recorded currents from the SAMU and the analog acquisition system. However, the phase
difference is too small to affect the performance of the transformer differential protection
in hybrid mode configuration.
The SAMU under test has a better transient performance than the conventional analog
acquisition system. The laboratory tests showed that the SAMU can close to perfectly
reproduce the DC component in transient currents without any distortion. The ability
of the analog acquisition system to reproduce DC components, on the other hand, was a
bit more limited, and the recorded currents were distorted at current levels significantly
above the rated current of the input current transformers of the protection IED.
101
CHAPTER 10. CONCLUSION OF LABORATORY TESTS AND FUTURE WORK
The differing transient responses of the SAMU and the analog acquisition system re-
sulted in fundamental frequency differential currents when employed in a hybrid mode
transformer differential protection configuration. In the tested case the resulting funda-
mental frequency differential currents were not big enough to affect the performance of the
transformer differential protection. However, one should be aware that different SAMUs,
NCITs and conventional protection IEDs can have different transient performances. In a
digital substation it should be made sure that differing transient responses will not affect
the performance of the protection functions.
10.2 Future Work

Due to limitations in time and the laboratory test setup, some tests were left for future
work. The SAMU could only be tested at a limited current level of 6.4 times its rated
input current due to limitations in the maximum current output from the test set. In
future tests a SAMU with a lower rated input current of 1 A should thus be tested. In
addition, the phase accuracy of the SAMU could not be investigated due to lack of time
synchronization between the output from the test set and the output from the SAMU. A
solution to this issue should be found and further tests conducted. The performance of
the voltage module of the SAMU should also be investigated. Only a rough estimate of
the response time of the protection IED was found in the performed laboratory tests. In
order to properly assess the performance of the transformer differential protection under
different operating conditions, the trip signal should be properly measured in future tests.
As mentioned in Section 5.3, SAMUs from different suppliers may employ different input
elements and thus have different transient responses. Interoperability between devices
from different vendors may also be an issue in digital substations, as mentioned in Section
6.2. Future tests should thus include SAMUs from different vendors and combinations of
SAMUs and protection IEDs from different vendors.
In the performed laboratory tests, the SAMU was connected with a direct fiber optic
link to the protection IED, representing an ideal process bus network. In an actual
digital substation, the process bus network is likely to include several switches and fiber
optic cables. There will also be SV streams from other SAMUs and the associated MUs
of NCITs in addition to other background traffic. The effects of a non-ideal process
bus network on the performance of the SAMU and the protection IED should thus be
investigated in future tests. A possible test is to introduce time delays of different length
or possibly background traffic in the process bus network and investigate the response of
the transformer differential protection. Another possible test is to investigate the response
of the transformer differential protection when SV messages are discarded.
102
Chapter 11
Discussion
This chapter provides a discussion of the benefits and drawbacks of applying OCTs in
digital substations compared to the use of CCTs in conventional substations. It is also
discussed whether OCTs are suitable for application in digital substations in the trans-
mission network.
11.1 Health, Safety and Environment

OCTs offer significant improvements in terms of health, safety and environment compared
to CCTs. In conventional substations the secondary circuits of CCTs are connected to the
back of the protection and control panels in the control room, as mentioned in Section 2.2.
Personnel may therefore get electrocuted if the contacts are touched. There is also a risk
that a secondary circuit of a loaded CCT can be open-circuited. As explained in Section
2.2, this will lead to a very high voltage across the secondary winding, which in turn can
result in arcing and potentially the CCT exploding. This is an unwanted situation which
presents a danger to both personnel and equipment in the substation.
The safety of personnel and equipment can be significantly improved by replacing the
CCTs with OCTs and substituting the analog measuring circuits with a process bus
realized by the use of fiber optic cables. As mentioned in Section 6.6, fiber optic cables
are non-conductive and provides galvanic isolation. This means that personnel cannot get
electrocuted from touching the contacts at the back of the protection and control panels.
In addition, OCTs do not have any secondary circuits which can be open-circuited and
do not contain any oil. There is thus no risk of an OCT exploding.
By replacing CCTs with OCTs, the environmental impact can also be reduced. Since
OCTs do not contain any oil or SF6 gas, as apposed to CCTs, the risk of oil or SF6 gas
leaking into the ambient environment is removed. OCTs also offer a significant weight
reduction compared to CCTs. Based on information from the product survey which
was conducted in this thesis, a weight reduction of approximately 80 % or more can be
103
CHAPTER 11. DISCUSSION
expected when replacing CCTs with OCTs. As a consequence of the reduced weight, it can
be expected that the CO2 emissions associated with the transport of CTs to the substation
will decrease. The 80 % weight reduction does not take into account the reduced weight
of cables due to the replacement of a large number of copper cables with a smaller number
of fiber optic cables. An even higher weight reduction can therefore be expected.
11.2 Performance
As mentioned in Section 2.2, the accuracy of CCTs is limited by the use of an iron
core. The literature review shows that OCTs can offer significant improvements in terms
of performance compared to CCTs. Thomas et al. [93] and Rahmamatian and Blake
[8] demonstrated that the accuracy of OCTs can satisfy and may even outperform the
measuring accuracy requirements of IEC instrument transformer standards, and that
OCTs can maintain this accuracy over a wide dynamic range. However, one should bear
in mind that these practical demonstrations are experimental tests of specific OCTs. As
described in Section 3.3, OCTs can be designed in many ways, and it is therefore likely
that the performance of OCTs from different vendors will differ to some degree.
The literature review also shows that OCTs have the advantage of an excellent transient
performance without saturation effects. As mentioned in Section 2.2, saturation of CCTs
may for instance result in incorrect operation of differential protection functions or differ-
ential protection functions not operating when required to. In Section 6.3 it was explained
how the sensitivity of differential protection has to be reduced in order to prevent incor-
rect operation due to saturation of CCTs during external faults. Since OCTs are linear
over a wide dynamic range and do not saturate, the sensitivity of differential protection
can be increased. However, it is still important to design the OCTs with the proper num-
ber of sensing fiber coil turns since distortion of the measured current can occur if the
short-circuit current exceeds the maximum detectable current of the OCTs.
11.3 Substation Layout and Footprint

The high weight of CCTs puts requirements on the supporting structures which are needed
in the substation yard. As mentioned in Section 3.7, the low weight and compact size
of OCTs open up the possibility for integration of the sensor heads into other primary
equipment. If such installation solutions are chosen, separate supporting structures for the
OCTs will not be needed. In turn, this reduces the required space in the substation yard,
which allows the footprint of future digital substations to be reduced. As mentioned
in Section 4.2, there are also products on the market which combine several primary
104
functions into a single device. Examples are the disconnecting circuit breaker and the
combined optical current and voltage transformer. Adoption of such solutions will also
reduce the required space in the substation yard.
All of the OCT which were considered in the product survey can be installed as free-
standing devices with a configuration similar to that of CCTs, as described in Section 4.2.
This means that one does not have to change the arrangement of the primary equipment in
present day transmission substations in order to apply OCTs. However, such free-standing
solutions does not offer the benefit of saving space in the substation yard. Nevertheless,
it should be possible to make the supporting structures lighter than those required today
since the weight of OCTs is much lower than the weight of CCTs.
The cores of CCTs and the dedicated wiring from each core to specific secondary devices
make the design of conventional substations inflexible. The CT cores limit the number of
associated applications, which means that a CCT might have to be replaced by one having
more cores if new applications are required and there are no free CT cores. This situation
changes in a digital substation with OCTs. Due to the ability of OCTs to maintain
a high accuracy over a wide dynamic range, the same sensing element can be used for
both measuring and protective applications, possibly two for redundancy. In addition, the
process bus makes all current and voltage measurements available to all protection, control
and monitoring devices. This makes the substation design more flexible. In addition, it
facilitates the integration of several application functions into the same IED, resulting in
fewer IEDs in the control room. In turn, this will reduce the required space in the control
building and possibly the need for cooling and power supply.
The present day solution with CCTs and dedicated wiring from each CT core to specific
secondary devices also makes it difficult to satisfy the regulatory requirements regarding
duplicated and physically independent routing of cables, which was mentioned in Section
2.5. With the application of a redundant process bus, these requirements may become
easier to satisfy. As mentioned in Section 6.6, there will also be a reduced need for large
cable trenches in digital substations since fewer fiber optic cables are needed in order
to realize the process bus than the number of copper cables which are required for the
measurement and control circuits in conventional substations. Instead of using large cable
trenches, the fiber optic cables can be laid in pipes or smaller cable trenches.
11.4 Installation and Replacement

Due to the large weight of CCTs, heavy cranes are typically required for the installation
and replacement of such CTs. With a weight reduction of approximately 80 % when
105
replacing CCTs with OCTs, it can be expected that there will be a reduced need for heavy
cranes during installation and replacement of OCTs in digital transmission substations.
However, at least for the higher voltage levels, it is reasonable to assume that some type
of lifting equipment will still be required based on the weight information provided by the
vendors during the product survey.
The workers which install CCTs today are probably unfamiliar with OCT technology and
how to install OCTs. It can therefore be assumed that some training will be required. In
addition, as mentioned in Section 6.6, the connection of fiber optic cables requires special
competence. In order to expand the competence of the workforce, resources have to be
put down. On the other hand, there will be fewer cables to install, which should result in
a reduction in the time spent on cable installations.
11.5 Maintenance and Expected Lifetime

It can be expected that OCTs require less maintenance than CCTs since the insulators do
not contain any oil. This assumption was confirmed by the vendors which were contacted
during the product survey. According to them, the OCTs require little or no maintenance.
As mentioned in Section 2.2, it is estimated that the periodical maintenance of a single
CCT takes approximately 30 minutes per month. As a result, around one an a half hour
of maintenance could be saved per month for a set of three CCTs which is replaced by a
three-phase OCT in a substation.
CCTs have proven to be reliable and show a good stability over a lifetime of approximately
25 to 40 years, as described in Section 2.2. Based on information from the product survey,
it can be expected that the lifetime of the sensor heads of an OCT is comparable to the
expected lifetime of a CCT. The sensor electronics, on the other hand, have an expected
lifetime which is about half as long as that of the sensor heads. The sensor electronics
will thus have to be replaced once during the lifetime of the OCT. Developments in the
technology are probable to occur within that time span, and it should be considered
whether compatibility issues between old and new parts could be a potential problem.
11.6 Interoperability and Compatibility

As mentioned in Section 2.2, CCTs have a standardized analog interface. This facilitates
the use of CCTs and secondary devices from different vendors in conventional substations.
Since the interface is simply an analog 1 A or 5 A connection, it may be assumed that
there will not be significant developments in the interface over time and that is relatively
106
unproblematic to replace a CCT from one vendor with a CCT from another vendor, and
to connect CCTs and secondary IEDs of different age or from different vendors.
When it comes to digital substations, on the other hand, it may be expected that there
will be more challenges related to interoperability and compatibility between NCITs with
associated MUs, SAMUs and secondary IEDs. Even though standards such as IEC 61850
shall ensure that devices from different vendors are interoperable, these standards often
leave some room for interpretation. One example of standards being interpreted differently
is the handling of the SV quality attribute, which was mentioned in Subsection 6.2.1.
Such differences in the interpretation of standards can prevent interoperability between
devices from different vendors and result in systems which are not working optimally or
possibly not working at all. There is no guarantee that devices from different vendors
will actually be interoperable in a digital substation, even though the specifications state
that the devices conform to the standards. Real-life testing of interoperability between
devices from different vendors is thus of great importance. In addition, it can be discussed
whether there is a need to make the standards more specific. However, this would leave
less freedom to the vendors in the design of their products.
One can also imagine that there will be greater changes in the digital interface between
MUs or SAMUs and secondary IEDs over time than what has been the case for the analog
interface of CITs. As a result, one may face greater difficulties in terms of compatibility
when replacing devices in a digital substation. Experiences from the laboratory tests
confirmed this assumption. The protection IED used in the laboratory tests was too old
to support time synchronization via PTP. There were also issues with the compatibility
of the contacts of different devices. Since the network clock was only equipped with
electrical output contacts for distribution of 1PPS signals, while the protection IED had
optical input contacts, an electrical to optical converter would have been needed in order
to connect these devices. Such a solution would most likely not be desirable in a digital
substation since it adds another possible point of failure and could introduce errors. In
addition, the vendors might not be able to guarantee that their equipment will work as
promised if such a solution is employed.
11.7 Maturity and Availability of Products

Since CCTs have been in wide use for a long time, utilities and manufacturers are well
acquainted with the technology and have learned how to deal with its downsides, such as
adapting the operate-restrain curve of differential protection in order to avoid incorrect
operation due to CT saturation. In addition, there is a large selection of commercially
available CCTs and compatible secondary devices from multiple vendors.
107
As mentioned in Section 3.7, OCT technology has also been known for quite some time
now. However, the technology has not been widely adopted in substations yet. Naturally,
the operational experience with such sensors is much more limited than for CCTs. As
described in Subsection 6.2.1, there are still several unanswered questions regarding the
application of OCTs, other NCITs and SAMUs in digital substations. Several examples
were mentioned. One example is if and how manufacturers of protection IEDs should
specify requirements for NCITs and SAMUs in order to ensure satisfactory performance
of the protection functions. Other examples are how the protection IEDs should respond
to missing SV, SV whose validity is set to questionable or invalid and SV which are not
sufficiently time synchronized. In addition, some standards are still being developed and
have not been released yet, such as IEC 61869-13, which will apply to SAMUs. All of this
may be interpreted as signs of immaturity. However, there is clearly a lot of work going
on and interest in this area at the present time, such as the work of IEC TC 95 AHG 3.
It is therefore reasonable to assume that solutions to the above-mentioned questions can
be found not too far ahead in the future.
The product survey in Chapter 4 showed that there are several vendors offering OCTs
at the present time. However, it was discovered that only one of the considered OCTs
supports network-based PTP time synchronization, as recommended by the new IEC
61869-9 standard. The other OCTs, which use 1PPS for time synchronization, will require
dedicated wiring for the distribution of time synchronization signals. Another attribute
which was investigated was the support of network redundancy protocols. Since it could
only be confirmed that one of the OCTs supports PRP and HSR, further inquires should
be made in order to establish whether the other OCTs support these network redundancy
protocols. The present day regulatory requirement regarding duplicated and physically
independent routing of cables, which was given in Section 2.5, will most likely apply in
digital substations as well. The requirement could be interpreted in such a way that the
process bus shall be redundant. Hence, OCTs should support PRP and/or HSR. Lack
of such attributes may limit the selection of OCTs which could actually be applied in a
digital transmission substation.
11.8 Vulnerability
There are several factors which may make digital substations with OCTs more vulnerable
than the present day conventional substations. Some examples are the placement of MUs,
SAMUs and other electronics in the substation yard and the need for highly accurate time
synchronization [122]. In digital substations with OCTs, several electronic devices such
as MUs and SAMUs will be placed outdoor in the substation yard. As described in
108
Section 6.2.3, the electronics have to be installed in suitable cabinets which are secured
against EMP and EMI. In addition, as can be seen from Table 4.1 in Section 4.8, the
sensor electronics of OCTs typically have a more limited operating temperature range
than the sensor heads. Comparing with the large outdoor temperature variations in
Norway, heating would be needed in order to keep the temperature inside the cabinet
within the operating temperature range of the MUs. Cooling could also be required. The
dependency on proper heating and cooling may make OCTs more vulnerable than CCTs,
which do not have such requirements. In addition, the lifetime of the electronics may be
affected by the outdoor conditions.
Highly accurate time synchronization is an important requirement of several protection

applications in digital substations, as described in Subsection 6.2.1. Inaccuracies in the
sampling synchronization will result in phase errors in the measured currents, and these
errors may for instance lead to incorrect operation of differential protection functions.
Digital substations can thus be expected to be more vulnerable than present day substa-
tions if inaccuracies in the time synchronization or a possible loss of time synchronization
were to occur.
The literature review shows that OCTs are inherently sensitive to environmental distur-
bances such as temperature variations and vibrations. However, as described in Section
3.5, there are several methods which can be used to mitigate the effects of such distur-
bances. Based on information from one of the vendors stating that the accuracy of the
OCT is maintained over the whole operating range and is not sensitive to vibrations, it
can be assumed that the vendors have managed to adopt successful methods which mit-
igates the effects of environmental disturbances. The practical demonstrations described
in Subsection 3.5.2 and Subsection 3.5.3 corroborates this assumption. An example is
the use of temperate sensors on the sensing fiber in conjunction with temperature com-
pensation. However, this implies that the accuracy of the sensor will be dependent on
a well functioning temperature sensor. As it was mentioned in Subsection 3.5.2, IEEE
recommends periodical verification of the accuracy of the temperature sensor if such a
solution is employed.
11.9 Costs
Based on information from the product survey which was conducted in this thesis, it
can be estimated that the price of a three-phase OCT with an associated MU is about
twice as high as the price of a set of three CCTs rated for a maximum system voltage of
420 kV. However, the application of OCTs in digital transmission substations can lead to
several other cost savings, which should be taken into account when comparing the two
109
solutions. Some examples are as follows. Since lighter or fewer supporting structures will
be required for the OCTs, savings can be made in the amount of foundation work which
is required when building new substations. In addition, due to the reduced cabling and
the replacement of copper cables by fiber optic cables, savings are expected to be made
on cabling, cable trenches and the installation of cables, as mentioned in Section 6.6.
If decided to make use of solutions where the OCTs are integrated with other primary
equipment, less space will be required in the substation yard, which may reduce the capital
expenses of new substations.
11.10 Hybrid Mode Differential Protection

From the specifications of the commercially available OCTs given in Table 4.1 in Section
4.8, it can be seen that the maximum system voltages which the OCTs are specified for,
do not cover the lowest voltage levels which are included in the transmission network,
according to the definition given in Section 2.1. Hence, it may for instance be desirable
to continue using CCTs at the low voltage side of some power transformers in transmis-
sion substations. In order to integrate these CCTs into the digital substation, SAMUs
can be applied. This is one of the reasons why differential protection is likely to be ap-
plied in configurations which involve combinations of NCITs and SAMUs in future digital
substations.
As mentioned in Section 5.3, different manufacturers may employ different types of current
input elements for SAMUs, which can have different transient performances. Some types
of input current transformers cannot reproduce the DC component in transient currents
properly and may go into saturation at current levels significantly above the nominal
value. This was demonstrated for the input current transformers of the protection IED
in the laboratory tests and reported by Holst and Zakonjšek [27]. Differences in the
transient response of different input elements could potentially cause unwanted operation
of differential protection when employed a hybrid mode configuration. The laboratory
tests showed that the combination of shunt input elements and input current transformers,
most likely of the LR type, did not affect the performance of the differential protection
in this case. However, the differing transient performances did result in a fundamental
frequency differential current, which could potentially be higher in other cases and with
other combinations of input elements.
110
11.11 Fulfillment of Application Requirements

CCTs cover the needs of the applications in present day transmission substations. This
includes energy metering, protection, busbar protection, disturbance recording, monitor-
ing, control, quality metering etc. However, as mentioned in Section 2.2, in order to
satisfy the requirements of these applications in terms of accuracy, dynamic range and
bandwidth, different types of CT cores are required. In addition, several cores are needed
in order to segregate different protection applications.
When it comes to the application of OCTs in digital substations, on the other hand, the
requirements of both protective and measuring applications can be covered with a single
sensing element, possibly two for redundancy. However, as mentioned in Section 6.5, it has
not yet been sorted how to perform accredited meter inspections in digital substations.
Since this is a regulatory requirement, CITs may have to be used for electrical energy
metering applications until the issue is worked out. In other words, OCTs are not able to
fulfill the requirements of electrical energy metering applications at the present time.
11.12 Supported Functionality in Digital Substations

In order to apply OCTs in transmission substations while fully realizing the potential
of digital substations, protection IEDs of all types needed in a transmission substation,
which support IEC 61850-9-2 process bus, have to be commercially available. Preferably,
there should be multiple vendors offering such protection IEDs. Since Statnett requires
two separate protection IEDs supplied by different vendors for distance protection as a
redundancy measure [116], there has to be at least two vendors offering distance protection
for digital substations.
As mentioned in Section 6.4, there are presently several vendors offering protection IEDs
with IEC 61850-9-2 LE process bus support, including the two suppliers Statnett has frame
agreements with. However, it is not clear yet whether all types of protection which will be
needed in digital substations are presently available. In particular, it is uncertain whether
good busbar protection solutions for bigger substations exists. Hence, investigations and
possibly testing will be required in order to find new solutions. Statnett will also have to
approve new protection IEDs in order to realize a digital substation since the protection
IEDs which support IEC 61850-9-2 LE process bus are newer versions of protection IEDs
which have been approved by Statnett.
111
11.13 Fulfillment of the Digital Substation Concept

Only one out of the four vendors which were contacted during the product survey offers
a solution where the MU is placed in an outdoor cabinet, as mentioned in Section 4.2.
Placing MUs in the protection and control panel does reduce the number of cables used
for measuring to some extent. However, the largest gain in terms of reduction of cables
running from the high voltage bays to the control building is achieved if the MUs are
placed in outdoor cubicles in the substation yard. It may be discussed whether the indoor
placement of MUs actually fulfills the concept of digital substations where measurements
are digitized close to the source, and if such a solution realizes the full potential of digital
substations.
When it comes to the issue regarding how to perform accredited meter inspections in
digital substations, the full potential of digital substations will not be realized if the
solution is to apply CITs for electrical energy metering applications. A solution where
one is forced to keep some parts of the substation analog and others digital will probably
not be desirable.
112
Chapter 12
Conclusion and Future Work
12.1 Conclusion
In this thesis a literature review and a product survey on OCTs have been performed in
order to evaluate whether OCTs are suitable for application in digital substations in the
transmission network. The analysis includes a discussion of the benefits and drawbacks
of applying OCTs in digital substations compared to the use of CCTs in conventional
substations in addition to an evaluation of potential performance limitations, maturity
of the products and solutions, and fulfillment of the substation application requirements
and the digital substation concept.
Experimental testing of a SAMU has also been conducted in order to investigate its steady
state and transient performance and compare with those of a conventional analog acqui-
sition system. In addition, it has been investigated whether a hybrid mode configuration
could have an impact on the performance of transformer differential protection. In the
rest of this section, answers to the main research questions stated in Section 1.2, will be
given and final conclusions drawn.
The literature review and product survey have shown that OCTs offer many benefits
compared to CCTs. Some of these are the improved safety of personnel and equipment in
the substation, the reduced environmental impact and the reduced need for maintenance.
OCTs also have a lower weight, which reduces the need for heavy cranes during installation
and replacement, and facilitates integration of the sensor heads into or in conjunction with
other primary equipment. This allows space to be saved in the substation yard, which in
turn can reduce the footprint of future digital substations. In addition, the application of
OCTs in combination with the process bus makes the design of digital substations flexible
and facilitates the integration of several application functions into the same secondary
IED, which allows space to be saved in the control building. Even though the price of
OCTs are presently about twice as high as that of a set of three CCTs, the application of
OCTs and the process bus allows many other costs to be reduced in a substation.
113
CHAPTER 12. CONCLUSION AND FUTURE WORK
On the other hand, it may be expected that there will be more challenges related to
interoperability and compatibility between NCITs with associated MUs, SAMUs and sec-
ondary IEDs in digital substations. CCTs have the benefit of a standardized and fairly
stable analog interface. The digital interface of NCITs and SAMUs is defined in the IEC
61850 standards. However, the standards leave some room for interpretation, which may
lead to lack of interoperability between devices from different vendors. Compatibility
issues when connecting different devices may also occur. Real-life testing of multivendor
digital substations is thus of great importance in order to ensure that the whole system is
working and that its performance is as desired. In addition, competence building will be
required since it is likely that most workers do not have experience with OCTs and digital
substations. Digital substations may also be expected to become more vulnerable than
the present day conventional substations due to factors such as the outdoor placement of
electronics and the increased need for highly accurate time synchronization.
The literature review has not uncovered any limitations in the performance of OCTs
themselves which may prevent their application in digital transmission substations. OCTs
can offer improvements in both steady state and transient performance compared to
CCTs. OCTs have the ability to maintain a high accuracy over a wide dynamic range,
can measure primary currents over a wide bandwidth and cannot saturate. However, the
application of OCTs in digital substations introduces some new considerations to take
into account. Inaccuracies in the sampling synchronization will result in phase errors
in the measured current, which may cause incorrect operation of differential protection
functions. The potential effects of the process bus network, such as time delays and
SV that are malformed, lost or arrive at the secondary IEDs out of sequence, can also
affect the performance of protection functions. It is important to ensure highly accurate
and reliable time synchronization and to properly design the process bus in order to
minimize the risk of the network affecting the performance of protection functions or
other application functions in digital substations.
Based on information from the literature review, it may be expected that different NCITs
and SAMUs will have different transient performances. Since combinations of NCITs and
SAMUs, possibly from different vendors, are likely to be applied in future digital substa-
tions, it is important to ensure that any differences in the transient performances will not
affect the performance of the protection functions. In particular, this applies to differen-
tial protection functions. The laboratory tests showed that differences in the ability of
the input current elements of a SAMU and a conventional protection IED to reproduce
the DC component in transient fault currents, resulted in fundamental frequency differ-
ential currents. Even though the resulting differential currents were too small to affect
the performance of the transformer differential protection, one should be aware of the
114
issue when combining different technologies in digital substations. Other combinations

of SAMUs and NCITs under different operating conditions could cause larger differential
currents and possibly result in incorrect operation of protection functions.
CCTs have the benefit of having been in wide use for a long time. CCTs have proven to
be reliable and stable over time, utilities and manufacturers are well acquainted with the
technology and there is a large selection of commercially available products and compatible
secondary equipment. When it comes to OCTs and SAMUs, on the other hand, there
are several signs that the products and solutions are not completely mature yet. There
are still several unanswered questions regarding the application of OCTs and SAMUs in
digital substations. In addition, parts of the IEC instrument transformer standard are
still being developed and have not yet been released. The product survey showed that
there are several vendors offering OCTs presently. However, several of these OCTs do not
support desirable and recommended features, such as PTP time synchronization, which
may limit the selection of suitable products at the current time.
CCTs meet the requirements of the applications in present day transmission substations.
However, in order to do so, several CT cores and a large number of copper cables are
required. OCTs, on the other hand, can fulfill the requirements of transmission substation
applications with just a single sensing element per phase, possibly two for redundancy.
However, there is one exception. Since it has not yet been figured out how to perform
accredited meter inspections in digital substations, which is a regulatory requirement,
OCTs cannot be said to meet the requirements of electrical energy metering at the present
time.
A possible solution to the above-mentioned issue is to apply CITs for electrical energy
metering applications. However, this solution does not fulfill the digital substation con-
cept. If this solution has to be employed, the full potential of digital substations will
not be reached. The product survey showed that most of the present day commercial
OCT solutions are based on an indoor placement of the MU in a protection and control
panel. It can be questioned whether this solution actually fulfills the digital substation
concept and if the full potential of digital substations will be realized if such solutions are
employed.
Based on the findings in this thesis, it can be concluded that OCTs seems to be suitable
for application in digital substations in the transmission network and can offer many
important benefits compared to CCTs. However, there are several signs that the products
and solutions are not completely mature yet. There are several investigations to be made,
questions to be answered and issues to be solved before a migration from the present day
conventional substations with CCTs to digital substations with OCTs can take place. In
115
this work real-life testing of digital substations to ensure that the different devices are
interoperable and that the performance of the resulting system is as desired, is of great
importance.
12.2 Future Work

Due to limitations in time and the absence of an actual OCT, some investigations were
left for future work. OCT technology was the only NCCT technology considered in this
thesis. This is the type of NCIT Statnett has installed in their pilot digital substation.
However, it should be investigated whether there are other types of NCCTs and NCVTs
commercially available which are suitable for digital substations in the transmission grid.
One point which has not been covered in this thesis is how to test OCTs and secondary
IEDs in digital substations. It should be investigated how to perform testing in a digital
substation and if there could be any potential problems here. The potential cost reductions
of applying OCTs in digital substations should also be investigated in further detail. It
was only briefly mentioned in this thesis that digital substations with OCTs allows costs
to be saved in several areas. Estimates of the cost savings compared to the increased price
of OCTs and other potential increased cost should be made in order to assess whether the
total solution will be cost-effective.
Due to the absence of an actual OCT, experimental tests of such a sensor could not
be performed. Thus, future work includes performing tests of an OCTs to corroborate
the findings in this thesis. Further investigations in terms of the performance of protec-
tion functions when combining NCITs and SAMUs with different transient performances
should also be performed.
116
Bibliography
[1] T. Buhagiar, J. P. Cayuela, A. Procopiou, and S. Richards, “Poste intelligent -

The next generation smart substation for the French power grid”, in 13th Inter-
national Conference on Development in Power System Protection 2016 (DPSP),
2016, pp. 1–4.
[2] K. Hinkley and C. Mistry, “First digital substation in TransGrid – Australia:
A journey, business case, lessons”, in The IET 14th International Conference on
Developments in Power System Protection 2018 (DPSP), 2018, pp. 1–6.
[3] C. Patterson, P. Mohapatra, C. Fundulea, C. Popescu-Cirstucescu, and P. New-
man, “Fitness - GB’s pilot multi-vendor digital substation - Architecture and design
philosophy”, in 2017 PAC World Conference, 2017, pp. 1–13.
[4] N. Hurzuk, S. Losnedal, S. Ingebrigtsen, H. H. Dietrichson, and R. S. J. Løken,
Digitale stasjoner i transmisjonsnettet. [Online]. Available: http://docplayer.
me/48638751- Digitale- stasjoner- i- transmisjonsnettet.html (visited on
08/28/2017).
[5] O. Kaspersen, Innovativ teknologi, 2017. [Online]. Available: http : / / www . sta
tnett . no / PageFiles / 12902 / Innovativ % 5C % 20Teknologi . pdf (visited on
06/07/2018).
[6] S. Weiss, P. Graeve, and A. Andersson, “Benefits of converting conventional instru-
ment transformer data into smart grid capable process data utilizing IEC 61850
merging unit”, in 21st International Conference on Electricity Distribution, 2011,
pp. 1–4.
[7] Network Protection & Automation Guide, 5th ed. Alstom Grid, 2011.
[8] F. Rahmatian and J. N. Blake, “Applications of high-voltage fiber optic current
sensors”, in 2006 IEEE Power Engineering Society General Meeting, 2006, pp. 1–6.
[9] I. U. Rian, “Digital substations in transmission networks”, Specialization project,
Norwegian University of Science and Technology, Trondheim, 2017.
[10] B. M. Buchholz and Z. Styczynski, “Modern technologies and the smart grid chal-
lenges in transmission networks”, in Smart Grids – Fundamentals and Technologies
in Electricity Networks. Springer, 2014, ch. 3, pp. 61–119.
[11] Lov om produksjon, omforming, overføring, omsetning, fordeling og bruk av energi
m.m. (energiloven), 1990. [Online]. Available: https://lovdata.no/dokument/
NL/lov/1990-06-29-50 (visited on 06/14/2018).
[12] R. Gore, H. Satheesh, M. Varier, and S. Valsan, “Analysis of an IEC 61850 based
electric substation communication architecture”, in 2016 7th International Con-
117
BIBLIOGRAPHY
ference on Intelligent Systems, Modelling and Simulation (ISMS), 2016, pp. 388–
393.
[13] E. Padilla, Substation Automation Systems: Design and Implementation. Wiley,
2015.
[14] S. Borlase, M. C. Janssen, M. Pesin, and B. Wojszczyk, “Role of substations in
smart grids”, in Electric Power Substations Engineering, J. D. McDonald, Ed.,
3rd ed. CRC Press, 2012, ch. 22, pp. 1–30.
[15] K. P. Brand et al., “The introduction of IEC 61850 and its impact on protection
and automation within substations”, pp. 1–102, 2007.
[16] M. Golshani, G. A. Taylor, and I. Pisica, “Simulation of power system substation
communications architecture based on IEC 61850 standard”, in 2014 49th Inter-
national Universities Power Engineering Conference (UPEC), 2014, pp. 1–6.
[17] H. Lei, C. Singh, and A. Sprintson, “Reliability modeling and analysis of IEC 61850
based substation protection systems”, IEEE Transactions on Smart Grid, vol. 5,
no. 5, pp. 2194–2202, 2014.
[18] S. Kunsman, S. Meier, and R. Hedding, “Protection and control system impacts
from the digital world”, in 2016 69th Annual Conference for Protective Relay En-
gineers (CPRE), 2016, pp. 1–13.
[19] J. W. Evans, “Interface between automation and the substation”, in Electric Power
Substations Engineering, J. D. McDonald, Ed., 3rd ed. CRC Press, 2012, ch. 6,
pp. 1–29.
[20] S. H. Horowitz and A. G. Phadke, “Current and voltage transformers”, in Power
System Relaying, 4th ed. John Wiley and Sons, 2014, ch. 3, pp. 46–73.
[21] Instrument transformers - Part 2: Additional requirements for current transform-
ers, IEC 61869-2:2012, 2012.
[22] E. Padilla, Substation Automation Systems: Design and Implementation. Wiley,
2015.
[23] D. H. Wilson, J. W. Wright, S. Richards, and P. Mohapatra, “Functionality, design
and benefit analysis of end-to-end digital substation”, in 13th International Con-
ference on Development in Power System Protection 2016 (DPSP), 2016, pp. 1–
5.
[24] J. W. Evans, “Interface between automation and the substation”, in Electric Power
Substations Engineering, J. D. McDonald, Ed., 3rd ed. CRC Press, 2012, ch. 6,
pp. 1–29.
[25] J. Ekanayake, K. Liyanage, J. Wu, A. Yokoyama, and N. Jenkins, Smart Grid:
Technology and Applications. Wiley, 2012.
[26] L. Hewitson, M. Brown, and R. Balakrishnan, “Instrument transformers”, in Prac-
tical Power System Protection. Elsevier Science, 2005, ch. 6, pp. 45–69.
118
BIBLIOGRAPHY
[27] S. Holst and J. Zakonjšek, “Transient behaviour of conventional current transform-

ers and its impact on performance of conventional and process bus based protection
systems”, in Protection, Automation & Control World (PAC World) Conference
2016, 2016, pp. 1–8.
[28] G. Ziegler, Numerical Differential Protection: Principles and Applications. John
Wiley & Sons, 2012.
[29] Instrument transformers - Part 6: Additional general requirements for low-power
instrument transformers, IEC 61869-6:2016, 2016.
[30] Application manual, transformer protection RET670 2.0, IEC, ABB, 2016. [On-
line]. Available: https : / / search - ext . abb . com / library / Download . aspx ?
DocumentID = 1MRK504138 - UEN & LanguageCode = en & DocumentPartId = &Action =
Launch (visited on 01/02/2018).
[31] J. D. P. Hrabluik, “Optical current sensors eliminate CT saturation”, in 2002 IEEE
Power Engineering Society Winter Meeting. Conference Proceedings, vol. 2, 2002,
pp. 1478–1481.
[32] K. P. Brand, C. Brunner, and I. D. Mesmaeker, “How to complete a substation
automation system with an IEC 61850 process bus”, Electra, no. 255, pp. 12–24,
2011.
[33] Instrument transformers - Part 9: Digital interface for instrument transformers,
IEC 61869-9:2016, 2016.
[34] K. P. Brand et al., “The introduction of IEC 61850 and its impact on protection
and automation within substations”, pp. 1–102, 2007.
[35] S. Kucuksari and G. G. Karady, “Experimental comparison of conventional and
optical current transformers”, IEEE Transactions on Power Delivery, vol. 25, no. 4,
pp. 2455–2463, 2010.
[36] J. Schmid and K. Kunde, “Application of non conventional voltage and currents
sensors in high voltage transmission and distribution systems”, in 2011 IEEE Inter-
national Conference on Smart Measurements of Future Grids (SMFG) Proceedings,
2011, pp. 64–68.
[37] Brocheure current transformers, Arteche. [Online]. Available: https://www.arte
che.com/en/products/current-transformers (visited on 04/22/2018).
[38] “IEEE guide for application of optical instrument transformers for protective re-
laying”, IEEE Std C37.241-2017, pp. 1–50, 2018.
[39] P. Roccato, A. Sardi, G. Pasini, L. Peretto, and R. Tinarelli, “Traceability of MV
low power instrument transformer LPIT”, in 2013 IEEE International Workshop
on Applied Measurements for Power Systems (AMPS), 2013, pp. 13–18.
[40] S. Richards et al., “Feedback on in-service deployment of the fully digital substa-
tion”, in 2014 CIGRE, 2014.
119
BIBLIOGRAPHY
[41] A. Apostolov, “Substation automation systems”, in Smart Grid Handbook, C.-C.

Liu, S. McArthur, and S. J. Lee, Eds. Wiley, 2016, vol. 1, ch. 20, pp. 333–356.
[42] K.-P. Brand and I. D. Mesmaeker, “Power system protection”, in Handbook of
Electrical Power System Dynamics: Modeling, Stability, and Control, M. Eremia
and M. Shahidehpour, Eds. John Wiley & Sons Inc., 2013, ch. 12, pp. 737–785.
[43] A. Apostolov, “Efficient maintenance testing in digital substations based on IEC
61850 edition 2”, Protection and Control of Modern Power Systems, vol. 2, no. 1,
pp. 1–37, 2017.
[44] Q. Yang, D. Keckalo, D. Dolezilek, and E. Cenzon, “Testing IEC 61850 merging
units”, in 44th Annual Western Protective Relay Conference, Spokane, Washington,
2017, pp. 1–9. [Online]. Available: https : / / cdn . selinc . com / assets / Liter
ature/Publications/Technical%5C%20Papers/6823_TestingIEC61850_QY_
20170907_Web.pdf?v=20180123-085831 (visited on 02/02/2018).
[45] G. Rzepka, S. Wenke, and S. Walling, “Choose simplicity for a better digital sub-
station design”, in 2017 70th Annual Conference for Protective Relay Engineers
(CPRE), 2017, pp. 1–9.
[46] COSI-CT optical current transformer, GE Grid Solutions, 2016. [Online]. Available:
https://www.gegridsolutions.com/app/Resources.aspx?prod=cosi&type=1
(visited on 02/22/2018).
[47] Optical current transformers for air insulated substations, Trench, 2016. [Online].
Available: http : / / www . trenchgroup . com / en / Products - Solutions / Instr
ument - Transformers / Optical - Current - Transformers / Optical - Current -
Transformers (visited on 02/22/2018).
[48] Arteche SDO OCT EN, Arteche. [Online]. Available: https://www.arteche.com/
en/products/optical-current-transformer (visited on 02/22/2018).
[49] Fiber optics current sensor - Free standing (FOCS-FS), ABB, 2015. [Online]. Avail-
able: https://search- ext.abb.com/library/Download.aspx?DocumentID=
2GJA708628 & LanguageCode = en & DocumentPartId = &Action = Launch (visited on
02/22/2018).
[50] J. Wang, F. Viawan, and T. Werner, “Effects of sensor technology on differen-
tial protection”, in 10th IET International Conference on Developments in Power
System Protection (DPSP 2010). Managing the Change, 2010, pp. 1–5.
[51] H. Lehpamer, “Introduction to power utility communications”, in. Artech House,
2016, ch. 2 Communications Fundamentals.
[52] S. Meier, T. Werner, and C. Popescu-Cirstucescu, “Performance considerations in
digital substations”, in 13th International Conference on Development in Power
System Protection 2016 (DPSP), 2016, pp. 1–9.
120
BIBLIOGRAPHY
[53] S. Meier, Webinar: We are bridging the gap. Enabling the ABB digital substation.
[Online]. Available: http : / / new . abb . com / network - management / webinars /
2016/abbs-digital-substation-webinar (visited on 10/05/2017).
[54] Communication networks and systems for power utility automation - Part 9-2:
Specific communication service mapping (SCSM) - sampled values over ISO/IEC
8802-3, IEC 61850-9-2:2011, 2011.
[55] F. Steinhauser, “Sampled values - History, status, and outlook”, in 2017 PAC
World Conference, 2017, pp. 1–5.
[56] Implementation guideline for digital interface to instrument transformers using
IEC 61850-9-2, UCA International Users Group, 2004.
[57] Instrument transformers - Part 1: General requirements, IEC 61869-1:2007, 2007.
[58] TC 38 Stability date of publications, International Electrotechnical Commission,
2018. [Online]. Available: http : / / www . iec . ch / dyn / www / f ? p = 103 : 21 :
2796706578222:::21:FSP_ORG_ID,FSP_LANG_ID:1241,25 (visited on 06/13/2018).
[59] TC 38 Work programme, International Electrotechnical Commission, 2018. [On-
line]. Available: http://www.iec.ch/dyn/www/f?p=103:23:2796706578222:::
23:FSP_ORG_ID,FSP_LANG_ID:1241,25 (visited on 06/14/2018).
[60] V. Skendzic and D. Dolezilek, “New and emerging solutions for sampled value
process bus IEC 61850-9-2 standard - an editor’s perspective”, in Southern African
Power System Protection & Automation Conference, Johannesburg, South Africa,
2017, pp. 1–10. [Online]. Available: https://cdn.selinc.com/assets/Literatu
re/Publications/Technical%5C%20Papers/6797_NewEmerging_DD_20170309_
Web2.pdf?v=20171115-101441 (visited on 02/02/2018).
[61] Instrument transformers - Part 8: Electronic current transformers, IEC 60044-
8:2002, 2002.
[62] Forskrift om elektriske forsyningsanlegg, 2005. [Online]. Available: https://lovd
ata.no/dokument/SF/forskrift/2005-12-20-1626 (visited on 06/12/2018).
[63] Forskrift om leveringskvalitet i kraftsystemet, 2004. [Online]. Available: https://l
ovdata.no/dokument/SF/forskrift/2004-11-30-1557 (visited on 06/12/2018).
[64] Forskrift om forebyggende sikkerhet og beredskap i energiforsyningen (beredskaps-
forskriften), 2012. [Online]. Available: https : / / lovdata . no / dokument / SF /
forskrift/2012-12-07-1157 (visited on 06/12/2018).
[65] Forskrift om systemansvaret i kraftsystemet, 2002. [Online]. Available: https://
lovdata.no/dokument/SF/forskrift/2002-05-07-448 (visited on 06/12/2018).
[66] Funksjonskrav i kraftsystemet, 2012. [Online]. Available: http : / / www . statnet
t.no/Global/Dokumenter/Kraftsystemet/Systemansvar/FIKS%202012.pdf
(visited on 06/12/2018).
[67] J. Skaar and T. Holtebekk, Polarisert lys, 2018. [Online]. Available: https://snl.
no/polarisert_lys.
121
BIBLIOGRAPHY
[68] S. Ziegler, R. C. Woodward, H. H. C. Iu, and L. J. Borle, “Current sensing tech-

niques: A review”, IEEE Sensors Journal, vol. 9, no. 4, pp. 354–376, 2009.
[69] A. K. Zvezdin and V. A. Kotov, Modern Magnetooptics and Magnetooptical Ma-
terials, J. M. D. Coey and D. R. Tilley, Eds., ser. Studies in Condensed Matter
Physics. Institute of Physics Publishing, 1997.
[70] J. Tioh, R. J. Weber, and M. Mina, “Magneto-optical switches”, in Optical Switches:
Materials and Design, S. J. Chua and B. Li, Eds. Elsevier, 2010, ch. 4, pp. 97–135.
[71] File:linearly polarized wave.svg, Wikimedia Commons, 2010. [Online]. Available:
https://commons.wikimedia.org/wiki/File:Linearly_Polarized_Wave.svg
(visited on 06/22/2018).
[72] File:circular.polarization.circularly.polarized.light with.components left.handed.svg,
Wikimedia Commons, 2010. [Online]. Available: https://commons.wikimedia.o
rg/wiki/File:Circular.Polarization.Circularly.Polarized.Light_With.
Components_Left.Handed.svg (visited on 06/22/2018).
[73] “Optical current transducers for power systems: A review”, IEEE Transactions on
Power Delivery, vol. 9, no. 4, pp. 1778–1788, 1994.
[74] M. H. Samimi, A. A. S. Akmal, and H. Mohseni, “Optical current transducers and
error sources in them: A review”, IEEE Sensors Journal, vol. 15, no. 9, pp. 4721–
4728, 2015.
[75] A. J. Rogers, “Elements of polarization optics”, in Polarization in Optical Fibers.
Artech House, 2008, ch. 3, pp. 75–112.
[76] R. Wang, S. Xu, W. Li, and X. Wang, “Optical fiber current sensor research: Review
and outlook”, Optical and Quantum Electronics, vol. 48, no. 9, pp. 1–22, 2016.
[Online]. Available: https://link.springer.com/article/10.1007/s11082-
016-0719-3 (visited on 04/17/2018).
[77] E. M. Esmail, N. I. Elkalashy, T. Kawady, and A. M. I. Taalab, “Experimental
implementation of optical current transducers”, in 2016 Eighteenth International
Middle East Power Systems Conference (MEPCON), 2016, pp. 276–281.
[78] W. Wang, X. Wang, and J. Xia, “The nonreciprocal errors in fiber optic current
sensors”, Optics & Laser Technology, vol. 43, no. 8, pp. 1470–1474, 2011. [On-
line]. Available: https : / / www . sciencedirect . com / science / article / pii /
S0030399211001277 (visited on 04/25/2018).
[79] R. M. Silva, H. Martins, I. Nascimento, J. M. Baptista, A. L. Ribeiro, J. L. San-
tos, P. Jorge, and O. Frazão, “Optical current sensors for high power systems:
A review”, Applied sciences, vol. 2, no. 3, pp. 602–628, 2012. [Online]. Available:
http://www.mdpi.com/2076-3417/2/3/602/htm (visited on 03/05/2018).
[80] F. Rahmatian, “Optical instrument transformers”, PAC World Magazine March
20018 Issue, pp. 18–25, 2018.
122
BIBLIOGRAPHY
[81] K. T. V. Grattan and Y. N. Ning, “Optical current sensor technology”, in Opti-

cal Fiber Sensor Technology: Applications and Systems, L. S. Grattan and B. T.
Meggitt, Eds., Springer, 1999, ch. 7, pp. 183–223.
[82] K. Bohnert, H. Brandle, M. G. Brunzel, P. Gabus, and P. Guggenbach, “Highly ac-
curate fiber-optic dc current sensor for the electrowinning industry”, IEEE Trans-
actions on Industry Applications, vol. 43, no. 1, pp. 180–187, 2007.
[83] K. Bohnert, P. Gabus, J. Nehring, and H. Brandle, “Temperature and vibration
insensitive fiber-optic current sensor”, Journal of Lightwave Technology, vol. 20,
no. 2, pp. 267–276, 2002.
[84] K. Bohnert, R. Thomas, and M. Mendik, “Light measures current”, no. 1, pp. 12–
17, 2014. [Online]. Available: https:/ /library.e.abb .com/public/0d948ce
db40451cec1257ca900532dd0 / 12 - 17 % 5C % 201m411 _ EN _ 72dpi . pdf (visited on
04/22/2018).
[85] K. Bohnert, C. P. Hsu, L. Yang, A. Frank, G. M. Müller, and P. Gabus, “Fiber-optic
current sensor tolerant to imperfections of polarization-maintaining fiber connec-
tors”, Journal of Lightwave Technology, vol. 36, no. 11, pp. 2161–2165, 2018.
[86] B. Li and L. Li, “An overview of the optical current sensor”, in 2012 Interna-
tional Conference on Computer Science and Electronics Engineering, vol. 3, 2012,
pp. 202–206.
[87] P. Ripka, “Electric current sensors: A review”, Measurement Science and Technol-
ogy, vol. 21, no. 11, pp. 1–24, 2010. [Online]. Available: http://iopscience.iop.
org/article/10.1088/0957-0233/21/11/112001/meta (visited on 03/05/2018).
[88] B. Lee, “Review of the present status of optical fiber sensors”, Optical Fiber
Technology, vol. 9, no. 2, pp. 57–79, 2003. [Online]. Available: https : / / www .
sciencedirect . com / science / article / pii / S1068520002005278 (visited on
04/25/2018).
[89] Y. Ding, Y. Dong, J. Zhu, B. Zheng, D. Zhang, and W. Hu, “Linear birefringence
and imperfect quarter wave plate effects on optic-fiber current sensor”, in 2011 Asia
Communications and Photonics Conference and Exhibition (ACP), 2011, pp. 1–6.
[90] J. Blake, P. Tantaswadi, and R. T. de Carvalho, “In-line sagnac interferometer
current sensor”, IEEE Transactions on Power Delivery, vol. 11, no. 1, pp. 116–
121, 1996.
[91] M. Lenner, W. Quan, G. Müller, L. Yang, A. Frank, and K. Bohnert, “Interferomet-
ric fiber-optic current sensor with inherent source wavelength shift compensation”,
in 2015 IEEE SENSORS, 2015, pp. 1–4.
[92] Instrument transformers - The use of instrument transformers for power quality
measurement, IEC TR 61869-103:2012, 2012.
[93] R. Thomas, A. Vujanic, D. Z. Xu, J. E. Sjödin, H. R. M. Salazar, M. Yang, and N.
Powers, “Non-conventional instrument transformers enabling digital substations
123
BIBLIOGRAPHY
for future grid”, in 2016 IEEE/PES Transmission and Distribution Conference

and Exposition (T D), 2016, pp. 1–5.
[94] M. Normandeau and J. Mahseredjian, “Evaluation of low-power instrument trans-
formers for generator differential protection”, IEEE Transactions on Power Deliv-
ery, vol. 33, no. 3, pp. 1143–1152, 2018.
[95] K. Bohnert, P. Gabus, J. Kostovic, and H. Brändle, “Optical fiber sensors for
the electric power industry”, Optics and Lasers in Engineering, vol. 43, no. 3-5,
pp. 511–526, 2005.
[96] Instrument transformers, Trench, Trench Group. [Online]. Available: http://www.
trenchgroup.com/en/Products-Solutions/Instrument-Transformers/Instr
ument-Transformers (visited on 04/22/2018).
[97] Product guide, process bus I/O system SAM600 version 1.2, ABB, 2017. [Online].
Available: https://search-ext.abb.com/library/Download.aspx?DocumentID
=1MRK511437-BEN&LanguageCode=en&DocumentPartId=&Action=Launch (visited
on 01/30/2018).
[98] Reason MU320 digital substation merging unit, GE Grid Solutions, 2017. [Online].
Available: https://www.gegridsolutions.com/app/ViewFiles.aspx?prod=
MU320&type=1 (visited on 04/22/2018).
[99] J. Jesus and A. Varghese, “Feeder differential protection in hybrid mode: Scheme
performance with mix of -9-2le sampled values and analogue inputs”, in 2017 In-
ternational Conference on Modern Power Systems (MPS), 2017, pp. 1–5.
[100] C. A. Dutra, I. H. Cruz, T. A. Franzen, R. R. Matos, F. C. Neves, L. B. Oliveira,
and G. Krefta, “Comparison of analogue measurements between merging units
and conventional acquisition systems”, in 12th IET International Conference on
Developments in Power System Protection (DPSP 2014), 2014, pp. 1–5.
[101] A. Bonetti, IEC 61850 standard and related iec committees as reference for relay
protection applications, 2018. [Online]. Available: http://gnu.ets.kth.se/~nt/
tmp/protaut/slides/ind_iec61850_standard_iec_committees_protection_
applications.pdf (visited on 06/13/2018).
[102] Communication networks and systems for power utility automation - Part 5: Com-
munication requirements for functions and device models, IEC 61850-5:2013, 2013.
[103] Communication networks and systems for power utility automation - Part 10: Con-
formance testing, IEC 61850-10:2012, 2012.
[104] S. Hutterer, R. S. J. Løken, and M. Schicklgruber, “Digital substation pilot: Sub-
station automation and protection system based on IEC 61850-9-2 process bus”,
in 2017 PAC World Conference, 2017, pp. 1–7.
[105] D. M. E. Ingram, P. Schaub, D. A. Campbell, and R. R. Taylor, “Evaluation of
precision time synchronisation methods for substation applications”, in 2012 IEEE
124
BIBLIOGRAPHY
International Symposium on Precision Clock Synchronization for Measurement,

Control and Communication Proceedings, 2012, pp. 1–6.
[106] V. Skendzic, I. Ender, and G. Zweigle, “IEC 61850-9-2 process bus and its impact
on power system protection and control reliability”, in 2007 9th Annual Western
Power Delivery Automation Conference, 2007, pp. 1–7.
[107] G. S. Antonova et al., “Standard profile for use of IEEE std 1588-2008 Precision
Time Protocol (PTP) in power system applications”, in 2013 66th Annual Confer-
ence for Protective Relay Engineers, 2013, pp. 322–336.
[108] X. Chen, H. Guo, and P. Crossley, “Interoperability performance assessment of mul-
tivendor IEC 61850 process bus”, IEEE Transactions on Power Delivery, vol. 31,
no. 4, pp. 1934–1944, 2016.
[109] X. Chen, H. Guo, and P. Crossley, “Performance testing of IEC 61850 based ar-
chitecture for UK national grid standardised substation automation solutions”, in
2015 IEEE Power Energy Society General Meeting, 2015, pp. 1–5.
[110] J. Haude et al., “Smart switchgear” using IEC standard 61850 - First experience
gained with a pilot project in a 380/110 kV substation”, in 2010 CIGRE, 2010,
pp. 1–8.
[111] C. R. del Castillo, J. F. Pozuelo, M. Goraj, R. Pereda, and A. Amezaga, “Re-
dundancy challenges on IEC 61850 systems and migration paths for IEC 61850
substation communication networks”, in 2012 CIGRE Paris Session, 2012, pp. 1–
14.
[112] D. M. E. Ingram, P. Schaub, R. R. Taylor, and D. A. Campbell, “System-level
tests of transformer differential protection using an IEC 61850 process bus”, IEEE
Transactions on Power Delivery, vol. 29, no. 3, pp. 1382–1389, 2014.
[113] O. Aune, “Ikke-konvensjonelle måletransformatorer i digital stasjon”, B.S. thesis,
Norwegian University of Science and Technology, Gjøvik, 2017.
[114] Communication networks and systems for power utility automation - Part 7-3:
Basic communication structure - common data classes, IEC 61850-7-3:2010, 2010.
[115] Technical manual, transformer protection RET670 2.0, IEC, ABB, 2016. [Online].
=1MRK504139-UEN&LanguageCode=en&DocumentPartId=&Action=Launch (visited
on 03/08/2018).
[116] R. Løken, M. W. Kristiansen, and S. Losnedal, “Expectations and specifications of
IEC 61850 based digital substation automation system used by Statnett”, in 2014
CIGRE, 2014.
[117] Z. Bajramović, I. Turković, A. Mujezinović, A. Čaršimamović, and A. Muhare-
mović, “Measures to reduce electromagnetic interferences on substation secondary
circuits”, in Proceedings ELMAR-2012, 2012, pp. 129–132.
[118] OMICRON CMC 356 reference manual, OMICRON electronics, 2014.
125
BIBLIOGRAPHY
[119] OMICRON getting started with OMICRON test universe 3.00, OMICRON elec-
tronics, 2014.
[120] Flyer SEL-2488 satellite-synchronized network clock, Schweitzer Engineering Labo-
ratories, 2014. [Online]. Available: https://cdn.selinc.com/assets/Literatur
e/Product%5C%20Literature/Flyers/2488_PF00327.pdf?v=20170728-141139
(visited on 01/02/2018).
[121] Product guide, transformer protection RET670 2.0, IEC, ABB, 2016. [Online].
=1MRK504141-BEN&LanguageCode=en&DocumentPartId=&Action=Launch (visited
on 01/30/2018).
[122] R. Løken and S. Losnedal, “Multifunctional protection and control IEDs in future
transmission systems”, in 2017 PAC World Conference, 2017, pp. 1–11.
126
Appendix A
E-Mail
A.1 Statnett
A.1.1 First E-Mail
Request:
Hei,
Sånn jeg har forstått det, benyttes det i dag godkjent utstyr for avregningsmåling i Stat-
netts stasjoner. I tillegg er det krav om at det må utføres en akkreditert målekjedekontroll
av noen med godkjenning etter at utstyret er satt i drift.
For digital stasjon lurer jeg da på hvilket utstyr som kan benyttes, evt. hvor funksjonen
avregningsmåling vil ligge? Foreligger det godkjent utstyr? Kan dette skape problemer
for innføringen av digital stasjon med optiske strømtransformatorer i Statnett? Hvem kan
utføre akkreditert målekjedekontroll for en digital stasjonsløsning?
Hvis det er slik at det ikke foreligger godkjent utsyr i dag, når forventes det at dette vil
foreligge?
Ønsker gjerne rask tilbakemelding med hensyn til fullføring av oppgaven.
Med vennlig hilsen
Ingvill Rian
127
APPENDIX A. E-MAIL
Reply:
Hei Ingvill,
Det stemmer at det i dag utføres akkreditert målekontroll for avregningsmåler etter at
den er satt på drift av godkjent akkrediterings selskap.
Det er ein avregningsmåler som kan benyttes i kontrollanlegg med prosessbus som kan
ta inn digitale målinger ved bruk av IEC 61850. Det er ikkje avklart korleis dei digitale
målingene skal akkrediteres. Dette er eit spørsmål som diskuteres internasjonalt. Eg er
usikker på når ei løysing er på plass, vi arbeider med problemstillinga i FoU Digital stasjon
i Statnett.
Ei løysing kan være å bruke konvensjonelle måletransformatorer for avregningsmåling,

men da tar vi ikkje i bruk teknikken heilt ut.
A.1.2 Second E-Mail
Hei,
Har du mulighet til å svare på disse spørsmålene i forbindelse med fullføring av master-
oppgaven? Skulle gjerne hatt svar så raskt som mulig.
1. Slik jeg har forstått det er Arteche’s løsning for optisk strømtransfor-
mator basert på innendørs montasje av merging unit inne i kontrollrom
sammen med vern- og kontroll-IED-er.
• Ønsker Statnett å plassere den tilhørende elektronikken til de optiske

strømtransformatorene ute i anlegget?
Det stemmer Arteche ønsker å plassere MU inne i kontrollrom, mens andre
leverandører må plassere dem i skap i ute i anlegget.
• Har Statnett sett på løsninger i forhold til utendørs plassering av

merging uniter i feltskap?
Statnett har planer om å ha en fleksibel løsning i forhold til plassering av MU,
slik at de enten kan plasseres inne i kontrollrom eller ute i anlegget.
• Finnes det egnede skap med oppvarming?

Statnett ser på løsninger andre har valgt og skal i år utforme skap som har
nødvendig oppvarming evt kjøling og EMP sikring for å plassere elektronikk
ute i apparatanlegget.
128
APPENDIX A. E-MAIL
2. Bruk av vern i digital stasjon.
• Støtter de vernene som Statnett benytter i dag prosessbuss, og kan

de benyttes i en digital stasjon?
Det er mange leverandører som støtter prosessbus, begge leverandørene vi i
dag bruker (ABB og Siemens) har vern som støtter prosessbus.
• Finnes det vern av alle typer som trengs i stasjoner, som støtter
prosessbuss og som Statnett har godkjent? Eller vil Statnett måtte
godkjenne nye vern for å kunne realisere en digital stasjon fullt ut?
Vernene som har prosessbus er nye versjoner som må godkjennes på nytt for de
kan tas i bruk. Det er uklart om det finnes gode løsninger for samleskinnevern
for større stasjoner, dette er en av sakene vi ser på i 2018.
3. Har Statnett gjort noen vurderinger i forhold til hvilke krav vernene vil
stille til de digitale målekretsene (optisk strømtransformator med tilhø-
rende merging unit, stand-alone-merging unit og prosessbuss/fiberoptis-
ke kabler) i forbindelse med prosjektet «Digital stasjon»?
• Det pågår arbeid i IEC TC95 AhWG3 som ser på krav som stilles til vern i
forhold til krav som stilles til måletransformatorer. I dag er det ikke samme
krav til transferkarakteristikken.
• Statnett arbeider med hvilke krav som skal stilles.
• Det diskuteres om typetest for godkjenning av nye produkt må inkluderer både
MU/SAMU og vern fra aktuelle leverandører i tillegg til krav som stilles.
4. Kunne jeg fått tilgang til teknisk dokumentasjon for Arteche’s optiske
strømtransformator, gjerne inkludert informasjon om hvilke tester som
har blitt utført?
Dette må eg sjekke.
5. Hvor stor datakapasitet (bit/s) har prosessbussnettverket i Statnetts pi-

lotstasjon? Tenker man å benytte samme datakapasitet i en fremtidig
digital stasjon?
Pilotstasjonen har 1 Gb prosessnettverk, det er og tenkt å bruke i framtida (vi har
og mulighet for 5 Gb). Verna har stort sett 100 Mb tilkobling til nettverket.
6. Bruk av fiber-optiske kabler i digital stasjon.
• Hvilke erfaringer har Statnett gjort vedrørende bruk av fiber-optiske

kabler i dagens anlegg?
129
APPENDIX A. E-MAIL
– Vi har i piloten kun lagt fibernettverk lokalt i kontrollrommet.

– Tilkobling til NCIT går på fiberoptisk forbindelse fra kontrollrom til ap-
paratanlegg.
– Det er behov for spesialkompetanse for å koble fibernettverk i Digital sta-
sjon.
• Har man erfart at det finnes støykilder eller andre problemer ved
bruk av slike kabler?
Vi har ikkje hatt denne type erfaring, men skal plassere utstyr i feltskap i
anlegg for å sjekke om det kan oppstå problem.
• Hvilke konsekvenser vil bruk av fiber-optiske kabler i stedet for kon-

vensjonelle kobberkabler få for kabelføringen og kabelleggingen i
Statnetts stasjoner?
– Statnett vil ha behov for mindre kabelføring fra kontrollrom til apparat-
anlegg.
– Det vurderes om vi endrer fra kabelkanaler til rør eller om vi legger ut små
kabel kanaler.
– Det vil trulig framleis være krav til to separate kanaler fra kontrollrom til
apparatanlegg.
– Vi ser for oss kostnadsreduksjon på kabling, kabelkanaler og moontasje.
Håper dette er til hjelp. Lykke til videre!
A.1.3 Third E-Mail
Hei,
Jeg har noen spørsmål angående konvensjonelle strømtransformatorer. Håper du kan

svare så raskt som mulig med tanke på fullføring av masteroppgaven.
1. Er det noen vedlikeholdsprosedyrer for de konvensjonelle strømtransfor-

matorene i Statnetts stasjoner?
Ja. Det er månedlige visuelle inspeksjoner, avlesing av evt. instrumenter og sjekk
av olje. Ikke særlig omfattende, men likevel noe som man da benytter litt tid på.
Anta eksempelvis 30 min per inspeksjon per måned per strømtrafo.
2. Hvor lang levetid har de konvensjonelle strømtransformatorene? Hvor

ofte bytter Statnett dem ut?
Variabelt, men avhengig av leverandør osv typisk 20-50 år.
130
APPENDIX A. E-MAIL
3. Omtrent hvor mye koster et sett med tre konvensjonelle strømtransfor-

matorer (f.eks. for systemspenning på 420 kV)?
Store variasjoner, men opptil 8500 Euro for en strømtrafo med tre vernkjerner og
to målekjerner.
A.2 ABB
Hi
Your Questions for our FOCS optical ct with replies from our factory for this:
1. What is the metering accuracy of FOCS-FS?

Class 0.2s for rated current above 400 A. For lower rated currents more prolonged
integration time is required.
2. Is FOCS-FS optimized for a specific application, such as metering or

protection, e.g. in terms of number of fiber coil turns or type of sensor
electronics? Or is a single FOCS-FS equally suited for both metering and
protection applications?
Same sensor covers both metering and protection applications.
3. Over which dynamic range are the metering and protection accuracy
specifications of FOCS-FS maintained?
Metering 0.2S from rated currents 400A to 164kA, class 0.2 from 80A to 164kA,
class 0.5 from 35A to 164kA. Protection 5P from rated currents 70A to 164kA.
4. How good is the ability of FOCS-FS to reproduce rapid changes in the

primary current?
The main limiting factor is the digital output protocol (IEC 61850-9-2LE). This
limits for 50Hz the output bandwidth to 2kHz or 6.4kHz. Internally the FOCS is
acquiring raw measurement data with more than 1MS/s, which means it will react
to changes faster than 500kHz.
5. What is the bandwidth of FOCS-FS? Does the opto-electronic module

impose any limitations on the bandwidth of the digital output?
As said in the point 4, the bandwidth is determined by the standard for process
bus protocol. Physical limitation of opto-electronic module is given by the length
of the sensor head optical cable which influences the time of flight of light from the
OE-unit to sensor head and back. This results in raw sampling rate of less than
2MS/s.
131
APPENDIX A. E-MAIL
6. How good is the ability of FOCS-FS to reproduce rapid changes in the

primary frequency?
FOCS is measuring how much current has flown through the sensor head aperture
during some short time of less than 1 microsecond. There is no direct influence of
the primary current signal form (i.e. frequency variations) to the FOCS. Again is
the final limit given with the output stream sampled values rate.
7. Does the opto-electronic module support other methods for time syn-
chronization than PPS?
No.
8. What is the expected lifetime of FOCS-FS?

Expected lifetime of opto-electronic units is comparable to the one for IEDs (15
years), lifetime of optical sensor is 30+ years.
9. Do you supply other non-conventional current transformers, or any non-

conventional voltage transformers?
ABB does not have optical voltage transformers with IEC61850-9-2 output. For
GIS application, there are capacitive voltage sensors with such output.
Let us know if you need any more information.
A.3 Arteche
A.3.1 First E-Mail
Dear Ingvill,
Thanks for your e-mail.
Most of your questions cover 2 main topics:
a) Questions 2,3 and 4 are related to the sample rate. Internally, we work with 192kHz
sampling rate. However, the output we provide (sampled values) is filtered so the
output is compliant with the recommendations of the IEC61869-9 standard.
b) Questions 1 and 7 are related to the accuracy and dimensioning of the OCT. There
are 2 basic principles here (see the attached document for more details) that explain
why the dynamic range is virtually unlimited:
i. You can see the measurement error as electronic noise → The lower the signal
132
APPENDIX A. E-MAIL
the more difficult to measure accurately

But, it is possible to increase the signal to noise ratio by adding fiber turns
at the sensing coil, so we can ensure accuracy at low currents by adding more
fiber turns.
ii. The response of the OCT is very linear and it does not saturate → High currents
are not an issue (from the point of view of accuracy)
You can find more detailed explanations on the attached document.
Please find my answers to each question below in red.
Should you need further clarifications, please do not hesitate to contact us.
1. Over which dynamic range are the metering and protection accuracy
specifications of SDO OCT maintained?
It is virtually unlimited. Designs for Cl0.2 accuracy are possible for any given
current levels.
2. How good is the ability of SDO OCT to reproduce rapid changes in the
primary current?
The performance is very good. This is related to the sample rate that is selected.
We have implemented the sample rates that the IEC61869-9 standard defines for
digital instrument transformers.
3. Does the SDO MU impose any limitations on the bandwidth of the digital
output?
No. The only constraint is the standard for the digital output (IEC61869-9), the
MU does not impose any additional constraints.
4. How good is the ability of SDO OCT to reproduce rapid changes in the
primary frequency?
This is related to the sample rate. Internally we 192kHz sample rate and the output
(sampled values) are as per the above mentioned standard. Please also see the above
answers.
5. What is the maximum permissible length of the fiber optic cables con-
necting the SDO ICT to the SDO MU?
No limitation for digital substation applications. For some special applications, we
have installations with distances as high as 20km
6. What is the expected lifetime of SDO OCT?
133
APPENDIX A. E-MAIL
Tricky question. There is no specific aging test that we conducted so far so I cannot
commit to a given period of time. However, there is no maintenance requirement in
the OCT, so its useful life should be at least as long as the useful life of equivalent
conventional CTs. This is applicable for the primary part, the Merging Unit is
an IED and it has a useful life that is equivalent to that of conventional digital
protection IED-s.
7. Is SDO OCT optimized for a specific application, such as metering or

electronics? Or is a single SDO OCT equally suited for both metering
and protection applications?
A single device covers both metering and protections applications. The response of
the OCT is very linear and there is no saturation, so at very high currents, we can
keep an excellent accuracy class (way better than protection accuracy requirements).
Please see the attached document for more details on this.
8. Do you supply other non-conventional current transformers other than

SDO OCT, or any non-conventional voltage transformers?
As of today, for HV applications we supply Capacitive VTs. For lower voltages (up
to 36kV) we are also supplying resistive dividers and Rogowski coils.
A.3.2 Second E-Mail
Dear Ingvill,
Please find below our answers.
1. What is the weight of SDO OCT (e.g. with a system voltage rating of
420 kV)? How does this compare to the weight of a conventional current
transformer with the same system voltage rating?
It is much lighter than a conventional CT. The sensor head is 15kg and it can be
used for any voltage level. For 420kV, I estimate that the total weight would be
approx. 160kg.
2. How much does SDO OCT cost (sensor heads and SDO MU, respec-
tively)? How does this compare to the cost of a comparable conventional
current transformer?
You can consider approx. 45k€ for a complete set of 3 CT-s and 1 MU. Once the
product is industrialized, we expect the optical CT to be more cost-effective than a
conventional one for voltages above 245kV.
134
APPENDIX A. E-MAIL
3. What is the operating voltage range of SDO OCT (e.g. with a system
voltage rating of 420 kV)? Can a SDO OCT with a system voltage rating
of 420 kV be operated at 300 kV?
Unlimited, the voltage level doesn’t influence on the operation of the OCT. It is
just a matter of selecting the proper insulator. So yes, a 420kV device can be used
for 300kV.
4. What is the processing delay time of SDO MU?

The “delay time” is less than 1,58 ms.
5. Does SDO MU support PRP and HSR network redundancy protocols?

Yes
6. According to the SDO OCT brochure, “the rotation of the polarization

state of the light is measured interferometrically”. What type of configu-
ration does the SDO OCT employ (Sagnac loop interferometer, reflective
interferometer, other)?
It is a 3x3 Sagnac interferometer
7. Does the accuracy of SDO OCT change over the given operating temper-
ature range (-40°C to +85°C)? Have any measures been taken in order
to reduce the sensitivity of SDO OCT to temperature variations? Does
SDO OCT employ temperature compensation?
No, we guarantee accuracy over the whole temperature range. We employ temper-
ature compensation by sensing the temperature of the sensor-fiber.
8. Is the accuracy of SDO OCT sensitive to vibrations and stray magnetic

fields? Have any measures been taken in order to reduce the sensitivity
of SDO OCT to vibrations and stray magnetic fields?
The design is immune to vibrations and stray magnetic fields.
9. Does SDO OCT require any maintenance? In that case, what are the
maintenance procedures and how often do they have to be performed?
No specific maintenance is required.
10. How is the procedure of replacing sensor heads and SDO MU? If a single
sensor head fails, is it possible to only replace this sensor head or does
all of the sensor heads and the SDO MU have to be replaced?
It is possible to replace any given sensor head or any given MU without on-site
calibration.
135
APPENDIX A. E-MAIL
11. What type of tests have SDO OCT been subjected to?
IEC 60044-8.
12. Does SDO OCT pose any requirements for the surrounding systems in a
digital substation (fiber-optic process bus, protection IEDs and measur-
ing IEDs) in order to perform satisfactory?
No special requirements other than being connected to the process bus using 2 x 100
Base FX Ethernet interface with LC type connectors and Multimode fiber optics.
We support both IEC61850-9-2LE and IEC61869-9 Sampled Values.
13. Lastly, I would like to ask permission to reuse the figure which displays
the SDO ICT head and insulator on page 4 of the Arteche SDO OCT
brochure, in my thesis.
You can do it with the only requirement of mentioning ARTECHE.
A.4 GE Grid Solutions

Hi Ingvill,
I am happy to respond to your questions related to our COSI-CT .
Please observe that we have a broad portfolio of non-conventional CTs & VTs and in
communication with Statnett we have expressed our interest in addition to the COSI-CT
to also pre-qualify our CMO/VTO/CTO range. See attached pdf’s for details.
Back to your questions , find my replay in red below.
1. What is the rated dynamic current of COSI-CT?

The COSI-CT has been tested up to 80 kA rms for 1 s and 210 kA pk (fully offset).
2. How good is the ability of COSI-CT to reproduce rapid changes in the

primary current?
Very good. dI/dt exceeds 70 A/µs.
3. Does the electronic module of COSI-CT impose any limitations on the

bandwidth of the digital output?
The digital output is filtered to avoid anti-aliasing in the digital output. The cut-off
frequency is programmable and is set based on the digital output sampling rate.
For 4800 Hz, we set the cut-off frequency to 1.5 kHz.
4. How good is the ability of COSI-CT to reproduce rapid changes in the
136
APPENDIX A. E-MAIL
primary frequency?
For protection it is very good. There are no limitation. For very precise revenue
metering applications the output is filtered, so there is a slower response to changes
in the power system frequency.
5. Is COSI-CT optimized for a specific application, such as metering or

electronics? Or is a single COSI-CT equally suited for both metering and
protection applications?
The number of fiber turns does affect the accuracy to some extent – especially for
very precise measurements such as revenue metering. Having said that, in general we
use a single fiber with a standard number of turns for both metering and protection.
One sensor can provide 0.15 % accuracy from <1 % to 200 % of the rated primary
current and IEC 5P accuracy up to 210 kApk. If the rated primary current exceeds
3000 A, we reduce the number of fiber turns to improve the metering accuracy as
we start to get into the non-linear portion of CT transfer function above 3000 A
with our standard number of fiber turns.
6. What is the maximum permissible length of the fiber optic cables con-
necting the sensor head of the COSI-CT to its electronic modules? Should
the electronic modules be installed in the in the switchyard or indoor in
a protection and control panel?
1000 m. For the COSI-CT, this limitation comes from the modulator cable (which
is a copper cable that controls the sensor’s phase modulator, i.e. the fiber optic
cable isn’t the limiting factor).
7. Which time synchronization methods does the electronic module of COSI-

CT support?
Today we support 1PPS and IRIG-B. IEEE 1588 PTP is planned for not-to-distant
future.
8. Can the electronic module of COSI-CT interface conventional current

and voltage transformers, or is a stand-alone merging unit required for
this purpose?
The COSI-MU (COSI-CT Electronics) has a 3-phase conventional secondary input
for digitizing VT signals. No CT inputs today.
9. What is the expected lifetime of COSI-CT?

The COSI-CT is designed for 30+ years. The limiting factor is the silicon sheds on
the insulator. The electronics are designed for 25 years, but in practise the actual
137
APPENDIX A. E-MAIL
life is lower (more like 15 years). We recommend planning to replace the electronics
once during the life of the COSI-CT (in general we recommend replacing the CT
electronics when the IEDs are replaced).
I hope our reply is according to your expectations. Feel free to contact me in case of
questions.
138
Appendix B
Formulas
B.1 Supply Grid

The short-circuit impedance of the supply grid was calculated by the use of Equation
(B.1),
UN 2
ZSC = (B.1)
SSC
where UN is the nominal voltage of the supply grid and SSC is the short-circuit capacity.
Based on Equation (B.2) and the assumption given by Equation (B.3), Equation (B.4)
can be derived,
ZSC 2 = R+ 2 + X+ 2 (B.2)
X+
= 10 (B.3)
R+
s
ZSC 2
X+ = (B.4)
1.01
where ZSC is the short-circuit impedance of the supply grid, R+ is the short-circuit pos-
itive sequence resistance and X+ is the short-circuit positive sequence reactance. The
assumptions given by Equation (B.5), Equation (B.6) and Equation (B.7) were used to
calculate the positive and zero sequence resistance and reactance of the supply grid at
different short-circuit capacities. The resulting values are given in Table B.1
139
APPENDIX B. FORMULAS
R+ = 0.1 · X+ (B.5)
X 0 = 3 · X+ (B.6)
R0 = 3 · R+ (B.7)
R+ is the short-circuit positive sequence resistance of the supply grid, X+ is the short-
circuit positive sequence reactance, R+ is the short-circuit zero sequence resistance and
X+ is the short-circuit zero sequence reactance.
Table B.1: Supply grid data at different short-circuit capacities.
Short-Circuit Capacity [MVA] R0 [Ω] X0 [Ω] R+ [Ω] X+ [Ω]

10 000 5.265 52.65 1.755 17.55
6 800 7.743 77.43 2.581 25.81
3 300 15.957 159.57 5.319 53.19
B.2 Current Transformers

The rated primary currents of the CT on the HV side of the power transformer and the
CT on the LV side of the power transformer were calculated by the use of Equation (B.8)
and Equation (B.9), respectively,
SN
IP N,HV = (B.8)
UHV N
SN
IP N,LV = (B.9)
ULV N
where SN is the rated power of the power transformer, UHV N is the rated voltage of the
HV winding and ULV N is the rated voltage of the LV winding.
140
APPENDIX B. FORMULAS
B.3 Load
The resistance of the load at different short-circuit capacities was calculated by the use
of Equation (B.10),
2
ULV
Rload = (B.10)
SSC
where ULV is the rated voltage of the LV winding of the power transformer and SSC is
the short-circuit capacity of the supply grid. The resulting resistances are given in Table
B.2.
Table B.2: Resistance of the load at different short-circuit capacities.
Short-Circuit Capacity [MVA] R [Ω]

10 000 55.37
6 800 54.86
3 300 52.94
141
Appendix C
MATLAB Scripts
C.1 Fundamental Frequency Phase Current
1 %% DFT, fundamental frequency

2
3 close all;
4 clear all;
5 clc;
6
7 R1 = 1;
8 C1 = 0;
9
10 B = csvread('simulation473.CSV', R1, C1);

11
12 t = B(763:4763,1);
13 CT1 = B(763:4763,2);
14 CT2 = B(763:4763,3);
15 CT3 = B(763:4763,4);
16 MU1 = B(763:4763,5);
17 MU2 = B(763:4763,6);
18 MU3 = B(763:4763,7);
19
20 t = t - 190.5;
21
22 for n = 1:80
23 cosDFT(n) = cos(2*pi*n/80);
24 sinDFT(n) = sin(2*pi*n/80);
25 end; clear n;
26
27 real_CT = sqrt(2) / 80 * filter(sinDFT, 1, CT1);

28 imag_CT = sqrt(2) / 80 * filter(cosDFT, 1, CT1);
29 real_MU = sqrt(2) / 80 * filter(sinDFT, 1, MU1);
30 imag_MU = sqrt(2) / 80 * filter(cosDFT, 1, MU1);
31
143
APPENDIX C. MATLAB SCRIPTS
32 abs_CT = (real_CT.^2 + imag_CT.^2).^(1/2) / 1000;

33 abs_MU = (real_MU.^2 + imag_MU.^2).^(1/2) / 1000;
34
35 figure;
36 plot(t, abs_MU, 'linewidth', 1.5);
37 hold on;
38 plot(t, abs_CT, 'linewidth', 1.5);
39 xlabel('Time [ms]');
40 ylabel('Primary Current [kA_{RMS}]');
41 xlim([0 1000]);
42 set(gca,'FontSize', 18);
43 legend('I_{rec\_samu\_fundamental}', 'I_{rec\_analog\_fundamental}');
144
C.2 Fundamental Frequency Differential Current
1 %% DFT, fundamental frequency differential current

2
3 close all;
4 clear all;
5 clc;
6
7 R1 = 1;
8 C1 = 0;
9

11
12 t = B(763:4763,1);
13 CT1 = B(763:4763,2);
14 CT2 = B(763:4763,3);
15 CT3 = B(763:4763,4);
16 MU1 = B(763:4763,5) / 1312*412;
17 MU2 = B(763:4763,6) / 1312*412;
18 MU3 = B(763:4763,7) / 1312*412;
19
20 t = t - 190.5;
21
22 for n = 1:80
23 cosDFT(n) = cos(2*pi*n/80);
24 sinDFT(n) = sin(2*pi*n/80);
25 end; clear n;
26

31
32 real = real_CT + real_MU;

33 imag = imag_CT + imag_MU;
34
35 abs = (real.^2 + imag.^2).^(1/2);

36
37 figure;
38 plot(t, abs, 'linewidth', 1.5);
40 ylabel('Primary Current [A_{RMS}]');
41 legend('I_{rec\_diff\_fundamental}');
42 xlim([0 1000]);
145
C.3 Second Harmonic Component
1 %% DFT, 2nd harmonic

2
3 close all;
4 clear all;
5 clc;
6
7 R1 = 1;
8 C1 = 0;
9

11
12 t = B(763:4763,1);
13 CT1 = B(763:4763,2);
14 CT2 = B(763:4763,3);
15 CT3 = B(763:4763,4);
16 MU1 = B(763:4763,5);
17 MU2 = B(763:4763,6);
18 MU3 = B(763:4763,7);
19
20 t = t - 190.75;
21
22 for n = 1:80
23 cosDFT(n) = cos(2*2*pi*n/80);
24 sinDFT(n) = sin(2*2*pi*n/80);
25 end; clear n;
26

31
32 abs_CT = (real_CT.^2 + imag_CT.^2).^(1/2);

33 abs_MU = (real_MU.^2 + imag_MU.^2).^(1/2);
34
35 figure;
36 plot(t, abs_MU, 'linewidth', 1.5);
37 hold on;
38 plot(t, abs_CT, 'linewidth', 1.5);
40 ylabel('Primary Current [A_{RMS}]');
41 xlim([0 1000]);
43 legend('I_{rec\_samu\_2nd\_harmonic}', 'I_{rec\_analog\_2nd\_harmonic}');
146

Fulltext

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Fulltext

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fulltext

Uploaded by

Copyright:

Available Formats

Application of Optical Current

Transformers in Digital Substations

Ingvill Urdal Rian

Master of Energy and Environmental Engineering

Norwegian University of Science and Technology

The following tasks are given:

• Describe a digital substation environment and the involved components.

This thesis finalizes my master’s degree in Energy and Environmental Engineering at

Trondheim, 22 Juni 2017

Ingvill Urdal Rian

I would like to thank the following persons:

It is concluded that optical current transformers seems to be suitable for application in

I denne masteroppgaven studeres optiske transformatorer for å vurdere hvorvidt slike

Laboratorietestene viser at stand-alone merging unit-en har en stasjonær oppførsel som

Optiske strømtransformatorer gir mange fordeler sammenlignet med konvensjonelle strøm-

Standardene som gjelder for ikke-konvensjonelle måletransformatorer og digitale stasjoner

3 Optical Current Transformers 19

5 Stand-Alone Merging Units 45

6 Considerations in Digital Substations 49

7 Laboratory Test Setup and Method 59

9 Discussion of Laboratory Tests 89

10 Conclusion of Laboratory Tests and Future Work 101

12 Conclusion and Future Work 113

C MATLAB Scripts 143

2.1 Layout of a conventional substation. . . . . . . . . . . . . . . . . . . . . . 6

3.1 Polarized light. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

4.1 A commercially available OCT. . . . . . . . . . . . . . . . . . . . . . . . . 42

5.1 Connection of CITs to secondary IEDs. . . . . . . . . . . . . . . . . . . . . 45

6.1 Total fault clearance time in a digital substation. . . . . . . . . . . . . . . 51

7.1 Transformer differential protection in hybrid mode configuration. . . . . . . 59

4.1 Overview of the specifications of some commercially available OCTs. . . . . 44

7.1 Settings of RET670 transformer differential protection. . . . . . . . . . . . 63

8.1 RMS values of phase A of Irec_samu and Irec_analog . . . . . . . . . . . . . . . 80

B.1 Supply grid data at different short-circuit capacities. . . . . . . . . . . . . 140

ALF Accuracy Limit Factor

CCT Conventional Current Transformer

CIT Conventional Instrument Transformer

CVT Conventional Voltage Transformer

DFT Discrete Fourier Transform

EMI Electromagnetic Interference

EMP Electromagnetic Pulse

GNSS Global Navigation Satellite System

GOOSE Generic Object Oriented Substation Event

HMI Human Machine Interface

HSR High-Availability Seamless Redundancy

IED Intelligent Electronic Device

MMS Manufacturing Message Specification

NCC Network Control Center

NCIT Non-Conventional Instrument Transformer

NCVT Non-Conventional Voltage Transformer

NTNU Norwegian University of Science and Technology

NTP Network Time Protocol

OCT Optical Current Transformer

PPS Pulse Per Second

PRP Parallel Redundancy Protocol

PTP Precision Time Protocol

SAMU Stand-Alone Merging Unit