APS Booklet 2015

Proceedings
August 18-19, 2015

Powerhouse Museum
Darling Harbour, Sydney, Australia
www.protectionsymposium.com.au
Australian
Protection Symposium
Proceedings of the
Australian Protection
Symposium 2015
Compiled by
Maryam Khallaghi
www.protectionsymposium.com.au
APS | 2015
Disclaimer
1 Content
The meeting organizer is not responsible for the accuracy, completeness, up-to-dateness’ or quality of
the information provided on behalf of third parties. The organizer is therefore not liable for any claims
regarding damage caused by the use of any information provided. The views and opinions expressed
therein are those of the individual contributors and do not necessarily represent the views and
opinions of the OMICRON Group or any of its employees.
2 Copyright
The material made available is intended for the customer’s personal use only. OMICRON (or its
licensor) reserves all rights to the documents and all parts thereof, including the translation, the reprint
and the reproduction. Therefore, third parties may not – for whatever purpose – reproduce or replicate,
or process or distribute with the help of electronic systems any part of the provided documents without
the author‘s written permission. This obligation continues to apply after the end of the meeting.
3 Request
If you identify any information in the meeting documentation which might be incorrect or misleading,
please advise suggested amendments to info@protectionsymposium.com.au
4 Terms and Conditions

Attendance at this meeting is subject to the General Sales Conditions of OMICRON including the
Special Terms for Training courses and Meetings.
These are available online on http://www.omicron.at/en/legal/terms/ or upon request.
II Australian Protection Symposium

Sydney, 18-19 August 2015
Welcome to the Australian Protection Symposium 2015
The Australian Protection Symposium is a forum for discussion and sharing of knowledge and first
hand experiences from Australia and around the world, whether its fundamentals, new approaches,
technology developments and standards, strategies and implementation experiences or even failure
case studies, there will be a discussion that will benefit you and your organization in the area of
protection, automation and control.
On behalf of the steering committee, I would like to thank all the presenters for bringing together their
knowledge and experiences to share with us and by doing so are investing their valuable expertise
back into the industry. Your willingness to volunteer your time, energy and support is greatly
appreciated.
Thanks you to the members of the 2015 steering committee. Your contribution has resulted in
achieving the quality of this event.
• Robert Bates | Ergon Energy

• Ian Young | Schneider Electrics
• John Ainsworth | Ausgrid
• Graeme Heggie | Electranet
• Michael Griffin | Australian Power Institute
• Sy Bui | Aurecon
• Warner Priest | Siemens
• Pascal Schuab | DT Partners
• Satendrea Bhola | TasNetworks
A special thanks to the APS sponsors, OMICRON electronics Australia Pty Ltd, PACWorld, CSE
Uniserve and Schneider Electric (Australia) Pty Ltd, whose support has helped us keep down the
costs for all participants.
A final word of thanks to all the attendees of this year’s APS, without you this event would not have
been possible. We hope that by attending this event you will have enhanced your technical knowledge
and gained a good perspective on how others in the industry are developing with the new industry
enhancements and dealing with problems protection engineers face every day. I hope you find this
symposium valuable and enjoyable.
Maryam Khallaghi
Technical Program Chair
OMICRON electronics Australia Pty Ltd
Australian Protection Symposium III

APS | 2015
APS | 2015
Table of Contents
Fundamental Principles and Requirements of Protection Systems

John Ainsworth
1
Calculation of Loadability Limits of Modern Numerical Relays

Maty Ghezelayagh 17
An Overview of D.C Traction Protection Systems

41
Zahra Bayat
The Digital Substation- Principles, Functions and Benefits

Alex Apostolov 57
Impact of IEC61850 ed 2 on Protection GOOSE

Ian Young 71
Case Study: Implementation of IEC61850 in Generator and

Transformer Protection System in NSW Power Stations 81
Sy Bui | Usman Mahmood
COMTRADE Analysis to Visualise Protection Operations 89

Greg Finlayson | Mitchell Eadie
Verification of Distance Protection Scheme of double Circuit transmission line

using Power disturbance records 103
Wing Chan | Si-Dieu Tran | Sameep Garti Chhetri
Cygnus Atratus and How It Causes Problems

Terry Foxcroft 113
IV Australian Protection Symposium

APS | 2015
Application of the Line Differential Protection Scheme for Radial Transmission Lines
Leonardo Torelli | LLia Voloh | Zhihan Xu 121
Modeling and Testing of High Impedance Differential Scheme Using Numerical Relays
Zarko Djekic 143
New Design of Distance Protection for Smart Grid Applications

Jorg Blumschein | Cezary Dzienis | Yilmaz Yelgin 151
Communication Aided Protection Schemes and their Implementation and Testing with
State of the Art Technologies 165
Fred Steinhauser
Implementation of Multi-Intertripping Protection Schemes in a complex EHV Industrial

Substation 173
Maty Ghazelayag | Chris Simmons
XML and UML – What they are and Why We Need to Know Them
Alex Apostolov
185
Old Lessons – Re learned

Graeme Heggie 199
330 KV Downed Conductor Protection and Successful Operation

Terrry Foxcroft 209
The Reality of Providing Energy in Australia – Bushfire Mitigation and the Benefits of
Power Quality Monitoring Outside Substations 223
Martin Van Der Linde
Enhancing Australian Protection System Testing Through Real Time Digital Simulations
231
Georgios Konstantinou and Colleagues
Australian Protection Symposium V

APS | 2015
VI Australian Protection Symposium

APS
APS| 2015
| 2015
Fundamental Principles and Requirements

of Protection Systems
John Ainsworth | Retired, formerly Ausgrid

johnainsworth@tpg.com.au
1 Introduction
The purpose of this paper is to provide as clear a statement as possible of the functionality which
Power System Protection has to fulfil in order to be considered ‘fit for purpose’, and the requirements
that this places on the concepts, design, components, and circuitry of protection systems. The
statement of functionality is made in such a way as to form the basis of concepts and designs
irrespective of the particular technology or combination of technologies used for its delivery.
The Electrical Protection of power systems is an essential and critical function. The level of
dependability and security required is the highest of any of the secondary circuit functions used in
substations. The requirements are stated in a way which is independent of the technology used, so as
to avoid restricting the development of new technology or applications. This approach is in the spirit of
IEC 61850 in providing base rules and a framework for developing solutions without restricting the
development of new solutions and improved functionality. The content includes key objectives of
protection, basic principles applying to any scheme, all of the components involved (not just the
relays), primary and back-up principles (Remote BU and Local BU) and their implications. It deals with
the requirements of secondary isolations of individual schemes required to permit safe work on a
scheme while the remainder of the substation is in service, and with the requirements for simulated
functional proving of protection schemes, again in a live substation. Segregation principles are also
covered.
There are two aspects to the functional requirements:
(a) The requirements which the protection has to fulfil in order to perform its functions of
protecting the power system and people,
(b) The requirements which the protection system places on whatever technology or mix of
technologies used in the delivery of these functions.
Protection is a specialized discipline and requires people qualified or appropriately trained
and competent in the discipline to carry out the various protection functions such as
protection planning, system analysis, design and specification of relay systems, circuit
design, relay settings, testing, installation, and commissioning
2 Safety
Safety of power systems is identified in legislation as paramount and is required to be given
priority over all other aspects of management and operation of the power system. This
requirement applies particularly to ‘safety in design’.
Protection systems, as well as being provided to disconnect failed cables, transformers, etc,
from the power system, are also essential for the safety of both staff and public. Protection has
a great influence on the consequences of electrical failure and on the level of injury sustained,
and is therefore a major consideration with respect to safety.
Safety of staff and the public is dependent on protection systems in three particular ways:
Australian Protection Symposium 1

APS | 2015
(a) Dependability. The protection system should always operate to clear a power system fault
(the safety outcome for people in the vicinity of a fault will be much worse if the protection
fails to operate).
(b) Co verage. The protection system should be capable of detecting all faults. (no fault
should be beyond the reach of the protection.)
(c) Speed of operation. Fast fault clearance limits the harm to people compared to slow
clearance. (There is a large difference to the safety outcome between clearance times of say
0.1 s and 1.0 s). Fast fault clearance can in many cases prevent splitting of equipment tanks,
and prevent or lower the incidence of fires.
3 Protection Objectives
Following are the basic objectives in the application of protection:
(a) To minimise the risk of injury to the public and personnel.
(b) To disconnect the faulted circuit or piece of equipment leaving the remainder of the
system in a healthy state.
(c) To interrupt the fault current with sufficient speed to ensure system stability is not
compromised.
(d) To maximise the electricity supply reliability by disconnecting only the failed equipment or
circuit.
(e) To prevent through-fault damage to upstream conductors and equipment.
(f) To preserve power system security by remaining secure against unintended tripping on
through-fault currents.
(g) To minimize the damage to network equipment at the fault location.
(h) To minimize damage to property.
4 Basic Principles
The objectives lead to the following basic principles which are generally applicable to all levels of the
system
4.1 Primary protection

A system of primary protection must be provided to detect all faults and disconnect all faulted
items of equipment from the power system. It should have no blind spots. That is, no fault
should be beyond the reach of the protection.
In principle, protection should be provided capable of detecting any fault on the power system,
particularly faults which can constitute a hazard. However, it is recognised that it is not always
technically practicable to detect certain very low level faults, such as may occur when an
overhead conductor falls to ground, or when vehicle tyres restrict the current. Nevertheless, the
industry is under obligation to apply the best practicable protection to these situations. In
substations it should always be practicable to provide protection for low level faults.
4.2 Back-up protection

A system of back-up protection should be provided which should also have no blind spots. Back-
up protection may utilise the principles of remote back-up or of local back-up as appropriate to
the level of the system.
2 Australian Protection Symposium

APS | 2015
Back-up should be provided for the failure of any element of the protection system to function,
that is :
(a) Relays and IEDs ( Intelligent Electronic Devices);
(b) Circuit breakers;
(c) D.C. supplies;
(d) Wiring and cabling including optical fibres; and
(e) Current transformers and voltage transformers.
4.3 Fault clearance times

Fault clearing times should be such that there is no damage (other than at the fault location) due
to the passage of through-fault current and that no additional faults are caused by it.
The type of protection chosen and the relay settings applied should limit temperatures on
overhead line conductors, cables, sheaths, earth conductors, and equipment so as to prevent
mechanical damage, annealing, thermal damage to insulation, and excessive sag.
4.4 System stability

Fault clearing times should be limited and coordinated to ensure network integrity and system stability
are maintained. The fault clearing time should meet the requirements of the relevant network
authorities.
4.5 Effects of faults

The protection system should limit the effects of faults and the damage at the fault location to
reduce the risk of harm to the public, personnel and property.
This is achieved by primary system design and equipment design as well as speed and
sensitivity of protection. Special precautions are taken in relation to major hazard facilities,
locations with potentially flammable or explosive atmospheres, swimming pools, communication
lines and cables, railways, mines, and high rise buildings.
4.6 Prevent / minimize loss of supply

The protection system should prevent or minimise loss of supply by confining the outage for a
particular fault to defined sections of the system consistent with planning reliability policies.
4.7 Compliance
Protection should comply with the legislation applying, being the Australian National Electricity
Rules, or the Western Australian Technical Rules.
5 Components of a protection scheme

A protection scheme involves all the components needed to detect when a fault occurs, to
decide whether to trip, and to isolate the faulted item from the system.
These components include the following:
(a) The current transformers and voltage transformers.
(b) The protection relays including Intelligent Electronic Devices (IEDs) used for
protection functions.
(c) Digital communications components such as network switches, routers, media
converters, merging units, multiplexers, teleprotection devices and other devices
used for protection functions.

APS | 2015
(d) Components such as ethernet switches, routers, merging units and multiplexers
used for protection functions.
(e) Batteries and d.c. supply arrangements.
(f) The circuit-breakers.
(g) The circuitry that functionally interconnects all these components.
(h) The physical wiring and cabling
6 Functional proving of Protection Relays and Systems

Consistent with the high degree of dependability required, protection relays and systems should
be proven to function correctly at commissioning and at defined intervals, after any alteration, and
after maintenance. These tests should involve the injection and measurement of operating
quantities and should totally prove the system by direct or simulated means.
It is necessary to prove that the scheme functions as designed and intended. That it does what
it is supposed to do. And that it does not do anything else.
The functional testing must prove all parts of the scheme. There are many component parts to
be proven:
(a) Relays, correct calibration, settings, logic.
(b) CTs and VTs—ratio, polarity, phase, class.
(c) Circuit-breakers—correct adjustment and operation.
(d) Wiring, cabling, and fibres—correct type, destination, and connections.
(e) D.C. supplies—correct source, range.
(f) Functioning as a complete scheme.
It is never possible to functionally prove a protection scheme by direct testing as this would
amount to putting it into service and throwing a primary system fault. Nevertheless the
functional proving must be carried out in a way which does not alter the scheme in any way
from what it would be when in service. A system of simulated tests on the scheme components
is therefore necessary, comprising all or some (depending on the nature of the work) of the
following sequence of steps:
(i) Prove CT ratios and polarities by primary current injection.
(ii) Prove CT classes and accuracies by a ‘saturation’ test.
(iii) Prove CT secondary resistances as intended.
(iv) Similar tests of ratio, class, etc of VTs.
(v) Verify relay settings and intended functions by injection of secondary current and volts
into relay.
(vi) Wiring checks to verify that the scheme is wired correctly to the schematic diagram
and that all connections are sound.
(vii) Prove any intertripping functions similarly.
(viii) Prove correct alarms for any functions that provide supervision, such as trip circuit
supervision or GOOSE subscription alarms.
(ix) Carry out balances of differential schemes.
(x) Prove correct phasing of CT and VT outputs.
(xi) Close all secondary links, VT fuses etc.

APS | 2015
(xii) Prove all tripping functions from relays to circuit-breakers by injection into relay
current and voltage inputs.
(xiii) First place the protection equipment into service and then energise the primary
equipment.
(xiv) Take in-service measurements of all a.c. currents and voltages to prove the integrity
of CT and VT inputs.
(xv) Carry out directional proving tests on directional overcurrent, directional earth fault,
and distance relay schemes.
The above simulated approach to functionally testing protection is quite different to setting to
work a control process scheme, e.g. a tapchanger control scheme, and requires a different
mind-set. In a control scheme one can set it to work and observe whether it is working correctly.
In the case of protection this approach cannot be taken. It must be proven to function correctly
beforehand because it is the protection, and a primary system fault may occur when the circuit
is first energised.
This simulated type functional testing has to be done on each protection scheme in a live
substation, while the remainder of the substation is energised and supplying load.
There has to be facilities incorporated in the design to allow this functional testing to be carried
out safely and with confidence that the scheme will work as intended.
There needs to be isolating and test links or an equivalent facility between each of these
component parts to allow this individual testing to be carried out in such a way as to not disturb
or temporarily alter any of the connections, so that the schemes are tested as they would be in-
service.
These links or equivalent are the test injection and measurement points for the individual
component parts of the protection scheme.
The design of protection schemes should incorporate a.c. and d.c. secondary isolating and
injection testing links or equivalent facilities to enable an individual scheme to be worked on for
maintenance, repair, or testing, while the remainder of the substation is in service. The facilities
should enable temporary connection of testing equipment to be readily made to the protection
scheme without the need to remove or disturb circuitry connections.
Finally, when the component testing is complete, the isolating links or equivalent are closed and
final trip checks carried out. The scheme is thus proven and authorised to go into service with
confidence that it will perform as intended when in service.
In addition, there will be a number of tripping and transfer tripping and CT circuits to and from
adjacent protection schemes which are in service. These have to be isolated for work on the
scheme under access, and also have their functions proven up to the open link.
In the case of hard wired schemes the use of sliding type links in DC and AC wiring allows the
opening of a circuit without disconnecting or disturbing wiring, so that there is confidence that
the circuit has not been altered during the testing process.
Relating these concepts to digital and fibre based protection schemes, it will be necessary to
provide equivalent means of carrying out the simulated functional proving of the scheme as it
would be when in service. That is, without having to unplug fibre cables or alter a relay’s
configuration by putting it into test mode. This also applies to all the other IEDs which might be
used in the scheme, such as Ethernet switches, multiplexers, etc.
At the present time the development of suitable equivalent facilities for this functional proving
has not reached a mature stage nor have standardized methodologies been developed.
7 Secondary Circuit Isolation

When an item of equipment such as a transformer, feeder or HV busbar is taken out of service for
work, it is necessary to isolate certain functions in its corresponding protection scheme, such as:

APS | 2015
(a) Tripping to circuit breakers.

(b) Transfer tripping or tripping functions to and from adjacent protection schemes.
(c) In some cases isolate CT inputs to the scheme or to adjacent schemes such as busbar
protection.
This secondary circuit isolation is in addition to the primary power system isolations necessary for
access permits. They often also serve as the point for injection and measurement for functional
proving, but not always.
The secondary circuit isolating points are necessary to prevent inadvertent operation of the circuit
being worked on (in particular unexpected tripping or closing of circuit-breakers -- a severe safety
hazard). They are also necessary to prevent unwanted operation or interference to adjacent protection
schemes of in-service equipment.
The isolations must be visible, self-evident, labelled as to function, unique, and obvious to all staff
involved.
There are strong safety reasons for these isolations to be clear, uniform, and well understood, and the
processes of their use documented.
Not all of these secondary isolations are performed by protection staff. Some have to be carried out by
operating staff to give access to other staff such as substation technicians, jointers, and linesmen.
The isolations must not alter the scheme from what it would be in service and from what has been
functionally proven. They must not be achieved by disconnecting wires or unplugging fibres.
Before restoring the HV equipment to service all secondary isolations have to be restored. This
process has to be rigorous so that HV equipment is not inadvertently re-energized with any of its
protection functions non-auto. Again this is a safety matter and the design of the isolation facilities
must enable this process without confusion.
The means of providing secondary circuit isolations in the digital environment must similarly not alter
the scheme from what it would be in service and from what has been functionally proven. This should
not be achieved by disconnecting wires or unplugging fibres; nor by altering the proven configuration
of a relay or IED by placing it in a test mode, nor by altering the structure of a GOOSE message to a
dummy message.
Safe and secure isolation of the ‘virtual’ trip signals and similar functions necessary for commissioning
and ongoing work on the live power system is fundamental to the design of such protection schemes
and should not be viewed as just a working practice to be resolved afterwards.
Digital systems provide the possibility of a centralised display of the status of all the isolations in the
substation, which would be of benefit, and of recording the history. However, it needs to be realised
that this introduces another level of functional proving; it needs to be proven that the status display of
each isolation point is in step with the actual isolation and that it is correctly identified. It also
introduces the possibility of automating isolation procedures, but again with the same issue of proving
the functions. Whilst these are benefits they are secondary to the main purpose of the isolations.
Further, failure of the isolation facility should be benign as far as the operational integrity of the system
and easily able to be replaced without any interference or risk to the system
8 Independence of A and B protection schemes

The ‘A’ and ‘B’ protection schemes should utilise relays from different manufacturers or relays using
different operating principles so as to avoid both schemes being disabled by a common failure mode.
Full independence of ‘A’ and ‘B’ protection schemes from each other is necessary so that no one
failure or event can render both inoperative.
In particular, the ‘A’ and ‘B’ protection schemes should utilise relays and IEDs from different
manufacturers and should utilise different operating principles so as to avoid both schemes being
disabled by a common failure mode. This concept applies also to the software
9 Segregation
Most substations are designed on an N–1 basis (except for single transformer distribution
substations); that is, the failure of any one item of primary equipment or outage for maintenance can
be accommodated while permitting the substation to continue to supply its rated load. This concept

APS | 2015
requires that segregation of the primary equipment items should be provided so that failure of one
does not damage another, e.g. a transformer fire should not damage adjacent transformers, or one
busbar damage another. Extending this concept, there should be no credible event possible which
affects more than one item of equipment.
This concept applies equally to the protection system of each of the primary items, particularly to the
secondary wiring and cabling. It would be inconsistent to provide segregation of the primary
equipment items but not carry this through to the corresponding protection equipments.
Physical segregation of the protection cabling and equipment belonging to one item of primary
equipment from that of other equipments is therefore necessary.
Further, for each item of primary equipment, physical segregation of the ‘A’ protection from the ‘B’
protection is necessary to prevent the functionality of both protection schemes being affected by a
single event or combination of events.
Optical fibres and fibre cables should be arranged to comply with these segregation principles.
Commercial grade fibre cables are not very robust and are susceptible to physical damage and fire
damage which heightens the need for segregation. Ruggedized fibre cables are available and it is
recommended that they be used in substations. They can be designed to be resistant to crushing, fire,
rodents, and termites.
An architecture where discrete process bus segments are set up to correspond to protection zones will
be necessary. Such an architecture is described in IEC 61850-9-2: 2004 as Alternative 4.
A station-wide process bus architecture would suit the concept of multi-casting, but would not fulfil
segregation requirements.
An Alternative 4 process bus segment architecture fulfils segregation requirements, but will limit the
station-wide sharing of information by multicasting.
The Ethernet network architecture used in a digital substation environment should be designed to
meet the end-user’s needs. If the Ethernet network is being used for protection purposes then it
should be designed for a very high overall availability. Guidelines for designing such systems can be
obtained from the following references:
a) IEC 61850-9-2: 2004
b) IEC 61850-90-4: 2013.
10 Electrical Interference
Protection systems should be designed to function correctly in the presence of various kinds of
electrical interference, including in particular any interference which may be generated by system
faults and switching operations. It should be recognized that the electric power system and
substations are a harsh environment for electronic equipment (surges and transients) and also a harsh
physical environment (shock and vibration).
The relays and protection equipment should be robust and immune to damage from these various
forms of interference. They should not be inhibited from operating correctly when required. They
should not be caused to operate incorrectly at any time, particularly during the presence of fault
current. They must not have settings, logic, software, or firmware corrupted by interference.
The design of protection equipment and circuits should comply with the type tests shown in the
following table, as specified in IEC 60255-5 and IEC 60255-26: They represent the types and
magnitude of electrical interference and surges that occur in substations.
For each of these tests the test voltage is selected from a range depending on the severity
class of the installation, as defined in IEC 60255. The values in brackets are typical test levels
used for protection equipment and schemes.
In addition, relays and equipment should be designed to not generate electrical interference,
conducted or radiated, of a level which could affect other relays and equipment in the
substation.
Intertrip systems using pilot cables between substations should be designed to withstand
without breakdown, maloperation, or failure to operate, the high levels of voltage, interference,
and noise, which are induced into pilot cables by power system faults and switching operations.
IEC 61850-3 defines the general requirements of communications equipment (including
protection relays, time servers and Ethernet switches).

APS | 2015
This requirement arises

Insulation 50 Hz from voltage induced into
Withstand (2kV RMS 1 minute) IEC 60255-5 : 2000 secondary wiring from
Test (1kV DC ) power system fault
currents.
This arises from lightning

currents in substation
Impulse
1.2 / 50 microsec earthing conductors
Withstand IEC 60255-5 : 2000
(5kV Peak) inducing corresponding
Test
impulse voltages into
secondary wiring.
This arises from HV

1 MHz bursts disconnector switching
Interference
(2.5 kV Longitudinal) IEC 60255-26 : 2013 generating high frequency
Test
(1kV Transverse) interference which couples
into secondary wiring.
This arises from relay,

circuit-breaker, and
contactor DC coils being
Fast Transient 1 MHz bursts
IEC 60255-26 : 2013 switched and generating
Test (3kV)
these surges directly in the
DC circuits in the
substation.
Gradual Shut-
For DC power supply Represents a battery going
down/Start-up IEC 60255-26 : 2013
to relays flat or being re-charged.
Test
Table 1: Type Test requirements
It is vitally necessary that all other devices used in conjunction with protection schemes and
process bus, such as merging units, routers, ethernet switches, multiplexers, etc, should
comply with the same impulse and disturbance immunity requirements and withstand levels.
Further, since these voltages exist in substations, all non-protection equipment associated with
the station bus concept should also comply with the same withstand levels.
The use of optical fibres, being non-conductive, is a major advantage in this regard, because
they are totally immune to electrical interference. However, there will still be need for copper
secondary wiring and cabling in substations to carry the DC supplies to all the electronic
devices necessary to interface with the fibres.
11 Sources of Energy for protection schemes

Protection systems and relays should use sources of energy for their operation which are independent
of the power system a.c. voltage. This is necessary because:
(a) The a.c. system voltage is depressed or disturbed during power system faults.
(b) The protection has to be operational prior to energization of the a.c. system which it
protects.
(c) Protection has to be operational throughout a supply interruption in readiness for restoration.

APS | 2015
The sources of energy will generally be in the form of battery supplies.

The D.C. supply to protection is a critical function and should always be available
12 DC supply Voltage Range

Battery and D.C. supply systems will inherently deliver a voltage which varies over a range, the
magnitude of which is dependent on several influencing factors.
Protection relays and systems including circuit-breaker trip coil circuits should be designed and proven
to function correctly over a voltage range which exceeds the D.C. supply range.
A substantial margin of safety at the lower end is necessary in view of the nature of the influencing
factors stated below. A margin of 15 to 20% is recommended.
The specification of the relay range is problematic for the following reasons.
IEC 60255–1: 2009 Measuring relays and protection equipment, currently specifies a range of
80 to 110% of relay rated voltage, which results in requirements on the range provided by the battery
system which are impracticable to meet, having regard to the following:
(a) The effects of depleted battery capacity near the end of an extended power system
interruption, the maximum current drawn by a protection operation plus other loading,
the battery internal resistance as affected by low temperature, age and condition, and
voltage drops in wiring, cabling, etc.
(b) The need to allow a margin of safety between the lowest voltage from the battery
system and the minimum operating voltage of the relays and other protection
components.
(c) The need to not exceed the relays’ upper limit when the battery is fully charged on
float voltage (typically about 115% of nominal).
(d) The need to not exceed the relays’ upper limit where boost or equalising charging is
employed. Some battery systems require boost or equalizing charges of at least 120%
nominal voltage from time to time to restore capacity and achieve expected life.
NOTE: The battery nominal voltage depends on the number of cells chosen and is not necessarily the
same as the relay rated voltage. The matter is further complicated if relays having slightly different
rated voltages are involved, e.g. 110 V and 125 V, in the one substation.
Circuit breaker standards, AS 62271.1:2012 and IEC 62271-1:2011 by contrast, specify a range of 70
to 110% of circuit breaker rated auxiliary voltage.
It is recommended that a range of 70 to 120% be specified for protection relays and systems, as well
as for circuit breakers. This is not a difficult requirement as modern digital relays incorporate a d.c. to
d.c. converter at their input which easily handle this range.
The factors which need to be considered in the design of the battery and d.c. supply system to deliver
a voltage within a given range include the following:
(i) Battery type and capacity (Ah).

(ii) The number of cells and therefore the nominal voltage.
(iii) The maximum current to be supplied during a protection trip or other switch
operation.
(iv) The standing load on the battery system.
(v) The required time period for which the d.c. supply system is required to maintain
supply capability without power system a.c. voltage.
(vi) The minimum d.c. voltage to be delivered to protection at the end of the battery
discharge time (while the battery charger is without a.c. voltage).
(vii) What load is connected to the battery other than protection.
(viii) The lower limit of the state of charge which is to be designed for. It is
recommended that it should not be allowed to fall below 30%.
(ix) Design life of battery.
(x) Minimum and maximum temperature.

APS | 2015
(xi) Voltage drops in wiring, cabling, and any series relay coils, when delivering the
highest current.
(xii) The voltage present when the battery is on floating charge.
(xiii) The highest voltage which may be delivered when the battery is on boost charge.
(xiv) Future loads that may be connected to the d.c. supply during the life of the
substation.
(xv) Battery internal resistance
13 Major System shutdown and Protection DC supplies

Relays and other electronic devices add permanent standing load to the protection batteries. Battery
capacity or the equivalent should be provided to safely ride through a wide and extended blackout or
power system shutdown. Supply to substation battery chargers will not be available from the power
system during this period.
The duration of a major system shutdown requiring a ‘black start’ depends on the network and the
location of the substation within the network, and is nowadays assessed to be in the range of 24 to 36
h. The availability of d.c. supply throughout and at the end of this period is critical. Protection, SCADA
and local control functions at all substations should be kept ready throughout this period to allow
progressive and controlled restoration of feeders and loads at the call of the system operator and at
any stage of the restoration process.
Similarly supply to substations can be interrupted for extended periods due to bushfires or flooding.
The designer should calculate the expected duration for a major system shutdown and the expected
time the substation equipment is to be supplied by the battery without external supply to the chargers.
The battery capacity should be sized to ensure correct operation of all protection devices for the
outage period and a minimum of two protection trip operations (e.g. bus zone trip) at the end of the
outage period.
Consideration should be given to provision of backup a.c. generation in the substation as a means of
maintaining the protection d.c. supply capability through an extended power system interruption. Other
options which could be considered in various different situations are:
(a) Use relays and devices having low standing d.c. drain.
(b) Use self-powered relays where appropriate (particularly in distribution substations).
(c) Use additional batteries to minimise the standing load on protection batteries.
(d) Use separate batteries for substation emergency lighting.
(e) Disable non-essential functions for the duration of such an event.
14 Safety in the design of protection systems

There are safety considerations for staff working on the protection equipment itself. Consideration of
the hazards must be central to the design of protection systems and equipment selection. The hazards
include the following:
(a) Working on electromagnetic current transformer secondary circuits which generate

dangerous voltages if open circuited.
(b) Induced transients in metallic secondary wiring and cabling, and pilot cables.
(c) Arc flash and burn hazards from secondary equipment, both d.c. circuits and CT
circuits.
(d) Working on circuit-breakers which have stored energy mechanisms capable of
causing serious bodily harm (fatalities have occurred).
(e) Exploding porcelain of current transformer housings, voltage transformer housings,
circuit-breaker housings.
(f) Insulation failure and flashover between primary HV conductors and secondary
circuits of CTs and VTs.
(g) The need to functionally test individual protection schemes while other parts of the
substation are in service.

APS | 2015
(h) The need to test protection schemes by injection of currents and voltages on to the
primary of HV equipment.
(i) The need to isolate protection secondary circuits from adjacent in-service sections of
the substation.
(j) The potential for the inadvertent or unexpected operation of primary or secondary
equipment because the functional isolations are either not implemented, not correct,
or not reliable.
(k) The potential for testing personnel to be exposed to dangerous voltages as a result of
an error in the primary or secondary isolations.
(l) Working in close proximity to live electrical equipment in the substation.
A key requirement is therefore that protection design and equipment specification are carried out in a
way which provides for safe working practices and does not give rise to safety traps and hazards.
15 Interlock Functions
Electrical interlocks between HV disconnectors, earthing switches, and circuit-breakers are an option
used in some cases.
These are for the purposes of safety to prevent accidental operation of the wrong switch, or operation
in the wrong sequence, either of which could result in a HV failure with explosive type consequences.
There are pros and cons about reliance on interlocks — they cannot cover all situations and
sometimes have to be defeated to carry out certain works.
Interlocks which are provided for prime safety reasons must be capable of being functionally proven in
the same way as protection functions.
Suitable commissioning and maintenance testing procedures must be developed by the designer to
ensure that the correct interlocking functionality is proven.
If they cannot be proven to function correctly they are worse than useless, in fact a hazard in the
sense that staff are relying on them.
There has been a trend towards providing these interlocks by means of GOOSE messages. This
approach also requires that facilities be developed to carry out the necessary functional proving.
16 Digital and Optical Fibre Based protection systems

This section deals with the application of digital and optical fibre based systems for protection and
control applications in substations in the context of IEC 61850.
It provides guidance in the application of digital and fibre based systems in substations, as to what
concept approaches are suitable.
It is important to clearly identify the functions which have to be performed by substation secondary
systems and to develop applications of the technology which are fit for purpose in performing these
functions. New ways of performing the functions will emerge and opportunities for additional,
previously not possible functions will also emerge. However the functional requirements of the
protection and the power system must remain the basis of developing and applying the digital
technology.
The principles set out in this document are the fundamental requirements which apply irrespective of
the technology used, whether conventional copper wire based systems, or fibre based systems. The
means of satisfying these requirements may change or vary with the technology, but the fundamental
requirements must remain the basis on which the changed concepts and applications are built.
Most substations have to remain in service at all times while providing for failure of individual items of
equipment and providing for disconnection of individual items from time to time for maintenance,
calibration, repair, refurbishment, or replacement. There is therefore need to functionally test individual
protection schemes while other parts of the substation are in service.
Two key points must be remembered with respect to protection systems:
(a) They are required for safety reasons with respect to both staff and the public.
(b) That total loss of the main and back-up protection systems in a substation or circuit
means that the substation or that circuit cannot be used.

APS | 2015
17 Safety in the design of protection systems

17.1 Power System to remain healthy
The protection system should leave the power system in a healthy state following a power system
fault or protection operation. That is—
(a) the remaining feeders or items of plant must not be left in an overloaded state;
(b) the remaining feeders and plant must be able to be satisfactorily operated;
(c) the configuration of the remaining network, feeders, or items of plant must be
capable of sustaining a further fault without sustaining damage or overload. This
may require operator intervention to re-configure the network or reduce load, or by
providing additional relaying for load shedding.
17.2 Fault protection and protection against overload

The protection system is intended to protect the network and its components against the effects of
faults. A distinction should be made between fault protection and overload protection. It is possible to
do both with the one relay, but requirements are often conflicting. It is essential that fault protection be
provided by relaying or other automatic means whereas overloading should be prevented by load
monitoring, system planning, and operating.
Nevertheless, there are certain specific cases where protection against overload is provided in the
form of load relief, or thermal relays. This, however, is additional to the normal fault protection and
must not interfere with the co-ordination of the normal fault protection. Protection of motors is a
particular case, generally requiring protection against overloading.
17.3 Rating of Primary equipment not to be limited by protection

Protection should not limit the rating of the primary equipment. The thermal rating of the relays,
current transformers, and the relay settings, should be arranged to be above the emergency rating of
the feeder, transformer, or busbar.
17.4 Circuit Breakers

A sufficient number of circuit-breakers should be provided on the primary system to:
(a) Enable automatic isolation of faults without complicated switching procedures.
(b) Enable Back-up protection principles to be applied.
(c) Enable protection circuits to be designed as far as possible in discrete
independent zones with a minimum of complicated secondary circuit
interconnections between zones.
(d) Facilitate straightforward operating procedures and avoid operating complications
and ‘traps’.
(e) Facilitate standard substation designs.
This should be balanced against the cost of installation and maintenance of circuit-breakers and the
need to provide protection of the circuit-breakers (and associated equipment) themselves.
17.5 Verification of Relay Settings

Verification of relay configurations and settings to ensure that they have taken effect in the device is
necessary at commissioning and at defined intervals thereafter. This should be carried out by
secondary current and voltage injection and measurement where appropriate, or by injection of
sampled values in the digital environment, or by primary injection in some cases. Reliance for correct
operation should not depend on settings established solely by downloading settings or by positioning

APS | 2015
dials and plugs without verification by measurement. Due to the inherent safety aspects of protection
systems, a relay settings change management process must be in place. The injected or simulated
current or voltage must have sinusoidal waveshape and the measurements must be by calibrated
instruments.
17.6 Balancing of differential protection schemes

It is necessary to prove correct balancing of differential schemes prior to commissioning and following
any alteration by injection of current at low voltage through the primary circuit. There are very many
connection errors which can result in wrong balance, some are as follows:
(a) One set of CTs having reversed connections.
(b) One or two individual CTs in a set having reversed connections.
(c) Wrong CT ratios or tap connections.
(d) Rolls or crosses in CT connections.
(e) Any of the above errors on interposing CTs.
(f) Wrong relay connections.
(g) Wrong pilot cable connections in the case of feeder differential.
(h) Any combination of the above.
(i) Any software equivalent of the above.
In addition, the following points should be verified by test as they can result in unwanted operation or
failure to operate of differential schemes;
(j) Different class CTs
(k) CT knee-points insufficient for external through current stability
(l) Ratio of CT knee-point voltage / excitation current differ sufficiently to cause false
differential currents for through current operation
(m) Additional burdens in series with CTs (bad connections or deliberate addition of
equipment) which upsets through-fault stability.
17.7 Directional testing and proving

Directional overcurrent, directional earth fault, and distance protections require the correct direction of
operation to be proven at commissioning and after any alteration. This involves proving first the
correct phasing of CTs and VTs by primary injection at low voltage, and then energizing the circuit
and using load current in a known direction to prove operation or restraint.
18 Conclusion
This paper has set out the requirements relating to power system protection schemes in a manner
which intentionally provides the base rules for the application of any particular technology or mix of
technologies to protection.

APS | 2015
About the Author
John Ainsworth has recently retired after a long career with Ausgrid
commencing in 1956. He has a Degree in Electrical Engineering with the
University of NSW as well as Certificates in Protection, HV Engineering
and Electrical Equipment Design, also from UNSW. His main life’s work
has been in Protection, in both design, field engineering and management,
and in policy areas, including all aspects of substation design. He serves
on several Standards Australia committees, EL043, EL024, EL023, EL7 as
well as having been a member of CIGRE Panels APB3 and APB5. He
currently gives seminars in Protection and Substation Design.

APS | 2015

APS | 2015

APS | 2015
APS | 2015
Calculation of Loadability Limits

of Modern Numerical Relays
Maty Ghezelayagh | Protection and Control Engineer,Tasmania,Australia

mghezelayagh@hotmail.com
Abstract
Conventional calculation of loadability limits of distance relays for different characteristics, such as Mho
and Lenz, is based on determining the minimum permissible load impedance which is equivalent to
maximum load current in order that it does not encroach to the last forward zone reach of distance relay
with a safety factor. In this paper it is shown that this philosophy, which is still used in some utilities for
calculation of relay load limit (RLL), is not applicable to modern numerical relays due to utilising of the
following elements for EHV transmission lines:
a) load encroachment element (LE)

b) power swing blocking (PSB).
The functionality of LE is to block the relay under normal loading conditions as long as the load
impedance is within the LE characteristic. Its characteristic consists of a circle cut by straight lines which
represents the load angle. The radius of the circle represents the magnitude of minimum load
impedance. Bearing in mind that LE responds only to balancing positive current, it is shown that LE
characteristic does not need to grade with any forward zones as long as the maximum load angle is less
than the line angle with sufficient margin.
The main objective of PSB is to block the operation of phase distance elements for stable swing. Its
characteristic consists of two blinders or circles, one inner and one outer. The inner blinder/circle
should encompass the outermost zone of phase distance protection that has been selected for blocking.
The outer blinder/circle should be set so that the closest minimum load impedance locus to be outside
the outer blinder/circle characteristic for all loading conditions.
In this paper, it is shown that the loadability limit of distance relays of numerical relays should be
calculated based on outer blinder/circle characteristics of PSB rather than the conventional method of
the last forward zone, otherwise the functionality of PSB will be ineffective and relay may trip on stable
swings. With this concept mathematical equations for calculation of RLL are developed for two cases,
one with blinder and one with circular characteristics for PSB.
In addition to impedance relays, methodology for calculation of RLL for current differential and
overcurrent protection of new numerical relays are given.
Finally, the results of analysis of system disturbances which have occurred under normal system
loading conditions due to wrong relay settings, design errors, relay failures and human errors during
testing are discussed. It will be shown how these disturbances could have been prevented if correct
design philosophy and test procedures had been implemented.

APS | 2015
1 Introduction
Generally unplanned tripping of primary equipment such as busbars, transmission lines and
transformers occurs by operation of protection devices to clear short circuit fault. However many
disturbances have also occurred under normal system configurations due to various reasons such as
wrong settings, design errors, human errors and failure of primary equipment due to age. Disturbances
due to short circuit faults are well known and have been documented in many publications. But,
publications for analysis of disturbances under normal system loading conditions are limited.
One of the contributing factors of the unwanted operation of protective relays under normal system
loading conditions, have been due to inappropriate settings of relays with respect to maximum loading
of primary equipment. In this respect distance and overcurrent relays have been the most vulnerable
protective devices.
For the last decade, advances in communication technology particularly in the area of fibre optic has
resulted to mpre application of line current differential protection than distance relays. In addition
modern numerical distance relays have more sophisticated algorithms [1-4] than old ones. These
factors have caused less unwanted operation of distance relays due to high loads at the cost of more
complex calculations for new functionalities of distance relays such as load encroachment and power
swing blocking elements.
Distance protection has been mainly used to protect sub transmission (66 kV) and EHV transmission
lines (above 110 kV) for protection against system short circuit faults. The main principal operation of
distance relays are based on measurement of line impedance during the fault [5]. If the measured
impedance lies in one of its zones, the protection trips the lines’ circuit breakers and isolates the fault.
In the past one of the most trouble making issues with old distance relays has been the encroachment
of the load to last forward zone and consequently operation of relay due to load [6]. To prevent this
scenario, loadability limit of the relay calculated based on the settings of the last forward zone reach.
These had caused limitations on maximum utilisation of primary equipment or provision of inadequate
protection.
With numerical relays the philosophy of calculation of loadability limit of distance relay is different from
old relays due to availability of power swing blocking (PSB) and load encroachment element (LEE).
The other type of relays which have frequently maloperated under normal loading condition is
overcurrent relays. The main reason has been due to lower pick-up settings other than the maximum
load. The low pick-up settings have been due to lack of periodical setting reviews with respect to load
growth or due to requirement of provision of adequate sensitivity.
Due to its principal design (unit protection), there are no significant numbers of unwanted operations of
current differential protection under normal system operation. However under some circumstances such
as low basic pick up setting or a faulty CT at one end, operation of current differential protection has
occurred. In order to avoid this, the basic pick-up of current differential relay should be set above the
maximum load + line charging current.
In this paper the methodology for calculation of loadability limit of modern numerical relays such as
distance, current differential and overcurrent protection is discussed. To achieve this, for distance
protection the required grading coordination between zone reach, minimum load impedance (maximum
load current) and PSB are clarified. For current differential protection, the required setting philosophy is
proposed and for overcurrent protection the required safety factor is calculated.
It is important to mention that loadability limits of different protection devices such as distance should be
calculated if the device has been set based on correct setting philosophy. Under this circumstance, the
fundamental question arises–what should be done if it is found the loadability limit of the relay is much
less than the rating of primary equipment or maximum required transmission line loading? Answer to
this question is not easy and engineering solutions should be found based on case by case.

APS | 2015
This paper also discusses system disturbances which have occurred due to unwanted operation of
protective devices under normal system loading conditions due to high load, wrong design, defective
relays and human errors are discussed and preventive measures are recommended.
2 Distance Relays Functionality and Characteristics

Most of today’s numerical relays have Mho and Quadrilateral characteristics for distance relays as
shown in Figures 1 and 2 [1]. The Mho characteristic is used for detection of multi-phase fault and
quadrilateral one for phase-to-ground fault. Generally for each characteristic, four zones are available to
be set. Depending on applications three zones are set forward and one zone as reverse. The
application of reverse zone can be for blocking signalling scheme or for back-up busbar protection
scheme.
Figure 1: Mho Characteristic of Distance Relays for Phase Element
Figure 2: Quadrilateral Characteristic of Distance Relays for Ground Element
The setting philosophy of zones reach of distance relays may depend on whether the distance element
is as back-up protection to current differential protection or whether it is as main protection. When it is
used as back up, Zone 1 will be enabled only when the current differential protection is out-of-service
(eg with communication channel fail). The general setting philosophy is as follows:
2.1 Zone Reach Setting
2.1.1 Phase-to-ground Quadrilateral characteristic
Zone 1 reach shall be set (as set in relay or in simulation program) in the range of 65%–85% of the
positive seq impedance of the line in order for fault within 65%–85% of line length. The relay operates in

APS | 2015
Zone 1 for each of the following three system conditions under presence of zero sequence mutual
impedance.
• Parallel line in-service (Must Criteria);
• Parallel line out-of-service (Must Criteria); and
• Parallel line out-of-service and earthed at its extremities (Desirable criteria).
The above criteria mainly should be met for solid phase-to-ground fault.
With this criteria conventional zero sequence compensation factor (K0=(Z0-Z1)/3Z1) can be used. The
alternative criteria is to set Zone 1 reach for phase-to-ground fault to 80%–85% of line impedance but
using a calculated zero sequence compensation factor instead of conventional value in order to ensure
Zone 1 does not under-reach/over-reach for single phase-to-ground fault. The calculated zero sequence
compensation factor can be obtained by a suitable protection computer program. A trial and error
method can be used to find the correct zero sequence compensation factor.
Zone 2 shall be set to reach 120% of the transmission line but shall not reach beyond 50% of the
shortest transmission line connected to the remote end busbar with all but the largest single infeed
removed from service. If it is found that these criteria cannot be met, the condition of the outage of the
infeed can be removed. It should also be checked that Zone 2 will see the fault at remote bus for single
phase-to-ground fault under presence of zero sequence mutual impedance.
Zone 3 (forward) shall reach 150% of the protected transmission line and less than inner characteristic
of PSB settings. If there is no reliable CBF scheme at the remote substation, then the reach of Zone 3
can be extended to cover remote CB failure to some extent, subject to studies with transient stability
program which indicates stable swing does not encroach to Zone 3 area or their durations are shorter
than Zone 3 time setting.
Zone 4 reach can be set reverse as a back-up protection if the duplicate busbar protection scheme
does not exist. The reach of Zone 4 should be 20% of the shortest line or smallest TF impedance in the
reverse protected Zone. If Zone 4 is used for blocking the signalling scheme, its reach should cover the
whole length of the line which Zone 4 sees fault as forward.
The Zones timings should be set based on general well known philosophy as practised in industry. For
example, Zone 1 timing is set instantaneous, Zone 2 to 0.4 sec and Zone 3 to 0.6 or 1 sec.
2.1.2 Phase Resistive Reach Setting of the Zones
Resistive reach is set as maximum as possible while avoiding PSB and abiding manufacturer R/X
constraint. It should be preferable in the following range:
Zone 1=5-10 ohm/prim, Zone 2=10-20 ohm/prim, Last Zone forward=20-30 ohm/prim
2.1.3 Multi-Phase element (Mho characteristic)
The general zone reach philosophy as for phase-to-ground is also applicable for multi-phase fault.
However since zero sequence mutual impedance does not affect a multi-phase fault, Zone 1 can be set
to conventional value of 80%–85% of positive sequence impedance of the line.
The fault arching resistance for ph-ph fault shall be calculated according to R Van Warrington’s formula
as:
×
= .
ohm/prim
Where:
I= Minimum phase-phase fault current (A)
L= Maximum phase conductor separation (m)

APS | 2015
Checks shall be made that over-reaching particularly for Zone 2 does not occur when ph-ph fault
arching resistance is taken into account. If over-reaching of Zone 2 with Zone 1 of the adjacent line
occurs, the calculated fault arching resistance shall be reduced.
2.1.4 Zone characteristic angle
All Zone characteristic angles shall be set to tan-1 (X1/R1). However, it can be set differently if the
required reach dictates.
3 Load Encroachment Element Functionality and Characteristics

Generally, load is represented either by current (I) or apparent power (MVA) or impedance ( ).
The most important values for these parameters for consideration in protection settings are the
maximum of I, MVA, and minimum load impedance ( ) and power factor (pf).
The relationships between these parameters are according to the following equations:
!" #
= (Eq 1) or
$ %&' !"
!" #
= & ()
!"
* = (Eq 2) or
√,×$ %&' !"
= - #/√3 × *_ 1
2 = cos 6 78
Where
- = 0.8 − 0.9# × - =>?

78 =Load power factor
Figure 3 refers to the characteristic of minimum load impedance with radius of Z= and angle
“β” on impedance plane.
Figure 3: Load Impedance Characteristic
For calculation of of a specified line, it is required to know maximum permissible of apparent

power or current. The actual permissible maximum power flow on any transmission line can be
determined based on one of the following criteria:

APS | 2015
1) Stability limit (both transient and dynamic stability limit)

2) Thermal emergency rating of the line
3) Maximum continuous rating of the line
4) Generation level limitation (eg for those radial lines connected directly to generation plants)
5) N-1 load flow contingencies
6) Actual measured maximum line current
Generally to determine the maximum permissible power flow on any transmission line based on stability
or N-1 require significant system studies. To avoid complexity, most utilities use maximum continuous
rating of the line conductor which can be obtained from conductor size and type.
It should be mentioned that appropriate calculations of minimum load impedance is important since it
affects directly the setting of the following elements of distance relay:
7) the reach of forward zones of distance relays;

8) the setting of power swing blocking, and
9) the setting of load encroachment element (LE).
Generally the LE element in relay is set equal to . It is not recommended to be

calculated based on the higher value than maximum rating of the line since it is preferred the distance
protection trips the line rather than be blocked by LE element when current exceeds the line rating.
It should be mentioned that the load encroachment element only responds to the three phase balanced
conditions (eg only to load and balanced three phase fault). Based on this, the important value of load
encroachment element is its angle with respect to line angle (3 phase fault angle). As long as there is a
sufficient margin between these two angles, it does not matter if load impedance encroaches to one of
the characteristic of the distance zone. Figure 4.6 [2] elaborates more on this concept. It shows how the
initial phase angle (load) is changed when a balanced fault occurs. Based on this, it is not needed to
limit the load due to the distance zone settings since load encroachment element will block the distance
elements for heavy load conditions.
If the relay does not have load encroachment elements (eg static relays) then the reach of the zone
should be set in order that the load does not encroach on any zone. This can be achieved either by
reducing the reach or changing the relay characteristic (eg using Lenticular instead of Mho
characteristics).
4 Power Swing Functionality and Characteristics

Power system swings occur after system disturbances such as short circuits fault, large load and
generation changes and switchings. With modern numerical relays [1-4], the power swing element can
be programmed to do one of the following functionalities:

APS | 2015
• Power swing blocking: blocks distance elements during power swings.

• Power swing tripping: trips the circuit breaker(s) during unstable swings.
The purpose of power swing blocking (or out-of-step blocking) elements is to prevent tripping of
distance relays for stable swing after a system short circuit fault. Generally disturbances with low
voltages and large currents causes small impedances which can be seen by zone distance protection
and as a result tripping the circuit.
The purpose of power swing tripping (PST) is to trip the line or generator on detection of unstable swing
in order to avoid large system black out and damage to generators.
The expected performance of power swing detection element when used for blocking is that the element
does not block the distance protection for short circuit fault. However, it should block the distance
protection for any other swings which may arise due to other events. Based on this the characteristics
and setting of PSB should be in such a way in order to be able to distinguish between different type of
swings which may arise due to different disturbances.
The characteristics used for both functionalities (PSB or PST) are generally the same and can be
blinder as shown in Figure 5 or circle as shown in Figure 9. Each one consists of inner and outer
characteristics.
Different relays have different algorithm for PSB. The latest PSB technique [1] employ two algorithms
that work in parallel.
1) Fast power swings (above 0.5 HZ): This algorithm does not have characteristic. The current
samples are compared to sample 2 cycles before. This is a continuous process. When the
difference between current sample and previous sample (that is taken two cycles back) exceed
the delta threshold the PSB detector picks up. If the PSB detector remains ON for more than
three cycles, it is classified as PSB condition. For this algorithm, there are no settings required.
The user needs only to enable the feature, and decides which zones are required to be blocked.
2) Slow power swing (below 0.5 HZ): This algorithm based as conventional method and having two
zones characteristics (Fig 5)
Power system swings may occur for both symmetrical and asymmetrical types of fault such as single
phase to ground or two phase faults. Consequently, any power swing detection should have a dedicated
measuring system for each phase. Measuring the rate of change of the impedance vectors is important
to distinguish swings due to different events. For short circuit fault the rate of change is high. The
detection algorithm starts when the impedance vector enters the power swing measuring range and
some other criteria have been met.
For different types of disturbances, swing trajectory will be different. For short circuit faults, the
impedance vector moves suddenly from load impedance to inner characteristic of PSB. Depending on
other type of events, the trajectory of swings may follow one of the following:
• The apparent impedance vector initially enters the outer and later enters inner characteristic
(dashed trajectory in Fig 5).
• The swing vector may enter to outer area and leave it again without entering inner characteristic.
• The trajectory may enter the characteristic and leaving it in opposite side (solid trajectory in Fig
5).
The first two swings represent stable system conditions but the last swing represents loss of
synchronism. The power swing element should block distance relays if the first two swings enters one of
the distance zones but it may be preferred to operate the relay for last swing.
Generally the swing trajectories for each type of events can be obtained by transient stability programs
with time stamp on trajectory for each quarter cycle intervals and be plotted on PSB characteristics.
From this information, proper settings for PSB can be achieved which can respond to the required
action such as tripping or blocking of distance relays.

APS | 2015
4.1 Setting Philosophy of Power Swing Blocking Element
The PSB is programmed to block phase distance protection during a swing when the measured
positive-sequence impedance enters the operating characteristics of the phase distance elements. In
some relays blocking occurs only for instantaneous Zone 1 and Zone 2 with communications-assisted
tripping scheme. The reason is that the operation time of these two zones is shorter than the time period
during which the impedance of a power swing resides in these protection zones.
The general setting philosophy of reactance and resistive reach of inner and outer characteristics is
based on the following criteria:
1) The inner zone resistance reach should be larger than the outmost characteristic of
distance zones with a safety margin. Its reactance reach should be higher than the Zone
which has been selected for blocking.
2) The outer right hand zone resistance reach should be less than minimum load impedance
at the closets point as required for starting the timer of the blocking logic.
The timer of out-of-step blocking logic starts when swing trajectory enters the outer characteristic and if
it expires before trajectory enters the inner characteristics, the relay detects a power swing blocking
condition. Bearing in mind that each trajectory starts from load impedance. In order for the timer to start
it is essential that the minimum load impedance be outside the outer characteristics. Otherwise the timer
will not start and the PSB element will be ineffective. The setting of timer depends on the positive
sequence source, line, remote impedances, and resistive reaches of PSB and swing frequency [2].
The PSB logic can be set self-resets or be latched for different type of swings. It can be set to blocking
mode for both stable and unstable swings. If a fault occurs during swing, the relay automatically
removes the PSB blocking signal.
Details of the principal operation of PSB/PST, and setting philosophy, can be found in most of the
manufacturer’s relays manuals [1-4] and are not discussed here. However, here attention is only made
to determine the required characteristic discrimination between PSB and load impedance for calculation
of distance relay’s loadability limit.
Figure 5: PSB Characteristic of Distance Relays (Micom)
5 Summary of Required Grading Coordination between Distance

Zones, LE and PSB
Based on the above concepts for distance zone reach, LE and PSB elements, these elements should
be set according to the following principles:
1) The distance zone forward reach should be set to provide line protection adequately and provide
CBF for next adjacent line if required.

APS | 2015
2) Minimum Load impedance to be calculated based on appropriate criteria as mentioned in Section

3.
3) Load encroachment element characteristics does not need to grade with any zone distance
characteristic as long as its angle is smaller than line angle with sufficient margin. Its magnitude
can be set equal to minimum load impedance as described in Section 3 and its angle should be
set based on worst load power factor+5 degrees margin.
4) PSB outer impedance characteristic to be set smaller than minimum load impedance with a
safety factor.
5) The inner zone impedance characteristic of PSB should be set larger than the outmost
characteristic of distance zones with a safety margin.
After implementation of the above principles, if it is found the loadability limit of the relay limits
unnecessary, the required transmission line loading, following the solutions should be envisaged:
• relax the criteria for calculation of minimum load impedance.

• reduce the last forward zone or select other relay characteristics such as Lenz instead of Mho,
and
• change the design for the CBF scheme.
6 Calculation of Loadability Limits of Distance Relays Based on Last

Forward Zone Reach
Loadability limits of distance relays have been calculated based on last forward zones setting with Mho
or Lenz characteristics [7-8]. The objectives of these calculations are to find what maximum load
current can flow on line without causing the relay operation. The procedure which has been outlined for
Mho and Lenz characteristics in [7] consists of several steps. The calculation process for Mho
characteristic is as follows:
Figure 6: Calculation of Loadability Limit for Mho Characteristic
1) Draw the Zone 3 impedance vector in the R-X diagram as shown in Figure 6.
2) Draw the load impedance vector at a specified load angle θ.
3) Draw a right triangle forming the 90° relay characteristic between the load impedance vector and
the difference vector that is made up of Z3 – Zload.
4) Calculate the interior angle β that is made between the load and line impedance vectors. This is
done by subtracting the line impedance angle minus the power factor angle.
∠β= ∠α - ∠θ
5) Calculate the load impedance that the relay will experience at the specified power factor using
right triangle properties:
Zload=Z3. Cos β

APS | 2015
6) Calculate the maximum loadability of the relay in MVA by:
S= P + jQ=KV2/Zload
Where kV is line to line voltage.
By applying the same procedure, the maximum loadability of the relay for different power factor angles
can be calculated and plotted on P&Q plane as shown in Figure 7. This will give “Trip” and “No Trip”
zones.
Figure 7: Calculation of Loadability Limit at Different Power Factor Angle
Similar procedure as above has been outlined in [7] for Lenz characteristics.
The above methodology for calculation of loadability limits of distance relays is only valid if distance
relays have not load encroachment element and power swing blocking. Consequently it is not applicable
for modern numerical relays which have these elements. Generally for applications of distance relays on
EHV transmission lines, utilising power swing blocking element is essential in order to avoid tripping of
the line due to system swings as mentioned in Section 4.
In Section 7 the methodology for calculation of RLL for modern numerical relays is discussed.
7 Calculation of Loadability Limit of Distance Relay Based on PSB

Characteristic
Based on the concepts of distance relay zones, PSB and LE as discussed in Sections 2-4, the
loadability limits of distance relays are calculated based on PSB characteristics. For this purpose
mathematical equations are developed for two cases. In one case for PSB with blinder, and in another
case with circle characteristics. In addition, calculation of the required safety factor which determines
the level which PSB characteristic should be below the minimum load impedance is given. For this
purpose, errors in current and voltage transformers and relays are considered and typical values are
provided.
7.1. General Equations for Calculation of Relay Load Limits (RLL)
As discussed previously, relay load limit are generally defined as the maximum apparent power
( ) or current (* # which can flow on primary equipment without any effect on operation or
functionality of the protective device. The relationship between * and minimum load
impedance is according to equations given in Section 3. From these equations, in order to
calculate RLL, it is required to calculate minimum load impedance .
7.2. Calculation of Minimum Load Impedance ( ABCDEFGH )
Based on the philosophy mentioned in Section 4, the outer characteristic of PSB (blinder/circle) should
be smaller than the minimum load impedance characteristic with a safety margin as shown in Figure 8.
This is because the swing impedance loci must start outside the PSB zones for the logic to detect a
power swing.
APS | 2015
In order to calculate minimum load impedance for calculation of RLL, it is assumed all zone distance
reach, PSB and LE have been set correctly based on grading coordination and philosophy as
mentioned in Sections 2, 3, 4 or any other appropriate philosophy. If with existing settings, it is found
that appropriate philosophy has not been implemented, the setting should be reviewed. The end result
is to have an acceptable and known setting for PSB in order to be able to calculate loadability limit.
7.3. Calculation of Minimum Load Impedance with Blinder Characteristic for

PSB
If it is assumed 6 (Fig 8) is the magnitude of resistive reach of outer blinder of PSB, then:
J≈ 6
$ > L !"
= MN
(SF=Safety Factor)
O
=
PQR S#
The above Equations give:

VW
= TU × PQR S (Eq 3)
Where,
α=cos-1pf +( 90-∟ZLANG )
ZLANG=line angle,
pf=worst load power factor (45 deg in figure)
After obtaining from EQ.3, RLL can be calculated in terms of apparent power or current as
given in Section 3 as:
!" #
= =$ %&' !"
OR
!"
=* =
√,×$ %&' !"
SF
X
ZL
Z
α R6
R
6
Figure 8: Calculation of Loadability Limit with Quadrilateral Characteristic of PSB

APS | 2015
7.4. Calculation of Minimum Load Impedance with Circle Characteristic for PSB
Generally most of today’s numerical relays have blinder characteristics for PSB. However, there are
relays with circle characteristics as shown in Figure 9. If it is assumed that the impedance of the PSB
(or the radius of the outer circle) is ZB, then:
= TU × X (Eq 4)
After obtaining from Eq.4, RLL can be calculated according to EQs 1 and 2 in Section 3 in
terms of apparent power or current as:
!" #
= =$ %&' !"
OR
!"
=* =
√,×$ %&' !"
Figure 9: Calculation of Loadability Limit with Circular Characteristic of PSB
7.5. Calculation of Loadability of Distance Relay without PSB
As discussed previously, it is necessary to enable PSB when distance protection is applied on EHV
transmission line (110 kV above) in an interconnected network. For EHV networks, voltage and angle
swings are noticeable after a short circuit fault. However, if the EHV interconnected network is large and
generators have noticeable inertia, then system swings on sub transmission and distribution networks
(below 66 kV) may not be significant. In this case it is possible not to enable PSB subjects to detail
transient and dynamic stability studies. Under this scenario, the philosophy as described in Section 6
can be applied. For example, instead of PSB characteristics, the last forward zone reach of phase
distance (Mho or Quad) should be used for calculation of loadability limit.
For old relays with starters, the characteristic of starter instead of last forward zone should be
considered.
If the relay has LE elements and has been set based on philosophy mentioned in Section 3, then the
minimum load impedance is obtained as in equation given below provided PSB is disabled:
= Y
Where: ZLE is the impedance setting of LE element
Based on this equation, RLL can be calculated according to EQs 1 and 2 in Section 3 in terms of
apparent power and current as:

APS | 2015
!"#
= = OR
$ %&' !"
-
=* =
√3 ×
7.6. Calculation of Safety Factor between PSB and Minimum Load Impedance
Characteristic
As discussed, in order, the load does not encroach the power swing characteristic, the minimum load
impedance should be higher than the PSB characteristic with a safety factor (SF). The safety factor
should be based on recommended value as in relay manufacturer manual. The safety factor generally
considers errors in CT, VT, relays and assumed data. The safety factor should apply to the closest
points of highest reach element characteristic (eg PSB) to load encroachment characteristic and not to
be only along the resistive reach. Generally a safety factor of 10-15 percent between closest points of
the two characteristics is acceptable if accurate data is not available.
The general equation for calculation of the safety factor is as given below:
TU = Z[\]]>] + [\]]>] + _` a\]]>]
a) Relay Error
Generally the relay error depends on types of relay and element and can be found on most of relays
Application Manuals [1-4]. Typical values for PSB element for some of the well-known relays are
referred to in Table 1.
Relay
Error % (5A relay) Error% (1A relay)
Type
Impedance : ±5% Same as 5 A rating

Type 1
Angle : ± 2°
Type 2 ±10% up to SIR=30 Same as 5A rating
1) Steady state accuracy 1) Steady state accuracy
a)±5% of the setting plus ± 0.01 A for SIR<30 a)±5% of the setting plus ± 0.05
b) )±10% of the setting plus ± 0.05 A for 30 A for SIR<30
<SIR<60 b) )±10% of the setting plus ±
Type 3
0.05 A for 30 <SIR<60
2) Transient overreach
<5% of the setting plus steady state accuracy 2) Transient overreach
<5% of the setting plus steady
state accuracy
Table 1: PSB Element errors of Modern Numerical Relays
b) CT and VT Error
Practically all types of CT and VT are in operation in various electricity companies. Some of them are
as old as 50 years.
The CT and VT errors and accuracy depends on type of equipment and its age.

APS | 2015
The specification and accuracy of modern EHV CTs and VTs have been shown in Table 2. Table 3
also shows the accuracy for older types of equipment.
Model CT VT
Type 1 - 1500/1000/750/400/1 - Rated Secondary voltage:110/√3 V

- IePXEKRct - Rated Ratio=2000
- Ie=0.1 - Accuracy class=3P
- EK-100(Rct+2)
- -
Error % Turns ratio error : ±0.25% Voltage ratio error=±3%
- Current error=[Ie/(Ip/N)] x100 - Phase displacement=±120
- Phase error=Im(Ie)/(Ip/N) minutes
≈
Current and Voltage ≈ 10%* 3%*
Error % at rated
secondary current and
Voltage
Table 2: Specification and error of modern protection EHV CTs and VTs
Model CT VT
-
Type 1 - Ratio=750/1 Rated Secondary voltage:110/√3 V
- 0.1PL500R1.5 - Rated Ratio=2000
- Accuracy class=3P
Error % turns ratio error : ±0.25% -Voltage ratio error=±3%

Current error=[Ie/(Ip/N)] x100 - Phase displacement=±120
Phase error=Im(Ie)/(Ip/N) minutes
Current and Voltage ≈10% * ≈ 3%*

error % at rated
secondary current and
Voltage
Table 3: Specification and Errors of old protection EHV CTs and VTs
*For distance relay, the composite error which considers phase angle error should be considered
Where:
Ie= Excitation current

N= ideal turns ratio
Ip= Primary current
8 Calculation of Loadability Limits of Current Differential Protection

With the development of communication technology particularly in the area of fibre optic, most of today’s
power systems have communication links between different substations for various purposes. One of
them is application for line current differential protection which has many advantages over distance
protection in terms of speed, selectivity and reliability. Based on this distance protection has lost
significantly its ground to current differential protection and these days are mostly used as back up
protection element in most of the numerical relays [1-4].
APS | 2015
The principal of current differential protection is well known and can be found in most relay manuals.
Consequently it is not discussed here.
Due to the unit type protection and percentage characteristics, current differential relays are stable to
high load. Any spurious current caused by CT accuracy and saturation will increase both differential and
restraint current. Based on this, the loadability limits of current differential protection under normal
loading conditions or N-1 contingency can be considered infinite if correct setting philosophy is to be
applied:
8.1. Setting of the basic pick up
a) Normal System Condition:
The basic pick up to be set above 1.1 times of maximum line charging current
b) N-1 contingency
The basic pick up setting to be set equal to 110 percent (max load + line charging current) in order the
element does not operate when the CT at one end is open or short circuited or to take out of service for
maintenance. With this setting it should be checked sensitivity criteria is met. The sensitivity criteria can
be based on detection of 80 percent of minimum internal fault under minimum system generation when
one end of the line is open. If sensitivity criteria cannot be met, the basic pick up should be set based on
normal system conditions (eg no contingency).
c) Transmission line with tapped load
The basic pick up setting should be set above 110 percent (tapped load+line charging current) provided
sensitivity criteria is met. If criteria is not met application of alternative protection scheme such as 3-
ended scheme should be investigated or alternatively a relay with sophisticated algorithm to be
implemented.
8.2. Slope Characteristic

The slope characteristic of differential protection should be sufficiently high in order that the relay
remains stable for maximum loads and through fault under partial CT saturation and maximum CT
errors.
If distance or overcurrent elements are used as back up protection for current differential element of a
numerical relay, then the loadability limits of the protection is determined by distance or overcurrent
elements.
9 Calculation of Loadability Limits of Overcurrent Protection
Overcurrent relays are generally used for protection of distribution feeders (below 33 kV) and incoming
feeders against multi-phase faults. In the past frequently overcurrent protection has operated due to
high loads due to several reasons as described in Section 11.1.1. One of the reasons for maloperation
has been lack of correct calculation of loadability limit of overcurrent relays. For example it has been
assumed that RLL is exactly the same as the pick-up setting of the relay. Bearing in mind that there are
CT and Relay error (Safety Factor), the actual RLL should be calculated based on following equation:
Zb\cc d
=
1 + TU
For example if the relay pick up setting is 200A, the RLL will be 181 A (with 10% for SF). The value of
SF depends on the type of relay and CT. Table 1 shows errors for overcurrent relays of modern
numerical relays. Table 2 shows typical errors for HV protection CTs.

APS | 2015
Relay Type Accuracy of pick up (5A) Accuracy of pick up (1A)
Type 1 a) Steady state : ±5% setting Same as 5A relay

b) Transient overreach: an
additional ±5%
Type 2 at < 2 x CT: ± 0.5% of 2 x CT Same as 5A

at 2 x CT: ± 1% of 20 x CT
0.01 A±3% of setting
Type 3 0.05 A±3% of setting (steady state)
(steady state)
Type 4 2% of setting or 50mA 2% of setting or 10mA
Table 4: Typical errors for modern numerical overcurrent relays
Relay Type 5A 1A
Type 1 7.5AVA 10P20 2.5VA 10P20
Table 5: Required CT class for a typical modern numerical overcurrent protection
Accuracy class Current error at Phase displacement at Composite error at

rated primary rated primary current rated accuracy limit
current % primary current %
5P ±1 ±60 minutes 5
10P ±3 ±60 minutes 10
Table 6: Error for HV protective current transformer
10 Loadability Limits of other Elements of the Protection Schemes and

Primary Equipment
The above calculations are to determine the maximum permissible load current which protective relays
can carry without any relay operation or adverse effect on relay functionalities (RLL) under normal
system operation. To determine what maximum load a transmission line can carry out other factors
such as relay rating, thermal limits of current transformers, transmission line
rating(continuous/emergency) and stability limits (transient and dynamic) should also be taken into
account. After considerations of all of these factors the final Maximum Safe Loading (MSL) of a line can
be determined, this can be used for operation in system control centre.
11 Operation of protective devices under normal system conditions

In the past, several unplanned outages have occurred by operation of protective devices under normal
system condition. The main reasons for these disturbances have been due to wrong relay settings and
unexpected high load, human errors during maintenance tests, relay failures and wrong design. These
disturbances have occurred on HV feeders (11, 22,33kV), EHV transformer and transmission lines. The
following are analysis of some these disturbances:

APS | 2015
11.1 Operation due to wrong setting and high loads
11.1.1 HV Feeders
Generally the protection of a HV feeder consists of a single overcurrent, earth fault and sensitive earth
fault (SEF).
Operation of overcurrent under normal system configuration due to load have been very common due to
one of the following factors:
1) The pick-up settings have been set lower than maximum load due to wrong
assumption of load magnitude and its short term forecast.
2) Errors due to CT and relay are not accounted during setting calculation (wrong RLL).
3) Unpredictable switchings and configuration along the feeder.
4) Lack of periodical setting review (3-5 years) to take into account of load increase and
configuration change.
5) Wrong setting for load encroachment element (both magnitude and power factor).
6) Cold load pick up
The earth fault setting of feeder protection is not affected by the magnitude of the load. However, the
setting of overcurrent protection of the HV feeder should meet the following criteria;
7) The pick-up setting is to be above 1.1-1.3 times of maximum feeder load (Stability
criteria).
8) It should detect 80 percent of minimum phase to phase fault at the end of protect
zone (sensitivity criteria).
9) It should grade with downstream protection such as recloser or fuse.
In some cases both sensitivity and stability criteria cannot be met, consequently, the following methods
have been used to rectify the problem
Method 1: Installation of additional recloser or fuse downstream

Method 2: Utilising of load encroachment element of modern numerical relays.
Method 1 is traditional one and costly. With Method 2 the pick-up setting of feeder overcurrent
protection can be set below maximum feeder load in order to meet sensitivity criteria. However, this
requires LE to be set to meet the following criteria:
• The angle of LE should be set well below the angle of the impedance of distribution
feeder and above the maximum load angle (worst power factor).
• Preferably, but not necessary the magnitude of LE to be set below minimum load
impedance.
In some cases the angle of LE could not be set properly to meet the above criteria due to the low angle
of the impedance of distribution feeders. In addition, maximum load angle is unpredictable since the
feeder can be tied to other feeders at different downstream locations. In some practical cases, the LE
angle was set to 40 degrees based on worst load power factor 0.8. However, feeder overcurrent
protection operated since the load angle increased above 45 degrees when the feeder was tied to other
feeders downstream. Investigation showed that raising the angle of load encroachment element could
not be a solution since the three phase fault impedance then lies on load encroachment area. Based on
this, depending on the impedance angle of the feeder, utilising the LE element cannot be appropriate to
prevent operation of overcurrent relays on high loads.

APS | 2015
11.1.2 Supply Transformer
Supply transformer (eg 110 kV/11 kV) can have overcurrent and impedance protection in addition to
differential protection. The overcurrent protection is installed on both the HV and EHV side of the
transformer. The purposes of overcurrent protection on the HV side of the transformer (11, 22 kV) are
the following:
a) To provide HV busbar protection

b) To provide back-up protection for CBF of the feeder since HV feeders have not
generally CBF scheme.
The purpose of overcurrent protection on the EHV side of transformer is as follows:
c) To provide back-up protection for the transformer where there is only one set of main
differential protection.
d) To provide back-up protection for the CBF on the HV side for HV busbar fault.
The philosophy of the pick-up setting of transformer’s overcurrent protection has been based on one of
the following criteria’s with consideration of sensitivity criteria:
1) To be set above maximum short term load as it may flow on each transformer under
normal system condition and N-1 contingency.
2) To be set based on maximum rating of the transformer.
3) To be set based on thermal overload rating of terminal’s cables.
In the past, frequent transformer overcurrent protection has operated due to load for the same reasons
that were given for HV feeders.
In some electricity companies, the HV and EHV busbar side of the transformer have distance protection
instead of overcurrent protection. The reach of zone has generally been set high in order to provide
back-up protection for fault at the end of the HV feeder. This necessitated enabling LE element in order
to avoid operation of distance protection under normal loading condition. In some cases, since the
setting of LE element had not been set based on N-1 load flow contingency, distance protection
operated when parallel transformer was taken out of service. This resulted to total substation black out.
11.1.3 EHV Transmission Lines
Generally protection of EHV transmission lines consists of current differential and distance relays. It is
not a good industry practice to use overcurrent protection for EHV lines although in the old days it has
been used for protection of sub transmission lines (66kV below).
Operation of distance relays due to load has occurred frequently and in some cases has caused
cascade tripping and large system black outs. The main reasons are as follow:
1) Encroachment of load to zone reach

2) Wrong setting of the load encroachment element (both magnitude and angle).
3) Errors due to CT/VT and relay not accounted during setting calculation.
4) Lack of periodical setting review (3-5 years).
5) VT selection switching.
Enabling load encroachment element on numerical relays have significant impact on prevention on
unwanted operation of distance relays due to high load.

APS | 2015
Investigation of 2003 North Eastern Blackout in North America [6] indicated that there were 11 instances
of Zone 3 trips on highly loaded transmission lines where load encroachment blocking hadn't been
configured. Subsequent investigations showed that the spread of disturbance could have been reduced
if relay load encroachment had been enabled correctly on all of these 11 transmission line protection
schemes. In addition the following were found:
1) Some of relays on these lines were old and did not have LEE. Some did not
have lenticular characteristic in order to reduce the possibility of load
encroachment to distance zone. Solutions which could be implemented were
replacement of relays, reducing Zone 3 reach or disabling it where back up
protection for multi-phase fault is not required.
2) Worst parameter assumption such as 0.85pu for nominal system voltage and 40
or 45 degrees for load power factor and effect of CT/VT errors were not
considered for calculation of load impedance and zone reach.
In addition to unwanted operation of distance relays, operation of TL current differential protection under
normal condition has occurred due to wrong settings. In one event, a long 220 kV line tripped by current
differential protection during commissioning. It was found that the basic pick up current setting of
differential protection was below the line charging current of the line. The charging current as recorded
by line digital relays as high as 60 A.
11.2 Operation of Relays due to human errors, design errors and relay failure
The other reasons which have caused operation of relays under normal system conditions and
consequently load interruptions were due to human errors during maintenance tests, relay failure and
wrong design. Some of them are described below:
1) Incident 1 : Operation of CBF during testing
This incident occurred on a large hydro power plant which is connected to interconnected network via
radial parallel 220 kV transmission lines. The transmission line has two main new numerical relays. One
configured as current differential protection and one as distance protection. The CBF scheme has been
programmed in differential protection. The scheme is initiated by protection of generator and its step up
transformer and sends direct intertrip to remote end via current differential communication link.
The event occurred during the running of the “Automated Relay Test System". This system isolates trip
contacts, field switch and master trip relay trips, but not CB fail from the particular relay (generator and
or transformer protection) under test and applies a secondary injection current and voltage. Following a
relay operation the injection is removed and the trip(s) are reinstated "automatically".
At the time of the event, the generator was operating at approximately 45 MW. This gives the 220 kV
side current well in excess of the 50 A CBF pickup level. The relay events indicate that automated test
sequencing held up the CB Fail initiate contact of TF protection in excess of 200ms prior to reset. This
condition allowed the CBF scheme of current differential protection of 220 kV line operates and sends
direct intertrip to remote end.
Before installing new transmission line protection, this incident did not occur since the CBF pickup level
was 450 A well above the generation current. However this pick up was lowered on new relays for
detection of high fault resistance for phase to ground fault since relay did not have an option to program
separately CBF for multi-phase and phase to ground fault.
In order to avoid the repeat of the incident, it was decided to operate the automated test when the
Generator CB is open (eg with no current to the CBF element). Alternatively to isolate physically the
output contact of the generator/transformer protection which initiate the CBF scheme of the
transmission line.
2) Incident 2 : Faulty Relays
Several disturbances have occurred due to defective output contact of numerical relays. One of these
failures tripped 110 kV bus and caused partial loss of supply. It was found the defective CBF contact is

APS | 2015
permanently closed instead of normally open. The current check of CBF used to prevent operation of
the CBF until the load current exceeded the pick-up. Investigation found that the defect resulted when a
ribbon cable terminal connector was damaged during a planned RAM upgrade[9].
Unwanted tripping’s have also occurred due to change of relay characteristic and component failures
mainly due to age.
3) Incident 3 : Misuse of Multimeter
This incident is related to accidental energisation of a contact associated to CBF scheme of an HV

capacitor bank. The CBF scheme has been programmed on numerical overcurrent relay of HV cap
bank and is externally initiated by neutral unbalance current protection using a binary input of
overcurrent protection of the capacitor bank.
During routine maintenance tests, a field operator was checking the integrity of external CBF initiating
binary input contact using multimeter.
It was checking by measuring the resistance across the contact by connecting the terminals of the
contact to multimeter. This caused short circuit across the contacts and subsequently the operation of
the CBF scheme and tripped HV busbar. Investigation of the incident revealed that the multimeter was
on the wrong setting. For example, instead of using ohm range, it was in volt range. To prevent a similar
incident, the following recommendations were made:
a) When using Multimeter, it should be ensured that it is with correct setting range.
b) The measurement should be made from each terminal to ground rather than across the
terminals.
c) During any measurement or test, the trip output contact of CBF to be isolated.
4) Incident 4 : Operation of CBF scheme
This incident occurred during routine maintenance tests of transformer guard relays. While checking
and testing the Buchhoz relay, the CBF of the 110 kV bus was initiated by Bucholz and tripped the bus.
5) Incident 5: Maloperation of Special Protection System
In this event Special Protection System (SPS) operated incorrectly. SPS has been designed to trip
industrial loads or generators on loss of a tie line depending on import/export conditions of tie lines. This
to ensure the system frequency remains within the acceptable level. Due to wrong SPS software
programming, on one occasion the loads were incorrectly tripped. Correct changes were made to the
software in order the tripping relays do not get unnecessary signal. The tripping relays should only be
armed or disarmed based on what SPS software determines in the last 4 second run.
6) Incident 6 : Tripping 220 kV TL during re-energisation
The 220 kV TL had already been tripped for a previous fault. However, during re-energisation it tripped
since the trip output contact had a latched contact while it should have been unlatched. In another
similar incident, the line tripped after energisation since the trip output contact had not reset due to high
allocation of reset time.
7) Incident 7 : Large worldwide system black outs
Many system black outs have occurred worldwide for different reasons such as high loads, human
errors and system mismanagement. Table 7 shows a summary of these black outs.

APS | 2015
Description of incident Cause of incident Mitigation Technique
Florida Coast to Coast black A field engineer isolated the two During work on any equipment at
out in 2008. 26 TLs and 38 protections of a line during trouble least one main protection and a
substations tripped and nine shooting of a failed 138 kV line switch. back-up protection should be
power plants including 2 Subsequent fault on switch caused available.
nuclear plants were shutdown. tripping of two nuclear plants due to
long depressed voltage. This caused
cascade tripping of other primary
plants.
Blackout in European The main causes of black outs were Design Criteria should consider
Networks: due to cascade tripping of transmission
lines after tripping of a heavy loaded N-1 contingency to cater for
• Italy and Switzerland line or a generation plant. The overload conditions. In these
2003 incidents generally resulted in system events maloperation of protection
• Germany and France voltage instability and operation of and control devices were not
2006 under frequency relays. involved.
• Sweden and Denmark
2003
1965 North-eastern United A relay with a faulty setting on a line Per the recommendations of
States black out tripped. The power shifted to three North American Electric
parallel lines, which quickly became Reliability Council (NERC), a
overloaded, triggering subsequent number of improvements to the
relay actions and cascade tripping network protection systems,
including the use of back-up
relays, and automatic load
shedding were made.
Three blackouts in West coast Mostly due to relay failure and Recommendation was made to
of America in 1996 vegetation. implement special protection
systems (SPS) to shed load
and/or generation automatically
based on a set of criteria and
regular testing of relays for
hidden failures, and managing
vegetation along transmission
corridors.
In 2003 in USA 62,000 MW It was initially caused by the failure to Planned vegetation system
were interrupted and 50 million trim trees under and near power lines. management to be implemented
people were affected. This brought a few power lines down and better training for operational
and cascading tripping occurred due to groups for handling disturbances.
lack of quick and correct action of the
local power company which could have
localised the problem.
There have been 529 events in Initiating causes have been: weather
USA between1984-2000 for and fire (40%), equipment failure
which major system (31%), human error (11%), insufficient
disturbances have occurred generation (15%) and sabotage (3%),
Table 7: Large Worldwide System Blackouts

APS | 2015
12 Conclusion
In this paper, the loadability limits of modern numerical relay for distance, current differential and
overcurrent protection were discussed. For distance relay, the effect of load encroachment element and
power swing blocking elements which are part of modern relays were considered. It was shown that
loadability limits of distance relays should be calculated based on characteristic of power swing blocking
element rather than traditional method of the last forward zone reach.
It was shown that LE characteristic does not need to grade with any forward zones as long as the
maximum load angle is less than the line angle with sufficient margin. The magnitude setting of LE can
be based on different criteria such as maximum load under N-1 contingency or simply based on
emergency rating of the line. Its angle setting should be set based on maximum load angle under
contingency.
The setting of power swing characteristic should be selected in order so the minimum load impedance
does not encroach the outer characteristic, otherwise the functionality of PSB will become ineffective.
Based on this concept mathematical equations were developed to calculate loadability limits of modern
numerical distance relays.
For current differential protection, loadability limit is theoretically infinite if correct setting philosophy is
applied. However, maximum permissible load current can be limited to the thermal ratings of CTs which
provide current signals to differential relays.
For overcurrent protection, loadability limit should be calculated based on the applied setting with
consideration of CT and relay’s errors. These errors depend on the type of numerical relays and the
CTs.
It is essential loadability limit of a protection device to be calculated under the condition that the relay
has been set based on sound philosophy. With this condition if it is found the loadability limit of the relay
limits unnecessary, the required transmission line loading and engineering solutions as discussed
should be implemented
Finally, system disturbances due to wrong relay’s settings, relay’s defects and human errors under
normal system conditions were discussed.
The major contributing factors for wrong settings were due to inaccurate assumption of maximum load
level (magnitude and angle) and lack of consideration of different system configurations. Enabling of the
Load Encroachment element for EHV transmission line could help to prevent disturbances while for
distribution feeder may cause disturbances.
Disturbances due to relay’s defects were mainly related to faulty output trip contacts. Frequent checking
of contacts are required to prevent these type of disturbances
Most of the disturbances due to human errors as discussed in this paper were due to inadvertent
energisation of CBF schemes during routine tests. Disabling the schemes before testing could prevent
the disturbances.

APS | 2015
13 References
[1] AREVA : Micom P54X; Current Differential Relay; Technical Manual
[2] Schweitzer Engineering Laboratories: SEL-421 Relay; Protection Automation Control;
Application Handbook
[3] GE Multilin : L90 Line Current Differential System; UR Series Instruction Manual
[4] Siemens : Multi Function Relays; Technical Manuals
[5] Alstom Grid : Network Protection & Automation Guide; 2011
[6] 2003 Northern Eastern Blackout: Interim Report; Causes of the August 14th Blackout in the
United States and Canada
[7] Rincon, C; Joe Perez; Midence, R : Calculating Loadability Limits of Distance Relays;
[8] IEEE Working Group D6 : Transmission Line Protective System Loadability ; A report to the
Power System Relay Committee; March 2001
[9] CIGRE AP B5 Melbourne Meeting : WIL 110 kV Bus B Tripping Wellington LOS Event; 2009
About the Author
Maty Ghezelayagh received his B.Sc. from Sharif University of Technology (Iran),
M.Sc. from University of Manchester (UK) and PhD from University of Wollongong
(Australia) all in electrical power systems. For the last 25 years he has worked with
five different companies in Australia and before that for four years for the main 400 kV
transmission company in Iran. His major work experiences have been in all areas of
protection and control of power systems namely EHV transmission and distribution
networks, substations and power stations.

APS | 2015

APS
APS| 2015
| 2015
An Overview of D.C Traction Protection Systems
Zahra Bayat | Yarra Trams

zahra.bayat@yarratrams.com.au
1 Introduction
DC electrical traction systems consist of the fixed components that enable power to be supplied to
trains and trams for their traction and auxiliary operations. The purpose of the system is to supply
electric power to rolling stock for traction and auxiliaries. It may also accept power back to the network
when regenerated through breaking by the train or tram. The electrical interface is the interaction of
the traction supply with the vehicle based systems and components.
Key factors that influence the effect on DC protection system are:
• Magnitude of electrical load – stationary and moving including the type of train, the current
required for powering, the current required for auxiliaries, the number of trains operating, the
timetable, and the speed of operation.
• Traction system capacity including fault levels, protection types, and rate of rise of current.
This paper describes the DC protection system on traction substation and limitations and implications
of the traction system.
2 Single Line Diagram of D.C traction system

The single line diagram of traction substation is shown in Figure 1. Components include incomer AC
switchgear, rectifier transformer, rectifier convertor, rectifier circuit breaker, positive feeder panels and
negative system. AC incomer protection is same as all MVAC feeder protections and is out of scope of
this paper. Other protection schemes will be described in the following sections.
Incoming Incoming Incoming
HV supply 1 HV supply 2 LV supply
HV AC
HV supply switchgear
circuit breaker
HV bus -tie
circuit breaker
HV rectifier
Isolating
circuit breaker transformer
Battery and battery

charger
Rectifier Auxiliary
transformer transformer
Rectifier
converter
Changeover
Negative system
LV AC distribution LV DC distribution
Rectifier circuit isolator board board
breaker
Positive bus (+)

Transfer bus (+)
Negative bus (+) Rail
59
Short -circuiting Bus -tie circuit breaker Frame leakage
switch protection relay
Feeder circuit
breaker
Feeder
Transfer switch
panel
External
isolator
Figure 1: Single line diagram-traction substation

APS | 2015
3 Scheme of protection function on transformer/rectifier circuit

The rectifier circuit consists of the rectifier transformer, rectifier and the rectifier DC breaker. A diagram
showing the circuit and the associated protection equipment on transformer/rectifier circuit is shown in
Figure 2. The objectives of the protection equipment are to isolate the fault immediately when a fault
occurs within the rectifier circuit and to protect the rectifier thermally.
Figure 2: Protection scheme on transformer/rectifier circuit
3.1 Rectifier transformer protection
Three protection devices are installed inside the rectifier transformer to measure the transformer oil
temperature, winding temperature and surge of gas. In addition, three overcurrent protective devices
are installed on the AC incomer breaker and measure AC current on primary side of the rectifier
transformer, which are Instantaneous, Inverse Definite Minimum Time (IDMT) and Earth Fault (E/F)
overcurrent protective relays.
An overcurrent protection coordination time current characteristic curve is shown in Figure 3. The aim
of overcurrent protection scheme is to trip AC breaker under transformer fault or rectifier internal fault.
The high set instantaneous overcurrent protective relay should trip the AC breaker instantaneously
under a short circuit at the primary side of rectifier transformer.
If the short circuit fault or earth fault happens at the secondary side of rectifier transformer the inverse
time overcurrent and earth fault relay should trip AC breaker according to the time characteristic of
relay. In addition, under voltage, CBF and negative sequence protection functions are considered on
AC breaker protection relay.

APS | 2015
Figure 3: Overcurrent coordination time current characteristic curve (rectifier transformer)
The rectifier fault relay has an inverse time current characteristic which matches the thermal withstand
characteristic of the diode, so the relay can protect the diode of the rectifier thermally by tripping the
AC breaker under overload condition. Rectifier load characteristic curve is shown in Figure 4.
Figure 4: Rectifier load characteristic curve

APS | 2015
The overcurrent protection scheme can also give a backup protection for the DC trackside fault.
Consider a close up fault occurs at the traction substation, the DC fault current is expected to be 30
KA when two rectifirers are in normal service. Each rectifier should carry a DC fault current of 15 KA.
In AC base, the AC fault current is about 616.09 A.
The calculation is based on the following equation:
750V * Idc = √3 * AC Voltage(KV) * Iac *0.96 (Power Factor) (1)
The settings of the three overcurrent protection devices are arranged such that the overcurrent
coordination time current characteristic can allow the system to carry overload capacity of the rectifier.
The overload capacities of the rectifier unit are stated as follows:
- Continuous at 100% of the rated capacity
- 120 minutes at 150% of the rated capacity
- 1 minute at 300% of the rated capacity
- 15 second at 450% of the rated capacity
Rectifier transformer rated current In = Power rating of rectifier transformer/(√3* voltage rating of AC
primary of transformer)
So,
Phase Time OC (51) > 300% of In (2)
Inverse operating time should be greater than the rectifier transformer inrush current duration time
which is estimated to be 6 cycles (0.12s).
High set overcurrent (50) has to be set to avoid unwanted operation inrush current and rectifier
transformer maximum through fault measured on the primary side.
Imaxtroughfault = Rectifier transformer secondary1 rating power /√3* secondary1 nominal voltage*
seconsary1 impedance persent (3)
The phase to ground element should be set to measure the residual currents. The earth fault and
sensitive earth fault should be in line with the following conditions:
• The earth fault (EF) and sensitive earth fault (SEF) shall discriminate against the existing
upstream settings.
• Both should be less than the minimum fault level on the HV side of the transformer.
The earth fault operating times should follow the below conditions:
• Sensitive earth fault operating time = Upstream sensitive earth fault operating - 1 s
• Earth fault inverse operating time = Upstream inverse earth fault operating time – Minimum
discrimination time
Because the rectifier transformer vector group’s LV winding is isolated from the ground, any earth
faults on the secondary side of transformer cannot cause enough fault current to be detected by the
relay on primary side and therefore, no coordination will be required for this case. The above setting is
to protect the transformer feeder against phase to ground faults up to the transformer HV bushing.
3.2 Rectifier protection
Two protection functions are installed in the rectifier cubicle to provide interlocking and protect the
rectifier against overtemperature and overvoltage.

APS | 2015
3.2.1 Rectifier overtemprature relay

When the temperature exceeds the trip setting, relay will operate and trip the incomer AC breaker.
Trip setting:110°C~ 120°C
3.2.2 Rectifier surge fuse failure relay

The diodes of the rectifier are protected against overvoltage by connecting suitable value of capacitor
across the DC output. Resistors are also connected in parallel with the capacitor to dissipate the surge
power. Each surge circuit is protected by a surge fuse. Surge fuse failure will operate the relay and
give out local and remote alarm.
3.3 DC incomer breaker protection
The rectifier DCCB will detect reverse current flow from the DC busbar to the rectifier and trip instantly
by its own mechanism. This will result in function 32 picking up due to proprietary circuitry on the
DCCB. A contact of 32 shall be used to initiate lockout of the rectifier, with consequent opening of the
ACCB and de-energization of the DCCB auxiliary closing circuitry. Since both line and rectifier circuit
breakers are of the series trip type, they will not discriminate for high fault currents. For this reason,
rectifier circuit breakers are not fitted with forward direct acting tripping.
4 Traction feeder protection scheme

Track feeder circuit consists of the 750 V DC track feeder breakers which feed traction power to the
overhead line. The purpose of the protection devices is to isolate the close up fault immediately,
provide overload protection for the overhead line and remote fault discrimination.
The basic protection function is the direct acting overcurrent tripping system. Depending on the train
characteristics and the line impedances, the load current of a line section can be greater than a distant
fault or a distant arcing fault and the instantaneous overcurrent level protection cannot achieve the
discrimination between service currents and distant fault currents. In this case the line circuit breaker
should be equipped with protection devices [2] using some or all of the following additional protection
functions which are shown in Figure 5:
• Protection examining the wave shape of the current (di/dt, ∆I)
• Inverse time overcurrent protection
• Inverse time overcurrent protection with thermal imaging
• Under/over voltage protection
Figure 5: Traction feeder protection functions

APS | 2015
4.1 Instantaneous overcurrent protection
This function can provide high speed operation on high line fault current. It is bidirectional and usually
set slightly lower than minimum fault current level and mechanical overcurrent setting on DCCB on
feeder panel. An overcurrent protection coordination time current characteristic curve on feeder panel
is shown in Figure 6.
Figure 6: Overcurrent protection coordination time current characteristic curve
4.2 Inverse time overcurrent protection
This characteristic is used on line circuit breakers to protect the feeder cable or contact line in the
circuit by choosing a characteristic similar to that of the cable or contact line. This element gives an
inverse time current setting to a variety of mathematical laws. The relay would have an inverse time
law and a range of time multipliers. Two types of STOC function for different relay types are described
below.
On Swartz relay [6] :
∆ISTOC (∆V,dial) = (50mv/∆V)*(dial+1)s (4)
STOC curve on Swartz relay is shown on Figure 6.
Inverse overcurrent protection formula on Mitre relay [5] :
0.14/ °˙°² 1 ∗ (5)
: Trip time is second

: measured current as a multiple of setting
: Time multiplier setting

APS | 2015
Note: some types of relays use IDMT function to detect current which last a long time and do not
accord with normal railway operating conditions. An example on Sitras Pro (Siemens relay) [4] and
cable damage curves is shown in Figure 7. In this example the setting for IDMT protection is 1800 A
with 10 second time delay.
Figure 7: IDMT function characteristic curve (Sitras Pro relay)
4.3 Inverse time overcurrent protection with thermal imaging
The maximum permissible operating temperature can be exceeded during overloads without the short
circuit monitoring function detecting a fault. For this reason, the operating temperature of the overhead
contact wire and feeder cable can be monitored at the same time. If a set temperature rise is
exceeded then the relay trips and the relay resets after a period of cooling set by the hysteresis value.
This function has different settings processes on different types of relays. However, the response of
the thermal element to current assumes square law heating and Newtonian cooling as below:
1 ) (6)
Temperature rise above ambient of the thermal image as a function of time

Aiming temperature rise
T Time constant of thermal image conductor
If the current applied is I1 then the square law on heating gives the aiming temperature rise as:
!
∗ 50 (7)
I1 Actual current
I50 Current which would cause Theta to be 50 K i.e. 50 K rating of thermal image conductor

APS | 2015
4.4 Protection examining the wave shape of the current (di/dt, ∆I)
The relay operates when the current wave shape detected by the relay exceeds the di/dt setting for
the selected time delay t. Train starting currents may exceed the di/dt setting, but would reset before
the time delay has elapsed.
∆I function operates when a step change of current occurs in a prescribed time. This step change
should be less than the distant fault current and with a time delay long enough to allow starting
currents to reduce below the step change of current. Having a di/dt feature coupled to the start of the
step characteristic can improve discrimination between trains starting current and fault current. Trip
range on ∆I and di/dt graph is shown in Figure 8.
Figure 8: Characteristic of ∆I and di/dt tripping
Considering how the relay should be set to operate on a particular distant fault it is assumed the fault
current against time can be represented as the following formula:
I (t) = (If-Istart)*(1- # ) (8)
I (t) Current as a function of time, with the fault assumed to occur at t=0
Istart Current at the initiation of fault
If Final value of fault current
$ Time constant of fault current
The fault current curve with an ideal DC source is shown in Figure 9.

APS | 2015
Figure 9: DC fault current without harmonic filters and with an ideal DC source
In case of having harmonic filters and reactor on outgoing side of rectifier, feeder fault current will be
the same as curve in Figure 10.
Figure 10: DC fault current with harmonic filter and reactor on outgoing side of rectifier
Current at initiation of fault (I0) and Initial rate of rise current is calculated through the following
equations:
%&'() %&'()
Io = (- ∆Imax +If*(1 - # ))/ (1- # ) (9)
di/dtinitial = (If – Io)/$ (10)
Note: From experience on different type of relay and recording the trips and values on ∆I, function
∆Imax should be selected at range of average max starting current of different type of trams or trains in
the electrical sections.

APS | 2015
Rate of rise curve without harmonic filters with an ideal DC source which is shown in Figure 11 is
provided according to the following formula:
%
di/dt = di/dtinitial * # (11)
Figure 11: di/dt Curve- without harmonic filters and with an ideal DC Source
It assumed harmonic filter parameters at each unit (rectifier):

300 Hz filter: 10 kΩ ǁ (1.12 mH+250µF)
600 Hz filter: 10 KΩ ǁ (0.7 mH+100µF)
di/dt curve with harmonic filters and reactors on outgoing side of rectifier is shown in Figure 12.
Figure 12: di/dt curve with harmonic filter and reactor on outgoing side of rectifier

APS | 2015
4.5 Under/Overvoltage protection
Regarding technical specification of railway published by Public Transport Victoria and IEC 60850/EN
50163 standards [1], undervoltage setting value on tram network is 400 V and overvoltage setting is
800 V. On both settings value of 1 second delay should be considered.
5 Rail to earth protection

In DC traction system, impermissible touch voltage may occur between the return circuit and the
structure earth due to operational and short circuit currents. SCD (Short Circuit Device) on VLD
(Voltage Limiting Device) panel on traction substation is designed to prevent these high touch
voltages. The limit of the touch voltage is defined in IEC 50122-1:2011[3] which is shown as a function
of time duration in table 1.
t Ub,max
s V
˃300 120
300 150
1 160
0.9 165
0.8 170
0.7 175
0.6 180
0.5 190
0.4 205
0.3 220
0.2 245
0.1 285
0.05 325
0.02 370
Key
t time duration
Ub,max permissible body voltage
Table 1: Maximum permissible body voltages Ub,max in DC traction systems as a function of time duration
6 DC power limitations and implications

DC electrical traction system has limitations and implications such as:
• Problem of single end fed electrical section specially on long distance feeding point condition
• The effect of different type of vehicles and their starting current
• Pantograph voltage limitations
• Underground cable screen fault
• Overhead conductor limits

APS | 2015
5.1 Problem of single end fed electrical section
Single end fed condition is a case when a section of the track is supplied from one substation only.
This situation may occur when one of DC track feeder breakers which feed a section is open with the
opening of associated isolator. If trains on the section start simultaneously, the total train’s starting
current will exceed the setting of the direct acting instantaneous overcurrent trip device of the single
end feeder breaker feeding the section. This results unnecessary tripping as this case is not a fault
condition.
One possible solution to this is to impose a restriction on the operation schedule such that
simultaneous starting of the trains cannot occur under single end fed condition.
5.2 The effect of different type of vehicles and their starting current
The tram motors and auxiliaries require a certain amount of voltage to operate properly. If the voltage
is below the minimum requirement, the motor may experience degraded performance and the
auxiliaries may switch off. The newer trams are not power limited. So, as the pantograph voltage
drops, their motors try to maintain the acceleration and they draw more current from the substation.
This makes the voltage even lower and eventually, the power system becomes overloaded and the
circuit breakers will trip to prevent damage to the equipment. This may result in a temporary power
loss to the section if both feeds trip. This effect is illustrated in Figure 13.
Figure 13: Various vehicle classes at a nominal pantograph voltage of 600 V at maximum acceleration
5.3 Pantograph voltage limitations
Because of the resistance in the wires, voltage is lost as power flows from the substation to the trams,
so the voltage at the tram pantograph may be much less than 600 VDC. As the section length
increases the resistance and number of trams of the section increases and more voltage is lost as a
result. There are various upgrade strategies for dealing with poor pantograph voltage:
1- Limit the current

a) Restrict speed and acceleration
b) Reduce service frequency
2- Decrease the resistance

a) Substation feeder enhancements
b) Side feeder enhancements

APS | 2015
3- Increase the substation power rating
4- Reduce the section length which inherently limits the current
5.4 Underground cable screen fault
Many of the feeds from the substation to the overhead wiring are underground cables. While the core
of these cables is typically a 400mm2 copper conductor, the outside of the cable is protected by a
120mm2 copper screen. One purpose of the screen is to ensure that inadvertent contact with the cable
will cause a short circuit to occur between the screen and the core thereby tripping the substation
protection. The drawback of the 120 mm2 screen is that its resistance is much higher than the
conductor core, and consequently, the longer the cable, the more rapidly the distant fault level
declines. For this reason, underground cable segments should be limited in length and the cable
screens broken into segments.
5.5 Overhead Conductor Limits

In general, the conductor thermal limits are less of a concern than tram pantograph voltage or
protection discrimination. However, it is possible to exceed conductor ratings if the average load of an
electrical section is high. The current limit for the primary conductors in the tram electrical network is
approximately the following:
Conductor Continues Current Rating Thermal Time Constant
129 Sn-Cu trolley wire 480 A per trolley wire ̴ 10 min
400 mm2 XLPE copper cable 1000 A ̴ 30 min

Table 2: List of primary conductor’s current rating in Melbourne tram network
7 Conclusions
Above DC protection functions are adequate for the operational requirement of the existing DC traction
system. However, this system cannot fully protect a section on abnormal service or single end fed
condition. In this condition, there is a risk of damaging of overhead line when the distant fault happens.
Improvement on supply capacity and reducing distance between substations and feed points to
sections will decrease noises tripping on protection system and reliable and secure power supply will
be predictable.
8 References
[1] IEC 60850/EN 50163 Standard
[2] IEC 61992-7-1 :2006-02 Standard
[3] IEC50122-1:2011Standard
[4] Sitras Pro Operating Instructions version1.4.0
[5] Mitre-MTR10 relay Manual
[6] Type 76 DC overcurrent Detector instraction Maual

APS | 2015
About the Author

Zahra Baya t received her BSc of Electrical Engineering from South Tehran
University in 2007. Currently, she is working as an Electrical Engineer
(secondary system) in the area of protective relays and power study
research on power supply upgrade division in Major Projects department in
Yarra Trams.

APS | 2015

APS | 2015

APS
APS| |2015
2015
The Digital Substation – Principles, Functions and Benefits
Alexander Apostolov | PAC World

alex.apostolov@pacw.org
1 Introduction
The IEC 61850 standard for Communication Networks and Systems for Utility Automation allows utilities
to consider new designs for substations applicable for both new substation and refurbishments. The
levels of functional integration and flexibility of communications based solutions bring significant
advantages in costs in all stages of a project. This integration affects not only the design of the substation
but almost every component and/or system in it such as protection, monitoring and control by replacing
the hardwired interfaces with communication links. Furthermore the design of the high voltage
installations can be reconsidered regarding the number and the location of switchgear components
necessary to perform the primary function of a substation in a high voltage network. The use of high-
speed peer-to-peer communications using Generic Substation Event (GSE) messages and sampled
values from non-conventional or conventional sensors allows the development of distributed
applications. In addition, the use of optical local area networks leads in the direction of copper-less
substations.
The paper focuses on the definitions of the digital substation communication busses such as the station
bus for the communication at the station level and between bays as well as the process bus for the
communication between the high voltage process and components interacting with it. It analyses the
substation communications architectures in substations with full implementation of IEC 61850, i.e. with
station and process bus.
The different types of devices required for the digital substation are described. Merging units that provide
the interface between the current and voltage sensors and the intelligent electronic devices at the
equipment, bay or substation level are described, as well as . The distribution of signal and data
processing functions between the different devices for both architectures are analyzed.
Some specific substation applications based on GOOSE and Sampled Analogue Values are described
later in the paper and demonstrate the advantages of the new technology. Improvements in functionality,
combined with practical elimination of performance or safety issues are covered in the paper.
2 Conventional Substation Design

Conventional substations are designed using design procedures for the high voltage switchgear in
combination with copper cables for all interfaces between primary and secondary equipment.
Several different types of circuits are used in the substation:
• Analog (current and voltage)

• Binary – protection and control signals
• Power supply – DC or AC

APS | 2015
Figure 1: Conventional substation design
Depending on the size of the substation the location of the switchgear components and the complexity
of the protection and control system there can be a huge number of cables with different lengths and
sizes that need to be designed, installed, commissioned, tested and maintained.
A typical conventional substation has multiple instrument transformers and breakers (Figure 1)
associated with the protection, control, monitoring and other devices being connected from the yard to
a control house with the individual equipment panels. The different types of cables described are used
for the connections.
Figure 2: Substation panel wiring
As can be seen from Figure 2, cables are cut to a specific length and bundled, which makes any required
future modification very labour intensive. This is especially true in the process of refurbishing old
substations where the cables insulation is starting to fail.
At the same time, there are many non-electrical interfaces that are based on different sensors with low
level analog interfaces, such as 0 ... 1 V DC, 0 ... 10 V DC, 0 ... 20 mA or 4 ... 20 mA.
The large amount of copper cables and the distances that they need to cover to provide the interface
between the different devices exposes them to the impact of electromagnetic transients and the
possibility for damages as a result of equipment failure or other events.
The design of the conventional substation needs to take into consideration the resistance of the cables
in the process of selecting instrument transformers and protection equipment, as well as their connection
to the instrument transformers and between themselves. The issues of CT saturation are of special
importance to the operation of protection relays under maximum fault conditions.
Failures in the cables in the substation may lead to misoperation of protection or other devices. In some
cases, such as an open CT circuit they represent a safety issue, especially when it occurs while the

APS | 2015
primary winding is energized. The induced secondary e.m.f. under these circumstances can be high
enough to present a danger to people’s life and equipment insulation.
The above is definitely not a complete list of all the issues that need to be taken into consideration in
the design of a conventional substation. It is intended only to provide some examples that will help us
better understand the impact of IEC 61850 in the design and operation of substations.
In order to take full advantage of any new technology, it necessary to understand what it provides.
The next part of the paper gives a short summary of some of the key concepts of the standard that have
the most significant impact on the substation design and the transition to a fully digital substation.
3 IEC 61850 Based Digital Substation Functional Hierarchy

A digital substation is a substation where all the interfaces between the primary equipment (the process)
and the Substation Protection, Automation and Control System (SPACS), as well as between all the
secondary equipment of the SPACS are communications based, using the object models and
communications services defined in the IEC 61850 standard.
The development of different functions in the substation protection and control system is possible only
when there is good understanding of both the problem domain and the IEC 61850 standard. It does not
only define how data is communicated between functions in the substation, but also describes the
functionality of the substation in an object-oriented approach. The concept of distributed functions is one
of the key elements of the standard that allows for utilities to rethink and optimize their substation
designs.
Figure 3: Process Bus definition
A function in an IEC 61850 based integrated protection and control system can be local to a specific
primary device (distribution feeder, transformer, etc.) or distributed and based on communications
between two or more IEDs over the substation local area network.
IEC 61850 defines several ways for data exchange between IEDs that can be used for different forms
of distributed applications. They introduce a new concept that requires a different approach in order to
define the individual components of the systems in substations.
As discussed earlier, the existing designs are based on hardwired interfaces between the high voltage
equipment – transformers, breakers, instrument transformers, etc. and the rest of the substation devices.
Considering the requirements for the reliability, availability and maintainability of functions, it is clear that
in conventional systems numerous primary and backup devices need to be installed and wired to the
substation. The equipment and their interfaces must then be tested and maintained.

APS | 2015
The interface requirements of many of these devices differ. As a result specific multi core instrument
transformers were developed that allow for accurate metering of the energy or other system parameters
on the one hand and provide a high dynamic range used by e.g. protection devices.
Figure 4: Logical interfaces in IEC 61850
With the introduction of IEC 61850 several different interfaces have been defined that can be used for
various substation applications using dedicated or shared physical connections - the communications
links between the physical devices. The allocation of functions between different physical devices
defines the requirements for the physical interfaces, and in some cases may be implemented into more
than one physical Local Area Network (LAN).
The functions in the substation can be distributed between Intelligent Electronic Devices (IEDs) on
the same, or on different levels of the substation functional hierarchy – Station, Bay or Process. These
levels and the logical interfaces are shown by the logical interpretation of Figure 4. The logical interfaces
of specific interest to distributed applications based on process bus are defined [1] as:
IF4: CT and VT instantaneous data exchange (especially samples) between process and bay
level.
IF5: control-data exchange between process and bay level
IF8: direct data exchange between the bays especially for fast functions such as interlocking
IF8 includes the use of GOOSE messages which can are also required for the interface between the
switching devices and the SPACS as part of the Process Bus.
A significant improvement in functionality and reduction of the cost of integrated substation protection
and control systems can be achieved based on the IEC 61850 based communications as described
below.
4 Process Bus Components

Non-conventional instrument transformers with digital interface based on IEC 61850-9-2 [3] (Process
Bus) result in further improvements and can help eliminate some of the issues related to the conflicting
requirements of protection and metering IEDs.
The interface of the instrument transformers (both conventional and non-conventional) with different
types of substation protection, control, monitoring and recording equipment is through a device called a
Merging Unit. This is defined in IEC 61850-9-1 as:
“Merging unit: interface unit that accepts multiple analogue CT/VT and binary inputs and produces
multiple time synchronized serial unidirectional multi-drop digital point to point outputs to provide data
communication via the logical interfaces 4 and 5”.
Existing IEC 61850 Merging Units have the following functionality:
• Signal processing of all sensors – conventional or non-conventional
• Synchronization of all measurements – 4 currents and 4 voltages
• Digital interface –IEC 61850-9-2 LE [3]

APS | 2015
The Implementation Guideline for Digital Interface to Instrument Transformers Using IEC 61850-9-2 is
a profile which was developed by the UCA International Users Group to insure interoperability between
merging units and multifunctional IEDs from different manufacturers. This profiled defined the object
model of the merging unit as a logical device with multiple instances of TCTR and TVTR logical nodes,
as well as the sampling rates and data sets as shown in Figure 5.
Figure 5: Object model of merging unit according to IEC 61850 9-2 LE
The use of process bus is expected to be further advanced by the availability of the new IEC 61869-9
standard for digital interface of instrument transformers developed by IEC TC 38.
It is important to be able to interface with both conventional and non-conventional sensors in order to
allow the implementation of the system in existing or new substations.
The stand-alone Merging unit has similar elements (as can be seen from Figure 6) and can be
considered as the analogue input module of a conventional protection or other multifunctional IED. The
difference is that in this case the substation LAN performs as the digital data bus between the input
module and the protection or functions in the device. They are located in different devices, just
representing the typical IEC 61850 distributed functionality.
There are several important differences between the data sampling in a microprocessor based relay
and the process bus as defined in IEC 61850:
• While in the relays the sampling is controlled by the IED and is usually using frequency tracking,
in IEC 61850 all interface or merging units are time synchronized with accuracy better than 1
microsecond and use a fixed number of samples per cycle at the nominal frequency
• The sampled values in the IED are exchanged directly between the A/D converter and the
processor, while in IEC 61850 they are transmitted using typically multicast from the merging
unit (publisher) to all IEDs (subscribers) that need these sampled values
Figure 6: Merging unit

APS | 2015
Interoperability between merging units and protection, control, monitoring or recording devices is
ensured through documents providing implementation guidelines. Two modes of sending sampled
values between a merging unit and a device that uses the data are defined. For protection applications
the merging units send 80 samples/cycle in 80 messages/cycle, i.e each Ethernet frame has the MAC
Client Data contain a single set of V and I samples. For waveform recording applications such sampling
rate may not be sufficient. That is why 256 samples/cycle can be sent in groups of 8 sets of samples
per Ethernet frame sent 32 times/cycle [3].
Since SNTP did not provide the required accuracy for the time synchronization of merging units, many
of the existing ones are using 1 pps. However, the precision time protocol (PTP) profiles based on IEEE
1588 are the direction in which our industry is going and we are expecting soon to have the merging
units being synchronized based on IEC 61850 9-3 Time Protocol Profile for Power Utility Automation
[4].
The sampled analog values model applies to the exchange of values of a DATA-SET. The difference in
this case is that the data of the data set are of the common data class SAV (sampled analogue value
as defined in part IEC 61850-7-3). A buffer structure is defined for the transmission of the sampled
values that are the output from the instrument transformer logical nodes TCTR and TVTR (Figure 7).
The information exchange for sampled values is based on a publisher/subscriber mechanism. The
publisher writes the values in a local buffer at the sending side (see Figure 5), while the subscriber reads
the values from a local buffer at the receiving side. A time stamp is added to the values, so that the
subscriber can check the timeliness of the values and use them to align the samples for further
processing. The communication system shall be responsible to update the local buffers of the
subscribers. A sampled value control (SVC) in the publisher is used to control the communication
procedure.
The currents and voltages from TCTR and TVTR accordingly are delivered as sampled values over the
substation LAN. In this case the network becomes the data bus that provides the interface between the
instrument transformer logical nodes and the different logical nodes that are used to model the functional
elements of the IED.
Figure 7: Bus differential based on Sampled Analog Values
Depending on the specific requirements of the substation, the user can design it with different
communications architectures as described in the next section of the paper.
Because of the requirements for use of IEC 61850 in other domains, such as hydro power plants or
distributed energy resources, the concept of process bus is being extended to cover many different non-
electrical process interfaces. The functional decomposition for any specific case can be based on the
abstract model shown in Figure 8.
The process interface is represented by the LN group T Logical Nodes for instrument transformers and
sensors. Their output is Sampled Analog Values that are published over the site’s local area network
and used by any function element that may need them.

APS | 2015
The example in Figure 8 shows logical nodes that may belong on some of the logical node groups
defined in IEC 61850 7-4:
LN group F Logical Nodes for functional blocks
LN group M Logical Nodes for metering and measurement
LN group P Logical Nodes for protection functions
LN group S Logical Nodes for supervision and monitoring
Figure 8: Abstract functional decomposition
The process interface functions are represented by several different logical nodes also defined in
IEC 61850 7-4:
LN Angle TANG
LN Axial displacement TAXD
LN Current transformer TCTR
LN Distance TDST
LN Liquid flow TFLW
LN Frequency TFRQ
LN Generic sensor TGSN
LN Humidity THUM
LN Media level TLVL
LN Magnetic field TMGF
LN Movement sensor TMVM
LN Position indicator TPOS
LN Pressure sensor TPRS
LN Rotation transmitter TRTN
LN Sound pressure sensor TSND
LN Temperature sensor TTMP
LN Mechanical tension / stress TTNS
LN Vibration sensor TVBR
LN Voltage transformer TVTR
LN Water acidity TWPH

APS | 2015
At this stage there are no implementation guidelines regarding process bus interfaces for non-
electrical signals based applications. However, these logical nodes allow the implementation of a digital
interface with any of the primary devices in the substation for the implementation of a fully digital
substation.
5 IEC 61850 Architecture

IEC 61850 is being implemented gradually by starting with adaptation of existing IEDs to support the
new communications standard over the station bus and at the same time introducing some first process
bus based solutions. The specifics of the two types of systems are described in the following two
sections of this part of the paper.
Full advantage of all the features available in the standard can be taken if both the station and process
bus are used.
Figure 9: Station and Process bus functional architecture
IEC 61850 communications based distributed applications involve several different devices connected
to a substation local area network as shown in the simplified block diagram in Figure 9.
A Merging Unit (MU) will process the sensor inputs, generate the sampled values for the 3 phase
currents and voltages, format a communications message and multicast it on the substation LAN.

APS | 2015
Figure 10: Substation design with Station and Process bus
A binary input/output unit (IOU) can be used to monitor the status of the breaker and trip or close it when
necessary based on the GOOSE messages it receives from the different IEDs.
The merging unit and the input/output unit can be combined in a single device – a process interface unit
(PIU) as shown in Figure 9. The integration of sensor technology with digital interface is the next phase
in the development of digital substations as shown in Figure 10
All multifunctional IEDs will then receive sampled values messages and binary status messages, the
ones that have subscribed to this data then process the data (including re-sampling in most of the
cases), make a decision and operate by sending a GSE message to the IOU to trip the breaker or
perform any other required action.
This is an illustration of how the substation design changes when the full implementation of IEC 61850
takes place. All copper cables used for analog and binary signals exchange between devices are
replaced by communication messages over fiber. If the DC circuits between the substation battery and
the IEDs or breakers are put aside, the digital “copper-less” substation is a fact (Figure 10).
Further integration of SPACS functions can be achieved by the implementation of a centralized system
that will use the 61850 process interfaces defined in the standard and using MMS, GOOSE and SV
communications. The system architecture in this case includes several substation computers running a
real time operating system with function elements interfacing based on the object models defined in the
standard.
Such an approach will accomplish the ultimate functional integration that results in significant savings
and improvements in the reliability of the substation protection, automation and control system.

APS | 2015
Figure 11: Substation design with centralized SPACS
6 IEC 61850 Digital Substation Benefits

Digitization of substations offers some significant benefits at all stages of the design, engineering,
commissioning and maintenance.
The reduction in costs is due to several different factors and includes material costs, as well as significant
time savings:
• Reduced costs due to the replacement of hundreds or even thousands control cables with a limited
number of fiber optic cables
• Reduced costs due to the replacement of the wiring of hundreds of copper wires to the panels’
terminal blocks and then from the terminal blocks to the relay terminals with the plugging in of a
few pairs of fiber cables
• Reduced costs due to the requirements for testing of all hard wired interfaces versus the testing
of the GOOSE messages based on advanced software tools
The above savings apply to both Sampled Values and GOOSE based functions.
Process bus based applications offer some important advantages over conventional hard wired analog
circuits. The first very important one is the significant reduction in the cost of the system due to the fact
that multiple copper cables are replaced with a small number of fiber optic cables.
Using digital analog interfaces based on IEC 61850 sampled values also results in the practical
elimination of CT saturation because of the elimination of the current leads resistance. Traditionally the
CT knee-point voltage is a function of the resistance of the different components of the current circuit:
VK = f (RCT, RL, RRP)

APS | 2015
where:
VK = Required CT knee-point voltage (volts)
RCT = Resistance of the current transformer secondary winding (ohms)
RL = Resistance of a single lead from relay to current transformer (ohms)
RRP = Impedance of a relay phase current input
In some cases RL is multiplied by 2 and plays a key role in determining the CT requirements.
In this case the CT secondary is connected to the phase current inputs of the Merging Units and RL is
practically equal to zero. The knee-voltage then will be only dependent on
VK = f (RCT, RRP)
The impedance of the merging unit current inputs RRP is very small, thus resulting in the significant
reduction in the possibility for CT saturation and all associated with it protection issues.
Process bus based solutions also improve the safety of the substation by eliminating one of the main
safety related problems - an open current circuit condition. Since the only current circuit is between the
secondary of a current transformer and the input of the merging unit is located right next to it, the
probability for an open current circuit condition is very small. It becomes non-existent if optical current
sensors are used.
The digital substation also improves the flexibility of the protection, monitoring and control systems.
Since current circuits cannot be easily switched due to open circuit concerns, the application of bus
differential protection, as well as some backup protection schemes becomes more complicated. The
above is not an issue with process bus, because any changes will only require modifications in the
subscription of the protection IEDs receiving the sampled analogue values over IEC 61850 9-2.
Another benefit of using communications based process interfaces is the ability to continuously monitor
them based on the streaming or repeated messages, thus improving the availability and reliability of the
protection, automation and control system.
Last, but not least, is the opportunity to introduce new methods for testing based on process bus
simulating test equipment which can be remotely controlled in order to perform functional testing based
on the testing features defined in IEC 61850 Edition 2.
7 Conclusions
IEC 61850 is a communications standard that allows the development of new approaches for the design
of new substations and refurbishment of old ones. A new range of protection and control applications
results in significant benefits compared to conventional hard wired solutions.
It supports interoperability between devices from different manufacturers in the substation which is
required in order to improve the efficiency of microprocessor based relays applications and implement
new distributed functions.
Sampled Measured Values communicated from Merging Units to different protection devices connected
to the substation Process bus replace the copper wiring between the instrument transformers in the
substation yard and the IEDs, while GOOSE messages replace the hard-wiring between the switching
devices and the SPACS.
Such systems provide some significant advantages over conventional protection and control systems
used to perform the same functions in the substations:
• Improved safety
• Reduced wiring, installation, commissioning and maintenance costs
• Optimization possibilities in the design of the high voltage system in a substation
Easy adaptation to changing bus configuration in the substation and practical elimination of CT
saturation and open circuit.

APS | 2015
8 References
[1] INTERNATIONAL STANDARD IEC 61850-9-1, Communication networks and systems in
substations – Part 9-1: Specific Communication System Mapping (SCSM) – Sampled values over
serial unidirectional multidrop point-to-point link, First edition 2003-05
[2] INTERNATIONAL STANDARD IEC 61850-9-2, Communication networks and systems in
substations – Part 9-2: Specific Communication System Mapping (SCSM) – Sampled values over
ISO/IEC 8802-3, First edition 2003-05
[3] IEC 61850-9-2 LE: Implementation Guideline for Digital Interface to Instrument Transformers
Using IEC 61850-9-2, UCA International Users Group
[4] IEC 61850 9-3 Time Protocol Profile for Power Utility Automation
About the Author
Dr. Alexander Apostolov received MS degree in Electrical Engineering,

MS in Applied Mathematics and Ph.D. from the Technical University in
Sofia, Bulgaria. He has 40 years’ experience in power systems protection,
automation, control and communications.
He is presently Principal Engineer for OMICRON electronics in Los
Angeles, CA. He is IEEE Fellow and Member of the Power Systems
Relaying Committee and Substations C0 Subcommittee. He is past
Chairman of the Relay Communications Subcommittee, serves on many
IEEE PES Working Groups and is Chairman of Working Groups C2 “Role
of Protective Relaying in Smart Grid”.
He is member of IEC TC57 working groups 10, 17, 18 and 19 and Convenor of CIGRE WG B5.53 ”Test
Strategy for Protection, Automation and Control (PAC) functions in a full digital substation based on IEC
61850 applications” and member of several other CIGRE B5 working groups. He is Distinguished
Member of CIGRE.
He holds four patents and has authored and presented more than 400 technical papers.
He is IEEE Distinguished Lecturer and Adjunct Professor at the Department of Electrical Engineering,
Cape Peninsula University of Technology, Cape Town, South Africa.
He is Editor-in-Chief of PAC World.

APS | 2015

APS | 2015

APS
APS| 2015
| 2015
Impact of IEC61850 ed2 on Protection GOOSE
Ian Young | Schneider Electric

ian.young@schneider-electric.com
1 Introduction
GOOSE itself brings many benefits which can be utilised for various protection signals, however,
isolation and testing are critical considerations which are often seen as deficiencies in edition 1 (ed1)
of the IEC61850 standard. A key driver for edition 2 (ed2) was to overcome these limitations. This
paper looks at various isolation and test methods used for edition 1 and compares them to using
edition 2.
A major drawback of edition 1 was the inability to use online simulation. This limitation was not obvious
and care is needed to ensure simulated or test messages are not used on a live system. This paper
looks at the new simulation modes available in edition 2 which greatly improves this situation. The
edition 2 simulation mode is quite powerful and this paper also looks at how this functionality can be
used in a practical application.
One of the stated objectives of edition 2 is to create solutions that can incorporate edition 1 relays.
This paper also looks at the practicality of doing this.
2 User requirements
All protection schemes require testing. This is generally to confirm the engineer has correctly
understood the relay settings and implemented their requirements accordingly. Tests are undertaken
during commissioning and usually when significant setting changes are required during maintenance.
Particularly at brown field sites it may be necessary to commission a new feeder with other feeders in
service. It is therefore critical to be able to confirm operation of all functions including GOOSE without
risking tripping in service loads. This requires reliable isolation techniques.
From a functional point of view, if we consider a function like circuit breaker fail (CBF) we need to be
able to confirm the relay correctly issues a CBF trip signal to other relays. We then need to check the
other relays have seen that signal. We want the relay under test to be operating normally without
disabling functions to perform this test. We also want to be able to ensure that the relay under test will
operate if it receives a CBF from another relay.
From an operational point of view we need to ensure that the isolation is obvious and can be applied
by all isolation staff. Depending on isolation practice this may mean isolations are applied by staff of
limited relay knowledge. When multiple manufacturers are used, not only is interoperability required,
but a consistent method of isolation will reduce errors. If non IEC61850 devices are used then isolation
techniques should be as similar as possible when moving to the new technology. One of the great
benefits of IEC61850 is the supervision of information. This should be extended to include isolation
with any device being tested being reported to SCADA systems.
There is also a performance requirement for GOOSE including its speed and availability. When testing
relays it is important that the relays and networks are operating as close as possible to normal
conditions. This ensures test results will be consistent with in service performance.
3 IEC61850 ed1 Protection GOOSE issues

GOOSE can be used for many functions within a substation including protection. What is unique about
protection applications is that the function requires maintenance testing since in service operation is
rare. When testing functions like CBF it is not possible to do a full scheme test since it would take a full
busbar out of service. Protection GOOSE therefore require adequate test facilities.
Edition 1 of IEC61850 had several items referred to as Test. Specifically for GOOSE there was a
“Test” bit in the message header.

APS | 2015
Whilst this would appear to be the intended method to use when testing GOOSE it had several
limitations:
• Applies to whole GOOSE message
• Not always reported (visible in subscriber)
• Signal status often not available
• Can be difficult to initiate
• Initiation can affect relay operation
Overall the use of the test bit provided isolation ensuring the receiver did not respond, however, it did
not provide visibility both in test equipment and receiving devices. Added with the difficulties in
initiation it was seldom used.
The other “Test” bit available is in the quality flags. These also had limitations:
• Not always implemented in data model
• Not always sent in GOOSE datasets
• Can be difficult to initiate
• Receivers do not always segregate data and quality
Here the lack of consistency between manufacturers meant this approach could not be globally used.
The above limitations are with respect to the GOOSE being published by devices. Subscriptions also
require testing and edition 1 did not clarify how this could be done. In general most devices should
only see one GOOSE for each subscription, either a normal GOOSE or a test GOOSE. Operation
when both GOOSE are present may cause the receiver may operate unexpectedly. Therefore on line
simulation should be avoided.
4 Typical ed1 Isolation Solution

The most common method of isolation is to include an additional “Test Mode” bit in the GOOSE
message [1][2][3][4].
Figure 1: Publisher Test Mode logic
In the above Programmable Scheme Logic (PSL) we are sending a CBF signal to remote relays via
Virtual Output 1. We are also sending an additional Test Mode status via Virtual Output 2. Initiation is
through an opto input connected to a test link. When the link is closed the relay is in service so the
Test Mode status is the inverted opto status. Any signal within the PSL could be used to initiate Test
Mode such as front panel push buttons or local/remote controllable inputs. Also shown is an LED
indicator which provides local feedback to the operator regarding the Test Mode status.
The receiver now needs to check the test mode before operating for remote trips
Figure 2: Subscriber Test Mode logic

APS | 2015
An AND gate is used by the receiver to check the Test Mode status. Here the remote trips are directed
to a separate output relay which can be used to provide further traditional isolation if required. It could
also be connected to the normal relay trip contact or the trip conditioner. The connection to external
trips allows initiation of CBF as well as maintenance recording.
The receiver now knows not only the status of the CBF but also whether the sending device is in test
mode. By providing separate signals we can highlight or report whether remote devices are in test
mode giving further confidence during testing and operation. This means that during testing the test
set can check the IED is in Test Mode before executing tests like CBF. For operations staff the Test
Mode status can be verified prior to commencing work to ensure isolation, and before returning to
service to ensure no links are left open. This can be done remotely and an example SCADA interface
is shown below
Figure 3: Subscriber Test Mode logic
In the above example the operators can see that Feeder 4 is Out Of Service (OOS) whilst Feeder 3 is
currently in Test Mode (TM).
The above example shows virtual outputs (VO) being used for GOOSE data. Many users adopt this
method as engineers and testers are used to a PSL representation of relay inputs and outputs. It also
allows the GOOSE signals to be amended without changing the IEC61850 configuration. The
disadvantage of this approach is that the testers need to know configuration of the VOs. In our
example, a tester sniffing the GOOSE needs to know that VO1 is CBF and VO2 is Test Mode.
Figure 4: GOOSE Message with CBF and TM using Virtual Outputs
Alternatively CBF, opto or LED status could be directly mapped to the GOOSE dataset. The use of
object oriented LN data has the advantage of being self-describing. In this case the tester would
directly see CBF in the GOOSE message as RBRF.OpEx.general, however, they would still need to
know that LED8 (LedGGIO.Ind8.stVal) indicates Test Mode.
Figure 5: GOOSE Message with CBF and TM using Logical Node data

APS | 2015
5 Edition 1 Subscription testing

The isolation mentioned in section 4 above meets most user requirements and, since it is based on
PSL, the logic is controlled by the user allowing full flexibility. Unfortunately a similar solution does not
exist for testing subscriptions. For testing purposes the relay needs to be removed from the network.
Figure 6: Subscription Testing
In ed1 an IED can send either a normal GOOSE message or a test GOOSE message. There is no
allowance for both messages to exist on the network. The exact operation under these conditions is
not specified, however, the subscriber is likely to attempt to respond correctly to the most recent
message. For example, if the relay was subscribing to a CBF and the normal message was No Trip it
would not trip. If it then received a test message with Trip it would also not trip since the message has
the Test bit in the header. However, if for example a CB status was being sent which was currently
closed, then a test message was received which also had CB closed, it may drop back to CB open
depending upon the default state of this input. Other interlocks and blocking signals could also change
and therefore additional test GOOSE should not be published on an ed1 system.
There is a risk during testing of accidentally publishing either normal or test GOOSE into the
substation network. This is actually quite likely since most test sets are configured to continuously
publish even when tests are not running. This means a tester could perform an offline test with
simulations present and once tests have been completed, connect to the substation network to
complete online tests. If they do not turn off or disable the GOOSE before connecting then an
unexpected state change could take place. Steps should be taken to ensure this does not occur such
as
• Utilise different test set ports for subscription and normal testing
• Block unused switch ports
• Have dedicated egress only test ports
6 Edition 2 Improvements
Edition 2 has many improvements over edition 1. Broadly these improvements were in four main
areas:
1. Interoperability
Instantiated IED Description (IID) and System Exchange Description (SED) files created to
assist in the configuration of substations and the exchange of information between them
2. Configuration
Input References (InRef) and Blocking References (BlkRef) created to allow the direct
mapping of data to logical node functions

APS | 2015
3. Testing
Clarification, simulation and control mirroring functions added
4. Modelling extensions
Many Logical Nodes (LN) were added to allow more equipment to be descriptively modelled
For protection GOOSE the main benefits were in the testing area with clarification and simulation
being the important changes. These are covered in section 7.8 of IEC61850-7-1 ed2 [5] and give three
optional features that can be used to facilitate testing
1. Simulation GOOSE.
Subscribers can choose between using normal or simulated GOOSE
2. Test GOOSE inputs
InRefs can have 2 inputs a normal or test
3. Subscription Test Mode
GOOSE will be processed according to the test mode of the sending and receiving IEDs
As with all of IEC61850 the facilities provided will vary and so the actual functions implemented should
be checked to ensure they meet the user requirements. The first two features provide a similar
functionality, the ability to test subscriptions and it is unlikely that both would be required
simultaneously.
7 Edition 2 Isolation Solution

Edition 2 clarified the correct method of providing isolation is to use the Test Bit in the quality status for
the data. Therefore to utilise this technique the publishing devices need to publish the Quality in the
GOOSE message and update the Quality Test bit. Whilst it is possible for Test Mode to be applied per
logical node or per logical device, most devices currently place the whole device into Test Mode. This
has the effect of making the Quality of all data “Test” and aligns well with current testing practices. The
logic is shown below
Figure 7: Publisher Test Mode Logic ed2
The PSL used is virtually identical to ed1 with Virtual Output 2 being replaced by Test Mode. Test
Mode will set all data quality bits to “Test”. The other difference to ed1 is LED8 is no longer used since
the relay alarms for Test Mode automatically. Of course, an LED or other indication can be added via
the PSL if required. The receiver logic is also very similar
Figure 8: SubscriberTest Mode Logic ed2
Here the second virtual input has been replaced by the Quality. Since the quality is good when not in
Test Mode this signal is no longer inverted.
Since the methodology is the same as that used for ed1 there is little impact on testers or operational
staff. The main difference would be for someone sniffing a GOOSE message who would now see the
quality directly associated with the data. For users directly publishing CBF, for example, they would be
able to see RBRF.OpEx.general as the data and directly know whether it is in Test Mode from
RBRF.OpEx.q.

APS | 2015
Figure 9: GOOSE Message with CBF and quality using Logical Node data
The other consideration is that the test set must now be able to detect the Quality to be able to check
on the IED Test Mode status.
The logic can be extended to provide extended logic such as that shown in IEC61850-7-1 ed2 7.8.3.
Figure 10: Extended Subscription Test Mode Logic ed2
In this mode the relay will trip for a CBF if neither relay is in test mode or if both relays are in test
mode.
It should also be noted that implementing the above logic has a significant impact on testing time since
four tests are required to cover the combinations of Test Mode and Normal of both the sending and
receiving devices. Most users prefer to use the simpler logic in Figure 8 providing the remote relays
which are in service can confirm the Test Mode is being received and indicate when CBF trip was
received even though they will not trip.
8 Edition 2 Subscription testing

For most ed1 users the changes in isolation provide minimal benefit and some may choose to stay
with their existing ed1 methods. However, for subscription testing edition 2 has the significant benefit
of allowing Simulation GOOSE. This has been achieved by replacing the confusing ed1 Test bit in the
GOOSE header with the ed2 Simulation Bit and allowing both a simulation and normal GOOSE
message to be present on the same network.
Figure 11: Subscription Testing ed2

APS | 2015
The receiving relays are then set to respond to either the normal or simulated GOOSE message via a
single setting (Sim in the LPHD). However, for the receiving relay it is not simply a matter of accepting
normal or simulated data. Instead, the Simulation setting tells the receiver to start looking for simulated
GOOSE and when they become available use their data instead of normal data. Once a Simulated
Signal is found the device will switch to using the simulated data only for that subscription. It will
continue to use this data until simulation subscription mode is disabled. Other subscriptions will
continue to use the normal GOOSE.
Whilst this may seem complex it does provide benefits over switching all GOOSE to simulated. For
example consider a feeder with a bay controller and X and Y protection.
Figure 12: HV Switchgear with X&Y Protection and Bay Controllers
The GOOSE may be used to exchange various signals as shown below
Figure 13: X protection GOOSE subscriptions
If we are just testing setting changes for the X protection we may have the Y protection and Bay
Controller still in service and we would want to simulate all subscriptions. However, if we are wanting
to confirm the operation of the autoreclose scheme with the circuit out of service we would just want to
simulate the CBF from the other circuits and use the actual status from the Bay Controller to provide
real breaker timings. The simulation solution given in IEC61850-7-1 ed2 7.8.1 provides the flexibility to
cover any mix of actual and simulated signals.
It should be noted that the risks highlighted in section 5 now present an even greater risk since we are
injecting simulation messages onto the substation network. This means we are now totally reliant on
the test set being in simulation mode before connecting.
We also need to ensure our network design allows for the extra bandwidth created by simulation
messages.

APS | 2015
9 Protection GOOSE ed2 issues

It would be incorrect to assume that an IED supporting IEC61850 ed2 will meet the user requirements.
Similar to any other feature of IEC61850, the availability of specific features will depend on what has
been implemented by the manufacturer. When considering protection GOOSE some of the main
considerations are
• Quality includes Test Mode
• GOOSE includes Quality
• Test Mode can be initiated as required
• Simulation with simultaneous messages supported
• Data status visible for received test messages
The last point is important since GOOSE subscriptions are often shown only after processing the local
and remote test mode. If one device is in test and the other is not, then a change of status cannot be
logged by the receiver.
10 Edition Interoperability
To provide interoperability the GOOSE message structure is identical between ed1 and ed 2 with the
following exceptions
• Optional security
• Longer Logical Device Names
• Fixed-length GOOSE
• Simulation
So assuming that security and fixed length GOOSE are not used then the normal GOOSE messages
are interoperable providing the device naming is restricted to ed1 lengths.
As mentioned in section 8 the Simulation Bit replaces the Test Bit in the GOOSE, header. There is
also an extra bit in the reserved section. This means that Simulation GOOSE would likely be read
correctly by ed1 devices, however, such messages should not be used simultaneously on systems
with ed1 relays as detailed in section 5. The other consideration would be for users of the GOOSE
header Test bit. As mentioned in section 3 this is rarely done, however, if it is then ed2 relays would be
unlikely to provide this function.
The real interoperability challenge is with regards to configuration. Edition 1 relays will reject ed2 SCL
files since they will not understand all the contents. Similarly ed2 relays may not directly accept ed1
SCL files. Today the only solution is to manually configure the GOOSE subscriptions which is quite
time consuming and likely to introduce errors. Looking ahead system configurators will need to
produce two versions of SCL files if ed1 and ed2 are both used at the same site. Ideally these
configurators would also need to check the ed2 configuration is compatible with ed1 devices.
A more practical solution today is to use only ed1 IEC61850 when all devices do not support ed2.
Ideally the relays installed should support both versions allowing the upgrade of the substation to ed2
when the ed1 devices are eventually replaced.
11 Conclusions
Protection GOOSE requires special consideration with regards to in service testing. The exact method
of achieving this was not clearly specified in ed 1, however, most users adopted a fairly similar logic
based approach. The testing methods were more clearly specified in ed2 and it aligns well with these
existing techniques.
The main advantage of ed2 is the simultaneous support of simulation and actual GOOSE. This allows
full scheme in service testing of devices mixing both actual and test set GOOSE.
Interoperability of ed1 and ed2 is possible but is quite difficult practically and will likely be avoided by
most users.
Overall the suitability of an IED, for edition 1 or edition 2, depends upon its implemented features and
its user interface to allow these features to be effectively used.

APS | 2015
12 References
[1] Practical Application of GOOSE Trip Circuit Isolation, Daniel Vidler et al. SEAPAC 2013
[2] GOOSE considerations when used in protection applications, Ian Young et al. SEAPAC 2013
[3] Testing of IEC61850 Process Bus Based Substation, MARTIN, D. et al. PS1 B5 Colloquium
2013
[4] Testing of a Digital Substation Based on IEC 61850, A.N. PODSHIVALIN et al. PS1 B5
Colloquium 2013
[5] IEC 61850-7-1 Ed.2:2011 Standard
About the Author
Ian Young is the Applications Manager for Automation Solutions within the
Energy Business of Schneider Electric, Australia providing application
support, training, commissioning services and IEC61850 engineering to
customers in the Pacific Region. He is the current Australian representative
on CIGRE study committee B5. He has presented at SEAPAC, Machines
and APTUG on protection topics ranging from motor protection to
substation network testing. He has been one of the key presenters at the
Australian APPS course for over 20 years.

APS | 2015

APS
APS| 2015
| 2015
Case Study: Implementation of IEC61850 in Generator and

Transformer Protection System in NSW Power Stations
Usman Mahmood | Energy Australia NSW

Usman.mahmood@energyaustraliansw.com.au
Sy Bui | Aurecon
Sy.bui@aurecongroup.com
1 Abstract
Reliability, Efficiency and Safety are the key objectives of any modern power system and these can be
achieved by the implementation of fully automated systems. The IEC61850 standard has
revolutionized the automation process by standardizing system specifications, configuration language,
naming convention, communication protocol, and conformance testing. Use of IEC61850 compliant
devices can simplify the engineering design process, increase flexibility and reduce engineering costs.
In order to operate and maintain IEC61850 based systems, plant engineers must acquire knowledge
about communication protocols, computer network, classes, and objects etc.
In New South Wales (NSW), the protection systems of the large turbo-generators (660MW units and
above) and associated generator transformers have reached the end of their operating life. Most NSW
generating companies have already implemented a program of replacing these generator and
generator transformer protection systems. The new protection systems used IEC61850 compliant
protection relays with an optical, peer to peer communication messaging system known as “GOOSE”.
Being a protection retrofitting project to existing generating units, there were a number of difficult
choices and design consideration in selecting the levels of IEC61850 implementation to these new
protection systems. This paper looks at the extent of IEC61850 implementation in these retrofitted
generator and generator transformer protection systems in NSW. It also provides a discussion on the
impacts of the IEC61850 system on the operation and maintenance of these protection systems and
possible future improvements.
2 Introduction
The major coal fired thermal power plants in NSW were built and commissioned in 1980’s. The original
generator and transformer protection systems installed in these stations were comprised of two
modular systems, designated as X and Y Protection. Each protection system had an electronic
modular type relay system comprised of a number of relay modules working into a tripping matrix
which were interfaced with the switchgear and other control systems via electro-mechanical multi-trip
relays and hardwired connections.
The X and Y protection systems were from different manufacturers and were designed to be “stand-
alone” systems to trip either the “transformer zone” or the “generator zone”. The transformer zone
protects equipment located on the transformer side of the generator circuit breaker. The generator
zone protects the equipment on the generator side of the generator circuit breaker. All control, tripping
and alarm signals between relays and external devices were hardwired connections with isolations via
isolating links.
As the X and Y protection systems reached the end of their economic life, these protection systems
were replaced with new intelligent electronic devices (IEDs) compliant to IEC61850 standard. While
this might present an excellent opportunity to make a radical replacement of the conventional
hardwired connections with an optical, peer to peer communication messaging system, there were
certain constraints to this concept due to the difficulty in interfacing with existing control systems and
electrical equipment such as the Current Transformers (CTs), Voltage Transformers (VTs) and
Switchgear. This paper looks at the use of IEC61850 in these retrofit projects, discusses various pros
and cons and recommends possible future improvements.

APS | 2015
3 System Design
Table 1 below shows typical relay models which have been used for the new numerical “X” and “Y”
generator protection systems:
Protection Description Relay Model

System Manufacturer
X 776MVA Generator Protection Relay GE Multilin G60

330/23kV Generator Transformer Protection Relay GE Multilin T60
23/11kV Auxiliary Transformer A Protection Relay Siemens 7UM6211
23/11kV Auxiliary Transformer B Protection Relay Siemens 7SJ6411
23/1kV Excitation Transformer Protection Relay Siemens 7UM6215
Y 776MVA Generator Protection Relay Siemens 7UM6225
330/23kV Generator Transformer Protection Relay Siemens 7UT6351
23/11kV Auxiliary Transformer A Protection Relay Siemens 7SJ6411
23/11kV Auxiliary Transformer B Protection Relay Siemens 7UM6211
Table 1: The new X and Y Protection systems at Mt. Piper Power Station
The original concept of reliability and redundancy between the ‘X’ and ‘Y’ protection systems was
maintained with the following considerations:
The IEDs for the “X” protection system with respect to the main protected equipment (generator
and generator transformer) were from different manufacturers to the corresponding IEDs for the
“Y” protection system. This is to ensure that any generic design fault due to one manufacturer is
confined to only one relay system.
For the auxiliary transformers (A or B), even though the IEDs are from the same manufacturer
(Siemens), these IEDs operate on different operating principles and design platforms.
All existing protection functions were retained and duplicated where possible with the availability
of current transformer (CT) and voltage transformer (VT) inputs.
Each IED within the ‘X’ and ‘Y’ protection system is a stand-alone relay, supporting the IEC61850
standard for substation automation and has an adequate number of inputs and outputs such that
taking any one relay out of service will not affect the operation of any other relay.
All the IEDs within each protection system connects to a ‘Ruggedcom’ Ethernet switch, which
communicates with the other Ethernet switches of other protection systems and the unit control
system, to form a closed ring network, which allows remote transfer of real time process variables,
protection settings, fault records and trip log events to the SCADA system.
Figure 1 on the next page shows a block diagram of the communication network between the IEDs
and the control system. Figure 2 is a typical communication topology which has been implemented for
Unit 1 protection system at Mt. Piper Power Station.
An IEC61850 station was created for each protection system (X and Y) and each IED was uniquely
identified with an IP address. All IEDs also have a time stamping capability for event and fault
recording. The clock for time stamping can be achieved in a number of ways; it can be via the global
positioning system (GPS) for some power generating plants, or from the central clock at the unit
control system via interface modules for other generating plants. The choice of clock system depends
on the compatibility between the unit control system and the IEC61850 station for generator protection
systems.
As these replacements have to be completed within fixed and planned unit outages, one of the key
criteria is to maintain all existing interfaces with the external devices such as the unit control system,
switchgear, automatic excitation system, transformer mechanical protection devices (Buchholz, Over

APS | 2015
Pressure devices, etc.). To achieve this criterion, a hardwired tripping arrangement via fast operating
multi-trip relays is used for both “X and ‘Y’ protection systems.
Figure 1: Communication Block Diagrams Between IEDs and the Control System

APS | 2015
Figure 2: Connection of IEDs for Unit 1 X & Y Protection at Mt. Piper Power Station
4 Discussion
4.1. Inter-operability between different relay manufacturers
While different relay manufacturers were used in the ‘X’ protection system, no problem were
encountered in integrating ICD files from the different relay vendors. Each vendor’s private tools export
IED capability description data in to standard format ICD files. The System Configuration Tool
integrates ICD files from different vendors into a common product as an SCD file.
Figure 3: IEC61850 File Standards [1]
4.2. Level of IEC61850 Utilization

The biggest advantage of IEC61850 based design is that it eliminates or significantly reduces wiring
between the relays, field devices such as breakers and control centers. Conventional wiring is
replaced with a communication infrastructure fulfilling the protection and control applications by
exchanging IEC61850 GOOSE messages over Ethernet.

APS | 2015
For these retrofitting projects, the allowable duration of plant outages imposed a significant constraint,
dictating the extent of modification work that could be done to relevant field devices to make them
compatible with the IEC 61850 platform. Because the generator protection systems are vital protection
systems in any power plant, any delay in completion of these retrofitting projects would undoubtedly
cause a flow-on effect in returning the generating units to service, which would have significant
impacts to the revenue side of these power plants.
There were no current commercially available merging units, which could be verified as compatible
and reliable to be used with these critical external devices. The generator protection systems were
essential in protecting the generators and generator transformers. Hence, there was a high level
concern in using any product which have not been proven in previous applications and projects.
In addition to the above constraints, the generator protection systems have to interface with a large
number of external devices such as the unit control system, switchgear, automatic excitation system,
and transformer mechanical protection devices (Bucholtz, Over Pressure devices, etc.). These
external devices were manufactured by different vendors in past years which makes more difficult in
interfacing with the IEC 61850 standard.
For the above reasons, it was decided to maintain hard-wired connections to most of these external
devices via binary inputs and fast operating multi-trip relays for both of “X and ‘Y’ protection systems.
For the interface with the unit control system, the level of utilization of IEC61850 was varied between
different power plants. In some power plants, all measurement variables, alarm and trip signals were
fully implemented in IEC61850 and utilized by the unit control systems. GOOSE messaging services
between relays were also implemented. In other power plants, there was no exchange of IEC61850
information between the relays and the unit control system. The level of utilization of IEC61850
depends on the compatibility between the vendors of the unit control system and the generator
protection systems. Where it is the same vendor then there is high probability that the IEC61850 will
be implemented for interfacing between the protection system and the unit control system.
The relay operating modes, setting values, fault records, event records and time stamps were also
available in web pages to engineering desks away from the generator protection panels. This
particular feature was also implemented in varying levels of complexity between power plants. Even
though the information was available in the “read’ mode only, there was a high level of concern about
the confidentiality and security of these types of information.
Notwithstanding the above, there was some commonality in the GOOSE messaging between relays
over fiber optical cables. Table 2 below shows a list of GOOSE messages configured between relays
over fiber optic cables:
GOOSE signal From Relay In panel To Relay In panel

Over-excitation V/f>> No1 Gen Diff G60 H01 No1 Gen Tx Diff T60 H01
Over-excitation V/f>> No1 Gen Diff G60 H01 No2 Gen Diff 7UM62 H03
Over-excitation V/f> No1 Gen Diff G60 H01 No2 Gen Diff 7UM62 H03
Neutral displacement No1 Gen Diff G60 H01 No2 Gen Diff 7UM62 H03
Neutral displacement No1 Gen Diff G60 H01 No2 Gen Diff 7UM62 H03
Under/Over Voltage Aux Tx A Diff 87T H02 No1 Gen Diff G60 H01
Under/Over Frequency Aux Tx A Diff 87T H02 No1 Gen Diff G60 H01
Generator inter-turn No2 Gen Diff 7UM62 H03 No1 Gen Diff G60 H01
Generator EF > Aux Tx A OC 7SJ64 H04 No2 Gen Diff 7UM62 H03
Table 2: GOOSE messages between relays via IEC61850 protocol

APS | 2015
5 Conclusions
Most original generator protection systems are reaching the end of their economic life and have to be
replaced by modern digital protection relays. While the IEC61850 platform can be specified and
incorporated into new generator protection systems, there are a number of pros and cons in the
implementation of the IEC61850 platform into this critical protection system of existing power plant.
The pros and cons can be summarized as follows:
There was a high level of compatibility in IEC61850 between different vendors. This would make
easier to achieve the essential requirement of different manufacturers for critical protective
functions.
The GOOSE messaging between relays could be fully utilized and beneficial in reducing the inter-
panel wirings. This would reduce the panel wirings which would have benefits in the cost and
construction time.
The level of GOOSE messaging between the generator protection systems and the unit control
systems depended on the compatibility between vendors for the unit control systems and the
generator protection systems. Where it is the same vendor for both systems, significant benefits
can be gained in reducing the wiring connections between these two systems.
There was not a great reduction in hard-wired connections to external devices such as the
switchgear, automatic excitation system, and transformer mechanical protection devices
(Bucholtz, Over Pressure devices, etc.). It is not expected that this shortcoming can be overcome
as there are no proven, reliable and compatible merging units which can be used for interfacing
with large electrical plants.
Notwithstanding the above points, the implementation of the IEC61850 standard in the generator
protection systems has achieved some levels of efficiency and cost reduction in retrofitting protection
projects.
6 References
[1] Wester, C. Baigent, D. Holbach, J. Roderiduez, J. Frisk, L. kunsman, S. Hossenlopp, L "First
IEC61850 Multivendor Project in the USA, PAC Magzine, AUTUMN 2007
[2] Mt Piper generator protection systems – units 1 and 2
[3] Eraring generator protection systems– units 1 to 4
[4] Vales Point generator protection systems– units 5 and 6

APS | 2015
About the Authors
Usman Mahmood received his Bachelor of Engineering (Electrical)

Degree from University of Technology, Sydney in 2007 and Master of
Engineering (Power Generation) degree from Queensland University of
Technology, Brisbane in 2012. He is a Chartered Professional Engineer
and has been working with Energy Australia NSW formerly Delta Electricity
Western since 2007. He has been working in Control, Instrument, and
Electrical group at Wallerawang and Mount Piper Power Station. Since
beginning of his career, he has spent considerable time in the Asset
Management of Generator Protection Systems, Excitation Systems, Fault
Recorders, DC Systems, Switchboards and Switchgear. He has been
proactive in implementation of condition monitoring and integrated
maintenance strategies. He presently holds a Plant Owner Position in
Asset Management Group of Energy Australia NSW.
Sy Bui received his Bachelor of Engineering (Electrical and Electronic)

Degree from University of Canterbury, New Zealand in 1984. He is a
Principal Electrical Engineer with Aurecon since 2003. He has previously
worked for the Electricity Commission of New Zealand and the Electricity
Commission of New South Wales, then Pacific Power International prior to
joining Aurecon in 2003. He has been working in transmission and
distribution group and power generation group, specializing in power
system, switchgear and protection. He has spent considerable time in the
design, testing and commissioning of protection systems. He presently
holds an Associate position in the Energy group at Aurecon.

APS | 2015

APS
APS| 2015
| 2015
COMTRADE Analysis to Visualise Protection Operations
Greg Finlayson | Schneider Electric

greg.finlayson@schneider-electric.com
Mitchell Eadie | Schneider Electric

mitchell.eadie@schneider-electric.com
1 Introduction
This paper discusses methods of analysing waveform recordings from numerical relays to diagnose
power system events and confirm relay operations. COMTRADE format recordings often only capture
the physical analogue and digital signals wired to the relay, while internal protection functions will use
algorithms to manipulate these signals before determining a trip condition.
With a basic understanding of the relay algorithms the operating quantity can be derived from the
recording and compared to an operating characteristic to verify settings and confirm relay operations.
Examples presented include analysis of harmonic content of inrush waveforms, trip locus of differential
and impedance based characteristics, high impedance based protection trips and fault finding
installation and application issues.
2 COMTRADE files
Numerical protection relays capture an extensive amount of data during a system disturbance. This
data is crucial to engineers to analyse the fault to determine:
• Sequence of events e.g circuit breaker tripping sequence and root cause analysis;
• Type of fault and/or fault evolution;
• Correct operation of protection;
COMTRADE (Common format for Transient Data Exchange for power systems) is a file format defined
by C37.111 for storing waveform and status data related to transient power system disturbances [1],
[2]. Files are generated by Intelligent Electronic Devices (IEDs) which are configured to record
analogue and digital signals when triggered by disturbances in the power system. Such data can be
used to enhance and automate the analysis, testing, evaluation, and simulation of power systems and
related protection schemes during fault and disturbance conditions [1], [2]. Analysis tools such as
Transview [3] and Wavewin [4] help to visualise and manipulate the information (waveforms,
harmonics and configured digital channels etc) in the COMTRADE files in a variety of ways; enabling
more efficient fault analysis to be performed.
3 Analysis and methodologies

3.1 Harmonic content of inrush waveforms
A straightforward example of COMTRADE file analysis is that of viewing the harmonic content of a
waveform, often to confirm a trip upon transformer energisation is due to inrush and not a genuine
fault. Figure 1 depicts a typical waveform for transformer inrush with the harmonic components of each
phase shown in the bar graph as a percentage of the fundamental. A typical restraint threshold setting
for differential protection based on second harmonic detection is ~15-20% and can be configured to
block single phase or cross block all three phases. Most COMTRADE viewers have this basic
functionality as a standard feature.

APS | 2015
Figure 1: Harmonic content analysis of an inrush waveform
3.2 Reverse power diagnosis

Figure 2 shows the disturbance record of a reverse power trip for a system with embedded generation.
The relay is configured with reverse power protection, set to trip when the current lags the voltage by
more than 90 degrees. It can be seen in the recorded file that shortly before the trip is issued, the A-
phase current lags it's voltage vector Van by 94.8°. It only occurs for 1-2 cycles during which time the
relay issues a trip to the embedded generation. The current then returns to the same pre-fault
magnitude and phase angle after about 200ms. Note the current waveforms are still present following
the trip showing that the site did not lose power.
Figure 2: Disturbance record of a reverse power trip
This behaviour is indicative of a power swing which is a variation in power flow that occurs when the
internal voltages of generators at different points of the power system slip relative to each other [8]. In
this case the system remains stable, and the reverse power element initiating the trip should be
reviewed as it would typically be applied with a definite time delay to prevent spurious operation during
transient power swings.
3.3 Line differential locus analysis

Another useful application that COMTRADE files enable is the manipulation of fault information to
convey the trip decision process of a relay. This is particularly useful for line differential schemes to

APS | 2015
visualise the differential locus evolution during a fault. Figure 3 shows the differential and bias current
characteristic for a line differential relay with equations 1 & 2 outlining the relay differential algorithm
trip conditions [6].
Figure 3: Bias characteristic implemented by the line differential relay [6]
For I bias < I s 2

I diff > k1 I bias + I s1 (1)
For I bias > I s 2
I diff > k 2 I bias − (k 2 − k1 )I s 2 + I s1 (2)
Where:
Is1 is the minimum set point of the differential;
k1 is the lower percentage bias slope, providing stability for CT errors;
Is2 is the setting above which the higher percentage bias k2 is used;
k2 is the higher percentage bias slope used to provide stability under heavy through faults.
Figure 4: Combined disturbance record for line differential scheme mal-operating during transformer inrush

APS | 2015
Figure 4 shows the disturbance record from both line ends for C-phase, recorded when the line
differential scheme tripped during a downstream transformer energisation. The current magnitudes
from either side of the differential scheme were manipulated to calculate the differential and bias
currents. Figure 5 shows the evolution of the differential locus of C phase during a transformer
energisation plotted against the bias characteristic. It can be seen that the scheme has operated as
the differential current exceeds the threshold of the characteristic. Further analysis of the CTs installed
at either end of the scheme showed that the scheme tripped on differential due to the saturation of
incorrectly sized CTs at one end of the scheme [5].
Figure 5: Differential locus during transformer energisation
Further to this, it can also be a useful exercise to see whether the relay would have tripped under the
same scenario if the transient biasing algorithm was enabled in the line differential relays. This is
shown in Figure 6, where the characteristic threshold is dynamically modified depending on the bias
current. The locus and transient bias evolution show that this trip would still occur even when the
transient biasing function is enabled. This was confirmed via bench testing, in which the COMTRADE
is played back to the relay scheme.
Figure 6: Differential locus with transient bias

APS | 2015
3.4 Transformer differential fault locus analysis

Plotting the fault locus of a transformer differential relay from the recorded current waveforms can also
be performed by calculating the restraint and differential values from the waveforms captured. As
these derived quantities are not recorded channels in the relay, the COMTRADE viewer was used to
calculate the values based on the algorithm described in the product manual.
An example for a relay under test is shown here. It is configured for protection of a 6.3MVA Dyn1,
66/11kV transformer and an internal B-C phase fault shot is applied from the secondary injection test
set. A restraint current of 5 Inom and a differential of 4 Inom is applied in the test set software, with the
resultant waveform captured by the relay for this test case shown in Figure 7.
Figure 7: Differential test shot and recorded waveforms
A number of steps are required to calculate the restraint and differential values, and were performed in
Wavewin using Software Analogue Channels. These allow the user to enter formulae to manipulate
existing channels and generate new channels with the derived quantity.
In summary, the following steps are required:

1. Filter harmonics from recorded channels
2. Divide by CT secondary to obtain per unit values (Divide LV end by 5 in this case)
3. Multiply by amplitude matching factor setting
4. Subtract zero sequence current (from LV end as this is earthed in zone)
5. Phase shift correction
6. Sum HV and LV to calculate restraint and differential currents
Firstly, the recorded channels are filtered to derive the fundamental 50Hz component of each
waveform. This removes the harmonics and DC component and is commonly the operating waveform
input into the protection functions of a relay. In the COMTRADE viewer this is performed by selecting
the channel and using the “h” operator. See calculation in Figure 8.
Figure 8: Deriving the fundamental of current waveforms
The CTs on the HV side are 200/1A, so the recorded quantities do not need to be modified. The LV
side however has 600/5A CTs, so the currents on this side are divided by 5, as shown in Figure 9.

APS | 2015
Figure 9: 5A inputs changed to per unit
Amplitude matching is performed on each set of current inputs to equalise the nominal load current to
the installed CT ratios, as this allows free selection of CTs in numerical relays. In this case, the HV
side is multiplied by a constant factor of 3.360, while the LV side is multiplied by 1.815, as shown in
Figure 10.
Figure 10: Amplitude matching
Zero sequence filtering removes the zero sequence current component of an input where an in zone
earthing point allows residual current to flow out of zone. In this case, as the LV side is earthed at the
neutral star point, filtering is performed by the relay to prevent maloperation of the differential
protection for external earth faults. Referring to the product manual, this is performed by subtracting
the lagging phase and dividing by √3, as shown in Figure 11.
Figure 11: Zero sequence filtering
For Dyn1 transformers, zero sequence filtering also corrects the 30° phase shift, so this step does not
have to be performed as a separate calculation.
Finally, the HV and LV currents can be used to determine the restraint and differential values
according to the formulae:
1
I restrain = I HV − I LV (3)
2
I diff = I HV + I LV (4)
These are shown in Figure 12.

APS | 2015
Figure 12: Calculating restraint and differential currents
The quantities are displayed graphically in Figure 13.
Figure 13: Calculated restrain and differential currents for test shot
Plotting the restraint and differential current against the bias characteristic shown in Figure 14 gives
the fault locus for a typical test shot. As this point matches that set in the test plan it confirms the
methodology and steps for calculating the restrain and differential current is correct for the transformer
in this application. This approach can then be used to analyse any future operations and easily
adapted for other transformer installations.
Figure 14: Locus of differential test shot

APS | 2015
3.5 Impedance protection locus analysis

A number of faults occurred on a section of overhead railway conductor, protected by a distance relay.
Faults on overhead conductors are not unusual due to their proximity to the ground, support structures
and movement due to contact with rolling stock. Two parallel sections of track are protected by a Zone
1 element intended to cover approximately 80% of the immediate track section with Zones 2 and 3
providing a time delayed protection for the remainder of line and a backup for the parallel track. A
small number of faults resulted in relays on both lines indicating a Zone 1 trip.
The relay measures current and voltage in both the catenary and feeder circuits and summates the
two circuit measurements internally, as shown in Figure 15. This results in a single current and voltage
value, which is used to determine the impedance of the track section.
Figure 15: Relay measurement of system voltage and current
A waveform recorded from one of the trip instances is shown in Figure 16. The channels labelled Icat-
Ifdr and Vcat-Vfdr have been derived in the COMTRADE viewer, from which the impedance Zmag has
then been calculated.
Figure 16: COMTRADE of fault with derived V, I & Z

APS | 2015
Exporting this data to Excel allows the impedance magnitude and phase angle between voltage and
current to be plotted on a graph to clearly indicate the fault location and locus against the set
characteristic for the distance protection zones. This trip event is shown in Figure 17.
Figure 17: Locus of distance Z1 trip
Review of all the trip occurrences and the original setting data highlighted a misinterpretation of the
line impedances used to determine the reach settings. The line impedances were given as Ω/km for
the complete parallel line loop, however the reach calculations used the value as line section lengths
resulting in reaches set to twice the intended value.
The settings were revised and the fault locations are plotted against the original and revised settings in
Figure 18. This shows a number of faults that originally operated in Zone 1, are now in Zone 2 and
similarly a few faults in Zone 2 now move out into Zone 3.
Figure 18: Original reach settings (left) and revised settings (right)

APS | 2015
3.6 Merging event logs into COMTRADE files

To better visualise recorded fault data a tool was developed to import event log entries into the
COMTRADE file. Relay event logs are an SOE record with time stamps recorded by the IED; alarms,
protection start and trip signals etc. Both the COMTRADE file and event log must be reviewed
simultaneously for an effective fault analysis, so combining these into one view aids in analysis.
An example of where the incorporation of event logs to the COMTRADE file has proven useful was for
a restricted earth fault protection mal-operation during an out of zone fault. The original COMTRADE
file and combined COMTRADE and relay event log files are shown in Figure 19 & Figure 20
respectively. The additional digital channels from the event log allows the engineer to more clearly
visualise the SOE against the waveforms at the time of operation. For example, the In>1 start, the
REF trip and the intertrip signal to the upstream transformer feeder via the IM64 output 1 signal.
Figure 19: Original COMTRADE file recorded for REF mal-operation
Figure 20: COMTRADE incorporating event log displaying additional signals
The root cause of the REF mal-operation was identified from the waveform in the IN sensitive current
input of the relay. High impedance differential scheme theory [7], explains that during an external fault
no current should flow through the current input of the relay, Figure 21 (top). The recorded waveform
APS | 2015
for an external fault confirms that the scheme is incorrectly installed, Figure 21 (bottom). An on-site
investigation quickly confirmed this.
Another interesting point to note is that as a consequence of the stabilizing resistor the CTs go into
saturation during a fault [7]. This was not seen in the recorded waveforms, further reinforcing that the
wiring on site was not correct.
Figure 21: Correct REF scheme (top) and incorrectly installed scheme (bottom)
4 Conclusions
This paper has presented the benefits of utilising COMTRADE files for effective fault analysis through
a variety of real world examples. The extensive amount of data stored within these files enables
almost endless investigation topics. Manipulation of the fault data from the COMTRADE files in the
form of waveforms and harmonics allow for the justification and/or confirmation of IED settings and
also trip decisions. Additional tools to combine information from various sources, such as SOE log files
into COMTRADE files enables one to better visualise and understand all the necessary points and
signals during fault analysis.
5 References
[1] Power Systems Relay Committee of the IEEE Power Engineering Society, IEEE Standard for
Common Format for Event Data Exchange ( COMFEDE ) for Power Systems. 1999.
[2] Working Group H-5 of the Relaying Channels Subcommittee of the IEEE Power System
Relaying Committee, “COMTRADE: A new standard dor common format for transient data
exchange,” IEEE Trans. Electron Devices, vol. 7, no. 4, p. 1920, 1992.
[3] Omicron, “TransView - Software for visualization and analysis of recorded signals,” 2015. [
Online]. Available: https://www.omicron.at/en/products/all/secondary-testing-
calibration/transview/.
[4] Schneider Electric, Wavewin User’s Guide. 2008.
[5] M. Eadie, “CT Requirements for Line Differential Schemes,” in APTUG, 2014.
[6] Schneider Electric, MiCOM P521 v13 Manual. 2012.

APS | 2015
[7] I. Young, “High Impedance Disturbance Records High Impedance circulating current scheme,”
in SEAPAC, 2015.
[8] Schneider Electric, Network Protection & Automation Guide, 2014.
About the Authors
Greg Finlayson is a Product Application Engineer for Automation

Solutions within the Energy Business of Schneider Electric, Australia. He
has over 15 years experience with roles in Engineering Systems,
Protection & Control and Substation Automation providing application
support, training, commissioning services and IEC61850 engineering to
customers in Australia and New Zealand.
Mitchell Eadie is a Product Application Engineer for Automation Solutions

within the Energy Business of Schneider Electric, Australia. He began his
career in Schneider Electric as a Graduate Engineer in 2012 and rotated
throughout the company over the next two years. He has experience in MV
project engineering and found his niche with protection relay and on-site
commissioning where he currently works as an application engineer for
Australian and New Zealand customers.

APS | 2015

APS | 2015

APS
APS| 2015
| 2015
Verification of Distance Protection Scheme of double circuit

transmission line using power disturbance records
Wing Chan | Power and Water Corp.

Wing.Chan@powerwater.com.au
Si-Dieu Tran | Power and Water Corp.

Si-Dieu.Tran@powerwater.com.au
Sameep Gharti Chhetri | Power and Water Corp.

Sameep.GhartiChhetri@powerwater.com.au
1 Introduction
The Channel Island Power Station is the major generation source for the Darwin-Katherine region.
Power delivery to the greater Darwin region is via two 132kV overhead lines to Hudson Creek
Substation which are installed on separate towers sharing a common corridor. Figure 1 below is a
simplified network diagram used in the model to study the impacts of the influencing factors including:
source impedance ratio, load-flow, mutual coupling, fault impedance, single pole tripping, auto-
reclosing and current reversal. The first part of this paper describes the selection of the proposed
replacement system. The second part describes the evaluation of the combined effects of the
influencing factors and the last part describes verification using available power system disturbance
records and the test scenarios generated by RelaySimTest.
2 Proposed Protection Scheme

The existing protection for Line 1 and Line 2 is provided by X and Y distance protection relays with a
Reverse Blocking scheme with single pole tripping and auto-reclosing for single phase to ground fault.
The replacement is based on the standardized protection scheme consisting of a distance protection
relay plus a line differential relay with distance protection. The communication assisted scheme for
the distance protection relay is to be implemented by hardwiring to the Dewar DM1200 Teleprotection
units which interface with the SDH communication system.
Zone 2 (Relay 4)
Zone 1 (Relay 4)
Zone 2 (Relay 2)
Zone 1 (Relay 2)
1 2 3 4 5
d 1-d
Channel Island Hudson Creek
Line 2
Vrelay2
Vrelay4
Irelay2 Irelay4 12
d· Zm 12
(1-d)·
Zm12
8 6
E E
Line 1
7
Vrelay1
Vrelay3
Irelay1 Irelay3
Zload
Zm23
Darwin
9 10 E Metropolitan
11
Vrelay5
Irelay5 Line 3
Vrelay6
Irelay6
Figure 1: Simplified Network Diagram

APS | 2015
2.1 Selection of the Communication Assisted Scheme

2.1.1 N-1 Critiria as applied to Protection System
In selecting the new protection scheme, the N-1 criteria is re-examined. There is lack of a
common definition of the ‘N-1’ security criterion [1]. Some definitions of N-1 criteria are:
• UCTE (OH-P3A) defines the ‘N-1’ security criterion as any probable single event
leading to a loss of power system element and that should not endanger the security
of interconnected operation (either cascade tripping or loss of load).
• … the system to be able to withstand the loss of any single component … is said to be
“N-1 secure” …[2]
Protection system is designed to clear any probable fault in the protected zone and should
discriminate from any probable fault outside the protected zone. Therefore an N-1 secure
protection system must be able to provide the fault clearance and discrimination functions in
case of any probable single event or loss of any single component.
2.1.2 Breaker and a Half Configuration
Breaker and a half configuration is

used at both Channel Island and
Hudson Creek substation. The CTs
are externally summated to obtain
the currents flowing into the lines
between Channel Island and Hudson
Creek. It is recognised [3] [4] that
saturation in one of the CTs during a
close in reverse fault will increase
the probability of incorrect detection
of the fault direction in distance
protection. This is one of the Figure 2: Breaker and a Half Configuration (extracted from
significant factors in the selection of GE – Transmission Line Protection Principles – Fig. 3)
the communication aided distance
protection scheme.
2.1.3 Selection of Communication Assisted Scheme
The existing Reverse Blocking scheme is not supervised. If the communication of the healthy
line has failed, then the blocking signal will not be received. In case of a fault close to the end
of one line, the distance relay of the healthy line will see the fault as a Zone 2 fault. As no
blocking signal is received, the healthy line will also trip in accelerated Zone 2 time before the
fault can be interrupted by the circuit breakers of the faulted line. An un-supervised Reverse
Blocking scheme is obviously not N-1 secure. This problem can be corrected by supervising
the communication which blocks the accelerated Zone 2 tripping. The other issue for Reverse
Blocking scheme is that the zero sequence current in the healthy line could be too low to
declare the fault is in the reverse direction and may cause false tripping for a fault in the
adjacent line. [5]
The other communication assisted scheme considered is Zone 2 Permissive Over-reach. In

this scheme, when the relay detected a Zone 2 fault, it will send a permissive signal to the
remote end. As stated in the previous section, if one of the CT in the breaker and a half
configuration saturates during close-in reverse fault, then a wrong permissive signal will be sent
and resulted in tripping in accelerated Zone 2 time before the fault is interrupted by the correct
protection system.

APS | 2015
The selected built-in communication assisted scheme is Zone 1 Permissive Under-reach. i.e.
the permissive signal is only sent when the distance protection relay sees a Zone 1 fault. The
receiving end then trip if a Zone 2 fault is detected. Therefore, comparing with the other two
schemes considered above, this scheme will provide higher security.
3 Factors Influencing Performance of Distance Protection

3.1 Influencing Factors
The factors influencing the performance of the distance protection for parallel circuit lines, such as
Source Impedance Ratio, Load Flow, Mutual Coupling and Fault Resistance are well documented in
various papers and books.
3.1.1 Source Impedance Ratio

There are varying opinions on the impact of source impedance ratio on the performance of
distance protection. [6] suggests that "the key advantage of distance protection is that its fault
coverage of the protected circuit is virtually independent of source impedance variations".
A more recent paper [7] suggests that the source impedance ratio could have significant
impact on the effective reach of distance protection.
3.1.2 Load-Flow
Load-flow significantly affects the "apparent impedance" measured by the distance protection
relays. It is well documented that the distance protection at the export end tends to over-reach
while the relay at the importing end tends to under-reach. The existing settings are dated
1999. Obviously the system and the load have grown significantly since then and therefore
the impact of load flow must be re-examined.
3.1.3 Double Circuit Lines

When there is a fault on the adjacent line, depending on the fault type (phase to phase, phase
to ground, etc.), load flow, fault location, fault impedance and system parameters, the relays
on the healthy line may see the fault as reverse, forward or cannot detect the fault at all. In the
process of isolating the fault, it is highly likely that the circuit breaker from one end will trip first
leaving the other end to continue to feed the fault. This will again abruptly change the current
distribution of the two lines and affect the distance protection operation.
3.1.4 Zero Sequence Mutual Coupling

The effect of mutual coupling on the phase fault is generally considered to be negligible and
therefore not considered in this paper. For phase to earth fault, some literature suggests that
the distance protection at the stronger source tend to under-reach while the distance
protection at the weaker source tend to over-reach. However, this is an over simplification.
The effect of the mutual coupling depends on whether the measured fault current and the
current in the un-faulted lines are in the same direction. [8]

APS | 2015
3.2 Evaluation of Combined Effects of Influencing Factors

A calculation model is developed based on the simplified network shown in Figure 1 to study the
combined effects of the influencing factors. The results are summed up in the X-R plots below.
Impedance calculated at CI ZSS Impedance calculated at HC ZSS

SIR = 2.16 SIR = 4.02
Load = 0 MVA Load = 0 MVA
Note 1: The solid markers are impedances seen by the Line 2 Distance Protection relays for fault locations between
0 pu to 1 pu of the line length in steps of 0.1 pu.
Note 2: The hollow markers are impedances seen by the Line 1 Distance Protection relays for the faults on Line 2.
Figure 3 – Impedance seen by the Distance Protection Relays for

Phase-to-Ground Fault at various location on Line 2

APS | 2015
4 Verification using Power System Disturbance Records

Power disturbance records are analyzed and the fault impedance trajectories are copied into the
setting calculation sheets to check if the proposed settings will operate correctly. As the recorded
disturbance record may not represent the worst case scenarios, these disturbance records will be
modified by applying a factor of 110% and 120% to the recorded line currents. These original and
modified COMTRADE files will then be played back into the relays to verify the performance of the
proposed protection scheme and to make improvement if required.
The following oscillograph is from Test Case 1. A lightning strike on Line 2 induced a 3 phase to
ground fault. The Circuit Breakers at CI ZSS tripped at around 70 ms and the Circuit Breakers at HC
ZSS tripped at around 90 ms. During the 20 ms between the tripping at the two ends, the fault
currents were fed via Line 1. This current reversal resulted in operation of phase A to ground element
of Line 1. Phase A of Line 1 at CI ZSS tripped at 144 ms and successfully reclosed at 539 ms.
Figure 4 – Test Case 1 – Lightning induced 3 Phase to Ground Fault

APS | 2015
Note: Impedance trajectories after the current Note: Impedance trajectories after the current
reversal are marked with solid markers. reversal are marked with solid markers.
Figure 5 – Test Case 1 – Impedance Trajectories Figure 6 – Test Case 1 – Impedance Trajectories
seen by CI ZSS Line Distance Protection Relay seen by HC ZSS Line 2 Distance Protection Relay
5 Single Pole Tripping, Auto-Reclose and Current Reversal

From the analysis of the power system disturbance records, it was noticed that the phase-to-phase
impedance may come into Zone 2 operating region while the fault is deemed to be phase-to-ground
fault. This phenomenon is exhibited in Figure 7 and Figure 8.
Impedance calculated at CI ZSS Impedance calculated at HC ZSS

SIR = 2.16 SIR = 4.02
Figure 7 – Calculation Model: Response of Phase -to-Phase (ab and ca) elements to
Phase-to-Ground (a-g) Fault at various location on Line 2

APS | 2015
Figure 8 – Test Case 2: Response of Phase -to-Phase (ab and ca) elements to
Phase-to-Ground (a-g) Fault on Line 2
As the intention of the protection scheme is to trip only the faulted phase for a single phase to ground
fault and auto-reclose is applied only to single phase to ground fault, this phenomenon is significant
and warrant thorough study. The complexity of single pole opened, current reversal and single pole
reclose onto fault are beyond the developed calculation model.
To verify the performance of the protection scheme under single pole operation, a model of the system
is set up in RelaySimTest [9] as shown in Figure 9. Test scenarios (see Figure 10) are then selected
to verify the performance of the proposed protection system.
Figure 9 – RelaySim model for verification of the protection system

APS | 2015
Figure 10 – RelaySim Test Scenarios for verification of the protection
6 Conclusions
Using distance protection to cover parallel transmission lines connecting two sources require detailed
studies to verify that the protection scheme will perform correctly under the combined effects of the
influencing factors such as varying source impedance ratio, varying fault impedance, zero sequence
mutual coupling, load-flow, single pole operation and current reversal. The best test cases are the
actual power disturbance records. However, most likely, they are not the worst case scenarios. Test
cases generated by simulation software such as RelaySimTest should be used to provide a thorough
verification of the protection scheme.

APS | 2015
7 References
[1] European Commission – Directorate General for Energy and Transport: Study on the
Technical Security Rules of the European Electricity Network – Final Report 62236A/001 REV
2, 2006
[2] Kirschen D., Strbac G. – Why investment do not prevent blackouts, 2003
[3] GE Multilin – Transmission Line Protection Principles, downloaded on 4 March 2013
[4] Steynberg G. – Special Consideration of Feeder Protection for Breaker-and-a-half
Configurations, 2008
[5] Stevens I., Fisher N., Kasztenny B. – Performance Issues with Directional Comparison
Blocking Schemes, 2010
[6] Alstom Grid's Network Protection & Automation Guide – 2011 - section 11.1
[7] Spoor D., Zhu J. - Selection of Distance Relaying Schemes when Protecting Dual Circuit Lines
– 2003
[8] Ziegler G. - Numerical Distance Protection – 3rd ed. – 2008 - section 3.5.3 - Parallel lines
[9] Omicron – RelaySimTest – Getting Started – 2014
About the Authors
Wing Chan obtained his B. Electrical / Computer Engineering and M. Engineering Management from
QUT in 1993 and 1994 respectively. He has more than 30 years experience in the power system
industry and is currently the Power Networks Protection Manager in Power and Water Corp. (Darwin).
He is interested in system fault investigation, development of calculation tools and in the application
and testing of protection and automation systems.
Si-Dieu Tran is a protection engineer at Power and Water Corp. (Darwin). He received his B.E (1st
Hons) in Electrical Engineering from Charles Darwin University in 2005. Si-Dieu has over 9 years
experience in the Darwin electricity supply industry. He has been involved in the design and
implementation of protection schemes for Darwin’s transmission and distribution systems since 2009.
Sameep Gharti Chhtri obtained his B. Electrical and Electronics Engineering from Katmandu
University in 2006 and MSc in Power System Engineering from Institute of Engineering in 2010. He
has extensive experience in power system and system studies upto 400kV. He is currently a
protection engineer at Power and Water Corp. (Darwin).

APS | 2015

APS
APS| 2015
| 2015
Cygnus Atratus, and How it Causes Problems
Terry Foxcroft | Snowy Hydro

terry.foxcroft@snowyhydro.com.au
1 Introduction (Arial font size 14 - bold)

Protection systems deal with two different types of faults, ones we expect and faults that are unique
and had not been considered.
This paper describes black swan events, and how we as protection people should make our system
black swan tolerant.
It looks at several events, not all within the power industry, and describes how different thinking may
have prevented incidents from occurring. It describes the effect of outliers on fault analysis and black
swan recognition.
The author has been involved in black swan style theory for many years. As a book by Nassim Nicholas
Taleb has created the terms used nowadays, those terms are used throughout this paper.
2 Murray 2 Unit 11 fault

The Unit 11 generator at Murray 2 Power Station tripped at full load. This happened on more than one
occasion.
The protection relays are old style electro-mechanical and solid state.
Figure 1: Murray 2 Unit 11 protection, alarm and control panels
As there are no numerical relays, all fault analysis relied upon relay flagging.

APS | 2015
2.1 Line side current transformers

The differential protection has current transformers on the line side, star point and stator dynamic
braking short circuiting switch. The line current transformers have an instantaneous over current relay
installed.
Figure 2: Murray 2 Unit 11 part of protection AC schematic
This relay is set to 1.1 per unit rated current and performs three functions.
• Circuit breaker failure
o When the protection operates and the circuit breaker fails to trip
• Accidental energisation
o Enabled when unit speed below 90%.
o Protects against closing a circuit breaker onto a stationary unit
• Loading error non urgent alarm
o Greater than full load current
2.2 Relay flagging
The flag for the instantaneous over current element was dropped.
This could have been due to any of the above functions. The only other alarm up was the loading
error. This is a non urgent alarm and not a trip.
It could not have been an accidental energisation as the machine was greater than 90% speed – it
was on line.
There was no circuit breaker failure initiated as the transformer was still in service.

APS | 2015
Figure 3: Instantaneous over current relay
There was no indication as to the cause of the fault.

2.3 Speed detection
The speed of a generator is often detected by a toothed wheel and a proximity probe. A proximity
probe is an electronic switch that operates when a ferrous item comes within its operating range. It is
effectively an electronic limit or position switch. Being electronic it can operate at high speed. The
operation of the proximity probe can be counted to determine the speed of a shaft.
Figure 4: Murray 2 Unit 11 exciter housing

APS | 2015
On the Murray 2 units there was insufficient space in the exciter housing to install a toothed wheel.
Figure 5: Murray 2 Unit 11 main and pilot exciters
As an alternative the proximity probe was pointed down to count bolt heads off the main exciter
armature.
Figure 6: Speed monitoring proximity probe and bolt heads
As the unit loads, thrust is applied to the turbine. This pulls the shaft down very slightly. This makes
the bolt heads to move away from the proximity probe. The result is the bolt head is now at the limit of
the detection zone.
The proximity probe then starts to miss bolt heads. The counter used to determine the unit speed now
starts to register a lower speed. When the loading gets high enough the speed registration gets below
90%.

APS | 2015
At this point the excitation is turned from auto to manual. This is supposed to be a bumpless transfer,
but it always has some small bump or change. At full load this bump is enough to make the stator
current exceed 1.1 per unit.
The current detector picks up. As the speed detection is less than 90% speed – even though the
generator is on line – the accidental energisation protection operates.
The generator then trips. This removes the load and the down thrust on the turbine. The speed
sensing now returns to normal and everything resets. So the only indication is the greater than full load
current flag, which could have occurred at any time whilst on line.
This is a black swan event.
3 Black swan events

In earlier times there used to be a saying rara avis in terris nigroque simillima cygno, which translated
is a rare bird in the lands and very much like a black swan.
This comes from the presumption that all swans are white. In Europe all swans were white. The saying
can be paraphrased as “The chances of that happening are the same as seeing a black swan”. A
similar saying today may be “that will happen when pigs fly”.
Then the Dutch sailors discovered the west coast of what is now Australia. And they saw black swans.
The captains in those days were similar to astronauts today. They were the peak of the technical and
leadership in society. Upon seeing the black swans they then realized that anything was now possible.
3.1 What is a black swan event?
A black swan event is a surprise to the observer. The event has a major effect. Typical comments
when the event first happens are “no one could have seen that coming”. In hindsight it was always
going to happen.
The data is all there but was ignored at the time. It is generally ignored because it doesn’t fit into the
desired perception. All swans are white…..
Some black swan events are positives, some a negatives.
Examples of black swan events include
• The rise of the internet
• The twin towers attack on September 11th 2001
• Mobile phone usage and saturation
• World War 1
• Almost all the failures presented at various groups over the years
3.2 The turkey example
A turkey lives on a farm. Every day he gets up and the farmer feeds him. This happens for 1000 days.
Then, on day 1001, the farmer grabs the turkey can cuts his head off.
This is a bit of a surprise for the turkey, but not for the farmer. The date is 18th December, and he was
being fattened for Christmas dinner.
It was a black swan event for the turkey, but not the farmer.
There was nothing in the turkey’s experience or knowledge to make it think he was going to have his
head chopped off. He may have noticed that there were no older turkeys, but this was ignored as
every day the food kept coming. His thoughts were constrained by his environment and experience.
3.3 How to avoid black swan events
By definition you cannot avoid black swan events. If you foresee it and put in a system to stop it, it is
no longer a black swan. You become the farmer instead of the turkey.
You need to make yourself strong against negative black swans, and able to take advantage of
positive black swans.
There is a saying that generals fight the last war. This is true in all aspects of life. When something
goes wrong, a system is usually put in place to stop it happening again. There is only marginal benefit
in doing this as we all know about the event now. The system installed often works against us as it
makes us blind to other black swan events. These systems make us think that another event can’t
happen.

APS | 2015
3.4 The danger of statistics

We will presume for this example that Osama Bin Laden is still alive. A facial recognition system is
implemented world wide to find him.
There are seven thousand million people in the world. Only one of them is Osama Bin Laden. A facial
recognition system that always responds “It’s not Osama” will be correct for 7,000,000,000 -1 faces. It
will be correct 99.9999999857% of the time. There are many systems implemented that crave this
amount of reliability. However, the statistic hides the fact that it can miss the one face it is trying to
locate.
It is impossible to detect very rare events with any level of accuracy. Either we will never detect the
event or problem, or we will have a lot of false positives. These reduce confidence in the system, so
when a true positive is detected it will more than likely be not treated correctly.
3.5 What if you see it coming?
If you see a black swan possibility, you put in a system to stop it. As the event hasn’t already
happened, your system will quite often be dismissed. If the system succeeds in being implemented,
people will not like it as “that has never happened before”.
The result is people will eventually bypass the system. If the system is removed and the event
happens, the general comment will be “no one could have seen that coming”. Again, a black swan
event.
3.6 Doors on planes 10th September 2001
Imagine someone mandated that on the 10th of September 2001 all planes had to have lockable
security doors for cockpit access.
No planes would have flown into building. There would have been no war in Afghanistan and Iraq as
we know them today.
And there would have been no evidence that they had changed history. They saw a need and
implemented it, stopping what is a world wide event. But as there would be no evidence of what they
had averted, they would be treated as obstructive and someone who makes things harder for no
reason.
4 What does this mean?
4.1 Look for outliers
If there is a black swan brewing, there are often indicators called outliers.
They will be minor, and out of the ordinary. They will be dismissed as a statistical aberration.
These are often the first sign of a black swan event. They are not in people’s normal view.
The more complex a system or device is, the more likely it is to have or cause a black swan event.
4.2 Fault finding
It is often very easy to find a fault. However, if you are fault finding, it is because something has gone
wrong.
Don’t dismiss something that doesn’t make sense. They are often indicators of future black swan
events.
4.3 Murray 2 Unit 11
In fault finding this event, I saw the outlier. It was the flag on the relay for instantaneous over current
which should have caused an accidental energisation trip.
I didn’t want to believe it as it couldn’t cause the trip. The unit was on line and loaded, and was
therefore greater than 90% speed. In fact it was at synchronous speed.
Once I acknowledged that it had to be the cause, the fault was located within 5 minutes.
4.4 Black swan tolerant
We do put in systems to be black swan tolerant. Protection is duplicated, and often from different
manufacturers or different designs. They are connected to different instrument transformers.
Hopefully this creates minimal to zero common mode failures. But this is not really possible.
Sometimes this security can hid black swan conditions.
For example, are both protections installed underneath a drain that can overflow?

APS | 2015
5 Conclusions
Randomness is random and cannot be predicted. We see this all the time at work and focus on the
solution to that event rather than making ourselves tolerant to negative events, and take advantage of
positive events.
We must always be careful with the false security provided to us by our system and models.
6 References
Taleb, Nassim Nicholas: The Black Swan: The Impact of the Highly Improbable. 2rd Ed. 2010;
New York, Random House.
Ibid.
Ibid.
About the Author
Terry Foxcroft has worked in the power industry for over 35 years. He has
worked in the protection commisioning, testing and design fields for over
29 years. Terry is responsible for protection design at Snowy Hydro.

APS | 2015

APS
APS| 2015
| 2015
Application of the Line Differential Protection Scheme for

Radial Transmission Lines
Leonardo Torelli| CSE Uniserve

leonardo.torelli@cse-uniserve.com.au
Ilia Voloh | GE Energy

Ilia.Voloh@ge.com
Zhihan Xu | GE Energy
Zhihan.Xu@ge.com
Abstract
Modern communication networks have dramatically increased the implementation of the line
differential scheme as one or both primary protection for transmission lines. The transmission network
is usually meshed which also provides fault current contribution from both ends of the line.
Considering the ongoing need to reduce CAPEX investment, it is expected that the supply to specific
loads, for instance mining in rural zones of Australia, the transmission network could be expanded
radially with single or double circuit applications. In a radial system, in the event of fault, fault current
contribution is drawn mainly from the source side. This scenario could potentially create issues to the
operation of supervisory elements of the line differential relay at the remote end. This paper reviews
this specific application with particular attention to solutions which balance the need for security, speed
and dependability of the line differential protection scheme.
1 Introduction
The transmission network has the crucial role to transport the bulk of the electrical energy from the
generators to the loads. The system is meshed as centre of loads and generators are geographically
dispersed. As a result, the transmission network is well interconnected and provides an inherent
redundancy of supply to each node of the network. Although the cost and engineering effort to
establish, operate and maintain such complex system, a meshed network is fundamental for the
operation of a reliable power system.
In Australia the majority of the transmission network and substations has been built decades ago. If
possible, new substations are usually located adjacent to existing transmission lines to avoid the
construction of new ones. Transmission lines are very expensive. For instance, for a single 220 kV line
on lattice towers, the average CAPEX cost in Australia is above AUS $ 900k per kilometre. In addition,
the construction of new lines has become rather problematic due to the difficulties to obtain public
opinion support, permits and land easement.
It is noted that there is an increasing demand to reduce the energy cost for customers and,
subsequently, to control investments in the electricity sector. As a result, it is likely that more radial line
will be considered and implemented in the near future to connect new large industrial or mining
customers, existing sites which require a higher capacity and radial transmission backbone in rural
areas currently fed by the subtransmission network.
Radial application in transmission network tends to be a solution mainly used in large countries like
Australia where relevant loads are sometimes located in remote areas. According to the latest
planning reports, there are several existing applications in South Australia, Queensland and Western
Australia. New radial lines are currently considered for future new connections. For the scope of this
exercise radial double circuit application are included in the analysis as are still dependent on the
same source. In practice even a small ring network could become a radial line if one of the line is open
for maintenance or outage. Last , we should consider also radial lines from this study point of view

APS | 2015
even a line that connects a small generator or inconsistent source of load, like a wind farm, ,which
does not generate or generate at small capacity during the line fault.
Figure 1: Examples of radial transmission lines
Transmission lines are usually protected by the line differential scheme and/or distance protection.
Benefits of the line differential scheme versus the distance scheme are widely documented in the
existing literature. In the last decades, with the advent of modern and more affordable communication
between substations, the line differential scheme has become more popular and is now used as one
or both primary protection.
Relay Utility 1 Utility 2 Utility 3 Utility 4 Utility 5 Utility 6 Utility 7
X Protection 87 and 21 87 and 21 87 87 and 21 87 and 21 87 and 21 87 and 21
Y Protection 87 and 21 21T 21T 87 and 21 21T 87 and 21 87 and 21

Table 1- Primary protection schemes used by Australian transmission utilities where modern communication is
available. 21 refers to time stepped distance. 21T refers to distance with teleprotection.
It is noted that in in the event of fault on a radial line, fault current contribution is drawn mainly from
the source side. Although the 87L scheme works satisfactory in this scenario, the supervisory
element at the remote end of the line could potentially no operate, hence inhibit or delay the 87L
operation.
This paper reviews the 87L scheme applied on radial line with particular attention to solutions for the
supervisory element. The paper investigates answers which balance the need for security, speed
and dependability of the line differential protection scheme in this challenging application. The study
targets transmission network and its peculiar needs of high reliability. However, the findings and
suggestions are also applicable at subtransmission level.
2 Overview of existing 87L and supervisory elements practice

The principle of the line differential is an extension of the First Kirchhoff’s Law applied to the full
length of the line as elaborated in 1845 by the young German physicist during its early studies. The
sum of the current entering the line must be equal to the current leaving the line. As a result, the 87L
scheme is a quite unique protection scheme as joins sophisticated electrical and communication
techniques to provide a highly reliable line protection.
Modern 87L relays work on a peer to peer architecture. Each relay is a master and processes the
87L algorithm as well as sending the required data to the remote end of the line. Although the
capability and high bandwidth of modern communications, line differential algorithm is still dictated by
the bandwidth of the communication channel of 64kbps. This limitation requires some specific design
in relation of the structure and contents of the data to be transmitted to the remote end. Another
major element in the design of the scheme is sampling synchronization which enables to compare
information at the ends of the line at the same time.
The implementation of the scheme as applied among modern relays has some obvious similarities
and some important differences. In the electrical part of the design, the key element is the restrain
current which then dictates the required operating current for the 87L trip. Restrain current can be
derived by the maximum current at each end of the line or from symmetrical components values.

APS | 2015
Some 87L schemes also apply a restraining adaptive component which caters for power system
errors such as noise, transients, harmonics, inaccuracy in the DSP measurements, clock and
asynchronous sampling. This adaptive stabilisation of the scheme is based on the principle that the
sum of all the currents entering the protected line is zero and, as result, the errors should be
compensated by applying some additional and dynamic restrain quantity.
Figure 2: Setting impact on restrain characteristic of one 87L relay
2.1 87L Supervisory function

Protection philosophy is a continuous research of finding the appropriate balance between
dependability, the capacity to operate when it is required, and security, the capacity to no operate
when does not have to do so. This equilibrium continuously changes as per available technology,
overarching philosophy and the specific application. In transmission lines, protection schemes are
traditionally biased for security. This is particularly true when we analyse the 87L in stressed
conditions for the protection relay.
Although the line differential scheme is intrinsically secure, there is a consensus between relay
vendors to improve its security by addressing its main challenging conditions: data corruption, data
misalignment, CT saturation, specific application requirements and incorrect settings.
Transmitted 87L packets implement BCH algorithm, checksum or CRC-32 bit data integrity check to
detect any corruption or missing data. Digital relays do not provide any data replacement or data
repair in order to preserve the security of the scheme. Corrupted packets are discarded and
protection relays hold the operation waiting for the next available and correct received packet. Data
corruption can arise due to communications noise, interference in the channel bearer, failure of
devices involved in the data transmission and conversion within the multiplexer, channel switching,
or within the relay itself when data is transmitted from communications port to CPU. These
algorithms have a very low probability of undetected data error even below one every 10 billion a
data. In a relay that generates and sends at 50 Hz two packets at every cycle, the total number of
packets in one second would be 100 or 3.15 billion every year with still clear risks for the security of
the 87L. It is important to bear in mind that even the most sophisticated data error algorithm cannot
eliminate the risk of an undetected error being accepted by the 87L and generating a relay
maloperation.
Large progresses have been done in the synchronisation area to align the current data at the local
and remote end of the line by using the well-known ping pong technique. GPS functionality is
provided for communication network prone to asymmetry in the relay data propagation. Data
synchronization is closely monitored to detect any misalignment and this information can be used to
improve 87L security.

APS | 2015
CT saturation is addressed by specific saturation algorithms either:
• Adding a portion of the signal distortion to increase the restraining current
• Switching the differential settings to more secure settings
• Use the transient bias as the additional restraining signal
There is also a consensus among relay manufacturers to provide extended 87L capacity to include
ground differential, multi CT input, Direct Transfer Trip to the remote ends of the line, line charging
compensation, stub bus protection, in zone transformer capability, hence, enhancing its inherent
security for specific and challenging application.
Human error is also a predominant cause of maloperation in secondary engineering. Incorrect
operation of personnel could happen during the design phase of the project or oversight during
commissioning. The multiplication of IEDs within the substation has also made more difficult the
integration of new protection and control scheme especially for a brown field substation. Incorrect
manipulation of the protection setting could appear during a routine maintenance or remote or local
engineering access to the relay to monitor online metering or the existing settings. These concerns
have led manufacturers to work extensively on the security of the relay in different forms not
excluding the specific algorithm of each protection element.
At the beginning of the modern 87L digital relays period, 1995-2000, the work to enhance the 87L
security brought relay manufacturers to also add an independent protection element to supervise the
operation of the scheme. One of the solutions to achieve this goal is by confirming that the fault is
taking place by implementing a disturbance detector element at each end of the line. If there is not a
disturbance in the network, it is very likely that there is not a fault. This section outlines the purpose
and functionality of the supervisory element of the line differential scheme as applied by six different
relay vendors:
Relay A uses a current sensitive algorithm that compares actual symmetrical component values to
the values of four cycles before. The threshold is set at 0.04 pu. If detected, the supervisory element
also resizes the operating window of the 87L packet to remove the pre-fault current, thus reducing
the operating time of the 87L. The supervisory element also provides an adaptive level detector
operating on zero and negative sequence absolute values. The scheme was designed to be flexible
to accommodate different application and provides seal in and no current supervision detection
functions.
Relay B does not have a specific 87L supervisory element but ensure additional stability upon
communication propagation changes which can be encountered in a SDH system. If successive
propagation delay exceed a user define settings( 0.25ms- 1ms), the relay in automatic change the
slope settings to a more conservative settings for a short period of time up to two seconds.
Relay C implements a supervisory scheme based on sensitive phase and zero sequence current
change detector set to 0.04. The relay also offers a settable phase overcurrent which runs in parallel
with the fixed two elements.
Relay D has been tailored to provide additional security for internal and external faults. The scheme
is fully integrated within the 87L. The external fault detection is designed to tackle fault scenarios
followed by high AC components or a long lasting DC component. The scheme is also used by the
87L element to switch the scheme to the extended security settings logic. The disturbance detection
logic is based on current and voltage and also integrates the detection logic from the remote end.
The detector is adaptive and based on full cycle filtered current phasors to adjust its threshold to
avoid asserting under normal load variations. Once asserted, the detector signal is maintained for 10
cycles.
Relay E monitors the stationary or slow building unbalance in the time propagation between transmit
and receive packets if GPS synchronisation is available. During normal operation the restrain and
differential quantities are also monitor and, in the absent of transmission time change typical of SDH
switching, an alarm based on “line differential too high” is raised.
APS | 2015
Relay F provides a flexible supervisory functionality based on current and voltage which is
articulated in four subsections operating with a parallel logic. Thresholds as well as timers are user
definable and provide the start for the protection element. Each subsection is independent and can
be disabled if required. Phase to phase current variation is the main element and provide sensitive
detection for the majority of the type of faults. The zero sequence current is monitored to target high
impedance fault scenarios. Instead, low current criterion is suited for switch onto fault scenarios
where the prefault current is below a settable threshold. The undervoltage element is tailored for
weak infeed scenarios with pre fault no load conditions. Phases to phase as well as phase to ground
voltages are monitored to supervise the differential element under any type of weak infeed fault.
All the relays have programmable logic capability to allow the implementation of specific supervisory
logics and are able to send this status to the remote end of the line. Current detectors are not usually
phase segregated. Hence, at current change detection on a phase could be used to assert the 87L
operation on a different phase.
3 Complications for the application of 87L on radial lines

Upon a fault on the protected radial transmission line, the remote line end has either no infeed or
weak infeed current contribution. At the source end phase and symmetrical current phasors depends
on the fault level, fault type, fault resistance, and location of the fault. In the event of an external fault
on the remote end side, current seen by both ends of the line will have similar phasors as in a
meshed network.
For an internal fault, voltage level at the source end will be reduced but still relevant in comparison to
the remote end where the voltage will collapse. The voltage at the source end depends to the source
impedance and the fault impedance. At both ends voltage level will increase in the event of a
resistance fault. For an external fault the voltage phasors at the two ends are more similar except for
long lines due to the voltage drop created by the fault current.
Typically, radial transmission lines are located in remote areas where the fault level is average to
low. In rural Australia it is not rare that the maximum fault level at 220 or 132 kV can be as low as 2.5
kA. Assuming I fault minimum/I fault maximum ratio of 70%, the fault level could decrease to 1.75 kA
during minimum fault level scenario. Other weak infeed scenarios include high resistance faults,
simultaneous faults, line or transformers out of service in the same network area.
Radial lines are often located in areas prone to bushfires. According to field experience, arc
resistance through smoke and fire dust can potentially reach 30-40 Ohms. These elements suggest
that this analysis should also include weak infeed scenario at the source end.
Voltage at any point of the line can also be calculated easily by using the Source Impedance Ratio,
SIR:
1
Vpu = (1)
(SIR + 1)
Where
Zsource
SIR = (2)
Zline
A high level representation is provided in Figure 3 to 6 for a three phase fault at the end of the line
with different fault levels, fault resistance values and line length. It should be noted that any current
contribution coming from the remote end will increase the voltage level at the remote substation.
Usually this increase is not relevant due to the minor current contribution.

APS | 2015
0 Ohm
0.5
5 Ohm
10 Ohm
0 40 Ohm
Infinite bus Line source end Line load end
Figure 3: 220 kV 10 km length – 10 Ω Z1 source- Three phase V fault profile
1
0 Ohm
0.5 5 Ohm
10 Ohm
0 40 Ohm
infinite bus source end load end
Figure 4: 220 kV 100 km length – 10 Ω Z1 source- Three phase fault V profile
1
0 Ohm
0.5 5 Ohm
10 Ohm
0 40 Ohm
1
0 Ohm
0.5 5 Ohm
10 Ohm
0 40 Ohm
One of the key advantages of 87L protection is being able to detect and operate for a very large
range of fault currents. The scheme is generally insensitive to how the current contribution is derived
from the two ends. Despite the specificities of each relay, the 87L should be treated as one scheme
which transfers current information between locations via communication. A zero current contribution
from the remote end is still a valid and meaningful data for the operation of the 87L algorithm.
For a transmission line located in a weak network area, the impact of the radial line on the restraining
current is minor and mainly depends on the 87L algorithm design, the specific protection settings and
the CT ratio selected. Referring to the fault scenario and results shown on Table 2, the operating
current of the 87L scheme is still well above the restraining current with a high level of dependability.

APS | 2015
Figure 7- Dependability of 87L for high resistance fault. Software simulation with 87L relay
Fault I I Restrain with I Restrain with fault Dependability?

current Operate fault current current from both ends.
(A) (A) from one end 50-50 % contribution
only. (A) share. (A)
250 250 158 150 Yes
500 500 200 173 Yes
1000 1000 316 245 Yes
2000 2000 583 425 Yes
4000 4000 2045 812 Yes

Table 2- Dependability of 87L for high resistance fault. 220 kV radial line. CT 500/1. Ip 0.2 pu, slope 1 20 %,
slope 2 40%, Breakpoint 4 pu
CT saturation. During an internal fault, if the fault level and the total CT burden are very high, there
is a higher risk of CT saturation at the source end as the full fault current is seeing by one end only.
However, with correct selection of the 87L high end differential slope setting, it is very likely that the
87L will operate, hence confirming a general tolerance of modern 87L relays towards CT saturation
for dependable operation.
The saturation caused by an external fault is more a concern for the 87L scheme, especially in
breaker and half application or where the CT performances between the two ends are rather
different. For an external fault on the remote end side, for instance on the HV side of the transformer,
a high fault level could develop CT saturation at one end and challenging the security of the 87L. The
relay maloperation will completely disconnect the load. This risk is minor for a fault on the LV side of
the transformer due to the reduced fault current drawn from the network.
In a double circuit, sharing the fault current contribution with the adjacent line will sensibly decrease
the risk of CT saturation. Similarly, for an external fault at the source end side, the two CTs on the
breaker and half will always see the same current on the diameter and, likely, will have similar
performance even during high through fault currents and reduced CT performance. We should
reiterate that CT performance is still one of the key points for the stability of any type of differential
scheme which affect in the same way meshed and radial applications.
In order to analyse the 87L reliability during CT saturation generated by an external fault, relay
manufacturers are developing specific tools tailored for the specific relay. The software tools are
used for CT selection and to check if it is possible to reduce the CT requirement. It could also
support the designer to determine the impact of changing the 87L settings. One of this tool
implements the CT model and saturation algorithm proposed by IEEE PSRC and simulates the
signal processing and data calculations as applied the 87L relay. As a result, the tool incorporates in
one application the analysis of the CT and the performance of the 87L algorithm.

APS | 2015
Figure 9: CT saturation analysis tool
Radial line and 87L sensitivity. Apparently in radial double circuit application, upon an internal
fault, the remote end will see some fault current contribution coming from the healthy adjacent line.
Assuming that the two circuits have the same line impedance, the highest current is for a fault near
the remote end with 50% of the total fault current seen by the relay. Vice versa, the lowest weak
infeed scenario is during faults towards the source end of the line. For instance, for a phase fault
located at 95% of the line, the remote end will see 2.5% of the total fault current. Assuming, a
minimum fault current of 1.5kA, the remote end contribution will be as low as 24A or 0.04 pu for a
600/5 CT ratio. However, upon a fault on one of the two lines, an 87L incorrect operation on the
healthy line will completely disconnect the supply to the load, hence a blackout to the remote
substation.
Figure 9: Double circuit with fault at 95% from remote end
These considerations for single and double circuit application are generally valid for three ended line
application with only one end acting as the source.
In relation to a long radial line and the line charging current compensation function, the radial line
itself does not add any additional challenge to the 87L scheme. For a long line line charging current
should be always considered as a standing differential current which impacts on the performance of
the scheme. If the compensation is not applied the security could be affected. If the designer prefers
to modify the protection settings with a more secure approach, for very long line say above 200 km,
the sensitivity for very high impedance fault could be marginally reduced. It is worth to note that the
line charging current compensation might be required at the source end and could be ignored at the
remote end of the line if there is not earthing reference to close the loop of the line charging current.
At this stage we have based the analysis considering that the remote end does not contribute to the
fault. In real application, however, the remote end of the line can supply some fault current
contribution due to the transformer windings, embedded generators and large motors.
For an internal phase to ground fault, the zero sequence path of the transformer with HV star
earthed and a tertiary winding will act as a fault current source. Even if the HV side is unearthed zero
sequence path can be created by the core and tank of the transformer. As a result, the magnitude of
the current contribution from the remote end depends on the transformer winding connection,
impedance and the network impedance connected behind the remote substation. Instead, the
transformer itself is not capable to supply current contribution for three phase or phase to phase line
faults but is still capable to contribute for phase to ground faults even if the CB on the LV side of the
transformer is open.

APS | 2015
Figure 10: Examples of transformers that provide zero sequence current path or shunt
Embedded generators are often located in large industrial or mining plant for emergency supply and,
in some cases, to offset the energy production cost. In this event, it is likely that the generator will be
connected during a fault on the line.
Synchronous machines contribute between 6-10 times the nominal current to the steady state fault
current, about 2-3 times in few cycles. For 87L protection consideration and protection coordination,
it is usually recommended to use either the subtransient or transient reactance according to the
related time constant.
Large motors could also generate large fault current contribution. Direct On Line induction motors
contribution is typically 5-6 times the nominal current. Initially, the contribution contains both AC and
DC components. Due to the voltage collapse the current quickly decays. For VSD and soft starter
motors the fault current contribution is lower and limited to 2-3 times the nominal current.
It is worth to note that if generators and large motors are connected at a lower voltage, the power
transformer impedances will considerably limit the three phase fault current contribution. For ground
faults instead, the various parallel zero sequence paths will still reduce the overall impedance of the
remote network. Fault current contribution from these sources can be determined by applying IEC or
IEEE standards. However, it is common that Utilities tailored the study to suit the specific application.
4 Solutions to mitigate impact of the disturbance detection at the

remote end
As reviewed in Section 2.1, 87L supervision operates mainly on current detectors or current change.
Upon a fault, the impact of the radial application to the 87L supervisory element has a dramatic
performance difference between the source and remote end of the line. If at the source end the
application does not create any specific difficulties compared to a traditional meshed network, the
remote end current detector is at risk of not operating, hence, preventing the fault clearing of the
remote circuit breaker.
Do we really need to trip the remote end for a line fault? Complete isolation of a faulty part of the
network is a cardinal rule of power system operation. Although the opening of the source side only
prevents damage to equipment and is common practice in radial feeder at distribution level, this
should be avoided at transmission level. Transmission substations contain duplicated, overlapping,
redundant protection and control schemes. The complexity of the system suggests that fault
outcomes should be unique and in line with standard practice in other substations. Double
contingencies will also add more difficulties to the operators. It is foreseen that at both substations
fault finding and fault investigation will be more complicated.
These substations are often located in remote areas which makes site visit from personnel more
onerous. Site inspection upon a fault is usually a stressful event which could take place at the most
unexpected time of the year with challenging weather conditions like bushfire or cyclonic conditions.
Remote end substations could be shared owned by the Utility and the mining owner which usually
adds more complexity to site operation and maintenance. Potentially, the level of knowledge and
experience on secondary engineering in the privately owned part of the substation could be lower.
Providing the least ambiguous design and event outcomes will decrease the likelihood of any safety
issue.
Technically, the absence of circuit breaker trip at the remote end allows current contribution from
motor to the fault during the dead time of the Autoreclose function. This decaying current could
potentially maintain the arc fault till AR recloses. The result is defeating the AR purpose of self-
healing transient faults.
A list of some existing and potential solutions for the 87L current detector for the remote end is below
provided.

APS | 2015
Do nothing. This “solution” implies no modification to the standard 87L supervisory current detector.
It relies on the deep knowledge of the remote substation load such as plant operation details, and
subsequent expected minimum and maximum fault current contribution. Motor data and duty cycle
are also required. Often this information is not available during feasibility study which forces utilities
to base the design on typical figures and other assumptions. The decision of how many transformer
and motors should be included in the system model is critical as reducing the impedance will
increase the expected fault current contribution.
Maintaining the same current detector design for all applications also embraces the need of
standardisation of protection schemes, simplification of design and cost reduction. A specific solution
for the current detector on transmission line could potentially be seen by some utilities as an
unwelcomed concession no in line with a modern and efficient approach to substation design.
Design changes following commissioning are expensive and, wherever is possible, should be
avoided. For mining plants, due to the market volatility and fluctuation of commodities price, it is not
uncommon to see plant operations and loads changes over time. Therefore, utilities tend to be very
cautious in using plant information too closely as these conditions might not be valid any more even
in the short/medium period.
Typically, selection of the CT ratio is based on the line rating and the CT performance during an
external fault. It is also noted that utilities usually select the same CT ratios to avoid using 87L CT
ratio compensation function. Sometime these factors lead utilities to select a higher CT ratio of what
it is really required. On radial line it is suggested to consider a CT ratio above maximum load plus a
safety margin for future use. At the remote end, the CT ratio could be further reduced for additional
50DD sensitivity based on avoiding onerous CT saturation and significant CT performance
differences between the two ends of the line. This solution is often feasible in remote areas with a
low fault level or in an application with a new CT with satisfactory specifications. This approach will
improve the sensitivity of the current detector as shown on Table 3.
Fault Line Rating Remote Source 87L current Suggested 87L current
level end max and detector with CT ratio (A) detector with
(A)
at load remote typical CT new CT
remote end ratio(0.04pu) ratio(0.04pu)
(A)
end
CT ratio (A) (A)
(KA)
(A)
6 550 130 600/1 24 400/1 16
6 550 260 600/1 24 400/1 16
3 550 130 600/1 24 300/1 12
3 550 260 600/1 24 300/1 12

Table 3- Examples of CT ratio selection and current detector sensitivity
For double circuit radial line there is also the consideration of the current contribution from the
healthy line: Should it be considered? For the dependability of the 87L supervisory element, it would
be prudent to keep in mind that the adjacent line could be out of service, hence, the consideration for
a single radial line still persists. This approach is in line with the distance scheme where several line
scenarios including adjacent circuit out of service and earthed at both ends are considered for
distance zone reach design.
The support of this solution also considers that the current change on the healthy phases during a
fault is often sufficient to enable the current detector. This is particularly true for 24/7 operation plant.
Further details are elaborated in Section 6. However, power system protection is for its nature
conservative and requires a high level of reliability, especially at transmission level. If a credible
current change or contribution from the remote end is not foreseen for likely fault scenarios, other
means to satisfy the 87L supervisory operation should be investigated.

APS | 2015
Sequential trip. This is a natural progression of the solution detailed above. The fault will be cleared
by the source end anyway which eventually drops on all the phases the current to zero.
Figure 11: Sequential trip example. The three typical phases
If the load was connected at that time, the current change even on the healthy phases is likely to
enable the current detector and, then clearing the fault with some delay. However, it is required to
latch the 87L operation to prevent the element dropout upon source end clearing, hence inhibit of the
sequential trip. In addition to relying on a minimum connected load, the other drawback of this
approach is delaying the fault clearing time of 80-100 ms at the remote end. Considering the small
current contribution from the remote end this issue appears to be acceptable. However, this delay
might not be tolerable from a National Electricity Rule point of view as, in the end, the radial line is
still part of the transmission network.
Disable 87L supervision. In this option the 87L remote end will trip without supervision which raises
the key question about the need of the 87L supervisory element for a radial line.
Figure 12: Disabling the 87L supervision logic
In fact, it is likely that X and Y protection will have dedicated redundant communication bearers with
often no practicable SDH rerouting. Hence, channel asymmetry is less likely or not possible in
comparison to a typical meshed network. However, extreme noise and data corruption are related to
the communication performance and are still possible as in a meshed application.
Further review of the 87L philosophy of the balance between security and dependability in a radial
line is also warranted. Is this balance for a radial line different from a meshed application? It is
noted that there are two schools of thought. The first approach leans towards dependability of the
scheme. The second one is biased for security.
Radial lines are often dedicated to one customer or to a rural area with less customers connected.
As a result, a lower degree of supply security might be acceptable. Besides, power supply will not be
interrupted to other substations and no thermal overload on adjacent line will be generated. Even in
a double circuit application, the rating of one line is capable to cater the full load of the remote
substation.
The concern is rather related to a load rejection which triggers dynamic changes on voltage and
frequency in the network. The opening of a transmission line, even in a radial application, always
creates transient overvoltages and surges which propagate to the adjacent lines with consequent
stress on the insulation of the system. Presence and performance of shunt capacitors, reactors or
other specific equipment at the source end should be also considered. In general, a maloperation on
a radial line has a low impact to the stability and supply to the overall transmission network and is
contained within the system specifications.
In terms of protection, 87L and 87T applied to adjacent lines and transformers are generally immune
from dynamic oscillations as the Kirchhoff’s first rule always applied even during transient conditions.
Modern distance relays also have satisfactory performance during these stressed conditions.

APS | 2015
The other approach considers the importance of supply customers fed from one source only. For a
large industrial plant financial impact of an incorrect line trip including interruption (loss production,
restart operation, risk of damages on connected drives) to large motors, conveyor belts, crush rocks
drives and large mining machines is also taken into consideration.
The importance and need to maintain the supply during transient faults has also lead some
Australian utilities to implement single pole tripping on some specific applications. These projects
remark the need of the continuity of supply and the security of the protection scheme applied to a
radial line.
Although disabling the 87L supervisory element completely resolves the issue of the current
detection, the authors suggest that, in general, the importance of continuity of supply and security for
the 87L scheme still applies to radial line as for meshed application. This is particularly true where
MUXes are implemented in the communication between the two ends of line. Considering that this is
becoming common practice for transmission utilities, the use of the 87L supervisory element is
obviously encouraged.
Decrease the threshold of the current detector. Current supervision threshold depends on the
minimum pick up, the conversion range of the relay and the accuracy of the current detector. It is
common to add a safety margin which is usually equal or above 100% of the minimum conversion
range. Although an increase sensitivity of the scheme will improve the dependability of the 87L
detector, this will not fully resolve the risk of no operation for extremely weak infeed or no current
change. It should be also noted that relay manufacturers do not allow the setting of the current
detector, hence, this option is not always possible. Moreover, it would be not advisable to have a
major sensitivity increase, let’s say from 0.04 pu to 0.01pu with the risk of an unwelcomed 50DD pick
up during normal load current fluctuation. For instance, considering an hypothetical threshold of 0.01
pu with a CT ratio of 300/1 A on a 220 kV line, the current detector would continuously pick up for
load change above 2.25MVA which is rather common in a large industrial plant.
Undervoltage detection. The use of undervoltage supervision appears to be a solution that could
work correctly for a large variety of fault scenarios. This element is already implemented for some off
the shelf 87L relays. Where the undervoltage supervision is not readily available, the undervoltage
check could be implemented and AND gated with the 87L element in the protection logic.
Figure 13: Use of the undervoltage element for the 87L supervision logic
The undervoltage element could be implemented either using a pick up threshold or monitoring the
change compared to the voltage of four cycle before fault inception. Design of the undervoltage pick
up should consider the maximum expected fault resistance on the line and voltage dip during the
worst credible motor starting scenario. Voltage depression during simultaneous faults on adjacent
line should be disregarded in this instance. Usually, an undervoltage pick up setting between 0.6 to
0.8 works correctly but needs to be confirmed during the power system study for the maximum
expected high resistance fault. Voltage elements are fast and will no delay the 87L operation.
It is noted that the use of voltage element does not utterly convince the protection engineering
audience for historical and technical reasons. One of the key points of the line differential over the
distance protection is being independent from the voltage input and overcome VT fuse failure issues.
Therefore, this solution has still a minor impact on the dependability of the 87L element.
87L supervision signal from the source end. The remote 87L supervision is here paralleled, OR
gate, with the identical signal from the source end which is deemed to detect the fault.

APS | 2015
Figure 14: Use of the 50DD SV from the source end for the 87L supervision logic
This signal will be contained in the first 87L packets with no risk of delay the 87L operation. This
solution appears to be in line with the nature of the 87L scheme which is, in the end, one scheme
only. As current and other data is used from the source end, there is not a strong reason not to use
the 87L supervisory signal. In theory, this solution does not perfectly align with the peer to peer 87L
architecture and independency of the operation of the relay upon fault detection. However, the
likelihood of simultaneous and undetected corruption of the differential data and the 87L current
detector supervision from the source end is extremely remote and, hence, should be disregarded.
Direct transfer trip. Modern 87L relays are provided with Direct Transfer Trip function, DTT. DTT
signal is generated by the other end 87L trip operation. In theory, the DTT signal could be also used
to map other protection elements operation. If the signal is enabled, the 87L trip at the source end
will trip the remote end.
Figure 15: Use of DTT from local ( source) end for the 87L supervision logic
In a typical application, this is a redundant tripping signal as the 87L at the remote end should clear
the fault anyway. We note that some utilities always enable this function all the time for redundancy
and simplicity purposes.
Although this solution includes advantages and disadvantages that appear to be similar to the 87L
supervision sent from the remote end, this option is not biased for security as the corruption of the
single DTT bit, the remote end will trip with no any other safeguard.
Use of current angle change detector. For the 87L supervision in very weak infeed application
other studies have investigated the use of angle change of the symmetrical components in parallel
with the traditional current magnitude change detector as an alternative method to overcome
challenging scenarios. Although the principle is correct and feasible, there is still a limitation in the
event that the current is still lower than the relay current threshold.
Use redundant protection element. Microprocessor based protection relays are provided with
multiple protection functions. Even where the 87L is the primary protection, it is common that the
distance or directional elements are enabled to provide either redundant protection or back
protection in the event of 87L unavailability. This “off the shelf” protection element should operate
continuously to provide dependable redundant supervisory function.
Figure 16: Use of redundant protection element for the 87L supervision logic

APS | 2015
Unfortunately, due to the weak infeed at the remote end, distance protection and directional
overcurrent do not have better dependability of the dedicated 87L current detector. Therefore, this
option is unconvincing.
Combination of two or more options. Multiple solution approach is already implemented by

several relay vendors or could be implemented by the user via relay logic. Where three options are
considered, decision should be made regards of any of the elements enable the 87L supervision or
rather a 2 out of 3 logic biased for security approach which is not recommended for a supervisory
scheme. It appears that a simple multiple approach suits the 87L supervision as provides enough
flexibility to cover different scenarios. However, there is a potential risk to overly complicate the
design, commissioning and maintenance of the scheme with subsequent impact on dependability.
5 Performance issues of 87L

There is a general understanding that if a minor load is connected at the time of the fault the current
detector will operate even for a three phase fault due to load drop current changes, hence,
disregarding the fault current contribution. Although this statement is statistically correct, there are
some scenarios where the load connected is low or below the current detector threshold, such as:
• Following a fault which has disconnected a large part of the remote substation load
• At a rural zone substation, for some very short period at night, in particular during mild
weather conditions, where air conditioning or electrical heating loads are at minimum
• At an industrial substation, during routine plant shutdown period, which usually takes place
every six months or one year
• At an industrial substation, where a major production reduction has been implemented
In any case, the extreme scenario of no load connected is during line energisation or during the
transitory phases between line energisation and load connection. The latter phase can take several
minutes or hours. In these scenarios and an internal fault, the source end will see enough fault
current to enable the 87L current detector and the remote end none which is not different from a
typical two source two ended line.
How does the 87L perform during simultaneous faults? For a radial line simultaneous faults are
not usually thoroughly investigated. Utilities are concerned whether the line shares the same towers
with another circuit. In this scenario, a simultaneous fault could take place for a broken conductor, a
dramatic pole or lattice tower collapse or a lightning event which hits or transfers the fault to the
adjacent line. We have reviewed the 87L performance upon a first fault on four different locations of
the network as shown on Figure 17. The 220 kV radial line supplies an industrial plant which
provides a small but still reasonable fault current contribution.
The first fault is on the following location:
1. Adjacent 220 kV transmission line

2. HV side of the 220/22 kV power transformer
3. LV side of the 220/222 kV power transformer
4. 22 kV switchboard within the plant
The second fault is obviously on the radial line under review. In particular, the focus is on the
undervoltage and 50DD current detector considering an internal three phase fault.

APS | 2015
Figure 17: Simulataneous fault analysis and impact on 87L on radial line. Location of first fault is 1 to 4
In any scenario the reduced fault current is sufficient to operate the 87L and the current detector at
the source end. At the remote end, fault current contribution the substation will be reduced until the
first fault is cleared. This could affect the 50DD supervision.
• For Scenario 1, it is expected that a reasonable fault current will be provided from the plant
anyway, with a reasonable chance to have the 50DD enabled well before the fault on the
radial line
• The impact on the 50DD is minor for Scenario 2, as the two fault locations are not
electrically very far away. In theory, if the faults are just before and after the 220 kV bus at
the remote substation with no fault resistance in both incidents, the fault current seen by the
50DD is 50% of the total current contribution from the plant
• For Scenario 3 and especially Scenario 4, the increased impedance between the two fault
locations, will limit the fault current drawn from the plant. It is likely that the 50DD will not
operate
• The use of the undervoltage element as 87L supervision will instead operate correctly for
Scenario 1 and 2 and partially for Scenario 3. For Scenario 4, there is a high risk that the
voltage collapse on the 220 kV line is minor, hence disabling the capacity of the 27 element
to operate
Faults in a distribution network are not always cleared instantaneously due to protection coordination
requirement. As a conservative approach we should consider that for Scenario 3 and 4, the scheme
supervision and the 87L operation at the remote end of the radial line could be delayed up to 0.5-1
second.
Academically, we should also mention that for Scenario 1, there is a minor risk for the fast 87L
operation upon a high resistance fault on the radial line connected to a breaker and half configuration
at the source end during high current transfer on the diameter. In the event that the other diameters
are interrupted, it is possible that a high current flowing on the affected diameter will restrain the 87L
until the first fault has been cleared. This scenario is unlikely and the drawback or delaying the fault
clearing of say 100 ms, is considered acceptable.

APS | 2015
Figure 18: Simulataneous fault. Scenario 1 with high fault transfer on the diameter feeding the radial line
Last, let’s now try to invert the scenarios by applying the fault on the 220 kV radial line first. In this
event and Scenario 1, we can safely state that the fault on the 220 kV radial line will not worsen the
operation of the 50DD on both ends of the other 220 kV line.
6 Fault analysis on a real radial application

To analyse the performance of the 87L supervision two genuine phase to ground fault events on the
same 132 kV line have been reviewed. The line is in the Electranet transmission network and
connects Hummocks Sustation, source end, to Kadina East Substation, remote end. The line is
situated in the northern Yorke Peninsula in South Australia.
Figure 19: Yorke Peninsula Transmission System Diagram
At Kadina East there are two 60 MVA transformers, YNd11yn group. The delta tertiary side is used
to supply the station service transformer. The LV side of the transformer is earthed via a Neutral
Earth Reactor, NEX, of 2.5 Ohms. Both transformers were in service at the time of the events. At
both substations there is a double circuit breaker or meshed configuration.
The fault level in the area is low. At Hummocks they are 4.2 kA and 4.1 kA for the three phase and
phase to ground fault. At Kadina East the fault level decreases to 2.3 and 2.6 kA respectively. The
132 kV line is 40 km long. The faults occurred on 17th of December 2011 and 12th of October 2014.
In both cases the 50DD current detector picked up correctly and, as expected, 10-20 ms before the
operation of the 87L.
The 50DD of the protection relay under review compares the change in magnitude of the negative,
zero and positive sequence current of the actual values to the values of four cycles before. The
threshold is fixed and set at 0.04 pu.
.

APS | 2015
Figure 20: 2011 Fault. Oscillography data
Figure 21: 2014 Fault. Oscillography data
Event CT Pre Fault Fault Fault Fault Remote. Remote Voltage Voltage
ratio fault current current current current at at
Io 50DD
at load from from from from source remote
change operation
both current Source remote remote remote (pu) (pu)
(A)
ends ( A) end (A) Phase Phase Phase
A (A) B (A) C (A)
2011 600/1 34 2950 480 410 480 455 Yes 0.31 0.25
A-G
fault
2014 600/1 31 2260 330 390 390 368 Yes 0.4 0.25
C-G
fault
Table 4- 132 kV Hummocks Kadina East faults analysis- Results
From the two events we can derive the following:
• At the source end the 50DD will definitely operate for any type of fault
• At the remote end the 50DD dependability is also very high due to the significant zero sequence
current provided by the tertiary winding of the two transformers
• The remote current contribution is above 400 A which is 0.6 pu. The 50DD would have operated
even with only one power transformer connected
• Total fault current and fault voltage shows that there was some minor resistance involved.
Considering the high current contribution, it is suggested that the 50DD would have operated even
with a higher fault resistance
• The fault current contribution ratio between the source and remote end is about 6 in both events
which suggest that the faults were located in approximately the same section of the line. Besides,
the ratio confirms the impact and importance of the transformer tertiary winding zero sequence
shunt for line internal fault

APS | 2015
• For the purpose of dependability, it would be prudent to analyse the 50DD operation for the worst
case scenario, which is a fault close to the source end, Hummocks substation, with a 40 Ohms
fault resistance. Assuming typical transformer and 132 kV line impedances, it is possible to
determine that the current contribution ratios between the two ends could be in the range of 7 to
13. Using a conservative ratio of 15 and considering a high impedance fault of 1.3 kA close to
Hummocks Substation, there is still a satisfactory 82 A of fault current drawn from Kadina East.
• It is noted that the current contribution from the remote end would have decreased in a strong
network area or in a much longer line. Realistically, it is believed that at Kadina East for a
ground fault the 50DD will likely operate for any credible fault scenario
• The tertiary winding of the transformer generates current changes even on the healthy phases. As
a result, even a 50DD current detector based on phase value magnitude change would have
operated. Figure 22 shows the post fault phase current increase, phasor rotation and similar
current contribution from the two power transformers as denoted by the CT source data
Figure 22: 2011 Fault. Pre and post fault phasors of the remote end breaker and half CTs
• Would the 50DD have operated for a three phase fault? The load at Kadina East is a
combination of rural and residential customers with no large industrial loads, HV drives or co-
generators. Considering the 132/33 kV transformers, other transformer impedances and feeder
distribution impedances, it is estimated that the ratio between the positive sequence impedances
of the source of the remote end is very high, hence reducing the fault contribution to a few
amperes. For the sake of this analysis we have arbitrary used a load current drop to 10-15A at the
time of the 87L pick up. Apparently, it is possible that the positive sequence current change would
have been just below 0.04 pu, 50DD pick up. Therefore, there are two possible scenarios:
o If the source end clears the fault within 80 ms, the opening of the circuit breakers would
drop the load to zero and enable the 50DD at the remote end. The fault clearing time would
have a delay by 50-60 ms
o If the source clears the fault after 80 ms from the fault inception, the dropping of the current
to zero in two phases will not trigger the 50DD
• If this fault clearing time delay could appear acceptable, the sequence of events also highlight that
with a lower load connected at the time of the events, the 50DD would not have operated for a
three phase fault. A similar conclusion can be draw for a phase to phase fault
• Considering that the maximum demand is about 30 MVA and the fault level is relatively low, a
higher sensitivity for the 50DD current detector could be achieved by reducing the CT ratio to , for
instance 300 or 400/1, based on CT ratio availability and no significant impact on the CT
performance during an external fault

APS | 2015
• The voltage collapse at the remote end also confirms that un undervoltage element could be used
as an alternative method for the 87L supervision
.
7 Practical solutions
The use of the undervoltage element in a modern line differential relay appears to be a practical
solution. As detailed earlier, there is a common concern that the VT fuse failure will block the 87L
supervision. Considering that the substation could be situated in a remote location, fuse replacement
and cable fault repair could require few hours or some days. There is a risk that the improvement
provided by the undervoltage element is ineffective. A practical solution is to bypass the element in
the event of VT fuse failure.
Figure 23: Bypass the undervoltage element if VTFF operates
A practical and simple fix to overcome very low load scenario would include the bypass of the 87L
supervision and its benefit in the event of phase current being below a user definable threshold. For
some radial line feeding large industrial customers, statistically, the impact on 87L security could be
contained. For instance, in a 220 kV application where the typical load is 100 A and CT 400/5, a load
current threshold of 0.04 pu to bypass the current detector would refer to 16 A equivalent to 6 MVA.
If the remote end the plant was a 24/7 industrial plant like a smelter, LNG processing site or large
mining, the percentage of time where the supervision is disabled could be low. This solution would
require a detailed risk assessment but, as elaborated throughout this paper, the use of the 87L
supervision is always recommended.
Figure 24: Bypass the 50DD element if load below user defined OC
A practical extension of the option above would include the bypass of the 87L supervision in the
event that the load current is below a certain threshold and the 87L DTT from the strong end has
been received for more than two consecutive packets. In the event that 87L computation is
completed every 10 ms, the tripping of the remote breaker would be potentially delayed by 10ms.
Figure 25: Bypass the 50DD element if load below user defined OC and DTT received
It is obvious that these solutions aim to improve the dependability of the 87L scheme for a marginal
impact on security or simplicity of the scheme.
Are these alternative methods fully convincing? The use of the current detector from the source end
to the remote end appears to be the most robust, simple, secure and dependable solution to

APS | 2015
maintain security on the line differential element in radial application. Latching of this signal should
be introduced to remove the risk of current detector signal resetting once the fault is cleared at the
source end.
Figure 26: 50DD SV from remote latched for the 87L supervision logic
It can be argued that the need to confirm that the 50DD from the source end is received for two
consecutive packets appears to be an over design measure which, in the end, could be disregarded.
For simplicity, the 50DD supervision “send and receive” could be implemented to any end of the line
and applied as standard feature for any line differential protection application. In addition, the facility
to use undervoltage detector as 87L supervision should still be maintained for product flexibility
purposes and to allow designers to increase the dependability of the scheme as required.
.
8 Conclusions
Line differential schemes can be implemented in a dependable and secure manner to either
transmission line in a meshed network as well as to radial application. It is obvious that the very
weak or absent current contribution from the remote end of the line presents a menace to the
traditional current detector supervision of the 87L. Especially in the event of low load and a three
phase fault or where the power transformers at the remote end do not provide a zero sequence path.
This paper has investigated several options. Overall, the implementation of the other ends’ 50DD
current detector as part of the packet to be transmitted to the remote end meets the need of
increasing the dependability of the 87L without affecting security, simplicity and speed of operation.
Considering that double circuits, open ring network and radial line with no generation at the remote
end could be exposed to the same issue, the results of this study envisages this option for any
application and any end of the line.
9 References
[1] GE, L90 Differential Relay Manual, 2015
[2] Bogdan Kasztenny, Ilia Voloh: Application of Modern Relays to Dual-Breaker Line Terminals,
GE, 2006
[3] Zhihan Xu, Ilia Voloh, Matt Proctor, Mike Lara:CT Saturation Tolerance for 87L Applications,
68th Annual Conference for Protective Relaying Engineers, GE, 2015
[4] Yiyan Xue, Bogdan Kasztenny, Douglas Taylor, Yu Xia: Line Differential Protection Under
Unusual System Conditions, 39th Annual Western Protective Relay Conference, Spokane,
Schweitzer, 2012
[5] Bogdan Kasztenny, Norman Fisher, Ken Fodero, Adrian Zvarych:Communication and Data
Synchronisation for Line Current Differential Schemes. Schweitzer, 2011
[6] Siemens, 7S5D Differential Relay Manual, 2014
[7] ABB, REL 670 Differential Relay Manual, 2014
[8] Toshiba, GRL100B Differential Relay Manual, 2014
[9] SEL, 411L Differential Relay Manual, 2014
[10] Alstom, P546 Differential Relay Manual, 2014

APS | 2015
About the Authors

Leonardo Torelli (02 December 1971) is a Senior Engineer at CSE-
Uniserve, Melbourne. Leonardo previously worked in consulting firms in
Australia and at ENEL Distribution, Florence, Italy. He currently provides
training and technical assistance for GE relays. He is also involved in
secondary design and teaches Power Protection at Melbourne Polytechnic.
In 2011 Leonardo received a Master by Research on distance protection at
RMIT University, Melbourne. He has presented four papers at protection
conferences. He is a chartered professional engineer in Australia.
Ilia Voloh received his Electrical Engineering degree from Ivanovo State
Power University, Russia. After graduation he worked for Moldova Power
Company for many years in various progressive roles in Protection and
Control field. He is currently an applications engineering manager with GE
Multilin in Markham Ontario, and he has been heavily involved in the
development of UR-series of relays. His areas of interest are current
differential relaying, phase comparison, distance relaying and advanced
communications for protective relaying. Ilia authored and co-authored more
than 30 papers presented at major North America Protective Relaying
conferences. He is an active member of the PSRC, member of the main
PSRC committee and a senior member of the IEEE.
Zhihan Xu received the B.Sc. and M.Sc. degrees in power engineering

from Sichuan University, the second M.Sc. degree in control systems from
the University of Alberta, and a Ph.D. degree in power systems from the
University of Western Ontario. He is a Lead Application Engineer with GE
Digital Energy in Markham. His areas of interest include power system
protection and control, fault analysis, modeling, simulation, and
automation. He is a professional engineer registered in the province of
Ontario.

APS | 2015

APS
APS| 2015
| 2015
Modeling and Testing of High Impedance Differential

Scheme Using Numerical Relays
Zarko Djekic | Ausgrid

Zdjekic@Ausgrid.com.au
1 Introduction
High impedance (HiZ) differential protection scheme has been used for busbar protection over many
decades [1] . The basic idea used in this scheme avoids problems related to CT saturation during
through faults by connecting high impedance in series with the operating relay. Relay with high
impedance burden and CT’s are connected in parallel to force the false differential currents (caused by
CT saturation) to go through the saturated CT’s rather than through the relay. As long as voltage
across summation point required to operate the relay is higher (with a certain margin) than the
maximum voltage developed during maximum through fault, scheme provides adequate security. This
requirement is usually easy to satisfy and very few utilities ever had an issue with scheme security
caused by CT saturation.
On the other hand, tripping for internal faults relies on CTs’ ability to provide enough secondary current
through very high impedance of the relay branch during an internal fault. Under an internal fault
condition, at least one CT with primary current drives secondary current into high resistance of the
relay branch. This creates high voltage across CT terminals. This voltage usually exceeds knee point
voltage of the CT and eventually saturates (from secondary terminal side) other CT’s connected into a
high impedance scheme. Low impedance of saturated CT’s reduce burden on source CT’s; but also
allows spill current to flow into saturated CT’s not into the relay. As a consequence, in steady state,
current through the relay flows only during short periods when voltage across summation point is
below knee point voltage of CT with the lowest knee point voltage - i.e. periods before the smallest CT
saturates.
Depending on fault current and relay impedance, maximum peak voltage developed across CT
terminals is usually much higher than the knee point voltage of CTs. In most instances it is necessary
to limit this voltage across CT to avoid insulation damage on CT terminals, wiring and protection panel.
Non-linear resistors (varistors) are often used to limit voltage below a safe value (usually under 2kV).
When varistor is connected across summation point (across resistor and relay), in addition to reducing
peak voltage it also diverts a portion of the spill current and limits maximum magnitude of the current
seen by relay. It limits magnitude of the spill current spikes but, due to reduction of burden resistance,
makes spill current spikes wider – Figure 1.
Figure 1: A Typical Steady State Differential Current – Internal Fault
Early schemes had been implemented using high (usually nonlinear) impedance voltage relays. These
relays were extremely fast because very sensitive operating coil in these relays (usually 5mA - 10mA)
would operate almost instantaneously when voltage increases over knee point of internal nonlinear

APS | 2015
resistors used to set the voltage pickup. Over time, the voltage relays had been replaced with
electromechanical (EM) instantaneous overcurrent relays (usually armature type) and burden
resistors. More recently, EM relays have been replaced with solid state and numerical overcurrent
relays. General understanding is that HiZ scheme performance is expected to remain satisfactory as
long as few original design “rules of thumb” are met. At the same time requirements for smaller CT
have been intensified due to switchgear size reduction. This paper presents methods and findings of
modelling and simulation of a HiZ differential scheme. Relay performance under internal fault
conditions were studied. Conclusions and some recommendations for scheme improvements based
on this findings are presented.
2 Modelling and Simulations

Analytical analysis of a circuit with multiple nonlinear components is extremely hard, so numerical
model of a high impedance busbar circuit was developed. To predict and test relay performance under
various conditions, a relay model was also developed to speed up the process. All modelling has been
done in EMTP RV software. Simple high impedance busbar scheme comprised of three sets of CT’s
connected in parallel was used – Figure 2.
Figure 2: High Impedance Scheme Model
CT’s are modelled using standard CT equivalent scheme with EMTP-RV inbuilt nonlinear inductor with
hysteresis connected into magnetizing branch – Figure 3. Real CT information has been used to
create model data. Silicon carbide varistor (Metrosil) was modelled as nonlinear resistor using voltage
and current data obtained from high current tests.
Figure 3: CT Model
CT model was verified by high current injection into set of two CT connected into high impedance
scheme. Currents up to 20kA were injected. Scheme with the models of the same CTs were modelled
and simulation results compared to recorded waveforms. Maximum instantaneous error was under 5%
while rms and 50Hz error was under 2%. Relay model was verified by injecting over 30 secondary spill
current waveforms obtained from simulation into four relays from various manufacturers (one armature

APS | 2015
type EM relay, one rms solid state relay and two new numerical relays). Rms and numerical relay trip
times were compared to simulation results. Actual relay trip times were longer by fixed delay
compared to simulated results; this delay was caused by relay processing time and output contact
delay. When adjusted for the fixed time simulation results were within 5ms of actual relay trip times.
Over one hundred internal and dozens of through fault cases were simulated with one and two CTs
connected to primary plant with a source behind. One CT was intentionally left without primary source
to simulate open points or radial load situations. Fault level, primary X/R ratio, CT resistance and knee
point voltage, CT remanence, burden resistor value and relay pickup were varied. Relay tripping times
were recorded and analysed.
3 Results
Two types of simulations were performed: steady state and transient. Protection engineers commonly
use steady state calculations to choose relay and resistor settings and to specify CTs. Steady state
simulations were performed to gain better understanding of the scheme and to explain effect of CT
size and relay setting selection on scheme performance. Transient simulations were used to check
scheme performance under DC transients and CT remanence conditions.
3.1 Steady state simulation results

Under internal fault condition, spill current drops to zero once the first CT in a circuit saturates. This
means that width and magnitude of spill current in steady state is highly dependent on size of the
smallest CT size connected in the scheme. Commonly, a rule of thumb, that CT knee point voltage
has to be at least twice the stability voltage setting, has been used to make sure relay sees enough
current that enables quick operation. Figure 3 shows how rms and 50Hz component of a steady state
spill current depends on CT size. Results have been simulated for a medium level fault (15kA).
`
Figure 4: Steady state spill current (rms and 50Hz component)
At knee point voltage equal to two times stability voltage, both rms and fundamental component of the
spill current are greater than 3 times pickup. High pickup has been used in simulations due to common
usage of CT supervision element. In this cases, high current pickup setting is used to avoid false trips
caused by load current when some CT's are, by mistake, excluded from the scheme (usually links left
open after CB maintenance).

APS | 2015
When smaller CT's are used, fundamental component of spill current can be substantially lower than
rms value. For the smallest acceptable CT's (Vk=2xVs) that difference is over 10%. This doesn't seem
to be an issue for HiZ scheme performance in steady state condition because both measurements are
well above pickup setting. These findings confirmed that common “rules of thumb” have been correctly
calculated for steady state conditions.
3.2 Transient simulation results

Transient behavior of high impedance scheme is almost impossible to calculate analytically due to
multiple nonlinear elements in a circuit. DC offset and CT remanence have significant impact on
current measurements during transients. Faults with high DC offset produce very low and distorted
spill currents during transient period. Depending on primary system X/R ratio, it may take up to several
cycles for spill current to reach steady state value (Figure 5). Due to high distortion of current
waveform during transient, fundamental component of spill current is significantly lower than peak or
rms value. Similar behaviour has been observed for cases with high CT remanence. Spill current has
very low value during transients even in cases when only one CT has high remanence. This is caused
by the fact that once one CT saturates, almost all spill current gets diverted from relay to secondary
winding of that CT. This makes spill current asymmetrical with very short spikes low in magnitude
during transient period. Things become much worse when several CTs have remanence with same
polarity as DC offset.
Figure 5: A typical spill current with DC offset (waveform, RMS and fundamental comp. magnitude)
Distorted transient spill current waveform has very high harmonic content; in some instances
fundamental component during transient is less than 30% of rms value. As the CTs recover from deep
asymmetrical saturation caused by DC offset and/or remanence, even harmonics die-off but odd
harmonics (3rd, 5th …) remain relatively high making fundamental component of the signal
significantly lower than rms [2]. This was usually not an issue in past when rms or instantaneous
relays have been used. When new fundamental component relays are used, this may delay tripping by
several cycles causing system stability issues and damage to the primary plant. This is especially the
case on busbars fed by bank of transformers with high X/R ratios. In some real world case simulations
(Vk/Vs = 2.5, X/R=25, If=10kA), tripping has been delayed by more than 100ms under not very
unusual conditions (CT remanence of 75% in a single CT and full fault current DC offset). Even for
more likely situations (50% remanence and 80% DC offset) tripping may be delayed by more than two
cycles.
4 Conclusions and Recommendations

Based on simulation results it appears that some of the existing busbar protection schemes with
numerical relays set using traditional “rules of thumbs” may be too slow under some fault conditions.
This maybe more the case since high pressure to reduce CT size has been going for some time.
Numerical relays require higher knee point voltage to stability voltage ratio to perform adequately
under all fault conditions. Increase in CT size for numerical relay implementations has also been
recommended by some relay manufacturers [3], [4] (e.g. Schneider/Alstom requires the ratio to be at
least 4). Unfortunately, this may not be an option in many cases because CT replacement is not
possible or it may cost too much. This may not be solution even in some green-field cases due to
switchgear size that prevents installation of larger CTs.

APS | 2015
The ratio of knee point voltage to stability voltage can be improved also by reduction of stability
voltage. Traditional rule of thumb for setting stability voltage was to double calculated steady state
voltage under maximum through fault condition (doubling was done to avoid tripping caused by spill
current with full DC offset. This was important to ensure stability of early instantaneous peak detection
type relays. Fundamental component relays measure significantly lower currents during through faults
conditions leaving space for stability voltage setting reduction. Based on simulations and waveforms
recorded during high current testing, fundamental component relays see about 30% less current than
rms relays and almost 50% less than older peak detection relays. This allows protection engineers to
reduce stability voltage safety factor k from traditional 2 to 1.4 as recommended by some
manufacturers [4] and even further to 1.2. This would improve ratio between knee point voltage and
stability voltage without expensive CT replacement.
It has been also observed that speed of operation of HiZ relay is very dependent on current pickup for
lower fault currents. Tripping time can be significantly longer for fault currents lower than 5 times the
pickup. It is recommended to set relay pickup to at most 1/5 of minimum fault current, in some
instances this may cause CT supervision scheme not work properly.
Another simple way to improve high impedance scheme performance is to modify varistor circuit. In
the older voltage relay schemes, varistor was traditionally wired across summation point (relay
terminals) to protect high impedance relay from overvoltage that develop during high internal faults.
New numerical relays have very low impedance (0.01-0.1 Ohm resistive) so the high voltages do not
develop across relay terminal but only across stabilising resistor. Varistor connected across stabilising
resistor only, still protects the circuit from overvoltage and at the same time it also prevents varistor
from “stealing” some of the spill current away from the relay.
Figure 6: Recommended high impedance varistor connection
In addition to varistor wiring changes, varistor with lower voltage characteristics should be used.
Varistor with lower breakdown voltage further reduces CT burden for internal faults, increases current
through the relay and makes CTs saturate later. This circuit is also better from a failure mode
perspective because it removes danger of relay failing to trip for an internal fault due to short circuit
failure mode of the varistor (the most common failure mode for varistors). Varistor characteristic
should be selected based on maximum leakage current at stability voltage setting. This should not be
greater than 10% of the relay pickup. To avoid high leakage current at lower voltages ZnO varistors
could be used. Separate relay element with lower pickup could be used to supervise health condition
of the varistor. Alarm could be set when leakage current increases following varistor partial
degradation.

APS | 2015
5 References
[1] H. T. Seeley, F. von Roeschlaub; Instantaneous Bus-Differential Protection Using Bushing

Current Transformers, AlEE Trans., 67 (1948), pp. 1709-1718. Discussions, p. 1719.
[2] E.O. Schweitzer, Daquing Hou, Filtering for Protective Relays, 47th Annual Georgia Tech
Protective Relaying Conference, 1993, Atlanta,
[3] Stanley E. Zocholl and David Costello, Application Guidelines for Microprocessor-Based, High-
Impedance Bus Differential Relays, Journal of Reliable Power vol 2 no 3, 2011
[4] Areva - MiCOM P12X High Impedance Protection Application Notes
About the Author
Zarko Djekic received his Dipl. Ing. Degree in electrical and computer
engineering from the University of Novi Sad, Serbia, in 2004 and Master of
Science degree in Electrical Engineering from Texas A&M University,
College Station, TX, USA in 2007. He is currently senior protection
engineer at Ausgrid, Sydney Australia. Previously he was with American
Electric Power Inc. Columbus OH, USA

APS | 2015

APS | 2015

APS
APS| 2015
| 2015
New Design of Distance Protection for Smart Grid Applications
Jörg Blumschein | Siemens AG

joerg.blumschein@siemens.com
Dr. Cezary Dzienis | Siemens AG

Cezary.dzienis@siemens.com
Yilmaz Yelgin | Siemens AG

Yilmaz.yelgin@siemens.com
1 Abstract
Smart grids of the future will have new challenging requirements for the protection elements regarding
selectivity and dependability. The load flow will be increasing, the magnitude and direction of load flow
may be changing frequently and even the network topology will be more complex than today.
This paper presents a new design of distance protection which perfectly fits to the requirements of the
smart grid of the future.
The impedance measurement is based on the calculation of the load compensated fault reactance X
and from line resistance separated fault resistance. This method is applied for phase to ground as well
as phase to phase faults. Separation of the fault resistance improves the accuracy of the impedance
calculation. The method reduces the negative influence of fault resistance during high load flow and
minimizes the risk of wrong pickup during high load condition.
Once a fault is detected it is very important to select the faulted loop to calculate the impedance to the
fault. For complicated faults in the complex network of a smart grid this can be a challenging task with
a certain risk of non-selective fault clearance.
In the past the loop selection was done by a so called decision tree, in which several criteria were
applied sequentially to find the faulted loop. Thereby, only the result from one criterion selects the
faulted loop.
The new approach is different. Several criteria based on magnitudes of voltages and currents,
changes in voltages and currents, symmetrical components or impedances are applied in parallel. The
results of each single criterion are weighted and combined to get a final result for the selection of the
faulted loop. With this principle the efficiency of the loop selection has been optimized to different
network topologies by changing the weights of each criterion.
The same principle is applied to the directional element. Multiple criteria based on actual voltages,
memorized voltages, symmetrical components or delta quantities are applied in parallel. The final
result is obtained as a weighted combination of the result of each single criterion.
2 Introduction
Distance protection is used worldwide to protect the lines for the transmission and distribution of
electrical energy against consequences of electrical faults. The distance protection has to detect these
faults and initiate a trip command to isolate the faulted line.
These faults can be phase to phase which are short circuits between different phases or phase to
ground which are short circuits between one or more phases and ground.
In most cases there will be an arc flash between the faulted phases or between phases and ground.

APS | 2015
Figure 1: Basic principle of distance protection
Figure 1 illustrates the basic principle of distance protection. The distance protection D determines the
fault impedance ZF from the voltage UA and the current IA measured at the relay location according to
Ohm's law:
UA
ZF = (1)
IA
The measured fault impedance ZF will be compared afterwards with the so called zone setting
resulting from the line impedance. If the fault impedance is less than the configured setting the fault is
on the line. In this case a trip command is issued to the local circuit breaker CB to isolate the faulted
line from system operation as shown in figure 1.
time
Z3
Z2
Z1 t3
t2
D1 D2 D3
distance
Figure 2: Principle of stepped distance protection
Figure 2 shows the principle of a stepped distance protection. For the selectivity it is important that
only the faulted line is separated from operating network. That means that the relay D1 in figure 2
should only issue a non-delayed trip command if the fault is inside the protected zone Z1. Due to some
uncertainty of the network parameters as well as measurement errors, this zone is normally configured
to 80% of the protected line. Faults on adjacent lines should be tripped by the related relays D2 or D3
in figure 2. The relay D1 works only as a backup function for faults on these lines.
Most faults on transmission and distribution lines are not pure metallic faults but faults with a so called
fault resistance like shown in figure 3. This fault resistance can be an arc flash for phase to phase
faults or the combination of an arc flash and the grounding resistance for phase to ground faults.
In reality the model of an arc flash is much more complicated but for simplicity it will be modeled as a
fault resistance RF in this paper.
Figure 3: Single phase diagram for a resistive fault
In case of faults including resistance RF the equation for the faulted loop can be obtained based on
figure 3 as follows:
U A = (m ⋅ Z L + RF ) ⋅ I A (2)

APS | 2015
According to (2) the impedance measured by the distance protection consists of two parts. The first
part m*ZL is the impedance of the line between the relay and the point of the fault. The second part RF
is the fault resistance, representing the resistance of the arc flash for phase to phase faults.
Figure 4 shows the graphical representation of both impedances in the complex plane.
The phasor of m*ZL is shown at the line angle which is close to 90° for high voltage overhead lines.
The phasor representing fault resistance RF is added in parallel to the R-axis.
X
RF
F
R
L
m* Z
ZL +
*
m
R
Figure 4: Graphical representation of a resistive fault in the complex plane
Depending on the fault position m the fault impedance m*ZL can vary between zero for faults close to
the relay and 0.8*ZL for faults at the end of the protected zone. The polygonal characteristic is the best
choice because this characteristic has a constant reach in R-direction for all faults on the protected
line.
Figure 5 represents a typical polygonal characteristic of a stepped distance protection. According to
the principle explained in figure 2 the red marked zone Z1 is responsible to detect faults up to 80% of
the protected line. In this case a non-delayed trip command is issued. The blue marked zone Z2 has
the same reach in R-direction but can detect faults up to 150% of the protected line. Normally zone Z2
is delayed by 300 ms to have a backup function. The green marked zone Z4 is a backup zone for the
reverse direction and the black marked zone Z5 is another unidirectional backup zone.
Figure 5: Typical polygonal characteristic of a stepped distance protection
The distance protection has the great advantage of selectivity which can be achieved by local
measurement only without any communication.
3 Impedance measurement on heavy loaded lines

The classical impedance measurement according to (2) is only accurate for lines with single ended
infeed as well as without significant load.
Figure 6 shows the single line diagram for a resistive fault, feeded from both ends of the line. In this
case the fault current IF is the sum of the fault current IA from the local end and the fault current IB from
the remote end.

APS | 2015
Figure 6: Single line diagram for a resistive fault, feeded from both ends of the line
The local relay at bus A can only measure the current IA from the local end. But with regard to figure 6
the fault current from the remote end causes an additional voltage drop at the fault resistance RF.
U A = (m ⋅ Z L + RF ) ⋅ I A + RF ⋅ I B (3)
This additional voltage drop caused by the fault current from the remote end has an impact on the
accuracy of impedance measured by the distance protection at bus A. This voltage drop results in an
additional impedance component which depends on the relation between the local fault current IA and
the fault current IB from the remote end.
IB
Z A = (m ⋅ Z L + RF ) + RF (4)
IA
According to (4) this additional impedance is increasing with increasing fault current contribution from
the remote end. The additional impedance can have a reactive component if the fault currents IA and IB
have different angles. This reactive component will result in a measuring error ∆X if the classical
impedance measuring method (2) is applied in distance protection:
IB
∆X = ⋅ RF ⋅ sin(∠( I B , I A )) (5)
IA
Transmission of active power requires a phase difference between the voltages of the equivalent
sources at sending and receiving end. During load flow the phase angle of the sending end leads the
phase angle of the receiving end. In case of a fault the phase angles of the fault current contributions
from both ends are related to the phase angles of the feeding voltages in a first approach.
Figure 7 illustrates the influence of the load flow on the calculated impedance of a resistive fault using
the classical impedance calculation method. In case of load export the phase angle of IA leads the
phase angle of IB. According to (5) ∆X becomes negative for a normal load angle of 30° which results
in an overreach of the distance protection.
Figure 7: Influence of the remote end infeed on the impedance calculation

of a resistive fault using the classical impedance calculation method

APS | 2015
A simulated case from the network of Colombia is used to explain the opposite case. A fault AB-N with
5 Ω fault resistance is applied at 50% line length of a heavy loaded line. This case was tested with a
distance protection relay using the classical impedance calculation method.
Figure 8: Fault record of a resistive fault at 50% of the line length which is
not seen in Z1 using the classical impedance calculation method
Referring to figure 8 there is a load import of 122 MW before the fault occurred. At fault inception the
load flow is changing the direction to feed the fault from the local source. The binary signals indicate
that this fault was not seen in zone Z1 even if the fault was applied at 50% of the line length, whereas
the reach of zone Z1 was configured to 80% of the line length. A trip command was issued by the
overreach zone Z1B supported by the receive signal from teleprotection. Figure 9 illustrates the
measured impedance in the complex plane. The fault was applied at 50% of the line length but
according to the principle explained in figure 7 the measured impedance is located outside the red
marked zone Z1 which is configured to 80% of the line length.
Figure 9: Impedance plane of a resistive fault at 50% of the line length which is
not seen in Z1 using the classical impedance calculation method
According to (5) the phenomenon of overreach or underreach for resistive faults only effects heavy
loaded lines with infeed from both ends. In the past this effect was compensated by changes to the
characteristic of the zones.
Today and in the smart grids of the future such a static compensation is not sufficient because the
load often changes in magnitude or direction.

APS | 2015
4 Impedance measurement with load compensation

The general idea of the impedance measurement with load compensation is the elimination of the
reactance measurement error ∆X given in (5). According to figure 6 equation (3) can be written in the
following form:
U A = m ⋅ Z L ⋅ I A + RF ⋅ I F (6)
To eliminate the reactance measurement error ∆X caused by the fault resistance RF multiplied with the
complex fault current IF equation (6) will be multiplied with a compensation quantity which is the
conjugate-complex value of the fault current IF. This compensation quantity consists a compensation
current ICmp which should be already very closed to IF and a compensation angle δCmp for the final
adjustment.
∗ − jδ Cmp ∗ − jδ Cmp ∗ − jδ Cmp

U A ⋅ I Cmp ⋅ e = m ⋅ Z L ⋅ I A ⋅ I Cmp ⋅ e + RF ⋅ I F ⋅ I Cmp ⋅ e (7)
Due to the multiplication of IF with its conjugate-complex replacement the term RF*IF*I*Cmp*e-jδCmp
becomes a real value. For the calculation of the fault reactance only the imaginary part of equation (7)
needs to be considered.
[ ∗
Im U A ⋅ I Cmp ⋅ e
− jδ Cmp
] = Im[m ⋅ Z L
∗
⋅ I A ⋅ I Cmp ⋅ e
− jδ Cmp
] (8)
Solved for X which is the imaginary part of m*ZL we get the following equation.
sin (ϕ ) ⋅ Im[U A ⋅ I Cmp ⋅ e

∗ − jδ Cmp
]
X= jϕ ∗ − jδ Cmp
(9)
Im[e ⋅ I A ⋅ I Cmp ⋅e ]
Figure 10 shows the current flow for a single phase to ground fault in the sequence network.
Figure 10: Sequence network for a single phase to ground fault

APS | 2015
It is shown, that the zero sequence current I0 or the negative sequence current I2 can be used as a
compensation current for the fault current IF because they are not influenced by load flow.
In case of different impedance angles in the zero or negative sequence network a final adjustment can
be done using the following formula as example for the zero sequence.
 Z 0 , A + Z 0, B + Z 0,L 
δ Cmp,0 = arg  (10)
 (1 − m) ⋅ Z 0,L + Z 0,B 
Using this compensation angle δCmp,0 the measured zero sequence current I0,A can be adjusted to the
angle of the fault current IF if the zero sequence impedances Z0,A, Z0,B and Z0,L have different angles.
5 Impedance measurement for phase to phase loops using reactance

method
The basic idea to compensate the influence of load flow can be extended to phase to phase faults.
Figure 11 shows the current flow for a phase to phase fault with the fault resistance RF in the
sequence network.
Figure 11: Sequence network for a phase to phase fault
For a phase to phase fault the negative sequence network is connected in parallel to the positive
sequence network. Since the zero sequence component does not exist for a phase to phase fault and
the positive sequence network includes the sources only the negative sequence current can be used
as compensation quantity.
The determination of the compensation quantity for phase to phase faults will be explained using an
AB fault as example. For an AB fault the fault current IF is the difference between the current of phase
A and the current of phase B.
IF = IA −IB (11)
Replacing the phase currents by the sequence components we get the following result:
I A = I1 + I 2
(12)
I B = a ⋅ I1 + a ⋅ I 2
2

APS | 2015
For the AB fault the relation between positive sequence current and negative sequence current is
given as:
a ⋅ I 1 = −a ⋅ I 2
2
(13)
Considering this transformation factor the compensation current for the AB fault based on the negative
sequence current can be derived.
I Cmp, AB = 2 ⋅ (1 − a) ⋅ I 2 (14)
For other fault types the compensation quantity will be determined accordingly.
In case of different impedance angles in the negative sequence network a final adjustment can be
done using the following formula.
 Z 2 , A + Z 2,B + Z 2 ,L 
δ Cmp, 2 = arg  (15)
 (1 − m) ⋅ Z 2,L + Z 2,B 
The exact reactance for a resistive fault AB on heavy loaded lines can be calculated using the
following equation.
sin(ϕ ) ⋅ Im[U AB ⋅ ( I Cmp, AB ⋅ e

jδ Cmp , 2 *
) ]
X= jϕ jδ Cmp , 2 *
(16)
Im[e ⋅ ( I A − I B ) ⋅ ( I Cmp, AB ⋅ e ) ]
This method was applied to the fault explained in chapter 3. As shown in figure 12 the fault is seen in
zone Z1 using the impedance calculation based on the reactance method.
Figure 12: Impedance plane for a resistive fault at 50% of the line length which is seen in Z1 using
impedance calculation according to the reactance method
Consequently with the new method of reactance compensation for resistive faults on heavy loaded
lines a trip command in zone Z1 for the fault simulated for a network in Colombia is issued as shown in
figure 13.

APS | 2015
Figure 13: Fault record of a resistive fault at 50% of the line length which is seen in Z1
using impedance calculation based on the reactance method
According to [2] the method of reactance compensation can also be applied for 3 phase fault using the
adequate compensation quantity.
6 Multi-criteria loop selection

The proper selection of the faulted loops is a precondition for each distance protection algorithm.
Different criteria are applied to select the faulted loops like:
• magnitude of currents and delta currents

• magnitudes of voltages and delta voltages
• impedances
• Symmetrical components.
In the past these different criteria were applied in form of a decision tree like illustrated in figure 14.
true false
true false
true false true false
Figure 14: Decision tree as classical structure for the loop selection
This decision tree was optimized to get the final result as fast as possible. Due to the structure of the
decision tree not all criteria are evaluated to get a final result. From figure 14 it is obvious that a wrong
loop can be determined if only one criterion give a wrong result. Sometimes the result of a single
criterion is not very strong but in the decision tree it must be reduced to binary information. Additional

APS | 2015
to this it was shown by simulations and analysis of real events that different criteria have different
importance for different network topologies.
All these considerations lead to another structure of the loop selection like shown in figure 15.
Figure 15: Multi-criteria loop selection
With the new concept all criteria of loop selection are applied in parallel creating the so called multi-
criteria loop selector. Each criterion can have an output quality which is a value between 0.0 which
means the criterion is not fulfilled and 1.0 which means the criterion is completely fulfilled.
Figure 16 gives an example of the output quality of the voltage criterion. A strong voltage drop in case
of a fault results in a high quality of the voltage criterion for the related loop. If the voltage drop is only
marginal the quality of the voltage criteria of the related loop will have a low quality.
Figure 16: Quality of criteria voltage
Figure 17 illustrates the output quality of the impedance criterion. If the measured loop impedance is
closed to the zone limit the quality of the impedance criteria is low. However if it is clear that the
impedance is in the zone the related quality is high.
Figure 17: Quality of criteria impedance
According to equation (17) a final quality Qloop for each loop will be calculated. The weighted quality
outputs of N criteria qi,loop will be summated to get the final quality Qloop.

APS | 2015
N
Qloop = ∑ qi ,loop ⋅ xi (17)
i =1
Figure 18 gives a graphical explanation of equation (17). It is shown that each criterion is weighted
according to its significance.
Figure 18: Calculation of loop quality
It was found by simulations and analysis of real events that depending on network topologies different
criteria have different significance.
Figure 19 gives an example for the best adaption of the weights for a network with strong infeed
compared to a network with weak infeed. It is obvious that the current based criteria have a higher
weight in a network with strong infeed however the voltage based criteria have a higher weight in
networks with weak infeed. By implementing the weighting factors between the boundaries shown in
figure 19 the loop selection is optimized for all system conditions.
Figure 19: Weights of multi-criteria loop selector depending on different infeed
Figure 20 illustrates the operating principle of the multi-criteria loop selector. After fault inception the
qualities for all loops are calculated. A loop is selected if the quality of the loop extends an adaptive
threshold like shown in figure 20.

APS | 2015
iA iB iC
I [A]
3
-1
-2
0.74 0.75 0.76 0.77 0.78 0.79 0.80 0.81 0.82 0.83 0.84 0.85 0.86 0.87
t/s
vA vB vC
V [V]
50
-50
0.74 0.75 0.76 0.77 0.78 0.79 0.80 0.81 0.82 0.83 0.84 0.85 0.86 0.87
t/s
quality AG quality BG quality CG quality AB quality BC quality CA quality ABC

%
75
50
25
0.74 0.75 0.76 0.77 0.78 0.79 0.80 0.81 0.82 0.83 0.84 0.85 0.86 0.87
t/s
Figure 20: Operation of the multi-criteria loop selector
7 Conclusions
It was shown that the reach of the classical impedance calculation method is significantly influenced
by resistive faults on heavy loaded lines. Using the reactance method this reach error can be
eliminated. Additionally a new method of loop selection was presented which is optimized for all
network topologies. The same philosophy is applied for directional element where different algorithm
are weighted and depended on network topology.
8 References
[1] G. Ziegler, “Numerical Distance Protection: Principles and Application”, Siemens
Aktiengesellschaft, 4th Edition, ISBN 978-3-89578-381-4, 2011.
[2] C. Dzienis, Y. Yelgin, G. Steynberg, M. Claus, “Novel Impedance Determination Method for
Phase-to-Phase Loops”, 18th Power Systems Computation Conference, Wroclaw, Poland,
August 2014
[3] S. E. Zocholl, “Three phase circuit analysis and the mysterious k0 factor”, 22nd Annual
Western Protective Relay Conference, Spokane, Washington, October 2011.
[4] E. O. Schweitzer and J. Roberts, „Distance relay element design“, 46th Annual Conference for
protective Relay Engineers, Texas A&M University College Station, April 1993 .
[5] M. M. Saha, J. Izyskowski and E. Rosolowski, “Fault location on Power Networks”, London:
Springer Verlag, 2010.
[6] L. Eriksson, M.M. Saha and G. D. Rockefeller, “An accurate fault locator with compensation
for apparent reactance in the fault resistance resulting from remote-end infeed”, IEEE
Transactions on Power Apparatus and Systems
[7] Manual SIPROTEC 5 Distance Protection, Line Differential Protection, and Breaker
Management for 1-Pole and 3-Pole Tripping 7SA87, 7SD87, 7SL87, 7VK87, Version V6.00
and higher, SIEMENS AG, C53000-G5040-C011-6

APS | 2015
About the Authors
Jörg Blumschein studied technical cybernetics and process measurement

at the University Magdeburg where he became a graduated engineer in
1992. Since 1992 he works with SIEMENS in the development department
of protection relays. Today he is the Principal Key Expert for Protection.
Cezary Dzienis graduated with a degree in Electrical Engineering from

Warsaw University of Technology in 2003. He then worked for university’s
Division of Industrial Electronics and Control System until 2004. Since then
he worked for Chair of Electric Power Networks and Renewable Energy
Sources at the University Magdeburg where he completed his Ph.D. in
2007. Since June of 2008 he has been a researcher in area of protection
algorithms by Siemens Berlin.
Yilmaz Yelgin studied computer science at the technical university in

Berlin and graduated in 2004. Since 2006 he works with Siemens in the
development department of protection relays. His main field of activity is
the analysis of reasons and effects of power swing.

APS | 2015

APS
APS || 2015
2015
Communication Aided Protection Schemes and their

Implementation and Testing with State-of-the-Art Technologies
Fred Steinhauser | OMICRON

fred.steinhauser@omicron.at
1 Introduction
Electrical power systems and the related protection and control systems evolve over time. Systems
get extended, and the existing parts have to be adapted to the new configuration. While extending the
primary system is more straightforward in most cases, the adaption of the protection system can be
more challenging if the topology of the primary system is essentially changed. The adaption of
protection settings is not sufficient to cope with such situations. New protections concepts or improved
information exchange between protection relays may be required.
This paper describes an electrical power system with five terminals that evolved over time and how
recent technologies were applied for upgrading the protection system and how the testing of the
system was performed.
2 The System and its Evolution

The electrical power system is located in the province of Styria in Austria. It is a 110 kV double line
system that spans over just a few dozen kilometers, but has five terminals of different nature.
Figure 1 shows the single line diagram of the system as it was drawn in the software that was used for
testing the system.
Figure 1: Single line diagram of the power system
The five terminals included in the scope are the Friesach Substation (FSS), the Laufnitzdorf
Substation (LAU), the Frohnleiten Substation (FRO), and the feeders Friesach Hydro Power Station
(FHP) and Deutschfeistritz (DFZ). As can be seen, the two feeder FHP and DFZ are only tapped on
line B of the double line system.
The buses left of FSS and right of LAU are also modeled to allow the placing of faults outside the
protected scope to test the stability of the protection during through faults.
2.1 The Evolution over Time
The line between FSS and LAU was built around 1990. According to the state-of-the-art at that time,
the protection in the two substations was equipped with static relays. The substation FRO did not exist

APS | 2015
as an explicit substation at that time. FRO was a bus "embedded" into the line. It had no own
protection relays and no circuit breakers, only disconnector switches were installed.
The feeder DFZ is a pure load and was tapped on line B of the double line system.
The power plant FPH was put into operation in 1998. The corresponding feeder was also tapped on
line B only.
2.2 The Communication System
The communication infrastructure at the time of the erection of the line was an SDH system. The
protection relays had no communication interfaces, the data were only exchanged via I/O contacts.
Dedicated teleprotection devices were used to convert the binary I/Os and to transfer the according
information over the SDH network. Just four bits in each direction could be transmitted and received
by such a device.
The solution was expensive and limited in features. Recently, the maintenance of this teleprotection
system became more and more a problem, also because it replacement parts were no longer
available.
2.3 The Protection System
The protection was performed by the distance relays in the substations FSS and LAU. A directional
comparison and transfer trip scheme assured fast fault clearing when the fault was seen in forward
direction on both sides. Depending on this protection decision, transfer trip commands were also send
to the breakers in FHP and DFZ.
3 A Major Modification
In 2014, the "embedded bus" in FRO was upgraded to a real substation. This led finally to the topology
shown in figure 1. Therefore, FRO obtained an own protection system that had to be integrated into
the remaining system.
3.1 The New Communication System
Around 2012, the communication infrastructure was upgraded with an MPLS system. This was a
precondition for the application of new communication techniques between the protection relays. The
MPLS system allows to engineer so-called layer-2 paths which can transfer IEC 61850 GOOSE
messages between substations. The network between the substations appears like a large local area
network, where such layer-2 messages can be normally exchanged. This transparent behavior of the
network essentially simplifies the implementation of the protection communication.
Although the utility itself has no explicit plan to deploy IEC 61850, the information exchange between
the protection relays in FSS, LAU, and FRO is now performed via GOOSE messages. The coverage
of the MPLS communication is indicated in figure 2.
Figure 2: MPLS communication with paths for transmission of IEC 61850 GOOE messages

APS | 2015
Fast binary I/O to GOOSE converters were wired up to the protection relays instead of the former
teleprotection signaling terminals. With the required paths provided by the IT department, the
protection specialist could easily configure the GOOSE communication by themselves.
The communication to the remaining two feeders FHP and DFZ is still based on the old system. The
dismantled teleprotection signaling terminals from the other locations serve as replacement parts to
keep this system alive. But the two feeders communicate only with the substation LAU as indicated in
figure 3.
Figure 3: Conventional communication between and the feeders FHP and DFZ
LAU serves as a kind of "master substation". In LAU, the transfer trip commands for FHP and DFZ is
generated by OR-ing the directional information from FSS, LAU and FRO and are then transmitted to
the feeders.
3.2 The New Protection System
The configuration of the distance protection in FSS and LAU had not to be changed, also because
changing parameters on static relays (which are partly set by external resistors) is complicated. But a
possible overreach of the distance protection, e.g. in LAU, is not so crucial since the main protection
decision is based on the directional comparison.
By upgrading FRO to a "real" substation, protection had to be installed there as well. The selection of
the protection also had to do with the conditions in the network "behind" FRO. Figure 4 shows some
details about this.
Figure 4: Weak infeed at FRO
Transformer T2 has a rather high relative short circuit voltage of 15%. Thus, the infeed from the 6kV
system from an industrial plant does not essentially contribute to the short circuit power when a fault
takes place on the double line.
But for the directional comparison scheme to work, a reliable behavior in FRO had to be assured. For
this reason a protection relay with weak infeed detection was installed in FRO. When it detects a weak
infeed condition, it "mirrors" the forward pickup information coming from the relays in FSS and LAU to
the other ends. FSS and LAU will then trip their breakers instantaneously and the transfer trip for FHP
and DFZ will be issued. Assumed that the fault was on line B and the corresponding breakers in the

APS | 2015
other feeders are tripped, only FRO remains feeding the fault, but now with a strong infeed via line A,
thus FRO will then trip as well.
The completion of the new system was the trigger for a final test of the overall protection system with
an advanced distributed protection testing and recording system.
4 Testing
4.1 The Test Topology
To test this distributed system, a distributed test system was set up. The assignment of the involved
equipment is indicated in figure 5.
Figure 5: Test equipment at the locations involved in testing
In each of the five involved locations, a protection test was placed for injecting test quantities and for
recording the reaction of the protection relays. In the three locations FSS, LAU, and FRO, which
communicate with each other through the MPLS network, additional equipment was used. First of all,
there are the ISIO 200 binary I/O to GOOSE converters, which are permanently installed and serve as
teleprotection interfaces.

APS | 2015
Figure 6: Teleptotection interface and test equipment in FSS
To observe and record all GOOSE data exchange and also other binary signals related to the
teleprotection, A DANEO 200 measurement device was used in each of the three locations. What is
not depicted in figure 5 are the GPS clocks in each location that were used to precisely time
synchronize all the test and measurement equipment.
Figure 7: GPS receiver and grandmaster clock for synchronizing the equipment in FHP
The entire test and measurement process was centrally controlled from FRO. There were the main
control PCs with the software for controlling the distributed protection testing and for controlling the
measurement devices. The test equipment in FSS and LAU could be directly accessed through the
MPLS network, the test sets in FHP and DFZ were accessed via the internet through the UMTS
network.

APS | 2015
4.2 Performing the Tests

The simulation based testing software made the process extremely efficient. The power system was
modeled once and the faults scenarios were added. The user does not need to calculate any test
quantities. The simulation calculates the test signals at all injection points in the distributed locations
from the network and fault parameters. The calculated signals are distributed to the connected test
sets and the injection is started time synchronized in all locations. All of this happens without explicit
user intervention.
Figure 8 shows where in the system faults were simulated to test the protection system
Figure 8: Fault locations for the different test cases
Of course, not all of these faults were applied at the same time. Fault locations 3 and 5 are outside of
the protected scope and are used to check the stability of the protection system during through faults.
Fault locations 7 and 8 were used together in a test case to simulate an evolving cross-country fault.
For some fault locations, multiple tests were performed with different fault types or different breaker
positions that affect the topology and thus the fault quantities.
Figure 9: Central test control in FRO

APS | 2015
The GOOSE Communication

CMC test sets can also subscribe to GOOSE messages and measure the reception time of such
messages relative to the fault inception. This was used for the protection test as such. But for
assessing the MPLS communication system and the performance of the teleprotection components,
the GOOSE traffic was explicitly recorded with the DANEO 400 devices. In addition, several binary
signals related to the protection communication were also measured by the DANEO 400 units.
The performance of the new communication system met the expectations and showed to be superior
to the old system. In addition, more information can now be transmitted between the relays. Changes
and extensions can be performed by the protection specialists without the involvement of the IT
department in every detail.
5 Conclusions
New technologies such as MPLS networks and IEC 61850 GOOSE messaging open new possibilities
for implementing communication aided protection schemes. Smart interface equipment allows
integration of existing protection relays in such schemes.
Thorough testing of a complex distributed protection system can be only done with a distributed test
system that covers the protection and the communication aspects together. Realistic test quantities
calculated from a network model give the highest confidence that the protection system will perform as
expected also in case of a real fault.
About the Author
Fred Steinhauser was born in Austria. He studied Electrical Engineering at

the Vienna University of Technology, where he obtained his diploma in
1986 and received a Dr. of Technical Sciences in 1991.
In 1998 he joined OMICRON, where he worked on several aspects of
testing power system protection. Since 2000 he worked as a product
manager with a focus on substation communication issues. Since 2014 he
is responsible for Power Utility Communication business of OMICRON.
Fred Steinhauser is a representative of OMICRON in the UCA International
Users Group. As a member of WG10 in the TC57 of the IEC he contributes
to the standard IEC 61850. He is one of the main authors of the UCA
Implementation Guideline for Sampled Values (9-2LE). He is also a
member of SC B5 of CIGRÉ and contributed to the synchrophasor
standards IEEE C37.118.1 and IEEE C37.118.2.

APS | 2015

APS
APS| 2015
| 2015
Implementation of Multi-Intertripping Protection Schemes in a

Complex EHV Industrial Substation
Maty Ghezelayagh | TasNetworks Pty Ltd

maty.ghezelayagh@tasnetworks.com.au
Chris Simmons | Entura, Hydro Tasmania

chris.simmons@entura.com.au
Abstract
This paper describes the design, installation and commissioning of a new numerical protection and
control and intertripping scheme at an EHV substation. The substation supplies a critical and large
industrial plant (paper mill). The connection consists of two radial 110 kV transmission lines
connecting four 25 MVA, 110/6.6 kV transformers, two of which are dual wound. Each transmission
line is directly connected to a transformer. The 6.6 kV sides of the transformers are connected to
different busbars via circuit breakers. For a fault on the 110 kV lines, or the transformers, the remote
end 110 kV circuit breaker and appropriate 6.6 kV circuit breakers must open.
This arrangement required complex intertripping schemes between protection devices of the
transformers within the substation and between the protection of the local substation and remote
substations.
The secondary work at the local, and remote EHV substation included installing new 110 kV
transmission line protection schemes, consisting of Main ‘A’ and Main ‘B’ new numerical relays for
each line. This included removal of existing inter-tripping schemes.
The project necessitated an innovative design, settings and logic programming to be implemented on
new relays. Existing secondary infrastructure including copper cables and intra substation fibres of the
single old protection scheme, were utilised for the now duplicated transmission line protection. This
involved the following.
• Removal of the existing multi-intertripping schemes and reconfiguration using new numerical
protection scheme with appropriate setting and logic programming.
• Utilising fibre optic cable and digital transceivers/receivers for provision of trip/close circuits.
• Implementing a reliable design and settings based on sound standard policy/philosophy.
• Installing, testing and commissioning the new protection and intertripping schemes with
minimum plant outages due to load critically.
Finally the paper discusses how the implementation of the project involved innovative solutions so that
new primary equipment and secondary copper cabling was not required while providing high system
reliability and security
1 Introduction
For many electricity companies, the drives for replacement of old electromechanical and static relays
with numerical relays has been mainly for enhancement of system reliability/security, reduction in
maintenance costs, and implementation of complex functionalities. However, in order to justify the
replacement of old secondary infrastructure with new technology it has been necessary to carry out
detailed condition assessment and risk analysis.

APS | 2015
One of the Main challenges for the replacement of old protection and control schemes with numerical
relays in brownfields substations, has been lack of sufficient secondary copper cables for new relays.
The Main reasons have been due to the following facts:
1. Numerical relays are programmed for other complex functionalities such as control, interlocking
and intertripping schemes in addition to protection functionalities. This requires new protection
and control panels to obtain significant numbers of signals from primary equipment as
compared to old ones; and
2. In the past many primary assets used to be protected only by one Main fast protection or by
simple protection devices. New national and international electricity standards/rules [1] requires
that each major primary asset should be protected by two Main protection devices capable of
fast fault clearing time within a specified time.
One obvious solution to solve the above challenge is to install new cables from switchyard to control
rooms, but this has proved to be costly and risky. In this paper, an efficient solution using fibre optic
instead of copper cable has been proposed and implemented.
This paper also demonstrates a practical case for justification of replacement of protection and control
of 110 kV transmission lines of a complex EHV industrial substation. The Main differences of this case
with other conventional transmission line protection upgrades is a maximum utilisation of logic
programming of new numerical relays for implementation of complex multi-intertripping protection
schemes within the industrial and remote substations. It has shown with the protection upgrade not
only an increase in reliability and security of 110 kV transmission has been achieved but also
significant simplicity for the operation and fault findings of intertripping schemes.
2 System configuration and control room locations

The single line diagram of the EHV industrial substation (paper mill) has been shown in Figure 1. The
connections between transmissions to distribution levels are via four 110 kV/6.6 kV, 25 MVA
transformers. Two transformers have dual secondary windings. Each 110 kV Transmission Line is
connected directly to two transformers as a tee-off without 110 kV transmission line CB and busbar.
The secondary windings of transformers supply three 6.6 kV busbars. Each 6.6 kV busbar is a single
bus bar arrangement with a bus coupler or section isolator. Each section is connected to a transformer
which is supplied from different 110 kV transmission lines. The total load of synchronous and induction
motors connected to 6.6 kV busbar is nearly110 MW (122 MVA at 0.9 pf). The size of the motor varies
from 1 to 12 MW with 6.6 kV voltage level. All customer loads are connected to 6.6 kV buses via
cables.
The substation has four different control rooms as shown in Figure 2. The protection and 6.6 kV
switchgear for T1 and T2 transformers, are located in switchhouse A (SWH A) and for T13 and T14 in
switchhouse B (SWH B). The old 110 kV TL protection panel and communication panels were located
in control room C. The location of the new 110 kV TL protection panels are located in control room D.
Copper cable is run from control room C to SWH A and B for tripping 6.6 kV CBs of TFs. Copper cable
also is run between the two SWHs A and B for protection intertripping between transformers. Fibre
cables are run between all control rooms for SCADA and monitoring purposes.
With this arrangement it is observed that due to the lack of 110 kV CBs, for fault on each TF, other
TFs should trip in order to remove fault contribution from both lines. In addition for faults on each of the
110 kV lines, 6.6 kV CBs of two associated transformers should trip. The details of these have been
described below:
The remote 110 kV substation is connected to this industrial substation with double busbar
arrangement. Each line connected to remote substation including the 110 kV lines connected to the
industrial substation has its own CB and each bay can be transferred to bus A or bus B via appropriate
isolator switchings.

APS | 2015
Figure 1: Single Line Diagram of the Industrial Substation.
Figure 2: Control room layout and location of Protection and Control panels at Industrial Substation
3 Old Protection Schemes at EHV Industrial Substation

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec ipsum enim, sodales dignissim
vulputate ac, laoreet a odio. Donec id nibh dui, rutrum euismod massa. Duis sed nibh sit amet urna
scelerisque pulvinar. Ut dapibus augue eget sapien molestie dictum.
3.1 Protection of 110 KV Transmission Lines

The old protection relays of each of 110 kV line from industrial to remote substation consisted of:
• One set of old current differential relay with fibre optic communication channel (this
tripped remote CB and respective TF 6.6 kV CBs at industrial sub)
• One old static distance relay at remote substation. (This tripped only CB at remote end.)
The old current differential protection panel was located in control room C in Figure 2.

APS | 2015
3.2 Protection of 110 kV/6.6 kV Transformers and 6.6 kV busbars and feeders
The 110/6.6 kV transformer protection consists of one set of static type relays for differential, restricted
earth fault and overcurrent and earth fault. The transformer guard protection includes, winding/oil
temperature, Buchholz (Main/tap changer) and pressure relief. The replacements of existing
transformer’s protection were not as part of this project but it is planned to be replaced with new
numerical relays within the next 3 years.
Each 6.6 kV CB on the LV side of the transformers have overcurrent and earth fault with static type
relays. Protections of T13 and T14 are located in SWH B and for transformers T1 and T2 in SWH A as
shown in Figure 2. The two switchhouses are separate from each other by a distance of 200 metres
with very limited room in existing cable ducts and trenches.
Each 6.6 kV busbar has one electromechanical high impedance unit busbar protection and each
6.6 kV feeder has overcurrent and earth fault (IDMT) protection with static type relay. Some of 6.6 kV
feeders have also current differential protection for fast tripping and for others instantaneous
overcurrent and earth fault in addition to IDMT element but instantaneous element is blocked by
protection devices at the customer site via signalling scheme
4 Old intertripping schemes between industrial and remote

substations
Due to absence of 110 kV TL CBs at the industrial substation, several intertrip schemes exist between
industrial and remote substations as follows:
• Intertrip scheme from remote to industrial sub for TL fault
• Intertrip scheme from remote to industrial sub when TL CB at remote sub is opened
• Intertrip scheme from industrial to remote sub for TF fault
The purpose of each scheme has been mentioned below:
4.1 Intertrip Scheme from Remote to Industrial Sub for TL fault

For TL fault, line current differential protection trips line CB at remote end and respective TF 6.6 kV
CBs at industrial sub (eg for fault on Line No 1 in Figure 1, 6.6 kV CBs of T1 and T13 trips).
In order to provide back-up to current differential relay for fault on TL, a distance relay in addition to
current diff exists only at the remote substation which trips only line CB at the remote end. However in
order to remove the fault contribution from healthy line via TFs, the distance relay will send an intertrip
signal over radio communication link to industrial sub via digital transceiver/receiver devices to trip
respective 6.6 kV CBs of TFs.
4.2 Intertrip Scheme from Remote to Industrial sub when TL CB at Remote sub is opened
In the past it was found that when one of the remote 110 kV TL circuit breakers is opened (not due to
fault); the transmission line is kept energised via generation from motors at the industrial plant. This
caused low frequency and resulted in the operation of the Under Frequency Load Shedding (UFLS)
scheme at a few substations. To prevent this, an auto opening intertripping scheme was installed to
open respective 6.6 kV breakers if any line CB at remote end is opened. The opening commands are
initiated by auxiliary contacts of the CB at the remote end and are sent via transmission current
differential relay (transfer trips) to trip respective 6.6 kV CBs. For example with the opening of the CB
of TL#1 at the remote end, the intertrip scheme trips also T1 CB and CB1 of T13 as shown in Figure 1.
4.3 Required Intertrip Scheme for Transformer Fault
With reference to Figure 1, it is observed that for THE transformer fault, remote CB in addition to
6.6 kV TFs’ CB should trip due to lack of transmission line CB at industrial substation. For example for
fault on T2, the fault current consisted of two components. The first component is from TL#2. The
second component is from TL #1 via T13 and T14. Contribution from TL#2 is interrupted by opening
the CB of TL at remote end via intertripping scheme between substation and remote end. Contribution

APS | 2015
from TL#1 is interrupted by opening CB1 of T13 via intertripping schemes between transformers. Each
intertripping scheme is discussed as follows:
A) Intertrip Scheme from Substation to Remote End for TF Fault
The Main objective of this intertrip scheme is to interrupt fault current contribution from the line which
is directly connected to faulty TF (eg T2) by tripping the associated line circuit breaker at the remote
end. It consists of two schemes.
Main A: The output trip contacts of each TF protection (eg T2) energise the binary input of the line
current differential protection. This in turn sends direct intertrip signals to remote end line CB. To
cover fault on any transformer, the output contacts of all protection elements of each TF are
paralleled together (in SWH A and B) and the output via copper cable energises the binary input of
line protection (in control room C). The output includes all protection of transformers including
differential, restricted earth fault, overcurrent and guard relays (Bucholz, pressure relief, winding
temperature).
Main B: This intertrip operates over voice frequency communications device and directly inter-trips
respective breakers at remote end for faults on transformers. It is energised by all TF protection.
B) Intertrip Scheme Between Transformers for TF fault
The Main objective of this scheme is to interrupt fault current contribution from healthy 110 kV line
(eg Line 1) via healthy TFs (eg T13 and T14) for fault on any transformer (eg T2). This requires
intertripping schemes between transformers. Since the CBs of T1 and T2 are in SWH A and for T13
and T14 on SWH B (Figure 1), the intertripping scheme is hardwired between T1/T2 and T13/T14
which is run between the two switchhouses. The scheme is energised by the operation of all
transformer protection as mentioned for item ‘A”.
The major deficiency of the intertripping scheme between transformers was lack of back-up of scheme.
This was rectified when the new protection schemes were installed as described in section 6.3(B) below
5 Installation of New Numerical Protection and Control for 110 kV

Transmission Lines
This section describes the challenges and solutions for replacement of the old transmission line
protection with new numerical relays:
110 kV TL Protection and Control at Industrial Substation
The new protection scheme of each of 110 kV TL consists of Main A and B current differential
protection. Both schemes are new numerical relays. The communication bearer for scheme A is radio
and for B is fibre optic.
The new relays at the remote end have been programmed for current differential, back-up distance,
CBF initiation and intertripping schemes to industrial substations. The reach of Zone 2 distance relay
has been set to see 6.6 kV fault at industrial substation since TFs’ 6.6 kV CBs have not CBF scheme.
The new relays at the industrial substation was programmed for current differential and intertripping
but not for distance protection since there is no major fault current contribution from industrial
substation for TL fault. Hence, the fault contribution from motors and healthy line are negligible due to
current flowing through transformers.
The Main challenges for installation of new relays at industrial substation were:
1. reusing the existing secondary copper cables of single old protection scheme for one of
the new scheme (Main A);

APS | 2015
2. provision of new Trip/close circuit for second scheme of new TL protection (Main B)
bearing in mind that spare copper cable did not exist and installation of new cables from
switchyard to control room were not allowed;
3. installing of new two protection panel in new control room D instead of existing room C
due to inadequacy of space;
4. removing existing intertripping scheme as discussed above and configuring them on new
transmission line protection scheme;
5. provision of a reliable design and settings policy; and
6. installing, testing and commissioning of the new protection and intertripping schemes with
minimum plant outages.
The solutions for each of these challenges are described below:

1. Reusing existing secondary copper cables for new panels on new control room
Since the location of the new protection and control panels are in new control room and the old TL
protection scheme was only a single scheme, significant numbers of drawings were checked along
with frequent site visits and cable conduit audit to find out that whether the existing secondary
copper cable and their length are sufficient for two new protection panels. It was found that existing
cables are sufficient only for Main A protection of each panel. For the other one (Main B) an
alternative solution should be found to provide trip/close circuits from protection panels (control room
D) to SWHs A and B (Figure 2) for tripping 6.6 kV CBs where the circuit breakers are.
The alternative solution which was implemented is based on using fibre optic cable and digital
transceiver and receiver instead of copper cable to provide trip and closed circuit for Main B
protection scheme of each transmission line.
The design of the scheme has been shown in Figure 3. It shows the installed fibre optic cables and
required transceivers/receivers on different control rooms. It is noticed that Main B protection of each
line is connected to both transceivers 1 and 2 which are for tripping of 6.6 kV CBs of T1/T2 and
T13/T14 CBs respectively. This is because as mentioned above protection of each line should also
trip 6.6 kV CBs associated to healthy line in order to remove its contribution.
Figure 3: Provision of Trip/close Circuits for Main B Protection using Transceiver/Receiver and fibre optic cable

APS | 2015
2. Protection and Control Upgrading at Remote Substation
With the protection upgrade of the line at industrial substation, concurrently all 110 kV lines (seven
lines) and bus zone protection at remote end substation were upgraded. This allowed full line current
differential protection at both ends and replacement of old intertripping schemes from the remote
substation to the industrial substation to be implemented on new relays.
6 Implementation of Intertripping Schemes on New Numerical Relays

of 110 kV Lines
As part of this project, all old intertripping schemes between industrial and remote substations as
mentioned in Section 4 were removed and incorporated on new transmission line protection. For this
purpose the following special design and configuration on each new transmission line protection was
implemented.
6.1 Intertrip Scheme from Remote to Industrial Sub for TL Fault

The old intertriping scheme for TL fault was intertrip via distance protection at remote end using
transceiver/receiver as mentioned in Section 4.1. With the installation of two new line current
differential protections, this scheme which was acting as back-up was not required and removed. The
new Main A and B schemes of line current differential protection provides full protection for TL fault by
tripping 110 kV line CB at remote end and associated 6.6 kV CBs at industrial substation.
6.2 Intertrip Scheme from Remote to Industrial sub when TL CB at Remote

sub is opened
The auto opening intertripping scheme as mentioned in Section 4.2 for tripping substation 6.6 kV
circuit breakers when the respective TL CB at remote end is open was removed and configured on
new TL numerical relays. For this purpose the relay at the remote end obtains the status of the line
CB via one of its binary inputs. This CB status is received at industrial substation via current
differential communication channel. By receiving the open status of the CB, the trip output contact of
relays at industrial substation is energised and trips respective 6.6 kV CBs.
6.3 Required Intertriping scheme for Transformer fault

As mentioned in Section 4.3, in order to interrupt fault current for TF faults, two intertripping schemes
is required. With new numerical relays of TL, the arrangement of these schemes is as follows:
A) Intertrip Scheme from Industrial Substation to Remote End for TF fault
The old intertrip as mentioned in Section 4.3 was removed and implemented on new TL’s protection.
For this purpose, operation of any TF protection energises a binary input of each of the new TLs
relay at industrial sub. By energisation of the relay input a direct intertrip signal via current differential
relay communication channel is sent to remote relay and trips associated line CB at remote end.
Since TF and TL protection are in different control rooms (Figure 2), the following
intercommunication arrangement between TF and TL protection was required:
1) Main A: The intertrip old circuit as mentioned in Section 4.3 (A) has been used to
energise the binary input of Main A line protection at industrial sub
2) Main B: the new circuit consists of using digital transceivers/receivers as installed

in control rooms of TF and TL protection panels. The transceivers and receivers
used for these schemes are the same as the one used in Figure 3 for provision of
trip/close circuit of Main B line protection. For example they operate on both
directions. In this scheme TF protection energises the transceivers (SWH A or B)

APS | 2015
via hard wire. Transceivers send the signal to receiver (room D) via fibre optic and
then receiver energises the binary input of Main B TL protection.
B) Intertrip Scheme Between Transformers for TF fault
As mentioned in Section 4.3 (B), the deficiency of the old scheme was lack of back-up intertripping
scheme between transformers for transformer fault. In order to provide back-up, a design based on
using transceiver/receiver devices via fibre optic has been implemented. This design is similar to the
design which was implemented for provision of trip/close circuit for Main B of TL protection as shown
in Figure 3.
In this scheme the trips from protective relays of T1 or T2 (in SWH A) for faults on these TFs are
sent via a digital transceiver device in SWH A over fiber optic cable to a digital receiver device in
switchhouse ‘B’ to trip respective 6.6 kV CBs of T13/T14 as mentioned in Section 4.3(B). Similar
intertripping scheme exists from SWH B to SWH A for fault on T13/T14 transformers.
7 Implementing a reliable design and settings based on sound

standard policy and philosophy
The protection and control setting and logics of the new TL’s numerical relays for Main functionalities
were based as per standard TasNetworks philosophy and general accepted principals [2, 3]. The few
exceptions are:
a) The reach of Zone 2 distance relay at remote end substation has been set to see 6.6 kV fault at
industrial substation since 6.6 kV CBs have not CBF scheme.
b) The new relays at industrial substation were not programmed for distance protection since there
is no major fault current contribution from industrial substation for TL fault. The fault contribution
from motors and healthy line are low due to current flowing through transformers.
c) Additional logic programming were implemented to incorporate all required intertripping
schemes as mentioned above.
d) Auto-reclose functionality was disabled due lack of 110 kV CB at industrial substation.
Also auto-reclose functionality at remote end was also set disabled due to asynchronous
conditions which may arise to large motor loads at industrial substation
8 Installing, testing and commissioning of the new protection and

intertripping schemes with minimum plant outages due to load
critically
Comprehensive Factory, site and commissioning tests were carried out on new panels to verify their
protection and intertripping functionalities in accordance with TasNetworks and international standards
[4, 5].
For Factory Acceptance Test (FAT), a temporary fibre communication link was established between
two relays in order to verify the integrity of communication scheme, setting, and all intertripping
functionalities and their associated logics. An automated test equipment used for all tests. The tests
included full slope characteristic of current differential relays (stability and sensitivity). Distance
characteristic, circuit breaker failure, direct intertrip and logics for interlocking schemes. The logic of
each intertriping scheme from remote to industrial substation and vice-versa was simulated and tested
by energisation of relevant binary inputs and measurement of associated binary outputs. The SCADA
functionality of relays (DNP3 points) were simulated by energising the associated protection element
and measurement the digital and analogue points on a simulated SCADA device.
Site Acceptance Tests (SAT) were carried out with all equipment and wiring such as trip/close circuits
and circuit breakers in service. The tests included functional tests and end to end scheme tests by
secondary injection tests using GPS to verify the integrity of all functionalities. It included simulation for
internal /external fault, DIT for CBF and each intertripping scheme. For SAT it was required to have
outage for each line and its associated connected transformers for two days. For example for
commissioning of TL# 1 protection panels, it was required to have outage of the line, T1 and T13

APS | 2015
transformers. All of the required TL and TF outages for testings purposes were coordinated carefully
with routine maintenance programs of customer load (motors). This necessitated all stages of the
project such as completion of design, FAT/SAT and commissioning to be programmed in detail and
accurate and be completed on time in order not to request any customer load interruption for
commissioning of the new panels. For this purpose some tests were required to be carried out during
Christmas holiday.
The SCADA functionality of relays (DNP3 points) were simulated by secondary injection test of the
associated protection element and measurement of the digital and analogue points as measured by
station HMI and system control centre. Generally DNP3 points were verified during other SAT tests
such as functionality and end to end scheme tests.
Commissioning tests were involved on load stability tests and metering measurement. The on load
stability tests were involved reading of restraint and operating current of current differential relays. The
correct directionality of distance relay was verified by measurement of direction of flow of MW/MVAR
as measured by relay against the actual primary value
9 Other Challenges in Industrial EHV Substations

In the past, some other issues were encountered on the industrial substation. Following describes
these issues and their solutions:
A) Tripping of 6.6 kV bus during motor start up
On several occasion, during energisation of motor loads, the 6.6 kV incomer reverse power relays on
CBs 1&2 in Figure 1 tripped. It was found that the measured current during energisation exceed the
pick up setting of incomer’s relays. With revised setting, the unwanted trippings were prevented.
B) Limitation of short circuit rating of 6.6 kV switchgears

The short circuit rating of 6.6 kV switchgear equipment is 31.5 KA for 3 second. Due to installation of
additional load of 2 x 12 MW new synchronous motors chippers and 2 x 1 MW induction motors, the
short circuit fault current on 6.6 kV bus 1 increased to 33.7 KA (Figure 1) with 8 KA fault contribution
from new motors. If BC 3 is opened, the short circuit fault current will be reduced to 20.6 KA. However
this will cause low voltage due to high normal loading current, consequently it is necessary the BC 3
reMain normally closed.
For 6.6 kV buses 2 where isolator 1 can normally be left open, the fault current is well within the short
circuit rating of the equipment.
In order to reduce short circuit fault level on 6.6 kV Bus 1, Fault Current Limiting (FCL) device were
installed in series with CBs 1 and 2 (Figure 1). This reduced fault current on Bus 1 to 27 KA which is
within the switchgear rating.
The FCL consist of an actuator with a fuse in parallel. The actuator is opened when current exceeds a
specified value. By opening the actuator, current is diverted into parallel fuse and then fuse is melted
and fault current is interrupted. The total interruption time is about 4 to 5 msec. The Interrupting
capability can be adjusted in the range from 4 kA-14 kA in precision of 2 kA. There is redundant
current sensor as back-up and the operation modes of the FCLs are SCADA adaptable. The Main
advantage of this type of FCL is its low price as compared to saturated reactor type. However its Main
disadvantages are non-operation due to power supply failure (DC system) and not” Fail Safe”
condition due to electronic components used in device.
With installation of FLC, replacement of switchgears was diverted.

APS | 2015
10 Conclusion
Challenges for installation of new digital relays for 110 kV transmission lines in an EHV industrial
substation which supplies a large paper mills were discussed. One was lack of sufficient secondary
copper cables from switch yard to protection and control rooms for trip/close circuit for one of
protection and control device of the 110 kV line. The other one was to remove existing old different
intertripping schemes between the substation and remote substation and within the plant and transfer
them to new numerical relays. The first challenge was solved by implementation of innovative
solutions using fibre optic cable and digital transceiver/receiver. The solution for other one was
achieved by appropriate design, setting and logic programming of the new numerical relays for
implementation of all intertripping schemes. The implementation of intertripping schemes were
required due to lack of EHV transmission line circuit breakers at industrial substation.
The paper also elaborated an effective application of a modern fault current limiter in an industrial EHV
substation in order to reduce the fault current within the switchgear rating.
The implementation of these innovative solutions highlighted the benefit and capability of digital relays.
It also demonstrated how significant cost savings in primary and secondary equipment can be
achieved by utilizing sophisticated protection and control functionalities of numerical relays.
11 References
[1] Australian National Electricity Rules; Version 66, December 2014
[2] Ghezelayagh; M: State of the Art Application Techniques of the Numerical Relays; CIGRE
Australia APB5 (SeaPac); Brisbane 2013
[3] Alstom Grid: Network Protection & Automation Guide; 2011
[4] CIGRE 34.1: Analysis and Guidelines for Testing Numerical Protection Schemes; Final Report
March 2000
[5] Ghezelayagh; M: State of the Art Testing Techniques and Required Tests of Numerical Relays;
Australian Protection Symposium; Sydney 2013

APS | 2015
About the Authors
Maty Ghezelayagh received his B.Sc. from Sharif University of Technology

(Iran), M.Sc. from University of Manchester (UK) and PhD from University of
Wollongong (Australia) all in electrical power systems. For the last 25 years he
has worked with five different companies in Australia and before that for four
years for the Main 400 kV transmission company in Iran. His major work
experiences have been in all areas of protection and control of power systems
namely EHV transmission and distribution networks, substations and power
stations.
Chris Simmons has 15 years’ experience in transmission systems initially

working as an Asset engineer with Transend Networks the owner/operator of
the Electricity Transmission network in Tasmania and then as a consultant with
Entura. Chris currently holds the position of Specialist Secondary Systems
Engineer with Entura and was the lead designer on the protection upgrade
project, a component of which is described in this paper.

APS | 2015

APS
APS| |2015
2015
XML and UML - What They are and Why We Need to Know
Them
Alexander Apostolov | PAC World

alex.apostolov@pacw.org
1 Introduction
The power system automation community is going through a period of transition from the world of hard
wired systems and proprietary configuration and analysis tools into the world of distributed IEC 61850
communications based systems and object oriented standards based engineering and analysis tools.
This requires the development of a new set of skills in order to help the specialists from our industry
understand and use the new technology.
The main goal of this paper is to introduce the Extensible Markup Language (XML) and the Unified
Modeling Language (UML) to the protection and control community and focus on the UML diagrams and
XML files used in the IEC 61850 standard and in some IEEE standards related to protection and control
in order to help the readers understand the diagrams and files included in or defined by the different
standards.
Figure 1 : XML version of part of the World file of this paper
Examples of the use of UML diagrams in IEC 61850 are given later in the paper.
Examples of the use of XML in the IEEE C37.239: IEEE Standard for Common Format for Event Data
Exchange (COMFEDE) standard are presented at the end of the paper.
2 What is XML?
XML is a markup language based on existing markup languages that have been used for different
applications for many years. Even the Word version of this paper and the corresponding Power Point
presentation use XML.
Some of the acronyms that are used later in the text of the paper are as follows:
SGML - Structured Generalized Markup Language
HTML – HyperText Markup Language

APS | 2015
XML - eXtensible Markup Language

W3C – World Wide Web Consortium
XSL - eXtensible Stylesheet Language
CSS - Cascading Style Sheets
XML was developed by members of the W3C and released as a recommendation by the W3C in
February 1998. It is a simplified version of SGML and a cousin of HTML.
SGML was developed to standardize the production process for large document sets and .is an
international standard (ISO 8879) that has been in use as a markup language primarily for technical
documentation and government applications since the early 1980s.
The growing popularity of XML is the result of its flexibility and strength. It is extensible, because it allows
you to extend the user’s ability to describe the domain specifics of the document. It is not only Microsoft
Office that uses XML. It has other uses, such as being the base language for the Extensible Messaging
and Presence Protocol (XMPP), as well as for Microsoft.NET Framework configuration.
In appearance, XML is quite similar to HTML. This similarity is due to the fact that they both use tags. In
HTML, there is a specified set of tags that defines the format of the data.
Figure 2 : XML version of an expense report spreadsheet
In XML, the user can create the tags required by the application domain. That is why XML is extensible
– it extends the ability to describe a document, letting you define meaningful tags for your applications.
For example, since any IED typically provides current measurements, for the phase A current
measurement that is available as a floating point we can create a tag called <PhsAf>. In a similar way,
we can create as few or as many tags as our document needs. It is obvious that we are extending the
tags to identify elements by what they are -- not by how they look.
The extensibility means flexibility, but flexibility requires planning. To make good use of XML, we want
to know and understand our documents: what pieces comprise them, how those pieces relate to each
other, and how do we want to identify the different pieces. This is where the object models defined by
IEC 61850 become the foundation for the development of XML files used in the standard.
XML is a Markup language because its purpose is to identify elements within the document. Without
markup, the computer sees any document as one long string of text, with each character having equal
importance to every other character. By marking up a document, we identify the bits and pieces in a
way that gives them value and context.
The big advantage of XML is that it allows extensible markup, i.e. we can mark up the document in ways
that match our substation protection, automation and control system needs.

APS | 2015
However, markup is nothing but a way of identifying information. It does not program the data to act in
a certain way, to display in a certain way, to do anything other than carry an identifying mark.
XML is a Language because it follows a firm set of rules. It allows us to create an extensible set of
markup tags, but its structure and syntax remain firm and clearly defined. This does not mean that it is
a programming language – it is not used to program a set of actions, but for a well-structured markup
definition.
XML applies structure to documents and data. Since SCL documents are sets of related information,
the structure is quite important. It is the way we put a skeleton behind the information, so that the pieces
of information work together and make sense as a whole.
The document structure defines the elements which make up a document, the information we want to
collect about those elements, and the relationship those elements have to each other. XML is used to
markup the document, following the structure of the model. That is why XML is appropriate for describing
the different aspects the substation configuration from the perspective of the substation automation
system.
The document structure is called the Document Tree. The main trunk of the tree is the parent. All the
branches and leaves are children.
Document trees are usually visually represented as a hierarchical chart.
Considering that the data object structure in IEC 61850 is also hierarchical, it is obvious that XML is very
well suited for use as a SCL file format.
A well-formatted document is not sufficient. It also has to meet some constraints in order to make sense
in the problem domain. The constraints enforce rules that determine the presence of elements and their
attributes, as well as the order of these elements. These rules are part of what is defined as a Data Type
Definition (DTD) or XML schema. This is the oldest schema inherited from SGML.
An important property of XML schemas is that they are also extensible, i.e. if necessary the schema can
be extended to meet new requirements.
XML Schemas describes the structure of XML documents. The XML Schema language is also referred
to as XML Schema Definition (XSD). XML Schemas are much more powerful than DTDs.
Any XML schema, including DTD, is a template for the markup of the document. That indicates the
presence, order and placement of elements and their attributes in an XML document. In this sense it is
similar to the object models of different elements of multifunctional IED’s. IEC 61850 defines the classes
that are later instantiated to build the object model of an IED. Different functional elements, for example
a time-delayed overcurrent element can have multiple instances in a multifunctional protection IED. DTD
has a “+” operator to indicate that more than one instance of an element is possible in the XML file.
HTML files appear to be very similar to XML files. This is due to the fact that both are defined from
SGML. HTML is functionally a specific subset of SGML limited to the description of web pages.
Since XML defines the data structure, it will not display a page by itself. We must use a formatting
technology, such as CSS or XSL to display XML-tagged documents in a Web browser.
Cascading Style Sheets (CSS), are the current way to display XML documents in a Web browser. CSS
is a means of assigning display values to page elements.
Extensible Stylesheet Language (XSL) is an XML-based language for expressing stylesheets. With XSL,
the user can make context-sensitive display decisions. XSL later split into the following XML languages:
• XSL Transformation (XSLT) for transforming XML documents

• XSL Formatting Objects (XSL-FO): for specifying the visual formatting of an XML document
3 The Unified Modeling Language

Object modeling is one of the foundations of IEC 61850. The models in the standard represent the
abstracts of the essential and communications visible parameters of the complicated real electric power
systems world. This process of virtualization requires the use of modeling tools that can present the
complex functionality of a substation and its protection and automation systems in a standardized way
that is also easy to represent and understand.

APS | 2015
In recent years UML (the Unified Modeling Language™) has become the common tool used in the
modeling of any process, system or device. It is a standard language for specifying, visualizing,
constructing and documenting different simple or complex systems.
Attempts to create an object-oriented modeling language began between mid-1970 and the late 1980s.
Different competing methods had advantages and disadvantages. The experience from their use led to
the development of UML that started in late 1994 when Grady Booch and Jim Rumbaugh of Rational
Software Corporation began their work on unifying the Booch and OMT (Object Modeling Technique)
methods. UML 1.0, a modeling language that was well defined, expressive, powerful, and generally
applicable was submitted to the Object Management Group (OMG) in January 1997 as an initial RFP
response. OMG has been since 1989 a non-profit, international, open membership, computer industry
consortium and UML is one of its standards. The current "Available" version of UML is 2.0. The UML
2.0 Superstructure specification (already built into vendor products) has been completed, and is
available to everyone for free download at http://www.uml.org/. Three separate parts of UML 2.0 - the
Infrastructure (meta-metamodel), Object Constraint Language, and Diagram Interchange - are still
undergoing their first maintenance and will become Available Specifications when this process is
completed.
Several modeling tools are covered under the heading of the UML. It uses mostly graphical notations to
express the design of software and other projects, systems or structures. Different types of diagrams
can be used to present data structures, device and operator interactions or any other substation
automation or protection related process. Using the UML helps project teams communicate, explore
potential designs, and validate the architectural design of the system.
UML 2.0 defines thirteen types of diagrams, divided into three categories:
Structure Diagrams include six diagram types that represent static application structure: Class Diagram,
Object Diagram, Component Diagram, Composite Structure Diagram, Package Diagram, and
Deployment Diagram.
Behavior Diagrams include three diagram types that represent general types of behavior: Use Case
Diagram, Activity Diagram, and State Machine Diagram.
Interaction Diagrams include four diagram types that represent different aspects of interactions. They
are derived from the more general Behavior Diagram and include the Sequence Diagram,
Communication Diagram, Timing Diagram, and Interaction Overview Diagram.
Figure 3: UML support in Microsoft Visio
The use of UML requires good understanding of both the tools available and the problem domain they
are applied to. In our case the problem domain is the substation, the substation protection and
automation systems, including the communications system used by both. Different tools are available
APS | 2015
on the market and can be selected depending on the requirements of the application. One available
option is the popular Microsoft Visio, which includes (among all other drawing tools) a set of UML
diagram modules (see Figure 3).
Since the UML is applicable to object-oriented problem solving, it is used in different parts of the IEC
61850 standard to represent in a standardized graphical way the complex models of multifunctional
substation IEDs and their interface with the primary substation equipment and the communications
network.
From the abstract modeling point of view, a model is an abstraction of the underlying problem, while the
domain is the actual world from which the problem comes. There are three main components of a model:
• Functional model
• Object model
• Dynamic model
The functional model describes the behavior of the system under different conditions from the point of
view of the user.
The object model represents the structure, including sub-layers and basic objects, data types, attributes,
services and associations.
The dynamic model covers the internal behavior of the system, including the interaction between sub-
systems and components, exchange of signals and conditions under which an action takes place.
The models of primary and secondary substation equipment consist of objects that interact by sending
each other messages. The IEC 61850 object models include all three above-mentioned components
and the communications visible attributes of what they represent. They have hierarchical structure that
corresponds to their functional hierarchy. At the same time, the different functional elements in the
system interact with each other to execute protection and control functions using specific services and
following rules of behavior defined by the standard..
4 UML Modeling Diagrams

UML uses different types of diagrams depending on the modeling requirements of the application – data
structure, interaction, etc. The following is a list of the more commonly used diagrams:
• Use case diagrams

• Class diagrams
• Object diagrams
• Sequence diagrams
• Collaboration diagrams
• Statechart diagrams
• Activity diagrams
• Component diagrams
• Deployment diagrams
In the following sections we are going to briefly describe the diagrams used in the IEC 61850 standard.
4.1 UML: Class Diagrams

Class diagram gives an overview of a system by showing its classes and the relationships among them.
Class diagrams are static - they display what interacts but not what happens as a result of the interaction.
UML class notation is a rectangle divided into three parts: class name, attributes, and operations.

APS | 2015
Figure 4: UML Class notation
Class diagrams show not only the relationship between the components of the model, but also their
interfaces. These are modeled using:
• Association - a relationship between instances of the two classes. There is an association

between two classes if an instance of one class must know about the other in order to perform
its work. In a diagram, an association is a link connecting two classes.
• Aggregation - an association in which one class belongs to a collection. An aggregation has a

diamond end pointing to the part containing the whole.
• Generalization - an inheritance link indicating one class is a superclass of the other. A

generalization has a triangle pointing to the superclass.
Multiplicity of an association end is the number of possible instances of the class associated with a
single instance of the other end. Multiplicities are single numbers or ranges of numbers.
Figure 5: UML Class diagram
The example in Figure 5 shows a generic UML Class diagram of the functional hierarchy of an IEC
61850 based server.
4.2 UML Object Diagrams

Object diagrams show instances instead of classes. They instantiate class diagrams as shown in Figure
4.

APS | 2015
Figure 6: Object diagram
4.3 UML: Sequence Diagrams

Class and object diagrams are static model views. They do not describe the behavior of the modeled
system.
Interaction diagrams are dynamic. They describe how objects collaborate.
Figure. 7: Sequence diagram
A sequence diagram is one of two types of interaction diagrams and details how operations are carried
out, i.e. what messages are sent and when.
Sequence diagrams are organized according to time and can help clarify a use case in order for it to be
realized in software.
The example in Figure 5 shows the use of a sequence diagram to define the exchange of messages
between an IEC 61850 client and a control object.
4.4 UML Statechart Diagrams

Objects have behaviors and states. The state of an object depends on its current activity or condition.
A UML statechart diagram shows the possible states of the object and the transitions that cause a
change in state. They are very similar to the well known state machines and are used in the design
process to help with the transition from the analysis of the system to its implementation.
The different states in these diagrams are represented as rounded rectangles.
Transitions are arrows from one state to another.
Events or conditions that trigger transitions are written beside the arrows.

APS | 2015
Unselected
Sel_req[Test not ok]^client.Sel_rsp-
Sel_req[Test ok]^client.Sel_rsp+
cancel_req^client.cancel_rsp+
Timeout
TimOper_req[Test not ok]^client.TimOper_rsp-
Oper_req[Test not ok]^client.Oper_rsp-
Ready Oper_req[Test ok, sboClass=OPERATE_ONCE]^client.Oper_rsp+
Oper_req[Test ok, sboClass=OPERATE_MANY]^client.Oper_rsp+
TimOper_req[Test ok]^client.TimOper_rsp+
timer expired[Test not ok] ^client.TimOper_rsp-

WaitForActivationTime timer expired[Test ok, sboClass=OPERATE_ONCE] ^client.TimOper_rsp+
entry / start timer
timer expired[Test ok, sboClass=OPERATE_MANY] ^client.TimOper_rsp+
Figure. 8 Statechart diagram
The example in Figure 6 shows the statechart diagram of the Select Before Operate (SBO) function as
implemented in IEC 61850 compliant systems.
5 UML and XML Use in IEC 61850

The different UML diagrams briefly described above are used in the IEC 61850 standard to present the
abstract models of the substation domain. They represent the foundation of the definitions of the object
models and services, as well as the Substation Configuration Language and the different types of files
used to describe the functional hierarchy of the system and data used for exchange between IEDs and
applications. Part 6 of the IEC 61850 standard specifies a description language for configurations of
electrical substation IEDs – the Substation Configuration Language (SCL), based on UML and XML
Version 1.0.
It is used to describe the substation connectivity, IED configurations and communication systems
according to parts 5 and 7 of this standard. Description of the relations between the substation
automation system and the substation (switchyard) itself
SCL was developed to support easier engineering of substation automation systems and application
functions. It allows the description of a substation or an IED’s configuration to be passed to a
communication and application system engineering tool
The SCL files need to meet requirements related to the support of different phases in the engineering
process. This is achieved through the use of UML and XML.

APS | 2015
Figure 9: Conceptual UML diagram of an IEC 61850 based system
Its main purpose is to allow the interoperable exchange of communication system configuration data
between an IED configuration tool and a system configuration tool from different manufacturers.
The substation configuration language supports the development of engineering tools that are capable
of describing:
• The substation one line diagram representing the different voltage levels, busses, transformers,
bays and switching devices. The functional requirements should also be included in terms of
allocation of logical nodes to the primary substation equipment.
• The IEDs to be used to perform the required functions based on a fixed number of logical nodes
(LNs)
• The communication interface of the different IEDs – specifically their connection to the
substation local area network
• The Client-Server and Peer-to-Peer communications for the specific substation automation
system implementation
It needs to be understood that the standard does not define any specific software tools that support the
intended engineering process. This is a task that the IED manufacturers, substation automation system
vendors or third party providers have to develop based on the requirements of the market using the
different types of files defined in the standard.
IEC 61850 defines four types of files required to support the intended engineering process. In order for
an IED or a system solution by a manufacturer to be compliant with the standard, they have to support
the use of the files described below directly from the IEDs or through tools delivered with the system.
The structure of the files is based on UML models that are converted to XML files with different
extensions used in the engineering of IEC 61850 based systems.
5.1 System Specification Description

The description of the system is the first step in the engineering process and until now has not been
based on any standardized approach. The IEC 61850 engineering process envisions the use of
substation specification tools that allow the user to describe the substation design and associated
functional requirements for the substation protection and automation systems.
The data exchange from such a system specification tool and other tools utilized in the process should
be based on the System Specification Description files defined in the standard. They have an SSD
extension.
The SSD file describes the single line diagram of the substation and the functional requi9rements
represented by logical nodes. The logical nodes can be abstract in the sense that they are not allocated
to specific IEDs.

APS | 2015
5.2 IED Capability Description

The default functionality of an IED in the substation configuration language is represented by the IED
Capability Description (ICD) file. It is used for data exchange from the IED configuration tool to the
system configuration tool.
This ICD file describes the capabilities of an IED. It contains exactly one IED section for the IED whose
capabilities are described. Since it represents the default functionality (i.e. before it has been
configured), the IED name in this file is TEMPLATE.
The file also includes the different logical node types as they are instantiated in the device.
The file extension shall be .ICD for IED Capability Description.
IEC 61850 does not specify where the ICD file comes from. In IEDs designed for IEC 61850 environment
and with large memory, this XML file may be available from the device itself.
For IEDs that are based on existing platforms that were adapted to support the standard, the
manufacturer is required to provide tools that output ICD files.
Figure 10: Part of IEC 61850 XML file of IED
5.3 Substation Configuration Description

The configuration of the system is represented by the Substation Configuration Description (SCD) file.
It contains:
• substation description section

• communication configuration section
• all IEDs
The IEDs in the SCD file are not anymore in their default configuration, but as they are configured to
operate within the substation protection and automation system. These files are then used to configure
the individual IEDs in the system.
5.4 Configured IED Description
The difference between the IED Capability Description (CID) file and the Configured IED Description file
is that the second includes the substation specific names and addresses instead of the default ones in
the first.
The CID file represents a single IED section of the SCD file described above.

APS | 2015
6 Common Data Format for IED Event Data

Another reason that we need to know XML is the standard data format developed by the IEEE PES
Power System Relaying Committee in order to meet the need for standardization of the event reporting.
It started a working group – H5b - in the Relay Communication subcommittee with the task to prepare a
Report on a Common Data Format for IED Event Data.
Fig. 12 Event data exchange process
The report defined a common XML-based file format for describing and exchanging event data records
collected from power systems. It addressed the fact that protection relays and other IEDs store in their
memory historical event data. The main categories of event data considered in this report were:
• Sequence of events (SOE)

• Fault reports
• Summary reports
• IED Status
• Other
The content and the format of the data recorded are vendor specific and therefore cannot be easily
integrated in a power network post analysis tool.
The main purpose of this file format is to facilitate power systems event data integration and analysis by
enabling event data exchange between multiple data sources from different vendor devices and vendor-
independent analysis tools.
The report was completed and published in 2008. A new working group (H16) was started with the task
to define a new standard. The development and balloting of the standard has been completed and now
it is available as C37.239-2010 IEEE Standard for Common Format for Event Data Exchange
(COMFEDE) for Power Systems.
Currently a new working group in the IEEE PSRC is working on the development of another XML based
file format for protection settings (COMSET).

APS | 2015
Figure 13: Example from an XML based COMFEDE file
7 Conclusions
XML and UML are some of the key technologies currently used in many domains and helping to improve
the efficiency of different business processes. With the wide spread acceptance of IEC 61850 and the
engineering process supported by it, it becomes essential for protection and control specialists to have
a good understanding of what XML and UML are, so they can use them in their everyday work.
UML (Unified Modeling Language) uses mostly graphical notations to express the design of software
and other projects, systems or structures. Different types of diagrams can be used to present data
structures, device and operator interactions or any other substation automation or protection related
process. Using the UML helps project teams communicate, explore potential designs, and validate the
architectural design of the system.
UML diagrams are widely used in different parts of IEC 61850 to present the abstract models of the
substation domain. They represent the foundation of the definitions of the object models and services,
as well as the Substation Configuration Language and the different types of files used to describe the
functional hierarchy of the system and data used for exchange between IEDs and applications.
XML applies structure to documents and data. Since SCL documents are sets of related information,
the structure is quite important.
Part 6 of the IEC 61850 standard specifies a description language for configurations of electrical
substation IEDs – the Substation Configuration Language (SCL), based on UML and XML.
The C37.239-2010 IEEE Standard for Common Format for Event Data Exchange (COMFEDE) for
Power Systems is also based on XML.Currently a new working group in the IEEE PSRC is working on
the development of another XML based file format for protection settings (COMSET).

APS | 2015
About the Author
Dr. Alexander Apostolov received MS degree in Electrical Engineering,

MS in Applied Mathematics and Ph.D. from the Technical University in
Sofia, Bulgaria. He has 40 years’ experience in power systems protection,
automation, control and communications.
He is presently Principal Engineer for OMICRON electronics in Los
Angeles, CA. He is IEEE Fellow and Member of the Power Systems
Relaying Committee and Substations C0 Subcommittee. He is past
Chairman of the Relay Communications Subcommittee, serves on many
IEEE PES Working Groups and is Chairman of Working Groups C2 “Role
of Protective Relaying in Smart Grid”.
He is member of IEC TC57 working groups 10, 17, 18 and 19 and Convenor of CIGRE WG B5.53 ”Test
Strategy for Protection, Automation and Control (PAC) functions in a full digital substation based on IEC
61850 applications” and member of several other CIGRE B5 working groups. He is Distinguished
Member of CIGRE.
He holds four patents and has authored and presented more than 400 technical papers.
He is IEEE Distinguished Lecturer and Adjunct Professor at the Department of Electrical Engineering,
Cape Peninsula University of Technology, Cape Town, South Africa.
He is Editor-in-Chief of PAC World.

APS | 2015

APS
APS| 2015
| 2015
Old Lessons - Re-learned
Graeme Heggie | ElectraNet

heggie.graeme@electranet.com.au
1 Introduction
Over national holiday breaks, weekends, especially during the spring and autumn, low loads and high
proportional wind generation could combine and result in fault levels on the network dropping to
unexpected values. In the event of a system fault this could adversely affect the operation of
protection schemes. If protection schemes do not detect faults the effects of system events will
become more widespread leading to unnecessary damage to assets and power supply disruption.
Looking at the operation of single phase tripping schemes, this paper will discuss how the schemes
work to clear a fault and the importance of ensuring tripping at all ends of the protected feeder with
particular emphasis on Wind Farm fault supplying capability and protection scheme sensitivity looking
at various differential, distance and distance signal-aided line protection schemes. For transformer
protection we look at how basic relay settings are modified by zero sequence filtering and ratio
compensation leading to differing sensitivity.
2 The Electricity Market, Renewables and Fault Levels

The Australian National Electricity Market comprises the interconnected transmission systems of
Queensland, New South Wales and ACT, Victoria, Tasmania and South Australia. The Australian
Energy Market Operator regulates generation across the NEM ensuring that generation matches
demand and that the system remains stable through step changes in demand and system faults. A
spot market is used to control the price paid to the generators with generators bidding to produce
electricity in 5 minute blocks. In general the cheapest bids are accepted first to meet the anticipated
demand.
Carbon pricing and other policies to reduce CO2 such as incentives to produce power from renewable
sources, such as wind and solar, has altered the mix of generation sources that produce power in
South Australia. Large coal and gas powered power stations are finding it more difficult to remain
viable when they can no-longer be characterized as “base load” generators and must compete with
renewable generators for access to the electricity market. We are seeing gas and coal fired power
stations shutting down or being “mothballed” as they become uneconomic to be in service and
available for production when wind generation takes a large proportion of the electricity demand.
With an installed capacity of almost 1.5GW, wind generation can easily supply South Australia’s
minimum loads during ideal conditions. In addition solar power, both residential and industrial, is
becoming a larger part of daytime generation further reducing system load and requirement for power
station generation. Some predictions are that due to upcoming advances in solar and battery storage
technology the minimum load in South Australia will approach zero at midday possibly as soon as
2024. In practice wind generation can approach 90% of total demand during a half hour period, with
excess power being exported to other states in the NEM.
Protection relays rely on there being substantial fault current to be able to detect the fault reliably and
remove the faulted parts of the network from the system leaving the rest of the system in service. A
traditional gas or coal fired synchronous generator will produce fault current into a system fault. A
synchronous machine fault supplying capability is characterized by its subtransient and transient
reactance and can produce fault current of at least 5x full load. In contrast a wind powered generator
or solar generation may only supply 1x present load to a fault. In determining the minimum fault level
on a system the available generation and its fault characteristics must be accurately modeled and
understood. A lot of work is currently being done to better understand how much fault current will be
available in the system when there is high levels of wind and solar power and the likely effects on the
system after a fault. Early results indicate that fault current may be reduced to 25% of its current level
with limited traditional synchronous generation connected to the network. This has a significant effect

APS | 2015
on power system protection operation where relay settings have been calculated for a time when fault
current was much higher.
Wind generators are also constructed based upon the available wind resource and the accessible
transmission network. In practice this results in wind farm connections being teed into existing
transmission lines rather than them being connected at bulk supply substations, leading to substations
with two feeders – see figure 1.
Figure 1: Typical wind farm connection to transmission network
Note that when a transmission line connecting the wind farm to the system is out of service the wind
farm will be the only source of fault current for the relays at the windfarm end of the remaining line.
During periods of low wind generation the available fault current at this end of the line may be
negligible.
3 Protection Schemes – Effects of Low Fault Levels

The following sections describe how low fault levels can affect the operation of different types of
protection schemes.
3.1 Single Pole Tripping
Transmission line protection is often configured to trip single pole for earth faults, using auto reclose
relays to issue a close pulse to the tripped circuit breaker to restore the system to normal a short time
after tripping. Single pole tripping for earth faults ensures that the faulty phase is tripped to clear the
fault and allows continued power flow through the line via the healthy phases. This allows higher
power transfer down the line whilst maintaining stability after a line fault. For single pole tripping to be
effective the faulty phase at each end of the line must be tripped, the faulty phase must be
disconnected from all other power system equipment. If the faulted phase CB at one end of the line
does not trip the fault arc may remain, damaging the transmission line insulation and preventing
successful re-energisation of the line.
3.2 Line Differential Protection
A typical line differential protection scheme will comprise a relay at each end of a line connected via a
communications channel, figure 2. Line current information is transmitted between the relays so each
relay can determine the current flowing through the line.

APS | 2015
Figure 2: Line Differential protection
An internal fault is detected if the current at each line end is not equal and opposite, leading to tripping
of both relays and opening CBs at each line end. Operation of the relays at each end is subject to the
sensitivity of the relays and their settings. In addition, relay manufacturers may add ‘security
algorithms’ that require the individual relays to detect a fault condition before tripping. Thus for single
end fed faults only one differential relay may trip, the relay at the weak end may not operate as it does
not detect any step change in line current at its terminals so the security algorithm may prevent
tripping. The inter relay communications channel may be used to provide transfer trip signals to ensure
both ends of the scheme operate for a line fault.
3.2 Line Distance Protection
Distance protection uses impedance zone characteristics to detect line faults. If the impedance
measured at a relay is less than the impedance zone setting the relay will trip. Tripping can be
instantaneous but, to prevent operation for faults not on the line, the reach of the instantaneous zone
must be set to no greater than approximately 85% of the line length. The last 15% of the line can be
detected by a second impedance zone that is time delayed to prevent it tripping for faults not on the
protected line that will be cleared by other protection, figure 3.
Z2
Z1
Z1
Z2
Z1 Z1
1 Trip Trip 1
Z2 T2 T2 Z2
Figure 3: Line Distance protection - simple

For auto reclosing to work successfully the initial fault must be tripped as quickly as possible this
means that auto reclose on transmission line can only occur after instantaneous tripping. Both ends of
the line must trip instantaneously to achieve this the instantaneous zone can be extended beyond the
line but this will result in tripping for external faults which may result in unnecessary power outages.
Alternatively, inter relay signaling scheme may be used to speed tripping for faults in the last 15% of
the line, as detailed below.

APS | 2015
3.2.1 Line Distance Blocking Scheme

A reverse looking impedance zone (Z3) is used to send a block signal to the remote end if the fault is
behind the relay. An impedance zone (Z2) set to cover impedances greater than the line length will trip
without further delay provided no block signal is received from the remote relay.
Z2
Z3 Z1
Z1 Z3 Send Logic : Z3
Z2 Trip Logic : Rx + Z2
Rx Rx
Tx Tx Z3
Z3
& &
Z1 1 Trip Trip 1 Z1
Z2 T2 T2 Z2
Figure 3: Line Distance Blocking Scheme – simple
Tripping will only occur if the relay can detect fault current, if there is little or no fault current from one
end of the line the relay at that end will not trip.
3.2.2 Line Distance Permissive Overeach Scheme
If a fault is detected within a large zone, set larger than the line impedance, a signal is sent to the
remote end. If a fault is detected within this large zone and a signal is received from the remote end
the relay will trip.
Z2
Z1
Z1 Send Logic : Z2
Z2 Trip Logic : Rx + Z2
Rx Rx
Tx Tx
& &
Z1 1 Trip Trip 1 Z1
Z2 T2 T2 Z2
Figure 4: Line Distance Permissive Overreach Scheme – simple
Permissive Overreach scheme may be augmented with Weak Infeed logic which allows the relay to
trip based on detection of a system fault via low voltage AND the remote relay on the line seeing a
forward fault and sending the permissive signal. This allows a relay to trip when there is no fault
current measured at its end of the line. A reverse looking zone prevents Weak Infeed tripping for
external faults. Some relay logic include the possibility to single phase trip for Weak Infeed conditions
– this may require single phase permissive signaling requiring additional signal channels.
The line protection and the features of the line protection should be carefully considered during the
design stage. It may be essential to ensure that both line ends trip for any line fault even when fault
infeed from a particular end is low or zero. Additional features may need to be used to ensure tripping
under all system conditions, for example using of multi-channel signaling, Weak Infeed features and/or
direct transfer tripping.

APS | 2015
3.2 Transformer Differential Protection

Transformer differential protection operates in the same basic manner as line differential protection.
Current is measured at the exit of each winding of the transformer and provided the sum of the
currents is zero then the relay will not trip.
IA IB
Relay
IA + IB = 0 Healthy
IA + IB ≠ 0 (= IF) Fault
Figure 5: Transformer Differential Protection – simple
If the sum of currents is not zero there could be an internal fault and the relay may trip. To provide
adequate sensitivity for internal faults on the transformer it is generally accepted to set the differential
protection pickup value to be between 20-50% of the transformer rating.
There are many complications in setting up a transformer differential protection to balance the currents
and ensure no tripping for normal load or external fault conditions. For this paper we consider two of
these and how they affect the basic pickup value of transformer differential protection.
1) Zero sequence filtering and/or phase shifting

Where transformer windings are earthed the transformer will act as a zero sequence source for
external earth faults. This zero sequence earth fault current will be measured at the winding supplying
the fault but may not be present on the other side of the transformer as the main transformer may act
as a zero sequence filter.
Similarly if there is a phase shift across the transformer then this phase shift must be compensated for
in the differential current measurement circuit, and this will alter the value of current measured by the
differential protection elements changing the pickup value of the element.
In traditional, pre numerical protection relay, this filtering and phase shifting was produced by
connecting interposing current transformers in the current measuring circuits. An example of the use of
interposing current transformers is shown in figure 6, this shows that for an external fault the current is
balanced in the differential protection circuit. If the interposing transformer were not fitted the currents
would not balance and the differential protection would trip for this external earth fault.

APS | 2015
Figure 6: Transformer Differential Protection, use of interposing transformer
In a modern transformer protection relay the function of the interposing transformers is provided by
algorithms based on the correct selection of filtering settings within the relay – these algorithms
produce the same affect as a physical interposing transformer. Relay manufacturers may provide a
table that indicates the affect that the zero sequence compensation settings have on the pickup setting
of the differential protection. The virtual interposing transformers and/or zero sequence filtering
algorithms can reduce the current seen at the differential measurement within the relay for any given
single phase injection. This is represented in Figure 6 where 3 units of current on the input to the
interposing CT is converted to 2 units of current on the differential element side of the ICT. When
testing, this effect should be borne in mind if testing differential pickup with single phase injection.
2) CT ratio adjustment
CT ratios used for transformer differential protection will be chosen based upon the MVA and voltage
rating of a transformer. CTs are produced with standard ratio tappings and it may not be possible to
match the CT ratio to the transformer full load current on each side of the transformer. In addition the
application may call for the transformer differential protection to protect not just the transformer but to
protect a section of bus or bay supplying the transformer, see Figure 7
Full Load = 1500A 800/1A 800/1A
25MVA
132/33kV
Relay Full load = 110A @132kV
= 440A @ 33kV
500/1A
Figure 7: Transformer Differential Protection, CT ratio selection

APS | 2015
Here the 132kV connections are protected by the transformer protection, in this case 132kV CT ratio
must be chosen to be able to carry the bus or diameter full load current which may be substantially
higher than the transformer full load current. In the example the 132kV bus full load current is 1500A
but the transformer full load current is approximately 110A. In order that the 132kV CT does not
become overloaded a CT ratio of 1000/1 has been chosen, this ensures a maximum continuous
secondary current of less than 2A when full load flows through the bus connections. The 33kV ratio
has been chosen to be 500/1 based on the available ratios on the standard 33kV CT. In the example
neither HV nor LV CT ratio is ideal and this mismatch must be compensated within the differential
relay before the currents can be compared by the differential protection element. This is acheieved by
adjusting the magnitude of the HV and LV currents before the quantities are compared in the
differential algorithm.
The effect of zero sequence filtering and CT ratio adjustment in a particular relay must be understood
and considered by the protection engineer if the initial basic requirement of a differential protection
pickup of 20-50% of transformer rating is to be achieved. The internal adjustment and the effect of this
adjustment will differ from relay to relay depending upon how the relay manufacturer has designed the
algorithm. For example the relay may automatically select which current (HV or LV) is closest to the
ideal ratio then adjust the other CT input to match – it may not be obvious whether HV or LV has been
chosen as the basis of this adjustment.
All relays will have a basic minimum sensitivity, irrespective of the CT ratio, transformer rating or
transformer voltage ratings. For example the minimum secondary current for correct differential
protection may be 100mA, meaning that the particular relay must be supplied with a minimum of
100mA at its terminals in order for the differential algorithm to operate correctly inclusive of all internal
zero sequence filtering and ratio compensations. For the example shown in Figure 7 this will mean
that the minimum allowable setting of the differential protection may be 100mA secondary or 100A
primary, meaning a basic pickup of 100/110 = 91% of transformer rating looking from the 132kV side.
This exceeds our target setting of 50% maximum before accounting for the additional affect of zero
sequence filtering. If fault current is low this could further hinder the correct operation of the
transformer protection as the effective setting could be higher than the available fault current.
In certain circumstances it may not be possible to include the HV bus connections in the transformer
protection zone. In which case the transformer protection must be fed from transformer HV bushing
CTs, the ratio for which can be chosen based on transformer full load current, with separate discrete
protection provided for the HV bus/connections zone.
Full Load = 1500A 1000/1A 1000/1A
Relay
100/1A
1000/1A
25MVA
132/33kV
Relay Full load = 110A @132kV
= 440A @ 33kV
500/1A
Figure 8: Use of Separate Connection Protection Transformer Differential Protection

APS | 2015
4 Conclusions
Where renewable energy generation provides the source for a large proportion of the supplied load,
fault current magnitudes may be substantially lower than has previously been assumed. Utilities must
be aware of this and check that protection relays remain sensitive to this new lower minimum fault
current. It is recommended that regular reviews of system fault levels and protection relay sensitivity
are undertaken.
When designing new power system connections the minimum fault infeed into any protection zone
must be considered to ensure that the protection schemes that are designed will operate correctly to
remove a fault and reclose correctly.
The minimum required pickup of a protection scheme should be calculated and this should be
confirmed during protection scheme commissioning.
Protection scheme designers, those responsible for calculating protection relay settings, and those
responsible for testing relays during commissioning must have thorough knowledge of the functional
capabilities of the protection relays used and the power system conditions that the relays will be
expected to operate under.
About the Author
Graeme Nicholas Heggie (21 Oct 1965 – present day, as of 1 July 2015)
is a Principal Engineer at ElectraNet, South Australia’s Electrical Power
Transmission Network Owner and Service Provider.
He is a member of the church of wittertainment and a regular and willing
contributor to the APS. He spends his working hours determining the cause
of power system faults and trying to reduce their effects, hoping one day to
make this particular endeavor redundant.
Optimistic that England will return to late noughties era form in the Ashes
(it’s still July 1st as I write this!), he can always take solace in reluctantly
supporting his nation of citizenship – being a fully passported subject of the
Commonwealth of Australia ….. in spite of only ever being convicted of
speeding.

APS | 2015

APS | 2015

APS
APS| 2015
| 2015
330 kV downed conductor protection and successful operation
Terry Foxcroft | Snowy Hydro

terry.foxcroft@snowyhydro.com.au
1 Introduction
Line protection systems were replaced between Transgrid's Upper Tumut Switching Station and
Snowy Hydro's Tumut 1 and Tumut 2 Power Stations.
A risk of a downed conductor before synchronising the line at the switching station was identified, and
a protection system designed to detect this failure. Traditional systems would not detect this unique
event until a significant quantity of fault current was reached with a possible bush fire.
The predicted event occurred less than a year after the protection was installed, and was correctly
cleared with no extra plant or environmental damage.
This paper examines the thought processes required to see the need, the need and implementation of
the protection, and the resulting fault and protection operation.
2 Power Station Configuration

2.1. Power station configuration
There are four transmission lines to power stations at Upper Tumut Switching Station. Two lines go to
Tumut 1 Power Station, and two lines to Tumut 2 Power Station. Each power station has four
generators. The power stations are underground 310 metres and 250 metres for Tumut 1 and Tumut 2
respectively.
Each power station line is a transformer ended feeder. The transformer is a three winding transformer
– 330 kV star, 12.5 kV delta, and another 12.5 kV delta. Each transformer is comprised of three single
phase transformers, and connects to two generators. There is one spare single phase transformer that
can replace any other phase.
2.2. Running up the first unit

The transformers are left de-energised when both generators are shut down. The generators may be
started and stopped several times a day, depending upon the load profile at the time.
To avoid transformer inrush, the first generator to be started on a transformer closes its generator 12.5
kV circuit breaker when the generator is still stationary. The generator then runs up to speed and is
excited. The first generator is then synchronised in the switchyard.
The synchronisation of the second generator occurs in the power station across the generator 12.5 kV
circuit breaker. Either generator can be the first generator on line on a transformer

APS | 2015
Figure 1: Tumut 1 Power Station Single Line Diagram
Figure 2: Power station transformer group first unit on line
.
APS | 2015
3 Transformer and line protection replacements

The transformer and line protections on the 330 kV plant in the Upper Tumut area required
replacement. The existing relays were showing failures. The intertrip systems were a custom design
and build from 1985 and were beyond their economic life, causing failures and outages.
The line protections were pilot wire schemes, and the copper pilot cables were aging.
Even though the transformer protections met NER requirements, the transformer protection systems
did not provide full duplication of all functions.
For these and other reasons, the transformer and line protection systems were replaced.
Figure 3: Transformer and line protection systems – duplication not shown for clarity
The Upper Tumut Switching Station end is owned by Transgrid. The power station end is owned by
Snowy Hydro. The design and implementation required significant liaison between both organisations.
The 330 kV line consists of a cable from the transformers in the power station to the cable yard above
ground, and a 330 kV transmission line to the switching station.
Figure 4: Tumut 1 330 kV cable yard
The cable and cable yard are owned by Snowy Hydro whilst the transmission line is owned by
Transgrid.

APS | 2015
3.1 Line protection functions

The line protection installed at Upper Tumut Switching Station has the following functions.
• Line differential
• Distance
• Intertripping
• Circuit breaker failure
The line protection in the power station has the following functions.
• Differential
• Emergency over current
• Cable earth fault – B protection only
• Emergency over current
• Station flooding
• Intertripping
• Zero sequence voltage with 330 kV circuit breaker open
Except for the cable earth fault protection, fully duplicated line protection systems were installed.
The A protection uses a Schneider P546.
Figure 5: Line A protection

The B protection uses a GE L90.
Figure 6: Line B protection

APS | 2015
4 Earth fault with 330 kV circuit breaker open

If there is no generation, the line is de-energised. This removes the cost of transformer magnetising
when the generator is not running. As a consequence the line can spend a lot of time de-energised,
depending upon the generating profile.
The transmission lines traverse some very steep terrain, and the replacement of the protection
systems allows a re-evaluation of probable faults and detection systems.
Figure 7: Transmission line traverse steep terrain
Conductor dropping was a high possibility. The transmission lines are the oldest 330 kV lines in the
country. They have snow loading in winter, and have very steep spans which place extra strain on the
connections.
A dropped conductor should be detected by the line differential protection when the line is
synchronised to the main power system. However, it can be quite different when the line is not
synchronised.
4.1 Conductor dropped from switchyard end

A dropped conductor cannot be detected if the line is de-energised.
A conductor dropped from the switchyard end would go undetected until the line was connected to the
system as the dropped conductor is essentially an open circuit.

APS | 2015
Figure 8: Conductor dropped from switchyard end
In summer switching an earth fault onto the 330 kV system can start a bush fire, with catastrophic
results. This is also true if the conductor drops onto a tree, particularly if it is a dead dry tree.
4.2 Conductor dropped from power station end
A conductor dropped from the power station end cannot be detected if the line is de-energised without
injecting a signal onto the line. A sub harmonic injection into the 330 kV transformer star point may be
able to detect this, similar to injection systems for generator stator earth faults. This would only work
for faults when the line was dropped from the power station end. It would have to be disabled when
the line was synchronised. The connection of the transformer neutrals and star points in the power
station precluded this installation.
If the conductor was dropped when the line was de-energised and was a solid earth, the fault current
would be limited as the generator impedance would be the synchronous reactance which is
approximately 1 per unit. If the conductor dropped after the unit was fully excited the generator
impedance would be the sub transient reactance, providing a greater fault current.
Figure 9: Conductor dropped from power station end

APS | 2015
In summer this can also cause a bushfire. In winter the line may not trip as pure snow is a good
insulator. If the conductor dropped onto a dead tree it may not conduct due to the dryness of the
wood. When it did conduct it would definitely burn.
There is a lot of granite in the area. Granite is also a good insulator. Any conductor dropped onto
granite will not pass much current, even at 330 kV.
If the conductor did drop and did not conduct, after synchronising the generator rotor would be subject
to negative phase sequence current heating. Whilst there is generator protection for this event, if it is
inevitable it is preferred not to allow it to happen.
4.3 Zero sequence voltage protection
A simple scheme was devised to detect this fault.
The line protection at Upper Tumut Switching Station had a zero sequence voltage element enabled.
This would operate whenever there was a system earth fault. For this reason it was not included in the
tripping logic.
However, the zero sequence element operated status, circuit breaker open status and circuit breaker
bypassed status signals were transmitted over the line differential signal from the switching station to
the power station.
Logic was coded within the relay at the power station to enable a zero sequence voltage trip when the
circuit breaker was open and not bypassed. When bypassed for maintenance purposes, the circuit
breaker could be in the open or closed position. Due to the complexities of the switchyard, the zero
sequence voltage protection is not enabled when the circuit breaker is bypassed.
Figure 10: Zero sequence votlage trip logic
The same logic is coded within the B protection relay.

5 The fault
On the Australia Day weekend a conductor on the U7 line between Upper Tumut Switching Station
and Tumut 2 Power Station broke and fell to earth.
This is exactly as predicted.
It is not known when it fell. There was no generation at the time and the line was de-energised.
Unit 7 at Tumut 2 Power Station was started. The local circuit breaker closed with the unit stationary to
run the unit up onto the transformer. The excitation was turned on at 90% speed and excited through
its field flashing and soft start circuitry.
The line and transformer then tripped.

APS | 2015
5.1 Dropped conductor

An operator was called out to investigate the fault. He stopped on the way down to the power station
when he saw this.
Figure 11: 330 kV line dropped across road
The 330 kV conductor had dropped from the switchyard end and was lying on the ground near the
road.
Figure 12: 330 kV line dropped across track
The conductor dropped at the steepest span.

APS | 2015
Figure 13: 330 kV tower with dropped line
The palm crimp broke at the downhill end.
Figure 14: Broken line palm crimp

These lines are the oldest 330 kV line in Australia.

APS | 2015
Figure 14: Broken line palm in situ

Due to the inaccessible location, a 110 tonne crane was required to restore the conductor.
Figure 15: Crane lifting fallen conductor

APS | 2015
After the U7 line was returned to service the other lines were checked using non destructive test
methods. Some other connections were found cracked and were repaired.
5.2 Protection operation
The zero sequence voltage protection worked exactly as designed.
The line tripped with no fault current, with the fault detected by unbalanced voltage at the switch yard
protection relays.
The following graph shows the current waveforms from the line protection relay at the power station.
The first three currents are the line currents. Red and white phase are approximately 5 amps. Blue
phase – which was open circuit from the power station end – was approximately 2.5 amps. This shows
the difference in line charging currents. The healthy phases had a higher current as they charged the
cable and transmission line to the switching station. Blue phase had less charging current as it was
only charging the cable and one short span of transmission line.
The bottom three currents are the charging currents from the cable screen.
Figure 16: Line currents from power station end

The events show that the trip was from the zero sequence voltage element 3V0. It operated on both
the A and B protection schemes.

APS | 2015
Figure 17: Power station end relay events

The voltages at the generator show balanced volts as from the power station there was no load on the
healthy phases and an open circuit on the faulted phase.
Figure 18: Generator voltages

At the time of the trip the generator was at 0.85 per unit voltage. The zero sequence element was set
to 0.47 per unit. It has a time delay to avoid any mal operation due to timing with circuit breaker
position changes. The generator excitation soft start increased the voltage between fault detection and
the trip element operating.
The voltages at Upper Tumut Switching Station show no blue phase voltage.

APS | 2015
Figure 19: 330 kV voltages

There is no blue phase voltage present at the switching station due to the open circuit blue phase
conductor.
Conclusions
The protection installed definitely stopped a fire and reduced the risk to the public.
Unique protection systems are often installed and they never operate as the fault they are protecting
against never actually happens. In this case the conductor dropped two months after the installation of
the new protection.
It is important to look at other faults and protection methods as well as the traditional methods.
Thinking differently and anticipating black swan events allowed detection of the fault before any
damage to people, the environment or plant could occur.
Acknowledgement
The author would like to thank Transgrid for providing the fault information from Upper Tumut
Switching Station.
About the Author
Terry Foxcroft has worked in the power industry for over 35 years. He has
worked in the protection commisioning, testing and design fields for over
29 years. Terry is responsible for protection design at Snowy Hydro.

APS | 2015

APS
APS| 2015
| 2015
The Reality of Providing Energy in Australia:

Bushfire Mitigation and the Benefits of Power Quality
Monitoring Outside Substations
Martin van der Linde, B.Eng, B.BusMan | NOJA Power Switchgear Pty Ltd
martinv@nojapower.com.au
1 Introduction
Australia’s electricity supply history, utilities have grappled with some unique issues. Australian utilities
are known for innovation, and the utility demands placed on manufacturers have been responsible for
some of the greatest steps in switchgear development. As one of the largest interconnected networks
in the world, distribution network service providers (DNSPs) in Australia face big issues presented by
the unique geography and climate.
One of the greatest causes for concern in the Australian Distribution community is the aspect of bushfire
risk. Australia has a long history of bushfire disasters, but in lieu of the devastating events of February
2009 the impact of utilities has been burned into the Australian psyche. The resilience of Australia’s
utilities and engineers prompts appropriate development to mitigate these risks, and this paper outlines
some of the features available using one of the most common protective distribution switchgear devices:
the automatic circuit recloser (ACR).
ACRs humble beginnings as hydraulic devices in the mid-1900s has evolved greatly through the years
through to semiconductor controlled switches. These switches have proliferated across networks all
over the globe, driven by the immediate reliability benefits and protection offered on a reasonable
budget. Australia itself has had a fair share of manufacturers, and today there are thousands of NOJA
Power reclosers in service providing reliability to customers.
Driven by continuous requirements from utilities, these semiconductor controlled reclosers have been
greatly developed, and this paper presents some solutions implemented to combat the challenges of
the Australian distribution network. Bushfire mitigation is at the forefront of development, and through
use of the NOJA Power RC10 control system it is now possible to update and integrate bushfire
mitigation strategies using commissioned NOJA Power assets, with a simple firmware update and a
network integration strategy.
By using the on board voltage and current sensing capabilities of these reclosers, the RC10 now has
the capability to conduct complete power quality monitoring and reporting – a technology long bound to
the granite yards of substations. This functionality is now available out on the feeders, using the exact
same protection devices installed years ago, providing power quality feedback from the shores of the
gold coast to Uluru.
2 Using Reclosers to Provide Bushfire Solutions
Bushfire risk is a fundamental issue of concern for most Australian utilities, with events such as the
February 2009 bushfires in Victoria being blamed on a local utility resulting in lengthy court proceedings
and fines. As responsible DNSPs, all utilities are interested in mitigating their risk of causing fires. Recent
developments in recloser technology allows for the simple integration of bushfire risk management
strategies using the current install base of NOJA Power Reclosers.

APS | 2015
Recloser strategy for system reliability basically relies on interrupting faults, and restoring supply after a
specified open time at the recloser. A reclose sequence may have multiple different close attempts, but
from a bushfire mitigation standpoint the more reclose operations in a sequence, the greater the risk of
ignition at a fault point on the feeder. Whilst for low fire risk days a longer recloser sequence will result
in less customer lost minutes, the multiple reclose attempts will each increase risk of ignition.
Previous strategies implemented in recloser schemes involved complete disabling the reclose
functionality on bushfire risk days. This can be achieved by remote SCADA control, and toggling of the
global control of the recloser converting it essentially to a single shot circuit breaker. This practice
compromises the economic performance, and is a brute force method of addressing the risk of bushfire
ignition. It is far more elegant to have a remote capability to modify the reclose sequence, by applying
global controls which can be toggled to reduce the length of the reclose sequence in different ways,
without completely compromising system performance like using the “Auto Reclose OFF” method.
Figure 1 outlines a basic four shot to lockout reclose sequence which is typical of a low fire risk recloser
system. From this figure, it is possible to see that there is a variation in tripping time and dead time
associated with each step in the reclose sequence. By granting the capability to selectively remove
portions of the reclose sequence, it is possible to customise the performance of the recloser to match a
particular bushfire mitigation need.
Figure 1: Reclose Sequence
NOJA Power has worked closely with utilities to establish global control points for the recloser system
which will allow for the mitigation of bushfire risk. The outcome of this development are the following
global control points:
• 79_2 and 79_3 Lockout

• Short Sequence Mode
• Disable Fast Trips
• Maximum Number of Trips
The ANSI standard for protection defines 79 lockout as being a protection interruption of a fault current.
Since the maximum trips to lockout in the C37.60-2003 standard is defined as being 4, the inclusion of
two new control points known as 79-2 and 79-3 lockout to the recloser control allows for blanket changes
to the reclose sequence without actually having to change internal settings. 79-2 and 79-3 lockout are
global controls which limit the reclose sequence in the device to the first two and three reclose operations
respectively. Figure 1 demonstrates the effects of 79-2 and 79-3 lockout. When these points are set to
off, the device returns to its usual configuration. By applying these points on bushfire risk days, the risk
is greatly minimized, all whilst using a simple process which can be easily integrated into a switching
scheme and protection configuration.
79-2 and 79-3’s reclose sequence reduction capabilities are very effective, but they are quite limiting
with regards to which steps to remove. The subject of recloser timing for each protection step merits a
paper of its own, but for the purposes of demonstration there exists a philosophy of “fault burning” and
“fuse burning” using reclosers. As reclosers are capable of withstanding fault current for short periods
of time, some reclosing philosophies intentionally include a slow trip response point in the fault
sequence. This allows the lines and recloser to attempt to “burn” a resilient fault away, such as a branch
APS | 2015
fallen across parallel overhead lines, or an unfortunate possum or bird suffering the same fate. Obviously
intentionally starting fires to attempt to improve reliability is a poor practice during high risk bushfire
periods, and this raises the requirement for an additional control point to remove other points in the
reclose sequence.
Figure 2: Short Sequence Mode
Figure 2 demonstrates the solution to this issue, known as Short Sequence Mode. Unlike 79-2 which
chops off the end recloser sequence to reach a limit of 2, Short Sequence Mode’s global control point
causes the recloser to use only the first and last part of the sequence. This functionality has applications
to both a 3 and 4 shot to lockout capability and it provides controllers with the capability to make all
different changes to the reclose sequence without actually having to manipulate settings. SSM is also
available as a SCADA control point, and the network integration method is exactly the same as per the
79-2/79-3 functionality.
Disable Fast Trips (DFT) is an additional global control which allow an operator to directly remove their
fuse burning implementation using a single control point. Similarly to the other control points, this is an
after effect on the reclose sequence, and does not edit the internal settings configuration, but it allows
for removing and enabling of the fuse burning strategy of a recloser with one point control.
Understanding the relevance of disable fast trips requires an understanding of the proprietary reclose
map implementation in the RC10 system, but essentially the OC2 element is solely responsible for the
fuse burning/saving strategy. DFT allows for a global disable of OC2, even if the setting is configured,
to allow for removal of the strategy. Figure 3 outlines the DFT effect zone. Maximum Number of Trips
(MNT) is purely a logic link between environmental factors and the existence of a fuse burning strategy.
MNT examines the number of nuisance trips over a period of hours, and if the threshold is exceeded
MNT will automatically disable the fuse burning strategy. This logic allows for automation of risk
mitigation on high fire risk/high wind days, where pecking faults such as a tree blowing up against a line
could cause ignition over constant fault events over the course of a day.
Figure 3: OC2 and Disable Fast Trips (DFT)
Finally, all of this protection modification is only as good as the communications method which supports
it. Whilst the RC10 has an onboard RTU like most good recloser controllers, the reliability of the
communications between the SCADA Master station and the recloser is the key in the reliability and

APS | 2015
bushfire mitigation capabilities outlined above. Good IT practice involve periodically resetting remote
communications peripherals, but if a connection is lost the remote device might not be able to receive a
reset command.
With this in mind, a communications watchdog was implemented within the RC10. This communications
watchdog monitors incoming SCADA traffic, and checks to see if communications with the master station
has been lost. There are two indicators which the watchdog examines, the binary control points, and the
master station polling. The watchdog within the controller will cause the communications peripheral to
reset if no binary control commands have been received within a certain window, and additionally to
reset the communications peripheral if no master station polls have been received within a certain
window of time. This allows the recloser to attempt to restore communications automatically when
communications are lost, saving time and money when successful and reducing risk to utilities.
Through embracing these controls, utilities can seamlessly integrate simple bushfire risk management
schemes into their current fleet of NOJA Power Reclosers. The logical step forward then is to evaluate
the power quality at these sites, to examine the effects of bushfire risk mitigation strategies and to
establish areas of concern for the network.
3 Power Quality
Power Quality is an important concept to understand in modern electrical service provision. By ignoring
this issue, we allow opportunity for devastating harmonics to freely travel through our networks,
destroying our assets and interrupting our customer service. Only through protection and monitoring of
these issues will it be possible to improve network performance, safety, reliability, and economic bottom
line.
New developments in reclosers allow for comprehensive Power Quality monitoring and protection
features using the current install base. These reclosers now have the ability to measure Harmonic
Distortion, Interruptions, and Sags and Swells, and it is important to develop an understanding of these
features for optimum use of the resource.
3.1 Harmonics
Within a Power Transmission System, all power is delivered at a set frequency, which in Australia is 50
Hz. Harmonics are “contaminants” within the power supply, which have a frequency which is a multiple
of the baseline or “Fundamental” Frequency. These contaminant harmonics enter the power system
through many different means, but ultimately the bottom line is they are a nuisance, and should be
protected against.
Harmonics on the network can be devastating. Since these harmonics are essentially carrying unusable
superfluous energy across the network, they put excess strain on any devices connected. These
harmonics cause damage to insulation and the very power electronics which cause them, along with
excess transmission losses. The major issue is that harmonic damage is insidious. There are usually
very limited symptoms of harmonic issues, until a catastrophic event such as the loss of a transformer
or motor, which usually is accompanied by an inherent fire risk. These risks are something which DNSPs
are taking a great interest in, and in Australia the harmonic limit of contamination is limited as low as 8%
at the point of common connection, it is our responsibility in the energy industry that harmonics are
prevented from travelling through the network. And this in turn means that any responsible DNSP needs
to be able to provide protection against these damaging harmonics.
The single greatest cause of harmonics within the power system is the results of the semiconductor.
Most modern loads which use some sort of power electronics to transform the grid energy to usable
energy for the device cause harmonic distortion. This is a result of what is known as “Non Linear current
draw”, meaning that the devices do not take in the full natural sine wave. There are other causes, such
as transformer saturation, or large industrial loads such as Arc Furnaces, or even fluorescent office
lighting. Additionally, the large scale installation of solar photovoltaic arrays and their semiconductor
inverters are a notorious source of harmonics. With the proliferation of power electronics into the
network, it’s easy to see how a minor issue of the past is becoming progressively more prevalent as
technology advances.

APS | 2015
Figure 4: A Waveform with High Harmonic Content
Simplicity of calculation is lost when starting to consider harmonics, which is the initial challenge of
interpreting harmonic content. Whilst these unattractive waves look terribly complicated to understand,
there are two major mathematical ideas which make the concept quite simple. These are:
• The Principle of Superposition

• Fourier Transforms
The basic idea is that any complex wave can be represented as a sum of individual simple waves. This
is the concept of superposition. Fourier Transforms are the mathematical method for working out what
these individual waves are. The main difference between these waves are their magnitude, and their
frequency – which just so happens to be the two main features of harmonic identification.
Harmonics are waves which have frequency multiples of the fundamental frequency. Since any complex
wave can be represented by a combination of these waves, it is then possible to understand what the
harmonic content of a power supply is. The RC10 system utilises what is known as a “Fast Fourier
Transform” or FFT algorithm to analyse the energy flowing through the device, to provide exact values
for both the harmonic frequency and its magnitude.
Harmonics are measured in two separate methods, known as Total Harmonic Distortion, and Total
Demand Distortion. THD is a ratio between the fundamental voltage wave, and all the voltage
harmonics. This is expressed as a percentage. Total Demand Distortion is calculated in a similar way,
except the ratio is based on the peak current demand, rather than instantaneous voltage used by THD.
TDD is used to calculate current distortion relative to the peak demand.
The NOJA Power RC10 uses FFT algorithms to guarantee protection against the first through the
fifteenth harmonic of the device, and allows for specific allowed limits of up to five of these individual
harmonics. The system records all the data for the harmonics within a separate log file, which can be
analysed to determine the power quality flowing through the device. The RC10 allows for user
configurable response times to harmonics, for both THD and TDD. All the information required in order
to calculate harmonics and provide protection is achieved through use of the inbuilt voltage sensors and
current transformers within the recloser. The raw values are passed through the FPGA of the controller
for parsing, and protection and monitoring is executed on the calculated results. This method allows for
all OSM reclosers in service to provide this additional functionality with no development cost. The only
cost to an end user is the process of implementing a data management scheme to follow the
implementation.
Within the RC10, the actual evaluation of harmonics is conducted through dead band monitoring. This
means that a harmonic record is logged each time the system deviates out of a dead band. This method
is similar to DNP3 analogue monitoring, and as such should be tweaked according to the individual
requirements of the DNSP. A dead band too small will result in excess data, but too broad and there will
be important transitions missed.
3.2 Interruptions
One of the greatest indicators of power quality issues is the measure of customer minutes lost. This
value is obviously of high interest to DNSPs as it directly relates to their economic bottom line, but it is
now possible to actually calculate feeder performance at an RC10 controlled recloser. Again, by using
the currently installed asset, it is much simpler to implement data gathering like this across a larger
install base whilst minimising the cost impact.

APS | 2015
The RC10 allows for user configurable settings to determine the difference between a short and long
interruption and can log all the information relevant to each of these interruptions separately.
Conveniently, the total time of interruption and all customisations to the metering can be read and
configured from the HMI panel on the device, along with a simple option to transfer all the field logs to a
USB to be analysed at the utilities discretion. Interruption tracking essentially fills the data void left by
our own protection systems. If a fault occurs which causes a loss of power to a customer, the device will
track the time. Our power quality management tool also has the capability to use this data to calculate
performance index values, such as SAIFI, SAIDI and MAIFI.
3.3 Sags/Swells
Sags and swells are characteristic of the ebb and flow of power system as the energy demand shifts
through the day. In times of low demand, the end user voltage can begin to creep up, and vice versa.
Just like the interruptions monitoring, sag and swell monitoring fills the data void left by our overvoltage
and under voltage protection. Also, since it evaluates a smaller deviation, data which is usually missed
by protection levels is still recorded. Sags and swells can be indicators of greater issues present in the
network, and also allow utilities to better prepare for the mitigation of ill effects caused by periods of
overvoltage and brownout. Sags and Swell gathering is implemented in similar fashion to interruptions
monitoring in the RC10, however instead of looking for periods of loss of supply, a comparison is made
between the registered system voltage and the witnessed voltage. Event durations are also captured
and timestamped, allowing utilities to understand what the trends in voltage are during the course of
days or weeks.
3.4 Oscillography
The final piece of the monitoring puzzle is to actually capture a direct copy of the current and voltage
waveforms passing through the recloser. By directly capturing the waveforms, a wealth of possibilities
for analysis, interpretation and network improvement are provided. As the reclosers already possess all
measuring devices required as well as the capability to interpret all data at a high sample rate, it is a
logical extension to be able to plot this data in the IEEE format of COMTRADE. This oscillography data
can then be retrieved and imported into many different software packages for analysis.
In order to capture data that is worthwhile, it is important to be able to trigger the correct capture point.
This depends greatly upon the installation, but selecting the correct Perrault capture and trigger point is
paramount in effective use of this technology.
One of the most interesting applications of this technology is to capture fault events at reclosers, and
import these COMTRADES into relay test sets such as an Omicron or Doble. These test sets allow for
simulation of the real faults present on a network, granting the capability for evaluation of performance
of network assets, along with the optimisation of network protection.
3.5 Remote Retrieving of Data

All power quality data including oscillography gathered from the NOJA Power RC10 can be gathered
remotely. Given the challenges of geographic distribution within Australia, many DNSPs have
progressed towards an engineering access approach to managing their smart reclosers. By using this
same port on the RC10 reclosers, it is possible to remotely gather the PQDIF and COMTRADE files
from the reclosers.
These logs can contain years of performance data, and through using a software package such as
Power Quality SoftwareTM from NOJA Power it is possible to filter and evaluate the information gathered.
In keeping with more traditional SCADA readings, most functionality mentioned for power quality within
this paper has associated SCADA Binary and Analogue points, allowing for instantaneous readings of
Harmonics, Sags/Swells, and interruption data. Through using familiar methods of data gathering,
integration of power quality monitoring into an existing recloser network can be seamless and simple.

APS | 2015
4 Conclusion
The only way to guarantee a continuous improvement in network performance is to embrace the new
technology on offer. This paper outlines the new capabilities available within an asset which already is
in use across the majority of Australia. Through bushfire mitigation and power quality monitoring, it is
possible to grow the performance and revenues of DNSPs, and in a time where efficient, safe network
operation is paramount it is negligent to disregard the capabilities available within the switchgear of
choice on distribution networks, the semiconductor controlled ACR.
About the Author

Martin van der Linde is a graduate of the University of Queensland, Australia, holding a
Bachelor of Engineering (Electrical) and a Bachelor of Business Management (Human
Resource Management). Martin is currently the Sales Engineer Australia and Pacific for
NOJA Power Switchgear Pty Ltd, Murarrie, Brisbane, QLD Australia 4172.

APS | 2015

APS | 2015
APS | 2015
Enhancing Australian Protection System Testing through

Real Time Digital Simulations
Georgios Konstantinou | UNSW Australia | g.konstantinou@unsw.edu.au

Guishi Allen Wang | UNSW Australia | allen.wang@unsw.edu.au
Baburaj Karanayil | UNSW Australia | k.baburaj@unsw.edu.au
Vassilios G. Agelidis | UNSW Australia | vassilios.agelidis@unsw.edu.au
1 Introduction
The modern power system is large and highly complex in nature. At the same time, there are a
number of new and potentially disruptive technologies that can redefine the way the power system is
designed, operated and managed, changing long-existing network dynamics. Examples of these
disruptive technologies include intermittent power generation (such as wind and large-scale solar PV
farms), centralized energy storage, and distributed generation (DG). Meanwhile, computational power
and sensing capacity is continuously increasing within the network. New generation network devices
(including flexible alternating current transmission systems (FACTS) such as synchronous static
compensators (STATCOMs), HVDC interconnections, multi-terminal dc networks and dc grids) extend
the available solutions but further complicate the optimal selection of technologies.
This increased complexity also translates into requirements for more sophisticated and highly selective
protection systems for faster fault detection and isolation, in order to minimize disruption and increase
network availability. Improper settings or co-ordination among protection devices can lead to major
network disruptions, as demonstrated in the 2003 North American Eastern Interconnection
Blackout [1]. Inadequate understanding of the power system and inadequate situational awareness
are considered among the major causes of such failures [1]. Recently, the aspect of grid resiliency has
also attracted attention, in addition to the existing focus on reliability, efficiency and cost-
effectiveness [2].
One of the main methods of demonstrating and verifying the performance of current and future power
systems, thoroughly testing the design and operation of protection schemes, and experimenting on
novel methods and approaches is through the use of real-time digital simulations. The concept of real-
time digital simulations was initially introduced in 1969 by Dommel [3] and is based on the calculation
of the system’s currents and voltages using the conductance matrix of an equivalent electrical network
at a time smaller than the defined time-step [4]. The time-step should be selected so that the real-time
behavior of the network is accurately emulated and provides trustworthy results.
Through real-time simulation, the behaviour of the power system can be observed in real-time. At the
same time, external hardware (e.g. digital controllers or protection relays) can be directly connected to
the simulator enabling studies on the response of hardware and on the response of the power system
following any operation of the external device. The major drawback of this approach is that it requires
large computational power and processing capabilities, which increase exponentially with the
complexity of the simulated system. The typical time-step for real-time digital simulations is in the
range of a few µs (up to 50µs).
Real-time digital simulation can be a very powerful tool for network planners, protection engineers and
everyone involved in the design and operation of a power system. The objective of this article is to
present the real-time digital simulator (RTDS) available at UNSW Australia, demonstrate its
capabilities and illustrate applications relevant to protection testing as well as the broad range of
simulations that can be performed.
2 UNSW System and Capabilities

The real-time digital simulation laboratory at UNSW Australia (Figure 1) includes the largest power
systems simulator in Australia and also the largest of its kind in any research institution globally,
providing a new way for utilities and universities to investigate the operation and management of

APS | 2015
networks. The RTDS is modular and highly reconfigurable meaning that one large or many smaller
simulations can run at the same system.
Figure 1: Real-time Digital Simulator at UNSW Australia. (Left) Racks and Communication, (Right) I/O boards and
CMS156 Amplifiers
The main purposes of the RTDS system are to provide a state-of-the-art real-time simulation
laboratory for general power system studies, allow hardware-in-the-loop testing of control and
protection systems to enable studies of renewable energy system integration (large-scale or
distributed) to the network, and to provide a platform for simulation and testing of the current
Australian transmission and distribution network.
At the core of UNSW’s RTDS system lie 18 “racks”, with a total of 90 PB5 processor cards and 180
PowerPC MC7448 RISC 1.7GHz CPUs (2 CPUs per processor card). Each of the racks has 5 PB5
cards and can accommodate up to 6. Each rack can use up to two processors for network solutions,
and each network solution can include up to 72 single-phase or 24 three-phase nodes. These
numbers can be further extended to 90 single-phase (30 three-phase) nodes through the use of an
enhanced fiber backplane which is also available in our system. The available processing power
allows simulation of up to 864 three-phase nodes in a single system and in hard real-time (solution
time smaller than the timestep) and can run continuously for several days with numerical stability.
Some examples of processing requirements for different networks are shown in Table 1.
Besides the processors used for network solutions, other processors within a rack are used for solving
the models of the components that comprise the system under study. These components include
transmission lines and cables, synchronous and asynchronous machines, passive elements and filters
as well as instrument transformers (i.e. CTs, PTs and CVTs). These components are typically
modelled at a timestep of 50µs. Additionally, sections of the system under study that include high-
frequency modulated power electronics, such as HVDC terminal converters & STATCOMs, can be
modelled at a smaller timestep of 2µs.
Analogue communication with external devices takes place through 18 dedicated input and 18 output
units with 12 optically isolated 16 bit D/A channels per unit providing a total of 216 analogue input and
output channels. Additionally, 1152 optically isolated channels (18 boards of 64 channels each) are
used for high-speed digital input and the same number for digital outputs from the RTDS. Input and
output devices are located at each rack as well as the two main I/O cabinets (Shown in Figure 1b).
The system also includes ten digital output interphase (up to 250V supplied by external power
supplies) and 4 Omicron CMS 156 three-phase current and voltage amplifiers (250V, 25A).
Interfacing of devices with the RTDS can also happen through network protocols implemented in the
GTNET (Gigabit Transceiver Network Interface) cards. The interfaces currently supported by the
UNSW RTDS include PMU for synchrophasor data streams, GSE supporting IEC61850 GOOSE and
GSSE messages, SV for sampled value data streams, as well as DNP and IEC60870-5-104 mainly for
SCADA communication. Two Playback cards that allow large COMTRADE files to be stored in a PC
rather than the RTDS are also available. The detailed specifications of network protocols are given in
Table 2.

APS | 2015
14 bus system – 3 Generators 1 rack – 2 PB5 cards

39 bus system – 9 Generators 1rack – 5 PB5 cards
108 bus system – 21 Generators 3 racks – 14 PB5 cards
CSC-HVSC 1 rack - 2 PB5 cards
Cigre B5 14 bus MTDC HVDC system 4 racks – 22 PB5 cards
Table 1: Typical simulation requirements
Network Protocol Capacity Network cards

Up to 8 three-phase currents and voltages
PMU 4
symmetrical components (1 – 60 Hz)
64 binary input - 64 binary output
GSE 4
Support for both GOOSE & GSE
IEC 61850-9-2 voltage and current sampled values
SV up to 8 signals (V & I) at 80 samples/cycle 4
1 signal at 256 samples/cycle
1024 status or output signals @ 1000Hz
512 control or input signals @ 1000Hz
DNP 3.0 2
500 analogue outputs @ 4Hz
100 analogue inputs @ 4Hz
1024 status or output signals @ 1000Hz
512 control or input signals @ 1000Hz
IEC60870-5-104 2
500 analogue outputs @ 4Hz
100 analogue inputs @ 4Hz
Playback COMTRADE file playback 2
Table 2: Network Interfaces of the RTDS at UNSW Australia
3 Applications for Protection Engineers

RTDS systems allow the real-time simulation of one or multiple fault scenarios and testing of a number
of relays. The capacity of UNSW’s RTDS is such that even complex fault scenarios and extremely
large network arrangements can be simulated in order to verify and confirm the optimal settings of the
relay. The setup allows testing of protection equipment either based on voltage and current secondary
injections [5] or through communication protocols for IEC 61850 compliant relays. These testing
procedures are briefly described in the following sections.
3.1 Secondary Voltage and Current Injections

In a typical secondary injection test of a relay, the model of the power system runs real-time in the
RTDS. The protection relay is connected in the loop through the available voltage and current
amplifiers (as shown in Figure 2) which, in turn, are connected to the D/A boards of the RTDS. The
commands to other equipment (such as breaker operation signals, reclose signals etc) are fed back to
the simulation through the A/D boards. The current and voltage transformers are also included in the
simulation and their characteristics and saturation behavior are also modelled. Up to four relays can
be simultaneously tested with this method in our lab. The testing can be automated so that a batch of
scenarios can be tested with minimal user interaction.

APS | 2015
Figure 2: Traditional Relay Testing methods based on secondary voltage and current injections
3.2 Relay Testing Based on IEC61850

Testing of relays based on IEC61850 is also possible with the RTDS. This testing does not require the
use of the I/O devices or the amplifiers but uses the GTNET cards (see Table 2), as shown in Figure
3, to establish the necessary communications between the CTs and VTs within the simulation, the
sampled values as per IEC61850 (GTNET SV card) and also communicate GSE through GOOSE or
GSSE messages to other relays or even breaker status back to the relay. Multiple relays, either
directly connected to the switches or implemented in the simulation environment of the RTDS can be
included. This allows for simulation of large and complex systems.
Figure 3: Testing of IEC61850 compliant relays with GTNET cards [6]
A unique capability of relay testing as Hardware-in-the-loop is that allows the evaluation of both the
relay performance as well as the response of the system to the actions and commands of the relay.
Additionally, both of the above methods represent a practical approach for testing the interaction
between multiple relays within the same area or across multiple terminals as well as optimizing their
settings and coordination.

APS | 2015
3.3 Protection Schemes with Synchrophasors and SIPS

The introduction of Class P synchrophasor measurements have enabled the use of PMUs as part of
protection schemes in what is known as WAMPC (Wide Area Monitoring, Protection and Control). A
detailed description of Synchrophasors, relevant IEC and IEEE standards, examples and applications
in protection can be found in [7]. Protection based on PMUs and synchrophasors can also be
performed in the RTDS, facilitated by the GTNET PMU protocol.
The application of PMUs is not limited to specific protection functions such as distance or busbar
protection schemes. Their measurement can be integrated in a System Integrity Protection Scheme
(SIPS, [8] with the general concept illustrated in Figure. 4) in order to provide an extended
management of disturbances and a wide-area scheme that reinforces the integrity of the network.
A SIPS integrates a number of functions and applications. When applied over extended areas (i.e.
sub-system and system level SIPS) the extend of the network, measurement and collection schemes
as well as multiple levels of decision making, can make the design of the system highly complex.
Through expansive real-time simulations, it is possible to assess and design SIPS, grid performance
based on multiple parameters (e.g. phase angle difference, frequency response etc) as well as
optimize the decision making process and overall operation.
Figure 4: A general concept of System Integrity Protection Schemes [8]
4 Other Applications
Real-time digital simulation has many other applications besides hardware-in-the-loop testing of
protection relays. Some of these applications are presented in the following sections:
4.1 Voltage and Frequency Control – AC Network Simulation

Enabling real-time simulation of large scale power systems enables network operators to investigate
voltage and frequency stability as well as the operation of controllers. Detailed modelling of
synchronous generators (as shown in Figure 5) also allows the hardware-in-the-loop (HiL) verification
of such controllers including typical excitation, governors and voltage controllers for synchronous
generators, wind-turbine farm modelling and network integration studies as well as co-generation
plants etc. The inclusion of additional power equipment in the loop with the simulation can be
facilitated through fully controllable four-quadrant power supplies acting as grid simulators.

APS | 2015
Figure 5: Simulation Environment of the RTDS and definition of generator, governor and exciter parameters.
4.2 Power Electronics and FACTS

Power electronics and FACTS devices are becoming more common in modern power systems and
with current and future expansion of transmission networks focusing around large HVDC corridors
(see Germany’s HVDC plans for example), as well as extensive HVDC interconnections (Northern
Europe and UK, Japan etc). The current vision for future networks also includes multiterminal [9] dc-
grids, voltage and network support with STATCOMs both at transmission and distribution level and
many others.
In this case, the modelling and simulation challenges are two-fold. On one hand, the size of the
network requires large processing power in order to provide accurate results. On the other hand,
power electronics converters (e.g. The modular multilevel converter – MMC – Figure 6 [10]) should be
included at an accuracy level higher than that of an average model [9], meaning that simulations
should be executed at very small timesteps to accommodate the PWM frequency, further increasing
the complexity and processing requirements.
It should be noted that the Real-time digital simulation lab at UNSW also has an OPAL-RT 5607 and
four OPAL-RT 4500 that are used for the real-time simulation of power electronics converters and
devices.
Figure 6: Modular Multilevel Converter [10] and a dc-grid testbench [9]

APS | 2015
4.3 Renewable Energy – Distributed Generation – Microgrids – Smart Grids

Real-time Digital Simulation can further enhance studies on high penetration of renewable and
distributed energy generation into the distribution networks, large-scale solar and wind farms behavior
and modelling and enhance their operation and economic evaluation as well as the modelling and
operation of energy storage devices at a distribution or transmission level. Modelling of microgrids,
optimal power sharing amongst local generation, islanding studies as well as studies on microgrid
protection can be performed in the RTDS while the inclusion of communication elements can facilitate
research on Smart Grids.
4.4 Training and Education
The use of real-time digital simulation provides a very powerful tool in the education of future
protection and power engineers in general and can enhance the learning experience of undergraduate
and postgraduate students. Additionally, there are many opportunities for further education of utility
engineers, introduction to new technologies as well as continuing professional development.
5 Sample Case Study

A simple case study including two distance relays is developed on RTDS to illustrate the capabilities of
the RTDS and demonstrate the open and closed-loop (HiL), real-time testing. As mentioned in
Section 3, simulations that include the relays in the loop can evaluate the performance of the
protection equipment, but also evaluate the response of the network following any operation of the
protection equipment.
The network of investigated case is a five bus system with the following components (as shown in
Figure 7): 2 sources models, 2 sets of CTs and CVTs for relays, 2 distance relays using RTDS relay
models, 2 Circuit Breakers to isolate the transmission lines, Bergeron Transmission line models, a
fault switch and a fault control block as means of configuring and applying different faults at various
locations along the transmission line.
BUS1 BUS2 BUS3 BUS4 BUS5
1.00000 /_ 0.00000 1.00000 /_ 0.00000 1.05449 /_ -0.85964 1.09187 /_ -1.59966 1.09187 /_ -1.59966
N3 N2 N1 N6 N5 N4 N7 N8 N9 N10 N11 N12 N13 N14 N15
control CT control CT
AC Type CT1 CT2 AC Type
BRK1A BRK2A
A A
RRL 1 T-LINE NAME: T-LINE NAME: 1 1 T-LINE NAME: T-LINE NAME: 1 RRL
VBURA LINE1 LINE1 LINE2 LINE2 VBURA2
BRK1B BRK2B
B B
RRL 2 2 2 2 RRL
SENDING END RECEIVING END SENDING END RECEIVING END
VBURB VBURB2
BRK1C TERMINAL NAME: TERMINAL NAME: TERMINAL NAME: TERMINAL NAME: BRK2C
C C
SRC1 RRL 3 LINE1SE LINE1RE 3 3 LINE2SE LINE2RE 3 RRL SRC2
VBURC VBURC2
Ell=228.424588kV kV252.195361Ell=
C
A
CVT1 CVT2
RL
RL
RL
control CVT control CVT
Zs FOR line
L-G FAULT POINT
=====================
Z1=35.19 @87.02
Z0=131.5 @73.48
Distance Protection Fault Distance Protection
Comm Scheme Control Comm Scheme
Figure 7: System diagram of distance relay testing.
The simulation results for the relay operation and the response of the rest system, reflected by the CT
and CVT measurements, in case of a three- or single-phase line-ground fault at random location are
presented in Figures 8 and 9, respectively. Figure 8a shows that the relay 1 trips based on the zone 1
protection settings and within 8ms of detecting the fault. As the recloser element is enabled, the
breaker recloses after the fault has been cleared. The voltages of two CVTs decrease and the currents
of two CT increase, as shown in Figures 8b and 8c. During the fault period and the operation of relay
1, the reaction at the location of relay 2 is also demonstrated in Figure 8c. Similar results for the case
of a single-phase fault are shown in Figure 9.

APS | 2015
Figure 8: Distance relay testing results in case of a three-phase fault: (a) operating states in case of three-phase fault;
(b) CVT 1 and CT 1 measurement in case of three-phase fault; (c) CVT 2 and CT 2 measurement in case of three-phase
fault;
The currents and the voltages can be sent in real-time to the current and voltage amplifiers which are
then connected to the relay under test. Fig. 10 shows the voltage and current as measured at the D/A
outputs of the RTDS. It should be noted that a proper conversion between the simulated values and
the output of the D/A converters should be done in order to extract accurate results.
6 Conclusions
The increased complexity of the power system and a number of disruptive technologies that are
expected to play an important role in the future require a comprehensive approach to network
modelling and simulation as well as a comprehensive means of testing power system protection and
control. Real-time digital simulation is widely considered as the optimal method for achieving the
above, enabling hardware-in-the-loop testing of protection and control, power electronics and their
interfaces, renewable and distributed generation and many others. The real-time digital simulation lab
at UNSW provides a unique tool for the university and any of the utilities to test and validate current
and future networks. It also provides the opportunity for UNSW researchers to work closely with the
manufacturing and utility sectors, and bridge the gap between the research community and end users

APS | 2015
Figure 9: Distance relay testing results in case of a single-phase line-ground fault: (a) operating states in case of single-
phase fault; (b) CVT 1 and CT 1 measurement in case of single-phase fault; (c) CVT 2 and CT 2 measurement in case of
single-phase fault.
Figure 10: Real-time outputs of the voltage and current from CT1 and CVT1 for phase A following an three-phase
fault at 10% of the length of the line.

APS | 2015
7 References
[1] Andersson, G., P. Donalek, R. Farmer, N. Hatziargyriou, I. Kamwa, P. Kundur, N. Martins et al.
"Causes of the 2003 major grid blackouts in North America and Europe, and recommended
means to improve system dynamic performance." IEEE Trans. Power Systems, vol. 20, no. 4,
(2005): 1922-1928.
[2] Ton, D.T., Wang, W.-T.P. “A More Resilient Grid: The U.S. Department of Energy Joins with
Stakeholders in an R&D Plan,” IEEE Power and Energy Magazine, vol. 13, no. 3, 2015.
[3] Dommel H.W. “Digital Computer Solution of Electromagnetic Transients in Single-and
Multiphase Networks”, IEEE Trans. on Power Apparatus and Systems, Apr. 1969
[4] Mosterman, PJ, & Popovici, K 2012, Real-time Simulation Technologies : Principles,
Methodologies, and Applications, CRC Press, Boca Raton, FL.
[5] Bastigkei, B., Pritchard, C. and Hensler, Th. “New possibilities in field testing of distributed
protection systems,” in Australian Protection Symposium, 2013, p. 115 – 121
[6] RTDS, www.rtds.com [Accessed 24 June 2015]
[7] Apostolov A. “Synchrophasors Application for Protection. Can we use them?” in Australian
Protection Symposium, 2013, p. 151-159
[8] Madani, V., Novosel, D., Begovic, M., Adamiak, M, “Application Considerations in System
Integrity Protection Schemes”
[9] Cigre WG B4.57, Guide for the Development of Models for HVDC Converters in a HVDC Grid
[10] Konstantinou, G. Pou, J. Ceballos, S.., Agelidis, V.G., “Active redundant sub-module
configuration inmodular multilevel converters," IEEE Trans. on Power Delivery, Vol. 28, No. 4,
pp. 2333-2341, Oct. 2013.”
8 Acknowledgment
This work was supported by funding from the Australian Government Department of Education
Education Investment Fund Program (AGL Solar PV Research Infrastructure Project: University of
New South Wales in partnership with AGL Pty Ltd, First Solar and the University of Queensland).

APS | 2015
About the Authors
Georgios Konstantinou received the B.Eng. degree in electrical and computer

engineering from the Aristotle University of Thessaloniki, Thessaloniki, Greece, in
2007, and the Ph.D. degree in electrical engineering from the UNSW Australia (The
University of New South Wales), Sydney, Australia, in 2012. Currently, he is a Senior
Research Associate with the Australian Energy Research Institute (AERI) and the
School of Electrical Engineering and Telecommunications, UNSW Australia. He is an
Associate Editor of IET Power Electronics. His research interests include hybrid and
modular multilevel converters and their applications, including renewable energy
systems, pulse-width modulation, and selective harmonic elimination techniques for
power electronics.
Guishi Allen Wang (Guishi Wang (S'10—M'15) received the B.Eng. degree in
electrical engineering and automation from the Northwestern Polytechnic University,
Xi`an, China, in 2007, and the Ph.D. degree in electrical engineering from the UNSW
Australia (The University of New South Wales), Sydney, Australia, in 2015. Currently,
he is a Research Associate with the Australian Energy Research Institute (AERI) and
the School of Electrical Engineering and Telecommunications, the UNSW Australia.
His research interests include the hybrid energy storage system supporting the
dispatch of large-scale renewable generation and providing ancillary services to the
electricity network.
Baburaj Karanayil received the B.Tech. (Hons) degree from the University of Calicut,
Malappuram, Kerala, India, the M.Tech. degree from Indian Institute of Technology,
Bombay, India, and the Ph.D. degree from the University of New South Wales
(UNSW Australia), Sydney, Australia, in 1984, 1986, and 2005, respectively, all in
electrical engineering. From 1986 to 1994, he was a Senior Engineer in the power
electronics and electrical drives industry in India. From 1995 to 2011, he was a
Professional Officer in power electronics and electric drives research at University of
New South Wales. Since February 2012, he has been a Research Fellow with the
Australian Energy Research Institute, UNSW, Sydney, Australia.
Vassilios G. Agelidis received the B.Eng. degree in electrical engineering from

Democritus University of Thrace, Thrace, Greece, the M.S. degree in applied science
from Concordia University, Montreal, QC, Canada, and the Ph.D. degree in electrical
engineering from Curtin University, Perth, WA, Australia, in 1988, 1992, 1997,
respectively. From January 2005 to December 2006, he was the inaugural Chair of
Power Engineering with the School of Electrical, Energy, and Process Engineering,
Murdoch University, Perth, Australia. From December 2006 to June 2010, he was the
Energy Australia Chair of Power Engineering, the University of Sydney, Sydney,
Australia. He is currently the Director of the Australian Energy Research Institute,
University of New South Wales (UNSW Australia), Sydney, NSW, Australia.

APS | 2015

APS | 2015

APS | 2015

Australian
Protection Symposium
Contact
For enquiries regarding the Australian Protection Symposium,
please do not hesitate to contact:
Maryam Khallaghi, Technical Program Chair maryam.khallaghi@omicron.at
Jessica Rizio, Organisational Coordinator jessica.rizio@omicron.at
Suite 22, Level 2, 240 Plenty Road, Bundoora VIC 3083, Australia
Phone: +61 3 9743 8400 | Fax: +61 3 9743 8401
info@protectionsymposium.com.au | www. protectionsymposium.com.au
Sponsored by:

APS Booklet 2015

Uploaded by

Document Information

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

APS Booklet 2015

Uploaded by

Copyright:

Available Formats

Proceedings

August 18-19, 2015

4 Terms and Conditions

II Australian Protection Symposium

• Robert Bates | Ergon Energy

Australian Protection Symposium III

Fundamental Principles and Requirements of Protection Systems

Calculation of Loadability Limits of Modern Numerical Relays

An Overview of D.C Traction Protection Systems

The Digital Substation- Principles, Functions and Benefits

Impact of IEC61850 ed 2 on Protection GOOSE

Case Study: Implementation of IEC61850 in Generator and

COMTRADE Analysis to Visualise Protection Operations 89

Verification of Distance Protection Scheme of double Circuit transmission line

Cygnus Atratus and How It Causes Problems

IV Australian Protection Symposium

New Design of Distance Protection for Smart Grid Applications

Implementation of Multi-Intertripping Protection Schemes in a complex EHV Industrial

Old Lessons – Re learned

330 KV Downed Conductor Protection and Successful Operation

Australian Protection Symposium V

VI Australian Protection Symposium

Fundamental Principles and Requirements

John Ainsworth | Retired, formerly Ausgrid

Australian Protection Symposium 1

4.1 Primary protection

4.2 Back-up protection

2 Australian Protection Symposium

4.3 Fault clearance times

4.4 System stability

4.5 Effects of faults

4.6 Prevent / minimize loss of supply

5 Components of a protection scheme

Australian Protection Symposium 3

6 Functional proving of Protection Relays and Systems

4 Australian Protection Symposium

7 Secondary Circuit Isolation

Australian Protection Symposium 5

(a) Tripping to circuit breakers.

8 Independence of A and B protection schemes

6 Australian Protection Symposium

Australian Protection Symposium 7

This requirement arises

This arises from lightning

This arises from HV

This arises from relay,

Table 1: Type Test requirements

11 Sources of Energy for protection schemes

8 Australian Protection Symposium

The sources of energy will generally be in the form of battery supplies.

12 DC supply Voltage Range

(i) Battery type and capacity (Ah).

Australian Protection Symposium 9

13 Major System shutdown and Protection DC supplies

14 Safety in the design of protection systems

(a) Working on electromagnetic current transformer secondary circuits which generate

10 Australian Protection Symposium

16 Digital and Optical Fibre Based protection systems

Australian Protection Symposium 11

17 Safety in the design of protection systems

17.2 Fault protection and protection against overload

17.3 Rating of Primary equipment not to be limited by protection

17.4 Circuit Breakers

17.5 Verification of Relay Settings